Marek J. Wójcik, Yukihiro Ozaki - Spectroscopy and Computation of Hydrogen-Bonded Systems-Wiley-VCH (2023)

Downloaded from https://onlinelibrary.wiley.com/doi/ by ibrahim ragab - Cochrane Germany , Wiley Online Library on [13/12/2022].
See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Downloaded from https://onlinelibrary.wiley.com/doi/ by ibrahim ragab - Cochrane Germany , Wiley Online Library on [13/12/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Spectroscopy and Computation of Hydrogen-Bonded Systems
Spectroscopy and Computation of
Edited by Marek J. Wójcik and Yukihiro Ozaki

Hydrogen-Bonded Systems
Editors All books published by WILEY-VCH are carefully
produced. Nevertheless, authors, editors, and
Prof. Marek J. Wójcik publisher do not warrant the information
Jagiellonian University contained in these books, including this book,
Faculty of Chemistry to be free of errors. Readers are advised to keep
Gronostajowa 2 in mind that statements, data, illustrations,
30-387 Krakow procedural details or other items may
Poland inadvertently be inaccurate.
Prof. em. Yukihiro Ozaki

Kwansei Gakuin University Library of Congress Card No.: applied for
Department of Chemistry
2-1 Gakuen British Library Cataloguing-in-Publication Data
Kobe Sanda Campus A catalogue record for this book is available
669-1337 Sanda, Hyogo from the British Library.
Japan
Bibliographic information published by
Cover Image: © Emre Terim/Shutterstock the Deutsche Nationalbibliothek
The Deutsche Nationalbibliothek lists
this publication in the Deutsche
Nationalbibliografie; detailed bibliographic
data are available on the Internet at
<http://dnb.d-nb.de>.
© 2023 WILEY-VCH GmbH, Boschstraße 12,

69469 Weinheim, Germany
All rights reserved (including those of

translation into other languages). No part of
this book may be reproduced in any form – by
photoprinting, microfilm, or any other
means – nor transmitted or translated into a
machine language without written permission
from the publishers. Registered names,
trademarks, etc. used in this book, even when
not specifically marked as such, are not to be
considered unprotected by law.
Print ISBN: 978-3-527-34972-2

ePDF ISBN: 978-3-527-83489-1
ePub ISBN: 978-3-527-83490-7
oBook ISBN: 978-3-527-83491-4
Typesetting Straive, Chennai, India

v
Contents
Part I Theory 1
1 Linear Response Theory Applications to IR Spectra of

H-Bonded Cyclic Dimers Taking into Account the Surrounding.
Updating Contributions Involving Davydov Coupling, Fermi
Resonances and Electrical Anharmonicity 3
Paul Blaise and Olivier Henri-Rousseau
1.1 Introduction 3
1.2 Dimer Strong Anharmonic Coupling Theory 3
1.2.1 Different Theoretical Situations 3
1.2.1.1 Strong Anharmonic Coupling Within Adiabatic Approximation For
Monomer 3
1.2.1.2 Introduction of Fermi Resonances 6
1.2.1.3 H-Bonded Centrosymmetric Dimer 8
1.2.1.4 Dimer Involving Damping, Davydov Coupling, and Fermi
Resonances 12
1.2.2 The Spectral Density 13
1.3 Comparison with Experiments 14
1.3.1 Carboxylic Acid Dimers Ignoring Fermi Resonances 14
1.3.1.1 Gaseous and Liquid Acetic Acid Dimers 14
1.3.1.2 Gaseous Acrylic and Propynoic Acids 15
1.3.2 Carboxylic Acids Taking Into Account Fermi Resonances 16
1.3.2.1 Crystalline Adipic Acid 16
1.3.2.2 Crystalline Polarized and Unpolarized Glutaric Acid Taking Into
Account Fermi Resonances 17
1.3.2.3 Crystalline Thiopheneacetic Acid and Thiopheneacrylic Acids 17
1.3.2.4 l.2-Naphtylacetic Acid (2-NA) Crystals 20
1.3.2.5 Crystalline Aspirin Dimers Involving Slow Mode Morse Potential 23
1.3.2.6 Phthalic and Terephthalic Acid Crystals 25
1.3.2.7 Liquid Formic Acid Mixing of Monomer and Dimer 27
vi Contents
1.3.2.8 Crystalline Furoic Acid Dimer with Slow Mode Morse Potential and
Fermi Resonances 28
1.3.2.9 Other Kinds of H-Bonded Compounds 31
1.3.2.10 Phosphinic Acid Dimer 31
1.3.2.11 Monomer of (CH3 )2 O · · · HCl 33
1.4 Conclusion 36
1.5 Acknowledgment 36
References 36
2 Dynamic Interactions Shaping Vibrational Spectra of

Hydrogen-Bonded Systems 39
Marek J. Wójcik, Mateusz Brela, Łukasz Boda, Marek Boczar, and Takahito
Nakajima
2.1 Introduction 39
2.2 Theoretical Model of the Infrared Spectra of Gaseous (CH3 )2 O-HCl and
(CH3 )2 O-HF Complexes 42
2.3 Simulation of the Cl–H(D) and F–H Stretching Bands in the
DME-H(D)Cl and DME-HF Complexes 45
2.4 Methodology of Molecular Dynamics 47
2.5 Spectroscopic Study of Uracil, 1-Methyluracil, and
1-Methyl-4-thiouracil 49
2.6 Hydrogen Bond Interaction Dynamics in the Adenine and Thymine
Crystals 50
2.7 Guanine and Cytosine Crystals 51
2.8 Spectroscopic Signature for Ferroelectric Ice 52
2.9 Conclusions 55
Acknowledgment 56
References 56
3 Trajectory On-the-Fly Molecular Dynamics Approach to

Tunneling Splitting in the Electronic Ground and Excited
States 67
Tetsuya Taketsugu and Yusuke Ootani
3.1 Introduction 67
3.2 Semiclassical Tunneling Approach 69
3.3 Results and Discussion 71
3.3.1 Umbrella Inversion of Ammonia 72
3.3.2 Intramolecular Hydrogen Transfer in Malonaldehyde 73
3.3.3 Excited State Intramolecular Hydrogen Transfer in Tropolone 75
3.4 Conclusions 79
Acknowledgments 79
References 80
Contents vii
Part II Spectroscopy 83
4 Spectroscopic Signatures of Low-Barrier Hydrogen Bonding in

Neutral Species 85
Lidor Foguel, Zachary N. Vealey, and Patrick H. Vaccaro
4.1 Introduction 85
4.2 Spectroscopic Metrics for Hydrogen Bonding 87
4.2.1 Continuum of Hydrogen Bonding 87
4.2.2 Relationship to Tunneling 92
4.2.3 Ground-State Properties of Model Systems 93
4.2.4 Excited-State Spectroscopy of 6-Hydroxy-2-Formylfulvene 98
4.2.5 Ground-State Spectroscopy of 6-Hydroxy-2-Formylfulvene 102
4.2.6 Excited-State Properties of Model Systems 105
4.3 Concluding Remarks 108
Acknowledgments 109
References 109
5 Hydrogen-Bonding Interactions Using Excess

Spectroscopy 123
Yaqian Wang and Zhiwu Yu
5.1 Introduction of Hydrogen Bond 123
5.1.1 Definition of Hydrogen Bond 123
5.1.2 The Criteria of the Existence of Hydrogen Bonds 124
5.1.3 The Strength of Hydrogen Bonds 125
5.2 Theory of Excess spectroscopy 126
5.3 Studies of Hydrogen Bonds by Excess IR 129
5.3.1 Classical Hydrogen Bonds 129
5.3.2 Charge Assisted Hydrogen Bonds 131
5.3.3 Cooperative Resonance-Assisted Hydrogen Bonds 134
5.3.4 Weak/Moderate Hydrogen Bonds 138
References 142
6 Intramolecular Hydrogen Bonding in Porphyrin Isomers 145

Jacek Waluk
6.1 Introduction 145
6.2 H-Bond Characteristics 146
6.2.1 Porphine (1) 147
6.2.2 Porphycene (2) 148
6.2.3 Hemiporphycene (3) 150
6.2.4 Corrphycene (4) 152
6.2.5 Isoporphycene (5) 154
6.2.6 Porphyrin-(2.2.0.0) (6) 156
viii Contents
6.2.7 Porphyrin-(3.1.0.0) (7) 157

6.2.8 Porphyrin-(4.0.0.0) (8) 158
6.2.9 Inverted/Confused Porphyrin (9) 162
6.2.10 Neo-confused Porphyrin (10) 164
6.3 Correlations Between Geometry and HB Strength 165
6.4 Parameters That Can Describe the HB Strength 167
6.5 Tautomerization Mechanisms 168
6.6 Summary 169
Acknowledgments 170
References 170
7 Isotope Effects in Hydrogen Bond Research 173

Poul Erik Hansen
7.2 Hydrogen Bond Potentials 173
7.3 Calculations 175
7.4 Hydrogen Bond Types 176
7.5 Deuterium Isotope Effects on Chemical Shifts 176
7.6 Intramolecular Hydrogen Bonds 177
7.6.1 Two-Bond Deuterium Isotope Effects on 13 C Chemical Shifts 178
7.6.2 Long-Range Isotope Effects 184
7.6.3 One-Bond Deuterium Isotope Effects on 15 N Chemical Shifts in
Solution 185
7.7 Biological Systems 185
7.7.1 Proteins 185
7.7.2 Deuterium Isotope Effects on 1 H Chemical Shifts 187
7.8 Intermolecular Hydrogen Bonds 187
7.9 Primary Isotope Effects 189
7.10 Isotope Effects and Acidity 191
7.10.1 Isotope Effects to Determine Protonation States 191
7.11 Solvent Isotope Effects and Exchange Rates 192
7.12 Exchange in the Solid-State 192
7.13 Hydrogen Bond Energies 193
7.14 Tautomerism 194
7.15 Solid-State NMR 197
7.15.1 Deuterium Isotope Effects on 15 N Chemical Shifts 199
7.16 Conclusions 202
References 203
8 Intramolecular Hydrogen Bonding: Shaping Conformers’

Structure and Stability 213
Gulce O. Ildiz and Rui Fausto
Contents ix
8.2 The Halogen-Substituted Acetic Acids CF3 COOH, CCl3 COOH, and
CBr3 OOH: Implications of IMHB on Structure and Conformers’
Stabilities 215
8.3 The Significance of IMHB in the ortho Chloro- and Fluoro-Substituted
Benzoic Acids 219
8.4 IMHB in Thiotropolone: Sculpturing the Bidirectional Infrared-Induced
Bond-Breaking/Bond-Forming Tautomerization 225
8.5 Conclusion 228
Acknowledgments 229
References 229
9 Hydrogen Bonding from Perspective of Overtones and

Combination Modes: Near-Infrared Spectroscopic Study 233
Mirosław A. Czarnecki, Yusuke Morisawa, and Yukihiro Ozaki
9.2 Investigation of Hydrogen Bonding of Water by NIR Spectroscopy 235
9.3 The Chain Length Effect on the Degree of Self-association of
1-Alcohols 237
9.4 Combined NIR and Dielectric Study on Association of 1-Hexanol in
n-Hexane 240
9.5 NIR Studies of Microheterogeneity in Alcohol/Alcohol and
Alcohol/Alkane Binary Mixtures 241
9.6 Overtones of νC≡N Vibration as a Probe of Molecular Structure of
Nitriles 244
9.7 Weak Hydrogen Bond in Poly(3-Hydroxybutyrate) (PHB) Studied by NIR
Spectroscopy 246
9.8 Studies of Hydrogen Bonding By Use of Higher Overtones 249
9.9 Comparison of Hydrogen Bonding Effects and Solvent Effects on Wave
numbers and Intensities of the Fundamental and First Overtone of the
N–H Stretching Mode of Pyrrole Studied By NIR/IR Spectroscopy and
One-Dimensional Vibrational Schrödinger Equation Approach 252
9.10 Summary 256
Acknowledgments 257
References 257
10 Direct Observation and Kinetic Mapping of Point-to-Point

Proton Transfer of a Hydroxy-Photoacid to Multiple
(Competing) Intramolecular Protonation Sites 261
Dina Pines, Dan Eliovich, Daniel Aminov, Mark Sigalov, Dan Huppert, and
Ehud Pines
10.2 From Intermolecular Proton Transfer to Solvent to Intramolecular
Point-to-Point Transfer in 1 : 1 Hydrogen-Bonding Complexes of Water
with Bifunctional OH Photoacids 270
x Contents
10.3 Water Is Able to Donate and Accept an H-bond as Demonstrated by IR

Absorption in 1 : 1 Water–(Acid or Base) Complexes 273
10.4 Proton Transfer Along with Water Bridges in Acetonitrile (ACN)
Spanning the Distance Between an Acidic and a Basic Side Groups of
Bifunctional Photoacids 274
10.5 Time-Resolved Fluorescence Measurements of Proton Transfer along
with Water Bridges 277
10.6 Isotope D/H Effect 283
10.7 Insights into the Mechanism of Proton Transfer Through One-Water
Bridge in Bifunctional 2-Naphthols 285
10.8 Summary 288
Acknowledgments 289
References 289
11 Spectroscopic Determination of Hydrogen Bond

Energies 293
Mausumi Goswami and Elangannan Arunan
11.2 Binding Energy Measurement Involving Infrared (IR) Excitation 296
11.2.1 Measurement of the Dissociation Energy of H-Bonded Complexes
Through Vibrational Pre-dissociation Dynamics via Infrared
Excitation 296
11.2.1.1 Optothermal Bolometric Determination 297
11.2.1.2 Velocity Map Imaging 299
11.2.2 Determination of Gibbs Free Energy of H-Bonded Complex Formation
By Infrared Spectroscopy 307
11.2.3 Measurement of Binding Energy of H-Bonded Complexes by IR–UV
Double Resonance Spectroscopy 314
11.3 Determination of the Binding Energy of H-Bonded Complexes Using
Spectroscopic Techniques Involving Electronic Excitation 316
11.3.1 Determination of H-Bond Dissociation Energy Through Multiphoton
Ionization Techniques 316
11.3.2 Determination of the Dissociation Energy of Cationic H-Bonded
Complexes Through Birge–Sponer Extrapolation 325
11.3.3 Determination of the Dissociation Energy of H-Bonded Complexes Using
SEP-REMPI Technique 328
11.4 Estimation of the Well Depth of H-Bonding Interactions Through
Microwave Spectroscopy 332
11.5 Conclusion 335
References 336
12 IR and NMR Spectral Diagnostics of Hydrogen Bond Energy

and Geometry 345
Peter M. Tolstoy and Elena Yu. Tupikina
Contents xi
12.1.1 Solving the Reverse Spectroscopic Problem 345

12.1.2 Spectral Markers for Proton Transfer and H-Bond Length 346
12.2 Spectral Characterization of Hydrogen Bond Geometry 348
12.2.1 Description of Hydrogen Bond Geometry 348
12.2.2 Averaging of NMR Parameters and Proton Tautomerism 350
12.2.3 NMR Hydrogen Bond Correlations 353
12.2.3.1 OHO Bonds – 1 H Chemical Shifts 353
12.2.3.2 OHO Bonds – 13 C and 31 P NMR Chemical Shifts 356
12.2.3.3 OHN Bonds 360
12.2.3.4 NHN Bonds 363
12.2.3.5 FHF, FHN, and FHO Bonds 365
12.2.3.6 Vicinal H/D Isotope Effects 369
12.2.4 IR Hydrogen Bond Correlations 371
12.2.4.1 Proton Donor Stretching Vibration 371
12.2.4.2 Proton Donor Deformational Vibrations 374
12.2.4.3 Carbonyl Stretching Vibration 375
12.3 Spectral Markers for Hydrogen Bond Energy 375
12.3.1 Defining Hydrogen Bond Energy 375
12.3.2 NMR Characterization of H-Bond Energy 377
12.3.3 IR Characterization of H-Bond Energy 378
12.3.3.1 Proton Donor Stretching Band Shift 378
12.3.3.2 Proton Donor Stretching Band Intensity 384
12.3.3.3 Proton Donor Deformational Vibrations 385
12.3.3.4 Low-Frequency Hydrogen Bond Stretching Frequency 385
12.3.3.5 Stretching Vibrations’ Force Constants 386
12.3.3.6 Carbonyl Stretching Vibration 387
References 387
13 ATR-Far-Ultraviolet Spectroscopy Holds Unique Advantages

for Investigating Hydrogen Bondings and Intermolecular
Interactions of Molecules in Condensed Phase 409
Yusuke Morisawa, Takeyoshi Goto, Nami Ueno, and Yukihiro Ozaki
13.2 Characteristics and Advantages of FUV Spectroscopy for the Studies of
Liquids and Solids 410
13.3 FUV Spectroscopic Studies of Hydrogen Bonds and Hydration Structures
of Electrolyte Aqueous Solutions 411
13.4 Quantum Chemical Calculations of the A ̃←X ̃ Transition of Hydrated
Group I Cations 412
13.5 Hydrogen Bonding States of Interfacial Water Adsorbed on an Alumina
Surface Studied by Variable Angle-ATR-FUV Spectroscopy 416
13.6 ATR-FUV and Quantum Chemical Calculation Studies of Hydrogen
Bondings in Amides 418
13.7 ATR-FUV and Quantum Chemical Calculation Studies of Hydrogen
Bondings in Nylons 422
xii Contents
13.8 An ATR-FUV Study for Poly(ethylene glycol) (PEG) and Its Complex
with Lithium Ion (Li+ ) 424
13.9 Summary and Perspective 427
References 429
14 Water–Hydrogen-Bond Network and Hydrophobic Effect 435

Barbara Zupančič and Jože Grdadolnik
Symbols and Abbreviations 435
14.2 Bulk Water 438
14.2.1 Temperature-Dependent Infrared Spectra of Bulk Water 439
14.3 Water Near Fully Hydrophobic Solutes 442
14.3.1 Verification of the Experimental Procedure 443
14.3.1.1 Effects of Temperature and Pressure on the OD-Stretching Band 446
14.3.1.2 Clathrate Formations 448
14.3.2 Pure Hydrophobic Solutes in Water Solution 449
14.3.3 MD Simulations of Purely Hydrophobic Solute in Water and the Origin
of Strengthened Water–Water–Hydrogen Bonds Near Methane
Molecule 453
14.4 IR Spectroscopy of the Water Hydrogen Bonding in the Alcohol–Water
Systems 455
14.4.1 Importance of Alcohol–Water Systems 455
14.4.2 IR Spectroscopy in the Study of Alcohol–Water Systems 455
14.4.2.1 Overview 455
14.4.2.2 Spectral Decomposition and Probes for Characterization 456
14.4.2.3 Influence of Alcohol Concentration and Temperature 464
14.5 Epilogue 470
Acknowledgments 470
References 471
15 Hydrogen Bond Chains in Foldamers and Dynamic

Foldamers 479
David T.J. Morris and Jonathan Clayden
15.1 Hydrogen-Bonded Foldamers 479
15.2 Hydrogen-Bonded Dynamic Foldamers 488
15.3 Reversible Hydrogen-Bond Directionality in Dynamic Foldamers 501
15.4 Cyclic Hydrogen Bond Chains 508
References 514
Index 521
1
Theory
Part I
3
Linear Response Theory Applications to IR Spectra of

H-Bonded Cyclic Dimers Taking into Account the
Surrounding. Updating Contributions Involving Davydov
Coupling, Fermi Resonances and Electrical Anharmonicity
Paul Blaise and Olivier Henri-Rousseau
Laboratory of Mathematics and Physics, 52 Av. Paul Alduy, 66100 Perpignan, France
1.1 Introduction
This chapter is devoted to the application of the Henri-Rousseau and Blaise model
[1] which has incorporated quantum mechanically the damping of the H-bond
bridge into the Maréchal and Witkowski model [2] to the experimental infrared
(IR) lineshapes of cyclic centrosymmetric dimers. In Figure 1.1, are depicted for
example linear and cyclic H-bonded carboxylic acids.
One may distinguish the length q of O—H bond and Q one of the H-bond. In
Figure 1.2 are recapitulated the connections between the present applied theory and
diverse older ones.
1.2 Dimer Strong Anharmonic Coupling Theory

1.2.1 Different Theoretical Situations
1.2.1.1 Strong Anharmonic Coupling Within Adiabatic Approximation For
Monomer
Let us consider a single H-bonded system where X and Y are nucleophilic sub-
stituents such as oxygen or nitrogen (See Figure 1.3). Define q and Q as the operators
corresponding to the lengths of X–H and X–Y bonds. Besides, both these lengths are
oscillating, the first one at high frequency and the last one H-bond bridge at low
frequency.
Now suppose that a strong anharmonic coupling may occur between the X-H
high-frequency mode q and the X· · ·Y low-frequency mode Q.
Within the strong anharmonic coupling theory, it is assumed a linear dependence
of the high-frequency mode 𝜔 (Q) on the H-bond bridge coordinate Q, according to:
𝜔 (Q) = 𝜔∘ + bQ with b < 0 (1.1)
where 𝜔∘ is the angular frequency of a isolated X–H bond and b some parameter.
Spectroscopy and Computation of Hydrogen-Bonded Systems, First Edition.
Edited by Marek J. Wójcik and Yukihiro Ozaki.
© 2023 WILEY-VCH GmbH. Published 2023 by WILEY-VCH GmbH.
4 1 Linear Response Theory Applications to IR Spectra
q Qa
Q R″ O H O
qa
O H O C R
R C
R C R′
O H O
qb
O Qb
(a) (b)
Figure 1.1 (a) H-bond monomer and the coordinates. (b) H-bond dimer and the
coordinates.
Blaise, Wójcik and Henri-

Rousseau
(2005) [1]
γ = γ° = 0
V=0 Maréchal and Witkowski V=0

γ° = 0 with Davydov coupling γ° = 0
(1968) [2]
V=0 V=0 V=0
Boulil et Maréchal and Witkowski Rösch-Ratner

al. γ=0 without Davydov coupling γ° = 0 (1974) [4]
(1988) [3] (1968) [2]
γ° = 0 γ° = γ = 0 γ=0
Quantum model of direct

and indirect dampings
(1998) [5]
Classical approximation
[Q,P] = 0
Semiclassical model
(2005) [6]
Sakun Commutation of stochastic Abramczyk

(1985) [7] average and integral on time (1985) [8]
Memory Robertson and Yarwood Rotation

neglect (1978) [9] neglect
Slow modulation limit
Bratos
(1957) [10]
[1] P. Blaise, M.J. Wójcik and Olivier Henri-Rousseau J. Chem. Phys. 122, (2005) 064360 ; [2] Y. Maréchal and A. Witkowski, J.
Chem. Phys. 48 (1968) 3677 ; [3] B. Boulil, O. Henri-Rousseau, P. Blaise, Chem. Phys. 126 (1988) 263 ; [4] N. Rösch, M. Ratner, J.
Chem. Phys. 61 (1974) 3344. ; [5] O. Henri-Rousseau, P. Blaise, Adv. Chem. Phys. Vol 103 in I. Prigogine, S.A. Rice, (Eds) John
Wiley&Sons, New York, 1998 p.1-186 ; [6] P. Blaise, P-M. Déjardin and O. Henri-Rousseau, Chem. Phys, 313 (2005) 177. ; [7] V. Sakun,
Chem. Phys. 99 (1985) 457. ; [8] H. Abramczyk, Chem. Phys.,94 (1985) 91. ; [9] G. Robertson, J. Yarwood, Chem. Phys. 32 (1978) 267.
; [10] S. Bratos and D. Hadzi, J. Chem. Phys. 27 (1957) 991. 40
Figure 1.2 Connections between the present theory and different older models.
The full Hamiltonian may be partitioned as follows:

Htot = Hfast + Hslow (1.2)
The Hamiltonian of the slow mode may be viewed as either harmonic or anharmonic
(Morse-like)
P2 1
Harmonic: Hslow = + MΩ2 Q2
2M 2
[ √ ]2
P2i MΩ
Morse like: Hslow = ℏΩ + De 1 − e−𝛽e Qi
2M ℏ
√ √
M ℏ
with 𝛽e = Ω
2De MΩ
Here, P is the momentum coordinate of the slow mode of reduced mass M and angu-
lar frequency Ω, whereas De is the dissociation energy of the Morse curve.
The Hamiltonian Hfast is corresponding to the 𝜈s (X–H) high-frequency mode.
Within the harmonic approximation and strong anharmonic coupling theory, it is:
p2 1 p2 1 1
+ M𝜔∘ q2 + m𝜔∘ bq2 Q + mb2 q2 Q2
2
Hfast = + m(𝜔 (Q))2 q2 =
2m 2 2m 2 2
(1.3)
whereas p is the momentum coordinates for the fast mode.
The eigenvalue equations of the fast and slow harmonic modes are given respec-
tively, neglecting the zero-point energy of the fast mode by:
( 2 )
p 1
+ M𝜔∘ q2 |{k}⟩ ≡ Hfree |{k}⟩ = ℏ𝜔∘ k |{k}⟩
2
2m 2
( 2 ) ( )
P 1 1
+ MΩ2 Q2 |(n)⟩ = ℏΩ n + |(n)⟩ (1.4)
2M 2 2
Within the adiabatic approximation the full Hamiltonian becomes simply:
[ ] ∑ [ {k} ]
Hadiab = HI |{k}⟩ ⟨{k}|
k
where
[ ] [ ]
P2 1 P2 1
H{0}
I
= + MΩ2 Q2 ; H{1}
I
= + MΩ2 Q2 + ℏbQ + ℏ𝜔∘ (1.5)
2M 2 2M 2
Figure 1.4 represents the absorption mechanism generating a coherent state.
It is possible to generalize the above approach by introducing together with the
coupling of the fast mode to the H-bond bridge, another coupling of the fast mode
with some bending mode according to:
H = H + H∘
tot fast + H∘
slow bend
with, by taking the H-bond bridge potential as Morse-like (See Table 1.1).
Figure 1.3 Coordinates of single q

H-bonded system.
X H Y
Q
∣(4)〉
∣(3)〉
∣α〉 ∣(2)〉
∣{1}〉
∣(1)〉
∣(0)〉
–∣α∣2 an
e 2 Σ ∣{m}〉
n!
∣{0}〉 ∣(0)〉
Figure 1.4 Physics of the absorption mechanism. The ground state of the slow mode
H-bond bridge (corresponding to the ground state situation of the fast mode) becomes a
coherent state 𝛼 = a | 𝛼⟩ after excitation towards the first excited state of the fast mode.
Table 1.1 Different sorts of Hamiltonians.

p2 ( )2
Hfast = + 12 m 𝜔∘ + 𝛼 ∘ Q + 𝛽Q𝛿 q2
2m [ √ M ]2
P2 −Ω
H∘ slow = 2M + De 1 − e 2De
Q
P2
H∘ bend = 2M𝛿
+ 12 M𝛿 Ω2𝛿 Q2𝛿
where Q𝛿 and P𝛿 are respectively the position and momentum coordinates of the
bending mode having Ω𝛿 as angular frequency and M𝛿 as reduced mass.
1.2.1.2 Introduction of Fermi Resonances

Now, there is the possibility to introduce Fermi resonance [11] in this physical model
as it is illustrated in Figure 1.5.
There is a coupling characterized by the parameter f1𝛿 between the two situations
evoked in Figure 1.5.
In the absence of damping, the full Hamiltonian involving Fermi resonances is:
[ ] [ ]
HFermi = HFree + Hslow + HInt + HBend + VBend (1.6)
Here, the three first right-hand side Hamiltonians are the components of the bare
H-bond Hamiltonians without Fermi resonance given respectively by equations
given in Table 1.1. Besides, the Hamiltonian HBend corresponding to the bend-
ing mode and the interaction VBend between the fast and bending modes are
respectively:
p2𝛿 1 ( )2
HBend = + m𝛿 𝜔𝛿 q2𝛿 ; VBend = l𝛿 qq2𝛿 (1.7)
2m𝛿 2
f1δ
∣(m)〉 ⊗ ∣{1}〉 ⊗ ∣[0]〉
f1δ
∣(m)〉 ⊗ ∣{0}〉 ⊗ ∣[2]〉
Figure 1.5 Fermi resonances interaction coupling parameters f1𝛿 between two situations of
the fast, slow, and bending modes. Source: Henri-Rousseau and Blaise 2008 [18]/John Wiley
& Sons.
where q𝛿 and p𝛿 are respectively the position and momentum coordinates of the
bending mode of reduced mass m𝛿 and 𝜔𝛿 its angular frequency, whereas l𝛿 is the
coupling parameter between the fast and bending modes. The eigenvalue equations
of the harmonic Hamiltonians corresponding respectively to the fast and slow modes
are respectively given by equations given by Eqs. (1.4) whereas that dealing with the
bending modes is, ignoring the zero-point energy:
|[ ]⟩ |[ ]⟩
HBend | l𝛿 = l𝛿 ℏ𝜔𝛿 | l𝛿 (1.8)
| |
Now, within the adiabatic approximation. The Hamiltonian (1.6) becomes:
[ Adiab ] [ ] [ ] [ ]
HFermi = HAdiab + VBend + HBend (1.9)
The different Hamiltonians are given as follows:

[ ] [ ] [ ]
HAdiab = H{1} I
|{1} ⟩ ⟨ {1}| + H{0}
I
|{0} ⟩ ⟨ {0}|
[ ] 2
P 1
H{k}
I
= + MΩ2 Q2 + kbQ + kℏ𝜔∘
[ ] 2 M 2
HBend = 2ℏ𝜔𝛿 |[2]⟩ ⟨[2]|
[ ] [ ]
VBend = |{0}⟩ |[2]⟩ ⟨[0]| ⟨{1}| ℏf 𝛿 + hc
Here, f1𝛿 is the anharmonic coupling parameter involved in the Fermi resonance
which is a function of l𝛿 .
As a consequence of the above equations, the full Hamiltonian describing the fast
mode coupled to the H-bond bridge (via the strong anharmonic coupling theory)
and the bending mode (via the Fermi resonance process) may be written within the
tensorial basis (1.10) according to [12]:
⟩
⎛||Ψa ({0} (m) [0])⟩ ⎞ ⎛|{0} (m) [0]⟩ ⎞
⎜|Ψ ({1} (m) [0]) ⎟ = ⎜|{1} (m) [0]⟩ ⎟ (1.10)
⎜|| b ⟩⎟ ⎜ ⎟
⎝ |Ψc ({0} (m) [2]) ⎠ ⎝|{0} (m) [2]⟩ ⎠
⎛ HII
{0}
0 0 ⎞
[ Adiab ] ⎜ ⎟
ℏf 𝛿1
{1}
HFermi = ⎜ 0 HII ⎟ (1.11)
⎜ 0 ℏf 1 𝛿 {0}
𝛿⎟
HII + 2ℏ𝜔1 ⎠
⎝
1.2.1.3 H-Bonded Centrosymmetric Dimer

Now, look at an H-bonded dimer. It will take place in a Davydov coupling [13].
Within the anharmonic coupling, the physics of the system may be viewed in
Figure 1.6.
It may be observed that because of the symmetry of the dimer, there is a C2 operator
̂ which exchanges the coordinates Q of the two slow modes H-bond
(with C22 = 𝟏), i
bridges of the cyclic dimer according to:
C2 Qa = Qb ; C2 Qb = Qa ; C2 Pa = Pb ; C2 Pb = Pa (1.12)
Ignoring for the present time the interaction between the two moieties and assuming
that, within each moiety, the adiabatic approximation may be performed as for a
V°
∣(m)2〉 ⊗ ∣{1}2〉 ∣{0}1〉 ⊗ ∣(m)1〉
V°
∣(m)2〉 ⊗ ∣(0)2〉 ⊗ ∣{1}1〉 ⊗ ∣(m)1〉
Figure 1.6 Davydov coupling interactions. Source: Henri-Rousseau and Blaise 2008
[18]/John Wiley & Sons.
single H-bond, the Hamiltonian of the symmetric dimer embedded in the thermal
bath, is:
[ {Adiab} ] [ {Adiab} ] [ {Adiab} ]
H = H + H (1.13)
a b
In Eq. (1.13), the two first right-hand side terms are the adiabatic Hamiltonians of
each moiety. They are given by an expression of the same form which is:
[ {Adiab} ] [ {0} ] ⟩ ⟨ [ ] ⟩⟨
H = HII |{0} {0} | + H{1} |{1} {1}i || with i = a, b
i
| i i| II
i
| i
i
⟩
|{k} are the eigenkets of the Hamiltonians of the fast modes harmonic oscillators,
| i
whereas the Hamiltonians of each moiety are respectively:
[ {0} ] [ {Slow} ]
H∘ II = H
i
[ {1} ] [ {Slow} ]i [ {Int} ]
∘
H II = H + bQi + ℏ𝜔∘ + ℍII
i i i
Here, the last term is the interacting coupling with the thermal bath that we shall
ignore in the present simplified exposition. The Hamiltonian of the cyclic dimer
involving Davydov coupling between the first excited state of the high-frequency
oscillator a of one moiety and the excited state of the oscillator b of the other moiety
and vice versa is,
[ ] [ {Adiab} ] [ {Adiab} ]
ℍDav = H + H + VDav (1.14)
a b
The Davydov coupling Hamiltonian VDav appearing in this equation may be written
either simply or as a function of the two slow modes coordinates [14]:
[ ⟩ ⟨ ⟩ ⟨ ]
VDav = V∘ || {1}a {0}b || + ||{0}a {1}b ||
( ) ( )[ ⟩ ⟨ ⟩ ⟨ ]
VDav Q1 , Q2 = VDav + Θ Q1 + Q2 || {1}a {0}b || + ||{0}a {1}b ||
where Θ is a dimensionless parameter governing the linear dependence of the Davy-

dov coupling operator on the H-bond bridge coordinates.
When ignoring the Θ coupling, then, within the following basis:
⟩
⎛|| 𝚽{0,0} ⎞ ⟩⨂ ⟩
⎜| {a,b} ⟩⎟ ⎛|| {0}a ⟩ || {0}b ⟩⎞
⎜|| 𝚽{1,0} ⎟ = ⎜| {1} ⨂ | {0} ⎟ (1.15)
⎜| {a,b} ⟩⎟ ⎜|| a⟩
⨂ || b⟩
⎟
|
⎜| 𝚽 {0,1}
⎟ ⎝| {0}a | {1}b ⎠
⎝| {a,b} ⎠
the Davydov Hamiltonian (1.14) takes the matrix form:
[ ]
⎛ H{0,0} 0 0 ⎞
[ ] ⎜ II [ {1,0} ] ⎟
ℍDav = ⎜ 0 HII VDav ⎟ (1.16)
⎜ [ {0,1} ] ⎟
⎜ 0 V HII ⎟⎠
⎝ Dav
with respectively:
( )
[ {0,0}
] ∑ P2i MΩ2 Q2i
HII = + with i = a, b
i
2M 2
( )
[ ] ∑ P2i MΩ2 Q2i
+ ℏ𝜔∘
{1,0}
HII = + + bQi
i
2M 2
Then, owing to the symmetry properties given by Eqs. (1.12), it appears that the
parity operator exchanges the two last Hamiltonians:
[ {1,0} ] [ {0,1} ]
C2 HII = HII (1.17)
To diagonalize the Davydov Hamiltonian, one may perform the following basis
change.
⟩ ⟩
⎛ || 𝚽{0,0} ⎞ ⎛ | {0,0}
| 𝚽 ⎞
⎜ | {a,b} ⟩⎟ ⎜ {1,0} ⟩| {a,b} ⟩⎟
⎜|| 𝛃(+) ⎟ = ⎜|| 𝚽 |
+ Ĉ 2 | 𝚽{a,b} ⎟
{0,1}
(1.18)
⎜| {1,0}↔{0,1} ⟩⎟ ⎜| {a,b} ⟩ | ⟩⎟
|
⎜| 𝛃 ⎟ ⎜ | {1,0}
̂ | {0,1}
⎟
⎝| {1,0}↔{0,1} ⎠ ⎝|| 𝚽{a,b} − C2 || 𝚽{a,b} ⎠
(−)
Then the Davydov Hamiltonian becomes:

[ {0,0} ]
⎛ HII 0 0 ⎞
⎜ [ {1,0} ] ⎟
ℍDav = ⎜ 0 HII + V∘ Ĉ 2 0 ⎟ (1.19)
⎜ [ {1,0} ]⎟
⎜ 0 0 HII − V∘ Ĉ 2 ⎟⎠
⎝
Moreover, to make tractable the action of the C2 operator, it is suitable to pass to
the symmetrized coordinates and their conjugate momenta according to Figure 1.7.
In Table 1.2, are given the symmetrized coordinates in the Davydov coupling
model.
Recall here, the improvement brought by Rekik et al. [15–17], by introducing the
electrical anharmonicity. As quoted above, the dependence of the Davydov coupling
on the slow mode coordinates reduces to one on Qg through:
( ) √
VD Qg = V ∘ + 2ΘQg
Now, the action of the parity operator on the symmetrized operators

[ (Slow) ] and the sym-
metrized ground state and first excited state of Hamiltonian H appearing in
i
Table 1.2, is depicted in Figure 1.7:
Table 1.2 Symmetrized coordinates.

1 [ ] 1 [ ] 1 [ ] 1 [ ]
Qg = √ Qa + Qb Qu = √ Qa − Qb Pg = √ Pa + Pb Pu = √ Pa − Pb
2 2 2 2
C2 Qg = Qg C2 Qu = −Qu C2 Pg = Pg C2 Pu = −Pu
Symmetric coordinate Antisymmetric coordinate

Qg = 1 [Qa + Qb] ∣(1)g〉 = 1 [∣(1)a〉 + ∣(1)b〉 Qu = 1 [Qa – Qb] ∣(1)u〉 = 1 [∣(1)a〉 – ∣(1)b〉
2 2 2 2
C2 C2
Ground states
Qa Qa
∣(0)a〉 ∣(0)a〉
∣(0)b〉 Q Qb
b ∣(0)b〉
C2 ∣(0)g〉 = ∣(0)g〉
C2 ∣(0)u〉 = ∣(0)u〉
C2 C2
First excited states
Qa
Qa ∣(1)a〉
∣(1)a〉
∣(1)b〉 –Qb
∣(1)b〉 Q
b
C2 ∣(1)u〉 = ∣(1)u〉
C2 ∣(1)g〉 = ∣(1)g〉
Figure 1.7 Action of the C2 operator on coordinates and eigenstates. Source:

Henri-Rousseau and Blaise 2008 [18] / John Wiley & Sons.
Next, we may consider the kets (1.18) as the result of the tensorial product of states,
according to:
⟩ ⟩ ⟩
⎛||Φ{0,0} ⎞ ⎛||{0} ⊗ | {0} ⎞
⎜ | ⟩⎟ ⎜| | g | u
⟩⎟
⟩ { }
{g,u}
⎜||𝛽 (+) ⎟ = ⎜|{1} ⊗ || 𝛽 (+) {1} ⎟ (1.20)

⎜| {1,0}↔{0,1} ⟩⎟ ⎜| g ⟩ |{ }
u+ ⟩
⎟
|
⎜|𝛽 (−) | |
⎟ ⎜ {1} ⊗ | 𝛽 (−) − ⎟ {1}
⎝| {1,0}↔{0,1} ⎠ ⎝|| g | u ⎠
with:
( ⟩ ) ( 1 [| ⟩ ⟩])
| {0}a + || {0}b
| {0}g √
2 |
| ⟩ = 1 | [ ⟩ ⟩] (1.21)
| {0}
| u
√
2 |
{0}a − || {0}b
within the framework of the symmetrized coordinates the Hamiltonian (1.19) takes
the form:
[ ]
⎛ H{0} 0 ⎞ ⟩
⎜ II g [
0
]
0 0
⎟ ⎛||{0}g ⎞
⎜ {1} ⎟ ⎜|| ⟩ ⎟
⎜ 0 HII 0 0 0 ⎟ ⎜ |{1} ⎟
[ ] ⎜ g [ ] ⎟ ⎜||
g
⟩ ⎟
ℍDav = ⎜ 0 {0}
{0}
0 HII 0 0 ⎟ within ⎜|{ u } ⟩⎟
⎜ u [ {1} ] ⎟ |
⎜| 𝛽 (+) + ⎟
{1}
⎜ 0 0 0 H(+) 0 ⎟ ⎜|{ u ⟩⎟
⎜ u+
[ {1} ] ⎟ ⎜|| 𝛽 (−) }{1} ⎟
⎜ 0 0 0 0 H(−) ⎟ ⎝| u− ⎠
⎝ u− ⎠
(1.22)
with respectively:
[ {0} ] P2g MΩ2 Q2g
HII = +
g 2M 2
[ {1} ] P2g MΩ2 Q2g 1
HII = + + √ bQg + ℏ𝜔∘
g 2M 2 2
[ {0} ] P 2 2 2
MΩ Qu
HII = u +
u 2M 2
[ {1} ] P2 MΩ2 Q2u ( )
1
and H(±) = u + + √ b Qu ± V∘ Ĉ 2 u
u 2M 2 2
1.2.1.4 Dimer Involving Damping, Davydov Coupling, and Fermi Resonances
Now, it is possible to introduce Fermi resonances in the precedent model taking into
account direct and indirect dampings, together with the relaxation of the bending
modes [18, 19].
For the special case of a single Fermi resonance, the physics related to this Hamil-
tonian is depicted in Figure 1.8.
Here, the right-hand side Hamiltonian is that dealing with the Fermi resonances
occurring between the g excited state of the fast mode and the g first harmonics of
the bending mode. For one Fermi resonance, this Hamiltonian is:
[ ]
[ {1} ] ⎛ ℍ{1} ℏf 𝛿 ⎞
ℍFermi = ⎜ II g ⎟
g ⎜ ℏf 𝛿 [H
{0}
] + E ⎟
⎝ II g ⎠
Besides, the fi is the coupling parameters involved in the Fermi resonances
expressed as angular frequencies, whereas the Δ are the angular frequency gap:
( )
E = ℏ𝜔∘ − 𝛼 ∘ ℏΩ + ℏΔ𝛿 − 𝜔∘ − 2𝜔𝛿
2
i
f f
Figure 1.8 Davydov coupling with an unique Fermi resonance. Source: Henri-Rousseau
and Blaise 2008 [18]/John Wiley & Sons.
As a consequence, the Hamiltonian of the dimer involving Davydov coupling,

Fermi resonances between the g excited state of the fast mode and the g first har-
monics of the bending mode is:
[ ]
⎛ ℍ{0} 0 0 0 0 ⎞
⎜ II
g [ ] ⎟
⎜ ⎟
ℍ
{1}
⎜ 0 0 0 0 ⎟
[ Fermi ] ⎜
Fermi
g [ ] ⎟
ℍDav = ⎜ 0
{0}
0 HII 0 0 ⎟
⎜ u [ ] ⎟
{1}
⎜ 0 0 0 H(+) 0 ⎟
⎜ u [ ] ⎟
⎜ 0 H(−) ⎟
{1}
0 0 0
⎝ u⎠
Now, proceed as to the corresponding situation when Fermi resonance is not taken
into account, it is possible to obtain the spectral density of the present system.
1.2.2 The Spectral Density

The transition dipole moment may be, without and with the electrical
anharmonicity, given by:
Without electrical anharmonicity: 𝜇 (q) = q
With electrical anharmonicity: 𝜇 (q, Q) = q (1 + 𝜁Q)
in which 𝜁 is the electrical anharmonicity parameter.

When ignoring the electrical anharmonicity, the ACFs for the IR and the Raman
absorptions may be given respectively by
(u)
[ [ ]† [ ]]
GDav (t) = trDav 𝜌Dav 𝜇(0)u 𝜇(t)u for pure IR
(g)
[ [ ]† [ ]]
GDav (t) = trDav 𝜌Dav 𝜇(0)g 𝜇(t)g for pure Raman
At time t the u transition operator is obtained by the Heisenberg transformation

involving the above Hamiltonian HDav :
1 iHDav t∕ℏ
𝜇 u (t) = √ e
2
[(( ) ⟩ ( ) ⟩) ⟨ ] −iHDav t∕ℏ
𝜇 b ||{0}a {1}b − 𝜇 ∘ a ||{1}a {0}b
∘ {0}b {0}a || e
1 iHDav t∕ℏ
𝜇 g (t) = √ e
2
[(( ) ⟩ ( ) ⟩) ⟨ ] −iHDav t∕ℏ
𝜇 b |{0}a {1}b + 𝜇 ∘ a ||{1}a {0}b
∘ | {0}b {0}a || e
It may be shown that the ACF takes the form:

[ ] [[ ] [ ] ]( ∘ )
GDav (t) = [G(t)]g G(+) (t) u + G(−) (t) u e−𝛾 t (1.23)
where 𝛾 ∘ is the relaxation parameter of the high-frequency mode. Besides, [G(t)]g

is the ACF of the g part of the system and where appears the damping parameter
𝛾 of the H-bond bridge and the function of it 𝛾̃ . The quantum calculation of the
damping is somewhat complex and appears only on the g part of the ACF. Besides,
[ ]
this last ACF and the G(±) (t) u which are the ACFs corresponding to the u parts, are
respectively:
( ( )[ 2 ]
2 ) 𝛼̃ ∘ [⟨n⟩+1∕2](2e−̃𝛾 t∕2 cos(Ωt)−e−𝛾t −1)
( ∘ )2 i𝜔∘ t i𝛼̃ ∘ e−̃𝛾 t∕2 sin Ωt
2
−i2𝛼̃ ∘ Ωt
G(t) = 𝜇g e e u e e
∑∑ −nu ℏΩ ∕k T
B
i𝜔±
𝜇t −inu Ωt (1.24)
G(±) (t)u = 𝜀 |B±𝜇 nu |2 e e e
𝜇 nu
[ ]2
[1 ± (−1) ] + 𝜂 ∘ [1 ∓ (−1)nu +1 ]
nu +1
[ {1} ] ⟩ ⟩ ⟩ ∑ ⟩
| (±) | | | (n)
H± | 𝛽𝜇 = ℏ𝜔±𝜇 | 𝛽𝜇(±) with | 𝛽𝜇(±) = B±nu 𝜇 |
u| | | u
nu
[ ] P2 MΩ2 Q2u ( )
1
+ √ b Qu ± V∘ Ĉ 2 u
{1}
H(±) = u +
u 2M 2 2
The lineshape of the H-bonded dimer is then the Fourier transform of the ACF:
[ ] ∞
−i𝜔t
( −𝛾 ∘ t )
IDav (𝜔) = GDav (t)e e dt (1.25)
∫−∞
[ ± ] ∑∑ [ ] ∑∑ [[ ] [ ]]2
IDav (𝜔) ≃ Pmg ng e−𝜆nu 1 ± (−1)nu +1 + 𝜂 ∘ 1 ∓ (−1)nu +1
mg ng nu 𝜇
[ ]
| ± |2
|Bnu ,𝜇 | I±mg ng nu 𝜇 (𝜔)
| |
where the components of the above formulas are given as follows:
[ ] 𝛾mg ng
I±mg ng nu 𝜇 (𝜔) ≃ ( )2 ( )2
𝜔 − Ω±mg ng nu 𝜇 + 𝛾mg ng
with
[( ) ]
Ω±mg ng nu 𝜇 = 𝜔∘ − mg − ng + nu Ω − 𝜔±𝜇 − 2 𝛼̃ ∘ Ω
2
( ) √
𝛾mg ng = mg + ng 𝛾̃ + 𝛾 ∘ with 𝛾̃ = 𝛾 2
[ ]n
[ ] 1 + ⟨n⟩ g 𝛼̃ ∘ 2(mg +ng )
Pmg ng =
mg !ng !
𝛼 ∘ 1 ℏΩ
𝛼̃ ∘ = √ , ⟨n⟩ = ̃ , and 𝜆̃ =
𝜆 k
2 e −1 BT
1.3 Comparison with Experiments

1.3.1 Carboxylic Acid Dimers Ignoring Fermi Resonances
1.3.1.1 Gaseous and Liquid Acetic Acid Dimers
In 2005 [1], it has been proposed an approach of the lineshape in which was
introduced the quantum theory of the lineshapes of H-bonded cyclic dimers. This
approach was applied to gaseous and liquid acetic acid dimers [19].
OD OH
Intensity
1900 2750 3600

(a) Wavenumbers (cm–1)
OD OH
Intensity
1900 2750 3600

(b) Wavenumbers (cm–1)
OD OH
Intensity
1900 2750 3600

(c) Wavenumbers (cm–1)
Figure 1.9 IR 𝜈XH lineshapes of gaseous cyclic acetic acid CD3 CO2 H/D dimers at room
temperature. Experimental lineshapes (grayed) of Novak and coworker [20].
Figure 1.9 reproduces the corresponding lineshapes involved in the spectra for
cyclic CD3 CO2 H/D dimers in the gas phase at room temperature.
The corresponding data are given in Table 1.3.
Moreover, the comparison for the experimental lineshapes dealing with CH3 CO2 H
in the gas and liquid phase measured by Flakus and the corresponding theoretical
ones [21] resulting from Eq. (1.25) are given in Figure 1.10.
1.3.1.2 Gaseous Acrylic and Propynoic Acids

Now, Figure 1.11a,b compares the experimental lineshapes of O–H and O–D gaseous
acrylic and propynoic acids as measured by Bournay and Maréchal [22] to the cor-
responding theoretical results of [21].
Table 1.3 Parameters used for fitting the experimental lineshapes of Figure 1.9.
Case Isotope 𝝎∘ (cm−1 ) 𝛀 (cm−1 ) 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) V∘ (𝛀) 𝜼∘
(a) O–H 3320 108 1.414 0.15 0 −1.10 0

O–D 2317 98 0.95 0.15 0 −0.96 0
(b) O–H 3320 108 1.414 0.15 1.10 −1.10 0
O–D 2317 98 0.95 0.15 1.10 −0.96 0
(c) O–H 3100 88 1.19 0.20 0.20 −1.50 0.30
O–D 2263 80 0.77 0.15 0.20 −1.15 0.20
Intensity
2000 Wavenumbers 3500 (cm–1)

Liquid phase
Intensity
2000 Wavenumbers 3500 (cm–1)

Gas phase
Figure 1.10 Gas and liquid 𝜈(X − H) IR lineshapes of CH3 CO2 H at room temperature.
Comparison with experiment. Data are given in Table 1.4. Source: Benmalti et al. [21], figure
1 (p. 272)/With permission of Elsevier.
1.3.2 Carboxylic Acids Taking Into Account Fermi Resonances

1.3.2.1 Crystalline Adipic Acid
In this section, we consider the lineshapes of O–H and O–D crystalline adipic acid.
Figure 1.12 gives the comparisons of the experimental lineshapes measured by
Auvert and Maréchal [23] of –OH and –OD crystalline adipic acids at different
temperatures with the corresponding theoretical ones where Fermi resonances are
absent (a) or present (b) in the model [21]. Effect of the 𝜂 parameter is also shown.
Table 1.4 Parameters used for fitting the experimental lineshapes of acetic acid given in
Figure 1.10.
𝝎∘ (cm−1 ) 𝛀(cm−1 ) 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) V ∘ (𝛀) 𝜼∘
Liquid 3100 88 1.19 0.10 0.24 −1.55 0.15

Gas 3100 88 1.19 0.24 0.24 −1.55 0.25
D H D H
Intensity
Intensity
2000 3000 2000 3000

(a) Wavenumbers (cm–1) (b) Wavenumbers (cm–1)
Figure 1.11 Gaseous acrylic (a) and propynoic acids. Comparisons of experimental
(grayed) and theoretical lineshapes for H and D isotopic species. Experiment (grayed). The
data are given in Table 1.5. Source: Based on Bournay and Maréchal [22].
Table 1.5 Parameters used for fitting the experimental lineshapes of Figure 1.11.
Case Isotope 𝝎∘ (cm−1 ) 𝛀(cm−1 ) 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) V ∘ (𝛀) 𝜼∘
(a) Acrylic O–H 3020 71 1.09 0.22 0.27 −1.66 0.19

acid O–D 2237 71 0.77 0.24 1.80 −1.08 0.15
(b) Propynoic O–H 3032 86 1.25 0.27 0.36 −1.94 0.16
acid O–D 2260 86 0.88 0.18 1.40 −1.25 0.05
1.3.2.2 Crystalline Polarized and Unpolarized Glutaric Acid Taking Into

Account Fermi Resonances
Figure 1.13 compares theoretical and experimental lineshapes of O–H and O–D crys-
talline glutaric acid as measured by Flakus and Miros [24] at 298 and 77 K for two
different polarizations, taking or not into account Fermi resonances.
Now, in presence of Fermi resonances, the theoretical lineshapes may be improved
(see Figure 1.14).
The corresponding parameters are given in Table 1.9 a and b.
1.3.2.3 Crystalline Thiopheneacetic Acid and Thiopheneacrylic Acids

Rekik et al. [31], have theoretically studied within the standard model, the line-
shapes of crystalline thiopheneacetic acid and thiopheneacrylic acids by taking into
account the coordinate dependence Θ of the Davydov coupling on slow mode coordi-
nates, and has compared their results to the experimental ones measured by Flakus
and Chelmecki [25] in the case where there the polarization is zero.
T(K) η≠0 η=0 T(K) η≠0
O–D O–H O–D O–H

D H
Intensity
I(ω)
I(ω)
10 1700 Wavenumbers (cm–1) 3400 1700 Wavenumbers (cm–1) 3400 10 1700 Wavenumbers (cm–1) 3400
O–H D H
O–H O–D
O–D
Intensity
I(ω)
I(ω)
1700 3400
100 1700 Wavenumbers (cm–1) 3400 Wavenumbers (cm–1) 100 1700 Wavenumbers (cm–1) 3400
D H
O–H O–D O–H
O–D
Intensity
I(ω)
I(ω)
1700 3400
200 1700 Wavenumbers (cm–1) 3400 Wavenumbers (cm–1) 200 1700 Wavenumbers (cm–1) 3400
O–D O–H D H
O–D O–H
Intensity
I(ω)
I(ω)
1700 3400
300 Wavenumbers (cm–1) 1700
Wavenumbers (cm–1)
3400 300 1700
3400
(a) (b)
Figure 1.12 Effects of temperature, isotopic substitution, and 𝜂 parameter on the lineshapes of crystalline adipic acid (a) without Fermi resonance (b)
with 4 Fermi resonances. Experiment (grayed). Theory (continuous line). The data are in Tables 1.6 and 1.7a,b. Source: Based on [23].
Table 1.6 Parameters used for fitting the experimental lineshapes of crystalline adipic
acid given in Figure 1.12a.
Case T(K) 𝝎∘ (cm−1 ) 𝛀 (cm−1 ) 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) V ∘ (𝛀) 𝜼∘
O–H 10 2990 115 1.00 0.40 0.00 −0.88 0.80

100 2970 115 1.00 0.40 0.00 −0.88 0.70
200 2990 115 1.00 0.33 0.20 −0.88 0.50
300 3020 115 1.00 0.33 0.40 −0.88 0.41
O–D 10 2199 115 0.39 0.22 0.00 −0.54 0.95
100 2199 115 0.39 0.22 0.00 −0.54 0.95
200 2203 115 0.39 0.20 0.20 −058 0.85
300 2199 115 0.35 0.20 0.40 −0.68 0.75
Table 1.7a Parameters used for fitting the experimental lineshapes of Figure 1.12b. in
presence of Fermi resonances.
case T(K) 𝝎∘ (cm−1 ) 𝛀 (cm−1 ) 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) V ∘ (𝛀) 𝜼∘
O–H 10 2865 108 0.88 0.11 0.16 −0.88 0.80

100 2850 108 0.88 0.11 0.18 −0.88 0.70
200 2870 108 0.88 0.10 0.34 −0.89 0.50
300 2890 108 0.88 0.12 0.30 −0.93 0.41
O–D 10 2170 108 0.29 0.22 0.00 −0.54 0.95
100 2170 108 0.29 0.22 0.10 −0.54 0.95
200 2170 108 0.29 0.20 0.20 −058 0.85
300 2172 108 0.29 0.20 0.35 −0.68 0.75
Table 1.7b Fermi resonances parameters for crystalline adipic acid in presence of Fermi
resonances given in Figure 1.12b.
Case T (K) f1v f2v f3v f4v 𝚫v1 𝚫v2 𝚫v3 𝚫v4 𝜸1v 𝜸2v 𝜸3v 𝜸4v
(cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (𝛀) (𝛀) (𝛀) (𝛀)
10
O–H 100 30 37 36 36 −155 −80 20 220 0.10 0.10 0.10 0.05
200
300
10
O–D 100 17 15 13 13 −33 −21 −0.84 126 0.07 0.04 0.17 0.10
200
300
Pol = 0° Pol = 90°

O–H O–D O–H
O–D
Intensity
Intensity
1700 2600 3500 1700 2600 3500
77 K Wavenumbers (cm–1) Wavenumbers (cm–1)
O–D O–H
Intensity
Intensity
1700 2600 3500 1700 2600 3500
Figure 1.13 Temperature and isotopic substitution effects at different polarizations for
crystalline glutaric acid without Fermi resonances. Grayed: Experimental lineshapes of
Flakus and Miros [24].
Table 1.8 Parameters used for fitting the experimental lineshapes of Figure 1.13 of
crystalline glutaric acid without Fermi resonance.
Pol (∘ ) Case T(K) 𝝎∘ (cm−1 ) 𝛀 (cm−1 ) 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) V ∘ (𝛀) 𝜼∘
O–H 77 3123 85 1.50 0.40 0.10 −1.20 0.60

0 O–H 298 3123 85 1.50 0.15 0.90 −1.20 0.20
O–D 77 2203 85 0.38 0.30 0.10 −0.85 1.30
O–D 298 2203 85 0.38 0.50 0.20 −0.85 0.60
O–H 77 3123 85 1.50 0.40 0.10 −1.20 0.75
90 O–D 298 3123 85 1.50 0.15 0.90 −1.20 0.30
O–H 77 2208 85 0.38 0.30 0.10 −0.85 1.30
O–D 298 2208 85 0.38 0.20 0.20 −0.85 0.70
Crystalline Thiopheneacetic Acid For the crystalline thiophene acetic acid they have
found the spectra given in Figure 1.15.
The corresponding parameters are given in Table 1.10.
Crystalline H(D)-3-Thiopheneacrylic Acid For the crystalline H(D)-3-thiopheneacrylic

acid they obtained the results given in Figure 1.16. (Table 1.11)
The corresponding parameters are given in Table 1.11.
1.3.2.4 l.2-Naphtylacetic Acid (2-NA) Crystals

In this chapter, Ghalla and coworkers [26] presents results of a theoretical study on
𝜈s (O–H(D)) band shapes in the polarized IR spectra of 2-naphtylacetic acid (2-NA)
crystals measured at the temperature of liquid nitrogen, based on our original theory.
Pol = 0° Pol = 90°

Intensity
Intensity
1700 2600 3500 1700 2600 3500
O–D O–H
O–D O–H
Intensity
Intensity
1700 2600 3500 1700 2600 3500
Wavenumbers (cm–1) Wavenumbers (cm–1)
298 K
Figure 1.14 Temperature and isotopic substitution effects at different polarizations for
crystalline glutaric acid with Fermi resonances. Grayed: Experimental lineshapes of Flakus
and Miros [24]. Source: Based on Flakus and Miros [24].
Table 1.9a Parameters used for fitting the experimental lineshapes of crystalline glutaric
in presence of Fermi resonance, given in Figure 1.14.
Pol (∘ ) Case T (K) 𝝎∘ (cm−1 ) 𝛀 (cm−1 ) 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) V ∘ (𝛀) 𝜼∘
O–H 77 3063 90 1.50 0.20 0.10 −115 0.55

0 O–H 298 3063 90 1.50 0.15 0.20 −1.15 0.12
O–D 77 2202 90 0.38 0.15 0.20 −0.82 1.00
O–D 298 2217 90 0.38 0.10 0.20 −0.82 0.40
O–H 77 3083 90 1.50 0.20 0.20 −1.15 0.70
0 O–D 298 3083 90 1.50 0.25 0.25 −1.15 0.40
O–H 77 2188 90 0.38 0.20 0.20 -0.82 1.00
O–D 298 2217 90 0.38 0.25 0.25 −0.82 0.50
The line shapes were studied within the frameworks of our original theory of strong
anharmonic coupling, Davydov coupling, Fermi resonance coupling, direct and indi-
rect damping, and a selection rule breaking mechanism for forbidden transitions
in IR.
The present approach (see Figure 1.17) correctly fits the experimental line shape
of the hydrogenated compound and predicts satisfactorily the evolution in the line-
shapes with isotopic substitution. Numerical calculations show that mixing of all
these effects allows one to reproduce satisfactorily the main features of the experi-
mental IR lineshapes of hydrogenated and deuterated 2-NA crystals and is expected
to confirm the importance of the Fermi resonances in reproducing the experimental
spectra. Parameters are given in Tables 1.12 and 1.13.
Table 1.9b (continued). Fermi resonances parameters for crystalline glutaric acid (cf. Figure 1.14).
Pol Case T (K) f1v f2v f3v f4v f5v 𝚫v1 𝚫v2 𝚫v3 𝚫v4 𝚫v5 𝜸1v 𝜸2v 𝜸3v 𝜸4v 𝜸5v
(cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (Ω) (Ω) (Ω) (Ω) (Ω)
O–H 77
0 O–H 298 40 25 20 30 30 −430 −120 30 100 200 0.10 0.10 0.10 0.10 0.10
O–D 77
O–D 298
O–H 77 23 0.05 0.05 0.05 0.05 0.05
90 O–H 298 40 0 20 30 30 −430 −430 30 310 200 0.10 0.10 0 10 0.10 0.10
O–D 77 25 0.10 0.10 0.10 0.10 0.10
O–D 298 0 0.10 0.10 0.10 0.10 0.10
77 K 300 K
1.0 1.0
Intensity
Intensity
0.0 0.0
2000 3000 2000 3000
H Wavenumbers (cm–1) Wavenumbers (cm–1)
1.0 1.0
Intensity
Intensity
0.0 0.0
2000 3000 2000 3000
D Wavenumbers (cm–1) Wavenumbers (cm–1)
Figure 1.15 Crystalline H(D)-3-thiophenacrylic acid experimental (H-3TAcetic) (grayed)

and theoretical lineshapes at different temperatures and H/D isotopic species. Source:
Modified from Rekik et al. 2015 [31].
Table 1.10 Parameters used for fitting the experimental lineshapes of the
3-thiopheneacetic (H-3TAcetic) acid crystals dimers and their deuterated analogs (Pol = 0).
Species T (K) 𝝎∘ (cm−1 ) 𝛀(cm−1 ) 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) V D∘ (𝛀) 𝚯
H-3TAcetic 77 3180 105 1.55 0.35 0.35 0.50 0.15

H-3TAcetic 300 3050 105 1.2 0.3 0.25 1.2 0.22
D-3TAcetic 77 2250 80 0.67 0.15 0.25 0.52 0.2
D-3TAcetic 300 2250 85 0.63 0.17 0.25 0.7 0.2
1.3.2.5 Crystalline Aspirin Dimers Involving Slow Mode Morse Potential

The application of our treatment for Davydov coupling has been performed with
aspirin by Rekik and coworkers [27] by accounting for the anharmonicity of the
slow mode which is described by a “Morse” potential with a dissociation energy of
De = 2100 cm−1 to reproduce the polarized IR spectra of the hydrogen and deuterium
bond in acetylsalicylic acid (aspirin) crystals.
Within the adiabatic approximation, the Hamiltonian of each moiety of the dimer
may be put on the form of sum of effective Hamiltonians which are depending on
the degree of excitation of the fast mode.
77 K 300 K
1.0 1.0
Intensity
Intensity
0.0 0.0
2200 2400 2600 2800 3000 3200 3400 2200 2400 2600 2800 3000 3200 3400
–1
Wavenumbers (cm ) Wavenumbers (cm–1)
H
1.0
1.0
Intensity
Intensity
0.0
1800 2000 2200 2400 2600 2800 3000 3200 3400 0.0
2000 2400 2800 3200
D Wavenumbers (cm–1)
Figure 1.16 Crystalline H(D)-3-thiophenacrylic acid experimental (black line); theoretical

(red line) lineshapes at different temperatures for H and D isotopic species. Source: Rekik
et al. 2015 [31]/With permission of Elsevier.
Table 1.11 Parameters used for fitting the experimental lineshapes of the crystalline
3-thiopheneacrylic (H-3TAcrylic) acid dimers and their deuterated analogs.
Pol = 0 T (K) 𝝎∘ (cm−1 ) 𝛀(cm−1 ) 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) V D∘ (𝛀) 𝚯
H-3TAcrylic 77 2950 85 1.1 0.3 0.3 0.9 0.23

H-3TAcrylic 300 3020 85 1.23 0.3 0.15 1 0.23
D-3TAcrylic 77 2125 75 0.65 0.22 0.2 0.65 0.12
D-3TAcrylic 300 2150 75 0.63 0.24 0.25 0.65 0.4
These corrections are introduced in the usual procedure for taking into account
the Davydov coupling. The theoretical lineshapes obtained are compared to the
experimental ones obtained by these authors for the two isotopic species at 77 and
300 K temperatures. Their results are given in Figure 1.18 and the coresponding
parameters in Table 1.14.
Pol = 0° Pol = 90°

Fermi
resonance
1.0 1.0
0.8
Intensity
0.6
0.4
0.2
0.0 0
2000 3200 2000 3200
NO Wavenumbers (cm–1) Wavenumbers (cm–1)

1.0 1.0
Intensity
Intensity
0
0
2000 3200 2000 3200
YES Wavenumbers (cm–1) Wavenumbers (cm–1)
Figure 1.17 2-Naphtylacetic Acid (2-NA). Comparison of experimental lineshapes with

theoretical ones for different polarizations and isotopic substitutions, without and with 3
Fermi resonances. Source: Issaoui et al. 2013 [26]/ Springer Nature.
Table 1.12 Parameters used to fit experimental H/D-2-NA spectra.
( ) ( )
Species Pol(∘ ) 𝝎∘ cm−1 𝛀 cm−1 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) V ∘ (ℏ𝛀) 𝜼
–H 0 3135 67 1.56 0.20 0.05 −1.55 0.45

–H 90 3110 65 1.60 0.22 0.002 −1.65 0.20
–D 0 2221 90 0.364 0.30 0.10 −0.857 0.98
–D 90 2221 90 0.364 0.17 0.10 −1.00 0.98
1.3.2.6 Phthalic and Terephthalic Acid Crystals

Phthalic (PAC) and terephthalic (TAC) acid crystals have been studied by Rekik et al.
[36].
They have studied two interacting cyclic dimers shown in Figure 1.19 in which the
qi and Qi are respectively the fast and slow modes position coordinates while V ∘ D1
and V ∘ are the Davydov coupling involved in each system of the superdimer. They
D2
have considered the full Hamiltonian of the superdimer for its diagonal part as the
Table 1.13 Values of Fermi coupling parameters used for fitting experimental H/D-2-NA
spectra.
( ) ( ) ( ) ( )
Pol ∘ f 1 (cm−1 ) f 2 (cm−1 ) f 3 (cm−1 ) 𝚫1 (cm−1 ) 𝚫2 (cm−1 ) 𝚫3 (cm−1 ) 𝜸𝚫1 𝚫1 𝜸𝚫2 𝚫2 𝜸𝚫3 𝚫3
–H 0 10 10 20 60 −80 42 0.4 0.4 10

–H 90 20 10 10 110 120 150 0.2 0.2 0.2
–D 0 20 30 30 140 110 −180 0.2 0.2 0.2
–D 90 20 30 30 140 110 −180 0.2 0.2 0.2
300 K 77 K
1 1
Intensity
Intensity
0 0
2200 2400 2600 2800 3000 3200 3400 2400 2600 2800 3000 3200 3400
-H Wavenumbers (cm–1) Wavenumbers (cm–1)
1
1
Intensity
Intensity
0 0
1900 2000 2100 2200 2300 2400 2500 2000 2100 2200 2300
–1
-D Wavenumbers (cm ) Wavenumbers (cm–1)
Figure 1.18 Comparison between the experimental (grayed) and theoretical (dashed line)
spectra for Aspirin-H (polycrystalline acetylsalicylic acid) at 300 and 77 K. Source: Ghalla
et al. 2010 [27]/ With permission of Elsevier.
Table 1.14 Parameters used for fitting experimental Aspirin-H and Aspirin-D acid dimer
spectra.
( ) ( )
Compound T(K) 𝝎∘ cm−1 𝛀 cm−1 𝜶∘ V∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) 𝜼
Aspirin-H 300 2910 76 0.95 −1.75 0.25 0.1 0.85

Aspirin-H 77 3095 77 1.62 −1.70 0.35 0.1 0.8
Aspirin-D 300 2185 86 0.743 −1.69 0.25 0.7 0.7
Aspirin-D 77 2228 77 1.238 −1.65 0.25 1.1 0.8
q1 q3
Q1 Q3
O H O O H O
V°D V°D
R C 1 C C 2 C R
O H O O H O
Q2 Q4
q2 q4
Figure 1.19 Structure of the superdimer and definition of the eight vibrational modes
involved in the superdimer dynamics.
77 K 298 K
1.0 1.0
–D –H –D –H
Intensity
Intensity
0.0 0.0
2000 2200 2400 2600 2800 3000 3200 3400 1800 2000 2200 2400 2600 2800 3000 3200 3400
–1
TAC
1.0 1.0
–D –H –D –H
Intensity
Intensity
0.0 0.0
1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600
PAC Wavenumbers (cm–1) Wavenumbers (cm–1)
Figure 1.20 Crytalline phthalic (PAC) and terephthalic (TAC) acids at 77 and 298 K.
Comparison of experimental (grayed) and theoretical (full line) lineshape. Source: Based on
Rekik et al. 2020 [36].
sum of the diagonal parts of each component and as for its off parts as the sum of
the Davydov couplings of each component. Their results are shown in Figure 1.20 for
the H6 and D6 isotopomers of Phthalic (PAC) and terephthalic (TAC) acid crystals
using the parameters given in Table 1.15.
1.3.2.7 Liquid Formic Acid Mixing of Monomer and Dimer

A full quantum-theoretical approach has been used by Fathi et al. [28] to study
the mOAH experimental IR lineshapes of liquid formic acid. For this purpose, the
Table 1.15 Theoretical parameters used for the fitting of the experimental lineshapes of
PAC and TAC.
𝝎∘ D1 𝜶 ∘ D2 𝛀D1 𝛀 D2
Cases T(K) (cm −1 ) (cm −1 ) (cm −1 ) (cm −1 ) 𝜶 𝜿 𝜸 ∘ (𝝎∘∘ D1 ) V ∘ D (𝛀D1 ) V ∘ D2 (𝛀D2 )
1
D6 PAC 77 2180 2193 69 62 0.85 0.95 0.32 0.84 0.8

D6 PAC 298 2150 2167 64 60 0.96 0.87 0.31 1.29 1.35
D6 TAC 77 2130 2142 70 66 1.15 1.04 0.27 1.76 1.8
D6 TAC 298 2130 2119 62 65 1.02 1.13 0.25 1.83 1.8
H6 PAC 77 3200 3188 68 61 1.26 1.36 0.37 0.68 0.6
H6 PAC 298 3050 3012 64 71 1.45 1.39 0.34 1.02 1.06
H6 TAC 77 2860 2895 66 68 1.49 1.57 0.24 1.06 1.10
H6 PTAC 298 2850 2887 63 65 1.52 1.48 0.28 1.46 1.41
Table 1.16 Parameters used in the theoretical fitting of formic acid species.
Species 𝝎∘ (cm−1 ) 𝛀(cm−1 ) 𝜶∘ V ∘D 𝜸 ∘ (𝛀) 𝜸(𝛀) 𝜼 r
HCOOH 3030 95 1.4 −1.2 0.5 0.1 0.0 0.25

HCOOD 2340 110 0.85 −1.0 0.5 0.1 0.3 0.12
DCOOH 3080 95 1.5 −1.2 0.7 0.03 0.1 0.15
DCOOD 2240 110 0.62 −0.75 0.3 0.02 0.08 0.25
authors use our original theory, based on the strong anharmonic coupling between
the high-frequency mode and the H-bond bridge, and including the Davydov cou-
pling between the excited states of the two moieties, multiple Fermi resonances
between the mOAH (Bu) mode and combinations of some bending modes, together
with the quantum direct and indirect dampings. They have studied the influence of
the proportion of dimers species with respect to monomers to obtain the best fitting
with experimental spectra. This model reproduces satisfactorily the main features of
the experimental lineshapes of liquid hydrogenated and deuterated formic acid, by
using a minimum set of independent parameters as it may be seen in Figure 1.21.
In Tables 1.16 and 1.17 are given the parameters used in the calculation. r is the
ratio Dimer/Monomer. The other parameters are those used in our original theory.
Parameters used to reproduce the experimental spectra of HCOOH and its deuter-
ated derivatives DCOOH, HCOOD, and DCOOD of Figure 1.21.
1.3.2.8 Crystalline Furoic Acid Dimer with Slow Mode Morse Potential
and Fermi Resonances
Ghalla et al. [29], have compared the experimental IR lineshapes of polarized
crystalline Furoic acid dimers [30] at 77 K with their theoretical ones calculated
HCOOD HCOOH
0.8
Intensity
0.4
0.0
2000 2200 2400 2600 2800 3000 3200 3400
DCOOD
DCOOH
0.8
Intensity
0.4
0.0
2000 2200 2400 2600 2800 3000 3200 3400
Figure 1.21 Liquid formic acid mixing of monomer and dimer of several H/D species.
Experiment: grayed: Dashed red: theory. The corresponding parameters are given in
Tables 1.16 and 1.17. Source: Modified from Fathi et al. 2017 [28].
Table 1.17 Fermi resonances parameters for crystalline glutaric acid (cf. Figure 1.14.
f1 f2 f3 f4 𝝎1 𝝎2 𝝎3 𝝎4 𝜸𝜹 1 𝜸 𝜹 1 𝜸 𝜹 3 𝜸 𝜹 4
Species (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (cm−1 ) (𝛀) (𝛀) (𝛀) (𝛀)
HCO2 H 30 85 95 – 3130 2660 2850 – 0.1 0.05 0.02 –

HCO2 D 60 75 75 50 2285 2670 2110 2495 0.01 0.01 0.01 0.01
DCO2 H – 75 135 – – 2710 2850 – – 0.01 0.01 –
DCO2 D 50 30 70 35 2060 2350 2680 2140 0.02 0.2 0.02 0.0
by introducing Morse potential for the slow modes. Their results are given in
Figure 1.22.
They have improved the agreement with experiment by introducing in the model
3 Fermi resonances. In both situations, the theoretical lineshapes appear as contin-
uous lines whereas the experimental ones are grayed.
In Figure 1.23 is given the comparison between the experiment (grayed) and the-
ory as computed by Eq. (1.25). The slow modes are described by a Morse potential.
The parameters used in the computations are given in Tables 1.18 and 1.19.
2-furoic (O–D) 2-furoic (O–H) 2-furoic (O–D) 2-furoic (O–H)
0.8 0.8
Intensity
Intensity
0.4 0.4
0.0 0.0
2000 2500 3000 2000 2500 3000
–1
Wavenumber (cm ) Wavenumber (cm–1)
Figure 1.22 Lineshapes of polarized 2-furoic acid when the Fermi resonances are ignored,
at 77 K. Source: Ghalla et al. [29]/John Wiley & Sons.
Table 1.18 Parameters involved for fitting the experimental spectra of 2-furoic acid.
( ) ( )
Species Pol(∘ ) 𝝎∘ cm−1 𝛀 cm−1 𝜶∘ V ∘ (ℏ𝛀) 𝜸 ∘ (𝛀) 𝜸(𝛀)
H 0 3030 80 1.45 1.10 0.2 0.1

D 0 2142 90 0.331 0.786 0.15 0.1
H 90 2995 85 1.36 1.13 0.15 0.1
D 90 2144 87 0.318 0.857 0.18 0.1
Pol = 0° Pol = 90°
2-furoic (O–D) 2-furoic (O–H) 2-furoic (O–D) 2-furoic (O–H)
0.8 0.8
Intensity
Intensity
0.4 0.4
0.0 0.0
2000 2500 3000 2000 2500 3000
Wavenumber (cm–1) Wavenumber (cm–1)
Figure 1.23 Lineshapes of polarized 2-furoic acid when there are three Fermi resonances
at 77 K. Source: Ghalla et al. [29]/John Wiley & Sons.
Table 1.19 Fermi coupling parameters (in cm−1 ) used for fitting experimental 2-furoic acid
spectra.
( ) ( ) ( ) ( )
Pol ∘ f 1 (cm−1 ) f 2 (cm−1 ) f 3 (cm−1 ) 𝚫1 (cm−1 ) 𝚫2 (cm−1 ) 𝚫3 (cm−1 ) 𝜸𝚫1 𝚫1 𝜸𝚫1 𝚫2 𝜸𝚫3 𝚫3
H0 95 95 80 30 20 10 0.2 0.2 0.2

D0 93 105 116 10 10 10 0.2 0.2 0.2
H 90 95 95 100 20 20 10 0.2 0.2 0.2
D 90 83.1 105 106 10 10 10 0.2 0.2 0.2
α-form 4
3a 3
4 5
3 2
3a
O 2.834 Å
5
2
O 2.979 Å 6
7a N1 H
6 7
7a N 1 H
7
H N
H N O
2.856 Å
2.797 Å O
β-form
Figure 1.24 The two forms of oxindol.
Table 1.20 Parameters used for fitting the experimental line shapes of the 𝜈S (N–H)
stretching band of (𝛼∕𝛽)-hydrogenated and (𝛼∕𝛽)-deuterated oxindole complexes at
T = 77 K and T = 293 K.
𝝎 ∘ 𝝎 ∘ 𝛀
HH∕DD
( 1 =−1𝛀)2
VDav
( 1 −1 ) ( 2 −1 ) ( −1 )
Compound T(K) cm cm 𝜶1 𝜶2 cm 𝜸 ∘ (𝛀) cm
𝛼-D oxindole 77 2350 2460 0.65 0.726 65 0.37 85.8

𝛽-D oxindole 77 2352 2293 0.50 0.518 67 0.40 93.4
𝛼-D oxindole 293 2350 2475 0.65 0.712 65 0.35 65
𝛽-D oxindole 293 2280 2367 0.70 0.714 65 0.36 91
𝛼- H oxindole 77 3160 3335 1.0 1.230 75 0.39 167.5
𝛽-H oxindole 77 3070 3159 1.0 1.068 70 0.38 179
𝛼-H oxindole 293 3110 3296 1.0 1.160 80 0.40 144
𝛽-H oxindole 293 3120 3230 1.0 1.040 70 0.40 176
1.3.2.9 Other Kinds of H-Bonded Compounds

Combined Crystalline Oxindole Acid Dimers Another application was done by Ghalla
et al. on oxindole crystals (2,3-dihydro-1H-indol-2-one) which form cyclic dimers
with two different H-bonds [34]. In these molecules, we find the group H-N-C=O
which is involved in the formation of cyclic dimer of H-bonds (𝛼-form; 𝛽-form).
They studies 2 forms of oxindol crystal, the 𝛼- and the 𝛽 forms. The oxindole acid
molecules form a cyclic, non-centrosymmetric acid dimer with two different H-bond
bridges because they have different bond lengths as it is shown in Figure 1.24.
They have introduced for the coupling of the fast mode angular frequency with
the slow mode coordinate of the hydogenated and deuterated compounds, a subtil
dependence on the slow mode coordinate The consequence is that there are two dif-
ferent coupling parameters. Using the parameters of Table 1.20 the following spectra
are given in Figure 1.25.
1.3.2.10 Phosphinic Acid Dimer

Hydrogen-bonded dimers of phosphinic acid (See Figure 1.26) and their deuterated
analogs [(R2 POOH(D), with R = CH2 Cl, CH3 ], IR lineshapes of phosphinic acids
77 K 293 K
1.0
1.0
Intensity
Intensity
0.0
0.0
2600 2800 3000 3200 3400
2600 2800 3000 3200 3400
α-H-OX
1.0
1.0
Intensity
Intensity
0.0 0.0
2600 2800 3000 3200 3400 2600 2800 3000 3200 3400
Wavenumbers (cm–1) Wavenumbers (cm–1)
β-H-OX
1.0 1.0
0.8 0.8
Intensity
Intensity
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
2000 2100 2200 2300 2400 2500 2600 2000 2100 2200 2300 2400 2500 2600
–1
α-D-OX
1.0 1.0
0.8 0.8
Intensity
Intensity
0.6 0.6
0.4 0.4
0.2 0.2
0.0 0.0
2000 2100 2200 2300 2400 2500 2600 2000 2100 2200 2300 2400 2500 2600
β-D-OX Wavenumbers (cm ) –1
Figure 1.25 Lineshapes of (𝛼∕𝛽 )-hydrogenated and (𝛼∕𝛽)-deuterated oxindole complexes

at T = 77 K and T = 293 K. Grayed: experimental spectra. Source: Rekik et al. 2020 [36] /
With permission of Elsevier.
Figure 1.26 Dimer of phosphinic O H O

acid.
R R
P P
O H O R
R
Table 1.21a Parameters used for fitting the experimental lineshapes of (CH2 Cl)2 PO2 H∕D
and (CH3 )2 PO2 H∕D.
T(K) 𝝎∘ (cm−1 ) 𝛀(cm−1 ) De (cm−1 ) 𝜶 𝜸 ∘ (𝛀) 𝜸(𝛀) VD (𝛀) 𝜼
(CH2 Cl)2 PO2 H 435 2300 205 2100 0.80 0.40 0.1 1.68 0.9
(CH2 Cl)2 PO2 D 475 1860 202 2100 0.25 0.35 0.1 0.55 0.49
(CH3 )2 PO2 H 530 2415 206 2100 0.95 0.55 0.1 1.9 0.6
(CH3 )2 PO2 D 515 1880 204 2100 0.30 0.65 0.1 0.78 0.29
R2 PO2 H dimers in the gas phase have been studied by Rekik and Alshammari [32]
and compared with experiment [35] (See Figure 1.27)
The theoretical model is based on a model for a centrosymmetric hydrogen-bonded
dimer that treats the high-frequency OH stretches harmonically and the
low-frequency intermonomer (i.e. O· · ·O) stretches anharmonically. This model
takes into account the following effects: anharmonic coupling between the OH
and O· · ·O stretching modes; Davydov coupling between the two hydrogen bonds
in the dimer; promotion of symmetry-forbidden OH stretching transitions; Fermi
resonances between the fundamental of the OH stretches and the overtones of the
in- and out-of-plane bending modes involving the OH groups; direct relaxation
of the OH stretches; and indirect relaxation of the OH stretches via the O· · ·O
stretches. Using a set of physically significative parameters into this model, the
authors reproduce the main features in the experimental OH(D) bands of these
dimers. By increasing the number and strength of the Fermi resonances and by
promoting symmetry-forbidden OH stretching transitions in our simulations, they
directly see the emergence of the ABC structure, which is a characteristic feature in
the spectra of very strongly hydrogen-bonded dimers. However, in the case of the
deuterated dimers, which do not exhibit the ABC structure, the Fermi resonances
are found to be much weaker.
The parameters corresponding to Figure 1.27 are given in Tables 1.21a and b.
1.3.2.11 Monomer of (CH3 )2 O · · · HCl

Taking Into Account Coupling Between Slow and Bending Modes In a recent paper,
→
−
Rekik et al. [33] have calculated the IR spectral density of the 𝜈S (Cl − H) band in
gaseous (CH3 )2 O· · ·HCl complex in order to fit the experimental spectra obtained
by Lassegues and Huong [37]. (See Figure 1.28) They have used a Morse curve for
the potential of the slow mode and introduced an additional bending mode effect
R=CH3 R=CH2Cl
Theo
1.0 1.0 Exp
Intensity
Intensity
0.0 0.0
1000 2000 3000 4000 1000 2000 3000 4000

–H
1.0 1.0
Intensity
Intensity
0.0 0.0
1000 1600 2000 2400 2800
1400 1800 2200 2600
Wavenumber (cm–1)
Wavenumber (cm–1)
–D
Figure 1.27 Comparison between experimental (grayed) and theoretical lineshapes of

dimeric CH3 − and CH2 Cl− phosphinic acids H/D analogs. Source: Rekik et al. 2012 [33]/
American Chemical Society.
Table 1.21b Fermi resonance parameters for used for fitting the experimental lineshapes
of (CH2 Cl)2 PO2 H∕D and (CH3 )2 PO2 H∕D (Continuation of Table 1.21a).
𝚫1 (cm−1 ) 𝚫2 (cm−1 ) f1 (cm−1 ) f2 (cm−1 ) 𝜸1𝜹 (𝛀) 𝜸2𝜹 (𝛀)
(CH2 Cl)2 PO2 H 380 320 96 120 0.02 0.02

(CH2 Cl)2 PO2 D 225 115 30 10 0.02 0.02
(CH3 )2 PO2 H −255 −220 110 120 0.02 0.02
(CH3 )2 PO2 D 5 25 15 20 0.02 0.02
for improving the fitting by the theoretical lineshape. This procedure writes the total
Hamiltonian of the complex as the sum:the parameter 𝛼 ∘ and 𝛽 are the coupling of
the fast mode with respectively the slow and bending modes. They have compared
their results with the experimental lineshape (grayed) with the parameters given in
Table 1.22.
Taking Into Account Electrical Anharmonicity On the same compound, Rekik et al. [16,
17] take into account the electrical anharmonicity and dampings in explaining the IR
spectrum of gaseous (CH3)2 O · · · HCl complex (see Figure 1.29 and corresponding
data in Table 1.23).
Exp
Theo
IR Intensity (Arb. units)
2300 2400 2500 2600 2700

–1
Wavenumbers (cm )
Figure 1.28 (CH3 )2 O· · ·HCl complex in gas phase at 226 K . Source: Rekik et al. 2019
[34]/With permission of Elsevier.
Table 1.22 Parameters used in the theoretical lineshape of Figure 1.28.
𝝎∘ (cm−1 ) 𝛀(cm−1 ) 𝜶∘ 𝜷 𝛀𝜹 (cm−1 ) 𝜸 ∘ (𝛀) 𝜸𝜹 (𝛀)
2600 66 0.7985 0.3275 30 0.125 0.275
I(ω) I(ω)
1.0 1.0
0.0 0.0
2300 2800
2300
ω (cm )
–1 2800 ω (cm–1)
Without electrical anharmonicity (ζ = 0) Without electrical anharmonicity (ζ = –0:15)
Figure 1.29 Gaseous (CH3 )2 O· · ·HCl complex. Effect of the 𝜁 electrical anharmonicity
parameter. Source: Rekik et al. 2017 [17]/ Royal Society of Chemistry.
Table 1.23 Parameters used for fitting the lineshapes of gaseous (CH3)2 O· · ·HCl complex.
𝝎∘ (cm−1 ) 𝛀(cm−1 ) 𝜶∘ 𝜸 ∘ (𝛀) 𝜸(𝛀) 𝜻
CH3 O··· HCl 2579 106 0.721 0.30 0.20 −0.15
They demonstrated the ability of a simple anharmonic model of the dipole

moment function of the X–H stretching band to explain a set of spectroscopic
features of hydrogen bonding formation.
1.4 Conclusion
In this chapter, before exposing the experimental tests of our theory of IR spectra
of cyclic H-bonded dimers, we have given the main theoretical elements to allow its
use, from the basic idea of an anharmonic coupling between a mode of low frequency
and a mode of high frequency, through the introduction of various effects such as the
Davydov effect, Fermi resonances or electrical anharmonicity.
The good results obtained for the adjustment of the experimental spectra by those
who have systematically used this theory can find in this work a good recognition of
their efforts towards a good understanding of the behavior of hydrogen bonds.
1.5 Acknowledgment
The authors are grateful to Ms. Joëlle Sulian for the preparation of several figures in
this manuscript.
References
1 Blaise, P., Wójcik, M.J., and Henri-Rousseau, O. (2005). J. Chem. Phys. 122:
064306.
2 Maréchal, Y. and Witkowski, A. (1968). J. Chem. Phys. 48: 3637.
3 Boulil, B., Henri-Rousseau, O., and Blaise, P. (1988). Chem. Phys. 126: 263.
4 Rösch, N. and Ratner, M. (1974). J. Chem. Phys. 61: 3344.
5 Henri-Rousseau, O. and Blaise, P. (1998). The infrared spectral density of weak
hydrogen bonds within the linear response theory. Adv. Chem. Phys., vol. 103
(eds. I. Prigogine and S.A. Rice), 1–186. New York: Wiley.
6 Blaise, P., Déjardin, P.-M., and Henri-Rousseau, O. (2005). Chem. Phys. 313: 177.
7 Sakun, V. (1985). Chem. Phys. 99: 457.
8 Abramczyk, H. (1985). Chem. Phys. 94: 91.
9 Robertson, G. and Yarwood, J. (1978). Chem. Phys. 32: 267.
10 Bratos, S. and Hadzi, D. (1957). J. Chem. Phys. 27: 991.
References 37
11 Chamma, D. and Henri-Rousseau, O. (1999). Chem. Phys. 248: 91–104.

12 Witkowski, A. and Wojcik, M. (1973). Chem. Phys. 1: 9–16.
13 Davydov, A. (1962). Theory of Molecular Excitons. New York: McGraw Hill.
14 Rekik, N., Al-Agel, F.-A., and Flakus, H.-T. (2016). Chem. Phys. Lett. 647:
107–113.
15 Rekik, N. and Wójcik, M.J. (2010). Chem. Phys. 369: 71–81.
16 Rekik, N., Suleiman, J., Blaise, P. et al. (2017). Phys. Chem. Chem. Phys. 19:
5917–5931.
17 Rekik, N. and Alshammari, M.F. (2017). Chem. Phys. Lett. 678: 222–232.
18 Henri-Rousseau, O. and Blaise, P. (2008). Adv. Chem. Phys. 139 (5): 245–496.
19 Blaise, P., El-Amine Benmalti, M., and Henri-Rousseau, O. (2006). J. Chem. Phys.
124: 024514.
20 Haurie, M. and Novak, A. (1965). J. Chim. Phys. 62: 146.
21 Benmalti, M.E.-A., Blaise, P., Flakus, H.T., and Henri-Rousseau, O. (2006). Chem.
Phys. 320: 267–274.
22 Bournay, J. and Maréchal, Y. (1971). J. Chem. Phys. 55: 1230.
23 Auvert, G. and Maréchal, Y. (1979). Chem. Phys. 40: 51; ibid, p. 61.
24 Flakus, H.T. and Miros, A. (1999). J. Mol. Struct. 484: 103.
25 Flakus, H.T. and Chelmecki, M. (2002). Spectrochim. Acta, Part A 58: 179.
26 Issaoui, N., Ghalla, H., and Oujia, B. (2013). J. Appl. Spectrosc. 80 (1): 14–24.
27 Ghalla, H., Rekik, N., Michta, A., Oujia, B., and Flakus, H.T. (2010). Spectrochim.
Acta, Part A 75-1: 37–47.
28 Fathi, S., Blaise, P., Ceausu-Velcescu, A., and Nasr, S. (2017). Chem. Phys. 492:
12–22.
29 Ghalla, H., Issaoui, N., and Oujia, B. (2012). Int. J. Quantum Chem. 112:
1373–1383.
30 Flakus, H.T., Jabonska, M., and Kusz, J. (2009). Vib. Spectrosc. 49: 174.
31 Rekik, N., Flakus, H.T., Jarczyk-Jedryka, A. et al. (2015). J. Phys. Chem. Solids 77:
68–84.
32 Rekik, N., Ghalla, H., and Hanna, G. (2012). J. Phys. Chem. A 116: 4495–4509.
33 Rekik, N., Salman, S., Suleiman, J. et al. (2019). Chem. Phys. 519: 110–125.
34 Rekik, N., Flakus, H.T., Hachula, B. et al. (2020). Spectrochim. Acta, Part A 237:
118302.
35 Asfin, R.E., Denisov, G.S., and Tokhadze, K.G. (2002). J. Mol. Struct. 608:
161–168.
36 Rekik, N., Alsaif, N., Flakus, H.T. et al. (2020). Spectrochim. Acta, Part A 242:
118728.
37 Lassegues, J.C. and Huong, P.V. (1972). Chem. Phys. Lett. 17: 444–446.
39
Dynamic Interactions Shaping Vibrational Spectra of

Hydrogen-Bonded Systems
Marek J. Wójcik 1 , Mateusz Brela 1 , Łukasz Boda 1 , Marek Boczar 1 , and
Takahito Nakajima 2
1 Jagiellonian
University, Faculty of Chemistry, Gronostajowa 2, 30-387 Kraków, Poland
2
RIKEN, Center for Computational Science, 7-1-26, Minatojima-minami-machi, Chuo-ku, Kobe, Hyogo
650-0047, Japan
2.1 Introduction
Hydrogen bonding in the gas phase is of great importance, especially for
quantum–chemical calculations, mainly due to their major application to iso-
lated systems. For such systems, it is important to compare theory and experiment,
i.e. infrared absorption spectra, which lack perturbations caused by additional
intermolecular interactions, with solvent molecules or with their environment in a
crystal lattice. Many theoretical and experimental studies have been performed for
gaseous hydrogen-bonded systems, concerning their structure, energetics, electric
properties, cooperativity, isotope effects, etc. [1–10].
Among binary complexes involving medium-strength hydrogen bonds (HBs), the
dimethyl ether (DME) with HCl or DME with HF have been popular systems for
model studies because of their amazing structures of the IR absorption rovibrational
bands of the 𝜈 s (Cl–H/F–H stretching) modes [11–17]. These bands, characteristic
of hydrogen bond formation, spread over several hundred cm−1 and reveal complex
fine structures. Their origin has been studied extensively, both theoretically and
experimentally [18–20]. Such structures are mainly interpreted as an effect of
anharmonic-type coupling between the high-frequency Cl/F–H stretchings and
the low-frequency intermolecular Cl· · ·O/F· · ·O modes [21–29]. However, these
interpretations were not sufficient to explain details of the spectra. Asselin et al.
[18–20] performed complex experimental and theoretical study of the DME-HCl
and DME-HF complexes and obtained new results. These authors have measured
the conventional gas-phase IR spectra, as well as supersonic jet-FTIR spectra for HCl
or HF, complexed with DME. Together with experiment they have also performed
quantum–chemical calculations of the structure and vibrational frequencies for the

40 2 Dynamic Interactions Shaping Vibrational Spectra of Hydrogen-Bonded Systems
(a) (b)
Figure 2.1 Equilibrium structure of the DME-HCl complex with C s symmetry, optimized at
the MP2/6-311++G(3d,3p) level. The structure of the DME-HF complex is very similar.
Source: Boda et al. [30]/Reproduced with permission of AIP Publishing.
DME-HCl and DME-HF complexes and presented theoretical reconstitution of the

fine structure of IR absorption bands in the H–Cl/H–F stretching region.
In our studies on the DME-HCl and DME-HF complexes [30, 31] we have filled
existing deficiencies of previous works [29], i.e. we presented phenomenologi-
cal model which took into account an anharmonic-type coupling between the
high-frequency H–Cl or H–F stretching and all-present low-frequency intermolec-
ular hydrogen bond vibrations and for the first time we presented the results of
quantum–chemical anharmonic structure and vibrational frequencies calcula-
tions performed by an ab initio MP2 and DFT double-hybrid dispersion-corrected
methods. Figure 2.1 presents two views of the equilibrium geometry of the DME-HCl
hydrogen-bonded complex, optimized at the MP2/6-311++G(3d,3p) level. The
structure of the DME-HF complex is similar. Results of quantum–chemical calcula-
tions of infrared spectra for both, DME-HCl and DME-HF complexes, are presented
in Sections 2.2 and 2.3.
Hydrogen bond plays a pivot role in chemical and biochemical systems. In nature,
hydrogen bonds (HBs) are responsible for the creation and properties of many orga-
nized complex structures such as DNA [32–34] and cell membranes [35, 36]. The
properties of hydrogen bonds have been widely investigated theoretically as well as
experimentally and have been subject of several monographs [37–41]. Many theoret-
ical models have been proposed to describe the unusual features of hydrogen bond
stretching bands in the IR spectra [42–54].
Due to their high biological importance as one of the four DNA nucleobases,
uracil and its derivatives 1-methyluracil and 1-methyl-4-thiouracil have been
extensively studied by various physicochemical spectroscopic techniques for more
than 50 years [55–73]. Additionally, these molecules were also extensively studied
in the condensed phases [74–77]. Interactions in crystals are in the spotlight from
decades, especially hydrogen bonds, whose directions and strengths are important
2.1 Introduction 41
in crystal phases. Biological and physicochemical properties of studied compounds

are strongly related to character of hydrogen bonds present in studied structures.
They also participate in π-stacking interactions [78–80]. It occurs that bonding
potential characteristics can be relatively easily influenced by various substituents
introduced into aromatic rings, as in the 1-methyluracil, or when oxygen atoms
are replaced by heavier analogs, i.e. sulfur, as in the 1-methyl-4-thiouracil. In
the present study, we focus on the changes in the strengths of hydrogen bond
interactions caused by such replacements.
Thymine and adenine as well as guanine and cytosine are nitrogen-containing
nucleobases. DNA can undergo structural changes, sometimes leading to muta-
tions, due to a variety of environmental factors which can either be chemical or
physical in nature [81–83]. It should be stressed that some chemical factors include
interactions with small molecules or enzymes leading to uncoiling of DNA or
breaking of hydrogen bonds [84, 85]. From this point of view, the HBs are the most
important interactions in nature. The HBs build the structure of DNA by means of
a lock-and-key mechanism [86–89]. Only complementary nucleobases may build
the DNA structure.
Calculations and interpretation of spectroscopic properties of hydrogen-bonded
systems are not obvious. Large anharmonicity, proton delocalization, and coupling
of fast and slow modes are present in hydrogen-bonded systems. During last few
decades, researchers improved considerably methodologies of theoretical studies.
Treatments of anharmonicity, proton delocalization, and coupling of fast and slow
modes, present in hydrogen-bonded systems, were shown in detail in Refs. [49–54].
The molecular dynamics (MD), presented in Section 2.4, allows us to describe impor-
tant intra- and intermolecular interactions. Conditions in the crystal approximate
in first approximation conditions present in the living cell. Therefore, MD simula-
tions of HBs in the crystal field and the analysis of the vibrational spectra became
recently very popular [90–100]. In Sections 2.5–2.7, we present some results of the
Born–Oppenheimer Molecular Dynamics spectroscopic studies of crystals of nucleic
bases.
One of the best examples of hydrogen bond networks are crystals of water that exist
in many phases. Recently, we performed Car–Parrinello molecular dynamics calcu-
lations for the analysis of infrared spectra of two forms of ice – ice Ih and ferroelectric
ice XI [90, 98]. Protons of ice under high-pressure become readily ordered upon
cooling forming ice XI [101]. Ice XI may exist in the outer solar system. Long-range
electrostatic forces, caused by the ferroelectricity, might be an important factor for
planet formation [102, 103]. The existence of ice XI on Pluto and Charon [104] and
the formation of ice XI in space have been predicted [102]; however, from infrared
observations, it is not clear whether ice XI exists or not. Our MD studies have shown
that librational region of the infrared spectra of ice Ih and ice XI exhibits especially
large differences in the simulated spectra of these two forms of ice. The results of
these studies can be used for the analysis of IR telescopes experimental data and the
differentiation of the ice forms in the Universe. They are presented in Section 2.8.
2.2 Theoretical Model of the Infrared Spectra

of Gaseous (CH3 )2 O-HCl and (CH3 )2 O-HF Complexes
Our theory assumes that the band structure of the infrared spectra of the Cl–H
stretching mode (𝜈 s ) in DME-HCl or DME-HF complex is composed of a progression
of the O· · ·Cl or O· · ·F stretching vibration (𝜈 σ ), in which each subband has its own
substructure due to the progressions of the two O· · ·Cl or O· · ·F bending vibrations
(𝜈 β1 and 𝜈 β2 , perpendicular to each other). The 𝜈 s band is composed from the set of
(l ) (m ) (n ) (l ) (m ) (n )
transitions |𝜈s(v=0) , 𝜈σ 0 , 𝜈β1 0 , 𝜈β20 ⟩ → |𝜈s(v=1) , 𝜈σ 1 , 𝜈β1 1 , 𝜈β21 ⟩, where v, l, m, and n are
the vibrational quantum numbers for 𝜈 s , 𝜈 σ , 𝜈 β1 , and 𝜈 β2 modes, respectively, and
the subscripts “0” and “1” denote the values of l, m, and n for the ground (v = 0)
and excited (v = 1) state of 𝜈 s , respectively.
Let q denote coordinates of the high-frequency Cl–H/F–H stretching vibration,
Q1 a coordinate of the low-frequency O· · ·Cl/O· · ·F stretching vibration, Q2 and Q3
coordinates of the two low-frequency O· · ·H—Cl/O· · ·H—F hydrogen bond bending
vibrations, and by 𝜔, Ω1 , Ω2 , Ω3 the corresponding frequencies (in cm−1 ), respec-
tively. Only couplings between these vibrations will be considered.
The vibrational Hamiltonian H of the considered system can be written as
̂ =T
H ̂ (q) + T
̂ (Q1 ) + T
̂ (Q2 ) + T
̂ (Q3 ) + V1 (q, Q1 ) + V2 (q, Q3 ) + V3 (q, Q3 ) (2.1)
where T and V i denote kinetic and potential energies, respectively.

Within the adiabatic approximation for the vibrational states, and assuming that
no couplings exist between the low-frequency modes 𝜈 σ , 𝜈 β1 , and 𝜈 β2 , the total vibra-
tional wave function Ψ has the form
Ψ(q, Q1 , Q2 , Q3 ) = 𝜒1 (Q1 )𝜒2 (Q2 )𝜒3 (Q3 )𝜑(q; Q1 , Q2 , Q3 ) (2.2)
where the wave function 𝜙 of the high-frequency 𝜈 s stretching vibration is defined

by the following eigenvalue equation:
̂ (q) + V1 (q, Q1 ) + V2 (q, Q3 ) + V3 (q, Q3 )]𝜑(q; Q1 , Q2 , Q3 )
[T
= [𝜀1 (Q1 ) + 𝜀1 (Q1 ) + 𝜀1 (Q1 )]𝜑(q; Q1 , Q2 , Q3 ) (2.3)
The eigenvalues of Eq. (2.3) can be written as the sum of three energies 𝜀1 , 𝜀2 , and
𝜀3 forming the potentials for the three low-frequency vibrations.
Physical meaning is as follows: the energy of the high-frequency 𝜈 s stretching
vibrations determines, on average, the potential governing the low-frequency inter-
molecular hydrogen bond vibrations (in terms of crude–adiabatic approximation).
The wave functions 𝜒 i of the low-frequency vibrations are solutions of the
equations:
̂ (Qi ) + 𝜀i (Qi )]𝜒i (Qi ) = Ei 𝜒i (Qi )
[T (2.4)
Within the crude–adiabatic approximation ([T ̂ (Qi ), 𝜑] = 0), the total vibrational
energy E is a simple sum of the energies being the solutions of Eq. (2.4)
E = E 1 + E2 + E3 (2.5)
2.2 Theoretical Model of the Infrared Spectra of Gaseous (CH3 )2 O-HCl and (CH3 )2 O-HF Complexes 43
The potentials for the low-frequency vibrations depend on the state of high-
frequency 𝜈 s stretching vibration (described by vibrational quantum number v).
It is usually assumed that the equilibrium value of the low-frequency vibration
coordinate Q(v) e,i
changes when going from the ground state (v = 0) to the excited
state (v = 1) of 𝜈 s vibration. Both the dissociation energy D0 and the bonding energy
De (depth of the potential) can be obtained from ab initio calculation. Therefore,
the potential 𝜀(Q1 ) for the low-frequency 𝜈 σ hydrogen bond stretching mode can be
described by the Morse curve
{ [ ( )]}2
𝜀(v)
1 (Q1 ) = D e 1 − exp −𝛼 Q 1 − Q(v)
e,1 (2.6a)
Due to the lack of the data about the potentials for the hydrogen bond bending vibra-
tions, we will use the harmonic form of their potentials
( )2
1
𝜀(v)
2
(Q 2 ) = k Q2 − Q(v)
e,2
(2.6b)
2
( )2
1
𝜀(v)
3
(Q3 ) = k Q3 − Q(v) e,3
(2.6c)
2
The vertical excitation energies in the excited Cl–H stretching state (for v = 1) have
been omitted in Eqs. (2.6a)–(2.6c).
With these assumptions, the energy of the vibrational states will be given by the
formula:
( ) ( ) ( ) ( )
1 1 2 1 1
G(v) (l, m, n) = Ω1 l + − xΩ1 l + + Ω2 m + + Ω3 n + (2.7)
2 2 2 2
The dipole moment of the complex 𝛍 is a function of vibrational coordinates and
can be expanded into a Taylor series:
( ) ( ) ( ) ( )
𝜕𝛍 𝜕𝛍 𝜕𝛍 𝜕𝛍
𝛍 = 𝛍0 + q+ Q + Q + Q +··· (2.8)
𝜕q 0 𝜕Q1 0 1 𝜕Q2 0 2 𝜕Q3 0 3
Assuming that the dipole moment is independent of the low-frequency intermolecu-
lar motions, and neglecting higher nonlinear terms (electrical anharmonicity), only
the first two terms retain in the expansion (2.8). With these assumptions, the Cl–H
stretching (𝜈 s ) IR absorption band will be composed of Franck–Condon-type pro-
gressions in the hydrogen bond low-frequency modes (𝜈 σ , 𝜈 β1 , and 𝜈 β2 ).
The intensity of the individual transition is expressed by the following formula
including Boltzmann factor modulating the intensity of hot transitions due to ther-
mal activation of higher vibrational states of 𝜈 σ , 𝜈 β1 , and 𝜈 β2 modes in the ground
vibrational state (v = 0) of 𝜈 s mode:
I(v=0,l0 ,m0 ,n0 )→(v=1,l1 ,m1 ,n1 ) ∼|⟨𝜑(v=0) |𝛍|𝜑(v=1) ⟩|2 (FCMl0 ,l1 )2 (FCHm0 ,m1 )2 (FCHn0 ,n1 )2
[ ]
hc
× exp − (l0 Ω1 + m0 Ω2 + n0 Ω3 ) (2.9)
kT
where FCMx0 ,x1 and FCHx0 ,x1 are the Franck–Condon integrals between the
vibrational states characterized by quantum numbers x0 and x1 and belonging to
two different states (v = 0 and v = 1) of 𝜈 s , for the Morse and harmonic oscillator,
respectively. The first term in Eq. (2.9), being the square of the IR transition moment
between the ground and the first excited vibrational state of 𝜈 s , can be omitted,
as it is constant (within crude–adiabatic approximation) for all transitions of type
(l ) (m ) (n ) (l ) (m ) (n )
|𝜈s(v=0) , 𝜈σ 0 , 𝜈β1 0 , 𝜈β20 ⟩ → |𝜈s(v=1) , 𝜈σ 1 , 𝜈β1 1 , 𝜈β21 ⟩.
For the case of Morse oscillator, the Franck–Condon integral is given by the
formula [105]
√
x0 !x1 !(j − x0 )(j − x1 ) j−x j−x
FCMx0 ,x1 =2 y 0y 1
Γ(2j − x0 + 1)Γ(2j − x1 + 1) 1 2
( )( ) ( )2j−x0 +k1 −x1 +k2
∑ ∑ (−1)k1 +k2 2j − x
x0 x1
2j − x1 k1 k2 2
0
× y1 y2
k =0k =0
k1 !k2 ! x0 − k1 x1 − k2 y1 + y2
1 2
× Γ(2j − x0 + k1 − x1 + k2 ) (2.10a)
where
( )
1
y1 = (2j + 1) exp 𝛼b (2.10b)
2
( )
1
y2 = (2j + 1) exp − 𝛼b (2.10c)
2
√
Γ is the Euler gamma function and j = 2𝜇De ∕𝛼ℏ − 1∕2 (𝜇 is the reduced mass of
the vibration). The value of ⌊j⌋, being the largest integer not exceeding j, gives the
number of bound vibrational states within the Morse potential.
For the harmonic oscillator, the formula is [106]
( 2) ( )2k+x1 −x0
b √ ∑x0
(−1)k b
FCHx0 ,x1 = exp − x0 !x1 !• • √ x0 ≤ x1
4 (x − k)!k!(x1 − x0 + k)!
k=0 0 2
(2.11)
The quantity b in Eqs. (2.10b), (2.10c), and (2.11) is so-called linear distortion param-
eter defined as
( )
b = − Q(1)
e − Qe
(0)
(2.12)
and it is related to the change in the value of the equilibrium coordinate Qe between
the ground and excited states of the high-frequency 𝜈 s vibration. The value of param-
eter b decreases when the Cl–H (F–H) proton is replaced by deuterium, by the ratio of
square roots of reduced masses for the Cl–H/Cl–D (F–H/F–D) stretching vibrations.
In our interpretation of the 𝜈 s infrared band shape for the DME-HCl and DME-HF
complexes, we also consider rotational transitions. Strictly, the structure of the com-
plex is asymmetric rotor; however, in first approximation, it can be treated as a pro-
late symmetric top. The values of the rotational constants B and C for the complex are
close together. In this study, we also limit to the rigid rotor approximation. Then the
rotational energy levels can be calculated using the following formula for a prolate
symmetric top rigid rotor:
[ ]
B+C 2
F(J, K) = BJ(J + 1) + A − K (2.13)
2
where: A, B, C are the rotational constants of the complex, and J, K are the rotational
quantum numbers (J = 1, 2, …; K = 0, ±1, ±2, …, ±J; J ≥ K).
2.3 Simulation of the Cl–H(D) and F–H Stretching Bands in the DME-H(D)Cl and DME-HF Complexes 45
The intensity of a pure rotational line for a J → J + 1 transition (since ΔK = 0) for

a symmetric top is given by [107]:
[ ( ) ]
hc B+C
S(J,K)→(J+1,K) =B(J + 1) exp − BJ(J + 1) + A − K2
kT 2
{ [ ]}
hc (J + 1)2 − K 2
× 1 − exp − 2B(J + 1) (2 − 𝛿0,K ) (2.14)
kT J+1
Joining Eq. (2.7) for the pure vibrational energy with Eq. (2.13) for the pure rota-
tional energy, we obtain the formula for the energies of rovibrational states:
T (v) (l, m, n, J, K) = G(v) (l, m, n) + F (v) (J, K) (2.15)
The energies of transitions (i.e. peak positions in absorption spectrum) between the
rovibrational states of 𝜈 s ground state (v = 0) and the first excited state (v = 1) of 𝜈 s
can be calculated from the formula
ΔEij(0)→(1) =r + Ti(1) (l1 , m1 , n1 , J1 , K1 ; A01 , B01 , C01 )−Tj(0) (l0 , m0 , n0 , J0 , K0 ; A00 , B00 , C00 )
(2.16)
where r is the vertical excitation energy; A00 , B00 , C00 are the rotational constants of
the complex in its totally ground vibrational state; and A01 , B01 , C01 are the rotational
constants of the complex in the excited state (v = 1) of 𝜈 s .
The transition intensities can be calculated from the formula
Iij(1)←(0) ∼(FCMl0 ,l1 )2 (FCHm0 ,m1 )2 (FCHn0 ,n1 )2 S(J0 ,K0 )→(J1 ,K1 )
[ ]
hc
× exp − (l0 Ω1 + m0 Ω2 + n0 Ω3 ) (2.17)
kT
taking into account the rotational selection rules for vibration–rotation transitions:
}
ΔK = 0, ΔJ = ±1, for K0 = K1 = 0
for parallel bands
ΔK = 0, ΔJ = 0, ±1, for K ≠ 0
ΔK = ±1, ΔJ = 0, ±1 for perpendicular bands
As a consequence, each vibrational band, within assumed model, will reveal rota-
tional structure consisting of three branches, P, Q, and R.
2.3 Simulation of the Cl–H(D) and F–H Stretching

Bands in the DME-H(D)Cl and DME-HF Complexes
To apply the model presented in Section 2.2, several parameters are needed to
be known. There are rotational constants (A, B, and C) for the vibrationally
averaged structures in the totally vibrational ground state (00), and in the excited
vibrational state of 𝜈 s (01), with the ground state for the remaining vibrational
modes, dissociation energy of the complex, and frequencies of the modes 𝜈 s , 𝜈 σ ,
𝜈 β1 , and 𝜈 β2 . Part of them can be taken from experiment (when available) and the
others from anharmonic second-order ab initio vibrational analysis. The relevant
values of parameters for both DME-HCl and DME-HF complexes, as well as for
the free DME, HCl, and HF molecules, were presented and discussed in detail in
Refs. [30, 31].
The series of calculations involving Eqs. (2.16) and (2.17) together with selection
rules were performed to calculate the energies and intensities of transitions between
the ground and excited states of the Cl–H/F–H stretching bands. Simulations of the
Cl–H stretching bands were performed using the rotational constants, and some
vibrational frequencies were obtained from the MP2/6-311++G(3d,3p) calculations
[30]. We used the experimental values 119 and 120 cm−1 of the frequency Ω1 for
the 𝜈 σ mode in the DME-HCl complex [19], whereas the frequencies Ω2 and Ω3
of the 𝜈 β1 and 𝜈 β2 modes were taken from ab initio anharmonic (PT2) calculations
due to lack of experimental data. In case of DME-HF complex, the experimental
frequency Ω1 for the 𝜈 σ was taken as equal to 185 cm−1 [16] and the remaining
parameters were taken from the results of ab initio MP2/6-311++G(2df,2pd)
calculations [31]. The parameter 𝛼 of the Morse potential (restitution constant in
harmonic approximation) and the anharmonicity constant xe have been derived
from the experimental value of Ω1 frequency, and from the bonding energy, De
and reduced mass of 𝜈 σ vibration were calculated by the MP2 method for both
complexes.
Theoretical absorption bands simulated in terms of proposed model are presented
in Figure 2.2 for the Cl–H stretching mode in the DME-HCl complex at 273 K (a)
and for the Cl–D stretching mode in the DME-DCl complex at 273 K (b) and com-
pared with the respective experimental spectra (taken from Ref. [19]). Theoretical
spectra are presented as solid lines obtained from calculated frequencies and inten-
sities of rovibrational transitions, without convoluting spectral line shape curves.
Due to the large number of calculated rovibrational transitions, the correspond-
ing points formed almost continuous lines. The reproduction of the experimental
bands is satisfactory. The calculated frequency and energy distributions agree well
with the experimental bands recorded by Asselin et al. [19]. The experimental band
for the deuterated complex (DME-DCl) is shifted toward lower wavenumbers and
it has smaller halfwidth and simpler fine structure, compared to the hydrogenated
isotopomer (DME-HCl). Our model calculations correctly fit the experimental line
shapes and fine structures, and the isotope effect is correctly reproduced.
Figure 2.3 presents the comparison of experimental F–H stretching absorption
band for the DME-HF complex at 188 K (taken from Ref. [20]) and the calculated
spectrum. The calculated spectrum presents only vibrational transitions (without
rotational structure) extracted from the model calculation to assign the most inten-
sive components. There are a large number of vibrational transitions; however, the
intensities reduce the number to a few meaningful, as presented in Figure 2.3.
Discrepancies between theory and experiment are related mainly to the assump-
tions of the model. The two low-frequency bending modes (𝜈 β1 and 𝜈 β2 ) were
assumed to be harmonic and the complex was assumed to be a rigid rotor of a
prolate top type. The model does not take into account electrical anharmonicity,
which reflects mainly in the intensity distribution.
2.4 Methodology of Molecular Dynamics 47
Experimental
Calculated
Absorbance (a.u.)
2200 2300 2400 2500 2600 2700 2800 2900

(a) Wavenumber (cm−1)
Experimental
Calculated
Absorbance (a.u.)
1600 1700 1800 1900 2000 2100

(b) −1)
Wavenumber (cm
Figure 2.2 Theoretical absorption bands (solid lines) for the Cl–H stretching mode in the
DME-HCl complex at 273 K (a) and for the Cl–D stretching mode in the DME-DCl complex
at 273 K (b) in comparison with the respective experimental spectra (dotted lines). Source:
Boda et al. [30]/Reproduced with permission of AIP Publishing.
2.4 Methodology of Molecular Dynamics
Molecular dynamics using potential based on independent electron structure calcu-

lations is a powerful theoretical tool for characterizing various systems [108, 109].
A broad yet accurate description of molecular dynamics has been presented in sev-
eral monographs [99, 110–112] and review papers [113–115]. The main assump-
tion of adiabatic molecular dynamics is to describe the motion of nuclei along the
potential energy surface (PES). The Born–Oppenheimer approximation gives the
|0 0 0 0〉 → |1 0 0 0〉
|0 0 1 0〉 → |1 0 1 0〉
|0 0 0 1〉 → |1 0 0 1〉
|0 1 0 0〉 → |1 1 0 0〉 |0 0 1 1〉 → |1 0 1 1〉
|0 1 0 1〉 → |1 1 0 1〉
|0 1 1 0〉 → |1 1 1 0〉
|0 2 0 0〉 → |1 2 0 0〉
|0 2 0 1〉 → |1 2 0 1〉
|0 0 0 1〉 → |1 0 1 1〉
|0 2 1 0〉 → |1 2 1 0〉 |0 0 1 0〉 → |1 0 2 0〉
|0 0 1 1〉 → |1 0 2 1〉
|0 0 1 0〉 → |1 0 0 0〉 |0 0 0 1〉 → |1 0 0 2〉
|0 0 2 0〉 → |1 0 1 0〉 |0 0 1 1〉 → |1 0 1 2〉
|0 1 0 0〉 → |1 1 1 0〉
|0 0 0 2〉 → |1 0 0 1〉 |0 0 1 1〉 → |1 0 2 1〉
|0 1 0 1〉 → |1 1 0 2〉
|0 0 0 3〉 → |1 0 0 2〉 |0 1 0 2〉 → |1 1 0 3〉
|0 0 0 0〉 → |1 1 0 0〉
3350 3400 3450 3500 3550 3600 3650

Wavenumber (cm−1)
Figure 2.3 Purely vibrational transitions with labeling of the most intensive ones were
calculated during the simulation and presented as Dirac 𝛿 peaks for DME-HF complex at
188 K in comparison with the experimental spectrum at the same temperature (solid line,
taken from Ref. [20]). Source: Boda et al. [31]/Reproduced with permission of Elsevier.
possibility to treat the motion of nuclei and electrons independently. The molecu-
lar dynamics method is based on the integration of the classical equation of motion
[108, 109]. This approach provides information about the electron structure using
quantum-chemistry methods in exactly the same way as in the static calculations.
PES calculated by ab initio methods is essential for systems in which there is a pos-
sibility of breaking and forming chemical bonds, e.g. systems with strong hydro-
gen bonds [116, 117]. It should be emphasized that the dynamics of proton motion
in hydrogen bonds is determined by a complex nature of interactions occurring in
the studied system. These interactions are responsible for the complex structure of
infrared and Raman spectra of hydrogen-bonded systems [51, 116–119]. It should
be noted that due to the quantum nature of the proton motion, its wave function
(wave packet) must be taken into account during MD simulations [120–123]. This
approach is very expensive and cannot be considered as a “standard” solution to this
problem.
We would like to point out the time cost of the MD calculation, it is approximately
equal to the cost of the (static) ab initio calculation multiplied by the number of tra-
jectory steps. Long trajectories are advantageous for spectroscopic studies, especially
in the low-frequency range, which is characteristic of slow motions. The time step
is an important parameter for molecular dynamics calculations. A decreasing time
step in the numerically calculated trajectory leads to more accurate results. Never-
theless, the computational cost increases with the number of steps. In practice, the
limit of the maximum time steps is determined by the period of the vibration with
the highest energy.
2.5 Spectroscopic Study of Uracil, 1-Methyluracil, and 1-Methyl-4-thiouracil 49
2.5 Spectroscopic Study of Uracil, 1-Methyluracil,

and 1-Methyl-4-thiouracil
In Ref. [124] we explored the nature of intermolecular hydrogen bonds in the

uracil derivatives in the crystal phase. In the crystal of uracil, 1-methyluracil,
and 1-methyl-4-thiouracil there are several hydrogen bonds stretching from
medium-strong to weak. However, substitution of methyl group changes com-
pletely the hydrogen bond network. In this study, anharmonic frequencies were
obtained by Fourier transform of the atom position functions obtained from the
Born–Oppenheimer trajectories.
The authors also calculated the spectra associated with the C–H and N–H
stretching modes (see Figure 2.4) that are very sensitive to the strength of hydrogen
bonding. This approach gave the possibility to discuss in detail the strengths of
hydrogen bonds in the considered crystals. The orbital interaction contributions
of hydrogen bonds in terms of the deformation densities elucidate the covalent
character of hydrogen bond in the studied systems. Results show the differences
H(1) H(3)
0.25 H(3) 0.25 H(5)
H(5) H(6)
H(6)
H3 H5
0.20 H3 H5 0.20
0.15 0.15
I (a.u.)
I (a.u.)
H6
H6
H1
0.10 0.10
0.05 0.05
0.00 0.00
2500 2750 3000 3250 3500 2500 2750 3000 3250 3500
(a) ν (cm−1) (b) ν (cm−1)
H(3)
0.25 H(5)
H(6)
H3 H5
0.20
0.15
I (a.u.)
H6
0.10
0.05
0.00
2500 2750 3000 3250 3500
(c) ν (cm−1)
Figure 2.4 Uracil (panel a), 1-methyluracil (panel b), and 1-methyl-4-thiouracil (panel c)
power spectra of the atoms in the N—H and C—H bonds obtained from the BOMD
trajectories. The blue and black colors correspond to the N—H bond: H(1) and H(3); black,
red, and green colors correspond to H(5) and H(6). Source: Brela et al. [124]/Reproduced
with permission of Elsevier.
between hydrogen bond networks in uracil and methylated derivatives. The

studied methylated derivatives 1-methyluracil as well as 1-methyl-4-thiouracil form
dimeric structures in the crystal phase, while uracil does not. The replacement
of oxygen atom by sulfur atom causes weakness of the hydrogen bonds that form
dimers.
2.6 Hydrogen Bond Interaction Dynamics in the

Adenine and Thymine Crystals
In Ref. [125] we compared hydrogen bond networks in the adenine and thymine
crystals. The used methodology was evaluated by comparing calculated IR spectra
with the experimental ones. The BOMD molecular dynamics simulations have been
performed for supercells of the studied crystals. The spectroscopic features of calcu-
lated bands are in the good agreement with experimental data. Based on this test, the
discussion of the intermolecular interactions has been done. The calculated widths
of the bands in the high-frequency regions have been slightly underestimated, which
has been discussed several times before [50, 72, 124].
In the next step, the decomposition of the power spectra in terms of domains
has been done for chemical groups and atoms. The results show the big differ-
ence between adenine and thymine. The thymine spectra have fewer bands in
the high-frequency region than adenine crystals. The analysis of positions and
intensities of the stretching bands show the difference in strengths of hydrogen
bonds. The characteristics of the adenine spectra reflect stronger hydrogen bond
network as compared to the thymine crystal [125]. To characterize hydrogen
bond networks the analysis of intermolecular interactions has been done. The
cluster of 27 molecules has been chosen for this analysis. The time courses of
the interaction energies show that the intermolecular interaction is more sta-
bilized in the adenine crystal than in the thymine crystal. The analysis of the
deformation densities, shown in Figure 2.5, as well as the electronic descriptors
along trajectories show that in the adenine crystal intermolecular interactions
have three directions and fluctuate, while in the adenine crystal have only two
directions and are stable. These results explain also the difference in the melting
temperature difference in studied crystals. Adenine, which has stronger hydrogen
bonds, has also higher melting temperature than thymine. The comparison of
the adenine and thymine exposes the electrostatic contribution impact on the
interactions present in the studied crystals. The BOMD simulations give us pos-
sibility to discuss the stability of the systems, as well as the influence of thermal
modes. Moreover, reactivity descriptors, based on electronic energy or electron
density, such as frontiers orbitals or Fukui functions may be easily determined.
Their fluctuations bring us information about influence of thermal modes on
the electronic structure. This methodology is still pioneering and has some
weaknesses, e.g. analyzed structures obtained from trajectories are not stationary
points.
2.7 Guanine and Cytosine Crystals 51
(a) (b)
(c) (d)
Figure 2.5 Analysis of the deformation densities, Δ𝜌 for the cluster of 27 molecules of
thymine (a, c) and adenine (b, d). The blue color shows the accumulation of the electron
density, the red color the depletion. Panels show the structures after 85 ps of the
simulation. Source: Brela et al. [125]/Reproduced with permission of Elsevier.
2.7 Guanine and Cytosine Crystals
The HBs networks in the guanine and cytosine crystals have been studied in Ref.
[126]. The used methodology was tested by comparing calculated IR spectra with
the experimental ones. The calculations have been performed for supercells of the
crystals. Positions of the calculated bands are in good agreement with the experi-
mental data, which gave us the possibility to discuss and compare properties of the
studied crystals. However, the calculated widths of the bands in the high-frequency
region have been underestimated, which has been discussed in Refs. [52, 54]
(Figure 2.6).
As in Ref. [125], we performed a similar analysis of the decomposition of the power
spectra in terms of component domains: chemical groups and atoms. The results
show the big difference between cytosine and guanine, especially in the NH group,
which forms HBs in both crystals. The analysis of positions and intensities of peaks
shows the difference in strengths of HBs. A redshift in the spectrum of the cytosine
crystal shows a weaker HBs network in this case than in the guanine crystal.
The analysis of the hydrogen bond networks shows the differences between
guanine and cytosine crystals. The time courses of the interaction energies show
that intermolecular interactions are more stabilized (on average) in the guanine
crystal than in the cytosine crystal. The analysis of the deformation densities
39 ps 61 ps
(a) (b)
(c) ΔEorb = –598.87 kcal mol−1 (d) ΔEorb = –431.14 kcal mol−1
Figure 2.6 Analysis of the deformation densities for the cluster of 27 molecules of
cytosine. The blue color shows the accumulation of the electron density and the red color
shows the depletion. Panel (a) shows the structure after 39 ps of simulation and panel (b)
after 61 ps. Panels (c, d) present increase in the deformation density around the considered
molecule in both snapshots, respectively. Source: Brela et al. [126]/Reproduced with
permission of American Chemical Society.
along trajectories shows that in both crystals the intermolecular interactions are
directional and stable. The authors conclude that differences in the melting point
originate from the hydrogen bond character [126].
2.8 Spectroscopic Signature for Ferroelectric Ice

The five unit cells with eight water molecules chosen for the calculations are pre-
sented in Figure 2.7.
Figure 2.8 presents simulation of infrared spectra of ice Ih at 60 K and ice XI at 4
and 40 K, using Car–Parrinello method of molecular dynamics.
To assign peaks to particular sets of water molecules, the power spectra of
atoms belonging to different atomic groups were calculated. The three prominent
bands are present in the IR spectra of ice XI at two temperatures 4 and 40 K: at
600–1100 cm−1 formed by the librational mode, at 1500–1750 cm−1 formed by the
HOH bending mode, and at 2700–3300 formed by the OH stretching mode 𝜈 s . In
addition, at 40 K there appear few weak bands in the range 1800–2400 cm−1 . They
are the overtone of the librational mode and the combination of the bending and
librational modes [127]. Increase in temperature does not significantly affect band
broadening. At lower temperature, the position of the 𝜈 s band is shifted to lower
frequency. Such shifts occur in systems with strong hydrogen bonds [128]. This
2.8 Spectroscopic Signature for Ferroelectric Ice 53
A B
C D
Figure 2.7 The five unit cells (A–E) with eight water molecules chosen for the calculations
for ice Ih and ice XI. The A structure corresponds to ice XI. Source: Wójcik et al.
[98]/Reproduced with permission of Elsevier.
result is also consistent with experimental temperature studies of infrared spectra

of ice XI [129].
Simulated spectra of ice Ih and ice XI show many differences. Especially sig-
nificant is the region of the librational motions of water molecules, presented in
Figure 2.9, because in this region there is a large difference in the spectra of ice Ih
α(ν~)
(a)
α(ν~)
(b)
α(ν~)
(c)
0 250 500 750 1000 1250 1500 1750 2000 2250 2500 2750 3000 3250 3500
Wavenumber ν~ (cm−1)
Figure 2.8 Comparison between theoretical IR spectrum of ice Ih calculated at 60 K (a)

and ice XI calculated at 4 K (b) and 40 K (c). Source: Wójcik et al. [98]/Reproduced with
permission of Elsevier.
and ice XI. Librational spectrum of ice Ih has more simple structure with only two
bands – one very broad at about 790–1080 cm−1 and the second with small intensity
at about 620 cm−1 . In the case of ice XI, the librational spectrum is different – it
is composed of two major bands, located at 945 and 1015 cm−1 at temperature
4 K, and at 930 and 1000 cm−1 at temperature 40 K. These bands can be treated as
spectroscopic signatures of ice XI.
Our results are in agreement with experimentally measured infrared spectra of
KOH-doped ice [128]. Infrared and Raman spectra of ice Ih have been previously
calculated using different water–water interaction potentials by Rice et al. [127, 128],
2.9 Conclusions 55
Ib 60 K
XI 40 K
XI 4 K
α(ν~)
600 700 800 900 1000 1100

Wavenumber ν~ (cm–1)
Figure 2.9 Comparison between librational spectra of ice Ih calculated at 60 K and ice XI
calculated at 4 and 40 K. Source: Wójcik et al. [98]/Reproduced with permission of Elsevier.
Wójcik et al. [129], and Skinner et al. [130, 131]. Our CPMD simulated spectra
of ice Ih are in agreement with these calculations. Recently, computational and
experimental studies of hydrogen-ordered ice have been also reported [132–141].
The authors of Refs. [135, 136] have shown that similar to the IR studies that
there are distinct differences in the Raman spectra between ices Ih and XI, with
ice XI showing much stronger peaks in the translational, librational, and in-phase
asymmetric stretch regions.
To find distant ice XI in outer space, we will need to investigate in detail the libra-
tional motions of water molecules on ice surfaces by satellites beyond the Earth’s
atmosphere. The evidence for ice XI in the doped ice and the simulated infrared
bands of ice XI hold out the hope that future telescope and planetary exploration
will find huge ferroelectric ice XI in the solar system. Simulated spectra of ice XI can
help in identifying ice XI in the outer solar system.
2.9 Conclusions
Theoretical model of vibrational couplings in hydrogen bonds in the (CH3 )2 O· · ·HCl

and (CH3 )2 O· · ·HF gas-phase complexes has been developed. Simulation of
rovibrational structure of the Cl–H and F–H stretching IR absorption bands
was performed for both complexes, and the results were compared with the
experimental spectra. The results were in good agreement with the experimental
bands. Presented model calculations correctly fitted the experimental line shapes
and fine structures at different temperatures and the isotope effect was correctly
reproduced.
Presented studies of nucleic acid bases [125, 126] are the first ones that combine
spectroscopic data with studies of the hydrogen bond network dynamics by per-
forming deformation density analyses along the BOMD trajectories. This approach
gives in the near future many possibilities to discuss the covered structures such as
metastable crystals or biological reaction intermediates.
Theoretical infrared spectra of ice Ih and ferroelectric ice XI were calculated using
Car–Parrinello molecular dynamics. The librational region exhibits especially large
differences in the simulated spectra of ice Ih and ice XI. Theoretical IR spectra of ice
XI can be an useful tool for analysis of experimental data obtained by IR telescopes
to investigate presence of ferroelectric ice in the universe.
Acknowledgment
This work was financially supported by the National Science Center, Poland, grant
2016/21/B/ST4/02102.
References
1 Da Silva, A.M., Chakrabarty, S., and Chaudhuri, P. (2015). Hydrogen-bonded

glycine–HCN complexes in gas phase: structure, energetics, electric properties
and cooperativity. Mol. Phys. 113: 447–462.
2 Rutkowski, K.S., Melikova, S.M., Linoka, O.V. et al. (2015). Infrared spec-
troscopy and ab initio study of hydrogen bonded Cl3 CD⋅N(CH3 )3 complex in
the gas phase. Spectrochim. Acta, Part A 136: 95–99.
3 Rekik, N., Ghalla, H., and Hanna, G. (2012). Explaining the structure of the
OH stretching band in the IR spectra of strongly hydrogen-bonded dimers of
phosphinic acid and their deuterated analogs in the gas phase: a computational
study. J. Phys. Chem. A 116: 4495–4509.
4 Durlak, P., Berski, S., and Latajka, Z. (2011). Car–Parrinello and path integral
molecular dynamics study of the hydrogen bond in the acetic acid dimer in the
gas phase. J. Mol. Model. 17: 2995–3004.
5 Roohi, H., Nowroozi, A.-R., and Eshghi, F. (2010). The gas phase
hydrogen-bonded dimers of HOCl: a high-level quantum chemical study.
Int. J. Quantum Chem. 110: 1489–1499.
6 Hippler, M. (2007). Quantum chemical study and infrared spectroscopy of
hydrogen-bonded CHCl3 –NH3 in the gas phase. J. Chem. Phys. 127: 084306.
7 Chung, S. and Hippler, M. (2006). Infrared spectroscopy of hydrogen-bonded
CHCl3 –SO2 in the gas. J. Chem. Phys. 124: 214316.
References 57
8 Du, D.-M., Fu, A.-P., and Zhou, Z.-Y. (2005). Theoretical study of the rotation
barrier of hydrogen peroxide in hydrogen bonded structure of HOOH–H2 O
complexes in gas and solution phase. THEOCHEM 717: 127–132.
9 Cole, G.C., Davey, J.B., Legon, A.C., and Lyndon, A.F. (2003). A hydrogen
bonded complex C2 H2 · · ·HBr isolated and characterized in the gas phase using
pulsed-jet, Fourier transform microwave spectroscopy. Mol. Phys. 101: 603–612.
10 Wugt, L.R., Hegelund, F., and Nelander, B. (2004). Studies of intermolecular
vibrations in hydrogen-bonded molecular complexes for the gas phase using a
synchrotron radiation source. Mol. Phys. 102: 1743–1747.
11 Lassègues, J.C. and Huong, P.V. (1972). Infrared spectrum of the gaseous
CIH· · ·O(CH3 )2 complex. Chem. Phys. Lett. 17: 444–446.
12 Bertie, J.E. and Falk, M.V. (1973). The infrared spectrum of the
hydrogen-bonded molecule dimethyl ether … hydrogen chloride in the gas
phase. Can. J. Chem. 51: 1713–1720.
13 Desbat, B. and Lassègues, J.C. (1979). Raman spectra of the dimethylether · · ·
hydrogen chloride hydrogen-bonded complex in the gas phase. J. Chem. Phys.
70: 1824–1829.
14 Millen, D.J. and Schrems, O. (1983). Comparative infrared study of
hydrogen-bonded heterodimers formed by HCl, DCl, HF and DF with (CH3 )2 O,
CH3 OH and (CH3 )3 COH in the gas phase. Assignment of vibrational band
structure in (CH3 )2 O· · ·HCl. Chem. Phys. Lett. 101: 320–325.
15 Arnold, J. and Millen, D.J. (1965). Hydrogen bonding in gaseous mixtures.
Part II. Infrared spectra of ether–hydrogen fluoride systems. J. Chem. Soc.
503–509. https://doi.org/10.1039/JR9650000503.
16 Thomas, R.K. (1971). Hydrogen bonding in the gas phase: the thermodynamic
properties of hydrogen fluoride-ether complexes and their far infrared spectra.
Proc. R. Soc. London, Ser. A 322: 137–146.
17 Thomas, R.K. (1971). Hydrogen bonding in the gas phase: the infrared spectra
of complexes of hydrogen fluoride with hydrogen cyanide and methyl cyanide.
Proc. R. Soc. London, Ser. A 325: 133–149.
18 Asselin, P., Dupuis, B., Perchard, J.P., and Soulard, P. (1997). The gas phase
infrared spectrum of HCl complexed with dimethyl ether revisited: assignment
of the fundamental transition from a jet-cooled experiment. Chem. Phys. Lett.
268: 265–272.
19 Asselin, P., Soulard, P., Alikhani, M.E., and Perchard, J.P. (1999). A new inter-
pretation of the IR spectrum of H(D)Cl complexed with dimethyl ether from a
supersonic jet-FTIR experiment. Chem. Phys. 249: 73–87.
20 Asselin, P., Soulard, P., Alikhani, M.E., and Perchard, J.P. (2000). Investigation
of the gas phase infrared spectrum of HF complexed with dimethyl ether from
both cell- and supersonic jet-FTIR experiments. Chem. Phys. 256: 195–205.
21 Coulson, C.A. and Robertson, G.N. (1975). A theory of the broadening of the
infrared absorption spectra of hydrogen-bonded species. I. Proc. R. Soc. London,
Ser. A 342: 167–197.
22 Coulson, C.A. and Robertson, G.N. (1975). A theory of the broadening of the
infrared absorption spectra of hydrogen-bonded species, II. The coupling of
anharmonic v(XH) and v(XH· · ·Y) modes. Proc. R. Soc. London, Ser. A 342:
289–315.
23 Robertson, G.N. (1977). A theory of a broadening of the infrared absorption
spectra of hydrogen-bonded species III. The kinematic and electronic coupling
mechanisms. Philos. Trans. R. Soc. London, Ser. A 286: 25–53.
24 Maréchal, Y. and Bouteiller, Y. (1974). Étude théorique du spectre infrarouge
du complexe formé par liaison hydrogène en phase gazeuse entre l’éther
diméthylique et le chlorure d’hydrogène: influence de la distorsion cetrifuge
sur le spectre. Proc. Acad. Sci. 279: 435–438.
25 Bouteiller, Y. and Maréchal, Y. (1976). Etude théorique du spectre infra-rouge
des complexes liés par liaison hydrogène à l’état gazeux: complexe ClH · · ·
O(CH3 )2 , son homologue deutérié et complexes voisins. Mol. Phys. 32: 277–288.
26 Bouteiller, Y. and Guissani, Y. (1979). Theoretical study of hydrogen bonded
complexes in the gas phase. Infra-red and Raman spectra of ClH· · ·O(CH3 )2
and of the deuterated species. Mol. Phys. 38: 617–624.
27 Wójcik, M.J. (1977). Temperature dependence of the infrared spectra of the
hydrogen bond in the gaseous ClH· · ·O(CH3 )2 complex. Chem. Phys. Lett. 46:
597–599.
28 Wójcik, M.J. (1986). Theoretical interpretation of infrared spectra of the Cl–H
stretching vibration in the gaseous (CH3 )2 O· · ·HCl complex. Int. J. Quantum
Chem. 29: 855–865.
29 Wójcik, M.J. (1986). Note on theoretical interpretation of infrared spectra of
Cl–H and Cl–D in gaseous (CH3 )2 O· · ·HCl and (CH3 )2 O· · ·DCl complexes. Int.
J. Quantum Chem. 30: 567–569.
30 Boda, Ł., Boczar, M., Gług, M., and Wójcik, M.J. (2015). Quantum-mechanical
study of energies, structures, and vibrational spectra of the H(D)Cl complexed
with dimethyl ether. J. Chem. Phys. 143: 204302.
31 Boda, Ł., Boczar, M., Brela, M.Z. et al. (2019). Quantum-mechanical study of
energies, structures and vibrational spectra of the HF complexed with dimethyl
ether. Chem. Phys. Lett. 731: 136590.
32 Zanuy, D. and Aleman, C. (2008). DNA-conducting polymer complexes: a com-
putational study of the hydrogen bond between building blocks. J. Phys. Chem.
B 112: 3222–3230.
33 Löwdin, P.O. (1969). Some aspects of the hydrogen bond in molecular biology.
Ann. N. Y. Acad. Sci. 158: 86–95.
34 Prabhu, V.V., Young, L., Awati, K.W. et al. (1990). Defect-mediated
hydrogen-bond melting in B-DNA polymers. Phys. Rev. B: Condens. Matter
41: 7839–7845.
35 Hermansson, M. and von Heijne, G. (2003). Inter-helical hydrogen bond forma-
tion during membrane protein integration into the ER membrane. J. Mol. Biol.
334: 803–809.
36 Arneson, L.S., Katz, J.F., Liu, M., and Sant, A.J. (2001). Hydrogen bond
integrity between MHC class II molecules and bound peptide determines
the intracellular fate of MHC class II molecules. J. Immunol. 167: 6939–6946.
References 59
37 Pimentel, G.C. and McClellan, A.L. (1960). The Hydrogen Bond. San Francisco,
CA: Freeman Co.
38 Schuster, P., Zundel, G., and Sandorfy, C. (ed.) (1976). Recent Developments in
Theory and Experiments, vol. I–III. New York: North Holland Pub. Co.
39 Hadži, D. (ed.) (1997). Theoretical Treatments of Hydrogen Bonding. Chichester:
Wiley.
40 Grabowski, S.J. (ed.) (2006). Hydrogen Bonding: New Insights. Dordrecht:
Springer.
41 Gilli, P. and Gilli, G. (2009). The Nature of the Hydrogen Bond. Oxford: Oxford
University Press.
42 Maréchal, Y. and Witkowski, A. (1968). Infrared spectra of H-bonded systems.
J. Chem. Phys. 48: 3697–3705.
43 Witkowski, A. and Wójcik, M. (1973). Infrared spectra of hydrogen bond a gen-
eral theoretical model. Chem. Phys. 1: 9–16.
44 Wójcik, M.J. (1976). Theory of the infrared spectra of the hydrogen bond in
molecular crystals. Int. J. Quantum Chem. 10: 747–760.
45 Henri-Rousseau, O. and Blaise, P. (1998). The infrared spectral density of weak
hydrogen bonds within the linear response theory. In: The Infrared Spectral
Density of Weak Hydrogen Bonds Within the Linear Response Theory, Advances
in Chemical Physics, vol. 103 (ed. I. Prigogine and S.A. Rice), 1–187. https://
doi.org/10.1002/9780470141625.ch1.
46 Yaremko, A.M., Ratajczak, H., Baran, J. et al. (2004). Theory of profiles
of hydrogen bond stretching vibrations: Fermi–Davydov resonances in
hydrogen-bonded crystals. Chem. Phys. 306: 57–70.
47 Blaise, P., Wójcik, M.J., and Henri-Rousseau, O. (2005). Theoretical interpre-
tation of the line shape of the gaseous acetic acid cyclic dimer. J. Chem. Phys.
122: 064306.
48 Sandorfy, C. (2006). Hydrogen bonding: how much anharmonicity? J. Mol.
Struct. 790: 50–54.
49 Brela, M.Z., Stare, J., Pirc, G. et al. (2012). Car–Parrinello simulation of the
vibrational spectrum of a medium strong hydrogen bond by two-dimensional
quantization of the nuclear motion: application to 2-hydroxy-5-nitrobenzamide.
J. Phys. Chem. B 116: 4510–4518.
50 Boczar, M., Kwiendacz, J., and Wójcik, M.J. (2008). Theoretical and spectro-
scopic study of infrared spectra of hydrogen-bonded 1-methyluracil crystal and
its deuterated derivative. J. Chem. Phys. 128: 164506.
51 Wójcik, M.J., Kwiendacz, J., Boczar, M. et al. (2010). Theoretical and spec-
troscopic study of hydrogen bond vibrations in imidazole and its deuterated
derivative. Chem. Phys. 372: 72–81.
52 Gług, M., Brela, M.Z., Boczar, M. et al. (2017). Infrared spectroscopy and
Born–Oppenheimer molecular dynamics simulation study on deuterium substi-
tution in the crystalline benzoic acid. J. Phys. Chem. B 121: 479–489.
53 Brela, M.Z., Boczar, M., Wójcik, M.J. et al. (2017). The Born–Oppenheimer
molecular simulations of infrared spectra of crystalline
poly-(R)-3-hydroxybutyrate with analysis of weak CH· · ·OC hydrogen bonds.

Chem. Phys. Lett. 678: 112–118.
54 Brela, M.Z., Wójcik, M.J., Boczar, M. et al. (2015). Car–Parrinello molecular
dynamics simulations of infrared spectra of crystalline vitamin C with analysis
of double minimum proton potentials for medium-strong hydrogen bonds.
J. Phys. Chem. B 119: 7922–7930.
55 Angell, C.L. (1961). An infrared spectroscopic investigation of nucleic acid con-
stituents. J. Chem. Soc. 504–515.
56 Susi, H. and Ard, J.S. (1971). Vibrational spectra of nucleic acid constituents-I:
planar vibrations of uracil. Spectrochim. Acta, Part A 27: 1549–1562.
57 Lord, R.C. and Thomas, G.J. Jr., (1967). Raman spectral studies of nucleic acids
and related molecules-I ribonucleic acid derivatives. Spectrochim. Acta, Part A
23: 2551–2591.
58 Bardi, G., Bencivenni, L., Ferro, D. et al. (1980). Thermodynamic study of the
vaporization of uracil. Thermochim. Acta 40: 275–282.
59 Nishimura, Y., Tsuboi, M., Kato, S., and Morokuma, K. (1981). In-plane vibra-
tional modes in the uracil molecule from an ab initio MO calculation. J. Am.
Chem. Soc. 103: 1354–1358.
60 Szcze˛śniak, M., Nowak, M.J., Rostkowska, H. et al. (1983). Matrix isolation
studies of nucleic acid constituents. 1. Infrared spectra of uracil monomers.
J. Am. Chem. Soc. 105: 5969–5976.
61 Bandekar, J. and Zundel, G. (1983). The role of C=O transition dipole–dipole
coupling interaction in uracil. Spectrochim. Acta, Part A 39: 337–341.
62 Bandekar, J. and Zundel, G. (1983). Normal coordinate analysis treatment on
uracil in solid state. Spectrochim. Acta, Part A 39: 343–355.
63 Chin, S., Scott, I., Szczepaniak, K., and Person, W.B. (1984). Matrix isola-
tion studies of nucleic acid constituents. 2. Quantitative ab initio prediction
of the infrared spectrum of in-plane modes of uracil. J. Am. Chem. Soc. 106:
3415–3422.
64 Barnes, A.J., Stuckey, M.A., and Le Gall, L. (1984). Nucleic acid bases stud-
ied by matrix isolation vibrational spectroscopy: uracil and deuterated uracils.
Spectrochim. Acta, Part A 40: 419–431.
65 Maltese, M., Passerini, S., Nunziante-Cesaro, S. et al. (1984). Infrared levels of
monomeric uracil in cryogenic matrices. J. Mol. Struct. 116: 49–65.
66 Brown, R.D., Godfrey, P.D., McNaughton, D., and Pierlot, A.P. (1988).
Microwave spectrum of uracil. J. Am. Chem. Soc. 110: 2329–2330.
67 Wójcik, M.J., Rostkowska, H., Szczepaniak, K., and Person, W.B. (1989). Vibra-
tional resonances in infrared spectra of uracils. Spectrochim. Acta, Part A 45:
499–502.
68 Nowak, M.J. (1989). IR matrix isolation studies of nucleic acid constituents: the
spectrum of monomeric thymine. J. Mol. Struct. 193: 35–49.
69 Graindourze, M., Smets, J., Zeegers-Huyskens, T., and Maes, G. (1990). Fourier
transform-infrared spectroscopic study of uracil derivatives and their hydrogen
bonded complexes with proton donors: Part I. Monomer infrared absorptions
References 61
of uracil and some methylated uracils in argon matrices. J. Mol. Struct. 222:
345–364.
70 Wójcik, M.J. (1990). Medium-frequency Raman spectra of crystalline uracil,
thymine and their 1-methyl derivatives. J. Mol. Struct. 219: 305–310.
71 Graindourze, M., Grootaers, T., Smets, J. et al. (1991). FT-IR spectroscopic
study of uracil derivatives and their hydrogen bonded complexes with proton
donors: Part III. Hydrogen bonding of uracils with H2 O in argon matrices.
J. Mol. Struct. 243: 37–60.
72 Leś, A., Adamowicz, L., Nowak, M.J., and Lapinski, L. (1992). The infrared
spectra of matrix isolated uracil and thymine: an assignment based on new
theoretical calculations. Spectrochim. Acta, Part A 48: 1385–1395.
73 Kyogoku, Y., Higouchi, S., and Tsuboi, M. (1967). Intra-red absorption spec-
tra of the single crystals of 1-methyl-thymine, 9-methyladenine and their 1:1
complex. Spectrochim. Acta, Part A 23: 969–983.
74 Jarzembska, K.N., Kubsik, M., Kaminski, R. et al. (2012). From a single
molecule to molecular crystal architectures: structural and energetic studies
of selected uracil derivatives. Cryst. Growth Des. 12: 2508–2524.
75 Stewart, R.F. and Jensen, L.H. (1967). Redetermination of the crystal structure
of uracil. Acta Crystallogr. 23: 1102–1105.
76 Hawkinson, S.W. (1975). 1-Methyl-4-thiouracil. Acta Crystallogr., Sect. B: Struct.
Sci. 31: 2153–2156.
77 McMullan, R.K. and Craven, B.M. (1989). Crystal structure of 1-methyluracil
from neutron diffraction at 15, 60 and 123 K. Acta Crystallogr., Sect. B: Struct.
Sci. 45: 270–276.
78 Whittleton, S.R., Hunter, K.C., and Wetmore, S.D. (2004). Effects of hydrogen
bonding on the acidity of uracil derivatives. J. Phys. Chem. A 108: 7709–7718.
79 Peral, F. and Troitino, D.I. (2010). Hydrogen-bonded dimers in self-association
of 5-substituted uracil derivatives and hetero-association with L-cysteine. A
density functional theory study. THEOCHEM 944: 1–11.
80 Sponer, J., Leszczynski, J., and Hobza, P. (1996). Hydrogen bonding and stack-
ing of DNA bases: a review of quantum-chemical ab initio studies. J. Biomol.
Struct. Dyn. 14: 117–135.
81 Watson, J.D. and Crick, F.H.C. (1953). A structure for deoxyribose nucleic acid.
Nature 171: 737–738.
82 Ghosh, A. and Bansal, M. (2003). A glossary of DNA structures from A to Z.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 59: 620–626.
83 Shabarova, Z.A. and Bogdanov, A.A. (2008). Advanced Organic Chemistry of
Nucleic Acids, 588. Wiley-VCH.
84 Klinman, J.P., Knapp, M.J., and Rickert, K. (2002). Temperature-dependent
isotope effects in soybean lipoxygenase-1: correlating hydrogen tunneling with
protein dynamics. J. Am. Chem. Soc. 124: 3865–3874.
85 Olsson, M.H.M., Mavri, J., and Warshel, A. (2006). Transition state theory can
be used in studies of enzyme catalysis: lessons from simulations of tunnelling
and dynamical effects in pipoxygenase and other systems. Philos. Trans. R. Soc.
London, Ser. B 361: 1417–1432.
86 Warshel, A., Mavri, J., Liu, H.B., and Olsson, M.H.M. (2008). Simulation of
tunneling in enzyme catalysis by combining a biased propagation approach
and the quantum classical path method: application to lipoxygenase. J. Phys.
Chem. B 112: 5950–5954.
87 Sarkar, S. and Singh, P.C. (2019). Mechanistic aspects of fungicide-induced
DNA damage: spectroscopic and molecular dynamics simulation studies.
J. Phys. Chem. B 123: 8653–8661.
88 Teo, R.D., Smithwick, E.R., Migliore, A., and Beratan, D.N. (2019). A single
AT–GC exchange can modulate charge transfer-induced p53–DNA dissociation.
Chem. Commun. 55: 206–209.
89 Sirover, M. and Loeb, L. (1974). Erroneous base-pairing induced by a chemical
carcinogen during DNA synthesis. Nature 252: 414–416.
90 Gług, M., Boczar, M., Boda, Ł., and Wójcik, M.J. (2015). Analysis of librational
modes of ice XI studied by Car–Parrinello molecular dynamics. Chem. Phys.
459: 102–111.
91 Pirc, G., Stare, J., and Mavri, J. (2010). Car–Parrinello simulation of hydrogen
bond dynamics in sodium hydrogen bisulfate. J. Chem. Phys. 132: 224506.
92 Kwiendacz, J., Boczar, M., and Wójcik, M.J. (2011). Car–Parrinello molecular
dynamics simulations of infrared spectra of crystalline imidazole. Chem. Phys.
Lett. 501: 623–627.
93 Jezierska, A., Panek, J.J., Koll, A., and Mavri, J. (2007). Car–Parrinello sim-
ulation of an O–H stretching envelope and potential of mean force of an
intramolecular hydrogen bonded system: application to a Mannich base in
solid state and in vacuum. J. Chem. Phys. 126: 205101.
94 Boczar, M., Kurczab, R., and Wójcik, M.J. (2010). Theoretical and spectroscopic
studies of vibrational spectra of hydrogen bonds in molecular crystal of β-oxalic
acid. Vib. Spectrosc. 52: 39–47.
95 Boczar, M., Boda, Ł., and Wójcik, M.J. (2006). Theoretical model of infrared
spectra of hydrogen bonds in molecular crystals and its application to interpre-
tation of infrared spectra of 1-methylthymine. J. Chem. Phys. 125: 084709.
96 Manin, A.N., Voronin, A.P., Shishkina, A.V. et al. (2015). Influence of sec-
ondary interactions on the structure, sublimation thermodynamics, and
solubility of salicylate: 4-hydroxybenzamide cocrystals. Combined experimental
and theoretical study. J. Phys. Chem. B 119: 10466–10477.
97 Zhurov, V.V. and Pinkerton, A.A. (2015). Inter- and intramolecular interactions
in crystalline 2-nitrobenzoic acid – an experimental and theoretical QTAIM
analysis. J. Phys. Chem. A 119: 13092–13100.
98 Wójcik, M.J., Gług, M., Boczar, M., and Boda, Ł. (2014). Spectroscopic signa-
ture for ferroelectric ice. Chem. Phys. Lett. 612: 162–166.
99 Stare, J., Panek, J.J., Eckert, J. et al. (2008). Proton dynamics in the strong
chelate hydrogen bond of crystalline picolinic acid N-oxide. A new compu-
tational approach and infrared, Raman and INS study. J. Phys. Chem. A 112:
1576–1586.
References 63
100 Dopieralski, P.D., Latajka, Z., and Olovsson, I. (2010). Proton transfer dynam-
ics in crystalline maleic acid from molecular dynamics calculations. J. Chem.
Theory Comput. 6: 1455–1461.
101 Kuhs, W.F., Finney, J.L., Vettier, C., and Bliss, D.V. (1984). Structure and
hydrogen ordering in ices VI, VII, and VIII by neutron powder diffraction.
J. Chem. Phys. 81: 3612–3623.
102 Iedema, M.J., Dresser, M.J., Doering, D.L. et al. (1998). Ferroelectricity in water
ice. J. Phys. Chem. B 102: 9203–9214.
103 Wang, H., Bell, R.C., Iedema, M.J. et al. (2005). Sticky ice grains aid planet for-
mation: unusual properties of cryogenic water ice. Astrophys. J. 620: 1027–1032.
104 McKinnon, W.B. and Hofmeister, A.M. (2005). Ice XI on Pluto and Charon?
Bull. Am. Astron. Soc. 37: 732–733.
105 Lopez V., J.C., Rivera, A.L., Smirnov, Yu., F., and Frank, A. (2002). Simple
evaluation of Franck–Condon factors and non-Condon effects in the Morse
potential. Int. J. Quantum Chem. 88: 280–295.
106 Koide, S. (1960). Über die Berechnung von Franck-Condon-Integralen.
Z. Naturforsch., A: Phys. Sci. 15: 123–128.
107 Malkmus, W. (1965). Intensities of pure rotational band systems of symmetric
top molecules. J. Quant. Spectrosc. Radiat. Transfer 5: 621–631.
108 Marx, D. and Hutter, J. (2009). Ab Initio Molecular Dynamics. Cambridge:
Cambridge University Press.
109 Krack, M., Parrinello, M., Quickstep: make the atoms dance Publication Series
of the John von Neumann Institute for Computing (NIC) series 25 (2004) 29.
110 Dopieralski, P., Latajka, Z., and Olovsson, I. (2010). Proton-transfer dynam-
ics in the (HCO3 − )2 dimer of KHCO3 from Car–Parrinello and path-integrals
molecular dynamics calculations. Acta Crystallogr., Sect. B: Struct. Sci. 66:
222–228.
111 Howard, D.L., Kjaergaard, H.G., Huang, J., and Meuwly, M. (2015). Infrared
and near infrared spectroscopy of acetylacetone and hexafluoroacetylacetone.
J. Phys. Chem. A 119: 7980.
112 Durlak, P., Latajka, Z., and Berski, S. (2009). A Car–Parrinello and path inte-
gral molecular dynamics study of the intramolecular lithium bond in the
lithium 2-pyridyl-N-oxide acetate. J. Chem. Phys. 131: 024308–024316.
113 Patrone, P.N. and Dienstfrey, A. (2018). Uncertainty quantification for molecu-
lar dynamics. In: Reviews in Computational Chemistry, vol. 31 (ed. A.L. Parrill
and K.B. Lipkowitz), 105. New York: Wiley-VCH.
114 Marx, D. (1998). Proton transfer in ice. In: Classical and quantum dynamics
in condensed phase simulations (ed. B.J. Berne, G. Ciccotti and D.F. Coker),
359–384. Singapore: World Scientific.
115 Tuckerman, M.E. and Hughes, A. (1998). Path integral molecular dynamics:
a computational approach to quantum statistical mechanics. In: Classical and
Quantum Dynamics in Condensed Phase Simulations (ed. B.J. Berne, G. Ciccotti
and D.F. Coker), 311–358. Singapore: World Scientific.
116 Śmiechowski, M. and Stangret, J. (2006). Proton hydration in aqueous solution:
Fourier transform infrared studies of HDO spectra. J. Chem. Phys. 125: 204508.
117 Dopieralski, P., Perrin, C.L., and Latajka, Z. (2011). On the intramolecular
hydrogen bond in solution: Car–Parrinello and path integral molecular dynam-
ics perspective. J. Chem. Theory Comput. 7: 3505–3513.
118 Berens, P.H., White, S.R., and Wilson, K.R. (1981). Molecular dynamics and
spectra. I. Diatomic rotation and vibration. J. Chem. Phys. 75: 515.
119 Luber, S., Marcella, I., and Hutter, J. (2014). Raman spectra from ab initio
molecular dynamics and its application to liquid S-methyloxirane. J. Chem.
Phys. 141: 094503.
120 Drukker, K. and Hammes-Schiffer, S. (1997). An analytical derivation of
MC-SCF vibrational wave functions for the quantum dynamical simulation
of multiple proton transfer reactions: initial application to protonated water
chains. J. Chem. Phys. 107: 363–374.
121 Sumner, I. and Iyengar, S.S. (2007). Quantum wavepacket ab initio molecular
dynamics: an approach for computing dynamically averaged vibrational spectra
including critical nuclear quantum effects. J. Phys. Chem. A 111: 10313–10324.
122 Iyengar, S.S., Sumner, I., and Jakowski, J. (2008). Hydrogen tunneling in an
enzyme active site: a quantum wavepacket dynamical perspective. J. Phys.
Chem. B 112: 7601–7613.
123 Li, X. and Iyengar, S.S. (2011). Quantum wavepacket ab initio molecular
dynamics for extended systems. J. Phys. Chem. A 115: 6269–6284.
124 Brela, M.Z., Boczar, M., Malec, L.M. et al. (2018). Spectroscopic study of uracil,
1-methyluracil and 1-methyl-4-thiouracil: hydrogen bond interactions in crys-
tals and ab-initio molecular dynamics. Spectrochim. Acta, Part A 197: 194–201.
125 Brela, M.Z., Klimas, O., Boczar, M. et al. (2020). A comparison of the hydrogen
bond interaction dynamics in the adenine and thymine crystals: BOMD and
spectroscopic study. Spectrochim. Acta, Part A 237: 118398.
126 Brela, M.Z., Klimas, O., Surmiak, E. et al. (2019). A comparison of the hydro-
gen bond interactions dynamics in the guanine and cytosine crystals: ab initio
molecular dynamics and spectroscopic study. J. Phys. Chem. A 123: 10757.
127 Bergren, M.S. and Rice, S.A. (1982). An improved analysis of the OH stretching
region of the vibrational spectrum of ice Ih. J. Chem. Phys. 77: 583–602.
128 Rice, S.A., Bergren, M.S., Belch, A.C., and Nielson, G. (1983). A theoretical
analysis of the hydroxyl stretching spectra of ice Ih, liquid water, and amor-
phous solid water. J. Phys. Chem. 87: 4295–4308.
129 Wójcik, M.J., Szczeponek, K., and Ikeda, S. (2002). Theoretical study of the
OH/OD stretching regions of the vibrational spectra of ice Ih. J. Chem. Phys.
117: 9850–9857.
130 Li, F. and Skinner, J.L. (2010). Infrared and Raman line shapes for ice Ih. I.
Dilute HOD in H2 O and D2 O. J. Chem. Phys. 132: 204505.
131 Li, F. and Skinner, J.L. (2010). Infrared and Raman line shapes for ice Ih. II.
H2 O and D2 O. J. Chem. Phys. 133: 244504.
132 Bertie, J.E. and Whalley, E. (1964). Infrared spectra of ices Ih and Ic in the
range 4000 to 350 cm−1 . J. Chem. Phys. 40: 1637–1645.
133 Bertie, J.E. and Whalley, E. (1964). Infrared spectra of ices II, III, and V in the
range 4000 to 350 cm−1 . J. Chem. Phys. 40: 1646–1659.
References 65
134 Arakawa, M., Kagi, H., and Fukazawa, H. (2009). Laboratory measurements of
infrared absorption spectra of hydrogen-ordered ice: a step to the exploration of
ice XI in space. Astrophys. J. Suppl. Ser. 184: 361–365.
135 Abe, K. and Shigenari, T. (2011). Raman spectra of proton ordered phase XI of
ICE I. Translational vibrations below 350 cm−1 . J. Chem. Phys. 134: 104506.
136 Shigenari, T. and Abe, K. (2012). Vibrational modes of hydrogens in the proton
ordered phase XI of ice: Raman spectra above 400 cm−1 . J. Chem. Phys. 136:
174504.
137 Parkkinen, P., Riikonen, S., and Halonen, L. (2014). Ice XI: not that ferroelec-
tric. J. Phys. Chem. C 118: 26264–26275.
138 Pamuk, B., Allen, P.B., and Fernandez-Serra, M.-V. (2015). Electronic and
nuclear quantum effects on the ice XI/ice Ih phase transition. Phys. Rev. B:
Condens. Matter 92: 134105.
139 Liu, Y. and Ojamäe, L. (2016). Raman and IR spectra of ice Ih and ice XI with
an assessment of DFT methods. J. Phys. Chem. B 120: 11043–11051.
140 Zhang, P., Wang, Z., Lu, Y.B. et al. (2016). The normal modes of lattice vibra-
tions of ice XI. Sci. Rep. 6: 29273.
141 Shi, R. and Tanaka, H. (2019). Homogeneous nucleation of ferroelectric ice
crystal driven by spontaneous dipolar ordering in supercooled TIP5P water. J.
Chem. Phys. 151: 024501.
67
Trajectory On-the-Fly Molecular Dynamics Approach to

Tunneling Splitting in the Electronic Ground and Excited
States
Tetsuya Taketsugu 1,2 and Yusuke Ootani 3
1
Hokkaido University, Faculty of Science, Department of Chemistry, N10-W8, Kita-ku, Sapporo 060-0810,
Japan
2
Hokkaido University, Institute for Chemical Reaction Design and Discovery (WPI-ICReDD), N21-W10,
Kita-ku, Sapporo 001-0021, Japan
3 Tohoku University, Institute for Materials Research, 2-1-1 Katahira, Aoba-ku, Sendai 980-8577, Japan
3.1 Introduction
A hydrogen bond is a non-covalent attractive interaction between a hydrogen atom

bonded to an atom with high electronegativity and a lone pair of a neighboring atom,
which is a relatively strong intermolecular force. Hydrogen bonds are also formed
within molecules, and structural changes can occur due to the transfer of hydrogen
atoms through hydrogen bonds between and within molecules. Since the hydro-
gen atom is the lightest element, quantum effects such as the tunneling effect are
important. In the tunneling phenomenon, a particle uses its wave nature to pene-
trate a barrier even if it does not have enough energy to cross the activation barrier,
and the lighter the mass of the particle and the lower the height of the barrier, the
greater the effect. The Wentzel–Kramers–Brillouin (WKB) approximation is one of
the semiclassical approximate solutions to the Schrodinger equation and is used
in the discussion of the tunneling effect [1]. For molecules exhibiting symmetric
double wells, the tunneling effect manifests itself as an energy splitting of the vibra-
tional ground state, which can be measured in spectroscopic experiments. Once
the one-dimensional tunneling path is defined, the tunneling amplitude along the
tunneling path can be calculated based on the WKB approximation to theoretically
estimate the tunneling splitting.
A polyatomic molecular reaction consisting of N atoms is described in a 3N − 6
dimensional coordinate space. For elementary reaction processes, the intrinsic reac-
tion coordinate (IRC) [2, 3] is used as a one-dimensional reaction pathway. Since
the IRC is defined as the minimum energy path connecting two minima (reactants
and products) and a first-order saddle point (transition state: TS) on the potential
energy surface (PES), the IRC can be regarded as a one-dimensional tunneling path
to estimate the tunneling effect. Taketsugu et al. applied the WKB approximation to
the IRC as a tunneling pathway (called IRC-WKB) and succeeded in reproducing the
68 3 Trajectory On-the-Fly Molecular Dynamics Approach
tunneling splitting of small water clusters [4–6]. In hydrogen transfer between heavy
atoms, it is known that the IRC is highly curved near the TS. In such a case, the IRC
is no longer a reasonable tunneling path and the IRC–WKB method underestimates
the tunneling splitting value. Truhlar, Garrett, et al. discussed and proposed various
methods for estimating the tunneling effect taking into account the curvature of the
IRC [7]. Taketsugu and Hirao also proposed a tunneling path that takes into account
the curvature of the IRC in multidimensional space [8]. As an approach to tunneling
pathways beyond the IRC, Mil’nikov and Nakamura have developed a semiclas-
sical instanton method for multidimensional tunneling splitting and successfully
reproduced the tunneling splitting of malonaldehyde in conjunction with high accu-
racy ab initio electronic structure calculations [9–11]. Richardson and Althorpe pro-
posed a semiclassical instanton formula for tunneling splitting based on the ring
polymer method [12, 13]. Kawatsu and Miura also proposed an efficient compu-
tational algorithm for semiclassical instanton approach based on discretized path
integrals [14].
Makri and Miller developed a semiclassical method to evaluate tunneling splitting
using the WKB approximation within the framework of molecular dynamics [15].
The computational cost of their method is much lower than that of the quantum
mechanical approach, although the implementation of this method requires pre-
determination of the tunneling pathway. Thompson and coworker [16] applied the
Makri–Miller method to the tunneling problems of real molecules, such as tunnel-
ing splitting in intramolecular hydrogen transfer in malonaldehyde and its isotopes
[17] and tunneling level shifts in cis–trans isomerization of HSiOH molecule [18].
Taketsugu et al. also discussed the tunneling dynamics by the Makri–Miller method
using the PES with all degrees of freedom taken into account, which was generated
by interpolation using the modified Shepard interpolation method of ab initio data
for HSiOH [19] and malonaldehyde [20].
In recent years, ab initio molecular dynamics or trajectory on-the-fly molecular
dynamics (TOF-MD) has been used to investigate reaction dynamics by using
on-the-fly information on the PES obtained from first-principles electronic struc-
ture calculations [21, 22]. Although the computational cost is higher than that
of MD simulation using potential functions, it can be applied to any polyatomic
molecule in principle because it does not require potential functions. It is now
widely applied to various types of reactions to discuss reaction mechanisms,
branching ratios, and quantum yields. With the progress in the last two decades,
the TOF-MD method has been extended to the reaction dynamics of excited states
where nonadiabatic transitions play an important role [23–25]. On the other hand,
few attempts have been made yet to introduce semiclassical tunneling methods into
TOF-MD simulation.
Ben-Nun and Martinez proposed to combine the Makri–Miller method with
a TOF-MD method based on density functional theory (DFT) calculations [26].
We have also developed a program that combines the Makri–Miller method with
the TOF-MD method (called TOF-MD-WKB) to calculate the tunneling splitting
in the electronic ground state of NH3 umbrella inversion and intramolecular
hydrogen transfer in malonaldehyde, and successfully reproduced the experimental
3.2 Semiclassical Tunneling Approach 69
values [27]. This method is a practical way to investigate tunneling effects in chem-
ical reactions, and the tunneling amplitudes are evaluated on the fly by ab initio
electronic structure calculations whenever a tunneling turning point is detected
along the trajectory. In addition, we have extended the TOF-MD-WKB method to
the electronically excited states and applied it to the evaluation of tunneling splitting
of the intramolecular hydrogen transfer process in the excited state of tropolone
and discussed it in detail [28]. In this chapter, we will introduce the TOF-MD-WKB
method and its application to the problem of tunneling splitting in the ground and
excited states, discuss the problems, and present future perspectives.
3.2 Semiclassical Tunneling Approach
Given a symmetric double well, the wave function of the vibrational ground state
delocalizes into two wells, resulting in eigenstates with in-phase and antiphase prob-
ability amplitudes on the one hand and the other. In general, the eigenvalue of an
in-phase eigenstate is more stable than that of an antiphase eigenstate, and this
difference in energy eigenvalues is called the energy splitting due to the tunneling
effect. The naming of the tunneling effect comes from the delocalization of particles
that do not have an energy corresponding to the height of the barrier in both wells.
In the one-dimensional WKB approximation, the tunneling splitting in a symmetric
double-well system is expressed as [1].
ΔE = 2ℏ𝜈 exp(−𝜃) (3.1)
where ℏ is the Dirac constant, 𝜈 is the frequency, and 𝜃 is the imaginary component
of the action integral along the tunnel path, defined as
xend √
1
𝜃= 2(V(x) − V(xinit ))dx (3.2)
∫
ℏ xinit
Here, x is the tunneling coordinate along the tunneling path, xinit and xend are the
initial and end points of the tunneling path, respectively. V(x) is the potential energy
at x, and the coordinates are all expressed in mass-weighted coordinates. The initial
and end points of the tunneling path are defined as the points where the poten-
tial energy equals the total energy of the system. When a particle is in vibrational
motion at the bottom of the well, the probability amplitude of tunneling through
the barrier from xinit to xend each time it hits the classical turning point (tunnel-
ing amplitude) is expressed as exp(−𝜃). This tunneling amplitude multiplied by the
number of vibrations hitting the turning point per unit time, 𝜈exp(−𝜃), means the
tunneling amplitude accumulated per unit time.
In the IRC–WKB method, the IRC is assumed to be a tunneling path, and the
calculation of the action integral given by Eq. (3.2) is performed along with the IRC.
The vibration frequency 𝜈 in Eq. (3.1) is the frequency of the vibration mode in the
direction along with the IRC at the potential minimum; since the IRC itself is a
one-dimensional reaction path in a multidimensional coordinate space, it is nec-
essary to take into account the multidimensional effect of the vibration degrees of
freedom orthogonal to the IRC. For this purpose, one can estimate the adiabatic
vibrational ground state energy V AG by adding the zero-point vibration energy of
3N − 7 vibration modes orthogonal to the IRC at each point along with the IRC to
the potential energy V [5–7] as,
∑1
3N−7
V AG (x) = V(x) + h𝜈 (x) (3.3)
i=1
2 i
where N, h, and 𝜈 i are the number of atoms in the system, Planck’s constant, and the
frequency of the i-th vibration mode orthogonal to IRC, respectively. By replacing
V(x) with V AG (x) in the action integral of Eq. (3.2), one can discuss the tunneling
effect considering the multidimensional effect.
In a one-dimensional system, the classical trajectories reach the turning point at
the same position with a constant period, the tunneling amplitude at each turning
point obtains a constant value, and the magnitude of the tunneling splitting is
uniquely determined. However, in an N-atom molecule, there are 3N − 7 vibrational
motions occurring simultaneously, so in a multidimensional system, the energy
transfer between the motion in the tunneling path direction and the vibrational
motions in the other vibrational degrees of freedom changes the turning point in
the tunneling direction and the time interval of hitting the turning point (this is also
a multidimensional effect). Therefore, the accumulated tunneling amplitude needs
to be averaged over several trajectories; in the TOF-MD-WKB method [26–28],
the trajectory is time-evolved in the potential well, and the turning point in the
tunneling path direction is sampled as the initial point of the tunneling path, and
each time the turning point is hit, the tunneling amplitude is calculated based
on Eq. (3.2), the action integral along the tunneling path is calculated, and the
tunneling amplitude is accumulated. This accumulated cumulative tunneling
amplitude, S(t), can be expressed by the following equation,
∑
S(t) = h(t − tn ) exp(−𝜃n ) (3.4)
n
{
0(t < tn )
h(t − tn ) = (3.5)
1(t ≥ tn )
where h(t − tn ) is the Heaviside step function and 𝜃 n is the action integral at the nth
tunneling event. The tunneling splitting is given by the time derivative of the cumu-
lative tunneling amplitude averaged over the classical trajectory, which corresponds
to 𝜈 exp(−𝜃).
d
ΔE = 2ℏ ⟨S(t)⟩ (3.6)
dt
Here, the bracket with S(t) in between shows the average by trajectories.
In the TOF-MD-WKB method, the choice of the turning point and tunneling path
is important. In principle, the tunneling path should be chosen so that the action
integral value is minimized, but in general, it is difficult to find such an optimal
tunneling path in a multi-dimensional coordinate space [29–31]. Therefore, it is nec-
essary to determine how to define an appropriate tunneling path before carrying out
TOF-MD simulation. For the turning point in the tunneling direction, we can choose
the point where the velocity component in the tunneling direction changes from
positive to negative. The time derivative of the atomic velocity along the tunneling
direction is expressed by the following equation:
d d d
(l(t)•v(t)) = l(t)•v(t) + l(t)• v(t)
dt dt dt
d
= l(t)•v(t) − l(t)•g(t) (3.7)
dt
where l(t), v(t), g(t) are the normalized tunneling direction vector, atomic velocity
vector, and potential gradient vector in the mass-weighted coordinates at t, respec-
tively. Since the tunneling direction vector l(t) varies with time [27, 28], the turning
point along the tunneling direction may satisfy l(t) ⋅ v(t) < 0. In such a case, this
turning point is not suitable as the initial point of the tunneling path because the
potential energy does not rise in the tunneling direction. Therefore, these points are
excluded from the tunneling calculation. The definition of the tunneling direction
vector l(t) will be described later.
3.3 Results and Discussion
In this section, the IRC-WKB and TOF-MD-WKB methods are applied to the
umbrella inversion of ammonia, the intramolecular hydrogen transfer of malon-
dialdehyde, and the intramolecular hydrogen transfer in the electronically excited
state of tropolone, where the intramolecular conformational change exhibits a sym-
metric double-well potential, to estimate the tunneling splitting [27, 28]. Figure 3.1
shows the structural changes associated with each reaction. For all the quantum
chemical calculations, we used GAMESS [32]. We discuss how to determine
the turning point of the trajectory that initiates the calculation of the tunneling
y O2 O1
H1
x
(b)
H1
O1 O2
(a)
C1 C2
y
x
(c)
Figure 3.1 Schematic illustration of (a) the umbrella inversion of ammonia, (b) the
intramolecular hydrogen transfer in malonaldehyde, and (c) the excited state intramolecular
hydrogen transfer in tropolone. Source: Based on Schmidt et al. [32].
amplitude and the tunneling path, and show the importance of multidimensional
effects through comparison with experimental values.
3.3.1 Umbrella Inversion of Ammonia

As a first example, we start with an application to the umbrella inversion of a simple
ammonia molecule, which has been studied in the past [27, 33, 34]. Intuitively, this
conformational change is described by a motion along a single vibrational mode, and
it is sufficient to consider the umbrella inversion vibrational mode as the tunneling
coordinate. The calculation of the equilibrium structure (C3v ) and the TS structure
(D3h ) using the Møller–Plesset second-order perturbation theory (MP2) with the
Sapporo-double zeta polarization (DZP) basis function [35] yields a barrier height of
5.0 kcal mol−1 , which is close to the more accurate coupled-cluster with singles, dou-
bles, and noniterative triples (CCSD(T))/quadruple zeta polarization (QZP) value of
5.3 kcal mol−1 , so we perform the calculations based on MP2/DZP [27]. Starting from
the TS structure, we first calculate the IRCs and then consider the IRC as a tunneling
path to calculate the tunneling splitting. To account for multidimensional effects, we
apply the vibrational adiabatic approximation and perform a normal mode analysis
in a five-dimensional coordinate space orthogonal to the IRC for selected structures
along with the IRC, and add the zero-point vibrational energy (ZPE) of the five
transverse vibration modes to the bare potential of the IRC (called intrinsic reaction
coordinate-vibrational adiabatic [IRC-VA]) [7]. For NH3 and isotopically substituted
ND3 , the tunneling splitting is calculated based on the IRC potential (IRC-bare) and
IRC-VA according to the IRC-WKB method. In the integral of Eq. (3.1) the lower and
upper limits of the tunneling coordinates are determined by the ZPE of the umbrella
inversion mode.
The calculated and experimental values of tunneling splitting for NH3 and ND3
are summarized in Table 3.1. For NH3 , the experimental value is 0.79 cm−1 [36],
while the calculated values by IRC-WKB are 2.18 cm−1 (IRC-bare) and 1.03 cm−1
(IRC-VA) [27]. The IRC-VA result is about half of the IRC-bare result and approaches
the experimental value, indicating the importance of multidimensional effects. Here
the ZPE of the vibrational mode orthogonal to the IRC increases near the TS. The
height of the effective barrier goes from 5.0 kcal mol−1 (IRC-bare) to 5.5 kcal mol−1
(IRC-VA), leading to a decrease in tunneling splitting. For ND3 , the calculated values
are 0.16 cm−1 (IRC-bare) and 0.072 cm−1 (IRC-VA) for IRC-WKB compared to the
experimental value of 0.053 cm−1 , and as in the case of NH3 , the calculated values
are closer to the experimental values by considering the multidimensional effect.
Next, we performed TOF-MD calculations for NH3 and ND3 at the MP2/DZP level.
The phase angle of each mode was determined by a random number, and 300 classi-
cal trajectories were run for NH3 and 100 for ND3 . The tunneling path was generated
by linear interpolation by picking up the classical turning point along the trajectory
that indicates the direction in which the umbrella reversal mode causes the umbrella
inversion as the initial point of the tunneling, and the terminal point of the tunneling
was generated by mirror operation, and the tunneling integral was calculated accord-
ing to Eq. (3.1). The results are shown in Table 3.1. TOF-MD-WKB gives calculated
Table 3.1 Calculated [27] and experimental [36] values of tunneling splitting due to
umbrella inversion in NH3 and ND3 (cm−1 ).
IRC-WKB
IRC-bare IRC-VA TOF-MD-WKB Experimental
NH3 2.18 1.03 0.71 0.79

ND3 0.16 0.072 0.078 0.053
tunneling splits of 0.71 and 0.078 cm−1 for NH3 and ND3 , respectively [27], which
are even closer to the experimental value of 0.79 cm−1 than the 1.03 cm−1 obtained
by IRC-WKB (IRC-VA) for NH3 . For NH3 , it is even closer to the experimental value
of 0.79 cm−1 than the 1.03 cm−1 obtained by IRC-WKB (IRC-VA).
3.3.2 Intramolecular Hydrogen Transfer in Malonaldehyde

Next, we consider the tunneling splitting due to intramolecular hydrogen transfer
in malonaldehyde, which is a 9-atom molecule with 21 degrees of freedom and
forms intramolecular hydrogen bonds between OH groups and O atoms, as shown
in Figure 3.1. When the H of the OH group is transferred to the O atom through
the hydrogen bond, the structures before and after the transfer are equivalent and
result in a symmetric double-well potential problem, and the vibrational wave
functions localized in each well interact with each other due to the tunneling
effect, resulting in the splitting of the energy levels. The hydrogen-transfer process
in malondialdehyde is a heavy–light–heavy system in which the H atom moves
between two O atoms, and it is known that there is a region before and after the
TS where the reaction path IRC is highly curved. In other words, the IRC includes
three stages of structural change: (i) OH and O approach each other, (ii) H moves
from one O to another and the bond alternation mode is reversed, and (iii) O and
HO leave each other. An IRC with a significant curvature in the coordinate space is
no longer appropriate as a tunneling path, and the tunneling effect estimated based
on the IRC is underestimated. In this context, the tunneling effect of malonaldehyde
has been studied using potential surfaces in two- or three-dimensional coordinate
space [37, 38], but recently, theoretical studies based on PESs considering the full
dimension have been performed [9, 10, 17, 20, 26, 27, 39, 40].
First, the IRC-WKB method is applied along with the IRC of intramolecular hydro-
gen transfer in malonaldehyde. To see the isotope effect, the calculations were also
performed for a molecular system in which the transferred H atom was replaced
by a D atom. For the electronic structure calculations, we used the MP2/6-31G(d,p)
level in consideration of computational cost and accuracy. Malonaldehyde is a planar
molecule with Cs symmetry at the equilibrium structure and C2v symmetry at the TS
structure. Since the reaction mode corresponding to the tunneling path direction at
the equilibrium structure is the O1 H1 · · ·O2 hydrogen bond stretching mode, the inte-
gral range for the tunneling amplitude was determined by the ZPE of this vibration
mode. As a result of IRC-WKB calculations, the tunneling splitting is 0.42 cm−1 (H)
and 0.15 cm−1 (D) when IRC-bare is used as the potential, and 4.9 cm−1 (H) and
0.8 cm−1 (D) when IRC-VA is used [27]. Since the experimental value of tunnel split-
ting is 21.6 cm−1 (H) and 2.9 cm−1 (D) [41], the calculated results are greatly under-
estimated compared to the experimental values, suggesting that IRC is not an appro-
priate tunneling path.
Next, we apply the TOF-MD-WKB method. In the case of malonaldehyde, the ZPE
is calculated to be 43.2 kcal mol−1 under the harmonic approximation, which is very
large compared to the barrier height of 3.6 kcal mol−1 , so intramolecular hydrogen
transfer may occur classically due to energy transfer between vibrational modes.
Therefore, instead of assigning ZPE to all the vibrational modes, ZPE was assigned
only to the vibrational modes related to O1 H1 · · ·O2 hydrogen bond stretching and
O1 H1 stretching modes, which are important for hydrogen transfer, and to apply
classical energy NkT to the other 19 vibrational modes as bath modes (N: num-
ber of bath modes, k: Boltzmann constant, and T: temperature). Hundred different
initial conditions were determined by randomly generating the initial phase angle.
To examine the energy dependence of the tunneling effect on the bath modes, the
temperature that determines the energy given to the bath modes was varied from
T = 0, 50, and 100 K, and each was run for 100 fs. To determine the turning point
in the tunneling direction along the trajectory, for each point on the trajectory, the
structure corresponding to the structure after hydrogen transfer was generated by
mirror operation, the direction obtained by linear interpolation of these two struc-
tures in Cartesian coordinates was defined as the tunneling direction, and the point
where the momentum in this direction evolves from positive to negative and the
point obtained by mirror operation was used as the initial and end points of the
tunneling path [27].
Table 3.2 shows the calculated and experimental values of tunneling splitting
[27]. While the IRC-WKB method underestimates the tunneling splitting, the
TOF-MD-WKB results overestimate the tunneling splitting when the bath mode
energy is 0 K. As the energy of the bath modes is increased, the tunneling splitting
decreases and approaches the experimental value. When the temperature of the
bath mode is raised above 100 K, the energy transfer between the modes leads to
a case where H-transfer is observed by classically crossing the barrier instead of
the tunneling effect. Regarding the isotope effect, tunneling splitting is reduced by
a factor of about 10 in deuterium-substituted materials, in good agreement with
experimental results [27].
Table 3.2 Calculated [27] and experimental [40] tunneling splitting due to intramolecular
hydrogen transfer in malondialdehyde and deuterium substituents (cm−1 ).
IRC-WKB TOF-MD-WKB
IRC-bare IRC-VA 0K 50 K 100 K Experimental
Malonaldehyde 0.42 4.9 54.5 22.7 19.8 21.6

Deuterium substituent 0.15 0.8 5.2 2.89 2.28 2.9
3.3.3 Excited State Intramolecular Hydrogen Transfer in Tropolone

Finally, an application to tropolone, which exhibits tunneling splitting in the
electronically excited state [28], is presented. Tropolone is a seven-membered ring
molecule with C7 H6 O2 and 39 degrees of internal freedom, as shown in Figure 3.1.
Like malonaldehyde, it has an intramolecular hydrogen bond between the OH and
O atoms, and intramolecular hydrogen transfer from OH to O atom generates a
symmetric double well, resulting in tunneling splitting of the vibrational ground
state [42]. From a theoretical point of view, Wojcik et al. have discussed tunneling
splitting based on quantum mechanical calculations using a two-dimensional
PES for the electronically excited state of tropolone [43, 44]. In our work, on
the other hand, we discussed the tunneling splitting based on the IRC-WKB and
TOF-MD-WKB methods considering all degrees of freedom [28]. First, the level
of the electronic structure calculation was determined by estimating the height of
the reaction barrier. In the framework of the time-dependent density functional
theory (TDDFT), the equilibrium and transition state structures in the excited state
are calculated using LC-BOP functional with 6-31+G* basis sets and the reaction
barrier was calculated to be 3.9 kcal mol−1 , which is a good agreement with the
value of 3.8 kcal mol−1 calculated by the EOM-CCSD(T)/aug-cc-pVDZ method
[45]. At the equilibrium and transition-state structures, the O–O interatomic
distances are calculated to be 2.49 and 2.32 Å for TDDFT (LC-BOP), respectively,
whereas they are 2.46 and 2.32 Å for EOM-CCSD, showing good agreement for the
structures. Based on the above results, the calculations are performed here at the
TDDFT(LC-BOP)/6-31+G* level.
Since tropolone forms a heavy–light–heavy system like malonaldehyde, the IRC
of the intramolecular hydrogen transfer reaction has a highly curved region before
and after the TS, and consists of three steps: first, O1 H1 and O2 approach each other,
then H1 is transferred from O1 to O2 , and finally O1 and H1 O2 leave each other.
A normal mode analysis was performed at each point along the IRC calculated
for the S1 excited state, and the vibration adiabatic ground state energy (V AG ) was
calculated, yielding an activation barrier of 2.5 kcal mol−1 , which is a significant
decrease in the barrier from 3.9 kcal mol−1 due to multidimensional effects. The
V AG -based IRC-WKB calculation estimates the tunneling splitting to be 1.4 cm−1 ,
which is a significant underestimate by a factor of 10 compared to the experimental
value of 18.9 cm−1 [42]. This means that to accurately evaluate the tunneling effect
in tropolone, it is necessary to apply a method that treats the multidimensional
effect more appropriately.
Next, TOF-MD simulation is performed on the excited state of tropolone around
the equilibrium structure of the S1 state. As initial conditions, three O1 C1 C2 /C1 C2 O2
angular vibrational modes and one O1 H1 stretching vibrational mode, which are
considered to be directly related to hydrogen transfer, are selected from the 39 ref-
erence vibrational modes of tropolone and given quantum ZPE, while the other 36
vibrational modes are given classical energy NkT as bath mode [28]. Here we con-
sider bath mode energies of T = 0 K (0.0 kcal mol−1 ), 10 K (0.70 kcal mol−1 ), 30 K
(2.09 kcal mol−1 ), and 50 K (3.48 kcal mol−1 ). The simulation at T = 0 K corresponds
to a simulation in four-dimensional space directly involved in the hydrogen transfer

reaction, although energy transfer to bath mode is also allowed, while the other cases
correspond to a simulation considering all degrees of freedom. For each tempera-
ture condition, 100 initial coordinates and initial velocities were generated, and the
classical trajectory was run for 100 fs, with a MD calculation time step of 0.2 fs.
In the TOF-MD-WKB method, it is necessary to decide how to determine the
tunneling path. According to the least action principle, the tunneling path is bet-
ter when the barrier height along the path is low and the path length is short [15].
In other words, as a tunneling path, the IRC path through the TS is preferred from
the energy point of view, but if the curvature of the IRC is large, the action inte-
gral becomes smaller if the path is shortened by corner-cutting. On the other hand,
given the initial and end points of the tunneling path, from the viewpoint of path
length, the linear path connecting the two structures is the extreme case with the
smallest path length. In this work, a practical method is introduced to determine
the tunneling path by interpolating the path through the TS and the linear path.
First, structures after intramolecular hydrogen transfer are generated for any point
along the classical trajectory by mirror operation, and the mass-weighted Cartesian
coordinates of each structure are denoted as x init and x end . These structures are chi-
ral to each other and have the same energy. Consider a linear path connecting these
two structures, and define the midpoint x mid as follows:
xmid = (xinit + xend )∕2 (3.8)
Further, interpolate x mid and TS structure x TS to introduce the following via point
x via .
xvia = (1 − p)xmid + pxTS (0 ≤ p ≤ 1) (3.9)
where p is the weight parameter. In this study, the tunneling path is the combination
of the linear path from x init to x via and the linear path from x via to x end as shown
in Figure 3.2. When p = 0 (x via = x mid ), the tunneling path is a linear path directly
connecting the initial and endpoints, and when p = 1 (x via = x TS ), the tunneling path
is a path through the TS structure. The p-value at which the tunneling amplitude is
minimized is considered to give the optimal tunneling path. For the structure on
the trajectory, the normalized vector l corresponding to the tunneling direction is
defined as
(xvia − xinit )
l= (3.10)
∣ (xvia − xinit ) ∣
Using vector l, the velocity component in the tunneling direction is calculated along
the trajectory according to Eq. (3.7), and once the turning point is found, the action
integral in Eq. (3.2) along the tunneling path is calculated from x init to x via .
Figure 3.3a shows the dependence of the tunneling splitting value on the weight
parameter p. The tunneling splitting shows a maximum for p = 0.6–0.8 at T = 0 K,
but the p that gives the maximum tunneling splitting value decreases as the bath
mode energy increases. The tunneling splitting at the optimum p for each bath mode
energy is shown in Figure 3.3b. When the bath mode energy is 0 K, the tunneling
splitting value is overestimated to 33.4 cm−1 , but when the energy is given in bath
Figure 3.2 Schematic view of the

tunneling path.
XTS
Xvia
H1
O1 O2
C1 C2
Xinit Xmid Xend
40 40
Tunneling splitting (cm )
0K Tunneling splitting (cm )

–1
–1
10 K
30 K
30 50 K 30
20 20
10 10
0 0
0.0 0.2 0.4 0.6 0.8 1.0 0 10 20 30 40 50
(a) p (b) Initial bath-mode energy (K)
Figure 3.3 (a) p-Dependence of tunneling splitting and (b) bath mode energy dependence
of tunneling splitting.
mode, the tunneling splitting decreases to 19.1–8.1 cm−1 , which is comparable to

the experimental value of 18.9 cm−1 [42]. Figure 3.4a shows the p-dependence of
the tunneling amplitude averaged over all tunneling events seen in the 100 trajecto-
ries calculated for each bath mode energy. The tunneling amplitude increases with
increasing p and shows a maximum at p = 0.6–1.0. The tunneling action integral
𝜃 n depends on the height of the barrier in the potential curve along the tunneling
path and the tunneling path length. Figure 3.4b,c show the energy barrier and path
length of the tunneling path averaged over tunneling events. The height of the acti-
vation barrier is independent of the bath mode energy, but as p increases and the via
point approaches the TS, the barrier decreases, and the tunneling splitting increases.
Therefore, the p-dependence of the tunneling amplitude is due to the change in the
barrier height, while the dependence on the bath mode energy is due to the change
in the tunneling path length. The dependence of the tunneling amplitude on the
bath mode energy due to the change in the tunneling path is a multidimensional
effect in the tunneling event that is not considered in the IRC-WKB method.
The tunneling splitting value is also related to the number of tunneling events per
unit time. Figure 3.5a shows the average number of tunneling events per 100 fs cal-
culated for 100 trajectories. It can be seen that as the weight parameter p increases,
the number of tunneling events gradually decreases, resulting in a decrease in the
path length (bohr amu1/2)

0.07 0 K 15 6
tunneling amplitude
Average tunneling
height (kcal mol–1)
0.06 3010 K
K
Average barrier
0.05 50 K 10 4
Average
0.04
0.03
0.02 5 2
0.01
0.00 0 0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0
(a) p (b) p (c) p
Figure 3.4 Changes in (a) tunneling amplitude, (b) barrier height, and (c) tunneling path
length, were averaged over all tunneling events in 100 trajectories. Source: Based on
Redington et al. [42].
10 2.6 2.6
O1–O2 distance (Å)
O1–O2 distance (Å)

Average number of
tunneling event
8
2.5 2.5
6
4
2.4 2.4
0K
2 10 K
30 K
50 K
0 2.3 2.3
0.0 0.2 0.4 0.6 0.8 1.0 0.8 1.0 1.2 1.4 1.6 0.8 1.0 1.2 1.4 1.6
(a) p (b) O1–H1 distance (Å) (c) O1–H1 distance (Å)
Figure 3.5 (a) p-Dependence of the number of tunneling events in 100 trajectories.
(b, c) Examples of trajectories and tunneling paths projected into two-dimensional space
with O1 –H1 and O1 –O2 distances ((b) p = 0, (c) p = 1). Solid lines indicate trajectories and
dashed lines indicate tunneling paths.
tunneling splitting value. Figure 3.5b,c show examples of projections of trajectories

into the two-dimensional coordinate space at O1 –H1 and O1 –O2 distances. When the
weight parameter is small (p = 0) in (b), tunneling occurs when the O1 –H distance
is at its maximum, indicating that the tunneling path is toward the O–H stretching
direction. Therefore, the frequency of tunneling occurrence is almost the same as the
high frequency of O1 –H stretching mode. On the contrary, when the weight param-
eter in (c) is large (p = 1), the tunneling events do not correlate with the timing of
the maximum O–H length because the tunneling path is toward the TS structure.
As a result, as the weight parameter p increases, the number of tunneling events
decreases, leading to a decrease in tunneling splitting.
To investigate the effect of out-of-plane vibration modes on tunneling splitting,
we also performed TOF-MD-WKB simulation for only the in-plane vibration modes
of tropolone. The number of bath modes was restricted to 23 in-plane vibration
modes with energies of T = 10 K (0.46 kcal mol−1 ), T = 30 K (1.37 kcal mol−1 ),
and T = 50 K (2.29 kcal mol−1 ). As in the full-degree-of-freedom simulation, the
tunneling splitting shows a maximum value for the weight parameter p = 0.6–0.8. In
the in-plane simulation, the tunneling splitting is nearly constant at ∼30 cm−1 and is
always larger than that obtained in the full simulation. This result indicates that the
out-of-plane vibrational mode reduces the tunneling splitting, which is consistent
with previous reports on ground-state tunneling splitting in malonaldehyde [46].
To further understand the role of out-of-plane vibrational motion, we examined
the tunneling amplitude. The average height of the barrier along the tunneling
path was similar to that of the full-degree-of-freedom simulation, but the tunneling
Acknowledgments 79
path length behaved differently. The tunneling path length in the energy range of
10–50 K is as short as the tunneling path length at 0 K in the full-degree-of-freedom
simulation and does not increase with increasing bath mode energy. Thus, the
tunneling splitting value is nearly constant, independent of the bath mode energy.
Note that the number of tunneling events is almost the same as in the case of
full-degree-of-freedom simulation. The out-of-plane oscillatory motion reduces the
tunneling splitting value by increasing the tunneling path length.
3.4 Conclusions
In this section, we introduce the TOF-MD-WKB method, which implements

the Makri–Miller method to estimate the tunneling effect semi-classically in the
TOF-MD method that can be applied to any molecular system without the need
to prepare a potential function. The present method was applied to the NH3
umbrella inversion, the intramolecular hydrogen transfer in malonaldehyde, and
the intramolecular hydrogen transfer in the excited state of tropolone, where the
potential profile for the conformational change of the molecule is a symmetric
double-well potential and was shown to be able to quantitatively reproduce the
experimental values of tunneling splitting. The intramolecular hydrogen transfer in
malonaldehyde is a typical heavy–light–heavy mass combination system, and the
IRC-WKB underestimates the tunneling effect because the IRC is sharply curved
before and after the TS. To avoid unphysical energy transfer between the vibrational
modes, the TOF-MD-WKB simulations were performed by dividing the internal
degrees of freedom into two parts, where the OH· · ·O hydrogen bond stretching
mode and the O–H stretching mode relevant to the reaction process were given
quantum ZPE and the other degrees of freedom were given classical energy. The
calculated results were in good agreement with the corresponding experimental
values, and the isotope effect was also successfully reproduced. The intramolecular
hydrogen transfer reaction in the electronically excited state of tropolone is also
a heavy–light–heavy mass combination system, where IRC-WKB underestimated
the tunneling effect, but TOF-MD-WKB gave results in good agreement with
experimental results. In this application, the optimal tunneling path was searched
by a weighted average of the linear path connecting the initial and end points of the
tunneling and the path through the TS structure. The dependence on the bath mode
energy was also analyzed in detail to clarify the importance of multidimensional
effects in the tunneling process.
Acknowledgments
We are sincerely grateful to Ms. Aya Satoh and Dr. Yu Harabuchi for their collabo-
rated work on the tunneling splitting of tropolone in the excited state.
References
1 Landau, L.D. and Lifshitz, E.M. (1981). Chapter 7, The quasi-classical case. In:
Quantum Mechanics: Non-relativistic Theory, 3e, 1153. Butterworth-Heinemann.
2 Fukui, K. (1970). J. Phys. Chem. 74: 4161.
3 Maeda, S., Harabuchi, Y., Ono, Y. et al. (2015). Int. J. Quantum Chem. 115: 258.
4 Taketsugu, T. and Wales, D.J. (2002). Mol. Phys. 100: 2793.
5 Watanabe, Y., Taketsugu, T., and Wales, D.J. (2004). J. Chem. Phys. 120: 5993.
6 Takahashi, M., Watanabe, Y., Taketsugu, T., and Wales, D.J. (2005). J. Chem.
Phys. 123: 044302.
7 Truhlar, D.G., Isaacson, A.D., Skodje, R.T., and Garrett, B.C. (1982). J. Phys.
Chem. 86: 2252.
8 Taketsugu, T. and Hirao, K. (1997). J. Chem. Phys. 107: 10506.
9 Mil’nikov, G.V., Yagi, K., Taketsugu, T. et al. (2003). J. Chem. Phys. 119: 10.
10 Mil’nikov, G.V., Yagi, K., Taketsugu, T. et al. (2004). J. Chem. Phys. 120: 5036.
11 Mil’nikov, G.V. and Nakamura, H. (2008). Phys. Chem. Chem. Phys. 10: 1374.
12 Richardson, J.O. and Althorpe, S.C. (2011). J. Chem. Phys. 134: 054109.
13 Richardson, J.O. (2018). J. Chem. Phys. 148: 200901.
14 Kawatsu, T. and Miura, S. (2014). J. Chem. Phys. 141: 024101.
15 Makri, N. and Miller, W.H. (1989). J. Chem. Phys. 91: 4026.
16 Guo, Y. and Thompson, D.L. (1998). A multidimensional semiclassical approach
for treating tunneling within classical trajectory simulations. In: Modern
Methods for Multidimensional Dynamics Computations in Chemistry (ed. D.L.
Thompson), 713. Singapore: World Scientific.
17 Sewell, T.D., Guo, Y., and Thompson, D.L. (1995). J. Chem. Phys. 103: 8557.
18 Guo, Y., Qin, Y., Sorescu, D.C., and Thompson, D.L. (1996). J. Chem. Phys. 104:
4041.
19 Taketsugu, T., Watanabe, N., and Hirao, K. (1999). J. Chem. Phys. 111: 3410.
20 Yagi, K., Taketsugu, T., and Hirao, K. (2001). J. Chem. Phys. 115: 10647.
21 Chen, W., Hase, W.L., and Schlegel, H.B. (1994). Chem. Phys. Lett. 228: 436.
22 Gordon, M.S., Chaban, G., and Taketsugu, T. (1996). J. Phys. Chem. 100: 11512.
23 Ben-Nun, M., Quenneville, J., and Martinez, T.J. (2000). J. Phys. Chem. A 104:
5161.
24 Taketsugu, T., Tajima, A., Ishii, K., and Hirano, T. (2004). Astrophys. J. 608: 323.
25 Ootani, Y., Satoh, K., Nakayama, A. et al. (2009). J. Chem. Phys. 131: 194306.
26 Ben-Nun, M. and Martinez, T.J. (1999). J. Phys. Chem. A 103: 6055.
27 Ootani, Y. and Taketsugu, T. (2012). J. Comput. Chem. 33: 60.
28 Ootani, Y., Satoh, A., Harabuchi, Y., and Taketsugu, T. (2020). J. Comput. Chem.
41: 1549.
29 Mil’nikov, G.V. and Nakamura, H. (2001). J. Chem. Phys. 115: 6881.
30 Tautermann, C.S., Voegele, A.F., Loerting, T., and Liedl, K.R. (2002). J. Chem.
Phys. 117: 1962.
31 Tautermann, C.S., Voegele, A.F., Loerting, T., and Liedl, K.R. (2002). J. Chem.
Phys. 117: 1967.
References 81
32 Schmidt, M.W., Baldridge, K.K., Boatz, J.A. et al. (1993). J. Comput. Chem. 14:
1347.
33 Brown, F.B., Tucker, S.C., and Truhlar, D.G. (1985). J. Chem. Phys. 83: 4451.
34 Lauvergnat, D. and Nauts, A. (2004). Chem. Phys. 305: 105.
35 Noro, T., Sekiya, M., and Koga, T. (2012). Theor. Chem. Acc. 131: 1124.
36 Spirko, V. (1983). J. Mol. Spectrosc. 101: 30.
37 Carrington, T. Jr., and Miller, W.H. (1986). J. Chem. Phys. 84: 4364.
38 Shida, N., Barbara, P.F., and Almlof, J.E. (1989). J. Chem. Phys. 91: 4061.
39 Meyer, R. and Ha, T.K. (2003). Mol. Phys. 101: 3263.
40 Wang, Y., Braams, B.J., Bowman, J.M. et al. (2008). J. Chem. Phys. 128: 224314.
41 Baughcum, S.L., Smith, Z., Wilson, E.B., and Duerst, R.W. (1984). J. Am. Chem.
Soc. 106: 2260.
42 Redington, R.L., Chen, Y., Scherer, G.J., and Field, R.W. (1988). J. Chem. Phys.
88: 627.
43 Wojcik, M.J. (2008). J. Mol. Liq. 141: 39.
44 Wojcik, M.J., Boda, L., and Boczar, M. (2009). J. Chem. Phys. 130: 164306.
45 Burns, L.A., Murdock, D., and Vaccaro, P.H. (2009). J. Chem. Phys. 130: 144304.
46 Yagi, K., Mil’nikov, G.V., Taketsugu, T. et al. (2004). Chem. Phys. Lett. 397: 435.
83
Spectroscopy
Part II
85
Spectroscopic Signatures of Low-Barrier Hydrogen Bonding

in Neutral Species
Lidor Foguel, Zachary N. Vealey, and Patrick H. Vaccaro
Yale University, Department of Chemistry, P.O. Box 208107, New Haven, CT 06520-8107, USA
4.1 Introduction
Noncovalent interactions [1–4], as epitomized by the ubiquitous donor–acceptor
motifs of hydrogen bonding, permeate the world around us, being responsible
for the binding forces that serve to coalesce soft matter (e.g. biomolecular assem-
blies) and condensed phases (e.g. water), as well as for the elementary reactions
(e.g. proton transfer) that mediate a multitude of chemical transformations [5, 6].
An extensive body of experimental and theoretical work has established the basic
paradigms that govern these intermolecular and intramolecular phenomena;
however, crucial questions regarding their origin and manifestation continue
to emerge, spawning renewed efforts to formulate a more comprehensive and
inclusive definition for the hydrogen bond [7, 8]. These assertions are bolstered by
the vigorous debate that has erupted over the biochemical relevance of so-called
low-barrier hydrogen bonds (LBHBs) [9, 10], where a small, yet nonvanishing,
potential-energy barrier for hydron transduction has been postulated (albeit not
without controversy) [11] to imbue great strength (40–80 kJ mol−1 ) over short
donor–acceptor distances (<2.6 Å) [12, 13]. These entities have been implicated in
diverse biocatalytic pathways, where putative formation of a lone LBHB is thought
to impart vital stabilization to an enzyme-bound transition state [9, 14–17]. Critics
of this hypothesis have adopted a radically different stance, claiming the resulting
complexes to be inherently anticatalytic and advancing alternative explanations for
biological catalysis [18–22]. Indeed, despite compelling evidence for the existence
of LBHBs in isolated molecules [23–25] and crystalline lattices [26–29], several
studies have challenged the resilience of such bonding motifs in the face of
extrinsic perturbations, such as solvation [11, 30–32]. By focusing on the distinct
laser-spectroscopic signatures arising from neutral model systems entrained in
rarefied (gas-phase) environments, the present chapter strives to elucidate the
unique structural properties and dynamical propensities that distinguish LBHBs
from their more conventional counterparts.
The modern concept of hydrogen bonding [7, 8] entails a net attractive interac-
tion between a hydrogen atom from a molecule or molecular fragment, A–H in
86 4 Spectroscopic Signatures of Low-Barrier Hydrogen Bonding in Neutral Species
which A (the proton donor) is more electronegative than H (the labile hydron),
and an atom or atomic assembly, B (the proton acceptor), located in the same or
a different molecule. Both directionality of binding [33, 34] and clear evidence
of bond formation (experimental and/or theoretical) are essential aspects of this
all-encompassing definition, which reflects the inherent complexity of a phe-
nomenon that conspicuously is more than the sum of its parts [35]. Indeed, the
collection of synergistic/antagonistic factors responsible for the manifestation of
a hydrogen bond spans the gamut of intramolecular and intermolecular forces,
with their relative extent of participation varying markedly subject to the intrinsic
chemical/electronic structure of the system as well as extrinsic perturbations
sustained from the surrounding environment. Although historically believed
to be primarily electrostatic in nature [36], both experiment and theory have
revealed notable contributions arising from dispersion, polarization/induction,
exchange–correlation, and charge-transfer processes [37–39], where the latter effects
can imbue significant covalent character throughout the resulting A–H· · ·B moiety,
especially in the case of ionic species [12, 40–42]. Likewise, early empirical criteria
for gauging the existence of a hydrogen bond, including the once-touted reduction
of the donor–acceptor distance, dA· · ·B , to values below the sum of their respective
van der Waals radii [43, 44], have been supplanted by a variety of more refined
metrics [7, 8], many of which are based on observable spectroscopic signatures.
An intimate relationship exists between hydrogen bonding and proton trans-
fer [6], where the former essentially sets the stage upon which the acts of the
latter are performed. From this perspective, the noncovalent interactions respon-
sible for hydrogen-bond formation afford a potent means of pre-organizing the
nascent reactant and product in an activated A–H· · ·B complex that ultimately
facilitates transduction of the labile hydron between donor (A) and acceptor
(B) sites [5, 45]. The intrinsic complexity of this multidimensional process is
compounded by the light mass of the hydrogen atom, which necessitates that
nuclear-quantum effects [46–49], included those ascribed to zero-point energy,
potential-barrier tunneling, and delocalization of probability density, be explicitly
taken into account to describe structure and dynamics quantitatively. Aside from
their ability to mediate properties of a reaction center and influence the ensuing
pathways of chemical transformation, quantum phenomena can be expected to
engender unique spectroscopic signatures and unusual kinetic isotope effects
[50, 51], each of which can serve to distinguish hydron-migration events from
their more-massive counterparts (viz, reactions where nuclei can be modeled
reliably as classical point particles). Evidence supporting this assertion can be found
in burgeoning ab initio molecular-dynamics simulations that directly highlight
the marked differences in behavior predicted when nuclear degrees of freedom
are treated by classical and quantal (e.g. path integral) techniques [52–54]. In
particular, such theoretical efforts have suggested that the strengths of relatively
short and strong (long and weak) intermolecular hydrogen bonds, as encoded in the
spectral transition frequencies of attendant A–H-stretching vibrations, tend to be
strengthened (weakened) even further by the action of competing nuclear-quantum
effects [55–57].
Numerous monographs [6, 58–64] and reviews [12, 65–70] have elaborated on the
structural and dynamic consequences of hydrogen bonding as well as the putative
properties engendered by short and strong variants of this important binding motif.
Rather than repeating this litany of concepts and paradigms, the present chapter
will focus on ongoing attempts to quantitatively characterize the low-barrier regime
of such noncovalent interactions by means of vibronically resolved electronic
spectroscopy performed in the near-ultraviolet region. Although the experimental
and theoretical efforts of other research groups will be discussed, primary emphasis
will be placed on recent work emerging from the authors’ laboratory that has
afforded the first laser-spectroscopic evidence for the existence of a LBHB in an
isolated neutral molecule (i.e. 6-hydroxy-2-formylfulvene or HFF) [71, 72]. To
highlight the unprecedented nature of tunneling-mediated signatures uncovered
during these studies, the potential-energy topographies and vibrational-energy
landscapes supported by two closely related constitutional isomers (viz, HFF and
tropolone or TrOH) will be contrasted, thereby revealing the origins of the markedly
different above-barrier and below-barrier behaviors that have been demonstrated
in these kindred species.
4.2 Spectroscopic Metrics for Hydrogen Bonding

4.2.1 Continuum of Hydrogen Bonding
The structural and dynamical consequences of hydrogen bonding reflect a con-
tinuum of donor–acceptor motifs that support noncovalent interactions with
effective strengths ranging from very weak to extraordinarily strong, where the
latter even can approach the bond energies of their covalent counterparts [12]. The
characteristic behaviors encountered across this span are illustrated schematically
in Figure 4.1 by projecting attendant potential-energy hypersurfaces along the
one-dimensional reaction coordinate for hydron transfer that has been assumed
(for simplicity) to coincide with displacement of the hydrogen bond. The minima
appearing in these model curves correspond to equilibrium geometries having the
hydron (H) bound primarily to the donor (A) or the acceptor (B) moiety (where
labels reflect correspondence with Brønsted definitions of acids and bases). These
species often are separated by a potential-energy barrier of finite height, the apex
of which defines the transition-state configuration (a first-order saddle point) for
the proton-transfer process. As discussed below, the intrinsic properties of a given
hydrogen-bonding motif (e.g. bond strength), as well as the efficacy of correlated
reactive events (viz, proton transfer), are encoded in the overall symmetry and
topography of the depicted energy profiles [73]. Of particular interest are the nature
of tunneling-induced signatures accompanying the migration of hydrons between
donor and acceptor sites involving oxygen atoms (viz, A ≡ O and B ≡ O) [74], as
well as related phenomena caused by isotopic modification of the “shuttling”
atom. Emphasis will be placed on primary isotope effects (e.g. alterations in A—H
bond properties and tunneling metrics) induced when the labile proton (H) is
B H–A B H A
A–H B X H X
A–H B A H B
Relative energy
Δ0 ( )
(+) ( )
Δ0 ( )
(+)
(+)
Hydron-transfer coordinate Hydron-transfer coordinate Hydron-transfer coordinate Hydron-transfer coordinate

(a) (b) (c) (d)
Relative hydrogen-bond strength
Figure 4.1 Continuum of hydrogen-bonding motifs. The progression of hydrogen-bonding motifs from weakest (a) to strongest (d) is illustrated by
projecting attendant potential-energy topographies along the one-dimension reaction coordinate for hydron transfer. In each panel, the proton donor (A)
and proton acceptor (B) are separated by a barrier of ΔE pt height such that the corresponding minimum-energy points reflect the crucial donor–acceptor
distance, dA· · ·B . The magnitude of ΔE pt decreases continuously as dA· · ·B shortens, ultimately leading to the single-well behavior of the last panel (where A
and B are designated by the common label of X). The lowest-lying eigenstates for each model potential are depicted, with the symmetrical
double-minimum forms in panels (b)-(d) also containing metrics that describe the magnitude of tunneling-induced bifurcation for the zero-point level,
Δ0 , and the parity of corresponding wavefunctions with respect to the reaction coordinate, +/−.
replaced by a more massive deuteron (D) [50, 75]. The origins and repercussions of
their secondary counterparts, such as the Ubbelohde effect [76] whereby isotopic
substitution of the bound hydron leads to a synergistic change in the A· · ·B
donor–acceptor distance (dA· · ·B ), will not be addressed.
Figure 4.1a illustrates the most common form of hydrogen bonding where the
donor and acceptor are chemically distinct moieties that possess different gas-phase
proton affinities, ΔPA = PAdonor − PAacceptor ≠ 0, and/or condensed-phase acid disso-
acceptor
ciation constants, ΔpKa = pKdonor
a − pKa ≠ 0 [77]. The resulting asymmetrical
shape for the potential-energy curve supports two inequivalent equilibrium configu-
rations, the lowest-lying of which entails primarily covalent A—H and noncovalent
H· · ·B bonds. Such circumstances give rise to the weakest manifestation of hydrogen
bonding,1 producing effective AH· · ·B bond strengths of Ehb ≤ 17 kJ mol−1 that
operate over donor–acceptor distances of 3.0 ≤ dA···B ≤ 3.5 Å [5, 60]. The changes
in electron density proximate to the labile hydron have measurable effects on a
variety of spectroscopic metrics, most notably causing a systematic deshielding (or
down-field shift) of the associated 11 H-NMR resonance as the strength of H· · ·B
coupling increases [80–84]. In the case of oxygen moieties, the O–H-stretching
fundamental, 𝜈̃ OH , undergoes a progressive shift of transition wavenumber toward
lower values (i.e. a redshift) as well as concurrent increases in the cross section,
spectral width, and aggregate complexity of the infrared-absorption profile [85],
with the attendant protium–deuterium
√ isotopic ratio, 𝜈̃ OH ∕𝜈̃ OD , dropping from the
harmonic expectation of ∼ 2 [12, 85–87]. This effective “softening” of the O–H
linkage will be accompanied by the creation of new vibrational features related to
nascent H· · ·O bond formation [88–91]. When properly calibrated and referenced
(typically to the properties of the free A–H moiety), each of these observable
signatures can afford quantitative response–structure correlations [82, 92–95] that
reflect the configuration of the A–H· · ·B reaction site (e.g. dA· · ·B ) and the nature of
adjoining interactions (e.g. Ehb ). The efficacy of the corresponding proton-transfer
process is severely encumbered by the nondegenerate nature of the donor and
acceptor (which limits the extent of overlap and interaction between wavefunctions
that are highly localized in the inequivalent potential wells), as well as by the
opacity of the sizable potential-energy barrier that separates them, thereby leading
to a pronounced quenching of tunneling-mediated dynamics [96–98].
The presence of equivalent donor and acceptor moieties, as engendered by the
intrinsic structural symmetry of an isolated molecular assembly (ΔPA = 0) or
the extrinsic properties of the surrounding environment (ΔpKa = 0) [77], can
lead to potential-energy surfaces that display a characteristic symmetrical double-
minimum topography along the proton-transfer (or hydrogen-bonding) coordinate.
Such scenarios are illustrated by the one-dimensional model forms found in
1 In keeping with the assertion that hydrogen bonds are “interactions without borders” [78], these
limits on Ehb and dA· · ·B are meant as guides rather than sharp delineations. For example, at longer
distances the directional forces of hydrogen bonding will smoothly be supplanted by their omnidi-
rectional van der Waals counterparts [79]. Although the (free) energy of formation for a hydrogen-
bonded complex formally should be negative, the common practice of specifying positive values for
quantities such as Ehb will be adopted throughout this chapter.
Figure 4.1b,c. In each case, two equilibrium configurations corresponding to degen-

erate A–H· · ·B and A· · ·H–B tautomers are separated by a potential barrier of height
ΔEpt , which progressively decreases as the donor–acceptor distance (reflected in
the separation of minimum-energy stationary points) becomes shorter [99–101].
Both experiment and theory have provided compelling evidence for a near-linear
relationship between dA· · ·B and ΔpKa in which the smallest values of dA· · ·B are
achieved in the limit of ΔpKa → 0 [83, 102]. This simple assertion, which appears
to be quite robust with respect to the nature of the surrounding medium, stands
in stark contrast to the complexity that besets the energetics of hydrogen-bond
formation [70]. Although isolated-molecule studies have tended to reinforce
long-standing beliefs of an inverse correlation between Ehb and dA· · ·B whereby the
effective hydrogen-bond strength grows as the donor and acceptor sites are brought
closer together [5], the general validity of this assertion, especially with respect
to spectroscopic measurements performed in condensed phases [11, 103, 104],
remains a topic of active debate. Nevertheless, the notably anharmonic nature
of resulting potential-energy profiles can be expected to yield nuclear eigenstates
that depart markedly from those predicted by conventional double-harmonic or
perturbative-anharmonic treatments of vibrational displacements [105, 106]. For
example, the atypical (i.e. inverse) anharmonicity of the resulting vibrational-energy
ladder was exploited in recent ultrafast two-dimensional infrared studies to identify
and characterize the potential-energy topographies of strong hydrogen-bonding
systems [107–109]. Likewise, the 𝜈̃ OH ∕𝜈̃ OD isotopic ratio for the O–H-stretching
fundamental has been found to fall steadily (albeit not linearly) to values approach-
ing unity as dO· · ·O lessens and ultimately to rise again when the donor–acceptor
separation drops below roughly 2.5 Å [12, 85, 101, 110].
For the model potential depicted in Figure 4.1b, the vibrationless (v = 0)
zero-point energy of the system is presumed to be substantially less than that of
the transition-state geometry, EZPE ≪ ΔEpt . Under these circumstances, transfer
of the labile hydron between donor and acceptor sites by classical above-barrier
motion will be hindered (viz, subject to the availability of sufficient thermal energy
to surmount the barrier crest), but still can take place by means of coherent
quantum mechanical tunneling [74]. Such barrier-penetration phenomena leave a
characteristic fingerprint on molecular eigenstates, causing all rovibronic features
to be bifurcated into components having wavefunctions that are symmetric (+)
and antisymmetric (−) with respect to the reaction coordinate (e.g. the paired
components for the vibrationless or v = 0 eigenstate commonly are designated as
0+ and 0− ). The magnitude of this tunneling-induced splitting scales in direct pro-
portion to the effective rate of proton transfer, kpt , and reflects not only the transit of
the “shuttling” hydron but also the concomitant displacement of other heavy atoms
constituting the assembly (which can serve to modulate dA· · ·B ) [111] and the redis-
tribution of electron charge about the donor and acceptor moieties [74]. Although
not captured by the one-dimensional forms of Figure 4.1, the emerging spectral
signatures are expected to be quantum-state specific, reflecting the differential
ability of selective vibronic excitation to influence the detailed pathway and overall
efficacy of proton-transfer dynamics. In the case of deep tunneling, where the
family of eigenstates residing substantially below ΔEpt retain nuclear probability
densities that are notably skewed toward the minimum-energy configurations

(cf. wavefunctions in Figure 4.1b), the state-specific bifurcation engendered in
vibronic state 𝜐, Δ𝜐 (e.g. Δ0 for the vibrationless zero-point level), typically is much
smaller than vibrational spacings and, thus, can be treated as a perturbative shift on
the resulting symmetric and antisymmetric manifolds of rovibronic eigenstates [74].
Nevertheless, the exponential dependence of tunneling probability on particle mass
implies that a large primary deuterium kinetic-isotope effect (DKIE) should exist,
Λ𝜐 = Δ𝜐 (H)/Δ𝜐 (D) = kpt (H)/kpt (D), with a metric of Λ𝜐 ≫ 7 often being used to
corroborate the involvement of hydron tunneling between oxygen centers during a
given reaction mechanism [75].
As the donor–acceptor distance shortens (dropping to 2.5 ≤ dO···O ≤ 2.6 Å in
the case of oxygen moieties) [5, 15, 112] and ΔEpt undergoes a parallel reduction
in height, the symmetric double-minimum potential along the proton-transfer
(or hydrogen-bonding) coordinate eventually attains the characteristic shape
depicted in Figure 4.1c. Here the zero-point energy of the system straddles the bar-
rier crest, EZPE ≈ ΔEpt , heralding the transition from below-barrier to above-barrier
dynamics that defines the onset of low-barrier hydrogen bonding (LBHBing).
The notable lessening of barrier opacity obtained under such conditions greatly
enhances the efficacy of tunneling, with the magnitude of bifurcation sustained by
molecular eigenstates approaching that of typical vibrational spacings. Quantitative
interpretation of the resulting spectral signatures encoded in the locations and
intensities of rovibronic transitions, as well as their alterations by kinetic-isotope
effects, thus will demand the simultaneous non-perturbative consideration of
large-amplitude (tunneling) and small-amplitude (vibrational) degrees of freedom.
Moreover, the nuclear probability distributions for the labile hydron (cf. wave-
functions in Figure 4.1c) now display substantial amplitude throughout the region
adjoining the donor and acceptor sites [113], with the implication of nascent hydron
delocalization reflecting the spread of covalent character across the A· · ·H· · ·B
moiety and the increase in hydrogen-bond strength ascribed to low-barrier
phenomena [12, 13].
In the case of chemically equivalent oxygen moieties (ΔPA = 0 and ΔpKa = 0),
reduction of the donor–acceptor distance [5, 60] to dO···O ≤ 2.5 Å causes the
potential barrier separating the previously distinct equilibrium configurations to
drop well below the zero-point energy of the molecular assembly, EZPE ≫ ΔEpt ,
and ultimately gives rise to the single-well model form depicted in Figure 4.1d.
Here the ground-state nuclear probability distribution encoded in the vibrationless
(v = 0) wavefunction is peaked midway along the proton-transfer coordinate,
in keeping with the nature of above-barrier dynamics that accompany complete
delocalization of the hydron from its amphoteric donor/acceptor sites (both of
which now are designated by the common label of X). This scenario produces
the strongest variant of hydrogen bonding, as epitomized by the hydrogen biflu-
oride anion, HF−2 ≡ [F · · · H · · · F]− , where the sizable bond strength [114] of
Ehb = 192 kJ mol−1 measured under isolated-molecule (vacuum) conditions reflects
a fluorine-atom separation [115] of dF···F = 2.30 Å that is considerably less than the
2 × 1.46 = 2.92 Å sum of their van der Waals radii [116]. The electronic structure
of the resulting short- and strong-hydrogen bond reveals substantial covalent

character to be spread throughout the linear [X· · ·H· · ·X]− framework, with the
bonding of such quasi-covalent assemblies often being interpreted in terms of
three-center, four-electron interactions [40, 117, 118]. The inherent symmetry of
this binding motif suggests the normal mode for collective symmetric stretching
of the X· · ·H· · ·X moiety to entail minimal displacement of the centrosymmetric
hydrogen atom and, therefore, a protium-deuterium ratio that approaches unity,
𝜈̃ XHX ∕𝜈̃ XDX = 1; however, the corresponding antisymmetric-stretching (or proton
shuttling) vibration, 𝜈̃ XH , will be affected markedly by the anharmonicity of actual
potential-energy profiles. Indeed, the latter tend to be more flat-bottomed than
parabolic in shape (cf. Figure 4.1d), with a tenable analogy to the well-known
energy spectrum for a one-dimensional infinite square-well potential predicting
𝜈̃ XH ∕𝜈̃ XD = 2 [100, 101].
4.2.2 Relationship to Tunneling

The discussion of hydrogen bonding in Section 4.2.1 has pointed to several possible
spectroscopic signatures for elucidating the nature of these noncovalent interac-
tions, many of which have been quantified and shown to be capable of extracting
pertinent structural information, including donor–acceptor distances, dA· · ·B , and
the lengths of accompanying A–H and H· · ·B linkages. Of particular note for the
present work are the related phenomena of quantum-mechanical proton tunneling,
which impose indelible spectral fingerprints in the form of energy shifts and
splittings that can reveal the topography of the encompassing potential-energy
hypersurface. For the important case of symmetrical or quasi-symmetrical
double-minimum wells (cf. Figure 4.1b,c) in which a barrier of height ΔEpt adjoins
two low-lying equilibrium configurations of nominally identical structure along a
multidimensional reaction coordinate, such processes will lead to the characteristic
bifurcation of all rovibronic eigenstates, the magnitude of which (Δ) can be equated
to the effective rate (or inverse timescale) of the ensuing chemical transformation,
k = 𝜏 −1 = 2Δ/h, where h is the Planck constant [98]. The putative role of tunneling
as a mediator of molecular structure and dynamics long has been a topic of interest
in the realm of high-resolution spectroscopy [119–123], where the quantum-state
specificity of the typically small tunneling-induced splitting, Δ𝜐 , measured for
a given vibronic level, 𝜐, has been interpreted in terms of the collective nuclear
displacements needed to engage and pursue the reaction path [74]. The size of this
dynamical bifurcation is expected to grow exponentially as the opacity (viz, height,
width, and/or asymmetry) of the impediment decreases, highlighting the intimate
relationship that exists between the processes of hydrogen bonding and hydron
transduction in both below-barrier and above-barrier regimes. Indeed, although
canonical tunneling commonly is relegated to the below-barrier milieu where it
affords a viable mechanism for affecting events that otherwise would be classically
forbidden (e.g. electron and proton transfer), an extensive literature exists for kin-
dred dynamical tunneling phenomena that can occur above barrier crests [124–128].
From this perspective, the aforementioned continuum of hydrogen-bonding motifs
will be accompanied by a parallel variation of hydron-tunneling behavior, the

consequences of which will be reflected in the correlated patterns and symmetries
(e.g. parity with respect to the reaction coordinate) of eigenstates as well as in the
distinct probability distributions defined by their attendant eigenfunctions.
4.2.3 Ground-State Properties of Model Systems

Figure 4.2 contrasts stationary points along the proton-transfer coordinates sup-
ported by the ground electronic potential-energy surfaces (X̃ A1 ) for two constitu-
1
tional isomers having the common molecular formula of C7 H6 O2 : tropolone (TrOH)

and HFF. The depicted structures follow from full-geometry optimizations per-
formed on the isolated (gas phase) species at the CCSD(T) level of coupled-cluster
theory [129–132] through use of an augmented correlation-consistent basis
of double-zeta quality, aug - cc - pVDZ ≡ apVDZ [133, 134], with analogous
results deduced from other quantum chemical methods being found elsewhere
[71, 72, 135, 136]. Each of the targeted compounds contains a strong intramolecular
hydrogen bond (IHB) that adjoins hydroxylic (proton donating) and ketonic
(proton accepting) oxygen moieties, the presence of which serves to mediate
hydron migration between equivalent minimum-energy or equilibrium config-
urations of Cs symmetry, EQ(Cs ), by a mechanism that putatively engages a
more-symmetrical and higher-lying transition state, TS(C2v ). The inherent non-
rigidity of the nuclear framework conferred by the presence of such large-amplitude
dynamics demands that rovibronic eigenstates be classified under the encompassing
G4 molecular-symmetry group [137], where a convenient isomorphism with the
C2v point group enables the four nondegenerate irreducible representations to be
labeled as a1 , b2 , b1 , and a2 . From this perspective, the attendant proton-transfer
reaction can be equated to a single sequence of pairwise nuclear exchanges involv-
ing equivalent carbon, oxygen, and hydrogen atoms (cf. the TS(C2v ) configuration)
that correspond to an antisymmetric and in-plane displacement of b2 character.
Although the hydron-transfer coordinates described above suggest that TrOH
and HFF each should support a symmetrical double-minimum topography of
potential energy within their ground electronic state, the structural differences
that distinguish respective stationary points allude to the fact that these species
occupy disparate positions along the continuum of hydrogen bonding depicted in
Figure 4.1. Most evident in the structures of Figure 4.2 is the nearly colinear nature
of the crucial O· · ·H· · ·O moiety in HFF, which reflects the enhanced flexibility
afforded by the spatially extended hydroxymethylene (proton donating) and formyl
(proton accepting) groups mounted on adjacent carbon atoms of the cyclopentadi-
ene ring. In contrast, steric constraints imposed by the more-compact neighboring
hydroxyl (proton donating) and carbonyl (proton accepting) appendages on the
cycloheptatriene framework of TrOH give rise to notably bent O· · ·H· · ·O arrange-
ments in the corresponding EQ and TS geometries. Indeed, the latter binding motif
is quite typical for IHBs, while their strong intermolecular counterparts tend to
adopt more linear forms [28]. In keeping with the tenants of resonant-assisted
hydrogen bonding (RAHBing) [138–143], the noncovalent interactions in both
‡
a
b
TrOH (+c)
(X1A1)
EQ(Cs) TS(C2v) EQ(Cs)
‡
a
(+c)
b
HFF
(X1A1)
Figure 4.2 Predicted reaction coordinates for ground-state TrOH and HFF. The rigorously
planar proton-transfer-reaction coordinates for the X̃ 1 A1 ground electronic states of
TrOH (top panel) and HFF (bottom panel) are depicted by using relaxed stationary-point
structures (color code: red ≡ O, black ≡ C, and gray ≡ H) corresponding to the minimum
energy, EQ(C s ), and transition state, TS(C 2v ), configurations of the molecular framework,
the latter of which have inertial-axis systems superimposed on them. These results
follow from full-geometry optimizations performed at the CCSD(T)/apVDZ level of
coupled-cluster theory.
of these molecules should be strengthened synergistically by the delocalization

of π - electron density that takes place in their quasi-aromatic reaction centers;
however, the nuances of the ensuing proton-transfer dynamics, as well as the
characteristic spectroscopic signatures that they engender, can be expected to vary
markedly.
Table 4.1 contains a summary of key structural parameters and energy metrics
deduced for the ground electronic states of TrOH and HFF at complementary
levels of coupled-cluster theory. Of particular note are the donor–acceptor dis-
tances, dO· · ·O , all of which reside in the range of 2.5 ≤ dO···O ≤ 2.6 Å that has been
postulated to give rise to LBHBing phenomena [15, 112]. Based on this solitary
measure, the fact that O· · ·O separations in TrOH are roughly 0.1 Å less than their
HFF complements also would suggest the former to have stronger intermolecular
interactions than the latter; however, such misconceptions can be allayed by
recalling the distinct binding motifs adopted by these kindred species. Indeed,
tabulated values of the O· · ·H· · ·O angle, 𝜃 HOH , for the EQ (120.0∘ ) and TS (139.0∘ )
forms of TrOH notably deviate from linearity and each other, indicating significant
displacement of the “shuttling” hydron from the axis connecting oxygen centers and
the need for substantial geometrical rearrangement to take place along the reaction
coordinate. In contrast, the analogous quantities for HFF (170.0∘ and 175.0∘ ) are
comparable and closely approach the straight angle of 180.0∘ expected for a linear
hydrogen bond. When combined with the uniformly longer O—H and shorter
Table 4.1 Electronic structure predictions for ground-state potential surfaces.
TrOH(X̃ 1 A1 ) HFF(X̃ 1 A1 )
Parameter EQ(C s ) TS(C 2v ) EQ(C s ) TS(C 2v )
̊
dO···O (A) 2.528 2.315 2.602 2.416
d ̊
(A) 0.989 1.236 1.011 1.209
O−H
̊
dH···O (A) 1.844 1.236 1.600 1.209
𝜃 OHO (deg) 120.0 139.0 170.0 175.0
Q(A)̊ −0.165 0.0 −0.137 0.0
𝜈̃ OH (cm−1 ) 3582.8 1598.4i 3157.8 1351.3i
(2605.8) (1191.6i) (2315.7) (1024.5i)
𝜈̃ OH ∕𝜈̃ OD 1.375 1.341 1.364 1.319
Ehb (kJ mol−1 ) 34.8 67.8
ΔEpt (cm−1 ) 2557.0 1009.2
′
ΔEpt (cm−1 ) 1683.3 (1952.8) 131.1 (362.6)
Key structural and energy parameters predicted for the X̃ 1 A1 ground electronic states of TrOH
and HFF are presented. These results follow from CCSD(T)/apVDZ calculations performed on
fully relaxed geometries corresponding to the rigorously planar minimum-energy (equilibrium),
EQ(Cs ), and transition-state, TS(C2v ), configurations (cf. Figure 4.2) of the corresponding
proton-transfer coordinates. Tabulated structural quantities include the donor–acceptor distance
(dO· · ·O ), covalent-bond length (dO − H ), hydrogen-bond length (dH· · ·O ), and hydrogen-bond angle
(𝜃 OHO ), as well as the π - electron delocalization parameter (Q) that measures the extent of
equalization achieved between opposing single/double bonds within the reaction site. In
addition to the predicted harmonic transition wavenumbers (𝜈̃ OH with 𝜈̃ OD in parentheses) and
protium–deuterium isotope ratios (𝜈̃ OH ∕𝜈̃ OD ) for the O − H-stretching fundamental, listed
energy metrics include the hydrogen-bond strength (Ehb ) and the height of the potential barrier
( )
′
impeding hydron migration, where the latter is given both without (ΔEpt ) and with Δ Ept
vibrational zero-point corrections.
H· · ·O bond lengths that HFF stationary points display relative to their TrOH coun-
terparts, such results imply that HFF supports stronger noncovalent interactions
and more efficient proton-transfer dynamics that are subject to minimal structural
encumbrance. These assertions are reinforced by the larger red shifts in harmonic √
O–H-stretching wavenumbers, 𝜈̃ OH , and greater deviations from harmonic ( 2)
behavior in isotopic ratios, 𝜈̃ OH ∕𝜈̃ OD , predicted for HFF, where the corresponding
EQ normal mode of b2 symmetry can be correlated directly with the lone imaginary
vibrational degree of freedom that characterizes the TS configuration. Similar
conclusions can be drawn from the tabulated π - electron delocalization metrics of
Gilli et al. [138]:
Q = q1 + q2 = (rC−O − rC=O ) + (rC−C − rC=C ), (4.1)
which rely on the differences in single/double bond lengths among C—O/C=O and
C—C/C=C linkages proximate to the reaction center to afford a measure for the
extent of RAHBing. This quantity vanishes (by construction) in the limit of complete
resonance stabilization, as epitomized by the highly conjugated and symmetrical

transition states of both model compounds. The predicted ∼20% decrease in the
magnitude of Q between the EQ configurations of TrOH and HFF, thus provides
further evidence for stronger hydrogen bonding and more efficient intramolecular
dynamics in the latter species.
The noncovalent forces responsible for much of the structure and dynamics in
molecules such as TrOH and HFF can be quantified by assessing the strength of
the hydrogen bond, Ehb , which bridges donor and acceptor sites. Although several
computational protocols have been developed for extracting this key metric in the
case of intramolecular interactions (where constituents cannot be separated) [144],
the present analyses have adopted the straightforward approach of comparing
quantum chemical energies predicted for fully relaxed geometries corresponding
to the minimum-energy or equilibrium configuration, EQ(Cs ), and a higher-lying
planar rotamer, ROT(Cs ), having the hydroxyl moiety removed from the O–H· · ·O
reaction center by means of a 180∘ rotation. At the CCSD(T)/apVDZ level of
theory, this procedure leads to an intramolecular hydrogen-bond strength of
Ehb = EROT − EEQ = 67.8 kJ mol−1 for the ground electronic state of HFF, which
is nearly a factor of 2 larger than the analogous 34.8 kJ mol−1 quantity deduced
for TrOH. These results suggest HFF falls within the range of 10–20 kcal mol−1
(42–84 kJ mol−1 ) proposed by Cleland and Kreevoy [10] for the manifestation of
LBHBing phenomena but relegates TrOH to the more typical regime of modestly
strong hydrogen bonds [60].
Aside from considerations based on the geometry of the proton-transfer-reaction
center and the relative strength of associated noncovalent interactions, the nature
of hydrogen bonding can be assessed by examining the location of the zero-point
energy, EZPE , with respect to the crest of the barrier that governs hydron migration,
ΔEpt (cf. Figure 4.1). In particular, prediction of ΔEpt follows directly from the
difference in quantum chemical energies for the fully relaxed equilibrium and
transition-state configurations, ΔEpt = ETS − EEQ , with the CCSD(T)/apVDZ
results compiled in Table 4.1 revealing the purely electronic impediment
of ΔEpt = 2557.0 cm−1 for TrOH to be roughly a factor of 2.5 larger in magni-
tude than the analogous ΔEpt = 1009.2 cm−1 quantity for HFF. Vibrationally
′
corrected barrier heights, Δ Ept , which reflect the relationship between EZPE and
ΔEpt , can be estimated from harmonic force fields computed for the EQ(Cs )
and TS(C2v ) stationary points, which lead to respective zero-point energies of 𝜀EQ
and 𝜀TS where the latter has excluded the lone imaginary frequency attributed to
proton transfer such that Δ𝜀pt = 𝜀TS − 𝜀EQ < 0. Under these assumptions, the desired
′
ΔEpt metric follows from [71, 72]:
′
ΔEpt = (ETS + 𝜀TS ) − (EEQ + 𝜀EQ ) = ΔEpt + Δ𝜀pt , (4.2)
where a positive (negative) value of this quantity indicates that the vibrationless level
resides below (above) the effective barrier crest. The intrinsic dependence of vibra-
tional motion (and therefore zero-point energy) on nuclear mass also implies that
′
ΔEpt must be calculated separately for parent (TrOH and HFF) and monodeuter-
ated (TrOD and HFF-d) isotopologs, the latter of which entail replacement of the
labile proton by a more massive deuteron.
′
The vibrationally corrected ΔEpt metric estimated for the ground electronic state
of TrOH places the zero-point level 1683.3 cm−1 below the effective barrier crest,
with the reduction in the magnitude of Δ𝜀pt obtained for TrOD, leading to an even
larger impediment of 1952.8 cm−1 . Given the continuum of hydrogen-bonding
motifs in Figure 4.1, these quantum chemical predictions suggest the symmetric
double-minimum topography of the X̃ 1 A1 potential surface to be described best
by the high-barrier model form of Figure 4.1b. Evidence for this assertion can be
found in the small tunneling-induced bifurcation experimentally measured for the
̃
vibrationless (v = 0) level of TrOH [145, 146], ΔX0 (TrOH) ≈ 0.9738 cm−1 , a quantity
̃
that drops precipitously for the monodeuterated isotopolog [146], ΔX0 (TrOD) ≈
̃ ̃ ̃
0.0508 cm−1 , to yield a corresponding DKIE of ΛX0 = ΔX0 (TrOH)∕ΔX0 (TrOD) ≈ 19.2
that clearly resides in the deep (below barrier) tunneling regime [75]. Likewise, a
variety of experimental [147–152] and theoretical [153–157] efforts have explored
the specificity of hydron migration in tropolone by quantifying the energy ( splitting,
)
X̃ ̃ ̃
Δ𝜐 , induced by selective excitation of vibration 𝜐. The enhancement ΔX𝜐 > ΔX0
( )
̃ ̃
or suppression ΔX𝜐 < ΔX0 of proton-transfer rate revealed by such spectroscopic
signatures can be correlated directly to the inherent ability of the attendant nuclear
displacements to couple with or decouple from the reaction coordinate, thus
highlighting the multidimensional nature of ensuing dynamics [74, 158, 159].
′
The vibrationally corrected ΔEpt parameters of 131.1 and 362.6 cm−1 predicted
for the ground electronic states of HFF and HFF-d are dramatically smaller than
their tropolone counterparts (e.g. by more than 10-fold in the case of HFF and
TrOH), but still indicate zero-point levels that are located slightly below the
effective barrier for hydron migration between donor and acceptor sites. However,
improved single-point CCSD(T)/apVQZ energy calculations performed on EQ(Cs )
and TS(C2v ) geometries optimized at the CCSD(T)/apVDZ level of theory reduce
the purely electronic impediment modestly to 870.5 cm−1 [71, 72]. When combined
with harmonic force-field estimates of 𝜀EQ and 𝜀TS (cf. Eq. (4.2)), this refined ΔEpt
′
estimate yields ΔEpt values of −7.5 cm−1 for HFF and 223.9 cm−1 for HFF-d, where
the negative sign of the former quantity implies EZPE > ΔEpt . The holistic inter-
pretation of all available quantum chemical results thus suggests the vibrationless
(v = 0) level of the parent (or monodeuterated) isotopolog to closely approach (or
lie marginally below) the critical threshold of energy defined by the proton-transfer
barrier crest, with the corresponding symmetrical double-minimum potential sur-
face being described best by the model form shown in Figure 4.1c. As such, the X̃ 1 A1
state of HFF represents a special locus on the continuum of hydrogen-bonding
motifs that is situated at the cusp of the transition between below-barrier and
above-barrier dynamics that often has been equated with the onset of LBHBing
phenomena. In keeping with this noteworthy stature, which appears to be unique
among all previously studied neutral species, the ensuing discussion will focus on
the characteristic spectroscopic signatures recently observed for HFF and HFF-d,
including tunneling-induced bifurcations of prodigious magnitude that display
seemingly anomalous kinetic-isotope effects.
4.2.4 Excited-State Spectroscopy of 6-Hydroxy-2-Formylfulvene

Although infrared absorption provides the most direct means for probing the
vibrational landscape of isolated (gas phase) HFF molecules in their X̃ 1 A1 ground
state, the laser-spectroscopic experiments that form the crux of the present chapter
have exploited the lowest-lying excited electronic manifold of singlet spin mul-
̃ 1 B2 (π∗ π), as a robust conduit for interrogating rovibronic features and
tiplicity, A
discriminating dynamical signatures. Aside from the practical benefits associated
with the use of visible (as opposed to infrared) wavelengths, the facile ability to
implement fluorescence-detection schemes, such as those depicted in Figure 4.3,
greatly enhances sensitivity and selectively relative to their direct absorption-based
counterparts, thus enabling background-free spectral information to be mea-
sured readily under rarefied conditions. As discussed below, the nature of the
richly structured A ̃ 1 B2 − X̃ 1 A1 (π* ← π) electronic transition also affords distinct
advantages for exploring the complexity of hydrogen-bonding and proton-transfer
processes.
3 3
A 2 A 2
1 1
0 0
Relative Energy
Relative Energy
Spectrally
LIF Signal
Dispersed
Spontaneous
Emission
3 Total 3
Unresolved
Spontaneous
X
2
X 2
Emission
1 1
0 0
DF Signal
Laser-Induced Dispersed
(a) Fluorescence (b) Fluorescence
Figure 4.3 Fluorescence-based spectroscopic probes. The techniques of (a) laser-induced

fluorescence (LIF) spectroscopy and (b) dispersed fluorescence (DF) spectroscopy are
shown schematically by highlighting vibronic transitions taking place between the ground
(X) ̃ electronic states of a given absorption system (Ã − X),
̃ and excited (A) ̃ with the levels
comprising each vibrational manifold being distinguished through use of primed and
double-primed labels. Initial laser excitation and subsequent spontaneous-emission
processes are indicated by solid and wavy arrows, respectively. Typical LIF spectra are
recorded by monitoring the total fluorescence emerging from the sample as the laser
frequency is scanned to access different Ã − X̃ absorption transitions, thus probing
rovibronic structure within the electronically excited manifold (assuming only a single
ground-state level to be thermally populated). In contrast, DF spectra are acquired by
tuning the frequency of the laser source to a preselected Ã − X̃ absorption transition and
spectrally resolving the emitted fluorescence to interrogated the vibrational landscape of
the ground electronic state.
Figure 4.4 highlights fluorescence-excitation spectra spanning the lowest-lying

portion of the 430 nm A ̃ 1 B2 − X̃ 1 A1 (π* ← π) absorption system for HFF and
HFF-d molecules isolated under both ambient (T ≈ 300 K) bulk-gas and “cold”
molecular-beam conditions [71, 72]. In particular, the dramatic reduction in
effective internal temperatures attained in the latter supersonic free-jet expansion
environment greatly alleviates the spectral congestion engendered by numerous
overlapping rovibronic lines (T rot < 5 K) and strongly suppresses “hot-band”
Shift from Origin Band (cm–1)

–100 0 100 200 300 400
1.0 1
4n0
Relative HFF Signal
1
Shift (cm–1) 7n0
~125cm–1 –2 –1 0 1 2
0++ Origin
Band
0.5 0––
23319 23322
Energy (cm–1)
0.0
23200 23300 23400 23500 23600 23700
Absolute LIF Excitation Energy (cm–1)
Shift from Origin Band (cm–1)

0 100 200 300 400
1.0 1
4n0
Relative HFF-d Signal
0–– 1
7n0
~36cm–1
0++
*
0.5
0.0
23200 23300 23400 23500 23600
–1
Absolute LIF Excitation Energy (cm )
Figure 4.4 Fluorescence excitation spectra of HFF isotopologs. The initial portions of the
laser-induced fluorescence (LIF) spectra recorded for the Ã 1 B2 − X̃ 1 A1 (𝜋 * ← 𝜋) absorption
system of HFF (top panel) and HFF-d (bottom panel) are depicted, with the lower and upper
abscissa scales on each graph, respectively, denoting the
( absolute
) excitation wavenumber
and the shift in term values relative to the origin band 000 . These data were acquired
under both ambient bulk-gas (shaded trace) and “cold” supersonic-expansion (unshaded
trace) conditions, with the assignments appearing on the latter highlighting the first
members of extensive vibronic progressions built on two excited-state vibrational degrees
of freedom, 𝜈 4 (a1 ) and 𝜈 7 (b2 ). The inset in the top panel shows an expanded view of the
rotational band contour supported by the jet-cooled origin band of the parent isotopolog.
Peaks labeled with asterisks in the bottom panel are residual HFF (rather than HFF-d)
features arising from incomplete in situ isotopic substitution. Source: Vealey et al.
[72]/Reproduced with permission from American Chemical Society.
transitions originating from thermally populated rovibrational manifolds that

reside above the vibrationless (v = 0) zero-point level of the X̃ 1 A1 potential surface
(T vib ≪ 50 K). These data were acquired through use of laser-induced fluorescence
(LIF) spectroscopy, where, as depicted in Figure 4.3a, a tunable laser source of
modest resolution (Δ𝜈̃ FWHM ≈ 0.06 cm−1 ) was scanned stepwise while concurrently
monitoring the total spontaneous emission generated by the target sample. The
temporal profile of the emerging fluorescence signal was found to closely follow
the shape of the impinging optical pulse (Δ𝜏 FWHM ≈ 6 ns), thereby suggesting the
pertinent A ̃ 1 B2 excited electronic state to be relatively short-lived, 0.1 < 𝜏Ã < 1.0 ns,
where the lower bound on lifetime 𝜏Ã reflects the laser-limited widths of spectral
features observed in other higher-resolution studies [160].
Individual vibronic bands comprising the ambient bulk-gas LIF datasets in
Figure 4.4 are dominated by sharp R-band heads (formed by coinciding R(J)
rovibronic lines with rotational quantum numbers of J > 30) that are accompanied
by red-degraded manifolds of overlapping P-branch and Q-branch structure. These
profiles, which are consistent with the hybrid type - a/b character predicted for the
Ã − X̃ transition-dipole moment [71, 72], collapse under “cold” molecular-beam
conditions to yield two closely spaced peaks that straddle the location of the putative
rotationless resonance (cf. inset in top panel of Figure 4.4). The bulk-gas spectra
also reveal repeating patterns of doublets in which a weaker rotational contour
is redshifted from its more pronounced counterpart by ∼125 cm−1 for HFF and
∼36 cm−1 for HFF-d, with the disappearance of these isotopolog-dependent features
in a free-jet expansion suggesting them to originate from thermally populated levels
of the X̃ 1 A1 state. As discussed below, the magnitude of this observed spectral bifur-
| Ã X̃ |
cation, Δobs𝜐 = ||Δ𝜐 − Δ0 ||, encodes the difference in tunneling dynamics between
pertinent vibronic eigenstates within the excited (𝜐) and ground (v = 0) electronic
surfaces.
The cryogenic environment afforded by a supersonic free-jet expansion has
enabled extensive vibronic progressions built upon A ̃ 1 B2 (π∗ π) vibrational degrees
of freedom to be identified [71, 72], with Figure 4.4 highlighting the lowest-
lying members of series involving the 𝜈 4 (a1 ) chelate-ring breathing and 𝜈 7 (b2 )
chelate-ring deformation modes, which, respectively, induce symmetric and
antisymmetric in-plane displacements of the molecular framework. Likewise,
( )
the isotopolog-specific origin bands 000 for the A ̃ − X̃ absorption system can
be assigned readily to the prominent peaks appearing at 𝜈̃ 00 = 23 321.1 cm−1
for HFF and 𝜈̃ 00 = 23 211.9 cm−1 for HFF-d. The isotope shift displayed by this
adiabatic excitation process, 𝛿 𝜈̃ 00 = 𝜈̃ 00 (HFF-d) − 𝜈̃ 00 (HFF) = −109.2 cm−1 , not
only embodies the mass-dependent differences in zero-point energies between the
two electronic manifolds but also betrays the marked changes in hydron-migration
dynamics induced by replacing the labile proton with a more massive deuteron.
Although jet-cooled LIF spectra exhibit canonical patterns of vibronic structure
that lack readily discernable signatures of intramolecular dynamics [72], their
bulk-gas counterparts show the aforementioned doubling of band contours that
are affected strongly by isotopic substitution. As discussed below, these observa-
tions reflect the radically different hydrogen-bonding motifs and proton-transfer
pathways supported by the ground and excited electronic states [71]. For HFF
rovibronic transitions taking place between the zero-point levels of the X̃ 1 A1
and Ã 1 B2 potential-energy surfaces, the paired features found in each ambient
( )
(T ≈ 300 K) dataset can be attributed to the symmetric 0++ and antisymmetric
( −)
0− members of the tunneling-split origin band (cf. labels in Figure 4.4), which
originate and terminate on vibrationless (v = 0) eigenstates that have the same
(+ or −) parity with respect to the reaction coordinate. Both of these excitation
paths are permitted by b - type selection rules for the Ã − X̃ absorption system,
with Figure 4.5 illustrating the G4 symmetry characteristics for all resonances
involving allowed a1 and b2 vibrational degrees of freedom, including those arising
from the a - axis component of the predominantly b - polarized transition-dipole
moment [72]. In keeping with the experimental results presented in Section 4.2.5
and the predictions of complementary quantum chemical calculations, the depicted
a1 vibration b2 vibration a1 vibration b2 vibration

( ev
= B2 a1 = b2) ( ev
= B2 b2 = a1) ( ev
= B2 a1 = b2 ) ( ev
= B2 b2 = a1)
( ) a1 ( ) b2 ( ) a1 ( ) b2
A1B2 b2 a1 b2
(+) (+) (+) (+) a1
Relative Vibronic Energy
b b b b
b b b b
a a a a
~
~ a a a a
( ) b2 ( ) b2 ( ) a1 ( ) a1
X1A1
(+) a1 (+) a1 (+) b2 (+) b2
a1 vibration a1 vibration b2 vibration b2 vibration

( ev
= A1 a1 = a1) ( ev
= A1 a1 = a1) ( ev
= A1 b2 = b2 ) ( ev
= A 1 b2 = b2 )
Figure 4.5 Allowed Ã − X̃ vibronic transitions in HFF. Schematic energy-level diagrams

are presented for vibronic transitions taking place between the X̃ 1 A1 ground and Ã 1 B2 (𝜋 ∗ 𝜋)
excited electronic states of HFF. Although the Ã 1 B2 − X̃ 1 A1 absorption system is
predominantly b - polarized (where the b - axis transition-dipole moment transforms
according to the b2 irreducible representation of the G4 molecular-symmetry group), a
significant transition-moment component also exists along the a - axis (transforming as a1 )
and leads to additional type - a resonances. Given that the hydron-migration process
transforms as b2 , each vibronic eigenstate of indicated ev Γ symmetry is bifurcated into
features of even (+) and odd (−) parity with respect to the tunneling coordinate that bear
the total G4 symmetry labels (viz, ev Γ ⊗ a1 and ev Γ ⊗ b2 ) shown adjacent to each level, the
latter of which have been partitioned according to a1 and b2 character for allowed
vibrational degrees of freedom. Transitions pertinent to the Ã − X̃ origin band of HFF
(where the attendant vibrationless or v = 0 level of each electronic manifold transforms
as a1 ) are depicted in the leftmost panel.
energy-level diagram has assumed the magnitudes of tunneling-induced splitting

in electronically excited HFF to be substantially smaller than those in the ground
̃ ̃
state, ΔA𝜐 ≪ ΔX𝜐 , thus implying that the sizable spectral bifurcations observed in
vibronically resolved LIF spectra should be dominated by the latter (ground state)
| Ã X̃ | X̃
𝜐 = ||Δ𝜐 − Δ0 || ≈ Δ0 .
effects, Δobs
4.2.5 Ground-State Spectroscopy of 6-Hydroxy-2-Formylfulvene

To more directly characterize the structure and dynamics of hydrogen bonding
within the ground electronic state of HFF, Figure 4.6 depicts initial portions of
the dispersed fluorescence (DF) spectra measured independently for the parent
(HFF) and monodeuterated (HFF-d) isotopologs under “cold” free-jet expansion
conditions [71]. As illustrated in Figure 4.3b, these data were acquired by selectively
( )
exciting the well-defined A ̃ − X̃ origin bands 00 in each species (cf. jet-cooled
0
LIF traces in Figure 4.4) and spectrally resolving the emerging spontaneous
emission through use of a scanning 0.75 m monochromator that has an effective
bandwidth of Δ𝜈̃ FWHM ≪ 5 cm−1 . The resulting set of spectral profiles affords a
trenchant glimpse of the isotopically dependent vibrational landscapes supported
by the X̃ 1 A1 potential surface, as mediated by the transition-moment properties
and Franck–Condon factors governing the utilized A ̃ 1 B2 − X̃ 1 A1 (π* ← π) absorp-
tion/emission system. Although the abscissa scale formally specifies the negative
(or redshifted) displacement of each observed feature from the fluorescence signal
coinciding with the excitation line (which established the zero reference point),
the use of 000 pump transitions that connect vibrationless (v = 0) levels of the
pertinent electronic manifolds enables desired ground-state term energies to be
obtained directly from the absolute values of their corresponding shifts.
The most conspicuous and revealing aspect of each DF dataset shown in Figure 4.6
is the lone peak located proximate to the excitation line, which undergoes a marked
drop from ∼125 to ∼36 cm−1 when the labile hydron of HFF is replaced by a
more massive deuteron to form HFF-d. Neither the term energies nor the isotopic
dependence of these resonances can be interpreted through the use of harmonic
or perturbative-anharmonic vibrational force-field predicted for the EQ(Cs ) and
TS(C2v ) structures of HFF (cf. Figure 4.2), with analogous normal-mode analyses
performed on various tautomers and rotamers of the molecular framework (all
of which are estimated to reside ≥5000 cm−1 above the global minimum-energy
configuration of the EQ(Cs ) stationary point) discounting involvement of these
species [160]. Such glaring discrepancies between experiment and theory, in
conjunction with the results of spectroscopic measurements performed on other
prominent LIF bands appearing in Figure 4.4 [161], argue for a different explanation
of these features that take into account the unique dynamical (tunneling) signatures
expected to be engendered by the onset of LBHBing.
The lone low-lying peak found in each DF spectrum of Figure 4.6 can be
attributed most readily to rovibronic lines terminating on the antisymmetric (−)
tunneling component of the vibrationless X̃ 1 A1 electronic state, the intensities of
which are governed by the a - axis contribution to the predominantly b - polarized
v7 (b2)
Relative DF Intensity (arb)
1.0 ∆Xv7 = 120.0cm–1 v4(a1)

v0(a1)
∆Xv4 = 155.7cm–1 HFF
∆X0 = 124.8cm–1
0.0
v4(a1)
v7 (b2) HFF-d
v0(a1)
∆Xv4 = 84.3cm–1
–1.0
∆X0 = 36.4cm–1
∆Xv7 = 32.0cm–1
– 600 – 500 – 400 – 300 – 200 – 100 0

Dispersed Fluorescence (DF) Spectral Shift (cm–1)
Figure 4.6 Dispersed fluorescence spectra of HFF isotopologs. The initial portions of
dispersed-fluorescence (DF) spectra recorded for the parent (positive-going trace) and
monodeuterated (negative-going trace, inverted for clarity) isotopologs of HFF are
presented. These ̃ ̃
( ) data, which follow from selective excitation of the symmetric A − X
origin band 0++ under “cold” free-jet expansion conditions, display spectral shifts and
bifurcations that depend on the nature of the associated ground-state vibrational motion,
̃
with the ΔX𝜐 tunneling metrics measured for several features being indicated. The
designation 𝜈 0 (a1 ) refers to the vibrationless (v = 0) zero-point level of the X̃ 1 A1
potential-energy surface. Source: Vealey et al. [71]/Reproduced with permission from
transition-dipole moment predicted for the A ̃ − X̃ absorption system (cf. transition

schemes in Figure 4.5). The separation of these features from their symmetric (+)
counterparts (which coincide with the strong resonances that constitute the zero
of the abscissa scale) thus provides a direct measure for the associated (v = 0)
̃
dynamical bifurcation. The resulting ΔX0 parameters of 124.8(2) cm−1 for HFF
and 36.4(2) cm−1 for HFF-d (where parentheses denote one-standard deviation
uncertainties imposed on final significant digits) are much greater in magnitude
than typical tunneling-induced signatures, reflecting a potential-energy topography
in which the zero-point level straddles the barrier crest separating two equivalent
equilibrium configurations such that EZPE ≈ ΔEpt (cf. Figure 4.1c). In keeping
with the tenants of this LBHBing motif, attendant hydron-migration events can be
̃
expected to occur on an ultrafast timescale, with the ΔX0 metric obtained for the par-
ent isotopolog – putatively the largest zero-point bifurcation known for the ground
electronic state of a neutral molecule – suggesting an effective proton-transfer
time (or inverse proton-transfer rate) of 𝜏pt = kpt −1
≈ 133 fs. This interpretation
of HFF structure and dynamics is bolstered further by the corresponding DKIE
̃ ̃ ̃
extracted for the vibrationless
√ ground state, ΛX0 = ΔX0 (HFF)∕ΔX0 (HFF-d) = 3.43,
which bridges the 2 ≈ 1.41 and ≫7 thresholds established for the above-barrier
(O–H stretching) and below-barrier (deep tunneling) regimes [75].
The low-barrier dynamics that distinguish the X̃ 1 A1 ground electronic state of
HFF produce telltale signatures in the form of prodigious spectral splittings, the
magnitudes of which approach the typical spacings observed for small-amplitude

vibrations. This leads to a striking alteration of the associated energy landscape,
imparting characteristics reminiscent of those encountered proximate to the
barrier crest of a reactive transition-state configuration [162] that cannot be
interpreted directly by reference to the results of canonical double-harmonic
or perturbative-anharmonic treatments of nuclear displacement [105, 106]. As
such, tunneling-mediated effects (viz, energy splittings and level shifts) must
be incorporated into quantitative analyses from the onset, with the cogent
assignment of each vibrational feature being based on the average or effectively
“unbifurcated” position of its widely separated (yet equally displaced) symmetric
(+) and antisymmetric (−) components [161].
By applying the procedures outlined above to the DF spectra in Figure 4.6,
the pronounced HFF (HFF-d) peaks at 270.8 cm−1 (238.1 cm−1 ) and 571.6 cm−1
(483.2 cm−1 ) can be assigned, respectively, to rovibronic transitions terminating on
a single tunneling component of the 𝜈 4 (a1 ) and 𝜈 7 (b2 ) ground-state fundamentals
[71, 161]. Identification of the opposite-parity partner for each of these features at
426.6 cm−1 (322.4 cm−1 ) and 451.9 cm−1 (451.5 cm−1 ) was facilitated by complemen-
( )
tary DF experiments that excited the antisymmetric 0−− rather than the symmetric
( +)
0+ member of the bifurcated A ̃ − X̃ origin band under bulk-gas conditions.
The vibrational term values obtained by averaging these correlated +/− pairs and
subtracting offsets for the zero-point bifurcation, 𝜈̃ 4 = 286.3 cm−1 (262.0 cm−1 )
and 𝜈̃ 7 = 449.4 cm−1 (449.3 cm−1 ), are in reasonable accord with the unscaled
predictions of 287.4 cm−1 (282.8 cm−1 ) and 444.8 cm−1 (442.5 cm−1 ) from harmonic
CCSD(T)/apVDZ force fields. In addition, the corresponding differences in these
̃
quantities yield mode-specific tunneling metrics of ΔX𝜈4 = 155.7 cm−1 (84.3 cm−1 )
̃
and ΔX𝜈4 = 119.7 cm−1 (32.0 cm−1 ) for HFF (HFF-d) that afford a measure of
̃ ̃
hydron-migration efficiency relative to that of the vibrationless level, 𝛿𝜐X̃ = ΔX𝜐 ∕ΔX0 .
From this perspective, the collective in-plane symmetrical nuclear displacements
of 𝜈 4 (a1 ), which entail substantial modulation of the crucial O· · ·O donor–acceptor
distance, produce a notable enhancement in dynamical tunneling, 𝛿𝜈X̃4 = 1.25
(2.31), especially in the case of the monodeuterated isotopolog. In contrast, the
in-plane yet antisymmetric C=O/HO—C wagging motions that characterize 𝜈 7 (b2 )
poorly coupled to the multidimensional reaction coordinate and lead to a modest
quenching of the proton-transfer process, 𝛿𝜈X̃7 = 0.96 (0.88). Further evidence
for the radically different behavior of these modes follows from their associated
̃ ̃
DKIE parameters of ΛX𝜈4 = 1.85 and ΛX𝜈7 = 3.75, which, respectively, approach the
thresholds for above-barrier ( and below-barrier
) dynamics more closely than their
̃
vibrationless counterpart ΛX0 = 3.43 despite the nearly twofold greater increase
in internal energy accompanying the latter (𝜈 7 ) degree of freedom.
Although early infrared-absorption studies of HFF suggested the existence
of a strong IHB consistent with either an asymmetric (Cs ) or symmetric (C2v )
equilibrium structure, the only prior spectroscopic evidence for a low-barrier motif
within the ground electronic state stems from the bulk-gas microwave measure-
ments of Pickett [163], where observed intensity patterns enabled a lower limit
of ∼150 cm−1 to be placed on the location of an excited rovibrational manifold

having the appropriate symmetry (b2 ) to serve as the antisymmetric component
of the vibrationless tunneling doublet. Subsequent theoretical work by Millefiori
and Alparone [164] supported this claim by providing a semiclassical estimate
̃
of ΔX0 = 142 cm−1 based on effective one-dimensional potentials calculated at
the MP2/6-31G** and MP3/6-31G** levels of Møller–Plesset perturbation theory.
Tayyari et al. [165] extended these analyses to two dimensions by considering the
influence of both O–H-stretching and C–O–H-bending degrees of freedom, leading
to MP2/6-31G** zero-point bifurcations of 181 cm−1 for HFF and 43 cm−1 for HFF-d,
which correspond to a DKIE of 4.2. More recent instanton-based treatments of HFF
and HFF-d dynamics performed with full dimensionality by Videla and Batista
[166] have predicted ground-state tunneling splittings and isotope effects that are
in near-quantitative agreement with the vibrationless (v = 0) results highlighted in
this chapter.
4.2.6 Excited-State Properties of Model Systems

In contrast to the X̃ 1 A1 ground state of HFF, where the sizable effects of LBHBing
permeate the energy landscape, the optically connected A ̃ 1 B2 excited state does not
display readily discernable spectroscopic signatures of proton-transfer dynamics
(vide supra). This dichotomy must reflect attendant changes in the topography of
the potential-energy surface and/or the nature of the reaction coordinate, both
of which can conspire to effectively quench hydron-migration events. To better
understand the origins and consequences of such behavioral differences, it again
proves useful to compare the properties of HFF to those of its more extensively
studied constitutional isomer, tropolone (TrOH). In particular, each of these
kindred species supports a low-lying singlet excited state of b2 symmetry that gives
rise to a strongly allowed A ̃ 1 B2 − X̃ 1 A1 absorption system, with the accompanying
redistribution of charge density being indicative of a 𝜋 * ← 𝜋 electron promotion
that is centered primarily on the conjugated ring [72, 136]. Table 4.2 contrasts key
structural parameters and energy metrics predicted for stationary points within the
Ã 1 B2 (𝜋 ∗ 𝜋) manifolds of HFF and TrOH at the EOM - CCSD/apVDZ level of theory.
Although the utilized equation-of-motion ansatz [167] does not include triple
excitations, the ensuing discussion will reference these results to the analogous
CCSD(T)/apVDZ ground-state quantities compiled in Table 4.1.
The reaction coordinate for the A ̃ 1 B2 (𝜋 ∗ 𝜋) potential surface of TrOH closely
parallels the general form depicted schematically in Figure 4.2 for the ground
electronic state; however, notable alterations in the planar equilibrium, EQ(Cs ),
and transition-state, TS(C2v ), configurations allude to a much more facile hydron-
migration mechanism. In particular, the crucial donor–acceptor distance, dO· · ·O ,
for the EQ(Cs ) stationary point is predicted to undergo a substantial reduction fol-
lowing the 𝜋 * ← 𝜋 electron promotion, with concomitant elongation (shortening) of
the O—H (H· · ·O) bond being reflected in a sizable redshift of vibrational wavenum-
ber for the associated stretching fundamental, 𝜈̃ OH . This effective displacement
of the minimum-energy structure toward that of the symmetric transition state
Table 4.2 Electronic structure predictions for excited-state potential surfaces.
TrOH(Ã 1 B2 ) HFF(Ã 1 B2 )
Parameter EQ(C s ) TS(C 2v ) EQ(C s ) TS(C 2 ) 2SP(C 2v )
̊
dO···O (A) 2.458 2.319 2.668 2.406 2.427
d ̊
(A) 1.006 1.235 0.992 1.206 1.214
O−H
̊
dH···O (A) 1.713 1.235 1.688 1.206 1.214
𝜃 OHO (deg) 127.4 139.7 168.6 172.3 175.9
𝜙OCCO (deg) 0.0 0.0 0.0 24.2 0.0
̊
Q(A) −0.042 0.0 −0.095 0.0 0.0
𝜈̃ OH (cm )
−1
3153.6 1197.2i 3297.1 917.7i 1113.3i
(2291.8) (868.6i) (2402.7) (712.4i) (847.6i)
𝜈̃ OH ∕𝜈̃ OD 1.376 1.378 1.372 1.288 1.313
−1
Ehb (kJ mol ) 62.8 53.1
ΔEpt (cm−1 ) 1270.6 1368.7 296.6a)
478.6 443.2
′
ΔEpt (cm−1 ) (745.2) (746.5) 231.7a)
(231.7)a)
a) Refers to skeletal-inversion barrier formed between TS(C2 ) and 2SP(C2v ) configurations.

Key structural and energy parameters predicted for the A ̃ 1 B (𝜋 ∗ 𝜋) excited electronic states of
2
TrOH and HFF are presented. These results follow from EOM - CCSD/apVDZ calculations
performed on fully relaxed geometries corresponding to the minimum-energy (or equilibrium),
EQ(Cs ), transition-state, TS(C2v ) or TS(C2 ), and second-order saddle point, 2SP(C2v ), configura-
tions (cf. Figures 4.2 and 4.7) of the corresponding proton-transfer coordinates, where the latter is
pertinent only in the case of HFF. Tabulated structural quantities include the donor–acceptor
distance (dO· · ·O ), covalent-bond length (dO − H ), hydrogen-bond length (dH· · ·O ), hydrogen-bond
angle (𝜃 OHO ), and torsional-deformation angle (𝜙OCCO ), as well as the π - electron delocalization
parameter (Q) that measures the extent of equalization achieved between opposing single/double
bonds within the reaction site. In addition to the predicted harmonic transition wavenumbers
(𝜈̃ OH with 𝜈̃ OD in parentheses) and protium-deuterium isotope ratios (𝜈̃ OH ∕𝜈̃ OD ) for the O − H-
stretching fundamental, listed energy metrics include the hydrogen-bond strength (Ehb ) and the
height of the potential
( barrier impeding hydron migration, where the latter is given both without
)
′
(ΔEpt ) and with Δ Ept vibrational zero-point corrections. Potential-energy parameters tabulated
′
for the 2SP(C2v ) saddle-point geometry of HFF (viz, ΔEinv and Δ Einv ) refer to the skeletal-inversion
barrier that is formed with respect to the nonplanar TS(C2 ) transition-state configuration.
̃
A X̃
greatly enhances the strength of intramolecular hydrogen bonding, Ehb ∕Ehb = 1.80,
( )
Ã X̃
and markedly decreases the electronic Δ Ept ∕ΔEpt = 0.50 and vibrationally corr-
̃ ̃
ected (ΔE′ Apt ∕ΔE′ Xpt = 0.28 for TrOH) barrier heights for proton transfer. The
ramifications of these changes are evident in the tunneling-induced bifurcations
measured for the A ̃ 1 B2 zero-point levels of TrOH [146, 168], ΔÃ = 19.846 cm−1 ,
0
̃
and TrOD [169], ΔA0 = 2.241 cm−1 , which are factors of 20.3 and 44.1 larger in
magnitude than their X̃ 1 A1 counterparts. Although these metrics yield a DKIE of
̃ ̃ ̃
ΛA0 = ΔA0 (TrOH)∕ΔA0 (TrOD) = 8.86 that still places electronically excited tropolone
HFF
(A1B2)
EQ(Cs) TS(C2) EQ(Cs)
Figure 4.7 Predicted reaction coordinate for electronically excited HFF. The notably
aplanar proton-transfer-reaction coordinate for the Ã 1 B2 (𝜋 ∗ 𝜋) excited electronic state of
HFF is depicted by using relaxed stationary-point structures (color code: red ≡ O, black ≡ C,
and gray ≡ H) corresponding to the minimum energy, EQ(C s ), and transition state, TS(C 2 ),
configurations of the molecular framework. These results follow from full-geometry
optimizations performed at the EOM - CCSD/apVDZ level of coupled-cluster theory.
in the realm of deep tunneling (cf. Figure 4.1b), the readily resolvable spectral
signatures of hydron dynamics within this electronic manifold have motivated a
variety of experimental [170–173] and theoretical [153–155, 174, 175] endeavors
designed to elucidate the multidimensional nature and accompanying mode
specificity of such processes [158, 159].
In contrast to TrOH, the hydrogen-bond strength for the planar minimum-energy
configuration of A ̃ 1 B2 HFF, EQ(Cs ), is predicted to be considerably weaker than
Ã X̃
that of the ground state, Ehb ∕Ehb = 0.78, an assertion reinforced by the concomitant
increase in the donor–acceptor distance, shortening (lengthening) of the O—H
(H· · ·O) bond, and blueshifting of the 𝜈̃ OH -stretching wavenumber. Although
quasi-linearity of the O – H· · ·O moiety is maintained, the partial loss of reso-
nance stabilization (as mediated by 𝜋 - electron conjugation) that accompanies
the 𝜋 * ← 𝜋 excitation profoundly alters the nature of the proton-transfer-reaction
coordinate [72]. As illustrated in Figure 4.7, the transition-state geometry, TS(C2 ),
becomes decidedly nonplanar owing to substantial displacement along the torsional
angle defined by the two oxygen centers and the adjoining carbon atoms of the
cyclopentadiene ring, |𝜙OCCO | = 24.6∘ . The associated A ̃ 1 B2 proton-transfer barrier
̃
A X̃
is significantly larger than its ground-state counterpart, Δ Ept ∕ΔEpt = 1.38, with
zero-point corrections suggesting the vibrationless (v = 0) levels of HFF and HFF-d
to reside well below the barrier crest. Moreover, the symmetrical arrangement of
the planar molecular framework (viz, akin to the X̃ 1 A1 transition state depicted
in Figure 4.2) now constitutes a second-order saddle point, 2SP(C2v ), which is
situated at the apex of a skeletal-inversion barrier, ΔEinv , separating equivalent, yet
oppositely contorted, TS(C2 ) structures distinguished by the algebraic sign of 𝜙OCCO .
Ã ̃
A ̃
A
The resulting electronic impediment of Δ Einv = E2SP − ETS = 296.6 cm−1 drops
̃
modestly to an essentially isotopolog-independent value of ΔE′ Ainv = 231.7 cm−1
when vibrational (zero point) effects are taken into account.
The aplanar proton-transfer mechanism postulated for the A ̃ 1 B2 (𝜋 ∗ 𝜋) potential
surface of HFF suggests that such processes will be substantially more encum-
bered than their ground-state counterparts, with the marked increase in effective
mass incurred by heavy-atom motion of the molecular framework exacting a sizable
kinematic penalty that adversely affects the probability of any tunneling-mediated
phenomenon (e.g. the reduced mass predicted for out-of-plane skeletal deformation
of the 2SP saddle point is 7.2 amu) [161]. Aside from precipitating a dramatic
reduction in the magnitude of dynamical bifurcation, the contorted (C2 ) configu-
ration of the transition state also implies that out-of-plane nuclear displacements,
which usually are detrimental for strictly planar unimolecular transformations
(e.g. as in the case of TrOH) [156], may instead facilitate coupling into the mul-
tidimensional hydron-migration pathway. Rotational band-contour analyses of
̃
jet-cooled LIF datasets have indicated a tunneling splitting of ΔA0 ≤ 0.12 cm−1
for the vibrationless (v = 0) A ̃ 1 B2 level [72], and complementary bulk-gas stud-
ies [160] of rotation-tunneling structure for individual vibronic bands of the
Ã 1 B2 − X̃ 1 A1 absorption system performed by means of polarization-resolved
degenerate four-wave mixing spectroscopy [176] have led to qualitatively similar
conclusions. These experimental results support an enormous decrease in the
efficacy of excited-state proton transfer relative to that taking place in the X̃ 1 A1
manifold of HFF and tend to corroborate the pronounced changes in the nature
of the reaction coordinate (viz, from rigorously planar to markedly nonplanar)
predicted to accompany the 𝜋 * ← 𝜋 excitation.
4.3 Concluding Remarks

Notwithstanding their manifest importance in diverse chemical and biochemi-
cal processes, LBHBs and their attendant pathways for hydron migration have
remained elusive targets to interrogate and classify spectroscopically. To address
this issue, the concerted experimental and theoretical analyses that form the
crux of this chapter have focused on a deceptively simple model compound,
HFF, which long has been earmarked as a prototypical example for low-barrier
phenomena in neutral species despite the lack of corroborating information for
the isolated molecule and the incongruous results emerging from solution-phase
measurements [177]. Vibronically resolved fluorescence techniques based upon the
Ã 1 B2 − X̃ 1 A1 (𝜋 * ← 𝜋) absorption system have enabled HFF and its monodeuterated
isotopolog (HFF-d) to be interrogated under both ambient bulk-gas and “cold”
supersonic-expansion conditions, with the distinct vibrational landscapes revealed
for the optically connected potential-energy surfaces suggesting them to possess
radically different structural and dynamical attributes. More specifically, the X̃ 1 A1
ground electronic state was found to display spectroscopic signatures (viz, large
tunneling-induced bifurcations and anomalous kinetic-isotope effects) consistent
with the onset of LBHBing, whereby the zero-point energy resides at the cusp of the
above-barrier and below-barrier regimes. In contrast, jet-cooled spectra acquired for
the electronically excited A ̃ 1 B2 (𝜋 ∗ 𝜋) species show conventional vibronic progres-
sions that do not betray readily discernable signs of intramolecular dynamics. The
origins of such disparate behavior were traced to profound differences in the nature
of the respective reaction coordinates, with the rigorously planar proton-transfer
mechanism of the ground state becoming notably aplanar in the excited state (thus
leading to substantial kinematic penalties from required heavy-atom motion) due
References 109
to the partial loss of 𝜋 - electron resonance stabilization incurred by the 𝜋 * ← 𝜋

excitation process. Direct comparison of structural parameters and energy metrics
predicted for the analogous X̃ 1 A1 and Ã 1 B2 (𝜋 ∗ 𝜋) manifolds of HFF and tropolone
(TrOH) highlighted the distinct potential-energy topographies supported by these
constitutional isomers.
The HFF studies discussed in this chapter have provided the first laser-spectro-
scopic evidence for LBHBing in an isolated neutral molecule; however, key
questions regarding the nature and resilience of such binding motifs remain to
be answered. For example, theoretical analyses of diverse model compounds have
suggested the presence of short and strong hydrogen bonds reflective of low-barrier
potential surfaces, with experimental corroboration emerging from X-ray/neutron
diffraction of crystalline lattices (that reveal near-symmetrical A· · ·H· · ·B struc-
tures) and 11 H-NMR chemical shifts of solvated species (that show labile protons to
be notably deshielded). In contrast, NMR isotopic-perturbation techniques applied
in both protic and aprotic media [11, 30–32, 178] have pointed to the involvement
of rapidly equilibrating mixtures of asymmetric tautomers that would seem to be
more in keeping with expectations of the high-barrier limit. This behavior has
been rationalized through the local disorder imparted by transient solvation events
[179, 180], which render the otherwise-degenerate donor and acceptor moieties
instantaneously inequivalent and lead to the formation of short-lived “solvatomers”
(i.e. differentially solvated isomers). More importantly, the prevalence of such
solution-phase findings, which include HFF [177] as well as a host of dicarboxylate
anions [181–184] and enolic diketones [185, 186] that are believed to support strong
intramolecular interactions, has been claimed to prove that LBHBs offer neither
configurational nor energetic benefits, thereby discounting their postulated ability
to mediate chemical/biochemical transformations. Although the interpretation
of these measurements in terms of tautomeric equilibria has been challenged
[187] and subsequently rebutted, [188, 189] it is clear that further spectroscopic
work is needed to characterize the intrinsic structural properties and dynamical
propensities of LBHBing, as well as the putative effects of extrinsic perturbations
sustained from the surrounding environment.
Acknowledgments
The work described in this chapter was completed under the auspices of grant
CHE-1464957 from the U.S. National Science Foundation (NSF). The authors wish
to thank The Yale Center for Research Computing for guidance and for provid-
ing computational infrastructure, specifically in the form of the shared-cluster
platforms designated as Grace and Farnam.
References
1 Müller-Dethlefs, K. and Hobza, P. (2000). Noncovalent interactions: a challenge

for experiment and theory. Chem. Rev. 100 (1): 143–167.
2 Stone, A.J. (2013). The Theory of Intermolecular Forces, 2nd edition. Oxford:
Oxford University Press.
3 Schneider, H.-J. (2009). Binding mechanisms in supramolecular complexes.
Angew. Chem. Int. Ed. 48 (22): 3924–3977.
4 Scheiner, S. (ed.) (2015). Noncovalent Forces. Cham, Germany: Springer Interna-
tional Publishing.
5 Hibbert, F. and Emsley, J. (1990). Hydrogen bonding and chemical reactivity.
Adv. Phys. Org. Chem. 26: 26255–26379.
6 Hynes, J.T., Klinman, J.P., Limbach, H.-H., and Schowen, R.L. (ed.) (2007).
Hydrogen-Transfer Reactions. Weinheim, Germany: Wiley-VCH Verlag
GmbH & Co.
7 Arunan, E., Desiraju, G.R., Klein, R.A. et al. (2011). Defining the hydro-
gen bond: an account (IUPAC Technical Report). Pure Appl. Chem. 83 (8):
1619–1636.
8 Arunan, E., Desiraju, G.R., Klein, R.A. et al. (2011). Definition of the hydrogen
bond (IUPAC Recommendations 2011). Pure Appl. Chem. 83 (8): 1637–1641.
9 Cleland, W.W. (1992). Low-barrier hydrogen bonds and low fractionation factor
bases in enzymatic reactions. Biochemistry 31 (2): 317–319.
10 Cleland, W.W. and Kreevoy, M.M. (1994). Low-barrier hydrogen-bonds and
enzymatic catalysis. Science 264 (5167): 1887–1890.
11 Perrin, C.L. (2010). Are short, low-barrier hydrogen bonds unusually strong?
Acc. Chem. Res. 43 (12): 1550–1557.
12 Emsley, J. (1980). Very strong hydrogen bonding. Chem. Soc. Rev. 9 (1): 91–124.
13 Perrin, C.L. and Nielson, J.B. (1997). ‘Strong’ hydrogen bonds in chemistry and
biology. Annu. Rev. Phys. Chem. 48: 48511–48544.
14 Gerlt, J.A., Kreevoy, M.M., Cleland, W.W., and Frey, P.A. (1997). Understanding
enzymic catalysis: the importance of short, strong hydrogen bonds. Chem. Biol.
4 (4): 259–267.
15 Cleland, W.W. (2010). The low-barrier hydrogen bond in enzymic catalysis.
Adv. Phys. Org. Chem. 44: 441–417.
16 Graham, J.D., Buytendyk, A.M., Wang, D. et al. (2014). Strong, low-barrier
hydrogen bonds may be available to enzymes. Biochemistry 53 (2): 344–349.
17 Dai, S., Funk, L.-M., Rabe von Pappenheim, F. et al. (2019). Low-barrier hydro-
gen bonds in enzyme coorperativity. Nature 573 (7775): 609–613.
18 Guthrie, J.P. and Kluger, R. (1993). Electrostatic stabilization can explain the
unexpected acidity of carbon acids in enzyme-catalyzed reactions. J. Am. Chem.
Soc. 115 (24): 11569–11572.
19 Warshel, A., Papazyan, A., Kollman, P.A. et al. (1995). On low-barrier hydrogen
bonds and enzyme catalysis. Science 269 (5220): 102–106.
20 Guthrie, J.P. (1996). Short strong hydrogen bonds: can they explain enzymic
catalysis? Chem. Biol. 3 (3): 163–170.
21 Shokri, A., Wang, Y., O’Doherty, G.A. et al. (2013). Hydrogen-bond net-
works: strengths of different types of hydrogen bonds and an alternative
to the low-barrier hydrogen-bond proposal. J. Am. Chem. Soc. 135 (47):
17919–17924.
References 111
22 Nadal-Ferret, M., Gelabert, R., Moreno, M., and Lluch, J.M. (2014). Are there
really low-barrier hydrogen bonds in proteins? The case of photoactive yellow
protein. J. Am. Chem. Soc. 136 (9): 3542–3552.
23 McAllister, M.A. (1997). Characterization of low-barrier hydrogen bonds. 3.
Hydrogen maleate. An ab initio and DFT investigation. Can. J. Chem. 75 (9):
1195–1202.
24 Chen, J., McAllister, M.A., Lee, J.K., and Houk, K.N. (1998). Short, strong
hydrogen bonds in the gas phase and in solution: theoretical exploration of pKa
matching and environmental effects on the strengths of hydrogen bonds and
their potential roles in enzymatic catalysis. J. Org. Chem. 63 (14): 4611–4619.
25 Grabowski, S.J. (2006). Theoretical studies of strong hydrogen bonds. Annu.
Rep. Prog. Chem. Sect. C: Phys. Chem. 102: 102131–102165.
26 Fillaux, F., Leygue, N., Tomkinson, J. et al. (1999). Structure and dynamics
of the symmetric hydrogen bond in potassium hydrogen maleate: a neutron
scattering study. Chem. Phys. 244 (2–3): 387–403.
27 Fillaux, F. (2000). Hydrogen bonding and quantum dynamics in the solid state.
Int. Rev. Phys. Chem. 19 (4): 553–564.
28 Steiner, T. (2002). The hydrogen bond in the solid state. Angew. Chem. Int. Ed.
41 (1): 48–76.
29 Belot, J.A., Clark, J., Cowan, J.A. et al. (2004). The shortest symmetrical O–HO
hydrogen bond has a low-barrier double-well potential. J. Phys. Chem. B
108 (22): 6933–6926.
30 Perrin, C.L. (1994). Symmetries of hydrogen bonds in solution. Science 266
(5191): 1665–1668.
31 Lau, J.S. and Perrin, C.L. (2006). Isotope effects and symmetry of hydrogen
bonds in solution: single- and double-well potential. In: Isotope Effects in Chem-
istry and Biology (ed. A. Kohen and H.-H. Limbach), 231–252. Boca Raton, FL:
CRC Press.
32 Perrin, C.L. (2009). Symmetry of hydrogen bonds in solution. Pure Appl. Chem.
81 (4): 571–583.
33 Hill, J.G. and Legon, A.C. (2015). On the directionality and non-linearity of
halogen and hydrogen bonds. Phys. Chem. Chem. Phys. 17: 17858–17867.
34 Santos-Martins, D. and Forli, S. (2020). Charting hydrogen bond anisotropy.
J. Chem. Theory Comput. 16 (4): 2846–2856.
35 Desiraju, G.R. (2011). A bond by any other name. Angew. Chem. Int. Ed. 50 (1):
52–59.
36 Hoja, J., Sax, A.F., and Szalewicz, K. (2014). Is electrostatics sufficient to
describe hydrogen-bonding interactions? Chem. Eur. J. 20 (8): 2292–2300.
37 Umeyama, H. and Morokuma, K. (1977). The origin of hydrogen bonding. An
energy decomposition study. J. Am. Chem. Soc. 99 (5): 1316–1332.
38 Tafipolsky, M. (2016). Challenging dogmas: hydrogen bond revisited. J. Phys.
Chem. A 120 (26): 4550–4559.
39 van der Lubbe, S.C.C. and Fonseca, G.C. (2019). The nature of hydrogen bonds:
a delineation of the role of different energy components on hydrogen bond
strengths and lengths. Chem. Asian J. 14 (16): 2760–2769.
40 Gilli, P., Bertolasi, V., Ferretti, V., and Gilli, G. (1994). Covalent nature of the
strong homonuclear hydrogen bond. Study of the O–H…O system by crystal
structure correlation method. J. Am. Chem. Soc. 116 (3): 909–915.
41 Grabowski, S.J. (2011). What is the covalency of hydrogen bonding? Chem. Rev.
111 (4): 2597–2625.
42 Meot-Ner, M. (2012). Update 1 of: strong ionic hydrogen bonds. Chem. Rev. 112:
PR/22–PR/103.
43 Hamilton, W.C. and Ibers, J.A. (1968). Hydrogen Bonding in Solids: Methods of
Molecular Structure Determination. New York: W. A. Benjamin.
44 Taylor, R. and Kennard, O. (1982). Crystallographic evidence for the exis-
tence of C–H· · ·O, C–H· · ·N, and C–H· · ·Cl hydrogen bonds. J. Am. Chem. Soc.
104 (19): 5063–5070.
45 Grabowski, S.J. (2020). Hydrogen bond and other Lewis acid-Lewis base
interactions as preliminary states of chemical reactions. Molecules 25 (20):
4668/1–4668/18.
46 Tuckerman, M.E., Marx, D., Klein, M.L., and Parrinello, M. (1997). On the
quantum nature of the shared proton in hydrogen bonds. Science 275 (5301):
817–820.
47 Schran, C., Marsalek, O., and Markland, T.E. (2017). Unraveling the influence
of quantum proton delocalization on electronic charge transfer through the
hydrogen bond. Chem. Phys. Lett. 678: 289–295.
48 Fang, W., Chen, J., Feng, Y. et al. (2019). The quantum nature of hydrogen.
Int. Rev. Phys. Chem. 38 (1): 35–61.
49 Schran, C. and Marx, D. (2019). Quantum nature of the hydrogen bond from
ambient conditions down to ultra-low temperatures. Phys. Chem. Chem. Phys.
21 (45): 24967–24975.
50 Kohen, A. and Limbach, H.-H. (ed.) (2006). Isotope Effects in Chemistry and
Biology. Boca Raton, FL: CRC Press.
51 Ceriotti, M. and Markland, T.E. (2013). Efficient methods and practi-
cal guidelines for simulating isotope effects. J. Chem. Phys. 138 (01):
0114112/1–0114112/13.
52 Habershon, S., Manolopoulos, D.E., Markland, T.E., and Miller, T.F. III, (2013).
Ring-polymer molecular dynamics: quantum effects in chemical dynamics from
classical trajectories in an extended phase space. Annu. Rev. Phys. Chem. 64 (1):
387–413.
53 Markland, T.E. and Ceriotti, M. (2018). Nuclear quantum effects enter the
mainstream. Nat. Rev. Chem. 2 (3): 0109/1–0109/14.
54 Curchod, B.F.E. and Martínez, T.J. (2018). Ab initio nonadiabatic quantum
molecular dynamics. Chem. Rev. 118 (7): 3305–3336.
55 Li, X.-Z., Walker, B., and Michaelides, A. (2011). Quantum nature of the hydro-
gen bond. Proc. Natl. Acad. Sci. U.S.A. 108 (16): 6369–6373.
56 Habershon, S., Markland, T.E., and Manolopoulos, D.E. (2009). Competing
quantum effects in the dynamics of a flexible water model. J. Chem. Phys.
131 (02): 024501/1–024501/11.
References 113
57 Ceriotti, M., Fang, W., Kusalik, P.G. et al. (2016). Nuclear quantum effects in
water and aqueous systems: experiment, theory, and current challenges. Chem.
Rev. 116 (13): 7529–7550.
58 Pimentel, G.C. and McClellan, A.L. (1960). The Hydrogen Bond. San Francisco:
W. H. Freeman & Co.
59 Jeffrey, G.A. and Saenger, W. (1991). Hydrogen Bonding in Biological Structures.
Berlin: Springer-Verlag.
60 Jeffrey, G.A. (1997). An Introduction to Hydrogen Bonding. New York: Oxford
University Press.
61 Scheiner, S. (1997). Hydrogen Bonding: A Theoretical Perspective, Topics in
Physical Chemistry (ed. D.G. Truhlar). Oxford: Oxford University Press.
62 Desiraju, G.R. and Steiner, T. (1999). The Weak Hydrogen Bond: In Structural
Chemistry and Biology. Oxford: Oxford University Press.
63 Gilli, G. and Gilli, P. (2009). The Nature of the Hydrogen Bond: Outline of a
Comprehensive Hydrogen Bond Theory. Oxford: Oxford University Press.
64 Grabowski, S.J. (2021). Understanding Hydrogen Bonds: Theoretical and Experi-
mental Views, vol. 19 (ed. J. Hirst). London: Royal Society of Chemistry.
65 Pimentel, G.C. and McClellan, A.L. (1971). Hydrogen bonding. Annu. Rev. Phys.
Chem. 22: 22347–22385.
66 Kollman, P.A. and Allen, L.C. (1972). The theory of the hydrogen bond. Chem.
Rev. 72 (3): 283–303.
67 Gilli, G. and Gilli, P. (2000). Towards a unified hydrogen-bond theory. J. Mol.
Struct. 552 (1–3): 1–15.
68 Buckingham, A.D., Del Bene, J.E., and McDowell, S.A.C. (2008). The hydrogen
bond. Chem. Phys. Lett. 463 (1–3): 1–10.
69 Weinhold, F. and Klein, R.A. (2012). What is a hydrogen bond? Mutually
consistent theoretical and experimental criteria for characterizing H-bonding
interactions. Mol. Phys. 110 (9–10): 565–579.
70 Herschlag, D. and Pinney, M.M. (2018). Hydrogen bonds: simple after all?
Biochemistry 57 (24): 3338–3352.
71 Vealey, Z.N., Foguel, L., and Vaccaro, P.H. (2018). Spectroscopic signatures of
proton-transfer dynamics at the cusp of low-barrier hydrogen bonding. J. Phys.
Chem. Lett. 9 (17): 4949–4954.
72 Vealey, Z.N., Foguel, L., and Vaccaro, P.H. (2019). Hydrogen-bonding
motifs and proton-transfer dynamics in electronically excited
6-hydroxy-2-formylfulvene. J. Phys. Chem. A 123 (30): 6506–6526.
73 de la Vega, J.R. (1982). Role of symmetry in the tunnelling of the proton in
double minimum potentials. Acc. Chem. Res. 15 (6): 185–191.
74 Redington, R.L. (2007). Coherent proton tunneling in hydrogen bonds of iso-
lated molecules: malonaldehyde and tropolone. In: Hydrogen-Transfer Reactions
(Chapter 1), vol. 1 (ed. J.T. Hynes, J.P. Klinman, H.-H. Limbach and R.L.
Schowen), 3–31. Weinheim, Germany: Wiley-VCH Verlag GmbH & Co.
75 Melander, L. and Saunders, W.H. Jr., (1980). Reaction Rates of Isotopic
Molecules. New York: Wiley.
76 Ubbelohde, A.R. and Gallagher, K.J. (1955). Acid-base effects in hydrogen

bonds in crystals. Acta Crystallogr. 8 (2): 71–83.
77 Gilli, P., Pretto, L., Bertolasi, V., and Gilli, G. (2009). Predicting hydrogen-bond
strengths from acid-base molecular properties. The pKa slide rule: toward the
solution of a long-lasting problem. Acc. Chem. Res. 42 (1): 33–44.
78 Desiraju, G.R. (2002). Hydrogen bridges in crystal engineering: interactions
without borders. Acc. Chem. Res. 35 (7): 565–573.
79 Steiner, T. and Desiraju, G.R. (1998). Distinction between the weak hydrogen
bond and the van der Waals interaction. Chem. Commun. 1998 (8): 891–892.
80 Smirnov, S.N., Golubev, N.S., Denisov, G.S. et al. (1996). Hydrogen/deuterium
isotope effects on the NMR chemical shifts and geometries of intermolecular
low-barrier hydrogen-bonded complexes. J. Am. Chem. Soc. 118 (17): 4094–4101.
81 Garcia-Viloca, M., Gelabert, R., González-Lafont, A. et al. (1997). Is an
extremely low-field proton signal in the NMR spectrum conclusive evidence
for a low-barrier hydrogen bond? J. Phys. Chem. A 101 (46): 8727–8733.
82 Mildvan, A.S., Harris, T.K., and Abeygunawardana, C. (1999). Nuclear magnetic
resonance methods for the detection and study of low-barrier hydrogen bonds
on enzymes. Methods Enzymol. 308: 308219–308245.
83 Sigala, P.A., Ruben, E.A., Liu, C.W. et al. (2015). Determination of hydrogen
bond structure in water versus aprotic environments to test the relationship
between length and stability. J. Am. Chem. Soc. 137 (17): 5370–5740.
84 Siskos, M.G., Choudhary, M.I., and Gerothanassis, I.P. (2017). Hydrogen atomic
positions of O–H· · ·O hydrogen bonds in solution and in the solid state: the
synergy of quantum chemical calculations with 1 H-NMR chemical shifts and
X-ray diffraction methods. Molecules 22 (3): 22030415/1–22030415/32.
85 Novak, A. (1974). Hydrogen bonding in solids. Correlation of spectroscopic
and crystallographic data. In: Large Molecules Structure and Bonding, vol.
18 (ed. J.D. Dunitz, P. Hemmerich, R.H. Holm, et al.), 177–216. New York:
Springer-Verlag.
86 Zeegers-Huyskens, T. (1990). Influence of the nature of the hydrogen bond on
the isotopic ratio 𝜈 AH /𝜈 AD . J. Mol. Struct. 217: 239–251.
87 Mielke, Z. and Sobczyk, L. (2006). Vibrational isotope effects in hydrogen
bonds. In: Isotope Effects in Chemistry and Biology (ed. A. Kohen and H.-H.
Limbach), 281–304. Boca Raton, FL: CRC Press.
88 Takahashi, M. (2014). Terahertz vibrations and hydrogen-bonded networks in
crystals. Crystals 4 (2): 74–103.
89 Bakker, D.J., Dey, A., Tabor, D.P. et al. (2017). Fingerprints of inter-
and intramolecular hydrogen bonding in saligenin-water clusters revealed
by mid- and far-infrared spectroscopy. Phys. Chem. Chem. Phys. 19 (31):
20343–20356.
90 El Khoury, Y. and Hellwig, P. (2017). Far infrared spectroscopy of hydrogen
bonding collective motions in complex molecular systems. Chem. Commun.
53 (60): 8389–8399.
References 115
91 Funke, S., Sebastiani, F., Schwaab, G., and Havenith, M. (2019). Spectroscopic
fingerprints in the low frequency spectrum of ice (Ih), clathrate hydrates, super-
cooled water, and hydrophobic hydration reveal similarities in the hydrogen
bond network motifs. J. Chem. Phys. 150 (22): 224505/1–224505/8.
92 Libowitzky, E. (1999). Correlation of O–H stretching frequencies and
O–H· · ·O hydrogen bond lengths in minerals. Monatsh. Chem. 130 (8):
1047–1059.
93 Rozenberg, M., Loewenschuss, A., and Marcus, Y. (2000). An empirical corre-
lation between stretching vibration redshift and hydrogen bond length. Phys.
Chem. Chem. Phys. 2: 22699–22702.
94 Boyer, M.A., Marsalek, O., Heindel, J.P. et al. (2019). Beyond Bager’s rule: the
origins and generality. J. Phys. Chem. Lett. 10 (5): 918–924.
95 Limbach, H.-H., Tolstoy, P.M., Pérez-Hernández, N. et al. (2009). OHO hydro-
gen bond geometries and NMR chemical shifts: from equilibrium structures to
geometric H/D isotope effects, with applications for water, protonated water,
and compressed ice. Isr. J. Chem. 49 (2): 199–126.
96 Flanigan, M.C. and de la Vega, J.R. (1974). Motion of the proton in an asym-
metric double minimum potential. J. Chem. Phys. 61 (5): 1882–1891.
97 Weiner, J.H. and Tse, S.T. (1981). Tunneling in asymmetric double-well poten-
tials. J. Chem. Phys. 74 (4): 2419–2426.
98 Hameka, H.F. and de la Vega, J.R. (1984). Intramolecular proton exchange in
near symmetric cases. J. Am. Chem. Soc. 106 (25): 7703–7705.
99 Samorjai, R.L. and Hornig, D.F. (1962). Double-minimum potentials in
hydrogen-bonded solids. J. Chem. Phys. 36 (8): 1980–1987.
100 McKenzie, R.H. (2012). A diabatic state model for donor-hydrogen vibra-
tional frequency shifts in hydrogen bonded complexes. Chem. Phys. Lett. 535:
535196–535200.
101 McKenzie, R.H., Bekker, C., Athokpam, B., and Ramesh, S.G. (2014). Effect
of quantum nuclear motion on hydrogen bonding. J. Chem. Phys. 140 (17):
174508/1–174508/13.
102 Garcia-Viloca, M., González-Lafont, A., and Lluch, J.M. (1997). On pKa match-
ing as a requirement to form a low-barrier hydrogen bond. A theoretical study
in gas phase. J. Phys. Chem. A 101 (21): 3880–3886.
103 Sørensen, J., Clausen, H.F., Poulsen, R.D. et al. (2007). Strong short hydrogen
bonds in 2-acetyl-1,8-dihydroxy-3,6-dimethylnapththalene: an outlier to current
hydrogen bonding theory? J. Phys. Chem. A 111 (2): 345–351.
104 Lu, J., Hung, I., Brinkmann, A. et al. (2017). Solid-state 17 O NMR reveals
hydrogen-bond energetics: not all low-barrier hydrogen bonds are strong.
Angew. Chem. Int. Ed. 56 (22): 6166–6170.
105 Barone, V. (2004). Vibrational zero-point energies and thermodynamic func-
tions beyond the harmonic limit. J. Chem. Phys. 120 (7): 3509–3065.
106 Barone, V. (2005). Anharmonic vibrational properties by a fully auto-
mated second-order perturbative approach. J. Chem. Phys. 122 (1):
014108/1–014108/10.
107 Dahms, F., Fingerhut, B.P., Nibbering, E.T.J. et al. (2017). Large-amplitude
transfer motion of hydrated excess protons mapped by ultrafast 2D IR spec-
troscopy. Science 357 (6350): 491–495.
108 Fournier, J.A., Carpenter, W.B., Lewis, N.H.C., and Tokmakoff, A. (2018).
Broadband 2D IR spectroscopy reveals dominant asymmetric H5 O+2 proton
hydration structures in acid solutions. Nat. Chem. 10: 10932–10937.
109 Dereka, B., Yu, Q., Lewis, N.H.C. et al. (2021). Crossover from hydrogen to
chemical bonding. Science 371 (6525): 160–164.
110 Sokolov, N.D., Vener, M.V., and Savel’ev, V.A. (1990). Tentative study of strong
hydrogen bond dynamics. Part II. Vibrational frequency considerations. J. Mol.
Struct. 222 (3–4): 365–386.
111 Tuckerman, M.E. and Marx, D. (2001). Heavy-atom skeleton quantization and
proton tunneling in “intermediate-barrier” hydrogen bonds. Phys. Rev. Lett.
86 (21): 4946–4949.
112 Frey, P.A. (2006). Isotope effects in the characterization of low barrier hydrogen
bonds. In: Isotope Effects in Chemistry and Biology (ed. A. Kohen and H.-H.
Limbach), 975–993. Boca Raton, FL: CRC Press.
113 Kong, X., Brinkmann, A., Terskikh, V. et al. (2016). Proton probability distri-
butions in the O· · ·H· · ·O low-barrier hydrogen bond: a combined solid-state
NMR and quantum chemical computational study of dibenzoylmethane and
curcumin. J. Phys. Chem. B 120 (45): 11692–11704.
114 Wenthold, P.G. and Squires, R.R. (1995). Bond dissociation energies of F−2 and
HF−2 . A gas-phase experimental and G2 theoretical study. J. Phys. Chem. 99 (7):
2002–2005.
115 Kawaguchi, K. and Hirota, E. (1986). Infrared diode laser study of the hydrogen
bifluoride anion: FHF− and FDF− . J. Chem. Phys. 84 (6): 2953–2960.
116 Alvarez, S. (2013). A cartography of the van der Waals territories. Dalton Trans.
42 (24): 8617–8636.
117 Landrum, G.A., Goldberg, N., and Hoffmann, R. (1997). Bonding in the tri-
( )
halides X−3 , mixed trihalides (X2 Y− ), and hydrogen bihalides (X2 H− ). The
connection between hypervalent, electron-rich three-center, donor-acceptor,
and strong hydrogen bonding. J. Chem. Soc., Dalton Trans. 1997 (19):
3605–3613.
118 Molina, J.M. and Dobado, J.A. (2001). The three-center-four-electron (3c-4e)
bond nature revisited. An atoms-in-molecules theory (AIM) and ELF study.
Theor. Chem. Acc. 105 (4–5): 328–337.
119 Townes, C.H. and Schawlow, A.L. (1955). Microwave Spectroscopy. New York:
McGraw-Hill.
120 Gordy, W. and Cook, R.L. (1984). Microwave Molecular Spectroscopy, 2nd
edition. New York: Wiley.
121 Saykally, R.J. and Blake, G.A. (1993). Molecular interactions and hydro-
gen bond tunneling dynamics: some new perspectives. Science 259 (5101):
1570–1575.
122 Kästner, J. and Kozuch, S. (ed.) (2021). Tunnelling in Molecules: Nuclear Quan-
tum Effects from Bio to Physical Chemistry. London: Royal Society of Chemistry.
References 117
123 Nguyen, H.V.L., Gulaczyk, I., Kre˛glewski, M., and Kleiner, I. (2021). Large
amplitude inversion tunneling motion in ammonia, methylamine, hydrazine,
and secondary amines: from structure determination to coordination chemistry.
Coord. Chem. Rev. 436: 213797/1–213797/21.
124 Heller, E.J. (1995). Dynamical tunneling and molecular spectra. J. Phys. Chem.
99 (9): 2625–2634.
125 Maitra, N.T. and Heller, E.J. (1996). Semiclassical perturbation approach to
quantum reflection. Phys. Rev. A: At. Mol. Opt. Phys. 54 (6): 4763–4769.
126 Maitra, N.T. and Heller, E.J. (1997). Barrier tunneling and reflection in the
time and energy domains: the battle of the exponentials. Phys. Rev. Lett. 78 (16):
3035–3038.
127 Heller, E.J. (1999). The many faces of tunneling. J. Phys. Chem. A 103 (49):
10433–10444.
128 Keshavamurthy, S. and Schlagheck, P. (ed.) (2011). Dynamical Tunneling:
Theory and Experiment. Boca Raton, FL: CRC Press.
129 Purvis, G.D. III, and Bartlett, R.J. (1982). A full coupled-cluster singles and
doubles model: the inclusion of disconnected triples. J. Chem. Phys. 76 (4):
1910–1918.
130 Scuseria, G.E., Janssen, C.L., and Schaefer, H.F. III, (1988). An efficient refor-
mulation of the closed-shell coupled cluster single and double excitation
(CCSD) equations. J. Chem. Phys. 89 (12): 7382–7387.
131 Crawford, T.D. and Schaefer, H.F. (2000). An introduction to coupled cluster
theory for computational chemists. In: Reviews in Computational Chemistry
(Chapter 2), vol. 14 (ed. K.B. Lipkowitz and D.B. Boyd), 33–136. New York:
VCH Publishers.
132 Bartlett, R.J. and Musiał, M. (2007). Coupled-cluster theory in quantum chem-
istry. Rev. Mod. Phys. 79 (1): 291–352.
133 Dunning, T.H. Jr., (1989). Gaussian basis sets for use in correlated molecular
calculations. I. The atoms boron through neon and hydrogen. J. Chem. Phys.
90 (2): 1007–1023.
134 Kendall, R.A., Dunning, T.H. Jr., and Harrison, R.J. (1992). Electron affinities of
the first-row atoms revisited. Systematic basis sets and wave functions. J. Chem.
Phys. 96 (9): 6796–6806.
135 Burns, L.A., Murdock, D., and Vaccaro, P.H. (2006). An exploration of elec-
tronic structure and nuclear dynamics in tropolone. I. The X̃ A1 ground state.
1
J. Chem. Phys. 124 (20): 204307/1–204307/15.

136 Burns, L.A., Murdock, D., and Vaccaro, P.H. (2009). An exploration of elec-
tronic structure and nuclear dynamics in tropolone. II. The A ̃ 1 B2 (𝜋 ∗ 𝜋) excited
state. J. Chem. Phys. 130 (14): 144304/1–144304/16.
137 Bunker, P.R. and Jensen, P. (1998). Molecular Symmetry and Spectroscopy, 2nd
edition. Ottawa: NRC Research Press.
138 Gilli, G., Bellucci, F., Ferretti, V., and Bertolasi, V. (1989). Evidence for
resonance-assisted hydrogen bonding from crystal-structure correlations on
the enol form of the β-diketone fragment. J. Am. Chem. Soc. 111 (3): 1023–1028.
139 Bertolasi, V., Gilli, P., Ferretti, V., and Gilli, G. (1991). Evidence for
resonance-assisted hydrogen bonding. 2. Intercorrelation between crystal struc-
ture and spectroscopic parameters in eight intramolecularly hydrogen bonded
1,2-diaryl-1,3-propanedione enols. J. Am. Chem. Soc. 113 (13): 4917–4925.
140 Gilli, G., Bertolasi, V., Ferretti, V., and Gilli, P. (1993). Resonance-assisted
hydrogen bonding. III. Formation of intermolecular hydrogen-bonded chains
in crystals of β-diketone enols and its relevance to molecular association. Acta
Crystallogr., Sect. B: Struct. Sci 49 (Part 3): 564–576.
141 Bertolasi, V., Gilli, P., Ferretti, V., and Gilli, G. (1997). Intramolecular O–H· · ·O
hydrogen bonds assisted by resonance. Correlation between crystallographic
data and 1 H NMR chemical shifts. J. Chem. Soc., Perkin Trans. 2 2: 945–952.
142 Bertolasi, V., Pretto, L., Gilli, G., and Gilli, P. (2006). π-Bond cooperativity and
anticooperativity effects in resonance-assisted hydrogen bonds (RAHBs). Acta
Crystallogr., Sect. B: Struct. Sci. 62 (5): 850–863.
143 Grabowski, S.J. and Leszczynski, J. (2006). Unrevealing the nature of hydrogen
bonds: π-electron delocalization shapes H-bond features. Intramolecular and
intermolecular resonance-assisted hydrogen bonds. In: Challenges and Advances
in Computational Chemistry and Physics, Hydrogen Bonding – New Insights,
vol. 3 (ed. S.J. Grabowski), 487–512. Dordrecht, The Netherlands: Springer.
144 Grabowski, S.J. (2004). Hydrogen bonding strength – measures based on geo-
metric and topological parameters. J. Phys. Org. Chem. 17 (1): 18–31.
145 Tanaka, K., Honjo, H., Tanaka, T. et al. (1999). Determination of the proton
tunneling splitting of tropolone in the ground state by microwave spectroscopy.
J. Chem. Phys. 110 (4): 1969–1978.
146 Keske, J.C., Lin, W., Pringle, W.C. et al. (2006). High-resolution studies of
tropolone in the S0 and S1 electronic states: isotope driven dynamics in the
zero-point energy levels. J. Chem. Phys. 124 (7): 074309/1–074309/12.
147 Frost, R.L., Hagemeister, F.C., Arrington, C.A. et al. (1996). Fluorescence-dip
infrared spectroscopy of tropolone and tropolone-OD. J. Chem. Phys. 105 (7):
2595–2604.
148 Redington, R.L. and Sams, R.L. (2002). State-specific spectral doublets in the
FTIR spectrum of gaseous tropolone. J. Phys. Chem. A 106 (33): 7494–7511.
149 Redington, R.L., Redington, T.E., and Sams, R.L. (2006). Quantum tunneling in
the midrange vibrational fundamentals of tropolone. J. Phys. Chem. A 110 (31):
9633–9642.
150 Redington, R.L., Redington, T.E., and Sams, R.L. (2008). Tunneling split-
tings for “O· · ·O stretching” and other vibrations of tropolone isotopomers
observed in the infrared spectrum below 800 cm−1 . J. Phys. Chem. A 112 (7):
1480–1492.
151 Murdock, D., Burns, L.A., and Vaccaro, P.H. (2007). Mode-specific tunneling
dynamics in the ground electronic state of tropolone. J. Chem. Phys. 127 (8):
081101/1–081101/5.
152 Murdock, D., Burns, L.A., and Vaccaro, P.H. (2010). Vibrational specificity of
proton-transfer dynamics in ground-state tropolone. Phys. Chem. Chem. Phys.
12 (29): 8285–8299.
References 119
153 Vener, M.V., Scheiner, S., and Sokolov, N.D. (1994). Theoretical study of hydro-
gen bonding and proton transfer in the ground and lowest excited singlet states
of tropolone. J. Chem. Phys. 101 (11): 9755–9765.
154 Takada, S. and Nakamura, H. (1995). Effects of vibrational excitation on mul-
tidimensional tunneling: general study and proton tunneling in tropolone.
J. Chem. Phys. 102 (10): 3977–3992.
155 Smedarchina, Z., Siebrand, W., and Zgierski, M.Z. (1996). Mode-specific hydro-
gen tunneling in tropolone: an instanton approach. J. Chem. Phys. 104 (4):
1203–1212.
156 Guo, Y., Sewell, T.D., and Thompson, D.L. (1998). Semiclassical calculations of
tunneling splitting in tropolone. J. Phys. Chem. A 102: 5040–5048.
157 Giese, K. and Kühn, O. (2005). The all-Cartesian reaction plane Hamiltonian:
formulation and application to the H-atom transfer in tropolone. J. Chem. Phys.
123 (5): 054315/1–054315/14.
158 Giese, K., Lahav, D., and Kühn, O. (2004). On the multidimensionality of
intramolecular hydrogen bond dynamics: hydrogen transfer and IVR in
3,7-dichlorotropolone. J. Theor. Comput. Chem. 3 (4): 567–597.
159 Giese, K., Petković, M., Naundori, H., and Kühn, O. (2006). Multidimensional
quantum dynamics and infrared spectroscopy of hydrogen bonds. Phys. Rep.
430 (4): 211–276.
160 Foguel, L. (2021). Spectroscopic probes of low-barrier proton-transfer dynamics.
PhD dissertation. Yale University, New Haven, CT.
161 Vealey, Z.N. (2019). Tunneling dynamics and hydrogen-bonding motifs in
model proton-transfer systems. PhD dissertation. Yale University, New Haven,
CT.
162 Baraban, J.H., Changala, P.B., Mellau, G.C. et al. (2015). Spectroscopic
characterization of isomerization transition states. Science 350 (6266):
1338–1342.
163 Pickett, H.M. (1973). Microwave studies on the structure and hydrogen bonding
in 6-hydroxy-2-formylfulvene. J. Am. Chem. Soc. 95 (6): 1770–1774.
164 Millefiori, S. and Alparone, A. (1994). Ab initio study of the molecular struc-
ture, polarizability and first hyperpolarizability of 6-hydroxy-1-formylfulvene.
J. Chem. Soc., Faraday Trans. 90 (19): 2873–2879.
165 Tayyari, S.F., Zahedi-Tabrizi, M., Rahemi, H. et al. (2005). A
two-dimensional potential function for bent hydrogen bonded systems. II.
6-Hydroxy-2-formylfulvene. J. Mol. Struct. THEOCHEM 730 (1): 17–21.
166 Videla, P.E. and Batista, V.S. (2021). Personal communication of work in
progress.
167 Stanton, J.F. and Bartlett, R.J. (1993). The equation of motion coupled-cluster
method. A systematic biorthogonal approach to molecular excitation ener-
gies, transition probabilities, and excited state properties. J. Chem. Phys. 98 (9):
7029–7039.
168 Bracamonte, A.E. and Vaccaro, P.H. (2003). Rotation-tunneling analysis of the
origin band in the tropolone 𝜋 * ← 𝜋 absorption system. J. Chem. Phys. 120 (10):
4638–4657.
169 Chew, K., Nemchick, D., and Vaccaro, P.H. (2013). Isotopic depen-
dence of excited-state proton-tunneling dynamics in tropolone probed by
polarization-resolved degenerate four-wave mixing spectroscopy. J. Phys.
Chem. A 117 (29): 6126–6142.
170 Tomioka, Y., Ito, M., and Mikami, N. (1983). Electronic spectra of tropolone in
a supersonic free jet. Proton tunneling in the S1 state. J. Phys. Chem. 87 (22):
4401–4405.
171 Redington, R.L., Chen, Y., Scherer, G.J., and Field, R.W. (1988). Laser fluo-
rescence excitation spectrum of jet-cooled tropolone: the A ̃ 1 B2 − X̃ 1 A1 system.
J. Chem. Phys. 88 (2): 627–633.
172 Sekiya, H., Nagashima, Y., and Nishimura, Y. (1989). Electronic spectra of
jet-cooled tropolone(-OD). Vibrational analysis for the A ̃ 1 B2 − X̃ 1 A1 transition.
Chem. Phys. Lett. 160 (5–6): 581–585.
173 Sekiya, H., Nagashima, Y., and Nishimura, Y. (1990). Electronic spectra of
jet-cooled tropolone. Effect of the vibrational excitation on the proton tunneling
dynamics. J. Chem. Phys. 92 (10): 5761–5769.
174 Wójcik, M.J., Nakamura, H., Iwata, S., and Tatara, W. (2000). Theoretical study
of multidimensional proton tunneling in the excited state of tropolone. J. Chem.
Phys. 112 (14): 6322–6328.
175 Wójcik, M.J., Boda, L., and Boczar, M. (2009). Theoretical study of pro-
ton tunneling in the excited state of tropolone. J. Chem. Phys. 130 (12):
164306/1–164306/5.
176 Bracamonte, A.E. and Vaccaro, P.H. (2003). Dissection of rovibronic band
structure by polarization-resolved degenerate four-wave mixing spectroscopy.
J. Chem. Phys. 119 (2): 887–901.
177 Perrin, C.L. and Ohta, B.K. (2002). Symmetry of O–H–O and N–H–N hydrogen
bonds in 6-hydroxy-2-formylfulvene and 6-aminofulvene-2-aldimines. Bioorg.
Chem. 30 (1): 3–15.
178 Perrin, C.L. and Ohta, B.K. (2003). Symmetries of NHN hydrogen bonds in
solution. J. Mol. Struct. 644 (1–3): 1–12.
179 Perrin, C.L. and Lau, J.S. (2006). Hydrogen-bond symmetry in zwitterionic
phthalate anions: symmetry breaking by solvation. J. Am. Chem. Soc. 128 (36):
11820–11824.
180 Dopieralski, P., Perrin, C.L., and Latajka, Z. (2011). On the intramolecular
hydrogen bond in solution: Car-Parrinello and path integral molecular dynam-
ics perspective. J. Chem. Theory Comput. 7 (11): 3505–3513.
181 Perrin, C.L. and Thoburn, J.D. (1989). Evidence for a double-minimum poten-
tial for intramolecular hydrogen bonds of aqueous hydrogen maleate and
hydrogen phthalate anions. J. Am. Chem. Soc. 111 (20): 8010–8012.
182 Perrin, C.L. and Thoburn, J.D. (1992). Symmetries of hydrogen bonds in
monoanions of dicarboxylic acids. J. Am. Chem. Soc. 114 (22): 8559–8565.
183 Perrin, C.L. and Nielson, J.B. (1997). Asymmetry of hydrogen bonds in
solutions of monoanions of dicarboxylic acids. J. Am. Chem. Soc. 119 (52):
12734–12741.
References 121
184 Perrin, C.L., Karri, P., Moore, C., and Rheingold, A.L. (2012). Hydrogen-bond
symmetry in difluoromaleate monoanion. J. Am. Chem. Soc. 134 (18):
7766–7772.
185 Perrin, C.L. and Kim, Y.-J. (1998). Symmetry of hydrogen bond in malonalde-
hyde enol in solution. J. Am. Chem. Soc. 120 (48): 12641–12645.
186 Perrin, C.L. and Wu, Y. (2019). Symmetry of hydrogen bonds in two enols in
solution. J. Am. Chem. Soc. 141 (9): 4103–4107.
187 Bogle, X.S. and Singleton, D.A. (2011). Isotope-induced desymmetrization can
mimic isotopic perturbation of equilibria. On the symmetry of bromonium ions
and hydrogen bonds. J. Am. Chem. Soc. 133 (43): 17172–17175.
188 Perrin, C.L. and Burke, K.D. (2014). Variable-temperature study of
hydrogen-bond symmetry in cyclohexene-1,2-dicarboxylate monoanion in
chloroform-d. J. Am. Chem. Soc. 136 (11): 4355–4362.
189 Perrin, C.L., Shrinidhi, A., and Burke, K.D. (2019). Isotopic-perturbation NMR
study of hydrogen-bond symmetry in solution: temperature dependence and
comparison of OHO and ODO hydrogen bonds. J. Am. Chem. Soc. 141 (43):
17278–17286.
123
Hydrogen-Bonding Interactions Using Excess Spectroscopy

Yaqian Wang and Zhiwu Yu
Tsinghua University, Department of Chemistry, Beijing 100084, China
5.1 Introduction of Hydrogen Bond

5.1.1 Definition of Hydrogen Bond
Hydrogen bond is probably the most important and extensively studied intermolecu-
lar interactions in nature. Although it had been hinted at as early as 1919 by Huggins
in his university course [1], the term “hydrogen bond” was formally proposed by
Pauling in 1931 [2]. Pauling claimed that “under certain conditions, an atom of
hydrogen is attracted by rather strong forces to two atoms, instead of only one, so
that it may be considered to be acting as a bond between them, which is called the
hydrogen bond.” He also supplemented that “hydrogen bond is formed only between
the most electronegative atoms” [3]. The widely accepted designation of hydrogen
bond is X—H· · ·Y, where X—H is a strongly polarized covalent bond, playing the
role of proton donor, and Y is the proton acceptor. Pauling also stated that Y should
contain at least one free electron pair.
Under extensive studies, substantial examples without characterized high
electronegativity atoms were found to behave similarly to hydrogen bonds [4].
For example, a carbon atom is a non-electronegative atom, but there are interac-
tions such as C—H· · ·Y, X—H· · ·C, and C—H· · ·C. These interactions nowadays
are also recognized as hydrogen bonds.
The current definition of the hydrogen bond by International Union of Pure and
Applied Chemistry (IUPAC) is: “the hydrogen bond is an attractive interaction
between a hydrogen atom from a molecule or a molecular fragment X—H in which
X is more electronegative than H, and an atom or a group of atoms in the same or a
different molecule, in which there is evidence of bond formation” [5].
The commonly used techniques/methods to study hydrogen bonds are infrared
(IR) spectroscopy, Raman spectroscopy, X-ray and neutron diffraction, proton
nuclear magnetic resonance (1 HNMR), and atomic force microscopy (AFM), as
well as quantum chemical computations [6]. After 80 years or so, many works on
the nature of hydrogen bonds have been published and numerous studies have
employed the concept to explain the molecular structures, most notoriously the
124 5 Hydrogen-Bonding Interactions Using Excess Spectroscopy
(a) (c)
(b) (d)
Figure 5.1 AFM measurements of 8-hydroxyquinoline assembled clusters on the surface

of Cu(111). (a, b) Constant height frequency shift images of typical molecule-assembled
clusters and (c, d) their corresponding structure models. The dashed lines in (c) and
(d) indicate likely hydrogen bonds. Green, carbon; blue, nitrogen; red, oxygen; white,
hydrogen. Source: Reproduced with permission from Zhang et al. [7], American Association
for the Advancement of Science – AAAS.
DNA double helix structure proposed about 70 years ago. Previous chapters of
this book must have provided many exciting new examples. Here we just want to
mention the work of Zhang et al., who reported a real-space visualization of the
formation of hydrogen bonds in 8-hydroxyquinoline molecular assemblies on a
Cu(111) substrate by employing a specially modified AFM technique (Figure 5.1).
This result is the first direct identification of hydrogen bonding interactions [7].
5.1.2 The Criteria of the Existence of Hydrogen Bonds

Hydrogen bonds exist between atoms, molecules, and ions (positive or negative)
in the gas, liquid, solid, and supercritical phases. Most hydrogen bonds are of
simple type involving one donor and one acceptor. More complicated ones are
regarded as bifurcated (three-center, one donor, and two acceptors) or trifurcated
(four-center, one donor, and three acceptors) [8]. There are also some kinds of
enhanced hydrogen bonds, including resonance assisted hydrogen bonds (RAHBs),
charge-assisted hydrogen bonds (CAHBs), and dihydrogen bonds. Specifically,
Zhang and his coworkers named the hydrogen bonds formed by the cations and
the anions of ionic liquids (ILs) (X+ —X· · ·Y− ) as Z-bonds, because they are zig-zag
shaped and coupled by electrostatic and hydrogen bonding interactions [9].
How to judge the presence of a hydrogen bond? This is a basic but not a
well-answered question. There are various necessary conditions for the hydrogen
bonding interactions, including geometrical, energetic, and spectroscopic ones.
In the case of Fourier transform infrared spectroscopy (FTIR), the formation
5.1 Introduction of Hydrogen Bond 125
Table 5.1 van der Waals radii (in Å) of selected atoms.
Pauling Bondi Bokii Zefirov Batsanov Kitaigorodsky

[3, 12] [13] [14] [15] [16, 17] [18]
H 1.10 1.20 1.17 1.16 1.20 1.17

C 1.72 1.70 1.70–1.90 1.71 1.70 1.80
N 1.50 1.55 1.57 1.50 1.60 1.58
O 1.40 1.52 1.38 1.29 1.55 1.52
F 1.35 1.47 1.35 1.40 1.50
Cl 1.80 1.75 1.80 1.90 1.80
Br 1.95 1.85 1.95 1.97 1.90
I 2.15 1.98 2.15 2.14 2.10
Si 2.10 2.10
P 1.90 1.80 1.90 1.95
S 1.85 1.80 1.80 1.84 1.80
As 2.00 1.85 2.00 2.05
Se 2.00 1.90 2.00 1.90
Te 2.20 2.06 2.20 2.10
of hydrogen bonds will cause red-shifting (e.g. O—H· · ·Y) or blue-shifting

(e.g. C—H· · ·Y) of the absorption peak of the respective X—H vibration (X=O/C).
The hydrogen bonds are referred to as proper (redshifting) and improper (blueshift-
ing) hydrogen bonds, respectively [10]. Because red and blueshifting may also occur
without the formation of hydrogen bonds, they are only necessary conditions for
the hydrogen bonding interactions.
The well-accepted sufficient conditions, however, are rare. In quantum chemical
calculations, the H· · ·Y distance and X—H· · ·Y bond angle are widely used as the
criteria to make the judgment. For the bond angle, it should be wider than 110∘ [5];
for the distance, it should be shorter than the sum of the corresponding van der
Waals (vdW) radii (of H and Y) [11]. Some selected vdW radii are listed in Table 5.1.
In X-ray crystallography, because H is unseen, the distance between X of the XH
group and Y is used to judge the presence of hydrogen bonding interactions [19, 20],
namely the distance should be shorter than a critical value. The principle is gener-
ally the same as in quantum chemical calculations. Taking hydroxyl groups as an
example, the critical O· · ·O distance should be the sum of the critical value of H· · ·O
distance in quantum chemical calculation and the O—H bond length, totaling 3.5 Å
for Pauling’s radii.
5.1.3 The Strength of Hydrogen Bonds

Hydrogen bonds are generally considered stronger than vdW interactions, while
much weaker than covalent bonds. For the convenience of discussion, Jeffrey
Table 5.2 Classification of interactions [1, 21].
Energy Distance Distance Distance

Interaction (kcal mol−1 ) (H· · ·O)/Å (H· · ·N)/Å (H· · ·Cl)/Å
vdW interaction 0.01–0.1

Hydrogen bond 1–40
Weak 1–4 2.0–2.5 2.1–2.6 2.4–2.9
Moderate 4–15 1.7–2.0 1.8–2.1 2.1–2.4
Strong 15–40 1.1–1.7 1.2–1.8 1.5–2.1
Covalent bond ∼100 0.98 1.01 1.27
classified hydrogen bonds into three categories according to the interaction ener-
gies [1]. As summarized in Table 5.2, hydrogen bonds with the interaction energies
ranging from 1–4, 4–15, and 15–40 kcal mol−1 are classified as weak, moderate, and
strong hydrogen bonds, respectively. The interaction energies are generally derived
from quantum chemical calculations. But they usually involve contributions from
factors other than hydrogen bonding interactions. This is particularly the case
when a hydrogen bond donor and/or acceptor are charged species. An alternative
approach is to classify the strength of hydrogen bonds through the H· · ·X distances,
which are listed in the third column of Table 5.2 when X is the most frequently
encountered O. Listed also in Table 5.2 are the distances when X = N, Cl proposed
in this work by simply justifying the differences of these elements in their vdW
radii. Typically, strong hydrogen bonds are characterized by short bond lengths
and are often named as short strong hydrogen bonds (SSHBs). Jeffrey claimed
that SSHBs are formed when the proton is shared between two strong bases or
between ions and molecules. It means that there should be a deficiency of electron
density in the donor group or an excess of electron density in the acceptor [20].
In addition, some hydrogen bonds with charge-transfer character can be attributed
to partial covalency, which is certainly in the category of strong hydrogen bonds [1].
It should be noted that the above classification is arbitrary, only for the convenience
of discussion.
5.2 Theory of Excess spectroscopy
Excess spectroscopy was proposed in 2003 by Yu and has found wide applications
in investigating hydrogen bonds ever since [22–25]. The concept was inspired by
excess thermodynamic functions, combining the advantages of both the sensitiv-
ity of excess thermodynamic functions and the information of spectroscopy. Excess
spectroscopy is defined as the difference between the spectrum of a real solution
and that of its corresponding ideal solution, which is similar to the definition of
excess thermodynamic functions.
5.2 Theory of Excess spectroscopy 127
0.6
1.2 1
0.4 Real 2 0.04
Absorbance (a.u.)
3 3
2 0.2 1 0.02
0.8
0 0
0.6
0.4 Ideal 2 –0.02
3 0.4
1
0.2 1 3 –0.04
2
0 0
3600 3400 3200 3000 3600 3400 3200 3000 3600 3400 3200 3000
Wavenumber Wavenumber Wavenumber
(a) (b) (c)
Figure 5.2 A schematic illustration of excess spectroscopy. In each panel, the thicker
black curve is the sum of the three sub-peaks representing three species. (a) IR band of a
pure compound. (b) The band after the pure compound is diluted with a transparent
solvent: in the ideal spectrum, the amounts of all the three associates decrease linearly
with a factor of 0.5; in the real spectrum, the decreasing factors are 0.68, 0.44, and 0.57,
respectively for the three associates. (c) Excess spectrum: the difference between the real
and the ideal spectrum. Source: Zhang et al. [23], figure 1 (p. 2)/Walter de Gruyter GmbH.
Conceptually, it can be understood as follows, here in taking FTIR as an example:

Let us assume that a broad IR band of a spectrum consists of several sub-peaks
representing various existing forms of a compound including possible monomers
and different self-associates/clusters, hereafter referred as species. By mixing with
another liquid which can be considered a perturbation operation, the amounts of
these species would change, normally in a nonlinear manner. Thus, by subtracting
the linear prediction, i.e. the ideal spectrum, we will get an excess spectrum. Most
likely, there will be positive and negative peaks in the excess spectrum. The nega-
tive peaks represent the species in the neat compound, where the amounts of them
decrease more than the linear prediction upon mixing. The positive peaks represent
either the species in the neat compound or newly formed interaction complexes,
referred still to as species. For the former, their amounts either decrease less than
the linear prediction or simply increase upon dilution. A schematic presentation
of this idea is shown in Figure 5.2. In practice, the spectrum in Figure 5.2a from
the pure compound and the real spectrum (Figure 5.2b, top) from a liquid mixture
are obtained experimentally. The ideal spectrum (Figure 5.2b, bottom) and excess
spectrum (Figure 5.2c) are from calculation.
In the theory of excess spectroscopy [25], as for FTIR, the absorbance of a pure
liquid i is defined by the form of Beer–Lambert Law,
A∗i = 𝜀∗i lCi (5.1)
where 𝜀∗i is the molar absorption coefficient (cm−1 l mol−1 ) at a specific wavenum-
ber, Ci is the molarity (mol l−1 ) of the pure compound, and l is the light pathlength
(cm). The above formula was assumed to be applicable to express the absorbance of
mixtures:
A = 𝜀lC (5.2)
where C is the total molarity of the mixed solution, and the absorption coefficient 𝜀
varies with wavelength and concentration at given temperature and pressure.
When two pure substances are mixed, the absorbance of the mixture is the summa-
tion of that of each substance. For an ideal mixture, the molar absorption coefficient
of each substance is considered a constant, not changing with the concentration.
Therefore, the absorbance of a binary solution can be written as:
A = 𝜀ideal lC = 𝜀∗1 lC1 + 𝜀∗2 lC2 (5.3)
where 𝜀ideal is the molar absorption coefficient of the mixture, C1 and C2 are the
molarities of the two components. As C = C1 + C2 , C = (n1 + n2 )/V, C1 = n1 /V,
C2 = n2 /V, V is the volume of the solution, n1 and n2 are the mole numbers of the
two components, the expression of 𝜀ideal of an ideal solution can be deduced readily.
𝜀ideal = x1 𝜀∗1 + x2 𝜀∗2 (5.4)
where x1 and x2 are mole fractions of the two components. Then excess IR absorption
spectroscopy, in the form of excess absorption coefficient, was defined as follows:
𝜀E = 𝜀 − 𝜀ideal (5.5)
It should be noted that 𝜀E is a function of wavelength (frequency), thus regarded

as an excess spectrum, at given temperature, pressure, and composition. By joining
Eqs. (5.2)–(5.5), we can get the following working equation:
A ( )
𝜀E = − x1 𝜀∗1 + x2 𝜀∗2 (5.6)
l(C1 + C2 )
In calculating the excess IR absorption spectrum of a mixture as a function
of wavelength using the equation, spectra of both neat compounds need to be
determined. Due to the high absorptivity of pure substances, light path has to be
short. This can be realized by using attenuated total reflection (ATR) spectroscopy
technique, where the light path is normally in the magnitude of micrometers. The
formula to evaluate the light path in an ATR setup is as follows [26]:
𝜆
l= [ 2 ]1∕2 (5.7)
2πn sin 𝛼 − (ns ∕n)2
where 𝜆 is the wavelength of the incident light, 𝛼 is the incident angle, and n and
ns are refractive indexes of the crystal and the sample, respectively.
The merits of excess spectroscopy include enhancing apparent spectral resolu-
tion, judging the non-ideality of mixtures, determining the selectivity of molecu-
lar interactions, identifying distinct species or clusters in solutions, and providing
information related to charge distributions in molecules. There have been many
investigations related to hydrogen bonds by employing excess spectroscopy, provid-
ing precious structural information on mixtures at molecular level [27–40].
It is worthy of clarifying several things. First of all, the excess spectrum can be
regarded as a difference spectrum, but it is a very special difference spectrum with
clear physical significance of the positive and negative peaks. Second, the idea
applies to FTIR, FT-Raman, and UV–vis, as long as the spectrum is overlapped,
actually better heavily overlapped. For FTIR, this method is widely suitable,
regardless of near-IR (0.78–2.5 μm), mid-IR (2.5–25 μm), or far-IR (25–1000 μm).
Third, the term “species” in the excess spectroscopy refers to a chemical iden-
tity that has a certain degree of stability and contributes to the appearance of
negative/positive excess peaks. It can be an associate/aggregate/cluster in liquid
mixtures and may exist in part or the entire concentration range.
Finally, to check whether the data processing is correct and the equipment
works properly, we suggest taking the binary system tert-butanol-CCl4 as the
standard system and compare the obtained excess spectra in the O—H stretching
wavenumber region with the literature results [25].
5.3 Studies of Hydrogen Bonds by Excess IR
5.3.1 Classical Hydrogen Bonds

The classical and widely known hydrogen bonds are that an H atom connected
with an electronegative atom interacts with another electronegative atom, noted
as X—H· · ·Y (X, Y = F, O, and N). These hydrogen bonds are usually very strong
and many researchers have concentrated on them. Here, we will introduce several
studies on the hydroxyl-related hydrogen bonding systems by applying excess
spectroscopy.
The first example is the hydrogen bonding interactions in tert-butanol-CCl4
system [25]. As shown in Figure 5.3, the IR spectra exhibit only two peaks, while
there are four peaks in excess IR spectra, two positive peaks, and two negative ones.
The limited number of excess peaks implies the structures of the liquid mixtures
are not totally disordered, demonstrating the microheterogeneity of the solution
80
x(CCl4)
0.0771
16
0.1491
60
εE (l mol–1 cm–1)
ε (l mol–1 cm–1)
12 0.2404
0.3357
0.4268
40 8
0.5383
0.6512
4
20 0.7621
0.8770
0
0
3750 3500 3250 3000 3750 3500 3250 3000
(a) (b)
Figure 5.3 (a) IR and (b) excess IR spectra in the O—H stretching region of
tert-butanol-CCl4 system. In (a), from bottom to top, the mole fraction of tert-butanol
increases from 0 to 1 with an increment of about 0.1. Source: Li et al. [25]/Optica
Publishing Group.
1000
εE (cm2 mol–1)
500
–500
–1000
–1500
1500 2000 2500 3000 3500
(a) Wavenumber (cm–1)
500
εE (cm2 mol–1)
–500
–1000
1500 2000 2500 3000 3500
(b) Wavenumber (cm–1)
Figure 5.4 The excess IR spectra of (a) CH3 OH-H2 O and (b) CD3 OH-H2 O systems. Source:
Tomza et al. [41], figure 4 (p. 4)/With permission of Elsevier.
structures. The nearly fixed positions of the excess peaks indicate relative stable
self-associates of the alcohol molecules through hydrogen bonding.
Binary systems containing water or alcohol are also good examples possessing
rich hydrogen bonding interactions. Tomza et al. studied the hydrogen bonding
interactions in the mixtures containing the smallest alcohol methanol and water
using excess spectroscopic technique [41]. Figure 5.4 shows the excess IR spectra of
CH3 OH-H2 O and CD3 OH-H2 O systems over quite wide wavenumber range. As can
be seen, in the O—H stretching region, there are two positive excess peaks and one
negative excess peak. Besides, they also defined a quantity to evaluate the degree of
deviation of a mixture from the respective ideal one, it is the integrated mean excess
molar absorptivity divided by the integrated mean molar absorptivity. By analysis,
the authors explained the excess peaks as that hydrogen bonding in methanol–water
mixture is weaker than that in pure alcohol and stronger than that in bulk water.
The addition of water disrupts the long-chain methanol associations. Other than
the OH-related bands, the CH and CD groups also show excess peaks, resulting
from the methyl groups subjecting interactions with water molecules.
Dimethyl sulfoxide (DMSO) is known to be a strong hydrogen bond acceptor. Li
et al. studied the role of methyl groups in the formation of O—H· · ·O=S hydrogen
bonds in DMSO–methanol mixtures [24]. As shown in Figure 5.5a, the IR spec-
tra show the increasing absorbance of the methyl rocking mode at 1115 cm−1 for
methanol and 952 cm−1 for DMSO with increasing concentration of the respective
components. In the excess IR spectra (Figure 5.5b), however, we see negative and
90 30
25
60 20
ε (l mol–1 cm–1)
15
30 10
0 0
1200 1150 1100 950 900 1200 1150 1100 950 900
(a) (b)
Figure 5.5 (a) IR and (b) excess IR spectra of the DMSO–methanol binary system. From
bottom to top, the mole fraction of DMSO increases from 0 (a) or 0.1 (b) with an increment
of about 0.1. Source: Li et al. [24], figure 1 (p. 1438)/American Chemical Society.
H3C
O
–e
H
H3C
S O
H3C
Figure 5.6 Schematic presentation of the direction of charge transfer from the methyl
groups of DMSO to the methyl group of methanol in the hydrogen-bonding complex.
Source: Li et al. [24], figure 2 (p. 1439)/American Chemical Society.
positive peaks of the vibration modes. This implies that the methyl groups in the
two molecules are subject to opposite influences upon mixing.
To unveil the mystery, the authors conducted quantum chemical calculations.
They found a negative charge loss in the DMSO methyl groups and negative
charge gain in the methanol methyl group upon the formation of DMSO–methanol
complex via hydrogen bond. This suggests a charge-donating effect of the DMSO
methyl groups and a charge-withdrawing effect of the methanol methyl group upon
the formation of a hydrogen bond between the two molecules. NMR measurements
supported such explanation, which is shown schematically in Figure 5.6. This
work reveals non-negligible influence of the secondary alkyl groups on hydrogen
bonding interactions and probably sheds light on the understanding of other more
complicated hydrogen bonds.
5.3.2 Charge Assisted Hydrogen Bonds

The CAHBs refer to the hydrogen bonds formed with the assistance of ions
or charges, exhibiting a binding strength stronger than conventional hydrogen
bonds [42, 43]. For example, [FHF]− is considered to have the strongest hydrogen
bond, due to the involvement of two most electronegative atoms and the assis-
tance of extra charge [44]. The strength of CAHBs is sometimes comparable to
that of covalent bonds. CAHBs exist widely in supramolecular systems, polymers,
biomacromolecules, and so on. Hence, they have aroused great interest of numerous
researchers.
Kalhor et al. reported a study related to CAHBs in deep eutectic solvents (DESs)
by using excess IR spectroscopy and quantum chemical calculations [45]. The DES
they selected is a mixture of choline chloride (ChCl) and ethylene glycol (EG) in
1 : 2 molar ratio, called ethaline (ETH). The pseudo-binary mixtures of ETH and
acetonitrile (CH3 CN) were investigated. Figure 5.7a,b shows the IR and excess IR
spectra in the region of 𝜐(C≡N) over the entire mole fraction range of the system.
As can be seen in Figure 5.7a, the absorbance of CH3 CN decreases monotonously
and redshifted with decreasing concentration, indicating there are interactions
between ETH and CH3 CN. The excess spectra in Figure 5.7b show a negative peak
40
Multi-state
0.5
35
Two-state x(CH3CN)
0.4 30 0.9002
0.7997
25
Absorbance
0.6995
0.3
20 0.5993
15 0.4999
0.2
0.4018
10 0.3006
0.1 5 0.2004
0.1028
0
0.0 Fixed Shifted
–5
2270 2260 2250 2240 2230 2270 2260 2250 2240 2230
(a) (b)
Excess spectrum B3LYP ETH-1CH3CN

10 Generated spectrum
2340 M06-2X
fitted peaks are colored
ETH-2CH3CN
Cal. wavenumber (cm–1)
8 CH3CN dimer
2250 cm–1 2350
6 CH3CN monomer
4 2360
2257 cm–1
2 2247 cm –1
0 2410
–2
–4 2420
–6 2254 cm–1
2430
2270 2260 2250 2240 2230 2258 2256 2254 2252 2250 2248 2246
Wavenumber (cm–1) Exp. wavenumber (cm–1)
(c) (d)
Figure 5.7 (a) IR and (b) excess IR spectra of ETH-CH3 CN system in the 𝜐(C≡N) region.
(c) Deconvolution results from the excess IR spectrum of the ETH-CH3 CN system at
equimolar composition. (d) Relationship between the averaged 𝜐(C≡N) of the deconvoluted
excess peaks and those of calculations from the respective species. The dash and
dash-dotted lines in (a) are spectra of pure CH3 CN and ETH, respectively. From top to bottom
in (a), the mole fraction of CH3 CN decreases from 1 to 0 with the exact values presented in
(b). Source: Kalhor et al. [45], figure 2 & 6 (pp. 1231–1234)/American Chemical Society.
with fixed position at 2254 cm−1 , a positive peak with fixed position at 2258 cm−1 ,
and another positive peak with non-fixed positions at around 2250 cm−1 . The latter
indicates a multistate transformation of the solution structure [46]. At least two
sub-peaks could be deconvoluted as shown in Figure 5.7c.
To assign the four excess peaks, quantum chemical calculations were used to
examine as many as possible interacting complexes. Shown in Figure 5.8 are selected
results. Among these optimized structures, we can see moderate X—H· · ·Cl− hydro-
gen bonds with the charge assisting (according to Table 5.2). Was it not due to the
negative charge, the H· · ·Cl distance would be much longer, corresponding to a
weak hydrogen bond. Four of the optimized structures were found to generate the-
oretical 𝜐(C≡N) with satisfactory linear relationship to the experimentally observed
results (Figure 5.7d). Thus, the negative excess peak was assigned to CH3 CN
dimer and the three positive excess peaks were assigned to CH3 CN monomer,
ETH-CH3 CN, and ETH-2CH3 CN. Here, ETH is discussed as an entity. This is
because Ch+ and Cl− form strong doubly ionic hydrogen bonds and, at the same
time, they also form hydrogen bonds with EG molecules. Therefore, the addition of
1.73
1.9
1.8
5
2.01
1.8
2.39
1.77
2.36
5
2.02
9
2.0
1.80
2.37 4 1.8
1.8 6
8
2 1.87
8
2.0
1
2.0
2.39
2.3
2.38 8
2.3 1.99
–394.90 kJ mol–1 –382.95 kJ mol–1 0.0 kJ mol–1 –31.08 kJ mol–1

–70.81 kJ mol–1 –100.82 kJ mol–1
–417.12 kJ mol–1 –396.52 kJ mol–1 0.0 kJ mol–1 –42.13 kJ mol–1
–103.52 kJ mol–1 –141.58 kJ mol–1
(a) (b) (c) (d) (e) (f)
2.25
1.1532
2.3
2.18
5
0
2.3
00
2.31
2.2
2.6
9 2.50 2. 44 2.7
7
1.15
2
1.1
1.97
0
2.79
5 2.31
2.2
2.6
27
2.24
2.27
2.62
2.
2.15
16 2.
2
2.
51
2. 14
26
1.1532 2.15 2.47

2
2.24 .35 2.13
0.0 kJ mol–1 –16.80 kJ mol–1 –503.56 kJ mol–1 –499.44 kJ mol–1 –538.73 kJ mol–1
0.0 kJ mol–1 –27.05 kJ mol–1 –568.64 kJ mol–1 –560.04 kJ mol–1 –625.46 kJ mol–1
(g) (h) (i) (j) (k)
2.11
1.1
2.2
34
53
2.43 8
2.
2.3
9
7
60
60
2.26 2.48
2.
2.
.58
2.48
1.1523
38
2.50
1.9
1.1539
2.4
(1)
1.153 2
2. 2. 7
5
2.8 84
2.1
72
1.1
8
2.37
23
4 4
6
2.3
2.
2.21 6 0 2 2.17
2.20 (2) 2.17 2.3 2.2 .23
6
2.3
2.30
(2) (1)
2.21
2.18
2.24
–535.96 kJ mol–1 –561.29 kJ mol–1 –559.34 kJ mol–1
–617.35 kJ mol–1 –663.91 kJ mol–1 –642.19 kJ mol–1
(l) (m) (n)
Figure 5.8 Optimized structures and interaction energies of (a, b) ChCl, (c–f) monomer to
tetramer of EG, (g, h) monomer and dimer of CH3 CN, (i, j) ETH, (k, l) ETH-CH3 CN, and
(m, n) ETH-2CH3 CN. The optimization was at the B3LYP/6-311++G(d,p) level of theory.
The interaction energies in bold were at the M06-2X/6-311++G(d,p) level of theory.
Distances between specified atoms are in the unit angstrom (Å). Source: Kalhor et al. [45],
figure 4 (p. 1233)/American Chemical Society.
CH3 CN can hardly break apart the CAHBs in ETH, and CH3 CN molecules mainly
surround the core structure of ETH. A scenario can be drawn like this: upon mixing
with ETH, the acetonitrile dimer was replaced with acetonitrile monomer and the
complexes of ETH-CH3 CN and ETH-2CH3 CN.
ILs are a family of environmentally friendly solvents with special physical
and chemical properties, thus has been attracting more and more attention [47].
For example, owing to the charge assistance, the anions (especially Cl− ) of ILs
can form strong hydrogen bonds with hydroxyl groups on the cellulose chain,
exhibiting much improved solubility of cellulose [48]. Generally, the hydrogen
bonds in ILs are doubly ionic in nature, which fall into the category of CAHBs. Xu
et al. studied the hydrogen bonding interactions and structures of a functionalized
IL, [C2 OHMIM][BF4 ], and its mixtures with acetonitrile [49]. The FTIR and the
respective excess spectra are shown in Figure 5.9. As can be seen in the IR spectra
(Figure 5.9a), with increasing x(CH3 CN), the peak centered at 3547 cm−1 redshifts,
while the shoulder peak is around 3400 cm−1 blueshifts. In the corresponding
excess IR spectra (Figure 5.9b), there are four positive peaks (3636, 3535, 3480, and
3460 cm−1 ) and a negative peak (3572 cm−1 ). Based on these results, the IR spectra
were best deconvoluted into six peaks, indicating the presence of six species.
Further, quantum chemical calculations were performed to assign the excess
peaks. All the promising self-associates and interaction complexes were examined
and their stability was judged by the interaction energies. Shown in Figure 5.10
are the selected calculation results. It was found that the calculated v(O—H) of
some optimized structures are in good linear relationship with the experimentally
observed ones, which are shown in Figure 5.9d. These structures were believed to be
the existing species in the solutions. Among the rational species, the authors found
hydrogen bonds not only formed between cations and anions but also between two
like-charged cations. This work will help understand the nature of CAHBs, which
is crucial for designing electrolytes for batteries or new reaction media.
5.3.3 Cooperative Resonance-Assisted Hydrogen Bonds

The cooperativity of hydrogen bonds was first reported by Salemme in 1982 [50] and
has been studied by many scientists ever since [51–58]. It refers to the extra energy
gain upon extension of the hydrogen bond units. Typical examples are the hydro-
gen bonding interactions (N—H· · ·O=C)n in the secondary structures of proteins
such as α-helixes and β-sheets. Gilli et al. named these kinds of hydrogen bonds as
RAHBs [54]. The idea of RAHBs has found applications in the area including the
construction of functional materials [55].
The nature of the cooperativity of hydrogen bonds has been explored for many
years, mostly by quantum calculations. There are various explanations for this
phenomenon [56]. In 1989, Gilli et al. put forward a hypothesis to relate RAHBs
to the special secondary structures of proteins [54]. In this hypothesis, the lone
pair electrons on the nitrogen atom and the π bond of the carbonyl group in a
peptide bond would resonate to an enol-like structure. There have been also other
works that attribute the cooperativity to long-range electrostatic interactions, vdW
1.2 50
3547 x(CH3CN)
0.10
v(O–H)
Increase by 0.1
40
0.8
30
Absorbance
20 0.90
Increase by 0.01
0.4
10
0 0.99
0.0
3700 3560 3420 3280 3700 3560 3420 3280
(a) (b)
0.4 3700
Observed spectrum
B2/G2
Generated spectrum
Fitted peaks
Calcd wavenumber (cm–1)
0.3 D2/D3
3600
Absorbance
3
F3/B3
0.2
2
3500
0.1 5 G1/E1
6 C3
4
3400 C2
0.0
3700 3600 3500 3400 3300 3400 3450 3500 3550 3600 3650
Wavenumber (cm–1) Obsvd wavenumber (cm–1)
(c) (d)
Figure 5.9 (a) IR and (b) excess IR spectra of [C2 OHMIM][BF4 ]-CH3 CN system in the O—H
stretching vibration region. (c) Deconvolution results of a spectrum at x(CH3 CN) = 0.5.
(d) The relationship between the deconvoluted wavenumbers and calculated
wavenumbers. From top to bottom, x(CH3 CN) increases from 0 to 1 in (a) and 0.1 to 0.99
in (b). The dash and dash-dotted lines in (a) display the spectra of pure [C2 OHMIM][BF4 ]
and CH3 CN, respectively. Source: Xu et al. [49], figure 4 & 5 (pp. 32–33)/With permission
of Elsevier.
interaction, σ-orbital interaction, electronic σ-framework rearrangement, or charge

interaction [57].
To study RAHBs, Zhou et al. took two small molecules, N-methylformamide
(NMF) and N-methylacetamide (NMA), as protein secondary structure analogs and
investigated their mixtures with deuterated dimethyl sulfoxide (DMSO-d6 ) [58].
Here DMSO-d6 was utilized rather than DMSO to avoid overlap of methyl C—H
stretching vibration peaks. The main research methods in this work are excess
IR spectroscopy and quantum chemical calculations. As shown in Figure 5.11a,c,
2.2
B2 B3 90
E1 F3
0
888
8
53 1. 30
1.9
0
2.1
1.768
2.4
2.042
2.13
5
22 2 1.859
25 2. 2.2
8
2.2 .888
17
1.9
1 2.153
00
2.0
–498.65 kJ mol–1 –493.17 kJ mol–1 –56.27 kJ mol–1 –406.91 kJ mol–1

8
2.12 G1
2.
C2 15 C3 92
2.2
2.3
0 8 75
2.30
5 1.9 2.13
50
2.2
06
1.772
68
1.7
1.99
2.297
1.8
6
1.7
2.275
1.7
1.96
24
1.721
18
50
48
9 19
2.0
29 2.19 1.846 1.9
2.2 2.0 1.824
44
2.1
40 36
2.0 2.2
73
2.373
–532.54 kJ mol–1
D2 –1286.25 kJ mol –1 –1639.52 kJ mol–1
D3 G2
1
2.19
07 0 1.79 1.869
1.8 2.3
6
2.29 2.441 8
80 2.4
7
6 8
2.1 14 2.27
8
3
1.7
2
1.2
2 2.4 2.1 2.
84
2.114
9
14 1.9 87
.2
1.901
2.24 83
1.97
90
2 4
1.9
9
2.13
9
2.
8
1.7
18
2.3
2.
3
1.8
34 5
5
2.35
4
3
00
1.7
4 5
–534.51 kJ mol–1
–1269.26 kJ mol–1 –1707.32 kJ mol–1
Figure 5.10 Optimized structures and the respective interaction energies of selected
complexes in [C2 OHMIM][BF4 ]-CH3 CN system. Source: Xu et al. [49], figure 1 & 2
(pp. 29–30)/With permission of Elsevier.
when adding DMSO-d6 to NMF, the N—H stretching IR absorption redshifts for
19.1 cm−1 . However, the N—H stretching IR absorption blueshifts 5.9 cm−1 for
NMA with introduction of DMSO-d6 . The excess IR spectra of these two systems
are shown in Figure 5.11b,d. Obviously, the excess spectra in the N—H stretching
region of NMF and NMA exhibit opposite features. For NMF, the wavenumber of
the negative excess peak (3380 cm−1 ) is higher than the positive one (3260 cm−1 )
at each concentration. While for NMA, the feature is reversed. Additionally, the
positions of both the negative and the positive peaks are fixed, demonstrating the
mixing processes resulting in the so-called “two-state” transformations between
stable complexes [46]. The authors assigned the negative peaks as NMF or NMA
associations, and the positive peaks as NMF/NMA-DMSO complexes. It is known
that the hydrogen bonds involving N—H are redshifted hydrogen bonds. Therefore,
the redshift and blueshift of N—H stretching vibration in the two binary systems
reflect the strengthening and weakening of the hydrogen bonds upon mixing.
The opposite shifting of the N—H stretching vibrations in the two systems is very
interesting. Particularly, the blueshift in the case of NMA is against the general
understanding that DMSO is capable of forming strong hydrogen bonds with proton
donors.
To investigate the nature of the opposite results, the authors investigated the
self-associations of NMF and NMA using quantum chemical calculations and
the main results are summarized in Figure 5.12. For the association of NMA
molecules, due to the steric effect of the methyl groups, the associates tend to
choose linear configurations. More importantly, the absolute value of the average
3289 NMF-DMSO-d6 NMA-DMSO-d6

3290
1.6 1.6
v(N–H)
v(N–H)
Absorbance
1.2 1.2
3066 2877 3099

2944 0.8 2944
0.8
2882
0.4 0.4
(a) 0.0 (c) 0.0
x(DMSO-d6) x(DMSO-d6)
40 0.0981 0.0996
24
0.2044 0.1992
0.3183 0.2952
30
0.4002 16 0.3996
0.5004 0.4992
20
0.5998 0.6002
8
0.6978 0.7003
10
0.8007 0.8018
0.8995
0.9001 0
0
(b) 3400 3200 3000 2800 (d) 3400 3200 3000 2800
–1
Wavenumber (cm )
Figure 5.11 The (a, c) IR and (b, d) excess IR spectra of (a, b) NMF-DMSO-d6 and
(c, d) NMA-DMSO-d6 systems in the range of N—H and C—H stretching vibration region.
The short-dashed line and dashed line in (a, c) depict the spectra of pure NMF/NMA and
DMSO-d6 , respectively. From top to bottom in (a, c) the mole fraction of DMSO-d6
increases from 0 to 1. Source: Zhou et al. [58], figure 2 (p. 2)/Springer Nature/Licensed
under CC BY 4.0.
relative hydrogen bond energies E is increasing with the prolonging of the linear
complexes. This implies the presence of RAHBs, making the energy levels in the
trimer and larger associates of NMA lower than that of NMA-DMSO complex. This
is in agreement with the blueshift of the N—H stretching vibration of NMA upon
addition of DMSO. Moreover, a closer look at the hydrogen bonding (N—H· · ·O)
distances reveals that they are all shorter than 2 Å, indicative of strong hydrogen
bonds according to Table 5.2; and the evolution of the distances from trimer to
hexamer further elaborates the details of RAHBs. In the case of NMF, there are two
conformers in its pure state, cis-NMF and trans-NMF. According to their analysis,
the cooperative effect, in this case, is not as strong as in NMA. Moreover, for both
conformers, the average hydrogen bond energy levels of NMF-DMSO complexes are
lower than that of NMA-DMSO. Particularly, the energy level of cis-NMF-DMSO
1.996
cis-trans
NMF
1.986
2 NMA
1.981
2 NMF
NMA-DMSO
–27
1.942 3 NMF
3 NMA
1.927
NMF-DMSO
4 NMF 4 NMA
1.886
1.926 5 NMA 1.946

1.939
1.886 1.926 –33 3 NMA
6 NMA 1.901
1.926 1.937
4 NMA
–36 1.921 1.880
1.882
1.934
E (kJ mol–1)
5 NMA
1.864 1.879
1.876 1.932
–39 1.918
6 NMA
cis-NMF-DMSO
–42
Figure 5.12 Calculated average relative hydrogen bond energies E and configurations
of different complexes. Dashed arrows indicate the energy change of different complexes
by adding DMSO to NMF (grey color) or NMA (black color). In the configurations, only
—N—H· · ·O=C— units are drawn in the ball-stick model, and N, H, O, and C atoms are
shown in blue, white, red, and grey balls, respectively. The value labeled in the models is
the hydrogen bond lengths in Å. All the NMF are trans conformers without specific
notification. Source: Zhou et al. [58], figure 4 (p. 4)/Springer Nature/Licensed under
CC BY 4.0.
is markedly lower than that of others (Figure 5.12), in agreement with the redshift
of the N—H stretching vibration of NMA upon addition of DMSO. To sum up,
the work reported the experimental evidence on the cooperativity of RAHBs in
NMA. It also reveals that NMA is a better model molecule for peptides/proteins
than NMF.
5.3.4 Weak/Moderate Hydrogen Bonds

Weak/moderate hydrogen bonds are nonconventional and were once accepted as
a kind of intermolecular interaction over a long period of time. With substantial
experimental and theoretical studies, scientists found such hydrogen bonds play
important roles in biological processes, self-assembly materials, and molecular
crystal engineering. Nowadays, weak hydrogen bonds are known when the proton
donor or acceptor is a non-electronegative atom such as Cl, S, and C [4]. The most
attractive one is carbon atom because it is ubiquitous in organic chemistry and life.
The hydrogen bonds formed by C—H may have the forms of C—H· · ·Y, X—H· · ·C,
C—H· · ·π, and C—H· · ·C. These interactions may also fall into the category of
moderate hydrogen bonds, particularly when they are also CAHBs at the same time,
as judged by the energy and distance criteria shown in Table 5.2. In the following,
we will describe two researches related to the weak/moderate hydrogen bonds.
The first example is ILs. Other than electrostatic interactions, hydrogen bonding
is known as the most important factor governing the physical properties of ILs and
their mixtures with molecular cosolvents [59, 60]. A great body of experimental and
theoretical investigations have revealed the important roles of hydrogen bonding
interactions in ILs [61, 62]. Zheng et al. reported a study focused on hydrogen
bonds between [Bmim][BF4 ] and acetonitrile a few years ago [63]. The IR and
excess IR spectra in v(C—H) region are shown in Figure 5.13. With the addition of
CH3 CN/CD3 CN, the shoulder peak of v(C2—H) at around 3105 cm−1 becomes more
pronounced (Figure 5.13a) and the bandwidth of v(C4,5—H) becomes broader
(Figure 5.13b). Moreover, v(C2—H) and v(C4,5—H) are redshifted for about
2.1 cm−1 , while v(alkyl C—Hs) are all blueshifted. As hydrogen bonds formed by
imidazolium ring C—H belong to redshift hydrogen bonds, while those formed by
alkyl C—H are blueshift hydrogen bonds, thus the hydrogen bonding interactions
involving the C—H in [Bmim]+ are strengthened upon mixing. It is worth stressing
that [Bmim]+ contains a charge, thus the hydrogen bonds are charge assisted. As
shown in the excess IR, we can see a position-fixed positive peak and a shoulder
peak for v(C2—H). For v(C4,5—H), there are two clear excess peaks with fixed
positions, one negative peak at higher wavenumber and a positive peak at lower
wavenumber. These excess peaks indicate the formation of distinct complexes. The
information was used to make spectral deconvolution, resulting in six sub-peaks,
three for v(C2—H), and the rest for v(C4,5—H).
By employing quantum chemical optimization, various possible species/
complexes were obtained. Following the rule that the theoretically calculated
vibrational wavenumbers of dominant species fall into a linear relationship with
the wavenumbers of experimentally observed sub-peaks, the authors identified a
few species in the solution, namely ion clusters, an ion cluster-CH3 CN complex,
an ion pair, and an ion pair-CH3 CN complex, all involving hydrogen bonds formed
by C—Hs (Figure 5.14). This study demonstrates hydrogen bonds formed by the
aromatic C—Hs of [Bmim]+ play a vital role in justifying the structures of ILs
and their mixtures with cosolvents. The underlying principles may shed light for
scientists to manipulate the properties of ILs.
Another example related to weak/moderate hydrogen bonds is the interactions
between tert-butanol and CCl4 or CHCl3 . This is the recent work from Kalhor et al.,
focusing on the so-called free O—H band (3640–3590 cm−1 ) of tert-butanol, which
is often recognized to be from the monomers of the alcohol [64]. Comparing the
spectra in Figure 5.15a,b, one can see the wavy feature of the excess spectra, showing
their higher resolution than the respective IR spectra. Accordingly, the bands were
best deconvoluted into four peaks (3622, 3616, 3610, and 3604 cm−1 ) as shown in
Figure 5.15c.
1.0 1.0
0.8 0.8 v (Alkyl C–H)

3161 2964
3122 v (C4,5–H)
2938
Absorbance
Absorbance
0.6 v (C2–H) 0.6

2877
0.4 0.4 x(CD3CN)
0.2 0.2
0.0 0.0
3150 3100 3050 3300 3200 3000 2900 2800
–1 –1
(a) Wavenumber (cm ) (b) Wavenumber (cm )
30 30
x(CH3CN) x(CD3CN)
0.1095 0.0848
0.1947
0.1597
20 20
0.3193 0.3182
0.4155 0.3902
0.5008 0.5089
10 0.6038 10 0.5998
0.6988 0.6991
0.8007 0.8032
0.9001 0.8998
0 0
3150 3100 3050 3300 3200 3000 2900 2800

(c) Wavenumber (cm–1) (d) Wavenumber (cm–1)
Figure 5.13 (a, b) IR and (c, d) excess IR spectra of (a, c) [Bmim][BF4 ]-CH3 CN and
(b, d) [Bmim][BF4 ]-CD3 CN systems in C—H stretching vibration region. From top to bottom,
the mole fraction of CH3 CN/CD3 CN increases from 0 to 1 with an increment of about 0.1
(a, b). The dashed and dashed-dotted lines represent spectra of pure [Bmim][BF4 ] and
CH3 CN/CD3 CN, respectively. Source: Zheng et al. [63], figure 2 (p. 7)/Royal Society of
Chemistry.
To assign the excess peaks, quantum chemical calculations were carried out
and the optimized structures of the monomer and dimer of the alcohol and some
complexes are shown in Figure 5.16. Among them, the calculated wavenumbers
of A (monomer), B (dimer), and complexes C, D, and F show good linearity with
the wavenumbers of the deconvoluted peaks, as demonstrated in Figure 5.15d. This
indicates that the five structures are the reasonable assignments to the O—H bond.
It is worth noting that both O and H can bond with hydrogen/halogen bond donors
and hydrogen bond acceptors, respectively. Thus, there exists quasi-free O—H,
which means a hydroxyl group that accepts but does not donate a hydrogen bond.
Accordingly, the peak at 3622 cm−1 is assigned as free O—H in the tert-butanol
monomer (Figure 5.16a). The peak at 3616 cm−1 is attributed to the quasi-free O—H
Observed spectrum 6

0.4 Generated spectrum 12 3280
Fitted peaks
Excess spectrum 5
0.3 9
Absorbance
3300 4
0.2 3 6
2 3
4 5 6
1 3320
0.1 3 2
1
0.0 0
3340
3250 3200 3150 3100 3050 3180 3160 3140 3120 3100
(a) Wavenumber (cm–1) (b) Obsd wavenumber (cm–1)
Ion cluster Peak 1 and 4
Ion cluster–CH3CN complex Peak 2 and 5
Ion pair
Peak 3 and 6
(c) Ion pair–CH3CN complex
Figure 5.14 (a) Deconvoluted IR spectrum in the aromatic C—H stretching region of
[Bmim][BF4 ]-CH3 CN system at x(CH3 CN) ≈ 0.5. (b) The relationship between the
wavenumbers of the deconvoluted peaks and the calculated wavenumbers of three
selected species/complexes. (c) The schematic display of the species in the solution.
Source: Zheng et al. [63], figure 4 & 5 (pp. 12–13)/Royal Society of Chemistry.
5
0.06
x(CCl4)
4
0.05 0.9005
Absorbance
0.7995
0.04 3 0.6997
0.03 0.6014
2 0.5008
0.02 0.3979
1 0.2996
0.01 0.1992
0.0997
0.00 0
3640 3630 3620 3610 3600 3590 3640 3630 3620 3610 3600 3590
(a) Wavenumber (cm–1) (b) Wavenumber (cm–1)
3840
t-BuOH dimer/Complex F
1.4 Excess spectrum

Generated spectrum
fitted peaks are colored

Complex I
1.2 3850
Complexes K
Complexes D/H/J/L
1.0
3860
t-BuOH monomer
0.8
3616 cm–1
3610 cm–1
0.6 3870
3604 cm–1
3622 cm–1
Complex C
0.4
3880
0.2
0.0 3890
3640 3630 3620 3610 3600 3590 3620 3610 3600 3590
(c) Wavenumber (cm–1) (d) Obsd wavenumber (cm–1)
Figure 5.15 (a) IR and (b) excess IR spectra of tert-butanol-CCl4 system in the O—H
stretching region of 3640–3590 cm−1 . (c) Deconvolution results from the excess spectrum
at equimolar composition. (d) Relationship between the wavenumbers of deconvoluted
peaks and those calculated from theoretically optimized structures. Source: Kalhor et al.
[64], figure 3 & 4 (pp. 13–15)/American Chemical Society.
2.93
9 6
1.8 2.9
2.9
6
(a) 0.0 kJ mol–1 (b) –32.67 kJ mol–1 (c) –10.52 kJ mol–1 (d) –10.18 kJ mol–1
2. 1.88
1.89
65
0 2.73
3.0 2.89
(e) –23.97 kJ mol–1 (f) –52.14 kJ mol–1 (g) –47.56 kJ mol–1
Figure 5.16 Optimized structures of (a) tert-butanol monomer, (b) tert-butanol dimer, and
(c–g) tert-butanol-CCl4 complexes with binding energies, as calculated by quantum
chemical methods. Hydrogen/halogen bonds are denoted by dashed lines in angstrom.
Source: Kalhor et al. [64], figure 2 (p. 10)/American Chemical Society.
in complex D (Figure 5.16d). For the peak 3610 cm−1 , it is quasi-free O—H in the
tert-butanol dimer (Figure 5.16b) and 2tert-butanol-CCl4 (Figure 5.16f). And the
peak at 3604 cm−1 is assigned to the hydrogen bond in complex C (Figure 5.16c).
Based on the interaction energies, the hydrogen bond O—H· · ·Cl in complex C is
and the halogen bond Cl· · ·O in complex D are weak hydrogen/halogen bonds. It is
also readily recognizable that the chlorine-involved interactions in complexes E, F,
and G are also in this category.
In addition to CCl4 , Kalhor et al. also investigated interactions between
tert-butanol and CHCl3 . Complexes H–K in Figure 5.15d are from that binary sys-
tem (data not shown). Through this work, CCl4 is determined not being a complete
inert solvent, and it can form weak hydrogen bonds with tert-butanol. Moreover,
the so-called free O—H bond include contributions from not only the monomer of
the alcohol but also the quasi-free O—H and weakly interacted O—H· · ·Cl.
References
1 Jeffrey, G.J. (1997). An Introduction to Hydrogen Bonding. New York: Oxford

University Press.
2 Pauling, L. (1931). J. Am. Chem. Soc. 53: 1367.
3 Pauling, L. (1960). The Nature of the Chemical Bond. Ithaca, NY: Cornell
University Press.
4 Desiraju, G.R. and Steiner, T. (1999). The Weak Hydrogen Bond in Structural
Chemistry and Biology. New York: Oxford University Press.
5 IUPAC (2011). IUPAC Technical Report PAC-REP-10-01-01.
6 Tuck, D.G. (1968). Prog. Inorg. Chem. 9: 161.
7 Zhang, J., Chen, P., Yuan, B. et al. (2013). Science 342: 611.
8 Aakeröy, C.B. and Seddon, K.R. (1993). Chem. Soc. Rev. 22: 397–407.
References 143
9 Dong, K., Zhang, S.J., and Wang, Q. (2015). Sci. China Chem. 58: 495.
10 Joseph, J. and Jemmis, D.E. (2007). J. Am. Chem. Soc. 129: 4620.
11 Taylor, R. and Kennard, O. (1982). J. Am. Chem. Soc. 104: 5063.
12 Pauling, L. and Pauling, P. (1975). Chemistry. San Francisco, CA: W. H. Freeman
Company.
13 Bondi, A. (1964). J. Phys. Chem. 68: 441.
14 Bokii, G.B. (1971). Kristallokhimiya (Crystal Chemistry). Moscow: Nauka.
15 Zefirov, Y.V. (2000). Russ. J. Inorg. Chem. 45: 1552.
16 Batsanov, S.S. (1991). Russ. J. Inorg. Chem. 36: 1694.
17 Batsanov, S.S. (2001). Inorg. Mater. 37: 871.
18 Kitaigorodsky, A.I. (1973). Molecular Crystals and Molecules. London: Academic
Press.
19 Matsushita, E. and Matsubara, T. (1982). Prog. Theor. Phys. 67: 1.
20 Ceccarelli, C., Jeffrey, G.A., and Taylor, R. (1981). J. Mol. Struct. 70: 255.
21 Kaplan, I.G. (1986). Theory of Molecular Interaction. Amsterdam: Elsevier.
22 Zhou, Y., Xu, J., Wang, N.N., and Yu, Z.W. (2016). Acta Phys. Chim. Sin. 32: 239.
23 Zhang, Y.Q., Wu, Z.W., Wang, Y.Q. et al. (2020). Pure Appl. Chem. 92: 1611.
24 Li, Q.Z., Wu, G.S., and Yu, Z.W. (2006). J. Am. Chem. Soc. 128: 1438.
25 Li, Q.Z., Wang, N.N., Zhou, Q. et al. (2008). Appl. Spectrosc. 62: 166.
26 Hansen, W.N. (1965). Spectrochim. Acta 21: 815.
27 Zheng, Y.Z., He, H.Y., Zhou, Y., and Yu, Z.W. (2014). J. Mol. Struct. 1069: 251.
28 Koga, Y., Sebe, F., Minami, T. et al. (2009). J. Phys. Chem. B 113: 11928.
29 Wrzeszcz, W., Tomza, P., Kwasniewicz, M. et al. (2016). RSC Adv. 6: 37195.
30 Idrissi, A., Polok, K., Marekha, B. et al. (2014). J. Phys. Chem. B 118: 1416.
31 Dong, Q., Yu, C., Li, L. et al. (2019). Spectrochim. Acta, Part A 222: 117183.
32 Zhou, Y., Zheng, Y.Z., Sun, H.Y. et al. (2014). J. Mol. Struct. 1069: 251.
33 Corsetti, S., Zehentbauer, F.M., McGloin, D., and Kiefer, J. (2015). Fuel 141: 136.
34 Wang, N.N., Li, Q.Z., and Yu, Z.W. (2009). Appl. Spectrosc. 63: 1356.
35 Chen, H., Wang, Z.H., Xu, X.Z. et al. (2021). J. Mol. Liq. 322: 114901.
36 Zhou, Y., Gong, S.D., Xu, X.Z. et al. (2020). J. Mol. Liq. 299: 112159.
37 Wallace, V.M., Dhumal, N.R., Zehentbauer, F.M. et al. (2015). J. Phys. Chem. B
119: 14780.
38 Kutsyk, A., Ilchenko, O., Pilgun, Y. et al. (2016). J. Mol. Struct. 1124: 117.
39 Zhang, Q.G., Wang, N.N., Wang, S.L., and Yu, Z.W. (2011). J. Phys. Chem. B 115:
11127.
40 Zheng, Y.Z., Zhou, Y., Deng, G., and Yu, Z.W. (2016). J. Mol. Struct. 1124: 207.
41 Tomza, P., Wrzeszcz, W., Mazurek, S. et al. (2018). Spectrochim. Acta, Part A
197: 88.
42 Braga, D. and Grepioni, F. (2000). Acc. Chem. Res. 33: 601.
43 Gilli, P., Bertolasi, V., and Ferretti, V. (1994). J. Am. Chem. Soc. 116: 909.
44 Kollman, P.A. and Allen, L.C. (1970). J. Am. Chem. Soc. 92: 6101.
45 Kalhor, P., Xu, J., Ashraf, H. et al. (2020). J. Phys. Chem. B 124: 1229.
46 Zhou, Y., Zheng, Y.Z., Sun, H.Y. et al. (2015). Sci. Rep. 5: 16379.
47 Dong, K., Liu, X.M., Dong, H.F. et al. (2017). Chem. Rev. 117: 6636.
48 Swatloski, R.P., Spear, S.K., Holbrey, J.D., and Roger, R.D. (2002). J. Am. Chem.
Soc. 124: 4974.
49 Xu, J., Deng, G., Zhou, Y. et al. (2018). J. Mol. Struct. 1161: 424.
50 Salemme, F.R. (1982). Nature 299: 754.
51 Kobko, N., Paraskevas, L., Rio, E.D., and Dannenberg, J.J. (2001). J. Am. Chem.
Soc. 123: 4348.
52 Rossmeisl, J., Nørskov, J.K., and Jacobsen, K.W. (2004). J. Am. Chem. Soc. 126:
13140.
53 Salvador, P., Kobko, N., Wieczorek, R., and Dannenberg, J.J. (2004). J. Am. Chem.
Soc. 126: 14190.
54 Gilli, G., Belluci, F., Ferretti, V., and Bertolesi, V. (1989). J. Am. Chem. Soc. 111:
1023.
55 Martí-Gastaldo, C., Antypov, D., Warren, J.E. et al. (2014). Nat. Chem. 6: 343.
56 Kobko, N. and Dannenberg, J.J. (2003). J. Phys. Chem. A 107: 10389.
57 Kurczab, R., Mitoraj, M.P., Michalak, A., and Ziegler, T. (2010). J. Phys. Chem. A
114: 8581.
58 Zhou, Y., Deng, G., Zheng, Y.Z. et al. (2016). Sci. Rep. 6: 36932.
59 Katayanagi, H., Nishikawa, K., Shimozaki, H. et al. (2004). J. Phys. Chem. B 108:
19451.
60 Fumino, K., Wulf, A., and Ludwig, R. (2008). Angew. Chem. Int. Ed. 47: 8731.
61 Zhang, Q.G., Wang, N.N., and Yu, Z.W. (2010). J. Phys. Chem. B 114: 4747.
62 Wang, N.N., Zhang, Q.G., Wu, F.G. et al. (2010). J. Phys. Chem. B 114: 8689.
63 Zheng, Y.Z., Wang, N.N., Luo, J.J. et al. (2013). Phys. Chem. Chem. Phys. 15:
18055.
64 Kalhor, P., Li, Q.Z., Zheng, Y.Z., and Yu, Z.W. (2020). J. Phys. Chem. A 124: 6177.
145
Intramolecular Hydrogen Bonding in Porphyrin Isomers

Jacek Waluk 1,2
1
Polish Academy of Sciences, Institute of Physical Chemistry, Department of Photochemistry and
Spectroscopy, Kasprzaka 44/52, 01-224 Warsaw, Poland
2
Cardinal Stefan Wyszyński University, Faculty of Mathematics and Science, Institute of Chemistry,
Dewajtis 5, 01-815 Warsaw, Poland
6.1 Introduction
Two factors that are crucial for hydrogen bond (HB) strength are (i) the donor–
acceptor distance and (ii) the angle formed by the three atoms that constitute the
HB. Intramolecular and intermolecular HBs differ significantly regarding these two
parameters. In the intermolecular case, the partners can usually assume the position
optimal from the point of view of the distance and the angle. This is not the case for
an intramolecular HB, where these values are determined by molecular geometry.
As a consequence, the same donor and acceptor can form HBs of different strengths,
depending on their location in the molecular skeleton. Comparison of the same type
of bond in structurally similar molecules can, therefore, be very useful for study-
ing the influence of distance and angle on the HB properties. One can also assess
the relative contributions of these two factors. In this respect, an important class of
such models is provided by porphyrin and its structural isomers (Scheme 6.1). These
macrocyclic compounds possess an internal cavity composed of four nitrogen atoms,
of which two – HB donors – are covalently bound to hydrogen, and the other two
possess a lone electron pair, which makes them the HB acceptors. For each isomer,
the geometry of the internal cavity has a different shape and size. The size can also
vary for different tautomeric forms. Moreover, for each particular isomer, the cavity
geometry can be modified in a controllable way by appropriate substitution. Finally,
the two HBs can interact via the aromatic framework in a cooperative or anticoop-
erative way. All these features make porphyrin isomers very attractive models for
systematic investigations of intramolecular HBs.
In this work, we will discuss the HB characteristics for all porphyrin isomers,
including those that have not been yet synthesized. Naturally, this comparison
must, therefore, be based on calculations. When possible, we will refer to the
experimental data, available for porphyrin and its two isomers, porphycene and
hemiporphycene. In addition, we will also describe the present state of knowledge
146 6 Intramolecular Hydrogen Bonding in Porphyrin Isomers
N N N N
H H H
H N N
N N
1 2
N N N N N N
H H HH
H H
N N N N N N
3 4 5
N N
N
N H H N N H H N N H
H
N N N
N
6 7 8
N
N
N N
H
H H
N N N N
9 10
Scheme 6.1 Free-base porphyrin (porphine, 1) and its isomers: porphycene (2),
hemiporphycene (3), corrphycene (4), isoporphycene (5), [18]porphyrin-(2.2.0.0) (6),
[18]porphyrin-(3.1.0.0) (7), [18]porphyrin-(4.0.0.0) (8), inverted/confused porphyrin (9),
neo-confused porphyrin (10).
regarding tautomerization in these molecules, since the thermodynamics, the rate,

and the mechanism of this process are usually predetermined by HB properties.
The geometry optimizations, followed by calculations of vibrational structure
and 1 H NMR shifts were performed using density-functional theory (DFT) with the
B3LYP functional and the triple zeta 6-311++G(d,p) basis set. Gaussian 16 software
package [1] was used. When presenting the results, no scaling factors are applied,
since our main goal was to compare different isomers with respect to each other
rather than focusing on comparison with fairly limited experimental data. In fact,
DFT calculations can now reproduce molecular properties, such as vibrational
frequencies with an accuracy comparable to the experimental one [2].
6.2 H-Bond Characteristics

Table 6.1 shows the relative energies calculated for the lowest energy forms of all the
isomers. Very similar results are obtained using double- and triple-zeta basis sets.
Porphine (1) and porphycene (2) are predicted to be the most stable ones. For other
isomers, the energy increases. The dates of the publications reporting the syntheses
Table 6.1 Calculated relative energies (kcal mol−1 ) of the

lowest energy tautomeric forms of porphyrin isomers.
B3LYP/6-31G(d,p) B3LYP/6-311++G(d,p)
1 0.97 (1.2) 0.00 (0.00)

2 0.00 (0.0) 0.15 (0.05)
3 6.0 (6.4) 5.6 (5.8)
4 13.3 (13.8) 12.7 (12.9)
5 21.0 (21.2) 20.9 (20.8)
6 31.4 (32.0) 30.8 (31.0)
7 41.9 (42.2) 41.1 (41.5)
8 41.4 (41.0) 39.9 (39.7)
9 25.9 (25.7) 23.4 (23.2)
10 26.6 (26.3) 27.2 (26.6)
In parentheses, zero-point energy-corrected values.
of 2 [3], 3 [4, 5], 4 [6, 7], 5 [8, 9], 9 [10, 11], and 10 [12] correlate with relative sta-
bilities. Three compounds that have not yet been obtained (6–8) correspond to the
highest energy structures.
Each isomer can exist in several tautomeric forms that differ in geometry and,
possibly, symmetry. One can, therefore, expect differences in the HB properties.
Comparison of HB characteristics is presented below, separately for each isomer.
The parameters that have been analyzed include the distance between H-bonded
nitrogen atoms, the NHN angle, the NH stretching, and out-of-plane (oop)-bending
frequencies, as well as the proton NMR shifts calculated for the inner hydrogens
(in case of 9, a structure with the peripheral NH group has also been considered).
These parameters provide a measure of the HB strength: shortening the N–N
distance and increasing the NHN angle should lead to a stronger bond. Experimen-
tally, this is detected, most often using infrared (IR) spectroscopy, by the shift of
NH-stretching bands to lower frequencies. The opposite shift is observed for the
oop modes. In the NMR spectra, the downhill shift of the inner proton signals is an
indicator of a stronger HB.
6.2.1 Porphine (1)

Doubly degenerate trans and quadruply degenerate cis tautomeric forms are
possible (Figure 6.1). The data presented in Table 6.2 show that the H-bonds are
stronger in the less stable cis structure. In the cis species, the distance between
H-bonded nitrogen atoms is considerably shorter, whereas the NHN angle is
larger than in the trans form. The NH-stretching frequency is redshifted in the cis
tautomer by c. 200 cm−1 . In contrast, the oop NH-bending modes are blueshifted in
the cis form. This tendency is best observed in the deuterated species, because in
the non-deuterated forms, the NH oop character is spread over several modes.
trans cis
0.0 (0.0) 8.34 (8.16)
Figure 6.1 Trans and cis tautomers of 1 and the calculated relative energies
(in kcal mol−1 ). In parentheses, zero-point energy-corrected values.
Table 6.2 HB characteristics of tautomers of 1 and their isotopologues.
NH–N NHN NH-stretching NH oop bend 𝜹 (1 H)

distance (pm) angle (∘ ) frequency (cm−1 ) (cm−1 ) (ppm)
trans 293 118 3538 a 626/667 s −5.61

3582 s 800/741 a
trans-d1 2616 495
3560 701/706
trans-d2 2602 a 467 s
2630 s 549 a
cis 272 130 3313 a 639/741 s −4.07
3347 s 836 a
cis-d1 2452 547
3331 755/808
cis-d2 2443 s 524 s
2464 a 584 a
Replacement of a single proton by a deuteron leads to well-separated NH and

ND vibrations, whereas double deuteration recovers the symmetric and anti-
symmetric combinations of the two NH/D stretching or oop-bending vibrations.
The NH/D-stretching frequencies in the −d1 isotopologues correspond well to the
average between the a and s transition energies of doubly protonated and deuterated
derivatives. This suggests that the splitting between the a and s bands can be used
as a measure of interaction between the two NH or ND oscillators.
The calculated structural and spectral parameters of the trans form of 1 agree well
with the X-ray [13] and vibrational [14–17] data.
6.2.2 Porphycene (2)

The tautomeric structures of 2 are shown in Figure 6.2. The cis1–trans energy
separation is about four times smaller than in the parent isomer. Both trans and
trans cis1 cis2

0.0 (0.0) 2.39 (1.80) 29.37 (29.32)
Figure 6.2 Trans and cis tautomers of 2 and their relative energies.
cis tautomers are doubly degenerate. The NHN angles are large and differ in the
trans and cis1 forms by only 2∘ . The N–N distances also are similar in these two
forms (Table 6.3), in contrast to the situation in porphyrin (5 vs. 21 pm difference),
but the separation of the NH-stretching frequencies is larger in 2 than in 1. The
∼650 cm−1 redshift with respect to porphyrin indicates a much stronger HB in 2.
The differences in oop frequencies and NMR 1 H shifts lead to the same conclu-
sion. It agrees well with the experimental data. The NH-stretching IR band lies
in 1 at 3324 cm−1 [14], whereas in 2 a very broad band is observed, centered at
∼2500 cm−1 [18]. The calculated difference in proton chemical shifts, 6.34 ppm,
compares well with the experimental values: 3.15 − (−3.76) = 6.81 [3, 19].
NH–N NHN NH-stretching NH oop bend 𝛅 (1 H)

trans 267 152 2942 s 951 s +0.73

2943 a 996 a
trans-d1 2188 650
2942 975
trans-d2 2188 s 637/644 s
2189 a 656 a
cis1 262 154 2656 a 1006 s +2.55
2691 s 1048 a
cis1-d1 2005 736
2673 1027
cis1-d2 1995 a 731 s
2015 s 740 a
cis2 270 97 3604 a 565 −4.46
3609 s 824
cis2-d1 2649 445/461
3606 620
cis2-d2 2647 425
2650 461
Similarly as in 1, the HB in the lowest energy trans form is weaker than in cis1.
On the other hand, the highest energy tautomer (cis2) does not reveal features
characteristic of HB. This can be explained by (i) nonplanarity caused by repulsion
between NH hydrogens, and (ii) an extremely small NHN angle.
Interestingly, calculations yield practically the same frequency for the s and a NH
stretches in the trans form, whereas these frequencies differ by 20–25 cm−1 in the
cis1 species, indicating that the NH oscillators “see each other” better in the latter.
This hypothesis is supported by the same frequencies calculated for the stretching
vibrations in trans and trans-d1 (NH) and in trans-d1 and trans-d2 (ND). The cis1
form of 2 does not exhibit such behavior, but the cis2 tautomer does.
6.2.3 Hemiporphycene (3)

Due to low symmetry (Cs ), all tautomeric forms of 3 are nondegenerate. How-
ever, very similar energies are obtained for three pairs (trans1/trans2, cis1/cis2,
and cis3/cis4, Figure 6.3). In each tautomer there exist two different types of
HB: one with a short N–N distance and a small NHN angle, and the other with
a long distance and a large angle. Such a situation provides a good model for
studying the relative contributions of distance and angle to the HB strength. It
was demonstrated for 3 that both factors are important [20]. As an illustration,
one may compare the NH-stretching frequencies in the trans1 form (Table 6.4).
The lower frequency (3313 cm−1 ), indicating a stronger bond, is obtained for the
Figure 6.3 Trans and cis tautomers of

3 and their relative energies.
cis3 19.37 (18.89) cis4 20.05 (19.55)
cis1 4.49 (4.04) cis2 5.12 (4.50)
trans1 0.0 (0.0) trans2 0.98 (0.97)


trans1 260 128 3362 815 −2.14

294 149 3313 879 −0.59
trans1-d1 a) 2446 592
3359 849
trans1-d1 b) 2475 588
3315 855
trans1-d2 2445 571
2476 616
trans2 260 132 3305 752 −3.67
293 142 3381 880 −1.25
trans2-d1 a) 2438 596
3380 756/810
trans2-d1 b) 2489 565
3307 862/865
trans2-d2 2437 534
2490 617
cis1 253 136 2988 866 −0.98
285 152 3104 949 +1.86
cis1-d1 a) 2216 635
3102 891
cis1-d1 b) 2298 613
2990 931
cis1-d2 2216 601
2298 652
cis2 252 134 3035 854 −2.27
283 146 3166 927 +0.72
cis2-d1 a) 2248 624
3164 875
cis2-d1 b) 2338 604
3037 911
cis2-d2 2247 593
2339 636
cis3 254 105 3618 558 −4.32
289 126 3472 694/712/722 −3.46
cis3-d1 a) 2653 439
3475 674
cis3-d1 b) 2555 467
3615 620/657
(continued)
Table 6.4 (Continued)

cis3-d2 2554 423

2655 523
cis4 253 108 3565 532 −5.24
289 131 3459 723/732 −4.72
cis4-d1 a) 2612 436
3467 692
cis4-d1 b) 2550 439
3557 625/690
cis4-d2 2546 402
2616 522
a) Deuteration of the shorter HB.

b) Deuteration of the longer HB.
longer distance and larger angle (294 pm/149∘ ). The other, higher-energy vibration
(3362 cm−1 ) is calculated for the much shorter HB (260 pm) which, however, is less
linear (128∘ ).
In trans1 and trans2 species, the N–N distances are the same (260 pm) or prac-
tically the same (293 and 294 pm). Therefore, comparison of the NH-stretching
frequencies in these two forms allows for assessing the role of the NHN angle in
HB strength. Making the bond more linear by changing the angle by 4∘ decreases
the frequency by 57 cm−1 ; in the other case, increase by 7∘ results in the frequency
lowered by 68 cm−1 .
The HBs in the trans forms of 3 are stronger than in 1 but weaker than in 2.
Based on comparison of N–N distances in 2 and 3, one would expect the opposite.
This again points to the influence of the angle, which is larger in 2, making the HB
more linear. The effect of the angle overrides that of the distance, pointing to the
crucial role of linearity.
Similarly as in 1 and 2, HBs in the cis1 and cis2 tautomers are stronger than
in the trans forms. It is easy to understand why: the bonds are shorter and more
linear. On the other hand, the NH-stretching frequencies in cis3 and cis4 are quite
high. In both forms, they differ by more than 100 cm−1 . The lower-frequency
vibration corresponds to a more linear structure, with a considerably larger HB
donor–acceptor distance (289 pm) than in the form that lacks the signatures of HB
(253/254 pm).
6.2.4 Corrphycene (4)

Corrphycene provides an interesting example of a structure with a doubly
degenerate trans tautomer, even though the two intramolecular HBs are not
equivalent (Table 6.5). The difference is not large, but sufficient to localize the

trans 274 129 3446 723 −2.66

275 122 3513 844 −1.57
trans-d1 a) 2539 563
3509 704
trans-d1 b) 2580 531
3450 751
trans-d2 2537 519
2582 590
cis1 266 135 3179 811 −0.18
3214 899
cis1-d1 2362 580
3196 857
cis1-d2 2351 552
2371 621/634
cis2 265 129 3283 755/809 −0.65
3309 869
cis2-d1 2429 573
3296 839
cis2-d2 2421 551
2438 607
cis3 244 110 3439 589 −4.48
321 144 3496 741 −3.21
cis3-d1 a) 2561 490
3449 703
cis3-d1 b) 2538 471
3487 721
cis3-d2 2533 454
2567 548

NH stretches on single NH bonds so that the description in terms of symmetric

and antisymmetric combinations of the two NH/D stretches no longer holds. The
lower NH-stretching frequency (by 67 cm−1 ) is predicted for the HB shorter by
just 1 pm and more linear by 7∘ (the bond on the right side of the trans structure
in Figure 6.4). The frequencies calculated for −d1 and −d2 isotopologues indicate
weak coupling between the NH/D oscillators.
The HBs in 4 are weaker than in 2 and 3, but somewhat stronger than in 1, in
agreement with the values of N–N distances and angles. One notes that the averages
of the NH-stretching frequencies in 4 and 1 differ by 80 cm−1 . This is not much,
trans 0.0 (0.0) cis1 5.36 (5.03) cis2 5.74 (5.25) cis3 16.28 (15.69)
given that the N–N distances in 4 are significantly shorter (274/275 pm compared to
293 pm in 1) and that the HBs in 4 are more linear (129∘ /122∘ vs. 118∘ in 1).
Comparison between trans and cis tautomers shows the same trend as in 1–3:
the HBs in the cis forms are stronger. Inspection of geometries shows clearly that
switching from trans to cis species in 4 should lead to stronger HBs, since both dis-
tance and angle effects act in the same direction: the HBs become shorter and more
linear.
In the cis3 tautomeric form, the geometries of the two HBs are completely
different, regarding both distances and angles (Table 6.5). However, the calculated
NH-stretching frequencies differ by less than 60 cm−1 . This can be explained by the
opposite effects of longer distances and larger angles almost canceling each other.
6.2.5 Isoporphycene (5)

We consider here the Z forms of isoporphycene, which are of lower energy than
the E isomers. It is known, however, that this ordering can be reversed upon specific
substitution [21].
The symmetry of 5 dictates double degeneracy of trans and cis1 tautomers
(Figure 6.5). In the trans form, both “horizontal” and “vertical” HBs are possible.
The former are energetically more favorable, because of linearity. In the vertical
arrangement, both bonds would be equivalent, whereas the two horizontal HBs
strongly differ (Table 6.6). The difference in the N–N distances (248 vs. 314 pm)
is too large to be offset by the more linear arrangement in the longer bond, so
it is weaker in the trans form of 5. The calculated NH-stretching frequencies of
the long and short HB differ by 240 cm−1 . The same ordering is encountered in
the cis1 form, but now, because the stronger bond gets even shorter (242 pm) and
trans 0.0(0.0) cis1 4.09 (3.19) cis2 21.59 (21.66) cis3.22.86 (22.53)

trans 248 134 3145 831 +2.88

314 160 3385 942 −2.30
trans-d1 a) 2325 642
3357 839
trans-d1 b) 2476 583
3145 935
trans-d2 2325 579
2476 632
cis1 242 140 2570 857 +6.48
306 161 3214 1044 −1.60
cis1-d1 a) 1927 662/667/684
3213 869
cis1-d1 b) 2374 605
2572 1040
cis1-d2 1927 602
2376 684/740/801
cis2 264 100 3584 605 −4.39
3596 760/848
cis2-d1 2637 447/464/522
3590 815
cis2-d2 2633 442
2641 594
cis3 264 100 3600 604 −4.45
3610 720
cis3-d1 2647 459/480
3605 687
cis3-d2 2644 455
2650 502

more linear (140∘ ), the difference is as large as 644 cm−1 . The calculated frequency
of 2570 cm−1 is the lowest among all isomers, indicating the strongest HB. An
interesting observation is that the displacement vector of the 2570 cm−1 vibration
no longer coincides with the direction of the N—H bond; it points toward the HB
acceptor nitrogen atom. It would thus be more appropriate to call such a vibration
a bending mode.
The calculations for 5 give a rather small cis1-trans energy difference
(3.19 kcal mol−1 ). A smaller value is obtained only for porphycene (1.80 kcal mol−1 ).
One can envisage that this difference should become zero for a very strong HB
in which the proton is delocalized so that the probabilities of finding it at a
specific distance from the nitrogen donor and acceptor are the same. Under such
conditions, the notions of trans and cis forms lose their meaning. Isoporphycene
and porphycene are getting closer to such arrangement. The reason for a smaller
cis1–trans energy difference in porphycene is that it has two very strong HBs,
whereas isoporphycene has only one (albeit stronger).
Neither cis2 nor cis3 reveals HB features. The N–N distances are favorable
(260 pm), but the formation of HBs is hindered by a small NHN angle of 100∘ .
6.2.6 Porphyrin-(2.2.0.0) (6)

Four tautomeric structures are possible in 6 (Figure 6.6), of which the cis
forms are doubly degenerate. Two equivalent inner hydrogens in the trans1
tautomer form a very weak HB (N–N distance of 323 pm, Table 6.7) with the
same nitrogen atom (upper nitrogen in Figure 6.6). Alternative, short HBs to
the bottom nitrogen are not realized because of the small NHN angle (99∘ ).
A similar situation is encountered in the trans2 form. In this tautomer, it is
the upper NH group that could possibly be engaged in the HB with the two
middle nitrogen atoms acting as acceptors. However, the combination of the
large N–N distance and the relatively small NHN angle makes such interac-
tion negligible. It is instructive to compare the NH-stretching frequencies in
trans1 and trans2: they are much higher in the latter, even though the N–N
distances are practically the same (323/325 pm). The difference in frequencies
is due to a more linear HB arrangement in trans1 (146∘ compared to 126∘ in
trans 2).
Both NH-stretching frequencies in cis1 and one in cis2 are definitely lower than
in the trans forms. Now, the inner two protons form HBs with different nitrogen
atoms. The lowest frequency, 3290 cm−1 in cis1, is associated with the long, but
very linear HB (303 pm, 156∘ ), whereas the highest one, 3557 cm−1 in cis2 may be
attributed to the proton located at the middle nitrogen (cf. Figure 6.6). This proton
can be considered to be non-hydrogen bonded, even though the distance to a possi-
ble HB acceptor is very short. However, a small NHN angle makes the HB interaction
inefficient.
trans1 0.0 (0.0) trans2 1.44 (1.60) cis1 8.99 (8.68) cis2 9.87 (9.54)
Figure 6.6 Trans and cis tautomers of 6 and the relative energies.

trans1 250 99 3457 a 741 s −4.02

323 146 3487 s 844 a
trans1-d1 2555 563
3472 748
trans1-d2 2545 a 545 s
2565 s 587 a
trans2 255 101 3574 686/720 −5.40
325 126 3628 746/806/817 −3.44
trans2-d1 a) 2629 514
3623 806/816
trans2-d1 b) 2658 555
3579 690/721
trans2-d2 2626 512
2661 559
cis1 242 110 3289 720 −3.02

303 156 3411 853 −2.15
cis1-d1 b) 2507 537
3292 748/803
cis1-d1 c) 2431 540
3408 749/820
cis1-d2 2429 516
2509 595
cis2 243 107 3365 714 −3.11
300 137 3557 807 −2.93
cis2-d1 c) 2611 509
3365 750/792/800
cis2-d1 a) 2479 548
3556 685/716
cis2-d2 2478 501
2612 567/570
a) Deuteration of the upper nitrogen (see Figure 6.6).

b) Deuteration of the bottom nitrogen.
c) Deuteration of the middle nitrogen.
6.2.7 Porphyrin-(3.1.0.0) (7)

Six possible tautomeric forms of the Z isomer of 7 are shown in Figure 6.7. The
geometry of the inner cavity resembles that of hemiporphycene (Figure 6.3). The
dimensions of the four, uneven sides of the quadrangle range from 250 to 364 pm
Figure 6.7 Trans and cis tautomers

of 7 and the relative energies.
cis3 10.77 (10.44) cis4 11.79 (11.26)
cis1 7.40 (7.11) cis2 9.70 (8.87)
trans1 0.0 (0.0) trans2 0.08 (–0.37)
in trans1 and from 245 to 366 nm in trans2. In the cis forms, the spread is even larger.
The shortest N–N distance (235 pm ) is obtained for cis3 and the longest one (388 pm)
for cis2. Trans2 and cis3 tautomers are planar, while the other ones are not.
The pattern of relative tautomer energies in 7 is also similar to that of hemi-
porphycene, with two almost degenerate trans tautomers being the most stable.
However, the calculated NH-stretching frequencies are notably higher for 7, which
means that the HBs are weaker than in 3. This result provides another demon-
stration of the crucial role of nonlinearity. Comparison of the HB geometries in 3
and 7 (Tables 6.4 and 6.8) shows very similar N–N distances in the two isomers,
but the angles are considerably smaller in 7 (116∘ and 101∘ vs 149∘ and 128∘ in 3).
As a result, the averaged NH-stretching frequencies in 3 and 7 differ by more than
250 cm−1 .
Similar to other isomers, the HBs are stronger in the cis forms. Comparison
with hemiporphycene reveals differences analogous to the case of trans species.
Nonlinearity of HBs in 7 makes them weaker in comparison with 3. In addition,
the differences between the two NH-stretching frequencies are much larger in 7.
This can be explained by large differences in the NHN angles in each tautomer that
make the two bonds dissimilar. In 3, the two angles do not differ that much.
6.2.8 Porphyrin-(4.0.0.0) (8)

E and Z isomers are possible, the latter calculated to be less stable by c. 30 kcal mol−1 .
We, therefore, focus here on the E forms (Figure 6.8). The least stable porphyrin

trans1 258 101 3568 a 658/707 −3.51

286 116 3616 s 748/813 −1.65
trans1-d1 a) 2649 568
3574 660/714/716
trans1-d1 b) 2626 520
3611 741/750/812
trans1-d2 2622 a 523
2652 s 570
trans2 253 99 3467 711 −4.65
282 128 3546 854 −2.39
trans2-d1 a) 2604 523
3469 847
trans2-d1 b) 2552 583
3544 719
trans2-d2 2550 518
2606 597
cis1 250 108 3249 709/710 −1.93
271 137 3506 848 −0.36
cis1-d1 a) 2575 557
3251 839
cis1-d1 b) 2400 571
3505 724
cis1-d2 2399 553
2576 591
cis2 248 105 3327 646/695 −4.52
268 127 3575 835 −2.24
cis2-d1 a) 2624 498
3328 829
cis2-d1 b) 2451 566
3375 699/709
cis2-d2 2451 495
2625 581
cis3 235 113 3402 673/715/823 −3.90
346 166 3429 879/899 −0.59
cis3-d1 a) 2508 580
3421 726
cis3-d1 b) 2520 522
3411 749/814
(continued)

cis3-d2 2504 520

2523 586
cis4 242 108 3444 721/724 −5.33
336 149 3496 734/802 −2.61
cis4-d1 a) 2566 556
3446 721/724
cis4-d1 b) 2536 524
3494 801
cis4-d2 2535 523
2567 556

trans1 0.0 (0.0) trans2 0.48 (0.41) cis1 3.79 (3.83)
cis2 9.37 (9.03) cis3 10.5 (10.52) cis4 13.40 (13.04)
isomer is characterized by irregular cavities, small N–N distances, and small NHN
angles. All tautomers are nonplanar.
Even though the N–N distances are extremely short, the calculations (Table 6.9)
indicate very weak or absent HBs. This is caused by small NHN angles in all
tautomers, spanning a narrow range between 99∘ and 107∘ . In two cases (cis3 and
cis 4), large angles are obtained, but the corresponding N atoms lie too far from
each other to enable HB formation. Only one calculated NH-stretching frequency is
smaller than 3500 cm−1 . This is the case of cis4 tautomer. Characteristically, it has
the largest (but still small) angle among all the tautomers.

trans1 252 104 3614 a 653 −3.18

263 99 3643 s 745 −2.61
trans1-d1 a) 2658 504
3635 739
trans1-d1 b) 2666 544
3623 661
trans1-d2 2653 500
2671 555
trans2 254 102 3565 616 −4.49
259 103 3633 832 −1.10
trans2-d1 a) 2616 595
3632 618
trans2-d1 b) 2666 466
3566 832
trans2-d2 2615 466
2666 596
cis1 252 104 3555 695 −2.66
257 104 3591 738 +0.01
cis1-d1 a) 2634 538/556
3557 709/727
cis1-d1 b) 2613 497/538
3588 692
cis1-d2 2612 495
2635 510/545
cis2 249 106 3527 639 −1.85
257 104 3613 742 −0.67
cis2-d1 a) 2591 550
3612 642
cis2-d1 b) 2652 492
3528 740
cis2-d2 2591 491
2652 556
cis3 248 105 3544 622/648 −2.92
408 154 3590 729 +0.64
cis3-d1 a) 2603 559
3589 636
cis3-d1 b) 2636 452
3545 719
cis3-d2 2602 451
2637 564
(continued)

cis4 246 107 3468 518/558/584 −5.03

416 159 3635 751/814 −0.13
cis4-d1 a) 2549 555/563
3634 590
cis4-d1 b) 2669 416
3468 749/813
cis4-d2 2549 415
669 556

6.2.9 Inverted/Confused Porphyrin (9)

All the molecules considered so far belong to “nitrogen-in” isomers. The inverted
porphyrin has different topology. Since one of the nitrogens is now located on
the periphery, some tautomers (D–F in Figure 6.9) contain the NH group that is
not intramolecularly H-bonded. The NH vibrations of this group can be treated
as a reference. Inspection of Table 6.10 shows very similar values obtained for
tautomers D–F. These values remain unchanged irrespective of whether the inner
NH proton is replaced by deuterium or not. Thus, the two NH oscillators do not
communicate.
A 0.0 (0.0) B 5.96 (5.67) C 6.33 (6.00)
D 6.91 (6.28) E 13.58 (12.83) F 14.04 (13.30)

A 295 125 3579 598 −5.51

3605 696 −5.50
A-d1 2639 465
3592 676
A-d2 2630 453
2647 491
B 284 128 3491 538 −7.32
3632 617 −4.14
B-d1 a) 2567 460
3631 552
B-d1 b) 2666 410
3491 608
B-d2 2566 408
2666 469
C 284 129 3484 524 −7.45
3632 622 −4.14
C-d1 a) 2562 459
3631 545
C-d1 b) 2665 402
3485 612
C-d2 2562 398
2666 473
D 296 116 3590 506c) +1.58
3649c) 619/683 +8.99c)
D-d1 a) 2635 488
3649c) 507c)
D-d1 b) 2679c) 381c)
3590 619/683
D-d2 2635 381c)
2679c) 488
E 279 128 3451 473/503c) +2.16
3650c) 622/695 +9.02c)
E-d1 d) 2537 494
3650 472/503
E-d1 c) 2680c) 374c)
3451 620/644/681
E-d2 2537 374c)
2680 495
F 280 128 3460 513c) +2.22
3648c) 629 +9.20c)
(continued)

F-d1 a) 2543 492

3648c) 515c)
F-d1 b) 2679c) 389c)
3460 623
F-d2 2543 389c)
2678 492

c) Peripheral NH.
d) Deuteration of inner NH.
The inner cavity geometries in form A and the trans tautomer of porphine 1 do
not differ much, which leads to similar values of the NH-stretching frequencies.
On the other hand, even though the cis forms B and C exhibit stronger HBs than the
trans tautomer, the difference is not so pronounced as in porphine. This is because
the N–N separation in 9 decreases upon passing from trans to cis forms not that
strongly as in 1. This is obviously caused by the presence of the third (CH) proton in
the cavity.
6.2.10 Neo-confused Porphyrin (10)

Figure 6.10 shows three tautomeric forms of 10. Although only one NH group is
present, one can consider tautomer A as an analog of a trans form and tautomers
B and C as analogs of the cis structure (with respect to the inner CH proton).
The “trans” structure has the lowest energy, as in porphine. The cis–trans energy
gap in 1 lies in between the values calculated for tautomers B and C.
The cavity parameters in 10 and in the parent porphyrin 1 are very similar for both
trans and cis tautomeric forms. It is, therefore, not surprising to find similar values of
the NH-stretching frequencies in both molecules (Tables 6.2 and 6.11). What should
A 0.0 (0.0) B 6.7 (6.6) C 10.3 (10.0)
Figure 6.10 Tautomers of 10 and the relative energies.

6.3 Correlations Between Geometry and HB Strength 165

A 292 116 3576 617 −2.26

293 118
A-d1 2625 504
B 274 131 3370 611 −0.63
B-d1 2470 511
C 278 129 3432 656 −2.01
C-d1 2524 513
be noted is the 62 cm−1 difference in the frequencies of B and C tautomers, which is

caused by relatively small changes in the distance and angle (4 pm and 2∘ ). However,
both factors are acting in the same direction upon conversion from B to C.
6.3 Correlations Between Geometry and HB Strength
Based on the values of the NH-stretching frequencies, one can now classify the iso-
mers according to the HB strength. Definitely, in the lowest energy form, which
always corresponds to the trans arrangement of the inner protons, porphycene (2)
is most strongly H-bonded. Moderate HBs found in four isomers can be ordered as
5 > 3 > 4 ≈ 6. Porphyrin 1 and its inverted and neo-confused isomers 9 and 10 can be
described as weakly H-bonded, whereas 7 and 8 are non-bonded.
In the cis form, all isomers strengthen their HBs because of more favorable dis-
tance and angle values. Making the bond shorter and more linear may lead to dra-
matic changes, as discussed above for the cis1 tautomer of isoporphycene.
The most general conclusion from the present simulation is that the intramolecu-
lar HB strength cannot be reliably predicted by considering only one parameter – the
bond length. This is illustrated in Figure 6.11. No correlation is observed between
the calculated frequency and the N–N distance or the NHN angle. However, the
projection of a 3D plot that simultaneously takes into account both parameters
clearly shows the region for which a strong HB can be expected. It corresponds
roughly to angles larger than 130∘ and distances smaller than 280 pm.
It has been postulated [22] that the strong HBs in porphycene can be responsi-
ble for its stability related to porphyrin and other isomers. In fact, the calculated
energies are practically the same for 1 and 2. However, knowing that the cis tau-
tomers are more stable than the trans forms, one may ask why the lowest energy
form in each isomer corresponds to the trans configuration. Apparently, the gain in
HB in the cis structure cannot compensate the effects related to electronic density
rearrangement upon trans–cis conversion.
3800
3800
3600
3600
NH-stretching frequency (cm–1)

NH-stretching frequency (cm–1)
3400
3400
3200
3200
3000 3000
2800 2800
2600 2600
2400 2400
90 100 110 120 130 140 150 160 170 220 240 260 280 300 320 340 360 380 400 420 440
(a) NHN angle (°) (b) N–N distance (pm)
1080
978.5
170
865.0
160
757.5
150 650.0
NHN angle (°)
140 542.5
130 435.0
327.5
120
220.0
110
112.5
100
5.000
90
220 240 260 280 300 320 340 360 380 400 420 440
(c) N–N distace [pm]
Figure 6.11 (a) Plot of the calculated NH-stretching frequencies vs. the corresponding NHN angles. (b) Plot of the calculated NH-stretching frequencies
vs. the corresponding N–N distances. (c) Projection of the 3D plot in which the values of (3650 cm−1 – calculated NH-stretching frequency) are presented
as the function of the N–N distance and the NHN angle. For the cases when the two frequencies correspond to symmetric and antisymmetric
combinations, the average of the two was taken. The value of 3650 cm−1 was selected as corresponding to a non-H-bonded NH group.
6.4 Parameters That Can Describe the HB Strength 167
6.4 Parameters That Can Describe the HB Strength
In the work that reported the tautomerization rates in a series of differently substi-
tuted porphycenes [23], three parameters were used as indicators of HB strength.
Nice correlation was found between the values of NH-stretching frequencies, N–N
distances, and proton NMR chemical shifts. The present results show that using
solely distances may not be correct. The good correlations obtained for porphycenes
without taking angles into account are most probably due to similar values of the
NHN angle in the derivatives of the same isomer. However, for different tautomers
of another isomer, hemiporphycene, it was shown that the correlation based only on
the distance fails [20].
The present results enable checking the correlation between the proton chemical
shift and the NH-stretching frequency. This is shown in Figure 6.12, which contains
the plot of the 1 H shifts vs. the frequencies calculated for all the isomers. The cor-
relation seems to work fairly well for some isomers, but breaks down completely
for nonplanar isomers 7 and 8. It is also poor for those tautomeric forms of 6 in
which the HB involves one donor and two acceptors or vice versa. The general con-
clusion is that the NMR parameters may provide a measure of the HB strength, but
a careful, case by case analysis is required.
Another parameter that may find use in estimating the HB strength is the NH
oop-bending frequency. The values obtained for different isomers span a large range,
from above 1000 cm−1 to less than 500 cm−1 (or 400 cm−1 in deuterated species).
The problem is that while the stretching vibrations are well-separated from other
1
8 2
3
4
6
5
6
H chemical shift (ppm)
4 7
8
2 9
10
–2
1
–4
–6
2600 2800 3000 3200 3400 3600
–1
NH stretching frequency (cm )
Figure 6.12 Correlation between the proton chemical shift and the NH-stretching
frequency. Different symbols are used for each isomer.
1
2
3
4
650 5
NH oop–bending frequency (cm–1)
6
9
600 10
550
500
450
400
1900 2000 2100 2200 2300 2400 2500 2600 2700
NH–stretching frequency (cm–1)
Figure 6.13 Correlation between the out-of-plane NH-bending and the NH-stretching
frequencies. Different symbols are used for each isomer. Nonplanar isomers 7 and 8 were
not included.
oscillations involving movement of other atoms, the oop modes may mix with other
ones. A similar difficulty is well recognized for the in-plane NH-bending modes:
after deuteration, many bands are shifted, so it is not possible to single out one
specific mode as an indicator. With oop vibrations, some hope comes from the obser-
vation that in the deuterated species the oop ND character is found for just one
mode, of which the frequency could serve, in addition to the frequency value of the
NH-stretching vibration, as a good measure of HB strength. We have checked this
using the values calculated for the singly deuterated isomers. The results presented
in Figure 6.13 are encouraging. Correlation between the stretching and oop-bending
modes is good in the region of strong HBs. It becomes worse for weak HBs, which is
understandable, since in this case the contributions from other effects become more
important.
6.5 Tautomerization Mechanisms
Knowledge of the HB properties is a prerequisite for understanding tautomeriza-

tion mechanisms, as the formation of HB usually precedes proton or hydrogen
transfer. Intramolecular tautomerization has been studied for porphine [24–26]
and porphycene [27, 28]. In 1, the stepwise mechanism involves thermally induced
population of a vibrationally excited level, from which a single hydrogen transfer
6.6 Summary 169
occurs (by tunneling), thus leading to a cis structure. Another single hydrogen
transfer brings the structure back to the initial structure or the equivalent trans form.
In porphycene, tautomerization proceeds via double hydrogen tunneling, either
from the vibrationless electronic ground state level (this is the dominant mecha-
nism at low temperatures) or via tunneling activated by vibrational excitation of
a totally symmetric low frequency mode (dominant mechanism at room tempera-
ture) [28]. The reaction can be promoted or hindered by specific vibrational excita-
tion, as demonstrated by the analysis of tunneling splittings in porphycene isolated
in supersonic jets [29]. This mode selectivity means that the effective barrier for
tautomerization depends on which vibrational modes are being excited. Keeping
in mind that the barrier to tautomerization and HB strength are related leads to the
conclusion that the HB strength also can vary with excitation of specific vibrations.
Tautomerizations in porphyrin and porphycenes occur on completely different
time scales. In 1, the reaction proceeds in microseconds, whereas in porphycenes
it requires, depending on the substituent, from tens of femtoseconds to hundreds of
picoseconds. This huge difference between rates in 1 and 2 is due to different charac-
teristics of HBs, strong in porphycene and weak in porphyrin. The results obtained
in the present work can serve as a starting point for estimation of the tautomeriza-
tion times and elucidation of mechanisms. For instance, in the case of isomers in
which the two HBs are strongly different, such as isoporphycene, one can expect
a stepwise transfer mechanism, with the kinetics controlled by the properties of a
weaker bond. On the contrary, for the isomers with identical or very similar HBs,
such as corrphycene, a concerted mechanism seems probable.
6.6 Summary
Porphyrin isomers are very good models for understanding the relationship
between geometry and intramolecular HB strength. The calculations presented in
this work show clearly that to properly characterize an intramolecular HB, two
parameters are inseparable and should be always taken into account. These are
the N–N distance and the NHN angle. In other words, distance and orientation
dependences are equally important. Contrary to intermolecular HBs, where the
partners often may adjust these parameters to an optimal value, the geometry of the
intramolecular bonds is given by the topology of the molecule.
The competing contributions of distance and orientation factors may lead to such
unusual phenomena as short HBs being weaker than longer ones due to nonlin-
earity. Our results indicate the range of distances and angles that are obligatory for
the formation of strong HBs.
Another general conclusion regarding porphyrin isomers is that the HBs in the
cis tautomers are stronger than in the trans forms. However, the latter are always
predicted to be the lowest energy structures, so stabilization by forming HBs is not
sufficient to overcome the unfavorable effects related to electron density redistribu-
tion upon trans–cis conversion.
Finally, some conclusions may be drawn regarding the cooperativity between

the two intramolecular HBs in porphyrin isomers. An effect that seems somehow
counterintuitive is that the interaction between the two oscillators need not increase
for stronger HBs. This was shown by comparison of 1 and 2. We hope that further
studies of HB dynamics and interactions between HBs of different strengths and
mutual orientations will exploit porphyrin isomers as valuable research objects.
Acknowledgments
This work was supported by the Polish National Science Centre (grant no.
2017/26/M/ST4/00872), the PL-Grid infrastructure, and the Interdisciplinary
Centre for Mathematical and Computational Modeling (grant no. G17-14).
References
1 Frisch, M., Trucks, G., Schlegel, H. et al. (2016). Gaussian 16, Revision A.03.
Wallingford CT: Gaussian, Inc.
2 Gorski, A., Gawinkowski, S., Herbich, J. et al. (2012). 1H-Pyrrolo[3,2-h]quinoline:
a benchmark molecule for reliable calculations of vibrational frequencies, IR
intensities, and Raman activities. J. Phys. Chem. A. 116: 11973–11986.
3 Vogel, E., Köcher, M., Schmickler, H., and Lex, J. (1986). Porphycene – a novel
porphin isomer. Angew. Chem. Int. Ed. 25: 257–259.
4 Callot, H.J., Metz, B., and Tschamber, T. (1995). A novel porphyrin isomer:
hemiporphycene. Formation and single-crystal X-ray diffraction structure deter-
mination of a hemiporphycene nickel complex. New J. Chem. 19: 155–159.
5 Vogel, E., Bröring, M., Weghorn, S.J. et al. (1997). Octaethylhemiporphycene:
synthesis, molecular structure, and photophysics. Angew. Chem. Int. Ed. 36:
1651–1654.
6 Sessler, J.L., Brucker, E.A., Weghorn, S.J. et al. (1994). Corrphycene – a new por-
phyrin isomer. Angew. Chem. Int. Ed. 33: 2308–2312.
7 Aukauloo, M.A. and Guilard, R. (1994). The "Etioporphycerin": synthesis and
characterization of a new porphyrin isomer. New J. Chem. 18: 1205–1207.
8 Vogel, E., Scholz, P., Demuth, R. et al. (1999). Isoporphycene: the fourth con-
stitutional isomer of porphyrin with an N4 core – occurrence of E/Z isomerism.
Angew. Chem. Int. Ed. 38 (2): 919–923.
9 Vogel, E., Bröring, M., Erben, C. et al. (1997). Palladium complexes of the new
porphyrin isomers (Z)- and (E)-isoporphycene – PdII -induced cyclization of
tetrapyrrolealdehydes. Angew. Chem. Int. Ed. 36: 353–357.
10 Furuta, H., Asano, T., and Ogawa, T. (1994). “N-confused porphyrin”: a new iso-
mer of tetraphenylporphyrin. J. Am. Chem. Soc. 116: 767–768.
11 Chmielewski, P.J., Latos-Grazy ̇ ński, L., Rachlewicz, K., and Glowiak, T. (1994).
Tetra-p-tolylporphyrin with an inverted pyrrole ring: a novel isomer of porphyrin.
Angew. Chem. Int. Ed. 33: 779–781.
References 171
12 Lash, T.D., Lammer, A.D., and Ferrence, G.M. (2011). Neo-confused por-
phyrins, a new class of porphyrin isomers. Angew. Chem. Int. Ed. Engl. 50:
9718–9721.
13 Devillers, C.H., Fleurat-Lessard, P., and Lucas, D. (2016). “Porphine,” the fully
unsubstituted porphyrin: a comprehensive overview. In: Handbook of Porphyrin
Science, vol. 37 (ed. K. Smith, K. Kadish and R. Guilard), 75–231. Singapore:
World Scientific.
14 Radziszewski, J., Nepraš, M., Balaji, V. et al. (1995). Polarized IR spectra of pho-
tooriented matrix-isolated free-base porphyrin isotopomers. J. Phys. Chem. 99:
14254–14260.
15 Kozlowski, P.M., Jarze˛cki, A.A., Pulay, P. et al. (1996). Vibrational assignment
and definite harmonic force field for porphine. 2. Comparison with nonreso-
nance Raman data. J. Phys. Chem. 100: 13985–13992.
16 Kozlowski, P.M., Jarze˛cki, A.A., and Pulay, P. (1996). Vibrational assignment
and definite harmonic force field for porphine. 1. Scaled quantum mechanical
results and comparison with empirical force field. J. Phys. Chem. 100: 7007–7013.
17 Verdal, N., Kozlowski, P.M., and Hudson, B.S. (2005). Inelastic neutron scat-
tering spectra of free base and zinc porphines: a comparison with DFT-based
vibrational analysis. J. Phys. Chem. A. 109: 5724–5733.
18 Gawinkowski, S., Walewski, Ł., Vdovin, A. et al. (2012). Vibrations and hydrogen
bonding in porphycene. Phys. Chem. Chem. Phys. 14: 5489–5503.
19 Scheer, H. and Katz, J.J. (1975). Nuclear magnetic resonance spectroscopy
of porphyrins and metalloporphyrins. In: Porphyrins and Metalloporphyrins
(ed. K.M. Smith), 399–524. Elsevier.
20 Ostapko, J., Nawara, K., Kijak, M. et al. (2016). Parent, unsubstituted hemipor-
phycene: synthesis and properties. Chem. Eur. J. 22: 17311–17320.
21 Gisselbrecht, J.P., Gross, M., Vogel, E. et al. (2001). Redox properties of hemi-
porphycene and isoporphycene comparison with those of porphyrin and other
porphyrin isomers, porphycene and corrphycene. J. Electroanal. Chem. 507:
244–249.
22 Wu, Y.D., Chan, K.W.K., Yip, C.P. et al. (1997). Porphyrin isomers: geometry, tau-
tomerism, geometrical isomerism, and stability. J. Org. Chem. 62: 9240–9250.
23 Cia˛ćka, P., Fita, P., Listkowski, A. et al. (2015). Tautomerism in porphycenes:
analysis of rate-affecting factors. J. Phys. Chem. B. 119: 2292–2301.
24 Schlabach, M., Wehrle, B., Rumpel, H. et al. (1992). NMR and NIR stud-
ies of the tautomerism of 5,10,15,20-tetraphenylporphyrin, including kinetic
HH/HD/DD isotope and solid state effects. Ber. Bunsenges. Phys. Chem. 96:
821–833.
25 Braun, J., Schlabach, M., Wehrle, B. et al. (1994). NMR study of the tautomerism
of porphyrin including the kinetic HH/HD/DD isotope effects in the liquid and
the solid state. J. Am. Chem. Soc. 116: 6593–6604.
26 Braun, J., Limbach, H.H., Williams, P.G. et al. (1996). Observation of kinetic
tritium isotope effects by dynamic NMR. The tautomerism of porphyrin. J. Am.
Chem. Soc. 118: 7231–7232.
27 Waluk, J. (2017). Spectroscopy and tautomerization studies of porphycenes.

Chem. Rev. 117: 2447–2480.
28 Cia˛ćka, P., Fita, P., Listkowski, A. et al. (2016). Evidence for dominant role of
tunneling in condensed phases and at high temperatures: double hydrogen
transfer in porphycenes. J. Phys. Chem. Lett. 7: 283–288.
29 Mengesha, E.T., Sepioł, J., Borowicz, P., and Waluk, J. (2013). Vibrations of
porphycene in the S0 and S1 electronic states: single vibronic level dispersed
fluorescence study in a supersonic jet. J. Chem. Phys. 138: 174201-1–174201-14.
173
Isotope Effects in Hydrogen Bond Research

Poul Erik Hansen
Roskilde University, Department of Science and Environment, DK-4000 Roskilde, Denmark
7.1 Introduction
Isotope effects in hydrogen bond research is not a new topic. Several fine books [1, 2]
and relevant reviews exist [3, 4]. The present chapter will primarily concentrate on
new findings during the last 15 years.
Considering a hydrogen bond, three sites are important to isotope substitution,
the heavy atom of the acceptor (A), that of the donor (B), and foremost the hydrogen
bonded hydrogen. The largest effects are found in the latter case. The largest effects
are seen when the mass difference between the “normal” and the isotope is large.
In the case of hydrogen, two heavier isotopes exist, deuterium 2 H(D) and tritium
3 H(T). The latter is radioactive and not studied so intensely recently.
A……..H-B
Isotope substitution especially replacing the hydrogen bonded hydrogen with deu-
terium or tritium is a very central way of gauging hydrogen bonding. Such sub-
stitutions will influence proton transfer, hydrogen bond strength, reaction rates of
both chemicals, and enzymatic reactions. Isotope substitution may also influence
a chemical equilibrium. Isotope substitution will influence average bond distances
and other physical parameters such as NMR chemical shifts and coupling constants,
and infrared stretching frequencies. All these changes in physical parameters offer
tools to investigate hydrogen bonds. Isotope effects may also be estimated, e.g. using
density functional theory (DFT) calculations.
Several interesting reviews exist [5–7].
7.2 Hydrogen Bond Potentials
Two-dimensional hydrogen bond potentials are shown in Figure 7.1. The one
to the right is shown the potential for a weak hydrogen bond. The one to the
left is a stronger but symmetric hydrogen bond. This theme can be developed
174 7 Isotope Effects in Hydrogen Bond Research
5000
4000
3000
V (cm–1)
2000
1000
ΔZPE = 200 cm–1 ΔZPE = 500 cm–1
0
O–H...N O...H–N
–1000
–20 0 20 40 60 80 –20 0 20 40 60 80
r – req. (pm) r – req. (pm)
Figure 7.1 Hydrogen bond potentials. A more complete potential is seen in Figure 7.2.
E (kcal mol–1)
70
60
50
40
30
20
10
0
3.2
2 3
2.8
0.8 2.6
1
1.2 2.4 r(O···O) (A)
1.4 2.2
r(O–H) (A) 1.6
1.8
Figure 7.2 Hydrogen bond potential for picolinic acid N-oxide in chloroform. Source: From
Ref. [10] with permission from the American Chemical Society.
further by varying the barrier. If this is very low we end up with what has been
termed low-barrier hydrogen bond (lbhb) potentials [8, 9]. Obviously, the potential
for slightly asymmetric hydrogen bonds can easily be imagined by lowering the
bottom of the potential for one of the wells. What is also obvious is that systems,
like presented to the left in Figure 7.1, are most likely tautomeric. What is also
important in the present discussion is the difference in zero-point energy (ZPE) in
Figure 7.1.
Over the years, a lot of discussion has occurred regarding lbhb. A comprehensive
review was given by Perrin [11].
7.3 Calculations 175
7.3 Calculations
Calculation of isotope effects can be of great help. Two schools regarding isotope
effects and geometries exist, one relies on the Born–Oppenheimer approximation,
which assumes that the hydrogen bond potential is unchanged by isotope substitu-
tion. The theories of Jameson et al. [12] are based on this idea. If the potential is
asymmetric an on the average shortening of the XD bond will be observed, X being
typically oxygen but also nitrogen or sulfur is a possibility.
The second school assumes that also the heavy atoms change position as a
consequence of isotope substitution. This idea is based on deuteration experiments
in the solid-state described by Ubbelohde and Gallagher [13]. This so-called
Ubbelohde effect usually results in a lengthening of the heavy atom distance, when
a deuterium is introduced in the hydrogen bond, but also the OH and OH· · ·O
distances will change. This is referred to as geometric isotope effects. Ubbelohde
and Gallagher studied a series of compounds among others KH2 PO4 . They found
that the O· · ·O distance increased from 40.5 × 10−3 Å to a slight contraction
upon deuteration [13]. A study of crystals of 2,3,5,6-tetrafluoroterephthalic acid
and the corresponding anion led to an elongation of the O· · ·O distance upon
deuteration. This study also demonstrated that no correlation could be established
between the O· · ·O distances and the OH· · ·O distances [14, 15]. Another recent
example of intermolecular geometric effects is seen in crystal of imidazolium
hydrogen terephthalate [16]. Tachikawa et al. studied a hydrogen bonded con-
ductor k-H3 (Cat EDT-ST)2 (see Figure 7.3). They found a slight shortening of the
heavy atom distance upon deuteration [17, 18]. The fragment molecular-orbital
–multicomponent molecular-orbital method was shown to be able to treat large
systems like proteins [19]. A theoretical study of the photoactive yellow protein
(PYP) using multicomponent DFT revealed that the hydrogen bonds between
the amino acids Glu-42 or Tyr-42 and the chromophore of PYP were not of
lbhb type.
The Limbach group studies geometric isotope effects in porphycene and con-
cluded that the two intramolecular hydrogen bonds in porphycene are behaving
cooperatively [20]. In contrast, no cooperativity is observed in porphyrin [21].
S Se S OX
S Se S
O
X = H : κ-H3(Cat-EDT-ST)2’ H-ST O S
S Se
X = D : κ-D3(Cat-EDT-ST)2’ D-ST
S Se
XO S
Figure 7.3 Structure of k-H3 (Cat EDT-ST)2 . Source: Taken from Ref. [17] with permission
from Elsevier.
Kawashima and Tachikawa [22] performed a path integral molecular dynam-

ics study on maleate anion. The result was a barrier-less proton transfer and a
C2 V symmetry of the hydrogen bonds. Deuteration led to rather small effects.
For experimental results see Section 7.14.
A series of intramolecular hydrogen bonded systems have been calculated using
the ab initio multi-component molecular orbital method. A shortening of the OH
bonds was obtained and an elongation of the O· · ·O distance was found upon deuter-
ation. However, no correlation was seen between the change in the OH bond lengths
and the O· · ·O distances [23]. As a matter of fact the latter are not varying very much.
Kanematsu and Tachikawa [24] have calculated deuterium isotope effects on
chemical shifts in picolinic acid N-oxide (PANO) and acetyl acetone using a
multicomponent hybrid DFT with the polarization continuum model. In case of
PANO, they managed to obtain values very close to the experimental ones, whereas
for acetyl acetone the primary deuterium isotope effect is calculated too high. In
case of PANO, the fit to experiment was better than in Ref. [10].
Gräfenstein has formulated a theory regarding vibrational holes and isotope
effects on chemical shifts and investigated o-hydroxybenzaldehydes. The changes
are mainly a direct response to the anharmonic molecular potential at the H site,
whereas the geometry change in the HC=O moiety is a response to the displacement
of the H atom [25, 26].
7.4 Hydrogen Bond Types
When discussing isotope effects hydrogen bonding must be differentiated into intra-
and intermolecular hydrogen bonds. Among intramolecular hydrogen bonds two
types exist, those with a conjugation between the donor hydrogen and the accep-
tor atom and those not. The latter is often akin to intermolecular hydrogen bonds.
Typical examples of the first type are o-hydroxy acyl aromatics also referred to as res-
onance assisted hydrogen bonds (RAHB). Other types can be charge-assisted hydro-
gen bonds. The RAHB type is an example of a case with electronic contact between
donor and acceptor, whereas in the protein the effect is usually, that the scaffold
makes sure the donor and the acceptor are close.
7.5 Deuterium Isotope Effects on Chemical Shifts
NMR is a very useful tool in the study of isotope effects. The effects can be on
chemical shifts or on coupling constants. However, the latter are not so relevant
in the hydrogen bond context. The most common isotope effect on chemical shifts
is that caused by deuterium. Deuterium, 2 H is normally given as D. Two types of
isotope effects can be studied, primary and secondary isotope effects. The primary
deuterium isotope effect is defined as:
P
Δ = 𝛿H − 𝛿D (7.1)
In case of tritium (T) D is changed to T. Secondary isotope effects are observed on

another nucleus. These are typically 1 H, 13 C, 15 N, 18 O, and 19 F. These are the more
common ones and the four latter are the preferred ones as they have large chemical
shift ranges. Secondary intrinsic isotope effects are defined as
n
ΔXint = 𝛿X(l) − 𝛿X(h) (7.2)
l is the light and h the heavier isotope, n being the number of bonds between the
isotope and the nucleus X. This description is for intrinsic isotope effects. Intrinsic
is opposite to equilibrium isotope effects (see later).
Isotope intrinsic effects on chemical shifts can be studied both in intermolecular
and intramolecularly hydrogen bonded systems. A prerequisite is slow exchange
of the label. Suitable solvents are CDCl3 , CD2 Cl2 , benzene-d6 , toluene-d8 , dry
acetone-d6 , and DMSO-d6 . However, it is possible to measure isotope effects in
protic solvents, but in that case, several measurements with varying deuterium
content have to be made and extrapolations to 0% and 100% deuterium content
are made [27]. The measurements can be made at very low temperatures and
intermolecular hydrogen bonded systems can be investigated as demonstrated
beautifully by the groups of Denisov and Limbach [1].
7.6 Intramolecular Hydrogen Bonds
Intramolecular hydrogen bonds in different systems can be investigated as shown in

Figures 7.4–7.6.
H H
X Y
X Y
C
C R
R ? R = H, alkyl, O or N-alkyl or
aryl
X = O or S
Y = O, S or NR
Z Z
(a) (b)
H
H – X Y +
+
X Y –
C
R
C
R
Z
Z
(c) (d)
Figure 7.4 Intramolecular hydrogen bonds of RAHB type in aromatic systems.

O H + O H
O H O H O
X N R X N N R
X N R X N R
R
α β ´R Rʺ ´R Rʺ
´R Rʺ ´R Rʺ
F H + O H
O H S H N R R C N X
X S N R X N R R N
´R Rʺ ´R
´R Rʺ ´R Rʺ
O H
X N C=O-R
α β
´R Rʺ
Figure 7.5 Intramolecular hydrogen bonds in systems having a NH donor (the double bond
could of course also be part of an aromatic system).
7.6.1 Two-Bond Deuterium Isotope Effects on 13 C Chemical Shifts

Two-bond deuterium isotope effects on 13 C chemical shifts (two-bond deuterium
isotope effects [tbdie]) have been used extensively to study hydrogen bond effects
[1, 4, 6] and have been related to hydrogen bond strength and hydrogen bond energy
(see Section 7.13).
Trends valid for tbdie can be found in Figure 7.7. The larger the bond order
between the donor and the acceptor, the larger the tbdie. This is very clear in a case
like 10 in Figure 7.7. It can also be seen, that the tbdie and the OH chemical shifts
are proportional. Steric compression plays a role as seen in the comparison of salicy-
laldehyde, tbdie = 0.227 ppm [32] and 6 of Figure 7.7. In this case, steric compression
leads to a stronger hydrogen bond. However, in cases in which the substituent at
carbon 6 is large, the C=O group may be twisted out of the ring plane. This will also
be the case of the OH group. This leads to a reduced tbdie and a reduction of the
hydrogen bond strength. Such cases can be identified by a plot of tbdie vs. OH chem-
ical shifts [6]. Two-bond deuterium isotope effect on 13 C chemical shifts, 2 ΔC(D),
have been studied in many systems (for structures see Figure 7.7). For OH as hydro-
gen bond donor the systems tend to become tautomeric except for aromatic systems
(see Section 7.14). Non-tautomeric systems are in this context called static. For
o-hydroxyacyl aromatics an order can be seen: ethyl salicylate: 2 ΔC(D) = 0.19 ppm,
salicylaldehyde 0.227 ppm; o-hydroxyacetophenone 0.276 ppm [32] and ultimately
2,4,6-trihydroxy-1,3,5-triacetylbenzene 2 ΔC(D) = 0.722 [33]. This illustrates the
span for o-hydroxyacyl aromatics. More examples are found in Figure 7.7 and
in Ref. [34]. The influence of substituents is complicated as demonstrated for
5-nitrosalicylaldehyde (Figure 7.8). It is obvious that the nitro group increases the
acidity of the OH group, which should lead to a stronger hydrogen bond and a larger
tbdie. However, the tbdie of 5-nitrosalicylaldehyde, 0.223 ppm, is approximately the
same as for salicylaldehyde (see above). The reason is that the second resonance
form leads to a reduced double bond character of the C1—C2 bond.
H
O H N
N R N
O O
N H N
N H O
H
H
H F
H H
S H
H H
N N H
H H H
(D)
H H H
H H
H
α
H(C )
H C'
N O
α
H(C )
RH
−O
ϕC'−N···O−C'
R N−
θN−H···O
O
O
H
θC'−N···O
θH−N···O
H(Cαi−1) Ni C'i
C'i−1 H(Cαi+1)
Cα i ψ Ni+1
Ф
O H
R H
Figure 7.6 Typical non-RAHB types of intramolecular hydrogen bonds. The NH2 C=OH
residue is supposed to be part of another peptide chain. Sources: (a) Sigalov et al. [28]/
American Chemical Society. (b) Sosnicki et al. [29]/American Chemical Society.
(c) Abildgaard et al. [30]/Springer Nature.
With multiple substituents, the influence on the double bond system is very obvi-
ous by a comparison of compounds 4 and 3 in Figure 7.7. For 4, the two hydrogen
bonded systems cannot both be linked by a “double bond.” This is the case for 3.
This results in a tbdie of only 0.240 ppm for 4, whereas the tbdies for 3 are 0.304 and
0.332 ppm. Isotope effects through hydrogen bonds are clearly seen in Figure 7.9.
10.03 10.45
H O H O O OH (D) O OH
10.89 11.88 10.40 304
OH (D) H3 C OH (D) H H
226 303 332
HO (D)HO
OH H3C OH
1 2 3 3
H O H 9.72 O H
H O O
10.24 10.44
11.80 10.10 H O H O
H O 11.95 11.96
O OH (D) 14.02
OH(D) H3C OH (D) OH (D)
240 HO
9.81 H 420 290 210
O H O 2N NO2
OH 6
4 CH3 7
H O H 5 OH O
10.90
10.08 H O 12.99
H H 10.77
OH (D)
O 390 O
170 640
OH (D) 10.32 OH OH (D) 14.60
8 9
H O 10
11.17 9.97
H H O 15.27 H O
O H (D)
O 850 OH (D) 10.94
220
H
H3C O 2 OH
CH3 12
H C
CH3
11
Figure 7.7 Two-bond deuterium isotope effects on 13 C chemical shifts given in ppb.
Source: Hansen et al. [31], figure 1 (p. 3)/MDPI/Licensed under CC BY 4.0.
H O Figure 7.8 Resonance forms of

H O 5-nitrosalicylaldehyde.
+ O H
O H
2 1
+
N O
N+ O –
– O
–
O
o-Hydroxy thio acetophenones [36] have also been studied as seen in Figure 7.10.
It is obvious that the isotope effects at the C=S carbon again are very negative. It
has been suggested that this four-bond deuterium isotope effect in these cases are
transmitted via the pi-bonds rather than via the hydrogen bond as the H· · ·S distance
is rather long. This is supported by the finding that for the 5-nitro derivative (see
Figure 7.10) the effect at the C=S carbon is small compared to the two-bond isotope
effect as the nitro group favors a resonance form with less double bond character
between C-1 and C-2. A plot of 4 ΔC(D) vs. the C-1, C-2 bond length (see Section 7.15).
Figure 7.9 Deuterium isotope effects on 13 C H3C CH3

chemical shifts in ppm. Data for similar –0.07 0.17
compounds and the corresponding oxygen H3C
C
analogs are also given. Source: Elias et al. [35]/
John Wiley & Sons. S 1.08
–0.19 C
S O
N –1.1 H (D)
S
–0.02
(D)
13.15 14.21 13.87
(D) (D)
H H
H
O S O S
O S
0.525 –0.599 0.34 –0.113
0.348 –0.219
2 1 CH3 CH3
CH3
H3CO OCH3
NO2
Figure 7.10 Deuterium isotope effects at 13 C chemical shifts of o-hydroxythio-

acetophenones. Source: Data from Ref. [35]. More examples are found in this reference.
The difference between C=O and C=S as acceptor can also be seen in flavones and
thioflavones [37].
Other systems investigated are 10-hydroxybenz[h]quinolines. The rather modest
tbdie of 0.286 ppm of 10-hydroxybenz[h]quinolone itself can be compared with
mesomerically electron-withdrawing in position 7: nitro 0.534 ppm, aldehyde
0.400 ppm and decanoyl 0.394 ppm [38]. This effect is very different from that seen
in the o-hydroxyacylaromatics (see above) and can only be ascribed to the fact that
the acceptor is a basic nitrogen.
For NH as donor, tautomerism is much less likely and more situations can be inves-
tigated as “static.” NH hydrogen bond donor can be of different kinds, enaminones,
o-hydroxySchiff bases, azo-hydrazo compounds, pyrroles, and other similar hetero-
cycles.
For examples, NH of o-hydroxy Schiff bases [39] enaminones [40, 41].
For the deuterated compounds, the tbdie at C-8 is 0.320 ppm for “a” of Figure 7.11,
but only 0.250 ppm for “b” as seen in Table 7.1. The value for “a” is similar to that
12 O
11
O
10 N
7 8 N N 13 H O
HH N
O O O O H
1
5 O O
O O
O
(a) (b )
Figure 7.11 Phenylenediamine derivatives of dehydracetic acid. Source: Jednacak et al.

[40]/Croatian Chemical Society.
Table 7.1 Deuterium isotope effects on 13 C chemical shifts in

ppb of a and b of Figure 7.11.
Atom 4 5a)
1(1′ ) b) 600.0
2(2 )′ b) 700.0
′
3(3 ) −193.1 −187.5
4(4′ ) 100.1 375.0
′
5(5 ) 108.0 −150.0
6(6′ ) b) 100.0
7(7 )′ b) −350.0
8(8′ ) 320.9 250.0
′
9(9 )
10(10′ ) b) 500.3
11(11′ ) −132.5 b)
′
12(12 ) −107.9 500.5
13(13′ ) b)
a) Prepared in a mixture of deuterated methanol and chloroform.

b) Broad signal.
Source: Jednacak et al. [40] with permission from the Croation
Chemical Society.
of the enaminone in Figure 7.12, whereas the value for “b” is much too small.
Furthermore, as seen in Table 7.1 many very long-range isotope effects are found,
which clearly indicates that b is tautomeric (for tautomerism see Section 7.14).
In 3-methyl-4-amino-3-penten-2-one the isotope effects of the hydrogen bonded
NH(D) is 0.29 ppm and that of the non-hydrogen bond one 0.10 ppm illustrating the
importance of hydrogen bonding. In the corresponding 4-amino-3-penten-2-one
Figure 7.12 Deuterium isotope effects on 13 C –0.06 0.112

chemical shifts in ppm of an enaminone. O H
H3C N
0.043
H CH3 0.355
0.03
0.02
O C2H5
O N
–0.10 0.04 0.03
N O C2H5
O
H N –0.06
0.14 N H N N
0.12
O 0.06 O 0.04
H3C H3C
Figure 7.13 Deuterium isotope effects on 13 C chemical shifts in ppm. Left major form,
right minor form. Source: Kurutos et al. [44]/John Wiley & Sons.
the values are 0.21 and 0.10 ppm [42]. In the former steric strain is present, but not
in the latter. This underlines the importance of steric strain.
The results of the dehydracetic acids of Figure 7.11 can also be compared to those
of derivatives of Meldrum and tetronic acids. In those cases, the tbdie are relatively
small, ∼0.15 ppm, but no large long-range effects are seen. For the tetronic acids two
different isomers are seen (see Figure 7.29 for the mother compound) [43].
For the hydrazo compounds in Figure 7.13, it is seen that the hydrogen bond to the
pyridine nitrogen is slightly stronger than in the minor one. In the latter, a typical
isotope effect over the hydrogen bond is seen.
Another study of deuterium isotope effects at 13 C in which the acceptor is a C=S is
found in thiophenoxyketenimines (Figure 7.14). The tbdie is rather small, whereas
the effect at the C=S carbon is large and negative [45]. This could be an indication
of a tautomeric equilibrium as shown in Figure 7.14. However, the fact that the both
the calculated 13 C chemical shifts and the deuterium isotope effects on 13 C could be
fitted based on a “static” structure shows that no tautomeric equilibrium is at play.
Other systems analyzed are 5-acyl-3-methylrhodanines. They exist as exocyclic enols
showing tbdie of ∼0.4 ppm [46].
0.063 0.14
CH2 CH3 CH3 CH3
+ H3C N H3C N
H3C N H ´R N H
H
–0.1
S S S S
– –0.67 H
–0.14
Figure 7.14 Resonance and tautomeric forms of thiophenoxyketenimines. Source: Data

from Ref. [44].
7.6.2 Long-Range Isotope Effects

Long-range isotope effects can be of several types. Some are transmitted via the
bonds, typically conjugated systems, others are transmitted via hydrogen bonds, and
finally, some are through space [29].
These can be of two different types. One type is seen in Figure 7.12 which is prob-
ably a combination of transmission via sigma-pi bonds combined with transmission
via the hydrogen bonds.
As seen in Figure 7.15 the long-range deuterium isotope effects on the 1 H aldehyde
protons do depend on the O· · ·O distance.
The second type is found in DNA and RNA in Figure 7.16. In this case, the NH(D)
isotope effect is seen at C-2 and transmitted via the hydrogen bond.
Hydrogen bond isotope effects on 13 C chemical shifts can be observed from H-3 to
C-2 between adenine and thymine respectively uracil. The isotope effects on chemi-
cal shifts are found to be sensitive to the N1-N3 distance suggesting that the isotope
effect is sensitive to hydrogen bond strength (see Figure 7.16) [47, 48].
Isotope effects may also be seen on other nuclei like 19 F. This is the case in a
5-fluorocytidine modified oligonucleotide when this is dissolved in a mixture of
1.712
3 1.701 (D)H 1.655
O O H O
H O (D)H
H(D) 2
H HO O
H3C O
O H
HO
H 3C OH
H H OH O
O
H O 5
3
2
6.9 1.665
3.2 1.729 H O 9 1.684
H O H(D) H O
H(D) (D)H
O
H3C O O
OH
O 20
CH3 H
6 9
20
H
1.633
O
H(D)
O
10
Figure 7.15 Deuterium isotope effects at the aldehyde proton chemical shift in ppb.
The numbers in red are OH· · ·O=C distances in Å. Source: Hansen et al. [31], figure 5
(p. 5)/MDPI/Licensed under CC BY 4.0.
7.7 Biological Systems 185
Figure 7.16 Adenine and thymine with hydrogen H

bond. O N
NH2
H N3
N H
1N
2 O
N N
H2 O/D2 O. This was ascribed as not a simple solvent effect but due to hydrogen
bonding to the NH2 (D2 ) group [49]. Long-range effects from a NH(D) group on 19 F
are also found for compounds like the one seen in Figure 7.6b [29].
7.6.3 One-Bond Deuterium Isotope Effects on 15 N Chemical Shifts

in Solution
One-bond deuterium isotope effects on 15 N chemical shifts, 1 Δ15 N(D), in resonance
assisted systems can be rather large. This is seen in enaminones (for the structure
of an enaminone see Figure 7.12). Values for the hydrogen bonded nitrogen is
typically above 1 ppm, whereas it is only ∼0.6 ppm for non-hydrogen bonded
ones [50]. Furthermore, the hydrogen bonded ones depend on the strength of the
hydrogen bond. 1 Δ15 N(D), vary somewhat depending on the protonation state.
This is seen by a comparison of values in DMANH+ (for a structure see later
Figure 7.23c). 1 Δ15 N(D) = ∼0.3 ppm [51]. This value is similar to ammonium ions
(see Figure 7.37), whereas for amines the value is typically ∼0.6 ppm. Examples
involving proteins are discussed in Section 7.7.
7.7 Biological Systems

One-bond deuterium isotope effects on 15 N chemical shifts are very useful in the
study of hydrogen bonding in biological systems, as the majority of the hydrogen
bonds have a NH or a NH+ donor. The former in peptides, DNA, and RNA, the latter
in side-chains of lysines and arginines.
7.7.1 Proteins
One-bond deuterium isotope effects on 15 N chemical shifts can be studied in proteins
to estimate hydrogen bonding. The exchange of the NH proton can normally be done
although some C=ONH units may be well protected in the inner part of the protein.
For those exposed to solvent, solvation may have to be taken into account. Theoret-
ical calculations (see Section 7.3) are very useful to gather information about which
parameters to take into account for proteins. Results for N-formylglycine amide with
a water attached (see Figure 7.17) are given in Table 7.2. It is clear that both the ϕ and
the Φ angles matters, but most of all the N–H· · ·OH2 distance. The shorter this dis-
tance (the stronger the hydrogen bond) the smaller is 1 ΔN(D). This finding is similar
to that observed for RNH3 + · · ·OH2 in which RNH3 + could be a lysine side chain [52].
OH2 Figure 7.17 N-Formylglycineamide with water

attached.
H O
O N
NH2
Table 7.2 Quantum–mechanical calculations for 1 ΔN(D) of N-formylglycineamide with

water attached.
Dihedral angles RN–H· · ·OH2 (Å)
𝚽(∘ ) 𝚿(∘ ) 1.8 1.9 2.0 2.2 2.4 2.6 2.8 ∞
180 180 0.53 0.56 0.59 0.63 0.66 0.68 0.79

180 90 0.43 0.48 0.53 0.94
90 180 0.43 0.48 0.54 0.94
90 90 0.43 0.49 0.54 0.97
Details of the QM calculations are provided in the “Materials and Methods.”

Source: Taken from Ref. [30] with permission from Springer.
Backbone CONH(D) isotope effects have been measured in ubiquitin and corre-
lated to backbone angles, but also to a term containing both electric field effects and
anharmonicity [30]. Based on these results Eq. (3) was formulated:
1
Δ15 N(D) = AΦ′ + BΨ′ + C cos(qN−H···O )∕RH···O + D (7.3)
where Φ′ = cos(Φ + 90∘ ), Ψ′ = cos(Ψ − 70∘ ), A = 0.058 ± 0.008 ppm, B = 0.041 ±
0.005 ppm, C = 0.05 ± 0.02 ppm Å, and D = 0.65 ± 0.01 ppm. Symbols are those of
Figure 7.6c.
It is important to notice that experimental distances were optimized using
DFT calculations of BPW91/6-31G(d) type. This led to the plot of Figure 7.18 for
non-charged amino acids. On the other hand, this indicates that a comparison of
isotope effects of a molecule with charged side chains vs. one at a suitable pH with
a minimum of charges could lead to structural information.
Zhang and Tugarinov [53] arrived at a different equation for 1 ΔN(D).
1
ΔN(D) = 689 + 13 sin(𝜙 + 117∘ ) + 43 cos(𝜓 − 60∘ ) + 16 cos (ΘN−H···O ) (7.4)
Isotope effects in ppb. Glycines are excluded. ΘN–H· · ·O is the angle formed between
N-H and the direction of a hydrogen bond in crystal structures of ubiquitin and GB1.
Tugarinov has also formulated equations for 4 ΔN(D) with deuterium at nitrogen
or carbon [54]. In addition, a recent extensive review covers several other correla-
tions [55].
For primary amides of the side chains of glutamine and asparagine, the
difference in 1 ΔN(D)E and 1 ΔN(D)Z can be correlated with hydrogen bond
7.8 Intermolecular Hydrogen Bonds 187
0.74
0.72
Δ N(D) (predicted), ppm
0.70
0.68
0.66 I
V
1 15
0.64 F
L
0.62 G
A
0.60
0.60 0.62 0.64 0.66 0.68 0.70 0.72 0.74
1 15
Δ N(D) (experimental), ppm
Figure 7.18 Plot of predicted one-bond deuterium isotope effects on 15 N chemical shifts
vs. experimental ones for amino acids with aliphatic side chains of ubiquitin. Source: Taken
from Ref. [30] with permission from Springer. Letters refer to amino acid notations.
interactions, especially if the acceptor is charged. The effects can also be related to
more complicated hydrogen-bonding networks that involve bifurcated hydrogen
bonds [56].
7.7.2 Deuterium Isotope Effects on 1 H Chemical Shifts

Deuterium isotope effects on chemical shifts are small but can be useful in cer-
tain cases. For primary amides, tbdie can be seen from one amide proton to the
other. This was used in the cytosine deaminase complex with the transition state
analog 5-fluoro-2-pyrimidone to determine bifurcated hydrogen bonds in which
asparagines were involved [57].
Histidines in the protonated state, as is the case in phospholipase A2 with a tran-
sition state analog inhibitor revealed different deuterium isotope effects at the NH’s
over four bonds supporting that one of the hydrogens were involved in a hydrogen
bond of low barrier hydrogen bond type. This was also supported by the high NH
chemical shift of 18 ppm [58].
7.8 Intermolecular Hydrogen Bonds

A simple scheme of intermolecular hydrogen bond motifs is seen in Figure 7.19.
It should be noticed that the motifs in Figure 7.19 are very similar to some
intramolecular hydrogen bonds except in that case the molecular scaffold makes
sure that the donor and the acceptor are in the same molecule.
O N O N N
H H H H H
O O O O N
C C C C CH
N N
H H
H
C
O O O N
O
H H H
H
O O O – N+
H
C C CH
C
O O O
Figure 7.19 Intermolecular hydrogen bond motifs.
The Limbach group has investigated isotope effects on chemical shifts in inter-
molecular hydrogen bonded systems using cryosolvent (CDF3 /CDF2 Cl) to slow
down exchange of the label (typically deuterium). An interesting and important
feature of freon cryosolvents is that the dielectric constant increases as the tem-
perature is lowered. This has led to a very large amount of data and development
of complex theories reviewed in Ref. [1]. Typical molecules involved have been
substituted with pyridines and simple carboxylic acids as seen in Figure 7.19.
This opens for studies of one-bond deuterium isotope effects on 15 N and tbdie on
13 C=O as well as primary deuterium isotope effects on chemical shifts.
The Limbach group has based its analysis on natural bond orders. The assump-
tions and terminology can be found in the review [1]. The terminology is used
differently from the present review, but it has been shown that the two approaches
lead to similar results [59]. The complexity of the situation is well described in
Ref. [60] as the equations formulated cannot be solved analytically. The question
of a continuous proton transfer vs. a tautomeric situation is described by the
authors as: “So now do these two descriptions fit together? We believe that for
many strong hydrogen bonds these two points of view are not a strict contradiction
and each of them describes a part of the reality” [58]. The present author has great
difficulty envisioning this statement. In the study of Ref. [58], tbdie on 13 C=O
of the system of figure are studied at low temperature. In addition to pyridine,
imidazole and alkyl amines are also included. In a previous study, one bond
deuterium isotope effect on 15 N chemical shifts of collidine has been studied
using acids of different strengths. An additional measure is one bond NH coupling
constants [61].
7.9 Primary Isotope Effects 189
O
O (D) 15
(D) O H N
– 13
H O H H3C R
H O
H (D) O
N +
(a) (b)
Figure 7.20 (a) Diisopropylamine-formate system. Source: [62]/John Wiley & Sons. (b)
Complex between acetic acid 13 C labelled at the carboxyl carbon and substituted 15 N
labelled pyridines.
A similar system was studied in relation to ionic liquids (see Figure 7.20a). In this
case, primary isotope effects at OH and NH were measured as 2.14 and 0.43 ppm at
243 K and as 1.08 and 0.30 ppm at 193 K [62].
Deuterium isotope effects on 1 H chemical shifts can be observed in cyclic dimers
and trimers of phosphinic acids. These can be used to reveal the stoichiometry of the
cyclic complexes and show cooperativity [63]. Also heterotrimers of phosphoric and
phosphinic acid have been investigated and anticooperativity was found ascribed to
steric factors [64].
Isotope on 1 H chemical shifts of a large number of monodeuterated benzoic acid
dimers have been studied both at low temperature and with the dimers encapsulated
in a capsule 1.24 1. Encapsulation leads to compression of the O· · ·O distances and
increased (numerically) isotope effects [65].
7.9 Primary Isotope Effects

For a definition see Section 7.5. Primary deuterium isotope effects have not been
studied so frequently due to the low sensitivity of deuterium and the often broad
lines. Primary deuterium isotope effects were early on suggested as a way to deter-
mine the hydrogen bond potential. Small positive values indicate an asymmetric
double-well potential of a weak hydrogen bond, a small negative value a single-well
potential and a large positive value a double-well potential of a strong hydrogen
bond. With this definition, the FHF – is a single well potential [66]. This is true
provided no tautomeric equilibrium takes place. In that case, one has to take into
account the chemical shifts between the two sites [4]. See also Section 7.14.
XH chemical shifts and H primary isotope effects both reflect on the hydrogen
bond except that XH chemical shift maybe have to be correct for ring current effects.
A plot of primary deuterium isotope effects vs. OH chemical shifts showed a rather
large spread (Figure 7.21).
As noted in Ref. [6], some of the outliers are compounds with the acceptor group
and the OH proton twisted out of the ring plane. This supports the mention that OH
chemical shifts may have to be corrected for ring current effects.
Figure 7.21 Plot of primary deuterium

Primary isotope
0.6
0.4 isotope effects vs. OH chemical shifts.
effects 0.2 Source: Hansen et al. [31], figure 10
0 (p. 2413)/MDPI/Licensed under CC
BY 4.0.
–0.2 0 10 20 30
OH chemical shifts
0.9
Primary deuterium isotope effect
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
9 11 13 15 17 19 21 23
NH chemical shifts
Figure 7.22 Primary deuterium isotope effects vs. NH chemical shifts. Source: Data from
Refs. [64, 67, 68].
In Figure 7.22 a plot of primary deuterium isotope effects vs. NH chemical shifts is
shown. The compounds studied are proton sponges, 2,3-dipyrrol-2-ylquinoxalines,
and non-tautomeric compounds of RAHB type (see Figure 7.23). The first mentioned
are positively charged, the second negatively charged, whereas the last group is neu-
tral. The last group fits nicely with the description of small primary isotope effects.
The 2,3-dipyrrol-2-ylquinoxalines fall clearly outside a possible correlation.
R R1
H 3C (D)
CH3
H
H3C
N N N N CH3
+
–
N N
H
(D)
(a) (b)
Figure 7.23 (a) 2,3-Dipyrrol-2-ylquinoxalines. R and R1 are H,H; H,NO2 , and NO2 ,NO2 . The
primary isotope effects are 0.86; 1.13 and 0.88 ppm (signs changed compared to original
paper). (b) DMANH+ . Source: Pietrzak et al. [69]/John Wiley & Sons.
7.10 Isotope Effects and Acidity 191
Series 1 Series 2 Series 3 Series 4

1.6
1.4
1.2
Primary isotope effects
1
0.8
0.6
0.4
0.2
0
–0.212 14 16 18 20 22
–0.4
OH chemical shifts
Figure 7.24 Primary deuterium isotope effects vs. C=OOH chemical shifts. Series 1. Mono
anions of dicarboxylic acids from Refs. [70, 71]. Series 2. Intramolecularly hydrogen bonded
compounds (see Ref. [34]). Series 3. Monoanions of succinic acid and derivatives [72]. Series
4. Citrinin (see Ref. [34]).
The data are part of the plot in Figure 7.24. In this figure, several primary iso-
tope effects for acids are plotted vs. C=OOH chemical shifts. In the plot are included
data for intramolecular hydrogen bonds, intermolecular effects from acid dimers,
and from acid–acid anion complexes. A spread is observed and nothing similar to
Figure 7.22.
7.10 Isotope Effects and Acidity

Perrin [73] has written a comprehensive review on secondary equilibrium isotope
effects on acidity demonstrating how isotope substitution will influence the pK a val-
ues of acids.
7.10.1 Isotope Effects to Determine Protonation States

Deuterium isotope effects on 15 N chemical shifts were investigated in amino acids
as a tool to estimate the protonation state of amino acid side chains. The idea
was to take advantage of the difference in one-bond deuterium isotope effects
on 15 N chemical shifts of NH vs. NH+ groups. The former are typically 0.65 ppm
vs. 0.3 ppm for NH+ of ammonium ions [74]. The study was done in a blocked
tripeptide acetyl-Gly-X-Glyamide. For N-eps at low Ph, the deuterium isotope
effects at 15 N were 1.0 ppm whereas at high pH, 15.25, it was 1.8 ppm. For N-η
it was 1.4 respectively 1.7 ppm. These effects, as noted by the authors, are a
combination of one-bond and long-range isotope effects. The results for N-η at low
pH are different from the 1 ΔN(D) values determined by Mackenzie and Hansen,
0.61 ppm (0.301 ppm per deuterium) [75]. For lysines McIntosh et al. [67] found
1.05 ppm at pH 4. This value is similar to those reported earlier [76]. However, when
saying similar one has in mind that these values depend on the counter ion [77].
At pH 15.25 a value of ∼1.9 ppm was obtained.
7.11 Solvent Isotope Effects and Exchange Rates

Solvent isotope effects can be seen on the 19 F chemical shifts of the MgF3 − fluorines
of the cAPK-ADP-MgF3 —SP20 TSA complex as well as at the corresponding AlF4 −
and BeF3 − complexes. The isotope effects for the MgF3 − complex are 1.4, 0.3, and
0.2 ppm for the three fluorines. For the other two, the effects are 0.8, 0.6, 0.4, and
0.2 ppm and 0.8, 0.2, and 0.1 ppm, respectively. For MgF3 − , the authors suggested
that the large isotope effects are due to fluorines hydrogen bonded to water [78] or it
could be due to hydrogen bonding to the NH of S53 and S17, which will be deuterated
in the H2 O/D2 O mixture.
Solvent isotope effects are seen in the 4-F-phenylalanine of hormone glucagon,
ΔF(D) = 0.15 ppm, which is due to full access and hydrogen bonding [79].
Shi et al. [80] introduced a trifluoromethyl phenylalanine giving a solvent isotope
effect of ∼0.1 ppm. As the trifluoromethyl group is less good to form hydrogen
bonds to the solvent, so smaller effects should be expected. Bai et al. [81] report a
value of 0.199 ppm for 4-F-phenylalanine of a-lactoalbumin as 99.3% exposure.
The fluorinated L-4,4,4,-trifluoroethylglycine was introduced into peptides and
solvent isotope effects were measured. The largest effect found was 0.11 ppm [82].
An acylamino-7-fluorofluorene attached at C-8 of a deoxyguanosine inserted into
DNA stands showed a maximal solvent isotope effect of 0.24 ppm [83].
Combined with the effects originally reported for 5-fluorotryptophan 0.23 ppm
and a series of substituted fluorobenzene [27] a pattern can be seen. The higher the
electron density the higher and larger the solvent isotope effect and the stronger the
hydrogen bond.
A change of solvent from pure H2 O or D2 O to a mixture of H2 O/D2 O may give
rise to observable isotope effects depending on the exchange rate. Typically, the OH,
NH, or SH protons are exchanged. Exchange rates have been measured for tyro-
sine OH protons [84, 85], OH protons of threonines, as well as SH protons of cys-
teines [78]. This method has also been used extensively for protein backbone NH
protons.
7.12 Exchange in the Solid-State

A microcrystalline sample of perdeuterated α-spectrin SH3 domain was investigated
using solid-state NMR the sample was crystallized from a buffer containing either
a 1 : 1 or 1 : 9 H2 O/D2 O mixture. Exchange was observed for six residues with
exchange rates from 0.2 to 5 s−1 . In case of N35 and N38 hydrogen bonds exist
between the side chain C=ONH2 NH protons and backbone C=O. Exchange
occurred also for the backbone NH or K60 and S36 and for the W41. 1 ΔN(D)’s
were measured for K60 as 0.65 ppm, for S36 as 0.36 ppm, and for the side chain NH
of N35 as 0.4 ppm. Values for R21 were not given but showed a very small value
according to the spectra [86]. Both the value for S36 and R21 are much smaller than
7.13 Hydrogen Bond Energies 193
expected for non-hydrogen-bonded NH groups [30]. This could possibly indicate

some exchange of labels.
7.13 Hydrogen Bond Energies

Reuben [87] formulated a correlation between hydrogen bond energies and tbdie on
13 C chemical shifts based on the work of Schaefer [88]. Schaefer pointed to some
caution. Reuben formulated the equation:

ln(2 Δ) = 2.783 + 0.354∗ E (7.5)
in which E is the hydrogen bond energy in kcal mol−1
and the two-bond isotope
effect is in ppb. The equation has been used in proteins [89, 90]. Recently, tbdie have
been correlated to hydrogen bond energies in o-hydroxy aromatic aldehydes [31] in
which the hydrogen bond energies were calculated by the hb and out method [91]
and also by electron densities at the ring critical points [92]. The use of tbdie has
the advantage that no reference is needed as is the case when using 1 H chemical
shifts [93]. Furthermore, as Eq. (7.5) is based on experimental data, this is a way to
relate calculated values to experimental ones.
The ring critical points were calculated using the B3LYP/6-311++G(d,p) func-
tional [94] and the AIM program [95, 96]. Considering the correlation between the
TBDIE on13 C and the electron density at the bond critical point (Figure 7.25) it is
obvious that the large isotope effect is correlated to a stronger hydrogen bond. A sim-
ilar plot was shown for enamino derivatives having ketones, esters, and nitro groups
as acceptors and both linear and cyclic structures and substituents at nitrogen both
aliphatic (methyl and t-butyl) and aromatic. In this plot, data for the linear and the
cyclic compounds fell on separate lines [97].
0.7
0.6
y = 15.666x –0.4223
0.5 2
R = 0.83
TBDIE
0.4
0.3
0.2
0.1
0
0.03 0.035 0.04 0.045 0.05 0.055 0.06 0.065
Electron density at bond critical points
Figure 7.25 Two-bond deuterium isotope effects of o-hydroxy aromatic aldehydes vs.
electron densities at bond critical points. Source: Hansen et al. [31], figure 8
(p. 7)/MDPI/Licensed under CC BY 4.0.
7.14 Tautomerism
Tautomerism and intramolecular hydrogen bonding are very closely connected as
seen in Figures 7.34 and 7.36. Using NMR and isotope effects have been reviewed
several times [98–101].
For tautomeric systems, the change in the equilibrium upon isotope substitution
has to be taken into account when discussing isotope effects on chemical shifts. This
can be expressed in the following way:
n
ΔX(D)int = (1 − xD )n ΔX(D)OH + xD n X(D)NH (7.6)
n
ΔX(D)eq = (𝛿XNH − 𝛿XOH )Δx (7.7)
n n n
ΔX(D)OBS = ΔX(D)int + ΔX(D)eq (7.8)
In Eqs. (7.6–7.8) xD is the mole fraction, Δx, the change in the mole fraction upon
deuteration, n ΔX(D) is the isotope effect for nucleus X due to deuteration n bonds
away. Int, eq, and OBS refer to intrinsic, equilibrium, and observed. An important
feature of equilibrium isotope effects is the fact that they can be found far from the
site of isotope substitution and they can be of both signs.
In case of 2-acetyl-1,8-dihydroxy-3,6-dimethylnaphthalene, a tautomeric equi-
librium is also seen in the liquid state [102]. Tautomerism is important in
biological systems. o-hydroxy Schiff bases have been investigated in great detail.
Limbach et al. [103] used N-(3,5-dibromosalicylidene)-methylamine and similar
Schiff bases as models in the study of transamination instead of pyridoxal-5′ -
phosphate. This was recently discussed in [104]. Tautomerism was also studied
in N-(pyridoxylidene)-methylamine. In this case, a deuterium isotope effect at N2
(Schiff base nitrogen) of −3.09* ppm and in the 1 : 1 complex with trifluoroacetic
acid an effect of 3.54* ppm. This was interpreted as deuterium being primarily at
oxygen in the former case and primarily on nitrogen in the latter [105]. Tautomerism
is also observed as Schiff base with a carboxylic acid ortho to the OH group, which
enables different hydrogen bond patterns (see Figure 7.26) [106].
Tautomerism of Schiff bases modified with amino acid ionic liquid structure [107]
or amino acids [108, 109] can be monitored based on deuterium isotope effects at
C-2 (Figure 7.27). The hydrogen bond with the acid group clearly influences the
equilibrium but also from the OH group of threonine.
H H Ph H H + Ph H H Ph
– – +
O O N O O N O O N
O O O
(a) (b) (c)
Figure 7.26 Schiff bases with a carboxylic acid group in ortho position.
7.14 Tautomerism 195
Figure 7.27 Example of a Schiff O O

base modified to contain an amino
acid. N O– O–
N
NH N +
NH N
H H
O O
–
H H
O O
12 –O N N +
H
7 9 N 11 N O
H
N N
S S
O O O O
Figure 7.28 Tautomerism of piroxicam.
Piroxicam (see Figure 7.28) is an NSAID drug. Deuteration studies showed pro-
portionality between isotope effects on 13 C chemical shifts and chemical shift dif-
ferences except for the carbons close to the site of protonation (C-7, C-9, C-11, and
C-12). See discussion about equilibrium isotope effects (Section 7.14) [110]. Interest-
ingly, the zwitterionic structure is the result of a reorientation and formation of new
hydrogen bonded structures.
Isotope effects have also been used to show that a compound (1,1′ ,1′′ -(2,4,6-
trihydroxybenzene-1,3,5-triyl) triethanone) was not tautomeric [111] although
tautomerism had been claimed [112]. A complex system is found in 3-acyltetronic
acids [113] (Figures 7.29 and 7.30).
The dicarboxylic acid anion system (see Figure 7.31) has been subject to much
discussion. Is the potential well a single or a double-well?
The systems analyzed are the cyclohexene-1,2-dicarboxylate monoanion
[114, 115], the phthalate monoanion [116], the difluoromaleate monoanion [107],
and the mono anion of a di-tertbutylsuccinate [117, 118]. Perrin et al. used the
method of isotopic perturbation originally suggested by Saunders et al. [119]. Perrin
et al. used 18 O substitution to create the isotopic perturbation. The isotope effect
can be expressed as an intrinsic effect Dint and an equilibrium effect Deq . They are
discussed in Section 7.15.
DOBS = Dint + Deq (7.9)

An important argument for the existence of an equilibrium is the finding that the
observed isotope effect change with temperature. Bogle and Singleton argued based
on calculations that desymmetrization could explain that the 18 O isotope effects
H O H O
O 6 R1 O R1
4 3
R3 5
1
2 R3
O O
R 2 O R2 O
R1 R2 R3
(a) ( b)
Me H H
Et Me Me
i
Pr Ph
t
Bu R1
R1
Ph
O O O O
Bz
H
R3 H R3
O O
R2 O R2 O
(c) (d)
Figure 7.29 Tautomeric equilibria of 3-acyltetronic acid. Source: From Ref. [113] with
permission from Elsevier.
Figure 7.30 Deuterium isotope on 13 C

0.114
0.110
0.109
0.109
D O 0.571
0.556
0.578
0.578
chemical shifts in ppm of 3-acyltetronic acid
0.574
0.566
0.570
0.519 in which R of Figure 7.29 is t-butyl. Source:
O –0.105
–0.129 –0.023
O –0.078
–0.073 O
–0.138
–0.150
–0.089
–0.105
–0.077
–0.077
From Ref. [113] with permission from Elsevier.
–0.117 –0.170
0.935
0.946
0.971
–0.166
–0.161 D
1.004 –0.166
0.106
0.117
0.0
0.0 O 0.154 0.542
0.530 O
0.121
0.134
O –0.029
–0.036
0.149
0.150
0.150
O 0.526
0.513
H Figure 7.31 Dicarboxylic acid anion. The

O O broken line indicates that this could be a
–
O double bond, a cyclic system, or an aromatic
system.
could be rationalized in terms of intrinsic isotope effects [109]. It has also been
suggested that symmetry can be broken by the nonidentical solvation of the two
identical sites [120–122].
Guo et al. [71] have measured 2 ΔC(D) of the following series of acid–acid
anion species and found the following values in ppm*: hydrogen bis-isobutyrate,
0.14; hydrogen bistrimethylacetate, 0.11; hydrogen bis-acetate, 0.20; hydrogen
bis-chloroacetate, 0.13; hydrogen bisdichloroacetate, 0.11; hydrogen bistriflu-
oroacetate, 0.10; hydrogen rac-dimethyl Succinate, 0.11–0.09 (lowest value at
lower temperature); hydrogen succinate, 0.12–0.11; hydrogen maleate, 0.07–0.05;
hydrogen phthalate, 0.04. Guo et al. proposed a correlation between H/D isotope
effects on carboxylic carbon chemical shifts and the proton transfer coordinate,
q1 = 1/2(rOH − rHO), which allowed an estimate of the hydrogen bond geometries
from the observed 13 C NMR parameters, taking into account the degenerate
TEA+ TMA+ TMA+ TEA+
H H H H H CH3 H CH3 H3C H H CH3 H CH3 H H
O O O OO O O O
– – – –
O H O O H O O H O O H O
1 2 3 4
Figure 7.32 Tetraethylammonium (TEA) hydrogen succinate (1), tetramethylammonium

(TMA) hydrogen meso-dimethyl succinate (2), hydrogen rac-dimethyl succinate (3), and
tetraethylammonium hydrogen R-(+)-methyl succinate (4). Source: Guo et al. [72]/American
Chemical Society.
H
H
O O
O O
H2N NH2
+
N
–
O O
N
(a) (b)
Figure 7.33 (a) β-Diketone. (b) Nitromalonamide.
proton tautomerism. However, for the maleate and phthalate anions, they found
a symmetric hydrogen bond potential, which is contrary to the findings of Perrin
et al., see above.
Guo et al. also studied tetraethylammonium (TEA) hydrogen succinate (1),
tetramethylammonium (TMA) hydrogen meso-dimethylsuccinate (2), hydro-
gen rac-dimethylsuccinate (3), and tetraethylammonium hydrogen R-(+)-
methylsuccinate (4) (see Figure 7.32) and its deuterated analogs [72].
β-Diketone systems have also been investigated 4-cyano-2,2,6,6-tetramethyl-3,5-
heptanedione enol (a) and nitromalonamide(b) (see Figure 7.33). Perrin et al.
used again 18 O to create isotopic perturbation of equilibrium. For a tautomeric
equilibrium was suggested whereas for b (nitromalonamide) a single well poten-
tial was suggested. This is contrary to the results of Hansen [123] based on the
deuterium isotope effect (OH is deuterated) at 13 C, who suggested a tautomeric
equilibrium.
7.15 Solid-State NMR
Isotope effects on chemical shifts are not studied so much mainly because lines are
broad in solid-state NMR spectra so diecs are more difficult to observe. A second
difficulty is that deuteration may lead to a different crystal structure. REF Results
are not so abundant. However, for compounds that are tautomeric in the liquid state,
H
H
C
C C C
C C
S O
H S O
H
Figure 7.34 Equilibrium of thiodibenzoylmethane.
0.0
(4)
(1)
–0.3 (2) (6)
(7)
–0.6 (3)
ΔCS(OD)
(5)
–0.9
4
–1.2
(8) y = 16.187x –23.686
R2 = 0.6704
–1.5
1.38 1.40 1.42 1.44 1.46
R(C1-C2) (Å)
Figure 7.35 Plot of 4 ΔCS(OD) vs. the R1-R2 distance for a series of o-hydroxythio-
acetophenones and thiodibenzoylmethane (8). Source: Based on Elias et al. [35].
it can be very useful to know the data for one of the tautomers, as this can be obtained
from the solid-state study.
Deuterium isotope effects on 13 C chemical shifts. In case of an equilibrium, a
large chemical shift difference is clearly an advantage (see Eq. (7.7)). A system
that has shown very large equilibrium isotope effects on 13 C chemical shifts is the
β-thioxoketones [124]. Thiodibenzoylmethane (Figure 7.34) has been studied in the
solid-state. However, the deuterium isotope effect at the C=S carbon of −1.2 ppm
could be explained as an intrinsic isotope effect (see Figure 7.35) [35].
A system somewhat akin to thiodibenzoyl methane (Figure 7.34) is the pyri-
doyl benzoylβ-diketones. These have been studied both in the solid and the
liquid state. In the liquid state, the 4-pyridoyl derivative is on the B-form (OH
on C-1) whereas the 2-and the 3-derivatives are on the A-form. However, in
the solid-state all three derivatives are on the A-form. The 4-pyridoyl deriva-
tive shows some unusual deuterium isotope effects on 13 C chemical shifts in
the solid-state, which is ascribed to a change in the crystal structure upon
deuteration [125].
H H
H H
O O O
O O O
(a) (b)
Figure 7.36 Equilibrium of 2-acetyl-1,8-dihydroxy-3,6-dimethylnaphthalene.
A complex hydrogen bonded system, 2-acetyl-1,8-dihydroxy-3,6-dimethyl-

naphthalene (Figure 7.36) has been deuterated at the OH protons and studied
in the solid-state. An isotope effect of 1.5 ppm at C-1 was found. Based on the
graph of Bolvig and Hansen [126] a mole fraction of 0.8 in favor of form a was
suggested [102].
7.15.1 Deuterium Isotope Effects on 15 N Chemical Shifts

A seminal paper on deuterium isotope effects on 15 N chemical shifts of ammonium
ions in the solid-state was published by Wasylishen et al. [127]. A later study of
ammonium halides in the solid-state resulted in a comparison also to water as seen
in Figure 7.37 [128]. It can be seen that water (coordinates 1.08, 2.78) has a shorter
distance to the ammonium ion than the halides.
1 ΔN(D) of Schiff bases of 2-hydroxynaphaldehyde (Figure 7.38) have been mea-
sured both in solution and in the solid state [129]. R = n-Pr gave a 1 ΔN(D) = 3.3 and
2.7 ppm, R = Ph, 1 ΔN(D) = −1.6 ppm; R = PhOCH3 , 1 ΔN(D) = −0.4 and 0.2 ppm; R =
PhN(CH3 )2 , 1 ΔN(D) = 0; R = quinoline, 1 ΔN(D) = 1.5 and 0; R = Naph, 1 ΔN(D) = 3.0
and 4.8 ppm. Based on the findings that 1 ΔN(D) for OH-form is close to −1.5 and
1.9
1.8
y = 0.6191x –0.4538
1.7
R2 = 0.9651
1.6
ΔN(D)4
1.5
1.4
1
1.3
1.2
1.1
1
2.5 2.7 2.9 3.1 3.3 3.5 3.7 3.9
RN···X
Figure 7.37 The sum of one-bond deuterium isotope effects on 15 N chemical shifts vs. the
N· · ·X distance. Data points are from Ref. The diamonds mark the halides and the squares
Sc-24 and water. These data were also presented in Ref. [128] but the axis titles have been
corrected. Source: Hansen [128], figure 3 (p. 3)/John Wiley & Sons/Licensed under CC BY 3.0.
R R
H N H N
H H
O O
Figure 7.38 Tautomeric equilibrium for Schiff bases of 2-hydroxynaphthalene.
Data from Refs. (27,28,31,32)

L (1-8)
Adducts (L+Rh*)
6
4
ΔN(D) (ppm)
0
n
–2
–4
0 1
Mole fraction of NH-form (χ)
Figure 7.39 Plot of 1 ΔN(D) vs. the mole fraction for o-hydroxy Schiff bases in solution.
Source: Data from Refs. [125, 131–133].
3 ppm for the NH-form [130], and from the graph of Figure 7.39, it could be
concluded that R = n-Pr the Schiff base is on the NH-form, whereas R = Ph is on the
OH-form.
For R = Naph the situation is more complex as two types of molecules exist in
the solid. One being on the NH-form, 1 ΔN(D) = 3.0 ppm and the other being in a
tautomeric equilibrium. Others being part of an equilibrium are compounds with
R = PhOCH3 and R = PhN(CH3 )2 .
The case of R = Quin is shown in Figure 7.40. The extra hydrogen bond as assumed
to give rise to the unusual isotope effect, when taking into account the 15 N chemical
shift [129].
The Schiff bases seen in Figure 7.41 have been analyzed in the same manner
[134]. 1, 2, and 3 exist almost exclusively at the OH-form. 4 and 5 are found to be
Figure 7.40 Quinoline derivative showing

bifurcated hydrogen bonds.
N
+
H N
H
O
–
Figure 7.41 Double Schiff

bases. Source: Dziembowska
et al. [134]/John Wiley & Sons.
(1) R = H
H H (2) R = 3,5-di-CI
N N (3) R = 4,6-di-OCH3
H H
O O
R R
H H
N N
HH (4)
OO
H H
N N
HH
(5)
OO
tautomeric in the solid-state and for 5 the equilibrium is shifted strongly towards the
NH form.
The Schiff bases N-(pyridoxylidene)tolylamine and N-(pyridoxylidene)methyl-
amine are investigated as models for pyridoxal-5′ -phosphate. The Schiff bases are
supposed to be tautomeric as described in Figure 7.42. The 1 ΔN(D) is 3.0* ppm
R R
+
N2 N2 H
H2 2
O2 –
O2
HO HO
N1 CH3 N1 CH3
1b: R = tolyl
2a: R = CH3
+ Acid (AcOH) + Acid (AcOH)
R R
N2 +
N2
H2 H2
–
O2 O2
HO HO
+ +
N1 CH3 N1 CH3
H1 H1
– –
AcO1 AcO1
Figure 7.42 Tautomeric scheme of Schiff bases N-(pyridoxylidene)tolylamine and

N-(pyridoxylidene)methylamine at the neutral and protonated forms. Source: Schagen et al.
[135]/American Chemical Society.
for 1b and 2.9* ppm for 2a of Figure 7.42 [135]. According to Figure 7.40 this is
in agreement with a tautomeric equilibrium close to K = 1. By protonation at the
pyridine nitrogen the isotope effect the isotope effects change to −5.0* ppm for 1b
acidified with 3,5-dinitrobenzoic acid, to −1.8* ppm for 1b acidified with HCl, and
to −0.8* ppm with trichloroacetic acid [130]. The acidification clearly induces a
dramatic change toward the enol form.
A pentachlorophenol-4-methylpyridine complex change crystal form from tri-
clinic to monoclinic when deuterated. This is ascribed to a weakening of hydrogen
bonds. [136]
7.16 Conclusions
Strong hydrogen bonds have been discussed vividly (see Section 7.14). Is the hydro-
gen bond potential a single or a double-well potential? Isotope substitution can be
a useful tool, but no unanimous answer has been reached. Theoretical calculations
can be a strong additional tool to isotope substitution in the description of hydrogen
bonds.
Primary deuterium isotope effects on chemical shifts can tell about hydrogen
bonds provided tautomerism has been excluded.
References 203
Two-bond deuterium isotope effects on 13 C chemical shifts can be related to

hydrogen bond energies.
Isotope effects on chemical shifts are a useful tool in detection and description of
tautomerism in hydrogen bonded systems.
Isotope effects have also been used to monitor the protonation state of amino acids
and to monitor exchange.
References
1 Kohen, A. and Limbach, H.-H. (ed.) (2006). Isotope Effects in Chemistry and
Biology. Boca Raton, FL, US: Taylor and Francis.
2 Grabowski, S.J. (2006). Hydrogen Bonding—New Insights. Dordrecht: Springer.
3 Dziembowska, T., Hansen, P.E., and Rozwadowski, Z. (2004). Studies based on
deuterium isotope effects on 13 C NMR chemical shifts. Prog. NMR Spectrosc.
45: 1–29.
4 Hansen, P.E. (2007). Isotope effect on chemic shifts in hydrogen bonded sys-
tems. J. Label. Compd. Radiopham. 50: 967–981.
5 Denisov, G.S., Mavri, J., and Sobczyk, L. (2006). Potential energy shape for
the proton motion in hydrogen bonds reflected in infrared and NMR spectra.
In: Hydrogen Bonding—New Insights (ed. S.J. Grabowski). Dordrecht: Springer.
6 Hansen, P.E. (2015). Isotope effects on chemical shift in the study of
intramolecular hydrogen bonds. Molecules 20: 2405–2424.
7 Sobczyk, L., Obrzud, M., and Filarowski, A. (2013). H/D isotope effects in
hydrogen bonded systems. Molecules 18: 4467–4476. https://doi.org/10.3390/
molecules18044467.
8 Frey, P.A. (1995). On low-barrier hydrogen bonds and enzyme catalysis. Science
269: 102–106.
9 Cleland, W.W. (2010). The low-barrier hydrogen bonding enzymatic catalysis.
Ad. Phys. Org. Chem. 44: 1–17.
10 Stare, J., Jezierska, A., Ambrožič, G. et al. (2004). Density functional calculation
of the 2D potential surface and deuterium isotope effect on 13C chemical shifts
in picolinic acid N-oxide. Comparison with experiment. J. Am. Chem. Soc.
126: 4437–4443.
Acc. Chem. Res. 43: 1550–1557.
12 Jameson, C.J. (1991). The dynamic and electronic factors in isotope effects
on NMR parameters. In: Isotopes in the Physical and Biomedical Sciences.
Isotopic Applications in NMR Studies (ed. E. Buncel and J.R. Jones).
Amsterdam, The Netherlands: Elsevier.
13 Ubbelohde, A.R. and Gallagher, K.J. (1955). Acid-base effects in hydrogen
bonds in crystals. Acta Cryst. 8: 71–83.
14 Jin, T. and Zhang, W. (2019). Geometric H/D isotope effect in a series of
organic salts involving short O-H…O hydrogen bonds. CrystEngComm
21: 4238–4242.
15 Miao, L.-P., Liang, B.-D., Tong; J. et al. (2021). H/D effects on the short
O-H· · ·O hydrogen bond geometries and temperature-dependent
polymorphism of two organic salts containing hydrogen 2,3,5,6-
tetrafluorophthalate monoanions. Cryst. Growth Des. 21: 2589–2595.
16 Shi, C., Zhang, X., Yu, C.-H. et al. (2018). Geometric isotope effect of
deuteration in a hydrogen-bonded host-guest crystal. Nat. Commun. 9: 481.
17 Yamamoto, K., Kanematsu, Y., Nagashima, U. et al. (2017). Multicomponent
DFT study of geometrical H/D isotope effect on hydrogen-bonded organic
conductor k-H3(Cat EDT-ST)2 . Chem. Phys. Lett. 674: 168–172.
18 Kita, Y., Kamikubo, H., Kataoka, M., and Tachikawa, M. (2013). Theoretical
analysis of the geometrical isotope effect on the hydrogen bonds in photoactive
yellow protein with multi-component density functional theory. Chem. Phys.
419: 50–53.
19 Ishimoto, T., Tachikawa, M., and Nagashima, U. (2006). A fragment molecular-
orbital-multicomponent molecular-orbital method for analyzing H/D isotope
effects in large molecules. J. Chem. Phys. 124: 014112.
20 Shibl, M.F., Piertzak, M., Limbach, H.-H., and Kühn, O. (2007). Geometric
H/D isotope effects and cooperativity of the hydrogen bonds in porphycene.
ChemPhysChem 8: 315–321.
21 Pietrzak, M., Shibl, M.F., Bröring, M. et al. (2007). 1H/2H NMR studies of
geometric H/D isotope effects on the coupled hydrogen bonds in porphycene
derivatives. J. Am. Chem. Soc. 129: 296–304.
22 Kawashima, Y. and Tachikawa, M. (2013). Nuclear quantum effect on
intramolecular hydrogen bond of hydrogen maleate anion: an ab initio path
integral molecular dynamics study. Chem. Phys. Lett. 571: 23–27.
23 Udagawa, T., Ishimoto, T., and Tachikawa, M. (2013). Theoretical study
of H/D isotope effects on nuclear magnetic shieldings using an ab initio
multi-component molecular orbital method. Molecules 18: 5209–5220.
24 Kanematsu, Y. and Tachikawa, M. (2014). Development of multicomponent
hybrid density functional theory with polarization continuum model for the
analysis of nuclear quantum effect and solvent effect on NMR chemical shift.
J. Chem. Phys. 140: 164111.
25 Gräfenstein, J. (2020). The structure of the “vibration hole” around anisotopic
substitution—implications for the calculationof nuclear magnetic resonance
(NMR) Isotopic Shifts. Molecules 25: 2915.
26 Gräfenstein, J. (2019). Efficient calculation of NMR isotopic shifts: difference-
dedicated perturbation theory. J. Chem. Phys. 151: 244120.
27 Hansen, P.E., Dettman, H.D., and Sykes, B.D. (1985). Solvent induced
deuterium isotope effects on 19 F chemical shifts of some substituted
fluorobenzenes. Formation of inclusion complexes. J. Magn. Reson. 62: 487–496.
28 Sigalov, M.V., Pylaeva, S.A., and Tolstoy, P.M. (2016). Hydrogen bonding in
bis(6-amino-1,3-dimehtyluracil-5-yl)-methane derivatives:dynamic NMR and
DFT evaluation. J. Phys. Chem. A 120: 2737–2748.
29 Sosnicki, J.G., Langaard, M., and Hansen, P.E. (2007). Long-range deuterium
isotope effects on 13 C chemical shifts of intramolecularly hydrogen bonded
References 205
N-substituted-3-(cycloamino) thiopropionamides or amides: a case of electric

field effects. J. Org. Chem. 72: 4108–4116.
30 Abildgaard, J., LiWang, A., Manolo, M.N., and Hansen, P.E. (2009). Peptide
deuterium isotope effects on backbone 15 N chemical shifts in proteins. J. Biol.
NMR 44: 119–124.
31 Hansen, P.E., Kamounah, F.S., Saeed, B.A. et al. (2019). Intramolecular
hydrogen bonds in normal and sterically compressed o-hydroxy aromatic
aldehydes. Isotope effects on chemical shifts and hydrogen bond strength.
Molecules 24: 4533.
32 Hansen, P.E. (1986). Deuterium isotope effects on the 13 C nuclear shielding of
Intramolecularly hydrogen-bonded system. Magn. Reson. Chem. 24: 903–910.
33 Abildgaard, J., Bolvig, S., and Hansen, P.E. (1998). Unravelling the elec-
tronic, steric and vibrational contributions to deuterium isotope effects on 13 C
chemical shifts by ab initio model calculations. Intramolecular hydrogen
bonded o-hydroxy acyl aromatics. J. Am. Chem. Soc. 120: 9063–9069.
34 Bolvig, S. and Hansen, P.E. (2000). Isotope effects on chemical shifts as an
analytical tool in structural studies of intramolecularly hydrogen bonded
compounds. Curr. Org. Chem. 4: 19–54.
35 Elias, R.S., Saeed, B.A., Kamounah, F.S. et al. (2020). Strong intramolecular
hydrogen bonds and steric effects. A NMR and computational study. Magn.
Reson. Chem. 58: 154–162.
36 Nguyen, T.T., Le, T.N., Hansen, B.V.K. et al. (2007). Hydrogen bonding of novel
o-hydroxythioacetophenones and related compounds studies by deuterium
isotope effects on 13 C chemical shifts. Magn. Reson. Chem. 45: 245–252.
37 Pham Nguyen, T., Nguyen, K.P.P., Zhang, W. et al. (2009). Synthesis and NMR
studies of novel hydroxyflavones, hydroxyflavothiones, hydroxyflavanones and
hydroxyflavanonethiones. Magn. Reson. Chem. 47: 1043–1054.
38 Hansen, P.E., Kamounah, F.S., and Gryko, D.T. (2013). Deuterium isotope
effects on 13 C chemical shifts of 10-hydroxybenzo[h]quinolones. Molecules
18: 4544–4560.
39 Hansen, P.E., Rozwadowski, Z., and Dziembowska, T. (2009). NMR studies of
hydroxy Schiff bases. Curr. Org. Chem. 13: 194–215.
40 Jednacak, T.L., Novak, P., Uzarevic, I.K. et al. (2011). Bioactive phenylenedi-
amine derivatives of dehydroacetic acid: synthesis, structural characterization
and deuterium isotope effects. Croat. Chem. Acta 84: 203–209.
41 Novak, P., Piculjan, K., Bilkjan, T. et al. (2007). Deuterium isotope effects
in 13 C NMR spectra of intramolecularly hydrogen-bonded salicylaldehyde-4-
phenylthiosemicarbazone. Croat. Chem. Acta 80: 575–581.
42 Seyedkatouli, S., Vakili, M., Tayyari, S.F. et al. (2019). Molecular structure and
intramolecular hydrogen bond strength of 3-methyl-4-amino-3-penten-2-one
and its NMe and N-Ph substitutions by experimental and theoretical methods.
J. Mol. Struct. 1184: 233–245.
43 Ullah, S., Zhang, W., and Hansen, P.E. (2010). Deuterium isotope effects on
13
C and 15 N chemical shifts of intramolecularly hydrogen-bonded β-enamine
derivatives of meldrum’s and tetronic acid. J. Mol. Struct. 976: 377–391.
44 Kurutos, A., Sauer, S.P.A., Kamounah, F.S., and Hansen, P.E. (2021).
Azo-hydrazone derived molecular switches: synthesis and conformational
investigation. Magn. Reson. Chem. https://doi.org/10.1002/mrc.5164.
45 Saeed, B.A., Elias, R.S., Kamounah, F.S., and Hansen, P.E. (2018). A NMR, MP2
and DFT study of thiophenoxyketenimines (o-thio Schiff bases). Magn. Reson.
Chem. 56: 172–182.
46 Smith, L.B. and Hansen, P.E. (2008). Intramolecular hydrogen bonding of
5-acyl-3-methylrhodanines. Z. Phys. Chem. 222: 1213–1223.
47 Kim, Y.-I., Manalo, M.N., Peréz, L.M., and LiWang, A. (2006). Computation and
empirical trans-hydrogen bond deuterium isotope shifts suggest that N1-N3 A:U
hydrogen bonds of RNA are shorter than those of A:T bonds of DNA. J. Biomol.
NMR 34: 229–236.
48 Manalo, M.N., Pérez, L.M., and LiWang, A. (2007). Hydrogen-bonding and
p_-_p base-stacking interactions are coupled in DNA as suggested by calculated
and experimental trans-H bond deuterium isotope shifts. J. Am. Chem. Soc.
129: 11298–11299.
49 Puffer, B., Kreutz, C., Rieder, U. et al. (2009). 5-Fluoro pyrimidines: labels
to probe DNA and RNA secondary structures by 1D 19 F NMR spectroscopy.
Nucleic Acids Res. 37: 7728–7740.
50 Hansen, P.E., Kawecki, R., Krowczynski, A., and Kozerski, L. (1990).
Deuterium isotope effects on 13 C and 15 N nuclear shielding in intramolecularly
hydrogen-bonded compounds. II. Investigation of enamine derivatives. Acta
Chem. Scand. 44: 826–832.
51 Grech, E., Klimkiewich, J., Nowicka-Scheibe, J. et al. (2002). Deuterium isotope
effects on 15 N, 13 C and 1 H chemical shifts of proton sponges. J. Mol. Struct.
615: 121–140.
52 Ullah, S., Ishimoto, T., Williamson, M.P., and Hansen, P.E. (2011). Ab initio
calculations of deuterium isotope effects on chemical shifts of salt-bridged
lysines. J. Phys. Chem. B 115: 3208–3215.
53 Zhang, D. and Tugarinov, V. (2013). Accurate measurements of the effects of
deuteration at backbone amide positions on the chemical shifts of 15N, 13Cα,
13Cβ, 13CO and 1Hα nuclei in proteins. J. Biomol. NMR 56: 169–182.
54 Tugarinov, V. (2013). Four-bond deuterium isotope effects on the chemical
shifts of amide nitrogens in proteins. Magn. Reson. Chem. 51: 722–728.
55 Tugarinov, V. (2014). Indirect use of deuterium in solution NMR studies of
protein structure and hydrogen bonding. Progr. NMR Spectrosc. 77: 49–68.
56 Liu, D., Wang, J., Lu, S. et al. (2008). Hydrogen-bond detection, configuration
assignment and rotamer correction of side-chain amides in large proteins by
NMR spectroscopy trough protium/deuterium isotope effects. ChemBioChem
9: 2860–2871.
57 Liu, A., Lu, S., Wang, J., and Yao, l., Li, Y. and Yan h. (2008). NMR detection of
bifurcated hydrogen bonds in large proteins. J. Am. Chem. Soc. 130: 2428–2429.
58 Yuan, C., Tu, S., Gelb, H.H., and Tsai, M.-D. (2005). Unusual four-bond
secondary H/D isotope effect supports a short-strong hydrogen bond between
References 207
phospholipase A2 and a transition state analogue inhibitor. Biochemistry

44: 4748–4758.
59 Filarowski, A. and Hansen, P.E. (2013). Secondary isotope effects on 13 C and
15
N chemical shifts of Schiff bases revisited. Z. Phys. Chem. 227: 917–927.
60 Tolstoy, P.M., Guo, J., Koeppe, B. et al. (2010). Geometries and tautomerism of
OHN hydrogen bonds in aprotic solution probed by H/D isotope effects on 13 C
chemical shifts. J. Phys. Chem. A 114: 10775–10782.
61 Tolstoy, P.M., Smirnov, S.N., Shenderovich, I.G. et al. (2004). NMR studies of
solid state-solvent and H/D isotope effects on hydrogen bond geometries of 1:1
complexes of collidine with carboxylic acids. J. Mol. Struct. 700: 19–27.
62 Hansen, P.E., Lund, T., Krake, J. et al. (2016). A reinvestigation of the ionic
liquid diisopropyl ethyl ammonium formate by NMR and DFT methods.
J. Phys. Chem. B 120: 11279–11286.
63 Mulloyarova, V.V., Giba, I.S., Kostin, M.A. et al. (2018). Cyclic trimers of
phosphinic acids in polar aprotic solvent: symmetry, chirality and H/D isotope
effects on NMR chemical shifts. Phys. Chem. Chem. Phys. 20: 4901–4910.
64 Mulloyarova, V.V., Ustimchuk, D.O., Filarowski, A., and Tolstoy, P.M. (1907).
(2020)H/D isotope effects on 1 H-NMR chemical shifts in cyclic heterodimers
and heterotrimers of phosphinic and phosphoric acids. Molecules 25.
65 Ajami, D., Tolstoy, P.M., Dube, H. et al. (2011). Encapsulated carboxylic acid
dimers with compressed hydrogen bonds. Angew. Chem. Int. Ed. 50: 528–531.
66 Gunnarson, G., Wennerstöm, H., Egan, W., and Forsén, S. (1976). Proton and
deuterium NMR of hydrogen bonds: relationship between isotope effects and
the hydrogen bond potential. Chem. Phys. Lett. 38: 96–99.
67 Chmielewski, P., Ozeryanskii, V.A., Sobczyk, L., and Pozharski, A.F. (2007).
Primary 1H/2H isotope effect in the NMR chemical hidt of HClO4 salts of
1,8-bis(dimethylamino)naphthalene derivatives. J. Phys. Org. Chem. 20: 643–648.
68 Bolvig, S., Hansen, P.E., Morimoto, H. et al. (2000). Primary tritium and
deuterium isotope effects on chemical shifts of compounds having an
intramolecular hydrogen bond. Magn. Reson. Chem. 38: 525–535.
69 Pietrzak, M., Try, A.C., Andrioletti, B. et al. (2008). The largest 15N-15N
coupling constant across an NHN hydrogen bond. Angew. Chem. Int. Ed. 47:
1123–1126.
70 Perrin, C.L. and Neison, J.B. (1997). Asymmetry of hydrogen bonds in solution
of monanions of dicarboxylic acids. J. Am. Chem. Soc. 119: 12734–12741.
71 Guo, J., Tolstoy, P.M., Koeppe, B. et al. (2012). H-H Hydrogen bond geome-
tries and proton tautomerism of homoconjugated anions of carboxylic acids
studied via H/D isotope effects on 13 C NMR chemical shifts. J. Phys. Chem. A
116: 11180–11188.
72 Guo, J., Tolstoy, P.M., Koeppe, B. et al. (2011). NMR study of conformational
exchange and double-well proton potential in intramolecular hydrogen bonds in
monoanions of succinic acid and derivatives. J. Phys. Chem. A 115: 9828–9836.
73 Perrin, C.L. (2010). Secondary equilibrium isotope effects on acidity. Adv. Phys.
Org. Chem. 44: 123–171.
74 Platzer, G., Okon, M., and McIntosh, L.P. (2014). pH-dependent random coil
1
H, 13 C and 15 N chemical shifts of the ionizable amino acids: a guide for
protein pKa measurements. J. Biomol. NMR 60: 109–129.
75 Mackenzie, H.M. and Hansen, D.F. (2017). A 13C-detected 15N
double-quantum NMR experiment to probe arginine side-chain guanidinium
15
Nη chemical shifts. J. Biomol. NMR 69: 123–132.
76 Tomlinson, J.H., Ullah, S., Hansen, P.E., and Williamson, M.P. (2009).
Characterisation of salt bridges to lysines in the protein G B1 domain. J. Am.
Chem. Soc. 131: 4674–4684.
77 Hansen, P.E. and Lycka, A. (1989). A reinvestigation of one-bond deuterium
isotope effects on nitrogen and on proton nuclear shielding for the ammonium
ion. Acta Chem. Scand. 43: 222–232.
78 Jin, Y., Cliff, J., Baxter, N.J. et al. (2012). Charge-balanced metal fluoride
complexes for protein kinase A with adenosine diphosphate and substrate
peptide SP20. Ang. Chem. Int. Ed. 51: 12242–12245.
79 Hou, Y., Hu, W., Li, X. et al. (2017). Solvent-accessibility of discrete residue
positions in the polypeptide hormone glucagon by 19 F-NMR observation of
4-fluorophenylalanine. J. Biomol. NMR 68: 1–6.
80 Shi, P., Li, D., Chen, H. et al. (2011). Site-specific solvent exposure analysis of
a membrane protein using unnatural amino acids and 19F nuclear magnetic
resonance. Biochem. Biophys. Res. Commun. 414: 379–383.
81 Bai, P., Luo, L., and Peng, X.-y. (2000). Side chain accessibility and dynamics
in the molten globule state of a-lactalbumin: a 19 F-NMR study. Biochemistry
39: 372–380.
82 Suzuki, Y., Buer, B.C., Al-Hashimi, H.M., and Marsh, E.N.G. (2011). Using
fluorine nuclear magnetic resonance to probe changes in the structure
and dynamics of membrane-active peptides interacting with lipid bilayers.
Biochemistry 50: 5979–5987.
83 Cho, B.P. and Zhou, L. (1999). Probing the conformational heterogeneity of
the acetylaminofluorene-modified 2′ -deoxyguanosine and DMA by 19F NMR
spectroscopy. Biochemistry 38: 7572–7583.
84 Takeda, M., Jee, J., Ono, A.M. et al. (2009). Hydrogen exchange rate of tyrosine
hydroxyl groups in proteins as studied by the deuterium isotope effect on C?
chemical shifts. J. Am. Chem. Soc. 131: 18556–18562.
85 Takeda, M., Miyanoiri, Y., Terauchi, T. et al. (2014). Use of H/D isotope effects
to gather information about hydrogen bonding and hydrogen exchange rates.
J. Magn. Reson. 241: 148–154.
86 Lopez del Amo, J.-M., Fink, U., and Reif, B. (2010). Quantification of protein
backbone hydrogen-deuterium exchange rates by solid state NMR spectroscopy.
J. Biomol. NMR 48: 203–212.
87 Reuben, J. (1987). Isotopic multiplets in the carbon-13 NMR spectra of aniline
derivatives and nucleosides with partially deuterated amino groups: effects of
intra- and intermolecular hydrogen bonding. J. Am. Chem. Soc. 109: 316–321.
References 209
88 Schaefer, T. (1975). A relationship between hydroxy proton chemical shifts

and torsional frequencies in some ortho-substituted phenol derivatives. J. Phys.
Chem. 79: 1888–1890.
89 Hansen, P.E. and Tüchsen, E. (1989). Deuterium isotope effects on carbonyl
chemical shifts of BPTI. Hydrogen-bonding and structure determination in
proteins. Acta Chem. Scand. 43: 710–712.
90 Tüchsen, E. and Hansen, P.E. (1991). Hydrogen bonding monitored by
deuterium isotope effects on carbonyl 13C chemical shift in BPTI: intra-residue
hydrogen bonds in antiparallel β-sheet. Int. J. Biol. Macromol. 13: 2–8.
91 Cuma, M., Scheiner, S., and Kar, T. (1998). Competition between
rotamerization and proton transfer in o-hydroxybenzaldehyde.J. Am. Chem.
Soc. 120: 10497–10503.
92 Sanz, P., Mó, O., Yáñez, M., and Elguero, J. (2007). Resonance-assisted
hydrogen bonds: a critical examination. Structure and stability of the enols
of β-diketones and β-enaminones. J. Phys. Chem. A 111: 3585–3591.
93 Afonin, A.V., Vashchenko, A.V., and Sigalov, M.V. (2016). Estimating the energy
of intramolecular hydrogen bonds from 1H NMR and QTAIM calculations.
Org. Biomol. Chem. 14: 11199–11211.
94 Becke, A.D. (1993). A new mixing of Hartree-Fock and local density-functional
theories. J. Chem. Phys. 98: 1372–1377.
95 Bader, R.W.F. (1990). Atoms in Molecules. A Quantum Theory. New York:
Oxford University Press.
96 AIMAll (Version 16.10.31), Kieth, T.A. (2016). TK Gristmill Software, Overland
Park KS, USA. aim.tkgristmill.com.
97 Hansen, P.E. (2021). A spectroscopic overview of intramolecular NH…O,N,S
hydrogen bonds. Molecules 26: 2409. https://doi.org/10.3390/molecules26092409.
98 Dziembowska, T. and Rozwadowski, Z. (2001). Application of the deuterium
isotope effect on NMR chemical shift to study proton transfer equilibrium.
Curr. Org. Chem. 5: 289–313.
99 Hansen, P.E. (2014). Isotope effects on chemical shifts as a tool in the study of
tautomeric equilibria. In: Equilibria. Methods and Theories (ed. L. Antonov).
Wiley-VCH.
100 Hansen, P.E. (2016). Methods to distinguish tautomeric cases from static ones.
In: Tautomerism: Ideas, Compounds, Applications (ed. L. Antonov). Wiley-VCH.
101 Hansen, P.E., Mortensen, J., and Kamounah, F.S. (2015). The importance of
correct tautomeric structures for biological molecules. JSM Chem. 3: 1014–1019.
102 Hansen, P.E., Kamounah, F.S., Hansen, B.V.K., and Spanget-Larsen,
J. (2007). Conformational and tautomeric eccentricities of
1,8-dihydroxy-2-acetylnaphthalenes. Magn. Reson. Chem. 45: 106–117.
103 Sharif, S., Denisov, G.S., Toney, M.D., and Limbach, H.-H. (2006). NMR studies
of solvent-assisted proton transfer in a biologically relevant Schiff base: towards
a distinction of geometric and equilibrium H-bond isotope effects. J. Am. Chem.
Soc. 128: 3375–3387.
104 Hansen, P.E. (2020). Isotope effects on chemical shifts in the study of hydrogen
bonded biological systems. Prog. NMR 120–121: 109–117.
105 Sharif, S., Denisov, G.S., Toney, M.D., and Limbach, H.-H. (2007). NMR
studies of low-and high-barrier hydrogen bonds in pyridoxal-5′ -phosphate
model systems in polar solutions. J. Am. Chem. Soc. 129: 6313–6327.
106 Golubev, N.S., Smirnov, S.N., Tolstoy, P.M. et al. (2007). Observation by NMR of
the tautomerism of an intramolecular OHOHN-charge relay chain in a model
Schiff base. J. Mol. Struct. 844-845: 319–327.
107 Ossowicz, P., Janus, E., Schroeder, G., and Rozwadowski, Z. (2013).
Spectroscopic studies of amino acid ionic liquid-supported Schiff bases.
Molecules 1: 4986–5004.
108 Rozwadowski, Z. (2006). Deuterium isotope effects on 13 C chemical shifts of
tetrabutylammonium salts of Schiff bases amino acids. Magn. Reson. Chem.
44: 881–886.
109 Rozwadowski, Z. (2006). Deuterium isotope effects on 13C chemical shifts of
lithium salts of Schiff bases amino acids. J. Mol. Struct. 753: 127–131.
110 Ivanova, D., Deneva, V., Nedeltcheva, D. et al. (2015). Tautomeric
transformations of Piroxicam in solution: a combined experimental and
theoretical study. RSC Adv. 5: 31852–31860.
111 Hansen, P.E., Kamounah, F.S., Zhiryakova, D. et al. (2014). 1,1′ ,1′′ -(2,4,6-
Trihydroxybenzene-1,3,5-triyl)triethanone non-tautomerism. Tetrahdron Lett.
55: 354–357.
112 Serdiuk, I.E., Wera, M., Roshal, A. et al. (2011). Tautomerism, structure and
properties of 1,1′ 1′′ -(2,4,6-trihydroxybenzene-1,3,5,tiy)triethanone. Tetrahedron
Lett. 52: 2737–2740.
113 Hofmann, J.P., Hansen, P.E., Bond, A.D., and Duus, F. (2006). Tautomerism in
3-acyltetronic acids revisited. A spectrochemometric approach to tautomerism
and hydrogen-bonding. J. Mol. Struct. 790: 80–88.
114 Perrin, C.L., Shrinidhi, A., and Burke, K.D. (2019). Isotopic-perturbation
NMR study of hydrogen-bond symmetry in solution: temperature dependence
and comparison of OHO and ODO hydrogen bond. J. Am. Chem. Soc. 141:
17278–17286.
115 Perrin, C.L. and Burke, K.D. (2014). Variable-temperature study of hydrogen-
bond symmetry in cyclohexene-1,2-dicarboxylate monoanion in chloroform-d.
J. Am. Chem. Soc. 136: 4355–4362. https://doi.org/10.1021/ja500174y.
116 Bogle, X.S. and Singleton, D.A. (2011). Isotope-induced desymmetrization can
mimic isotopic perturbation of equilibria. On the symmetry of bromonium ions
and hydrogen bonds. J. Am. Chem. Soc. 133: 17172–17175. https://doi.org/10
.1021/ja2084288.
117 Perrin, C.L., Lau, J.S., Kim, Y.-J. et al. (2009). Asymmetry of the “strongest”
OHO hydrogen bond, in the monoanion of (±)-α,α′ -di-tert-butylsuccinate.
J. Am. Chem. Soc. 131: 13548–13554. https://doi.org/10.1021/ja905806h.
118 Perrin, C.L., Lau, J.S., Kim, Y.-J. et al. (2010). Asymmetry of the “strongest”
OHO hydrogen bond, in the monoanion of (±)-α,α′ -di-tert-butylsuccinate.
J. Am. Chem. Soc. 132: 2099–2100. https://doi.org/10.1021/ja9104733.
References 211
119 Saunders, M. and Kates, M.R. (1977). Isotopic perturbation of resonance.

Carbon-13 nuclear magnetic resonance spectra of deuterated cyclohexenyl
and cyclopentenyl cations. J. Am. Chem. Soc. 99: 8071–8072.
120 Perrin, C.L. (1994). Symmetries of hydrogen bonds in solution. Science
266: 1665.
121 Perrin, C.L. and Lau, J.S. (2006). Hydrogen-bond symmetry in zwitteri-
onic phthalate anions: symmetry breaking by solvation. J. Am. Chem. Soc.
128: 11820–11824.
81: 571–583.
123 Hansen, P.E. (2008). Deuterium isotope effects on 13 C chemical shifts of
nitromalonamide. Magn. Reson. Chem. 46: 726–729.
124 Andresen, B., Duus, F., Bolvig, S., and Hansen, P.E. (2000). Variable
temperature 1 H and 13 C NMR spectroscopic investigation of the enol-enethiol
tautomerism of β-thioxoketones. Isotope effects due to deuteron chelation.
J. Mol. Struct. 552: 45–63.
125 Hansen, P.E., Borisov, E.V., and Lindon, J.C. (2015). Determination of the
tautomeric equilibria of pyridoyl benzoyl β-diketones in the liquid and solid
state through the use of deuterium isotope effects on 1 H and 13 C NMR
chemical shifts and spin coupling constants. Spectrochim. Acta 136: 107–112.
126 Bolvig, S. and Hansen, P.E. (1996). Deuterium isotope effects on 13 C chemical
shifts as a probe for tautomerism in enolic β-diketones. Magn. Reson. Chem.
34: 467–478.
127 Stringfellow, T.C., Wu, G., and Wasylishen, R.E. (1997). An experimental
study of isotope effects on NMR parameters in the solid state. J. Phys. Chem. B
101: 9651–9656.
128 Hansen, P.E. (2011). Deuterium isotope effects on 14,15 N chemical shifts of
ammonium ions. A solid state NMR study. Int. J. Inorg. Chem. https://doi.org/
10.1155/2011/696497.
129 Rozwadowski, Z., Schilff, W., and Kamin′ nski, B. (2005). Solid–state NMR
study of Schiff base derivatives of 2-hydroxynaphthaldehyde. Deuterium
isotope effects on 15 N chemical shifts in the solid state. Magn. Reson. Chem.
43: 573–577.
130 Hansen, P.E., Sitkowski, J., Stefaniak, L. et al. (1998). One bond deuterium
isotope effects on 15 N chemical shifts, 1 ΔN(D), in Schiff bases. Berichte
Bunsengesel. Chem. Phys. 102: 410–413.
131 Dziembowska, T., Rozwadowski, Z., Filarowski, S., and Hansen, P.E. (2001).
A multinuclear NMR study of proton transfer equilibrium in Schiff bases
derived from 2-hydroxy-1-naphthaldehyde. Deuterium isotope effects on 13 C
and 15 N chemical shifts. Magn. Reson. Chem. 39: S67–S80.
132 Dziembowska, T., Jagodzinska, E., Rozwadowski, Z., and Kotfica, M. (2001).
Solvent effect on intramolecular proton transfer equilibrium in some
N-(R-salicylidene)-alkylamines. J. Mol. Struct. 598: 229–234.
133 Rozwaddowski, Z. and Nowak-Wydra, B. (2008). Chiral recognition of Schiff
bases by 15 N NMR spectroscopy in the presence of a dirhodium complex.
Deuterium isotope effect on 15 N chemical shift of the optically active Schiff

bases and their dirhodium tetracarboxylate adducts. Magn. Reson. Chem.
46: 974–978.
134 Dziembowska, T., Ambroziak, K., Rozwadowski, Z. et al. (2003). Deuterium
isotope effects on 15 N chemical shifts of double Schiff bases in the solid state
and solution. Magn. Reson. Chem. 41: 135–138.
135 Schagen, S.D., Toney, M.D., and Limbach, H.H. (2007). Coupling of functional
hydrogen bonds in pyridoxal-5 ’-phosphate-enzyme model systems observed by
solid-state NMR spectroscopy. J. Am. Chem. Soc. 129: 4440–4455.
136 Ip, B.C.K., Shenderovich, I.G., Tolstoy, P.M. et al. (2012). NMR studies of
solid pentachlorophenol-4-methylpyridine complexes exhibiting strong OHN
hydrogen bonds: geometric H/D isotope effects and hydrogen bond coupling
cause isotopic polymorphism. J. Phys. Chem. A 116: 11370–11387.
213
Intramolecular Hydrogen Bonding: Shaping Conformers’

Structure and Stability
Gulce O. Ildiz 1,2 and Rui Fausto 1
1
University of Coimbra, CQC-IMS, Department of Chemistry, P 3004−535 Coimbra, Portugal
2
Istanbul Kultur University, Department of Physics, Faculty of Sciences and Letters, Atakoy Campus, Bakirkoy,
34156 Istanbul, Turkey
8.1 Introduction
Intramolecular hydrogen bonds (IMHBs) are those which occur within one single
molecule. For this to happen, both a hydrogen donor and a hydrogen acceptor
must be present within the molecule, and they must be close to each other.
Being close, however, does not mean that they have to occupy positions in the
molecular framework that are close, since conformational flexibility within a
molecule may allow groups placed remotely to establish hydrogen bonds. What
is relevant is that the interacting groups become spatially close to each other
due to a specific tridimensional arrangement adopted by the molecule. In this
regard, γ-aminobutyric acid (usually abbreviated as GABA), which is a well-known
major inhibitory neurotransmitter in the mammalians' central nervous system
[1, 2], is an interesting example. In the gas phase, GABA exists as the molecular,
non-zwitterionic form, and some of its most stable conformers are intramolecularly
hydrogen-bonded, despite the interacting groups (–NH2 ; –COOH) are located at
the extreme positions of the molecule's carbon chain (Figure 8.1). As determined
by microwave spectroscopy [3], the experimentally observed conformers gG2 and
GG1 of the GABA molecule (the used designation of the conformers is that of the
original study [3]) are stabilized by O—H· · ·N IMHBs that force a trans-COOH
arrangement, and conformer GG3 is stabilized by an N—H· · ·O=C hydrogen bond.
Nevertheless, the large majority of molecules bearing IMHBs have the participating
groups located nearby in the molecule's framework, this being mandatory when
they are conformationally rigid systems.
While the coining of the expression “hydrogen bonding” has been attributed
by Pauling, in his 1939 seminal book “The Nature of the Chemical Bond” [4],
to Latimer and Rodebush, as early as 1920 [5], the phenomenon had been first
described almost 20 years before by Werner [6], who refer to it by the German word
nebenvalenz (“secondary or weak valence”). To the best of our knowledge, the first
reference to an IMHB was authored by Pfeiffer et al. [7] to denote the intramolecular
214 8 Intramolecular Hydrogen Bonding: Shaping Conformers’ Structure and Stability
gG2 GG1 GG3
Figure 8.1 Experimentally observed low-energy intramolecularly hydrogen-bonded

conformers of GABA. Source: Based on Blanco et al. [3].
O—H· · ·O=C interaction in 1-hydroxyanthraquinone (the author used the German

expression “innere komplexsalzbildung,” or “internal complex salt-bridge”).
The structural importance of IMHBs has been recognized early [7, 8], and it is
nowadays clear that they play a fundamental role in shaping potential energy sur-
faces of molecules [9–11]. In particular, for conformationally flexible molecules, they
are important in determining the relative stability of the possible conformers as well
as their specific properties and reactivity [9–12].
IMHBs are accepted also as an important feature shaping the structures of
biological molecules and the biological activity of compounds [13, 14]. For example,
a bioactive conformation of a molecule can be stabilized by an IMHB, limiting
the conformational entropy upon binding and resulting in a stronger association
[15–18], or the accessibility of polar atoms in a molecule can be reduced, influencing
desolvation equilibria and eventually facilitating the passage of molecules through a
low dielectric environment [19–21]. Accordingly, intramolecular hydrogen bonding
was recently found to be crucial in the process of diffusion of cyclic peptides through
cell membranes [22].
Experimental energetic data for IMHBs are relatively scarce because they are an
intrinsic feature of the ground state structure of the molecule. Breaking an IMHB
leads necessarily to a different conformer or tautomer, which has a different set of
intramolecular interactions compared to the original one, so that the precise direct
experimental determination of the energy of IMHBs is prevented. On the other hand,
structural data for IMHBs are most of times easily to obtain experimentally, either for
the isolated molecule (for example by microwave spectroscopy or gas phase electron
diffraction) or in the crystalline state (e.g. using X-ray diffraction or, preferentially,
neutron diffraction, which surpasses the difficulties of X-ray experiments in deal-
ing with the small hydrogen atoms). Indirect methods for characterizing IMHBs
regarding their structural and energetic features exist. In this regard, vibrational
spectroscopy has been playing a major role, since spectral parameters like the fre-
quency shifts of the vibrations of both hydrogen atom donor and acceptor moieties
upon hydrogen bond formation can be correlated with the hydrogen bond length
and enthalpy [23–25]. Empirical correlations that use band intensity changes in the
infrared spectra as spectroscopic probes of structural and/or thermodynamic data of
hydrogen bonds have also been successfully developed [26].
Theoretically, the estimation of structural characteristics of IMHBs is also a
straightforward procedure. On the other hand, direct access to energetic data
8.2 The Halogen-Substituted Acetic Acids CF3 COOH, CCl3 COOH, and CBr3 OOH 215
through theoretical methods is limited by the same fundamental reason that

prevents their experimental determination. Most of time, the energy of an IMHB is
estimated theoretically by comparing the energy of two conformers that essentially
differ by the presence of one IMHB. The energy (or the enthalpy) of the bond is
approximated as the difference between their energies. However, as stated above,
this ignores the fact that the intramolecular overall potentials are essentially
different in the two conformers. Other suggested approaches to access energetic
data for IMHBs involve the definition of isodesmic reactions leading to bond
breaking/formation related to specific hydrogen bonds [27, 28], but these are also
rough approximations that ignore the global differences in the intramolecular
potentials of the different species.
The precise determination of the energetics of IMHBs is in fact still an open
problem. However, most of time we are not interested in determining the specific
energetic contribution of an IMHB, but instead we want to evaluate the global
changes in the potential energy surface of the molecule its formation implies (and
its consequences). This is the subject of the present chapter. In the next sections,
we will present some illustrative cases where IMHBs impact the potential energy
surfaces in a profound way, strongly influencing the properties and the chemical
behavior of the molecules.
8.2 The Halogen-Substituted Acetic Acids CF3 COOH,

CCl3 COOH, and CBr3 OOH: Implications of IMHB
on Structure and Conformers’ Stabilities
The conformational preferences exhibited by carboxylic acids and the relative
stability of their different possible conformers have been experimentally addressed
very successfully by using a combination of matrix isolation infrared spectroscopy
and quantum chemical calculations with vibrationally induced selective photo-
chemistry. Conformers of the target carboxylic acid that are present in gas phase
are initially trapped into a cryogenic matrix (e.g. solid Ne, Ar, Kr, Xe, and N2 )
kept at a few Kelvin and characterized structurally by using infrared spectroscopy.
Subsequent near-infrared selective vibrational excitation of the trapped low-energy
conformers allows for their conversion into higher-energy conformers, which can
then also be characterized structurally and investigated regarding their stability. The
identification of the precise nature of the conformers and the rationalization of their
relative stabilities are usually done with help of quantum chemistry calculations
of their vibrational spectra and of the potential energy surface landscape of the
molecule.
The prototype systems first studied using this approach are formic (HCOOH)
and acetic (CH3 COOH) acids [29–37]. These molecules have two conformers each,
which differ in the configuration of the carboxylic acid group. In the most stable
form, the carboxylic group assumes the cis arrangement (O=C—O—H dihedral
equal to 0∘ ), while the less stable conformer has the carboxylic group in the trans
configuration (O=C—O—H dihedral equal to 180∘ ). Only the cis conformer of
Phonon assisted intramolecular

vibrational energy relaxation
V=1
Internal rotation
V=2
V=3 Tunneling
Torsional reactive coordinate

Excited coordinate
Figure 8.2 General mechanism for vibrationally induced selective conformational

isomerization, as exemplified for formic acid, and subsequent decay of the higher-energy
trans conformer into the more stable cis form by tunneling (green arrow). The
photoisomerization involves vibrational excitation of a high absorption cross section
high-energy mode, followed by phonon-assisted intramolecular vibrational energy
relaxation and internal rotation. The over-the-barrier thermal back-isomerization is not
accessible. Source: Reproduced from Nunes et al. [38] with permission from the Royal
Society of Chemistry. Copyright © 2021.
formic and acetic acids can be efficiently trapped from the gas phase into the cryo-
genic matrices so that the infrared spectra of the matrix-isolated compounds match
those of their most stable conformers [29–37]. Irradiation of the matrix-isolated
cis form at the frequency of its first 𝜈(OH) stretching overtone allows for selective
introduction of energy in the molecule, which is then redistributed among their
vibrational coordinates following preferential relaxation channels [32]. Ultimately,
the energy is deposited in the 𝜏(C–O) torsional coordinate, leading to internal
rotation and conversion of the cis conformer into the higher-energy trans form
(Figure 8.2).
The vibrationally induced selective conformational isomerization taking place
through this mechanism has been achieved for many chemical systems possessing
the carboxylic group [39–44], as well as for molecules where the carboxylic group
appears together with other substituents, such as α-hydroxy and α-amino acids, for
example [45–54].
The reasons for the general higher stability of the cis conformation of the car-
boxylic acid moiety in relation to the trans conformation are well known [55–57].
The major factor is the favorable bond-dipole/bond-dipole interaction established
between the strongly polarized C=O and O—H bonds that exist in the cis config-
uration (where these two bond-dipoles are antiparallel), which is replaced by the
repulsive interaction due to the nearly parallel alignment of these bond-dipoles in
the trans configuration.
Very interestingly, very often the higher-energy trans conformers of simple
carboxylic acids were found to be unstable, converting spontaneously by quantum
mechanical tunneling to their most stable cis counterparts. For example, after their
8.2 The Halogen-Substituted Acetic Acids CF3 COOH, CCl3 COOH, and CBr3 OOH 217
production in an argon matrix by vibrational excitation of the corresponding cis

conformer, the trans conformers of formic and acetic acids convert back to the cis
form by tunneling, with a decay time (t1/2 ) of 8 and 0.8 minutes, respectively [31].
The same phenomenon has been observed to take place for many other systems, the
lifetime of the higher-energy conformers spawning from fractions of a second to sev-
eral hours, depending on both characteristics associated with the specific molecule
(in particular the intrinsic height and width of the energy barriers separating the
higher and lower energy conformers) and experimental conditions (e.g. matrix
type, trapping site local microenvironment) [29, 48, 58–61]. Regarding the influence
of the matrix type on the stability of the conformers, it shall be highlighted that N2
matrices show substantially different properties compared to noble gases matrices.
In fact, N2 molecules can establish specific stabilizing intermolecular hydrogen
bonds with the carboxylic moiety, while the interactions between the matrix noble
gases atoms and the carboxylic group are negligible [29, 44]. In practical terms, this
usually leads to an increased stabilization of the in situ generated high-energy trans
conformers of carboxylic acids in an N2 matrix compared to noble gases [29, 44].
The trihalogeno-substituted acetic acids, CF3 COOH (trifluoroacetic acid, TFA),
CCl3 COOH (trichloroacetic acid, TCA), and CBr3 COOH (tribromoacetic acid,
TBA) have recently been studied in our laboratory by using the research strategies
described above [62–64]. As shown below, the relevance of IMHB in TFA is of prime
significance in determining the unique structural characteristics of this compound
within the studied series of compounds (and also compared to the parent acetic
acid). In addition, it plays also an important role in determining the stability of the
trans-TFA conformer in an N2 matrix.
As for formic and acetic acids, TFA, TCA, and TBA may exist in two conformers
(Figure 8.3), with the cis form being more stable than the trans form by 10.9, 10.7,
and 10.4 kJ mol−1 , respectively [62–64]. The similarity between these values is
striking, and it implies cancellation of the effects due to various intramolecular
interactions present in the two conformers upon changing the halogen atoms.
In the cis conformer, the major effects occurring in the three compounds are (i) the
destabilizing interaction due to electron charge repulsion between the negatively
charged halogen atoms staying out of the molecular plane and the acid (OH) oxygen
atom (see Figure 8.3), which increases in importance in going from TBA to TFA,
and (ii) the steric repulsion between these atoms, which shall follow the reverse
trend. For the trans conformer, (iii) the interactions between the halogen atoms
and the acid hydrogen are attractive and more important in TFA than in TCA and
TBA. On the other hand, (iv) the geometry adopted by the –CF3 substituent in TFA,
where this group is staggering the carbonyl bond (see Figure 8.3) [62], is known to
be unfavorable in comparison to the one adopted by the –CCl3 and –CBr3 groups
in TCA and TBA (where one of the halogen atoms eclipses the C=O bond [63, 64];
see Figure 8.3) in terms of group orbital interactions [55, 56, 65–67]. The observed
major structural difference in trans-TFA compared to trans-TCA and trans-TBA
is a consequence of the existence in trans-TFA of an IMHB between the –OH
group and the –CF3 moiety, which is favored in the different geometry adopted
by the trans-TFA molecule. The narrow range of values for the energy difference
cis-TFA cis-TCA cis-TBA
trans-TFA trans-TCA trans-TBA
Figure 8.3 Conformers of TFA, TCA, and TBA. Source: Based on Apóstolo et al. [62].
between the trans and cis conformers of the three molecules demonstrates that on
the whole, these interactions result in a very similar energy balance in the three
compounds.
Particularly interesting are the results obtained for ΔG∘ (trans–cis) in the series
of studied compounds. ΔG∘ (trans–cis) decreases from 11.2 kJ mol−1 in TFA [62]
to 7.5 kJ mol−1 in TBA [64] and to only 5.38 kJ mol−1 in TCA [63]. The entropy
effects stabilizing the trans conformer should then gain importance in the order
TFA < TBA < TCA, which is in fact in agreement with calculated ΔS∘ (trans–cis)
entropy differences for TFA, TBA, and TCA that amount to 0.05, 6.58, and
16.30 J mol−1 K−1 , respectively [62–64]. The increase of the entropy of the trans
conformer along the series TFA < TBA < TCA is essentially due to the increased
conformational flexibility associated with the torsion around the C—C bond in the
vicinity of the trans minimum, which is a measure of the flatness of the potential
energy surface along the torsional coordinate for this conformation. For trans-TFA,
the existence of the O—H· · ·F IMHB strongly restricts the torsional movement
and leads to a considerably steeper torsional potential and lower entropy. In the
remaining two molecules, no such interaction exists, and the torsional potential is
significantly shallower [62–64].
By comparing the observed infrared spectra of the as-deposited matrices of the
compounds with the quantum mechanically predicted infrared spectra for both
cis and trans conformers [62–64], the sole presence in the matrices of the more
stable cis conformer could be doubtlessly confirmed. In all cases, subsequent
near-infrared irradiation of the cis conformer at its 𝜈(OH) stretching first overtone
frequency allowed generation of the trans conformer. This was easily noticed by the
appearance in the spectra of the irradiated matrices of the characteristic spectrum
of this species [62–64]. As anticipated, the trans conformer of all three compounds
was found to decay spontaneously to the most stable cis conformer by tunneling.
8.3 The Significance of IMHB in the ortho Chloro- and Fluoro-Substituted Benzoic Acids 219
Spectroscopic following of the band intensity changes with time allowed to extract
kinetical data for the tunneling transformations [62–64].
In the studied noble gas matrices (Ar, Kr), the population of trans-TBA was found
to be reduced to half in ∼2 minutes, while in the N2 matrix this requires ∼10 times
more time, evidencing the stabilization of the conformer in this latter matrix, as
expected [64]. The decay rates of trans-TBA and trans-TFA in Ar and Kr matrices
are identical [62, 64], and do not differ much from those of trans acetic acid in the
same matrices (0.8 minutes) [31] (no data are available for TCA). However, they are
considerably shorter than those measured for formic acid (8 and 14 minutes in Ar
and Kr, respectively) [31]. The results obtained for the Ar and Kr matrices are in
agreement with the relative heights of the trans → cis energy barriers in the various
compounds: 34.5, 32.2, 32.9, and 38.5 kJ mol−1 , respectively for TBA, TFA, acetic,
and formic acids [31, 62, 64].
In the N2 matrix, the panorama changes considerably. In fact, while the
lifetimes of trans-TBA and trans-TCA in this matrix are not very different (23 and
12 minutes, respectively) [63, 64], they are significantly shorter than that of
trans-TFA (90 minutes) [62] and much shorter than those found for both acetic and
formic acids (5 and 7 hours, respectively) [29]. The reduction of the lifetime of the
trans-TFA conformer in the N2 matrix in comparison to trans acetic and formic
acids can be ascribed to the reduced accessibility of the –OH group in trans-TFA
to the matrix N2 molecules due to the presence of the intramolecular OH· · ·F
interaction. A somewhat similar explanation can be given to justify the additional
decrease of stability observed for trans-TCA and trans-TBA since for these two
molecules the bulkier chlorine or bromine atoms impose severe steric hindrance to
the approach of the N2 molecules to the –OH group. Note that, contrarily to what
happens for trans-TBA and trans-TCA, the presence of the IMHB in trans-TFA is
critical to justify its reduced stability in the N2 matrix compared to trans acetic and
formic acids, since the fluorine and hydrogen atoms are of comparable size.
8.3 The Significance of IMHB in the ortho Chloro-

and Fluoro-Substituted Benzoic Acids
Another family of compounds where intramolecular hydrogen bonding plays a sig-
nificant structural role is that of ortho chloro- and/or fluoro-substituted benzoic
acids (BAs). The whole family of compounds was investigated recently in our lab-
oratory [12], including the parent BA for comparison, and the ortho mono- and
disubstituted chloro- and/or fluorobenzoic acids. This example highlights the effect
of the different abilities of chlorine and fluorine to participate in hydrogen bonds
and signals the relative impact of O—H· · ·F and O—H· · ·Cl hydrogen bonds on the
structure and stability of the intramolecularly hydrogen-bonded conformers of the
studied substances.
The structures and energies of the conformers of these molecules, together
with the values of the conformationally relevant dihedral are shown in Figures 8.4
and 8.5.
Benzoic acid 2-Fluorobenzoic acid 2-Chlorobenzoic acid
13 13 13
11 11 12
12 11 12
10 10 10
1 14 1
1 15
14 15 14 15
6 2 6 2 6 2
3 5 3
5 3 5
7 9 7
9 7 9 4
4 4
8 8
8
cis cis-I cis-II cis-I cis-II

(0.0; 180.0, 0.0) (0.0; 180.0, 0.0) (0.0; 0.0, 180.0) (1.2; –18.0, 163.5) (1.9; –155.9, 23.5)
[0.00] [2.82] [0.00] [0.00] [0.24]
trans trans-I trans-II trans-I trans-II

(–171.4; –157.4, 24.0) (–172.0; –132.5, 44.3) (180.0; 0.0, 180.0) (–173.7; –54.9, 122.1) (–178.3; –172.0, 8.2)
[28.56] [29.36] [7.69] [22.02] [9.92]
Figure 8.4 B3LYP/6-311++G(d,p) calculated minimum energy structures of benzoic acid,

2-fluorobenzoic acid, and 2-chlorobenzoic acid, with adopted atom numbering. The values
of the O=C—O—H, C6—C—C=O, and C2—C—C=O dihedral angles (in ∘ ) are given in
parentheses in this order. Relative energies are given in square parentheses, in kJ mol−1 .
Each non planar structure depicted has a symmetry-related form. The cis benzoic acid
conformer is also twofold degenerate, while the trans conformer is fourfold symmetry
degenerate. Source: Ildiz and Fausto [12], figure 1 (p. 4)/MDPI/Licensed under CC BY 4.0.
All studied molecules have two conformational degrees of freedom, corresponding

to the internal rotations around the exocyclic C—C and C—O bonds, which define
the orientation of the carboxylic group relatively to the aromatic ring and the internal
configuration of the carboxylic group, respectively.
BA has two conformers (Figure 8.4). The lowest energy cis conformer is planar
and corresponds to the experimentally observed species [68–72]. The planarity of
the molecule ensures maximum conjugation between the carboxylic substituent
and the aromatic ring and is favored by the stabilizing attractive interactions
between the ortho hydrogen atoms and the oxygen atoms of the carboxylic group
(carbonyl, O=, and acid, OH). On the other hand, the trans conformer has a non
planar structure and has a predicted energy ∼29 kJ mol−1 above that of the cis
conformer. In this conformer, the repulsion between the carboxylic hydrogen atom
and the nearby located ring ortho hydrogen atom is the main reason leading to
the observed tilt of the carboxylic moiety out of the plane of the ring. The trans
conformer has never been experimentally observed.
All the other molecules investigated also have at least one cis and one trans con-
former. The structures of their cis conformers are mostly determined by the size of
the substituents present in the molecules, being planar in the case of the mono-
substituted fluorine compound, and having the carboxylic group deviated from the
aromatic ring plane in all the other cases, with the dichloro-substituted compound
representing the extreme case where the two molecular fragments (aromatic ring
and carboxylic substituent) are perpendicular to each other to minimize repulsions.
2,6-Difluorobenzoic acid 2,6-Dichlorobenzoic acid
13 13
12
11 12
11
10 10
1 14 1
14 15 15
6 2 6 2
5 3 5 3
9 7 9 7
4 4
8 8
cis trans cis trans

(2.5; –130.6, 47.1) (–173.3; –150.1, 26.6) (0.0; –90.0, 90.0) (180.0; –90.0, 90.0)
[0.00] [14.87] [0.00] [17.05]
2-Chloro-6-fluorobenzoic acid
13
12
11
10
14
1
15
6 2
5
3
9 7
4
cis trans trans′

(1.6; –67.6, 110.2) (–176.5; –48.9, 127.1) (–179.5; –103.0, 81.2)
[0.00] [17.07] [17.31]
Figure 8.5 B3LYP/6-311++G(d,p) calculated minimum energy structures of

2,6-difluorobenzoic acid, 2,6-dichlorobenzoic acid, and 2-chloro-6-fluorobenzoic acid, with
adopted atom numbering. The values of the O=C—O—H, C6—C—C=O, and C2—C—C=O
dihedral angles (in ∘ ) are given in parentheses in this order. Relative energies are given
in square parentheses, in kJ mol−1 . Each conformer of 2,6-dichlorobenzoic acid and
2-chloro-6-fluorobenzoic acid has a symmetry-related identical form; the conformers
of 2,6-difluorobenzoic acid have four symmetry-equivalent structures. Source: Ildiz and
Fausto [12], figure 5 (p. 5)/MDPI/Licensed under CC BY 4.0.
A detailed analysis of the reasons for the relative energies of the two different cis
conformers in 2-fluorobenzoic acid (2FBA) and 2-chlorobenzoic acid (2CBA) can be
found in our original manuscript [12].
More relevant for the topic of this chapter are the trans conformers of the studied
series of compounds. As for the trihalogeno-substituted acetic acids described in
Section 8.2, the trans conformers of the investigated BAs correspond to higher-
energy forms, due to the intrinsically less stable configuration of the trans car-
boxylic moiety [55–57]. As already mentioned, in BA the trans conformer is higher
in energy than the cis form by ∼29 kJ mol−1 . For the disubstituted compounds, the
trans conformers are 15–17 kJ mol−1 above the energy of the cis conformer. No
intramolecularly hydrogen-bonded conformers exist for these molecules because
the required geometries would also imply strong repulsions between the ring
halogen atoms and the oxygen atoms of the carboxylic moiety (Figure 8.5). On
the other hand, each one of the monosubstituted compounds has two distinct
trans conformers, one with a higher relative energy (over 29 and 22 kJ mol−1 ,
respectively for 2FBA and 2CBA; Figure 8.4) where the –OH fragment of the
carboxylic substituent is roughly directed to the ortho hydrogen atom, and the
other with a comparatively much lower relative energy that is stabilized by an
IMHB established between the –OH moiety and the ortho halogen substituent (see
Figure 8.4). The energies of the hydrogen-bonded conformers in 2FBA and 2CBA
are respectively ∼7.6 and 9.9 kJ mol−1 above those of the corresponding most stable
cis conformer, these values correlating with the expected relative strengths of the
O—H· · ·F and O—H· · ·Cl IMHBs. Note also that the hydrogen-bonded conformers
of these two compounds have planar (for 2FBA) or very close to planarity (for
2CBA) geometries, which help to satisfy the typical directional requirements of a
hydrogen bond. The non-intramolecularly hydrogen-bonded conformers, on the
other hand, have geometries strongly deviated from planarity, since in these cases
repulsive interactions predominate.
IMHB in the trans conformers of the BAs has a strong impact not just on the
geometries of the conformers, as pointed out above, but also on the barriers for
conformational isomerization and, consequently, on the kinetical stability of the
conformers.
One interesting fact is that the profiles for internal rotation around the exocyclic
C—C bond are considerably different for the two arrangements of the carboxylic
group. This fact is observed even in the case of the parent BA, where for the cis
carboxylic group geometry the energy barrier (∼28 kJ mol−1 ) is more than twice
larger than for the trans arrangement (∼13 kJ mol−1 ) (Figure 8.6). The explanation
for this remarkable difference relies on IMHB. In fact, when the carboxylic group is
cis, the high energy of the transition state (TS) for rotation around the C—C bond
is a consequence of two main factors: (i) the breakdown of the above-mentioned
stabilizing interactions between the ring ortho hydrogen atoms and the oxygen
atoms of the carboxylic acid substituent (that operate in the minimum energy struc-
ture), and (ii) the lack of conjugation between the carboxylic substituent and the
aromatic ring at the TS. On the other hand, in the TS for the C–C internal rotation
when the carboxylic group is trans there is a favorable interaction between the
π-system of the aromatic ring and the carboxylic group (O—H· · ·π IMHB), which
considerably lowers the internal rotation barrier. This type of O–H· · ·π stabilizing
interaction has been described for other, structurally related systems [73].
Even more remarkable is the effect of the IMHB established between the O–H
moiety and the ortho halogen substituent on the potential energy barrier for con-
formational isomerization about the C—O bond, i.e. carboxylic group trans → cis
isomerization. By reducing the energy of the trans conformer, the intramolecular
hydrogen bonding increases substantially the isomerization barrier, which leads to
a kinetic stabilization of the intramolecularly hydrogen-bonded trans conformer.
In BA, the trans → cis isomerization barrier is ∼23 kJ mol−1 , while in 2CBA and
2FBA the corresponding barriers are respectively, c. 37 and 40 kJ mol−1 , i.e. for the
fluoro-substituted compound the barrier is almost twice that found in BA.
It is instructive to examine the available experimental data on BA, 2FBA, and
2CBA to conformational isomerization in light of the above discussed potential
30
27
24
21
18
ΔE (kJ mol–1)
15
12
6
13.4
trans
3
1:1
0
–180 –120 –60 0 60 120 180
(a) C(H)–C–C=O (°)
30
27
24
21
13
12 11
18
ΔE (kJ mol–1)
10
1
14 15
15 6 2
5 3
12 9
4
7 28.2
8
6
cis
3
0
–180 –120 –60 0 60 120 180
(b) C(H)–C–C=O (°)
Figure 8.6 B3LYP/6-311++G(d,p) calculated potential energy profiles for internal

rotation about the exocyclic C–C for trans (a) and cis (b) arrangement of the carboxylic
group in benzoic acid. Source: Ildiz and Fausto [12], figure 3 (p. 6)/MDPI/Licensed under
CC BY 4.0.
energy landscapes. For the three compounds, only the lower energy cis con-
formers were observed in low-temperature infrared matrix isolation experiments
[70, 71, 74, 75]. The cis conformer of BA was also observed in the gas phase
both by electron diffraction experiments and microwave spectroscopy [68, 69],
and the gas phase infrared spectrum of this species has also been reported [72].
On the other hand, as already mentioned above, the trans conformer of BA has
never been observed experimentally. In the case of 2FBA, the intramolecularly
hydrogen-bonded trans conformer was detected in gas phase (together with the two
cis conformers) by microwave spectroscopy [76], and it has been produced in argon
and N2 cryomatrices as result of vibrational excitation of the cis conformers (using
infrared narrowband in situ irradiation) [74]. The existence of the highest-energy
trans conformer of 2FBA was inferred indirectly in the matrix isolation studies
reported in [74], but it could not be directly observed due to its fast spontaneous
conversion into lower energy conformers. For 2CBA, in situ UV irradiation of the
matrix-isolated cis conformers was shown to lead to their conversion into their trans
counterparts, which were then found to decay by quantum mechanical tunneling
back to the cis forms [75]. In the studies described in [75], the intramolecularly
hydrogen-bonded trans conformer was directly observed, while the highest-energy
trans conformer was only possible to observe experimentally upon OH → OD
isotopic substitution, to make its tunneling conversion into the cis form slow
enough.
The experimental observation of the cis forms was expected for all molecules, con-
sidering their low energies (see Figure 8.4) and the fact that they correspond to deep
minima on the potential energy surfaces. The intramolecularly hydrogen-bonded
trans conformer of 2FBA is also a well-defined minimum with a high energy
barrier (∼40 kJ mol−1 ) of conversion into its cis counterpart, and has an expected
population in the gas phase at room temperature of c. 2% [74], justifying its exper-
imental observation in the microwave experiment [76]. This trans form was not
detected in the deposited cryogenic matrices [74] most probably because its expected
population in the matrix is below the sensitivity of the technique. However, once
it is produced in situ by vibrational excitation of the matrix-isolated cis conformer,
the high energy barrier for the trans → cis conversion precludes this transformation
to take place (the over the barrier thermal process is not possible at all, and the
experiments showed that quantum mechanical tunneling is also inefficient in
this case). It can then be concluded that the intramolecularly hydrogen-bonded
trans conformer of 2FBA is a stable species. In its turn, the highest-energy trans
conformer of 2FBA is unstable due to the low energy barriers for conformational
isomerization, decaying fast into other forms [74].
In the case of 2CBA, as mentioned above, the highest energy non-hydrogen
bonded trans conformer could only be detected [75] upon deuteration of the
carboxylic group, which reduces its tunneling decay rate substantially allowing its
experimental detection under cryogenic conditions. The barrier for the trans → cis
conversion is, in this case, ∼27 kJ mol−1 [12] which is low enough to allow fast
tunneling. On the other hand, the 2CBA intramolecularly hydrogen-bonded trans
conformer could be experimentally detected after its in situ photoproduction in a
8.4 IMHB in Thiotropolone 225
cryogenic matrix [75], but found to decay slowly by quantum mechanical tunneling
into its cis counterpart (half-live of ∼10–30 minutes, at 9–45 K in a Xe matrix).
In this case, the trans → cis energy barrier is ∼37 kJ mol−1 , i.e. intermediate between
that of the hydrogen-bonded trans conformer in 2FBA, which is a stable species
(barrier: ∼40 kJ mol−1 ) and the barriers for all other trans → cis isomerizations in
BA, 2FBA, and 2CBA (between ∼27 and 14 kJ mol−1 ) that lead to fast tunneling and
unstable trans conformers.
Taking into account the discussion above for 2FBA and 2CBA, the reasons why
the trans conformer of BA was not experimentally detected both in the gas phase
experiments [68, 69, 72] and in the matrix isolation studies [70, 71] also become
clear: the barrier for its conversion into the cis form is only ∼23 kJ mol−1 , and stays
within the range of values which allows fast tunneling. The trans conformer of BA
is then also kinetically unstable, even at cryogenic temperatures. Note also that the
trans → cis barriers for all disubstituted compounds are of c. 31 kJ mol−1 [12], still
in the range of values for which fast tunneling is expected so that the trans con-
formers of these compounds can also be predicted to convert to the corresponding
cis form spontaneously by tunneling. This conclusion is in agreement with the data
reported in Ref. [12] for the chloro/fluoro-substituted compound.
In a summary, it can be concluded that the intramolecularly hydrogen bonds
present in the trans conformers of 2FBA and 2CBA considerably change the poten-
tial energy landscape so that these forms appear as unique species regarding chem-
ical stability among the trans conformers of the whole series of BAs investigated.
8.4 IMHB in Thiotropolone: Sculpturing

the Bidirectional Infrared-Induced
Bond-Breaking/Bond-Forming Tautomerization
Thiotropolone (TT) may exist in two tautomeric forms: the hydroxy (TT-OH) and
thiol (TT-SH) tautomers, each one having two conformers (Figure 8.7). However,
only the most stable conformer of TT-OH (s-OH-TT) can be trapped from the
gas phase into a cryogenic matrix. According to quantum chemical electronic
structure calculations, this species is more stable than the most stable SH-TT
conformer (s-SH-TT) by about 8 kJ mol−1 , while the higher-energy conformers of
each tautomer have considerably higher relative energies (∼57 and ∼20 kJ mol−1 ,
respectively for a-OH-TT and a-SH-TT; see Figure 8.7) [77]. The reason for the low
energies of the s-OH-TT and s-SH-TT conformers is the presence in these forms of
an IMHB: a strong OH· · ·S=C IMHB in s-OH-TT, and a weaker SH· · ·O=C IMHB
in a-SH-TT. The strong OH· · ·S=C IMHB characteristic of the s-OH-TT conformer
is especially relevant, since in practical terms it reduces the conformational space
of the molecule, under equilibrium conditions, to this conformer. As already
mentioned, it is the sole conformer that can be trapped from the gas phase into a
cryogenic matrix, and even in the gas phase room temperature equilibrium it is
predicted to account for more than 96% of the total conformational population [76].
O
H H
O O H O Figure 8.7 Tautomers of thiotropolone
S S S S H and its conformers, with their
B3LYP/6-311++G(2d,p) calculated
relative energies (in kJ mol−1 ).
Source: Nunes et al. [77]/American
s-OH-TT a-OH-TT s-SH-TT a-SH-TT
Chemical Society.
0.0 56.8 8.1 20.5
× 10−3
λ = 5940 cm−1
6.0
∆ Absorbance
0.0
–6.0
(a)
Computed
Relative intensity
40
–40
(b)
λ = 5994 cm−1
0.2
∆ Absorbance
0.0
–0.2
1600 1400 1200 1000 800 600

(c) Wavenumber (cm−1)
Figure 8.8 Experimental difference infrared (IR) spectrum (spectrum after matrix
irradiation “minus” spectrum of the same matrix before irradiation): (a) irradiation at
𝜆 = 5940 cm−1 (60 mW, one hour); (c) irradiation at 𝜆 = 5994 cm−1 (30 mW, one hour).
Bands marked with circles (●) are due to a-SH-TT, which is produced by irradiation at
𝜆 = 5940 cm−1 (a) and consumed by irradiation at 𝜆 = 5994 cm−1 (c). (b) Simulated
difference IR spectrum based on B3LYP/6 311+G(2d,p) vibrational data considering
quantitative (1 : 1) transformation of s-OH-TT into a-SH-TT. The symbols of tilde (∼)
designate truncated bands. Source: Reproduced from Nunes et al. [77] under permission
of the American Chemical Society. Copyright © 2020.
The narrowband vibrational excitation of the matrix-isolated intramolecularly

hydrogen-bonded s-OH-TT form, at frequencies corresponding to the first overtone
of the 𝜈(CH) modes or combination modes involving the 𝜈(CH) vibrations of the
molecule (e.g. 5980 and 5940 cm−1 ), was found to induce its tautomerization into
a-SH-TT (Figure 8.8) [77]. The identification of the photoproduced species was
8.4 IMHB in Thiotropolone 227
5994, 5947 cm–1
TS-OH(rot)
80
5980, 5940 cm–1
70 TS-SH(rot)
Relative energy (kJ mol–1)
60 (Filter)
a-OH 2200 cm–1
50
40
TS-(taut)
30
ΔEZPE
20
a-SH
10
s-SH
0
s-OH
–2 –1 0 1 2 3 –1 0 1 –3 –2 –1 0 1 2 3 4
OH-rotamerization Tautomerization SH-rotamerization
Intrinsic reaction coordinates (bohr)
Figure 8.9 B3LYP/6-311+G(2d,p) computed intrinsic reaction coordinate (IRC) profiles for
tautomerization (H-shift) and rotamerizations (OH- and SH-torsion) in thiotropolone. The
horizontal bars at each stationary point (minima and transition states) show the
corresponding computed zero point energy-corrected relative energy value. In the
experiments, a filter was used that did not allow IR radiation from the spectrometer IR
source with energy above 2200 cm−1 to reach the sample, to guarantee that the observed
reactions were not a result of broadband IR irradiation, but resulted exclusively from the
narrowband selective IR pumping. Source: Reproduced from Nunes et al. [77] under
permission of the American Chemical Society. Copyright © 2020.
unequivocal, considering the excellent match between the calculated infrared

spectrum for a-SH-TT and the observed characteristic bands of the photogenerated
species, as shown in Figure 8.8.
The experimental observations can be rationalized by taking into account the
potential energy surface profiles for tautomerization and rotamerizations of the rele-
vant isomers of TT (Figure 8.9). Excitation of s-OH-TT at 5980 or 5940 cm−1 provides
∼71 kJ mol−1 to the molecule. Such energy is below the TS for OH rotamerization
(∼76 kJ mol−1 ) but considerably above the TS for tautomerization to both s-SH-TT
(∼17 kJ mol−1 ) and a-SH-TT (∼53 kJ mol−1 ), so that both SH-TT conformers are a
priori possible to be formed (but not the a-OH-TT conformer). However, s-SH-TT
is a kinetically unstable form because of the low energy (∼9 kJ mol−1 ) and thin
(∼0.63 Å) barrier separating this form from s-OH-TT (see Figure 8.9). Such a small
and narrow barrier allows for occurrence of extremely fast quantum mechanical
tunneling even under cryogenic conditions (𝜏 1/2 = ∼1 × 10−11 seconds, according to
the Wentzel–Kramers–Brillouin [WKB] model [77]). Under these circumstances,
the a-SH-TT isomer becomes the only possible to observe photoproduct, as found
experimentally.
It shall be noticed that the observed process implies that the excited 𝜈(CH) over-
tone/combination modes are coupled to some extent to the reaction coordinates
for tautomerization and SH rotamerization, leading to a-SH-TT through an

intramolecular vibrational relaxation (IVR) process crossing two reaction coordi-
nates (see Figure 8.9). The alternative occurrence of a two-step process, involving
the formation of a thermalized s-SH-TT at an intermediate stage, is not possible,
because s-SH-TT has no significant bright transitions near 5980 or 5940 cm−1
that could activate this species for conversion into the observed a-SH-TT form
[77]. It shall also be emphasized that this is one of the very first examples of
bond-breaking/bond-forming infrared-induced reactions reported hitherto.
It was also found to be possible to revert the tautomerization reaction using the
same experimental strategy, i.e. converting a-SH-TT back to s-OH-TT by vibrational
excitation (5990 or 5947 cm−1 ) of a-SH-TT [77].
Since, the a-SH-TT isomer is not stabilized by an IMHB, this isomer sits on a
considerably less profound potential energy well along with the SH rotamerization
coordinate (∼33 kJ mol−1 ) compared to s-OH-TT along the OH rotamerization
coordinate (∼76 kJ mol−1 ). Accordingly, its conversion to s-SH-TT is easily induced
upon irradiation at 5990 or 5947 cm−1 (∼71 kJ mol−1 ). In turn, once formed, s-SH-TT
rapidly tautomerizes by tunneling to s-OH-TT. On the whole, this easily achieved
energetically sequence of processes leads to an observed much more efficient
a-SH-TT → s-OH-TT tautomerization, compared to the inverse process [77]. Hence,
it can be concluded that the OH· · ·S=C IMHB present in the s-OH-TT isomer
shapes the potential energy surface in such a way that, on one side, this form reacts
much less efficiently than the non-intramolecularly hydrogen-bonded a-SH-TT
isomer, and on the other side, it follows a substantially different tautomerization
mechanism of tautomerization.
8.5 Conclusion
In this chapter, the significant role of IMHBs in shaping the potential energy
surface of molecules and their physical and chemical properties has been high-
lighted. Examples have been provided, illustrating the major role of IMHB in
determining the structures and relative stability/reactivity of the conformers of
several compounds. In these examples, infrared spectroscopic data obtained for
the matrix-isolated compounds, combined with quantum chemical calculations
and in situ selective narrowband vibrational excitations, have been used to explore
the selected molecular systems and shed light on the influence of IMHB on their
chemical behavior.
The significance of IMHB in determining the structural characteristics of
halogen-substituted acetic acids and the stability of their conformers was described
in detail, in particular, the consequences of the intramolecular O—H· · ·F hydrogen
bond present in the trans conformer of the fluorinated compound, which appears
as a unique species regarding its structure and stability when compared with the
analogous trans conformers of both the chloro- and bromo-substituted compounds.
The distinctive capability of chlorine and fluorine to participate in intramolecular
hydrogen bonding in the series of ortho chloro- and fluoro-substituted BAs, and
References 229
the strong influence of IMHB on the preferred geometries and the barriers for
conformational isomerization in these molecules (and, as a result of the relative
heights of the barriers, also on the kinetical stability of the different conformers)
were addressed in detail. A rationale has been presented to explain a series of
experimental results previously reported in the literature on this type of compound
[12, 70, 71, 74, 75].
A rare example of a bond-breaking/bond-forming infrared-induced reaction,
the reversible tautomerization of TT, has been discussed, highlighting the role
of IMHB to determine the relative efficiency of the tautomerization in the direct
(s-OH-TT → a-SH-TT) and reverse (a-SH-TT → s-OH-TT) directions. The IMHB
that exists in the s-OH-TT isomer shapes the potential energy surface of the
molecule in such a way that it causes a substantial decrease in the tautomerization
reaction efficiency compared to the reverse process where the non-intramolecularly
hydrogen-bonded a-SH-TT isomer acts as the reactant species. Moreover, the
presence of the IMHB in s-OH-TT also results in a different tautomerization
mechanism.
Acknowledgments
The authors thank the Portuguese Science Foundation (Fundação para a Ciência e
a Tecnologia [FCT]), Lisbon (Portugal), for the financial support through Projects
UIDB/QUI/0313/2020, UIDP/QUI/0313/2020, and PTDC/QUI-QFI/1880/2020,
co-funded by COMPETE-UE. We also thank all members of the Laboratory
of Molecular Criospectroscopy and Biospectroscopy (LMCB) of the Coimbra
Chemistry Centre (present and past) as well as our coauthors from other research
laboratories who participated in the studies mentioned in this chapter.
References
1 Watanabe, M., Maemura, K., Kanbara, K. et al. (2002). GABA and GABA recep-
tors in the central nervous system and other organs. In: International Review of
Cytology, vol. 213 (ed. W.J. Kwang), 1–47. New York: Academic Press.
2 McCormick, D.A. (1989). J. Neurophysiol. 62: 1018–1027.
3 Blanco, S., López, J.C., Mata, S., and Alonso, J.L. (2010). Angew. Chem. Int. Ed.
49: 9187–9192.
4 Pauling, L. (1960). The Nature of the Chemical Bond. Ithaca, NY: Cornell
University Press (The first edition was published in 1939).
5 Latimer, W.M. and Rodebush, W.H. (1920). J. Am. Chem. Soc. 42: 1419–1433.
6 Werner, A. (1902). Liebigs Ann. Chem. 322: 261–296.
7 Pfeiffer, P., Fischer, P., Kuntner, J. et al. (1913). Liebigs Ann. Chem. 398: 137–196.
8 Pauling, L. (1936). J. Am. Chem. Soc. 58: 94–98.
9 Desiraju, G.R. and Steiner, T. (1999). The Weak Hydrogen Bond. Oxford: Oxford
University Press.
10 Grabowski, S.J. (ed.) (2006). Hydrogen Bonding: New Insights. Dordrecht,

Netherlands: Springer.
11 Gilli, P. and Gilli, G. (2009). The Nature of the Hydrogen Bond. Oxford: Oxford
University Press.
12 Ildiz, G.O. and Fausto, R. (2020). Molecules 25: 4908.
13 Giordanetto, F., Tyrchan, C., and Ulander, J. (2017). ACS Med. Chem. Lett. 8:
139–142.
14 Kuhn, B., Mohr, P., and Stahl, M. (2010). J. Med. Chem. 53: 2601–2611.
15 Davoren, J.E., O'Neil, S.V., Anderson, D.P. et al. (2016). Bioorg. Med. Chem. Lett.
26: 650–655.
16 Sakamoto, T., Koga, Y., Hikota, M. et al. (2014). Bioorg. Med. Chem. Lett. 24:
5175–5180.
17 de Vicente, J., Lemoine, R., Bartlett, M. et al. (2014). Bioorg. Med. Chem. Lett. 24:
4969–4975.
18 Miah, A.H., Copley, R.C.B., O'Flynn, D. et al. (2014). Org. Biomol. Chem. 12:
1779–1792.
19 Ettorre, A., D'Andrea, P., Mauro, S. et al. (2011). Bioorg. Med. Chem. Lett. 21:
1807–1809.
20 Labby, K.J., Xue, F., Kraus, J.M. et al. (2012). Bioorg. Med. Chem. 20: 2435–2443.
21 Hickey, J.L., Zaretsky, S., St. Denis, M.A. et al. (2016). J. Med. Chem. 59: 5368.
22 Rezai, T., Bock, J.E., Zhou, M.V. et al. (2006). J. Am. Chem. Soc. 128:
14073–14080.
23 Rozenberg, M., Shoham, G., Reva, I., and Fausto, R. (2005). Phys. Chem. Chem.
Phys. 7: 2376–2383.
24 Rozenberg, M., Shoham, G., Reva, I., and Fausto, R. (2004). Spectrochim. Acta,
Part A 60: 2323–2336.
25 Rozenberg, M., Loewenschuss, A., and Marcus, Y. (2000). Phys. Chem. Chem.
Phys. 2: 2699–2702.
26 Iogansen, A.V. (1999). Spectrochim. Acta, Part A 55: 1585–1612.
27 Rozas, I., Alkorta, I., and Elguero, J. (2001). J. Phys. Chem. A 105: 10462–10467.
28 Estácio, S.G., Cabral do Couto, P., Costa Cabral, B.J. et al. (2004). J. Phys. Chem.
A 108: 10834–10843.
29 Lopes, S., Domanskaya, A.V., Fausto, R. et al. (2010). J. Chem. Phys. 133:
144507.
30 Maçôas, E.M.S., Kriachtchev, L., Pettersson, M. et al. (2003). J. Am. Chem. Soc.
125: 16188–16189.
31 Maçôas, E.M.S., Khriachtchev, L., Pettersson, M. et al. (2004). J. Chem. Phys. 121:
1331–1338.
32 Fausto, R., Khriachtchev, L., and Hamm, P. (2010). Conformational changes in
cryogenic matrices. In: Physics and Chemistry at Low Temperatures (Chapter 3)
(ed. L. Khriachtchev), 51–84. World Scientific.
33 Pettersson, M., Lundell, J., Khriachtchev, L., and Räsänen, M. (1997). J. Am.
Chem. Soc. 119: 11715–11716.
34 Maçôas, E.M.S., Fausto, R., Pettersson, M. et al. (2000). J. Phys. Chem. A 104:
6956–6961.
References 231
35 Hakala, M., Marushkevich, K., Khriachtchev, L. et al. (2011). J. Chem. Phys. 134:
054506.
36 Maçôas, E.M.S., Lundell, J., Pettersson, M. et al. (2003). J. Mol. Spectrosc. 219:
70–80.
37 Paulson, L.O., Anderson, D.T., Lundell, J. et al. (2011). J. Phys. Chem. A 115:
13346–13355.
38 Nunes, C.M., Reva, I., and Fausto, R. (2021). Direct observation of tunnelling
reactions by matrix isolation spectroscopy. In: Tunnelling in Molecules: Nuclear
Quantum Effects from Bio to Physical Chemistry, Theoretical and Computational
Chemistry Series No. 18 (ed. J. Kästner and S. Kozuch), 1–53. London: The Royal
Society of Chemistry.
39 Maçôas, E.M.S., Kriachtchev, L., Fausto, R., and Räsänen, M. (2004). J. Phys.
Chem. A 108: 3380–3389.
40 Lapinski, L., Reva, I., Rostkowska, H. et al. (2013). J. Phys. Chem. A 117:
5251–5259.
41 Maçôas, E.M.S., Khriachtchev, L., Pettersson, M. et al. (2005). J. Phys. Chem. A
109: 3617–3625.
42 Bazsó, G., Góbi, S., and Tarczay, G. (2012). J. Phys. Chem. A 116: 4823–4832.
43 Lopes, S., Nikitin, T., and Fausto, R. (2019). J. Phys. Chem. A 123: 1581–1593.
44 Kuş, N. and Fausto, R. (2014). J. Chem. Phys. 141: 234310.
45 Reva, I.D., Stepanian, S., Adamowicz, L., and Fausto, R. (2001). J. Phys. Chem. A
105: 4773–4780.
46 Reva, I.D., Jarmelo, S., Lapinski, L., and Fausto, R. (2004). J. Phys. Chem. A 108:
6982–6989.
47 Kuş, N., Sharma, A., Peña, I. et al. (2013). J. Chem. Phys. 138: 144305.
48 Nunes, C.M., Lapinski, L., Fausto, R., and Reva, I. (2013). J. Chem. Phys. 138:
125101.
49 Bazsó, G., Magyarfalvi, G., and Tarczay, G. (2012). J. Phys. Chem. A 116:
10539–10547.
50 Bazsó, G., Najbauer, E.E., Magyarfalvi, G., and Tarczay, G. (2013). J. Phys. Chem.
A 117: 1952–1962.
51 Bazsó, G., Magyarfalvi, G., and Tarczay, G. (2012). J. Mol. Struct. 1025: 33–42.
52 Najbauer, E.E., Bazsó, G., Apóstolo, R. et al. (2015). J. Phys. Chem. B 119:
10496–10510.
53 Halasa, A., Lapinski, L., Reva, I. et al. (2014). J. Phys. Chem. A 118: 5626–5635.
54 Najbauer, E.E., Bazsó, G., Góbi, S. et al. (2014). J. Phys. Chem. B 118:
2093–2103.
55 Fausto, R., Batista de Carvalho, L.A.E., Teixeira-Dias, J.J.C., and Ramos, M.N.
(1989). J. Chem. Soc., Faraday Trans. 2 85: 1945–1962.
56 Fausto, R. (1994). J. Mol. Struct. THEOCHEM 315: 123–136.
57 Teixeira-Dias, J.J.C. and Fausto, R. (1986). J. Mol. Struct. 144: 199–213.
58 Maçôas, E.M.S., Kriachtchev, L., Pettersson, M. et al. (2003). J. Chem. Phys. 119:
11765–11772.
59 Pettersson, M., Maçôas, E.M.S., Khriachtchev, L. et al. (2002). J. Chem. Phys. 117:
9095–9098.
60 Reva, I., Nunes, C.M., Biczysko, M., and Fausto, R. (2015). J. Phys. Chem. A 119:
2614–2627.
61 Lapinski, L., Reva, I., Rostkowska, H. et al. (2014). J. Phys. Chem. B 118:
2831–2841.
62 Apóstolo, R.F.G., Bento, R.R.F., Tarczay, G., and Fausto, R. (2016). J. Mol. Struct.
1125: 288–295.
63 Apóstolo, R.F.G., Bento, R.R.F., and Fausto, R. (2015). Croat. Chem. Acta 88:
377–386.
64 Apóstolo, R.F.G., Bazsó, G., Ildiz, G.O. et al. (2018). J. Chem. Phys. 148: 044303.
65 Wiberg, K.B. (1986). J. Am. Chem. Soc. 108: 5817–5822.
66 Kanters, J.A., Kroon, J., Peerdeman, A.F., and Schoone, J.C. (1967). Tetrahedron
23: 4027–4033.
67 Fausto, R., Batista de Carvalho, L.A.E., and Teixeira-Dias, J.J.C. (1990).
THEOCHEM 207: 67.
68 Aarset, K., Page, E.M., and Rice, D.A. (2006). J. Phys. Chem. A 110: 9014–9019.
69 Onda, M., Asai, M., Takise, K. et al. (1999). J. Mol. Struct. 482/483: 301–303.
70 Stepanian, S.G., Reva, I.D., Radchenko, E.D., and Sheina, G.G. (1996). Vib. Spec-
trosc. 11: 123–133.
71 Reva, I.D. and Stepanian, S.G. (1995). J. Mol. Struct. 349: 337–340.
72 Bakker, J.M., Mac Aleese, L., von Helden, G., and Meijer, G. (2003). J. Chem.
Phys. 119: 11180.
73 Arendorf, J. R. T. (2011). A study of some non-covalent functional group interac-
tions. PhD thesis. University College, London.
74 Kuş, N. and Fausto, R. (2017). J. Chem. Phys. 146: 124305.
75 Nishino, S. and Nakata, M. (2007). J. Phys. Chem. A 111: 7041–7047.
76 Daly, A.M., Carey, S.J., Pejlovas, A.M. et al. (2015). J. Chem. Phys. 142: 144303.
77 Nunes, C.M., Pereira, N.A.M., Reva, I. et al. (2020). J. Phys. Chem. Lett. 11:
8034–8039.
233
Hydrogen Bonding from Perspective of Overtones and

Combination Modes: Near-Infrared Spectroscopic Study
Mirosław A. Czarnecki 1 , Yusuke Morisawa 2 , and Yukihiro Ozaki 3,4
1
University of Wrocław, Faculty of Chemistry, F. Joliot-Curie, 14, 50-383 Wrocław, Poland
2
Kindai University, School of Science and Engineering, Department of Chemistry, Kowakae, Higashi-Osaka,
Osaka 577-8502, Japan
3
Kwansei Gakuin University, School of Biological and Environmental Sciences, 1, Gakuen-Uegahara, Sanda,
Hyogo 669-1330, Japan
4
Toyota Physical and Chemical Research Institute, Yokomichi, Nagakute, Aichi 480-1192, Japan
9.1 Introduction
Applications of near-infrared (NIR) spectroscopy to the studies of hydrogen
bonding started in 1950s, and in 1960s, there was significant progress in this field
[1–4]. For example, in 1963 Bujis and Choppin [2] measured NIR spectra of pure
water and investigated its structure in relation to hydrogen bonds. In early 1970s,
Sandorfy’s group [3, 4] found that the relative intensities of the free OH or NH
stretching bands of alcohols and amines are much stronger for the overtones than
fundamentals, demonstrating that NIR spectroscopy holds unique characteristics
in hydrogen bonding research. Undoubtedly, Sandorfy is a pioneer in basic studies
of NIR spectroscopy. He is famous, particularly in the research on relation between
anharmonicity and hydrogen bonding.
NIR spectroscopy has the following characteristics, which are important for
hydrogen bond studies [1, 5, 6]:
(1) Bands due to the vibrational modes in the NIR region are derived from overtones
and combinations. There are many overlapping bands in the NIR spectra, and
thus, the assignment of the NIR bands is generally not easy.
(2) Most bands in the NIR region originate from functional groups containing a
hydrogen atom (e.g. OH, CH, NH). Hence, NIR spectroscopy is called “an XH
spectroscopic method.” Besides XH vibrational modes, the second overtones of
C=O stretching modes appear in the NIR region. Recently, bands due to the first
and second overtones of C≡N stretching modes of acetonitrile and its derivatives
were also observed (see Section 9.6).
(3) The first overtones of XH stretching bands show a redshift upon the hydro-
gen bonding formation as in the case of the corresponding fundamental bands.

234 9 Hydrogen Bonding from Perspective of Overtones and Combination Modes
This shift is much larger for the first overtones (nearly twice) than that for the
corresponding IR bands.
(4) In the NIR region, OH and NH stretching bands of monomeric and polymeric
species are better separated compared with the IR region. Moreover, in the NIR
spectra, one can differentiate bands due to terminal free OH or NH groups
(acceptor only) in the open chain associates from those originating from their
free OH or NH groups in the monomeric species.
(5) Because of the larger anharmonicity, the first overtone bands of OH and NH
stretching modes of monomeric species are more enhanced compared with the
corresponding bands of polymeric species. On the other hand, fundamental
bands originating from the OH and NH stretching modes of polymeric species
are more enhanced than those of the monomeric species because of a larger
charge separation of X–H (δ–X–Hδ+ · · ·: Y) in a hydrogen bonding.
It is noted that almost all of the above characteristics originate from the fact that
NIR spectroscopy is concerned with forbidden transitions within the harmonic oscil-
lator approximation [1].
Figure 9.1a,b compares an IR spectrum of methanol in CCl4 in the 3800–3000 cm−1
region and the corresponding NIR spectrum in the 7600–6000 cm−1 region. A peak
at 3630 cm−1 and that at 7090 cm−1 derives from a fundamental and a first overtone
of the stretching mode of free OH group of methanol, respectively, while broadbands
in the regions of 3500–3100 and 6800–6150 cm−1 are due to a fundamental and a first
overtone of the stretching mode of hydrogen-bonded OH groups of methanol dimer,
trimer and, higher oligomers. As can be seen, there is the correspondence between
Absorbance
3800 3600 3400 3200 3000

(a)
(b) 7600 7200 6800 6400 6000

Wavenumber (cm–1)
Figure 9.1 IR (a) and NIR (b) spectra of methanol in CCl4 in the 3800–3000 cm−1 and
7600–6000 cm−1 regions, respectively.
9.2 Investigation of Hydrogen Bonding of Water by NIR Spectroscopy 235
the NIR and IR spectra, but the relative intensities of the bands from the free and
hydrogen-bonded OH are significantly different.
In Chapter 10, we describe the unique characteristics of NIR spectroscopy in the
studies of hydrogen bonding by introducing several interesting examples.
9.2 Investigation of Hydrogen Bonding of Water by NIR

Spectroscopy
Figure 9.2 shows NIR spectra in the 12 000–4000 cm−1 region of liquid water mea-
sured in three different cells (1, 0.1, and 0.01 cm). Note that band intensities change
largely with the path lengths, and an intense foot appears near 4000 cm−1 originat-
ing from the fundamentals of the OH stretching modes. Water bands become weaker
and weaker step wisely as the wavelength goes to a shorter wavelength. Two strong
bands at 5235 and 6900 cm−1 are due to the combination of H–O–H antisymmetric
stretching mode (v3 ) and bending mode (v2 ) and that of H–O–H symmetric (v1 ) and
antisymmetric (v3 ) stretching modes, respectively.
Figure 9.3a shows temperature-dependent NIR spectra of water from 5 to
85 ∘ C [7]. As can be seen, the intensity at 7050 cm−1 increases while that at
6844 cm−1 decreases with temperature increase but the presence of the band shift
in the 7300–6000 cm−1 region is not obvious. To answer this question, we calculated
the difference spectra of water (Figure 9.3b) by subtracting the spectrum at 5 ∘ C
from the other spectra shown in Figure 9.3a [7]. It can be seen from Figure 9.3b
that the broad water feature consists of two bands at 7089 and 6718 cm−1 and that
there is no significant band shift. We assigned the peaks at 7089 and 6718 cm−1
to weakly hydrogen-bonded (WHB) and strongly hydrogen-bonded (SHB) water
2
Absorbance
(a)
(b)
(c)
0
12 000 10 000 8000 6000 4000
Wavenumber (cm–1)
Figure 9.2 NIR spectra of water in the 12 000–4000 cm−1 region measured in three
different cells of 1 cm (a), 0.1 cm (b), and 0.01 cm (c).
7050
6844
1.5
5–85 °C
5–85 °C
1.0
Absorbance
0.5
ν1 + ν2 + ν3 ν1 + ν3
0
9000 8000 7000 6000
0.6
7089
85 to 5 °C
0.3
Absorbance
15 to 5 °C
0
–0.3 6718
7500 7000 6500 6000

Figure 9.3 (a) Temperature-dependent NIR spectra of water from 5 to 85 ∘ C. (b) The
difference spectra obtained by subtracting the spectrum at 5 ∘ C from other spectra shown in
Figure 9.3a. Source: Reproduced from Ref. [7] with permission.
species, respectively [7]. From Figure 9.3b it is evident that the WHB species
gain in intensity while the SHB species decreases in intensity as the temperature
increases. Sasic et al. carried out a self-modeling curve resolution (SMCR) study
of temperature-dependent NIR spectra of water to investigate hydrogen bonding
of water [8]. Figure 9.4a,b shows the spectra and concentrations of SHB and WHB
after rescaling, respectively. It was found that a quasi-lattice model with broken
hydrogen bonds is highly supported. This study also showed that the two-state
model approximation of water spectra is very satisfactory but the residuals that
cover as small as 0.4% of the starting data are not noise. The most important
features of the residuals are the peak at 1400 nm and significant response in
the temperature range of 6–25 ∘ C. On this basis, one can conclude that the NIR
spectra of water do not represent a purely two-component system but a very close
system.
9.3 The Chain Length Effect on the Degree of Self-association of 1-Alcohols 237
2.5
2.0
1.5
Absorbance
SHB
1.0
0.5
WHB
0.0
(a) 1300 1350 1400 1450 1500 1550 1600

Wavelengths (nm)
0.85
0.80
0.75
0.70 SHB
0.65
0.60
Molar ratio
0.55
0.50
0.45
0.40
0.35
WHB
0.30
0.25
0.20
0.15
0 10 20 30 40 50 60 70 80 90
(b) Temperature (°C)
Figure 9.4 Pure component NIR spectra (a) and concentrations (b) of two kinds of species
(SHB and WHB) of pure water calculated by Simplisma. Source: Reproduced from Ref. [8]
with the permission.
9.3 The Chain Length Effect on the Degree

of Self-association of 1-Alcohols
Our chemical intuition says that an increase in the chain length should reduce
the degree of self-association of alcohols as a result of growing importance of the
hydrophobic interactions and better separation of the OH groups. In the litera-
ture, one can find indirect evidence of this commonly accepted assumption [9],
however, as yet any direct proofs are available. What is more interesting, some
of the authors suggest that the extent of self-association of 1-alcohols in CCl4
solutions does not depend on the alcohol’s chain length [10]. To elucidate the effect
15
C
10
5
4 5 6 7 8 9 10
(a) Chain length
0.075
0.05
S
0.025
6000 6500 7000 7500 8000 8500 9000

(b) Wavenumbers (cm–1)
Figure 9.5 Contribution (a) and spectral (b) profiles determined from MCR of NIR spectra
of alcohols from 1-butanol to 1-decanol. Source: Mirosław Czarnecki.
of the chain length on the extent of self-association of 1-alcohols from methanol

to 1-decanol we applied mid-infrared (MIR) and NIR spectroscopy together with
two-dimensional correlation spectroscopy (2DCOS) and chemometric methods
(cluster analysis – CA, multivariate curve resolution – MCR) [11]. The variable
part (perturbation) in this experiment was the chain length. The spectra of all
alcohols in the pure liquid phase were recorded under the same experimental
conditions and used for creation of the chain length ordered data matrix. In
Figure 9.5 are shown the results of application of MCR approach to the NIR spectra
of 1-alcohols. As can be seen (Figure 9.5b), the spectral profile of the methylene
group includes also the absorption from the free OH group. This means that the
population of both groups follows the same spectral pattern. The contribution
profiles (Figure 9.5a) reveal that the population of the methylene group and the
nonbonded OH groups increase. Of course, the population of the free OH groups
is growing at the expense of the bonded ones. Hence, we have a direct proof
that the degree of self-association of 1-alcohols decreases with the chain length
increase.
Another proof supporting above thesis provides the synchronous 2D correlation
spectrum shown in Figure 9.6. One can see a clear negative synchronous peak
between the first overtone of the associated OH (near 6300 cm−1 ) and the second
overtone of the methylene group (near 8225 cm−1 ).
Hence, one can conclude that the populations of the corresponding species are
changing in the opposite direction. In other words, an increase in the population
of the methylene groups (related to the chain length) decreases the popula-
tion of the associated OH groups. As can be seen (Figure 9.7), the MIR–NIR
hetero correlation synchronous spectrum develops a series of peaks. The most
9.3 The Chain Length Effect on the Degree of Self-association of 1-Alcohols 239
8500
8000
Wavenumber (cm–1), ν2
7500
7000
6500
6000 6500 7000 7500 8000 8500

Figure 9.6 Synchronous 2DCOS spectrum constructed from NIR spectra of alcohols from
1-butanol to 1-decanol. Positive peaks are drawn by solid lines, while the negative peaks
are drawn by dashed lines. Source: Mirosław Czarnecki.
9000
8500
8000
7500
7000
6500
2800 2900 3000 3100 3200 3300 3400 3500 3600

Figure 9.7 Synchronous 2D hetero-correlation spectrum constructed from MIR and NIR
spectra of alcohols from 1-butanol to 1-decanol. Positive peaks are drawn by solid lines,
while the negative peaks are drawn by dashed lines. Source: Mirosław Czarnecki.
prominent peaks occur between the first overtone of the associated OH group
(near 6300 cm−1 ) and the fundamental bands from the methyl and methylene
groups (3000–2800 cm−1 ). It is of note that the positive correlation peaks are
developed with the bands from the methyl group, while the negative correlation
peaks originate from the correlation with the methylene bands. This indicates
that the population of the bonded OH and the methylene groups are changing in
the opposite direction. An increase in the chain length increases the number of
methylene groups. As a result, the OH groups are better separated and decrease
the degree of self-association of alcohols. For long-chain alcohols, the ordering
of the chains may impact the OH· · ·OH interactions, but this effect is not well
recognized as yet.
9.4 Combined NIR and Dielectric Study on Association

of 1-Hexanol in n-Hexane
In the literature, one can find numerous works devoted to NIR studies on hydrogen
bonding and association in solutions of alcohols in hydrocarbons [12–15]. However,
the study of alcohol/hydrocarbon mixtures in the whole range of mole fractions is
occasional [16], despite great meaning of these mixtures in the science and indus-
try. Hence, we applied nonlinear dielectric effect (NDE) and NIR spectroscopy to
examine the process of association of 1-hexanol in n-hexane for mole fractions of
1-hexanol (X 1-hexanol ) from 0 to 1 with a step of 0.1 [17]. The experimental results
were supported by the theoretical (density functional theory [DFT]) calculations of
the structures and dipole moments of selected associates. Numerical fitting of the
dielectric data permitted for determination of the population of all species present in
the mixture. From these calculations, it results that, contrary to commonly accepted
dominant role of the cyclic tetramers, the most populated cyclic clusters are trimers.
At low values of X1-hexanol dominates the cyclic species, but increasing concentra-
tion of 1-hexanol shifts the equilibrium towards the open chain associates. These
associates dominate in the mixture for X1-hexanol > 0.6. Deconvolution of NIR spec-
tra provided the total concentration of the free OH groups, which may originate
from both the monomers and the terminal free OH in the linear associates. These
values were used for estimation of the mean size of the linear associates (nlin ). As
expected, the values of nlin increase with growing alcohol content but at all concen-
trations nlin determined from NIR data were significantly lower than those obtained
from the dielectric measurements (Figure 7 in ref. 17). One has to remind that NIR
data includes contributions from both the monomers and the free terminal OH in
the linear associates. In principle, the values determined from NIR spectra corre-
spond to the sum of both contributions obtained from the dielectric measurements.
It seems that one of the reasons for these large differences may result from the pres-
ence of the bifurcated associates, not considered in the calculations. However, the
more controversial explanation is that the band at 7100 cm−1 , assigned to the “free”
OH, includes only the contributions from the monomers, while the linear associates
absorb in a different spectral range. As can be seen (Figure 9.8), the population of
the monomers determined from the dielectric measurements is close to the values
obtained from the NIR spectra, supporting this supposition. Confirmation of this
interesting suggestion requires additional studies. By averaging the values of nlin
9.5 NIR Studies of Microheterogeneity in Alcohol/Alcohol and Alcohol/Alkane Binary Mixtures 241
0.2
0.15
XOH,free
0.1
0.05
0.2 0.4 0.6 0.8 1

X1-hexanol
Figure 9.8 Relationship between the mole fraction of 1-hexanol (X 1-hexanol ) and mole
fraction of the free OH (X OH,free ). Data obtained from dielectric measurements: OH in
monomers – ●, free OH in linear associates – ◼. Data obtained from NIR – ⧫. Source:
Mirosław Czarnecki.
from the dielectric and NIR measurements one can obtain a mean value of the size
of the linear associates as a function of the alcohol content. This value is changing
from 3 at X 1-hexanol = 0.1 to 15 in the pure liquid 1-hexanol (X 1-hexanol = 1). Proba-
bly, these values are overestimated, however, it seems that the higher associates are
stabilized by favorable interaction of the chains.
9.5 NIR Studies of Microheterogeneity

in Alcohol/Alcohol and Alcohol/Alkane Binary Mixtures
Recently, it has been found that fully miscible liquids like methanol and ethanol,
may form different kinds of microscopic clusters and reveal the microheterogeneity
at a molecular level [18]. This phenomenon seems to be particularly important
in the mixtures where exists the possibility of creation of the hydrogen bonding
between the components of the mixture. In particular, the mixtures of water with the
aliphatic alcohols were intensively studied by using different experimental and the-
oretical methods [19–22]. In contrast, examinations of the microheterogeneity in the
binary mixtures of alcohols are rare [18]. In reference [22], we reported combined
ATR-IR/NIR, chemometric, and DFT studies on the microheterogeneity in binary
mixtures of methanol with a series of aliphatic alcohols of different chain lengths
and structures (linear, cyclic). Our results evidence that all studied mixtures deviate
from the ideal mixture and the largest deviation was found in equimolar mixture
(X = 0.5). This conclusion was nicely confirmed by theoretical (DFT) calculations
of binding energies of different kinds of clusters. As can be seen from Figure 9.9,
in the whole range of mole fractions in the methanol/ethanol mixture coexists
the clusters of pure methanol, pure ethanol, and heteroclusters (mixed clusters).
0.8
0.6
C
0.4
0.2
0 0.2 0.4 0.6 0.8 1

Xethanol
Figure 9.9 Contribution profiles (in mole fractions) obtained from MCR of
composition-dependent NIR spectra of the methanol/ethanol binary mixtures.
Homoclusters of methanol – ◼, homoclusters of ethanol – ⧫, heteroclusters (mixed
clusters) – ●. Source: Mirosław Czarnecki.
0.05
0
Excess absorbance
–0.05
–0.1
–0.15
6000 6500 7000 7500 8000 8500 9000

Wavenumber (cm–1)
Figure 9.10 Composition-mean excess NIR absorption spectra of methanol/ethanol (solid,

upper), methanol/1-propanol (dotted), methanol/2-propanol (dash dotted), methanol/
cyclopentanol (dashed) and methanol/tert-butanol (solid, lower) binary mixtures. Source:
Mirosław Czarnecki.
Similar results were obtained for the other studied mixtures. An analysis of the
excess absorption spectra (Figure 9.10) reveals that the degree of nonideality
increases in going from methanol/ethanol to methanol/tert-butanol mixture. It
appears that the degree of deviation from the ideality in methanol/alcohol binary
mixtures is correlated with the alcohol’s order and the chain length. Interestingly,
these two factors directly impact the degree of association of alcohols [15]. Hence,
one can conclude that the structure of the binary mixtures of alcohols at a molecular
level and the degree of deviation from the ideality are determined by the strength of
9.5 NIR Studies of Microheterogeneity in Alcohol/Alcohol and Alcohol/Alkane Binary Mixtures 243
the OH· · ·OH hydrogen bonding, whereas the interactions of the alkyl chains are
of minor importance. However, for long-chain alcohols, one can expect increasing
impact of the chain on behavior of the mixture.
In binary mixtures of alcohols, both components may participate in the hydrogen
bonding. On the other hand, in alcohol/alkane mixtures only one component can
create the hydrogen bonding and associate through the OH· · ·OH interactions.
To examine the possibility of heterogeneity at a molecular level in alcohol/alkane
binary mixtures, we recorded ATR-IR and NIR spectra of the mixtures in the
whole range of mole fractions [23]. For studies, we used two alcohols (1-hexanol
and cyclohexanol) and two alkanes (n-hexane and cyclohexane). This way, we
obtained four binary mixtures of different compositions and properties. It has been
demonstrated that in most of the mixtures the homoclusters of alcohol and the
heteroclusters exist in the whole range of mole fractions, while the population of the
homoclusters of alkanes rapidly decreases and as a result they are present only up to
mole fraction of X alkane ≈ 0.6 (Figure 9.11). It is of note that the contribution profile
of the heterocluster has the maximum at cyclohexanol mole fraction of X ≈ 1/3.
This means that the most stable are the clusters with the alcohol: alkane molar ratio
of 1 : 2. In Figure 9.12 are displayed the excess NIR absorption spectra for all studied
mixtures. As can be seen, the largest deviation from the ideality appears for cyclo-
hexanol/n-hexane mixture, whereas the smallest one is for 1-hexanol/n-hexane
mixture. In the case of alcohol/alkane mixtures, the extent of microheterogeneity
depends on two factors: (i) the degree of association of alcohol and (ii) the similarity
of the molecular structures of the alcohol and solvent. Based on these results, one
can conclude that the stronger OH· · ·OH hydrogen bonding between molecules of
alcohol and more similar are the structures of alcohol and alkanes, the mixture is
more ideal.
0.8
0.6
C
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
Xcyclohexanol
Figure 9.11 Contribution profiles (in mole fractions) obtained from MCR of composition-
dependent NIR spectra of the cyclohexanol/cyclohexane binary mixtures. Homoclusters of
cyclohexane – ◼, homoclusters of cyclohexanol – ⧫, heteroclusters (mixed clusters) – ●.
Source: Mirosław Czarnecki.
0.12
Excess absorbance
0.08
0.04
6000 6500 7000 7500 8000 8500 9000

Wavenumber (cm–1)
Figure 9.12 Composition-mean excess NIR absorption spectra of 1-hexanol/n-hexanol

(solid), 1-hexanol/cyclohexane (dashed), cyclohexanol/n-hexane (dotted) and
cyclohexanol/cyclohexane (dash dotted) binary mixtures. Source: Mirosław Czarnecki.
9.6 Overtones of 𝛎C≡N Vibration as a Probe

of Molecular Structure of Nitriles
Overtones of the double bonds are usually not observed in NIR region, with
exception of the second overtone of the C=O bond [24]. However, by combination
of NIR spectroscopy with the high-level quantum mechanical calculations, for the
first time, we identified the first and second overtones of the 𝜈C≡N vibration in
the NIR spectra of liquid acetonitrile and its derivatives (CD3 CN, CCl3 CN) [25].
It appears that the intensity of these overtones provides valuable information on
the state of molecules in the liquid phase. The best anharmonic spectra of the cyclic
dimers were predicted at GVPT2//M06-2X/6-311++G(3df,3dp)+CPCM level of
theory. The vibrational spectra included the anharmonic transitions up to the third
order (first and second overtones as well as the binary and tertiary combinations).
The obtained results clearly demonstrate that the intensity changes in going from
the fundamentals to the first and second overtones of the 𝜈CN vibration are closely
related to the symmetry of the dimers. The theoretical calculations reveal that
the dimers of CH3 CN and CD3 CN have C2h symmetry and therefore, the first
overtone of the 𝜈C≡N vibration should be inactive in IR/NIR spectra. In contrast,
the dimer of CCl3 CN has C1 symmetry and therefore its first overtone should be
IR/NIR active. The difference in the symmetry of the dimers is due to larger size
of Cl atom as compared to H or D atoms. According to the theoretical predic-
tions, the second overtones of all three acetonitriles are IR/NIR active. A careful
examination of the NIR spectra in the region of the first (Figure 9.13) and second
(Figure 9.14) overtones of the 𝜈C≡N vibration confirms the conclusions derived
from the theoretical calculations. The first overtone of this vibration for CCl3 CN was
predicted at 4472 cm−1 , while in the experimental spectrum it appears at 4478 cm−1 .
Hence, one can observe an excellent agreement between the experimental and
9.6 Overtones of νC≡N Vibration as a Probe of Molecular Structure of Nitriles 245
4
Absorbance
4450 4460 4470 4480 4490 4500 4510 4520

Wavenumber (cm–1)
Figure 9.13 NIR spectra of CH3 CN (solid), CD3 CN (dashed), and CCl3 CN (dash dotted) in
the 4450–4520 cm−1 region. The arrow indicates the position of the first overtone of the
νC≡N vibration. Source: Mirosław Czarnecki.
×10–3
20
15
Absorbance
10
6650 6700 6750 6800

Wavenumber (cm–1)
Figure 9.14 NIR spectra of CH3 CN (solid), CD3 CN (dashed), and CCl3 CN (dash dotted) in
the 6650–6800 cm−1 region. The arrows indicate the positions of the second overtones of
the νC≡N vibration. Source: Mirosław Czarnecki.
theoretical data. Also, good agreement was observed for the second overtones
of the 𝜈C≡N vibration. The theoretical/experimental positions were as fol-
lows: 6712/6724 cm−1 (CH3 CN), 6726/6732 cm−1 (CD3 CN), and 6708/6677 cm−1
(CCl3 CN). Again, the theoretical predictions were very close to the experimental
band positions. Concluding, exploration of the overtone bands coupled with
the high-level theoretical calculations is a powerful tool for obtaining valuable
information on the molecular structure and interactions.
9.7 Weak Hydrogen Bond in Poly(3-Hydroxybutyrate)

(PHB) Studied by NIR Spectroscopy
Hu et al. [26] investigated a weak hydrogen bond in poly(3-hydroxybutyrate) (PHB)

by NIR spectroscopy. They compared NIR spectra of PHB with the corresponding
IR spectra obtained by Sato et al. [27, 28]. PHB is one of the most often investigated
biodegradable polymers. Sato et al. found that a crystalline C=O stretching band
appears at 1723 cm−1 , is located at a much lower wave number by about 20 cm−1
compared with an amorphous C=O stretching band at around 1740 cm−1 and
that a crystalline CH3 asymmetric stretching band appears at an anomalously
high frequency (3009 cm−1 ). The IR results were complemented by X-ray crys-
tallographic measurements of PHB. Based on these results they concluded that
the CH3 and C=O groups of PHB form a weak C—H· · ·O=C hydrogen bond-
ing and a chain of C—H· · ·O=C bonds pair combines the two parallel helical
structures [27, 28].
H O
H3C
O O
C C
c
b
a
C
O
H2
n C atom O atom H atom
Figure 9.15 shows chemical and lamellar structures of PHB and (a) time-
dependent changes in IR spectra in the 3050–2850 cm−1 region of a PHB film during
the melt crystallization process at 125 ∘ C [26]. In Figure 9.15b the second derivatives
of the spectra shown in (a) for 0 and 180 minutes are depicted. The second derivative
spectrum for 180 minutes shows many bands due to semicrystalline PHB, while
only four bands are identified in the spectrum of the amorphous PHB. A band at
3007 cm−1 originates from crystalline CH3 asymmetric stretching band. Figure 9.16
shows time-dependent NIR spectra in the 6050–5650 cm−1 region of a PHB film
during the melt crystallization process at 125 ∘ C [26]. The second derivative of the
spectra measured at 0 and 180 minutes are shown in Figure 9.16b. As can be seen,
the spectrum collected at 0 minutes yields only four amorphous bands at around
5954, 5913, 5828, and 5768 cm−1 , while the spectrum obtained at 180 minutes shows
at least seven bands at 5973, 5952, 5917, 5900, 5811, 5757, and 5681 cm−1 in the same
spectral region. It is of note that the spectral variations in the 6050–5650 cm−1 region
and the corresponding changes in the 3050–2840 cm−1 region show significant
similarities. For example, the highest wave number NIR band at 5973 cm−1 and the
corresponding IR band at 3007 cm−1 show similar time-dependent changes. It is
9.7 Weak Hydrogen Bond in Poly(3-Hydroxybutyrate) (PHB) Studied by NIR Spectroscopy 247
Figure 9.15 Chemical and 0.12

lamellar structures of PHB.
(a) Time-dependent 0.10
variations in IR spectra in
the 3050–2850 cm−1 region
Absorbance
0.08
of PHB film during the melt
crystallization process at
125 ∘ C. (b) Second derivative 0.06
of the spectra at 0 and
180 minutes. Source: 0.04
Reproduced from Ref. [26]
with the permission. 0.02
0.00
3050 3000 2950 2900 2850
× 10–3
1
Amorphous (0 min)
0
Second derivative
2852
2878
–1 2937
2986
Semi-crystalline
–2 (180 min)
2987
2966
2853
2873
–3
3007
2997
2927
2935
2976
–4
3050 3000 2950 2900 2850
very likely that the former band is the first overtone of the latter band. It is of note
that like the fundamental band, the first overtone is located at unusually high wave
numbers. Therefore, the NIR spectra also suggest the existence of the C—H· · ·O=C
hydrogen bonding [26].
Figure 9.17a,b shows temporal variations in the NIR spectra in the 5200–5060 cm−1
region and the corresponding IR spectra in the 1780–1670 cm−1 region of a PHB
film during the melt crystallization process at 125 ∘ C, respectively [26]. It is noted
that a band at 5127 cm−1 in the NIR spectra gradually gains in intensity during the
crystallization process, while a broad feature centered at 5160 cm−1 decreases in
intensity with time, suggesting that the former band is due to the crystalline sample
and the latter band to the amorphous one. In the corresponding IR spectra, bands
at 1722 and 1743 cm−1 are assigned to C=O stretching modes of the crystalline and
amorphous states, respectively. Of note is that the NIR and IR spectra show very
0.05
0.04
5973
Absorbance
0.03
0.02
0.01
0.00
6000 5900 5800 5700
× 10–4
2
Amorphous (0 min)
0
5828
Second derivative
5768
5954
5913
–2
Semi-crystalline (180 min)
–4
5681
5811
5952
5917
–6
5973
5900
5757
6000 5900 5800 5700

Figure 9.16 (a) Time-dependent variations in NIR spectra in the 6050–5650 cm−1 region
of a PHB film during the melt crystallization process at 125 ∘ C. (b) Second derivative of the
spectra measured at 0 and 180 minutes. Source: Reproduced from Ref. [26] with the
permission.
clear correspondence. On this basis Hu et al. [26] assigned the band at 5127 cm−1
to the second overtone of the C=O stretching mode of the C—H· · ·O=C hydrogen
bonding in the crystalline state and the broad feature near 5160 cm−1 to the
corresponding amorphous state. This study showed that NIR spectroscopy can be
used for the investigations of weak hydrogen bonding. This method is particularly
useful for studies of weak hydrogen bonds in bulk or thick samples. Our results
also demonstrated that a band originating from the second overtone of the C=O
stretching mode is very useful for investigation of the hydrogen bonding of the C=O
group [26].
0.015
Absorbance
0.010
0.005
0.000
5200 5180 5160 5140 5120 5100 5080 5060
1.2
1.0
Absorbance
0.8
0.6
0.4
0.2
0.0
1780 1760 1740 1720 1700 1680
Figure 9.17 (a) Time-dependent variations in the NIR spectra in the 5200–5060 cm−1
region of a PHB film during the melt crystallization process at 125 ∘ C. (b) The corresponding
IR spectra in the 1780–1670 cm−1 region. Source: Reproduced from Ref. [26] with
permission.
9.8 Studies of Hydrogen Bonding By Use of Higher

Overtones
Gonjo et al. [29] employed the first, second, and third overtones of OH stretch-
ing mode of phenol and 2,6-dihalogenophenols to investigate the inter- and
intramolecular hydrogen bonding. Figure 9.18 displays Vis/NIR/IR spectra in
the OH fundamental and its first, second, and third overtone regions of phenol,
2,6-difluorophenol, 2,6-dichlorophenol, and 2,6-dibromophenol in n-hexane, CCl4 ,
CHCl3 , and CH2 Cl2 , respectively [29]. The authors selected phenol and three
kinds of 2,6-dihalogenated phenols because phenol may form an intermolecular
300 Phenol Phenol

n-Hexane 4
n-Hexane
CCl4
200 CCl4
CHCl3
CHCl3
CH2Cl2 2
100 CH2Cl2
0 0
300 2,6-Difluorophenol 2,6-Difluorophenol
4
200
2
ε (M–1cm–1)
ε (M–1cm–1)
100
0 0
300 2,6-Dichlorophenol 2,6-Dichlorophenol
4
200
2
100
0 0
300 2,6-Dibromophenol 4 2,6-Dibromophenol
200
2
100
0 0
3700 3600 3500 3400 7200 7000 6800 6600
Phenol 0.008 Phenol

0.15 n-Hexane
n-Hexane
CCl4
0.10 CCl4
CHCl3
CHCl3 0.004
CH2Cl2
0.05 CH2Cl2
0.00 0.000
0.15 2,6-Difluorophenol 0.008 2,6-Difluorophenol
0.10
0.004
ε (M–1cm–1)
ε (M–1cm–1)
0.05
0.00 0.000
2,6-Dichlorophenol 0.008 2,6-Dichlorophenol
0.15
0.10
0.004
0.05
0.00 0.000
2,6-Dibromophenol 2,6-Dibromophenol
0.15 0.008
0.10
0.004
0.05
0.00 0.000
10500 10200 9900 9600 13600 13200 12800 12400
(c) Wavenumber (cm–1) (d) Wavenumber (cm–1)
Figure 9.18 Vis/NIR/IR spectra in the OH stretching band region of phenol,

2,6-difluorophenol, 2,6-dibromophenol in n-hexane, CCl4 , CHCl3 , and CH2 Cl2 .
(a) Fundamental region. (b) First overtone region. (c) Second overtone region. (d) Third
overtone region. Source: Reproduced from Ref. [29] with the permission. Copyright (2011)
OH· · ·OH hydrogen bonding and the OH· · ·Cl hydrogen bonding with the solvent.
In contrast, 2,6-dihalogenated phenols may create a weak to strong intramolec-
ular hydrogen bond and a weak to very weak intermolecular hydrogen bond
with the solvent. Among selected samples, 2,6-difluorophenol has much weaker
intramolecular hydrogen bonding and relatively stronger interactions with the
solvent. Figure 9.18 demonstrates the solvent-dependent spectral changes for the
fundamentals and overtones. The area of the fundamental bands decreases upon
going from n-hexane to CCl4 , CHCl3 , and CH2 Cl2 , while for the first overtones an
opposite trend is observed.
Figure 9.19 shows the relative intensities of the OH stretching band of phenol,
2,6-difluorophenol, 2,6-dichlorophenol, and 2,6-dibromophenol in CCl4 , CHCl3 ,
and CH2 Cl2 for the fundamentals, first, second, and third overtones [29]. The
intensity of the OH band of phenol in n-hexane was used as an internal standard.
Interestingly, the relative intensities of the OH stretching vibrations of phenol in
CCl4 , CHCl3 , and CH2 Cl2 increase for the fundamentals and the second overtones
but decrease in the first and third overtones. The relative intensities of the OH
stretching vibrations show the so-called “parity” effect [29]. The parity effect is more
prominent for phenol than for 2,6-dihalogenated phenols probably because phenol
has a stronger intermolecular OH· · ·Cl hydrogen bonding with the solvents in con-
trast to its derivatives that have a weaker intermolecular hydrogen bond. Due to this
reasoning, Gonjo et al. concluded that the intermolecular phenol–solvent OH· · ·Cl
hydrogen bond is responsible for the tendencies observed in Figure 9.19 [29].
This interesting phenomenon was successfully explained using an explicit con-
sideration of solute–solvent interactions in combination with modern grid-based
methods to solve the time-independent Schrödinger equation [30]. It was suggested
that the electrical anharmonicity of the system, manifested as a nonlinear depen-
dence of the transition dipole moments with respect to the nuclear coordinates, may
contribute to the parity effects observed by Gonjo et al. [29].
2.5
Phenol 2,6-Difluoro 2,6-Dichloro 2,6-Dibromo
phenol phenol phenol
Relative intensity
2.0
CH2Cl2
CHCl3
CCl4
1.5
1.0
0.5
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
V V V V
Figure 9.19 Relative intensities of the OH stretching band of phenol, 2,6-difluorophenol,

2,6-dichlorophenol, and 2,6-dibromophenol in CCl4 , CHCl3 , and CH2 Cl2 for the
fundamentals, first, second, and third overtones. Source: Reprinted from Ref. [29] with
permission. Copyright (2016) American Chemical Society.
9.9 Comparison of Hydrogen Bonding Effects

and Solvent Effects on Wave numbers and Intensities
of the Fundamental and First Overtone of the N–H
Stretching Mode of Pyrrole Studied By NIR/IR
Spectroscopy and One-Dimensional Vibrational
Schrödinger Equation Approach
Futami et al. studied effects of hydrogen bonding and solvent on wave numbers and
intensities of the fundamental and first overtone of the N–H stretching mode of pyr-
role by NIR/IR spectroscopy and one-dimensional vibrational Schrödinger equation
approach [31–33]. They revealed that NIR spectroscopy can discriminate spectral
variations induced by hydrogen bonding formation from those due to solvent
changes. Figure 9.20a–c display NIR/IR spectra in the region of 8000–3000 cm−1
of pyrrole, pyridine, and pyrrole–pyridine complex in CCl4 , respectively [31].
Figure 9.20d shows the difference spectrum, (c) – (a) – (b). Figure 9.21a–d shows
enlarged spectra in the region of 7500–5500 cm−1 . From Figures 9.20d and 9.21d
is seen that the intensity of the first overtone of hydrogen-bonded N–H stretching
mode decreases on the formation of pyrrole–pyridine complex (almost missing)
while the corresponding fundamental band significantly gains in intensity.
To explore these interesting results, Futami et al. investigated the changes in
the vibrational potential, vibrational energies, wave functions, and transition
dipole moment of N–H stretching mode on the formation of hydrogen bonding
1.0
Pyrrole
0.5
(a) 0.0
1.0
Pyridine
0.5
Absorbance
(b) 0.0
1.0
Pyrrole and pyridine
0.5
(c) 0.0
0.10
0.05 (c)–(a)–(b)
0.00
–0.05
–0.10
(d) 8000 7000 6000 5000 4000 3000
Wavenumber (cm–1)
Figure 9.20 FT-NIR/IR spectra of (a) pyrrole, (b) pyridine, and (c) pyrrole + pyridine in CCl4
solutions. (d) Difference spectrum (c) − (a) − (b). Concentrations of pyrrole and pyridine
were 0.04 M for all solutions. Source: Reproduced with permission from Ref. [31]. Copyright
2009 Elsevier.
9.9 Comparison of Hydrogen Bonding Effects and Solvent Effects on Wave numbers and Intensities 253
0.015
Pyrrole
0.010
0.005
(a) 0.000
0.015
Pyridine
0.010
0.005
Absorbance
(b) 0.000
0.015 Pyrrole and pyridine
0.010
0.005
(c) 0.000
0.002
(c)–(a)–(b)
0.001
0.000
–0.001
(d) –0.002
7500 7000 6500 6000 5500
Wavenumber (cm–1)
Figure 9.21 FT-NIR spectra in the NH stretching overtone region of (a) pyrrole,
(b) pyridine, and (c) pyrrole + pyridine in CCl4 solutions. (d) Difference spectrum
(c) − (a) − (b). Concentrations of pyrrole and pyridine were 0.04 M for all solutions. Source:
Reproduced with permission from Ref. [31]. Copyright 2009 Elsevier.
between pyrrole and pyridine using the one-dimensional Schrödinger wave

equation [31]. Figure 9.22 displays vibrational potential, vibrational wave functions,
and dipole moment functions along potential energy curves of the NH stretching
mode of (a) pyrrole monomer and (b) pyrrole–pyridine complex calculated at the
DFT//B3LYP/6-311++G(3df,3pd) level. The results shown in Figure 9.22 revealed
that the vibrational potential, vibrational energies, wave functions, and transition
dipole moment of N–H stretching mode are substantially changing upon the
formation of the hydrogen bonding [31]. The marked variation in the vibrational
potential curve and the overlap of wave functions suggest an increase in the relative
(nonbonded/hydrogen-bonded) intensity of the overtone mode versus the funda-
mental one in this case. It is of particular note that the dipole transition moment of
N–H stretching mode largely decreases on the hydrogen-bonded complex formation
(Figure 9.22), in contrast to nonbonded pyrrole molecule [31]. Accordingly, Futami
et al. [31] concluded that the dipole moment changes play a key role in the intensity
decrease of the first overtone of NH stretching mode of pyrrole on the formation of
hydrogen-bonded complex with pyridine.
Futami et al. also investigated solvent effect on the intensities and wave numbers
of the fundamental and first overtone modes of N–H stretching vibration of pyr-
role [32]. Figure 9.23 shows solvent dependence of intensities and wave numbers of
the (a) fundamental and (b) first overtone modes of N–H stretching vibration of pyr-
role in CCl4 , CHCl3 , and CH2 Cl2 [32]. The spectra show clear solvent dependence
50 000 5
z
40 000 0
Dipole moment (Debye)

x,y
Energy (cm–1)
30 000 –5
20 000 –10
10 000 –15
0 –20
–0.4 –0.2 0.0 0.2 0.4 0.6 0.8
(a) Normal coordinate qM / qM0
50 000 5
x,y
40 000 0

z
Energy (cm–1)
30 000 –5
20 000 –10
10 000 –15
0 –20
–0.4 –0.2 0.0 0.2 0.4 0.6 0.8
(b) Normal coordinate qC / qC0
Figure 9.22 Vibrational wave functions and dipole moment functions along potential
energy curves of the NH stretching mode of (a) pyrrole monomer and (b) pyrrole–pyridine
complex calculated at the DFT//B3LYP/6-311++G(3df,3pd) level. Symbols qM0 and qC0
denote units of the normal coordinates for the NH stretching mode in the monomer and the
complex, respectively. Source: Reproduced from Ref. [31] with the permission. Copyright
(2009) Elsevier.
for both the fundamental and first overtone of the NH stretching band of pyrrole
(Figure 9.23). It was found that the static relative dielectric constant 𝜀r (CCl4 – 2.2,
CHCl3 – 4.8, and CH2 Cl2 – 8.9) plays a key role in the solvent dependences of the
IR and NIR spectra. With increasing 𝜀 of the solvent, the wave numbers of both the
fundamental and first overtone of the NH stretching bands decrease. Furthermore,
the corresponding intensities increase and this increase is more pronounced for the
fundamental band in contrast to the first overtone band. It was noted that the wave
numbers of fundamental and overtones of the NH stretching mode decrease in the
following order: CCl4 > CHCl3 > CH2 Cl2 . Moreover, the intensity of these two bands
9.9 Comparison of Hydrogen Bonding Effects and Solvent Effects on Wave numbers and Intensities 255
CCl4
1.2
CHCl3
1.0 CH2Cl2
0.8
Absorbance
0.6
0.4
0.2
0.0
3550 3500 3450 3400
0.015
CCl4
0.012
CHCl3
Absorbance
0.009
CH2Cl2
0.006
0.003
0.000
6950 6900 6850 6800 6750 6700
(b)
–1)
Wavenumber (cm
Figure 9.23 Solvent dependence of absorption intensities and wave numbers of the
(a) fundamental and (b) first overtone modes of N–H stretching vibration of pyrrole in CCl4 ,
CHCl3 , and CH2 Cl2 . Source: [32]. Figure 2 (p. 1196)/Reproduced with permission of American
Chemical Society.
increases in the same order, and this effect is more pronounced for the fundamen-
tal than the overtone band. The study suggested that the solvent-dependent shift on
𝜀r results from the anharmonicity of the vibrational potential. However, the inten-
sity variations originate from changes in the slope of the dipole moment function
(Figure 9.24) [32]. Therefore, it was found that both the mechanical and electri-
cal anharmonicity have a distinct and nontrivial impact on the NIR spectra [32].
Interestingly, the spectral variability of the NH stretching bands of pyrrole caused
by the solvent effect is quite different from the that resulting from the hydrogen
bonding. This way, one can discriminate between the hydrogen bonding and solvent
effects by comparing the intensities of the fundamentals and the corresponding first
overtones.
ε = 10 1.2
0.8
0.4
(c) 0.0
ε=1
300
200
100
ε=1
0
–100
(b) –200
ε = 10
50,000 4
45,000

40,000 3
Energy (cm–1)
35,000
30,000 2
25,000
20,000 1
15,000
10,000 0
5,000
0 –1
–0.4 –0.2 0.0 0.2 0.4 0.6 0.8
(a) Normal coordinate q/q0
Figure 9.24 (a) Dependences on 𝜀 of the potential energy curve, dipole moment function
(𝜀 = 1–10), and wave function (𝜀 = 1) of the NH stretching mode (𝜀 = 1–10). (b) Difference
in the potential energy curve between the calculation result for 𝜀 = 1 and a variety of 𝜀. (c)
Difference of the dipole moment function between the calculation result for 𝜀 = 1 and a
variety of 𝜀. Source: Reproduced with permission from Ref. [32]. Copyright 2011 American
Chemical Society.
9.10 Summary
This chapter is dedicated to investigation of hydrogen bonding and various manifes-

tations of this phenomenon by NIR spectroscopy. This technique offers some ben-
efits for hydrogen bonding studies, as compared with the IR spectroscopy. The free
and weakly bonded OH or NH groups are more pronounced in the NIR spectra as
compared with bonded groups, while in the IR spectra an opposite situation occurs.
Hence, many aspects of the hydrogen bonding are better seen in the NIR spectra.
However, a complementary application of the IR and NIR spectroscopy guided by
theoretical calculations provides more comprehensive information on the hydrogen
bonding. For example, the combination of information from both spectral ranges
enables the studies of the hydrogen bonding effect on anharmonicity, vibrational
potential, and dipole moments.
Temperature-dependent NIR spectra of liquid water support quasi-lattice model
with broken hydrogen bonds and suggest that the spectral data do not represent a
References 257
purely two-component system but a system that is very close [7]. Analysis of IR and
NIR spectra provided direct evidence that the degree of self-association of 1-alcohols
decreases with the increasing size of the alcohol [11]. Combination of NIR spec-
troscopy with NDE measurements allows us to estimate the size and population
of all species present in the 1-hexanol/n-hexane mixture [17]. Other NIR/ATR-IR
studies have shown that the hydrogen bonding is responsible for the heterogene-
ity at a molecular level in seemingly homogeneous binary mixtures of alcohols [22]
and alcohols with alkanes [23]. An additional important factor responsible for the
microheterogeneity and deviation from the ideality is the molecular shape of the
components of the mixture. Interesting results offer the theoretical analysis of the
anharmonic vibrational spectra of the dimers of acetonitrile and its derivatives [25].
It has been shown that the overtones of the 𝜈C≡N vibration may provide infor-
mation on the molecular structure of nitriles in the liquid phase. Examination of
PHB by NIR spectroscopy has shown that the second overtone of the C=O stretch-
ing mode can be used to monitor the state of the C—H· · ·O=C hydrogen bond-
ing in the crystalline and amorphous state [26]. Hydrogen bonding in phenol and
2,6-dihalogenophenols in different solvents (CCl4 , CHCl3 , and CH2 Cl2 ) was studied
by use of the higher overtones (first to third) of the OH stretching vibration [29].
The intensity of the successive overtones reveals so-called “parity effect” resulting
from the solute–solvent OH· · ·Cl interaction. A series of papers [31–33] has exam-
ined the solvent and hydrogen bonding effect on intensities and positions of the
fundamental and the first overtones of the N–H stretching vibration of pyrrole. As
shown, NIR/IR spectroscopy supported by one-dimensional vibrational Schrödinger
equation approach allows for discrimination of the hydrogen bonding and the sol-
vent effect by comparison of the intensities of the fundamental and first overtone of
the N–H stretching vibration.
Acknowledgments
M. A. C. acknowledges financial support from the National Science Center, Poland,

grant 2017/27/B/ST4/00948. Theoretical calculations have been carried out in
Wrocław Centre for Networking and Supercomputing under grant no. 163.
References
1 Ozaki, Y., Huck, C., Tsuchikawa, S., and Engelsen, S.B. (2020, 2021).
Near-infrared Spectroscopy, Theory, Spectral Analysis, Instrumentation, and
Applications, 297–330. Springer.
2 Bujis, K. and Choppin, G.R. (1963). Near-infrared studies of the structure of
water. I. Pure water. J. Chem. Phys. 39: 2035–2041.
3 Bernard-Houplain, M.C. and Sandorfy, C. (1974). On the similarity of the relative
intensities of Raman fundamentals and infrared overtones of free and hydrogen
bonded X-H stretching vibrations. Chem. Phys. Lett. 27: 154–156.
4 Sandorfy, C., Buchet, R., and Lachenal, G. (2007). Chapter 2: Principle of

molecular vibrations for near-infrared spectroscopy. In: Near-Infrared
Spectroscopy in Food Science and Technology (ed. Y. Ozaki, W.F. McClure and
A.A. Christy), 11–46. Wiley-Interscience.
5 Czarnecki, M.A., Morisawa, Y., Futami, Y., and Ozaki, Y. (2015). Advances in
molecular structure and interaction studies using near-infrared spectroscopy.
Chem. Rev. 115: 9707–9744.
6 Ozaki, Y., Huck, C.W., and Beć, K.B. (2017). Near-infrared spectroscopy and its
applications. In: Molecular and Laser Spectroscopy: Advances and Applications
(ed. V.P. Gupta), 11–38. Elsevier.
7 Maeda, H., Ozaki, Y., Tanaka, M. et al. (1995). Near infrared spectroscopy and
chemometrics studies of temperature-dependent spectral variations of water:
relationship between spectral changes and hydrogen bonds. J. NIR Spectrosc.
3: 191–201.
8 Sasic, S. and Segtnan, H. (2002). Self-modeling curve resolution study
temperature-dependent near-infrared spectra of water and the investigation of
water structure. J. Phys. Chem. A 106: 760–766.
9 Czarnecki, M.A. and Ozaki, Y. (1999). The temperature-induced changes
in hydrogen bonding of decan-1-ol in the pure liquid phase studied by
two-dimensional Fourier transform near-infrared correlation spectroscopy. Phys.
Chem. Chem. Phys. 1: 797–800.
10 Wilson, L., Alencastro, R.B., and Sandorfy, C. (1985). Hydrogen bonding of
n-alcohols of different chain lengths. Can. J. Chem. 63: 40–45.
11 Kwaśniewicz, M. and Czarnecki, M.A. (2018). The effect of chain length on
mid-infrared and near-infrared spectra of aliphatic 1-alcohols. Appl. Spectrosc.
72: 288–296.
12 Fletcher, A.N. and Heller, C.A. (1967). Self-association of alcohols in nonpolar
solvents. J. Phys. Chem. 71: 3742–3756.
13 Iwahashi, M., Hayashi, Y., Hachiya, N. et al. (1993). Self-association of octan-1-ol
in the pure liquid state and in decane solutions as observed by viscosity,
self-diffusion, nuclear magnetic resonance and near-infrared spectroscopy
measurements. J. Chem. Soc. Faraday Trans. 89: 707–712.
14 Czarnecki, M.A., Maeda, H., Ozaki, Y. et al. (1998). Resolution enhancement and
band assignments for the first overtone of OH stretching modes of butanols by
two-dimensional near-infrared correlation spectroscopy. 2. Thermal dynamics of
hydrogen bonding in n- and tert-butyl alcohol in the pure liquid states. J. Phys.
Chem. A 102: 9117–9123.
15 Tomza, P., Wrzeszcz, W., and Czarnecki, M.A. (2019). Tracking small het-
erogeneity in binary mixtures of aliphatic and aromatic hydrocarbons: NIR
spectroscopic, 2DCOS and MCR-ALS studies. J. Mol. Liq. 276: 947–953.
16 Asprion, N., Hasse, H., and Maurer, G. (2001). FT-IR spectroscopic investigations
of hydrogen bonding in alcohol–hydrocarbon solutions. Fluid Phase Equilibria
186: 1–25.
References 259
17 Orzechowski, K. and Czarnecki, M.A. (2019). Association of 1-hexanol in

mixtures with n-hexane: dielectric, near-infrared and DFT studies. J. Mol. Liq.
279: 540–547.
18 Mello, C., Mello, T., Sevéri, E. et al. (2009). Microstructures formation in a
seemingly ideal homogeneous mixture of ethanol and methanol: an experimental
evidence and two-dimensional correlation spectroscopy approach. J. Chem. Phys.
131: 084501.
19 D’Angelo, M., Onori, G., and Santucci, A. (1994). Self-association of monohydric
alcohols in water: compressibility and infrared absorption measurements.
J. Chem. Phys. 100: 3107–3113.
20 Dixit, S., Crain, J., Poon, W.C.K. et al. (2002). Molecular segregation observed in
a concentrated alcohol-water solution. Nature 416: 829–832.
21 Požar, M., Lovrinčević, B., Zoranić, L. et al. (2016). Micro-heterogeneity versus
clustering in binary mixtures of ethanol with water or alkanes. Phys. Chem.
Chem. Phys. 18: 23971–23979.
22 Wrzeszcz, W., Tomza, P., Kwaśniewicz, M. et al. (2016). Microheterogeneity in
binary mixtures of methanol with aliphatic alcohols: ATR-IR/NIR spectroscopic,
chemometrics and DFT studies. RSC Adv. 6: 37195–37202.
23 Wrzeszcz, W., Tomza, P., Kwaśniewicz, M. et al. (2016). Microheterogeneity in
binary mixtures of aliphatic alcohols and alkanes: ATR-IR/NIR spectroscopic and
chemometric studies. RSC Adv. 6: 94294–94300.
24 Chen, Y., Morisawa, Y., Futami, Y. et al. (2014). Combined IR/NIR and density
functional theory calculations analysis of the solvent effects on frequencies and
intensities of the fundamental and overtones of the C=O stretching Vibrations of
acetone and 2-hexanone. J. Phys. Chem. A 118: 2576–2583.
25 Beć, K.B., Karczmit, D., Kwaśniewicz, M. et al. (2019). Overtones of νC≡N vibra-
tion as a probe of structure of liquid CH3 CN, CD3 CN, and CCl3 CN: combined
infrared, near-infrared, and Raman spectroscopic studies with anharmonic
density functional theory calculations. J. Phys. Chem. A 123: 4431–4442.
26 Hu, Y., Zhang, J., Sato, H. et al. (2006). C-H center dot center dot O=C hydro-
gen bonding and isothermal crystallization kinetics of poly(3-hydroxybutyrate)
investigated by near-infrared spectroscopy. Macromolecules 39: 3841–3847.
27 Sato, H., Nakamura, M., Padermshoke, A. et al. (2004). Thermal behavior and
molecular interaction of poly(3-hydroxybutyrate-co-3-hydroxyhexanoate) studied
by wide-angle X-ray diffraction. Macromolecules 37: 3763–3769.
28 Sato, H., Murakami, R., Padermshoke, A. et al. (2004). Infrared spectroscopy
studies of CH⋅⋅⋅O hydrogen bondings and thermal behavior of biodegradable
poly(hydroxyalkanoate). Macromolecules 37: 7203–7213.
29 Gonjo, T., Futami, Y., Morisawa, Y. et al. (2011). Hydrogen bonding effects
on the wavenumbers and absorption intensities of the OH fundamental and
the first, second, and third overtones of phenol and 2,6-dihalogenated phenols
studied by visible/near-infrared/infrared spectroscopy. J. Phys. Chem. A 115:
9845–9853.
30 Schuler, M.J., Hofer, T.S., and Huck, C.W. (2017). Assessing the predictability
of anharmonic vibrational modes at the example of hydroxyl groups – ad hoc
construction of localised modes and the influence of structural solute–solvent

motifs. Phys. Chem. Chem. Phys. 19: 11990–12001.
31 Futami, Y., Ozaki, Y., Hamada, Y. et al. (2009). Frequencies and absorption
intensities of fundamentals and overtones of NH stretching vibrations of pyrrole
and pyrrole–pyridine complex studied by near-infrared/infrared spectroscopy and
density-functional-theory calculations. Chem. Phys. Lett. 482: 320–324.
32 Futami, Y., Ozaki, Y., Hamada, Y. et al. (2011). Solvent dependence of
absorption intensities and wavenumbers of the fundamental and first overtone of
NH stretching vibration of pyrrole studied by near-infrared/infrared spectroscopy
and DFT calculations. J. Phys. Chem. A 115: 1194–1198.
33 Futami, Y., Morisawa, Y., Ozaki, Y. et al. (2012). The dielectric constant
dependence of absorption intensities and wavenumbers of the fundamental
and overtone transitions of stretching vibration of the hydrogen fluoride studied
by quantum chemistry calculations. J. Mol. Struct. 1018: 102–106.
261
10
Direct Observation and Kinetic Mapping of Point-to-Point

Proton Transfer of a Hydroxy-Photoacid to Multiple
(Competing) Intramolecular Protonation Sites
Dina Pines 1 , Dan Eliovich 1 , Daniel Aminov 1 , Mark Sigalov 1 , Dan Huppert 2,† ,
and Ehud Pines 1
1
Faculty of Natural Sciences, Ben-Gurion University of the Negev, Department of Chemistry, P.O. Box 653,
Beer-Sheva 84105, Israel
2
Tel Aviv University, Raymond and Beverly Sackler Faculty of Exact Sciences, School of Chemistry,
Tel Aviv 69978, Israel
10.1 Introduction
Proton-transfer (PT) reactions are one of the most common reactions in nature, yet
the mechanism of these reactions in complex molecular systems, such as commonly
found in biochemical processes [1], or when coupled to their solvating environ-
ment [2–5] is still unsettled in many cases. Generally, the mechanism of PT depends
on the way the proton is solvated in the particular environment, the amount of avail-
able water molecules solvating the reaction center and assisting in the reaction, and
the properties and the solvation patterns of the proton-reactive acid and base pairs
that undergo a mutual PT reaction. Such Brønsted-type PT reactions may be by ran-
dom diffusion, or directed by some specific structural arrangement that causes the
proton to shuttle essentially point-to-point between two specific acidic and basic
sites. We have termed such PT reactions point-to-point PT reactions. PT reactions
that overall result in a single PT may be the outcome of several PT reactions, which
are coupled to each other either in a synchronized or in unsynchronized way. When
considered mechanistically, a multi-PT event may happen coherently, in a concerted
way or step-wise, when each PT is well separated in time from the other PT events.
To study synchronized or directed PT reactions, one needs to synchronize between
many PT reactions, which otherwise would occur at random times. The timing of
the reaction, start time, is usually achieved by transferring an otherwise nonreactive
system to a reactive system within a trigger time that is shorter than the PT reac-
tion time. To achieve timing in PT reactions, one usually uses photoacids that are
weak organic Brønsted acids [6] in their ground electronic state that upon electronic
excitation becomes strong (photo)acids. Photoacids may have a second (remote) pro-
tonation site secondary to the primary photoacidic site, which will react with the
† Deceased.

262 10 Direct Observation and Kinetic Mapping of Point-to-Point Proton Transfer
proton following the proton dissociation reaction from its primary photoacid site
[7–18].
Protonation of a secondary basic group of a photoacid, which is coupled to the
PT reactions of the primary site, may be monitored directly by following the time
evolution of the population of the proton-donor group or/and the population of the
proton-acceptor group. One can study such coupled PT reactions under conditions
when it is clear that the proton, which arrives at the proton acceptor group, has
originated from the proton-donating photoacidic group. Population dynamics
may be studied by time-resolved fluorescence (TRF) or by transient absorption
spectroscopy since PT affects both the fluorescence and the absorption spectra of the
molecule. Fluorescence spectroscopy is by far more sensitive than other detection
methods and allows observing mechanistic details in PT reactions not attainable
by other conventional detection methods. We have identified several types of
secondary protonation reactions following proton dissociation of a photoacid by
analyzing changes occurring in the dissociation profiles of photoacids undergoing
reversible proton dissociation–recombination reaction (reversible “geminate”
proton recombination, GEM reactions) at the primary photoacidic OH group. The
benchmark for the study of GEM reactions is the photodissociation profiles of
8-hydroxypyrene 1,3,6-trisulphonate (HPTS) photoacid, R*OH where “*” indicates
the photoacid being in its electronic excited-state.
The GEM of HPTS was first observed by Pines and Huppert [19, 20] and occurred
entirely on the excited-state (ES) potential surface without quenching [19–25].
The long-time asymptote of the R*OH decaying population monitored by TRF
approaches a t−3/2 dependence over time [23] and was well characterized both
experimentally and theoretically by Pines, Huppert, and Agmon et al. [23–25] (see
Figure 10.1a).
The benchmark GEM reaction of HPTS serves to illustrate the extreme sensi-
tivity of our kinetic analysis. HPTS has only one reactive side group, the primary
photoacidic OH group, and undergoes reversible proton dissociation from the OH,
where the proton back-recombines reversibly to the O− group. The proton disso-
ciation rate of HPTS to water is about 100 ps but the fluorescence emerging from
the photoacid R*OH undissociated protonated state is still detectable after about
10 ns due to the GEM reaction, which reversibly regenerates from the R*O− · · ·H+
ion-pair state, the proton-bound R*OH state. The back GEM is by bulk diffusion
and occurs when the molecular anion of HPTS, R*O− , and the proton still forms
an essentially solvent-separated isolated ion pair. A transition to the homogeneous
bimolecular reaction regime is observed under very acidic pH conditions or when
the concentration of the initially formed contact ion pairs is very high; so, the
solvent-separated ion pairs cannot be considered isolated from each other [24].
We have simulated the kinetics of the coupled reaction–diffusion reactive system
using the spherical symmetric diffusion program (SSDP) [21, 26]. In the SSDP rou-
tine, the proton is assumed to reversibly dissociate from the photoacid and then dif-
fuse away from the reaction center formed by its geminate R*O− anion in a spherical
symmetric space modeled by a radial one-dimension (1D) grid. From any grid point,
the proton may either move back toward the reaction center or move further away
100
–0.75
Normalized counts
HPTS HPTS
d(log(P))/d(log(t))
10–1
–1.00
10–2
–1.25
10–3
–1.50
10–4
0.1 1 10 20 0.1 1 10 100
(a) Time (ns) (b) Time (ns)
Figure 10.1 (a) Time-resolved fluorescence of HPTS in water measured at 420 nm at 20 ∘ C,

log–log plot of the normalized TCSPC data, I(t) for the excited-state proton-transfer
reaction R*OH ⇌ [R*O− ⋅⋅⋅H+ ]a ⇌ RO− + H+ , after excited-state lifetime correction (gray
dots) as compared with the numeric solution of the SSDP program after it was convoluted
with the instrument function (solid black line) [25]. Total time range of the TCSPC
measurement was 10 ps–50 ns, S/N > 5000 : 1 at the maximum channel. (b) The
logarithmic derivative curve of the numeric solution of the reaction-diffusion problem is
shown as an overlay on the experimental data (dots) in (a) extended to 100 ns with crossing
of the long-time −1.5 slope, which, on later times, is approached asymptotically from
below. The average slope of the log–log plot between 1 and 9 ns is −1.50, an observation
expanding on earlier reports that the long-time asymptote is apparently reached already
after several nanoseconds. The SSDP simulation predicts that the true asymptotic slop of
−1.5 is approached from below only after about 100 ns. The first oscillation is due to the
onset of the back proton geminate recombination reaction which temporary increases the
survival probability of the bound R*OH state while the second oscillation is due to the
establishment of a quasi-equilibrium between the proton flux arriving at the reaction
center from long-separation distances and the proton dissociation reaction which
continuously depletes the product R*OH state. Source: Based on Simkovitch et al. [25].
from it each time randomly changing its position by one grid point. The movement
of the proton is thus modeled by a random walker on an explicit spheric-symmetric
grid with the probabilities of moving forward or backward from a specific grid point
scaled by the distance-dependent Coulomb potential of the charged walker with
respect to the origin of the grid where a counter-ion is assumed to be positioned
stationary.
The numeric data fit of the SSDP routine [21, 26] requires several reaction
parameters: the reaction radius a, the Coulomb interaction RD and the mutual dif-
fusion coefficient between the proton and its counter ion D. The GEM process that
reversibly reforms the photoacid is modeled by two kinetic rate constants on the con-
tact radius of the reaction for proton association and proton dissociation ka1 and kd1 ,
respectively. The log of R*OH population over log of time d(log(P))/d(log(t)) [25]
for the excited-state GEM of HPTS is revealing (Figure 10.1b). It contains kinetic
details practically unattainable by inspecting the conventional log–log plots of
the decaying concentration profiles of photoacids undergoing reversible proton
dissociation.
Introducing a second basic side group on the aromatic system of the photoacid
competes for the proton with the primary OH photoacidic group. The problem of
multiple reversible binding sites does not have a known analytic solution even when
the Coulomb potential V(r) = 0.
A second reversible site means adding a second differential equation for the bind-
ing probability to the second site and inputting into the SSDP routine additional
kinetic terms for proton association and proton dissociation of the second site with
rate parameters ka2 and kd2 . The protonation reaction of a secondary protonation
site may be reversible or irreversible (ka2 = 0) on the timescale of the excited-state
lifetime of the photoacid depending on the basicity of the secondary protonation
site. Strong basic side groups, such as a carboxylate group, capture the proton for
tens of microseconds, which makes it practically an irreversible protonation reaction
within the excited-state ns lifetime of the photoacid. Slightly basic groups recombine
reversibly with the proton and retain it before dissociating on a similar time scale as
the reversible GEM reactions of the primary OH group. The two types of protona-
tion reactions to a remote (secondary) protonation site, reversible and irreversible,
are kinetically distinguishable from each other by the different effects they have on
the time profiles of the GEM reaction of the primary photoacidic group.
A direct illustration of the effect of remote protonation on the primary GEM
reaction is found in 2-naphthol-6-carboxy photoacid [11]. The decay profiles of 6-
carboxy-2-naphthol are shown in Figure 10.2 on a linear (a) and on a log-log scale (b).
In neat water, the carbonate group reacts irreversibly with the ejected proton
from the primary –OH side group. TRF decay curves of 6-carboxy-2-naphthol
exhibit almost an exponential decay due to this irreversible capture of the geminate
proton by the –CO2 − side group. In these PT reactions, the geminate proton
originating from the excited-state proton dissociation of the OH side group, having
1.0 1
HOOCR*OH pH2.89 @ 392 nm
0.8 𝜏d1 = 0.41 ns (91.7%)
Normalized counts
Normalized counts
–
OOCR*OH pH6.2 @ 365 nm
𝜏f = 1.19 ns (100%) 0.1 t–3/2
0.6
10 0
10–1
2N6C –2
10
0.4 10
–3
–4
10
0.01 10
–5
0.2 10
–6
0.0
1 2 3 4 5 6 1
Figure 10.2 (a) Dissociation profiles of 6-carboxy 2-naphthol showing the threefold
increase in the proton dissociation rate from the primary OH group of the photoacid upon
protonation of the CO2 − group demonstrate the effect of remote protonation on the
reactivity of the primary photoacid group. (b) log–log of the data shown in (a) revealing the
effect of the irreversible protonation of the secondary basic site of the photoacid, the
–COO− group. The solid line is SSDP simulation (shown in the inset for − OOCR*OH)
convoluted with the instrument function. The best biexponential fit curve (gray line) fails to
reproduce the experimental data for HOOCR*OH when the secondary protonation site of
the photoacid is blocked by ground-state protonation at pH = 2.8 which allows full
back-protonation of the primary O− site, which results in a nonexponential long-time
fluorescence tail, the trademark of a diffusion-assisted reversible GEM process. In contrast,
the − OOCR*OH fluorescence profile at pH = 6.2 when the second protonation site is not
occupied is almost monoexponential because the irreversible secondary protonation
reaction terminates the GEM reaction of the primary OH group.
pK a * = 1.2, is transferred irreversibly (on the timescale of the fluorescence lifetime

of the photoacid of about 5 ns) to the carboxylate side group, which becomes a
strong base, the electronic excited-state having pK a * = 8 [11]. The PT reaction is
analogous to a bimolecular PT reaction between an acid and a base in a system
where the acid and base molecules are at a fixed mutual separation distance and
the proton is transferred between them by bulk diffusion through the solvent that
separates them.
An additional aspect of PT from a primary side group to a secondary side group
of the photoacid is that the fast change in the protonation state of the primary func-
tional side group switches the acid–base reactivity of the secondary side group and
vise-versa in a similar to the effect of a ground-state protonation of a secondary side
group on the acidity of the primary OH group [11]. We have coined the protonation
of a secondary side group “remote protonation” and named the effect of transient
remote protonation on the acid–base reactivity of a primary functional group a reac-
tivity switch. The effect of remote protonation stems from the electronic coupling of
the two side groups through the π electronic system of their mutual aromatic back-
bone (Figure 10.3).
Sulfonate substituted naphthols behave in a different way than carboxy-
substituted naphthols. The sulfonate side group is much weaker base than the
carboxylate side group and reacts reversibly with the proton on the timescale of the
GEM reaction of the primary O− /OH side groups. In this case, the SO3 − side group
acts like a reversible proton collector that keeps the proton in the vicinity of the
primary O− group over longer times than otherwise determined by bulk diffusion.
The existence of a reversible secondary protonation site increases the overall
NH2 CH3
10 H
Br SO3–
pK 0a = 9.55 - 1.71 𝜎p CN NH3+
COO–
R = 0.990 COOH
8 COOCH3
pKa (ROH, R*OH)
6 NH2
4 CH3
H
pK*a = 2.74 - 4.37 𝜎p
R = 0.990 Br SO –
2 COO– 3
COOCH3
COOH NH3+
0 CN
–0.8 –0.6 –0.4 –0.2 0.0 0.2 0.4 0.6 0.8

𝜎p (benzoic acid)
Figure 10.3 pK a 0 (balls) and pK a * (squares) [27–35] values vs. Hammett’s 𝜎 p [36–38]
values for 2-naphthol substituents at the C-6 position [11]. The 𝜌 and R values of the
correlations are indicated in the figure. The pK a value of conjugate acid of HOOC–RO−
monoanion in the ground-state (star) was found by using the shown correlation. The 𝜎 p
values correlate with the effect of remote protonation of selected protonatable basic side
groups on the acidity of the primary OH group shown in Figure 10.2 are highlighted in
Table 10.1.
Table 10.1 List of 𝜎 p [36, 37] values of relevant

protonatable side groups of aromatic acids giving rise to
reactivity switch of the photoacid by a change in the
protonation state of a secondary (remote) reactive site.
Substituent 𝝈 p (benzoic acid)
H 0.0
O− −0.81
OH −0.37
CO2 − 0.02
COOH 0.41
SO3 − 0.37
probability of the GEM reaction of the primary photoacidic OH group. The basicity
of the SO3 − group in the electronically excited-state of SO3 − substituted 2-naphthol
depends on its position. For the 6 positions, we do not detect considerable secondary
protonation reaction of the sulfonate group, while at the 8-position, the sulfonate
group acts as a reversible proton scavenger [15]. At first, the scavenging reaction
takes the proton away from the solution and by doing so decreases the probability of
the GEM reaction of the proton with the primary O− side group and then increases
the GEM probability by releasing the proton back into the solution after a delay
time, which is equal to the retention time of the proton by the SO3 − side group.
This was first studied in 2N68dS, where the reactive side group out of the two side
groups was found to be the 8-position sulfonate while the sulfonate group at the
6-position only affect the overall reactivity (increasing it) of the primary OH group.
For the 2N68dS photoacid, the OH side group was shown to transfer the proton
reversibly to the sulfonate group at the 8-position by bulk diffusion or/and transfer
the proton point-to-point by a direct water bridge linking between the two side
groups in acetonitrile (ACN)–water solutions [15]. The relative importance of each
PT channel depends on the number of water molecules solvating the photoacid.
In bulk water, we only observe PT between the primary and the secondary side
groups by bulk diffusion (Figure 10.4a, dots) while in water-poor solvents, such as
can, with traces of water we only observe PT through one-water-molecule bridge
with no parallel PT reaction to the ACN solvent [15]. To simulate the GEM reaction
of 2N68dS with two competing sites for the proton, we use the SSDP routine with
double reactive boundary (DB) conditions (Figure 10.4).
According to the SSDP simulation routine, the two reversible proton reactive sites
are coupled to each other through a common diffusion space grid. Both sites are
assumed to be at the origin of the grid and every time the proton arrives there, it
fractionates between the two proton recombination sites and the diffusion space
according to the relative magnitude of the two competing proton recombination rate
constants and the bulk diffusion coefficient of the proton, which acts to move it back
into the solution.
100
10–1
Normalized intensity
Normalized counts
10–1
10–2
10–2
10–3 2N68dS 10–3 2N68dS

DB
DB
DB Irrevers
SB
10–4 10–4
0.1 1 10 0.1 1 10
–0.5
d(log(I))/d(log(t))
–1.0
2N68dS
–1.5
SB
DB
DB irrevers
–2.0
0.1 1 10
(c) Time (ns)
Figure 10.4 (a) Oscillations are evident in the fluorescence intensity of

2-naphthol-6,8-disulfonate (2N68dS, dots) plotted on a log–log scale after instrument
function convolution and lifetime correction (12.5 ns) [15]. The oscillations are due to the
reversible capture of the geminate proton by the SO3 − group at the 8-position of the
naphthalene ring. Also plotted are the numeric simulation of the reaction by the SSDP
program. DB – solid line: two competing reversible proton-transfer reactions. SB – dash
line: assuming only reversible geminate recombination reaction to reform the OH group
occurs without a competing secondary site of protonation. (b) Comparison between numeric
simulations by the SSDP for irreversible (short dash line) and the reversible DB conditions
(solid line) that were used to best fit the experimental data shown in (a). (c) The logarithmic
derivatives of the two curves in (b) and the SB curve shown in (a) shown on a logarithmic
timescale amplify the differences in the possible three reaction scenarios. The dash gray
line marks the asymptotic slope of −1.5 (Pines et al. [23–25]) which is first crossed at
intermediate times and then approached asymptotically from below at very long times. The
derivatives of the SB and DB curves for the two fully reversible protonation scenarios
exhibit three extreme points corresponding to the short, intermediate, and long-time
phases of the coupled dissociation reversible GR kinetic system. DB with irreversible second
protonation site exhibits only one extreme point. The line symbols are as plotted in (a)–(c).
Park and Agmon [39] have considered the problem of multiple competing
recombination sites and showed the long-time asymptotic behavior of the binding
probability to any of the molecular sites, and for any initial location of the diffus-
ing particle on the molecule, always decays as t−3/2 . It follows that a secondary
proton-binding site does not change the time asymptotic of the R*OH dissociation
profile P(t) of a reversible photoacid. However, we observe a drastic change in the
preasymptotic behavior of P(t) in 2N68dS. Park and Agmon [39] solved this problem
numerically using “double boundary” (DB) solution of the SSDP [39]. The SSDP
routine may be thus applied for both single-boundary (SB) and double-boundary
(DB) reactive conditions. The reaction radius a, the Coulomb interaction RD, and
the mutual diffusion coefficient D are assumed to be the same for the two types of
boundary conditions.
The kinetic analysis of 2N68dS using DB conditions for the GEM reaction in pure
water clearly demonstrates that the sulfonate groups act as a secondary reversible
protonation site for the primary protonation site of the O− group following proton
dissociation of the primary photoacid OH group.
The SO3 − group at the 8-position of the 2N68dS photoacid is likely to be the pro-
tonation site while the SO3 − group at the 6-position does not appear to do so [15].
To clarify the mechanism of the two competing protonation sites of 2N68dS, we
have investigated the 8-position monosulfonated 2-naphthol 2N8S photoacid where
only one secondary protonation site is possible at the 8-position. Figure 10.5 shows
the TRF of 2N8S in H2 O at pH 6 in neat water and 90%/10% (vol%) water/ACN. The
unbuffered 10−5 M solutions of 2N8S are measured at pH = 6.0 at room temperature
t = 20 ∘ C.
To simplify the complexity of the kinetic analysis of the TRF decay, we solved the
GEM problem with two competing reversible reactions occurring at the same con-
tact radius, a. Such an approximation is fully justified within the framework of the
spherical symmetric diffusion problem SSDP routine already at intermediate reac-
tion times of several hundreds of picosecond when the proton returns to the reaction
center from relatively long-separation distances.
Huppert et al. have reported that it was not possible to get a good fit of the R*OH
signal of 2N8S measured at 370 nm using the conventional SB GEM model [13]. To
achieve a good fit of the TRF decay signal of the photoacid, Huppert et al. assumed an
additional exponentially decaying component in the TRF signal of 2N8S, which was
attributed to a small concentration of an unidentified impurity or a photoproduct.
1.0
2N8S
0.8 90% H2O
Normalized counts
Neat H2O
0.6
0.4
0.2
0.0
2 4 6 8 10 12
Time (ns)
Figure 10.5 TCSPC-normalized fluorescence of the R* OH form of 2N8S in neat water at

360 nm with 10% ACN (49.9 M water) excited at 290 nm. Time-resolved fluorescence of
2N8S in water was measured at 370 nm for the R*OH form at 20 ∘ C. Data have shown
without lifetime correction so at long times the graph of the fluorescence decay is
dominated by population loss due to the finite fluorescence lifetime of R*O− in the
excited-state (𝜏f′ (R*O− ) = 13.5 ns).
Figure 10.6 Comparison 0

between the numeric solutions of
SSDP-SB and SSDP-DB in neat
water for 2N8S fluorescence 10–1 2N8S
Normalized counts
decay. DB – solid line and Neat water
SB – dash line. Reaction
10–2
parameters used for the numeric
simulations are summarized in
Table 10.2 and the considerable 10–3
difference between the two
simulations is evident to the eye.
10–4
t–3/2
10–5
0.1 1 10 100
Time (ns)
100
100 90% H2O
Neat H2O SSDP SB
SSDP SB SSDP DB
Normalized counts
Normalized counts
SSDP DB
10–1 10–1
10–2 10–2
10–3 10–3
1 10 1 10
Figure 10.7 Time-resolved fluorescence decay of 2N8S after lifetime correction (dots) in
neat water (a) and with 10% ACN (b). The experimental data is well-fitted by the SSDP
program with DB conditions after it was convoluted with the instrument function (full-line).
For comparison, the SSDP simulation using SB conditions is also shown and clearly does
not fit well with the measured fluorescence signal. The reaction parameters used in the
numeric simulations are summarized in Table 10.2.
Below we demonstrate an excellent fit of the decaying R*OH population of 2N8S by

utilizing the DB conditions [39] within the SSDP routine utilized in a similar way
to that used for simulating the TRF kinetics of 2N68dS, as shown in Figures 10.6
and 10.7.
The explicit pKa∗ values of the hydroxyl and sulfonate groups are directly calcu-
lated from the kinetic parameters of Table 10.2 because the numeric solutions are
carried out assuming fully reversible protonation reactions. To find the numeric val-
ues of the pKa∗ ’s of the OH and the SO3 H side groups we have used the following
relation:
( )
R
exp − aD kd
Ka∗ = 4πa2 ka 10−27 NA
(10.1)
kd and ka are the proton dissociation and proton recombination rate constants on
the contact radius a, which are listed in Table 10.2 for the two protonation equilibria,
Table 10.2 Best-fit parameters used in generating the simulated decay curves of 2N8S in
neat water (Figures 10.6 and 10.7a) and with 10% ACN shown in Figure 10.7b calculated
using the SSDP program under SB and DB conditions.
k d1 (ns−1 ) ka1 (Å ns−1 ) k d2 (ns−1 ) ka2 (Å ns−1 ) D(cm2 s−1 ) ̊
RD (A) ̊
a(A)
SB (10% ACN) 2.4 30.0 — — 7.5 × 10−5 14.0 5.1

−5
DB (10% ACN) 2.4 30.0 0.8 30.0 7.5 × 10 14.0 5.1
SB (neat H2 O) 3.0 30.0 — — 8 × 10−5 14.2 5.1
−5
DB (neat H2 O) 3.0 30.0 0.9 0.9 8 × 10 14.2 5.1
The meanings of the various kinetic parameters are as indicated in the text.
Table 10.3 The calculated pKa∗ values of the functional groups in

different concentrations of water.
pKa∗ of the pKa∗ of the

OH photoacidic SO3 H side
Water concentration side group group
49.9 M (90%) 1.6 2.1

Neat water 1.5 2.0
N A is the Avogadro number and 10−27 N A converts the units from molecules/Å3 to
molar. The pKa∗ is summarized in Table 10.3.
Using Table 10.3 it is evident that the pKa∗ of the sulfonate group at the 8-position
of 2N is slightly more basic than the pKa∗ of the photoacidic OH group and that the
value of the two pKa∗ ’s slightly increases (as expected for regular Brønsted acids)
when moving from pure water environment to the less polar aqueous environment
containing 10% by volume ACN.
10.2 From Intermolecular Proton Transfer to Solvent

to Intramolecular Point-to-Point Transfer in 1 : 1
Hydrogen-Bonding Complexes of Water with Bifunctional
OH Photoacids
The OH groups of aromatic phenols are excellent hydrogen-bond donors (HBD) as

indicated by their very high Kamlet–Taft alpha values [40], which are much larger
than the alpha values of aliphatic alcohols. OH photoacids are known to form in
their electronic ground-state stable hydrogen bonds with proton bases in aprotic
solvents, which do not compete with the high hydrogen-bond donicity of the OH
group. When in the excited electronic state, pre-existing ground-state hydrogen
bonds in the form ROH⋅⋅⋅B become much stronger in the form of R*OH⋅⋅⋅B bonds.
An example of 1 : 1 hydrogen-bond interaction of an OH photoacid, 8-hydroxypyrene
10.2 From Intermolecular Proton Transfer to Solvent to Intramolecular Point-to-Point Transfer 271
1.6 1.6
1.4 1.4
HPTA in DCM HPTA in DCM
Absorption (OD)
1.2 [DMSO] = 1.0 x 10–3 M 1.2 [DMSO] = 2.0 x 10–2 M
Absorption (OD)
1.0 1.0
87.3%
0.8 0.8 HPTA–DMSO
HPTA
0.6 73.5% 0.6
0.4 HPTA–DMSO 0.4
26.5%
0.2 0.2 12.7%
HPTA
0.0 0.0
340 360 380 400 420 440 460 340 360 380 400 420 440 460
(a) Wavelength (nm) (b) Wavelength (nm)
Figure 10.8 A demonstration of the strong hydrogen bond, which is formed in the
ground-state between a photoacid and a strong hydrogen-bond-acceptor DMSO. The
H-bond affects the absorption of the uncomplexed HPTA, which is clearly distinguishable
from the absorption of the HPTA–DMSO hydrogen bonding complexes in dichloromethane
(DCM) solvent. From the ratio between complexed and uncomplexed HPTA absorption as a
function of the DMSO concentration, the value of photoacid-DMSO equilibrium constant of
the complexation K compl = 358 was extracted. (a) In conditions when 26.5% of the photoacid
was complexed with |DMSO. (b) In conditions when 87.3% of the photoacid was complexed
with DMSO. Source: Based on Pines et al. [41].
1,3,6-trisdimethylsulfonamide (HPTA), and the considerable effect of the H-bonds

on the optical spectroscopy of the photoacid is shown in Figure 10.8 [41].
In the excited-state, the effect of the hydrogen-bonding interaction on the flu-
orescence spectra is dramatic both in terms of the change in the position of the
steady-state fluorescence spectra and the change in the time-resolved absorption
spectra (Figure 10.9) [41]. The time-resolved Stokes shift in the fluorescence emis-
sion clearly demonstrates the redshift in the steady-state fluorescence spectra, as
shown in Figure 10.9a, is due to a dynamic Stokes shift when the 1 : 1 ground-state
hydrogen-bond ROH–DMSO becomes much stronger in the exited state and adjusts
1.0
1.0
Normalized fluorescence
0.8 HPTA in DCM

HPTA–DMSO
0.6 0.8
HPTA
0.4
0.6
0.2
HPTA–DMSO
0.0 0.4
400 450 500 550 600 650 0 100 200 300 400 500
(a) Wavelength (nm) (b) Time (fs)
Figure 10.9 (a) Fluorescence of HPTA and HPTA–DMSO complex in DCM. (b) Transient
absorption data of HPTA and HPTA–DMSO in DCM after subtracting the long (3.5 ns)
excited-state decay component and renormalizing the data. The relaxation dynamics of
the O−H⋅⋅⋅B H-bond to its new equilibrium geometry following charge redistribution in
the electronic excited-state is 56 fs and are faster than the bulk solvation response of the
solvent to the charge redistribution in the chromophore.
ES
2N-methanol GS
300
(∆ν)OH(cm–1)
200 2N-water
2N-water
Ph-water
100
0
140 160 180 200
∆PA (kcal mol–1)
Figure 10.10 Correlation between the change in the IR absorption of the stretch vibration
of ROH Δ𝜈 OH and ΔPA = PA(photoacid) − PA(base) for gas phase H-bonded complexes of
2-naphthol (2N) 2N⋅⋅⋅B and of phenol Ph⋅⋅⋅B relative to 𝜈 OH of the uncomplexed photoacid.
Large additional redshifts are observed in 𝜈 OH of the complexed acids upon electronic
excitation from the electronic ground-state GS to the first electronic excited-state ES
rationalized by the 14 kcal mol−1 reduction in the proton affinity of the 2-naphtholate anion
when electronically excited The correlation line is only meant for guiding the eye pointing
to the nonlinearity of the correlation. Source: Based on Keinan et al. [42].
its geometry to reflect the stronger H-bonding interaction on an ultrafast timescale

of several tens of femtosecond, still resolvable experimentally (Figure 10.9b).
Similar observations were made in the gas phase for several OH photoacids.
Unlike in the solution phase, in the gas phase, one directly probes the 1 : 1
H-bonding complex of OH photoacids, such as formed between 2-naphthol and
a single base molecule. In particular, it was demonstrated that a single-water
molecule acts like a regular weak Brønsted base shifting the IR absorption of the
1 : 1 water-complexed photoacid, ROH· · · OH2 to the red with the redshift in the
excited electronic state of the photoacid being considerably larger than the redshift
observed for the ground-state complex (Figure 10.10) [42]. The spectral redshift in
the O–H stretch vibration for both the ground and the excited-state of the photoacid,
when complexed with a single-water molecule, correlates with the effect of other
1 : 1 H-bonding complex of the photoacid using the gas-phase basicity (PA) of the
complexing base molecule as a single (nonlinear) correlating factor.
The above experimental observations strongly suggest that water can form 1 : 1
hydrogen bonds with OH photoacids in aprotic solvents and that the site of com-
plexation is the OH group of the photoacid, which donates hydrogen bond to the
oxygen atom of a water molecule. A most intriguing case arises when a single-water
molecule forms a hydrogen-bonding bridge between an acid and a base molecule,
thus enabling PT along with the water bridge. It was found that bimolecular PT
between an acid and a base through one-water bridge may be extremely efficient
depending on the reaction exothermicity [43, 44]. A similar situation occurs when
a water bridge is formed between an acidic and a basic side group of a photoacid.
In particular, as we demonstrate below, the proton may be exclusively transferred
10.3 Water Is Able to Donate and Accept an H-bond as Demonstrated by IR Absorption 273
from the primary OH side group via a water bridge to a secondary basic SO3 −
side group when it is present within an appropriate distance from the OH group.
Generally, short-separation distances allow the formation of a stable water bridge
made of one or two water molecules spanning the distance between two reactive
sites on the aromatic skeleton of the photoacid. This is achieved by dissolving
a bifunctional OH photoacid in ACN with very small concentrations of water,
which form a water bridge between the OH and the secondary functional group
of the photoacid. In such cases, water acts in amphoteric way: Acting on one side
of the bridge as a base accepting a hydrogen bond from the OH group and the
other side of the bridge as an acid when donating a hydrogen bond to the SO3 −
group to form the ROH⋅⋅⋅OH2 ⋅⋅⋅−O3 S bridge. Similar to 1 : 1 hydrogen-bonding
complexes of OH photoacids, the hydrogen-bonding interaction, which forms the
water bridge, becomes stronger in the electronic excited-state of the photoacid and
supports full ESPT from one side to the other side of the water bridge, resulting in a
proton-transferred hydrogen-bonding complex of the type R*O⋅⋅⋅H2 O⋅⋅⋅HO3 S.
10.3 Water Is Able to Donate and Accept an H-bond

as Demonstrated by IR Absorption in 1 : 1
Water–(Acid or Base) Complexes
Water can either donate or accept hydrogen bonds, as shown in Figure 10.11.
When water is hydrogen-bonded to the weak base DMSO, acting as an HBD to a
base, the water asymmetric O–H stretch transition is slightly redshifted by about
20 cm−1 , while the intensity of the symmetric O–H stretch greatly decreases and a
new much broader absorption redshifted from the symmetric stretch absorption
by about 200 cm−1 appears and marks the absorption of OH oscillators which are
H-bonded to DMSO. In contrast, when water acts as an H-bond acceptor (HBA)
accepting an H-bond from a weak acid chlorophenol (Figure 10.11b), the position
of the asymmetric and symmetric O–H stretches are practically unaffected by the
1.0
1.0
Asymmetric OH stretch of H2O
Asymmetric OH stretch of H2O
0.8 Symmetric OH stretch H2O
Normalized absorbtion
0.8
Normalized absorbtion
H-OH stretch of H2O-DMSO complex

Symmetric OH stretch of H2O
0.6 H3C
0.6
S O H O 4-Cl-phenol OH stretch
H3C H
H-donor
0.4 0.4 H
O H O
H
0.2 0.2 H-acceptor
0.0 0.0
3300 3400 3500 3600 3700 3200 3300 3400 3500 3600 3700 3800
Figure 10.11 (a) IR spectra of H2 O + DMSO complexes in DCM: 0.1 M H2 O in DCM and
0.3 M DMSO + 0.1 M H2 O in DCM. (b) IR spectra of chlorophenol–H2 O complexes in DCM:
0.05 M H2 O in DCM (solid-line) and 0.01 M 4-chlorophenol + 0.05 M H2 O in DCM
(dashed-line).
HBA interaction. In contrast to the water HBD case, the intensity of the asymmetric
vibration, which involves the oxygen atom, is more affected than the intensity of
the symmetric stretch and is reduced by about 20%. The symmetric water O–H
stretch absorption is narrower than the phenolic O–H stretch transition and is little
blueshifted from it, which helps to distinguish between the two absorptions. The
absorption of the H-bonded phenolic OH oscillators is redshifted by about 200 cm−1
from the position of the uncomplexed oscillators and is much broader than the
uncomplexed ones.
10.4 Proton Transfer Along with Water Bridges

in Acetonitrile (ACN) Spanning the Distance Between an
Acidic and a Basic Side Groups of Bifunctional Photoacids
The advantages of using ACN solutions for studying protonated water solvates in
the liquid state were demonstrated by Kolthoff and Chantooni [45, 46], which are
described below: First, ACN is an aprotic nonassociating solvent that can be mixed
in any ratio with water to form highly polar solution. This allows rich acid–base
chemistry in ACN/water mixtures. Second, liquid ACN is a very weak Brønsted base,
with basicity considerably lower than the basicity of H2 O clusters in ACN promis-
ing that ACN does not directly protonate to form CH3 CNH+ in presence of water.
Third, mineral acids, such as trifluoromethanesulfonic acid (triflic acid CF3 SO3 H|),
are strong acids in ACN and are excellent protonation agents in ACN/water mix-
tures [47]. Finally, ACN cannot participate in the building of the solvent bridges, as
depicted in Figure 10.12, which require amphoteric molecules, such as water and
alcohols.
Figure 10.13 shows the absorption of 2N8S in various solvents. In pure water,
the absorption spectra are wider and less structured than in pure ACN and are
also blueshifted by about 1 nm. Upon addition of 0.28 M of water to the ACN
solution, the absorption is blueshifted by about 0.5 nm and is slightly wider and
less structured. In presence of the very strong acid CF3 SO3 H, the absorption
spectrum becomes considerably wider and less structured and further blueshifts by
3–335.7 nm. These minor changes observed in the absorption spectra could have
been easily overlooked if not being magnified in the fluorescence spectra of the
photoacid, as shown in Figure 10.14. The R*OH and R*O- fluorescence bands in
pure water are at 375 and 460 nm, respectively, indicating a rapid proton-transfer
process to the solvent, which competes with the fluorescence decay lifetime of the
R*OH photoacid. In pure ACN, only the R*OH forms fluoresce at 360 nm, indicat-
ing the absence of an ESPT reaction within the lifetime of R*OH. The emission
spectrum of the photoacid in ACN with 0.028 M of water shows a new emission
band at 520 nm much redshifted from both the R*OH and the R*O− fluorescence
bands. We observe three regions for the emission spectra of the photoacid as a
function of water concentration of which only two of them were well-characterized
before (Huppert et al. [13]). In the low-water-concentration region, not resolved
before (Figure 10.14a), the intensity of the 520 nm band increases almost linearly
10.4 Proton Transfer Along with Water Bridges in Acetonitrile (ACN) Spanning the Distance 275
W W1
W2
Figure 10.12 A pictorial sketch of the one- (a) and two- (b) water-bridge arrangements of
2-naphthol 8-sulfonate (2N8S) in acetonitrile. The water-bridge oxygen atom accepts a
hydrogen bond from the OH group of the photoacid and one of the water-bridge molecules
donates a hydrogen bond to one of the oxygens of the SO3 − group at the 8-position and
thus forms hydrogen-bonding “bridge” along which the proton may be transferred in the
otherwise aprotic solvent environment. For the one-water bridge case W, a single-water
molecule acts both as proton donor and as proton acceptor when forming a bridge between
the acidic and the basic side groups of the photoacid. Similar model for the photoacid–
water complex in ACN was suggested by Huppert et al. [13]. Source: Based on Gajst
et al. [13].
1.0
2N8S
0.8
Normalized absorption
0.6
0.4 ACN
ACN + 0.28 M H2O
H2O
0.2 ACN + 0.28 M H2O + 10–3 Triflic acid
0.0
300 310 320 330 340 350 360
Wavelength (nm)
Figure 10.13 UV absorption of 2N8S in pure water (gray full line), ACN (dashed line), in
ACN + 0.28 M H2 O (full line), and in ACN with 0.28 M water and 0.001 M triflic acid CF3SO3 H
(dashes-dotted line). The absorption maxima are at 332.5, 333.5, 334.0, and 335.7 nm
respectively.
with the water concentration and the fluorescence band of R*OH concomitantly
decreases resulting in an iso-emissive point at about 430 nm. In the intermediate
water concentration region between about 0.5 and 5.5 M of water, the band at
520 nm gradually decreases until it disappears while the intensity of the R*OH
band gradually increases, and blueshifts toward its value in pure water with no
apparent iso-emissive point (Figure 10.14b). In the high-water-concentration region
up to the pure water limit (Figure 10.14c), the R*OH band gradually decreases
600
–
– 600 O3S-R*OH
O3S-R*OH
ACN
Fluorescence intensity (a.u)

500 370 nm 0.55 M H2O
360 nm 0.028 M H2O 500
0.055 M H2O 1.11 M H2O
400 2.77 M H2O
0.11 M H2O 400
0.28 M H2O 365 nm 5.54 M H2O
300
300
200 200
517 nm HO3S-R*O–
100 HO3S-R*O– 100
470 nm
0 0
350 400 450 500 550 600 350 400 450 500 550 600
–
800 O3S-R*O–
458 nm
ACN
– 0.28 M
600 O3S-R*OH
39.1 M
44.3 M
400 Neat water
200 375 nm
H+---–O3S-R*O–
520 nm
0
350 400 450 500 550 600
(c) Wavelength (nm)
Figure 10.14 (a) Fluorescence emission spectra of 2N8S in pure ACN (359.5 nm) and the
low-water-concentration region from 0.028 to 0.28 M H2 O (up to 0.5% H2 O by volume,
361.5 nm), under conditions of one-water bridge in ACN solvent as discussed in the text. The
intensity of the 520 nm emission band goes up practically linearly with water concentration
(b). Fluorescence emission spectra of 2N8S in the intermediate water concentration region
are shown for 0.55, 1.11, 2.77, and 5.54 M H2 O in ACN solvent. The intensity of the
520-emission band goes down with the rise in the water content. (c) Fluorescence emission
spectra of 2N8S in the high-water-concentration region 44.3 M in ACN and neat H2 O. The
fluorescence spectra with 0 and 0.28 M of water in ACN are shown for comparison.
and concomitantly a new band, the R*O− band at 458 nm, appears establishing
an iso-emissive point at 412 nm typical of an excited-state proton-transfer (ESPT)
reaction from R*OH to the water solvent R*OH → R*O− + H+ .
Figure 10.15 shows the effect of protonating the sulfonate side group with triflic
acid already in the ground-state and by doing so blocking the ESPT process from
the OH group to the sulfonate side group. When protonation of the sulfonate group
occurs without the presence of water there is large 25 nm redshift in the fluorescence
of the photoacid. In presence of 0.28 M water with no protonating mineral acid,
the R*OH fluorescence redshifts by 2.5–362 nm, and a new emission band from the
proton-transferred complex is detected at 520 nm. When a protonating acid is added
to the ACN/0.28 M water solution, the 520 nm band of the proton-transferred com-
plex disappears and the fluorescence from the protonated H3 OS-R*OH photoacid
appears at practically an identical position and shape as of the protonated photoacid
fluorescence in pure ACN with no added water. This proves that the sulfonate group
is coupled to the aromatic electronic structure of the photoacid and that the new
520 nm fluorescence band is indeed associated with intramolecular ESPT process
10.5 Time-Resolved Fluorescence Measurements of Proton Transfer along with Water Bridges 277
600
ACN
500 ACN + 0.001 M triflic acid
ACN + 0.28 M H2O
Fluorescence intensity
400 ACN + 0.28 M H2O + 0.001 M triflic acid
300
200
100
0
350 400 450 500 550 600
Wavelength (nm)
Figure 10.15 Fluorescence spectra of 2N8S in ACN (full line), in ACN with 0.001 M
CF3 SO3 H (dashed line), in ACN with 0.28 M of water (dots), and with 0.28 M water with
0.001 M CF3 SO3 H in ACN (dashed-dotted line).
between the OH group and the SO3 − group. The reaction is facilitated by a water
bridge, which is formed at low-water concentrations as a weak H-bonding complex
in the ground-state of the photoacid. The concentration of water is low enough to
be considered a solute in ACN and not a co-solvent. The weak H-bonding complex
becomes tighter complex in the electronic excited-state of the photoacid and sup-
ports full PT from the OH group to the sulfonate group a reaction that is likely to
involve rearrangement of the H-bonding complex prior to PT.
TRF measurements detailed below provide strong additional evidence for the
intramolecular proton-transfer process occurring at the low-water-concentration
region within an already existing ground-state one-water bridge.
10.5 Time-Resolved Fluorescence Measurements

of Proton Transfer along with Water Bridges
TRF measurements of 2N8S in ACN/H2 O solutions are shown in Figure 10.16.
The ground-state H-bonding complex of the photoacid complexed with water
was excited at 290 nm away from the emission band of the photoacid using 1 ps
laser pulses. The fluorescence emission of the H-bonding complex prior to PT was
collected at 360 nm at the peak emission of the fluorescence and the emission from
the proton-transferred complex was collected at 520 nm (Figures 10.13–10.15) using
TCSPC (time-correlated single-photon-counting) system.
The decay of the photoacid–water complex fluorescence is governed by the fluo-
rescence decay lifetime of the photoacid in absence of PT. With increasing popula-
tion of the photoacid–water complex increasing fractions of the photoacid undergoes
through-water-bridge ESPT reaction to the sulfonate side group, which shortens
the excited-state lifetime of the photoacid. The risetime of the proton-transferred
100 1.0
0.8 0.028 M H2O

Normalized counts
Normalized counts
0.75 ACN 0.055 M H2O
0.028 M H2O 0.11 M H2O
0.055 M H2O 0.6
0.50 0.11 M H2O
0.4
0.25
0.2
0.00 0.0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 35 40
Figure 10.16 (a) TCSPC-normalized fluorescence of the R* OH form of 2N8S in neat ACN
(black dots) (5 ns lifetime) and in ACN/H2 O mixtures under conditions of one-water bridge
with 0.028, 0.055, and 0.11 M water, collected at 360 nm. (b) TCSPC-normalized fluorescence
of the proton-transferred form of 2N8S in ACN/H2 O mixtures collected at 520 nm. The
lifetime of the proton-transferred form of 2N8S in the one-water molecule-bridge complex
is 8.5 ns considerably longer than the lifetime of the uncomplexed photoacid.
H-bonding complex of the photoacid does not depend on the fraction of the pho-
toacid, which undergoes ESPT and so is also the fluorescence decay lifetime of the
proton-transferred complex. We, thus, conclude that the dynamics of the generation
and decay of the proton-transferred complex do not depend on the water concentra-
tion in the low-water-concentration region. In Figure 10.16b, we superimpose the
TRF spectra collected at 520 nm for three different water concentrations and it is
clear that the dynamics exactly match each other although the water concentration
in the three photoacid ACN solutions differs by as much as a factor of 4. This obser-
vation also negates the possibility of the water bridge forming in the excited-state of
the photoacid following a bimolecular reaction with water, which, in this case, could
have become the rate-determining step for the ESPT reaction.
We also observe up to about 5% of immediate risetime monitored at 520 nm due
to an overlap of the R*OH fluorescence band. This overlap is noticeable in the
steady-state emission spectra of the system (Figure 10.14b). To avoid the overlap,
we have measured the TRF fluorescence band of the proton-transferred complex at
560 nm away from the fluorescence peak, which reduces the overlap to less than
1%, as shown in Figure 10.17.
We have carried out in the low-water region a full kinetic analysis of the
time-resolved spectra assuming irreversible two-population dynamics between the
nonproton-transferred and the proton-transferred photoacid–water complexes.
The analysis of the kinetic of the TRF signal of the R*OH form in the
low-water-concentration region 𝜒 water = 0.025–0.3% (vol%) is shown in Figure 10.17
fits very well the experimental data assuming irreversible ESPT kinetics of a fraction
of the total concentration of 2N8S linearly increasing with water concentration.
The data in Figures 10.16a and 10.17 exhibit a nearly biexponential decay. The
shorter decaying component is of variable amplitude. The amplitude increases with
water concentration with a concentration invariant decay time of about 4.0 ns. The
amplitude of the longer decaying component of about 5.0 ns is also independent of
water concentration. We assigned these two components to the water-complexed
and the nonwater-complexed fractions of the photoacid and carry out the kinetic
1.0 1.0
2N8S 2N8S
0.11 M H2O
Normalized counts
Normalized counts
0.8 0.014 M H2O 0.8
0.6 0.6
517 nm 560 nm
0.4 0.4
0.2 360 nm 0.2 360 nm
0.0 0.0
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35
Figure 10.17 TCSPC-normalized fluorescence of the R* OH and proton-transferred forms of

2N8S in ACN solvent (a) with 0.014 M water, taken at 517 nm (dots) and showing an
immediate risetime due to a residual spectral overlap with the photoacid emission band at
360 nm (squares) and (b) for 0.11 M H2 O taken at 560 nm (dots) with no apparent immediate
risetime, the photoacid emission band at 360 nm (squares) is also shown. Solid lines are fits
of the experimental data using the kinetic model of irreversible proton transfer from the OH
side group to the SO3 − side group along a preformed single-water molecule
hydrogen-bonded bridge.
analysis accordingly. The fraction of the population with the longer decay time
is assigned to the non-complexed photoacid population, and the second popula-
tion fraction is assigned to the water-complexed photoacid. It means that in the
low-water-concentration region, only a fraction (water-concentration dependent)
of the 2N8S molecules undergoes irreversible intramolecular ESPT process. In
such two independent (on the timescale of the excited-state lifetime) population
dynamics that overlap spectrally, one population being reactive and the other
not, the time-dependence behavior of the reactive (water-complexed) population
directly reflects the ESPT reaction that it undergoes. We find by this analysis that
the ESPT bridge process occurs with a relatively long 20 ns lifetime. Monitoring the
rise of the proton-transferred complex gives additional valuable information, which
confirms this assignment: the 4.0 ns risetime of the proton-transferred complex
measured at the 520–560 nm is found to be identical within our experimental
uncertainty (about 0.05 ns) with the decay of the 360 nm band emission of the
water-complexed 2N8S. More specifically, we find that the population fraction
of the water-complexed 2N8S emitting at 360 nm, which disappears by both
proton-transfer and by excited-state decay, matches the kinetic behavior revealed
by analyzing the risetime of the proton-transferred emission band at 520–560 nm.
The lifetime of the proton-transferred complex was found to be 8.5 ns for 2N8S and
9.5 ns for 2N68dS. Table 10.4 summarizes our kinetic analysis of 2N8S and 2N68dS.
Next, we use the population amplitudes as extracted from the kinetic analysis to
determine the ground-state complexation equilibrium constant K between 2N8S and
water,
ROH + H2 O ⇌ ROH · · · H2 O, K = [ROH · · · H2 O]∕[ROH][H2 O] (10.2)
ROH · · · H2 O ⇌ complex, [H2 O]0 − [complex] = [H2 O]free ,

[ROH]0 − [complex] = [ROH]free K = [complex]∕([ROH]free [water]free )
(10.3)
Table 10.4 Time constants of R*OH, R*OH–water complex and the ESPT reaction for 2N8S
and 2N68dS.
𝝉 (ns) 2N8S 𝝉 (ns) 2N68dS
−
O3 S R*OH 5.0 5.5
R*O− · · ·(H2 O)· · ·HO3 S (proton-transferred complex) 8.5 9.5
HO3 S R*OH (two protonated side groups) 5.5 6.0
Proton transfer between OH and SO3 − (water complex) 20 20
Table 10.5 The ratio between [ROH]free and ROH–water complex

in ACN as found in the kinetic analysis of the time resolved
spectra (TRS) data and the water concentration, which was
introduced into the ACN-2N8S solutions.
[H2 O]0 [M] [ROH]free Complex Complex/[ROH]free
0.014 0.8 0.2 0.25

0.028 0.7 0.3 0.428 57
0.055 0.55 0.45 0.818 18
0.11 0.33 0.67 2.030 3
0.165 0.25 0.75 3
The concentration of photoacid is 10−5 M and the concentration of water is three to

four orders of magnitude greater than that, which allows to do the excellent approx-
imation [H2 O]free = [H2 O]0 and to write:
K[water]0 = [complex]∕[ROH]free (10.4)
Plotting [complex]/[ROH]free as a function of [water]0 (Table 10.5) gives a straight
line with a slope equal to K, the ground-state complexation equilibrium constant
between 2N8S and water.
We find by the graphic analysis, as shown in Figure 10.18, that the water–photoacid
complexation constant K is K = 18.4 M−1 . The small complexation constant is in
accord with the relatively low basicity of a single-water molecule in ACN and the
geometric constraints for the creation of the H-bonding complex, which only allow
for the creation of a weak H-bonding complex between 2N8S and water, see below.
Further increase in the water concentration in the intermediate water concen-
tration region causes a gradual change in the ESPT process. The single-water
bridge stops to be the exclusive water bridge structure and the ESPT reaction is
gradually transformed from being intramolecular through a single-water bridge to
being intramolecular through a two or more water molecules that cluster around
the polar side groups of the photoacid and then to being a purely intermolecular
ESPT reaction to bulk water. As a result, the through bridge ESPT rate becomes
water concentration-dependent, as shown in Figure 10.19, where the most evident
3
[complex]/[ROHfree]
K = 18.4 M–1
R2 = 0.994
2
0
0.00 0.05 0.10 0.15 0.20
[H2O] (M)
Figure 10.18 Plot of the data of Table 10.5. The slope of the linear correlation equals K,
the single-water-photoacid complexation constant.
1.0 1.0
2N8S
0.8 ACN 0.8
Normalized counts
2N8S
Normalized counts
0.28 M H2O 0.28 M H2O

0.6 0.56 M H2O 0.6 0.56 M H2O
1.1 M H2O 1.11 M H2O
0.4 0.4
0.2 0.2
0.0 0.0
0 2 4 6 8 10 12 14 16 18 20 0 5 10 15 20 25 30 35
Figure 10.19 (a) TCSPC-normalized fluorescence of the R* OH form of 2N8S in neat ACN
(5 ns lifetime) and ACN/H2 O mixtures with 0.028, 0.56, 1.1 M of water at 360 nm. (b) TCSPC
normalized fluorescence of the proton-transferred form of 2N8S in ACN/H2 O mixtures
collected at 520 nm.
observation is the increase in the ESPT rate as a function of water concentration.

Good fits of the TRF data in the intermediate water concentration region are
only possible when reversibility in the ESPT is introduced to the kinetic analysis
because, in this range, the kinetics of the ESPT reaction turns to be biexponential
instead of monoexponential as observed in the lower water concentration region
(Figure 10.20).
The TRF of the R*OH form of 2N8S in the intermediate H2 O concentration
region shown for the range 0.28–2.8 M water exhibits a nearly biexponential decay
of about 2 ns and a second slower decaying component of about 5–7 ns. We attribute
the longer time component in the TRF of 2N8S to ESPT from the SO3 H group back
to the deprotonated O− group, the two side groups having almost an identical pKa∗
values in water-rich solutions, as listed in Table 10.2. We thus suggest that in the
intermediate water concentration region, the proton is able to shuttle reversibly
1.0 1.0
2N8S
Normalized counts
Normalized counts
0.8 0.8 0.28 M H2O
2N8S
0.6 1.1 M H2O 0.6
0.4 0.4
0.2 0.2
0.0 0.0
0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40
Figure 10.20 TCSPC-normalized fluorescence of the R*OH (squares) and proton-

transferred forms of 2N8S in ACN solvent (dots) with 0.28 M H2 O (a) and 1.1 M H2 O (b). Solid
lines are data fits using a reversible intramolecular kinetic model for the ESPT.
between the 2-OH and the 8-sulfonate side groups through water bridges of various
sizes. This also explains why in the intermediate water concentration region, the
long decay component observed in the decay of the R*OH state and the decay of
the proton-transferred state is practically identical to within experimental error.
Solving the kinetic system and fitting the TRF profiles for both the R*OH state and
the proton-transferred state, we find the percentage of the protons that take part
in a reversible intramolecular ESPT process. The short component of the reversible
reaction kinetics marks the relaxation of the kinetic system to its equilibrium
concentrations and the long component is roughly the average lifetime of the two
coupled reactive states, as shown in Table 10.6.
In summarizing, the analysis of the TRF measurements unequivocally shows that
the 2N8S does not dissociate in pure ACN and the low-water-concentration region
undergoes irreversible state-to-state ESPT population dynamics. The dynamics fit
a slow proton-transfer reaction along with a pre-formed one-water bridge from
the OH side group to the SO3 − side group in 2N8S occurring with a time constant
of 20 ns. At higher water concentrations, water–water interactions compete with
the water–bridge interactions and the structurally well-defined one-water and
two-water bridge structures become less favorable.
Table 10.6 Time constants of R*OH, R*OH–water complex and the ESPT reactions for
2N8S and 2N68dS and the fraction of the water-complexed population of the photoacid in
%.
[H2 O]
𝝉 (ns), % 0.28 M 2N8S 0.56 M 2N8S 1.1 M 2N8S 0.56 M 2N68dS
R*OH 5.0 ns 5.0 ns 5.0 ns 5.5 ns

–
R*O–H (H2 O)x 8.5 ns 8.5 ns 8.5 ns 9.5 ns
Forward proton transfer 6 ns 3.5 ns 2 ns 3.0 ns
Backward proton transfer 12 ns 4.4 ns 1.2 ns 5.5 ns
% complex 90% 80% 50% 80%
10.6 Isotope D/H Effect 283
10.6 Isotope D/H Effect

The isotope effect of the various ESPT processes of 2N8S was studied by replacing
H2 O with D2 O in ACN solutions of the photoacid. The steady-state absorption and
emission spectra with various concentrations of D2 O (up to 0.55 vol%) exhibit similar
spectral features to that of 2N8S in ACN/H2 O, as depicted in Figure 10.21.
The TRF data of the deuterated solutions of 2N8S in the low-D2 O-concentration
region was fitted assuming an irreversible ESPT process of D+ from the OD side
group to the SO3 − side group (ESDT) along with a one-D2 O bridge similar to
the kinetic analysis done for H+ transfer in the low-H2 O-concentration region
(Figure 10.22).
While the spectral features of 2N8S in water and D2 O are similar, we find consider-
able kinetic isotope effects (KIE) for the ESDT rate constants compared to the ESPT
rate constants (Tables 10.7–10.9).
The KIE, we find for the ESPT reaction, is 30 ns/20 ns = 1.5 formally within the
range of the KIE found for many ESPT reactions, for example, the KIE reported
1.0 1.0
2N8S
0.8 ACN 0.8 2N8S
Normalized counts
Normalized counts
0.28 M H2O 0.28 M H2O

0.6 0.56 M H2O 0.6 0.56 M H2O
1.1 M H2O 1.11 M H2O
0.4 0.4
0.2 0.2
0.0 0.0
0 2 4 6 8 10 12 14 16 18 20 0 5 10 15 20 25 30 35
Figure 10.21 (a) UV absorption spectra of 2N8S with up to 0.28M D2 O (up to 0.5% D2 O by
volume) do not show absorption changes under conditions of one D2 O bridge in the ACN
solvent. (b) Fluorescence emission spectra of 2N8S with various concentrations of D2 O up
to 0.28 M of D2 O.
Figure 10.22
TCSPC-normalized 1.0 2N8S
0.028 M D2O
fluorescence of the R*OD and @ 360 nm
deuteron-transferred forms of 0.8 @ 520 nm
Normalized counts
2N8S in ACN solvent with @ 570 nm

0.028 M D2 O, collected at
360 nm (− O3 S R*OD) and 520 0.6
and 570 nm R*O− · · ·(D2 O)· · ·
DO3 S (deuteron transferred 0.4
complex).
0.2
0.0
0 5 10 15 20 25 30 35 40
Time (ns)
Table 10.7 Excited-state lifetimes parameters for deuterated 2N8S and its intramolecular
deuteron transfer.
𝝉 (ns)
−
O3 S R*OD 5
R*O− · · ·(D2 O)· · ·DO3 S (deuteron-transferred complex) 9
−
Deuteron transfer between OD and SO3 (water complex) 30
Table 10.8 Kinetic parameters used in generating the simulated decay curves of 2N8S in
D2 O are shown in Figure 10.23 using the SSDP program under SB and DB conditions.
k d1 (ns−1 ) ka1 (Å ns−1 ) k d2 (ns−1 ) ka2 (Å ns−1 ) D(cm2 s−1 ) ̊
RD (A) ̊
a(A)
SB 1.15 19.5 — — 5.0 × 10−5 14.0 5.0

DB 1.15 19.5 0.45 19.5 5.0 × 10−5 14.0 5.0
The meaning of the various kinetic parameters is as indicated in the text.
Table 10.9 Calculated isotope effects for 2N8S in H2 O/D2 O for processes involving H+ /D+
transfer with data obtained for both isotopes by analyzing the TRF profiles of 2N8S using
DB conditions.
H2 O D2 O KIE
kd1 (ns−1 ) 3.00 1.15 2.61

−1
kd2 (ns ) 0.90 0.45 2.00
( ̊ )
A
ka𝟏 ns 30.00 19.50 1.54
( ̊ )
A
ka𝟐 ns 30.00 19.50 1.54
pK a * of OH/D side group 1.5 1.7
*
pK a of SO3 H/D side group 2.0 2.1
The meaning of the various kinetic parameters is as indicated in the text.
by Pines et al. [48] for various photoacids and H. Yu for 3-hydroxyquinoline pho-
toacid [49]. However, as we detail below, the relatively slow ESPT rate and the fact
that it is likely to be the overall KIE for a double-PT reaction and not for an elemen-
tary one-PT reaction make such a comparison questionable. To get better insight
into the mechanism of the double-PT along with the one-water bridge, we have also
measured the KIE of the various ESDT reactions of 2N8S in pure D2 O and analyzed
the data using the SSDP program with double-boundary conditions, as described for
the ESPT case.
The SSDP program provided the fitting parameters of the SSDP for single- and
double-boundary conditions (SB and DB, respectively), as shown in Figure 10.23.
10.7 Insights into the Mechanism of Proton Transfer Through One-Water Bridge 285
100 1
SSDP-SB
SSDP-DB
Normalized counts
D2O
Normalized counts
10–1 SSDP-DB
10–2 0.1
10–3
10–4 t–3/2 0.01

0.1 1 10 100 1 10
Figure 10.23 (a) Comparison between the numeric solutions of SSDP-SB and SSDP-DB for
the deuterated photoacid case. (b) Time-resolved fluorescence of 2N8S in neat D2 O,
compared with the instrument function convoluted numeric solution of SSDP-DB.
The reaction radius a, the Coulomb interaction RD, and the mutual diffusion coef-
ficient D for the SB SSDP run were kept the same for the SSDP-DB run. As in the
H-photoacid case, the DB simulation for the D-acid results in much better data fit
than the SB simulations. Table 10.8 summarizes the kinetic parameters used in the
SSDP simulation.
Table 10.9 summarizes the best-fit kinetic parameters using the SSDP program
with DB conditions for 2N8S in water and D2 O with the calculated KIE values.
The analysis shows that the KIE for the proton dissociation reactions of 2N8S is
larger for the proton dissociation reaction than for the proton recombination reac-
tions in agreement with previous observations. The KIE for the dissociation reaction
of the OH group is larger than for the dissociation of the SO3 H group following the
order of their pKa∗ in water; the OH group is more acidic than the SO3 H group in
water, and so exhibits a larger KIE [48].
For the two recombination reactions, a more complex picture is obtained. The two
KIEs are relatively small, about 1.5, and identical in magnitude pointing out that the
proton recombination process is unlikely to be an elementary reaction and proba-
bly involves multiple PTs. PT along with a chain of water molecules is many times
described as a coupled multi-reactive event as was suggested for the recombination
of H+ with OH− in water [50]. The isotope effect in the equilibrium constant is very
small as expected for medium-strong acids, for example the discussion in Ref. [51].
10.7 Insights into the Mechanism of Proton Transfer

Through One-Water Bridge in Bifunctional 2-Naphthols
We focus in our discussion on the one-water-bridge case, which was accessed for the
first time and quantitatively characterized in our low-water-region experiments in
ACN. The ESPT bridge reaction of the 1 : 1 2N8S: water complex is pictorially drawn
in Figure 10.24.
Calculations in Ref. [13] have pointed out that the excited-state proton dissocia-
tion from the OH group to protonate the water bridge in 2N8S is endothermic and
R W IS W P W
Figure 10.24 Scheme of the step-wise double-ESPT reaction, which we envision taking
place in the one-water-bridge arrangement of 2N8S. R is the initial state, IS is the
protonated bridge state, which we envision as the intermediate state of the step-wise PT
reaction, and P is the final product state following the second proton transfer from the
bridge to the SO3 − group. The structure of the hydrogen-bonding complexes was calculated
for the photoacid in its electronic ground-state using the optimized structures generated by
Gaussian 09 with B3LYP//6-311G++(d,p) functional and is generally similar to the
structures calculated for the 1 : 1 complex of photoacid in the ES. Source: Kiefer et al.
the second PT from the water bridge to protonate the SO3 − group is exothermic with
the overall reaction being slightly up-hill. Usually, such a reaction path implies a
two-step PT reaction through a reaction intermediate: first step of slow protonation
of the water bridge by the dissociation of the photoacidic OH group followed by the
second step of fast deprotonation of the water bridge, which protonates the SO3 − side
group. We now expand on this picture by using the analysis developed by Hynes et al.
[52] for PT between an acid and a base (developed for a bimolecular acid–base neu-
tralization reaction in aqueous solutions) to qualitatively analyze the reaction mech-
anism. We first summarize the main results of our kinetic analysis. We find the ESPT
rate along with the one-water bridge to be relatively slow, 20 ns−1 , much slower than
the ESPT rate reported for 2N8S in ACN in presence of higher concentrations of
water and for 2N8S in aqueous solutions, which are at least an order of magnitude
faster. The reaction occurs between the two side groups of the bifunctional pho-
toacid, which we find (Table 10.3) to have similar pK a *’s, the overall ESPT reaction
being down-hill by at least 0.5 pK a units in agreement with our observation that this
ESPT reaction along with the one-water bridge is irreversible in absence of addi-
tional water molecules, which are likely to further solvate the two polar side groups
of the photoacid. The irreversibility of the reaction measured within the several ns
timescales of the excited-state lifetime of 2N8S is also in general agreement with
the free-energy calculations reported in Ref. [13]. According to Ref. [52], the ESPT
reaction is irreversible because of dynamic reasons emerging from the high-energy
barrier for the back reaction and not because of thermodynamic ones. In addition,
we find an identical ESPT rate for the bridge reaction of 2N8S and the bridge reaction
of 2N68dS although the latter is a stronger photoacid by about 1 pK a unit. Finally,
we find an identical small KIE of about 1.5 for the one-water-bridge ESPT reac-
tion for the two 2-naphthol derivatives when under similar low-water-concentration
conditions.
Relying on our experimental findings, we now discuss the mechanism of the
one-water-bridge ESPT reaction. Clearly, the reaction R to P, in Figure 10.24,
10.7 Insights into the Mechanism of Proton Transfer Through One-Water Bridge 287
involves double PT and the two ESPT reactions must be coupled to each other.
In other words, the presence of the water bridge makes PT possible by connecting
through the water bridge, the proton acceptor with the proton donor. In such a
case, the best way to describe the reaction mechanism is by a concerted two-step
PT mechanism. Using Hynes vision of acid–base PT [52], the rate-determining
step for the PT is a structural rearrangement, which brings the two heavy atoms
closer together and so compresses the relevant H-bond distance along which the
proton is transferred between them bringing the oxygen atoms within a short
distance suitable for PT. This distance is shorter than the equilibrium H-bond
distance between the proton-donor and proto-acceptor oxygen atoms; so, the first
reorganization step in the R state which allows PT is a fluctuation that compresses
the distance between the OH side group and the oxygen atom of the water-bridge
molecule. PT to the water bridge molecule W is then followed by a compression
of the protonated water bridge toward the proton-acceptor SO3 − oxygen atom, the
most productive fluctuation being the one that compresses both bridge distances
almost simultaneously. It follows that the activation of the overall PT reaction
lies in establishing the proper (short) position of the bridge water in between the
proton-donor and proton-acceptor oxygens, of about 2.5 Å [52], which allows for
rapid double PT. Our findings point to an additional ingredient that is likely to
be dominant in the one-water-bridge ESPT process, namely, vibrational modes of
the two-side groups, which are coupled to ring-vibration modes of the photoacids.
These vibrational modes involve charges and are affected by fluctuations in the
solute–solvent polar interactions. Put together, the configuration suitable for the
bridge PT reaction is created after relatively long waiting times as compared to
the actual PT event once a suitable configuration allowing PT is formed. The
small invariant KIE in 2N8S and 2N68S point out the importance of the first PT
reaction, which is activated by both intramolecular vibrational ring modes and
the solvent, both being insensitive to the isotope substitution and control the
distance between the OH group and the water-bridge molecule along which the
PT reaction occurs, which is isotope sensitive. The above-suggested mechanism
for the water-bridge-ESPT reaction is complex and is not based on concrete
simulations of the reactive system, which are clearly required. Isotope-sensitive
vibrational modes, such as the ones governing the various distances in the
hydrogen-bonding complex between the OH (OD) group and bridge-H2 O (D2 O)
molecule, apparently lead to a small KIE of 1.5, which is about the square root
of the mass ratio of the D/H isotopes. We suggest based on the above arguments
that the first PT reaction is from R to W and the second PT is from W to P
Figure 10.24. The two PT events are likely to occur concertedly almost at the
same time with the first PT R to W clocking both PT events. Our kinetic model
for the double through one-water-bridge ESPT process is schematically outlined
in Figure 10.25. Using the notations of Figure 10.25, the reactant state, state
1, is a one-water-hydrogen-bonding bridge formed in the ground-state of the
sulfonate-substituted 2N photoacids, the water residing in between the OH and
the 8-position SO3 − group but not necessary residing in equal distance from the
two side groups. An activated fluctuation in the OH· · · OH2 distance in the ES
1 2 3
O O O
R O H O H O S R R O O H O S R R O H O H O S R
H O H O H O
5 4
O O
R O H O H O S R R O H O H O S R
H O H O
Figure 10.25 Intramolecular sequential proton transfer through one-water bridge

between the two side groups of 2N8S, ΔpK a * = pK a *(ROH) − pK a *(RSO3 H) = 1.5 − 2.0 = −0.5.
of the photoacids brings the two oxygen atoms closer to within a short distance
appropriate for PT to the bridge-water depicted in state 2. ESPT then follows and
forms state 3 and the positive charge on the water bridge very rapidly compress
the distance between the protonated bridge water and the negatively charged
SO3 − group state 4. The second PT reaction follows almost immediately and
forms the product state, state 5, which is the proton-transferred state of the 1 : 1
water-photoacid H-bonding complex. This kinetic model is potentially important
for biological systems where proton gradients are many times known to be too small
to allow for the observed reactivity of the proton at its biological target assuming
a direct diffusion-controlled bimolecular reaction process. Secondary basic side
groups may interact reversibly with the proton and by doing so act to transiently
increase the local concentration of the proton in the vicinity of the primary O− side
group and thus may also greatly increase the probability of the main protonation
reaction.
Mechanistic considerations of proton availability led biophysicists to suggest the
existence of proton antennas made of basic groups, such as imidazole [53, 54],
which collects the proton and delivers it to its final biological target [55]. Our study
of the PT reactions in bifunctional photoacids opens the way for directly studying
proton-antennas systems [55].
10.8 Summary
The study gives an insight into the molecular mechanism of PT in aprotic or
low-water-content systems, such as those often found in biological systems.
Long-range PTs occupy an important place in biology and are assisted by the exis-
tence of multitude of deprotonation and protonation sites, and water-chains arrays,
which efficiently transfer the proton and proton-collecting antennas. We were able
to demonstrate how multiple protonation sites keep the local concentration of the
proton higher over extended period of time and by doing so enhance the probability
of the proton to react at its main protonation target, which may be a target having a
biological function. This study thus allows to approach the rule of local reversible
proton-binding sites in biological systems, such as found for an example, in the
carbonic anhydrase enzyme.
References 289
Acknowledgments
This work was supported in part by a grant from the Israel Science Foundation
(Grant No. 1587/16).
The paper is dedicated to the memory of the late Professor Dan Huppert.
References
1 Wraight, C.A. (2006). Chance and design – proton transfer in water, chan-
nels and bioenergetic proteins. Biochim. Biophys. Acta-Bioenerg. 1757 (8):
886–912.
2 Ando, K. and Hynes, J.T. (1995). HCl acid ionization in water: a theoretical
molecular modelling. J. Mol. Liq. 64: 25.
3 Ando, K. and Hynes, J.T. (1997). Molecular mechanism of HCl acid ionization in
water: ab initio potential energy surfaces and Monte Carlo simulations. J. Phys.
Chem. B 101: 10464.
4 Daschakraborty, S., Kiefer, P., Miller, Y. et al. (2016). Reaction mechanism
for direct proton transfer from carbonic acid to a strong base in aqueous
solution I: acid and base coordinate and charge dynamics. J. Phys. Chem. B
120 (9): 2271–2280.
5 Daschakraborty, S., Kiefer, P.M., Miller, Y. et al. (2016). Direct proton trans-
fer from carbonic acid to a strong base in aqueous solution II: solvent role in
reaction path. J. Phys. Chem. B 120 (9): 2281–2290.
6 Brønsted, J.N. (1923). Some observations about the concept of acids and bases.
Recl. Trav. Chim. Pays-Bas 42 (8): 718–728.
7 Lee, S.I. and Jang, D.J. (1995). Proton transfers of aqueous 7-hydroxyquinoline in
the first excited singlet, lowest triplet, and ground-states. J. Phys. Chem. 99: 7537.
8 Bardez, E. (1999). Excited-state proton transfer in bifunctional compounds. Isr. J.
Chem. 39: 319.
9 Cohen, B. and Huppert, D. (2001). Excited state proton-transfer reactions of
coumarin 4 in protic solvents. J. Phys. Chem. A 105: 7157–7164.
10 Kwon, O.H., Lee, Y.S., Yoo, B.K., and Jang, D.J. (2006). Excited-state triple pro-
ton transfer of 7-hydroxyquinoline along a hydrogen-bonded alcohol chain:
vibrationally assisted proton tunneling. Angew. Chem. Int. Ed. 45: 415.
11 Ditkovich, J., Mukra, T., Pines, D. et al. (2015). Bifunctional photoacids: remote
protonation affecting chemical reactivity. J. Phys. Chem. B 119: 2690.
12 Ditkovich, J., Pines, D., and Pines, E. (2016). Controlling reactivity by remote
protonation of a basic side group in a bifunctional photoacid. Phys. Chem. Chem.
Phys. 18: 16106.
13 Gajst, O., Pinto da Silva, L., Esteves da Silva, J.C.G., and Huppert, D. (2018).
Excited-state proton transfer from the photoacid 2-naphthol-8-sulfonate to ace-
tonitrile/water mixtures. J. Phys. Chem. A 122: 6166.
14 Gajst, O., Pinto da Silva, L., Esteves da Silva, J.C., and Huppert, D. (2018).
Enhanced excited-state proton transfer via a mixed water-methanol molecular
bridge of 1-naphthol-5-sulfonate in methanol-water mixtures. J. Phys. Chem. A

122: 4704.
15 Pines, E., Pines, D., Gajst, O., and Huppert, D. (2020). Reversible intermolecular-
coupled-intramolecular (RICI) proton transfer occurring on the R reaction-radius
a of 2-naphthol-6,8-disulfonate photoacid. J. Chem. Phys. 152 (7): 074205.
16 Ekimova, M., Hoffmann, F., Bekçioğlu-Neff, G. et al. (2019). Ultrafast proton
transport between a hydroxy acid and a nitrogen base along solvent bridges
governed by the hydroxide/methoxide transfer mechanism. J. Am. Chem. Soc.
141 (37): 14581–14592.
17 Hoffmann, F., Ekimova, M., Bekçioğlu-Neff, G. et al. (2016). Combined
experimental and theoretical study of the transient IR spectroscopy of
7-hydroxyquinoline in the first electronically excited singlet state. J. Phys. Chem.
A 120: 9378.
18 Codescu, M.-A., Weiß, M., Brehm, M. et al. (2021). Switching between pro-
ton vacancy and excess proton transfer pathways in the reaction between
7-hydroxyquinoline and formate. J. Phys. Chem. A 125 (9): 1845–1859.
19 Pines, E. and Huppert, D. (1986). Observation of geminate recombination in
excited state proton transfer. J. Chem. Phys. 84: 3576.
20 Pines, E. and Huppert, D. (1986). Geminate recombination proton-transfer
reactions. Chem. Phys. Lett. 126: 88.
21 Pines, E., Huppert, D., and Agmon, N. (1988). Geminate recombination
in excited-state proton transfer reactions: numerical solution of the
Debye-Smoluchowski equation with backreaction and comparison with experi-
mental results. J. Chem. Phys. 88: 5620.
22 Agmon, N., Pines, E., and Huppert, D. (1988). Geminate recombination in
proton-transfer reactions. II. Comparison of diffusional and kinetic schemes.
J. Chem. Phys. 88: 5631.
23 Huppert, D., Pines, E., and Agmon, N. (1990). Long time behavior of reversible
geminate recombination reactions. J. Opt. Soc. Am. B: Opt. Phys. 7: 1545.
24 Pines, D. and Pines, E. (2001). Direct observation of power-law behavior in
the asymptotic relaxation to equilibrium of a reversible bimolecular reaction.
J. Chem. Phys. 115 (2): 951.
25 Simkovitch, R., Pines, D., Agmon, N. et al. (2016). Reversible excited-state proton
geminate recombination: revisited. J. Phys. Chem. B 120: 12615.
26 Krissinel’ EB, Agmon N. (1996). Spherical symmetric diffusion problem.
J. Comput. Chem. 17: 1085.
27 Weller, A. (1961). Fast reactions of excited molecules. Prog. React. Kinet.
1: 187–213.
28 Rosenberg, J.L. and Brinn, I. (1972). Excited state dissociation rate constants in
naphthols. J. Phys. Chem. 76: 3558.
29 Ireland, J.F. and Wyatt, P.A.H. (1976). Acid-base properties of electronically
excited states of organic molecules. Adv. Phys. Org. Chem. 12: 131–221.
30 Rosenberg, J.L. and Brinn, I.M. (1976). Excited states of naphthols. Part
2 – Molecular orbital calculations on substituted naphthols. J. Chem. Soc.,
Faraday Trans. 72: 448–452.
References 291
31 Krishnan, R., Lee, J., and Robinson, G.W. (1990). Isotope effect on weak acid
dissociation. J. Phys. Chem. A 94: 6365.
32 Tolbert, L.M. and Haubrich, J.E. (1990). Enhanced photoacidities of
cyanonaphthols. J. Am. Chem. Soc. 112: 8163.
33 Tolbert, L.M. and Haubrich, J.E. (1994). Photoexcited proton transfer from
enhanced photoacids. J. Am. Chem. Soc. 116: 10593–10600.
34 Huppert, D., Tolbert, L.M., and Linares-Samanieg, S. (1997). Ultrafast
excited-state proton transfer from cyano-substituted 2-naphthols. Phys. Chem.
A 101: 4602–4605.
35 Genosar, L., Leiderman, P., Koifman, N., and Huppert, D. (2004). Effect of pres-
sure on proton transfer rate from a photoacid to a solvent. 3. 2-Naphthol and
2-naphthol monosulfonate derivatives in water. J. Phys. Chem. 108: 1779–1789.
36 Hammett, L.P. (1937). The effect of structure upon the reactions of organic
compounds. Benzene Derivatives. J. Am. Chem. Soc 59 (1): 96–103.
37 Hansch, C., Leo, A., and Taft, R.W. (1991). A survey of Hammett substituent
constants and resonance and field parameters. Chem. Rev. (Washington, DC, US)
91 (2): 165–195.
38 Ritchie, C.S. (1990). Physical Organic Chemistry. The Fundamental Concepts
(Studies in organic chemistry)., vol. 4. M. Dekker.
39 Park, S. and Agmon, N. (2009). Multisite reversible geminate reaction. J. Chem.
Phys. 130: 074507.
40 Kamlet, M.J., Abboud, J.L.M., Abraham, M.H., and Taft, R.W. (1983). Linear
solvation energy relationships. 23. A comprehensive collection of the solva-
tochromic parameters, pi-star, alpha and beta, and some methods for simplifying
the generalized solvatochromic equation. J. Org. Chem. 48: 2877.
41 Pines, E., Pines, D., Ma, Y.-Z., and Fleming, G.R. (2004). Femtosecond
pump-probe measurements of solvation by hydrogen-bonding interactions.
ChemPhysChem 5 (9): 1315–1327.
42 Keinan, S., Pines, D., Kiefer, P.M. et al. (2015). Solvent-induced O–H vibration
red-shifts of oxygen-acids in hydrogen-bonded O–H⋅⋅⋅base complexes. J. Phys.
Chem. B 119 (3): 679–692.
43 Mohammed, O.F., Pines, D., Dreyer, E. et al. (2005). Sequential proton transfer
through water bridges in acid base reactions. Science 310: 83.
44 Mohammed, O.F., Pines, D., Nibbering, E.T.J., and Pines, E. (2007). Base-induced
solvent switches in acid-base reactions. Angew. Chem. Int. Ed. 46: 1458.
45 Kolthoff, I.M. and Chantooni, M. (1968). Protonation in acetonitrile of water,
alcohols, and diethyl ether. J. Am. Chem. Soc. 90: 3320.
46 Chantooni, M.K. and Kolthoff, I.M. (1970). Reevaluation of the formation con-
stants of the hydrated proton in acetonitrile. J. Am. Chem. Soc. 92: 2236.
47 Raamat, E., Kaupmees, K., Ovsjannikov, G. et al. (2013). Acidities of strong neu-
tral Brønsted acids in different media. J. Phys. Org. Chem. 26: 162.
48 Prémont-Schwarz, M., Barak, T., Pines, D. et al. (2013). Ultrafast excited-state
proton-transfer reaction of 1-naphthol-3,6-disulfonate and several 5-substituted
1-naphthol derivatives. J. Phys. Chem. B 117 (16): 4594–4603.
49 Yu, H., Kwon, O.-H., and Jang, D.-J. (2004). Migration of protons during the
excited-state tautomerization of aqueous 3-hydroxyquinoline. J. Phys. Chem. A
108 (28): 5932–5937.
50 Natzle, W.C. and Moore, C.B. (1985). Recombination of hydrogen ion (H+) and
hydroxide in pure liquid water. J. Phys. Chem. 89 (12): 2605–2612.
51 Pines, E. (2006). The kinetic isotope effect in the photo-dissociation reaction. In:
Isotope Effects in Chemistry and Biology (ed. A. Kohen and H.H. Limbach). Boca
Raton: CRC Press.
52 Kiefer, P.M., Daschakraborty, S., Pines, D. et al. (2021). Electron flow charac-
terization of charge transfer for carbonic acid to strong base proton transfer in
aqueous solution. J. Phys. Chem. 125 (41): 11473–11490.
53 Shenderovich, I.G., Lesnichin, S.B., Tu, C. et al. (2015). NMR studies of
active-site properties of human carbonic anhydrase II by using 15N-labeled
4-methylimidazole as a local probe and histidine hydrogen-bond correlations.
Chem. Eur. J. 21: 2915.
54 Jagoda-Cwiklik, B., Slavicek, P., Cwiklik, L. et al. (2008). Ionization of imidazole
in the gas phase, microhydrated environments, and in aqueous solution. J. Phys.
Chem. A 112: 3499.
55 Bondar, A.-N. (2021). Proton-binding motifs of membrane-bound proteins: from
bacteriorhodopsin to spike protein S. Front. Chem. 9 (413): 685761.
293
11
Spectroscopic Determination of Hydrogen Bond Energies

Mausumi Goswami 1 and Elangannan Arunan 2
1
Vellore Institute of Technology, School of Advanced Sciences, Department of Chemistry, Vellore,
632014, India
2
Indian Institute of Science, Department of Inorganic and Physical Chemistry, Bangalore, 560012 India
11.1 Introduction
In this chapter, we discuss various experimental spectroscopic techniques applied
for the determination of the hydrogen bond (H-bond) energies. Clearly, to deter-
mine the H-bond energies, we need to define what a H-bond is. We refer to the recent
IUPAC definition [1] and the accompanying technical report [2] for the readers inter-
ested in the recent definition and the history of the H-bond. According to the IUPAC
definition, a typical H-bond may be depicted as X–H· · ·Y–Z, where the three dots
denote the bond. The dash (–) denotes the covalent bond. It sounds strange today
that chemists did not worry about a symbol for ionic bonds. In any case, the def-
inition of hydrogen bond follows: “The hydrogen bond is an attractive interaction
between a hydrogen atom from a molecule or a molecular fragment X–H in which
X is more electronegative than H, and an atom or a group of atoms in the same or a
different molecule, in which there is evidence of bond formation.”
A careful look at the definition of H-bond already points out some operational
difficulties. The H- bonds could be inter or intramolecular. This chapter discusses
intermolecular H-bonds exclusively. An intermolecular H-bond has two molecules
that are bound together through a H-atom from one of them connecting to an elec-
tron donor in the other molecule. One could envisage the breaking of this H-bond
leading to the two molecules parting ways. For example, a textbook example of the
bond is water dimer, having O–H· · ·O hydrogen bond. The H-bond energy for water
dimer has been accurately determined recently by Reisler’s group using the veloc-
ity map imaging (VMI) experiment in which, the rotational and vibrational energies
of the two H2 O fragments could be measured using spectroscopy [3]. The H-bond
energy for (H2 O)2 is related to the energy change in the following reaction:
(H2 O)2 → 2H2 O ΔE = 1105 ± 10 cm−1 (11.1)
Reisler’s measurement is in excellent agreement with an accurate theoretical
estimate by Bowman’s group 1103 ± 4 cm−1 [4]. More accurately, this dissociation
294 11 Spectroscopic Determination of Hydrogen Bond Energies
Energy
D0
De
Hydrogen bond dissociation coordinate
Figure 11.1 Schematic showing De (equilibrium dissociation energy) and D0 for a

hydrogen-bonded complex.
energy, D0 , is the difference in energies between the reactant (water dimer) and
products (two water molecules), all being in their ground vibrational state. The dif-
ference in energies between the equilibrium potential energies of a molecule and
the two fragments is denoted as De . Figure 11.1 shows the difference between
De and D0 for clarity. Experimental determination usually leads to D0 . However,
determining the energies of more vibrational and rotational levels does allow
experimental determination of De [5].
Beyond the distinction between the De and D0 , there is yet another complication
in determining the intermolecular H-bond energy, which can be illustrated by
the example of hydrogen sulfide dimer. The microwave spectrum of this dimer
was published by Das et al. only in 2018, confirming that the structure is indeed
H-bonded [6]. This was followed by some theoretical work by Fernández-Alarcón
et al. that questioned the characterization of H2 S dimer as H-bonded [7]. The
critical observation by the authors is that there are two minima in the intermolec-
ular torsional potential surface for (H2 S)2 with a barrier of less than 1 kcal mol−1
separating them. The difference between the two structures of (H2 S)2 is mainly
in the orientation of the acceptor H2 S in the dimer. It is important to realize that,
though there is only one minimum for (H2 O)2 , there are three possible tunneling
motions that lead to eight indistinguishable minima [8]. We have been aware of this
and proposed a universal criterion for a H-bond in 2009 [9]. The criterion is that the
zero-point energy along any torsional degree of freedom that can break a H-bond be
significantly below the energy barrier along that coordinate so that there is at least
one bound level. The criterion is explained in Figure 11.2. Clearly, if the thermal
energy along any of these coordinates exceeds the barrier along that coordinate, the
characterization as H-bond is no longer reasonable and the molecule would be freely
rotating and the H-bond would no longer have the strength needed to freeze the
orientation. In fact, the difference between ice and solid H2 S may have contributed
significantly to the debate and controversies in this field. The detailed discussion in
Refs. [9] and [10] might convince the readers about this criterion [9, 10]. We point
out that the torsional motion that led to two different minima in (H2 S)2 does not
V=0 Case II. No Hydrogen bond
H-X---Y H-X---Y
X
θ X-H---Y
H
Energy
V=0 Case I. Hydrogen bonded

Y
0 90 180 270 360

θ°
Figure 11.2 Schematic showing the effect of zero-point motion on a H-bonded complex.
If the zero-point energy for a motion which takes the hydrogen away from the acceptor is
above the barrier along that coordinate, the averaged geometry of the complex will not be
H-bonded. If the zero-point energy is below the barrier, the averaged geometry of the
complex will be H-bonded. Source: Goswami and Arunan [9]/Royal Society of Chemistry.
break the hydrogen bond [7]. The experimental binding energy, De , for (H2 S)2
has been estimated from pressure broadening coefficients as 607 ± 8 cm−1 , again
in excellent agreement with theoretical estimates [11, 12]. For comparison, the
torsional barriers, reported for (H2 S)2 are typically less than 350 cm−1 .
One more general critique about H-bond (and hence, H-bond energy) is that it is
a group phenomenon and so defining the bond energy is not practical. To highlight
this point further, we point out that there may not be a unique vibrational mode that
involves only intermolecular stretching as a normal mode (NM). We contend that
it is still possible to define H-bond energy as defined by Eq. (11.1). It may help to
consider n-pentane for example. In this molecule, no molecular orbital (MO) or NM
of vibration would exist corresponding to individual C—C bonds, i.e. there are only
MOs/vibrations that involve all C—C bonds behaving as a group. However, one can
still define individual bond energies as shown below:
n-C5 H12 → CH3 + C4 H9
n-C5 H12 → C2 H5 + C3 H7
The intermolecular H-bond energies that we discuss in this chapter refer to such
bond energies.
We should point out that there are examples in which more than one
H-bond can stabilize a complex or solid. Two examples from our laboratory
are phenylacetylene· · ·water and propargyl alcohol· · ·water complexes, both of
which have H2 O accepting and donating one hydrogen bond each [13, 14]. It is not
easy to define the bond energy for individual H-bonds in these examples. These
examples have similarities to intramolecular H-bonds. Defining and determining
intramolecular H-bond energy is more complicated and we refer to two articles
discussing this in detail, both employing theoretical methods [15, 16]. Direct
experimental determination of intramolecular hydrogen bond energy appears
impossible. Electron density at a bond critical point, both from the experiment [17]
and Atoms in Molecules theory [18], could help in giving reliable bond energies in
these cases. This chapter does not cover intramolecular hydrogen bond energies.
In this chapter, we discuss experimental methods based on the spectroscopic
measurements of H-bonded complexes which provide the binding energy or the
H-bond energies (dissociation energy including the zero-point energy, henceforth
D0 ) of these complexes. Experimental techniques described in the chapter cover
different spectral regions including infrared, ultraviolet-visible, and microwave.
Barring the room-temperature determination of the free energy of the complex
formation described in Section 11.2.2, other techniques described in the chapter
employ the supersonic molecular beam method for the preparation of H-bonded
complexes in the gas phase. Interested readers are pointed toward Ref. [19] for a
detailed description of these techniques [19]. Different techniques, e.g. vibrational,
electronic, and rotational spectroscopic methods are known to provide highly pre-
cise geometrical parameters, rotational constants, and vibrational frequencies of the
H-bonded clusters, but only a handful of the techniques enable the measurement
of D0 of these complexes. Examples covered in this chapter are only representative
and have been chosen for a better understanding of the techniques. Techniques
described in the chapter also have a variable level of accuracy based on the method
of choice. For example, the VMI technique following the infrared pre-dissociation
measurements (described in Section 11.2.1.2) provides a highly accurate value of
D0 whereas the Birge–Sponer extrapolation method discussed in Section 11.3.2 pro-
vides only an estimate. Accurate values of D0 of H-bonded complexes determined
from the experimental measurements not only can validate the nature of PES
obtained from the theoretical calculations, but also can expand the understanding
of co-operative interactions or many-body interactions in these complexes and can
also serve as an input to other branches of science, e.g. atmospheric chemistry.
11.2 Binding Energy Measurement Involving Infrared

(IR) Excitation
11.2.1 Measurement of the Dissociation Energy of H-Bonded
Complexes Through Vibrational Pre-dissociation Dynamics via Infrared
Excitation
The focus of this section is to highlight the techniques which were used to deter-
mine the dissociation energy of H-bonded complexes by breaking the H-bond
through the excitation of an intramolecular vibration, a phenomenon also known
as vibrational pre-dissociation. Since the intermolecular binding energies asso-
ciated with the H-bonded complexes are less than the vibrational energy of the
monomers, the phenomenon of vibrational pre-dissociation is common in these
complexes. By determining the final state of photofragments, information on the
bond breaking process including the dissociation energy (D0 ) can be obtained.
The extent of coupling of the intramolecular and the intermolecular modes would
determine the lifetime of pre-dissociation. Because of the frequency difference
of the intramolecular and intermolecular vibration, the pre-dissociation process

following the excitation of the former is slow and thus allows enough time for the
excitation of different internal states of the IR photofragments [20]. By measuring
the translational energy distribution or the angular distribution for each internal
state of one of the photofragments, the state distributions for the cofragment
can be determined from the conservation of energy. From the fits of the angular
distribution, the dissociation energy (D0 ) can be determined. Discussion in this
section is mainly focused on extracting D0 from the experimental techniques and
the details of the energy disposal mechanism in the fragments are not addressed
here. Interested readers are referred to the relevant reviews on this topic [20–24].
11.2.1.1 Optothermal Bolometric Determination

Pioneering studies by Miller and coworkers on the vibrational pre-dissociation
dynamics of several H-bonded complexes helped determine their binding energies
[20, 21]. As the photofragments created in the pre-dissociation process are in
different internal states, their recoil velocities are different resulting in the spatial
separation of the IR photofragments in the molecular beam. The resulting spatial
separation of the IR photofragments due to their difference in the translational
energy can be imaged and was first recorded by Miller and coworkers during the
vibrational pre-dissociation dynamics studies of the H-bonded complex (HF)2 and
later extended to other complexes listed in Table 11.1 [20, 25, 26]. In their experi-
ments of (HF)2 by Miller and coworkers, the intensity of the photofragment as a
function of the recoil angle was measured by employing an optothermal bolometer
detector. The resulting angular distribution is reflective of the translational energy
distribution of the photofragments. Such method is suitable where the density
of states in the photofragments is low enough to allow for significantly different
kinetic energy releases in different channels leading to their independent detection
and was used for (HF)2 [25]. Angular distributions obtained for (HF)2 were fit and
the fitting involved varying the dissociation energy so that the translational energy
released in the process matches the intensity profile in the angular distribution.
The released translation energy (ETrans ) in the final states is given by
ETrans = hν + Eint − Erot − D0 (11.2)
Eint is the internal energy of the dimer and Erot is the rotational energy of the
fragments and h𝜈 corresponds to the IR frequency at which the dimer is excited.
From the fits, the dissociation energy D0 for (HF)2 was determined as 1062 ± 1 cm−1 .
In their experiments with (HF)2 , while both the photofragments were detected
together, a large electrical field was applied in the case of DF–HF and HF–DF
complexes to orient the parent molecules using brute force so that the two
photofragments recoil in the opposite directions in the laboratory frame and thus
can be detected separately using pendular state spectroscopy. The dissociation
energies for HF· · ·DF and DF· · ·HF complexes were determined as 1157 (2) cm−1
and 1082 (2) cm−1 [26].
The D0 of an analogous H-bonded system (HCl)2 was determined by a different
technique where the energy in v = 1 levels of the H–Cl stretch of both donor
Table 11.1 Experimental (determined via optothermal bolometric detection or VMI)

values of D0 listed for a few H-bonded complexes. Theoretically calculated values of
Dissociation energies (D0 or De ), wherever available, are provided for comparison.
D0 (cm−1 ) D0 (cm−1 ) De (cm−1 )

Complex (experimental) (theoretical) (theoretical)
(HF)2 1062(2) [25] 1609 [50]

(HCl)2 439 ± 1 [27] 431 ± 1 [24]
HF· · ·DF 1157(2) [26]
DF· · ·HF 1082(2) [26]
C2 H2 · · ·HF 1088(2) [51] 655 [52]
C2 H2 · · ·HCl 830(6) [53] 393 [52]
700 ± 10 [54]
HF· · ·HCl 642(2) [55] 654 [56]
C2 H2 · · ·DCl 755 ± 10 [54]
C2 H2 · · ·NH3 900 ± 10 [57] 777 [58]
OCO· · ·HF 672(4) [59]
HCN· · ·HF 1970(10) [60] 1945 [61]
NO· · ·HF 448 [62]
OC· · ·HF 732(2) [28] 1211 [63]
H2 O· · ·H2 O 1105 ± 10 [3] 1105 ± 4 [3]
ClH· · ·OH2 1334 ± 10 [42] 1334.63 [64]
HOH· · ·NH3 1538 ± 10 [65]
NH3 · · ·NH3 660 ± 20 [47]
(H2 O)3 → H2 O + (H2 O)2 2650 ± 150 [44] 2726 ± 30 [44]
(HCl)3 → HCl + (HCl)2 1142 ± 20 [46] 1133 ± 2 [46]
(HCl)3 → 3HCl 1545 ± 10 [46] 1564 ± 1 [46]
HCl(H2 O)3 → HCl + (H2 O)3 2100 ± 300 [49]
HCl(H2 O)3 → H2 O + HCl(H2 O)2 2400 ± 100 [48]
(HF)5 <2941 [30]
(HF)6 <2854 [30]
and acceptor molecules was deposited by Stimulated Raman scattering and the
resonance-enhanced multiphoton excitation scheme was used for the state-specific
detection following the vibrational pre-dissociation [27]. The derived dissociation
energy for (HCl)2 is included in Table 11.1. As the bolometer in itself is unable to
detect the photofragments in a state-specific way, a probe CW-wave IR laser was
used by Miller coworkers for the state-specific detection of photofragments in some
cases including N2 –HF, CO2 –HF, and CO–HF [28, 29].
This section would be incomplete without touching upon the vibrational
pre-dissociation studies of large (HF)n (n = 5–7) clusters using the bolometer,
which led to the determination of the upper limit of D0 for these large clusters [30].
As mentioned earlier, resolution in the angular distribution of the photofragments

depends on the recoil energies in different channels which in turn depends on the
density of states in the photofragments. Pre-dissociation of large clusters (trimers
and larger clusters) leads to larger size photofragments that have higher density of
states due to a relatively larger number of low-frequency vibrations (than those of
the photofragments created in the vibration pre-dissociation of dimers) and smaller
values of rotational constants. (HF)n clusters are an interesting class of H-bonded
system as the H-bonding in solid HF has a zigzag chain pattern and in gas phase
when n ≥ 3, global minimum of (HF)n has a cyclic configuration [31]. Concerning
the size of the photofragments, (HF)n is a favorable system as the excitation of
the H-bonded H–F stretch in large HF clusters (n ≥ 3) leads to the dissociation
of one HF monomer fragment which then can be probed to obtain state-to-state
information. When the probe laser is off, the angular distribution arises from
both the cofragments. The angular distribution of H–F monomer photofragments
obtained by exciting the H–F stretch frequencies of (HF)5 (3302 cm−1 ), (HF)6
(3240 cm−1 ), and (HF)7 (3213 cm−1 ) lacked structures unlike what was seen in the
case of the vibrational pre-dissociation of the dimer fragment. To detect the HF
monomer fragment state selectively, a second F-center laser was used by Miller and
coworkers which led them to separate the contribution of the monomer cofragment
in the total angular distribution. The photolysis laser was fixed at the H–F monomer
frequency of the selected cluster size and the probe laser was scanned through a
specific HF monomer transition. In the case of the pentamer, because of the factor
of four difference in the mass of monomer and tetramer fragments, the latter would
appear at smaller laboratory angles. At the time when these experiments were
performed in Miller’s group, there was a controversy regarding the H–F stretch
frequencies in large HF clusters. However, the ability to separate the angular
distribution for the monomer and the tetramer fragments and their simultaneous
fitting for (HF)5 provided confirming evidence that the observed signal was indeed
from (HF)5 . Similar experiments were performed for (HF)6 and (HF)7 . Since the
correlated internal energy of the cluster fragments could not be determined in these
experiments, an upper bound for D0 was estimated and is listed in Table 11.1.
11.2.1.2 Velocity Map Imaging

The pioneering work by Roger Miller’s group set the path for the techniques
which could be used for the imaging studies of photofragments produced in
the pre-dissociation of non-covalent complexes. The technique of VMI was first
proposed by D.W. Chandler and Paul Houston in 1987 and later developed by
Eppink and Parker in 1997 [32, 33]. The first application of VMI to measure
dissociation energies of van der Waals’ complexes of rare gas atoms with aromatic
systems including the rare gas complexes of difluorobenzene was pioneered by
Lawrance and coworkers and the study was later extended to the H-bonded
complex difluorobenzene· · ·H2 O [34–36]. The first VMI studies published by
Lawrance and coworkers were the measurements of the dissociation energies of
the S0 and S1 states of diflurobenzene· · ·Ar complex and the dissociation energy
for the ground state of the cationic complex, and the dissociation energy of
(C6 H6 · · ·Arn )+ (n = 1 and 2) complexes [34]. As the discussion in this section is

limited to vibrational pre-dissociation measurements through IR excitation, the
technique adopted by Lawrance and coworkers for difluorobenzene· · ·H2 O is not
described here.
Although Lawrance and coworkers were the first to apply VMI to the study
of van der Waals’ complexes, Reisler and coworkers were the first to use VMI
for the determination of experimental dissociation energies of several H-bonded
complexes by obtaining the pair-correlated product state distributions following
vibrational predissociation [22–24]. Determination of D0 using VMI is pertinent to
the existence of a state-selected detection scheme for at least one of the monomer
fragments. Reisler and coworkers used the resonance enhanced multi-photon ion-
ization (REMPI) detection scheme for the state-selectivity of product states [22, 24].
The pair-correlated product state distributions were measured by monitoring
one of the fragments for which a REMPI detection scheme exists. The phrase
“pair-correlated product state distributions” points towards the state distributions
of the undetected cofragment 2 correlated with the monitored internal states of the
detected cofragment 1. In the detection scheme used by Reisler and coworkers, the
IR photon was used to excite the intramolecular vibration of one of the fragment
monomers. As the energy deposited by the infrared photon is above the dissociation
energy of the complex, it leads to vibrational predissociation. Upon dissociation, the
fragments are produced in the distribution of vibrational and rotational states and
the fragments in selected rovibrational states are ionized by tunable UV radiation.
The ionized fragments are detected mass-selectively by a position-sensitive detector.
Following the ionization, the ionized fragments are passed through an ion optics
assembly, a field-free drift tube and then detected by a multichannel plate which
is coupled to a phosphor screen which is monitored by a charge-coupled device
camera. The ion hitting events are thus imaged as the 2-D projections and the 3-D
images are reconstructed from the 2-D images. After summing the angular distri-
bution for each radius, the velocity distributions were obtained and then converted
to the center-of-mass translational energy (ET ) distribution. The schematic of the
experimental setup for the VMI detection scheme is shown in Figure 11.3.
Assuming that the H-bonded complex under investigation is a dimer, the conser-
vation of energy would require:
Eint (dimer) + hν = D0 + ET + Evib (detected cofragment 1)

+ Erot (detected cofragment 1) + Evib (cofragment 2 )
+ Erot (cofragment 2) (11.3)
where Eint is the internal energy of the dimer and in a molecular beam experiment,
typically, the complexes are created in their ground vibrational and electronic states
whereas several rotational states are populated in the expansion (corresponding to
a temperature of a few Kelvin), h𝜈 corresponds to the IR frequency at which the
complex is excited, D0 is the dissociation energy of the complex, Evib and Erot are
the vibrational and rotational energies of the cofragment 1 (detected fragment) and
2 (undetected fragment). ET is center-of-mass translational energy. While the Evib,rot
UV radiation
Pulsed nozzle
CCD
Ion optics Camera
Sample
Cryopumps
MCP–phosphor
assembly
IR radiation
Figure 11.3 Schematic of the experimental setup for VMI measurement. Source: Samanta
et al. [23]/with permission of Elsevier.
of the cofragment 1 is detected and defined via the REMPI technique, ET is deter-
mined from the images, Evib,rot of the cofragment 2 and D0 are determined from the
fits of the reconstructed images. Accuracy of D0 relies on having unique assignments
of the structures of velocity distributions to the internal states of the cofragment. The
velocity distributions obtained by monitoring different internal states of the cofrag-
ment must fit to a single value of D0 . The best resolution in the structure of velocity
distribution is obtained for the fragments having low recoil velocity which is true for
most of the clusters. In fact, according to the energy disposal guideline proposed by
Ewing which is known as “Energy gap law,” the fragments produced in the vibra-
tional predissociation tend to minimize the transfer of quanta in the dissociation
channel and thus have a higher propensity of vibrational and rotational excitation
resulting in a smaller translational energy release (i.e. lower recoil velocity) [37–39].
In the following paragraph, we describe the determination of D0 of (H2 O)2 to famil-
iarize the reader with this technique in detail.
The D0 of (H2 O)2 was determined by Reisler and coworkers [3, 40]. Structure of
(H2 O)2 is shown in Figure 11.4. Briefly, the (H2 O)2 was formed by expanding seeded
Helium gas through a pulsed valve nozzle. For the experiments, two modes were
used for data collection, the time-of-flight (TOF) mode was used for the spectro-
scopic investigation and the VMI mode was used for imaging the velocity distribu-
tion. The first step of the experiment is to record the IR spectrum of the dimer in the
region of interest which is done to separate the dimer spectrum from the spectra of
monomer and other clusters. The spectrum, which is also called Infrared “action”
spectrum was recorded by scanning the IR laser in the region of the bound O–H
Figure 11.4 Structure of water dimer.

Figure 11.5 Infrared spectrum of

1.1
(H2 O)2 recorded while monitoring
1.0 the C̃ 1 B1 (000)←X̃ 1 A1 (010) REMPI
Intensity (arb.units)
transition of H2 O. Rotational
0.9
assignment of the monitored state
0.8 of H2 O photofragments is JK′′ k = 321 .
a c
The red trace indicates the
0.7
enhancement of REMPI signal from
0.6 the dimer and the black trace is the
background signal. Source:
0.5 Rocher-Casterline et al. [3]/with
0.4 permission of AIP Publishing.
3585 3590 3595 3600 3605 3610

IR energy (cm–1)
stretch while monitoring the H2 O monomer in the selected rovibrational state by

(2 + 1) REMPI. Thus, one of the prerequisites of the experiment is to have a known,
well-resolved REMPI spectrum of one of the cofragment. For the (H2 O)2 experiment,
a specific rovibrational state of the H2 O monomer fragment was monitored via the
C̃ 1 B1 (000) ← X̃ 1 A1 (000 and 010) bands of H2 O while scanning the IR laser. A rep-
resentative Infrared action spectrum of (H2 O)2 is shown in Figure 11.5. Next, the IR
laser was tuned to the maximum in the absorption spectrum of (H2 O)2 (3602 cm−1 )
while the UV laser was scanned to record the REMPI spectrum of H2 O fragment for
both “IR on” and “IR off” conditions. The water monomer fragments formed in the
vibrationally excited states (000) and (010) were detected via C̃ 1 B1 (000) ← X̃ 1 A1
(000) and C̃ 1 B1 (000)← X̃ 1 A1 (010) transitions respectively. The enhancement of
the REMPI signal pertaining to the “IR on” condition provided information on the
rovibrational state distribution of the H2 O fragment. Next, by monitoring the H2 O
fragment in specific internal states, the 2-D projections were measured which were
converted to 3-D velocity distribution. From the conservation of energy,
Eint (H2 O · · · H2 O) + h𝜈 = D0 + ET + Evib (H2 O) + Erot (H2 O)
+ Evib (H2 Ocofragment ) + Erot (H2 Ocofragment ) (11.4)
By putting Eint (H2 O· · ·H2 O) = 5 cm−1 , h𝜈 = 3602 cm−1 , and by determining ET from
the velocity distributions, while Evib (H2 O) and Erot (H2 O) are defined by the REMPI
selection scheme, the dissociation energy D0 of the (H2 O)2 and the Evib,rot of the
cofragments were determined from the fits by monitoring the transitions from X
1 A (000) and X 1 A (010) states. It is worth noting that these experiments could
1 1
not distinguish between the donor and acceptor fragments in the dimer. Quasiclas-
sical trajectory calculations of (H2 O)2 and (D2 O)2 by Bowman and coworkers using
full-dimensional ab initio potential showed that there is no significant difference in
the rovibrational distributions of the donor and acceptor fragments for both (H2 O)2
and (D2 O)2 [41]. A similar experiment was also performed for (D2 O)2 by exciting the
deuterium bonded O–D stretch at 2634 cm−1 [40]. Because of the fast predissociation
of the high J and K levels of C 1 B1 (000) excited state of H2 O, higher rotational levels
could not be monitored whereas, for D2 O, smaller spacing between the consecutive
Figure 11.6 Experimental ~

180
velocity distributions (in red) of H2O X(000); 32,1
H2 O and D2 O photofragments Cofrag X(000)
produced in the vibrational Cofrag X(010)
predissociation of (H2 O)2 and
120
(D2 O)2 . The REMPI selected states
Counts
are specified in the respective
figures. The Gaussian curves
indicate the correlated states of
the cofragment. Source: Ch’ng 60
et al. [40]/with permission of
(a) 0
~
H2O X(010); 32,1
60
Gaussians
Counts
40
20
(b) 0
~
D2O X(000); 13*1,12
150 Gaussians
Counts
100
50
0
0 200 400 600 600
(c) Velocity (m s–1)
rotational levels gave rise to congestion in the spectrum. The dissociation energy
of (H2 O)2 and (D2 O)2 were determined to be 1105 ± 10 and 1244 ± 10 cm−1 . The
velocity distributions of H2 O and D2 O fragments produced in the vibrational predis-
sociation of (H2 O)2 and (D2 O)2 is shown in Figure 11.6. Although (H2 O)2 is probably
one of the most extensively studied hydrogen-bonded systems, prior to the studies
by Reisler’s group, only the dissociation energy values obtained by indirect methods
were known. Diffusion Monte Carlo simulation of (H2 O)2 and (D2 O)2 on an ab initio
potential energy surface (PES) showed an excellent match of theoretical and expe-
riential values for D0 of (H2 O)2 and (D2 O)2 [4, 41]. The theoretical values of D0 are
1103 ± 4 and 1244 ± 5 cm−1 for (H2 O)2 and (D2 O)2 respectively. Such an excellent
match of theory and experiments for (H2 O)2 makes it an ideal system to be used as
a benchmark for the theoretical calculations on the dissociation energy of the com-
plexes. A difference of 141 cm−1 between (H2 O)2 and (D2 O)2 , where (D2 O)2 is more
stable than (H2 O)2 , is consistent with the consensus that deuterium bonds are more
stable than hydrogen bonds.
As the acceptor and donor fragments cannot be detected separately in the VMI
studies of (H2 O)2 and (D2 O)2 , H2 O· · ·HCl represents an interesting example where
the velocity distribution of both the fragments could be detected in separate experi-
ments [42, 43]. Minimum energy structure of H2 O· · ·HCl complex has a Cl−H· · ·O
geometry where HCl is the H-bond donor and H2 O is the H-bond acceptor. The H−Cl
stretch in the dimer is redshifted from HCl monomer by 162 cm−1 . During the vibra-
tional predissociation measurement, the Cl–H stretch in the dimer was excited and
vibrational predissociation measurement was performed subsequently. In the first
experiment by Reisler and coworkers, upon excitation of the Cl–H stretch, the HCl
photofragment was detected via the (2 + 1) REMPI excitation and in a later experi-
ment, (2 + 1) REMPI experiment monitored specific J′′ levels of H2 O. Both studies
agreed to a dissociation energy value of 1334 ± 10 cm−1 and showed a preference
for the rotational excitation in HCl over the translational energy release. Although
the REMPI detection of H2 O photofragment is tricky [43], the signal-to-noise ratio
was adequate to determine the D0 and the correlated state distribution with enough
accuracy for H2 O· · ·HCl. The REMPI detection scheme for H2 O was later used to
determine the dissociation energy of (H2 O)2 as described above.
Apart from the (H2 O)2 , dissociation energies of several other H-bonded complexes
have been determined by the VMI technique through the REMPI detection scheme
and are included in Table 11.1. For complexes that show a well-resolved structure
in the velocity distribution, D0 can be determined within ±10 cm−1 [24]. The VMI
studies were extended to trimers like (HCl)3 and (H2 O)3 and these experiments
are challenging because of the higher density of states in the cofragments (e.g. a
dimer cofragment created in the predissociation of the trimer) leading to poorly
resolved structures in the velocity distribution [44]. The (H2 O)3 has a cyclic
structure with the global minimum having a configuration of “up-up-down” (uud)
which denotes the orientations of the free H atoms related to the plane defined by
three oxygen atoms [44]. The structure of the global minimum (uud) is included
in Figure 11.7. Two types of H-bonded O–H stretch fundamentals separated by
12–15 cm−1 are present for (H2 O)3 [45]. By exciting the H-bonded O–H stretch
at 3536 cm−1 , D0 of (H2 O)3 was determined for (H2 O)3 → H2 O + (H2 O)2 channel.
Since the IR excitation at 3536 cm−1 was not enough to dissociate the trimer into
individual monomers, the D0 for simultaneous breaking of all H-bonds in (H2 O)3
could not be determined experimentally. The trimer has two different H-bonded
OH stretches which lie within 12–15 cm−1 and within the experimental resolution,
these two frequencies could not be separated. Thus, the monomer fragment created
in the predissociation process was formed by breaking both types of H-bonds
in the trimer. Considering that the energy required to break a single H-bond of
(H2 O)2 is 1105 cm−1 and assuming the pair-wise additive interaction, the energy
required to break two H-bonds in (H2 O)3 is expected to be (1105 × 2) = 2210 cm−1 .
upp
uud uuu
upb
upd upu ppp
Figure 11.7 Low lying stationary points of (H2 O)3 . Source: Ch’ng et al. [44]/with
The experimental binding energy D0 for (H2 O)3 (for the dissociation to (H2 O)2
and H2 O monomer) was determined to be 2650 ± 150 cm−1 which is close to the
calculated value of 2726 ± 30 cm−1 obtained by Bowman and coworkers using
diffusion Monte Carlo calculation on an ab initio full-dimensional PES. The dif-
ference in the expected value and the experimental value of D0 for (H2 O)3 points
toward a cooperative or nonadditive contribution or many-body contributions
of 450–500 cm−1 to the binding energy in (H2 O)3 . The calculated value of D0
for breaking all three H-bonds in (H2 O)3 is 3855 ± 20 cm−1 , ∼1129 cm−1 greater
than the D0 for breaking two H-bonds in (H2 O)3 . As pointed out by Reisler and
coworkers, the difference in D0 for the two pathways, i.e. the dissociation of the
trimer into individual monomers, and into the dimer and monomer, is close to the
experimental value of 1105 cm−1 for the D0 of (H2 O)2 [23]. Hence, this difference
infers that the cooperative effect is operative in the cyclic structure of (H2 O)3 and
is thus captured in the (H2 O)3 → H2 O + (H2 O)2 dissociation channel. While the
translational energy distributions of the cofragments monitored in the VMI studies
of the dimer had a well-resolved structure, the velocity distributions monitored
in the dissociation of the trimer were smooth and did not have a well-resolved
structure due to the higher density of states in (H2 O)2 (Figure 11.8). Such smooth
structure in the velocity distribution led to lower accuracy in the determination of
D0 for the trimer.
In a similar experiment with the cyclic trimer of (HCl)3 , where it was possible
to break H-bonds trimer simultaneously (upon the excitation of the H–Cl stretch),
D0 could be determined for two channels following the vibrational predisso-
ciation: the first channel is (HCl)3 → HCl + (HCl)2 and the second channel is
(HCl)3 → 3HCl [46]. Experimental dissociation energies for the former and the
latter channel were determined as 1142 ± 20 and 1545 ± 10 cm−1 in excellent
agreement with two sets of calculated values including anharmonic zero-point
energies with many body PES, and complete basis set calculations. Using the
Ionization 800 Figure 11.8 Schematic depicting

continuum the higher density of states for the
dimer photofragment created in
the vibrational predissociation of
~
C 600 (H2 O)3 . Source: Ch’ng et al.
[44]/with permission of American
REMPI Chemical Society.
Energy (cm–1)
400
EExcess
~
ν=0–X
hνIR 200
D0
ν=0 0
(HCl)2 D0 determined by Valentini and coworkers [27], the cooperativity con-

tribution to (HCl)3 trimer energy was determined to be ∼250 cm−1 . As Reisler
and coworkers point out, such a large contribution of cooperativity to the D0
for the dissociation of the trimer to monomer and dimer for both (H2 O)3 and
(HCl)3 (19% and 22% for H2 O and HCl, respectively) is surprising considering
that the dimer geometry is significantly distorted in the trimer. Unlike (H2 O)3
and (HCl)3 , accurate determination of D0 is yet to be performed for (NH3 )3 . An
estimate of D0 for (NH3 )3 dissociating into the monomer (NH3 ) and dimer (NH3 )2
channel indicates that the D0 for this channel is twice the value of Do for the
dimer thus indicating a large cooperative contribution similar to (H2 O)3. and
(HCl)3 . The dissociation energy of (NH3 )2 measured by IR action spectrum is
listed in Table 11.1 [47]. Recently, the D0 for the tetrameric complex HCl–(H2 O)3
was determined for the dissociation channel HCl–(H2 O)3 → H2 O + HCl(H2 O)2
by Reisler’s group [48]. The complex HCl–(H2 O)3 has two O–H· · ·O H-bonds,
one O–H· · ·Cl H-bond and one Cl–H· · ·O bond. The D0 for the channel was
determined by exciting the O–H stretch fundamental of H2 O and then monitoring
the H2 O photofragments by the REMPI excitation scheme. The D0 for the channel
HCl–(H2 O)3 → H2 O + HCl(H2 O)2 was determined as 2400 ± 100 cm−1 . In a later
experiment by the Reisler’s group, the D0 for HCl–(H2 O)3 → HCl + (H2 O)3 was
determined as 2100 ± 300 cm−1 [49].
Not only do the VMI experiments help shed light on the cooperativity effect
or the many-body interaction contribution to the binding energy in H-bonded
complexes, but these studies also reveal the energy transfer pathways following
the predissociation. We will not discuss the energy transfer pathways in detail
as these are beyond the scope of this chapter but will include the C2 H2 · · ·HCl as
an example to illustrate the point [66]. The complex C2 H2 · · ·HCl has a T-shaped
geometry with a Cl–H· · ·π H-bond. Pair correlated product state distributions after
the predissociation were measured following two different excitation schemes:
(i) excitation in the asymmetric C–H stretch and (ii) excitation in the H–Cl stretch.
Before these studies, it was shown that the asymmetric C–H stretch fundamental
in the dimer is redshifted by 11 cm−1 from the monomer whereas the HCl stretch
is redshifted by 79 cm−1 in the complex [67, 68]. Hence, in the complex, the HCl
stretch is more strongly coupled to the H-bond whereas the coupling of acetylenic
C–H stretch to the H-bond is small. Following this fact, different vibrational
predissociation dynamics was observed upon the excitation of C–H asymmetric
stretch and H–Cl stretch in the dimer. While the C–H asymmetric stretch is excited,
vibrational predissociation occurs through initial vibrational energy transfer
from the C–H stretch to the C–C stretch and then subsequent coupling to the
intermolecular dissociation modes. Thus, the predissociation dynamics following
C–H asymmetric stretch excitation is dominated by the initial coupling of C–C
stretch to C–H stretch and the bending modes in the dimer. Excitation in the C–H
stretch of the acetylene cofragment is preserved during the predissociation and
the rest of the energy is distributed statistically among all other vibrational modes.
Excitation in the H–Cl mode led to a nonstatistical distribution of vibrational and
rotational energies in the acetylene cofragment with one or two quanta in the
bending modes. From the pair-correlated acetylene cofragment state distributions
following the acetylenic C–H stretch excitation and assuming the excitation of
one quantum of C–C stretch, the dimer dissociation energy was determined to be
700 ± 10 cm−1 .
11.2.2 Determination of Gibbs Free Energy of H-Bonded Complex

Formation By Infrared Spectroscopy
While the binding energy extracted from the molecular beam experiments as
described above serves as a benchmark for the theoretical calculations and sheds
light on many-body effects in smaller dimers and trimers, a knowledge of the
Gibbs free energy of the complex formation is essential for the determination of
equilibrium constant. Many atmospheric processes depend on the concentration of
the complexes which in turn depends on the equilibrium constant at the relevant
temperature of the atmosphere [69]. For a complex A⋅B, the formation equilibrium
is given by:
A+B⇋A⋅B (11.5)
The equilibrium constant (K) is the ratio of the forward and backward rate con-
stants and can be related to the pressures of the complex and the monomers as given
below:
PAB
K= × P⊝ (11.6)
(PA × PB )
where P⊝ is the standard pressure at 1 bar.
The free energy of complexation (ΔG⊖ ) is related to the equilibrium constant as
ΔG⊖ = − RT ln K and ΔG⊖ = ΔH ⊖ − TΔS⊖ . ΔH ⊖ is the enthalpy and ΔS⊖ is the
entropy of the complex formation. It is to be noted that at lower temperatures in
the molecular beam, different degrees of freedom (i.e. translation, rotation, and
vibration) of a H-bonded complex are cooled to different extent and the Boltzmann
distribution of population in these modes pertain to different temperatures. Hence,
in such a nonequilibrium situation, the measured binding energy cannot be directly
used to determine the ΔG⊖ for the complex formation. As the complex formation
from the monomers is accompanied by a decrease in entropy, the value of ΔG⊖
would increase as the temperature rises and accordingly, the equilibrium constant
decreases. Hence, as the temperature increases, measuring the equilibrium con-
stant for the complex formation becomes challenging because the complex signal
is much weaker than the monomer signal at higher temperatures. For example,
at 300 K, the value of equilibrium constant for (H2 O)2 formation is ∼0.01 favoring
the monomer [70]. In a H-bonded complex D–H· · ·A, the IR vibrational frequency
of D–H stretch is redshifted compared to the monomer along with a characteristic
intensity enhancement which is used as a signature of H-bond formation. The
characteristic redshift in the IR spectrum is related to the strength of D–H· · ·A bond
in the complex although because of the presence of other van der Waals interactions
including any secondary H-bonding interactions, ΔG⊖ values from these room
temperature measurements may not necessarily correlate with the redshift in the
D–H stretch. At room temperature, as the complex formation favors the monomer
over the dimer, the characteristic redshift in the D–H stretch frequency is often
weak to spot against the backdrop of strong monomer signal [70]. Although the
theoretical calculations of ΔG⊖ are viable, accuracy of such calculations rely
on the accuracy of low-frequency vibrations which is subject to the choice of
computational methods. Moreover, the use of rigid rotor approximation in the
calculation of partition functions also limits the accuracy of the calculations of
ΔG⊖ . Although there are several purely experimental methods that are used for the
determination of the equilibrium constants, discussion in this section is limited to
hybrid experimental and theoretical approach used by Kjaergaard and coworkers
[70–74]. Kjaergaard and coworkers use the experimental infrared intensities and
vibrational theory to determine ΔG⊖ near room temperature. Abundance of the
complex is determined from the ratio of the measured and calculated intensity
of the D–H stretching fundamental of the H-bond donor, which is then used to
determine the equilibrium constant K and ΔG⊖ . To determine the equilibrium
constant (K), pressures of the monomers and the complex are determined at a
given temperature. Known amounts of monomers are mixed and the IR spectrum
of the mixture is recorded. The reference IR spectra of the monomers are then
subtracted from the spectrum of the mixture to get the spectrum of the complex.
As the monomer pressure in the mixture can vary because of different reasons, the
monomer reference spectrum will have to be scaled before subtraction. The refer-
ence monomer spectrum is adjusted by a scaling factor such that near the isolated
monomer features in the mixture spectrum, the baseline is flat. The pressure of
the monomer in the mixture is obtained by multiplying the pressure that is used to
obtain the reference monomer spectrum by the scaling factor. The complex pressure
(PAB ) in the mixture is related to the intensity of a measured IR transition of the
complex and the latter is proportional to the integrated absorbance. Intensity of the
transition can be expressed as the dimensionless oscillator strength (f ) and is given

by the following expression:
( )
4me c𝜀0
f = A (11.7)
NA e2
me is the electronic charge, c is the speed of light, 𝜀0 is the vacuum permittivity, and
N A is Avogadro’s constant, A is integrated absorption coefficient and is related to the
pressure of the complex as:
c ln(10) RT
A= A( 𝜈)d̃
̃ 𝜐 (11.8)
Pl ∫
∫ A( 𝜈)d̃
̃ 𝜐 is the integrated absorbance and l is the optical path length. Combining
Eqs. (11.7) and (11.8), we obtain:
( )
4me c𝜀0 c ln(10)RT
f = A(̃v)d̃𝜐
NA e2 Pl ∫
T ∫ A(̃v)d̃𝜐
= 2.6935 × 10−9 (K−1 Torr m cm) (11.9)
Pl
Hence, pressure of the complex is given by:
T ∫ A(̃v)d̃𝜐
PAB = 2.6935 × 10−9 (K−1 Torr m cm) (11.10)
lfcalc
The integrated absorbance (∫ A( 𝜈)d̃̃ 𝜐) is obtained from the experimentally mea-
sured IR transition whereas fcalc is the theoretically calculated oscillator strength.
The oscillator strength is calculated as follows:
fv←0 = 4.702 × 10−7 cm D−2̃
𝜐v←0 |𝜇⃗v←0 |2 (11.11)
̃
𝜐 is the transition frequency in cm−1 , 𝜇⃗v←0 is the transition dipole moment,
⟨v←0
v|𝜇⃗v←0 | 0⟩ , in Debye (D). The dipole moment function (DMF) is expressed as a
Taylor series expansion in the displacement coordinates such that
∑ 1 𝛿 i 𝜇⃗ ||
𝜇(q)
⃗ = 𝜇⃗i qi with 𝜇⃗i = | (11.12)
i
i! 𝛿qi ||q=0
To calculate the oscillator strength, the matrix elements ⟨v |qn |0⟩ are calculated.
Unlike the equilibrium constants calculated by ab initio method, the calculated oscil-
lator strength shows less deviation based on the level of theory.
Using Eq. (11.10), the complex pressure (PAB ) is determined for various values of
the monomer pressures PA and PB . Next, PAB is plotted as a function of PA × PB which
is fit into a straight line passing through zero, thus indicating the formation of a 1 : 1
complex. The slope of the fitted line is multiplied by the standard pressure to yield
the equilibrium constant K.
Spectral intensities of the IR transitions in the complex spectrum are accurately
calculated using vibrational local mode (LM) theory combined with ab initio PESs
and DMFs [70, 75]. A full-dimensional calculation for the determination of the
intensity of the H-bonded X–H stretching vibration requires large sampling of the
PES and DMF, and a higher level of electronic structure theory in such cases is
expensive. Hence, instead of a full-dimensional NM model, a reduced dimension

LM model is used for the calculation of intensity [76, 77]. In the reduced dimension,
local oscillators are used which are the functions of internal curvilinear coordinates
like bond lengths and angles. In the LM model, the zeroth-order Hamiltonian
consists of the decoupled anharmonic oscillators and any coupling between these
anharmonic modes are treated as a perturbation. The mode of interest and other
modes which strongly affect the mode of interest are included in the LM model.
The simplest LM Hamiltonian is a one-dimensional (1D) Hamiltonian that consists
of only one single oscillator i.e. the hydrogen-bonded X–H stretch. Since the
hydrogen-bonded X–H stretch is much higher in frequency than other vibrational
modes in the complex, an adiabatic separation of this mode from the rest of
the vibrational frequencies is considered to be a good approximation. Hence, an
approximate form of the LM Hamiltonian considering a single X–H stretch is used.
For a single oscillator, the adiabatic 1-D LM Hamiltonian is given by:
̂ = 1 p̂ G(q, q)̂
H p + V(q) (11.13)
2
where q is the internal displacement coordinate, ̂ p is the momentum operator, V(q)
is the PES, and G(q, q) is the element of Wilson’s G matrix. Elements of the Wilson’s
G matrix are given by:
∑ 1 𝜕qi 𝜕qj
G (qi , qj ) = (11.14)
𝛼
m𝛼 𝜕x𝛼 𝜕x𝛼
where m𝛼 and x𝛼 are the atomic masses and the cartesian coordinates of each atom
respectively. By solving the 1-D LM Hamiltonian either analytically or numerically,
transition intensities in terms of dimensionless oscillator strengths are calculated.
It was observed that the 1-D LM model underestimates the vibrational frequencies
and the discrepancy increases as the H-bond strength, i.e. the redshift increases
(for a detailed discussion see Ref. [70] and the references therein). Accuracy of 1-D
LM model for the H-bonded complexes depends on the coupling of X–H mode
with other intramolecular and intermolecular vibrational modes. Discrepancy
in the vibrational frequencies is found to arise mainly from the coupling of the
intermolecular low-frequency vibration to the H-bonded X–H stretch. Although the
low-frequency vibrational modes are adiabatically separable from the H-bonded
X–H stretches, the effective potential of the low-frequency mode changes as the
geometry of the molecule changes upon vibrational excitation.
The local mode perturbation theory (LMPT) model, which involves modification
of the 1-D LM model through perturbative treatment of the intermolecular vibra-
tional modes, is used for the calculation of X–H stretch transition frequencies of
the donor [78]. The zeroth-order Hamiltonian includes the coupled anharmonic
intramolecular modes and the harmonic intermolecular modes, whereas the
first-order perturbation operator includes the coupling of the intramolecular
and intermolecular modes. An example of the LMPT model Hamiltonian which
incorporates two intermolecular modes of (H2 O)2 among the six intermolecular
modes in H2 O is given by Eq. (11.15) [70]:
∑ (1 1
) ∑∑
̂ LMPT = H
H ̂ LM + Gss p̂ 2 + Fss s2 + V(q, s) (11.15)
s=𝛽,𝜒
2 2 q s=𝛽,𝜒
In Eq. (11.15), q and s are the displacement coordinates for the intramolecular
and intermolecular modes, respectively. For example, if q is the O–H donor stretch
frequency in (H2 O)2 , then, s is the coordinate of the two intermolecular modes 𝛽
and 𝜒 in (H2 O)2 which results in partial breaking of the H-bond for both positive
and negative displacements [79]. Accuracy of the transition frequency calculated
by the LMPT model is improved over the frequencies calculated by LM model.
The required oscillator strengths using the LMPT model are calculated using
Response theory and the details are available in Ref. [70]. Accuracy of LMPT
model can be enhanced by incorporating other intermolecular or intramolecular
modes as required and also by including the anharmonic frequencies for the
intermolecular modes. For complexes where alcohol O–H is a H-bond donor, a
2-D LM Hamiltonian which includes the COH bending vibration is preferred,
whereas for complexes where H2 O is a H-bond donor, a 3-D LM Hamiltonian which
includes both HOH bending and the free O–H stretch is required. Thus, for the
complex of CH3 OH with dimethylamine, the 1-D LM Hamiltonian underestimates
the O–H stretch frequency by 100 cm−1 or more resulting in an overestimation of
the redshift. Inclusion of the COH bending vibration in the Hamiltonian (i.e. 2-D
LM Hamiltonian) improves the accuracy of the O–H stretch transition frequency
by 30 cm−1 , whereas including the intermolecular modes which partially break the
H-bond in a 2D+2D LMPT Hamiltonian increases the O–H stretch frequency by
up to 70–95 cm−1 . Accordingly, the oscillator strength, which is required for the
calculation of ΔG⊖ , decreases by ∼30% in the LMPT model as compared to the
value obtained in 1-D LM model [70]. Recently, it was shown by Kjaergaard and
coworkers that in addition to the inclusion of other intramolecular modes which
strongly influence the O–H/O–D stretch, use of molecule-fixed Eckart frame leads
to a more accurate prediction of O–H and O–D stretch frequencies of a series of
isolated alcohols [80].
Next, we discuss a recent representative study of the determination of ΔG⊖ of
water· · ·amine complexes by Kjaergaard and coworkers [74].
Water· · ·amine complexes are important in the context of atmospheric aerosol
nucleation. Accurate ΔG⊖ values of these complexes are required for the quan-
titative calculations of the aerosol nucleation process. In a recent experiment by
Kjaergaard and coworkers, the average value of ΔG⊖ was determined for both
H2 O· · ·(CH3 )2 NH (water· · ·DMA) and H2 O· · ·(CH3 )3 N (water· · ·TMA) com-
plexes. Both these complexes contain O–H· · ·N H–bonds as shown in Figure 11.9.
In the room temperature IR spectra, five distinct bands were observed for both
the complexes. The five bands observed in the spectra were HOH bending
overtone bend (2𝜈 HOH ), the H-bonded OH-stretching fundamental band (𝜈OHb ),
free O–H stretching fundamental (𝜈OHf ), combination band of O–H stretch-
ing fundamental and HOH bending (𝜈OHb + 𝜈HOH ), and the combination band
of free O–H stretching fundamental and HOH bending (𝜈OHf + 𝜈HOH ). One
approach to increase the accuracy of ΔG⊖ is to determine its value for multiple
bands of the complex. Determination of the intensities for different vibra-
tional bands reduces the chances of systematic errors in the transition intensity
calculations.
Figure 11.9 Water complexes with dimethylamine (DMA) and trimethylamine (TMA).
Source: Kjaersgaard et al. [74]/with permission of American Chemical Society.
For the measurement, at first, the IR spectrum of a mixture of the amine and
water was collected. Next, the reference monomer spectrum was subtracted from
the spectra of the mixture to get the spectrum of the complex and the subtraction
process is explained in Figure 11.10. The reference spectrum of the amine monomer
was collected in a separate experiment, whereas the reference spectrum of the water
(a) Mixture spectrum (H2O & TMA)
(b) TMA spectrum

Absorbance
(c) Mixture after TMA subtraction
(d) Simulated H2O spectrum
(e) Complex Spectrum
3100 3200 3300 3400 3500 3600 3700 3800 3900 4000
Figure 11.10 Room temperature infrared spectrum of H2 O· · ·(CH3 )3 N (water· · ·TMA)

obtained by subtracting the reference monomer spectra from the mixture spectrum. In the
specific experiment by Kjaergaard [74], the reference TMA spectrum was obtained in a
separate experiment, whereas the reference H2 O spectrum was simulated. Source:
Kjaersgaard et al. [74]/with permission of American Chemical Society.
35 35
ν*OHb ν*OHb
νOHf νOHf
30 H2O·DMA 30 H2O·TMA
νOHb + νHOH νOHb + νHOH
νOHf + νHOH νOHf + νHOH
Integrated absorbance (cm–1)
Integrated absorbance (cm–1)

25 25
× 50
20 20
× 50
× 50
15 15
× 50
10 10
5 5
0 0
0 2000 4000 6000 8000 0 2000 4000 6000 8000
PH O × PDMA (Torr2) PH O × PTMA (Torr2)
2 2
Figure 11.11 Linear dependence of the calculated integrated absorbance on the product
of monomer pressures. Source: Kjaersgaard et al. [74]/with permission of American
Chemical Society.
monomer was simulated. For each band in the spectrum, the integrated absorbance
was calculated and was found to increase linearly (Figure 11.11) with the product
of the monomer pressures thus confirming that each band is from a bimolecular
complex.
To calculate the transition wave numbers and intensities, a six-dimensional LM
Hamiltonian including the three intramolecular modes localized on the donor
H2 O and three intermolecular modes was constructed. The intramolecular modes
based on the H2 O donor mode included free O–H stretch, H-bonded O–H stretch,
and the HOH bending. The intermolecular modes included in the Hamiltonian
consisted of donor rocking, twisting and the H-bonded stretch, and other inter-
molecular modes were found to have negligible effect on the intramolecular
transitions of interest. For both H2 O· · ·(CH3 )2 NH and H2 O· · ·(CH3 )3 N complexes,
the H-bonded O–H stretch fundamental and the overtone of HOH-bending state
are nearly degenerate. The LMPT model cannot be used in such cases because
the intramolecular modes in the LMPT model are coupled before the inclusion of
the effect of the intermolecular modes and this approximation breaks down if the
intramolecular modes are degenerate. Hence, for these complexes, the intensities
were calculated by using a variational LM model. By solving the LM Schrodinger
equation, transition intensities as the dimensionless oscillator strength for each
band were calculated using the method described above. Using the values for the
oscillator strength and the integrated absorbance, the pressure of the complex was
calculated, and subsequently, ΔG⊖ was determined for each of the observed bands.
The best estimates of ΔG⊖ were determined as 5.0 ± 0.2 kJ mol−1 (418 cm−1 ) for
Table 11.2 Room temperature free energy of formation ΔGobs (cm−1 ) and the O–H stretch
redshift Δ𝜈̃ (cm−1 ) and the H-bond distances (R) for the complexes of alcohol with dimethyl
ether, dimethyl sulfide, and dimethyl selenide.
̃ (cm−1 )
(𝚫𝝂) 𝚫G ⊖ (cm−1 ) R (H· · ·A)Å
Dimethyl ether· · ·t-butyl alcohol 99 652 ± 75 1.89

Dimethyl sulfide· · ·t-butyl alcohol 107 844 ± 92 2.39
Dimethyl selenide· · ·t-butyl alcohol 104 803 ± 159 2.54
Dimethyl ether· · ·2,2,2-t-fluoroethanol 198 217 ± 84 1.79
Dimethyl sulfide· · ·2,2,2-t-fluoroethanol 201 552 ± 75 2.28
Dimethyl selenide· · ·2,2,2-t-fluoroethanol 197 552 ± 92 2.46
H2 O· · ·(CH3 )2 NH and 3.8 ± 0.2 kJ mol−1 (318 cm−1 ) for H2 O· · ·(CH3 )3 N with an
estimated error of 0.4 kJ mol−1 (∼33 cm−1 ) for ΔG⊖ .
In similar experiments by Kjaergaard and coworkers, room temperature ΔG⊖
for the H-bonded complexes of alcohols (t-butyl alcohol and 2,2,2-t-fluoroethanol)
with dimethyl ether, dimethyl sulfide, and dimethyl selenide complexes were
determined [81]. In these complexes, alcohols act as H-bond donors, and the
associated redshift and the ΔG⊖ values are listed in Table 11.2. While the redshift
in the O–H stretch frequency indicates the strength of the H-bond interaction,
the ΔG⊖ values correspond to the overall strength of the complex including
any secondary interactions, if present in the complex (see the introduction for
the discussion). The redshift is higher for 2,2,2-t-fluoroethanol, the stronger
H-bonded donor and smaller for the complexes of t-butyl alcohol, the weaker
H-bond donor. Experimental redshifts are similar for the complexes of dimethyl
ether, dimethyl sulfide, and dimethyl selenide complexes of t-butyl alcohol and
2,2,2-t-fluoroethanol, suggesting that the oxygen, sulfur, and selenium atoms have
similar H-bond acceptor strength. However, the Gibbs free energy shows a different
trend. Although the room temperature values of ΔG⊖ is similar for the complexes of
t-butyl alcohol, the free energy for dimethyl ether· · ·2,2,2-t-fluoroethanol is much
smaller than those of the complexes of 2,2,2-t-fluoroethanol with dimethyl sulfide
and dimethyl selenide indicating a higher stability for the former. Interestingly,
in a recent experiment by Kjaergaard and coworkers, the authors studied a series
of H-bonded and deuterium bonded (D-bonded) complexes and from their ΔG⊖
values at room temperature concluded that the stability of the H-bonded complexes
and the corresponding D-bonded complexes is similar at higher temperatures and
attributed it to the higher entropy of formation of the H-bonded complexes at higher
temperatures [82].
11.2.3 Measurement of Binding Energy of H-Bonded Complexes

by IR–UV Double Resonance Spectroscopy
Two-color pump–probe techniques such as IR–UV double resonance spec-
troscopy have been used to measure the binding energy of H-bonded complex
9-hydroxy-9-fluorenecarboxylic acid (9-HFCA) with water. The details of the

experimental measurements performed by J.L. Knee’s group using IR–UV double
resonance spectroscopy are included in Ref. [83]. Briefly, the experiment involves
three different steps. In the first step of the experiments, the mass-resolved REMPI
spectrum of the complex or the monomer was acquired by scanning the UV laser.
The UV laser pulse used for the REMPI experiment intersected the molecular
beam within a TOF mass spectrometer. The ions produced in the REMPI excitation
were detected by a microchannel plate (MCP) detector. In the next step, IR action
spectrum in the O–H stretch region of the complex was collected by tuning the
UV laser at a fixed wavelength while the IR laser was scanned. The UV laser
was tuned into one of the transitions of either the monomer or the complex and
the ion signal was monitored using the REMPI technique. As the IR laser is
introduced before the UV laser, as soon as there is a resonant transition in the
IR, a corresponding depletion in the REMPI spectrum was observed. Sensitivity
in the depletion spectrum was enhanced by taking the difference between the
IR on and IR off signals for multiple laser shots. The final experiment involves
scanning the UV laser while the IR laser was tuned to a specific wavelength in
the complex spectrum and the dissociation of the complex to the monomer was
probed by monitoring the REMPI spectrum in the monomer channel. As shown in
Figure 11.12, hot bands due to carboxylic acid group rocking motion were observed
in the S0 → S1 vibronic spectrum of the monomer. Zero-electron kinetic energy
(ZEKE) spectroscopy of the acid monomer confirmed the vibrational frequency for
the rocking motion as 67 cm−1 . As the IR laser was tuned to the monomer peaks at
3207 (carboxylic acid O–H stretch) and 3145 cm−1 , hot bands corresponding to the
monomer were observed in the REMPI spectrum. For the frequency at 3207 cm−1 ,
three peaks at 32759, 32751, and 32 742 cm−1 were observed corresponding to
the hot bands arising from the states v = 1, v = 2, and v = 3, respectively. As the
laser was tuned to 3145 cm−1 , only the hot bands corresponding to v = 1 and
v = 2 were observed and the band for v = 3 could not be observed. From these
Figure 11.12 REMPI –1

spectrum of 9HFCA IR at 3207 cm 0
–1
monomer. The signals were IR at 3145 cm
collected after the excitation
at two different IR 1
2
REMPI (arb.units)
frequencies. Source: Gu and 3

Knee [83]/with permission of
AIP Publishing.
32 700 32 720 32 740 32 760 32 780 32 800

–1
S0 → S1 : cm
observations, the binding energy of the complex was determined to be between

[3145 − (67 × 2)] = 3011 and [3145 − (67 × 3)] = 2944 cm−1 . A slightly lower value
for the upper limit of 3006 cm−1 was determined for the IR excitation at 3207 cm−1 .
Based on these observations, dissociation energy of 2975 ± 30 cm−1 was determined
for the 9-hydroxy-9-fluorenecarboxylic acid· · ·H2 O complex.
11.3 Determination of the Binding Energy of H-Bonded

Complexes Using Spectroscopic Techniques Involving
Electronic Excitation
The focus of this section is to describe the measurement techniques which utilize
the transitions in Ultraviolet (UV) regions for the determination of D0 . One of the
requirements for the complex is thus at least one of the binding partners in the
complex is a UV chromophore. These techniques involving electronic transitions of
the molecules explicitly rely on the Frank–Condon (FC) overlap of different elec-
tronic states of the molecules and thus, have their limitations when it comes to
the molecules of choice. Several of these techniques include exciting the complex
M–S (where M is the chromophore) near its ionization threshold via an interme-
diate state and thus, the dissociation energies of complexes having one partner as
the aromatic molecule can be conveniently studied by these techniques because the
aromatic molecules have one electronic state located midway from their ionization
potential. Hence, many complexes whose D0 has been measured using electronic
excitation involve an aromatic molecule as one of the binding partners.
11.3.1 Determination of H-Bond Dissociation Energy Through

Multiphoton Ionization Techniques
In this section, we focus on techniques based on laser ionization coupled with mass
spectrometry and also outline the principles with examples from literature. In mul-
tiphoton ionization (MPI) experiments, the dissociation energies of both complex
and the complex cation can be determined in cases where the adiabatic ionization
energies (IEad ) of the monomer and the complex are known. For a complex M–S,
the appearance potential or the appearance energy of the fragment cation M+ can
be measured by experiments and using the known value for the adiabatic ionization
energy (IEad ) of M+ , the dissociation energy of the neutral cluster is calculated using
the following relation:
D0 = AE (M+ − S → M+ + S) − IEad (M → M+ ) (11.16)
In addition, if IEad of the complex is also known, the dissociation energy of the
ionic complex (E0 ), i.e. the complex cation is determined from the following relation:
E0 = AE (M+ − S → M+ + S) − IEad (M − S → M+ − S) (11.17)
In cases where the IEad of the parent neutral molecule cannot be measured (i.e. if the
measured ionization potential is nonadiabatic), only the lower limits of the dissoci-
ation energy can be obtained from the experiments. Hence, the experiments provide
11.3 Determination of the H-Bond energy through electronic excitation 317
the threshold energy for the parent cluster ion fragmentation and the D0 for the
neutral cluster is obtained from the energy balance. Determination of D0 from these
multiphoton resonance ionization experiments is feasible only where the fragmenta-
tion limit resides within the FC distribution of vibrational levels of the ionic complex
fragment.
At the beginning of this section, we describe a pioneering study by Neusser and
coworkers, conducted during late 1980s where the authors determined the dissocia-
tion energy of a prototypical weakly bound complex (C6 H6 )2 [84]. Benzene dimer is
the prototypical example of C–H· · ·π interaction in gas phase and its structure was
determined as T-shaped by Microwave spectroscopy by Arunan and Gutowski [85].
The parallel-displaced π-stacked structure of (C6 H6 )2 , which does not have a dipole
moment, is yet to be detected in gas-phase despite having similar binding energy as
the T-shaped structure (only 0.17 kcal mol−1 or 60 cm−1 less stable than the T-shaped
structure) [86]. The determination of D0 of (C6 H6 )2 is elaborated here as a represen-
tative example of the technique. During late 1980s, Neusser and coworkers prepared
neutral benzene clusters (C6 H6 )n (n ≤ 40) in a supersonic beam and determined the
first experimental value of D0 for the benzene dimer and benzene dimer cations [84].
Benzene clusters up to n = 40 were prepared in a supersonic expansion and were ion-
ized by a two-photon absorption method within the acceleration region of a reflec-
tion time-of-flight (RETOF) mass spectrometer. During the experiment, the photon
energy was chosen such that the generated clusters decay slowly on a microsec-
ond time scale. The RETOF mass spectrometer was used to monitor the dissociation
pathways of the clusters near the threshold. A RETOF mass spectrometer consists of
an acceleration region, a field-free drift region, the reflector, and an ion detector. For
the detection of all cluster cations simultaneously by the mass spectrum, the laser
was operated off-resonance thus making the ion signal weaker by one order of mag-
nitude. Within the laser focus, the ions are produced by rapid fragmentation or fast
ionization and these ions are termed as “stable” ions. In addition to the stable ions,
there are metastable ions that dissociate into the daughter ions in the drift region.
The daughter ions which are produced in the drift region have a different kinetic
energy than the ones produced in the laser focus regions. Typically, in a RETOF
mass spectrometer, any difference in flight time due to differing kinetic energy is
corrected by choosing different reflector voltages. To detect the ions produced from
the dissociation of metastable ions in the drift region, the time focusing was used in
“partial correction” mode. Thus, in case of a metastable ion M+ dissociating into a
daughter ion m+ in the drift region, the resulting daughter ion will have a different
kinetic energy than the ions which were created in laser focus region. In a “partially
corrected” mode, voltages on the reflector plates are varied such that the masses orig-
inating from the dissociation of the metastable ions appear between the stable ion
peaks of masses M and m. Strong signals for the ions produced by the dissociation
of metastable ions were observed for n = 1–8 as shown in Figure 11.13. By tweaking
the experimental conditions (by varying the reflector potential), mass spectrum con-
sisting of peaks only from the metastable fragmentations was acquired and is shown
in Figure 11.13. For the clusters of size n = 8, the signals decaying in the drift region
were found to be greater than the number of ions reaching the detector indicating an
n=1 +
(C6H6)n
5
2
× 10
6
3 7 8
(a) 0 200 400 600
2
3 4
5
6 7 8
0 200 400 600
(b) Mass (amu)
Figure 11.13 (a) Partially corrected multiphoton mass spectrum of (C6 H6 )n showing both
stable and metastable ion peaks. The clusters were ionized by two-photon ionization at
259.4 nm. Marked peaks in the spectrum indicate the stable ions peaks and the arrows
point from the peaks corresponding to the stable ions to the peaks obtained from the
metastable decay. (b) Mass spectrum showing the signal for ions produced by the
dissociation of metastable ions. Small asymmetric peaks in the spectrum are due to the
decay in the acceleration region of RETOF. Source: Kiermeier et al. [84]/with permission of
efficient dissociation of the metastable ions for larger clusters implying a decrease
in the dissociation threshold. For the metastable ions in the drift region, the loss of
one neutral benzene molecule as shown below was found to be the dominant loss
mechanism.
(C6 H6 )+n → (C6 H6 )+n−1 + C6 H6 (11.18)
From these measurements by Neusser and coworkers, dissociation energies for both
neutral (D0 ) and cationic (E0 ) benzene dimer were determined by measuring the
Relative metastable intensity (%)
+ +
(C6H6) → C6H6 + C6H6
2
10
–1
(2hν)0 = 75 120 (cm )
5
0
77 000 76 000 75 000
Two photon energy (cm–1)
Figure 11.14 Plot showing the experimental signal intensity (points) of C6 H6 + (normalized
to the parent ion signal) arising from the dissociation of metastable (C6 H6 )2 + ions in the
drift region of RETOF vs the two-photon energy. The fitted line is shown as a black solid
line. The dissociation threshold of the cationic dimer (C6 H6 )2 + is determined from these
measurements. Source: Kiermeier et al. [84]/with permission of American Chemical Society.
“breakdown curve”. The breakdown curve was generated by monitoring the extent of
metastable fragmentation with decreasing photon energy. As the two-photon energy
decreases, the excess energy above the ionization threshold also reduces. To measure
the dissociation energies of the neutral and cationic dimers, conditions in the super-
sonic molecular beam were adjusted to maximize the concentration of the monomer
and dimer cations in the beam. Normalized intensities of the ions created in the
metastable fragmentation of (C6 H6 )2 + vs the two-photon energy were plotted and
fit into a straight line as shown in Figure 11.14. From the intercept of the fitted
line with the X-axis, the value for the dissociation threshold of the cationic dimer
(C6 H6 )2 + was derived. Linear extrapolation of the fitted straight line as shown in
Figure 11.14 provided the dissociation threshold value (2hν0 ) for (C6 H6 )2 + (or the
appearance energy of the monomer cation) as 75 120 ± 160 cm−1 . Relation between
the appearance energy and the dissociation energies of the dimer in the cationic state
and the neutral states is shown in Figure 11.15. Binding energies of the neutral (D0 )
and cationic (E0 ) dimers are derived from the relations:
D0 = (2h𝜈)0 − IP (monomer) (11.19)
E0 = (2h𝜈)0 − IP (dimer) (11.20)
Using the ionization potential of C6 H6 and (C6 H6 )2 as 9.243 eV and 8.86 eV,
respectively, D0 and E0 were calculated as 70 ± 20 meV and 448 ± 20 meV (565 and
3614 cm−1 ), respectively.
In two-color photodissociation experiments, such as during two-color REMPI
(2C-REMPI) measurements, at first, the complexes (M–S) are selectively excited
by the first laser and the second laser is scanned over the dissociation threshold of
the complex cation (M–S+ ). Subsequently, the dissociation of the parent complex
cation (M–S+ ) into the fragment cations (M+ ) is monitored by mass spectrometric
+ Figure 11.15 Schematic showing the

C6H6 + C6H6
relation between the appearance
energy (2h𝜈 0 ) and the dissociation
energies of (C6 H6 )2 in the cationic and
E0
the neutral states. The ionization
potential is adiabatic. The scheme was
+ used as a model for the determination
[C6H6]
2 of D0 for the dimer. Source: Kiermeier
et al. [84]/with permission of American
Chemical Society.
(2 hν)0
IPC6H6
IP[C H ]
6 6 2
B0
C6H6 + C6H6
[C6H6]
2
techniques. The signal in 2C-REMPI experiments depends on the cumulative sum

of FC accessible vibration levels of the ions having energy less than the photon
energy. As the photon energy is increased, the parent stable ion signal increases
until the dissociation limit is reached beyond which a plateau is observed. Beyond
the dissociation limit, all parent complex ions dissociate into the fragment cation
M+ and the onset of the fragment signal is marked as the fragmentation threshold
or the appearance potential of the M+ ion. The 2C-REMPI technique was used
by Mons and coworkers to determine the D0 of several H-bonded complexes and
those of the corresponding cations including the dimer of phenol with H2 O and
CH3 OH [87, 88]. In the complexes with H2 O and CH3 OH, phenol acts as a H-bond
donor and forms O–H· · ·O hydrogen bond with the oxygen atoms of the binding
partners. For these sets of complexes, the ionization process is nonadiabatic and
the schematic for the experimental principle is shown in Figure 11.16. During
the experiment, the complex ion is excited to a highly vibrationally excited state
close to the dissociation threshold. Either the intermolecular mode in the complex
ion is directly excited or the intramolecular mode is excited with subsequent
distribution of vibrational energy into the intermolecular mode. Excitation of
both intramolecular or intermolecular modes depends on the FC overlap between
the neutral S0 or S1 state with the ground state of the ion. Mons and coworkers
measured the appearance potential or the fragmentation threshold of phenol+
fragment ion for C6 H5 OH· · ·H2 O and C6 H5 OH· · ·CH3 OH complexes. From the
ionization energy of phenol monomer, the best estimates of the dissociation
energy of C6 H5 OH· · ·H2 O and C6 H5 OH· · ·CH3 OH complexes were determined as
5.6 ± 0.11 and 6.11 ± 0.18 kcal mol−1 (1960 and 2139 cm−1 ), respectively.
Figure 11.16 Experimental

schematic used for the
determination of D0 of
C6 H5 OH· · ·H2 O and Energy M+(v) – S
C6 H5 OH· · ·CH3 OH by Mons and
coworkers. The ionization process Fragmentation
+
is nonadiabatic. Source: Courty M +S
et al. [88]/with permission of
American Chemical Society. Ion
M+ – S
+
AP(M )
Franck− IP(M)
Condon IP(M – S)
zone for
S1
intermol.
modes
M+S
D0 S0 Neutral
state
O
M–S {q1}
The two-color two-photon resonant ionization techniques were also used for
the determination of D0 of X–H· · ·π H-bonded complexes of C6 H6 with H2 O,
NH3 [89, 90]. The complexes of C6 H6 with H2 O, NH3, and HCl are π–H-bonded
as determined by microwave spectroscopy where the C6 H6 π-cloud acts as the
H-bond acceptor and accepts hydrogen from H2 O, NH3 , and HCl [91–93]. Dis-
sociation energies of both C6 H6 · · ·H2 O and C6 H6 · · ·NH3 were determined as
2.44 ± 0.09 and 1.84 ± 0.11 kcal mol−1 (854 and 634 cm−1 ), respectively whereas a
bracket of D0 for C6 H6 · · ·HCl in the range of 1.8 kcal mol−1 ≤D0 ≤ 3.8 kcal mol−1
(630 cm−1 ≤ D0 ≤1330 cm−1 ) was obtained by Zwier and coworkers from the analysis
of the dispersed fluorescence spectrum [94].
Mass-selected two-color MPI technique has also been used to measure the
dissociation energies of clusters having C–H· · ·π bonds such as C6 H6 · · ·HCCH and
C6 H6 · · ·C2 H4 clusters [95]. The clusters were prepared using the molecular beam
technique and were then pumped into the S1 –S0 vibronic band of the cluster. After
the excitation of the cluster to the S1 state by the first laser (𝜈 1 ), the clusters were
successively ionized by the second laser (𝜈 2 ). The generated ions were first selected
by a mass spectrometer and then subsequently detected by a multichannel plate
detector. The dissociation threshold of the cluster fragment cation was identified by
scanning the second laser wavelength (𝜈 2 ) over the threshold when a corresponding
increase in the fragment ion intensity was observed. The generated ions were
extracted and accelerated to a TOF mass spectrometer. Before collecting the two
color multi photon ionization spectrum, 1-C multi photon ionization spectrum of
the complex was recorded and analyzed. The 1-C multi photon ionization spectrum
of C6 H6 · · ·C2 H4 recorded in S1 –S0 610 region shows that the 610 band of the complex
appears at 38 542 cm−1 and the band is 64 cm−1 low-frequency shifted compared to
bare benzene. The 610 band of the cluster is accompanied by a long progression in
the intermolecular vibration with a harmonic frequency of 23 cm−1 and is attributed
to either a torsional or bending motion in the complex. The two-color-multi photon
ionization spectrum of C6 H6 · · ·C2 H4 collected while monitoring the benzene and
C6 H6 · · ·C2 H4 cations is plotted against (𝜈 1 + 𝜈 2 ) is shown in Figure 11.17. The
spectrum was recorded by keeping the 𝜈 1 wavelength fixed at the combination
band 38 631 cm−1 (+89 cm−1 from S1 –S0 610 band) while 𝜈 2 was scanned. As the
total excitation energy crosses the dissociation threshold of the cluster cation,
the fragment ions are observed in the spectrum which rises with an increasing
excitation energy. Above the dissociation threshold, while monitoring the parent
cluster cation, increasing the excitation energy contributes only to the fragmented
channel and a plateau is observed beyond the dissociation threshold. However, as
(Benzene−ethylene)+
(a)
+
Benzene
74 990
74 400 74 600 74 800 75 000 75 200 75 400

Figure 11.17 Mass-selected two-color multi-photon ionization spectrum of C6 H6 · · ·C2 H4

recorded by exciting at 38 631 cm−1 . Signals were measured for both cluster cation and the
fragment ion channels. These experiments are used to determine the dissociation threshold
for the cluster cation. Source: Shibasaki et al. [95]/with permission of American Chemical
Society.
can be seen in case of C6 H6 · · ·C2 H4 spectrum shown in Figure 11.17, rise of the
signal for the fragment ion and the start of the plateau do not occur at the same
excitation energy. The reason was attributed to longer dissociation time of the
cluster cation just above the dissociation threshold. Thus, some of the parent cluster
cations which survive the extraction region contribute to the parent cluster cation
signal. From these measurements, the dissociation threshold (appearance energy of
C6 H6 + ) of (C6 H6 · · ·C2 H4 )+ was determined to be in the range of 74 990 ± 50 cm−1 .
Equation (11.19) is now modified to include the field correction term (ΔE)
D0 = (2h𝜈)0 − IP (monomer) + ΔE (11.21)
From the observed value of wavelength (𝜈 1 + 𝜈 2 ) at the dissociation threshold (74 990
±50 cm−1 ), using the field correction term ΔE = 60 cm−1 and from the adiabatic
ionization potential of benzene monomer as 74 556 cm−1 , the dissociation energy of
C6 H6 · · ·C2 H4 in the neutral ground state was determined as D0 = 490 ±60 cm−1 . In
a similar experiment, the dissociation energy of C6 H6 · · ·C2 H2 was measured as D0
= 930±60 cm−1 .
Mass-analyzed threshold ionization (MATI) technique has been used for the
determination of the dissociation energies of complexes including C6 H6 · · ·CH4
complex [96]. The MATI method is a variation of ZEKE spectroscopic technique.
In ZEKE, the system is excited to a high-n Rydberg state following which a pulsed
electric field is applied after certain delay to ionize the Rydberg state. These high-n
Rydberg states reside within a very narrow band just below the ionization threshold
of each ionic state. The ZEKE spectrum is obtained by scanning the excitation laser
over the ionization threshold and by subsequently measuring the yield of electrons
generated by the delayed pulsed field ionization. In MATI technique, instead of the
photoelectrons, the ions produced by pulsed field ionization is analyzed by mass
spectrometry thus facilitating their unambiguous detection. The D0 of C6 H6 · · ·CH4
complex was measured by MATI technique [96]. Initially, a supersonic molecular
beam of neon seeded by benzene and methane was prepared and was introduced
into the ion extraction stage of the TOF mass spectrometer. Two-color two-photon
′
(1 + 1 ) techniques were used to pump the complex to a Rydberg state just below
the ionization threshold of the cluster via S1 61 level of benzene monomer. The ions
which were produced by fast autoionization or direct ionization (prompt ions) were
spatially separated from the neutral Rydberg states by applying a weak electric field.
A pulsed acceleration field was applied to ionize the long-lived Rydberg states after
7 μs from the laser excitation and the generated ions were extracted to the TOF tube.
Both prompt ions and the cations generated by the pulsed field ionization (MATI
ions) were mass analyzed and the flight time of the spatially separated MATI ions
and the prompt ions are different and thus can be distinguished temporally. While
acquiring the MATI spectrum, 𝜈 1 is fixed, whereas 𝜈 2 is scanned. The second laser
takes the cluster from the S1 state to the Rydberg state. In the Rydberg state, the
Rydberg electron is located far away from the ion core and thus, the dissociation
threshold of the Rydberg state is same as the cluster cation. When the vibrational
energy of the ion core in the Rydberg state surpasses the dissociation threshold,
the appearance channel in the MATI spectrum switches from the parent cluster
[+675]
*
Benzene−CH4
(c) (fragment channel)
[+640]
74 276 [0]
Benzene−CH4
(b) (parent ion channel)
1
(74 556) [0] 6 (±3/2)
16161(±3/2)
(a) Bare benzene 61(±1/2) 62(±1/2)
1 1
1 1161(±3/2)
4
0 200 400 600 800 1000 1200 1400

Wavenumber (cm–1)
Figure 11.18 (a) MATI spectrum recorded for the benzene monomer excited via S1 61 level
of benzene monomer. MATI spectrum recorded for the complex in the (b) parent ion channel
and in the (c) fragment ion channel excited via S1 61 level of benzene monomer through
′
two-color two-photon (1 + 1 ) techniques. Source: Shibasaki et al. [96]/with permission of
cation to the daughter fragment ion. For both parent and daughter fragment cation
channels, the MATI spectrum was measured via the intermediate S1 61 state.
Because of the difference in the energy of the applied field used for the ionization
of the Rydberg state and the actual ionization potential of the cluster, the ionization
potential measured by MATI spectroscopy will have to be subjected to a correction
in order to be extrapolated to the field free value. Field free IP0 of C6 H6 · · ·CH4 was
determined as 74 276 cm−1 . The field-free ionization potential (IP0(monomer cation) )
of benzene monomer was determined as 74 556 cm−1 . Limits on the dissociation
energies were determined from the MATI spectrum shown in Figure 11.18. Charac-
teristic feature of the MATI spectrum is the switching of the appearance potential
with the increase in the vibrational energy. Thus, for the parent cation channel, no
peak was observed above 675 cm−1 and whereas for the fragmented channel only the
bands higher than 640 cm−1 was observed in the MATI spectrum. Hence, 675 cm−1
is considered the upper bound of the dissociation energy (of the cluster cation)
whereas 640 cm−1 is the lower bound for the same. Switching of the spectrum from
the parent cation channel to the fragment cation channel is a characteristic of the
MATI spectrum and thus this technique provides a precise bracketing of the dissoci-
ation energy. As described previously, the dissociation energy of the neutral cluster
is evaluated from the ionization energy of the monomer fragment and from the
measured dissociation threshold of the cluster cation using the following expression:
D0 (S0 )(neutral cluster) + IP0(monomer cation) = D0 (cluster cation) + IP0(cluster). (11.22)
From the range of the dissociation energies measured above, the dissociation D0 (S0 )
for the neutral C6 H6 · · ·CH4 cluster is estimated in the range of 360–395 cm−1 .
11.3.2 Determination of the Dissociation Energy of Cationic H-Bonded

Complexes Through Birge–Sponer Extrapolation
The Birge–Sponer (BS) extrapolation technique has been long known as a method
for the determination of the dissociation energies of diatomic molecules and applies
to systems where the vibrational energy levels converge at a finite value of the vibra-
tional quantum number [97, 98]. The extrapolation method was discussed in detail
by Gaydon for the ground and excited states of different classes of molecules includ-
ing the molecules with ionic bonds [97]. Recently, the BS linear extrapolation was
used by Wategaonkar and coworkers for the estimation of D0 of the cationic com-
plexes with O–H· · ·S interactions via ZEKE spectroscopy [99, 100]. The BS extrap-
olation is an indirect method and in cases where the successive difference between
the experimentally observed bands (either electronic bands approaching the ion-
ization limit or the vibrational bands within a given electronic state) decreases, the
dissociation energy of a system can be estimated from a linear extrapolation to the
convergence limit. Neglecting the terms beyond the quadratic terms in the vibra-
tional energy level expression, successive separations between the vibrational energy
levels (Δ𝜀) in cm−1 are given by:
Δ𝜀 = 𝜀v+1 − 𝜀v = [−2𝜔e xe ]v + [𝜔e − 2𝜔e xe ] (11.23)
where 𝜔e is the oscillation frequency and 𝜔e xe is the anharmonicity constant.

Hence, a plot of Δ𝜀 vs. v is linear and thus, from the slope and intercept of the
plot, the values of 𝜔e and 𝜔e xe can be determined. At the dissociation threshold,
the separation between the successive energy levels is assumed to be zero and
thus, by equating the left-hand side of Eq. (11.23) to zero, the value of vmax for
the uppermost vibrational level is determined. The dissociation energy D0 is thus
determined from the value of vmax . Wategaonkar and coworkers were the first to
apply BS linear extrapolation in the determination of D0 of p-fluorophenol· · ·H2 S
cationic complex from its ZEKE spectrum [99]. The photoionization efficiency
curve for p-fluorophenol and the complexes p-fluorophenol· · ·H2 S were recorded
by scanning the ionization laser while the excitation laser was fixed at S1 –S0
band origin. The IEad (i.e. the adiabatic ionization energy) for the complex
p-fluorophenol· · ·H2 S was determined as 68 542 ± 2 cm−1 although the slow rise
of the signal in the photoionization efficiency curve made the determination
of its ionization threshold difficult. For ZEKE spectroscopy, an excitation laser
pumps the molecule to an intermediate state, whereas the probe laser excites the
system to a Rydberg state. In the Rydberg state, the electron is loosely bound to the
ionic core and a weak electric field causes the ionization of the Rydberg neutral
states. The ionic core resembles the cationic state and by scanning the probe laser
over the Rydberg manifold, the long-lived Rydberg states close to the ionization
threshold are identified and the associated rovibrational levels of the cations are
thus probed. Wategaonkar and coworkers recorded the ZEKE spectrum of the
complex p-fluorophenol· · ·H2 S via the S1 –S0 band origin (S1 000 ) and an extended
ZEKE spectrum of the p-fluorophenol· · ·H2 S complex recorded beyond 2400 cm−1
above its IEad is shown in Figure 11.19. The ZEKE spectra were collected for the
Intramolecular modes
σ and its overtones
(a) ZEKE via S1000
Intensity (arb.units)
(b) ZEKE via S1σ10
65 500 66 000 66 500 67 000 67 500

Two-photon energy (cm–1)
Figure 11.19 An extended ZEKE spectrum of p-fluorophenol· · ·H2 S complex showing

progression in the intermolecular stretch vibration. Source: Bhattacharyya and Wategaonkar
[99]/with permission of American Chemical Society.
intermediate states S1 000 and S1 σ10 . As the authors pointed out, because of the poor
S/N ratio, the scan could not be performed up to the dissociation limit. The ZEKE
spectrum recorded via the S1 000 intermediate state showed a long progression in the
intermolecular stretching vibration up to v′ = 5 at 146, 288, 428, 567, and 701 cm−1 .
Thus, successive difference between the vibrational levels decreases indicating
significant anharmonicity along the intermolecular stretch mode. Interestingly, the
intermolecular stretch for the cationic p-fluorophenol· · ·H2 O complex was found
to be harmonic owing to a greater H-bond strength. Other transitions shown in the
spectrum of the cationic p-fluorophenol· · ·H2 S were assigned to the intramolecular
modes and the progression of the intermolecular stretch was also observed in
combination with other fundamental intramolecular vibrational modes. The
intermolecular stretch fundamental transition of the cationic p-fluorophenol· · ·H2 S
was identified as 145 cm−1 . The ZEKE spectrum obtained by exciting the 𝜎 1
(intermolecular fundamental stretch) in the S1 state also had similar features.
Energy level spacing (Δ𝜀) between the successive states (of the intermolecular
stretching modes) was determined by averaging the corresponding differences
seen in the different progressions in the recorded spectrum and was found to be
145, 144, 141, 138, and 133 cm−1 for the first five vibrational levels and the BS
extrapolation plot for fluorophenol· · ·H2 S in the cationic state is generated from
these experimental data (Figure 11.20). From the slope and the intercept of Δ𝜀 vs. V,
𝜔e and 𝜔e xe were determined as 149.41 and 1.6062 cm−1 , respectively. By equating
the left-hand side of Eq. (11.23) to zero (at the dissociation limit), vmax is determined
as 45. The D0 for the cationic p-fluorophenol· · ·H2 S was thus determined as
150
135
Δε (cm–1)
16
1
0 3 6 42 45
v
Figure 11.20 The Birge–Sponer extrapolation plot for fluorophenol· · ·H2 S in the cationic
state. Source: Bhattacharyya and Wategaonkar [99]/with permission of American Chemical
Society.
9.72 ± 1.05 kcal mol−1 (3402 cm−1 ). The D0 for the cationic p-fluorophenol· · ·H2 S
calculated using MP2/wB97X-D theory was found to be about 18% higher than
the extrapolated value. Interestingly, in a follow-up paper, Wategaonkar and
coworkers validated the BS extrapolation technique by comparing the directly
determined D0 with the BS extrapolated value for the cationic complexes having
O–H· · ·S interaction [100]. The D0 for the cationic complexes of phenol· · ·H2 S
and p-cresol· · ·H2 S complexes were determined directly using photofragmentation
spectroscopy and were also estimated using the BS extrapolation. By combining
the ab initio calculations, experimental data, and the BS extrapolation for other
complexes including the cationic indole· · ·H2 O complex, the authors concluded
that the D0 values estimated by the BS extrapolation technique were consistently
within 20% of the best known value for the dissociation energy (either derived
experimentally or theoretically), but on the lower side. Such deviation from the
true value has been addressed by Gaydon and it was pointed out that a decrease
in the rate of convergence of the vibrational levels at longer interatomic distances
would make the estimated dissociation energy lower than the true value [97].
Also, in cases where the dissociation coordinate is not a pure stretching vibration,
the one-dimensional representation of the dissociation coordinate based on the
stretching vibration would be an oversimplification. In cases, where the motion is
a pure stretching vibration, the complex is taken as a pseudo diatomic molecule.
Moreover, as pointed out by Wategaonkar and coworkers, the application of the BS
extrapolation for the complex would also require to have little or no deformation
in the monomer structures in the complex as the original assumption used in the
application of BS extrapolation is that the dissociation products are either produced
in the ground state or the state of the products are definitively known. In conclusion,
in cases where direct determination of D0 is hindered, BS extrapolation can be used
if anharmonicity in the dissociation coordinate is observed in the first few levels of

the intermolecular stretches. Note that the BS extrapolation technique could not
be used for p-fluorophenol· · ·H2 O complex as the progression in its intermolecular
stretch was found to be harmonic.
11.3.3 Determination of the Dissociation Energy of H-Bonded

Complexes Using SEP-REMPI Technique
The SEP-REMPI technique is a triply resonant technique where the first laser
excites the complex to an excited electronic state (S1 ) via S0 –S1 transition and
the complex is then subsequently dumped to a vibrationally excited level of S0
by the second laser (stimulated emission pumping [SEP]) [101]. The SEP by the
second laser produces the complex M–S in a vibrationally excited state. Change in
geometry between the S0 and S1 electronic states determines the vibrational modes
in the complex which are produced with excitation after the complex returns to the
S0 state following SEP. The intramolecular vibrational modes in the complex are
produced with high excitation in the dumping process and after the M–S complex
returns to a vibrationally excited level (v′′ ) of S0 state, the state evolves into a super-
position of many low-lying vibrational states (including the intermolecular modes)
by intramolecular vibrational relaxation (IVR). In Section 11.2.1, the vibrational
predissociation process following direct IR excitation of the intramolecular stretch
was described. For the SEP-REMPI technique described in this section, the IVR
from the intramolecular to the intermolecular modes precedes the vibrational
predissociation dynamics. Because the intramolecular vibrational frequencies are
3–50 times higher than the lowest intermolecular frequencies, the IVR occurs on
a slow time scale of 5–100 ps. In cases where the SEP target state (v′′ ) energy is
smaller than D0 of the complex, the dissociation of the complex will not occur and
the vibrationally excited M–S complex (hot M–S complex or M# S) can be detected
using the third laser via Resonant-2-photon ionization (R2PI) technique. When
the SEP target state energy is greater than D0 of the complex, the complex would
dissociate and its absence can be detected via R2PI. For the complex to dissociate
at the dissociation threshold or just above it, the total amount of vibrational energy
that is available to the complex should flow into the intermolecular stretch mode
which is statistically rare and thus, the vibrational predissociation in these cases
occurs on a much slower timescale (100 ns–10 μs) than IVR. Thus, by bringing in
the third laser after a delay of 100 ns to 10 μs,the absence of the complex is detected
via R2PI. The experimental scheme is thus referred to as SEP-R2PI. Figure 11.21
shows a schematic of the SEP-R2PI experiments used for the determination of D0
reproduced from Ref. [102]. The first laser (pump) excites the M–S complex to v′ = 0
level of the excited state S1 via S0 (v′′ = 0) → S1 (v′ = 0) 000 transition (process 1 in
Figure 11.21). After certain time delay, the second laser (dump laser) is scanned
over energies less than the energy of the pump laser and when the dump laser is
resonant with S1 → S0 transition of the complex, the complex returns to a FC allowed
excited vibrational level of the S0 state (process 2 in Figure 11.21). Subsequently, the
complex undergoes IVR and the vibrational predissociation occurs in cases where
Stimulated emission pumping (SEP) resonant two photon ionization (R2PI)
M* + S
Ion
M+• S complex
3
3'
1
M* + S
0
Energy
S1
M* • S complex
2'
2
3
1 3'
2
vib(intra)
IVR + VP 1
M+S
vib(inter)
0
D0(S0)
IVR
S0 M • S complex
100−15 000 ns Time
Figure 11.21 SEP-REMPI experimental scheme for the determination of D0 for M–S
complex. Source: Frey et al. [102]/with permission of American Chemical Society.
the magnitude of dissociation energy is smaller than the energies of the vibrational
levels populated by SEP. Thus, the presence or absence of the vibrationally hot M–S
complex is probed by R2PI by bringing in the third laser (process 3 in Figure 11.21).
Different schemes shown in Figure 11.21 are utilized to collect the mass-selected
ion signal and the spectra. To determine the D0 of the M–S complex, the hot band
SEP-R2PI spectrum is collected by tuning the third laser (the probe laser) to the
hot band of the M–S complex (process 3) while the dump laser is scanned. In
absence of a resonant downward transition stimulated by the dump laser, the
hot band SEP-R2PI background signal is contributed by the fluorescence signal
arising from the S1 (v′ = 0) state of the M–S complex. As the dump laser is resonant
with any vibrational level higher than v′′ = 0 of S0 state (but having energy less
than the D0 of the complex), an increase in the hot SEP-R2PI signal is observed.
On the other hand, if the dump laser is resonant with a vibrational level of S0
which is above the dissociation energy of the complex, vibrational predissociation
occurs and the SEP-R2PI spectrum disappears. Thus, the highest energy band
observed in the SEP-R2PI spectrum provides the lower bound for the dissociation
energy.
For the determination of the upper bound of D0 , the “dump spectrum” or the
“origin-probed SEP-R2PI” spectrum is collected. A 1-Color REMPI (1-C-REMPI)
spectrum of the cold M–S complex can be monitored via process 1 while the sec-
ond laser (dump laser) is scanned. As the second laser becomes resonant with any
downward transition to S0 state, a corresponding decrease in the 1-C-REMPI spec-
trum is observed as the v′ = 0 state of S1 is depopulated and this depletion spectrum
is called the “dump spectrum.” On the other hand, the origin-probed SEP-R2PI spec-
trum is collected by tuning the probe laser into the S0 (v′′ = 0) → S1 (v′ = 0) 000 tran-
sition as shown in process 3′ and as soon as the dump laser is resonant with a level
other than v′′ = 0 in the S0 state, a corresponding depletion in the SEP-R2PI spec-
trum is observed. The dump spectrum, the origin-probed SEP-R2PI spectrum, and
the fluorescence spectrum of the complex have transitions that are not dependent
on the dissociation of the complex. Therefore, the vibronic band (observed in the
dump spectrum, origin-probed SEP-R2PI or the fluorescence spectrum) which is
next higher to the one observed in the hot-band SEP-R2PI spectrum provides the
upper bound for the D0 of the complex. As the technique provides bracket to the
value of D0 , the precision of D0 depends on the intensity and densities of the vibronic
bands in the region near the dissociation energy.
D0 of the complexes of 1-naphthol with acetylene and 2-butyne having O–H· · ·π
bonds were measured using SEP-R2PI techniques [103]. After the initial pump and
dump steps, the third laser was brought after a time delay to probe the vibrational
predissociation in the hot M–S complex. The vibrational predissociation in the hot
complex was detected via probing the hot vibronic transitions close to the 000 band
of S0 → S1 levels through R2PI. Figure 11.22 shows the hot-band SEP spectra of
1-naphthol· · ·acetylene complex collected by pumping both S0 (v′′ = 0) → S1 (v′ = 0)
000 and (000 + 41) cm−1 transitions along with its fluorescence spectrum pumped
at 000 band. The dump spectra as described earlier was collected by pumping at
(000 + 41) cm−1 . For 1-naphthol· · ·acetylene, the vibrationally excited complex (with
intermolecular excitation) or the hot complex was detected at 43 cm−1 below the 000
band. The highest energy vibronic band, observed in both the hot-band detected
SEP spectra, was identified at 1191 cm−1 which provided the lower bound for the
dissociation energy whereas the next higher vibronic band observed in the dump
spectrum was identified at 1206 cm−1 . Although a band at 1204 cm−1 was observed
in the hot-band SEP spectrum obtained by pumping at S0 (v′′ = 0) → S1 (v′ = 0) 000 ,
the signal-to-noise ratio was too small to consider this band as the lower bound
for the dissociation energy. From the upper bounds and the lower bounds of the
dissociation energy obtained from the observed bands in the spectra, the D0 (S0 )
for the complex was determined at 1198.5 ± 7.5 cm−1 . For 1-naphthol· · ·2-butyne
complex, the hot-band SEP spectrum and the fluorescence spectra overlapped
for several vibronic bands. The lower bound of D0 was given by the highest wave
number narrow band at 2271 cm−1 observed in the hot-band SEP spectrum. Inter-
estingly, the vibronic bands in the frequency range of ∼2440–2480 cm−1 observed in
the hot-band SEP spectrum were broad and weaker than the corresponding peaks
1-Naphthol•acetylene
1191
(a) (0 + 41 cm–1) hot-band SEP
Intensity (arb. units)
(b) Origin hot-band SEP
(d) Origin fluorescence
1206 (d) (0 + 41 cm–1) dump spectrum

1216
1050 1100 1150 1200 1250 1300

–1
Relative wavenumber (cm )
Figure 11.22 The hot-band SEP spectra and the fluorescence spectra of
1-naphthol· · ·acetylene complex showing the upper limit and the lower limits of
dissociation energy of the complex. Source: Knochenmuss et al. [103]/with permission of
AIP Publishing.
in the fluorescence spectrum and were attributed to the metastable vibrational

states with predissociation lifetime longer than the timescale of the experiments.
However, a more conservative limit on the upper bound of D0 was chosen by
selecting the first band in the fluorescence spectrum which was not seen in
the hot-band SEP spectrum, and thus, the upper bound of D0 was identified at
2671 cm−1 . From the upper bound and the lower bound for D0 obtained from
the spectrum, the D0 (S0 ) for 1-naphthol· · ·2-butyne complex was determined as
2471 ±200 cm−1 .
Before we conclude the section on electronic spectroscopy, we would like
to mention that the dissociation energies of difluorobenzene complexes were
determined by Lawrance and coworkers by exciting the complex to S1 state
where the complex dissociated [36]. Subsequently, difluorobenzene absorbed one
photon and was ionized, then mass-selected and imaged by VMI. Dissociation
energies of ortho-difluorobenzene· · ·H2 O, meta-difluorobenzene· · ·H2 O, and para
difluorobenzene· · ·H2 O complexes were determined as 891 ± 4, 945 ± 10, and
922 ± 10 cm−1 , respectively for the experiments. In another experiment, using a
combination of resonant ionization and VMI techniques, dissociation energies of
aniline· · ·(CH4 )n=1,2 were measured as 6.6 kJ mol−1 (552 cm−1 ) (1,1 complex) and
5.7 kJ mol−1 (476 cm−1 ) by Reid and coworkers [104]. The larger value of D0 for the
1 : 1 complex compared to benzene· · ·methane was attributed to the presence of
both C–H· · ·π and C–H· · ·N interactions in the 1 : 1 complex of aniline· · ·CH4 .
11.4 Estimation of the Well Depth of H-Bonding

Interactions Through Microwave Spectroscopy
Microwave spectroscopy has been exclusively used for the determination of the
structure of near-equilibrium geometry of weakly bound van der Waals and
H-bonded complexes. Rotational spectrum of a molecular system is uniquely
dependent on the mass and position of each atom and the rotational constants
extracted from the spectra are inversely dependent on the moments of inertia
about the principal axes. By fitting the rotational spectrum of a parent H-bonded
complex and that of its isotopologues, spectroscopic parameters such as rotational
constants and distortion constants are determined. From the rotational constants
obtained for the parent complex and the isotopologues, the distance of the atoms
from the center of mass (c.m.) of the complex is deduced. Besides the structures
of the complex, the spectral splitting observed in the rotational spectrum because
of the internal rotation and the tunneling motion provides indirect information
on the strength of the intermolecular potential. One such example is illustrated by
the comparison of C2 H4 · · ·H2 S with C2 H4 · · ·H2 O complexes [105–107]. Owing to a
weaker H-bonding interaction in the H2 S complex compared to the corresponding
H2 O complex, C2 H4 · · ·H2 S exhibits spectral splitting pattern due to the internal
motion of both C2 H4 and H2 S monomers in the complex [106, 107]. In C2 H4 · · ·H2 O
complex, spectral splitting was observed because of the internal motion of H2 O
alone and C2 H4 motion was quenched due to the stronger H-bonding interaction
than the corresponding H2 S complex [105]. Although the binding energy (D0 )
of the complex cannot be directly determined from its rotational spectrum, the
centrifugal distortion constants obtained from the spectral fitting can be used to
determine the quadratic force constant or the H-bond stretching force constant ks .
For the evaluation of force constant, in case of a linear or symmetric rotor, DJ is used
whereas ΔJ is used for the planar asymmetric rotor. In cases where the moments of
inertia of the constituent monomers are small, a pseudo-diatomic approximation
considering the monomers as the point masses is used for the determination of the
stretching force constant from DJ /ΔJ [93, 107, 108]. The inherent assumptions are
that the intermolecular stretch is decoupled from other vibrational modes in the
complex and only the intermolecular stretch vibration contributes to the constant
DJ or ΔJ . Under this approximation, the stretching force constant ks for a planar
asymmetric rotor is given by:
64 π4 (𝜇Rc.m. )2 (B4 + C4 )
ks = (11.24)
hΔJ
In Eq. (11.24), Rc. m. is the c.m–c.m distance, 𝜇 is the reduced mass and is given by
M M
𝜇 = M B+MHX . In this diatomic model, the radial potential for the H-bonding interac-
B HX
tion can be approximated as the Lennard–Jones 6/12 potential and is thus given by
the following expression:
[( ) ( )6 ]
Re 12 Re
V(R) = 𝜀 −2 (11.25)
R R
11.4 Estimation of the Well Depth of H-Bonding Interactions Through Microwave Spectroscopy 333
A Taylor series expansion of V(R) about R = Re is given by:

36𝜀 252𝜀
V(R) = −𝜀 + (R − Re )2 − (R − Re )3 (11.26)
Re 2 Re 3
In Eq. (11.26), 𝜀 is the well depth and Re is the equilibrium intermolecular distance.
Note that the well depth 𝜀 is synonymous with the equilibrium dissociation energy
De . Thus, the well depth 𝜀 is deduced from the stretching force constant ks as follows:
36𝜀 1 72𝜀
= ks or ks = (11.27)
Re 2 2 Re 2
By approximating Re as R0 (i.e. the c.m–c.m distance between the two monomers as
determined by the experiment) and by substituting the ks values calculated from the
experimental distortion constants in Eq. (11.27), the well depth 𝜀 for the H-bonding
interaction in the complex is evaluated. Table 11.3 lists the stretching force constants
for a few H-bonded complexes determined from the microwave spectroscopic data
using the method described above.
As the pseudo-diatomic model considered above was found to be satisfactory only
when the moments of inertia of the monomers about the principal axes of rotation
were small, D.J. Millen later came up with expressions relating the distortion
constants and the quadratic force constants where this criterion was relaxed [116].
Millen derived expressions relating the force constant with microwave spectroscopic
parameters e.g. centrifugal distortion constants and the rotational constants for
various classes of complexes including the asymmetric top dimers. In cases where
B· · ·HX is a linear or symmetric-top complex, the value of ks is derived from the
following expression:
{( )}
16π2 𝜇B3 B B
ks = 1− − (11.28)
DJ BB BHX
In case where the complex B· · ·HX is an asymmetric top dimer such that the axis of
HX lies perpendicular to the nuclear plane base of B, where B and HX are considered
as rigid subunits, the ks is given by the following equation:
8π2 𝜇 3
ks = {B (1 − b) + C3 (1 − c)} (11.29)
ΔJ
where b= BB + BB and c = CC + CC and 𝜇 is the reduced mass. The rotational con-
B HX B HX
stants B and C in Eqs. (11.28) and (11.29) should ideally be the equilibrium rotational
constants. However, microwave spectroscopic studies provide the zero-point vibra-
tionally averaged rotational constants and not the equilibrium rotational constants.
The zero-point averaged rotational constants can be used for B and C in Eqs. (11.28)
and (11.29), respectively.
Interestingly, recently, by considering a series of halogen-bonded and H-bonded
complexes having generic formula B· · ·XY and B· · ·HX where X and Y are halogens,
Legon showed that the equilibrium dissociation energy (De ) and the stretching
force constant ks is related as De = Cs ks where ks is the stretching force constant,
Cs is the proportionality constant and is evaluated as 1.5 (3) × 103 m2 mol−1 . The
equilibrium dissociation energy (De ) was calculated using ab initio method at
Table 11.3 The stretching force constant (k s ) for several H-bonded complexes determined
from rotational spectroscopic data.
k s (mdyn Å−1 ) 𝜺/De a) (cm−1 )
C2 H4 · · ·HCl 0.066 [108] 627 (𝜀) [108]

C2 H4 · · ·H2 O 0.046 [105] 375 (𝜀)b) [105]
C2 H4 · · ·H2 S 0.019 [107] 217(𝜀)b)/227.5c) [107]
C2 H4 · · ·HBr 0.052 [109] 807c) [110]
C2 H2 · · ·HCl 0.069 [111] 643 (𝜀)
C2 H2 · · ·HBr 0.0538 [112] 772.3c) [110]
HF•••HCl 0.057 [113] 451 (𝜀) [113]
OC· · ·HCl 0.045 [114] 569 (𝜀) [114]
OC· · ·HBr 0.0113 [115] 469 (𝜀) [115]
H2 S· · ·HF 0.120b) [110] 1702c) [110]
H2 S· · ·HCl 0.0681d) [110] 1064c) [110]
H2 S· · ·HBr 0.0586d) [110] 920c) [110]
H2 S· · ·HI 0.0402d) [110] 650c) [110]
H2 O· · ·HF 0.2451d) [110] 2954c) [110]
H2 O· · ·HCl 0.1272d) [110] 1784c) [110]
H2 O· · ·HBr 0.1006d) [110] 1465c) [110]
H2 O· · ·HI 0.0664d) [110] 1008c) [110]
The well depth or equilibrium dissociation energy values (𝜀/De ) are listed. The well depth 𝜀 is
calculated using the experimental data whereas De is theoretically calculated. When the values
correspond to well depth 𝜀, it is indicated using a bracket beside the listed values. Appropriate
references are included in brackets.
a) The well depth (𝜀) calculated using the microwave spectroscopic data and the Lennard–Jones
potential, has been reproduced from the references included in the table.
b) Calculated using Lennard–Jones potential from the experimental data in reference [105] and
[107]
c) Theoretical value of equilibrium dissociation energy De .
d) ks values obtained from microwave spectroscopic data are taken from Table 3 in Ref. [110]
CCSD(T)(F12*)/cc-pVDZ-F12 level of theory whereas ks was calculated using

appropriate Eq. (11.28) or (11.29). Some of the H-bonded complexes considered
by Legon are included in Table 11.3. The equilibrium dissociation energy De
calculated by ab initio method vs. the stretching force constant ks determined from
the microwave spectroscopy experiments were plotted for B· · ·HX and the plot was
found to be reasonably linear, although the line did not pass precisely through the
origin (Figure 11.23). The plot was fitted using the following expression
De = 1.53 (3)(ks ) − 1.8(3) (11.30)
The slope obtained for De vs. ks (note, Legon quoted ks as kσ ) plot provides the value
for the proportionality constant Cs . Within the experimental error bar, the slope
11.5 Conclusion 335
60
50 B...HX (X = F, CI, Br, or I)
40
Dσ /(kJ mol–1)
30
20
10
–10
0 10 20 30
–1
kσ /(N m )
Figure 11.23 Plot of equilibrium dissociation energy (De ) vs. the stretching force constant
(k s ) evaluated from microwave spectroscopic experimental data for H-bonded complexes of
type B· · ·HX where B = N2 , CO, C2 H2 , C2 H4 , H2 S, H2 O, PH3 , or NH3 and X = F, Cl, Br, and I. The
black solid line is the line fitted by linear regression. Source: Legon [110]/with permission
of Royal Society of Chemistry.
obtained for the H-bonded complexes B· · ·HX was the same as the halogen-bonded
B· · ·XY complexes. Hence, Legon proposes that for such complexes where a direct
proportionality is found between De and ks , the one-dimensional functions such
1
as the reduced Morse function V(r) = De [1 − e−𝛼(r−re ) ]2 where a = (2cs )− 2 or
the reduced Rydberg function such as V(r) = −De [1 + b (r − re ) ]e−b(r−re ) , where
1
b = Cs − 2 can be used to describe their radial intermolecular potential energy
functions.
11.5 Conclusion
The experimental spectral data on the binding energy measurements are necessary
for benchmarking the PESs calculated by high-level theoretical calculations.
In this chapter, we have put together myriad of high-resolution spectroscopic
techniques which serve as a tool for the determination of the dissociation energy
of hydrogen-bonded complexes. The techniques which have been covered in
this chapter include microwave, infrared spectroscopy, and different excitation
schemes in the UV-visible region. Basic principles of the techniques have been
highlighted with representative examples to familiarize the reader with the scope of
the techniques. The representative examples include a variety of hydrogen bonding
interactions ranging from the strong O–H· · ·O interaction in the water dimer to the
complexes having weaker interactions such as C–H· · ·π, O–H· · ·S. Accurate data
on the binding energy measurements of hydrogen-bonded complexes are relatively
scarce due to the limitation posed by various factors including the spectral selection
rules, the density of states in the fragments, the number density of states near
the dissociation threshold, and the shape of the intermolecular potential. While
the methods such as the VMI of vibrational predissociation phenomenon provide
highly accurate binding energies, the method is less suitable for larger systems.
On the other hand, the electronic excitation schemes, although can be applied
for larger size complexes are restricted to the presence of a chromophore in the
complex. Knowledge of the accurate binding energies of the hydrogen-bonded
complexes is crucial for the understanding of physical and chemical processes in
the atmosphere, acid dissociation, and the many-body effects of the intermolecular
interactions. Hence, going forward, a synergistic play of theory and experiment
along with an expansion in the available experimental techniques to include larger
and complicated system for binding energy measurement is desired.
References
bond (IUPAC Recommendations 2011). Pure Appl. Chem. 83 (8): 1637–1641.
2 Arunan, E., Desiraju, G.R., Klein, R.A. et al. (2011). Defining the hydrogen
bond: an account (IUPAC Technical Report). Pure Appl. Chem. 83 (8):
1619–1636.
3 Rocher-Casterline, B.E., Ch’ng, L.C., Mollner, A.K., and Reisler, H. (2011).
Communication: determination of the bond dissociation energy (D0 ) of the
water dimer,(H2 O)2 , by velocity map imaging. J. Chem. Phys. 134: 211101.
4 Shank, A., Wang, Y., Kaledin, A. et al. (2009). Accurate ab initio and “hybrid”
potential energy surfaces, intramolecular vibrational energies, and classical ir
spectrum of the water dimer. J. Chem. Phys. 130 (14): 144314.
5 Bernath, P. (1995). Spectra of Atoms and Molecules, 339–343. New York: Oxford
University Press.
6 Das, A., Mandal, P.K., Lovas, F.J. et al. (2018). The H2 S dimer is hydrogen-
bonded: direct confirmation from microwave spectroscopy. Angew. Chem. Int.
Ed. 57 (46): 15199–15203.
7 Fernández-Alarcón, A., Guevara-Vela, J.M., Casals-Sainz, J.L. et al. (2021). The
nature of the intermolecular interaction in (H2 X)2 (X = O, S, Se). Phys. Chem.
Chem. Phys. 23 (16): 10097–10107.
8 Fraser, G.T. (1991). (H2 O)2 : spectroscopy, structure and dynamics. Int. Rev. Phys.
Chem. 10 (2): 189–206.
9 Goswami, M. and Arunan, E. (2009). The hydrogen bond: a molecular beam
microwave spectroscopist’s view with a universal appeal. Phys. Chem. Chem.
Phys. 11 (40): 8974–8983.
10 Arunan, E. and Mani, D. (2015). Dynamics of the chemical bond: inter-and
intra-molecular hydrogen bond. Faraday Discuss. 177: 51–64.
11 Ciaffoni, L., Cummings, B., Denzer, W. et al. (2008). Line strength and
collisional broadening studies of hydrogen sulphide in the 1.58 μm region
using diode laser spectroscopy. Appl. Phys. B 92 (4): 627–633.
References 337
12 Lemke, K.H. (2017). Structure and binding energy of the H2 S dimer at the
CCSD (T) complete basis set limit. J. Chem. Phys. 146 (23): 234301.
13 Goswami, M. and Arunan, E. (2011). Microwave spectroscopic and theoretical
studies on the phenylacetylene· · ·H2 O complex: C–H· · ·O and O–H· · ·π
hydrogen bonds as equal partners. Phys. Chem. Chem. Phys. 13 (31):
14153–14162.
14 Gnanasekar, S.P. and Arunan, E. (2021). Structure and internal motions of a
multifunctional alcohol–water complex: rotational spectroscopy of the propargyl
alcohol⋅⋅⋅H2 O dimer. J. Phys. Chem. A 125 (33): 7138–7150.
15 Deshmukh, M.M., Bartolotti, L.J., and Gadre, S.R. (2008). Intramolecular
hydrogen bonding and cooperative interactions in carbohydrates via the
molecular tailoring approach. J. Phys. Chem. A 112 (2): 312–321.
16 Afonin, A.V. and Vashchenko, A.V. (2019). Benchmark calculations of
intramolecular hydrogen bond energy based on molecular tailoring and
function-based approaches: developing hybrid approach. Int. J. Quantum Chem.
119 (21): e26001.
17 Coppens, P. (1997). X-Ray Charge Densities and Chemical Bonding. International
Union of Crystallography.
18 Bader, R. (1990). Atoms in Molecule. A Quantum Theory. Clarendon Press.
19 Scoles, G. (1988). Atomic and Molecular Beam Methods. Oxford university press.
20 Oudejans, L. and Miller, R. (2001). Photofragment translational spectroscopy
of weakly bound complexes: probing the interfragment correlated final state
distributions. Annu. Rev. Phys. Chem. 52 (1): 607–637.
21 Miller, R.E. (ed.) (1998). Pendular state spectroscopy in photodissociation
experiments of hydrogen-bonded complexes. In: Laser Techniques for
State-Selected and State-to-State Chemistry IV . International Society for Optics
and Photonics.
22 Samanta, A.K., Wang, Y., Mancini, J.S. et al. (2016). Energetics and
predissociation dynamics of small water, HCl, and mixed HCl–water clusters.
Chem. Rev. 116 (9): 4913–4936.
23 Samanta, A.K. and Ch’ng LC, Reisler H. (2013). Imaging bond breaking and
vibrational energy transfer in small water containing clusters. Chem. Phys. Lett.
575: 1–11.
24 Samanta, A.K., Gb, C., Wang, Y. et al. (2014). Experimental and theoretical
investigations of energy transfer and hydrogen-bond breaking in small water
and HCl clusters. Acc. Chem. Res. 47 (8): 2700–2709.
25 Bohac, E., Marshall, M.D., and Miller, R. (1992). Initial state effects in the
vibrational predissociation of hydrogen fluoride dimer. J. Chem. Phys. 96 (9):
6681–6695.
26 Oudejans, L. and Miller, R. (1997). Dissociation dynamics of oriented DF−HF
and HF−DF complexes: evidence for direct and indirect dissociation. J. Phys.
Chem. A 101 (41): 7582–7592.
27 Ni, H., Serafin, J.M., and Valentini, J.J. (2000). Dynamics of the vibrational
predissociation of HCl dimer. J. Chem. Phys. 113 (8): 3055–3066.
28 Oudejans, L. and Miller, R. (2000). The state-to-state predissociation dynamics

of OC–HF upon HF stretch excitation. J. Chem. Phys. 113 (11): 4581–4587.
29 Bohac, E. and Miller, R. (1993). Intermolecular V-V energy transfer in the
photodissociation of weakly bound complexes: a new experimental approach.
Phys. Rev. Lett. 71 (1): 54.
30 Oudejans, L. and Miller, R. (2000). Photodissociation of cyclic HF complexes:
pentamer through heptamer. J. Chem. Phys. 113 (3): 971–978.
31 Song, C., Tian, Z., Wang, C. et al. (2021). Growth behavior and properties of
(HF)1–16 clusters. Struct. Chem. 32 (1): 395–403.
32 Chandler, D.W. and Houston, P.L. (1987). Two-dimensional imaging of
state-selected photodissociation products detected by multiphoton ionization.
J. Chem. Phys. 87 (2): 1445–1447.
33 Eppink, A.T. and Parker, D.H. (1997). Velocity map imaging of ions and elec-
trons using electrostatic lenses: application in photoelectron and photofragment
ion imaging of molecular oxygen. Rev. Sci. Instrum. 68 (9): 3477–3484.
34 Bellm, S.M., Gascooke, J.R., and Lawrance, W.D. (2000). The dissociation energy
of van der Waals complexes determined by velocity map imaging: values for S0
and S1 p-difluorobenzene–Ar and D0 (p-difluorobenzene–Ar)+ . Chem. Phys. Lett.
330 (1-2): 103–109.
35 Bellm, S.M., Moulds, R.J., and Lawrance, W.D. (2001). The binding energies of
p-difluorobenzene–Ar,–Kr measured by velocity map imaging: limitations of
dispersed fluorescence in determining binding energies. J. Chem. Phys. 115 (23):
10709–10717.
36 Bellm, S.M., Moulds, R.J., van Leeuwen, M.P., and Lawrance, W.D. (2008).
A velocity map ion imaging study of difluorobenzene-water complexes: binding
energies and recoil distributions. J. Chem. Phys. 128 (11): 114314.
37 Ewing, G.E. (1979). A guide to the lifetimes of vibrationally excited van der
Waals molecules: the momentum gap. J. Chem. Phys. 71 (7): 3143–3144.
38 Ewing, G.E. (1980). Vibrational predissociation in hydrogen bonded complexes.
J. Chem. Phys. 72 (3): 2096–2107.
39 Ewing, G.E. (1987). Selection rules for vibrational energy transfer: vibrational
predissociation of van der Waals molecules. J. Phys. Chem. 91 (18): 4662–4671.
40 Ch’ng, L.C., Samanta, A.K., Gb, C. et al. (2012). Experimental and theoretical
investigations of energy transfer and hydrogen-bond breaking in the water
dimer. J. Am. Chem. Soc. 134 (37): 15430–15435.
41 Czakó, G., Wang, Y., and Bowman, J.M. (2011). Communication: quasiclassical
trajectory calculations of correlated product-state distributions for the dissocia-
tion of (H2 O)2 and (D2 O)2 . J. Chem. Phys. 135 (15): 151102.
42 Casterline, B.E., Mollner, A.K., and Ch’ng LC, Reisler H. (2010). Imaging
the state-specific vibrational predissociation of the hydrogen chloride−water
hydrogen-bonded dimer. J. Phys. Chem. A 114 (36): 9774–9781.
43 Rocher-Casterline, B.E., Mollner, A.K., and Ch’ng LC, Reisler H. (2011).
Imaging H2 O photofragments in the predissociation of the HCl−H2 O
References 339
44 Ch’ng, L.C., Samanta, A.K., Wang, Y. et al. (2013). Experimental and theoretical
investigations of the dissociation energy (D0 ) and dynamics of the water trimer,
(H2 O)3 . J. Phys. Chem. A 117 (32): 7207–7216.
45 Ceponkus, J., Karlström, G., and Nelander, B. (2005). Intermolecular vibra-
tions of the water trimer, a matrix isolation study. J. Phys. Chem. A 109 (35):
7859–7864.
46 Mancini, J.S., Samanta, A.K., Bowman, J.M., and Reisler, H. (2014). Experiment
and theory elucidate the multichannel predissociation dynamics of the HCl
trimer: breaking up is hard to do. J. Phys. Chem. A 118 (37): 8402–8410.
47 Case, A.S., Heid, C.G., Kable, S.H., and Crim, F.F. (2011). Dissociation energy
and vibrational predissociation dynamics of the ammonia dimer. J. Chem. Phys.
135 (8): 084312.
48 Zuraski, K., Kwasniewski, D., Samanta, A.K., and Reisler, H. (2016). Vibra-
tional predissociation of the HCl–(H2 O)3 tetramer. J. Phys. Chem. Lett. 7 (21):
4243–4247.
49 Zuraski, K., Wang, Q., Kwasniewski, D. et al. (2018). Predissociation dynamics
of the HCl–(H2 O)3 tetramer: an experimental and theoretical investigation.
J. Chem. Phys. 148 (20): 204303.
50 Peterson, K.A. and Dunning, T.H. Jr., (1995). Benchmark calculations with cor-
related molecular wave functions. VII. Binding energy and structure of the HF
dimer. J. Chem. Phys. 102 (5): 2032–2041.
51 Oudejans, L., Moore, D., and Miller, R. (1999). State-to-state vibrational pre-
dissociation dynamics of the acetylene-HF complex. J. Chem. Phys. 110 (1):
209–219.
52 Viana, M.A., Araújo, R.C., Neto, J.A.M. et al. (2017). The interaction strengths
and spectroscopy parameters of the C2 H2 · · ·HX and HCN· · · HX complexes
(X= F, Cl, CN, and CCH) and related ternary systems valued by fluxes of
charge densities: QTAIM, CCFO, and NBO calculations. J. Mol. Model. 23 (4):
110.
53 Oudejans, L. and Miller, R. (1999). State-to-state vibrational predissociation
dynamics of the acetylene−HCl complex. J. Phys. Chem. A 103 (25): 4791–4797.
54 Pritchard, M., Parr, J., Li, G. et al. (2007). The mechanism of H-bond rupture:
the vibrational pre-dissociation of C2 H2 –HCl and C2 H2 –DCl. Phys. Chem. Chem.
Phys. 9 (47): 6241–6252.
55 Oudejans, L. and Miller, R. (1995). State-to-state photodissociation of ori-
ented HF-HCl complexes: isotopic and isomeric effects. J. Phys. Chem. 99 (37):
13670–13679.
56 Johnson, S.N. and Tschumper, G.S. (2018). Hydrogen bonding in the mixed
HF/HCl dimer: is it better to give or receive? J. Comput. Chem. 39 (14):
839–843.
57 Parr, J.A., Li, G., Fedorov, I. et al. (2007). Imaging the state-specific vibrational
predissociation of the C2 H2 −NH3 hydrogen-bonded dimer. J. Phys. Chem. A
111 (31): 7589–7598.
58 Hartmann, M. and Radom, L. (2000). The acetylene−ammonia dimer as a

prototypical C−H· · ·N hydrogen-bonded system: an assessment of theoretical
procedures. J. Phys. Chem. A 104 (5): 968–973.
59 Oudejans, L. and Miller, R. (1998). Intermolecular V–V energy transfer in the
photodissociation of CO2 –HF (v= 1). J. Chem. Phys. 109 (9): 3474–3484.
60 Oudejans, L. and Miller, R. (1998). Mode dependence of the state-to-state vibra-
tional dynamics of HCN–HF. Chem. Phys. 239 (1-3): 345–356.
61 Sexton, T.M., Van Benschoten, W.Z., and Tschumper, G.S. (2020). Dissociation
energy of the HCN· · ·HF dimer. Chem. Phys. Lett. 748: 137382.
62 Shorter, J.H., Casassa, M.P., and King, D.S. (1992). Fragment state correlations
in the dissociation of NO⋅ HF (v= 1). J. Chem. Phys. 97 (3): 1824–1831.
63 Legon, A.C. (2021). A test of ab initio-generated, radial intermolecular poten-
tial energy functions for five axially-symmetric, hydrogen-bonded complexes
B· · ·HF, where B= N2 , CO, PH3 , HCN and NH3 . Phys. Chem. Chem. Phys.
23 (12): 7271–7279.
64 Liu, Y., Li, J., Felker, P.M., and Bačić, Z. (2021). HCl–H2 O dimer: an accurate
full-dimensional potential energy surface and fully coupled quantum calcula-
tions of intra- and intermolecular vibrational states and frequency shifts. Phys.
Chem. Chem. Phys. 23 (12): 7101–7114.
65 Mollner, A.K., Casterline, B.E., Ch’ng, L.C., and Reisler, H. (2009). Imag-
ing the state-specific vibrational predissociation of the ammonia−water
66 Li, G., Parr, J., Fedorov, I., and Reisler, H. (2006). Imaging study of vibrational
predissociation of the HCl–acetylene dimer: pair-correlated distributions. Phys.
Chem. Chem. Phys. 8 (25): 2915–2924.
67 Dayton, D., Block, P., and Miller, R.E. (1991). Spectroscopic evidence for
near-resonant intermolecular energy transfer in the vibrational predissocia-
tion of acetylene-HX and acetylene-DX (X= Cl, Br and I) complexes. J. Phys.
Chem. 95 (7): 2881–2888.
68 Carcabal, P., Broquier, M., Chevalier, M. et al. (2000). Infrared spectra of the
C2 H2 –HCl complexes: an experimental and ab initio study. J. Chem. Phys.
113 (12): 4876–4884.
69 Sennikov, P., Ignatov, S., and Schrems, O. (2005). Complexes and clusters
of water relevant to atmospheric chemistry: H2 O complexes with oxidants.
70 Hansen, A.S., Vogt, E., and Kjaergaard, H.G. (2019). Gibbs energy of complex
formation–combining infrared spectroscopy and vibrational theory. Int. Rev.
Phys. Chem. 38 (1): 115–148.
71 Hansen, A.S., Maroun, Z., Mackeprang, K. et al. (2016). Accurate thermody-
namic properties of gas phase hydrogen bonded complexes. Phys. Chem. Chem.
Phys. 18 (34): 23831–23839.
72 Hansen, A.S. and Kjaergaard, H.G. (2017). Dimethyl sulfoxide complexes
detected at ambient conditions. J. Phys. Chem. A 121 (32): 6046–6053.
References 341
73 Du, L., Tang, S., Hansen, A.S. et al. (2017). Subtle differences in the hydro-
gen bonding of alcohol to divalent oxygen and sulfur. Chem. Phys. Lett. 667:
146–153.
74 Kjaersgaard, A., Vogt, E., Hansen, A.S., and Kjaergaard, H.G. (2020). Room
temperature gas-phase detection and gibbs energies of water amine bimolecular
complex formation. J. Phys. Chem. A 124 (35): 7113–7122.
75 Henry, B.R. and Kjaergaard, H.G. (2002). Local modes. Can. J. Chem. 80 (12):
1635–1642.
76 Meyer, R. and Günthard, H.H. (1968). General internal motion of molecules,
classical and quantum-mechanical Hamiltonian. J. Chem. Phys. 49 (4):
1510–1520.
77 Pickett, H.M. (1972). Vibration—rotation interactions and the choice of rotating
axes for polyatomic molecules. J. Chem. Phys. 56 (4): 1715–1723.
78 Vogt, E., Hansen, A.S., and Kjaergaard, H.G. (2019). Local modes of vibration:
the effect of low-frequency vibrations. In: Molecular Spectroscopy: A Quantum
Chemistry Approach, vol. 2, 389–424. Wiley.
79 Mackeprang, K., Kjaergaard, H.G., Salmi, T. et al. (2014). The effect of large
amplitude motions on the transition frequency redshift in hydrogen bonded
complexes: a physical picture. J. Chem. Phys. 140 (18): 184309.
80 Vogt, E., Bertran Valls, P., and Kjaergaard, H.G. (2020). Accurate calculations of
OH-stretching intensities with a reduced-dimensional local mode model includ-
ing Eckart axis embedding. J. Phys. Chem. A 124 (5): 932–942.
81 Kjaersgaard, A., Lane, J.R., and Kjaergaard, H.G. (2019). Room temperature
gibbs energies of hydrogen-bonded alcohol dimethylselenide complexes. J. Phys.
Chem. A 123 (39): 8427–8434.
82 Kjaersgaard, A., Vogt, E., Christensen, N.F., and Kjaergaard, H.G. (2020).
Attenuated deuterium stabilization of hydrogen-bound complexes at room
temperature. J. Phys. Chem. A 124 (9): 1763–1774.
83 Gu, Q. and Knee, J. (2012). Communication: spectroscopic measurement of
the binding energy of a carboxylic acid-water dimer. J. Chem. Phys. 136 (17):
05B401.
84 Kiermeier, A., Ernstberger, B., Neusser, H., and Schlag, E. (1988). Multiphoton
mass spectrometry of clusters: dissociation kinetics of the benzene cluster ions.
J. Phys. Chem. 92 (13): 3785–3789.
85 Arunan, E. and Gutowsky, H. (1993). The rotational spectrum, structure and
dynamics of a benzene dimer. J. Chem. Phys. 98 (5): 4294–4296.
86 Dinadayalane, T. and Leszczynski, J. (2009). Geometries and stabilities of var-
ious configurations of benzene dimer: details of novel V-shaped structure
revealed. Struct. Chem. 20 (1): 11–20.
87 Mons, M., Dimicoli, I., and Piuzzi, F. (2002). Gas phase hydrogen-bonded com-
plexes of aromatic molecules: photoionization and energetics. Int. Rev. Phys.
Chem. 21 (1): 101–135.
88 Courty, A., Mons, M., Dimicoli, I. et al. (1998). Ionization, energetics, and
geometry of the phenol−S complexes (S = H2 O, CH3 OH, and CH3 OCH3 ).
J. Phys. Chem. A 102 (25): 4890–4898.
89 Mons, M., Dimicoli, I., Tardivel, B. et al. (2002). Energetics of a model NH–π
interaction: the gas phase benzene–NH3 complex. Phys. Chem. Chem. Phys.
4 (4): 571–576.
90 Courty, A., Mons, M., Dimicoli, I. et al. (1998). Quantum effects in the thresh-
old photoionization and energetics of the benzene−H2 O and benzene−D2 O
complexes: experiment and simulation. J. Phys. Chem. A 102 (33): 6590–6600.
91 Suzuki, S., Green, P.G., Bumgarner, R.E. et al. (1992). Benzene forms hydrogen
bonds with water. Science 257 (5072): 942–945.
92 Gutowsky, H., Emilsson, T., and Arunan, E. (1993). Low-J rotational spectra,
internal rotation, and structures of several benzene–water dimers. J. Chem.
Phys. 99 (7): 4883–4893.
93 Read, W., Campbell, E., and Henderson, G. (1983). The rotational spectrum and
molecular structure of the benzene–hydrogen chloride complex. J. Chem. Phys.
78 (6): 3501–3508.
94 Gotch, A.J., Garrett, A.W., Severance, D.L., and Zwier, T.S. (1991). The structure
and photophysics of clusters of immiscible liquids: C6 H6 -(H2 O)n . Chem. Phys.
Lett. 178 (1): 121–129.
95 Shibasaki, K., Fujii, A., Mikami, N., and Tsuzuki, S. (2007). Magnitude and
nature of interactions in benzene−X (X= Ethylene and Acetylene) in the gas
phase: significantly different CH/π interaction of acetylene as compared with
those of ethylene and methane. J. Phys. Chem. A 111 (5): 753–758.
96 Shibasaki, K., Fujii, A., Mikami, N., and Tsuzuki, S. (2006). Magnitude of the
CH/π interaction in the gas phase: experimental and theoretical determina-
tion of the accurate interaction energy in benzene-methane. J. Phys. Chem. A
110 (13): 4397–4404.
97 Gaydon, A.G. (1946). The determination of dissociation energies by the
birge-sponer extrapolation. Proc. Phys. Soc. 58 (5): 525.
98 David, C.W. (2008). The Birge Sponer ExtrapolationChemistry Education
Materials Paper 63.
99 Bhattacharyya, S. and Wategaonkar, S. (2014). ZEKE photoelectron spectroscopy
of p-fluorophenol⋅⋅⋅H2 S/H2 O complexes and dissociation energy measure-
ment using the Birge–Sponer extrapolation method. J. Phys. Chem. A 118 (40):
9386–9396.
100 Ghosh, S., Bhattacharyya, S., and Wategaonkar, S. (2015). Dissociation ener-
gies of sulfur-centered hydrogen-bonded complexes. J. Phys. Chem. A 119 (44):
10863–10870.
101 Knochenmuss, R., Sinha, R.K., and Leutwyler, S. (2020). Benchmark experi-
mental gas-phase intermolecular dissociation energies by the SEP-R2PI method.
Annu. Rev. Phys. Chem. 71: 189–211.
102 Frey, J.A., Holzer, C., Klopper, W., and Leutwyler, S. (2016). Experimental
and theoretical determination of dissociation energies of dispersion-dominated
aromatic molecular complexes. Chem. Rev. 116 (9): 5614–5641.
103 Knochenmuss, R., Sinha, R.K., and Leutwyler, S. (2019). Face, notch, or edge?
Intermolecular dissociation energies of 1-naphthol complexes with linear
molecules. J. Chem. Phys. 150 (23): 234303.
References 343
104 Makuvaza, J.T., Loman, J.L., Kokkin, D.L., and Reid, S.A. (2020). Probing
cooperativity in C–H· · ·N and C–H· · ·π interactions: dissociation energies of
aniline· · ·(CH4 )n (n= 1, 2) van der Waals complexes from resonant ionization
and velocity mapped ion imaging measurements. J. Chem. Phys. 153 (4):
044303.
105 Andrews, A.M. and Kuczkowski, R.L. (1993). Microwave spectra of C2 H4 ⋅ H2 O
and isotopomers. J. Chem. Phys. 98 (2): 791–795.
106 Goswami, M., Mandal, P., Ramdass, D., and Arunan, E. (2004). Rotational
spectra and structure of the floppy C2 H4 –H2 S complex: bridging hydrogen
bonding and van der Waals interactions. Chem. Phys. Lett. 393 (1-3): 22–27.
107 Goswami, M., Neill, J., Muckle, M. et al. (2013). Microwave, infrared-microwave
double resonance, and theoretical studies of C2 H4 · · ·H2 S complex. J. Chem.
Phys. 139 (10): 104303.
108 Aldrich, P., Legon, A., and Flygare, W. (1981). The rotational spectrum,
structure, and molecular properties of the ethylene–HCl dimer. J. Chem. Phys.
75 (5): 2126–2134.
109 Fowler, P., Legon, A., Thumwood, J., and Waclawik, E. (2000). Geometry and
binding strength of a π-type hydrogen-bonded complex of ethene and hydro-
gen bromide determined by rotational spectroscopy. Coord. Chem. Rev. 197 (1):
231–247.
110 Legon, A.C. (2014). A reduced radial potential energy function for the halogen
bond and the hydrogen bond in complexes B· · ·XY and B· · ·HX, where X and Y
are halogen atoms. Phys. Chem. Chem. Phys. 16 (24): 12415–12421.
111 Legon, A., Aldrich, P., and Flygare, W. (1981). The rotational spectrum and
molecular structure of the acetylene–HCl dimer. J. Chem. Phys. 75 (2): 625–630.
112 Cole, G., Davey, J., Legon, A., and Lyndon, A. (2003). A hydrogen bonded
complex C2 H2 …HBr isolated and characterized in the gas phase using
pulsed-jet, Fourier transform microwave spectroscopy. Mol. Phys. 101 (4-5):
603–612.
113 Janda, K.C., Steed, J.M., Novick, S.E., and Klemperer, W. (1977). Hydrogen
bonding: the structure of HF–HCl. J. Chem. Phys. 67 (11): 5162–5172.
114 Soper, P., Legon, A., and Flygare, W. (1981). Microwave rotational spec-
trum, molecular geometry, and intermolecular interaction potential of the
hydrogen-bonded dimer OC–HCl. J. Chem. Phys. 74 (4): 2138–2142.
115 Keenan, M., Minton, T., Legon, A. et al. (1980). Microwave spectrum and
molecular structure of the carbon monoxide-hydrogen bromide molecular
complex. Proc. Nat. Acad. Sci. 77 (10): 5583–5587.
116 Millen, D. (1985). Determination of stretching force constants of weakly bound
dimers from centrifugal distortion constants. Can. J. Chem. 63 (7): 1477–1479.
345
12
IR and NMR Spectral Diagnostics of Hydrogen Bond Energy

and Geometry
Peter M. Tolstoy and Elena Yu. Tupikina
St. Petersburg State University, Institute of Chemistry, Department of Physical Organic Chemistry,
Universitetskii pr. 26, 198504 St. Petersburg, Russia
Hydrogen bond is like your spouse: immediately recognizable by the back of the
hair in a crowd of a thousand people but you don’t know how you did it.
Prof. Piero Ugliengo speaking at XIV Horizons in Hydrogen Bond Research
conference, Torino, 2003
12.1 Introduction
12.1.1 Solving the Reverse Spectroscopic Problem
The formation and strengthening of hydrogen bonds (H-bonds) manifests itself
in striking geometric and spectral characteristics of complexes. So much so
that these manifestations constitute the majority of H-bond formation criteria [1].
The inclusion of spectral manifestations in the criteria of a chemical bond formation
might be viewed as a peculiarity of hydrogen bonding among other non-covalent
interactions. Indeed, it stands to reason that H-bond, being a chemical bond in
every sense of the word, is fully described by its electronic structure while all other
parameters – energetic, geometric, and spectroscopic – are consequences. However,
in real-world examples, it is often quite challenging to extract detailed electronic
information, so that spectroscopic parameters become essential probes and markers
for geometric and energetic H-bond properties. Extracting these properties con-
stitutes an inverse spectroscopic problem. Generally speaking, such a problem is
posed incorrectly and, in principle, there are no laws of Nature that guarantee that
it should have a solution or have a unique solution. Luckily, in many practical cases,
it appears to be possible to establish one-to-one correspondences between spectral
features and geometric/energetic properties of hydrogen bonds. Some of these
correlations are quite robust and work on broad sets of complexes, so it is tempting
to call them functional dependencies. For example, bridging proton deshielding in
NMR spectra and redshift of proton donor stretching frequency in IR spectra (if one
forgets for a second about so-called blueshifted H-bonds) are among such reliable

346 12 IR and NMR Spectral Diagnostics of Hydrogen Bond Energy and Geometry
H-bond predictors, explicitly included in the list of H-bond criteria [1]. Many other
correlations seem to perform reasonably well only for limited classes of complexes
and/or in certain ranges of external conditions. Some other correlations are flimsy
and constitute correlations is a less noble sense of the word. Nevertheless, within
the limits of their applicability, each one of established correlations is a useful
analytical tool, and construction of new predictive correlations is in high demand,
especially when direct information on interatomic distances and complexation
energies is hard to obtain.
What is necessary for the reverse spectroscopic problem to be solvable at all?
First of all, the existence of a spectral parameter is required which is sensitive
primarily to the H-bond formation and less affected by other molecular processes.
The existence of sensitive spectral parameters is perhaps trivial: the formation
of a hydrogen bond necessarily changes the electronic structure of interacting
moieties, which in turn affects nuclear shieldings, spin–spin couplings, vibrational
frequencies, and band intensities for atoms and functional groups directly involved
in H-bond formation, located in spatial proximity to or in electronic conjugation
with the hydrogen bridging site. The selectivity of a spectral parameter is somewhat
more tricky, as small spectroscopic changes often can be caused by several phenom-
ena, not relevant to hydrogen bonding in question (small conformational changes,
solvation/packing changes, and secondary interactions). This makes the study of
weak hydrogen bonds – which are characterized by small changes in spectroscopic
parameters – especially challenging. Second of all, for the solution of the reverse
spectroscopic problem, the effective dimensionality of the parameter that is being
evaluated should match the dimensionality of the spectral parameter. Observable
spectral parameters are usually one-dimensional quantities and so is the hydrogen
bond energy, thus they could be straightforwardly correlated. In contrast, hydrogen
bond geometry is at its core a multidimensional property, as it takes at least three
numbers to define the relative positions of three atoms in space (even if one forgets
about the hydrogen bond dynamics and delocalization of nuclear wave functions).
This means that to correlate spectral observables with the hydrogen bond geometry
the latter should be simplified in such a way that it could be expressed as a
number. More on the reduction of the number of geometric dimensions which
is necessary for the solution of reverse spectroscopic problem will be given in
Section 12.2.1.
12.1.2 Spectral Markers for Proton Transfer and H-Bond Length

All spectroscopic parameters useful for construction of hydrogen bond correla-
tions could be roughly divided into two major groups, schematically depicted in
Figure 12.1. The first group (Figure 12.1a) contains spectral observables that tend
to change monotonously upon gradual proton transfer within a hydrogen bond.
These parameters are more suited to serve as markers for the degree of proton
transfer. It is hard to make an exhaustive list of such parameters, but among the
most often used ones are NMR chemical shifts of atoms directly participating in
the H-bond formation (13 C, 15 N, 19 F, 31 P), adjacent ones or those further away but
Monotonous With an extremum

reflect the proton transfer degree reflect the H-bond strength/length
X··H··Y
Spectral parameter
Spectral parameter
X + HY
X··H··Y
XH + Y
XH + Y X + HY
Proton transfer coordinate Proton transfer coordinate

(a) (b)
Chemical shifts
X H Y (X, Y= F, N, P, etc.) X H Y
XO H Y (X= C, P, etc.)
1J 1J
Coupling constants 2hJ
XH HY XY
X H Y
Vibrational frequencies
ν(CO), (PO), etc. ν(XH)/ν(HY)
ν(XY)
Figure 12.1 Schematic representation and selected examples of two types of spectral
parameters for an XHY hydrogen bond: (a) parameters that change monotonously with the
proton transfer coordinate (reflecting the degree of proton transfer) and (b) parameters that
exhibit an extremum in the region of strong hydrogen bonds (reflecting the bond length).
in electronic conjugation with the atom forming the H-bond, one-bond spin–spin
coupling constants, as well as characteristic vibrational frequencies of functional
groups containing these atoms. The second group (Figure 12.1b) contains spectral
observables that tend to pass through an extremum along the proton transfer
pathway. Usually, an extremum is reached for short strong (low-barrier) hydrogen
bonds, which makes such parameter more suitable for construction of correlations
with hydrogen bond length or with complexation energy. The second-group param-
eters are less applicable for the determination of the proton transfer coordinate,
as they give multiple values for it. Examples include the NMR chemical shift of
the bridging proton, spin–spin coupling constant across the hydrogen bridge, the
frequency, and intensity of low-frequency hydrogen bond stretching vibration and
proton donor stretching vibration. It is worth mentioning that upon proton transfer
the proton donating and proton accepting groups switch roles and thus the nature
of the “proton donor stretching” vibration qualitatively changes, which could lead
to discontinuities in the frequency of this vibration as a function of proton transfer
coordinate, as well as to other problems in assignment and characterization of the
corresponding bands. If one excludes the complexes with proton transfer, then
the distinction between two groups of parameters (monotonous ones and those
passing an extremum) essentially disappears. Finally, it stands to reason that the
division of all spectral parameters into two groups as shown in Figure 12.1 paints a
crude picture, as there are numerous exceptions. For example, a more complicated
behavior is often shown by various composite quantities, including differential ones
(such as primary and secondary H/D isotope effects), and quotient ones (such as
isotopic spectral ratios).
Obviously, before their practical usage, spectral correlations have to be first
established for more or less broad set of model complexes for which there are
independent ways to verify the hydrogen bond geometry or energy. How valuable
the resulting correlation will be, depends on the limits of its applicability (the
broadness of the selected set of complexes and experimental conditions), predictive
power (the quality of the observed correlation), the precision of the measurement,
and the experimental accessibility of the chosen spectral parameter in general.
In Sections 12.2 and 12.3, our main objective is to review several new and outline
several well-established correlations between hydrogen bond geometry (inter-
atomic distances) as well as hydrogen bond energy (strength) and spectroscopic
observables in NMR and IR spectra. On average, every two hours a new research
paper appears mentioning H-bonds, thus assembling a comprehensive review is
a Herculean labor, which we wouldn’t be able to complete. Instead, we focus on
selected relevant publications published by our work group and by others over the
last couple of decades, that mostly help the main narrative of this chapter. Many
of the findings described in this chapter were published by the work groups of
T. Steiner, H.-H. Limbach, G.S. Denisov, and N.S. Golubev as some of the most
prominent specialists in the field of geometric and spectroscopic hydrogen bond
correlations.
12.2 Spectral Characterization of Hydrogen Bond

Geometry
12.2.1 Description of Hydrogen Bond Geometry
H-bonds are “soft”: being weaker and less directional than classical covalent bonds
(though many hydrogen bonds have significant covalent character), hydrogen
bonds can adjust to external conditions – temperature, electric fields, and steric
constraints – by changing interatomic distances and angles, a phenomenon that was
early on called “hydrogen bond polarizability” [2]. In systems with nonzero thermal
energy, the dynamic, fluxional nature of H-bonded complexes leads to creation of
ensembles of hydrogen bond geometries each being characterized by nuclear wave
functions of certain shape and width. For the bridging particle, it is sometimes
referred to as “the shape of the proton” [3]. Nevertheless, often it is instructional
to represent the XHY bond geometry (instantaneous or averaged one) by point
positions of three constituting atoms in space. This can be done by giving three
interatomic distances (Figure 12.2a), two distances, and one angle (Figure 12.2b).
At the first glance, three independent geometric dimensions are prohibitive for the
construction of meaningful correlations with one-dimensional spectral parameters.
H
rXH rHY
X rXY Y
(a)
H
rXH rHY
X ∠XHY Y (d)
(b)
rXH
X + HY
H
X Y
q1 = ½ (rXH – rHY) X··H··Y

XH + Y
q2 = rXH + rHY
(c) (e)
rHY
Figure 12.2 (a, b) Parameters that describe hydrogen bond geometry: (a) three interatomic
distances; (b) two distances and one angle. (c) Definition of hydrogen bond coordinates
q1 = 1/2(r XH − r HY ) and q2 = r XH + r HY . (d, e) Two ways to schematically depict proton transfer
pathway.
Fortunately, it is possible to decrease the number of independent geometric

parameters. By analyzing the results of neutral scattering experiments T. Steiner
and others have observed the interdependence between r XH and r HY distances for
hydrogen bonds of various types (X, Y = O, and N) [4–8]. These interdependences
seem to be quite robust: they hold for very broad sets of complexes [9, 10] and they
are reproduced by various quantum-chemical calculations [9]. In essence, for a
series of complexes, i.e. upon change of the proton accepting/donating abilities of
interacting fragments, this could be expressed by the mnemonic rule: “when the
shorter distance gets shorter, the longer one gets longer even more.” As a result, the
shortest heavy atom distance, r XY , is reached for bridging proton positions close to
the hydrogen bond center, r XH ≈ r HY . In the absence of prototropic tautomerism,
the overall proton transfer pathway is shown in Figure 12.2d and as a schematic
plot in Figure 12.2e.
In summary, for intermolecular complexes which tend to have more or less lin-
ear hydrogen bonds and in other cases when the deviations from the linearity could
be neglected or are otherwise inconsequential the dimensionality of the geometric
problem is reduced from three down to two, and further down to one because r XH
and r HY distances are well correlated. This opens the possibility to construct cor-
relations between spectroscopic observables and one or the distances because the
second distance could be reconstructed using previously published purely geometric

correlations. The latter is convenient to express via the linear combinations
q1 = 1∕2(rXH − rHY ) and
q2 = rXH + rHY (12.1)
For a linear hydrogen bond, q1 represents the distance from the bridging particle to
the hydrogen bond center and q2 the distance between atoms A and B (Figure 12.2c),
and the sign of q1 tells on which side the bridging proton is located. Analytically,
the r XH (r HY ) or q2 (q1 ) dependence could be expressed via so-called valence bond
orders pXH and pHY , which are essentially exponential interatomic distances [4]:
rXH −rXH 0 rHY −rHY 0
− −
pXH = e bXH
, pHY = e bHY
(12.2)
so that the following expression holds:
pXH + pHY = 1 (12.3)
The parameters bXH and bHY are found empirically, by fitting, so that Eq. (12.3)
describes best the interdependence of r XH and r HY . The term “valence bond order”
is used here rather loosely, just because Eq. (12.3) looks like the single valency of the
bridging proton is split between its bonds with two heavy atoms. In turn, the param-
eters r XH 0 and r HY 0 have the meaning of interatomic distances in free hypothetical
molecules XH (HY) and thus could in principle be estimated independently, as far
as these distances are conserved in extended series of molecules featuring the same
proton-donating (or proton-accepting) atoms. In other words, the numerical values
of parameters in Eq. (12.2) depend slightly on the data set used for fitting Eq. (12.3)
and are subject to revision, depending on the task at hand. Some empirical modifi-
cations of Eq. (12.3) were proposed in the literature, allowing for better description
of strong short hydrogen bonds and H/D isotope effects on hydrogen bond geometry
[10, 11], but the essence of Eq. (12.3) seems to withstand the test of time.
12.2.2 Averaging of NMR Parameters and Proton Tautomerism

In Section 12.2.1 we were assuming that H-bonded complexes could well be
described by point positions of atoms in space and single values of NMR parameters.
However, fluxional nature of many H-bonds means that their geometry changes
in time and so do the NMR parameters. In case of a fast chemical exchange
between several H-bond configurations, the observed value of a spectroscopic
parameter 𝛿 obs (chemical shift or coupling constant) becomes weighted average of
the intrinsic ones:
∑
𝛿obs = xi ⋅ 𝛿i (12.4)
i
where index i corresponds to particular exchanging configuration, 𝛿 i is its intrin-

sic NMR parameter and xi – weighting coefficient (molar fraction of configuration
i in the exchange process). For H-bonded complexes in solutions, the index i also
enumerates solvation shell configurations, sometimes called “solvatomers” [12, 13].
Furthermore, each atom Z (where Z is the bridging proton, the proton donating or
accepting atom, or any other atom) in each H-bond configuration is characterized by
vibrational wave functions Ψi Z (r), where r is the vibrational coordinate. In this case,
the observed NMR parameter 𝛿 i (Z) could be expressed as the following integral:
𝛿i (Z) = Ψi Z (r) ⋅ 𝛿i (r) ⋅ Ψi Z∗ (r) ⋅ dr (12.5)

∫
where * denotes the complex conjugation and 𝛿 i (r) is the dependence of 𝛿 i on r.
While the averaging of NMR parameters given by Eq. (12.4) occurs due to the ther-
mal motion of interacting moieties and/or the surrounding medium, the averaging
given by Eq. (12.5) remains present even at absolute zero due to the existence of
zero-point vibrations. In principle, the combination of Eqs. (12.4) and (12.5) allow
one to predict temperature dependences and isotope effects in NMR spectra, though
rarely the variables in these equations are known explicitly.
Quite often, in experimental solution-state NMR of intermolecular hydrogen-
bonded complexes the fast processes of proton and hydrogen bond exchanges “wash
out” from NMR spectra information about individual complexes. Indeed, 𝛿 obs are
dependent not only on intrinsic parameters of exchanging species, 𝛿 i , but also – via
xi – on absolute and relative concentrations in the sample, temperature, enthalpy,
and entropy of complexation. The number of unknowns quickly becomes larger
than the number of equations. One of the ways to overcome this is to lower the
temperature in an attempt to reach the slow exchange regime, which is especially
effective if low-freezing solvents are used, such as, for example, liquefied gases
CDF3 /CDF2 Cl/CDFCl2 (freezing temperature below 100 K). This methodology was
pioneered by N.S. Golubev and G.S. Denisov [14] and it was successfully used in
several work references in this chapter.
One of the most prominent and well-studied types of H-bond dynamics is
proton tautomerism, i.e. proton transfer within the hydrogen bridge, X − H· · ·Y
⇌ X· · ·H − Y (charges are omitted for brevity) [15]. Formally, this is a thermally
activated two-state chemical exchange characterized by an equilibrium constant
K. However, as mentioned above, in reality, each tautomeric state is defined not
by a point geometry but by a distribution of geometries and as a result, the clear
distinction between tautomers becomes more or less blurred, as visualized in
Figure 12.3.
For H-bonded complexes in solution, the driving force for the proton transfer and
the main cause of the distribution of H-bond geometries within each tautomer are
the same – the thermal fluctuations of the solvation shell. Using the generalized sol-
vent coordinate (a hypothetical collective numerical parameter that describes the
structure of the solvation shell), this could be visualized as shown in Figure 12.4.
The dual maximum proton distribution function is realized along the reaction path-
way which includes the motion of the solvent. Cross section for fixed solvent con-
figurations (horizontal dashed lines in Figure 12.4) corresponds to single maximum
proton distribution functions.
The change of the temperature or the proton donating/accepting ability of
the interacting fragments leads to the change of the equilibrium constant K.
However, the shift of the equilibrium usually does not occur as depicted in
Well defined
Figure 12.3 Schematic representation of
tautomers proton distribution functions in case of
gradual disappearance of proton
tautomerism in a hydrogen-bonded
complex (from top to bottom).
Proton distribution function
Poorly defined
tautomers
Undefined
tautomers
No
tautomerism
Proton transfer coordinate
A···H–B
Generalized solvent coordinate
(fixed bridging proton position)
A··H··B
A–H···B
Proton transfer coordinate

(fixed solvent configuration)
Figure 12.4 Schematic representation of proton distribution function including

generalized solvent coordinate. Dashed lines and red curves – 1D cross sections of 2D
function along different pathways.
Figure 12.5 Two forms of proton

transfer pathway. (a) “Classical” shift
of tautomeric equilibrium; (b) shift of
tautomeric equilibrium including
continuous changes in the structures
of exchanging tautomers.
(a) (b)
Figure 12.5a. It was shown for a series of anionic OHO-bonded complexes of

2-chloro-4-nitrophenol with various carboxylates (using combined NMR/UV–vis
spectroscopy [16]) and for a series of neutral OHN-bonded complexes of chloroacetic
acid with various pyridines (using a combination of NMR and FTIR spectroscopy,
as well as ab initio molecular dynamics [MD] [17]) that the intermolecular proton
transfer pathway is similar to the one depicted in Figure 12.5b. One could consider
the pathway in Figure 12.5b as a combination of “fully tautomeric” pathway in
Figure 12.5a and “fully mesomeric” one in Figure 12.2d. The key point is that
the H-bond geometries of individual tautomers (however vague this term is
defined for a given system) do not stay constant but change together with the
equilibrium constant K, so that the interdependence between r XH and r HY distances
(see Section 12.2.1) is fulfilled for individual configurations (for intramolecular
hydrogen bonds – as far as allowed by the rigidity of the molecular skeleton).
12.2.3 NMR Hydrogen Bond Correlations

As mentioned in Section 12.1.2, many NMR spectral observables are sensitive to the
hydrogen bond formation and strengthening and thus can be used to elucidate the
degree of proton transfer or the overall length of the hydrogen bridge. In this section,
we will group NMR spectral observables according to the type of hydrogen bond,
limiting ourselves to hydrogen bonds formed between three most electronegative
atoms: O, N, and F.
12.2.3.1 OHO Bonds – 1 H Chemical Shifts

OHO hydrogen bonds are ubiquitous, as interacting partners in this class include
such abundant groups of organic compounds as alcohols, phenols, and carboxylic
acids as proton donors and carbonyls, ethers, esters, and anions of alcohols and
carboxylic acids as proton acceptors, as well as numerous organoelement and
inorganic O/OH containing molecules. As 17 O is a rare isotope (0.038% natural
abundance) with substantial nuclear quadrupole moment, the measurements of
high-resolution 17 O NMR spectra are quite challenging, making bridging proton
chemical shift 𝛿H an indispensable source of information about the hydrogen bond
geometry [18, 19]. Since the first systematic NMR studies on hydrogen-bonded
complexes it became a common knowledge that 𝛿H grows with H-bond shortening,
reaches maximal value for strongest H-bonds, and falls back for complexes with
proton transfer. There were several attempts to correlate δH with interatomic
distances, using sets of H-bonded complexes for which solid-state 1 H NMR spectra
and structural data are available for the same crystal. Some of the first dependences
were presented without explicit equations [20, 21], but later on more refined
correlations were constructed. For example, in Ref. [22] the following equation was
proposed:
46.5
𝛿H(ppm) = − 17.4 (12.6)
̊
rH···O (A)
where numerical coefficients were found by fitting to a set of data points for
carboxylic acids, selenites, phosphates, and several other compounds, c. 25 entries
in total. Equation (12.6) works better for medium–strong hydrogen bonds, but
its behavior for very long and very short r H· · ·O distances has some peculiari-
ties. For example, it predicts unrealistic 𝛿H value of −17.4 ppm for monomeric
(non-hydrogen-bonded) OH proton donors. It is convenient to demonstrate the
performance of Eq. (12.6) using q1 as the proton transfer coordinate (r H· · ·O can be
converted into q1 using Eqs. (12.2) and (12.3) and the numerical parameters from
Ref. [4]), see blue curve in Figure 12.6.
In Ref. [23] the r O· · ·O distance, which coincides with q2 parameter for linear
H-bonds, was expressed as a function of 𝛿H (in ppm) as
̊ = 5.04 − 1.16 ⋅ ln(𝛿H) + 0.0447 ⋅ 𝛿H
rO···O (in A) (12.7)
The fitting of Eq. (12.7) was done for a set of 59 hydrogen bond lengths measured for
molecular crystals for which solid-state NMR was obtained as well [24]. In contrast
to Eq. (12.6), Eq. (12.7) describes a smooth 𝛿H(q1 ) function in the range of short
strong H-bonds (around q1 = 0, see green curve in Figure 12.6; shortest r O· · ·O
distance is c. 2.42 Å) but tends to somewhat exaggerate the 𝛿H values for weaker
H-bonds.
25
Equation (12.6)
20
δH (ppm)
Equation
15 O··H··O (12.7)
10 Equation (12.10)
Equation (12.11)
O·H···O O···H·O
5
–0.6 –0.4 –0.2 0 0.2 0.4 0.6
q1 = ½ (rOH – rHO) (Å)
Figure 12.6 Dependence of OHO bridging chemical shift on hydrogen bond geometry
expressed as a function of hydrogen bond asymmetry parameter q1 . The curves correspond
to Eqs. (12.6), (12.7), (12.10), and (12.11). Source: Adapted from Limbach et al. [9], Sternberg
and Brunner [22], and Harris et al. [23].
Narrowing the set of considered complexes allows one to do either some fine-
tuning or simplification of the correlational function, as it was done, for example,
in Ref. [25], where for a series of phenols with intramolecular OH· · ·O=C hydrogen
bonds the following linear dependence was proposed:
𝛿H = [A + B(rH···O − r0 )] ⋅ cos2 𝜑 (12.8)
where B = −7.6 ppm Å−1 was found by fitting; values of A and r 0 were found to be
0 ppm and 2.05 Å for all the strong intramolecular hydrogen bonds studied; 𝜑 is
the dihedral angle H· · ·O=C—R. Equation (12.8) describes a linear function with
the negative slope: the shorter the hydrogen bond, the higher the 𝛿H value. Simi-
larly, in Ref. [26] using quantum-chemical calculations for a set of complexes with
intramolecular OH· · ·O=C hydrogen bonds another linear function with the nega-
tive slope was proposed:
𝛿H = a ⋅ rH···O + b (12.9)
where coefficients a and b were slightly dependent on the level of theory:
a = −19.83 Å−1 , b = 46.49 ppm for M06-2X/6-31+G(d) and a = −20.49 Å−1 ,
b = 47.49 ppm for B3LYP/6-31+G(d).
Alternatively, in Ref. [9], using the set of complexes with CO—H· · ·OC hydro-
gen bonds (partially overlapping with the sets previously used in Refs. [22, 23])
and following the original ideas of Hans–Heinrich Limbach, outlined, for
example, in Ref. [27], the 𝛿H(q1 ) dependence was expressed parametrically, via
valence bond orders pOH and pH· · ·O as the following expression (see red curve in
Figure 12.6):
𝛿H = 𝛿H0 + Δ ⋅ (4pOH ⋅ pH···O )m (12.10)
where m = 1.2 is an empirically fitted parameter; 𝛿H0 represents the chemical shift
of free OH group (taken as 0.73 ppm for water and aliphatic OH groups and 6.0 ppm
for OH groups bound to unsaturated carbon atoms); Δ is the maximal proton signal
shift with respect to 𝛿H0 , i.e. Δ + 𝛿H0 corresponds to the maximal bridging pro-
ton chemical shift, reached for the central-symmetric OHO hydrogen bonds, which
was set to 21.3 ppm, as reported in Ref. [28] for protonated water dimer (similar
values close to 21 ppm was also reported for intramolecular hydrogen bond in hydro-
gen maleate [29]). Thus, by design Eq. (12.10) has realistic asymptotic behavior for
monomeric species and for short strong H-bonds, both of which in principle could
be measured independently. It is worth mentioning, however, that determination of
𝛿H for monomeric species is not an easy task, as it requires extrapolation to infinite
dilution, where such factors as exchange with water traces have significant impact
on the shape and position of the OH signal.
Besides, in Ref. [9] a simpler and equally well-performing dependence was pro-
posed (black curve in Figure 12.6):
2
𝛿H = 𝛿H0 + Δ ⋅ e−6.2⋅q1 (12.11)
where q1 is in Å, while other parameters are the same as in Eq. (12.10). The advan-
tage of Eq. (12.11) is that it can be solved analytically with respect to q1 .
In summary, Eqs. (12.6), (12.7), (12.10), and (12.11) perform similarly for strong
H-bonds, while for weaker ones the sensitivity of 𝛿H to the chemical nature of the
proton donor starts to play a role and all general dependencies approach the limits of
applicability. Another potential difficulty stems from the fact these correlations do
not take into account various averaging effects, such as exchange within the ensem-
ble of solvatomers or proton tautomerism (see Section 12.2.2 above). Two-state pro-
ton tautomerism, O − H· · ·O ⇌ O· · ·H − O, does not affect the 𝛿H values only in case
of degenerate proton transfer between a pair of point geometries in formally symmet-
ric complexes, such as homo-conjugated anions of carboxylic acids [30, 31], while in
all other cases it is difficult to extract hydrogen bond geometries of individual non-
degenerate tautomers from the single average 𝛿H value. Thermal motions within
the complex further complicate the situation. For example, it was shown by ab initio
MD simulations [32] that the distribution of 𝛿H values in liquid water is c. 10 ppm
broad with the average value (∼5 ppm) depends on the number of H5 O2 + -like struc-
tures at each given moment in time. Such a broad distribution of H-bond geome-
tries is not an exclusive property of aqueous solutions. For the complex formed by
4-nitrophenol and acetic acid dissolved in CD2 Cl2 ab initio, MD simulations predict
a similar 10 ppm-wide distribution of 𝛿H values (from 14 to 24 ppm with the aver-
age value of 18.3 ppm practically coinciding with the experimental one) caused by
formation/breakage of weak additional CH· · ·O hydrogen bonds with surrounding
solvent molecules [33].
The primary H/D isotope effect on it, i.e. Δ𝛿(H/D) = 𝛿D − 𝛿H is of high diagnostic
value for hydrogen bond characterization as well. This spectroscopic isotope effect
reflects the isotope effect on H-bond geometry. Skipping most of the details, it goes
like this: if OHO bond is asymmetric, then ODO bond is even more so; if OHO
bond is central-symmetric, then ODO bond gets even shorter (which is called
anomalous Ubbelohde effect [34]). For more on geometric H/D isotope effects see
Refs. [9, 10, 35]. The sign of Δ𝛿(H/D) was proposed as a criterion to distinguish
asymmetric hydrogen bonds (or double-well H-bonds; negative Δ𝛿(H/D)) from
symmetric ones (low-barrier or no barrier H-bonds; positive Δ𝛿(H/D)) [29]. The
overall dependence of Δ𝛿(H/D) on proton transfer coordinate is non-monotonous:
upon H-bond strengthening, Δ𝛿(H/D) first decreases (becomes more negative) and
then increases to positive values for the shortest H-bonds [31].
12.2.3.2 OHO Bonds – 13 C and 31 P NMR Chemical Shifts

Apart from 𝛿H, the NMR chemical shifts of nuclei adjacent to the OHO bridge could
be used to elucidate the H-bond geometry. Most common examples include 13 C and
31 P nuclei in COHO- and POHO-bonded complexes, respectively (𝛿(COHO) and
𝛿(POHO), observed nuclei are underlined). Complexation effects on 13 C NMR chem-

ical shifts of amides (though usually involved in C=O· · ·HN hydrogen bonds) were
studied by Ando et al. [36–38]. The effects of hydrogen bonding and deprotonation
on 13 C NMR chemical shifts of carboxylic carbons of Aspartic and Glutamic acid side
chains were extensively studied by McDermott et al. [39–42]. It was shown that car-
boxylic group deprotonation shifts its isotropic 13 C NMR signal by several ppm to the
low field, though this trend could be masked by the effects of peptide conformation,
Carboxylic carbon COO–

δ(COHO)
H-bond dependent R-dependent
H
δ(COHO) D
COOH
(a) Isotope effect
q1
0
Bridging particle
ΔδC(H/D)
H-bond dependent R-dependent
δH H 0
D
δD
(b) Isotope effect (c) 0 q1
Figure 12.7 (a, b) Relative contribution of substituent effects (green) and hydrogen
bonding effects (black and blue) to the resulting NMR chemical shifts of (a) carboxylic
carbon and (b) bridging particle for H-bonded complexes formed by an RCOOH(D) molecule.
(c) Sigmoidal dependence of 𝛿(COHO) on proton transfer coordinate q1 produces a
dispersion-like dependence of the H/D isotope effect Δ𝛿C(H/D) on q1 .
motional averaging, and multiple H-bonding to the carboxylic groups. The absolute
values of 𝛿(COHO) of different carboxylic acids are hard to compare due to the elec-
tronic influence of substituents R in RCOOH molecule. In this respect, the usage
of H/D isotope effects on NMR chemical shifts, Δ𝛿C(H/D) = 𝛿(CODO) − 𝛿(COHO),
becomes especially informative, because in this differential quantity the substituent
effects are canceled and the remaining value is determined primarily by the geomet-
ric differences between deuterated and protonated forms of the complex, as depicted
in Figure 12.7a. The same reasoning is valid also for the primary isotope effect on
bridging proton chemical shift, Δ𝛿(H/D), though in this case, the influence of R is
much smaller in the first place (Figure 12.7b).
The isotope effects on carboxylic carbon chemical shifts were used in Ref. [31] to
study H-bond geometries in formally symmetric complexes of carboxylic acids with
their conjugated bases (see structures in Figure 12.8a). Homo-conjugated anions
of carboxylic acids are of special importance, as they exhibit some of the strongest
hydrogen bonds [43–46] and occur as functional H-bonds in cofactors of proteins
[47] and in enzymes [48, 49]. In Ref. [31] the following expression was proposed,
linking Δ𝛿C(H/D) with the H-bond asymmetry:
e−B|q1 |
ΔC(H∕D) = A ⋅ q1 ⋅ (12.12)
1 + e−Cq1
where A = 31 ppm Å−1 , B = 13.7 Å−1 , and C = 6.1 Å−1 are fitting parameters.
Equation (12.12) describes a slightly asymmetric dispersion-like curve, which one
would expect in case the absolute value of carboxylic carbon chemical shift behaves
in a sigmoidal fashion with respect to q1 .
Ref. [16] reports the H/D isotope effects on 13 C NMR chemical shifts of C-1 carbon
of 1-13 C-2-chloro-4-nitrophenol in H-bonded complexes with various carboxylates
and other proton acceptors (Figure 12.8b). It was shown that while the overall proton
transfer pathway looks as shown in Figure 12.5b, the isotope effects Δ𝛿C(H/D) as
O H O K=1 O H O Cl O H A
R R R R 13
C
O O O O
R′ R′ O2N
R = –CH3 O O
–CH2Cl O O
OH O
O
–CHCl2
A= O
OH O R R′
–CF3 O
–CH(CH3)2 R′ = –H
O O
–C(CH3)3 –CH3
OH O Cl Br I BF4
(a) (b)
Figure 12.8 Complexes were studied to establish a Δ𝛿C(H/D) − q1 correlation. (a) Homo-
conjugated anions of carboxylic acids and mono-anions of dicarboxylic acids. Source: Guo
et al. [31]/American Chemical Society. (b) Hetero-conjugated anions of 1-13 C–2-Cl–4-NO2 –
phenol with carboxylates, phenolates, and several inorganic acids. Source: Koeppe et al.
a function of H-bond geometry have the dispersion-like behavior: negative when

proton is closer to C-1, positive when proton is closer to carboxylate and reaching
zero either for central-symmetric hydrogen bonds of asymptotically for complexes
with complete proton transfer toward a carboxylate. The weakly bonded phenols
Δ𝛿C(H/D) isotope effects are negative and small in absolute value.
It is worth mentioning that H/D substitution in a hydrogen bridge affects 13 C
NMR chemical shifts of not only the neighboring carbon atom but also of several
more remote carbon atoms, especially if there is an electronic conjugation with the
proton-donating/proton-accepting group. The long-range H/D isotope effects are
usually of high diagnostic value and allow one to draw conclusions concerning the
hydrogen bond geometries for systems of different nature [50–55].
Another way to study the dependence of 13 C NMR chemical shifts of H-bond
geometry and to minimize the substituent effects is to consider a homologous series
of acid–base complexes in which the proton-donating carboxylic acid is kept the
same, while the proton accepting ability of the base is increased systematically, caus-
ing more or less gradual displacement of the bridging proton toward the base. This
approach was used in Ref. [56] to study complexes formed by acetic acid with substi-
tuted pyridines. We will return to this case in Section 12.2.3.3 when discussing OHN
hydrogen bonds.
The 31 P NMR spectroscopy, despite potentially being a very informative tool
for H-bond characterization [57–59], has yielded significantly fewer results so far,
compared with 1 H and 13 C NMR spectroscopy. One of the reasons for that seems
to be high sensitivity of 31 P NMR chemical shifts to several factors, H-bonding
being only one of them, thus making construction of predictive correlations much
harder [60]. Staying within the class of OHO bonds, there are two main types
of relevant complexes: those in which proton-donating POH group belongs to
phosphinic, phosphoric, or phosphonic acids (Figure 12.9a) and those in which
proton-accepting P=O group belongs to phosphine oxides (Figure 12.9b).
For example, in complexes formed by phosphinic acids (RR′ POOH) the 31 P NMR
chemical shifts, 𝛿(POHO), are comparably sensitive to the formation of hydrogen
R
O H B O H B O H A
R′ O R
R P O P R′
P
O R′ O R″
(a) (b)
R R
P
O H O O O
R
R P P R H
H
R
O H O O O
(c) R P P
O H O R
R R R′ R′
P
R′ R′ O O
P H
H
O H O O O
R R O O
P R′ H
R P R′ H R P P R
O H O O O O R
O
R P P H
(d) O H O R H
R R
O O
P
R P O H O R
P R′ R′
O H O
O O
H H
H
O
H
P R
OH O
O H
P O
(e) R
Figure 12.9 H-bonded complexes are formed by (a) phosphinic and phosphoric acids,
(b) phosphine oxides. (c) Cyclic dimers and trimers of POOH-acids. Source:
(a–c) Mulloyarova et al. [61]/Royal Society of Chemistry. (d) Cyclic hetero-dimers/trimers/
tetramersPOOH-acids. Source: Mulloyarova et al. [62]/MDPI/CC BY 4.0. (e) Tetrahedral
tetramer of phosphonic acid [63] (shown as two approaching cyclic dimers to emphasize
the H-bond network). Source: Giba and Tolstoy [63]/MDPI/CC BY 4.0.
Figure 12.10 Schematic representation of Direct Hydrogen bonding

substituent
the main factors influencing the 31 P NMR effects (R)
as proton donor (···)
chemical shift of phosphinic/phosphoric/ R1 O H

phosphonic acids.
P Angles and dihedral angles ( )
2 O
R
Internal
degrees of Hydrogen bonding
freedom ( ) as proton acceptor (···)
bonds by both POH and P=O groups [64], to the internal rotation of substituents
R1 and R2 [65], and to other factors, such as angles and dihedral angles within the
“tetrahedron” of bonds formed by 31 P atom [66], which is schematically summarized
in Figure 12.10.
Nevertheless, 31 P NMR was successfully used to identify and assign signals
of various H-bonded species: homo-conjugated anions of phosphinic acids [67],
cyclic dimers and trimers of phosphinic, and phosphoric acids (Figure 12.9c) [61],
cyclic hetero-dimers, hetero-trimers, and hetero-tetramers, involving molecules of
two different acids (Figure 12.9d) [62], as well as tetrahedral tetramers of phospho-
nic acids (Figure 12.9e) [63]. Despite the substantial scattering of data points, for
cyclic self-associates (dimers, trimers, etc.) the following expression was proposed
in Ref. [66] for the change of 31 P NMR chemical shift, Δ𝛿(POHO), upon change of
the hydrogen bond length, Δq2 :
Δ𝛿(POHO) = k ⋅ Δq2 (12.13)
where q2 is defined in Eq. (12.1) as the overall length of the hydrogen bridge.
The proportionality coefficients, k, were found to be larger for phosphinic acids
(110–170 ppm Å−1 ) than for phosphoric acids (60–100 ppm Å−1 ). It is likely that
Eq. (12.13) is valid only for cases when POH and P=O groups form H-bonds which
become equivalent in the NMR time scale due to fast reversible multiple (double,
triple, etc.) proton transfer processes in self-associates.
The large span of 31 P NMR chemical shift changes of phosphine oxides due to
their complexation (up to 85 ppm) was used previously to characterize compounds
exhibiting Lewis acidity [68–70] (the so-called Gutmann–Beckett scale of acceptor
numbers AN), Brønsted acidity [71–74] or combined Lewis and Brønsted acidity
[75–77]. Recently the 31 P NMR chemical shifts of triethylphosphine oxide were
correlated with pK a values of added Brønsted acids in a wide range of acidities [78].
Though these correlations do not allow one to explicitly estimate the hydrogen
bond geometry, they demonstrate unequivocally that the Δ𝛿P values could serve as
a measure of strength/length of hydrogen bonds formed by phosphine oxides.
12.2.3.3 OHN Bonds

When it comes to NMR characterization of hydrogen bonds, there are two impor-
tant things that separate OHN bonds from OHO ones. The first thing is the
intrinsic lack of symmetry in OHN bonds which creates an asymmetry in the
𝛿H dependence on the proton transfer coordinate q1 (this is true for all kinds of
hetero-nuclear H-bonds). Indeed, the limiting chemical shifts of hypothetical OH
and HN monomers, 𝛿(OH)0 , and 𝛿(HN)0 , respectively (here and below the charges
often will be omitted for simplicity), are different and as a result Eq. (12.10) should
be modified to reflect this. For example, in Ref. [79] for a series of intermolec-
ular complexes formed by benzoic acids with 2,4,6-trimethylpyridine (collidine;
Figure 12.11a) the following equation was proposed:
𝛿H = 𝛿(OH)0 ⋅ pOH + 𝛿(HN)0 ⋅ pHN + Δ ⋅ (4pOH ⋅ pHN )m (12.14)
where m was set to 1. The parameter 𝛿(OH)0 was set to −4 ppm for fictious
carboxylic acid monomer, while parameter 𝛿(HN)0 was taken as 7 ppm for free
collidinium cation. As it was mentioned above, these limiting values are subject
to revision and taking into account more recent publications, the 𝛿(OH)0 = 6 ppm
[9] and 𝛿(HN)0 = 10 ppm [83] values would seem to be more realistic. Finally, in
Ref. [79] the excess term Δ was set to 19.5 ppm to reproduce maximum 𝛿H values
of about 21 ppm. Later on, to describe better the geometry of short strong OHN
H
N
R O H N O H
R
O O R′
(a) (b)
δ3
δ2
δ1 O H N O H N
(c) (d)
Figure 12.11 Intermolecular complexes with OHN hydrogen bonds: (a) benzoic
acid–collidine complexes. Source: Lorente et al. [79]/John Wiley & Sons. (b) Carboxylate–
imidazolium complexes. Source: Viragh et al. [80]/American Chemical Society. (c) The
schematic orientation of principal components of 15 N CSA tensor for heterocyclic
compounds. Source: Based on data from Refs. [79, 81, 82]. (d) Proton tautomerism between
molecular and zwitterionic forms in OHN hydrogen bonds.
hydrogen bonds and to describe H/D isotope effects on the geometry, in Ref. [10]
some modifications of the valence bond order definitions were proposed, but the
form of Eq. (12.14) was preserved.
In Ref. [80], after an analysis of the X-ray and NMR data for 12 substances with
carboxylate-imidazolium H-bond (Figure 12.11b), the following equation was pro-
posed to determine the interatomic distance r O· · ·N from the NMR chemical shift of
the bridging proton:
( )5
̊ = 1.99 + 0.198 ⋅ ln(𝛿H) + 10.14
rO···N (in A) (12.15)
𝛿H
Equation (12.15) was derived for zwitterionic complexes (O− · · ·H–N+ ) and it was
not meant to describe molecular complexes (O–H· · ·N). Nevertheless, for both
Eqs. (12.14) and (12.15) one can safely state that for hydrogen bonds of intermediate
strength an increase in the proton chemical shift is associated with a hydrogen
bond shortening, while the maximal value of 𝛿H does not correspond to the proton
located in the geometric center of the hydrogen bond, as it was for OHO bonds.
The second important property that distinguishes OHN bonds from OHO ones
is the existence of additional spectral parameters for the hydrogen bond character-
ization: the 15 N NMR chemical shift, 𝛿N, and the 15 N–1 H one-bond spin–spin cou-
pling constant, 1 J(NH) (the latter is usually directly measurable in high-resolution
liquid-state NMR only). In this work, we deliberately put aside the three-bond cou-
pling 3h J(NC) across C=O· · ·H—N hydrogen bonds; for more on its utility in H-bond
description see Refs. [84, 85] and papers cited therein.
One of the advantages of using 𝛿N as a hydrogen bond descriptor is that the
chemical shift changes upon protonation/deprotonation of the nitrogen acceptor
atom could be quite substantial: c. 100–130 ppm for pyridines [79, 83], c. 80 ppm
for imidazoles/histidines [86], and 10–15 ppm for amines [87]. The larger ranges
of chemical shifts could be translated into more precise estimations of H-bond

geometries. Of course, first, the relations between 𝛿N and interatomic distances have
to be established. In Ref. [79], for a series of polycrystalline benzoic acid–collidine
complexes (Figure 12.11a) the values of the dipolar coupling constants 2 H–15 N were
measured. This coupling is proportional to the average inverse cubic D· · ·N distance,
⟨r DN −3 ⟩. Strictly speaking, the latter is not identical to the inverse cubic average
distance, (⟨r DN ⟩)−3 , especially for hydrogen bonds with shallow and anharmonic
proton stretching potentials. Neglecting this difference, the values of ⟨r DN −3 ⟩ could
be used to estimate ⟨r DN ⟩. The following experimental relation between 𝛿N and r DN
(expressed as pDN ) has been established [79]:
𝛿N = 𝛿N∞ − (𝛿N∞ − 𝛿N0 ) ⋅ pDN (12.16)
where 𝛿N0 and 𝛿N∞ are the 15 N NMR chemical shifts of the fictive isolated colli-
dinium cation and free collidine. Thus, (𝛿N∞ − 𝛿N0 ) is the protonation shift of the
selected base. In Ref. [79] 𝛿N∞ was set to 0 and 𝛿N0 was set to 130 ppm. Essentially,
the 𝛿N value falls exponentially when H· · ·N (D· · ·N) distance increases, asymptot-
ically approaching the 𝛿N∞ value. Equation (12.16) was used to estimate H-bond
geometries of carboxylic acid substituted pyridines complexes in polar aprotic solu-
tions [11, 56, 83], to evaluate interatomic distances in cofactor–enzyme functional
hydrogen bond in the active site of aspartate aminotransferase (AspAT) [88] and to
characterize the surface of mesoporous materials, using 15 N-labelled pyridine as a
probe [89].
In solid-state NMR, individual components of the 15 N NMR chemical shift
anisotropy (CSA) tensor are quite informative as well, as their changes upon
protonation/deprotonation could be larger than the change of the isotropic
value 𝛿N. For a heterocyclic nitrogen atom – in pyridines [79], imidazoles [81], and
pyrazoles [82] – the directions of the CSA principal axes are schematically shown
in Figure 12.11c. Though for less symmetric structures there are small deviations,
the axes’ directions are essential as follows: the first axis is along with the NH bond
(or along the lone pair direction), and the second axis is perpendicular to it in the
heterocycle’s plane, and the third axis is perpendicular to that plane.
Calculations of 𝛿N generally reproduce the behavior described in Eq. (12.16)
[90–92], though one should keep in mind that for the comparison with the
experiment, apart from the chemical shift rescaling, which is sometimes advis-
able [93], one should pay attention to the medium effects [90] and the proper
calibrations of 15 N NMR chemical shifts [94].
Most of the one-bond spin–spin coupling constants 1 J(NH) are negative and
their values depend on the hybridization of the nitrogen atom. This being said,
the absolute values of 1 J(NH) – from here on we will omit the sign of the coupling
for simplicity – are very sensitive to the hydrogen bond geometry, namely, to the
NH distance, ranging from 0 (an obvious limit for the deprotonated 15 N nitrogen,
i.e. infinite N· · ·H distance) to more than 100 Hz [92]. For example, one-bond NH
coupling constants could be used to probe hydrogen bonds in proteins [95]. As the
1 J(NH) value depends primarily on the Fermi contact interaction between electrons
and nuclei [96, 97], the overall dependence of 1 J(NH) on r NH could be roughly
described by a decaying exponential function. In Refs. [11, 98] for carboxylic

acid–pyridine type of complexes the following expression was proposed:
1
J(NH) = 1 J(NH)0 ⋅ pHN − 8ΔJ ⋅ pHN ⋅ pOH 2 (12.17)
where 1 J(NH)0 represents the limiting NH coupling constant of the fictive free
pyridinium and ΔJ is the excess term describing the deviation of the 1 J(NH) param-
eter of the strongest OHN hydrogen bond from the half of 1 J(NH)0 . In Eq. (12.17)
the 1 J(NH) values are taken with the sign. The following numerical values were
proposed: 1 J(NH)0 equal to −100 Hz in [99] or −115 Hz in [98]; ΔJ equal to 12.5 Hz
in [99] or 14.4 Hz in [98]. Equation (12.17) expressed as a function of proton
transfer coordinate q1 describes an asymmetric sigmoidal curve. Note that upon the
gradual proton transfer across the hydrogen bond center there are no qualitative
changes of 1 J(NH), only quantitative ones. The coupling across the hydrogen bond,
1h J(NH), while reflecting the covalent character of the bond [100, 101], does not
differ in nature from any other one-bond coupling. Quantum-chemical calculations

of 1 J(NH) couplings can reproduce the couplings reasonably well [97, 102–105],
though sometimes a rescaling is advised [106].
Complexes formed by carboxylic acids and various nitrogen bases (pyridines,
amines, etc.) are classical examples of systems showing proton tautomerism
between molecular and zwitterionic forms (Figure 12.11d) [56, 107–110]. Among
intramolecular OHN complexes, the most prominent examples of proton tau-
tomerism include Schiff bases [111, 112], Mannich bases [113, 114], enaminones
[115], and others [116]. Though fast proton tautomerism complicates the direct
interpretation of the observed NMR spectral parameters in terms of H-bond
geometry due to averaging, the values of 1 J(NH) coupling constants could be used
to study the tautomeric equilibrium [53, 117, 118].
12.2.3.4 NHN Bonds

The class of NHN hydrogen bonds includes examples of such systems as anionic
complexes formed between —C≡NH and − N≡C— molecules (Figure 12.12a),
cationic complexes such as pyridine–pyridinium or proton sponges (Figure 12.12b),
neutral complexes formed between amines or imidazoles (Figure 12.12c), cyclic
self-associates of pyrazoles and amidines (Figure 12.12d), intramolecularly bound
porphyrin and porphycene (Figure 12.12e), as well as RNA and DNA base pairs
(Figure 12.12f), though this list is far from being exhaustive.
Spectroscopic characterization of NHN hydrogen bonds has several advantages.
First of all, the symmetry of homonuclear NHN bonds leads to simplification of
some equations, as it was in the case of OHO bonds (see, for example, Eqs. (12.10)
or (12.11); for NHN bonds see the 𝛿H correlations with the H-bond geometry in
Refs. [27, 119]). There are two 𝛿N values for an NHN hydrogen bond and both
of them change in a predictable way when the H-bond geometry is modified
(see Eq. (12.16)). For example, for hydrogen bonds between imidazoles in crystal
state when the N· · ·N distance decreases one of the N· · ·H distances gets longer,
while the other one gets shorter and thus the corresponding 𝛿N values change in
the opposite directions [120]. In other words, for each spectroscopic observable
(a) (b)
(c) (d)
(e) (f)
Figure 12.12 Some examples of NHN-bonded systems. (a) Anionic —C≡NH· · ·− N≡C—
complexes; (b) cationic pyridine–pyridinium complex and 1,8-bis(dimethylamino)
naphthalene; (c) neutral amine–amine and imidazole–imidazole complexes; (d) self-
associates of pyrazoles and amidines; (e) porphyrin and porphycene; (f) DNA base pairs.
that depends primarily on the positions of three atoms forming the NHN bridge,
the dependence on the proton transfer coordinate will be symmetric with respect
to the H-bond geometric center. Of course, there are limits to this symmetry
advantage in case proton donor group NH and proton acceptor atom N belong to
chemically very dissimilar moieties. Note also that many NHN hydrogen bonds
in formally symmetric systems exhibit proton tautomerism (symmetric two-state
chemical exchange process), which is typically fast in the NMR time scale, leading
to averaging of spectroscopic observables.
Second of all, to the set of previously discussed 𝛿H, 𝛿N, and 1 J(NH) observables
one new NMR parameter of high diagnostic value is added, namely the spin–spin
scalar coupling 2h J(NN). Here, following the commonly used nomenclature, the
letter h in the superscript indicates that this coupling is across the hydrogen bond.
The value of 2h J(NN) is determined predominantly by the Fermi contact term [121]
and as a consequence, it is strongly distance-dependent: it increases when the N· · ·N
distance gets shorter and the largest value reported so far is 16.5 Hz, reached for
6-nitro-2,3-dipyrrol-2-ylquinoxaline [122]. Note that for symmetric systems 2h J(NN)
coupling still could be measured if one breaks the symmetry using, for example,
asymmetric 13 C labeling [123].
In Ref. [27] the following dependence of 2h J(NN) of the bond orders (and thus on
interatomic distances) was proposed:
2h
J(NN) = 2h J(NN)0 ⋅ (4pNH ⋅ pHN )m (12.18)
where 2h J(NN)0 corresponds to the value of the coupling constant for central-
symmetric H-bonds, i.e. when pNH = pHN = 0.5. It was pointed out that the 2h J(NN)0
depends on the hydrogen bond angle. In Ref. [124] based on quantum-chemical
calculations it was concluded that the hybridization state of the nitrogen atoms – i.e.
the chemical nature of interacting fragments – while playing a certain role [121],
does not have too large of an impact on the overall correlation between 2h J(NN)
and interatomic distances within the hydrogen bridge, and reliable N· · ·N distances
can be extracted from the 15 N NMR spectra with the precision of c. ±0.05 Å [125].
Perhaps the most practical application of 2h J(NN) to the structural studies was
found in the case of DNA base pairs. In Refs. [126, 127] both experiment and calcu-
lations linear dependences between 2h J(NN) (in Hz) and bridging proton chemical
shift 𝛿H (in ppm) were found. For example, in Ref. [127] the following expression
was proposed:
2h
J(NN) = 1.32 ⋅ 𝛿H − 10.1 (12.19)
The quality of the linear fit for the computed data set was almost perfect, while for the
experimental data set there was some noticeable scattering. Equation (12.19) illus-
trates that both 𝛿H and 2h J(NN) are symmetric bell-shaped functions of the proton
transfer coordinate (the type of parameter shown in Figure 12.1b) and to a certain
degree remain proportional to the other.
Fitting the calculated data points for 62 neutral and cationic NHN-bonded com-
plexes, in Ref. [128] an exponential expression linking 2h J(NN) and N· · ·N distances
was presented:
2h
J(NN) = 795 579 ⋅ e−3.9868⋅rN···N (12.20)
Later, a different correlation was proposed by some of the same authors specifically
for intramolecular NHN bonds in proton sponges [129]:
2h
J(NN) = 96.26 ⋅ rN···N 2 − 593.27 ⋅ rN···N + 917.71 (12.21)
Equations (12.18)–(12.21) deliver essentially the same message, namely, the 2h J(NN)
coupling constant sensitively depends on the N· · ·N distance, though – as it usually
happens in the field of spectroscopic hydrogen bond characterization – the numeri-
cal details of the interdependence are subject to frequent revisions, depending on the
considered set of complexes, i.e. on how general (and thus less precise) the desired
correlation should be.
12.2.3.5 FHF, FHN, and FHO Bonds

One of the first systems for which hydrogen bond correlations were expressed via
valence bond orders were F− (HF)n clusters (n = 1–4; see Figure 12.13a), studied in
F
F
H
H
F F
F H F F H
H H H F
F F H H F H
F F
F
D∞h C2v D3h Td
(a)
R=H Alkali metal Alkaline earth metal Metalloid

methyl (Li, Na, K) (Be, Mg, Ca) (B)
F H F R
ethyl
iso-propyl
tert-butyl Transition metal Post-transition metal Halogen
(b) (Sc, Cu, Zn) (Al, Ga) (F, Cl, Br)
Figure 12.13 Complexes with FHF hydrogen bonds. (a) Schematic structures of F− (HF)n
clusters (n = 1–4), studied experimentally in Refs. [130, 131]. Source: Adapted from
Shenderovich et al. [130, 131]. (b) Complexes are studied computationally in
Refs. [130, 131]. Source: Adapted from Shenderovich et al. [130, 131].
a series of pioneering works by Limbach, Shenderovich, et al. [130, 131]. In these

clusters, all of the nuclei are magnetically active and directly involved in H-bond
formation.
The (FHF)− anion exhibits the strongest known hydrogen bond [132, 133] and it
was the subject of several spectroscopic studies, demonstrating the central position
of the bridging proton (D∞h symmetry) [134], the anomalous Ubbelohde effect [135],
positive primary H/D isotope effect on chemical shift [29], one of the first examples
of scalar coupling across the hydrogen bridge [130] (later on confirmed computation-
ally [136, 137]), very low frequency of proton stretching vibration [138], and other
interesting features.
For F− (HF)n clusters, the experimentally measured chemical shifts (𝛿H, 𝛿F) and
coupling constant (1 J(FH), 1h J(H· · ·F), and 2h J(FF)) were linked to the proton trans-
fer coordinate q1 as polynomial functions of bond orders. For the chemical shifts the
following equations were proposed:
𝛿H = 𝛿H0 + Δ(4pFH ⋅ pH···F )2
𝛿F = r ⋅ 𝛿F0 ⋅ pFH + 𝛿F∞ ⋅ pH···F (12.22)
where 𝛿H0 and 𝛿F0 represent chemical shift of free HF molecule; Δ + 𝛿H0
corresponds to the maximal bridging proton chemical shift, reached for the
central-symmetric (FHF)− anion; 𝛿F∞ is the fluorine chemical shift of free F− anion
and r is the parameter that allows to differentiate between 19 F NMR chemical shifts
of limiting species HF, H2 F+ , and H3 F2+ . Equation (12.22) coincides with Eq. (12.10)
for OHO hydrogen bonds, but with somewhat different fitting coefficients. In turn,
𝛿F changes monotonously between two asymptotic values (free FH and free F− ),
giving a sigmoidal curve as a function of q1 .
For spin–spin couplings the dependence on the H-bond geometry was
expressed as
1
J(FH) = 1 J(FH)0 ⋅ pFH − 8ΔJ(FH) ⋅ pFH ⋅ pH···F2
2h
J(FF) = 2h J(FF)0 ⋅ (4pFH ⋅ pH···F )2 (12.23)
where 1 J(FH)0 = 600 Hz is the coupling in free FH molecule, while ΔJ(FH) =

162.5 Hz and 2h J(FF)0 = 225 Hz are adjusted so that Eq. (12.23) give correct (experi-
mental) values for the (FHF)− anion [131]. The bell shape of 2h J(FF) dependence on
the proton transfer coordinate classifies it as the type of spectral parameter shown in
Figure 12.1b. Interestingly, 1 J(FH) does not behave strictly monotonously between
two asymptotes (coupling in free FH molecule and zero at infinite r FH distance).
Instead, 1 J(FH) falls from c. 600 Hz to 0 for complexes with proton transfer, becomes
negative, reaches an extremum at c. −50 Hz, and only after that asymptotically
approaches 0 for large r FH distances. This non-monotonous behavior was later on
confirmed computationally [139].
Recently, Eqs. (12.22) and (12.23) were demonstrated to be applicable for a
computational series of FH· · ·FR complexes (Figure 12.13b), there R were alkyls,
alkali metals, alkaline earth metals, transition metals, post-transition metals,
halogens, and metalloids [140]. However, in this case, Eq. (12.22) for 𝛿F had to be
modified, as there are as many asymptotic fluorine chemical shifts for FR molecules
as there are different R fragments. In contrast, Eq. (12.22) for 𝛿H seemed to be
robust and valid for vacuum conditions, for solutions in polar aprotic solvents
and – by extension – potentially in other condensed media as well. In other words,
while the interactions between the complex and its medium could substantially
change the hydrogen bond geometry, the overall 𝛿H dependence on the bridging
proton position stays the same, i.e. all the changes happen by sliding data points
along the 𝛿H(q1 ) curve. This was additionally confirmed in Ref. [141], where the
spectral manifestations of solvent–solute interactions were studied by ab initio
MD simulations for central-symmetric (FHF)− anion in vacuum and dissolved in
CH2 Cl2 or CCl4 . It was shown that the F· · ·F distance and bridging proton position
are strongly influenced by multiple additional non-covalent bonds formed between
fluorine atoms and solvent molecules (C—H· · ·F hydrogen bonds in case of CH2 Cl2
and C—Cl· · ·F halogen bonds in case of CCl4 ). Nevertheless, the set of 𝛿H values
gathered at random snapshots of the MD trajectory was correlated with the bridging
proton position as described by Eq. (12.22). In all media, the width of instantaneous
distribution of 𝛿H values was c. 5 ppm [141].
When it comes to other types of hydrogen bonds involving fluorine, it should be
noted that organic fluorine atom is usually a weak proton acceptor and hydrogen
bonding to it is rather rare [142, 143], though by no means impossible [144]. For
this reason, the majority of spectrum–structure correlations were constructed for
complexes with fluoride anion. For example, the first case of heteronuclear FHO
hydrogen bond between an OH group and F− was reported in Ref. [145] for the com-
plex formed by dissolution of potassium fluoride in glacial acetic acid. In the mono-
solvate, the OH chemical shift was strongly downfield shifted to c. 17.4 ppm [146].
Three decades later, the 1 : 1 and 2 : 1 complexes of acetic acid with fluoride anion
(Figure 12.14a) in solution in polar aprotic solvent were studied by low-temperature
(160–110 K) 1 H and 19 F NMR in Ref. [147], and the formation of very strong FHO
hydrogen bond for confirmed. The hydrogen bond geometries in these complexes
appeared to be sensitive to the polarity of the surrounding medium, as evidenced
by the strong temperature dependence of the chemical shifts and coupling constant.
In Ref. [147] the interpretation of NMR parameters – 𝛿H, 𝛿F, 1 J(FH) – in terms of
Figure 12.14 Complexes of acetic acid

δ– δ– F with fluoride anion, studied in Ref. [147]
O H F
H H in solution in CDF3 /CDF2 Cl mixture
O O
(110 K). Countercation –
O tetrabutylammonium. (a) Structures of
O O
(a) 1 : 1 and 1 : 2 complexes; (b) E-/Z-
isomerization of the 1 : 1 complex.
δ– Source: Golubev et al. [147]/Elsevier.
δ– δ– F
O H F H δ–
O
O
(b) O
interatomic distances was done in a very similar way as for FHF-bonded complexes,
though some changes to Eqs. (12.22) and (12.23) had to be done to account for the
intrinsic asymmetry of the heteronuclear FHO bonds:
𝛿H = 𝛿(FH)0 ⋅ pFH + 𝛿(HO)0 ⋅ pHO + 4Δ ⋅ pFH ⋅ pHO
𝛿F = 𝛿F0 ⋅ pFH + 𝛿F∞ ⋅ pHO + 4ΔF ⋅ pFH ⋅ pHO 2 (12.24)
1
J(FH) = J(FH) ⋅ pFH − 8ΔJ(FH) ⋅ pFH ⋅ pHO
1 0 2
where 𝛿(FH)0 , 𝛿F0 , and 1 J(FH)0 are NMR parameters of free FH molecule, 𝛿F∞ is the
chemical shift of free F− anion (all these parameters were the same as in Eqs. (12.22)
and (12.23)); 𝛿(HO)0 is the OH chemical shift of free acetic acid molecule; all other
parameters were adjusted by fitting (for the numerical values see Ref. [147]).
For 1 : 1 complex, Eq. (12.24) helped to establish a quasi-central-symmetric
geometry of the FHO hydrogen bond, while additional dynamic processes visible
in NMR spectra indicated that upon lowering the temperature there might be also
E-/Z-isomerization of the complex (Figure 12.14b).
When it comes to FHN hydrogen bonds, their spectral characterization is
facilitated by the fact that all three atoms forming the heteronuclear hydrogen
bridge could be magnetically active (in practice it sometimes requires 15 N-labeling).
Otherwise, the treatment of such H-bonds could be done by a combination of
approaches described above for NHN, FHF, and FHO bonds. A large portion of
the FHN class of hydrogen bonds is constituted by complexes formed between
NH groups of organic molecules and fluoride anion. Various NH ligands are used
for anion recognition [148–150] and usage in catalysis [151]. While the fact of
complexation could be detected by the appearance of 1h J(H· · ·F) couplings and the
magnitudes of these couplings reflect the relative strength of the bonding, there are
not so many predictive correlations, which would be useful for quantitative solution
of the inverse spectroscopic problem.
Perhaps the most developed case is presented by FH complexes with 2,4,6-
trimethylpyridine (collidine) (Figure 12.15), studied experimentally by Shen-
derovich et al. in Refs. [152, 153] (1 : 1 complex) and [99] (2 : 1, 3 : 1, and 1 : 2
complexes). The corresponding correlational equations could be readily con-
structed by combining Eqs. (12.24) (changing HO to HN) and (12.16). For the exact
expressions of these equations and the numerical values of the fitted parameters in
Figure 12.15 Structures of

1 : 1, 2 : 1, 3 : 1, and 1 : 2 δ− δ+ δ++
δ−
complexes of FH with F H N F H N
collidine, studied in Refs. δ− H
[99, 152, 153]. For charged F
complexes counter cation –
tetrabutylammonium. δ− −
Source: Adapted from F δ+ F δ+
Shenderovich et al. [99], H H H
− + N N
Golubev et al. [152], and F H N
Shenderovich et al. [153]. H
F
them, we refer the reader to Ref. [99], while here we will explicitly mention only one
so far unexplored spectroscopic observable, namely, 2h J(FN) coupling. In analogy
with 2h J(FF) coupling, in Ref. [99] the following expression was presented:
2h
J(FN) = 2h J(FN)0 ⋅ (4pFH ⋅ pH···F )2 (12.25)
where 2h J(FN)0 = 97.5 Hz was found by fitting. According to Eq. (12.25), the 2h J(FN)
coupling reaches its maximum value (97.5 Hz) when the bridging proton is equally
shared by F and N atoms, i.e. when pFH = pHN = 0.5, though this does not corre-
spond to the geometric center between F and N, because pFH and pHN bond orders
fall with different rates when the corresponding interatomic distances increase. In
other words, Eq. (12.25) describes slightly asymmetric bell-shaped function of the
proton transfer coordinate, which is characteristic of all 2h J(XY) couplings. Such
behavior was reproduced by quantum chemical calculations [154, 155] and it was
noted that the value of 2h J(FN) is largely determined by the Fermi contact term and
as the s-character of nitrogen hybridization increases, so does the absolute value of
2h J(FN).
12.2.3.6 Vicinal H/D Isotope Effects

A quite informative though qualitative NMR spectral feature that could be used to
characterize complexes with multiple hydrogen bonds is the change of 1 H NMR
chemical shift of a bridging proton due to H/D substitution in the neighboring
hydrogen bond. This phenomenon is called vicinal H/D isotope effect [67] and its
origin is as follows: H/D substitution weakens the hydrogen bond and changes its
geometry, i.e. average r XH and r HY distances (for more on that see Refs. [9, 10]; basi-
cally, these are zero-point vibrational effects in an anharmonic potential). In turn,
this causes electron redistribution in the proton accepting and proton donating
atoms, which changes their ability to participate in additional hydrogen bonds
(“σ-polarization”) or the electron effects propagate along the chain of conjugated
bonds (in “resonance-assisted” systems [156]) and affect proton accepting/donating
abilities of other electronegative atoms. In either case, the geometries of neighboring
hydrogen bonds become affected, which changes the 1 H NMR chemical shifts of
remaining bridging protons.
For a system with two coupled hydrogen bonds (denoted as HH), the sign of
the vicinal H/D isotope effect, 𝛿(HD) − 𝛿(HH), depends on whether H-bonds are
1H NMR
Cooperatively coupled H-bonds
(head-to-tail binding) HH HH
H H
HD
H D
(a) δ(HD) – δ(HH) < 0
Anti-cooperatively coupled H-bonds

(head-to-head binding) HH HH
H H
HD
H D
(b) δ(HD) – δ(HH) > 0
Figure 12.16 Vicinal H/D isotope effects on 1 H NMR chemical shifts in systems with
(a) cooperatively coupled and (b) anti-cooperatively coupled H-bonds.
cooperatively or anti-cooperatively coupled. The cooperative coupling usually

occurs in head-to-tail complexes and mutual weakening of H-bonds leads to nega-
tive vicinal isotope effects (Figure 12.16a). In anti-cooperatively coupled head-to-
head systems, in which two hydrogen bonds compete for one proton acceptor,
lengthening of one hydrogen bond upon deuteration shortens the other and vicinal
isotope effects change its sign to positive (Figure 12.16b). In turn, the overall number
of vicinal H/D isotope effects depends on the number of interacting hydrogen bonds
and allows one to count the number of chemically nonequivalent isotopologs, thus
distinguishing between complexes of different stoichiometry.
Figure 12.17 depicts several H-bonded complexes for which vicinal H/D isotope
effects were used to either establish the stoichiometry or to elucidate the relative
positions of the bridging protons (the coupling scheme of H-bonds). The examples
include intermolecular 1 : 2 and 1 : 3 complexes of carboxylic acids with substituted
pyridines [98, 157] (Figure 12.17a), 1 : 2 complex of 3,5-dimethylpyrazole with
trichloroacetic acid [158] (Figure 12.17b), cyclic dimers of acetic acid or other
carboxylic acids either dissolved in polar aprotic solvent [30] (Figure 12.17c)
or trapped in supramolecular capsules [159] (Figure 12.17d), cyclic dimer and
trimers of phosphinic and phosphoric acids [61, 67] (Figure 12.9c), cyclic dimers
of dimethylarsinic acid [160] (Figure 12.17e), homo-conjugated anions of car-
boxylic [30] or phosphinic acids [67] (Figure 12.17f), F− (HF)n clusters [FHF]
(Figure 12.13), as well as intramolecular H-bonded systems, such as substi-
tuted Schiff bases [161] (Figure 12.17g) and models of the H-bonded triad in
the active site of Photoactive Yellow Protein [162] (PYP; Figure 12.17h; for the
latter case the vicinal H/D isotope effects were also detected for the protein
itself [163]).
′ ′
(a) (b)
(c) (d) (e)
(f) (g)
(h)
Figure 12.17 Complexes for which vicinal H/D isotope effects on 1 H NMR chemical shifts
were used to either establish the stoichiometry or to elucidate the relative positions of the
bridging protons. (a) 1 : 2 and 1 : 3 complexes of substitutes pyridines with carboxylic acids;
(b) 1 : 2 complex of trichloroacetic acid with 3,5-dimethylpyrazole; (c, d) cyclic dimers of
carboxylic acids; the dashed circle in (d) indicates encapsulation of the dimer in self-
assembled molecular capsules; (e) cyclic dimers of arsinic acids; (f) homo-conjugated
anions of carboxylic and phosphinic acids; (g) Schiff base with intramolecular hydrogen
bonds; (h) intramolecular models of hydrogen bond chains in the active site of Photoactive
Yellow Protein. See text for the corresponding references.
12.2.4 IR Hydrogen Bond Correlations

Among many vibrational modes that are sensitive to the hydrogen bond formation,
five major ones are directly associated with the motion of the atoms constituting the
H-bond, which are schematically depicted in Figure 12.18.
12.2.4.1 Proton Donor Stretching Vibration

According to the hydrogen bond definition [1], the stretching frequency of proton
donor group 𝜈 XH (Figure 12.18a) experiences an increasing redshift (wave number
decrease), Δ𝜈 XH , as the strength of the complex increases. This statement remains
true as long as one stays within the range of “usual” hydrogen bonds (i.e. not
blueshifting ones) and the bridging proton does not pass the hydrogen bond center
(i.e. the XH stretching remains characteristic vibrational mode). The 𝜈 XH vibrational
H Figure 12.18 Major vibration with participation of

atoms forming hydrogen bond. (a) Proton donor
νXH
X Y stretching; (b, c) proton donor bending vibrations;
(d) low-frequency hydrogen bond vibration;
(a) H (e) hydrogen bond bending.
δXH
X Y
(b)
H
γXH
X Y
(c)
H
νσ
X Y
(d)
H
νβ(XY)
X Y
(e)
frequency is linked directly to the X—H bond force constant, which in turn depends
on the r XH interatomic distance. Thus, it seems reasonable to expect a correlation
between 𝜈 XH and r XH . Furthermore, since interatomic distances r XH , r HY , and r XY
are interdependent, correlations between 𝜈 XH and r HY or r XY should exist as well.
Historically, the first mentions of sensitivity of 𝜈 XH to the interatomic distance
between heavy atoms r XY determined by X-ray methods in XH· · ·Y hydrogen bonds
appeared in Refs. [164–166], where some empirical curves for the rapid evaluation
of hydrogen bond geometry were proposed. It was shown that in a wide range of r XY
distances (2.3–3.4 Å) the 𝜈 XH (r XY ) dependence is roughly an exponential decay for
close-to-linear intermolecular hydrogen bonds, but the decay rate depends on the
types of X and Y atoms. For example, for FH· · ·F hydrogen bonds the correlational
curve is flatter, while for NH· · ·N and OH· · ·N ones the curves are the steepest [166].
In Ref. [166] the difference in the 𝜈 XH (r XY ) decay rates was explained in terms of
electronegativity of atom X. Later it was shown that for strong hydrogen bonds, the
shape of the X–H band changes significantly (becomes increasingly “blurred,” see
cartoons in Figure 12.19), so the reliable determination of band position is compli-
cated [167, 168].
One of the best-known results in this field were published in the seminal works
of Novak and Mikenda [169–172], where the 𝜈 XH -geometry relations were examined
mostly on examples of OH· · ·O and NH· · ·N hydrogen bonds. For OH· · ·O hydrogen
bonds it was proposed to distinguish weak, moderate, and strong hydrogen bonds by
using three different steepness factors Δ𝜈 OH /Δr OO of the piecewise-linear 𝜈 OH (r OO )
Figure 12.19 Schematic

representation of evolution of a XH
band shape upon hydrogen bond
shortening (decrease of a
wavenumber).
Wavenumber
dependence: 1500, 5000, and 12 000 cm−1 Å−1 respectively. At the same time, it was
shown that for NH· · ·N hydrogen bonds the steepness Δ𝜈 NH /Δr NN remains constant
at c. 1850 cm−1 Å−1 for a wide range of complexes. It was also proposed to evalu-
ate the exact vibrational frequency of broadbands corresponding to strong hydrogen
bonds as the center of gravity of a band or using additional information from Raman
spectra [170]. Weak hydrogen bonds with participation of strongly coupled systems
such as XH2 can be treated by using a partial isotopic substitution XH2 → X(D)H,
which makes the remaining XH stretching more characteristic.
For OH· · ·O hydrogen bonds, the dependence of 𝜈 OH (in cm−1 ) on the r OH distance
(in Å) was found to be linear and the following correlational equation was proposed
[169]:
Δ𝜈OH∕𝜈
OH free
= 3.6 ⋅ ΔrOH (12.26)
Another correlation between the changes of 𝜈 CH (in cm−1 ) and the r XH distance
(in Å) for CH· · ·X hydrogen bonds was proposed in Ref. [173]:
Δ𝜈CH = 2.69 − 17.91 ⋅ ΔrCH (12.27)
The advantage of Eq. (12.27) in comparison with Eq. (12.26) is its applicability to
both redshifted and blueshifted hydrogen bonds.
In a series of works published by Rozenberg et al. [174–176] using a wide set of
binary gas-phase hydrogen-bonded complexes of carbohydrates and nucleosides, as
well as hydrogen-bonded networks in polycrystalline amino acids and peptides it
was proposed to link the redshift Δ𝜈 XH (in cm−1 ) and interatomic distance r HY (in
nm) using a reverse power function:
Δ𝜈XH = 0.011 ⋅ [ΔrHY ]−6.11 (12.28)
which is claimed to be applicable for hydrogen bonds with ∠XHY > 140∘ . It was
also noted, that the position of 𝜈 XH band is temperature-dependent and correspond-
ing correction of low-temperature spectral data is important for using Eq. (12.28).
For example, a linear dependence of the peak wave number on temperature was
established in Ref. [174]. Authors of Ref. [165] mention, that Eq. (12.28) can be
applied not only for solving the inverse spectroscopic problem: in case of a com-
plicated network of hydrogen bonds, the bands in spectra can be assigned based on
the experimentally available geometric information.
In the last decade or so, with the progress made in the development of compu-
tational methodologies and availability of highly efficient computational resources,
it became possible to consider a wide series of homologs complexes to establish the
applicability of earlier proposed equations or clarify numerical coefficients [177]. For
example, in Ref. [178] for a series of complexes with OH· · ·O=C hydrogen bond the
following linear equation was proposed linking the frequency of stretching vibration
𝜈 OH (in cm−1 ) and interatomic distance r OH (in Å):
𝜈OH = 23 065 − 19 966 ⋅ rOH (12.29)
The authors suggested using Eq. (12.29) for fast evaluation of a wavenumber 𝜈 OH
based on the known r OH value. The error of such estimation was stated to be
±18 cm−1 .
With the increasing amount of available spectroscopic data, it becomes clear, that
even in a range of complexes with a particular type of hydrogen bond it is hard
to establish a single and simple correlational equation for most pairs of spectro-
scopic and geometric/energetic parameters. In some cases, for OH· · ·O hydrogen
bonds the linear character of the 𝜈 OH (r OO ) dependence was found for complexes
with 𝜈 OH > 2500 cm−1 , while for stronger complexes there are deviations from lin-
earity [177, 179]. In some other cases, for OH· · ·X− and OH· · ·O hydrogen bonds
the linearity seems to hold in the wide range of Δ𝜈 OH (0–3000 cm−1 ) [180].
Investigations of FH· · ·F hydrogen bonds presented in Ref. [140] allowed to con-
struct a parametric dependence between Δ𝜈 FH and the hydrogen bond orders pFH
and pHF :
Δ𝜈FH = 1912 ⋅ (4pFH pHF )2.8 (12.30)
which allows to plot Δ𝜈 FH as a function of hydrogen bond asymmetry q1 using the
approach described in Section 12.2.1. In Ref. [181] for XH· · ·Y (X = O, N, C, and
Y = N, O) hydrogen bonds it was proposed to link the redshift Δ𝜈 XH linearly with
the r XH interatomic distance, but with different slopes for OH, NH, and CH proton
donors and comparable intercepts.
12.2.4.2 Proton Donor Deformational Vibrations

It was shown [169] that the shift of stretching 𝜈 XH band upon formation of
hydrogen-bonded complexes are also accompanied by a shift of bending bands
(in plane 𝛿 XH and out of plane 𝛾 XH , Figure 12.18b,c) as well. For example, for
OH· · ·O hydrogen bonds the interrelation between the change of 𝛾 OH and 𝛿 OH band
positions was expressed as
Δ𝜈OH∕Δ𝛾 = 5.8 (12.31)
OH
Due to the stronger vibrational coupling with other vibrations 𝛿 XH is considered

to be less suitable than 𝛾 XH for building spectrum–structure correlations. At the
moment, two equations linking the 𝛾 OH vibrational frequency and r OO distance are
used:
rOO = 3.01 − 4.4 ⋅ 10−4 ⋅ 𝛾OH (12.32)
rOO = 2.17 + 2.1 ⋅ 10−4 ⋅ 𝛾OH (12.33)
Equation (12.32) was constructed based on IR spectral data [182], while Eq. (12.33)
for based on inelastic neutron scattering (INS) spectra [183]. Although both
equations are linear and slopes and intercepts are of the same order of magnitude,
there is still a considerable difference in their absolute values and the different
signs of the slopes. In our opinion, the INS data can be more accurate because INS
spectra are not directly perturbed by electrical fields in comparison with IR [184].
It is also important to note that identification of deformational bands in experi-
mental IR spectra is rather difficult [185] even with the help of deuteration of the
sample or using additional computational data [186]. The temperature dependence
of 𝛿 XH and 𝛾 XH should be kept in mind as well. Nevertheless, deformational bands
can be a useful tool for probing strong hydrogen bonds with significantly broadened
stretching bands.
12.2.4.3 Carbonyl Stretching Vibration

Yet another informative band in IR spectra is C=O stretching vibration, 𝜈 CO
[187–189]. This band has several advantages: it is characteristic, intensive, and
located in relatively unobscured spectral region and the corresponding group
is often directly involved in H-bond formation. Moreover, in case of electronic
conjugation, as in RCOOH groups, 𝜈 CO band remains informative even if C=O
group is not participating in H-bonding, because upon gradual deprotonation
of the carboxylic group the 𝜈 CO band shifts from 𝜈 CO (COOH_free) to asymmet-
ric stretching band in COO− , 𝜈 as (COO− ). This property was used, for example,
in Ref. [17], where for a series of complexes formed by chloroacetic acid and
substituted pyridines dissolved in CD2 Cl2 it was shown by low-temperature IR
spectroscopy that in case of proton tautomerism, O − H· · ·N ⇌ O− · · ·H − N+ , the
𝜈 CO band appear as a dual maximum band. Both components of this dual-band shift
when the equilibrium between molecular and zwitterionic form shifts, and at the
same time the intensity between components is being redistributed, confirming the
proton transfer pathway as shown in Figure 12.5b, though no explicit dependence of
𝜈 CO on proton transfer coordinate was proposed. Ab initio MD simulations carried
out in Ref. [17] have demonstrated that the bands of O–H· · ·N and O− · · ·H–N+
tautomers are inhomogeneously broadened due to thermal fluctuations of the
surrounding solvent or, in a broader sense, the overall shape of the 𝜈 CO band reflects
the distribution of the hydrogen bond geometries in the ensemble of solvatomers.
12.3 Spectral Markers for Hydrogen Bond Energy

12.3.1 Defining Hydrogen Bond Energy
Generally speaking, hydrogen bond energy evaluation is a complicated task. In case
of intermolecular complexes with a single hydrogen bond (Figure 12.20a), the energy
of formation of a hydrogen-bonded complex ΔE in a gas phase could be determined
experimentally (by calorimetric measurements, chemical kinetics, etc. [190–194])
or calculated using quantum chemistry methods as the energy required to separate
X–H and Y fragments onto infinite distance:
ΔE = EX−H···Y − (EX−H + EY ) (12.34)
In a solution or in solid-state solvation and packing effects make direct determina-

tion of ΔE trickier [195].
For crystals, for example, contribution of the lattice energy is essential [196].
For hydrogen bonds with proton transfer (Figure 12.20b) the ΔE value depends on
the way the interacting fragments are separated: either into XH and Y or into X−
and HY+ . For intramolecular hydrogen bonds (Figure 12.20c) the separation of a
single molecule into two monomeric fragments is impossible without destruction
of the molecule, so arbitrary definitions of what constitutes monomers and other
methodologies are usually employed with varying accuracy [197–200]. The sep-
aration into two monomeric fragments is also a key problem for systems with
X Y Figure 12.20 H-bonded systems that

(a) require different definitions of H-bond
energy: intermolecular complexes without (a)
and with (b) proton transfer,
X Y (c) intramolecular H-bond, (d) multiple
(b) H-bonds.
X Y
(c)
X1 Y1
Y2 X2
(d)
multiple hydrogen bonds (Figure 12.20d), as it is not clear how to divide the total
complexation energy into parts corresponding to each H-bond [201, 202], especially
taking into account the possible mutual influence of hydrogen bonds, cooperative
or anti-cooperative. Due to aforementioned problems, determination of H-bond
energy can be polysemantic and requires additional clarification in many cases.
However, as far as H-bond energy reflects the electronic structure of the XHY
moiety, the estimation of ΔE from spectroscopically measurable parameters could
serve as a unified approach, because many spectroscopic observables reflect just
that – the electronic features of the complex.
In many research papers, devoted to the evaluation of hydrogen bond strength,
different quantities are assumed under the term “strength.” The most common
terms are hydrogen bond energy ΔE, hydrogen bond dissociation energy D0 ,
and enthalpy of hydrogen bond formation ΔH. However, other thermodynamic
characteristics – such as free energy ΔG and association/dissociation constants
K – can be used, but they relate to ΔH through standard thermodynamic equations.
For the sake of clarity, we would like to discuss briefly the difference between the
most commonly used terms.
In computational works, hydrogen bond energy ΔE is usually defined as the
electronic energy required for the separation of monomers to an infinite distance.
It is usually defined at 0 K, unless otherwise specified. Enthalpy of hydrogen bond
formation ΔH is defined similar to ΔE (Eq. (12.34)): as a difference between the
enthalpy of a hydrogen-bonded complex and a sum of monomers’ enthalpies.
Dissociation energy D0 is usually defined exactly as ΔH. To understand the
difference between ΔH and ΔE it is useful to remember that ΔH consists of the
following temperature-dependent terms: a change of electronic energy upon com-
plexation ΔE, change of energy of rotational ΔErot , translational ΔEtrans , vibrational
motions ΔEvib (which includes the change of zero-point energy ΔEZPE ), and the
term which depends on the change of the number of moles ΔnRT:
ΔH = ΔE + ΔErot + ΔEtrans + ΔEvib + ΔnRT (12.35)

Thus ΔE and ΔH are not strictly equal even at absolute zero: when ΔErot = ΔEtrans =
ΔnRT = 0, they differ from each other by the change of the nonvanishing zero-point
energy, 𝛥Evib
T=0 K
= 𝛥EZPE
T=0 K
. For higher temperatures, the difference becomes more
significant. However, the term ΔE in Eq. (12.35) is dominant, and the numerical
difference between ΔE and ΔH usually does not exceed a couple of kcal mol−1 .
Quantum-chemical calculation of a hydrogen bond energy ΔE, as well as an
enthalpy of hydrogen bond formation ΔH, usually include an additional term called
basis set superposition error (BSSE), ΔEBSSE , which is caused by the usage of finite
basis sets in ab initio calculations – basis sets used from monomers’ and complex’s
calculations are different [203, 204]. Taking ΔEBSSE into account usually leads to a
decrease in the final value of hydrogen bond strength. For weak hydrogen bonds,
the magnitude of ΔEBSSE can be comparable with ΔE [205, 206]. There is an opinion
that calculation of ΔEBSSE using counterpoise procedure with the full basis set of
the complex for monomers’ calculations is excessive, so taking into account only a
part of ΔEBSSE is sometimes recommended [207, 208].
In further discussion, we will use the term ΔE/ΔH/D0 following its usage in the
discussed paper, taking in mind the difference between them.
12.3.2 NMR Characterization of H-Bond Energy

By their nature, chemical shifts and coupling constants in NMR spectra are
determined by the electron distribution resulting from a particular nuclear con-
figuration. Thus, it could be argued that NMR parameters are potentially well
suited for the extraction of molecular geometry, including interatomic distances
for hydrogen-bonded complexes. Following this logic, NMR parameters should be
suited for the estimation of hydrogen bond energy as well, to the same extent and
with the same limitations as the energy and geometry correlate with each other.
The latter is not always the case, as outlined in Section 12.3.1 (the major problems
being presented by proton transfers, steric hindrances, and multiple interactions in
a given complex). Nevertheless, in many circumstances, the H-bond energy appears
to be a “well-behaving” quantity and useful correlations with the NMR observables
could be constructed. This is especially true if the NMR observable belongs to the
type shown in Figure 12.1b, i.e. reflects the hydrogen bond length. The changes
in hydrogen bond length and energy often go hand-in-hand, leading, for example,
to the appearance of the term “short strong hydrogen bonds” (SSHB), which is
convenient despite being not rigorous and sometimes even self-contradicting [209].
Perhaps the most developed NMR energy correlation is the interdependence
between the bridging proton chemical shift 𝛿H and the hydrogen bond energy ΔE
(or ΔH). Several decades ago linear ΔE(𝛿H) correlations were noticed [210–212],
which in the general form could be written as
ΔE = a ⋅ Δ𝛿H + b (12.36)
where Δ𝛿H is the change of the bridging proton chemical shift upon complexation,
Δ𝛿H = 𝛿H complex − 𝛿H free ; a and b are empirical parameters, which could be adjusted
for a selected set of complexes. In Eq. (12.36) ΔE is written for simplicity, while in
original experimental works the enthalpy of formation, ΔH, was often measured
instead. The value 𝛿H free is not always easy to obtain, so sometimes it is more con-
venient to express Eq. (12.36) as
ΔE = a ⋅ 𝛿Hcomplex + b (12.37)
In Table 12.1, we have collected the fitting parameters of Eqs. (12.36) and (12.37)
for several previously published correlations. One can see that the values of a and b
depend on the chemical nature of interacting moieties and the fitted energy range.
In Ref. [119] it was also pointed out that parameters a differ for neutral and charged
complexes. In other words, while the linearity of the ΔE(𝛿H) dependence seems to
be robust and well established, there is no single set of parameters that fits large
sets of complexes. Very roughly speaking, for typical OHO, OHN, and NHN hydro-
gen bonds the slope of the linear dependence is usually 4–8 kcal mol−1 ppm−1 , i.e.
the shift of 𝛿H on 1 ppm downfield means the enhancing of a hydrogen bond on
1–2 kcal mol−1 .
The meaning of parameters a and b deserves a brief discussion. At the first glance,
the parameter b should be 0, as in the absence of complexation (ΔE = 0) there is no
chemical shift change (Δ𝛿H = 0). However, the best fits often require non-zero val-
ues of b. The reasons for this are not entirely clear. Sometimes it might compensate
for the nonlinearity of ΔE(𝛿H) function, sometimes it might be a result of shield-
ing/deshielding effects caused by the proximity of the interacting partner group even
without complexation (so-called magnetic anisotropy effects), sometimes it might be
a workaround to account for difficulties with ΔE definition. In turn, if b is set to 0,
then the parameter a could be estimated independently, as it was proposed in [119].
Indeed, the change of the proton chemical shift has a limit: the maximal value,
Δ𝛿H max , is reached for symmetric (or quasi-symmetric) hydrogen bonds, which also
happen to be the strongest, exhibiting complexation energy ΔEmax . Then one could
write
ΔEmax
a= (12.38)
Δ𝛿Hmax
(Of course, to use Eq. (12.38) with Eq. (12.37) one should first change Δ𝛿H max
to 𝛿H max .) Generally speaking, in the absence of parameter b any data point
(ΔE, Δ𝛿H) could be used to estimate a, but record strong and central-symmetric
H-bonded complexes give a data point farthest away from (0, 0), thus increasing
the precision of the correlation. Besides, strongest complexes often are better
investigated and sometimes more reliable studies estimating their characteristics
are available (OHO: [9]; NHN: [219]; FHF: [29, 130]; FHN and ClHN [220]).
12.3.3 IR Characterization of H-Bond Energy

12.3.3.1 Proton Donor Stretching Band Shift
The idea of a correlation between IR spectroscopically observed parameters and the
hydrogen bond energy was proposed for the first time in the paper of Badger and
Bauer published in 1937 [221]: the relative shift of OH stretching band upon hydro-
gen bond formation, Δ𝜈 OH /𝜈 OH (where Δ𝜈 OH = 𝜈 OH_free − 𝜈 OH ), was linked linearly
Table 12.1 Parameters of linear correlations between H-bond energies and bridging proton chemical shifts, expressed by Eqs. (12.36) and (12.37).
Range of
Fitting energies a b
H-bond Proton donor Proton acceptor equation (kJ mol−1 ) (kJ mol−1 ppm−1 ) (kJ mol−1 ) References
Experimental data
CHO Halothane or Alicyclic monoethers (12.36) 0–10 14.0 0 [213]
trichloromethane
CHN Halothane or Piperidine (12.36) 0–13 6.8 0 [213]
trichloromethane
NHN 2′ -Deoxyuridines 2′ -Deoxyuridines (12.37) 52–56 3.0 8.6 [214]
OHO, OHN Phenol Pyridine or carbonyls (12.37) 0–40 5.6 26.2 [210]
OHO, OHN 1,1,1,3,3,3-Hexafluoro- Various bases (12.36) 20–50 3.7 15.0 [211]
2-propanol
OHO Phenols or carboxylic Phenols or carboxylic (12.36) 22–109 4.6 2.0 [212]
acids acids
Computational data
OHO Formic acid or vinyl Anions of formic acid (12.37) 65–125 6.1–6.3 –a) [215]
alcohol or vinyl alcohol
XHY (X,Y = O,N,F,C) Various acids Various bases (12.36) 0–190 9.0 −5.1 [216]
ClHN, ClHCl HCl Various bases (12.37) 20–80 3.6 –a) [217]
OHO, OHN, NHN Various acids Various bases (12.37) 20–140 7.2 –a) [217]
FHF charged FH F–(HF)n (n = 0–5) (12.36) 75–200 9.8 47 [218]
FHF neutral FH Various bases (12.36) 0–100 10.9 0 [218]
NHN Aniline Various bases (12.36) 0–22 7.5 0 [140]
a) Intercept value was not reported.

with the enthalpy of hydrogen bond formation ΔH. Nowadays, any linear relation
between hydrogen bond strength (ΔE or ΔH) and the shift of a XH stretching band
upon complexation is known as the Badger–Bauer rule:
ΔH = a ⋅ Δ𝜈XH + b (12.39)
In the following decades, such correlations were reported both experimentally and
computationally for various hydrogen-bonded systems formed by OH, NH, and FH
hydrogen bond donors [222–234]. In some of the works, it was observed that the rela-
tion between Δ𝜈 XH and hydrogen bond strength deviates from linearity [235–237].
Besides, there were several systems that do not obey the Badger–Bauer rule – a shift
of stretching vibration frequency Δ𝜈 XH does not seem to correlate with the hydro-
gen bond strength at all [238–241]. For example, in Ref. [238] on the example of
hydrogen-bonded complexes of phenol with alkyl halides or alkyl chalcogenides, it
was shown that an increase of Δ𝜈 XH cannot be strictly associated with an increase
in hydrogen bond strength, due to steric factors influencing the enthalpy of hydro-
gen bond formation ΔH. The steric argument of disobeying the Badger–Bauer rule
was used in other works as well [239]. Another source of problems with usage of the
Badger–Bauer rule is blueshifting hydrogen bonds [242]. It should be mentioned that
Δ𝜈 XH values also depend on environmental effects, e.g. temperature [171, 176, 243].
It stands to reason that ΔH(Δ𝜈 XH ) dependence should pass the point of origin
because in the absence of complexation the zero change of 𝜈 XH should match zero
value of ΔH. However, often the best fitting results of experimental or computa-
tion data were obtained with non-zero intercept coefficient b (i.e. non-zero value of
ΔH at zero Δ𝜈 XH ). The summary of a and b coefficients for various experimentally
studied complexes with OH proton donating group is given in Table 12.2 (which is
partially reproduced from Ref. [252] and was supplemented with additional data in
this work).
It can be seen in Table 12.2, that even within a certain class of complexes with a
particular type of hydrogen bonds, for example, OH· · ·X bonds formed by alcohols
and various bases, a slope and an intercept for Badger–Bauer correlation depends on
the set of complexes chosen to construct the correlation. In part, it can be explained
by the statistical errors due to the limited number of complexes within each set (most
of the sets presented in Table 12.2 consist of less than two dozen of complexes) and
narrow energetic ranges (usually ΔH < 30 kJ mol−1 ). However, most likely the main
reason is the different “nature” of interacting fragments between sets. For example,
from Table 12.2 it seems that for complexes with OH· · ·Y hydrogen bonds for weaker
proton donors coefficients a tend to be larger, and coefficients b tend to be smaller.
In Ref. [252] authors state that the slope is different for liquid and gas phases (being
larger in a gas) and the quality of a correlation decreases upon change from a solution
phase to a gas phase.
In Ref. [253] on the example of a set with more than 250 complexes with XH· · ·Y
hydrogen bonds (X = C, Si, N, P, O, S, F, Cl, Br; Y = C, N, P, O, S, F, Cl, Br), hydro-
gen bond energies ΔE were successfully correlated with Δ𝜈 XH using intercept b = 0.
It was pointed out that the largest deviations from the fitted function are observed
for weak hydrogen bonds, whose vibrational mode 𝜈 XH can be coupled to other
Table 12.2 Coefficients a and b for Badger–Bauer equation (12.39).
Range of
energies a b
H-bond Proton donor Proton acceptor (kJ mol−1 ) (kJ cm mol−1 ) (kJ mol−1 ) References
Experimental data
OHY Methanol Various bases 8–26 0.0438 7.73 [244]
OHY (Y = O, N, π) Methanol Various bases 1–30 0.0479 6.57 [243]
OHY (Y = O, N) Butanol Various bases 8–21 0.044 35 6.904 [245]
OHF 4-Fluorophenol Fluoroalkanes 10–15 0.13 5.4 [228]
OHCl 4-Fluorophenol Chloroalkanes 6–10 0.12 −0.4 [228]
OHBr 4-Fluorophenol Bromoalkanes 5–10 0.1 −0.9 [228]
OHI 4-Fluorophenol Iodoalkanes 5–9 0.09 −1.45 [228]
OHY (Y = S, N, O) Hexafluoropropanol Various bases 20–42 0.0481 15.06 [211]
OHS 2,2,2-Trifluoroethanol Organosulfur compounds 15–27 0.0506 11.29 [246]
OHO Cholesterol Oxygen bases 10–25 0.055 7.45 [247]
OHY (Y = O, N) Phenol Various O/N bases 13–38 0.0669 2.635 [248]
OHY (Y = O, N, F, Cl) Phenol Various O/N/F/Cl bases 19–38 0.046 11.67 [249]
OHY (Y = O, N) Substituted phenols Various O/N bases 21–39 0.0439 12.88 [250, 251]
3,5-Dichlorophenols Ketones, ethers 29–70 0.108 −9.11 [240]
(Continued)
Range of
energies a b
H-bond Proton donor Proton acceptor (kJ mol−1 ) (kJ cm mol−1 ) (kJ mol−1 ) References
Computational data
OHY (Y = O, N) Alcohols Various bases 5–20 0.077(0.053a)) 2.02(2.15a)) [252]
CHY (Y = N, P, O, S, F, Cl, Br) Various acids Various bases 0–42 0.130 0 [253]
SiHY (Y = N, P, O, S, F, Cl, Br) Various acids Various bases 0–5 −0.007 99 0 [253]
NHY (Y = N, P, O, S, F, Cl, Br) Various acids Various bases 0–15 0.0529 0 [253]
PHY (Y = N, P, O, S, F, Cl, Br) Various acids Various bases 0–10 −0.0453 0 [253]
OHY (Y = N, P, O, S, F, Cl, Br) Various acids Various bases 0–55 0.0705 0 [253]
SHY (Y = N, P, O, S, F, Cl, Br) Various acids Various bases 0–22 0.0727 0 [253]
FHY (Y = N, P, O, S, F, Cl, Br) Various acids Various bases 0–80 0.0719 0 [253]
ClHY (Y = N, P, O, S, F, Cl, Br) Various acids Various bases 0–60 0.0631 0 [253]
BrHY (Y = N, P, O, S, F, Cl, Br) Various acids Various bases 0.0568 0 [253]
a) Data for polarizable continuum model (PCM) (CCl4 ).

vibrations. Interestingly, blueshifting hydrogen bonds, such as SiH· · ·Y and PH· · ·Y,
also obey the Badger–Bauer rule, but with a negative slope.
Upon expansion of the energetic range of investigated complexes, it became clear,
that the character of the ΔH(Δ𝜈 XH ) dependence for moderate and strong hydrogen
bonds is different from that for weak bonds – the second derivative of the over-
all ΔH(Δ𝜈 XH ) function is non-zero [235, 254]. Besides the “classical” linear form
of the Badger–Bauer rule, the possibilities of applying other empirical dependen-
cies have been discussed in the literature. One of them is proposed in the work of
Iogansen [255]:
ΔE = 0.33 ⋅ (Δ𝜈XH − 40)1∕2 (12.40)
where ΔE is in kcal mol−1 and Δ𝜈 XH is in cm−1 .
Equation (12.40) was used suc-
cessfully for estimations of strengths of NH· · ·O hydrogen bonds in amino acids
heterodimers [256], intramolecular CH· · ·O, and OH· · ·O hydrogen bonds in some
nucleoside analogs [257–260], in substituted pyrroles [261], ionic liquids [262], and
other systems.
The other functional dependencies – parabolic, power, etc. – are used for
building Badger–Bauer-like correlations as well [140, 252]. In Ref. [140] for a set
of hydrogen bonded-complexes formed by FH molecule and various fluorine-
containing acceptors (Figure 12.13) in the energy range 0.2–47 kcal mol−1 based
on quantum-chemical calculations (MP2/6-311++G(d,p)) it was proposed to link
hydrogen bond energy ΔE with Δ𝜈 FH by a power function:
ΔE = a ⋅ Δ𝜈FH 𝛾 (12.41)
where 𝛾 is fitting coefficient. The Δ𝜈 FH values were calculated using the local mode
approach [263–265], which is a quite useful and robust method for calculations
of stretching vibrations in systems with multiple identical anharmonic hydrogen
bonds (for example FH· · ·F− hydrogen bonds in (FH)n F− clusters). It was shown,
that for anionic FH· · ·F− and neutral FH· · ·F hydrogen bonds the curvatures
of ΔE(Δ𝜈 FH ) dependencies are different. The following fitting functions were
proposed:
ΔE(kcal mol−1 ) = 3.65 ⋅ Δ𝜈FH 0.337 (cm−1 ), for anionic hydrogen bonds
ΔE(kcal mol−1 ) = 0.18 ⋅ Δ𝜈FH 0.756 (cm−1 ), for neutral hydrogen bonds
(12.42)
It should be pointed out, however, that all of the abovementioned correlations
between Δ𝜈 FH and ΔH or ΔE actually contain an assumption that the frequency of
the stretching vibration of a free hydrogen bond donor is known, which is an issue
for some systems (for example for complexes with intramolecular hydrogen bonds).
To amend this, in Refs. [199, 266] for complexes formed by aliphatic compounds it
was proposed to use the absolute value of proton donor stretching vibration 𝜈 OH as
a descriptor of a hydrogen bond strength. The linear function
ΔE(kcal mol−1 ) = a ⋅ 𝜈OH (cm−1 ) + b (12.43)
was suggested as a fitting function. For hydrocarbonyl aliphatic compounds with
intramolecular OH· · ·O hydrogen bonds in the range 1.4–13.7 kcal mol−1 the
following proportionality was found by gas-phase quantum-chemical calculations

(B3LYP/6-311++G(d,p)) [266]:
ΔE(kcal mol−1 ) = −0.010 ⋅ 𝜈OH (cm−1 ) + 41.1 (12.44)
For an extended set of similar complexes (with hydrogen bonds strength up
to 26.3 kcal mol−1 ) calculated at MP2/6-311+G(d,p) level of theory, the following
equation was proposed [199]:
ΔE(kcal mol−1 ) = −0.084 ⋅ 𝜈OH (cm−1 ) + 40.8 (12.45)
The accuracy of Eqs. (12.44) and (12.45) can be estimated as ±(2–3) kcal mol−1 ,
while for complexes with 𝜈 OH < 2700 cm−1 the deviation of the estimated ΔE from
the experimental one can reach 5 kcal mol−1 . The deviations are also observed for
nonlinear hydrogen bonds and resonance assisted hydrogen bonds, for which cal-
culated ΔE values are respectively higher and lower than those predicted by linear
Eqs. (12.44) and (12.45) [199].
12.3.3.2 Proton Donor Stretching Band Intensity

In practice, the determination of 𝜈 XH band position is often difficult, due to its
overlapping with other bands, e.g. combination bands, hot bands, etc. [267, 268].
This evokes to search for alternative IR spectral parameters sensitive to the strength
of hydrogen-bonded complexes. Among a large number of spectral parameters, that
change upon a complexation [269], special attention of researchers was given to the
intensity of the 𝜈 XH band, which – in comparison to the frequency – can in many
cases be obtained unequivocally from experimental IR spectra.
The idea to estimate hydrogen bond strength from the intensity A of the vibra-
tional band 𝜈 XH originates from the work of Becker [229] and was further devel-
oped in a series of works by Iogansen [236, 270–272]. Becker proposed that the
enthalpy change upon formation of a hydrogen-bonded complex is proportional to
the intensity of the absorption band of proton donor stretching vibration at least
in the range of 2–6 kcal mol−1 . In works of Iogansen for complexes with ΔH in the
range 1–15 kcal mol−1 , it was found that the ΔH(A) dependence is not linear, so as
the measure of ΔH a band intensification
√ √ √
Δ A = A − A0 (12.46)
was proposed, where A is the intensity of hydrogen bond donor stretching vibration
of a complex and A0 is the intensity of stretching vibration of a free hydrogen bond
donor. In Ref. [272] the following proportionality was proposed:
√
ΔH(kJ mol−1 ) = −12.2 ⋅ 102 ⋅ Δ A (12.47)
where A and A0 are expressed in km mol−1 , proportionality coefficient in
kJ (km⋅mol)−1/2 . Estimated accuracy of Eq. (12.47) is about 2 kcal mol−1 for
moderately strong OH· · ·O hydrogen bonds in crystals (ΔH < 60 kJ mol−1 ) [255].
Iogansen’s equation has been actively used in several experimental works for
estimations of a hydrogen bond strength [273–277], however, the limits of its
applicability have not been defined until recently. It can be partially caused by the
problem with accumulation of systematical experimental data for a large set of

similar complexes or high computational costs for accurate calculations of dipole
moment functions and vibrational functions.
In Ref. [278] Iogansen’s rule was examined for complexes with FH· · ·X (X = F,
N, O) hydrogen bonds in the widest range of energies (0.1–49.2 kcal mol−1 ) using
coupled cluster singles and doubles (CCSD)/complete basis set (CBS) calculations.
√
It was shown that for hydrogen bonds with ΔE > 15 kcal mol−1 the ΔE(𝛥 A) depen-
dence increases significantly faster than a linear function, leading to underestima-
tion of a hydrogen bond strength. It was proposed to use two additional parameters
for estimations of ΔE. First of them,
√ √ √
A A A0
Δ = − (12.48)
𝜈XH 𝜈XH 𝜈XH free
provides a linear dependence in the whole range of hydrogen bond energies:
√ ( )
[ ] 1∕2
A km 1
ΔE(kcal mol ) = 2.94 ⋅ Δ
−1
⋅ (12.49)
𝜈XH mol cm−1
The estimated accuracy of Eq. (12.49) is about 2.5 kcal mol−1 . However, the propor-
tionality coefficient of 2.94 is more likely to be valid only for FH· · ·Y hydrogen bonds
and should be independently determined for other types of hydrogen bonds using
√
calibration set of complexes. The second parameter, 𝛥 A ⋅ mR (mR is the reduced
mass of vibrating atoms), reflects the change of the dipole moment function along
the internal coordinate and is expected to be more universal for a wide variety of
hydrogen bond types. The following equation was proposed:
( √ )2.06
ΔE(kcal mol−1 ) = 5.8 ⋅ 10−6 ⋅ Δ A(km mol−1 ) ⋅ mR (a.u.) (12.50)
12.3.3.3 Proton Donor Deformational Vibrations

Upon formation of a hydrogen-bonded complex R1 –X–H· · ·Y–R2 the rotation of X–H
group around R1 –X axis is transformed into two deformational vibrations: 𝛿 XH (in
plane) and 𝛾 XH (out of plane). Both these vibrations are coupled with low-frequency
vibration 𝜈 σ , proton-donating group stretching vibration 𝜈 s , and vibrational modes of
groups neighboring the hydrogen bridge [279]. For example, the correlation between
H–O–H bending mode and O–H stretching for a water network was recently pro-
posed [280, 281]. Thus it is reasonable to expect that frequencies, intensities, and
bandwidth of deformational vibrations can be correlated with the hydrogen bond
strength in analogy with r OO (𝛾 OH ) dependency discussed in Section 12.2.4.
12.3.3.4 Low-Frequency Hydrogen Bond Stretching Frequency

The low-frequency stretching vibration 𝜈 σ arises with the formation of a H-bonded
complex and its frequency increase with the strengthening of a hydrogen
bond, reaching a maximal value for centrally symmetric hydrogen bonds (see
Figure 12.1b). It is located in IR spectra in the wave number region 100–500 cm−1 .
The importance of this vibration for estimations of hydrogen bond strength was first
noted in the book by Pimentel and McClellan [282]. Due to the lack of experimental
capabilities to obtain high-resolution IR and Raman spectra in the far-infrared
region, the accumulation of necessary data was progressing slowly for decades.
The first experimental breakthrough was made when commercial spectrometers
with a transparent dispersing element in the far-infrared region and high precision
diffraction gratings became available. The investigation of low-frequency hydrogen
bond vibration started from pioneering works of Zundel and coworkers [283–287].
The next step was the development of low-temperature (1–10 K) gas-phase spec-
troscopic techniques (supersonic molecular beams [288–290], inert gas matrices
[291–293], and inelastic neutron scattering technique in crystals [294–296]). These
techniques are suitable for obtaining high-resolution vibrational spectra without
the transitions from excited internal degrees of freedom, the influence of solvent
effects, and Doppler broadening. However, despite the obvious success in updating
experimental techniques, at the moment the only published paper that considers
the correlation between the energy of a hydrogen bond ΔE and the low-frequency
stretching vibration 𝜈 σ is a single computational work [297]. In Ref. [297] it is
shown that for complexes with FHF hydrogen bonds in a range up to 50 kcal mol−1
ΔE(𝜈 σ ) dependency is non-smoothly increases with an increase of a hydrogen bond
strength.
12.3.3.5 Stretching Vibrations’ Force Constants

Keeping in mind numerous spectrum–energy correlations proposed for many types
of hydrogen bonds it becomes clear, that it is virtually impossible to construct a uni-
versal correlation in the whole energetic range of hydrogen bond strength, from a
few tenths of kcal mol−1 up to more than 40 kcal mol−1 . One of the disrupting fac-
tors is the dependency of frequencies and corresponding intensities on the mass of
vibrating atoms. The important step toward solution of this problem is transfer from
frequency to corresponding force constant as an abscissa axis for Badger–Bauer-like
correlations. As such, in Ref. [298] on the example of more than 60 complexes with
XH· · ·Y hydrogen bonds (X = F, Cl, Br, I; Y = O, N, C, S, P) in the energy range
4–51 kJ mol−1 it was proposed to use the force constant kσ that corresponds to the
low-frequency hydrogen bond vibration 𝜈 σ for estimations of a H-bond dissociation
energy:
D0 (kJ mol−1 ) = 1.53 ⋅ kσ (N m−1 ) − 1.8 (12.51)
Besides, the universal equation that is suitable for both hydrogen and halogen
bonds was proposed:
D0 (kJ mol−1 ) = 1.50(3) ⋅ kσ (N m−1 ) (12.52)
Equation (12.52) is the first example of an extension of Badger–Bauer-like cor-
relations to different non-covalent interactions. (Another example was given in
Ref. [299], where for a series of 128 halogen-bonded (CH3 )3 P=O· · ·Hal complexes a
linear correlation was proposed between shift of the P=O stretching band and the
complexation energy.) A few years later, the slope and intercept in Eq. (12.52) were
refined using a higher level of theory CCSD(T)/aug-cc-pVTZ and the respective
References 387
values of 1.38 and −0.6(12) were found (units are omitted for brevity, for details see
Ref. [300]).
For complexes with FH· · ·F hydrogen bonds with the ΔE > 25 kcal mol−1 it
was found that dependences of ΔE on force constants of proton donor stretching
vibration, ks , and low-frequency hydrogen bond vibration, kσ , are nonlinear [297],
so the following quadratic functions were proposed for evaluation of a hydrogen
bond strength:
ΔE(kcal mol−1 ) = −18 ⋅ Δks2 + 114 ⋅ ks (12.53)
ΔE(kcal mol−1 ) = −1156 ⋅ Δkσ2 + 490 ⋅ kσ (12.54)
where ks and kσ are in mdyne Å−1 .
12.3.3.6 Carbonyl Stretching Vibration

Among various possible vibrations of groups located close to the hydrogen bond
and their sensitivity to the hydrogen bond strength, one should note the work [267],
where for complexes CO· · ·HX (X = F, Cl, I, Br, CN, CCH), the blueshift of the car-
bonyl stretching band, 𝜈 CO , was considered as a marker of dissociation energy with
the following correlational equation:
D0 (cm−1 ) = 27.942 ⋅ Δ𝜈CO (cm−1 ) + 69.016 (12.55)
References
bond (IUPAC Recommendations 2011). Pure Appl. Chem. 83: 1637–1641.
2 Janoschek, R., Weidemann, E.G., Pfeiffer, H., and Zundel, G. (1972).
Extremely high polarizability of hydrogen bonds. J. Am. Chem. Soc. 94:
2387–2396.
3 Benoit, M. and Marx, D. (2005). The shapes of protons in hydrogen bonds
depend on the bond length. ChemPhysChem 6: 1738–1741.
4 Steiner, T. and Saenger, W. (1994). Lengthening of the covalent O–H bond
in O–H· · ·O hydrogen bonds re-examined from low-temperature neutron
diffraction data of organic compounds. Acta Crystallogr., Sect. B: Struct. Sci.
50: 348–357.
5 Steiner, T. (1995). Lengthening of the N–H bond in N–H· · ·N hydrogen bonds.
Preliminary structural data and implications of the bond valence concept.
J. Chem. Soc., Chem. Commun. 3: 1331–1332.
6 Steiner, T. (1998). Lengthening of the covalent X–H bond in heteronuclear
hydrogen bonds quantified from organic and organometallic neutron crystal
structures. J. Phys. Chem. A 102: 7041–7052.
7 Steiner, T. (2002). The hydrogen bond in the solid state. Angew. Chem. Int. Ed.
41: 48–76.
8 Gilli, P., Bertolasi, V., Ferretti, V., and Gilli, G. (1994). Evidence for resonance-
assisted hydrogen bonding. 4. Covalent nature of the strong homonuclear
hydrogen bond. Study of the O–H· · ·O system by crystal structure correlation
methods. J. Am. Chem. Soc. 116: 909–915.
9 Limbach, H.-H., Tolstoy, P.M., Perez-Hernandez, N. et al. (2009). OHO
hydrogen bond geometries and NMR chemical shifts: from equilibrium
structures to geometric H/D isotope effects with applications for water,
protonated water and compressed ice. Isr. J. Chem. 49: 199–216.
10 Limbach, H.-H., Pietrzak, M., Benedict, H. et al. (2004). Empirical corrections
for anharmonic zero-point vibrations of hydrogen and deuterium in geometric
hydrogen bond correlations. J. Mol. Struct. 706: 115–119.
11 Limbach, H.-H., Pietrzak, M., Sharif, S. et al. (2004). NMR parameters and
geometries of OHN and ODN hydrogen bonds of pyridine-acid complexes.
Chem. Eur. J. 10: 5195–5204.
12 Perrin, C.L. and Lau, J.S. (2006). Hydrogen-bond symmetry in zwitterionic
phthalate anions: symmetry breaking by solvation. J. Am. Chem. Soc.
36: 11820–11824.
81: 571–583.
14 Golubev, N.S. and Denisov, G.S. (1981). Kinetics of hydrogen bond formation
and breaking the life time of the complex (CF3 )3 COH – hexamethylphosphoric
triamide in freone solution. Dokl. Akad. Nauk 258: 546–549.
15 Limbach, H.-H., Denisov, G.S., Shenderovich, I.G., and Tolstoy, P.M. (2016).
Tautomerism in systems of increasing complexity: examples from organic
molecules to enzymes. In: Tautomerism: Concepts and Applications in Science
and Technology (ed. L. Antonov), 329–372. Wiley.
16 Koeppe, B., Tolstoy, P.M., and Limbach, H.-H. (2011). Reaction pathways of
proton transfer in hydrogen bonded phenol-carboxylate complexes explored by
combined UV–vis and NMR spectroscopy. J. Am. Chem. Soc. 133: 7897–7908.
17 Koeppe, B., Pylaeva, S.A., Allolio, C. et al. (2017). Polar solvent fluctua-
tions drive proton transfer in hydrogen bonded complexes of carboxylic acid
with pyridines: NMR, IR and ab initio MD study. Phys. Chem. Chem. Phys.
19: 1010–1028.
18 Siskos, M.G., Choudhary, M.I., and Gerothanassis, I.P. (2017). Hydrogen atomic
positions of O–H· · ·O hydrogen bonds in solution and in the solid state: the
synergy of quantum chemical calculations with 1 H-NMR chemical shifts and
X-ray diffraction methods. Molecules 22 (3): 415/1–415/32.
19 Wei, Y. and McDermott, A.E. (1999). Effects of hydrogen bonding on
1
H chemical shifts. In: ACS Symposium Series, vol. 732 (ed. J.C. Facelli and
A.C. de Dios), 177–193. Oxford University Press.
20 Jeffrey, G.A. and Yeon, Y. (1986). The correlation between hydrogen-bond
lengths and proton chemical shifts in crystals. Acta Crystallogr., Sect. B: Struct.
Sci. 42: 410–413.
References 389
21 Berglund, B. and Vaughan, R.W. (1980). Correlations between proton chemical

shift tensors, deuterium quadrupole couplings, and bond distances for hydrogen
bonds in solids. J. Chem. Phys. 73: 2037–2043.
22 Sternberg, U. and Brunner, E. (1994). The influence of short-range geometry
on the chemical shift of protons in hydrogen bonds. J. Magn. Reson., Ser. A
108: 142–150.
23 Harris, T.K., Zhao, Q., and Mildvan, A.S. (2000). NMR studies of strong hydro-
gen bonds in enzymes and in a model compound. J. Mol. Struct. 552: 97–109.
24 McDermott, A. and Ridenour, C.F. (1996). Proton chemical shift measure-
ments in biological solids. In: Encyclopedia of Nuclear Magnetic Resonance
(ed. D.M. Grant and R.K. Harris), 3820. Wiley.
25 Abraham, R.J. and Mobli, M. (2007). An NMR, IR and theoretical investigation
of 1 H chemical shifts and hydrogen bonding in phenols. Magn. Reson. Chem.
45: 865–877.
26 Siskos, M.G., Iqbal Choudhary, M., and Gerothanassis, I.P. (2017). Hydrogen
atomic positions of O–H· · ·O hydrogen bonds in solution and in the solid state:
the synergy of quantum chemical calculations with 1H-NMR chemical shifts
and X-ray diffraction methods. Molecules 22 (3): 415.
27 Benedict, H., Shenderovich, I.G., Malkina, O.L. et al. (2000). Nuclear scalar
spin–spin couplings and geometries of hydrogen bonds. J. Am. Chem. Soc.
122: 1979–1988.
28 Golubev, N.S. (1984). Study of H3 O+ and H5 O2 + ion formation with
NMR-spectra of H-1 high-resolution. Khim. Fiz. 3: 772–774.
29 Schah-Mohammedi, P., Shenderovich, I.G., Detering, C. et al. (2000). Hydrogen/
deuterium isotope effects on NMR chemical shifts of formally symmetric
complexes with a strong intermolecular hydrogen bond in liquid solutions
at 100–150 K. J. Am. Chem. Soc. 122: 12878–12879.
30 Tolstoy, P.M., Schah-Mohammedi, P., Smirnov, S.N. et al. (2004). Character-
ization of fluxional hydrogen-bonded complexes of acetic acid and acetate
by NMR: geometries and isotope and solvent effects. J. Am. Chem. Soc.
126: 5621–5634.
31 Guo, J., Tolstoy, P.M., Koeppe, B. et al. (2012). Hydrogen bond geometries
and proton tautomerism of homo-conjugated anions of carboxylic acids
studied via H/D isotope effects on 13 C NMR chemical shifts. J. Phys. Chem.
A 116: 11180–11188.
32 Murakhtina, T., Heuft, J., Meijer, E.J., and Sebastiani, D. (2006). First principles
and experimental 1 H NMR signatures of solvated ions: the case of HCl(aq).
33 Pylaeva, S., Allolio, C., Koeppe, B. et al. (2015). Proton transfer in a short
hydrogen bonded complex caused by solvation shell fluctuations: an ab initio
MD and NMR/UV study of an (OHO)-bonded system. Phys. Chem. Chem. Phys.
17: 4634–4644.
34 Ubbelohde, A.R. and Gallagher, K.J. (1955). Acid–base effects in hydrogen
bonds in crystals. Acta Cryst. 8: 71–83.
35 Engelen, B. and Lutz, H.D. (2002). Hydrogen bonds in inorganic solids. Trends
Appl. Spectrosc. 4: 355–375.
36 Ando, S., Ando, I., Shoji, A., and Ozaki, T. (1988). Intermolecular hydrogen-
bonding effect on carbon-13 NMR chemical shifts of glycine residue carbonyl
carbons of peptides in the solid state. J. Am. Chem. Soc. 110: 3380–3386.
37 Asakawa, N., Kuroki, S., Kurosu, H. et al. (1992). Hydrogen-bonding effect
on carbon-13 NMR chemical shifts of L-alanine residue carbonyl carbons of
peptides in the solid state. J. Am. Chem. Soc. 114: 3261–3265.
38 Asakawa, N., Kurosu, H., Ando, I. et al. (1994). A structural study of peptides
and proteins containing L-alanine residues by 13 C NMR spectroscopy combined
with ab initio chemical shift calculations. J. Mol. Struct. 317: 119–129.
39 McDermott, A. and Gu, Z. (2012). Carbon and nitrogen chemical shifts of solid
state enzymes. In: Encyclopedia of Magnetic Resonance (ed. R.K. Harris and
R.E. Wasylishen), 1–11. Wiley.
40 Gu, Z., Zambrano, R., and McDermott, A. (1994). Hydrogen bonding of
carboxyl groups in solid-state amino acids and peptides: comparison of
carbon chemical shielding, infrared frequencies, and structures. J. Am. Chem.
Soc. 116: 6368–6372.
41 Gu, Z. and McDermott, A. (1993). Chemical shielding anisotropy of
protonated and deprotonated carboxylates in amino acids. J. Am. Chem. Soc.
115: 4282–4285.
42 Facelli, J.C., Gu, Z., and McDermott, A. (1995). Carbon-13 chemical shift
tensors of carboxylic acids: GIAO calculations in acetic acid + methylamine
dimer. Mol. Phys. 86: 865–872.
43 Kumar, G.A., Pan, Y., Smallwood, C.J., and McAllister, M.A. (1998). Low-
barrier hydrogen bonds: ab initio and DFT investigation. J. Comput. Chem.
19: 1345–1352.
44 Brueck, A., McCoy, L.J., and Kilway, K.V. (2000). Hydrogen bonds in carboxylic
acid–carboxylate systems in solution. 1. In anhydrous, aprotic media. Org. Lett.
2: 2007–2009.
45 Harmon, K.M., Shaw, K.E., and Sadeki, S.M. (2000). Hydrogen bonding.
Part 75. Infrared, titrametric and molecular orbital study of the structure,
solubility and reaction with fluoride ion of 2,3-dicarboxypyrazine. J. Mol. Struct.
526: 191–199.
46 Perrin, C.L. and Nielson, J.B. (1997). Asymmetry of hydrogen bonds in
solutions of monoanions of dicarboxylic acids. J. Am. Chem. Soc.
119: 12734–12741.
47 Luecke, H., Schobert, B., Richter, H.T. et al. (1999). Structure of bacteri-
orhodopsin at 1.55 Å resolution. J. Mol. Biol. 291: 899–911.
48 Smith, R., Brereton, I.M., Chai, R.Y., and Kent, S.B.H. (1996). Ionization states
of the catalytic residues in HIV-1 protease. Nat. Struct. Biol. 3: 946–950.
49 Mildvan, A.S., Massiah, M.A., Harris, T.K. et al. (2002). Short, strong hydrogen
bonds on enzymes: NMR and mechanistic studies. J. Mol. Struct. 615: 163–175.
References 391
50 O’Leary, D.J., Hickstein, D.D., Hansen, B.K., and Hansen, P.E. (2010).
Theoretical and NMR studies of deuterium isotopic perturbation of hydrogen
bonding in symmetrical dihydroxy compounds. J. Org. Chem. 75: 1331–1342.
51 Dziembowska, T., Hansen, P.E., and Rozwadowski, Z. (2004). Studies based
on deuterium isotope effect on 13 C chemical shifts. Prog. Nucl. Magn. Reson.
Spectrosc. 1: 1–29.
52 Grech, E., Klimkiewicz, J., Nowicka-Scheibe, J. et al. (2002). Deuterium isotope
effects on N-15, C-13 and H-1 chemical shifts of proton sponges. J. Mol. Struct.
615: 121–140.
53 Filarowski, A., Koll, A., Rospenk, M. et al. (2005). Tautomerism of sterically
hindered Schiff bases. Deuterium isotope effects on C-13 chemical shifts.
J. Phys. Chem. A 109: 4464–4473.
54 Hansen, P.E. (2008). Deuterium isotope effects on C-13 chemical shifts of nitro-
malonamide. Magn. Reson. Chem. 46: 726–729.
55 Sosnicki, J.G. and Hansen, P.E. (2005). New deuterium isotope effects on
C-13 and F-19 chemical shifts across intramolecular hydrogen bonds of
non-resonance assisted systems. Tetrahedron Lett. 46: 839–842.
56 Tolstoy, P.M., Guo, J., Koeppe, B. et al. (2010). Geometries and tautomerism of
OHN hydrogen bonds in aprotic solution probed by H/D isotope effects on 13 C
NMR chemical shifts. J. Phys. Chem. A 114: 10775–10782.
57 Kühl, O. (2008). Phosphorus-31 NMR Spectroscopy: A Concise Introduction for
the Synthetic. Springer Science & Business Media.
58 Shenderovich, I.G. (2013). Effect of non-covalent interactions on the 31 P
chemical shift tensor of phosphine oxides, phosphinic, phosphonic, and
phosphoric acids and their complexes with lead(II). J. Phys. Chem. C
117: 26689–26702.
59 Begimova, G.U., Tupikina, E.Yu., Denisov, G.S. et al. (2016). Effect of hydrogen
bonding to water on the 31 P chemical shift tensor of phenyl- and trialkylphos-
phine oxides and α-amino phosphonates. J. Phys. Chem. C 120: 8717–8729.
60 Alkorta, I. and Elguero, J. (2017). Is it possible to use the 31 P chemical shifts of
phosphines to measure hydrogen bond acidities (HBA)? A comparative study
with the use of the 15 N chemical shifts of amines for measuring HBA. J. Phys.
Org. Chem. 30 (11): e3690/1–e3690/7.
61 Mulloyarova, V.V., Giba, I.S., Kostin, M.A. et al. (2018). Cyclic trimers of
phosphinic acids in polar aprotic solvent: symmetry, chirality and H/D isotope
effects on NMR chemical shifts. Phys. Chem. Chem. Phys. 20: 4901–4910.
62 Mulloyarova, V.V., Ustimchuk, D.O., Filarowski, A., and Tolstoy, P.M. (2020).
H/D isotope effects on 1 H-NMR chemical shifts in cyclic heterodimers
and heterotrimers of phosphinic and phosphoric acids. Molecules
25 (8): 1907/1–1907/17.
63 Giba, I.S. and Tolstoy, P.M. (2021). Self-assembly of hydrogen-bonded cage
tetramers of phosphonic acid. Symmetry 13: 258/1–258/12.
64 Giba, I.S., Mulloyarova, V.V., Denisov, G.S., and Tolstoy, P.M. (2019). Influence
of hydrogen bonds in 1:1 complexes of phosphinic acids with substituted
pyridines on 1 H and 31 P NMR chemical shifts. J. Phys. Chem. A 123: 2252–2260.
65 Giba, I.S., Mulloyarova, V.V., Denisov, G.S., and Tolstoy, P.M. (2020). Sensitivity
of 31 P NMR chemical shifts to hydrogen bond geometry and molecular
conformation for complexes of phosphinic acids with pyridines. Magn. Reson.
Chem. 59: 465–477.
66 Mulloyarova, V.V., Giba, I.S., Denisov, G.S. et al. (2019). Conformational
mobility and proton transfer in hydrogen-bonded dimers and trimers of
phosphinic and phosphoric acids. J. Phys. Chem. A 123: 6761–6771.
67 Detering, C., Tolstoy, P.M., Golubev, N.S. et al. (2001). Vicinal H/D isotope
effects in NMR spectra of complexes with coupled hydrogen bonds: phosphoric
acids. Dokl. Phys. Chem. 379: 191–193.
68 Mayer, U., Gutmann, V., and Gerger, W. (1975). The acceptor number – a
quantitative empirical parameter for the electrophilic properties of solvents.
Monatsh. Chem. 106: 1235–1257.
69 Pahl, J., Brand, S., Elsen, H., and Harder, S. (2018). Highly Lewis acidic
cationic alkaline earth metal complexes. Chem. Commun. 54: 8685–8688.
70 Beckett, M.A., Brassington, D.S., Coles, S.J., and Hursthouse, M.B. (2000). Lewis
acidity of tris(pentafluorophenyl)borane: crystal and molecular structure of
B(C6 F5 )3 ⋅OPEt3 . Inorg. Chem. Commun. 3: 530–533.
71 Kolling, O.W. (1984). Triethylphosphine oxide as a probe of weak hydrogen
bond donor behaviour. Trans. Kansas Acad. Sci. 87: 115–118.
72 Byrne, E., O’Donnell, R., Gilmore, M. et al. (2020). Hydrophobic functional
liquids based on trioctylphosphine oxide (TOPO) and carboxylic acids. Phys.
Chem. Chem. Phys. 22: 24744–24763.
73 Gilmore, M., McCourt, É., Connolly, F. et al. (2018). Hydrophobic deep eutectic
solvents incorporating trioctylphosphine oxide: advanced liquid extractants. ACS
Sustainable Chem. Eng. 6: 17323–17332.
74 Diemoz, K. and Franz, A. (2019). NMR quantification of hydrogen-bond-
activating effects for organocatalysts including boronic acids. J. Org. Chem.
84: 1126–1138.
75 Zheng, A., Bin Liu, S., and Deng, F. (2017). 31 P NMR chemical shifts of
phosphorus probes as reliable and practical acidity scales for solid and liquid
catalysts. Chem. Rev. 117: 12475–12531.
76 McCune, J.A., He, P., Petkovic, M. et al. (2014). Brønsted acids in ionic
liquids: how acidity depends on the liquid structure. Phys. Chem. Chem. Phys.
16: 23233–23243.
77 Osegovic, J.P. and Drago, R.S. (2000). Measurement of the global acidity of solid
acids by 31 P MAS NMR of chemisorbed triethylphosphine oxide. J. Phys. Chem.
B 104: 147–154.
78 Pires, E. and Fraile, J.M. (2020). Study of interactions between Brønsted acids
and triethylphosphine oxide in solution by 31 P NMR: evidence for 2:1 species.
Phys. Chem. Chem. Phys. 22: 24351–24358.
79 Lorente, P., Shenderovich, I.G., Buntkowsky, G. et al. (2001). 1 H/15 N NMR
chemical shielding, dipolar 15 N,2 H coupling and hydrogen bond geometry cor-
relations in a novel series of hydrogen-bonded acid–base complexes of collidine
with carboxylic acids. Magn. Reson. Chem. 39: S18–S29.
References 393
80 Viragh, C., Harris, T.K., Reddy, P.M. et al. (2000). NMR evidence for a short,
strong hydrogen bond at the active site of a cholinesterase. Biochemistry
39: 16200–16205.
81 Wei, Y., De Dios, A., and McDermott, A.E. (1999). Solid-state 15 N NMR
chemical shift anisotropy of histidines: experimental and theoretical studies
of hydrogen bonding. J. Am. Chem. Soc. 121: 10389–10394.
82 Hoelger, C.G., Aguilar-Parilla, F., Elguero, J. et al. (1996). 15 N chemical shift
tensors and hydrogen bond geometries of 3,5-substituted pyrazoles and their
orientation in the molecular frame. J. Magn. Reson., Ser. A 120: 46–55.
83 Andreeva, D.V., Ip, B., Gurinov, A.A. et al. (2006). Geometrical features of
hydrogen bonded complexes involving sterically hindered pyridines. J. Phys.
Chem. A 110: 10872–10879.
84 Dingley, A.J., Cordier, F., and Grzesiek, S. (2001). An introduction to hydrogen
bond scalar couplings. Concepts Magn. Reson. 13: 103–127.
85 Cornilescu, G., Ramirez, B.E., Frank, M.K. et al. (1999). Correlation between
3h
J NC′ and hydrogen bond length in proteins. J. Am. Chem. Soc. 121: 6275–6279.
86 Bachovchin, W.W. (1986). 15 N NMR spectroscopy of hydrogen-bonding inter-
actions in the active site of serine proteases: evidence for a moving histidine
mechanism. Biochemistry 25: 7751–7759.
87 Semenov, V.A., Samultsev, D.O., and Krivdin, L.B. (2015). Theoretical and
experimental study of 15 N NMR protonation shifts. Magn. Reson. Chem.
53: 433–441.
88 Sharif, S., Fogle, E., Toney, M.D. et al. (2007). NMR localization of protons in
critical enzyme hydrogen bonds. J. Am. Chem. Soc. 129: 9558–9559.
89 Shenderovich, I.G., Buntkowsky, G., Schreiber, A. et al. (2003). Pyridine-15 N a
mobile NMR sensor for surface acidity and surface defects of mesoporous silica.
J. Phys. Chem. B 107: 11924–11939.
90 Alkorta, I. and Elguero, J. (2003). GIAO calculations of chemical shifts in
heterocyclic compounds. Struct. Chem. 14: 377–389.
91 Kuroki, S., Ando, S., Ando, I. et al. (1990). Hydrogen-bonding effect on 15 N
NMR chemical shifts of the glycine residue of oligopeptides in the solid state
as studied by high-resolution solid-state NMR spectroscopy. J. Mol. Struct.
240: 19–29.
92 Marek, R., Lycka, A., Kolehmainen, E. et al. (2007). 15 N NMR spectroscopy in
structural analysis: an update (2001–2005). Curr. Org. Chem. 11: 1154–1205.
93 Xin, D., Sader, C.A., Fischer, U. et al. (2017). Systematic investigation of
DFT-GIAO 15 N NMR chemical shift prediction using B3LYP/cc-pVDZ:
application to studies of regioisomers, tautomers, protonation states and
N-oxides. Org. Biomol. Chem. 15: 928–936.
94 Bertani, P., Raya, J., and Bechinger, B. (2014). 15 N chemical shift referencing in
solid state NMR. Solid State Nucl. Magn. Reson. 61–62: 15–18.
95 Xiang, S.Q., Narayanan, R.L., Becker, S., and Zweckstetter, M. (2013). N–H
spin–spin couplings: probing hydrogen bonds in proteins. Angew. Chem. Int. Ed.
52: 3525–3528.
96 Cremera, D. and Gräfenstein, J. (2007). Calculation and analysis of NMR

spin–spin coupling constants. Phys. Chem. Chem. Phys. 9: 2791–2816.
97 Hansen, P.E., Saeed, B.A., Rutu, R.S., and Kupka, T. (2020). One-bond 1 J(15 N,H)
coupling constants at sp2 hybridized nitrogen of Schiff bases, enaminones and
similar compounds. A theoretical study. Magn. Reson. Chem. 58: 750–762.
98 Tolstoy, P.M., Smirnov, S.N., Shenderovich, I.G. et al. (2004). NMR studies of
solid state – solvent and H/D isotope effects on hydrogen bond geometries of
1:1 complexes of collidine with carboxylic acids. J. Mol. Struct. 700: 19–27.
99 Shenderovich, I.G., Tolstoy, P.M., Golubev, N.S. et al. (2003). Low-temperature
NMR studies of the structure and dynamics of a novel series of acid–base
complexes of HF with collidine exhibiting scalar couplings across hydrogen
bonds. J. Am. Chem. Soc. 125: 11710–11720.
100 Grzesiek, S., Cordier, F., Jaravine, V., and Barfield, M. (2004). Insights into
biomolecular hydrogen bonds from hydrogen bond scalar couplings. Prog. Nucl.
Magn. Reson. Spectrosc. 45: 275–300.
101 Alkorta, I. and Elguero, J. (2003). Review on DFT and ab initio calculations of
scalar coupling constants. Int. J. Mol. Sci. 4: 64–92.
102 Del Bene, J.E. and Elguero, J. (2004). Predicted signs of one-bond spin–spin
coupling constants (1h J HY ) across X–H–Y hydrogen bonds for complexes with
Y = 15 N, 17 O, and 19 F. J. Phys. Chem. A 108: 11762–11767.
103 Zarycz, N. and Aucar, G.A. (2008). Theoretical NMR spectroscopic analysis
of the intramolecular proton transfer mechanism in ortho-hydroxyaryl
(un-)substitued Schiff bases. J. Phys. Chem. A 112: 8767–8774.
104 Kjær, H., Sauer, S.P.A., and Kongsted, J. (2010). Benchmarking NMR indirect
nuclear spin–spin coupling constants: SOPPA, SOPPA (CC2), and SOPPA
(CCSD) versus CCSD. J. Chem. Phys. 133: 144106/1–144106/14.
105 Yachmenev, A., Yurchenko, S.N., Paidarová, I. et al. (2010). Thermal averaging
of the indirect nuclear spin–spin coupling constants of ammonia: the
importance of the large amplitude inversion mode. J. Chem. Phys.
132: 114305/1–114305/16.
106 Del Bene, J.E., Alkorta, I., and Elguero, J. (2009). A systematic comparison of
second-order polarization propagator approximation and equation-of-motion
coupled cluster singles and doubles C–C, C–N, N–N, C–H, and N–H spin–spin
coupling constants. J. Phys. Chem. A 113: 12411–12420.
107 Denisov, G.S. and Schreiber, V.M. (1972). Infrared study of the interaction of
pentachlorophenol with secondary amines. Spectrosc. Lett. 5: 377–384.
108 Denisov, G.S. and Schreiber, V.M. (1974). The evolution of the potential surface
of chlorophenols – amines interaction. Dokl. Akad. Sci. 215: 310–313.
109 Dega-Szafran, Z. and Dulewicz, E. (1981). 1 H NMR studies of solvent effects
on hydrogen bonding in some pyridine trifluoroacetates. Org. Magn. Reson.
16: 214–219.
110 Perrin, C.L., Lau, J.S., and Ohta, B.K. (2003). Low-barrier hydrogen bonds in
pyridine–dichloroacetic acid complexes. Pol. J. Chem. 77: 1693–1703.
111 Sharif, S., Denisov, G.S., Toney, M.D., and Limbach, H.-H. (2006). NMR studies
of solvent-assisted proton transfer in a biologically relevant Schiff base: toward
References 395
a distinction of geometric and equilibrium H-bond isotope effects. J. Am. Chem.

Soc. 128: 3375–3387.
112 Takeda, S., Inabe, T., Benedict, C. et al. (1998). Proton dynamics in interacting
hydrogen bonds in the solid state: proton tunneling in the NHO hydrogen
bonds of N,N ′ -di(2-hydroxy-1-naphthylmethylene)-p-phenylenediamine. Ber.
Bunsen Ges. Phys. Chem. 102: 1358–1369.
113 Rospenk, M., Sobczyk, L., Schah-Mohammedi, P. et al. (2001). Dimerization
and solvent assisted proton dislocation in the low-barrier hydrogen bond of a
Mannich base. A low-temperature NMR study. Magn. Reson. Chem.
39: S81–S90.
114 Jezierska, A., Tolstoy, P., Panek, J., and Filarowski, A. (2019). Intramolecular
hydrogen bonds in selected aromatic compounds: recent developments.
Catalysts 9 (11): 909/1–909/19.
115 Elias, R.S. (2012). Theoretical study of the proton transfer in enaminones. Am.
J. Appl. Sci. 9: 103–106.
116 Gilli, P., Bertolasi, V., Ferretti, V., and Gilli, G. (2000). Evidence for intramolecu-
lar N–H· · ·O resonance-assisted hydrogen bonding in β-enaminones and related
heterodienes. A combined crystal-structural, IR and NMR spectroscopic, and
quantum-mechanical investigation. J. Am. Chem. Soc. 122: 10405–10417.
117 Rozwadowski, Z. and Dziembowska, T. (1999). Proton transfer equilibrium in
Schiff bases derived from 5-nitrosalicylaldehyde. A study of deuterium isotope
effects on 13 C NMR chemical shifts. Magn. Reson. Chem. 37: 274–278.
118 Dziembowska, T., Rozwadowski, Z., Filarowski, A., and Hansen, P.E. (2001).
NMR study of proton transfer equilibrium in Schiff bases derived from
2-hydroxy-1-naphthaldehyde and 1-hydroxy-2-acetonaphthone. Deuterium
isotope effects on 13 C and 15 N chemical shifts. Magn. Reson. Chem. 39: S67–S80.
119 Tupikina, E.Yu., Sigalov, M., Shenderovich, I.G. et al. (2019). Correlations
of NHN hydrogen bond energy with geometry and 1 H NMR chemical
shift difference of NH protons for aniline complexes. J. Chem. Phys.
150: 114305/1–114305/10.
120 Ueda, T., Nagatomo, S., Masui, H. et al. (1999). Hydrogen bonds in crystalline
imidazoles studied by 15 N NMR and ab initio MO Calculations. Z. Naturforsch.,
A: Phys. Sci. 54: 437–442.
121 Del Bene, J.E., Ajith Perera, S., and Bartlett, R.J. (2001). What parameters
determine N–N and O–O coupling constants (2h J X–X ) across X–H+ –X hydrogen
bonds? J. Phys. Chem. A 105: 930–934.
122 Pietrzak, M., Try, A., Andrioletti, B. et al. (2008). Angew. Chem. Int. Ed.
47: 1123–1126.
123 Pietrzak, M., Benedict, C., Gehring, H. et al. (2007). J. Mol. Struct.
844–845: 222–231.
124 Del Bene, J.E. and Bartlett, R.J. (2000). N–N spin–spin coupling constants
[2h J(15 N–15 N)] across N–H· · ·N hydrogen bonds in neutral complexes: to what
extent does the bonding at the nitrogens influence 2h J N–N ? J. Am. Chem. Soc.
122: 10480–10481.
125 Del Bene, J.E., Ajith Perera, S., and Bartlett, R.J. (2001). 15 N,15 N spin–spin
coupling constants across N–H–N and N–H+ –N hydrogen bonds: can coupling
constants provide reliable estimates of N–N distances in biomolecules? Magn.
Reson. Chem. 39: S109–S114.
126 Dingley, A.J., Masse, J.E., Peterson, R.D. et al. (1999). Internucleotide scalar
couplings across hydrogen bonds in Watson–Crick and Hoogsteen base pairs of
a DNA triplex. J. Am. Chem. Soc. 121: 6019–6027.
127 Barfield, M., Dingley, A.J., Feigon, J., and Grzesiek, S. (2001). A DFT study
of the interresidue dependencies of scalar J-coupling and magnetic shielding
in the hydrogen-bonding regions of a DNA triplex. J. Am. Chem. Soc.
123: 4014–4022.
128 Del Bene, J.E. and Elguero, J. (2006). Systematic ab initio study of 15 N–15 N and
15
N–1 H spin–spin coupling constants across N–H+ –N hydrogen bonds: pre-
dicting N–N and N–H coupling constants and relating them to hydrogen bond
type. J. Phys. Chem. A 110: 7496–7502.
129 Del Bene, J.E., Alkorta, I., and Elguero, J. (2008). Spin–spin coupling across
intramolecular N–H+ –N hydrogen bonds in models for proton sponges: an ab
initio investigation. Magn. Reson. Chem. 46: 457–463.
130 Shenderovich, I.G., Smirnov, S.N., Denisov, G.S. et al. (1998). Nuclear magnetic
resonance of hydrogen bonded clusters between F− and (HF)n : experiment and
theory. Ber. Bunsen Ges. Phys. Chem. 102: 422–428.
131 Shenderovich, I.G., Limbach, H.-H., Smirnov, S.N. et al. (2002). H/D isotope
effects on the low-temperature NMR parameters and hydrogen bond geometries
of (FH)2 F− and (FH)3 F− dissolved in CDF3 /CDF2 Cl. Phys. Chem. Chem. Phys.
4: 5488–5497.
132 Wenthold, P.G. and Squires, R.R. (1995). Bond dissociation energies of F2 −
and HF2 − . A gas-phase experimental and G2 theoretical study. J. Phys. Chem.
99: 2002–2005.
133 Stein, C., Oswald, R., Sebald, P. et al. (2013). Accurate bond dissociation ener-
gies (D-0) for FHF− isotopologues. Mol. Phys. 111: 2647–2652.
134 Kawaguchi, K. and Hirota, E. (1986). Infrared diode laser study of the hydrogen
bifluoride anion: FHF− and FDF− . J. Chem. Phys. 84: 2953–2960.
135 Almlöf, J. (1972). Hydrogen bond studies. 71. Ab initio calculation of the
vibrational structure and equilibrium geometry in HF2 − and DF2 − . Chem.
Phys. Lett. 17: 49–52.
136 Del Bene, J.E., Jordan, M.J.T., Perera, S.A., and Bartlett, R.J. (2001). Vibrational
effects on the F–F spin–spin coupling constant (2h J(F–F)) in FHF− and FDF− .
J. Phys. Chem. A 105: 8399–8402.
137 Hirata, S., Yagi, K., Perera, S.A. et al. (2008). Anharmonic vibrational frequen-
cies and vibrationally averaged structures and nuclear magnetic resonance
parameters of FHF− . J. Chem. Phys. 128: 214305.
138 Kawaguchi, K. and Hirota, E. (1987). Diode laser spectroscopy of the ν3 and ν2
bands of FHF− in 1300 cm−1 region. J. Chem. Phys. 87: 6838–6841.
References 397
139 Del Bene, J.E., Bartlett, R.J., and Elguero, J. (2002). Interpreting 2h J(F,N),
1h
J(H,N) and 1 J(F,H) in the hydrogen-bonded FH–collidine complex. Magn.
Reson. Chem. 40: 767–771.
140 Tupikina, E.Yu., Denisov, G.S., Melikova, S.M. et al. (2018). New look at the
Badger–Bauer rule: correlations of spectroscopic IR and NMR parameters
with hydrogen bond energy and geometry. FHF complexes. J. Mol. Struct.
1164: 129–136.
141 Pylaeva, S.A., Elgabarty, H., Sebastiani, D., and Tolstoy, P.M. (2017). Symmetry
and dynamics of FHF− anion in vacuum, in CD2 Cl2 and in CCl4 . Ab initio MD
study of fluctuating solvent–solute hydrogen and halogen bonds. Phys. Chem.
Chem. Phys. 19: 26107–26120.
142 Dunitz, J.D. and Taylor, R. (1997). Organic fluorine hardly ever accepts
hydrogen bonds. Chem. Eur. J. 3: 89–98.
143 Cormanich, R.A., Freitas, M.P., Tormena, C.F., and Rittnera, R. (2012). The
F· · ·HO intramolecular hydrogen bond forming five-membered rings hardly
appear in monocyclic organofluorine compounds. RSC Adv. 2: 4169–4174.
144 Mishra, S.K. and Suryaprakash, N. (2017). Intramolecular hydrogen bonding
involving organic fluorine: NMR investigations corroborated by DFT-based
theoretical calculations. Molecules 22: 423.
145 Emsley, J. (1971). Glacial acetic acid as a non-aqueous solvent for metal
fluorides. J. Chem. Soc. 2511–2512.
146 Clark, J.H. and Emsley, J. (1974). Very strong hydrogen bonds formed between
carboxylic acids and anions: trifluoroacetic and acetic acids with acetates and
fluoride. J. Chem. Soc., Dalton Trans. 11: 1125–1129.
147 Golubev, N.S., Tolstoy, P.M., Smirnov, S.N. et al. (2004). Low-temperature NMR
spectra of fluoride–acetic acid hydrogen-bonded complexes in aprotic polar
environment. J. Mol. Struct. 700: 3–12.
148 Cametti, M. and Rissanen, K. (2009). Recognition and sensing of fluoride anion.
Chem. Commun. 20: 2809.
149 Evans, N.H. and Beer, P.D. (2014). Advances in anion supramolecular
chemistry: from recognition to chemical applications. Angew. Chem. Int. Ed.
53: 11716–11754.
150 Busschaert, N., Caltagirone, C., Van Rossom, W., and Gale, P.A. (2015).
Applications of supramolecular anion recognition. Chem. Rev. 115: 8038–8155.
151 Ibba, F., Pupo, G., Thompson, A.L. et al. (2020). Impact of multiple hydrogen
bonds with fluoride on catalysis: insight from NMR spectroscopy. J. Am. Chem.
Soc. 142: 19731–19744.
152 Golubev, N.S., Shenderovich, I.G., Smirnov, S.N. et al. (1999). Nuclear scalar
spin–spin coupling reveals novel properties of low-barrier hydrogen-bonds in a
polar environment. Chem. Eur. J. 5: 492–497.
153 Shenderovich, I.G., Burtsev, A.P., Denisov, G.S. et al. (2001). Influence of the
temperature-dependent dielectric constant on the H/D isotope effects on the
NMR chemical shifts and the hydrogen bond geometry of the collidine-HF
complex in CDF3 /CDClF2 solution. Magn. Reson. Chem. 39: S91–S99.
154 Del Bene, J.E. and Elguero, J. (2005). Ab initio study of complexes with two
cations as N–H donors to F− : structures and spin–spin coupling constants
across N–H–F hydrogen bonds. J. Phys. Chem. A 109: 10753–10758.
155 Del Bene, J.E. (2004). Characterizing two-bond NMR 13 C–15 N, 15 N–15 N, and
19
F–15 N spin–spin coupling constants across hydrogen bonds using ab initio
EOM-CCSD calculations. In: Calculation of NMR and EPR Parameters: Theory
and Applications (ed. M. Kaupp, M. Bühl and V.G. Malkin), 353–370. Wiley.
156 Gilli, G. and Gilli, P. (2009). The Nature of the Hydrogen Bond: Outline of a
Comprehensive Hydrogen Bond Theory. Oxford University Press.
157 Smirnov, S.N., Golubev, N.S., Denisov, G.S. et al. (1996). Hydrogen/deuterium
isotope effects on the NMR chemical shifts and geometries of intermolecular
low-barrier hydrogen-bonded complexes. J. Am. Chem. Soc. 118: 4094–4101.
158 Bureiko, S.F., Golubev, N.S., Denisov, G.S. et al. (2005). 3,5-Dimethylpyrazole
complexes with strong carboxylic acids. Russ. J. Gen. Chem. 75: 1821–1829.
159 Ajami, D., Tolstoy, P.M., Dube, H. et al. (2011). Compressed hydrogen bonds
isolated in encapsulation complexes. Angew. Chem. Int. Ed. 50: 528–531.
160 Mulloyarova, V.V., Puzyk, A.M., Efimova, A.A. et al. (2021). Solid-state and
solution-state self-association of dimethylarsinic acid: IR, NMR and theoretical
study. J. Mol. Struct. 1234: 130176.
161 Golubev, N.S., Smirnov, S.N., Tolstoy, P.M. et al. (2007). Observation by NMR of
the tautomerism of an intramolecular OHOHN-charge relay chain in a model
Schiff base. J. Mol. Struct. 844: 319–327.
162 Koeppe, B., Tolstoy, P.M., Guo, J. et al. (2021). Combined NMR and UV–Vis
spectroscopic studies of models for the hydrogen bond system in the active
site of photoactive yellow protein: H-bond cooperativity and medium effects.
J. Phys. Chem. B 125: 5874–5884.
163 Sigala, P.A., Tsuchida, M.A., and Herschlag, D. (2009). Hydrogen bond dynam-
ics in the active site of photoactive yellow protein. Proc. Natl. Acad. Sci. U.S.A.
106: 9232–9237.
164 Rundle, R.E. and Parasol, M. (1952). O–H stretching frequencies in very short
and possibly symmetrical hydrogen bonds. J. Chem. Phys. 20: 1487–1488.
165 Lord, R.C. and Merrifield, R.E. (1953). Strong hydrogen bonds in crystals.
J. Chem. Phys. 21: 166–167.
166 Nakamoto, K., Margoshes, M., and Rundle, R.E. (1955). Stretching frequencies
as a function of distances in hydrogen bonds. J. Am. Chem. Soc. 77: 6480–6486.
167 Ratajczak, H. and Orville-Thomas, W.J. (1968). Hydrogen-bond studies. J. Mol.
Struct. 1: 449–461.
168 Falk, M. and Wójcik, M.J. (1979). Effect of temperature and hydrogen-bond
strength on the ir absorption of isolated OH oscillators: crystalline L-ascorbic
acid. Spectrochim. Acta, Part A 35: 1117–1123.
169 Novak, A. (1974). Hydrogen bonding in solids correlation of spectroscopic
and crystallographic data. In: Large Molecules (ed. J.-H. Fuhrhop, G. Blauer,
T.J.R. Weakley and A. Novak), 177–216. Berlin, Heidelberg: Springer.
References 399
170 Lautié, A., Froment, F., and Novak, A. (1976). Relationship between NH
stretching frequencies and N· · ·O distances of crystals containing NH· · ·O
hydrogen bonds. Spectrosc. Lett. 9: 289–299.
171 Mikenda, W. (1986). Stretching frequency versus bond distance correlation of
O–D(H)· · ·Y (Y = N, O, S, Se, Cl, Br, I) hydrogen bonds in solid hydrates. J. Mol.
Struct. 147: 1–15.
172 Lampert, H., Mikenda, W., and Karpfen, A. (1997). Molecular geometries and
vibrational spectra of phenol, benzaldehyde, and salicylaldehyde: experimental
versus quantum chemical data. J. Phys. Chem. A 101: 2254–2263.
173 Kryachko, E.S. and Zeegers-Huyskens, T. (2001). Theoretical study of the
CH· · ·O interaction in fluoromethanes⋅H2 O and chloromethanes⋅H2 O
complexes. J. Phys. Chem. A 105: 7118–7125.
174 Rozenberg, M., Loewenschuss, A., and Marcus, Y. (2000). An empirical
correlation between stretching vibration redshift and hydrogen bond length.
Phys. Chem. Chem. Phys. 2: 2699–2702.
175 Rozenberg, M., Shoham, G., Reva, I., and Fausto, R. (2005). A correlation
between the proton stretching vibration red shift and the hydrogen bond
length in polycrystalline amino acids and peptides. Phys. Chem. Chem. Phys.
7: 2376–2383.
176 Rozenberg, M., Jung, C., and Shoham, G. (2003). Ordered and disordered
hydrogen bonds in adenosine, cytidine and uridine studied by low temperature
FT infrared spectroscopy. Phys. Chem. Chem. Phys. 5: 1533–1535.
177 Hansen, P.E. and Spanget-Larsen, J. (2017). NMR and IR investigations of
strong intramolecular hydrogen bonds. Molecules 22: 552/1–552/21.
178 Vojta, D., Dominković, K., Miljanić, S., and Spanget-Larsen, J. (2017).
Intramolecular hydrogen bonding in myricetin and myricitrin. Quantum
chemical calculations and vibrational spectroscopy. J. Mol. Struct.
1131: 242–249.
179 Bertolasi, V., Gilli, P., Ferretti, V., and Gilli, G. (1991). Evidence for resonance-
assisted hydrogen bonding. 2. Intercorrelation between crystal structure
and spectroscopic parameters in eight intramolecularly hydrogen bonded
1,3-diaryl-1,3-propanedione enols. J. Am. Chem. Soc. 113: 4917–4925.
180 Boyer, M.A., Marsalek, O., Heindel, J.P. et al. (2019). Beyond Badger’s rule: the
origins and generality of the structure–spectra relationship of aqueous hydrogen
bonds. J. Phys. Chem. Lett. 10: 918–924.
181 Sen, S. and Patwari, G.N. (2018). Electrostatics and dispersion in X–H· · ·Y
(X = C, N, O; Y = N, O) hydrogen bonds and their role in X–H vibrational
frequency shifts. ACS Omega 3: 18518–18527.
182 Howard, J., Tomkinson, J., Eckert, J. et al. (1983). Inelastic neutron scattering
studies of some intramolecular hydrogen bonded complexes: a new correlation
of 𝜈(OHO) vs. R(OO). J. Chem. Phys. 78: 3150–3155.
183 Tomkinson, J. (1992). The vibrations of hydrogen bonds. Spectrochim. Acta,
Part A 48: 329–348.
184 Rekik, N. and Wójcik, M.J. (2010). On the influence of electrical anharmonicity
on infrared bandshape of hydrogen bond. Chem. Phys. 369: 71–81.
185 Hartl, M., Daemen, L., Hartl, H. et al. (2013). Short hydrogen bonds in
2,4-dinitrobenzoic acid complexed with pyridine. Chem. Phys. 427: 87–94.
186 Wójcik, M.J., Kwiendacz, J., Boczar, M. et al. (2010). Theoretical and
spectroscopic study of hydrogen bond vibrations in imidazole and its deuterated
derivative. Chem. Phys. 372: 72–81.
187 Drichko, N.V., Kerenskaia, G.Y., and Schreiber, V.M. (1999). Medium and
temperature effects on the infrared spectra and structure of carboxylic acid–
pyridine complexes: acetic acid. J. Mol. Struct. 477: 127–141.
188 Takei, K., Takahashi, R., and Noguchi, T. (2008). Correlation between the
hydrogen-bond structures and the CQO stretching frequencies of carboxylic
acids as studied by density functional theory calculations: theoretical basis for
interpretation of infrared bands of carboxylic groups in proteins. J. Phys. Chem.
B 112: 6725–6731.
189 Gusakova, G.V., Denisov, G.S., and Smolyansky, A.L. (1972). On 𝜈CO
frequencies in IR spectrum of non-symmetrically perturbed carboxylate ion.
Opt. Spectrosc. 32: 922–925.
190 Lau, Y.K., Ikuta, S., and Kebarle, P. (1982). Thermodynamics and kinetics of
the gas-phase reactions H3 O+ (H2 O)n−1 + water = H3 O+ (H2 O)n . J. Am. Chem. Soc.
104: 1462–1469.
191 Larson, J.W. and McMahon, T.B. (1982). Gas-phase bifluoride ion. An ion
cyclotron resonance determination of the hydrogen bond energy in FHF−
from gas-phase fluoride transfer equilibrium measurements. J. Am. Chem. Soc.
104: 5848–5849.
192 Asfin, R.E., Denisov, G.S., and Tokhadze, K.G. (2002). The infrared spectra
and enthalpies of strongly bound dimers of phosphinic acids in the gas phase.
(CH2 Cl)2 POOH and (C6 H5 )2 POOH. J. Mol. Struct. 608: 161–168.
193 Libri, S., Jasim, N.A., Perutz, R.N., and Brammer, L. (2008). Metal fluorides
form strong hydrogen bonds and halogen bonds: measuring interaction
enthalpies and entropies in solution. J. Am. Chem. Soc. 130: 7842–7844.
194 Tupikina, E. Yu., Denisov, G.S., and Tolstoy, P.M. (2015). NMR study of
CHN hydrogen bond and proton transfer in 1,1-dinitroethane complex with
2,4,6-trimethylpyridine. J. Phys. Chem. A 119: 659–668.
195 Pollice, R., Bot, M., Kobylianskii, I.J. et al. (2017). Attenuation of London
dispersion in dichloromethane solutions. J. Am. Chem. Soc. 139: 13126–13140.
196 Aakeröy, C.B. and Seddon, K.R. (1993). The hydrogen bond and crystal engi-
neering. Chem. Soc. Rev. 22: 397–407.
197 Jabłoński, M., Kaczmarek, A., and Sadlej, A.J. (2006). Estimates of the energy
of intramolecular hydrogen bonds. J. Phys. Chem. A 110: 10890–10898.
198 Deshmukh, M.M., Suresh, C.H., and Gadre, S.R. (2007). Intramolecular
hydrogen bond energy in polyhydroxy systems: a critical comparison of
molecular tailoring and isodesmic approaches. J. Phys. Chem. A 111: 6472–6480.
199 Rusinska-Roszak, D. (2015). Intramolecular O—H· · ·O=C hydrogen bond
energy via the molecular tailoring approach to RAHB structures. J. Phys. Chem.
A 119: 3674–3687.
References 401
200 Afonin, A.V. and Vashchenko, A.V. (2019). Benchmark calculations of

intramolecular hydrogen bond energy based on molecular tailoring and
function-based approaches: developing hybrid approach. Int. J. Quantum Chem.
119: e26001/1–e26001/13.
201 Liu, C., Zhao, D.-X., and Yang, Z.-Z. (2012). Direct evaluation of individual
hydrogen bond energy in situ in intra- and intermolecular multiple hydrogen
bonds system. J. Comput. Chem. 33: 379–390.
202 Dong, H., Hua, W., and Li, S. (2007). Estimation on the individual hydrogen
bond strength in molecules with multiple hydrogen bonds. J. Phys. Chem. A
111: 2941–2945.
203 Gutowski, M. and Chalasiński, G. (1993). Critical evaluation of some computa-
tional approaches to the problem of basis set superposition error. J. Chem. Phys.
98: 5540–5554.
204 Jensen, F. (1996). The magnitude of intramolecular basis set superposition
error. Chem. Phys. Lett. 261: 633–636.
205 Estarellas, C., Lucas, X., Frontera, A. et al. (2010). Erroneous behaviour of the
widely used MP2(full)/aug-cc-pVXZ (X = D, T) level of theory for evaluating the
BSSE in ion–π complexes. Chem. Phys. Lett. 489: 254–258.
206 Frey, J.A. and Leutwyler, S. (2005). Comment on “Strength of the N—H· · ·OC
bonds in formamide and N-methylacetamide dimers”. J. Phys. Chem. A
109: 6990–6990.
207 Alagona, G. and Ghio, C. (1995). Basis set superposition errors for Slater vs.
Gaussian basis functions in H-bond interactions. J. Mol. Struct. THEOCHEM
330: 77–83.
208 Plumley, J.A. and Dannenberg, J.J. (2011). A comparison of the behavior
of functional/basis set combinations for hydrogen-bonding in the water
dimer with emphasis on basis set superposition error. J. Comput. Chem.
32: 1519–1527.
Acc. Chem. Res. 43: 1550–1557.
210 Eyman, D.P. and Drago, R.S. (1966). Nuclear magnetic resonance studies of
hydrogen bonding. J. Am. Chem. Soc. 88: 1617–1620.
211 Purcell, K.F., Stikeleather, J.A., and Brunk, S.D. (1969). Linear enthalpy-spectral
shift correlations for 1,1,1,3,3,3-hexafluoro-2-propanol. J. Am. Chem. Soc.
91: 4019–4027.
212 Odinokov, S.E., Mashkovsky, A.A., Dzizenko, A.K., and Glazunov, V.P. (1975).
Complexes of AH acids with triethylamine: specificities of spectroscopic
characteristics. Spectrosc. Lett. 8: 157–164.
213 Dohnal, V. and Tkadlecová, M. (2002). A simple relation between 1 H NMR data
and mixing enthalpy for systems with complex formation by hydrogen bonding.
J. Phys. Chem. B 106: 12307–12310.
214 Ishikawa, R., Kojima, C., Ono, A., and Kainosho, M. (2001). Developing model
systems for the NMR study of substituent effects on the N—H· · ·N hydrogen
bond in duplex DNA. Magn. Reson. Chem. 39: S159–S165.
215 Kumar, G.A. and McAllister, M.A. (1998). Theoretical investigation of the rela-
tionship between proton NMR chemical shift and hydrogen bond strength.
J. Org. Chem. 63: 6968–6972.
216 Weinhold, F. and Klein, R.A. (2012). What is a hydrogen bond? Mutually
consistent theoretical and experimental criteria for characterizing H-bonding
interactions. Mol. Phys. 110: 565–579.
217 Del Bene, J.E., Perera, S.A., and Bartlett, R.J. (1999). Hydrogen bond
types, binding energies, and 1 H NMR chemical shifts. J. Phys. Chem. A
103: 8121–8124.
218 Denisov, G.S., Bureiko, S.F., Kucherov, S.Y., and Tolstoy, P.M. (2017).
Correlation relationships between the energy and spectroscopic parameters
of complexes with F· · ·HF hydrogen bond. Dokl. Phys. Chem. 475: 115–118.
219 Gurinov, A.A., Lesnichin, S.B., Limbach, H.-H., and Shenderovich, I.G. (2014).
How short is the strongest hydrogen bond in the proton-bound homodimers of
pyridine derivatives? J. Phys. Chem. A 118: 10804–10812.
220 Shenderovich, I.G. (2006). Maximum value of the chemical shift in the
1
H NMR spectrum of a hydrogen-bonded complex. Russ. J. Gen. Chem.
76: 501–506.
221 Badger, R.M. and Bauer, S.H. (1937). Spectroscopic studies of the hydrogen
bond. II. The shift of the O–H vibrational frequency in the formation of the
hydrogen bond. J. Chem. Phys. 5: 839–851.
222 Badger, R.M. (1940). The relation between the energy of a hydrogen bond and
the frequencies of the O–H bands. J. Chem. Phys. 8: 288–289.
223 Gramstad, T. and Fuglevik, W.J. (1965). Studies of hydrogen bonding – XIV
Hydrogen bonding of indole to phosphoryl compounds. Spectrochim. Acta
21: 503–509.
224 Zeegers-Huyskens, T. (1977). Correlation between the vibrational properties and
the energy of the hydrogen bonds OH· · ·N and OH· · ·O=C systems. Bull. Soc.
Chim. Belg. 86: 823–832.
225 Van Ness, H.C. (1967). Infrared spectra and the thermodynamics of alcohol-
hydrocarbon systems. J. Phys. Chem. 71: 1483–1494.
226 Purcell, K.F. and Drago, R.S. (1967). Theoretical aspects of the linear enthalpy
wavenumber shift relation for hydrogen-bonded phenols. J. Am. Chem. Soc.
89: 2874–2879.
227 Kleeberg, H. and Luck, W.A.P. (1983). Infrared study of anion–water interac-
tions in dichloromethane. J. Solution Chem. 12: 369–381.
228 Ouvrard, C., Berthelot, M., and Laurence, C. (2001). An enthalpic scale
of hydrogen-bond basicity, part 1: halogenoalkanes. J. Phys. Org. Chem.
14: 804–810.
229 Becker, E.D. (1961). Infrared studies of hydrogen bonding in alcohol–base
systems. Spectrochim. Acta 17: 436–447.
230 Liddel, U. and Becker, E.D. (1957). Infra-red spectroscopic studies of hydrogen
bonding in methanol, ethanol, and t-butanol. Spectrochim. Acta 10: 70–84.
References 403
231 Arnett, E.M., Joris, L., Mitchell, E. et al. (1970). Organic and biological
chemistry studies of hydrogen-bonded complex formation. III. Thermody-
namics of complexing by infrared spectroscopy and calorimetry. J. Am. Chem.
Soc. 92: 2365–2377.
232 Cohen, D.T., Zhang, C., Pentelute, B.L., and Buchwald, S.L. (2015). An umpol-
ung approach for the chemoselective arylation of selenocysteine in unprotected
peptides. J. Am. Chem. Soc. 137: 9784–9787.
233 Allerhand, A. and Schleyer, P.v.R. (1963). A survey of C–H groups as proton
donors in hydrogen bonding. J. Am. Chem. Soc. 85: 1715–1723.
234 Gonjo, T., Futami, Y., Morisawa, Y. et al. (2011). Hydrogen bonding effects
on the wavenumbers and absorption intensities of the OH fundamental and
the first, second, and third overtones of phenol and 2, 6-dihalogenated phe-
nols studied by visible/near-infrared/infrared spectroscopy. J. Phys. Chem. A
115: 9845–9853.
235 Shubina, E.S., Belkova, N.V., Krylov, A.N. et al. (1996). Spectroscopic evidence
for intermolecular M—H· · ·H—OR hydrogen bonding: interaction
of WH(CO)2 (NO)L2 hydrides with acidic alcohols. J. Am. Chem. Soc.
118: 1105–1112.
236 Iogansen, A.V. and Rassadin, B.V. (1969). Dependence of intensity and shift in
the infrared bands of the hydroxyl group on hydrogen bonding energy. J. Appl.
Spectrosc. 11: 1318–1325.
237 Ratajczak, H., Orville-Thomas, W.J., and Rao, C.N.R. (1976). Charge transfer
theory of hydrogen bonds: relations between vibrational spectra and energy of
hydrogen bonds. Chem. Phys. 17: 197–216.
238 West, R., Powell, D.L., Whatley, L.S. et al. (1962). The relative strengths of
alkyl halides as proton acceptor groups in hydrogen bonding. J. Am. Chem. Soc.
84: 3221–3222.
239 Chitale, S.M. and Jose, C.I. (1986). Infrared studies and thermodynamics of
hydrogen bonding in 2-halogenoethanols and 3-halogenopropanols. J. Chem.
Soc., Faraday Trans. 82: 663–679.
240 Abraham, M.H., Prior, D.V., Schulz, R.A. et al. (1998). Hydrogen bonding Part
44. Thermodynamics of complexation of 3,5-dichlorophenol with ketones and
ethers in cyclohexane: the Badger–Bauer relationship. J. Chem. Soc., Faraday
Trans. 94: 879–885.
241 Borisenko, V.E., Blinkova, G.Y., Osipova, L.L., and Zavjalova, Y.A. (1996).
Temperature effects on absorption bands 𝜈(NH) and thermodynamics of
hydrogen-bonded complexes of NH-donors with proton acceptors. J. Mol. Liq.
70: 31–54.
242 Li, X., Liu, L., and Schlegel, H.B. (2002). On the physical origin of blue-shifted
hydrogen bonds. J. Am. Chem. Soc. 124: 9639–9647.
243 Koné, M., Illien, B., Laurence, C., and Graton, J. (2011). Can quantum-
mechanical calculations yield reasonable estimates of hydrogen-bonding
acceptor strength? The case of hydrogen-bonded complexes of methanol.
J. Phys. Chem. A 115: 13975–13985.
244 Laurence, C. and Gal, J.-F. (2009). Lewis Basicity and Affinity Scales: Data and
Measurement. Wiley.
245 Drago, R.S., O’Bryan, N., and Vogel, G.C. (1970). A frequency shift-enthalpy
correlation for a given donor with various hydrogen-bonding acids. J. Am.
Chem. Soc. 92: 3924–3929.
246 Sherry, A.D. and Purcell, K.F. (1972). Linear enthalpy-spectral shift correlation
for perfluoro-tert-butyl alcohol. J. Am. Chem. Soc. 94: 1853–1857.
247 Góralski, P. (1993). Hydrogen bonds between cholesterol and oxygen bases: a
thermodynamic study. J. Chem. Soc., Faraday Trans. 89: 2433–2435.
248 Joesten, M.D. and Drago, R.S. (1962). The validity of frequency shift–enthalpy
correlations. I. Adducts of phenol with nitrogen and oxygen donors. J. Am.
Chem. Soc. 84: 3817–3821.
249 Epley, T.D. and Drago, R.S. (1967). Calorimetric studies on some
hydrogen-bonded adducts. J. Am. Chem. Soc. 89: 5770–5773.
250 Drago, R.S. and Epley, T.D. (1969). Enthalpies of hydrogen bonding and
changes in hydroxy frequency shifts for a series of adducts with substituted
phenols. J. Am. Chem. Soc. 91: 2883–2890.
251 Vogel, G.C. and Drago, R.S. (1970). Hydrogen bonding of sulfur donors with
various phenols. J. Am. Chem. Soc. 92: 5347–5351.
252 Bhatta, R.S., Iyer, P.P., Dhinojwala, A., and Tsige, M. (2014). A brief review
of Badger–Bauer rule and its validation from a first-principles approach.
Mod. Phys. Lett. B 28: 1430014/1–1430014/16.
253 Wendler, K., Thar, J., Zahn, S., and Kirchner, B. (2010). Estimating the hydro-
gen bond energy. J. Phys. Chem. A 114: 9529–9536.
254 Solomonov, B.N., Novikov, V.B., Varfolomeev, M.A., and Klimovitskii, A.E.
(2005). Calorimetric determination of hydrogen-bonding enthalpy for neat
aliphatic alcohols. J. Phys. Org. Chem. 18: 1132–1137.
255 Iogansen, A.V. (1999). Direct proportionality of the hydrogen bonding energy
and the intensification of the stretching 𝜈(XH) vibration in infrared spectra.
256 Brovarets’, O.O. and Hovorun, D.M. (2011). IR vibrational spectra of H-bonded
complexes of adenine, 2-aminopurine and 2-aminopurine+ with cytosine and
thymine: quantum-chemical study. Opt. Spectrosc. 111: 750–757.
257 Ponomareva, A.G., Yurenko, Y.P., Zhurakivsky, R.O. et al. (2012). Complete
conformational space of the potential HIV-1 reverse transcriptase inhibitors
d4U and d4C. A quantum chemical study. Phys. Chem. Chem. Phys.
14: 6787–6795.
258 Yurenko, Y.P., Zhurakivsky, R.O., Ghomi, M. et al. (2008). Ab initio
comprehensive conformational analysis of 2′ -deoxyuridine, the biologically
significant DNA minor nucleoside, and reconstruction of its low-temperature
matrix infrared spectrum. J. Phys. Chem. B 112: 1240–1250.
259 Yurenko, Y.P., Zhurakivsky, R.O., Samijlenko, S.P. et al. (2007). The whole of
intramolecular H-bonding in the isolated DNA nucleoside thymidine. AIM
electron density topological study. Chem. Phys. Lett. 447: 140–146.
References 405
260 Yurenko, Y.P., Zhurakivsky, R.O., Ghomi, M. et al. (2007). How many conform-
ers determine the thymidine low-temperature matrix infrared spectrum? DFT
and MP2 quantum chemical study. J. Phys. Chem. B 111: 9655–9663.
261 Afonin, A.V., Sterkhova, I.V., Vashchenko, A.V., and Sigalov, M.V. (2018).
Estimating the energy of intramolecular bifurcated (three-centered) hydrogen
bond by X-ray, IR and 1 H NMR spectroscopy, and QTAIM calculations. J. Mol.
Struct. 1163: 185–196.
262 Katsyuba, S.A., Vener, M.V., Zvereva, E.E. et al. (2013). How strong is
hydrogen bonding in ionic liquids? Combined X-ray crystallographic, infrared/
Raman spectroscopic, and density functional theory study. J. Phys. Chem. B
117: 9094–9105.
263 Halonen, L. and Child, M.S. (1982). A local mode model for tetrahedral
molecules. Mol. Phys. 46: 239–255.
264 Halonen, L. (2007). Local mode vibrations in polyatomic molecules. Adv. Chem.
Phys. 104: 41–179.
265 Shchepkin, D.N. and Melikova, S.M. (1992). Theoretical vibrational study of
hydrogen-bonded complexes: a simple anharmonical model. J. Chim. Phys.
Phys.- Chim. Biol. 89: 607–613.
266 Rusinska-Roszak, D. and Sowinski, G. (2014). Estimation of the intramolecu-
lar O—H· · ·O=C hydrogen bond energy via the molecular tailoring approach,
Part I: aliphatic structures. J. Chem. Inf. Model. 54: 1963–1977.
267 Rivera-Rivera, L.A., McElmurry, B.A., Scott, K.W. et al. (2013). The Badger–
Bauer rule revisited: correlation of proper blue frequency shifts in the OC
hydrogen acceptor with morphed hydrogen bond dissociation energies in
OC–HX (X = F, Cl, Br, I, CN, CCH). J. Phys. Chem. A 117: 8477–8483.
268 Dahms, F., Costard, R., Pines, E. et al. (2016). The hydrated excess proton in
the Zundel cation H5 O2 + : the role of ultrafast solvent fluctuations. Angew.
Chem. Int. Ed. 128: 10758–10763.
269 Kundu, A., Dahms, F., Fingerhut, B.P. et al. (2019). Hydrated excess protons
in acetonitrile/water mixtures: solvation species and ultrafast proton motions.
J. Phys. Chem. Lett. 10: 2287–2294.
270 Iogansen, A.V. and Litovchenko, G.D. (1965). The characteristic bands of the
stretching vibrations of the nitro group in infrared absorption. II. Correlation
of the frequencies and intensities with the structure of the molecules. J. Appl.
Spectrosc. 3: 404–411.
271 Iogansen, A.V., Kurkchi, G.A., and Rassadin, B.V. (1969). Intensity of
𝜈(A–H) infrared bands and N—H· · ·B hydrogen bonds. J. Appl. Spectrosc.
11: 1487–1492.
272 Iogansen, A.V., Kurkchi, G.A., and Levina, O.V. (1973). The enhancement and
displacement of the IR band of hydrogen chloride as a function of the energy of
the hydrogen bond. J. Appl. Spectrosc. 18: 499–503.
273 Vener, M.V., Egorova, A.N., Churakov, A.V., and Tsirelson, V.G. (2012).
Intermolecular hydrogen bond energies in crystals evaluated using electron
density properties: DFT computations with periodic boundary conditions.
J. Comput. Chem. 33: 2303–2309.
274 Lu, J. and Scheiner, S. (2020). Relationships between bond strength and
spectroscopic quantities in H-bonds and related halogen, chalcogen, and
pnicogen bonds. J. Phys. Chem. A 124: 7716–7725.
275 Katsyuba, S.A., Zvereva, E.E., Vidiš, A., and Dyson, P.J. (2007). Application
of density functional theory and vibrational spectroscopy toward the rational
design of ionic liquids. J. Phys. Chem. A 111: 352–370.
276 Howard, D.L., Jørgensen, P., and Kjaergaard, H.G. (2005). Weak intramolecular
interactions in ethylene glycol identified by vapor phase OH-stretching overtone
spectroscopy. J. Am. Chem. Soc. 127: 17096–17103.
277 Jarmelo, S., Reva, I., Carey, P.R., and Fausto, R. (2007). Infrared and Raman
spectroscopic characterization of the hydrogen-bonding network in L-serine
crystal. Vib. Spectrosc. 43: 395–404.
278 Tupikina, E.Yu., Tolstoy, P.M., Titova, A.A. et al. (2021). Estimations of
FH· · ·X hydrogen bond energies from IR intensities: Iogansen’s rule revisited.
J. Comput. Chem. 42: 564–571.
279 Nibbering, E.T., Dreyer, J., Kuhn, O. et al. (2007). Vibrational dynamics of
hydrogen bonds. In: Analysis and Control of Ultrafast Photoinduced Reactions
(ed. O. Kühn and L. Wöste), 619–687. Springer.
280 Falk, M. (1984). The frequency of the H–O–H bending fundamental in solids
and liquids. Spectrochim. Acta, Part A 40: 43–48.
281 Seki, T., Chiang, K.-Y., Yu, C.-C. et al. (2020). The bending mode of water: a
powerful probe for hydrogen bond structure of aqueous systems. J. Phys. Chem.
Lett. 11: 8459–8469.
282 Pimentel, G.C. and McClellan, A.L. (1960). The Hydrogen Bond. Freeman.
283 Zundel, G. (1998). The far infrared vibration of hydrogen bonds with large
proton polarizability. J. Mol. Struct. 381: 23–37.
284 Brzezinski, B., Rabold, A., and Zundel, G. (1994). Far-IR study of
the hydrogen-bond vibration of intramolecular bonds in substituted
2-diethylaminomethylphenol N-oxides, as a function of the pK a of the phenolic
group. J. Chem. Soc., Faraday Trans. 90: 843–844.
285 Langner, R. and Zundel, G. (1995). FT-IR investigation of polarizable, strong
hydrogen bonds in sulfonic acid-sulfoxide, -phosphine oxide, and -arsine
oxide complexes in the middle- and far-infrared region. J. Phys. Chem.
99: 12214–12219.
286 Keil, T., Brzezinski, B., and Zundel, G. (1992). Far-infrared investigation of
the proton transfer with substituted phenol + N-mono- and N,N ′ -dioxides as a
function of the pK a of the phenols. J. Phys. Chem. 96: 4421–4426.
287 Langner, R. and Zundel, G. (1998). FT-IR investigation of OH· · ·N ⇌ O− · · ·HN+
hydrogen bonds with large proton polarizability in phosphinic acid + N-base
systems in the middle and far infrared region. J. Phys. Chem. A 102: 6635–6642.
288 Snels, M., Horká-Zelenková, V., Hollenstein, H., and Quack, M. (2011). High
resolution FTIR and diode laser spectroscopy of supersonic jets. In: Handbook
of High-Resolution Spectroscopy (ed. M. Quack and F. Merkt), 1021–1067. Wiley.
References 407
289 Oswald, S. and Suhm, M.A. (2019). Soft experimental constraints for soft
interactions: a spectroscopic benchmark data set for weak and strong hydrogen
bonds. Phys. Chem. Chem. Phys. 21: 18799–18810.
290 Kollipost, F., Wugt Larsen, R., Domanskaya, A.V. et al. (2012). The highest
frequency hydrogen bond vibration and an experimental value for the
dissociation energy of formic acid dimer. J. Chem. Phys. 136: 151101/1–
151101/4.
291 Asfin, R.E., Kolomiitsova, T.D., Shchepkin, D.N., and Tokhadze, K.G.
(2017). Infrared studies of the symmetry changes of the 28 SiH4 molecule in
low-temperature matrixes. Fundamental, combination, and overtone transitions.
J. Phys. Chem. A 121: 5116–5126.
292 Ignatov, S.K., Kolomiitsova, T.D., Mielke, Z. et al. (2006). A matrix isolation and
theoretical study of SiF4 dimers spectra. Chem. Phys. 324: 753–766.
293 Tokhadze, I.K., Kolomiitsova, T.D., Shchepkin, D.N. et al. (2009). Influence of
resonance interactions and matrix environment on the spectra of SF6 dimers in
low-temperature nitrogen matrixes. Theory and experiment. J. Phys. Chem. A
113: 6334–6341.
294 Rachwalska, M., Natkaniec, I., Zborowski, K. et al. (2014). Inelastic neutron
scattering (INS) study of low frequency vibrations and hydrogen bonding of
(E)-benzil monoxime. Z. Phys. Chem. 228: 63–97.
295 Albers, P.W., Glenneberg, J., Refson, K., and Parker, S.F. (2014). Inelastic inco-
herent neutron scattering study of the molecular properties of pure hydrogen
peroxide and its water mixtures of different concentration. J. Chem. Phys.
140: 164504/1–164504/10.
296 Kwocz, A., Panek, J.J., Jezierska, A. et al. (2018). A molecular roundabout:
triple cyclically arranged hydrogen bonds in light of experiment and theory.
New J. Chem. 42: 19467–19477.
297 Tupikina, E.Yu., Tokhadze, K.G., Karpov, V.V. et al. (2020). Stretching force
constants as descriptors of energy and geometry of F· · ·HF hydrogen bonds.
Spectrochim. Acta, Part A 241: 118677/1–118677/6.
298 Legon, A.C. (2014). A reduced radial potential energy function for the halogen
bond and the hydrogen bond in complexes B· · ·XY and B· · ·HX, where X and Y
are halogen atoms. Phys. Chem. Chem. Phys. 16: 12415–12421.
299 Ostras’, A.S., Ivanov, D.M., Novikov, A.S., and Tolstoy, P.M. (2020). Phosphine
oxides as spectroscopic halogen bond descriptors: IR and NMR correlations
with interatomic distances and complexation energy. Molecules 25: 1406.
300 Alkorta, I. and Legon, A. (2018). Strengths of non-covalent interactions in
hydrogen-bonded complexes B· · ·HX and halogen-bonded complexes B· · ·XY
(X, Y = F, Cl): an ab initio investigation. New J. Chem. 42: 10548–10554.
409
13
ATR-Far-Ultraviolet Spectroscopy Holds Unique Advantages

for Investigating Hydrogen Bondings and Intermolecular
Interactions of Molecules in Condensed Phase
Yusuke Morisawa 1 , Takeyoshi Goto 2 , Nami Ueno 1 , and Yukihiro Ozaki 3,4
1
Kindai University, School of Science and Engineering, Department of Chemistry, Kowakae, Higashi-Osaka,
Osaka 577-8502, Japan
2
High Energy Accelerator Research Organization, Center for Applied Superconductive Accelerator,
Accelerator Laboratory, Ooho, Tsukuba, Ibaragi 305-0801, Japan
3
Kwansei Gakuin University, School of Biological and Environmental Sciences, Department of Medical
Sciences, Gakuen-Uegaharu, Sanda, Hyogo 669-1337, Japan
4
Toyota Physical and Chemical Research Institute, Yokomichi, Nagakute, Aichi 480-1192, Japan
13.1 Introduction
This chapter describes far-ultraviolet (FUV) spectroscopy studies of hydrogen

bondings and intermolecular interactions of molecules in condensed phase. In this
chapter, the FUV region is defined as the region of 120–200 nm (10 to 6 eV)
[1–7]. Vacuum ultraviolet (VUV) region (10–120 nm; 124 to 10 eV) requests a
vacuum-evaporated spectrometer while a spectrometer used in the FUV region is
evacuated or purged with nitrogen gas. In the deep-ultraviolet (DUV; 200–300 nm)
and ultraviolet (UV; 300–400 nm) regions spectrometers vacuum-evaporated or
purged with nitrogen gas are not needed. FUV spectroscopy is concerned with
electronic spectroscopy, and in the FUV region bands due to the transitions to
valence orbitals, such as π–π* and n–σ* transitions and those to Rydberg orbitals are
observed [1–10].
FUV spectroscopy has a long history as spectroscopy for gas molecules [1–3, 8–10].
It was found from the studies on gas molecules that various kinds of molecules have
strong absorptions due to electronic transitions to low-lying Rydberg states in the
FUV region until their vertical ionized energy. FUV spectroscopy studies of liquid
and solid samples were much delayed in comparison with those for gas molecules
because absorptivity of condensed matters is very high in the FUV region. To over-
come this problem, we have introduced ATR (attenuated total reflection) technique
into the FUV region [1–3, 11]. We succeeded in measuring the FUV spectra in the
145–250 nm region of water without band saturation [1, 2, 11, 12]. One of the key
points in the development of the ATR-FUV spectrometer was the design of a very
small internal reflection element (IRE) probe [1, 2, 11]. We employed a vacuum UV
grade sapphire. The penetration depth of ATR spectroscopy is around 100 nm in the

410 13 ATR-Far-Ultraviolet Spectroscopy Study of Hydrogen Bondings
FUV region so that one can investigate the surfaces of materials, their thin films,
and adsorption of liquids and solids on an IRE using the ATR-FUV spectroscopy.
We reported the details of the ATR spectrometers which we developed in the
literatures [1, 2, 11].
13.2 Characteristics and Advantages of FUV

Spectroscopy for the Studies of Liquids and Solids
Spectroscopy in the FUV region is concerned with the transitions between electronic
energy levels including Rydberg states [1, 3, 7–10]. Samples in the condensed phase
show very intense absorption in the FUV region. Various kinds of molecules that
do not have a clear absorption in the region longer than about 200 nm show intense
absorptions in this region. Water and alkanes are good examples. Band assignments
in the FUV region are not straightforward and were not well established until we
have introduced quantum chemical calculations [3, 7, 13–18].
The advantages of FUV spectroscopy for the studies of electronic transitions and
structure of liquid and solid phases are summarized as follows [1, 3, 7]:
1. The FUV region contains unique information about the electronic transitions and
structure of a molecule, which one cannot get from ordinary UV spectroscopy.
For example, it is possible to investigate Rydberg transitions of liquids and solids
[3, 15, 19].
2. FUV spectroscopy is suitable to investigate hydrogen bondings and inter-
molecular interactions of various molecules and materials in condensed phase
[1, 3, 7, 13, 17, 18, 20–27].
3. It is possible to apply FUV spectroscopy to qualitative analysis and discrimination
analysis of various samples because each molecule shows a characteristic FUV
spectrum, and FUV spectra are very sensitive to a variety of variations such as
those in chemical bondings, electronic structures, molecular conformations, and
molecular environments [1, 4, 5a, 7, 9, 26, 27].
4. FUV spectroscopy is useful also for highly sensitive quantitative analysis because
almost all molecules give rise to strong absorption in the FUV region, and the
intensity and wavelength of an FUV band are very sensitive to variations in con-
centration, temperature, pH, and so on [24–27].
5. One can investigate not only structure but also functions of various kinds of mate-
rials such as polymers, inorganic semiconductors, graphenes, and ionic liquids
using ATR-FUV spectroscopy [28–34].
6. ATR-FUV spectroscopy enables to explore thin films and adsorbed molecules on
a surface. One can apply variable angle-ATR-FUV technique for these kinds of
applications [20].
7. Time-resolved FUV spectroscopy has considerable potential for studying kinetics
of photochemical reactions [34, 35].
In the FUV region, one can observe bands due to transitions of π electrons,
lone pair electrons, and σ electrons, and thus, ATR-FUV spectroscopy allows to
13.3 FUV Spectroscopic Studies of Hydrogen Bonds 411
investigate electronic states of almost all kinds of molecules in condensed phases.

We have established a band assignment method of FUV spectra based on quantum
chemical calculations. For example, a band near 153 nm of alkanes is assigned to
a transition from HOMO-2 or HOMO-1 to 3p Rydberg [3, 16]. Since alkanes are
formed only by single bonds and do not have any lone pair electrons, this transition
reflects the state of σ electrons of alkanes. Moreover, it was found that the 153 nm
band of tetradecane shifts to 200 and 230 nm at the phase transition temperature,
showing that σ electrons are affected by an intermolecular interaction in condensed
phase [3, 19].
ATR-FUV spectroscopy has been used for a variety of applications such as those
to materials research like polymers [18, 22, 32], graphene [32, 33], inorganic semi-
conductors [28–30], and ionic liquids [31], electroanalytical chemistry [34], surface
adsorption studies [31], water research [26], laser photolysis [35], process analysis
[36], radical cleaning [37], and FUV surface plasmon resonance (SPR) [38]. Recently,
it has been applied to biological samples [39].
13.3 FUV Spectroscopic Studies of Hydrogen Bonds and

Hydration Structures of Electrolyte Aqueous Solutions
The development of ATR-FUV spectrometers enabled us to observe the entire first
̃ ← X)
electronic transition (A ̃ absorption band of liquid water, including the band
̃←X
maximum directly [2, 11, 12]. The A ̃ transition of water is strongly affected by
the hydrogen bonding and hydration states of water molecules because the A ̃←X ̃
transition of water is mostly the excitation of the nonbonding electrons of the oxygen
atom (1b1 ) to the 4a1 orbital that consists of σ* and the 3s Rydberg orbitals [2, 9, 12].
The Ã←X ̃ energy of water increases with increasing hydrogen bond interactions:
7.4 (vapor) [40], 8.3 (liquid) [2, 11, 12, 20, 41], and 8.7 eV (solid) [42].
With the addition of salts to liquid water, the electric field of an ion induces the
rearrangement of the hydrogen bonding state of water molecules from bulk to liquid
state. The high charge density ions electrostatically bind to the surrounding water
molecules and form the first and even second coordination sphere [43], which con-
tributes to the hydration energy of the ion [44, 45]. ATR-FUV spectroscopy provides
new insights into the hydration states of cations from the perspective of the elec-
tronic transition of water molecules [21, 23, 24].
Figure 13.1a shows the A ̃←X ̃ bands of Group I cation nitrate aqueous solutions
(1 M) [23, 24]. All the bands appear at higher energies relative to that of pure water.
Because the counter anion of all the solutions is the same (nitrate), the extent of
the band shift depends on the cation species. Figure 13.1b depicts the correlation
between the A ̃←X ̃ energies of Group I, II, and XIII nitrate electrolyte solutions and
the hydration energies of the cations (ΔGhyd ) [24], showing that the correlation is lin-
ear for each Group. These linear correlations indicate that the electrostatic interac-
tion between the nonbonding electrons of the oxygen atom of water and the cations
determines the A ̃←X ̃ energies of the electrolyte solutions. However, there are some
apparent deviations from the linear correlations for the small cations; specifically,
Wavelength (nm) Figure 13.1 (a) ATR–FUV spectra

160 180 200 220 240 of pure water and 1 M alkali metal
cation nitrate aqueous solutions.
0.4 A–X (H2O) ̃←X ̃ energies of
Water (b) A plot of the A
H+ Group I, II, and XIII nitrate aqueous
ATR Abs. (logl/l0)
0.3 Li+ solutions versus the hydration

Na+ energies of the cations (DGhyd).
K+ Source: Goto et al. [21]/Royal
0.2 Rb+
Society of Chemistry.
Cs+
0.1 π–π* (NO3–)
0.0
8 7 6 5
(a) Photon energy (eV)
8.05
Al3+ I
Ga3+ II
Transition energy (eV)
XIII
8.04
Be2+
Mg2+
8.03
Ca2+
H+
In3+ Sr2+ Li+
8.02
Na+
K+
Rb+ +
Cs
8.01
–4000 –3000 –2000 –1000 0
(b) ΔGhyd (kJ mol–1)
H+ and Be2+ deviate to lower transition energies and Li+ to higher energy. Small
cations with high charge densities strongly disrupt the water–water hydrogen bond-
ing by forming condensed hydration structures [45–47]. Thus, these deviations indi-
cate that the hydration structures around the cations affect the electronic ground
and excited states of the hydrating water molecules in a complicated manner.
̃ ←X
13.4 Quantum Chemical Calculations of the A ̃
Transition of Hydrated Group I Cations
FUV spectra of Group I, II, and XIII cation nitrate electrolyte aqueous solutions show
̃←X
that the A ̃ energies of water hydrating the cations are linearly dependent on
the hydration energies of the cations but those of water hydrating the small cations
are not [24]. To understand the cation size effects on the electronic states of each
hydration shell of water molecule, the A ̃←X ̃ transitions of Group I cation–water
clusters holding the first and second shell water molecules around the cations
(M+ (H2 O)6 , M+ : H+ , Li+ , Na+ , and K+ ) were investigated using quantum chemical
calculations [21].
̃←X
13.4 Quantum Chemical Calculations of the A ̃ Transition of Hydrated Group I Cations 413
The optimized geometries and partial charges of the M+ (H2 O)6 clusters were cal-
culated at the DFT-M062X/6-311++G(d,p) level of theory, and those show that the
sizes and the charge densities of the cations strongly affect the condensed cluster
structures [21]. As the cation size decreases from K+ to H+ , the cluster size decreases.
The bond length between the cation and the oxygen atom of the first shell water (r M )
becomes shorter from K+ (H2 O)6 (2.54 Å) to H+ (H2 O)6 (1.19 Å). The proton induces
more distortion in the cluster structure than the other cations. The smallest OM+ O
angle (𝜃 M ) of H+ (H2 O)6 is 176.7∘ (180∘ for the other M+ clusters). The calculated
partial charge on each atom in the clusters shows that the charge distribution of
H+ (H2 O)6 is more delocalized than those of the other clusters.
The natural transition orbitals (NTOs) of the A ̃←X ̃ transition of each cluster
was calculated using the transition density matrices based on Time-dependent
density-function theory (TDDFT) calculations with the M06HF functional [48]
to show the electron density distributions for the ground and excited states [21].
Figure 13.2 Ground (bottom) and excited state (top) orbitals of the first shell water
molecules of H+ (H2 O)6 (s) (left), (l) (center), and Li+ (H2 O)6 (right) calculated with the NTO
method. Source: Goto et al. [21]/Royal Society of Chemistry.
Figure 13.3 Ground (bottom) and excited state (top) orbitals of the second shell of
H+ (H2 O)6 (s) (left), (l) (center), and Li+ (H2 O)6 (right) calculated with the NTO method. Source:
Goto et al. [21]/Royal Society of Chemistry.
0.08 Figure 13.4 The energy

First shell Li+ and oscillator strength per
Na+
Second shell water molecule of calculated
H+ (DFT opt) K+
0.06 with EOM-CCSD method
Oscillator strength
(6-311++G(d,p)). The
H+1st
Li+ oscillator strengths of the
H+2nd Na+
H+s K+ first shale water molecules
0.04 were divided by two and
H+/ those of the second shell by
H+s H+/ four. Source: Goto et al.
0.02 [21]/Royal Society of
Chemistry.
0.00
10.5 10.0 9.5 9.0 8.5 8.0
Electronic transition energy (eV)
Figures 13.2 and 13.3 show the calculated NTOs of H+ (H2 O)6 (left) and Li+ (H2 O)6
(right) for the first shell water molecules and the second shell, respectively. The
̃←X
A ̃ transitions of all the clusters correspond to the excitation of nonbonding
electrons on the oxygen atoms of water (1b1 ). The excited electrons of the first shell
water molecules occupy the σ* orbitals of the M+ O bonds that are strongly localized
around the central cations, indicating that the A ̃←X ̃ transitions of the first shell
water molecules are CT transitions. Those of the second shell water molecules
occupy the σ* orbitals of OH bonds of the second shells, corresponding to the 4a1
orbital of the water monomer, as shown in Figure 13.3.
Figure 13.4 shows the A ̃←X ̃ energies and oscillator strength per water molecule
+
of M (H2 O)6 calculated with the EOM-CCSD method [21]. There are analogous
trends in the transition energies and oscillator strengths among the cations for each
shell. As the cation size decreases from K+ to H+ , the A ̃←X ̃ energies of the clus-
ters increase from 8.16 to 9.95 eV for the first shell water molecules and from 8.09 to
8.41 eV for the second shell, and the oscillator strengths of both the shells become
larger except for that of H+ (H2 O)6 . These cation size dependences indicate that the
stronger orbital overlaps between the cations and the hydrating water molecules
cause higher transition energies and probabilities, and these are more significant
for the first shell water molecules than the second shell.
Figure 13.5a depicts the FUV spectra of M+ (H2 O)6 clusters calculated with
EOM-CCSD method (bandwidth: 0.333 eV), and Figure 13.5b plots the calculated
̃←X
A ̃ energies of M+ (H2 O)6 vs. the ΔGhyd values of the corresponding cations [21].
This correlation analysis provides qualitative information concerning the A ̃←X ̃
transitions of the electrolyte solutions. The correlation shows an analogous trend
to that determined experimentally (Figure 13.1b). The A ̃←X ̃ energy of H+ (H2 O)6
deviates from the linear relationship to a lower transition energy, while that of
Li+ (H2 O)6 deviates slightly to the higher energy side. These deviations arise from
the calculated results that the differences in the A ̃←X ̃ energies between the first
and second shell water molecules (ΔE1st−2nd ) are very large for the small cations.
̃←X
Specifically, the A ̃ band of the nitric acid solution is associated with the second
shell of water only, that of lithium nitrate solution is associated with the second
shell and partially from the first shell, and those of sodium nitrate and potassium
̃←X
13.4 Quantum Chemical Calculations of the A ̃ Transition of Hydrated Group I Cations 415
Wavelength (nm)
Molar absorption coefficient (mol–1 l cm–1)
130 140 150 160 170
H+ (H2O)6
Li+ (H2O)6
10 000
Na+ (H2O)6
K+ (H2O)6
5000
0
10.0 9.5 9.0 8.5 8.0 7.5 7.0
(a) Photon energy (eV)
Calculated transition energy (eV)
8.8
Li+
8.6
H+
8.4
Na+
8.2
K+
–100 –800 –600 –400 –200

(b) ∆Ghyd (kJ mol–1)
Figure 13.5 (a) FUV spectra of the M+(H2O)6 clusters calculated with the EOM-CCSD
method (6-311++G(d,p), bandwidth: 0.333 eV). (b) A plot of the calculated Ã ←X̃ energies
of M+ (H2 O)6 vs. DGhyd of the corresponding cations. The transition energy of Li+ (H2 O)6 was
taken at the midpoint between the two overlapping bands. The fitting line was drawn,
based on analogy with the experimental data presented in Figure 13.1b. Source: Goto et al.
[21]/Royal Society of Chemistry.
nitrate solutions are associated with both the shells. The FUV spectra of 1 M Group
I cation nitrate solutions (Figure 13.1a) show corresponding results.
Figure 13.6 shows the ground (bottom) and excited state (top) orbital energies
versus the inverse of r M for the first (black line) and second shell (red line) water
[21]. The calculated electronic energies show that the cation size effects on the
ground and excited states of M+ (H2 O)6 are different between the first and second
shell water molecules. As the r M lengths decrease from K+ (H2 O)6 to H+ (H2 O)6 ,
the ground state energies decrease. Those dependences of each shell on the cation
species arise mostly from the electrostatic interactions between the nonbonding
electrons of the water molecules and the cations. The correlation between the
ground state energy and the inverse of r M for both the shells is not completely
First shell
H+ s
–8 Second shell
K+ Na+ H+/
Li+
Orbital energy (eV)
K+ +
Na
–9 Li+ H+
/
H+s
(0.987)
–17
(0.971)
(0.987)
(0.929)
–18
(0.559) (0.559)
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

rM–1 (Å–1)
Figure 13.6 Ground and excited state energies of the first (right) and second shell (left)
water of M+ (H2 O)6 calculated with the EOM-CCSD method. The numbers in parentheses are
the calculated charges of the cations with the NPA method. The vertical lengths between
the ground and excited states correspond to the Ã←X ̃ energies. Source: Goto et al.
[21]/Royal Society of Chemistry.
linear, because the localized charges on the cations are very different (q(H+ ) = 0.530
and q(K+ ) = 0.983). The excited state orbital energies of each shell water molecule
show different dependences on the inverse of r M , which reflects that the excited
state orbitals of each shell are completely different (Figures 13.2, top and 13.3,
top). The energy dependence on the cation species is very small for the first shell
water molecules. The interactions of the excited electrons of the first shell water
molecules with the cations are comparable among the cation species because they
are strongly localized around the cations. For the second shell water molecules,
the excited state orbital energies become lower as the r M length becomes shorter
because the occupied σ* orbitals (4a1 ) are detached from the cations.
13.5 Hydrogen Bonding States of Interfacial Water

Adsorbed on an Alumina Surface Studied by Variable
Angle-ATR-FUV Spectroscopy
The hydrogen bonding states of interfacial water adsorbed on solid surfaces are
̃←X
very different from those of bulk liquid water [49, 50]. Then, the A ̃ transition
of interfacial water on a solid surface must be different from that of bulk liquid
water. The ATR-FUV spectroscopic technique is very suitable to study interfacial
water on a solid surface because of the very short penetration depth of the FUV
probe light. The Ã←X ̃ transition and the hydrogen bonding state of liquid water
on an α-alumina substrate were investigated with variable angle (VA)-ATR-FUV
spectroscopy [20]. It was possible to determine individual FUV spectra associated
with bulk water (distance from the alumina surface >2 nm) and interfacial water
(<2 nm) based on a variation in the penetration depth of the evanescent wave of a
probe light (25–19 nm). The penetration depth (dp ) of evanescent wave, defined as
13.5 Hydrogen Bonding States of Interfacial Water Adsorbed on an Alumina Surface Studied 417
Figure 13.7 FUV absorption Wavelength (nm)

spectra of H2 O on the aluminum 150 160 170 180
with incident angles from 58.4∘ to
71.8∘ . Source: Goto et al. [20]/with
Absorption coefficient (cm–1)

8.48 eV (71.8°) H2O
permission of American Chemical
Society. 2 × 105
1 × 105 8.41 eV
(58.4°)
0
8.5 8.0 7.5 7.0
Photon energy (eV)
the distance at which the electric field intensity of probe is attenuated to 1/e from
the surface, is given by;
𝜆
dp (𝜆, 𝜃) = ( (13.1)
( )2 )1∕2
2πn1 sin 𝜃 − n2∕n1
2
where 𝜆 is the wavelength of the probe light, 𝜃 is the incident angle, and n1 and n2
are the refractive indices of the alumina prism and sample, respectively. One can
tune the penetration depth of the evanescent wave of the probe light by varying the
incident angle of an ATR spectrometer.
Figure 13.7 shows the absorption coefficient (𝛼) spectra of liquid water on the
α-alumina with the incident angle from 58.4∘ to 71.8∘ [20]. As 𝜃 increases from
58.4∘ to 71.8∘ , the band maximum shows a higher energy shift from 8.41 to 8.48 eV,
the 𝛼 value increases from 2.01 × 105 cm−1 to 2.39 × 105 cm−1 , and the half-width at
half-maximum (HWHM) on the longer wavelength side becomes broader from 0.598
to 0.630 eV. As a variation in 𝜃 is reflected in that in the dp (Eq. (13.1)), the observed
higher energy shift, increasing absorptivity, and broadening bandwidth with larger
𝜃 (shortening dp ) show the electronic nature of the interfacial water.
Figure 13.8 shows the optimum spectra of the bulk and interfacial phases calcu-
lated with the alternating least squares (ALS) method with the assumption that the
thickness of the interfacial water layer on the alumina surface is ∼2 nm [20]. The
calculated spectra show that the A ̃←X ̃ transitions of the bulk water (dp > 2 nm)
and the interfacial water (<2 nm) were quite different. The 𝜆max value of the inter-
facial water was larger than that of the bulk water by 0.26 eV, and the red tailing of
the Ã←X ̃ band around 7.5–8.0 eV was observed only for the interfacial water. The
higher A ̃←X ̃ energy of the interfacial water relative to the bulk water was associ-
ated with the stronger hydrogen bond network of the interfacial water than the bulk
liquid water. Figure 13.8 also shows the normalized A ̃←X ̃ bands of the solid H2 O
in the I h and amorphous phases [43]. The band maximum position of the interfacial
water was much closer to those of the solid states than that of the bulk water.
̃←X
The red tailing of the A ̃ band around 7.5–8.0 eV appears only in the interfa-
cial water, not in the ice state spectra (Figure 13.8). FUV studies of ice [43] and the
Wavelength (nm) Figure 13.8 Optimum spectra

140 150 160 170 of the bulk and interfacial
phases calculated with the
Ice Liquid alternating least squares (ALS)
method with the assumption
Normalized α (arb. unit)
Broken H-bond acceptor that the thickness of the

interfacial water layer on the
alumina surface is ∼2 nm. FUV
spectra of Ih and amorphous
ice states obtained from Ref.
[19]. Source: Goto et al.
H2O [20]/with permission of
Bulk American Chemical Society.
Interface (2 nm)
Ice (lh)
Ice (amorphous)
9.0 8.5 8.0 7.5 7.0
Photon energy (eV)
theoretical studies of water clusters suggest that the red tailing derives from the par-
tially broken hydrogen-bonded water molecules. The electronic state calculations of
water pentamer clusters by the EOM-CCSD method show that the excitation energy
of the nonbonding electron of a fully hydrogen bond acceptor water molecule located
at the cluster center is higher than that of the monomer by 0.7 eV, while those of
the broken hydrogen bond acceptor ones located on the cluster surface are lower by
0.2 eV [52]. Consequently, the interfacial water spectrum indicates that a denser and
more structured hydrogen bond network accompanied with broken hydrogen bond
acceptors is formed 2 nm below the alumina surface.
13.6 ATR-FUV and Quantum Chemical Calculation

Studies of Hydrogen Bondings in Amides
Morisawa et al. investigated electronic transition and structure and hydrogen
bondings of several kinds of amides and nylons [13, 17, 18]. The amides are good
model compounds of the nylons. Figure 13.9a,b shows ATR-FUV spectra and their
absorption index spectra, respectively, in the 140–260 nm region for five kinds
of amides in the liquid phase, which are formamide (FA), N-methylformamide
(NMF), N-methylacetamide (NMA), N,N-dimethylformamide (NdMF), and
N,N-dimethylacetamide (NdMA) [17]. The absorption index spectra were obtained
by the Kramers–Kronig transformation. All absorption index spectra show a peak
due to the π–π* transition of an amide group in the 180–200 nm region. The first
moment of the band decreases in the order of FA (6.88 eV), NMA (6.81 eV), NMF
(6.67 eV), NdMA (6.44 eV), and NdMF (6.44 eV) as the number of methyl groups on
the N atom increases.
Table 13.1 shows the excitation energies of the first momentum in the gas
and liquid states and the shift between them. The redshift was observed for all
molecules except for NdMF. Notably, it is particularly large for FA (−0.56 eV) and
13.6 ATR-FUV and Quantum Chemical Calculation Studies of Hydrogen Bondings in Amides 419
Photon energy (eV) Photon energy (eV)

8 7 6 5 8 7 6 5
0.6
1.2
FA FA
NMF 0.5 NMF
1.0 NMA
NMA
Absorption index
ATR absorbance
NdMF 0.4 NdMF

0.8 NdMA NdMA
0.3
0.6
0.4 0.2
0.2 0.1
0.0 0.0
160 180 200 220 240 260 160 180 200 220 240 260
Figure 13.9 (a) ATR-FUV spectra of FA, NMF, NMA, NdMF, and NdMA in the liquid phase
and (b) their absorption spectra obtained by Kramers–Kronig transformation. Source:
Morisawa et al. [17]/with permission of AIP Publishing.
Table 13.1 Excitation energies (eV) of the center of balance of the

absorption bands in the gas and liquid phases and the shift between
them.
Gas Liquid Gas-to-liquid
FA 7.44 6.88 −0.56

NMF 7.31 6.67 −0.64
NMA 6.91 6.81 −0.10
NdMF 6.25 6.31 +0.07
NdMA 6.55 6.44 −0.11
Source: Morisawa et al. [17]/with permission of AIP Publishing.
NMF (−0.64 eV). These shifts reflect various factors such as the effects of hydro-
gen bondings, the dipole–dipole interactions, and the polarization by electronic
transition.
To explore the origin of the gas-to-liquid shift, theoretical calculations were car-
ried out for reproducing the shifts. Simulation spectra, which were calculated by
TD-DFT/aug-cc-pVTZ with Gaussian (FWHM = 0.33 eV) convolution, are compared
for the five kinds of amides in a vacuum state and polarized continuum model (PCM)
as shown in Figure 13.10a,b. We employed the polarization using the linear response
(LR)-PCM-TD-DFT calculations which account for the effect of the LR of surround-
ing molecules. The π–π* absorption band of amides appears in the lower energy in
the LR-PCM, which describes that the excitation energy in the liquid phase is lower
than that in the gas phase except for NdMF.
Another important effect in the liquid phase is the direct intermolecular inter-
action via hydrogen bonding. Figure 13.11a,b shows the optimized structures of
the dimers of five amides and their simulation spectra calculated in vacuum by
TD-DFT/aug-cc-pVTZ [17]. In the geometry optimization, FA, NMF, and NMA
have hydrogen-bonded ring structures, while for NdMF and NdMA, structures

9 8 7 6 5 9 8 7 6 5
0.6
0.5 FA
FA
0.5 NMF
NMF
0.4 NMA NMA
Absorbance (a.u.)
Absorbance (a.u.)
NdMF 0.4 NdMF

NdMA NdMA
0.3
0.3
0.2 0.2
0.1 0.1
0.0 0.0
140 160 180 200 220 240 260 140 160 180 200 220 240 260
Figure 13.10 Theoretical spectra of the amides (a) in vacuum calculated by TD-DFT and
(b) in the liquid phase by LR-PCM-TD-DFT. Source: Morisawa et al. [17]/with permission of
AIP Publishing.
Photon energy (eV)

9 8 7 6 5
0.9
0.8 2FA
2NMF
0.7 2NMA
Absorbance (a.u.)
0.6 2NdMF
2NdMA
0.5
FA NMF NMA 0.4
0.3
0.2
0.1
0.0
140 160 180 200 220 240 260
NdMF NdMA
(a) (b) Wavelength (nm)
Figure 13.11 (a) The optimized structure of dimers for five kinds of amides and (b) their
simulation spectra calculated in vacuum by TD-DFT. Source: Morisawa et al. [17]/with
permission of AIP Publishing.
enforced by dipole–dipole interactions between each C=O group are consid-

ered. Comparing Figures 13.10a and 13.11b, it was found that the peak of NMF
appears in the lower energy than those of the other amides upon the formation of
the dimer.
Table 13.2 presents the peak maximum of the simulation spectra calculated by
TD-DFT for five amides. The observed gas-to-liquid shift of FA was −0.56 eV, while
the calculated shift from the vacuum to the LR-PCM-TD-DFT and that from the
monomer to the dimer model were −0.35 and −0.22 eV, respectively. This means
that the contributions of hydrogen bonding and polarization of the solute to the
band shift are both important. In the case of NMF and NMA, the calculated band
shifts for the dimer model, −0.50 and −0.12 eV, well reproduced the experimental
values, −0.64 and −0.10 eV, respectively. The results by LR-PC did not show signif-
icant shifts (−0.02 and +0.02 eV) in the liquid phase of NMF and NMA. For NdMA
and NdMF, neither the dimer model (−0.02 and −0.01 eV) nor the LR-PCM (+0.13
and +0.01 eV) reproduces the experimental results (+0.07 and −0.11 eV). As shown
13.6 ATR-FUV and Quantum Chemical Calculation Studies of Hydrogen Bondings in Amides 421
Table 13.2 Excitation energies (eV) of the π–π* transition of FA, NMF, NMA, NdMF, and
NdMA calculated by TD-DFT.
Monomer Dimer Monomer Gas
Gas PCM Gas Gas → PCM Monomer → dimer

FA 7.80 7.45 7.58 −0.35 −0.22
NMF 7.30 7.28 6.80 −0.02 −0.50
NMA 7.18 7.20 7.06 +0.02 −0.12
NdMF 6.62 6.75 6.60 +0.13 −0.02
NdMA 6.91 6.92 6.90 +0.01 −0.01
Table 13.3 Excitation energies (eV) of the π–π* transition of FA, NMF, NMA, NdMF, and
NdMA calculated by SAC-CI.
Monomer Dimer Monomer Gas
Gas PCM Gas Gas → PCM Monomer → dimer

FA 7.47 7.22 7.17 −0.25 −0.30
NMF 6.47 6.47 6.12 0.00 −0.35
NMA 6.66 6.61 6.77 −0.05 +0.11
NdMF 5.84 5.82 5.63 −0.02 −0.21
NdMA 6.38 6.27 6.44 −0.11 +0.06
in the calculated result of dimer model, small contributions of dipole interaction to

the band shift are trivial. However, the LR-PCM does not take into account nonlinear
effects such as relaxation of the electronic structure of the solvent, and the electronic
spectra of polarized NdMA and NdMF cannot be evaluated.
For these five kinds of amides, the Symmetry Adapted Cluster/Configuration
(SAC-CI) calculations were performed for the excited states of these molecules in
vacuum and in the liquid phase [17]. Table 13.3 summarizes the peak positions
and band shifts of five kinds of amides calculated by the SAC-CI method in the
gas and liquid phases. For the liquid phase, dimer models and the nonequilibrium
SS-PCM-SAC-CI were performed. The trends for FA and NMF are similar to those
obtained by LR-PCM-TD-DFT. As to NMA, the SS-PCM-SAC-CI shows the redshift
of −0.05 eV, which is similar to the experimental value of −0.10 eV. It is of note
that in the TD-DFT results the hydrogen bonding is suggested to be more impor-
tant. Infrared spectroscopy studies have shown that intermolecular C=O…N—N
hydrogen bonds exist in liquid NMA [17]. Using SAC-CI it was found for NdMF
and NdMA that the nonequilibrium SS-PCM model better reproduced the exper-
imental observations than those by the dimer model. The redshifts of NdMF and
NdMA were estimated as −0.02 and −0.11 eV, respectively, by the SS-PCM, in better
agreement with the experimental values of +0.07 and −0.11 eV compared with those
by the dimer model, −0.21 and +0.06 eV. This suggests that the nonlinear effect, the
polarization of the solute, included in the SS-PCM model plays an important role.
13.7 ATR-FUV and Quantum Chemical Calculation

Studies of Hydrogen Bondings in Nylons
Since nylons have both amide group and alkyl chain, their FUV spectra have charac-
ters of these groups observed in the FUV spectra of n-alkanes and amides [18]. FUV
spectra of nylon 6, nylon 11, nylon 12, nylon 6/6, and nylon 6/12, whose chemical
structures are shown in Figure 13.12, were obtained by ATR-FUV spectra as shown
in Figure 13.13.
Figure 13.13a depicts ATR-FUV spectra in the 145–260 nm region of the five kinds
of nylons in the cast films and liquid NMA. For all the nylons two bands are observed
near 150 and 200 nm. Since the FUV spectra of liquid n-alkanes and liquid NMA give
only a band at around 155 and 200 nm, respectively, the 150 and 200 nm features of
O
O O
N (CH2)11 C
N (CH2)5 C N (CH2)10 C
H
H H n
n n
(a) (b) (c)
O O O O
N (CH2)6 N C (CH2)4 C N (CH2)6 N C (CH2)10 C
H H H H
n n
(d) (e)
Figure 13.12 Chemical structures of (a) nylon 6, (b) nylon 11, (c) nylon 12, (d) nylon 6/6,
and (e) nylon 6/12. Source: Morisawa et al. [18]/with permission of American Chemical
Society.

8 7 6 5 8 7 6 5
0.7 1.4
Nylon6
0.6 Nylon6 1.2 Nylon11
Nylon11
Normalized absorbance
Nylon12
0.5 Nylon12 1.0
ATR absorbance
Nylon6/6
Nylon6/6
Nylon6/12
Nylon6/12
0.4 NMA (amide)
0.8
0.3 0.6
0.2 0.4
0.1 0.2
0.0 0.0
160 180 200 220 240 260 160 180 200 220 240 260
Figure 13.13 (a) ATR-FUV spectra of nylon 6, nylon 11, nylon 12, nylon 6/6, nylon 6/12,
and liquid NMA in cast films and (b) their normalized spectra with the band near 200 nm.
Source: Morisawa et al. [18]/with permission of American Chemical Society.
13.7 ATR-FUV and Quantum Chemical Calculation Studies of Hydrogen Bondings in Nylons 423
Figure 13.14 Calculated Photon energy (eV)

spectra of model 10 9 8 7 6 5
compounds of nylon 6, 0.7
nylon 11, and nylon 12 in 1.6 Nylon 6
vacuum. Source: Morisawa 1.4 0.6
σ-3py Rydberg
Oscillator strength (au)

et al. [18]/with permission Nylon 11
Absorbance (au)
1.2 0.5
of American Chemical
Society. 1.0 0.4
Nylon 12
0.8
0.3
0.6
π–π* 0.2
0.4
0.2 0.1
0.0 0.0
120 140 160 180 200 220 240
Wavelength (nm)
nylons are assigned to the alkane’s σ-Rydberg 3p and amide’s π–π* band, respec-
tively [18]. Figure 13.13b shows the normalized spectra with the intensity of the
peak of π–π* band. The normalized intensity of the 150 nm band increases in the
order of nylon 6 < nylon 6/6 < nylon 6/12 < nylon 11 < nylon 12, which is followed
to the relative ratio of CH2 groups.
To interpret ATR-FUV spectra of nylons, we performed TD-DFT calculations for
monomer units of nylon 6, nylon 11, and nylon 12. Figure 13.14 shows the simulation
spectra of the model compounds which are convoluted by Gaussian with the width of
0.33 eV, and calculated oscillator strength and excitation wavelength are also yielded
in Figure 13.14 (as shown by stick diagram). The simulation spectra yield bands near
140 and 170 nm, supporting the above band assignments.
The effects of hydrogen bonding on the FUV spectra of nylons were discussed
using TD-DFT calculation of the monomer, dimer, and trimer models of nylon 6,
nylon 11, and nylon 12. The calculated structures of trimer models of nylon 6 and
nylon 6/6 are shown in Figure 13.15 with calculated transition dipole strengths. In
nylon 6 model, the structure is the α-form of the crystal in which hydrogen bonds
are perfectly constituted by inverting alternate chains, and the progressive shear
3.856 au (177.0 nm)
3.770 au (171.7 nm) 8.112 au (176.7 nm)

2.325 au (170.8 nm)
(a) (b)
Figure 13.15 Computational trimer models of hydrogen-bonded systems for (a) nylon 6
and (b) nylon 6/6 with the transition dipole strengths. Source: Morisawa et al. [18]/with
structure is adopted in the α-form of nylon 6/6. These trimer models correspond
to the hydrogen-bonded system of unit cell for the π–π* transitions. The TD-DFT
results of the π–π* and few σ-Rydberg transitions are summarized in Ref. [18].
The peak positions of the π–π* bands which are convoluted as a band show a longer
wavelength shift as 171, 172, and 173 nm for the monomer, dimer, and trimer mod-
els of nylon 6, respectively. Those for the monomer, dimer, and trimer models of
nylon 6/6 present the more significant redshift as 171, 175, and 176 nm than nylon 6.
The absorption intensity is considerably distributed to many transitions in the wide
range of spectra in the dimer or trimer models by hydrogen bonding effects. The
deviation in the calculated redshift of 3 nm (0.12 eV) between nylon 6 and nylon
6/6 qualitatively matches the experimental value of 1.7 nm (0.06 eV). In the nylon
6 trimer model, three dominant transitions at 170.8, 171.7, and 177.0 nm with simi-
lar oscillator strength exist because of its crystal structure, which is, the inversion of
alternating chains. On the other hand, in the nylon 6/6 trimer model, an intense tran-
sition at 176.7 nm is enhanced as seen in Figure 13.15. The differences in transition
dipole coupling scheme can be attributed to the deviation of the redshift between
nylon 6 and nylon 6/6.
13.8 An ATR-FUV Study for Poly(ethylene glycol) (PEG)

and Its Complex with Lithium Ion (Li+ )
Concept of hydrogen bonding is interactions of coordination bonds between non-
bonding electron of Lewis bases and terminus hydrogen of Lewis acids. Similar inter-
action is found in the coordination bonds between nonbonding electron and alkali
metal cations [22]. Especially for Li+ /poly(ethylene glycol) (PEG) complex, because
of characteristic properties such as electronic conductivity, the complexes have been
employed for various purposes in electrochemistry, electric devices, and lithium-ion
battery [22]. In this section, the effects on the ATR-FUV spectra by the coordination
bond between Li+ and PEG molecules are addressed. In the FUV spectra, we can
easily observe the electronic transitions from nonbonding electrons which reflect
the difference in the interaction from environment of the molecules.
Since the alkali–metal complexes do not have charge transfer (CT) transitions,
electronic states of alkali–metal complexes have not been investigated. Moreover,
since electronic states of ethers compose only single bonds, the electronic struc-
tures determining the energies of ground and excited orbitals in the condensed
phase have not been understood previously, too. The ATR-FUV spectra of PEG
were measured as shown in Figure 13.16a [22]. The transitions calculated in the
145–220 nm region for diethylene glycol diethyl (DGDE) and triethylene glycol
(TG) by TD-DFT are shown along with the convoluted spectra in Figure 13.16b.
TG is an oligomer of the ethylene glycol. There are two types of n-orbitals that are
found in the oxygen of the ether, n(CH2 OCH2 ) and hydroxyl terminal n(OH). In the
case of DGDE, the transitions from nonbonding orbitals at three oxygen atoms of
ether, n(CH2 OCH2 ), to the Rydberg 3s and 3p orbitals were observed. On the other
hand, in the case of TG, n(OH)-Rydberg 3p transitions were observed in addition
13.8 An ATR-FUV Study for Poly(ethylene glycol) (PEG) and Its Complex with Lithium Ion (Li+ ) 425
Excitation energy (eV)

8.5 8.0 7.5 7.0 6.5 6.0
DG 12
0.30 n(OH)-3p Ryd. 0.10
TG DGDE
PEG200
ε/103 mol–1 dm3 cm–1

0.25 TG
Oscilator strength
ATR–absorbance
PEG300 0.08
0.20 8 n(CH2OCH2)-3p Ryd.
PEG400
n(CH2OCH2)-3s Ryd. 0.06
0.15
0.10 4 0.04
0.05 0.02
0.00
0 0.00
160 180 200 150 160 170 180 190 200 210 220
Figure 13.16 (a)ATR-FUV spectra in the 145–200 nm region for DG, TG, DGDM, DGDE, and
TGDM in the liquid phase. (b) Simulation spectra of TG and DGDE calculated using the
TD-CAM-B3LYP/6–311++G(2d,p) level of theory. Source: Ueno et al. [22]/with permission of
to the n(CH2 OCH2 )-Rydberg 3s and 3p. As seen in Figure 13.16a, PEGs have three
bands at about 155, 163, and 177 nm. The assignments of the transitions were; 155
(n(OH) – 3p Ryd), 163 (n(CH2 OCH2 ) – 3p Ryd), and 177 nm (n(CH2 OCH2 ) – 3s
Ryd) [22]. As the result, the n orbitals in the PEG can be distinguished between the
OH and ether groups in PEG.
To analyze the effect of coordination bond between Li+ and PEG on the electronic
structure of the complex, Li+ -concentration dependence of ATR-FUV spectra
for solutions composed of PEG400 and LiTFSI was investigated as shown in
Figure 13.17 [22]. A value for composition of the mixture, C, denotes the ratio of
ether unit to Li ion, [COC]/[Li+ ], and the small value of C indicates the high concen-
tration of lithium salt in the solutions. As shown in Figure 13.18, ATR-absorbance
0.35
0.30
0.25
ATR-absorbance
0.20
0.15
0.10
PEG400
C = 48
0.05
C=8
C = 0.5
0.00
160 180 200
Wavelength (nm)
Figure 13.17 Dependence of absorbance on the concentration of Li+ for mixture samples
composed of PEG400 and LiTFSI. Source: Ueno et al. [22]/with permission of American
Chemical Society.
Figure 13.18 (a) Wave-

length shift (𝛿𝜆) of three
0 bands at 155, 163, and
177 nm vs. the
concentration of Li
complexes. (b) Intensity
–2
ratio (Asample /Aneat PEG) of
the three bands at 155, 163,
δλ (nm)
and 177 nm for the binary

mixture and PEG vs. the
concentration. Source: Ueno
–4 et al. [22]/with permission
of American Chemical
155 nm Society.
163 nm
–6
176 nm
(a) 1.0
0.8
0.6
Asample/Apeg
0.4
155 nm
0.2
163 nm
176 nm
0.0
0 2 4 6 8 10 12 14 16 24 32 48
(b) Concentration (C, molar ratio)
of the solutions decreases with decreasing C in which the concentration of LiTFSI

increases [22]. Figure 13.18a,b shows the peak shift and intensity ratio versus C,
respectively. The 163 nm band did not show the shift in the entire concentration
range. Shorter wavelength shifts of 155 and 177 nm bands occur according to a
decrease in C. The 155 nm band shows a large shift in the low-C region (C < 8),
while, the 177 nm band rapidly shifts from the high-C region (C < 16).
To analyze the observed variations in the ATR-FUV spectra of PEG/Li+ complex
caused by Li+ coordination, TD-DFT calculations were performed for the three mod-
els of complexes as shown in Figure 13.19 [22].
All the models calculated have intercalation coordination structure. The Int 1
and 2 form the coordination between one Li+ ion and two PEGs each having five
oxygen atoms. A difference between model Int 1 and Int 2 is the terminal groups
of the ligands. Int 1 has two OH and two methyl ether groups as terminal, while
13.9 Summary and Perspective 427
(a) (b) (c)
Figure 13.19 The complex structure models; (a) Int 1, (b) Int 2, (c) Int 3. Source: Ueno et al.
[22]/with permission of American Chemical Society.
Figure 13.20 Simulation

spectra of neat PEG and 1.6
complexes (Int 1, Int 2, and Int
ε / 103 mol–1 dm–1 cm–1
3) structure model by TD-DFT.

Source: Ueno et al. [22]/with 1.2
permission of American
Chemical Society.
0.8
PEG
0.4 Int 1
Int 2
Int 3
0.0
150 160 170 180 190 200
Wavelength (nm)
the Int 2 has four OH as the terminal groups of ligands. In Int 3, two Li+ ions are
contained in the ligand of Int 2. It is important that the differences in numbers
of ether O atoms that do not coordinate with Li+ . For Int 1, Int 2, and Int 3, the
numbers of the ether without coordination are 4, 2, and 0, respectively. Figure 13.20
displays simulation spectra of the above three models and pure PEG. Differences in
simulation spectra between PEG and Li+ /PEG complex were explained by the varia-
tion of concentration dependences on the ATR-FUV spectra of Li+ /PEG. The origin
of the intensity decrease around 177 nm with decreasing C was clearly understood
as the decrease of ether oxygen atom without coordinate since only the transition
from the un-coordinate ether is calculated in the region. We assign a band in the
145–155 nm region observed in the low-C region to the new transitions from the
Li+ /PEG complex because the simulation spectra of the complex models also give a
new band in this wavelength region. As the result, the blueshift around 155 nm is
due to the appearance of a new band of the complex.
13.9 Summary and Perspective
This chapter has reported that ATR-FUV spectroscopy holds unique advantages for
investigating hydrogen bondings and intermolecular interactions of molecules in
condensed phase. In the FUV region, one can observe bands due to transitions of
π electrons, lone pair electrons, and σ electrons, and thus, ATR-FUV spectroscopy
allows to investigate electronic states of almost all kinds of molecules. Therefore,

FUV spectroscopy is suitable to investigate hydrogen bondings and intermolecular
interactions of various molecules and materials from the points of electronic struc-
ture and transitions. In this chapter, hydrogen bondings studies of water, aqueous
solutions, adsorbed water, and the two kinds of polymers, nylons, and PEG and its
complex with lithium ion (Li+ ) have been reviewed.
In this section, perspective of FUV spectroscopy will be discussed from the
points of several directions. First of all, one should note that the region of ATR-UV
spectroscopy is expanding to both shorter and longer wavelengths sides. From the
starting point of ATR-FUV spectroscopy, FUV spectroscopy has been combined
with DUV spectroscopy, partly because the FUV and DUV regions are seamless and
partly because the first-generation ATR-FUV spectrometers could be used in the
145–300 nm region. ATR measurements in the whole FUV–DUV–UV–Vis region
(140–800 nm) will become more popular to collect various information about the
electronic structure and transitions of a variety of compounds. It may be possible to
expand the measurement region to 120 nm by purging a spectrometer with Ar gas
instead of N2 gas. In the 120–145 nm region, one can observe bands due to the σ–σ*
transitions.
There are many possibilities for the development of novel FUV spectroscopy
instrumentation. An electrochemical-attenuated total reflection (EC-ATR) spec-
trometer in the FUV–DUV–UV–Vis region is very promising. Tanabe et al. have
already demonstrated a good example of application of EC-ATR to the electronic
transition spectra of interfacial ionic liquids [34]. Time-resolved ATR-FUV spec-
troscopy has already been proposed [34, 35], and its further development and
applications are expected. ATR-FUV imaging may emerge soon; high special resolu-
tion imaging for top surfaces can be expected. The ATR-FUV imaging may provide
a novel possibility for hydrogen bonding studies. The development of higher energy
UV light sources and optical elements is expected to advance instrumentation in
the FUV region.
Quantum chemical calculations, such as TD-DFT and SAC-CI, have been used
extensively in the band assignments and studies of electronic transitions and states
in ATR-FUV spectroscopy. It is important to deepen the application of these quan-
tum chemical calculation methods. As described in this chapter, ALS method is
very useful to analyze VA-dependent spectral variations. Self-modeling curve reso-
lution methods should probably be used more for band decomposition of ATR-FUV
spectra. It may be good idea to apply various chemometrics and two-dimensional
correlation analysis methods for the spectral analysis of ATR-FUV spectra.
ATR-FUV spectroscopy is basically a top surface analysis method. The penetra-
tion depth of an ATR-FUV spectrometer is approximately 100 nm, and thus, it is
possible to explore the electronic structure and molecular structure of a top sur-
face. VA-ATR-FUV technique may be used more for surface studies. It is noted that
ATR-FUV spectroscopy may allow us to disclose even the top surface of a biologi-
cal cell.
ATR-FUV spectroscopy has paved the way toσchemistry. It may be used to explore
intermolecular interactions between n-alkanes, such as a CH· · ·HC interaction in
References 429
the condensed phase. The origin of the unique behavior in the phase transition
of alkanes may be another target of ATR-FUV spectroscopy. These studies are
very much concerned with hydrogen bonding studies. Chemical reactions con-
taining a C—H bond can also be investigated through ATR-FUV spectroscopy,
which may shed light on condensed matter physics through electronic transition
absorption.
One of the most exciting new application fields is the application of FUV spec-
troscopy to biomedical science. One can use ATR-FUV spectroscopy to investigate
the structure and electronic state of biological molecules. Proteins, carbohydrates,
nucleic acids, and lipids are expected to provide characteristic bands in the FUV
region. Based on these basic data, one can analyze the FUV spectra of more com-
plicated biological molecules and materials, such as lectin, glycoproteins, biomem-
branes, erythrocyte, blood, and cultured cells. It is also possible to investigate the
very surface of biological materials.
In this way, ATR-FUV spectroscopy holds considerable promise in a variety of
exciting research fields.
References
1 Ozaki, Y. and Kawata, S. (2015). Far- and Deep-Ultraviolet Spectroscopy. Springer.

2 Ozaki, Y., Morisawa, Y., Ikehata, A., and Higashi, N. (2012). Far-ultraviolet
spectroscopy in the solid and liquid states: a review. Appl. Spectrosc. 66: 1–25.
3 Morisawa, Y., Tanabe, I., and Ozaki, Y. (2018). Advances in far-ultraviolet
spectroscopy in the solid and liquid states. In: Frontiers and Advances in Molecu-
lar Spectroscopy (ed. J. Laane), 251–285. Elsevier.
4 Morisawa, Y., Tanabe, I., and Ozaki, Y. (2020). Far-ultraviolet (FUV) spectroscopy
in the solid and liquid states, principle, instrumentation, and application. Ency-
clopedia Anal. Chem., https://doi.org/10.1002/9780470027318.a9279.pub2.
5 (a) Ozaki, Y. and Tanabe, I. (2016). Far-ultraviolet spectroscopy of solid and
liquid states; characteristics, instrumentations, and applications. Analyst 141:
3962–3981. (b) Tanabe, I. and Ozaki, Y. (2016). Far- and deep-ultraviolet spec-
troscopic investigations for titanium dioxide: electronic absorption, Rayleigh
scattering, and Raman spectroscopy. J. Mater. Chem. C 4: 7706–7717.
6 Ozaki, Y. (2019). Recent advances in molecular spectroscopy of electronic and
vibrational transitions in condensed phase and its application to chemistry. Bull.
Chem. Soc. Jpn. 92: 629–654.
7 Ozaki, Y., Morisawa, Y., Tanabe, I., and Bec, K.B. (2021). ATR-far-ultraviolet
spectroscopy in the condensed phase-The present status and future perspectives.
Spectrochim. Acta Part A 253: 119549.
8 Robin, M.B. (1974). Higher Excited States of Polyatomic Molecules (ed. I.I. Vol).
New York, NY: Academic Press Inc.
9 Robin, M.B. (1985). Higher Excited States of Polyatomic Molecules, vol. III, 176.
Orlando, FL: Academic Press.
10 Sándorfy, C. (ed.) (1999). The Role of Rydberg States in Spectroscopy and

Photochemistry-Low and High Rydberg States. Dordrecht: Kluwer Academic
Publishers.
11 Higashi, N., Ikehata, A., and Ozaki, Y. (2007). An attenuated total reflectance
far-UV spectrometer. Rev. Sci. Instrum. 78: 103107.
12 Ikehata, A., Ozaki, Y., and Higashi, N. (2008). Direct observation of the absorp-
tion bands of the first electronic transition in liquid H2 O, D2 O by attenuated
total reflectance far-UV spectroscopy. J. Chem. Phys. 129: 234510.
13 Ehara, M. and Morisawa, Y. (2019). Theoretical and experimental molecular
spectroscopy of the far-ultraviolet region. In: Molecular Spectroscopy: A Quantum
Chemistry Approach, vol. 1 (ed. Y. Ozaki, M.J. Wojcik and J. Popp), 119–145.
Wiley-VCH.
14 Morisawa, Y., Ikehata, A., Higashi, N., and Ozaki, Y. (2009). Attenuated total
reflectance–far ultraviolet (ATR–FUV) spectra of CH3 OH, CH3 OD, CD3OH and
CD3 OD in a liquid phase ∼Rydberg states∼. Chem. Phys. Lett. 476: 205–208.
15 Morisawa, Y., Ikehata, A., Higashi, N., and Ozaki, Y. (2011). Low-n Rydberg
transitions of liquid ketones studied by attenuated total reflection far-ultraviolet
spectroscopy. J. Phys. Chem. 115: 562–568.
16 Morisawa, Y., Tachibana, S., Ehara, M., and Ozaki, Y. (2012). Elucidating
electronic transitions from σ orbitals of liquid n- and branched alkanes by
far-ultraviolet spectroscopy and quantum chemical calculations. J. Phys. Chem.
116: 11957–11964.
17 Morisawa, Y., Yasunaga, M., Fukuda, R. et al. (2013). Electronic transitions in
liquid amides studied by using attenuated total reflection far-ultraviolet spec-
troscopy and quantum chemical calculations. J. Chem. Phys 139: 154301.
18 Morisawa, Y., Yasunaga, M., Sato, H. et al. (2014). Rydberg and π–π* transitions
in film surfaces of various kinds of nylons studied by attenuated total reflection
far-ultraviolet spectroscopy and quantum chemical calculations: peak shifts in
the spectra and their relation to nylon structure and hydrogen bondings. J. Phys.
Chem. B 118: 11855–11861.
19 Morisawa, Y., Tachibana, S., Ikehata, A. et al. (2017). Changes in the electronic
states in low-temperature solid n-tetradecane: decrease in the HOMO-LUMO
gap. ACS Omega 2: 618–625.
20 Goto, T., Ikehata, A., Morisawa, Y., and Ozaki, Y. (2015). Surface effect of alu-
mina on the first electronic transition of liquid water studied by far-ultraviolet
spectroscopy. J. Phys. Chem. Lett. 6: 1022–1026.
̃←X
21 Goto, T., Beć, K.B., and Ozaki, Y. (2017). Interpretation of the A ̃ transition
of hydrated solutions observed in the far-UV region with quantum chemical
calculations. Phys. Chem. Chem. Phys. 19: 21490.
22 Ueno, N., Wakabayashi, T., Sato, T., and Morisawa, Y. (2019). Changes in the
electronic transitions of polyethylene glycol upon the formation of a coordinate
bond with Li+ studied by ATR far-ultraviolet spectroscopy. J. Phys. Chem. A 123:
10746–10756.
23 Morisawa, Y., Tanimura, E., Ehara, M., and Sato, H. (2021). Attenuated total
reflection-far-ultraviolet spectroscopy and quantum chemical calculations of
References 431
electronic structure of the top surface and bulk of polyethylenes with different
crystallinities. Appl. Spectrosc. 75: 971–979.
24 Ikehata, A., Mitsuoka, M., Morisawa, Y. et al. (2010). Effect of cations on absorp-
tion bands of first electronic transition of liquid water. J. Phys. Chem. 114:
8319–8322.
25 Goto, T., Ikehata, A., Morisawa, Y. et al. (2012). The effect of metal cations on
the nature of the first electronic transition of liquid water as studied by atten-
uated total reflection far-ultraviolet spectroscopy. Phys. Chem. Chem. Phys. 14:
8097–8104.
26 Goto, T., Ikehata, A., Morisawa, Y. et al. (2012). Effects of lanthanoid cations
on the first electronic transition of liquid water studied using attenuated total
reflection far-ultraviolet spectroscopy: ligand field splitting of lanthanoid hydrates
in aqueous solutions. Inorg. Chem. 51: 10650–10656.
27 Goto, T., Ikehata, A., Morisawa, Y., and Ozaki, Y. (2013). Electronic transitions
of protonated and deprotonated amino acids in aqueous solution in the region
145–300 nm studied by attenuated total reflection far-ultraviolet spectroscopy.
J. Phys. Chem. A 117: 2517–2528.
28 Tanabe, I. and Ozaki, Y. (2014). Consistent changes in electronic states and pho-
tocatalytic activities of metal (Au, Pd, Pt)-modified TiO2 studied by far-ultraviolet
spectroscopy. Chem. Commun. 50: 2117–2119.
29 (a) Tanabe, I., Ryoki, T., and Ozaki, Y. (2014). Significant enhancement of photo-
catalytic activity of rutile TiO2 compared with anatase TiO2 upon Pt nanoparticle
deposition studied by far-ultraviolet spectroscopy. Phys. Chem. Chem. Phys.
16: 7749–7753. (b) Tanabe, I., Ryoki, T., and Ozaki, Y. (2015). The effects of
Au nanoparticle size (5–60 nm) and shape (sphere, rod, cube) over electronic
states and photocatalytic activities of TiO2 studied by far- and deep-ultraviolet
spectroscopy. RSC Adv. 5: 13648–13652.
30 (a) Tanabe, I., Yamada, Y., and Ozaki, Y. (2016). Far- and deep-UV spectroscopy
of semiconductor nanoparticles measured based on attenuated total reflectance
spectroscopy. ChemPhysChem 17: 516–519. (b) Tanabe, I. and Kurawaki, Y.
(2018). Far-ultraviolet spectral changes of titanium dioxide with gold nanoparti-
cles by ultraviolet and visible light. Spectrochim. Acta, Part A 197: 103–106.
31 (a) Tanabe, I., Kurawaki, Y., Morisawa, Y., and Ozaki, Y. (2016). Electronic
absorption spectra of imidazolium-based ionic liquids studied by far-ultraviolet
spectroscopy and quantum chemical calculations. Phys. Chem. Chem. Phys. 18:
22526–22530. (b) Tanabe, I., Suyama, A., Sato, T., and Fukui, K. (2018). System-
atic analysis of various ionic liquids by attenuated total reflectance spectroscopy
(145–450 nm) and quantum chemical calculations. Analyst 143: 2539–2545.
32 Bec, K.B., Morisawa, Y., Kobashi, K. et al. (2018). Rydberg transitions as a
probe for structural changes and phase transition at polymer surfaces; as
ATR-FUV-DUV and quantum chemical study of poly(3-hydroxybutyrate) and
its nanocomposite with graphene. Phys. Chem. Chem. Phys. 20: 8859–8873.
33 Bec, K.B., Morisawa, Y., Kobashi, K. et al. (2018). Electronic structure of
graphene in far- and deep-ultraviolet region: attenuated total reflection
spectroscopy and quantum chemical calculation study. J. Phys. Chem. C 122:

28998–29008.
34 Tanabe, I., Suyama, A., Sato, T., and Fukui, K. (2019). Potential dependence of
electronic transition spectra of interfacial ionic liquids studied by newly devel-
oped electrochemical attenuated total reflectance spectroscopy. Anal. Chem. 91:
3436–3442.
35 Goto, T., Morisawa, Y., Higashi, N. et al. (2013). Pulse laser photolysis of aque-
ous ozone in the microsecond range studied by time-resolved far-ultraviolet
absorption spectroscopy. Anal. Chem. 85: 4500–4506.
36 Higashi, N., Yokota, H., Hiraki, S., and Ozaki, Y. (2005). Direct determination
of peracetic acid, hydrogen peroxide, and acetic acid in disinfectant solutions by
far-ultraviolet absorption spectroscopy. Anal. Chem. 77: 2272–2277.
37 (a) Higashi, N. and Ozaki, Y. (2004). Potential of far-ultraviolet absorption spec-
troscopy as a highly sensitive quantitative and qualitative analysis method for
aqueous solutions. Part I: Determination of hydrogen chloride in aqueous solu-
tions. Appl. Spectrosc. 58: 910–916. (b) Higashi, N., Ikehata, A., Kariyama, N.,
and Ozaki, Y. (2008). Potential of far-ultraviolet absorption spectroscopy as a
highly sensitive analysis method for aqueous solutions. Part II: Monitoring
the quality of semiconductor wafer cleaning solutions using attenuated total
reflection. Appl. Spectrosc. 62: 1022–1027.
38 (a) Tanabe, I., Tanaka, Y.Y., Ryoki, T. et al. (2016). Direct optical measurements
of far- and deep-ultraviolet surface plasmon resonance with different refrac-
tive indices. Opt. Express 24: 21886–21896. (b) Tanabe, I., Tanaka, Y.Y., Watari,
K. et al. (2017). Far- and deep-ultraviolet surface plasmon resonance sensors
working in aqueous solutions using aluminum thin films. Sci. Rep. 7: 5934. (c)
Tanabe, I., Shimizu, M., Kawabata, R. et al. (2020). Far- and deep-ultraviolet
surface plasmon resonance using Al film for efficient sensing of organic thin
overlayer. Sens. Actuators A 301: 111661.
39 Hashimoto, K., Morisawa, Y., Tortora, M. et al. ATR-far-ultraviolet spectroscopy
is a sensitive tool for investigation of protein adsorption. Appl. Spectrosc.
40 Watanabe, K. and Zelikoff, M. (1953). Absorption coefficients of water vapor in
the vacuum. Ultraviolet J. Opt. Soc. Am. 43: 753–755.
41 Painter, L., Hamm, R., Arakawa, E., and Birkhoff, R. (1968). Phys. Rev. Lett. 21:
282–284.
42 Seki, M., Kobayashi, K., and Nakahara, J. (1981). Optical spectra of hexagonal
ice. J. Phys. Soc. Jpn. 50: 2643–2648.
43 Ohtomo, N. and Arakawa, K. (1979). Bull. Chem. Soc. Jpn. 52: 2755.
44 Vinogradov, E., Smirnov, P., and Trostin, V. (2003). Structure of hydrated com-
plexes formed by metal ions of groups I—III of the periodic table in aqueous
electrolyte solutions under ambient conditions. Russ. Chem. Bull. Int. Ed. 52:
1253–1271.
45 Marcus, Y. (1991). Thermodynamics of solvation of ions. Part 5. Gibbs free
energy of hydration at 298.15 K. J. Chem. Soc. Faraday Trans. 87: 2995–2999.
46 Grigoriew, H. and Siekierski, S. (1993). J. Phys. Chem. 97: 5400–5402.
References 433
47 Kameda, Y. and Uemura, O. (1993). Neutron diffraction study on the structure of

highly concentrated aqueous LiBr solutions. Bull. Chem. Soc. Jpn. 66: 384–389.
48 Zhao, Y. and Truhlar, D.G. (2006). Density functional for spectroscopy:
no long-range self-interaction error, good performance for Rydberg and
charge-transfer states, and better performance on average than B3lYP for ground
states. J. Phys. Chem. A 110: 13126–13130.
49 Catalano, J.G., Park, C., Zhang, Z., and Fenter, P. (2006). Termination and
water adsorption at the α-Al2 O3 (012)−aqueous solution interface. Langmuir
22: 4668–4673.
50 Sung, J., Zhang, L., Tian, C. et al. (2011). Effect of pH on the water/α-Al2O3
(1102) interface structure studied by sum-frequency vibrational spectroscopy.
J. Phys. Chem. C 115: 13887–13893.
435
14
Water–Hydrogen-Bond Network and Hydrophobic Effect

Barbara Zupančič and Jože Grdadolnik
National Institute of Chemistry, Theoretical Department, Hajdrihova 19, SI-1000, Ljubljana, Slovenia
Symbols and Abbreviations

Symbol Description
addALC_ci Additional portion of bulk water component to
compensate negative bands
̂
bALCci Bulk water spectral component coefficient from
decomposition
bALCci Bulk water portion in alcohol-aqueous solutions
BW ALC_c and BW ALC_Ti Bulk water spectral component from decomposition
̂cALC ci Solute-correlated spectral component coefficient from
decomposition
ci Volume concentration
Hw Band halfwidth
M Molarity
m Molality
P Pressure
p Subtraction parameter
̂rALC ci Residual spectral component
SALC_ci and SCALC_Ti Spectrum of alcohol-water solution at a given
concentration and temperature
̂ ALC
SC c Solute-correlated spectral component from
decomposition
SCALC_ci and SCALC_Ti Solute-correlated spectrum at a given concentration and
temperature
T Temperature
𝛿 HOH HOH bending frequency
ΔCp Change in heat capacity at constant pressure
ΔG Change in Gibbs free energy
Δ𝜈 OD Shift of OD stretching
ΔS Change in entropy
436 14 Water–Hydrogen-Bond Network and Hydrophobic Effect
𝜈 CH CH stretching frequency
𝜈 CO CO stretching frequency
𝜈 OD OD stretching
𝜈 OH OH stretching
Abbreviation Long name

ALS Alternating Least Squares (algorithm)
ATR Attenuated Total Reflection
EtGOH Ethylene glycol
EtOH Ethanol
EXAFS Extended X-ray Absorption Fine Structure
FT Fourier Transformation
GlOH Glycerol
IR Infrared
isoBtOH Isobutanol
MCR Multivariate Curve Resolution (method)
MD Molecular dynamics
MeOH Methanol
NMR Nuclear Magnetic Resonance
PCA Principal Component Analysis
PLS Partial Least Square
PT diagram Pressure, temperature phase diagram
SC Solute-correlated
S/N Signal to noise ratio
tertBtOH Tert-butyl alcohol
1BtOH 1-butanol
1PrOH 1-propanol
2BtOH 2-butanol
2PrOH 2-propanol
14.1 Introduction
Water represents a contradictory liquid. There is not a single matter which possesses
so many ambiguous properties. Indeed, it is much more difficult to find one that
obeys the properties found in comparable compounds. All these unusual properties
originate in an extreme density of proton donor and proton acceptor groups of the
water molecule. Moreover, it forms a nearly tetrahedral yet flexible hydrogen-bond
network in the condensed phase. This labile network is subject to incessant bond
breaking and formation on a picosecond time scale. This unique property makes the
study of the local structure and dynamics of water a major challenge and, due to the
fundamental importance of water in the chemical and biological sciences, generates
strong interest in the development of microscopic images of the bulk water and its
response to a hydrophobic and/or hydrophilic perturbation.
Understanding the changes in the local network structure and dynamics of water
in the vicinity of solutes is crucial for unraveling the subtle mechanisms responsible
for numerous phenomena, such as the formation of the three-dimensional struc-

ture of proteins, the segregation of polar and nonpolar liquids, and the formation of
micelles, to name a few. The structural pictures of hydration are strongly influenced
by what is known about the structure of ice. The most stable form of solid water is
hexagonal ice, in which each water molecule forms four tetrahedrally coordinated
hydrogen bonds, two of which are acceptor and two donor bonds. This tetrahedral
coordination is maintained in the liquid phase, but not completely due to the ther-
mal motion of the molecules. The formation of an interface upon solute addition
further modifies the local bonding in water.
The basic arrangement of water molecules is tetrahedral and the number of
hydrogen bonds is similar to that of covalent bonds, which is an exceptional
situation. Infrared (IR) spectroscopy clearly shows that in liquid water almost
all OH groups form hydrogen bonds; consequently, the number of “free” OH
groups remains very small [1, 2]. This suggests that liquid water, together with
this basic tetrahedral structure, which may, however, be more disordered, retains a
hydrogen-bond network almost as dense as ice. So the question arises: how is this
hyper-dense hydrogen-bond network in liquid water modified to allow the fluidity
that, in combination with this density of hydrogen bonds, is the origin of all the
extraordinary properties of water? Many experimental methods and theoretical
approaches have been used to answer this question, but so far, no definitive
conclusion has been reached.
IR (and Raman) spectroscopy is known to be the method of choice in the study
of hydrogen-bonded systems because it is particularly sensitive to the presence of
hydrogen bonds. Moreover, it is commonly used as a most reliable sensitive and
model-free approach for determining the relative strength of hydrogen bonds [3–9].
Two characteristic features can be observed in the stretching band of the proton
donor group. When hydrogen bond is formed, the OH (OD)-stretching frequency
shifts to a lower frequency [10, 11] with a simultaneous increase in the integral inten-
sity of the stretching band [12]. On contrary, the shift of the OH (OD) deformation
and HOH (DOD) libration bands on formation or strengthening of hydrogen bond-
ing occurs in the opposite direction [2, 13] to higher wavenumbers. Moreover, the
frequency shift can be correlated with the strength of the hydrogen bond formed [3].
Since the OH-stretching frequency is sensitive to both the geometry of the water
molecule and its environment, vibrational spectroscopy provides an excellent
probe of the properties of the hydrogen-bond network in water. In general, in the
vibrational spectrum of water solution, the H2 O bands are altered by the solutes
compared to those of pure water. Therefore, the shift in the OH stretching, defor-
mation and libration bands of water, as well as the band shape could be the most
informative probe for the strengthening/weakening of hydrogen bonds.
However, the application of IR spectroscopy in the study of the structure and
dynamics of water solution or water itself is not trivial work. Due to the very high
absorption index of the water molecule in the infrared region, it is very easy, espe-
cially in the transmission mode, to saturate the most intense band(s) [14]. To avoid
saturation, thin spacers should be used, usually thinner than 1 μm, but their appli-
cation can produce interference fringes that completely distort the spectrum. The
solution represents the application of attenuated total reflection (ATR), where only
the evanescent wave emerging from the crystal is applied, in which the infrared
beam undergoes total reflection. The evanescent beam represents only a small part
of the original infrared beam, avoiding the saturation of intensity.
The aim of this chapter is to demonstrate the power and efficiency of IR spec-
troscopy in revealing distortion of the hydrogen-bonding network of water, espe-
cially with respect to the hydrophobic effect. It is organized as follows. First, IR
spectroscopy of bulk water is briefly discussed, primarily to define the spectral prop-
erties used in interpreting the spectra of water solutions of pure hydrophobic species
and various types of alcohols. This introduction to water spectroscopy is followed by
a presentation of the infrared spectra, upgraded with a theoretical treatment describ-
ing the physical origin(s) of the hydrophobic effect. The chapter ends with a com-
prehensive study of various alcohol-aqueous solutions with a detailed description of
the perturbation in the water-hydrogen network when alcohol is added.
14.2 Bulk Water

We begin a brief overview of infrared water spectroscopy with an introduction to the
infrared spectra of various isotopes of bulk water (Figure 14.1). The most prominent
band in the spectrum of liquid water (H2 O or D2 O) belongs to the strongly inter-
and intramolecularly coupled OH (OD)-stretching modes. The OH-stretching band
is located in the spectrum of liquid H2 O (T = 298 K) at 3408 cm−1 , followed by
the combination band at 2128 cm−1 and the HOH-bending band at 1644 cm−1 .
The broad band in the low-frequency region of the spectrum of liquid H2 O is
assigned to the libration mode at 665 cm−1 . The exchange of hydrogen with heavier
deuterium atoms significantly changes the corresponding bands. OD stretching is
shifted mainly due to the mass effect and is found at 2503 cm−1 , combination band
at 1553 cm−1 , DOD bending at 1210 cm−1 , and vibration at 484 cm−1 . In partially
isotopically exchanged bulk water, the additional band is found at 1456 cm−1 , which
is assigned to DOH bending.
The OH (OD)-stretching mode is the most commonly used indicator for the
analysis of spectra of water. The width of the band reflects the broad distribution
of the different types of hydrogen bonds as well as the intense coupling between
the symmetric- and antisymmetric-stretching vibrations. The decoupling can be
achieved very efficiently by adding a small amount of the isotope D2 O to the bulk
H2 O. If the concentration of the added D2 O is low enough, we can observe in the
infrared spectrum predominantly only the HOD molecules surrounded by H2 O
molecules and thus the fully decoupled OD stretching. The spectrum of a 1.4%
solution of D2 O in H2 O is shown in Figure 14.1.
The changing band shape of the OD stretching, shown on right of Figure 14.1,
clearly demonstrates the effects of coupling on the shape and bandwidth of
OD(H)-stretching band. The stretching band is now highly symmetric with no
shoulders as in the case of corresponding coupled band. Although the halfwidth is
significantly reduced, the frequency of the band maxima remains the same. Even
more dramatic changes are observed when the solution of 1.4% of D2 O in H2 O is
frozen. The blue spectrum represents the spectrum of hexagonal ice. The frequency
of the band maxima is redshifted and the bandwidth is significantly reduced. The
14.2 Bulk Water 439
0.4
0.4
0.3
0.3
a.u.
0.2
a.u
0.2
0.1
0.1
0.0
0.0
4000 3500 3000 2500 2000 1500 1000 500 2800 2600 2400 2200
(a) (b)
Figure 14.1 The infrared spectra of pure D2 O (red line), mixture of 1.4% D2 O in H2 O as
liquid at room temperature (black line) and as ice (blue line). (b) The same spectra but
normalized in the region of OD stretching.
redshifted frequency can be related to the occurrence of enhanced hydrogen bonds

and the reduced halfwidth to the more ordered structure of hexagonal ice with
respect to the structures found in liquid water.
14.2.1 Temperature-Dependent Infrared Spectra of Bulk Water

By changing the temperature, we can effectively disturb the bulk water. Therefore,
the temperature measurements are very suitable for probing the water structure
and dynamics, as well as for testing spectral processing methods. We analyzed the
series of infrared spectra recorded at different temperatures in the temperature range
between 0∘ and 80∘ . The spectra shown in Figure 14.2 were taken from the SI of the
1.2
1.0
0.8
Absorbance
0.6
0.4
0.2
0.0
4000 3500 3000 2500 2000 1500 1000 500 0
Wavenumber (cm–1)
Figure 14.2 Infrared spectra of bulk water as a function of temperature and the spectrum
of ice (blue line). The spectra were taken from Maréchal’s paper. Source: Based on
Maréchal [2].
article published by Maréchal [2]. They represent the entire mid- and far-infrared
range, making them up to date with the most comprehensive infrared spectra avail-
able. The mid-infrared spectra were recorded in ATR mode and then converted to
absorbance by applying the Kramer–Krönig transformation [15–17].
However, broad and intense water bands are not a promising system to detect
small spectral changes due to the increase in temperature. Therefore, we must use
mathematical methods that eliminate the constant part of the spectra or decompose
the raw spectra into components in which the perturbation of the spectral properties
is more pronounced. Most commonly, two different methods are applied very effi-
ciently; the spectral subtraction [18] or the spectral decomposition [2, 12, 19] of the
recorded spectra on orthogonal basis vectors. The main purpose common to both
methods mentioned above is to focus only on the spectral changes induced by the
perturbation (temperature, pressure, solute concentration, etc.).
Among spectra decompositions, one of the most powerful and promising
approaches is vibrational MCR spectroscopy, which belongs to the group of
quantification methods of multivariate analyses [20]. It is generally considered
a supervised approach, but can also be performed in an unsupervised regime if
concentration and spectral constraints are omitted [20]. Essentially, it is difference
spectroscopy, which yields SC spectra that exhibit vibrational properties of two
(or more) different groups of solvent (in the case of bulk water) or the solute itself
and the solvent molecules perturbed by the solute (e.g. the alcohol–water solutions
shown later). Each SC spectral component satisfies the nonnegative minimum
area difference condition between the component spectra that may belong to two
or more different sets (populations) of water molecules or the solution and the
pure (bulk) solvent [21]. This method has one very important advantage over the
subtraction method. It is free from the subjective errors that may occur in the sub-
traction procedure. This was the main reason why we implement it in the analysis
of water spectra. The application of the MCR spectral decomposition program was
performed by the Matlab software MCR-ALS GUI v4c [22, 23] developed by the
authors from Umea University, Sweden. Two calculated components, which can
explain more than 99% of the total variance of bulk water spectra measured at
different temperatures, resulting from the application of MCR analysis to the series
of temperature-dependent infrared spectra are shown in Figure 14.3. With these
two components, we are able to construct all temperature-dependent spectra of
bulk water with minimized differences between the calculated and experimental
ones. However, the question arises as to the origin of these two components.
There is a long and heated debate about whether water is homogeneous (con-
tinuous distribution of hydrogen bonds) [24–27] or inhomogeneous (or two-state
model) [28–33] liquid. The purpose of this chapter is not to become a part of that
discussion, so both principles are only briefly mentioned to understand the mean-
ing of two orthogonal spectral components. Moreover, we are mainly interested in
the microscopic picture of water at physiological temperatures and pressures, where
it seems that both theories culminate [2, 34, 35]. The first hypothesis asserts that the
14.2 Bulk Water 441
0.9
0.8
0.7
0.6
Absorbance
0.5
0.4
0.3
0.2
0.1
0.0
4000 3500 3000 2500 2000 1500 1000 500 0
Wavenumber (cm–1)
Figure 14.3 The spectral components computed by the multivariant curve resolution
program MCR-ALS.
hydrogen bonds between water molecules are continuously distributed with respect
to their strength. Water is thus a homogeneous liquid with a wide variety of hydrogen
bonds in terms of strength and angle. In contrast, the second hypothesis is based on
the main experimental results of IR and Raman spectroscopy, where several isos-
bestic points were found in the spectra of water as a function of temperature [2, 30].
These points can be explained by the presence of two different clusters of water,
differing in density, geometry, and hydrogen bonding strength. The cluster with the
lower density should have water molecules arranged in higher symmetry, as found in
hexagonal ice, and bound with stronger hydrogen bonds. The second cluster should
have higher density, lower symmetry, and weaker hydrogen bonds.
At first glance, the components determined by the MCR analysis (Figure 14.3) can
be assigned to these two clusters. Both components differ substantially in the fre-
quencies and band halfwidths of the stretching, bending, and libration modes. Thus,
the black spectrum can be assigned to the clusters with low density and high symme-
try with stronger hydrogen bonds due to its similarity with the spectrum of hexag-
onal ice, while the red component then obviously corresponds to the cluster with
lower symmetry, weaker hydrogen bonds, and higher density. But as we have already
mentioned, these two approaches are not mutually exclusive, at least at ambient
physiological temperatures [2]. The stretching bands of both components are broad,
unlike those found in the gaseous state. This shows the continuous distribution of
frequencies that implies a continuous distribution of hydrogen bonding strength
between the different water molecules. The existence of two different classes of
water molecules or clusters is, therefore, consistent with a continuous distribution
of hydrogen bonds, since each type of water molecule (cluster) entails a continuous
and wide distribution of hydrogen bonds with respect to energy and geometry.
14.3 Water Near Fully Hydrophobic Solutes

Similar to various hypotheses about bulk water structure, the origin of the effects
of fully hydrophobic solutes on the surrounding water structure is far from clearly
understood [36–40]. This is due to subtle and complex changes in the surrounding
aqueous medium caused by the presence of hydrophobic solutes, which are difficult
to access using experimental techniques. However, experiments have shown that
small, purely hydrophobic solutes (alkanes and noble gases) in water increase the
order [41, 42] and restrict the mobility [43] of neighboring water molecules. Two
opposing views are able to explain these data. The first, so-called “classical view,”
suggests that a solute alters water structure by forming the transient semi-ordered
clathrate-like clusters (“icebergs”) around it with enhanced hydrogen bonding of
water [41, 42]. The “classical view” explains positive ΔG and ΔCp and negative
ΔS changes in hydrophobic hydration as well as restricted water mobility [43].
However, diffraction studies [44–46] and EXAFS experiments [47] show that water
molecules in the vicinity of hydrophobic solutes do not differ significantly from
bulk water molecules. These observations form the basis for the so-called “dynamic
view,” which explains hydrophobic hydration by slowing down the dynamics of
water molecules in the vicinity of hydrophobic solutes. The origin of the reduced
dynamics is the hindrance of the jump mechanism of rotational relaxation without
significant change of the water-hydrogen-bond energy [39, 48]. The main problem
with both views has been the lack of accurate experimental data to confirm or
refute the strengthening of the water hydrogen bonds in the vicinity of hydrophobic
solutes. Besides two already presented, some alternative views have also appeared.
Baldwin [40] proposed that the arrangement of water molecules near hydrophobic
solutes is caused by van der Waals forces between hydrophobic carbon and water
oxygen, while Lee [49, 50] stated that hydrophobic hydration can be explained by
the difficulty of the hydrophobic solute to accommodate itself in a cavity in water
due to the small size of the water molecule. Filling the gap of missing experimental
evidence that unequivocally confirms or refutes the classical view has been a
challenging task, and we showed that this could be done by studying the infrared
spectra of solutions of methane, ethane, krypton, and xenon in water [51]. We
took full advantage of IR spectroscopy to show whether the hydrogen bonds of
water are strengthened near the fully hydrophobic solute or whether these water
molecules are similar to bulk molecules, at least in terms of hydrogen-bond energy.
Two major obstacles had to be overcome: the expected low concentration of fully
hydrophobic solutes and the high absorption index of the water itself. These are
two mutually exclusive conditions that must be taken into account. Therefore, we
constructed the high-pressure experiment shown in Figure 14.4 and used 1.4% D2 O
solution in H2 O instead of isotope-free water. The application of pressure increases
the concentration of the gases dissolved in water, and the low concentration of the
deuterium isotope in the water significantly reduces the intensity of the stretching
band. Namely, we analyzed the OD (H)-stretching band, which is completely inter-
and intra-molecularly decoupled due to the low concentration of D2 O (only HDO
species are present) and has a low band intensity far from the saturation level of the
Gas cylinder
Needle valves (inlet)
Manometer Water pump
IR beam probing Gas-water interface

gas-water interface
IR beam probing
liquid water
Teflon spacer
Thermocouple
HPL-TC-13-3
Needle valve (outlet)
Figure 14.4 Schematic presentation of transmission high-pressure cell. Source: Based on

Grdadolnik et al. [51].
detector. This allows studies of band shape and band frequency as a function of the
presence of poorly soluble apolar substances, such as methane, ethane, krypton,
and xenon, in water.
14.3.1 Verification of the Experimental Procedure

We expected small amount of perturbed water molecules with corresponding small
impact on the OD-stretching band shape and frequency. We applied difference
spectroscopy, which allows stepwise elimination of spectral features from the
original [51]. The whole subtraction procedure is shown in Figure 14.5, where we
trace the discrete steps on the well-known spectrum of the NaCl solution in water.
A similar subtraction procedure to determine the influence of ions on the water–
water-hydrogen-bond network was developed by Lindgren and coworkers [52, 53].
To avoid misinterpretations and methodological errors, a series of tests were
carried out to verify whether the method of spectra measurement by application of
the high-pressure cell and the subtraction method used here are suitable for mea-
suring the effect of apolar solutes on neighboring water molecules. There are two
critical factors that could prevent obtaining correct results for apolar solutes. First,
the solubilities of apolar solutes in water are much smaller than those of NaCl ions
presented in Figure 14.5. Second, the pressure dependence of the thickness of the
2.8% HDO in H2O 2.8% HDO in H2O

Absorbance and diluted solute
Absorbance
H2 O H2O
2800 2600 2400 2200 2000 2800 2600 2400 2200 2000
(a) (b)
Subtractoin of H2O spectrum

1.0 1.0
2505 cm–1 2510 cm–1
0.8 0.8 HDO and diluted
Absorbance
Absorbance
HDO solute
0.6 0.6
0.4 0.4
0.2 0.2
2800 2600 2400 2200 2000 2800 2600 2400 2200 2000
(c) (d)
Subtractoin of HDO spectrum
D- p*C
0.05
2538 cm–1
0.04
Absorbance
HDO perturbed
0.03 by solute
0.02
0.01
0.00
2800 2600 2400 2200 2000

Wavenumber (cm–1)
(e)
Figure 14.5 Schematic representation of the subtraction procedure: 0.5 M solution of

NaCl; (a) Spectra of pure H2 O and diluted D2 O in H2 O; (b) Spectra of pure H2 O and diluted
D2 O and solute in H2 O; (c) Spectrum of HDO; (d) Spectrum of HDO and diluted solute; (e)
Difference spectrum as a result of subtraction between the HDO spectra presented in (d)
and (c) graphs. Source: Based on Grdadolnik et al. [51].
sample in the high-pressure cell may prevent accurate performance of the subtrac-
tion procedure. Therefore, we needed to test the applicability of the high-pressure
cell probe at expected low solute concentrations and investigate the effects of pres-
sure on the water solution. The effect of Na+ and Cl− ions on neighboring water
molecules is well-known and documented by Stangret et al. [53]. The Na+ and Cl−
ions cause a shift in the OD-stretching mode of HDO molecules perturbed by solute
to higher frequencies, suggesting that the water-to-water-hydrogen bonds near Na+
and Cl− ions are weaker than those in bulk water.
NaCl was added to 0.5 M in the solvent, which consisted of 2.8% HDO in H2 O.
The OD-stretching regions of the pure solvent and NaCl solution are shown in
Figures 14.5 A, B. First, the spectrum of the pure H2 O is subtracted from these two
spectra to eliminate the combination band at 2128 cm−1 that partially obstructs
the OD-stretching band. All spectra were recorded at identical temperatures and
pressures (300 K and 1 bar). The resulting spectra are shown in the second row
in Figures 14.5 C, D. The spectrum of pure HDO is then subtracted from the
water-solute spectrum. Since only a small number of water molecules are affected
by the ions, the other (unaffected) water molecules are removed by subtraction. The
resulting spectrum (Figure 14.5 E) is that of the water molecules perturbed by the
solute. The factor p represents the subtraction parameter.
Determination of the subtraction parameter p, which is generally not equal to 1,
is required for both subtractions. If the concentrations, optical path, and absorp-
tivity are known, the subtraction factors can be calculated using the Beer–Lambert
law. However, these quantities are very difficult to obtain. To attain the subtrac-
tion parameter, a procedure was used in which the value of the subtraction factor
is gradually increased until the first signs of the appearance of negative bands in
the difference spectrum, i.e. a flat line. The procedure determines the subtraction
factor in an unambiguous and reproducible manner and all differences in the cor-
responding bands become clearly visible. The remaining band is that of the water
molecules perturbed by the solute. Since the first-solvation layer may contain water
molecules that are not different from the bulk, the area of the residual band in the
difference spectrum does not correlate with the total number of water molecules in
the first-solvation shell. Rather, it correlates with those that are different from the
bulk, i.e. those that are affected by the solutes.
Following this procedure, the spectrum of the OD-stretching mode of the per-
turbed water molecules in the hydration shell of Na+ and Cl− ions was extracted.
The peak is symmetric and blueshifted to 2538 cm−1 , in close agreement with the
frequencies reported in the literature. Stangret and Gampe [53] have shown that
the OD bands for Na+ and Cl− ions are at maxima of 2534 ± 8 and 2530 ± 8 cm−1 ,
respectively.
The spectra shown in Figure 14.5 were recorded under atmospheric pressure and
with high solute concentrations (0.5 M). Since the same approach should be used for
apolar solutes at challenging conditions (high pressures and low concentrations), we
verified the experimental method and the subtraction procedure using the solutions
of NaCl under such conditions. As mentioned above, the peak maximum should be
observed at 2538 cm−1 . The spectrum of NaCl was recorded at low concentration
(0.03 M) and pressure of 53 bar, similar to the ones measured at the most extreme
conditions applied for fully hydrophobic gases. Since both ions contribute to the
Figure 14.6 OD stretching of

2536 cm–1
Δv subtracted spectrum of 0.03 M NaCl
OD solution in HDO/H2 O (bold line) and
corresponding OD stretching of bulk
Absorbance
HDO/H2 O. The low S/N ratio on the

right side of the OD band is due to
nonideal subtraction of atmospheric
CO2 bands. Source: Based on
2700 2600 2500 2400 2300 2200

Wavenumber (cm–1)
band at 2536 cm−1 , the corresponding concentration is 0.06 M. The subtraction pro-
cedure described above was used. The resulting spectrum is shown in Figure 14.6.
The reduced signal-to-noise ratio and the traces of CO2 bands on low-frequency side
are the consequences of using a very low solute concentration. However, the fre-
quency and shape of the OD-stretching mode are very similar to those obtained at
higher concentrations of Na+ and Cl− and at ambient pressure (Figure 14.5 E). Based
on this validation, we conclude that the applied experimental method provides reli-
able differences in the OD-stretching mode, which are used to account for the effects
of the perturbation by dissolved apolar solutes.
Spectra obtained with the high-pressure transmission cell may be further distorted
by interference from multiple reflections at the windows of the thin sample cell.
To test whether this and other effects could cause artifacts in the spectrum of the
OD-stretching band under challenging conditions (high pressure and low concen-
tration), we compared the transmission spectra of methanol solutions in the mixture
of 2.8% HDO in H2 O at 56 bar with the corresponding ATR spectra at 1 bar. Three
concentrations of methanol were used to record the transmittance and ATR spectra:
0.08, 0.15, and 0.2 m. The ATR spectra were recorded using the horizontal multi-
ple reflection ATR cell (Graseby, Specac) equipped with a ZnSe reflection crystal.
The ATR spectra of methanol solutions, 2.8% HDO in H2 O and H2 O were recorded
at 298 K and 1 bar. The transmission spectra of methanol solutions, 2.8% HDO in
H2 O and H2 O were recorded at 298 K and 56 bar. We used the same solvent subtrac-
tion protocol as for NaCl solutions in water (Figure 14.5). Figure 14.7 shows that the
difference spectra obtained with the transmission cell and ATR are very similar, con-
firming that there are no artifacts in the transmission spectra of the OD-stretching
bands. Small differences between the transmission and ATR spectra are attributed
to band distortions due to the influence of reflection and the lower sensitivity of the
horizontal ATR cell.
14.3.1.1 Effects of Temperature and Pressure on the OD-Stretching Band

Errors in the measurement of temperature and pressure in the high-pressure trans-
mission cell could produce artifacts in the difference spectra of the OD-stretching
band. To determine how sensitive the results are to changes in temperature and
pressure, we measured the difference spectra of pure solvent (without dissolved
gasses) at different temperatures and pressures. Figure 14.8 A shows that a decrease
2521 cm–1
2518 cm–1
0. 002 a.u.
0. 005 a.u.
Absorbance
Absorbance
2592 cm–1 2428 cm–1 2596 cm–1 2423 cm–1

2700 2600 2500 2400 2300 2700 2600 2500 2400 2300
(a) (b)
Figure 14.7 Difference spectra of methanol dissolved in 2.8 % HDO in H2 O after

subtraction of H2 O and HDO recorded in the transmission (a) and ATR (b) mode (bottom
spectra: 0.08 m, middle spectra: 0.15 m, and upper spectra: 0.2 m). Source: Based on
1.0 1.0
0.8 0.8
Absorbance
0.6 0.6
1.05 1.05
0.4 1.00 0.4 1.00
Absorbance
Absorbance
0.95
0.95
0.90
0.2 0.85
Δv 0.2 0.90
0.80 0.85
0.75 0.80
0.0 2550 2525 2500
Wavenumber (cm–1)
2475 2450 0.0 2550 2525 2500
Wavenumber (cm–1)
2475 2450
2700 2600 2500 2400 2300 2700 2600 2500 2400 2300
(a)
(b)
0.03
0.006
0.02
0.01 0.004
Absorbance
0.00 0.002
–0.01 0.000
–0.02
–0.002
–0.03
–0.004
–0.04
2700 2600 2400 2500 2300 2700 2600 2500 2400 2300
(c) (d)
Figure 14.8 Infrared spectra of OD-stretching band νOD of the mixture of 2.8 % HDO in
H2 O as a function of temperature and pressure. (a) νOD at temperatures 285K and 279K.
Pressure is 56 bar. Spectra are scaled to the same height. (b) νOD at pressures 56.2 bar, 51.6
bar, 45.4 bar, 40.4 bar, 36.7 bar, and 25.4 bar. Temperature is 285 K. Spectra are scaled to
the same height. (c) The difference spectra at the pressure 56 bar from less to more intense
in OD-stretching range: (284 K) - (285 K), (283 K) - (285 K), (281 K) - (285 K), and (280 K) -
(285 K), and (279 K) - (285K). (d) The difference spectra at the temperature 285 K from less
to more intense: (56.2 bar) - (51.6 bar), (56.2 bar) - (45.4 bar), (56.2 bar) - (40.4 bar), (56.2
bar) - (36.7 bar), and (56.2 bar) - (25.4 bar). Source: Based on Grdadolnik et al. [51].
in temperature causes a large redshift in the OD-stretching band; however, its

halfwidth remains essentially the same. The effect is highlighted in the difference
spectra shown in Figure 14.8 C. The redshift of the frequency maximum of the band
is accompanied by a positive peak in the low-frequency region and a negative peak
in the high-frequency region. The influence of pressure on the OD band is much
smaller than the influence of temperature. Figure 14.8 B shows the normalized
spectra recorded at pressures between 56 and 26 bar. The effect of increasing
pressure is a slight broadening of the OD band, but the frequency of the band
maximum remains essentially the same. This is evident in the difference spectra
by the two broad positive wings and the near-zero intensity at the position of the
OD band maximum (Figure 14.8 D). The intensities of the peaks in the pressure
difference spectra are almost 10 times smaller than in the temperature difference
spectra.
14.3.1.2 Clathrate Formations

To test the effectiveness of the solvation of gases in water solution, we recorded a
series of clathrates by the gases used in the experiment. Apolar gases form clathrates
under appropriate conditions [54]. Clathrate formation can dissolve a large amount
of gas in water. The clathrates and hexagonal ice were prepared from a mixture
of 2.8% HDO in H2 O. These spectra were recorded to determine if the spectra of
aqueous solutions of apolar gases and those of clathrates could be distinguished
(see below). The effectiveness of dissolving the various gases in water in the
high-pressure-sampling cell was determined. It should be emphasized here that no
bulk water is subtracted from the experimental clathrate spectra. In other words, this
means that all water molecules are involved in the clathrate structure (Figure 14.9).
The spectra of the clathrates were measured using the same experimental setup
as presented in Figure 14.4. However, the spectra were recorded at temperatures
and pressures indicated on the left side of the equilibrium line in the PT phase
diagram [55, 56], i.e. under conditions where clathrates are stable. The cell was
filled with water using a water pump. The gas pressure in the sample cell was then
increased to the preselected value. The temperature in the sample cell was close to
300 K and well above that of the phase transition (i.e. on the right side of the phase
transition line). The temperature was then slowly lowered, moving along the isobaric
line toward the phase transition. The spectra of saturated gas solutions can be mea-
sured at any desired temperature after equilibration for several hours. Constant pres-
sure was maintained in the sample cell by the needle valve and the pressure-reducing
valve connected to the gas cylinder. The pressure remains nearly constant until the
temperature of the phase transition, i.e. the solution–clathrate phase line, is reached,
then it drops abruptly. Crossing this line requires a large amount of the gases due to
the formation of clathrate. A sudden drop in gas pressure can, therefore, be used as
a good indication of the onset of clathrate formation. Spectra were recorded after
at least six hours of equilibration. Spectra of clathrates are presented as they were
recorded since it is not necessary to subtract the spectra of bulk water.
Absorbance
Absorbance
2600 2500 2400 2300 2600 2500 2400 2300
(a) (b)
Absorbance
Absorbance
2600 2500 2400 2300 2600 2500 2400 2300

(c) (d)
Figure 14.9 Infrared spectra of the OD-stretching mode of clathrates. (a) Methane
clathrate. (b) Ethane clathrate. (c) Xenon clathrate. (d) Krypton clathrate. The spectrum of
HDO ice (dashed, at 271 K, 1 bar) is shown for comparison. The OD-stretching bands are
scaled to equal heights. Source: Based on Grdadolnik et al. [51].
14.3.2 Pure Hydrophobic Solutes in Water Solution

The presented tests show that the applied setup would give us reliable and accurate
experimental data. Therefore, we applied it to the solution of apolar solutes in
water. The spectra of methane, ethane, xenon, and krypton were recorded using the
same experimental setup as for the recording of clathrates with only one exception.
Considering the pressure and temperature of the recorded solution, we have to stay
strictly on the right side of the equilibrium line in the PT diagram for the respective
gas. The analysis of the OD-stretching bands was performed in two steps. In the first
analysis, we determine the halfwidths of the OD band of temperature-dependent
solution spectra. We expected broader OD stretch in solution spectra as well as
different temperature behavior of halfwidths with respect to bulk water. The
halfwidth response to temperature is presented in Figure 14.10 for all solutes used
in the experiment.
Even the first analysis of solutions shows that pure hydrophobic solutes have
impact on the halfwidths of the OD-stretching band. It is small but consistent in the
whole interval of temperature measurements. Moreover, at higher temperatures,
176
176 Bulk HDO Bulk HDO
172 Methane 172 Ethane
168
168
Hw (cm–1)
164
164
160
160
156
156
152 152
270 280 290 300 310 320 330 270 280 290 300 310 320 330
176 176
Bulk HDO Bulk HDO
172 Xenon 172 Krypton
168 168
Hw (cm–1)
164 164
160 160
156
156
152 152
270 280 290 300 310 270 280 290 300
Temperature (K) Temperature (K)
Figure 14.10 The halfwidths of OD stretching of methane, ethane, xenon, and krypton as
a function of temperature. The spectra of corresponding bulk water were recorded at the
same pressures as solutions, i.e. methane 56 bar, ethane 14 bar, xenon 50 bar, and krypton
16 bar, respectively.
the difference of the halfwidths between solutions and bulk water recorded at the
same pressures becomes substantially smaller, and in the case of ethane, xenon
and krypton almost disappears. Nevertheless, the analysis of the bandwidths of the
unprocessed spectra is subject to a rather large error due to the fluctuation of the
background. The origin of the broadening of the OD-stretching band clearly shows
the subtraction of bulk water according to the procedure presented in Figure 14.5.
The appearance of a small redshifted band can be observed in all the studied
solutions (Figure 14.11). The redshifts and even the halfwidths of these bands can
be compared to the bands of HDO ice or corresponding gas clathrates (Figure 14.12,
Table 14.1).
The number of disturbed water molecules was calculated only for the methane
solution, the only one with known concentration of the solute [51]. It was estimated
by comparing the areas of perturbed OD stretching and OD stretching of bulk water
and/or ice. The number of strengthened hydrogen bonds is between 10 and 15, which
corresponds to 5–8 hypothetical frozen water molecules if we consider the similar-
ity of the corresponding stretching bands. This amount is only two or three times
lower than the number of waters in the first-solvation layer of methane measured
by neutron scattering [46] but in very good agreement with the number estimated
by Kauzmann [42] who predicted that highly crystalline iceberg contains less than
half a dozen water molecules.
Methane
Xenon
Absorbance
Ethane
Krypton
HDO
2700 2600 2500 2400 2300

Wavenumber (cm–1)
Figure 14.11 The resulting spectra after subtraction procedure applied to the solution
spectra of methane, ethane, krypton, and xenon. The last spectrum (HDO) shows the
reproducibility of the measurements. The decrease in the S/N ratio below 2400 cm−1 is due
to the nonideal removal of atmospheric CO2 bands [51]. The recording parameters are
presented in Table 14.1. Source: Grdadolnik et al. [51]/with permission of National Academy
of Sciences.
ΔvOD
Methane
in water
HDO liquid
Absorbance
Methane
clathrate
HDO ice
2700 2600 2500 2400 2300

Wavenumber (cm–1)
Figure 14.12 The normalized spectra of the decoupled OD-stretching bands of liquid
water (bold line), methane in water (bold line, peak frequency is redshifted), methane
clathrate (dotted line), and HDO ice (dashed line). Source: Grdadolnik et al. [51]/with
permission of National Academy of Sciences.
Table 14.1 Summarized recording and spectral parameters of the OD-stretching band of
liquid solvent, solutions of gases, NaCl solution, clathrates, and ice.
T (K) P (bar) 𝝂 OD (cm−1 ) 𝚫𝝂 OD (cm−1 ) H w (cm−1 )
Pure solventa) 293 1 2502 (±1) 0 164 (±2)

Pure solventa) 293 56 2502 (±1) 0 (±2) 160 (±2)
Pure solventa) 285 1 2498 (±1) 4 (±2) 158 (±2)
Pure solventa) 285 56 2498 (±1) 4 (±2) 158 (±2)
CH4 solutionb) 285 56 2445 (±5) 57 (±8) 92 (±8)
C2 H5 solutionb) 291 14 2430 (±5) 78 (±8) 86 (±8)
Xe solutionb) 299 16 2442 (±5) 60 (±8) 96 (±8)
Kr solutionb) 289 50 2445 (±5) 57 (±8) 87 (±8)
NaCl solutionc) 285 53 2536 (±5) −34 (±8) 166 (±8)
CH4 clathrated) 275 40 2446 (±1) 56 (±2) 60 (±2)
C2 H5 clathrated) 276 7 2442 (±1)e) 60 (±2) 80 (±2)
Xe clathrated) 292 18 2442 (±1) 60 (±2) 79 (±2)
Kr clathrated) 274 56 2446 (±1) 56 (±2) 55 (±2)
Ice Ih 271 1 2445 (±1) 57 (±2) 37 (±2)
Temperature T, pressure P, frequency at maximum intensity 𝜈 OD , frequency redshift Δ𝜈 OD , and

halfwidth H w are shown. H w is defined as the width of a peak at half height. Error in measuring T
and P are estimated at 0.5 K and 0.5 bar. All samples contain a solvent mixture of 2.8% HDO in H2 O.
a) Solvent mixture (2.8% HDO in H2 O) under four different conditions.
b) Spectral parameters of the water molecules perturbed by solutes.
c) Solution of 0.03 M NaCl used for verification of the subtraction procedure.
d) Solid state.
e) Shoulder at 2472 (±1) cm−1 .
Source: Grdadolnik et al. [51]/with permission of National Academy of Sciences.
Since we have removed the spectral features of bulk water, the redshifted
bands that are shown in Figure 14.11 clearly show the occurrence of enhanced
water–water-hydrogen bonds induced by the presence of pure hydrophobic solutes.
The reduced halfwidths of the subtracted OD-stretching bands with respect to the
stretching band of bulk water (Figure 14.12) also indicate that these water molecules
with enhanced hydrogen bonds are less disordered than in bulk water. The com-
parisons of the redshifts and halfwidths of all the systems studied are summarized
in Table 14.1. It is obvious that the OD-stretching band of the subtracted spectra,
shown in Figure 14.11, has more similarities with the OD-stretching bands of ice
and clathrates and thus might also have some similarities in the hydrogen bond
strengths and the general structural ordering of the water molecules. In general,
the spectral bandwidth of OH stretching is used to assess structural order [8, 57];
a decrease of halfwidth indicates an increase in structural order. The halfwidths
of OD of water molecules perturbed by hydrophobic solutes are only slightly
larger than those of clathrates and much smaller than those of liquid solvent
(Figure 14.12, Table 14.1). It is also evident that temperature and pressure do not
affect halfwidths significantly (Table 14.1). The similarity of halfwidth values of

water molecules perturbed by hydrophobic solutes and that of clathrates indicates
a relatively small range of hydrogen bond strengths and thus a similar structural
ordering of water molecules near the strengthened hydrogen bonds. The halfwidths
reflect the structural ordering of only those water molecules that are close enough
to affect the values of OD-stretching frequency. Auer et al. [8] found that OD of a
hydrogen-bonded water dimer is predominantly influenced by the water molecules
closer than 4 Å from the labeled hydrogen-bonded atom. This local environment
of an ice-like hydrogen bond consists of a water dimer and ≈6 nearest-neighbor
hydrogen-bonded water molecules arranged in two fused tetrahedrons. Note that
the formation of an ordered local environment around individual ice-like hydrogen
bonds is quite different from the formation of fully formed clathrate-like clusters
of water molecules since the number of ice-like hydrogen bonds per individual
methane molecule is relatively small.
14.3.3 MD Simulations of Purely Hydrophobic Solute in Water and the

Origin of Strengthened Water–Water–Hydrogen Bonds Near Methane
Molecule
Monte Carlo and MD simulations at a variety of theoretical levels have been used
extensively to rationalize the effect of purely hydrophobic solutes on water struc-
ture and dynamics [58–65]. However, the results have been inconsistent. We have
performed ab initio MD simulations of methane [51] in D2 O at three temperatures,
283, 293, and 300 K, using the dispersion-corrected DFT approach [51, 66]. The most
suitable exchange-correlation functional was found to be revPBE [67, 68] in combi-
nation with the Grimme D3 correction [69], as shown in a series of test simulations
of bulk water according to the ability to best reproduce experimental radial distri-
bution functions, diffusion constants, line shape of OD and solvation enthalpy of
liquid water at ambient conditions [51]. Time-dependent OD-stretching frequen-
cies of individual water molecules were derived using the wavelet transform of the
corresponding OD distance time series.
The analysis of the MD was focused on comparing the structural and energetic
properties of the hydrogen bonds of water molecules in close proximity to methane
and relative to those of bulk water. The details of evaluation of MD results can be
found elsewhere [51]. The results from MD simulations clearly show that the water
molecules near methane have stronger, more numerous, and more tetrahedrally ori-
ented hydrogen bonds than those in bulk water and that their mobility is restricted,
which is consistent with the experimental results presented above and with the NMR
data of Haselmeier [43].
The presented experimental results reproduced by the ab initio MD simulations
show that the water hydrogen bonds strength in the neighborhood of purely
hydrophobic is comparable to ones in ice or clathrates. The formation of ice-like
hydrogen bonds causes a stronger ordering of water by the formation of “icebergs”
as postulated by Kauzmann [42]. Determining the physical origin of ice-like
hydrogen bonds in the vicinity of hydrophobic solutes is, therefore, critical to
understanding hydrophobicity at the most fundamental level. Theoretical sim-

ulations show that water spans hydrogen bonds near small hydrophobic solutes
similar to hydrogen bonds in clathrates to maximize the number of hydrogen
bonds [62]. Such a constraint on the hydrogen bonding network may per se
order the water molecules; however, it does not necessarily make the hydrogen
bonds stronger. The lack of a clear relationship between steric constraint and
hydrogen bond strength led us to focus on electrostatic interactions since hydrogen
bonds in water are predominantly electrostatic in nature [70]. Hydrogen bonds in
water are known to be strongly cooperative [8, 71–73]; therefore, their strength
depends on electrostatic interactions of hydrogen-bond dimers with neighboring
water molecules.
The analyze the effects of neighboring water molecules on the strength of the
hydrogen bond between a water dimer during ab initio MD simulations [51] shows
that three different classes of water molecules can be identified in the neighborhood
of a water-hydrogen-bond dimer in liquid water. The first two classes include
the water molecules bound to the hydrogen-bonded dimer; their contributions
to the value of electric field along the hydrogen bond are not significantly dif-
ferent from the contributions of the corresponding waters in ice. The first class
includes the hydrogen-bond-enhancing water molecules that increase the value
of electric field along hydrogen-bond direction. The second class includes the
hydrogen-bond-weakening waters that decrease the value of electric field along
hydrogen-bond direction. The average numbers of these two classes per hydrogen
bond are 3.4 and 1.7, respectively; thus, the net effect is a strengthening of hydrogen
bond. The corresponding values in ice are 4 and 2 [51].
Most interesting is the third class of intercalating water molecules. These water
molecules are able to temporarily occupy the space closest to atom H of the
hydrogen-bonded dimer, which is empty in ice. The intercalating waters represent
a subset of the so-called fifth, interstitial, non-tetrahedral, or mismatched water
molecules [58, 74, 75] that interact with the labeled water molecule predominantly
through van der Waals forces. These water molecules are specific to liquid water
and are not present in ice. The contributions of the intercalating water molecules
to this electric field are generally negative and primarily responsible for the lower
strength of hydrogen bonds in liquid water than in ice.
When the hydrogen-bonded water dimer is transferred from bulk to the first
solvation layer of methane, the fraction of intercalating water molecules per
hydrogen-bonded dimer is reduced from 6% to 2% because methane and intercalat-
ing water molecules tend to occupy the same space in the side of the direction of the
hydrogen bonds [51]. Methane molecule prefers this position because the hydrogen
bonds of water tend to span small hydrophobic solutes, similar to clathrates.
Removal of the intercalating water molecules increases the value of electric field
along hydrogen-bond direction. A higher value of the electric field means that a
larger force pushes positively charged hydrogen toward the acceptor oxygen, which
strengthens the hydrogen bond. This mechanism explains why the hydrogen bonds
in water near methane and other fully nonpolar solutes are expected to be stronger
than those in bulk water [51].
14.4 IR Spectroscopy of the Water Hydrogen Bonding in the Alcohol–Water Systems 455
14.4 IR Spectroscopy of the Water Hydrogen Bonding

in the Alcohol–Water Systems
14.4.1 Importance of Alcohol–Water Systems

Alcohol–water mixtures are considered to be the simplest aqueous amphiphilic solu-
tions [76]. They serve as very illustrative and useful prototypical systems to gain
deeper insight into hydrophilic and hydrophobic hydration and the delicate compe-
tition between intermolecular and intramolecular hydrogen bonding. Namely, their
functional properties can be tuned through changes in the hydrophobic alkyl tail
of an alcohol molecule [77] and the number of hydroxyl groups [78]. Hydrogen
bonding is extensively analyzed as one of the fundamental secondary interactions
in chemistry, biology, and materials science, with water taking the versatile role as a
solvent and/or biological medium governing a variety of biochemical processes and
biomolecular interactions [38, 79]. The study and deciphering of specific features of
aqueous solutions of alcohols provide an indispensable basis for a better understand-
ing of the distortion of the hydrogen-bonding network of water molecules when it
is perturbed by more complex biomolecule solutes (mostly amphiphilic in charac-
ter). Alcohol solution systems are also of research interest because of their important
role in fuel cells, as industrial solvents for separation processes [76, 80, 81], in food
science and technology [82, 83], and as cryoprotectants in technology and medicine
[84–86].
The perturbation of the hydration shell structure is the result of a combined influ-
ence of the sizes and relative positions of the hydrophobic and hydrophilic parts of
the solute [87, 88]. On the other hand, the dynamics of water molecules within the
hydration shell of an amphiphile, such as a biomolecule, play an important role in
biological binding, folding, and self-assembly with respect to its functions [89–91].
The hydrogen-bonding structure of water at hydrophilic sites of the amphiphilic
solute is mainly dictated by the spatial arrangement of hydrogen-bond donors,
acceptors, and possible polar surface species [92]. In contrast, the elucidation of
hydrogen-bond structuring with respect to the effect of hydrophobic moieties is
much more challenging [93].
As a continuation of our earlier study [51] on the hydrophobic hydration of small
hydrophobic solutes, also summarized in the previous section, we pursue the inves-
tigation of the structural changes of water near the hydrophobic moieties of alcohols.
At the beginning of the following section, we give a brief literature review of IR spec-
troscopy used in the study of the alcohol–water system. This part is then followed
by our main findings on the alcohol–water mixtures analyzed by ATR FT-IR spec-
troscopy, which are presented in more detail in ref. [94].
14.4.2 IR Spectroscopy in the Study of Alcohol–Water Systems

14.4.2.1 Overview
Invaluable information on hydrogen bonding in aqueous solutions of alcohols can
be provided by vibrational spectroscopy, which encompasses several techniques, of
which IR and Raman spectroscopy are the most important and widely used. Indeed,
there are a large number of publications reporting results of Raman spectroscopy
[9, 21, 81, 95–103] as well as complementary IR spectroscopic studies [78, 81–86,
104–110] on the hydration of alcohols. Since IR spectroscopy is the focus of this
chapter, we first provide a brief literature review of the various IR spectroscopic tech-
niques used for the analysis of alcohol–water systems (see Table 14.2), followed by
a summary of our recent results on this topic.
In our recent study [94], we investigated aqueous solutions of ten aliphatic alco-
hols using FT mid-IR spectroscopy in ATR mode for the range of 4000 to 650 cm−1 .
We analyzed aqueous solutions of seven alcohols with unlimited miscibility with
water, which had different lengths of the alkyl chain and the position and num-
ber of hydroxyl groups, i.e. MeOH – the smallest amphiphilic organic solute [111],
EtOH, EtGOH, 1PrOH, 2PrOH, GlOH, and tertBtOH – the largest molecule in the
series of monohydric alcohols with unlimited miscibility with water [112]. In addi-
tion, three alcohols with four hydrocarbons and limited miscibility were included
in the study, i.e. 1BtOH, 2BtOH, and isoBtOH. The solutions were prepared at dif-
ferent concentrations ranging from 2.5 to 40 vol% (except for 1BtOH and isoBtOH,
where the highest concentration being miscible with water is ∼10 vol%) and stud-
ied at 298 K. Temperature-dependent spectroscopy was performed with the 5 vol%
alcohol–water solutions in the temperature range between 298 and 349 K. Further
technical details on the implemented mid-IR spectroscopic measurements can be
found in ref. [94].
14.4.2.2 Spectral Decomposition and Probes for Characterization

Post-processing of our ATR FT-IR spectroscopic data on the solutions of different
concentrations at 298 K was performed in the form of MCR spectral decomposition
using the Matlab software MCR-ALS GUI v4c [22, 23]. The preliminary PCA analysis
showed that the two principal components can explain more than 99% of the total
variance of the solution spectra measured for different alcohol concentrations.
Therefore, using the MCR approach, we decomposed the concentration set of the
solution spectra, {SALC_ci ; for ci = 2.5, 5, 7.5, 10, 15, 20, 25 and 40 vol%}, of each
analyzed alcohol (ALC) into the component related to the solute part providing
the minimum area solute-correlated spectral component, SC ̂ ALC c , weighted by
the coefficient ̂cALC ci , and the component fixed to the bulk water spectrum,
BW ALC_c , weighted by the coefficient ̂ bALC ci . The deviation of the linear combina-
tion of these two spectral components from the real spectra of the solution for a
given alcohol concentration is compensated by the residual ̂rALC ci , i.e.
SALC ci
̂ ALC c + ̂
= ̂cALC ci SC bALC ci BWALC c + ̂rALC ci
for ci = 2.5, 5, 7.5, 10, 15, 20, 25, and 40 vol% (14.1)
The SC spectra sought, representing the vibrational characteristics of the solute itself
and the water molecules perturbed by the solute at a given concentration, were then
constructed as follows,
SCALC ci
̂ ALC c + ̂rALC
= ̂cALC ci SC ci + addALC ci BWALC c
Table 14.2 Brief literature review of IR spectroscopic methods used to study alcohol-water systems and hydrogen bonding.
Source Method of IR spectroscopy / Alcohol used in alcohol-water solution / Type of analysis and characteristics analyzed / Observed results
Ahmed et al., 2012 [104] Method: FT mid-IR in ATR mode; Alcohol: MeOH and EtOH
Analysis and characterization:
● Analysis of spectra of solutions at different concentrations and room temperature via the CO stretching frequency shift and
CO stretching band area

Results:
● Shifts in the CO stretching frequency are related to the changes in the hydrogen-bond associations between alcohol and water
molecules with changing the concentration.

Coldea et al., 2013 [82] Method: FT mid-IR in ATR mode; Alcohol: MeOH and EtOH
● Post-processing of spectra of solutions at different concentrations and room temperature by the PCA analysis in the frequency
range from 1000 to 1170 cm−1 and the PLS regression to predict the content of MeOH based on the absorbance at 1020 cm−1 ;
CO stretching peak absorbance, intensity and area
Results:
● FT-IR supported by the PCA and PLS approaches appears to be a useful method to determine the MeOH and EtOH content
and the ratio between the two to discriminate between the biological and geographical origin of brandies.
Habuka et al., 2020 [85] Method: FT mid-IR in ATR mode; Alcohol: GlOH
● Difference spectra obtained by subtraction of the pure water or pure alcohol spectrum from the spectrum of solution at
different concentrations and room temperature; OH stretching, HOH bending, CH stretching, COH rocking and CO
stretching band shifts and areas
–1
● Three-component Gaussian curve fitting for the 3800–2800 cm region in the ATR spectra
Results:
● Redshifts of CH2 stretching, COH rocking and CO stretching of CHOH at higher concentrations suggest interactions of outer
CH2 OH with inner CHOH groups of surrounding glycerol molecules rather than with water molecules.
● The OH band of glycerol solutions after subtracting pure alcohol could be fitted by three Gaussian components corresponding
to long, medium, and short hydrogen bonds of water molecules (the short ones possibly bound between inner CHOH and
outer CH2 OH groups, and the long ones loosely surrounding outer CH2 OH groups).
Sun and Petersen, 2017 Method: FT mid-IR in ATR mode; Alcohol: tertBtOH
[110] Analysis and characterization:
● Post-processing of spectra of solutions at different concentrations and room temperature by the MCR-ALS decomposition in
the whole frequency range (from cca. 750 to 4000 cm−1 ) to obtain the solute-correlated spectra (corresponds to the solute and
solvent molecules affected by the solute); ration between the OH stretching and CH stretching areas for the solution spectra
and the solute-correlated spectra
Results:
● MCR method is a powerful tool for obtaining the solute-correlated IR spectrum for the analysis of the vibrational modes of
solute and the solvation shell molecules perturbed by the solute.
● Ratio calculations yield 1.4 water molecules in the hydration shell perturbed by the molecule of tertBtOH in ∼9.5 vol%
aqueous solution.
Kwaśniewicz and Method: FT mid-IR in ATR mode, FT near-IR in transmission mode; Alcohol: ten aliphatic 1-alcohols from MeOH to
Czarnecki, 2018 [108] 1-decanol distilled and dried under freshly activated molecular sieves
● Post-processing of alcohol spectra at 303 K by the PCA analysis of mean-centered MIR+NIR spectra of 1-alcohols and
MCR-ALS decomposition (into the two spectral components) of the 1-alcohols spectra from 1-butanol to 1-decanol at 303 K
● Post-processing of spectra by the 2D correlation spectral analysis (algorithm based on Noda’s algorithm) of 1-alcohols from
1-butanol to 1-decanol at 303 K; OH and CH stretching

Results:
● Spectra of methanol, ethanol, and 1-propanol (unlimited miscibility with water) are noticeably different from the spectra of
higher 1-alcohols (limited miscibility with water); similarity between the spectra increases with an increase in the alkyl chain
length.
● One of the two spectral components from the MCR-ALS decomposition includes contributions from the associated OH and
from methyl groups, while the other mainly the contribution of the methylene units and free OH.
● Spectral intensity changes of the OH group with the length of an alkyl chain are nonlinear (which is related to the hydrogen
bonding changes).
● Hydrogen bonding in neat 1-alcohols weakens with the chain length increase; the population of the bonded OH decreases,
while that of the free OH increases.

● Results suggest that for 1-butanol and higher 1-alcohols, the hydrophobic interactions dominate.
Shimoaka and Hasegawa, Method: FT mid-IR in ATR mode, Time-resolved IR in ATR mode; Alcohol: EtGOH
2016 [86] Analysis and characterization:
● Post-processing of spectra of solutions at different concentrations and room temperature by the PCA analysis and MCR-ALS
decomposition in the range from 1150 to 4000 cm−1 ; HOH bending peak shift; CH stretching shift
Results:
● Indication of three component composition of an EtGOH-aqueous solutions involving a water-EtGOH complex, as well as
bulk water and bulk EtGOH component.
● The water/EtGOH complex component reveals that the EtGOH molecules keep gauche conformation, and the terminal
hydroxyl groups are hydrogen bonded by water molecules; an EtGOH molecule in the complex is hydrated by about four
water molecules; the strong hydrogen bonds are assigned to this intermolecular interaction.
● EtGOH molecule in the complex seems to have more hydrogen bonds than that in bulk alcohol, and it is suggested that no
EtGOH molecules should have all the lone pairs on oxygen hydrogen-bond-free in bulk alcohol.
Hu et al., 2010 [83] Method: FT mid-IR in transmission mode; Alcohol: EtOH
● Post-processing of spectra of solutions at a given concentration and room temperature by the MCR-ALS decomposition in the
range from 2000 to 4000 cm−1 into the four spectral components, and the analysis of the mole ratio of water-rich ethanol
hydrate to water (introduction to the structurability parameter)
Results:
● Prevalence of a particular water-rich hydrate implying a cage-like morphology, whereas it is argued the hydrate structure and
its content are related to the perception of vodka taste.
Shao et al., 2019 [109] Method: FT near-IR in transmission mode; Alcohol: EtOH
● Analysis of the fourth derivative spectra at different concentrations and 303 K; consideration of the OH and CH overtone and
combination bands
Results:
● High order derivative enhances the resolution to reveal the detail information in the NIR spectra.
● New peaks related to the hydrogen bonding and the interactions are observed in the derivative spectra of the mixtures
compared to pure alcohol and water.

● In the water–rich mixture, EtOH prefers to interact with water through hydrogen bonding, whereas with the increase of
EtOH content, alcohol tends to assemble due to the hydrophobic effect.

● Water destroys the network of EtOH, i.e., the dissociation of EtOH clusters occurs through breaking the large clusters into
small ones or free molecules.

● Effect of the interaction between the hydrocarbon chain and water on the spectrum is very weak.
(continued)
Chen et al., 2013 [84] Method: FT near-IR in transmission mode; Alcohol: EtGOH
● Post-processing of spectra of solutions at different concentrations and temperatures by the generalized 2D correlation spectral
analysis (algorithm based on Noda’s algorithm), the moving window approach and chemometric methods like PCA and
MCR-ALS analysis; consideration of the OH overtone and combination bands
Results:
● Structures of pure liquid EtGOH and its mixtures with water are determined by the intermolecular hydrogen bonding
through the OH group; intramolecular bonds in the pure EtGOH are not observed.
● In the alcohol-rich solutions the water molecules are bonded with two molecules of alcohol and this cooperative hydrogen
bonding is stronger than that in bulk water.

● Small water clusters surrounding hydroxyl group of alcohol are formed with increasing water content which weakens the
average strength of the water hydrogen bonding.

Dong et al., 2019 [107] Method: FT near-IR in transmission mode; Alcohol: EtOH
● Calculation and analysis of the excess near-IR spectrum in form of excess absorption coefficient (the difference between the
spectrum of a real solution and that of the respective ideal solution under the same conditions) with Gaussian fitting for the
lower concentrations (0–10 %) of alcohol at room temperature; consideration of the OH overtone and combination bands
● Post-processing of spectra by the 2D correlation spectral analysis (algorithm based on Noda’s algorithm) for higher
concentrations (10–100 %) of alcohol at room temperature; consideration of the OH overtone and combination bands
Results:
● For concentrations lower than 5%, the main form of the intermolecular interaction in the mixtures is hydrophobic hydration,
while for the concentrations from 5 to 10 % the OH from water and ethanol form the fixed interaction by hydrogen bonds.
● The maximum participation of water in hydrogen bonding in EtOH-water mixtures is reached at 40% concentration of EtOH.
● In 40–100% solutions EtOH molecules tend to initiate the self-association which leads to the weakening of the interaction
between alcohol and water molecules.
Katsu et al., 2019 [78] Method: FT near-IR in transmission mode; Alcohol: eleven polyhydric alcohols including EtGOH and GlOH
● Post-processing of spectra of solutions at different concentrations and temperatures by the MCR-ALS decomposition into the
three spectral components

Results:
● Spectra of solutions can be decomposed into three spectral components regardless of the type of alcohol, i.e., component
related to the weakly hydrogen-bonded (or non-hydrogen-bonded) water, component related to water strongly
hydrogen-bonded with other water molecules, and component related to water interacting with the alcohol and the alcohol
itself.
● The component related to water strongly hydrogen-bonded with other water molecules increases initially but then decreases
with the increase in alcohol concentration; the initial increase is attributed to the enhancement of hydrogen bonding by
hydrophobic interactions, while the subsequent decrease to an increase in water-alcohol interactions and a decrease in water
concentration; the maximum increase in the abundance of this component of water is greater for alcohols with larger alkyl
groups while smaller for alcohols with more hydroxyl groups.
Nedić et al., 2011 [81] Method: FT mid-IR in supersonic jet; Alcohol: MeOH and EtOH
● Averaged spectra from 150–200 pulses; studying dimers and trimers composed of water and either methanol or ethanol for
different concentrations and room temperature

Results:
● In mixed dimers, water assumes the role of the hydrogen bond donor, due to its inferior role as acceptor in comparison to the
two alcohols.
● Hydrogen bond induced red shifts of OH modes behave non-linearly as a function of composition.
● Effects induced by the weak hydrogen bond between the CH group and the oxygen lone electron pair are detected.
● Water forces EtOH into its less stable gauche conformation upon dimerization.
(continued)
Deng et al., 2019 [106] Method: FT mid-IR, Femtosecond 2D mid-IR in transmission mode; Alcohol: MeOH, EtOH, 1PrOH, 1BtOH and 1-pentanol
● Analysis of FT mid-IR spectra of solutions at different concentrations and temperatures via the HOH bending and OH
stretching modes
● KSCN rotational decay (anisotropy decays of the SCN- vibration) at different concentrations and room temperature
Results:
● Blue shift of the water HOH bending frequency at room temperature seems to be caused by stronger hydrogen bond networks,
where the water structure enhancement is observed on the hydrophobic surface, however it cannot be truly ice-like.
● Ordered water structure towards hydrophobic surface disappears with the increase in temperature and the new disordered
water structure arises in dependence of surface topography (transformation from a more ordered to a less ordered water
structure is only observed in the large-alcohol solutions).
Bakulin et al., 2011 [105] Method: Femtosecond 2D mid-IR; Alcohol: tertBtOH
● Two-dimensional correlation spectra measured at different waiting times; OH stretching vibration in solution of tertBtOH in
isotopically diluted water

● Slope value as a function of delay time for different concentrations at room temperature
Results:
● At low concentrations, the amplitude of the slow water component increases linearly with the solute concentration.
● Slow water molecules hydrate the hydrophobic groups of the solute (slow hydrogen bond dynamics in the hydration shell of
methyl groups).
● The effect of hydrophobic moieties on the dynamics of water appears to be the hindrance of the formation of bifurcated
hydrogen-bonded structures.
for ci = 2.5, 5, 7.5, 10, 15, 20, 25, and 40 vol% (14.2)
with the additional term addALC_ci BW ALC_c , which served to compensate for the
bands dipping below the surrounding baseline in the frequency ranges of the water
OH-stretching band and the HOH-bending band. Therefore, the actual bulk water
fraction in the solutions was calculated as
bALC ci =̂
bALC ci − addALC ci for ci = 2.5, 5, 7.5, 10, 15, 20, 25, and 40 vol%
(14.3)
The SC spectra at different temperatures for the 5 vol% aqueous solutions of the
alcohols were derived by subtracting the same fraction of the bulk water compo-
nent at a given temperature, BW ALC_Ti , from the spectrum of the solution, SALC_Ti ,
as determined for the ci = 5 vol% solution at 298 K by the calculated weighting
coefficient bALC_5 vol% in Eq. (14.3), i.e.
SCALC Ti = SALC Ti − bALC5 vol% BWALCTi
for Ti = 298, 301, … , 349 K (the step of 3 K) (14.4)
Further details of the decomposition can be found in our paper [94].
The SC spectra thus obtained were further used to investigate the extent to
which the hydrogen-bonding network of the H2 O molecules is perturbed by the
alcohol solutes and differs from that of the bulk water, as well as how the aqueous
environment affects the vibrational characteristics of the alcohol molecules. We
chose the HOH-bending mode as a probe to study the changes in hydrogen bonding
of the perturbed water, which has been shown to be an adequate indicator of
the microscopic structure of H2 O, providing even more direct information than the
most commonly studied OH-stretching mode [13]. The main advantage of the
HOH-bending mode is that it belongs exclusively to H2 O molecules and does
not interfere with other modes, unlike the OH stretching, which also covers the
vibrational contribution of the hydroxyl group of alcohols [13]. Inspection of the
shifts in the frequency of the HOH-bending peak from the SC spectra with respect
to the position of the HOH-bending peak of the bulk water spectrum served us
to identify the perturbation of the water molecules upon addition of alcohol in
terms of changes in hydrogen-bonding strength. The ratio between the integral
area under the HOH-bending peak of the SC spectra and that of the bulk water
spectrum for a given concentration was calculated to estimate the number of
perturbed water molecules per one molecule of alcohol. The shifts in the CH- and
CO-stretching peak frequencies observed in the SC spectra in comparison to the
corresponding peak frequencies from the IR spectrum of the pure alcohol were
analyzed to identify the changes in the molecular vibrations of the alcohol when
mixed with water. Figure 14.13 shows the 10 vol% spectra of the MeOH–water
solution, the corresponding bulk water and SC spectral components from the MCR
decomposition, and pure MeOH. The insets in Figure 14.13 show the magnified
spectral regions of the probes we used to analyze the alcohol hydration phenomena.
The HOH-bending mode as the probe for consideration of the alcohol-aqueous
solutions was previously used and reported in refs. [85, 86, 106], while the CH- and
10 vol% MeOH–water solution spectrum νCO CO

SC spectral component × 5 stretching
Bulk water spectral component
Pure MeOH
ATR absorbance (a.u.)
CH HOH
stretching bending
νCH δHOH
2950 1640 1020
4000 3500 3000 2500 2000 1500 1000 650

Wavenumber (cm–1)
Figure 14.13 Experimentally determined vibrational spectra of 10 vol% MeOH solution

and pure alcohol at 298 K and atmospheric pressure, and the spectral components of bulk
water and SC spectrum (magnified 5 times) from the decomposition of the solution
spectrum in the entire frequency range with the regions of interest in the magnified insets,
i.e. the CH-stretching mode region (𝜈 CH ), the HOH-bending mode region (𝛿 HOH ), and the
CO-stretching mode region (𝜈 CO ) with the spectral peak frequency shifts indicated.
CO-stretching modes were used in refs. [85, 86, 108–110] and [82, 85, 104], respec-
tively (see Table 14.2).
14.4.2.3 Influence of Alcohol Concentration and Temperature

As shown in Table 14.2, various aspects of alcohol hydration in dependence on alco-
hol type, concentration, and temperature have been investigated by several authors
using IR spectroscopic methods. Our contribution to this topic concerns the iden-
tification of changes in the hydrogen-bonding network of water upon addition of
alcohol at different concentrations and temperatures and the underlying phenom-
ena that explain these changes.
The HOH-bending peak frequency, which we chose as a probe to identify changes
in the hydrogen-bonding network of water from the SC spectral component,
is shown in Figure 14.14 in dependence of alcohol volume concentration (left)
and dependence of temperature (right). The filled symbols on the left represent
the same points as the filled symbols on the right, i.e. the frequency of the
HOH-bending peak identified from the SC spectral component obtained from the
MCR decomposition of the 5 vol% alcohol–water solution at 298 K and atmospheric
pressure. The blueshift of the HOH-bending frequency relative to the bulk water
indicates the embedding of the perturbed water molecules in the hydrogen-bonding
network with stronger hydrogen bonds compared to the bulk [13]. A similar
observation of the blueshift in HOH bending was previously reported by Deng
et al. [106], based on the FT mid-IR measurements in transmission mode (see
Table 14.2). However, their analysis of the shifts is performed using the spectra of
solutions that include the vibrational properties of both the bulk water and the
perturbed water, whereas we use the SC spectral components that include only
the perturbed water molecules. Thus, the shifts we observe relate exclusively to
MeOH EtOH 2PrOH 2BtOH Bulk water Bulk water

1PrOH 1BtOH MeOH EtOH 2PrOH 2BtOH
1670 1PrOH 1BtOH
1655
HOH bending frequency (cm–1)
1640 Bulk Bulk

water water
isoBtOH tertBtOH EtGOH GlOH Bulk water Bulk water

isoBtOH tertBtOH EtGOH GIOH
1670
1655
1640 Bulk Bulk

water water
0 5 10 20 40 0 5 10 20 40 298 313 328 343 298 313 328 343

Alcohol concentration (vol%) Temperature (K)
(a) (b)
Figure 14.14 HOH-bending mode frequency, 𝛿 HOH , obtained from the SC and bulk water
spectral components of: (a) alcohol–water solution spectra at different volume
concentrations at 298 K and atmospheric pressure, and (b) 5 vol% alcohol–water solution
spectra at different temperatures and atmospheric pressure.
the distorted part of the water. Complementary Raman MCR spectroscopy per-
formed by Davis et al. [9] also supports our findings by showing that the hydration
shells of linear alcohols at ambient temperature and pressure exhibit fewer weak
hydrogen bonds and enhanced water structure compared to the surrounding
bulk water.
A decrease in the blueshifted HOH-bending frequency is observed with the
increase in concentration (for monohydric alcohols in Figure 14.14 – left)
and temperature (Figure 14.14 – right), which tends toward the bulk water
vibrational frequency. Exceptions to this picture are the two polyhydric alco-
hols (Figure 14.14 – left), which show a slight increase in the blueshift of the
HOH-bending frequency as the alcohol concentration is increased. Although the
decreasing trend in HOH-bending frequency, where observed, indicates a decrease
in the strength of the perturbed water hydrogen bonds with concentration and
temperature, it remains blueshifted, at least in the concentration and temperature
ranges studied.
Moreover, as the number of hydrocarbon moieties in the alkyl chain of alcohols
increases, the indicated slope of the blueshifted HOH-bending frequency gets
steeper (see dashed lines: guide for the eyes in Figure 14.14), particularly in
dependence of concentration (Figure 14.14 – left) and less evidently in dependence
of temperature (Figure 14.14 – right).
6 Data obtained in our research
Sun and petersen, 2017

Data from the literature
Böhm et al., 2017 &

5
No. of perturbed H2O molecules(/)
Böhm et al., 2017 &

4
Böhm et al., 2017
Davis et al., 2012

Böhm et al., 2017
Böhm et al., 2017
H
H
H
H
H
H
H
tO
H
O
rO
tO
rO
tO
eO
IO
O
Bt
tB
G
Et
1P
1B
2P
2B
G
iso
M
Et
r
te
Figure 14.15 Number of perturbed H2 O molecules per one molecule of alcohol calculated
in our study for different concentrations (color-filled boxes and diamond symbols), and the
corresponding numbers of perturbed H2 O molecules found in the literature (dashed line
boxes).
Further analysis of the HOH-bending mode probe is concerned with the number
of perturbed water molecules per one molecule of alcohol. For this purpose, we cal-
culated the integral areas under the HOH-bending peak from the bulk water and SC
spectral components. The observed linear dependence of the integral areas on the
volume concentration up to 40 vol% [94] provided a reasonable argument for calcu-
lating the number of perturbed water molecules from the ratio of the HOH integral
area to the sum of the HOH integral areas of the two spectral components, and the
known nominal concentration of the alcohol–water solution. Further details on the
deeper elaboration of the HOH-bending mode areas can be found in our work [94].
Color-filled boxes and diamond symbols in Figure 14.15 show the calculated num-
ber ranges of perturbed water molecules per one molecule of alcohol based on the
SC component from the decomposition of the measured spectra at different alcohol
concentrations and the temperature of 298 K. The dashed-line boxes in Figure. 14.15
show the ranges of numbers available in the work of Böhm et al. [113], Davis et al.
[9], and Sun and Petersen [110]. These relatively low numbers of perturbed water
molecules indicate no evidence of any bigger distorted water structural units sur-
rounding the alcohol molecule and are consistent with previously observed indica-
tions that small alcohols, such as methanol, ethanol, 1-propanol, and 2-propanol,
are commonly known as inhibitors of clathrate hydrate formation [114].
Looking at the number of perturbed water molecules from our study (color-filled
boxes and diamond symbols in Figure 14.15), there is an indication of an unexpected
influence of the hydrocarbon moieties as the alkyl chain grows from propanol to
butanol. Indeed, the number remains almost the same for 1PrOH and 1BtOH and
even drops when the chain grows from 2PrOH to 2BtOH, while intuitively one
would expect an increase. As discussed by Kwaśniewicz and Czarnecki [108] (see
Table 14.2), this counterintuitive phenomenon can be attributed to the tendency
for more pronounced hydrophobic interactions as alkyl chain length increases.
Another aspect of the observed phenomenon may be the change in cavity formation
behavior, leading from volume dependence for small cavities to area dependence
for larger cavities (i.e. the small-to-large length scale crossover) [115]. Data in the
literature suggest that this crossover occurs at lengths on the order of 1 nm (the esti-
mated chain length of 1BtOH [9]) near-standard thermodynamic conditions (see,
e.g. [9, 115–117]). However, some authors attribute this to the change in the physical
mechanism of hydration from an entropy-dominated to an enthalpy-dominated
mechanism [116, 118] by invoking interfacial physics phenomena for the hydration
process of large solutes [117]. In contrast, other authors explain that crossover
behavior for cavities with a radius of at least 2 nm does not require different physical
mechanisms to describe it, and that cavity formation can be considered a purely
entropy-driven process [119, 120], accompanied by the compensating changes
in enthalpy and entropy of other sub-processes when a hydrophobic molecule is
transferred to water [120–122].
The vibrational characteristic of an alcohol molecule when added to the water are
investigated via the CH- and CO-stretching frequencies at different concentrations
compared to the frequency of pure alcohol (see Figure 14.16a and Figure 14.17a), and
at different temperatures (see Figure 14.16b and Figure 14.17b). As in Figure 14.14,
the filled symbols on (a) represent the same points as the filled symbols on (b) of
Figures 14.16 and 14.17, which correspond to a 5 vol% alcohol–water solutions at
298 K.
The CH-stretching frequency decreases with concentration and approaches that
of pure alcohol (Figure 14.16a). It remains more or less the same in the tempera-
ture range between 298 and 349 K for all alcohols studied (Figure 14.16b). On the
other hand, the CO-stretching frequency has an increasing tendency toward the CO
frequency of pure monohydric alcohols and a decreasing tendency for polyhydric
alcohols (Figure 14.17a). It also shows temperature independence up to 349 K for
monohydric alcohols and a slight decrease in CO-stretching frequency with temper-
ature for the two polyhydric alcohols, i.e. EtGOH and GlOH (Figure 14.17b).
The frequency shifts observed in our study for the vibrational modes of alcohol
molecules can be supported by the phenomenology of vibrational solvatochromism
induced by the surrounding environment in solutions [123]. They can be attributed
to the changes in solute–solvent intermolecular interaction potential along its
normal coordinate, charge transfer, dispersion and electrostatic interactions, the
affected vibrational dipole moment and polarizability, and the coupling between
MeOH EtOH 2PrOH 2BtOH MeOH EtOH 2PrOH 2BtOH

2985 1PrOH 1BtOH 1PrOH 1BtOH
2970
Pure
alcohol
2955
CH stretching frequency (cm–1)
Pure
2940
alcohol
isoBtOH tertBtOH EtGOH GlOH isoBtOH tertBtOH EtGOH GlOH

2985
2970
2955 Pure
alcohol Pure
alcohol
2940
05 20 40 100 05 20 40 100 298 313 328 343 298 313 328 343
(a) (b)
Figure 14.16 CH-stretching mode frequency, 𝜈 CH , obtained from the SC and bulk water
modes [123, 124]. In this context, the CH-stretching frequency shifts (Figure 14.16a)
may indicate the increased number of contacts between alcohol molecules with
increasing concentration, whereas the alcohol aggregates are not expected to be
significantly affected by hydrophobic interaction, but mostly resemble contacts
in random mixtures [100]. A decrease in vibrational frequency with increasing
concentration could be related to the decrease in electron density of the CH bond
due to partial electron transfer to oxygen, which was previously observed by Mizuno
et al. [125] using NMR. Consequently, this affects the CO-stretching frequency shift
as a function of alcohol concentration (Figure 14.17a). Nevertheless, temperature
does not seem to have an obvious influence on this mechanism in the studied range.
The other mechanism that may affect the CO-stretching frequency is different
hydrogen-bonding possibilities for the alcohol molecules, which change with the
composition of the alcohol–water solution [126]. Kabisch and Pollmer [126] found
that if the ratio of occupied protons (i.e. those participating in hydrogen bonding)
of the alcohol OH group to the occupied lone pairs in pure alcohol tends to one,
an increase in this ratio due to any possible reason imposed by the change in the
environment of the alcohol molecule will increase the CO-stretching frequency,
whereas a decrease in this ratio will cause a decrease in the CO-stretching fre-
quency. Considering this mechanism in relation to our results, the ratio between
the occupied protons of the alcohol OH protons and the occupied alcohol lone
MeOH EtOH 2PrOH 2BtOH MeOH EtOH 2PrOH 2BtOH

1PrOH 1BtOH 1PrOH 1BtOH
1080 1130 1080 1130
1070 1126 1070
1055 1055 1126
1050 1050
CO stretching frequency (cm–1)
1108
1045 Pure Pure
1045
1108
alcohol alcohol
1020 1020
1010 1102 1010 1102
1200 isoBtOH tertBtOH EtGOH GlOH 1200 isoBtOH tertBtOH EtGOH GlOH
1045 1045
1195 1195
1040 1040
1040 1040
Pure
1030 alcohol 1030
Pure
1020 1035 alcohol
1020 1035
05 20 40 100 05 20 40 100 298 313 328 343 298 313 328 343
(a) (b)
Figure 14.17 CO-stretching mode frequency, 𝜈 CO , obtained from the SC and bulk water
pairs appears to decrease in aqueous solutions of monohydric alcohols, while it

increases in solutions of polyhydric alcohols (Figure 14.17a). However, it appears to
be relatively insensitive to temperature (Figure 14.17a).
According to Kabisch and Pollmer [126], the altered ratio between the occupied
OH protons and the lone pairs of alcohol in aqueous environments, in turn, leads
to changes in the same ratio concerned with the water molecules, resulting in
changes in the fraction of lone pairs of water not involved in hydrogen bonding
compared to bulk water. We suggest that this mechanism controls the evolution of
the blueshifted HOH-bending frequency in dependence on alcohol concentration
(Figure 14.14 – left). Further details are available in ref. [94].
The presence of the perturbed H2 O molecules confirms that there is some
water restructuring induced by the alcohol solute and, together with other studies
[9, 112, 113, 127, 128], proves that the structural perturbation is dictated by the
hydrophobic alkyl group. Based on the studies and elaborations of Dixit et al.
[127, 128] and Bowron et al. [112], the observed perturbation in the solvent water
network can be thought of as appearing beyond the first coordination shell of a
water molecule, presumably at the level of the second-neighbor water correlations.
Thus, should the structural perturbation of water upon addition of alcohol occur
beyond the first coordination shell, this implies that the immediate water–water
structural correlation remains bulk like. Finally, this explains the presence of the
notable bulk-like water component even in the alcohol–water systems with high
alcohol concentrations (above 65 vol%) that we studied recently [94]. From what
we have learned so far from our ATR FT-IR spectroscopic data is that there is no
evidence for larger distorted structural units different from those found in bulk
water at ambient pressure conditions and temperatures at and above ambient.
14.5 Epilogue
We have shown that the application of standard IR spectroscopy, upgraded with
reliable spectral processing and theoretical support, can challenge the most subtle
processes that have been unanswered for years. By using methods of the highest
accuracy to measure and calculate subtle effects of hydrogen bonding in water, we
have obtained evidence that clearly supports the “iceberg” view of hydrophobicity
proposed by Frank and Evans [41] and Kauzmann [42]. Our experimental results
show that water hydrogen bonds are strengthened in the vicinity of purely hydropho-
bic solutes to the level observed in ice and clathrates. The origin of hydrophobicity
is based on electrostatic screening that combines the hydrophobic effect with
electrostatic interaction. However, this may not be the only mechanism involved
in hydrophobicity. The above explanation for the strengthening of water hydrogen
bonds near hydrophobic solutes is consistent with the general concept that electro-
static interactions are stronger near hydrophobic groups. Berry and coworkers have
shown that water molecules near a hydrophobic solute have lower dielectric suscep-
tibility [129] or, in microscopic terms, less effective electrostatic screening [130]. To
find out more about the phenomenology of the hydrophobic effect of amphiphilic
compounds, we analyzed aqueous solutions of the simplest amphiphilic molecules,
i.e. alcohols. Our results, based on the infrared spectra, show that the hydrophobic
moiety of the alcohol governs perturbation of the hydrogen-bonding water network.
However, the number of perturbed water molecules embedded in the network with
stronger hydrogen bonds than in bulk water ranges from 1 to 3 per one molecule
of alcohol. Therefore, there is no evidence for clathrate-like structures or other
larger distorted structural units that would differ from those in bulk water, at
least not under ambient conditions. This intuitively less expected result for the
alcohol-aqueous solutions compared to the results for the purely hydrophobic
compounds testify to a complex interplay of various molecular and atomistic mech-
anisms that need to be further explored to fully elucidate the influence of water on
the conformational preferences of larger amphiphiles, such as biomolecules.
Acknowledgments
The authors gratefully acknowledge financial support from the Slovenian Research
Agency (ARRS), Slovenia, through the project “The influence of intermolecular
interactions on the structure of peptides and proteins,” No. J1-1705, and research
core funding No. P1-0010. In addition, the authors would like to thank Franci
Merzel and Franc Avbelj for stimulating discussions.
References 471
References
1 Max, J.-J. and Chapados, C. (2002). Isotope effects in liquid water by infrared
spectroscopy. J. Chem. Phys. 116 (11): 4626–4642.
2 Maréchal, Y. (2011). The molecular structure of liquid water delivered by
absorption spectroscopy in the whole IR region completed with thermodynam-
ics data. J. Mol. Struct. 1004 (1): 146–155.
3 Novak, A. (1974). Hydrogen bonding in solids correlation of spectroscopic
and crystallographic data. In: Large Molecules. Structure and Bonding, vol. 18.
Berlin, Heidelberg: Springer https://doi.org/10.1007/BFb0116438.
4 Stiopkin, I.V., Weeraman, C., Pieniazek, P.A. et al. (2011). Hydrogen bonding at
the water surface revealed by isotopic dilution spectroscopy. Nature 474 (7350):
192–195.
5 Hecht, D., Tadesse, L., and Walters, L. (1992). Defining hydrophobicity –
probing the structure of solute-induced hydration shells by Fourier-transform
infrared-spectroscopy. J. Am. Chem. Soc. 114 (11): 4336–4339.
6 Hecht, D., Tadesse, L., and Walters, L. (1993). Correlating hydration shell struc-
ture with amino-acid hydrophobicity. J. Am. Chem. Soc. 115 (8): 3336–3337.
7 Sharp, K.A., Madan, B., Manas, E., and Vanderkooi, J.M. (2001). Water struc-
ture changes induced by hydrophobic and polar solutes revealed by simulations
and infrared spectroscopy. J. Chem. Phys. 114 (4): 1791–1796.
8 Auer, B., Kumar, R., Schmidt, J.R., and Skinner, J.L. (2007). Hydrogen bonding
and Raman, IR, and 2D-IR spectroscopy of dilute HOD in liquid D2 O. Proc.
Natl. Acad. Sci. U.S.A. 104 (36): 14215–14220.
9 Davis, J.G., Gierszal, K.P., Wang, P., and Ben-Amotz, D. (2012). Water struc-
tural transformation at molecular hydrophobic interfaces. Nature 491 (7425):
582–585.
10 Marechal, Y. (1983). A quantitative analysis of the νs (IR) bands of H-bonds. I.
Theory. Chem. Phys. 79 (1): 69–83.
11 Marechal, Y. (1983). A quantitative analysis of the νs (IR) bands of H-bonds. II.
Adipic acid crystals. Chem. Phys. 79 (1): 85–94.
12 Maréchal, Y. (2007). The Hydrogen Bond and the Water Molecule: The Physics
and Chemistry of Water, Aqueous and Bio Media. Amsterdam; Oxford: Elsevier.
13 Seki, T., Chiang, K.-Y., Yu, C.-C. et al. (2020). The bending mode of water: a
powerful probe for hydrogen bond structure of aqueous systems. J. Phys. Chem.
Lett. 11 (19): 8459–8469.
14 Grdadolnik, J. (2003). Saturation effects in FTIR spectroscopy: investigation of
amide I and amide II bands in protein spectra. Acta Chim. Slov. 50: 777–788.
15 Bertie, J.E. and Eysel, H.H. (1985). Infrared intensities of liquids I: determi-
nation of infrared optical and dielectric constants by FT-IR using the CIRCLE
ATR cell. Appl. Spectrosc. 39 (3): 392–401.
16 Bertie, J.E. and Lan, Z. (1996). An accurate modified Kramers–Kronig transfor-
mation from reflectance to phase shift on attenuated total reflection. J. Chem.
Phys. 105 (19): 8502–8514.
17 Grobelnik, B. and Grdadolnik, J. (2008). Calculation of the absorption spectrum

from an ATR infrared experiment. Acta Chim. Slov. 55 (4): 978–984.
18 Grdadolnik, J. (2003). Infrared difference spectroscopy: Part I. Interpretation of
the difference spectrum. Vib. Spectrosc. 31: 289–294.
19 Grdadolnik, J. and Maréchal, Y. (2003). Infrared difference spectroscopy: Part II.
Spectral decomposition. Vib. Spectrosc. 31: 279–288.
20 Liu, Y.J., Kyne, M., Wang, C., and Yu, X.Y. (2020). Data mining in Raman imag-
ing in a cellular biological system. Comput. Struct. Biotechnol. J. 18: 2920–2930.
21 Ben-Amotz, D. (2019). Hydration-shell vibrational spectroscopy. J. Am. Chem.
Soc. 141: 10569–10580.
22 MCR-ALS GUI v4c, Umeå Univ. Sweden. (2016). https://www.umu.se/en/
research/infrastructure/visp/downloads/ (accessed 18 November 2020).
23 Gorzsás, A. (2016). Vibrational Spectroscopy Core Facility. Sweden: Umeå
University.
24 Smith, J.D., Cappa, C.D., Messer, B.M. et al. (2005). Response to comment
on “Energetics of hydrogen bond network: rearrangements in liquid water”.
Science 308 (5723): 793–793.
25 Smith, J.D., Cappa, C.D., Wilson, K.R. et al. (2005). Unified description of
temperature-dependent hydrogen-bond rearrangements in liquid water. Proc.
Natl. Acad. Sci. U.S.A. 102 (40): 14171–14174.
26 Smith, J.D., Cappa, C.D., Wilson, K.R. et al. (2004). Energetics of hydrogen
bond network rearrangements in liquid water. Science 306 (5697): 851–853.
27 Niskanen, J., Fondell, M., Sahle, C.J. et al. (2019). Compatibility of quantitative
X-ray spectroscopy with continuous distribution models of water at ambient
conditions. Proc. Natl. Acad. Sci. U.S.A. 116 (10): 4058–4063.
28 Nilsson, A. and Pettersson, L.G.M. (2015). The structural origin of anomalous
properties of liquid water. Nat. Commun. 6 (1): 8998.
29 Kim, K.H., Späh, A., Pathak, H. et al. (2017). Maxima in the thermodynamic
response and correlation functions of deeply supercooled water. Science 358
(6370): 1589–1593.
30 Walrafen, G.E., Hokmabadi, M.S., and Yang, W.H. (1986). Raman isosbestic
points from liquid water. J. Chem. Phys. 85 (12): 6964–6969.
31 Chaplin, F.M. (2019). Structure and properties of water in its various states. In:
Encyclopedia of Water: Science, Technology, and Society (ed. A.P. Maurice), 1–19.
Wiley.
32 Martinez Maestro, L., Marques, M., Camarillo, E. et al. (2016). On the existence
of two states in liquid water: impact on biological and nanoscopic systems. Int.
J. Nanotechnol. 13: 667.
33 Gallo, P., Amann-Winkel, K., Angell, C.A. et al. (2016). Water: a tale of two
liquids. Chem. Rev. 116 (13): 7463–7500.
34 Soper, A.K. (2011). Water: two liquids divided by a common hydrogen bond.
J. Phys. Chem. B 115 (48): 14014–14022.
35 Soper, A.K. (2019). Is water one liquid or two? J. Chem. Phys. 150 (23): 234503.
36 Blokzijl, W. and Engberts, J.B.F.N. (1993). Hydrophobic effects – opinions and
facts. Angew. Chem. Int. Ed. 32 (11): 1545–1579.
References 473
37 Southall, N.T., Dill, K.A., and Haymet, A.D.J. (2002). A view of the hydrophobic
effect. J. Phys. Chem. B 106 (3): 521–533.
38 Ball, P. (2008). Water as an active constituent in cell biology. Chem. Rev. 108
(1): 74–108.
39 Laage, D., Stirnemann, G., Sterpone, F. et al. (2011). Reorientation and
allied dynamics in water and aqueous solutions. Annu. Rev. Phys. Chem. 62:
395–416.
40 Baldwin, R.L. (2014). Dynamic hydration shell restores Kauzmann’s 1959 expla-
nation of how the hydrophobic factor drives protein folding. Proc. Natl. Acad.
Sci. U.S.A. 111 (36): 13052–13056.
41 Frank, H.S. and Evans, M.W. (1945). Free volume and entropy in condensed
systems III. Entropy in binary liquid mixtures; partial molal entropy in dilute
solutions; structure and thermodynamics in aqueous electrolytes. J. Chem. Phys.
13 (11): 507–532.
42 Kauzmann, W. (1959). Some factors in the interpretation of protein denatura-
tion. Adv. Protein Chem. 14: 1–63.
43 Haselmeier, R., Holz, M., Marbach, W., and Weingartner, H. (1995). Water
dynamics near a dissolved noble-gas – first direct experimental-evidence for a
retardation effect. J. Phys. Chem. 99 (8): 2243–2246.
44 Buchanan, P., Aldiwan, N., Soper, A.K. et al. (2005). Decreased structure on dis-
solving methane in water. Chem. Phys. Lett. 415 (1–3): 89–93.
45 Pertsemlidis, A., Saxena, A.M., Soper, A.K. et al. (1996). Direct evidence for
modified solvent structure within the hydration shell of a hydrophobic amino
acid. Proc. Natl. Acad. Sci. U.S.A. 93 (20): 10769–10774.
46 Koh, C.A., Wisbey, R.P., Wu, X.P. et al. (2000). Water ordering around methane
during hydrate formation. J. Chem. Phys. 113 (15): 6390–6397.
47 Bowron, D.T., Filipponi, A., Lobban, C., and Finney, J.L. (1998).
Temperature-induced disordering of the hydrophobic hydration shell of Kr
and Xe. Chem. Phys. Lett. 293 (1–2): 33–37.
48 Stirnemann, G., Hynes, J.T., and Laage, D. (2010). Water hydrogen bond
dynamics in aqueous solutions of amphiphiles. J. Phys. Chem. B 114 (8):
3052–3059.
49 Lee, B. (1985). The physical origin of the low stability of non-polar solutes in
water. Biopolymers 24: 813–823.
50 Lee, B. (1991). Solvent reorganization contribution to the transfer thermody-
namics of small nonpolar molecules. Biopolymers 31 (8): 993–1008.
51 Grdadolnik, J., Merzel, F., and Avbelj, F. (2017). Origin of hydrophobicity and
enhanced water hydrogen bond strength near purely hydrophobic solutes. Proc.
Natl. Acad. Sci. U.S.A. 114 (2): 322–327.
52 Bergstrom, P.A., Lindgren, J., and Kristiansson, O. (1991). An IR study of the
hydration of perchlorate, nitrate, iodide, bromide, chloride and sulfate anions in
aqueous solution. J. Phys. Chem. 95 (22): 8575–8580.
53 Stangret, J. and Gampe, T. (2002). Ionic hydration behavior derived from
infrared spectra in HDO. J. Phys. Chem. A 106 (21): 5393–5402.
54 Sloan, E.D. and Carolyn, K.A. (2008). Clathrate Hydrates of Natural Gases
(ed. J.G. Speight). Boca Raton; London; New York: CRC Press.
55 Ballard, A.L. and Sloan, E.D. (2001). Hydrate phase diagrams for
methane + ethane + propane mixtures. Chem. Eng. Sci. 56 (24): 6883–6895.
56 Adisasmito, S., Frank, R.J., and Sloan, E.D. (1991). Hydrates of carbon-dioxide
and methane mixtures. J. Chem. Eng. Data 36 (1): 68–71.
57 Hamm, P. and Zanni, M.T. (2011). Concepts and Methods of 2D Infrared Spec-
troscopy. Cambridge: Cambridge University Press.
58 Sharp, K.A. and Madan, B. (1997). Hydrophobic effect, water structure, and
heat capacity changes. J. Phys. Chem. B 101 (21): 4343–4348.
59 Raschke, T.M. and Levitt, M. (2005). Nonpolar solutes enhance water structure
within hydration shells while reducing interactions between them. Proc. Natl.
Acad. Sci. U.S.A. 102 (19): 6777–6782.
60 Titantah, J.T. and Karttunen, M. (2012). Long-time correlations and
hydrophobe-modified hydrogen-bonding dynamics in hydrophobic hydration.
J. Am. Chem. Soc. 134 (22): 9362–9368.
61 Kirchner, B., Stubbs, J., and Marx, D. (2002). Fast anomalous diffusion of small
hydrophobic species in water. Phys. Rev. Lett. 89 (21): 215901-1–215901-4.
62 Montagna, M., Sterpone, F., and Guidoni, L. (2012). Structural and spectro-
scopic properties of water around small hydrophobic solutes. J. Phys. Chem. B
116 (38): 11695–11700.
63 Godec, A., Smith, J.C., and Merzel, F. (2011). Increase of both order and disor-
der in the first hydration shell with increasing solute polarity. Phys. Rev. Lett.
107 (26): 267801-1–267801-4.
64 Godec, A. and Merzel, F. (2012). Physical origin underlying the entropy loss
upon hydrophobic hydration. J. Am. Chem. Soc. 134 (42): 17574–17581.
65 Godec, A., Smith, J.C., and Merzel, F. (2013). Soft collective fluctuations gov-
erning hydrophobic association. Phys. Rev. Lett. 111 (12): 127801-1–127801-5.
66 Kresse, G. and Hafner, J. (1993). Ab initio molecular-dynamics for
liquid-metals. Phys. Rev. B: Condens. Matter 47 (1): 558–561.
67 Zhang, Y.K. and Yang, W.T. (1998). Comment on “Generalized gradient approxi-
mation made simple”. Phys. Rev. Lett. 80 (4): 890.
68 Perdew, J.P., Burke, K., and Ernzerhof, M. (1996). Generalized gradient approxi-
mation made simple. Phys. Rev. Lett. 77 (18): 3865–3868.
69 Grimme, S., Antony, J., Ehrlich, S., and Krieg, H. (2010). A consistent
and accurate ab initio parametrization of density functional dispersion
correction (DFT-D) for the 94 elements H-Pu. J. Chem. Phys. 132 (15):
154104-1–154104-19.
70 Mitchell, J.B.O. and Price, S.L. (1990). The nature of the N—H· · ·O=C
hydrogen-bond – an intermolecular perturbation-theory study of the formamide
formaldehyde complex. J. Comput. Chem. 11 (10): 1217–1233.
71 Ohno, K., Okimura, M., Akai, N., and Katsumoto, Y. (2005). The effect of coop-
erative hydrogen bonding on the OH stretching-band shift for water clusters
studied by matrix-isolation infrared spectroscopy and density functional theory.
Phys. Chem. Chem. Phys. 7 (16): 3005–3014.
References 475
72 Bartha, F., Kapuy, O., Kozmutza, C., and Van Alsenoy, C. (2003). Analysis of
weakly bound structures: hydrogen bond and the electron density in a water
dimer. J. Mol. Struct. THEOCHEM 666–667: 117–122.
73 Dannenberg, J.J. (2002). Cooperativity in hydrogen bonded aggregates. Models
for crystals and peptides. J. Mol. Struct. 615 (1): 219–226.
74 Eisenberg, D.S. and Kauzmann, W. (1969). The Structure and Properties of
Water. Oxford University Press.
75 Kusalik, P.G. and Svishchev, I.M. (1994). The spatial structure in liquid water.
Science 265 (5176): 1219–1221.
76 Ballal, D. and Chapman, W.G. (2013). Hydrophobic and hydrophilic interactions
in aqueous mixtures of alcohols at a hydrophobic surface. J. Chem. Phys. 139:
114706.
77 Salamatova, E., Cunha, A.V., Shinokita, K. et al. (2017). Hydrogen bond
and lifetime dynamics in diluted alcohols. Phys. Chem. Chem. Phys. 19:
27960–27967.
78 Katsu, S., Ito, S., Yoshimura, N., and Takayanagi, M. (2019). Variation in
near-infrared spectra of water containing polyhydric alcohol. J. Solution Chem.
48: 1564–1575.
79 Ball, P. (2017). Water is an active matrix of life for cell and molecular biology.
Proc. Natl. Acad. Sci. U.S.A. 114: 13327–13335.
80 Fujii, A., Sugawara, N., Hsu, P.J. et al. (2018). Hydrogen bond network struc-
tures of protonated short-chain alcohol clusters. Phys. Chem. Chem. Phys. 20:
14971–14991.
81 Nedić, M., Wassermann, T.N., Larsen, R.W., and Suhm, M.A. (2011). A
combined Raman- and infrared jet study of mixed methanol–water and
ethanol–water clusters. Phys. Chem. Chem. Phys. 13: 14050–14063.
82 Coldea, T.E., Socaciu, C., Fetea, F. et al. (2013). Rapid quantitative analysis of
ethanol and prediction of methanol content in traditional fruit brandies from
Romania, using FTIR spectroscopy and chemometrics. Not. Bot. Horti. Agrobot.
Cluj-Napoca 41: 143–149.
83 Hu, N., Wu, D., Cross, K. et al. (2010). Structurability: a collective measure of
the structural differences in vodkas. J. Agric. Food. Chem. 58: 7394–7401.
84 Chen, Y., Ozaki, Y., and Czarnecki, M.A. (2013). Molecular structure and
hydrogen bonding in pure liquid ethylene glycol and ethylene glycol–water
mixtures studied using NIR spectroscopy. Phys. Chem. Chem. Phys. 15:
18694–18701.
85 Habuka, A., Yamada, T., and Nakashima, S. (2020). Interactions of glycerol,
diglycerol, and water studied using attenuated total reflection infrared spec-
troscopy. Appl. Spectrosc. 74: 767–779.
86 Shimoaka, T. and Hasegawa, T. (2016). Molecular structural analysis of
hydrated ethylene glycol accounting for the antifreeze effect by using infrared
attenuated total reflection spectroscopy. J. Mol. Liq. 223: 621–627.
87 Bakker, H.J. (2012). Water’s response to the fear of water. Nature 491: 533–535.
88 Grdadolnik, J. and Maréchal, Y. (2002). Urea and urea–water solutions – an
infrared study. J. Mol. Struct. 615: 177–189.
89 Fogarty, A.C., Duboué-Dijon, E., Sterpone, F. et al. (2013). Biomolecular hydra-

tion dynamics: a jump model perspective. Chem. Soc. Rev. 42: 5672–5683.
90 Matvejev, V., Zizi, M., and Stiens, J. (2012). Hydration shell parameters of aque-
ous alcohols: THz excess absorption and packing density. J. Phys. Chem. B 116:
14071–14077.
91 Wu, X., Lu, W., Streacker, L.M. et al. (2018). Temperature-dependent hydropho-
bic crossover length scale and water tetrahedral order. J. Phys. Chem. Lett. 9:
1012–1017.
92 Rimola, A., Costa, D., Sodupe, M. et al. (2013). Silica surface features and
their role in the adsorption of biomolecules: computational modeling and
experiments. Chem. Rev. 113: 4216–4313.
93 Monroe, J., Barry, M., DeStefano, A. et al. (2020). Water structure and proper-
ties at hydrophilic and hydrophobic surfaces. Annu. Rev. Chem. Biomol. Eng. 11:
523–557.
94 Zupančič, B. and Grdadolnik, J. (2021). Solute-induced changes in the water
H-bond network of different alcohol-aqueous systems. J. Mol. Liq. 341: 1–11.
95 Ahmed, M., Singh, A.K., and Mondal, J.A. (2016). Hydrogen-bonding and
vibrational coupling of water in a hydrophobic hydration shell as observed by
Raman-MCR and isotopic dilution spectroscopy. Phys. Chem. Chem. Phys. 18:
2767–2775.
96 Bredt, A.J. and Ben-Amotz, D. (2020). Influence of crowding on hydrophobic
hydration-shell structure. Phys. Chem. Chem. Phys. 22: 11724–11730.
97 Davis, J.G., Rankin, B.M., Gierszal, K.P., and Ben-Amotz, D. (2013). On the
cooperative formation of non-hydrogen-bonded water at molecular hydrophobic
interfaces. Nat. Chem. 5: 796–802.
98 Dolenko, T.A., Burikov, S.A., Dolenko, S.A. et al. (2015). Raman spectroscopy
of water–ethanol solutions: the estimation of hydrogen bonding energy and
the appearance of clathrate-like structures in solutions. J. Phys. Chem. A 119:
10806–10815.
99 Gong, Y., Xu, Y., Zhou, Y. et al. (2017). Hydrogen bond network relaxation
resolved by alcohol hydration (methanol, ethanol, and glycerol). J. Raman
Spectrosc. 48: 393–398.
100 Rankin, B.M., Ben-Amotz, D., Van Der Post, S.T., and Bakker, H.J. (2015). Con-
tacts between alcohols in water are random rather than hydrophobic. J. Phys.
Chem. Lett. 6: 688–692.
101 Starciuc, T., Guinet, Y., Hedoux, A., and Shalaev, E. (2021). Water content
thresholds in glycerol/water system: low- and high-wavenumber Raman spec-
troscopy study. J. Mol. Liq. 321: 114678.
102 Wilcox, D.S., Rankin, B.M., and Ben-Amotz, D. (2013). Distinguishing aggrega-
tion from random mixing in aqueous t-butyl alcohol solutions. Faraday Discuss.
167: 177–190.
103 Yang, B., Cao, X., Lang, H. et al. (2020). Study on hydrogen bonding network in
aqueous methanol solution by Raman spectroscopy. Spectrochim. Acta, Part A
225: 117488.
References 477
104 Ahmed, M.K., Ali, S., and Wojcik, E. (2012). The C–O stretching infrared
band as a probe of hydrogen bonding in ethanol–water and methanol–water
mixtures. Spectrosc. Lett. 45: 420–423.
105 Bakulin, A.A., Pshenichnikov, M.S., Bakker, H.J., and Petersen, C. (2011).
Hydrophobic molecules slow down the hydrogen-bond dynamics of water.
J. Phys. Chem. A 115: 1821–1829.
106 Deng, G.H., Shen, Y., Chen, H. et al. (2019). Ordered-to-disordered transfor-
mation of enhanced water structure on hydrophobic surfaces in concentrated
alcohol–water solutions. J. Phys. Chem. Lett. 10: 7922–7928.
107 Dong, Q., Yu, C., Li, L. et al. (2019). Near-infrared spectroscopic study of
molecular interaction in ethanol–water mixtures. Spectrochim. Acta, Part A
222: 117183.
108 Kwaśniewicz, M. and Czarnecki, M.A. (2018). The effect of chain length on
mid-infrared and near-infrared spectra of aliphatic 1-alcohols. Appl. Spectrosc.
72: 288–296.
109 Shao, X., Cui, X., Wang, M., and Cai, W. (2019). High order derivative to
investigate the complexity of the near infrared spectra of aqueous solutions.
110 Sun, Y. and Petersen, P.B. (2017). Solvation shell structure of small molecules
and proteins by IR-MCR spectroscopy. J. Phys. Chem. Lett. 8: 611–614.
111 Primorac, T., Požar, M., Sokolić, F. et al. (2018). A simple two dimensional
model of methanol. J. Mol. Liq. 262: 46–57.
112 Bowron, D.T., Soper, A.K., and Finney, J.L. (2001). Temperature dependence of
the structure of a 0.06 mole fraction tertiary butanol–water solution. J. Chem.
Phys. 114: 6203–6219.
113 Böhm, F., Schwaab, G., and Havenith, M. (2017). Mapping hydration water
around alcohol chains by THz calorimetry. Angew. Chem. Int. Ed. 56:
9981–9985.
114 Alavi, S., Takeya, S., Ohmura, R. et al. (2010). Hydrogen-bonding alcohol–water
interactions in binary ethanol, 1-propanol, and 2-propanol + methane structure
II clathrate hydrates. J. Chem. Phys. 133: 074505.
115 Lum, K., Chandler, D., and Weeks, J.D. (1999). Hydrophobicity at small and
large length scales. J. Phys. Chem. B 103: 4570–4577.
116 Chandler, D. (2005). Interfaces and the driving force of hydrophobic assembly.
Nature 437: 640–647.
117 Garde, S. and Patel, A.J. (2011). Unraveling the hydrophobic effect, one
molecule at a time. Proc. Natl. Acad. Sci. U.S.A. 108: 16491–16492.
118 Pratt, L.R. and Chandler, D. (1977). Theory of the hydrophobic effect. J. Chem.
Phys. 67: 3683–3704.
119 Graziano, G. (2006). Comment on “Global thermodynamics of hydrophobic cav-
itation, dewetting, and hydration” [J. Chem. Phys. 123: 184504 (2005)]. J. Chem.
Phys. 125: 037101.
120 Graziano, G. (2006). Scaled particle theory study of the length scale depen-
dence of cavity thermodynamics in different liquids. J. Phys. Chem. B 110:
11421–11426.
121 Graziano, G. (2019). Contrasting the hydration thermodynamics of methane

and methanol. Phys. Chem. Chem. Phys. 21: 21418–21430.
122 Lee, B. (1994). Enthalpy–entropy compensation in the thermodynamics of
hydrophobicity. Biophys. Chem. 51: 271–278.
123 Baiz, C.R., Błasiak, B., Bredenbeck, J. et al. (2020). Vibrational spectroscopic
map, vibrational spectroscopy, and intermolecular interaction. Chem. Rev. 120:
7152–7218.
124 Błasiak, B., Londergan, C.H., Webb, L.J., and Cho, M. (2017). Vibrational
probes: from small molecule solvatochromism theory and experiments to
applications in complex systems. Acc. Chem. Res. 50: 968–976.
125 Mizuno, K., Miyashita, Y., Shindo, Y., and Ogawa, H. (1995). NMR and FT-IR
studies of hydrogen bonds in ethanol–water mixtures. J. Phys. Chem. 99:
3225–3228.
126 Kabisch, G. and Pollmer, K. (1982). Hydrogen bonding in methanol–organic
solvent and methanol–water mixtures as studied by the νCO and νOH . J. Mol.
Struct. 81: 35–50.
127 Dixit, S., Crain, J., Poon, W.C.K.K. et al. (2002). Molecular segregation observed
in a concentrated alcohol–water solution. Nature 416: 829–832.
128 Dixit, S., Soper, A.K., Finney, J.L., and Crain, J. (2002). Water structure and
solute association. Europhys. Lett. 59: 377–383.
129 Despa, F., Fernandez, A., and Berry, R.S. (2004). Dielectric modulation of
biological water. Phys. Rev. Lett. 93: 228104-01–228104-04.
130 Warshel, A. and Russell, S.T. (1984). Calculation of electrostatic interactions in
biological systems and in solutions. Q. Rev. Biophys. 17: 283–422.
479
15
Hydrogen Bond Chains in Foldamers and

Dynamic Foldamers
David T.J. Morris and Jonathan Clayden
School of Chemistry, Cantock’s Close, Bristol BS8 1TS, UK
15.1 Hydrogen-Bonded Foldamers

Hydrogen bonding is one of the most powerful aggregating interactions available in
nature. With an array of donors and acceptors of varying strengths and geometric
parameters, biomolecules use hydrogen bonds to exert precise control over protein
structure, exploit the unique solvent properties of water, and process information
through the highly selective association of DNA nucleobases [1]. Hydrogen bonds
vary in strength (8–170 kJ mol−1 ) depending on the chemical nature, proximity, and
geometry of the components, but relative to covalent and ionic bonds, they are weak
interactions [2]. This lability of hydrogen bonds confers rich dynamism to systems
they operate in, and allows them to modulate biological processes at physiologically
relevant temperatures [3].
Hydrogen-bonded chains are prevalent in supramolecular and biological systems.
Hydrogen-bond cooperativity – the synergistic bonding arising from contiguously
bound hydrogen-bonded units, is thought to contribute to protein folding and
structural fidelity in hydrogen-bonded polymers. While using cooperativity to
modulate hydrogen-bond strength remote from a binding site is a useful tool that
both nature and chemists have used or adapted for controlling macromolecu-
lar structure, it is not well-understood, particularly in intramolecular contexts.
Cockroft and coworkers used N,N-diarylformamides as molecular balances for
quantifying hydrogen-bond cooperativity in intramolecularly hydrogen-bonded
chains [4]. Rotameric formamides were appended with hydrogen-bond-donating
phenols, which biased the amide to hydrogen bond with them (Figure 15.1).
More alcohols were added to the hydrogen-bonding phenol, and the effect of the
additional hydrogen bonds was quantified by monitoring the perturbation of the
rotameric ratio with 19 F NMR. When two alcohols were appended (two hydrogen
bonds), the ortho-phenol became twice as strong a hydrogen-bond donor as the
single alcohol (after being corrected for the change in electronic structure using
Hammett plots). Interestingly, upon extension to the third alcohol (three hydrogen
bonds), hydrogen bonding with the formamide changed minimally, indicating

480 15 Hydrogen Bond Chains in Foldamers and Dynamic Foldamers
O H H O
H
H K
O O
N N
X X
F F
H O H O H O H O
H H H
O O H O H
N N N N
O O
X H
F F F F O
No hydrogen bonds One hydrogen bond Two hydrogen bonds Three hydrogen bonds
H O H O
H H
O H O H
N O N
O
H O
F F
Two hydrogen bonds Three hydrogen bonds
Figure 15.1 N,N-Diarylformamide balances for the quantifying hydrogen-bond

cooperativity.
that cooperativity has a minimal effect beyond two hydrogen bonds. Similarly,
molecular phenol was added to intermolecularly hydrogen bond to the balances.
Improved hydrogen bonding was observed when one alcohol was incorporated into
the balance (one intermolecular and one intramolecular hydrogen bond), but gave
no further improvement upon introduction of further intramolecular hydrogen
bonds. From this, short-range cooperative effects can have a significant effect on
hydrogen bonding strength, meaning that longer hydrogen-bond chains can be
globally strengthened by mutual cooperativity between adjacent units. This affects
oligomers of dual hydrogen-bond donors/acceptors, but cooperativity may also
arise when one hydrogen bond increases the effective molarity of another hydrogen
bond by preorganization. The remainder of this chapter focuses on hydrogen-bond
chains comprising at least two hydrogen bonds in oligomeric structures.
Chemists have recognized nature’s use of molecules that adopt specific folded
conformations and precisely orientate chemical functionality to achieve useful
function and have sought to make artificial molecules that can mimic these proper-
ties. These molecules form part of a family of synthetic molecules called foldamers.
Coined by Gellman in 1996, the term “foldamer” is defined as a polymer that has a
strong tendency to adopt a specific compact conformation [5]. Many foldamers have
been designed with the goal of making biologically orthogonal architectures that
can exhibit the same spatially specific orientation of chemical functionality as their
natural counterparts. For example, Gellman and coworkers simulated and synthe-
sized β-amino acid oligomers derived from trans-2-aminocyclohexanecarboxylic
acid (ACHC) and trans-2-aminocyclopentanecarboxylic acid (ACPC). It was found
that these oligomers formed particularly stable 14- and 12-helices (the number
denotes how many atoms are in the hydrogen-bonded rings) (Figure 15.2) [6].
These families of foldamers were discovered as part of an investigation to
make oligomers having well-defined secondary structures but with a minimum
number of residues. Oligomers derived from β-amino acids open up the option
O O
O O H
O
N O NO N NO
N O H H
N N
O H H O O
H H
HN O
ACHC O
O O
O O
O O
N N O O
N N
O H H O N H NN H
H H
H
O H O O
ACPC N
O
Figure 15.2 ACHC- and ACPC-derived foldamers (arrows indicate the atoms between
which hydrogen bonds form).
of incorporating ring structures into the foldamer, which can impose significant
conformational restriction and bias the formation of secondary structures [6].
Indeed, X-ray crystallography revealed well-defined helical conformations in the
solid state for ACHC oligomers as short as the tetramer, where only two hydrogen
bonds can form. It was also shown that the ACHC and ACPC helices were stable in
methanol and pyridine solutions. Treatment with a readily exchangeable deuterated
solvent is a very common way to determine the conformation of proteins and is
similarly applicable to the field of foldamers [7]. Periodic analysis by 1 H NMR
spectroscopy after exposure to the deuterated solvent results in attenuation of amide
protons signals at a rate commensurate with the strength of their internal hydrogen
bonds and how well the secondary structure conceals them from the solvent. In
the case of the ACHC and ACPC foldamers, deuteration studies showed that the
amide protons were involved in strong hydrogen bonds, as they deuterated slowly
relative to the exposed N-terminal amido protons. Interestingly, the ACHC and
ACPC foldamers show hydrogen-bond donation in opposite directions. Therefore,
the conformation-dependent dipoles and hydrogen-bond directionality of the two
helices also point in opposite directions despite the difference of single methylene
in the foldamer backbone subunit. This switch in hydrogen-bond directionality is
unknown in natural oligopeptides and will be contextualized in detail later in this
chapter with other foldamer families.
ACHC and ACPC-type foldamers and hybrid α/β-amino acid foldamers bind to
proteins, such as Bcl-2 proteins, which act as apoptosis regulators [8]. The overex-
pression of some of these proteins can result in cancer, making foldamers a very
promising area for the development of anticancer agents. They are also known to
inhibit other damaging protein–protein interactions, giving them further medical
applications [9]. This lends weight to the supposition that artificial molecules that
adopt specific and compact conformations can regulate biological processes in the
same way that nature does [8]. The preorganization in these foldamers has also
been used in other biomimetic contexts, such as the action of a designed aldolase
O
H
N
O S
N R
N N
R R' H H H
R O O O
N N N N
H H H H
R
O
R O O
N S
H N
R O H
N
H
R
Figure 15.3 Variations of residue and hydrogen-bonding group from the amide in
foldamers.
foldamer [10]. They have more recently been further extended to high-yielding aldol
macrocyclizations [11].
Since the 1990s, comprehensive research has uncovered the vast conformational
space that foldamers can adopt. Predominantly, this research has been centered
around modification of the peptides that nature uses to make secondary struc-
tures. Amides are typically used as they act as powerful dual hydrogen-bond
donors/acceptors for the formation of continuous hydrogen bond chains. Often, the
hydrogen-bond-donating and accepting ability are linked as the groups are conju-
gated, resulting in synergic hydrogen bonding. This function has been altered to
incorporate other dual hydrogen-bonding functionality, such as N,N ′ -disubstituted
(thio)ureas [12, 13] and sulfonamides (Figure 15.3) [14]. The position of the amino
acid residue has also been altered by further substitution in quaternary amino
acids, [15] as well as additional methylene groups as is the case in β- and γ-peptide
foldamers [16–18]. Substituted acrylamides have been used to make vinylogous
peptides [19, 20] as well as the introduction of other rigidifying functionality, such
as cyclopropanes [21] as well as the aforementioned cyclopentanes, cyclohexanes,
and pyrrolidines [22].
X-ray crystallography remains the apex of molecular characterization and can
be used to inform likely stable structures in solution [23]. Indeed, it is used wher-
ever possible to gain information on the precise structure that hydrogen-bonded
foldamers can adopt. By definition, foldamers are typically designed to be rigid
and therefore adopt conformations that often lend themselves to analysis by X-ray
crystallography [24]. However, this is often not possible and can sometimes be
misleading in terms of characterization in solution, where multidimensional

nuclear magnetic resonance (spectroscopy) (NMR) and computational simulation
are better analytical and predictive tools [25]. Analysis of nuclear overhauser
effect spectroscopy (NOESY) and ROESY data is particularly powerful in these
assignments – they can be used to quantify interproton distances between
hydrogen-bonded protons and/or residue protons, providing information on
hydrogen-bonding ring size, helical tightness, and helical pitch. Analysis by
correlated spectroscopy (COSY) and one-dimensional NMR not only reveals
crucial details on molecular constitution but can also be used for determining
dihedral angles between protons hydrogen-bonded and/or residue protons in the
foldamer scaffold, further divulging precise structural detail. Typically, all of these
experimentally determined parameters are then used in computational simulation,
which generates precise structures subject to restraints provided by the empirical
data. By cyclic refinement, the structures determined by computation can then
have their NMR and circular dichroism (CD) spectra predicted and compared to the
experimental data. From this, and potential comparison with crystallographic data,
extremely precise structural information about the whole length of the foldamer can
be determined and used to predict how the foldamer will orientate chemical func-
tionality and therefore inform their potential applications. Although less common,
Fourier transform infrared spectroscopy (FTIR) can be used to gain information
on hydrogen-bonding strength and geometry [26, 27]. In some instances where
appropriate chromophores are employed, UV–Vis and fluorescence spectroscopies
can be used to gain further insight into local conformations of the chromophore
and surrounding groups [28, 29].
With greater understanding of foldamers, more applications have been discov-
ered and have necessitated other methods of analysis. As such, variable-temperature
NMR (VT NMR) and CD have been used to assess the dynamics of foldamers, partic-
ularly when multiple similar conformations are populated in solution [30]. Finally,
solid-state and 15 N NMR have been used to analyze them in the solid phase [31].
Much of the characterization of foldamer families was done in the 1990s and early
2000s, and so the rules that govern how various foldamers form their secondary
structure have been well established. This has allowed foldamer research to shift
more to their applications in recent decades, such as interactions with crystals, mem-
branes, and proteins [8, 32–34].
The highlighted foldamers use attraction between hydrogen-bond donor and
acceptor groups to impose a helical conformation in the resultant oligomer. Evi-
dently, this natural phenomenon has served as inspiration for a lot of foldamer
research, but it is a significant restriction on the parameters that foldamers can
access. Many other foldamers have been made that instead use hydrogen bonding
to enforce turns in sections of an oligomer. With appropriate levels of rigidity in
the oligomer, helical conformations result. One of the largest classes of foldamers
with designed hydrogen-bond-induced turn structures is the aromatic oligoamide
foldamers developed by Huc and coworkers (Figure 15.4) [35].
In these cases, nitrogen heterocycles have been used as a hydrogen-bond acceptor
that lies coplanar with a conjugated amide acting as a hydrogen-bond donor. To
form this very favorable hydrogen bond, the conformation of the amide is restricted
O
R R
ON
H N
O
R
N N ON H O
H OMe = O H N
n
R N O N
N
O O
H
NH O OR
N
N
O
R
Figure 15.4 Helical aromatic oligoamide foldamers.
i O
BuO
H
O2N HN N N NH NO2
O Oi Bu H
i BuO O
H
O2N HN N N NH NO2
O Oi Bu H
Figure 15.5 Foldamers aggregated by π-stacking interactions.
to an approximately planar and kinked structure. Several of these continuously

bound units then result in helices with a large variety of helical diameters and
pitches [36–39].
As discussed, hydrogen bonds induce a foldamer scaffold to restrict the num-
ber and nature of stable conformations across its potential energy surface. Other
attractive interactions can similarly be used to promote compact conformations,
such as dipole–dipole interactions in aryl-linked imidazolidine-2-one oligomers [40]
and π-stacking interactions in aromatic oligoamide foldamers (Figure 15.5) [41–43].
Conversely, much as attractive interactions can be used to stabilize certain
conformations, other foldamers exploit repulsive interactions, which destabilize
certain conformations and therefore preclude their population at equilibrium
(Figure 15.6) [44–46].
For example, Aggarwal and coworkers used their iterative enantiospecific
lithiation-borylation methodology to make polypropionates with either an all-anti
relationship between the methyl groups protruding from the hydrocarbon chain, an
all-syn relationship, or alternating syn and anti-relationships (Figure 15.6a) [45]. In
the all-anti polypropionate, destabilizing syn-pentane interactions were present to
some degree in all low-energy conformations, giving rise to an effectively random
molecular shape. However, syn-pentane interactions were completely avoidable
in the all-syn and alternating syn and anti polypropionates, resulting in a strong
bias toward helical and linear conformations, respectively. Clayden and coworkers
also used a combination of dipole–dipole interactions and steric encumbrance in
xanthene-1,8-dicarboxamides to make all-anti-relationships between the individ-
ual benzamides (Figure 15.6b). When one terminus was appended with a chiral
All-anti All-syn Alternating syn and anti
NO2 NO2
O O O
B
O O O
Random conformation Helical conformation Linear conformation

(a)
N O O N N O O N N O O N N O O N
PhMgBr Ph
Me Me Me Me
N O O N O O
Ph O O Ph O OH
O >95 : 5 dr O
N O O N N O O N
(b)
Figure 15.6 Foldamers exploiting repulsive interactions: (a) Aggarwal’s polypropionates; (b) Clayden’s long-range stereocontrol of
oligo(1,8-xanthene-dicarboxamides).
Figure 15.7
β-Hairpin-mimetic trimeric
MAMBA foldamers.
O O O
H
O O O O
N H H
H
N N N
O O
O
O O
cis trans
oxazolidine group in a trimeric xanthene-1,8-dicarboxamide, this resulted in a chain

of oppositely orientated amide groups, which resulted in an induced asymmetric
environment adjacent to the axially chiral amide group at the opposite end of the
trimer that could be used for asymmetric reaction. Finally, non-hydrogen-bonded
conformational control is evident in peptoids – peptide analogs where the side
chain has instead been placed on the nitrogen atom, leading to tertiary amide
linkages with no hydrogen bonds between residues. In some cases with appropriate
substitution patterns, helical conformations can result [46, 47].
While the most common biological secondary structure to have been mimicked
by foldamers has been the helix, other secondary structures have also been studied.
Hydrogen bonds and rigidifying functionalities have been used to introduce turns
of particular angle and radius to promote conformations reminiscent of β-strands,
β-hairpins, and β-sheets [48–51]. For example, Hamilton and coworkers have
published broad research on mimicking multiple elements of secondary structure
with foldamers. Most recently, the 2,4-dialkoxy-meta-aminomethylbenzoic acid
(MAMBA) scaffold has been shown to act as a β-hairpin mimetic. Specifically, X-ray
crystallographic studies showed that MAMBA trimers can form cis conformations
whereby the turn induced by intramolecular hydrogen bonds mimics the turn of
a β-hairpin, and the substituents of the alkoxy groups fit as mimics of amino acid
residue positions in a β-hairpin (Figure 15.7) [50].
The MAMBA trimers were shown to be stable β-hairpin mimics even in strongly
polar solvents, such as dimethylsulfoxide (DMSO). While X-ray crystallography
showed predominant occupancy of the cis conformation in the solid state, it was
supposed that a trans conformation could also be occupied in solution with each of
the intramolecular salicylamide hydrogen bonds.
Nowick and coworkers investigated the use of di- and triureas derived from
ethylenediamine-linked oligomers for β-sheet mimicry [52, 53]. These oligoureas
were appended with various urea substituents and termini, using both aromatic
groups and smaller aliphatic groups; and their behavior and dynamics in solution
were studied using 1 H NMR and FTIR spectroscopy. Interestingly, when terminated
with a phenyl group, the conformational preference of arenes to lie s-trans in ureas
and amides orientated the phenyl urea’s carbonyl group toward the adjacent urea
[54]. The second urea then hydrogen bonded to the first urea through its ureido
proton, imposing same directionality on both ureas. This effect propagated through
the oligourea, meaning that the phenyl group biased the global hydrogen-bond
NC NC
O O
N N s β-Strand mimic
N R′′ N O c
H H H a
O N
O Ph f
N ≡ Peptide strand
N O f
N R′ o
H N H
l
O H N O Peptide strand
d
N O
N R HN
H
Figure 15.8 β-Sheet-mimetic oligourea foldamers.
directionality across all three ureas, and exert global conformational control on
the structure. The conformational preference imposed by the terminating phenyl
group and urea–urea hydrogen bonds rigidified the scaffold, allowing each urea to
be appended with amino acid oligomers for β-sheet mimicry (Figure 15.8) [55, 56].
As has been proven customary in foldamers, the precise structure of Nowick’s
β-sheets was gleaned by detailed 1 H NMR analysis and comparison with compu-
tational simulations. The analysis was further supported by FTIR analysis that
gave information about the hydrogen bond environments in the oligomer. In these
ethylene-bridged oligoureas, multiple conformations are populated in solution with
respect to the flexible ethylene linker. The ethylene bridge may adopt gauche or
anti-conformations, reflected in the vicinal-coupling patterns in 1 H NMR. To gain
further insight into these two conformations, an analogous diurea was prepared
with 1,2-trans-disubstituted cyclohexane in place of an ethylene linker, which
enforced a gauche conformation due to the preference of the two substituents
to orientate equatorially in the cyclohexane [53]. By comparison of the 1 H NMR
spectra and computational simulations of the cyclohexane analog with the parent
ethylene-bridged diurea, it was determined that the gauche conformer was the
minor conformer present in solution, and formed a weaker hydrogen bond between
the ureas than the anti-conformer. An additional complication of these foldamers
is the conformations of the linkers relative to each other. The anti-conformation is
the more stable due to the formation of a stronger hydrogen bond, but sequential
anti-conformations can propagate in different ways and result in different orienta-
tions of the ureas and therefore strengths of their hydrogen bonds (Figure 15.9) [55].
Interestingly, it was found by analysis of the vicinal-coupling patterns on the
ethylene bridges that the major conformer of the triurea was that where the
anti-relationships oscillated between + and − conformations, introducing destabi-
lizing syn-pentane interactions. This, however, was compensated for as the foldamer
adopted the best available geometry for forming urea–urea hydrogen bonds. From
these examples, it is pertinent to consider the most stable conformations of sin-
gle units, but also to consider the energetic discrepancies that may arise in the
analogous oligomer.
The hydrogen bond has been used extensively in foldamer chemistry to precisely
access twists and turns with wide-ranging chemical functionality. Numerous
N
CN CN CN
N
O O O
N N N N
N PhN PhN NPh
H H
N N N CN O O OH
N N N
H H H N NPh NPh NPh
N O N O N O
H H
Ph Ph Ph O O OH
N N N
N PhN Ph PhN Ph PhNPh
H H H
syn-Pentane –/+/– –/+/+ +/+/+
interaction
Major Minor Not observed
conformer conformer
Figure 15.9 Accessible conformations of ethylene-bridged triureas.
examples exist to mimic many elements of proteinogenic secondary structure,

where they have already shown great promise in interfacing with biological sys-
tems. It seems reasonable that one of the next steps for foldamers chemistry is to
mimic nature further by combining these elements of artificial secondary structure
to create sophisticated tertiary structures from foldamers. These tertiary structures
may then have potential for more advanced applications than the elements of
secondary structure alone.
15.2 Hydrogen-Bonded Dynamic Foldamers

While Gellman’s definition implies the population of a single compact conforma-
tion, the definition has been extended to foldamers that can adopt more than one sta-
ble conformation, particularly in response to stimuli. This deviates from the original
definition of a foldamer, but it can be said that if the different stable conformations
can interconvert in response to a stimulus, they are rendered dynamic foldamers [30].
Many examples of using attractive or repulsive interactions in a functionalized
oligomer have been successfully used to promote conformations that can precisely
orientate chemical functionality in a biomimetic fashion. These examples are
molecules with tailored potential energy surfaces for the control of conformer
population. Commonly, these molecules carry functionality that can respond to
stimuli, and so their conformer distributions and potential energy surfaces can be
modified to select between alternative conformations [42, 44, 57–59]. Foldamers
that adopt specific conformations by mitigation of steric encumbrance within
the oligomer are typically difficult to alter without synthetic elaboration. This
stimulus-responsive regulation of conformation is the basis for most biological
processes, and so serves as a pertinent next step for foldamers research.
Hamilton and coworkers have shown that oligomers derived from 2,6-bis
(N-imidazolidin-2-onyl)pyridines act as β-sheet mimics rigidified by dipole–dipole
interactions [48]. The group later showed that the basicity of the pyridines in the
oligomer could be exploited to introduce and delete hydrogen-bond-induced turns
using trifluoroacetic acid and trifluoromethanesulfonic acid (Figure 15.10) [60].
O O
H+ R
N N N N N N N N
H R
N O O N
R R
Dipole–dipole interactions – linear Hydrogen bonding – curved
Figure 15.10 Control of topology in 2,6-bis(N-imidazolidin-2-onyl)pyridine oligomers.
When protonated, the imidazolidinones reorientated to form hydrogen-bonding

rings with the resultant pyridinium, introducing a kink. This changed the tra-
jectory of the imidazolidinones and introduced curvature to the oligomer. In
longer oligomers, this curvature and the enantioenrichment of the imidazolidi-
nones gave rise to a P-helical conformation. Additionally, subtle changes to the
electron-richness of the pyridines in the oligomer allowed site-selective protona-
tion, enabling spatial control of curvature. Finally, the basicities of the pyridines
decreased with each subsequent protonation, allowing control of general curvature
dependent on the acidity of the medium.
One of the most remarkable ways that chemists have recently taken inspiration
from nature is by looking at how it handles information. Although restricted to
a limited pool of amino acids, sugars, and other prebiotic molecules, nature has
evolved to use those molecules for informational self-replication and developed
networks for the long-range communication of information in a variety of forms
[61]. Dynamic foldamers provide an ideal class of synthetic molecules to do the
same, provided that they can read information, communicate it through the length
of the oligomeric structure, and export it again at the opposite end. Evidently,
many stimulus-responsive dynamic foldamers interconvert between conformations
with very different shapes [42, 60]. In these cases, this has been made possible
because the stimulus interacts with multiple units in the oligomer, reinforcing
the imposed conformational preference at several points [62, 63]. This is much
more difficult when communicating information spatially, as the stimulus must
interact exclusively with a single point in the oligomer (often a terminus), and
this interaction must be sufficiently strong to effect conformational restriction
throughout the entire oligomer and compensate for whatever energetic penalty that
restriction imposes.
Most synthetic molecular communication devices operate through control of
chirality in a dynamically racemic structure – opposing enantiomers of a structure
are necessarily identical in energy, meaning that the only energetic cost to enan-
tioenrichment is entropic, and can, therefore, be easily overcome by introduction
of a single attractive interaction [30]. Clayden and coworkers have published exten-
sively on hydrogen-bonded dynamic foldamers that communicate information by
dynamic enantioenrichment of racemic helices. Oligomers of the achiral quaternary
amino acid aminoisobutyric acid (Aib) adopt 310 -helical conformations. As the helix
is made of achiral subunits, the helix becomes a rapidly equilibrating racemate,
but the introduction of a chiral influence can induce a global conformational bias
toward either the P or the M screw sense [64].
Quantification of screw-sense preference can be done by comparison to known or
modeled CD spectra of known levels of screw-sense preference [65]. However, this
only provides information about the helix as a whole, as opposed to information
about local screw-sense preference at differing distances from the chiral influence.
Local screw-sense preference can be determined and quantified as “helical excess”
(by analogy with enantiomeric excess) [66]. In a dynamically interconverting
mixture of two isomers, one nucleus will give rise to one NMR signal for each
of the isomers when they are in slow exchange. When in fast exchange, these
signals will change their chemical shift to reflect the averaged environment of the
exchanging isomers for that nucleus. In the case of a geminal dimethyl group in
an Aib foldamer, the two methyl groups are diastereotopic by virtue of the chiral
environment of the helix, and so there is an associated chemical shift difference
between the 13 C signals for each of the methyl groups at slow exchange, Δ𝛿 slow . At
fast exchange, the 13 C signals of the two methyl groups shift to reflect their averaged
environments in the helix, and now have a different chemical shift difference,
Δ𝛿 fast . This relationship holds:
Equation for Determination of Helical Excess
Δ𝛿fast [P] − [M]

= (15.1)
Δδslow [P] + [M]
The quantity Δ𝛿 fast /Δ𝛿 slow is, therefore, the helical excess for a particular pair of
dynamically interconverting nuclei. If values for Δ𝛿 fast and Δ𝛿 slow can be attained
for the two methyl groups of a geminal dimethyl group in an Aib oligomer, the
screw-sense preference can be quantified at that particular point at a particular tem-
perature (Figure 15.11).
Aib nonapeptides were synthesized and appended either with an N-terminal azide
or a Cbz-protected L-α-methylvaline unit, making the foldamers configurationally
achiral or enantioenriched helices, respectively. The fifth Aib unit in the helix was
isotopically enriched for identification of this residue amongst signals from the
other Aibs. For the configurationally achiral helix (Figure 15.11a), the exchanging
methyl groups are in the same chemical (but enantiomeric) environment, and so
their 13 C signals coalesce to a single signal at fast exchange. As Δ𝛿 fast is therefore
0, the helical excess is also 0 (it is racemic). When appended with an N-terminal
L-α-methylvaline (Figure 15.11b), the two exchanging methyl groups are in different
environments and are diastereotopic even at fast exchange, corresponding to a
helical excess of 52% at 296 K. As Δ𝛿 slow is the same for the configurationally
achiral helix and the L-α-methylvaline helix, the diastereotopicity of the methyls at
fast exchange can be attributed to the environment of the helix as opposed to any
local effect of the L-α-methylvaline. Several other L-α-methylvaline-terminated Aib
nonapeptides were made with several instances of isotopic enrichment throughout
the oligomer to quantify helical excess at different points within the same molecule.
Determination of helical excess using VT NMR in (a) a configurationally achiral helix (P helix depicted) (b) an L-α-methylvaline-terminated
22
24
26
30 ppm 28
273 K
233 K
273 K
233 K
O
BuO
O
O
t
N
BuO
O
H
O
H N
t
N
O
H
O
H N
N
O
O
H313C 13CH3
H
N
N
O
H
H313C 13CH3
O
H N
N
O
H
O
H N
N
O
O
H
N
N
O
H
H
O
H N
N
O
P enriched
H
Racemic
O
H N
N
O
(S)
H
N
N3
BnO
(b)
(a)
Figure 15.11
helix.
For example, one other L-α-methylvaline-terminated Aib nonapeptide was synthe-

sized with 100% isotopic enrichment at the third Aib residue, 50% at the sixth, and
25% at the ninth, meaning that these different Aib units could be distinguished
by relative integration and analyzed separately using the same molecule. From
this and other isotopically enriched Aib oligomers, it was found that helical excess
decays by 2.5% per residue removed from the chiral influence in acetonitrile.
A subsequent study revealed that helical excess can be bolstered by using
two L-α-methylvaline units at the N terminus, where the helical excess at the
fifth Aib residue away from the nearest L-α-methylvaline was 97% at −50 ∘ C
in THF [67]. Furthermore, interspersing the helix with another achiral amino
acid, 1-aminocyclohexanecarboxylic acid could be used to mitigate decay of
helical excess. The communication of information in the form of screw-sense
preference was then used to perform chiral functions remotely from the input of
chiral information (Figure 15.12). When the fifth Aib residue was replaced with
Z-didehydrophenylalanine, >95 : 5 diastereoselectivity was observed when this
unsaturated residue was hydrogenated to give the corresponding L-phenylalanine
residue, despite being 15 bonds away from the nearest chiral center. This concept
of asymmetric reaction as a result of communication of screw-sense preference
was then extended to a helical polypeptide with two N-terminal L-α-methylvaline
residues, the rest of the helix being made up of iterations of four Aib units and a
1-aminocyclohexanecarboxylic acid unit. An acyliminium ion was formed at the
C terminus and diastereoselective addition of an electron-rich arene nucleophile
proceeded with an 88 : 12 diastereomeric ratio despite being 61 bonds and 4 nm
away from the nearest chiral center.
As exemplified by these studies, hydrogen-bonded helices are excellent communi-
cators of information. They maintain a screw-sense preference that can be commu-
nicated from one terminus of an oligomer to the other. Clayden and coworkers then
extended this work by replacing the N-terminal L-α-methylvaline with an achiral
binding site that allowed a chiral guest to induce a screw-sense preference that could
be reversibly communicated down the length of the foldamer depending on the chi-
rality of the bound guest [68]. The helix was terminated with a basic pyridine unit,
and a variety of chiral Brønsted acids were used as binding partners (Figure 15.13).
A chiral carboxylic acid, phosphoric acid, and N-triflylphosphoramide were used
to hydrogen bond both to the pyridinium and the N-terminal amide of the helix,
with binding of the chiral guests resulting in a bias of screw-sense preference toward
one conformational enantiomer, identified using an enantio selectively labeled
Aib residue. A carboxylic acid and an N-triflylphosphoramide were identified that
induced an M helix, and a phosphoric acid that induced a P helix. These were then
combined into a four-component system whereby the handedness of the helix was
dictated by the relative strengths of the acidic ligands available for binding. Sequen-
tial addition of the carboxylic acid, phosphoric acid, and N-triflylphosphoramide
resulted in switches from racemic to M to P to M helices. After binding of the
carboxylic acid, the phosphoric acid protonates the carboxylate to generate the
phosphate, removing the carboxylic acid from the binding site and replacing it with
the phosphate, which induces the opposite screw-sense preference. Acid and base
OMe
OMe
MeO
O
H N
O
N N
O
H
H
O
H N
O
H N
O
N
H
O
H N
O
N
O
H
N
H
Screw-sense preference of a 310 helix induced by L-α-methylvaline.

O
H N
O
H N
O
N
H
O
H N
O
N
O
H
N
H
O
H N
O
N
O
H
N
H
O
H N
O
N
O
H
N
H
O
H N
Figure 15.12
BnO
Isotopically labelled
spectroscopic reporters
75%13C
25% 13C
N
H H
H O N O N O Me Me
H P helix
N O H N O H N CO2Me
O/NTf H N H N H N ppm
O O O
X*
O
Ph Ar
O O O O
(R) P (R) P OH
CbzHN (S)
O NHTf O OH M helix
O
ppm
Ph Ar
Ar = 2,4,6-(i Pr)3C6H2
(M-inducing) (P-inducing) (M-inducing)
pK a
Figure 15.13 Binding of chiral Brønsted acids with a configurationally achiral Aib helix
(P helix depicted).
could then be added to the complex mixture to select which anion would bind to
the helix through protonation of the highest pK a conjugate bases.
These examples of dynamic foldamers have shown introduction of a chiral
guest to an achiral foldamer for conformational control, foldamers can also be
controlled by modulating an already chiral influence in the foldamer. This is
particularly useful as it follows that the stimuli available for controlling dynamic
foldamers are not restricted to chiral stimuli. Inai and coworkers have investigated
Aib-derived foldamers containing both chiral and achiral units [69–71]. With an
embedded L-phenylalanine in an otherwise achiral 310 helix, the chirality of the
L-phenylalanine was sufficient to induce screw-sense preference to a P helix, as
detected by CD spectroscopy (Figure 15.14) [72]. The degree of screw-sense induc-
tion by the L-phenylalanine was increased by decreasing the temperature. When the
foldamer was appended with an N-terminal primary amine, Boc-D-proline inter-
fered with the screw-sense preference, inducing an M helix despite the influence
of the L-phenylalanine. Binding of the Boc-D-proline could also be augmented by
decreasing the temperature: by cooling a mixture of Boc-D-proline and the 310 helix,
the P helix could be converted to an M helix as determined by CD spectroscopy.
The temperature range through which the Boc-D-proline controlled the screw-sense
preference was altered by changing the equivalents and concentration of the
foldamer and ligand. The process was fully reversible and the helix went through
several cycles of screw-sense inversion by heating and cooling of the sample. Due to
the chirality of the L-phenylalanine, the two helical screw senses are diastereomers
and are different in energy, but they are sufficiently close to be reversibly biased by
an external stimulus.
H2N H H OMe
O N O N O
H O H N H N O
O
H N
N
H N H N
8 53 °C
N O * O O 36 °C
OtBu P helix 27 °C
O H 6 20 °C
OH O 10 °C
2 °C
4 –22 °C
Δε
0
Increasing Decreasing
temperature temperature
–2
ε × 10–4
1
H
N O O O O 0
O tBuO H N H N* H N
O H N O H N O
H
N O 260 280 300 320 340
H
N O
H
N O
H
N O
OMe Wavelength (nm)
H H
M helix
Figure 15.14 Regulation of screw-sense preference by temperature-dependent binding of

a chiral ligand.
With its external “coating” of geminal dimethyl groups, a 310 -helical Aib foldamer
is hydrophobic, with internal polar chains of hydrogen bonds running along the inte-
rior [73]. This shields the foldamer from the surrounding medium and makes it par-
ticularly suited to communicating information in hydrophobic environments. Aib
foldamers and their parent peptaibols insert into the membranes of vesicles, open-
ing the potential to use them to communicate information in the form of screw-sense
preference in the membrane phase. Chiral information on the outside of a mem-
brane may be communicated through the helix and into the interior of a mem-
brane, vesicle, or cell. This brings Aib foldamers and other synthetic communication
devices closer to applications where information can be communicated to and from
physically separated compartments.
To investigate the efficacy of Aib foldamers in communicating informa-
tion through artificial membranes, Clayden and coworkers appended a bis(2-
quinolylmethyl)(2-pyridylmethyl)amine (BQPA)-ligated copper complex to an Aib
helix at the N terminus (Figure 15.15) [33]. This copper- or zinc-binding site could
be used to form adducts with chiral amino acid carboxylates, such as Boc-L-proline,
inducing a chiral propeller-like conformation in the BQPA. This chiral BQPA
conformation, along with hydrogen bonding from the Boc-L-proline to N-terminal
amide protons, enforced a screw-sense preference in the Aib foldamer depending
on the chirality of the bound ligand that propagated through the helix. Importantly,
the Coulombic interaction between the ligand and the metal persists even in water,
allowing such structures to function in aqueous environments.
Isotopic labeling and analysis by NMR were not suitable for determining the
level of screw-sense preference in the membrane phase, so at the C terminus, the
foldamer was ligated to a fluorescent chiral bis(pyrenyl)ethylenediamine unit.
Different screw-sense preferences of the Aib helix induced different levels of
N OtBu
H N H N H N O
O O O H N
O
ON O H O H O H O H N
N N N
Cu2+ N N N
N O H H H
N O O O
N
Figure 15.15 Screw-sense induction by association of a chiral ligand to a Cu(II)-BQPA

binding site.
excimer (E) and monomer (M) emission: the E/M ratio of fluorescence emission
in the pyrenes was correlated with established methods in solution and used to
quantify screw-sense preference in the membrane phase [29].
The foldamer was embedded in dipalmitoylphosphatidylcholine (DOPC)-derived
vesicles, where the polarity of the Cu(II)-BQPA-binding site preferentially orientates
it external to the vesicle, poised for interaction with ligands. Boc-L-proline binds to
the copper and biases the helical conformation toward a P screw sense, inducing
a conformation at the C terminus where the pyrenes do not associate as strongly
to form an excimer, lowering the E/M ratio. This foldamer, therefore, acts as a
G-protein-coupled receptor (GPCR) mimic capable of binding a ligand that results
in a conformational change that propagates through a lipid. To emphasize this
biomimicry, the foldamer was ligated to endogenous opioid peptide neurotransmit-
ter, Leu-enkephalin, which binds to μ- and δ-opioid GPCRs. The neurotransmitter
induced an M helix in the vesicles but interacted more weakly than Boc-L-proline.
Initial binding of the Leu-enkephalin “agonist” thus induced an M helix in the
vesicle, promoting excimer formation and increasing the E/M ratio. Boc-L-proline
displaces the Leu-enkephalin as an “antagonist,” resulting in an inversion of
screw-sense preference and thus lowering the E/M ratio (Figure 15.16). This
interface with artificial membranes represented a fully functioning GPCR mimic
that could bind different chiral ligands to a Cu(II) cofactor. The chiral information
was communicated down the length of the foldamer while embedded in the vesicle
and elicited a fluorescence response over 2.6 nm away from the source of chiral
information.
This work in vesicles was extended by Clayden and coworkers to Raman
optical activity (ROA) and vibrational circular dichroism (VCD) analysis in the
DOPC-derived vesicles [74]. The polarity of chloroform is similar to that of the
hydrophobic interior of a phospholipid bilayer, so analysis in chloroform could be
extrapolated to and compared with behavior in the bilayer. Comparison between
VCD spectra in chloroform and the bilayers indicated that the relevant 310 helix
VCD signals in the amide I region were more intense in the vesicles for an
L-phenylalanine-capped Aib tetramer. This result showed that the bilayer environ-
ment either increases foldamer rigidity or increases the proportion of 310 -helical
conformers. In the enantiomeric D-phenylalanine-capped Aib tetramer, the spec-
trum was of the opposite sign, but with significant changes to the amide I region as
1.36
Agonist
1.34 Leu-enkephalin
1.32
1.30
1.28
E/M
1.26
1.24
1.22
1.20
1.18 Antagonist
Boc-L-proline
1.16
0.0 0.2 0.4 0.6
0.0 0.4 0.8 1.2 1.6 2.0
Guest concentration (mM)
Guest concentration (mM)
Figure 15.16 Reversal of screw-sense preference in DOPC-derived vesicles.

Source: Lister et al. [33]/Springer Nature.
a result of the enantiopure environment of the phospholipid bilayer. Additionally,

both enantiomers displayed several new bands in the amide I region corresponding
to some α-helical character. Crucially, comparison of the signs of the spectra
indicated that the screw-sense preference induced in the solution phase is the same
as that induced in bilayers when the chiral influence is inherent to the foldamer.
A similar Aib oligomer with an N-terminal azobenzene photoswitch and a
C-terminal geminal bis(fluoromethyl) spectroscopic reporter was used by Clayden
and coworkers to transmit information in a phospholipid bilayer, in a structure
that has functional analogy with the vision proteins or opsins. The photoswitching
process was used to modulate the extent to which the azobenzene hydrogen bonded
to an L-valine unit, which in turn affected the degree of screw-sense induction
imposed by the L-valine (Figure 15.17) [32]. The screw-sense induction from the
L-valine was detected at the C terminus using solid-state 19 F NMR analysis, where
similar principles were followed to the studies of screw-sense induction and helical
excess with 13 C labels. In its E configuration, the nitrogen of the azobenzene
group hydrogen bonds weakly to the L-valine residue, which consequently interacts
strongly with the next Aib unit and propagates a preference for an M helix along
the length of the helix in the phospholipid bilayer. Upon photoisomerization to the
Z configuration, the phenyl groups of the azobenzene deviate from coplanarity to
accommodate their steric encumbrance, increasing the sp3 -character and therefore
the basicity of the azobenzene nitrogen. This strengthens the hydrogen bond
between the azobenzene and the L-valine, drawing the L-valine away from the
remainder of the foldamer and therefore switching off its screw-sense induction.
This effect was demonstrated in the membrane phase over lengths of up to 2 nm.
The azobenzene chromophore operates as an on/off switch over several cycles for
the transmission of stereochemical information through an Aib oligomer.
Communication over long distances necessitates a “noiseless” communication
channel. In Shannon’s information and communication theory, this requires an
extended medium that can transmit information between its termini without
noise changing or attenuating the signal [75]. In molecular terms, this requires an
chemical shift (ppm)
–232 –235
19F
–229
Light-modulated screw-sense induction in phospholipid bilayers. Source: Based on De Poli et al. [75].
M helix induced
Racemic helices
Figure 15.17
Matched oligomer
O O O O O
H H H H
CbzHN (S) N N N N (S)
N (S) N N N (S) NHMe
H H H H
O O O O
Residue 1 2 3 4 5 6 7 8 9
number
Mismatched oligomer
O O O O O
H H H H
CbzHN (S) N N N N (R)
N (S) N N N (R) NHMe
H H H H
O O O O
Figure 15.18 Tendril perversions in Aib foldamers.
oligomeric structure with a well-defined single global conformation. In a mixture of

helical foldamers comprised exclusively of helical conformers, helical excess would
maintain the same value at each residue in the molecule. However, isotopic labeling
of Aib residues in 310 -helical foldamers shows that the magnitude of screw-sense
induction wanes as the Aib residues become more distant from a chiral influence
[66]. This signal decay is caused by random “tendril perversions” in the helix, where
an alternative non-helical conformation induces a communication fault and inverts
the helical screw-sense preference, attenuating the signal.
Clayden and coworkers assessed the prevalence of tendril perversions and their
molecular manifestations in Aib helices by preparing two nonapeptides, with chi-
ral influences at both the N terminus and the C terminus (Figure 15.18) [76]. One
“matched” oligomer had a pair of L-α-methylvaline units at both termini, which
formed a very stable P helix as enforced by both ends. The other “mismatched”
oligomer was terminated with a pair of L-α-methylvaline units at the N terminus,
and with a pair of D-α-methylvaline units at the C terminus, resulting in a chirality
clash. An enforced tendril perversion is required to accommodate the screw-sense
preference of both termini.
Isotopic labeling of several of the geminal dimethyl groups with various degrees of
enrichment allowed analysis of helical excess at each point in the matched and mis-
matched oligomers in solution. Predictably, the matched oligomer displayed almost
complete helical excess throughout the oligomer apart from a small dip at residue
5, which was attributed to the trace amounts of conformers exhibiting two tendril
perversions. Helical excess throughout the mismatched oligomer was lower than
that observed in the matched oligomer, with helical excess being greatest (80% P)
closest to the N terminus. This helical excess then decayed to 60% P before declining
to 30% M between residues 6 and 7. This indicated that the tendril perversion tended
to be localized closer to the C terminus than the N terminus.
When 310 helices form, they exhibit β-turns that form hydrogen bonds between
residues i and i + 3. A less common turn is the γ-turn, which instead forms a hydro-
gen bond between residues i and i + 2 and can induce tendril perversions in helical
oligopeptides (Figure 15.19).
N Me
O
H
O
N
O
H
γ-turn
N
H
O
H N
O
N
O
H
N
H
O
H N
(a) Deleted hydrogen bonds in tendril perversions and (b) γ-turns in Aib foldamers.
N
O
H
N
H
O
H N
BnO
(b)
Me
O
H N
Deleted hydrogen bond

O
N
O
H
N
H
O
H N
O
Tendril perversion
N
O
H
N
H
O
H N
O
N
O
H
N
H
O
H N
(a)
BnO
Figure 15.19
Computational simulations indicated that a γ-turn was a relatively stable con-

formation in the mismatched oligomer that was a likely platform for the tendril
perversion. In agreement with the NMR and CD data, the highest concentration
of these γ-turns was predicted to be between the center and the C terminus. Inter-
estingly, these conformations were more populated when simulated in methanol,
which is in agreement with the fact that helical excess decays in configurationally
achiral helices more quickly in polar media. This indicates that γ-turns are more
stable in polar solvents, promoting their formation and their accompanying tendril
perversion.
Evidently, Aib oligomers adopt a conformational ensemble, some of which
exhibit some non-310 -helical content. Clayden and coworkers conducted fur-
ther studies into this using Raman spectroscopy to determine the persistence of
310 -helical conformations in solution [74]. Analysis in chloroform showed that
an Aib dimer, which cannot form a 310 helix, exhibited an amido carbonyl band
at 1681 cm−1 corresponding to Aib amides in non-helical conformations. The
homologous Aib tetramer, which can form a 310 helix, showed a main band at
1688 cm−1 and a shoulder at 1667 cm−1 , the latter of which was attributed to an
amide in a hydrogen-bonded 310 -helical conformation. Longer Aib oligomers
showed a proportional increase of hydrogen-bonded 310 -helical amide bands,
with the octamer showing exclusive occupancy of a 310 -helical conformation in
chloroform. These experiments were repeated in d6 -DMSO, which was shown
to bolster the 310 -helical conformations, presumably by strengthening the helix
through hydrogen bonding to the N terminus. This effect could be attributed to the
hydrogen-bond-accepting nature of d6 -DMSO, where hydrogen-bond cooperativity
reinforces the 310 helix. While Aib-derived foldamers have been shown to function
as communication devices both in homogenous and non-homogenous states,
they come with platforms for communication faults to occur unless bolstered with
appropriate functionality [67]. This has revealed a lot of the prerequisites and factors
to consider when designing new artificial molecular communicators, and can be
used to inform new devices that can communicate information in different forms.
15.3 Reversible Hydrogen-Bond Directionality

in Dynamic Foldamers
A hydrogen bond, as well as being an adhesive interaction, is inherently direc-
tional – a hydrogen-bond donor receives electron density from a hydrogen-bond
acceptor. There is an associated dipole from the hydrogen-bond donor to the
hydrogen-bond acceptor, and the same is also true for oligomers of functional
groups containing both donors and acceptors [77]. For example, the hydrogen-bond
directionality of an oligopeptide always runs from the N terminus to the C terminus.
If a hydrogen bond can be orientated in two different directions, its directionality
becomes a form of binary information. Nature has recognized this, and uses the
patterns of hydrogen-bond donors and acceptors in the nucleobases for storing,
communicating, and processing information [78].
N O N O N O O N O N O N
H H H H H H
R R R R
(a)
H N H N H N O H N O H N H N
H H H
O O O O O S S R P O R P O
H N H N H N O H N O H N H N
N H N H O N H O N H N H N H
H H H
O O O O O S S O P R O P R
N H N H O N H O N H N H N H
(b)
Figure 15.20 Symmetry elements in dual hydrogen-bond donor/acceptor monomers: (a) in

oligoamides where the amide bonds are not part of the backbone; (b) other functional
group oligomers with hydrogen-bond directionality reversibility.
Any functional group that can act as a dual hydrogen-bond donor/acceptor

can, in principle, form oligomers with reversible hydrogen-bond directionality.
However, when the functional group is incorporated into the backbone of the
oligomer, reversibility of hydrogen-bond directionality is possible only if certain
symmetry elements are present to maintain in the oligomer. For example, due to the
constitution of the amide in oligopeptides, opposing hydrogen-bond directionalities
are typically very different in energy. Different oligomers may have different
directionalities, as was noted by Gellman and coworkers in the case of ACHC and
ACPC-derived helical foldamers, but they do not lend themselves to reversibility
[5]. If the amide does not form part of the foldamer backbone, then hydrogen-bond
directionality reversibility becomes possible again (Figure 15.20a). If subunits are
used that contain other elements of symmetry, then hydrogen-bond directionality
reversibility is possible even when the functional group is part of the foldamer
backbone (Figure 15.20b). Numerous dual hydrogen-bond donors/acceptors, such
as alcohols, ureas, phosphates, imides and sulfamides, amongst many others,
can display opposing and enantiomeric hydrogen-bond directionalities in their
oligomers. As these directionalities are enantiomeric, they are isoenergetic and
therefore have the potential to be reversible and switchable, and could even be used
for the communication of information.
Guichard and coworkers have worked extensively on urea-derived helical
foldamers. These oligourea foldamers have been successfully employed for enan-
tioselective catalysis and anion binding and as bactericides [13, 79, 80]. Many
of these foldamers incorporate point chirality into the linkages connecting the
ureas, promoting one screw-sense preference and also rendering the opposing
hydrogen-bond directionalities diastereomeric. In collaboration with Clayden
t t
Bu
H N H N H N Bu N H N H N H tBu
t
Bu
O O O O O O
H N H N H N N H N H N H N H N H N H
H N O H N O H N O O O O
N H N H N H
H N H N H N
N terminus P Helix C terminus C terminus M helix N terminus
Figure 15.21 Hydrogen-bond directionality reversal in helical oligoureas (212/14 helix

depicted).
and coworkers, analogous foldamers were made with meso linkers. The resultant
helices were, therefore, conformationally racemic and the opposing hydrogen-bond
directionalities were enantiomeric (Figure 15.21) [81].
Specifically, the oligoureas were constructed from oligomers of meso cyclohexane-
1,2-diamine, which formed 2.512/14 -helical conformations whereby two ureido pro-
tons hydrogen bond with the carbonyl of the urea two units preceding it. This
pattern resulted in two enantiomeric helices, both of which had two ureas, each
with a pair of pendant ureido protons at the N terminus, and a pendant ureido
carbonyl at the C terminus. Due to the incorporation of the chiral centers in the
meso cyclohexane-1,2-diamine linkers (despite being meso subunits), the urea that
is set as the N terminus enforces a certain screw-sense preference. When the urea at
the opposite end sets itself as the N terminus, the opposite screw-sense preference
is induced. Therefore, any switch in hydrogen-bond directionality of these helices
is associated with a concomitant switch in screw-sense preference.
1 H NMR analysis of the oligomer in various solvents revealed different signals
corresponding to the two tert-butyl groups, indicating that the helix was in slow
exchange between its two equally populated screw-sense preferences on the NMR
timescale, placing one tert-butyl group at the hydrogen-bond N terminus and the
other at the C terminus. VT NMR and Eyring analysis (described in detail later) indi-
cated that, depending on the solvent, the barrier to screw-sense inversion was around
ΔG‡ 298K = 70.0 kJ mol−1 . Adding the tetrabutylammonium salt of Boc-D-proline to
the sample (as a hydrogen-bond acceptor) induced preference for a P helix, provid-
ing the first example of reversible and controllable hydrogen-bond directionality in
hydrogen-bonded foldamers.
Clayden and coworkers extended this work to make molecular “spring torsion bal-
ances” by desymmetrizing the helical oligourea (Figure 15.22) [82]. Thioureas are
much stronger hydrogen-bond donors than their analogous ureas due to the greater
diffusivity of sulfur’s orbitals [83]. Therefore, when one of the terminal ureas was
replaced with a thiourea, the equilibrium shifted to ensure the thiourea acted as
a hydrogen-bond donor, placing the thiourea at the C terminus and enforcing a P
helix. Conversely, as carbamates can only act as single hydrogen-bond donors, they
are much weaker donors than ureas and so when they terminated the helix, they
were placed at the N terminus to ensure they acted as a hydrogen-bond acceptor.
W O O O O O Z
X NH HN NH HN NH HN NH HN NH HN NH HN Y
Ratio P:M
O S
t H N H N H N >95 : 5
BuO NH HN NHBn
H N O H N O H N O
O X H N H N H N O
W O O Z
tBuO 90 : 10
NH H N H N H N Y HN NHBn
O O
N terminus P Helix C terminus
tBuO 65 : 35
NH HN OAllyl
O O
tBuHN 60 : 40
NH HN NHBn
O O
tBuHN 40 : 60
NH N H N H N H HN OAllyl
O O O
O N H N H N H N H N H N H Y O
W O O Z
40 : 60
BnHN NH X N H N H N H HN OAllyl
S O
C terminus M Helix N terminus 10 : 90
BnHN NH HN NHBn
150
100 0 equiv
θ (103 deg cm2 dmol–1)
50
O equiv TBAP O
tBuO NH 0 HN NHBn
–50
–100 30.5 equiv
–150
190 200 210 220 230
Wavelength (nm)
Figure 15.22 Helical oligoureas as a molecular spring torsion balance.
When a carbamate and thiourea were placed at opposite ends of the helix, 1 H NMR
analysis revealed that the oligomer exclusively adopted a P helical conformation.
When termini of similar hydrogen bond-donating/accepting ability were used,
the helix adopted P- and M-helical conformations in approximately equal measure.
When thiourea terminated the helix, the conformational distribution was biased
toward the thiourea acting as a hydrogen-bond donor. The thiourea then hydro-
gen bonded to the best available hydrogen-bond acceptor, that is, the urea two units
away in the oligomer. However, when presented with an alternative acceptor, such as
that of tetrabutylammonium phosphate, the thiourea would reverse its directionality
to hydrogen bond intermolecularly with the phosphate ligand. Therefore, as tetra-
butylammonium phosphate was titrated into the thiourea-terminated oligomer, the
hydrogen-bond directionality reversal of the oligomer, along with its screw-sense
R R R R
N N N N N N
H H H H H H
N O N O N O O N O N O N
Ar Ar Ar Ar Ar Ar
Figure 15.23 Hydrogen-bond directionality reversibility in ethylene-bridged oligourea

foldamers.
reversal, resulted in an inversion of the CD spectrum depending on the degree of

phosphate present.
These examples of hydrogen-bonded dynamic foldamers have each required
the use of chirality as a platform for the communication of information. This
complicates the use of interfacing these foldamers with chiral environments and
often restricts the inputs and outputs of information to those that can exhibit
chirality. To overcome this limitation, Clayden and coworkers repurposed the
ethylene-bridged oligourea β-sheet mimics used by Nowick to make hydrogen-bond
directionality-reversible dynamic foldamers that were free of the constraints of
chirality and could, therefore, take information in the form of hydrogen-bond
directionality alone and communicate it down the length of a foldamer [84, 85].
When the phenyl group was removed from the foldamer, the first urea was no longer
conformationally restricted and could, therefore, access both of its hydrogen-bond
directionalities (Figure 15.23).
VT NMR, line shape, and Eyring analysis allowed the quantification of the rate of
hydrogen-bond directionality reversal in oligoureas with varying numbers of ureas.
In a dynamic system where two interconverting nuclei A and B are enantiotopic
and necessarily isoenergetic, their rate of interconversion can be determined by use
of the Eyring–Polanyi equation (Eq. (15.1)) [86].
The Eyring–Polanyi Equation
kB T −ΔG‡
k= e RT (15.2)
h
Substituting the Gibbs free energy equation into the Eyring–Polanyi equation, and
subsequent rearrangement gives the Eyring–Polanyi equation in its linear form
(Eq. (15.2)).
The Linear Form of the Eyring–Polanyi Equation

k −ΔH ‡ 1 k ΔS‡
=ln + ln B + (15.3)
T R T h R
From this, plotting ln(k/T) against (1/T) produces a straight line whose gradient is
equal to (−ΔH ‡ /R) and intercept is equal to (ln(kB /h) + ΔS‡ /R). In a dynamic system
where two interconverting nuclei A and B are diastereotopic and therefore of differ-
ent energy, the rate constants for the forward reaction and backward reaction will
Ph
CF3
Ph Ph Ph
N N N N N N
H H H H H H H
O N O N O N O N O N O N O N CF3
S Ar Ar Ar Ar Ar Ar H
N S
F3C CF3
Figure 15.24 Hydrogen-bonding additives associating with ethylene-bridged oligourea

foldamers.
be different. If K and the rate are known, then the rate can be partitioned into the
different rate constants of the forward and backward reactions. If rate constants can
be determined for different temperatures, then the ΔH ‡ , ΔS‡ , and therefore ΔG‡ can
be determined for the dynamic process in question at any temperature. These data
can be used to determine the values of other useful parameters in dynamic processes,
such as half-life and rate constants at unmeasured temperatures, and provide insight
into associative vs. dissociative exchange mechanisms. In the ethylene-bridged
oligoureas, a dibenzyl monourea had a barrier to rotation of the (Bn)2 N–C urea bond
and therefore reversal of directionality of ΔG‡ 298 K (CD2 Cl2 ) = 48.9 kJ mol−1 . Upon
introducing a second urea and an associated hydrogen bond between them, the
barrier rose to ΔG‡ 298 K (CD2 Cl2 ) = 59.4 kJ mol−1 , indicating the energetic penalty
associated with breaking the inter-urea hydrogen bond. Interestingly, the addition of
further ureas and therefore further hydrogen bonds did not change the barrier, show-
ing that during global hydrogen-bond directionality reversal, the hydrogen bonds
break and reform one at a time as opposed to all together. The hydrogen-bonded
chain was stable in the presence of strong hydrogen-bond additives, as
neither Schreiner’s thiourea nor DMSO was sufficiently strong hydrogen-bond
donors/acceptors to break the inter-urea hydrogen bonds, instead complementing
the available hydrogen-bonding groups (Figure 15.24). These additives did, however,
alter the kinetics of the hydrogen-bond directionality-reversal process.
When the foldamer was terminated with a strong hydrogen bond-accepting
2-pyridyl group, a strong hydrogen bond was formed with the first urea, resulting
in global control of hydrogen-bond directionality. The basicity of the pyridine
was then exploited by treating it with tetrafluoroboric acid, which protonated the
pyridine and converted it into a hydrogen-bond donor (Figure 15.25). This formed
a new hydrogen-bonded ring with the adjacent urea where its hydrogen-bond
directionality was switched, and this propagated down the length of the tetraurea
foldamer. This switch in control could be inferred by the shifting of the ureido
protons, as well as introduction of an nuclear overhauser effect (NOE) correlation
between terminating benzylic methylene and the adjacent ureido proton when it
was switched into proximity. The pyridine retained global control of hydrogen-bond
directionality even when the ethylene-bridged oligourea was extended to seven
urea units.
The fluorescence emission of the weakly hydrogen bond-accepting dithiomale-
imide fluorophore is modulated when the fluorophore is bound to an intramolecular
pH-modulated hydrogen-bond directionality switching in ethylene-bridged oligoureas. Source: Morris et al. [85]/Elsevier/CC BY 4.0.
6.0
6.5
7.0
7.5
8.0
8.5
9.0
δ (ppm)
9.5
10.0
10.5
11.0
11.5
12.0
Pyridinium
12.5
Pyridine
Pyridine
13.0
NOE
H
H
H
H
Ph
Ph
Ar
O
N
N
N
Ar
O
N
H
Ar
O
N
HBF4
N
N
Ar
O
N
H
H
Et3N
Ar
O
N
N
N
Ar
O
N
H
Ar
O
N
N
N
Ar
O
N
Figure 15.25
H
H
N
H
hydrogen-bond donor (Figure 15.26). When the basic pyridine was protonated with
tetrafluoroboric acid, and deprotonated with triethylamine over several cycles,
the reversal of the hydrogen-bonded chain was communicated down the length
of the foldamer and resulted in different fluorescent responses remote from the
pyridine.
These examples demonstrate that communication devices in hydrogen-bonded
dynamic foldamers do not require chirality to transmit information. Hydrogen-
bonded foldamers can be used for many more applications free of this restriction,
such as interaction with polar molecules or ions, which may open up avenues in
fields such as remote activation of catalysis or uptake of a guest. By incorporating
chirality in tandem with reversible hydrogen-bond directionality, artificial commu-
nication devices that can handle several types of information simultaneously may
be envisaged.
15.4 Cyclic Hydrogen Bond Chains

Hydrogen bond chains both in nature and foldamers chemistry most commonly
propagate as helices, sheets, or strands. Much rarer are hydrogen bond chains that
form closed loops. While the uses of these cyclic hydrogen bond chains are more
restricted than more conventional elements of secondary structure, they are still
responsible for some functions in nature. Cyclic-hydrogen-bonding networks can
also lead to more sophisticated structures if several of them stack on top of each
other. For example, guanine-rich DNA strands can aggregate and orientate four
guanines in a coplanar orientation for them to form a cyclic-hydrogen-bonding
network by Hoogsteen base-pairing [87]. Often, this guanine tetrad secondary
structure is further stabilized by the incorporation of alkali metal cations, such as
potassium into the cavity in the center of the cyclic-hydrogen-bonding network.
Several stacked guanine tetrads form tertiary structures called G-quadruplexes,
which are responsible for regulating the transcription of DNA into RNA, and for
regulating the length of telomeric DNA by inhibiting the action of DNA poly-
merases. Interestingly, there is uniform hydrogen-bond directionality in guanine
tetrads, which may serve to further stabilize the secondary and resultant tertiary
DNA structures. In artificial systems, cyclic hydrogen-bond networks using dual
hydrogen-bond donors/acceptors are much more common, a simple example of
which is the campestarenes (Figure 15.27) [88].
Campestarenes are symmetrical Schiff-base macrocycles with interesting tau-
tomeric behavior. They are formed from the polycondensation of 2-amino-6-
formylphenols to form the corresponding macrocyclic polyimine [89]. This poly-
merization occurs spontaneously, therefore precursors are required to synthesize
these structures with selectivity. MacLachlan and coworkers showed that, in the
case of 2-nitro-6-formylphenols, reduction of a nitro group to acquire the amine
resulted in exclusive formation of the C5h campestarene. The campestarene could
exist in two different tautomeric forms, a keto-enamine form and an enol-imine
form (Figure 15.27). In different solvents, the energy difference between the two
O Equiv Equiv
1.0
N N N N N N N N N Et3N HBF4
SBu
H H H H H H H
N O N O N O N O N O N O N O O 0.8 9.0 8.0
Ar Ar Ar Ar Ar Ar Ar SBu
0.6 6.5 5.5
Et3N HBF4 4.0 3.0

0.4
1.5 1.0
O 0.2
N N N N N N N N N
0.0 0.0
SBu
H H H H H H H H 0.0
O N O N O N O N O N O N O N O 300 340 380 420 460 λ (nm)
Ar Ar Ar Ar Ar Ar Ar SBu
Figure 15.26 Remote fluorescence response from communication of protonation state in ethylene-bridged oligoureas. Source: Morris et al.
[85]/Elsevier/CC BY 4.0.
H H
N N
H O H H O H
N O N
H O H H H
N H O N H O
O O H
H H
H O N O N
N N
H H
H H
Figure 15.27 Tautomerism in campestarenes.
tautomeric forms differs greatly, allowing solvent-dependent modulation of the

equilibrium. Crucially, 1 H NMR, computational and X-ray crystallographic evidence
suggested that the phenolic protons are in identical chemical environments and
have very strong hydrogen bonding. To satisfy this criterion, in the enol-imine form,
all of the hydroxyl groups have to be aligned in either a clockwise or anticlockwise
fashion. These structures are degenerate due to their horizontal plane of symmetry,
but they would be enantiomers of each other if two different groups protruded from
either side of the campestarene plane.
Aida and coworkers used a cyclic-hydrogen-bonding network to slow the
bowl-to-bowl inversion of corannulenes as a means of accessing the corannulene
family’s unique spectroscopic properties and functions (Figure 15.28). The group
functionalized the corannulene with both enantiopure and achiral amide-appended
thioalkyl residues [90].
The hydrogen-bonding network gave rise to some unexpected chiroptical prop-
erties. When achiral side chains were used on the corannulene, all of the amides
hydrogen bonded in an intramolecular fashion to form a cyclic-hydrogen-bonding
network on top of the corannulene. Due to the helical chirality of the corannulene,
which arose from its bowl shape and substitution pattern, the two hydrogen-bond
directionalities were diastereomeric, and therefore one was favored over the other.
As desired, the strength of the cyclic-hydrogen-bonding network slowed down the
bowl-to-bowl inversion of the corannulene significantly. The corannulene was then
examined in enantiomerically pure (R)-limonene as a chiral solvent. Due to the
chiral environment the (R)-limonene imposed on the corannulene, a clear Cotton
effect was observed in the absorbance range of the corannulene motif. The presence
of the chiral solvent induced a particular handedness of the corannulene, which
in turn induced a particular hydrogen-bond directionality in the amides in the
cyclic-hydrogen-bonding network. The same effect was observed upon introduction
of chiral residues on the corannulene. Analogs were synthesized with (R) and
(S)-2,6-dimethyloctane residues. Studies of these in an achiral solvent, methylcyclo-
hexane resulted in opposite Cotton effects. This indicated that the chiral residues
induced preferential directionality both in the cyclic-hydrogen-bonding network
and helical chirality in the corannulene. This effect was further exemplified by addi-
tion of ethanol to the mixture. The ethanol disrupted the cyclic-hydrogen-bonding
H nHept
nHept N
N O
H O
S
S
S H
O N
n
S Hept
O
n
Hept N
S
H
O
N H
nHept
Figure 15.28 Hydrogen-bond directionality in amide-appended corannulenes.
network by competitive-hydrogen-bonding interactions, which increased the rate

of bowl-to-bowl inversion, and thus no Cotton effect was observed. This supported
the notion that the cyclic-hydrogen-bonding network was responsible for the
corannulene’s chiroptical properties.
Cyclic-hydrogen-bonding networks have found use in forming cages for
host–guest chemistry. Diederich and coworkers developed methodology to syn-
thesize enantiopure diethynylallenes [91], which was used to create tertiary
alcohol-substituted quaternary ethynylallenes that were coupled to a resor-
cin[4]arene scaffold to give an alleno-acetylenic cage (AAC) (Figure 15.29) [92].
Due to the axial chirality of the allenes in the structure, the AAC was chiral.
When recrystallized from a small, polar solvent, such as THF or MeCN, the alcohols
in the AAC hydrogen-bonded with the solvent, keeping the structure in an “open”
form (Figure 15.30). However, when recrystallized from a bulky, hydrophobic
solvent, such as benzene or cycloheptane, the alcohols could not form hydrogen
bonds with the solvent, and they, therefore, aggregated at the top of the AAC into
a cyclic-hydrogen-bonding network to intramolecularly form hydrogen bond with
each other. This stabilized a “closed” form.
These effects also dominated in solution as evidenced by 1 H NMR, FTIR, and
2D ROESY data recorded in a variety of solvents. The aggregation of the alcohols
at the top of the AAC in the “closed” form also gave rise to interesting chiroptical
Ar
H
O O
H H O H
O
Ar
H O
O H
=
H O
O H
Ar
O OO O O OO O
Ar
n nHex nHex
Hex n
Hex
Figure 15.29 Cyclic hydrogen bond networks in alleno-acetylenic cages.

Source: Gropp et al. [92]/John Wiley & Sons.
Figure 15.30 The “open” and “closed” form of the AAC. Source: Gropp et al. [92]/
John Wiley & Sons.
properties. By comparison to a monomeric system without the resorcin[4]arene

scaffold, the AAC exhibited approximately the same molar extinction coefficient
per ethynylacetylene, but the Cotton effect of the AAC was over 100 times greater
than that of the monomer. This showed that the aggregation of the alcohols into
a cyclic-hydrogen-bonding network provided synergy in enforcing the chirality
expression of the structure. Interestingly, the enantiomer of the ethynylallene used in
the AAC dictated the hydrogen-bond directionality of the cyclic-hydrogen-bonding
network. This effect was further exemplified by permethylation of the alcohols,
which destroyed the cyclic-hydrogen-bonding network. Permethylation of the AAC
ensured that the “closed” conformation of the AAC was unstable and resulted in an
additive Cotton effect as opposed to a synergic one.
The AAC can accommodate a number of small, hydrophobic molecules and com-
plement their size by orientating the methyl groups of the tertiary alcohol in or out of
the cavity in the “closed” structure. The structure also found use in enantioselective
Clockwise Hydrogen bond Anticlockwise

donation
R R
N H O R R O H N
O N N O
H H H H
N N O O N N
R O H N N H O R
R R
N N
Figure 15.31 Chirality arising from a dissymmetric cyclic hydrogen bond network in
4-pyrrolidinopyridines.
complexation of trans-1,2-dimethylcyclohexane. Recrystallizing the P enan-

tiomer of the AAC from a racemic mixture of trans-1,2-dimethylcyclohexane
encapsulated exclusively R,R-trans-1,2-dimethylcyclohexane. The opposite was
observed when recrystallized from the M enantiomer of the AAC. The same was
repeated with cis-1,2-dimethylcyclohexane, where no enantioselective binding
was observed. However, recrystallization did afford two equally populated occu-
pancies of the enantiomeric conformers. This served as the first experimental
observation of the two enantiomers of cis-1,2-dimethylcyclohexane, which nor-
mally racemise too quickly to be individually observed. This study showed that
cyclic-hydrogen-bonding networks can give rise to interesting solvent-dependent
synergic chiroptical properties and host–guest chemistry.
These examples incorporate a cyclic arrangement of hydrogen bonds whose
directionality is dynamically enriched by chirality elsewhere in the molecule.
However, the 4-pyrrolidinopyridines (PPY) synthesized by Kawabata and cowork-
ers served as the first example of a compound that is chiral by virtue of its
cyclic-hydrogen-bonding network alone, or “cyclochiral” (Figure 15.31) [93].
In this structure, the four amides on the pyrrolidine lie approximately coplanar
to form a series of alternating six- and eight-membered hydrogen-bonding rings in a
cyclic-hydrogen-bonding network. Two different sets of 1 H NMR signals correspond-
ing to the amide protons are observed at room temperature corresponding to amide
protons in either six- or eight-membered hydrogen-bonding rings, indicating that
the enantiomeric hydrogen-bond directionalities in the cyclic hydrogen-bond net-
work are in slow exchange on the NMR timescale. Heating the compound caused
the signals corresponding to the amide protons to coalesce into one signal, with a
barrier to inversion of ΔG‡ 273 K (C6 D4 Cl2 ) = 84 kJ mol−1 , where R = Ph. This high
barrier to racemization meant the enantiomers of the PPY could be separated by
preparative chiral HPLC at 0 ∘ C.
PPYs with a wide variety of amide substituents were synthesized, all show-
ing similarly large barriers to racemization, with a maximum of ΔG‡ 273 K
(C6 D4 Cl2 ) = 100 kJ mol−1 when R = dicyclohexylmethyl. The amides were
then synthesized with chiral substituents so that the conformers with opposite
hydrogen-bond directionalities would be diastereomeric. The chiral group

introduced was an (R)-1-(1-naphthyl)ethyl group and in this case, the compound
was present as a 73 : 27 mixture of diastereomers. The pyridine of the PPY was
then protonated using HCl, which interestingly resulted in an inversion of the
equilibrium to 32 : 68. The cyclochiral PPY functions as a chiral nucleophilic
catalyst in an acylative kinetic resolution of a chiral but racemic secondary alcohol.
With a 10 mol% loading of the PPY, enantioselective acylation gave a modest 59%
ee with no racemization of the catalyst.
Hydrogen bonding is a stimulus-responsive and directional interaction with inher-
ent compatibility with biology, and it is no wonder that it has featured so extensively
in artificial replications of secondary structures, communication devices, and cages.
With increasing understanding of how hydrogen bond chains can be used to mod-
ulate processes in biomimicry, it is reasonable to assume that they will form a key
component of organic nanotechnology in the future.
References
1 Harding, S.E., Channell, G., and Phillips-Jones, M.K. (2018). The discovery of
hydrogen bonds in DNA and a re-evaluation of the 1948 Creeth two-chain model
for its structure. Biochem. Soc. Trans. 46 (5): 1171–1182.
2 Frey, P.A. (2004). Encyclopedia of Biological Chemistry, 594–598. San Diego:
Elsevier.
3 Godbey, W.T. (2014). An Introduction to Biotechnology, 9–33. Woodhead
Publishing.
4 Dominelli-Whiteley, N., Brown, J.J., Muchowska, K.B. et al. (2017). Strong
short-range cooperativity in hydrogen-bond chains. Angew. Chem. Int. Ed. 56
(26): 7568–7662.
5 Gellman, S.H. (1998). Foldamers: a manifesto. Acc. Chem. Res. 31 (4): 173–180.
6 Appella, D.H., Christianson, L.A., Karle, I.L. et al. (1996). β-Peptide foldamers:
robust helix formation in a new family of β-amino acid oligomers. J. Am. Chem.
Soc. 118 (51): 13071–13072.
7 Engen, J.R. (2009). Analysis of protein conformation and dynamics by hydro-
gen/deuterium exchange MS. Anal. Chem. 81 (19): 7870–7875.
8 Lee, E.F., Sadowsky, J.D., Smith, B.J. et al. (2009). High-resolution structural
characterization of a helical α/β-peptide foldamer bound to the anti-apoptotic
protein Bcl-x L. Angew. Chem. Int. Ed. 48 (24): 4318–4322.
9 Checco, J.W., Lee, E.F., Evangelista, M. et al. (2015). α/β-Peptide foldamers tar-
geting intracellular protein–protein interactions with activity in living cells.
J. Am. Chem. Soc. 137 (35): 11365–11375.
10 Müller, M.M., Windsor, M.A., Pomerantz, W.C. et al. (2009). A rationally
designed aldolase foldamer. Angew. Chem. Int. Ed. 48 (5): 922–925.
11 Girvin, Z.C., Andrews, M.K., Liu, X., and Gellman, S.H. (2019).
Foldamer-templated catalysis of macrocycle formation. Science 366 (6472):
1528–1531.
References 515
12 Semetey, V., Rognan, D., Hemmerlin, C. et al. (2002). Stable helical secondary
structure in short-chain N,N′ -linked oligoureas bearing proteinogenic side chains.
Angew. Chem. Int. Ed. 41 (11): 1893–1895.
13 Antunes, S., Corre, J.-P., Mikaty, G. et al. (2017). Effect of replacing main-chain
ureas with thiourea and guanidinium surrogates on the bactericidal activity of
membrane active oligourea foldamers. Bioorg. Med. Chem. 25 (16): 4245–4252.
14 Vertesaljai, P., Biswas, S., Lebedyeva, I. et al. (2014). Synthesis of
taurine-containing peptides, sulfonopeptides, and N- and O-conjugates. J. Org.
Chem. 79 (6): 2688–2693.
15 Leonard, D.J., Zieleniewski, F., Wellhöfer, I. et al. (2021). Scalable synthesis and
coupling of quaternary α-arylated amino acids: α-aryl substituents are tolerated
in α-helical peptides. Chem. Sci. 12 (27): 9386–9390.
16 Kolesinska, B., Eyer, K., Robinson, T. et al. (2015). Interaction of β3/β2-peptides,
consisting of val-ala-leu segments, with popc giant unilamellar vesicles (GUVs)
and white blood cancer cells (U937) – a new type of cell-penetrating peptides,
and a surprising chain-length dependence of their vesicle- and cell-lysing activity.
Chem. Biodivers. 12 (5): 697–732.
17 Seebach, D., Overhand, M., Kühnle, F.N.M. et al. (1996). β-Peptides: synthesis
by Arndt-Eistert homologation with concomitant peptide coupling. Structure
determination by NMR and CD spectroscopy and by X-ray crystallography. Heli-
cal secondary structure of a β-hexapeptide in solution and its stability towards
pepsin. Helv. Chim. Acta 79 (4): 913–941.
18 Seebach, D., Beck, A.K., and Bierbaum, D.J. (2004). The world of β- and
γ-peptides comprised of homologated proteinogenic amino acids and other
components. Chem. Biodivers. 1 (8): 1111–1239.
19 Hagihara, M., Anthony, N.J., Stout, T.J. et al. (1992). Vinylogous polypeptides: an
alternative peptide backbone. J. Am. Chem. Soc. 114 (16): 6568–6570.
20 Legrand, B. and Maillard, L.T. (2021). α,β-Unsaturated γ-peptide foldamers.
Chempluschem 86 (4): 629–645.
21 De Pol, S., Zorn, C., Klein, C.D. et al. (2004). Surprisingly stable helical confor-
mations in α/β-peptides by incorporation of cis-β-aminocyclopropane carboxylic
acids. Angew. Chem. Int. Ed. 43 (4): 511–514.
22 Hayen, A., Schmitt, M.A., Ngassa, F.N. et al. (2004). Two helical conforma-
tions from a single foldamer backbone: “split personality” in short α/β-peptides.
Angew. Chem. Int. Ed. 43 (4): 505–510.
23 Urushibara, K., Ferrand, Y., Liu, Z. et al. (2021). Accessing improbable foldamer
shapes with strained macrocycles. Chem. Eur. J. 27 (43): 11205–11215.
24 Meunier, A., Singleton, M.L., Kauffmann, B. et al. (2020). Aromatic foldamers
as scaffolds for metal second coordination sphere design. Chem. Sci. 11 (44):
12178–12186.
25 Huc, I. and Hecht, S. (2007). Foldamers: Structure, Properties, and Applications.
Weinheim: Wiley.
26 Blodgett, K.N., Zhu, X., Walsh, P.S. et al. (2018). Conformer-specific
and diastereomer-specific spectroscopy of αβα synthetic foldamers:
Ac–ala−βACHC–ala–NHBn. J. Phys. Chem. A 122 (14): 3697–3710.
27 Bonnel, C., Legrand, B., Bantignies, J.-L. et al. (2016). FT-IR and NMR struc-
tural markers for thiazole-based γ-peptide foldamers. Org. Biomol. Chem. 14 (37):
8664–8669.
28 Jeon, H.-G., Jang, H.B., Kang, P. et al. (2016). Helical aromatic foldamers func-
tioning as a fluorescence turn-on probe for anions. Org. Lett. 18 (17): 4404–4407.
29 Lister, F.G.A., Eccles, N., Pike, S.J. et al. (2018). Bis-pyrene probes of foldamer
conformation in solution and in phospholipid bilayers. Chem. Sci. 9 (33):
6860–6870.
30 Le Bailly, B.A.F.F. and Clayden, J. (2016). Dynamic foldamer chemistry. Chem.
Commun. 52 (27): 4852–4863.
31 Aisenbrey, C., Pendem, N., Guichard, G., and Bechinger, B. (2012). Solid state
NMR studies of oligourea foldamers: interaction of 15N-labelled amphiphilic
helices with oriented lipid membranes. Org. Biomol. Chem. 10 (7): 1440–1447.
32 De Poli, M., Zawodny, W., Quinonero, O. et al. (2016). Conformational photo-
switching of a synthetic peptide foldamer bound within a phospholipid bilayer.
Science 352 (6285): 575–580.
33 Lister, F.G.A.A., Le Bailly, B.A.F.L., Webb, S.J., and Clayden, J. (2017).
Ligand-modulated conformational switching in a fully synthetic
membrane-bound receptor. Nat. Chem. 9 (5): 420–425.
34 Rinaldi, S. (2020). The diverse world of foldamers: endless possibilities of
self-assembly. Molecules 25 (14): 3276.
35 Huc, I. (2004). Aromatic oligoamide foldamers. Eur. J. Org. Chem. 2004 (1):
17–29.
36 Kolomiets, E., Berl, V., and Lehn, J.-M. (2007). Chirality induction and
protonation-induced molecular motions in helical molecular strands. Chem.
Eur. J. 13 (19): 5466–5479.
37 Garric, J., Léger, J.-M., and Huc, I. (2005). Molecular apple peels. Angew. Chem.
Int. Ed. 44 (13): 1954–1958.
38 Singleton, M.L., Pirotte, G., Kauffmann, B. et al. (2014). Increasing the size of an
aromatic helical foldamer cavity by strand intercalation. Angew. Chem. Int. Ed. 53
(48): 13140–13144.
39 Merlet, E., Moreno, K., Tron, A. et al. (2019). Aromatic oligoamide foldamers
as versatile scaffolds for induced circularly polarized luminescence at adjustable
wavelengths. Chem. Commun. 55 (32): 9825–9828.
40 Lockhart, Z. and Knipe, P.C. (2018). Conformational analysis conformationally
programmable chiral foldamers with compact and extended domains controlled
by monomer structure. Angew. Chem. Int. Ed. 57 (28): 8478–8482.
41 Sebaoun, L., Maurizot, V., Granier, T. et al. (2014). Aromatic oligoamide β-sheet
foldamers. J. Am. Chem. Soc. 136 (5): 2168–2174.
42 Gole, B., Kauffmann, B., Maurizot, V. et al. (2019). Light-controlled conforma-
tional switch of an aromatic oligoamide foldamer. Angew. Chem. Int. Ed. 131
(24): 8147–8151.
43 Lamouroux, A., Sebaoun, L., Wicher, B. et al. (2017). Controlling dipole orien-
tation through curvature: aromatic foldamer bent β-sheets and helix-sheet-helix
architectures. J. Am. Chem. Soc. 139 (41): 14668–14675.
References 517
44 Clayden, J., Lund, A., Vallverdú, L., and Helltwell, M. (2004). Ultra-remote stere-
ocontrol by conformational communication of information along a carbon chain.
Nature 431 (7011): 966–971.
45 Burns, M., Essafi, S., Bame, J.R. et al. (2014). Assembly-line synthesis of organic
molecules with tailored shapes. Nature 513 (7517): 183–188.
46 Kirshenbaum, K., Barron, A.E., Goldsmith, R.A. et al. (1998). Sequence-specific
polypeptoids: a diverse family of heteropolymers with stable secondary structure.
Proc. Natl. Acad. Sci. 95 (8): 4303–4308.
47 Chongsiriwatana, N.P., Patch, J.A., Czyzewski, A.M. et al. (2008). Peptoids that
mimic the structure, function, and mechanism of helical antimicrobial peptides.
Proc. Natl. Acad. Sci. 105 (8): 2794–2799.
48 German, E.A., Ross, J.E., Knipe, P.C. et al. (2015). β-Strand mimetic foldamers
rigidified through dipolar repulsion. Angew. Chem. Int. Ed. 54 (9): 2649–2652.
49 Wyrembak, P.N. and Hamilton, A.D. (2009). Alkyne-linked
2,2-disubstituted-indolin-3-one oligomers as extended β-strand mimetics. J. Am.
Chem. Soc. 131 (13): 4566–4567.
50 Meisel, J.W., Hu, C.T., and Hamilton, A.D. (2018). Mimicry of a β-hairpin turn
by a nonpeptidic laterally flexible foldamer. Org. Lett. 20 (13): 3879–3882.
51 Atcher, J., Nagai, A., Mayer, P. et al. (2019). Aromatic β-sheet foldamers based
on tertiary squaramides. Chem. Commun. 55 (70): 10392–10395.
52 Nowick, J.S., Powell, N.A., Martinez, E.J. et al. (1992). Molecular scaffolds. I.
Intramolecular hydrogen bonding in a family of di- and triureas. J. Org. Chem. 57
(14): 3763–3765.
53 Nowick, J.S., Abdi, M., Bellamo, K.A. et al. (1995). Molecular scaffolds. 2.
Intramolecular hydrogen bonding in 1,2-diaminoethane diureas. J. Am. Chem.
Soc. 117 (1): 89–99.
54 Clayden, J., Lemiègre, L., Pickworth, M., and Jones, L. (2008). Conformation and
stereodynamics of 2,2′ -disubstituted N,N′ -diaryl ureas. Org. Biomol. Chem. 6 (16):
2908–2913.
55 Nowick, J.S., Holmes, D.L., Mackin, G. et al. (1996). An artificial β-sheet com-
prising a molecular scaffold, a β-strand mimic, and a peptide strand. J. Am.
Chem. Soc. 118 (11): 2764–2765.
56 Nowick, J.S. (2006). What I have learned by using chemical model systems
to study biomolecular structure and interactions. Org. Biomol. Chem. 4 (21):
3869–3885.
57 Khan, A., Kaiser, C., and Hecht, S. (2006). Prototype of a photoswitchable
foldamer. Angew. Chem. Int. Ed. 45 (12): 1878–1881.
58 Berl, V., Krische, M.J., Huc, I., and Lehn, J. (2000). Template-induced and molec-
ular recognition directed hierarchical generation of supramolecular assemblies
from molecular strands. Chem. Eur. J. 11: 1938–1946.
59 Jeon, H.-G., Lee, H.K., Lee, S., and Jeong, K.-S. (2018). Foldamer-based heli-
cate displaying reversible switching between two distinct conformers. Chem.
Commun. 54 (45): 5740–5743.
60 Knipe, P.C., Thompson, S., and Hamilton, A.D. (2016). Acid-mediated topological
control in a functionalized foldamer. Chem. Commun. 52 (39): 6521–6524.
61 Patel, B.H., Percivalle, C., Ritson, D.J. et al. (2015). Common origins of RNA,
protein and lipid precursors in a cyanosulfidic protometabolism. Nat. Chem. 7
(4): 301–307.
62 Nagata, Y., Nishikawa, T., Suginome, M. et al. (2018). Elucidating the solvent
effect on the switch of the helicity of poly(quinoxaline-2,3-diyl)s: a conforma-
tional analysis by small-angle neutron scattering. J. Am. Chem. Soc. 140 (8):
2722–2726.
63 Shimomura, K., Ikai, T., Kanoh, S. et al. (2014). Switchable enantioseparation
based on macromolecular memory of a helical polyacetylene in the solid state.
Nat. Chem. 6 (5): 429–434.
64 Mazzier, D., Crisma, M., De Poli, M. et al. (2016). Helical foldamers incorporat-
ing photoswitchable residues for light-mediated modulation of conformational
preference. J. Am. Chem. Soc. 138 (25): 8007–8018.
65 Inai, Y., Kurokawa, Y., and Kojima, N. (2002). Screw sense preference of
non-polar L-amino acid residues second from the N-terminal position. J. Chem.
Soc., Perkin Trans. 2 (11): 1850–1857.
66 Sol, J., Morris, G.A., and Clayden, J. (2011). Comparison of 13 C NMR signal
separation at slow and fast exchange. J. Am. Chem. Soc. 133 (11): 3712–3715.
67 Byrne, L., Solà, J., Boddaert, T. et al. (2014). Foldamer-mediated remote stereo-
control: >1,60 asymmetric induction. Angew. Chem. Int. Ed. 53 (1): 151–155.
68 Brioche, J., Pike, S.J., Tshepelevitsh, S. et al. (2015). Conformational switching
of a foldamer in a multicomponent system by pH-filtered selection between
competing noncovalent interactions. J. Am. Chem. Soc. 137 (20): 6680–6691.
69 Inai, Y., Ishida, Y., Tagawa, K. et al. (2002). Noncovalent Domino effect on
helical screw sense of chiral peptides possessing C-terminal chiral residue. J. Am.
Chem. Soc. 124 (11): 2466–2473.
70 Fujii, T., Shiotsuki, M., Inai, Y. et al. (2007). Synthesis of helical
poly(n-propargylamides) carrying azobenzene moieties in side chains. Reversible
arrangement-disarrangement of helical side chain arrays upon photoirradiation
keeping helical main chain intact. Macromolecules 40 (20): 7079–7088.
71 Ousaka, N., Inai, Y., and Kuroda, R. (2008). Chain-terminus triggered chiral
memory in an optically inactive 310-helical peptide. J. Am. Chem. Soc. 130 (37):
12266–12267.
72 Komori, H. and Inai, Y. (2007). Control of peptide helix sense by temperature
tuning of noncovalent chiral domino effect. J. Org. Chem. 72 (11): 4012–4022.
73 Jones, J.E., Diemer, V., Adam, C. et al. (2016). Length-dependent formation of
transmembrane pores by 310-helical α-aminoisobutyric acid foldamers. J. Am.
Chem. Soc. 138 (2): 688–695.
74 Lizio, M.G., Andrushchenko, V., Pike, S.J. et al. (2018). Optically active vibra-
tional spectroscopy of α-aminoisobutyric acid foldamers in organic solvents and
phospholipid bilayers. Chem. Eur. J. 24 (37): 9399–9408.
75 Shannon, C.E. (1948). A mathematical theory of communication. Bell Syst. Tech.
J. 27 (3): 623–656.
References 519
76 Tomsett, M., Maffucci, I., Le Bailly, B.A.F. et al. (2017). A tendril perversion in
a helical oligomer: trapping and characterizing a mobile screw-sense reversal.
Chem. Sci. 8 (4): 3007–3018.
77 Kollman, P.A. (1972). A theory of hydrogen bond directionality. J. Am. Chem.
Soc. 94 (6): 1837–1842.
78 Cherezov, V., Rosenbaum, D.M., Hanson, M.A. et al. (2007). High-resolution
crystal structure of an engineered human β2-adrenergic G protein-coupled recep-
tor. Science 318 (5854): 1258–1265.
79 Bécart, D., Diemer, V., Salaün, A. et al. (2017). Helical oligourea foldamers as
powerful hydrogen bonding catalysts for enantioselective C–C bond-forming
reactions. J. Am. Chem. Soc. 139 (36): 12524–12532.
80 Diemer, V., Fischer, L., Kauffmann, B., and Guichard, G. (2016). Anion recogni-
tion by aliphatic helical oligoureas. Chem. Eur. J. 22 (44): 15549.
81 Wechsel, R., Raftery, J., Cavagnat, D. et al. (2016). The meso helix: symme-
try and symmetry-breaking in dynamic oligourea foldamers with reversible
hydrogen-bond polarity. Angew. Chem. Int. Ed. 55 (33): 9657–9661.
82 Wechsel, R., Žabka, M., Ward, J.W., and Clayden, J. (2018). Competing
hydrogen-bond polarities in a dynamic oligourea foldamer: a molecular spring
torsion balance. J. Am. Chem. Soc. 140 (10): 3528–3531.
83 Tarai, A. and Baruah, J.B. (2017). Conformation and visual distinction between
urea and thiourea derivatives by an acetate ion and a hexafluorosilicate cocrystal
of the urea derivative in the detection of water in dimethylsulfoxide. ACS Omega
2 (10): 6991–7001.
84 Nowick, J.S., Mahrus, S., Smith, E.M., and Ziller, J.W. (1996). Triurea deriva-
tives of diethylenetriamine as potential templates for the formation of artificial
β-sheets. J. Am. Chem. Soc. 118 (5): 1066–1072.
85 Morris, D.T.J., Wales, S.M., Tilly, D.P. et al. (2021). A molecular communication
channel consisting of a single reversible chain of hydrogen bonds in a confor-
mationally flexible oligomer. Chem https://doi.org/10.1016/j.chempr.2021.06
.022.
86 Laidler, K.J. and King, M.C. (1983). The development of transition-state theory.
J. Phys. Chem. 87 (15): 2657–2664.
87 Spiegel, J., Adhikari, S., and Balasubramanian, S. (2020). The structure and func-
tion of DNA G-quadruplexes. Trends Chem. 2 (2): 123–136.
88 Nam, S., Ware, D.C., and Brothers, P.J. (2018). Campestarenes: new building
blocks with 5-fold symmetry. Org. Biomol. Chem. 16 (35): 6460–6469.
89 Chen, Z., Guieu, S., White, N.G. et al. (2016). The rich tautomeric behavior of
campestarenes. Chem. Eur. J. 22 (49): 17657–17672.
90 Kang, J., Miyajima, D., Itoh, Y. et al. (2014). C5 -Symmetric chiral corannu-
lenes: desymmetrization of bowl inversion equilibrium via “intramolecular”
hydrogen-bonding network. J. Am. Chem. Soc. 136 (30): 200–203.
91 Siemsen, P., Livingston, R.C., and Diederich, F. (2000). Acetylenic coupling:
a powerful tool in molecular construction. Angew. Chem. Int. Ed. 39 (15):
2632–2657.
92 Gropp, C., Trapp, N., and Diederich, F. (2016). Alleno-acetylenic cage

(AAC) receptors: chiroptical switching and enantioselective complexation of
trans-1,2-dimethylcyclohexane in a diaxial conformation. Angew. Chem. Int. Ed.
55 (46): 14444–14449.
93 Mishiro, K., Furuta, T., Sasamori, T. et al. (2013). A cyclochiral conformational
motif constructed using a robust hydrogen-bonding network. J. Am. Chem. Soc.
135 (37): 13644–13647.
521
Index
a ATR-FUV spectrometer 409, 411, 428

ab initio molecular dynamics 68, 86, 355 attendant protium-deuterium isotopic
ab initio potential energy surface (PESs) ratio 89
303 attenuated total reflection (ATR)
acetonitrile (ACN) 132, 134, 139, 233, spectroscopy technique 128
244, 257, 266, 268–270, 273–283, attractive or repulsive interactions 488
285, 286, 492
2-acetyl-1,8-dihydroxy-3,6- b
dimethylnaphthalene 194, 199 backbone CONH(D) isotope effects 186
adenine 41, 50, 51, 184 Badger-Bauer correlation 380
adenine crystal 50 barrier-penetration phenomena 90
adiabatic ionization energy 316, 325 basis set superposition error (BSSE) 377
Aib foldamer 490, 495, 499, 500 Beer-Lambert law 127, 445
alcohol-water systems benzene dimer 317
alcohol concentration and temperature benzoic acid (BA) 219–225, 360–362
464–470 bifunctional 2-naphthols 285–288
importance of 455 binding energy of H-bonded complexes
literature review of 457–462 314–331
spectral decomposition and probes for Birge–Sponer extrapolation method 296
characterization 456–464 Birge–Sponer (BS) extrapolation
alleno-acetylenic cage (AAC) 511–513 technique 325
alternating least squares (ALS) method bis(2-quinolylmethyl)(2-pyridylmethyl)
417, 418, 428 amine (BQPA)-ligated copper
amide-appended corannulenes 511 complex 495
γ-aminobutyric acid 213 2,6-bis(N-imidazolidin-2-onyl)pyridines
aminocyclohexanecarboxylic acid 488
(ACHC) 480, 481, 492, 502 blueshifted H-bonds 345
aminoisobutyric acid (Aib) 489, 490, bond strength 87, 89–91, 95, 96, 107,
492, 494–497, 499–501 173, 178, 184, 376, 377, 380,
anions of ionic liquids 124 383–387, 441, 452–454, 463, 479,
anomalous Ubbelohde effect 356 480, 483
aplanar proton-transfer mechanism 107 Born–Oppenheimer approximation 47,
apolar gases 448 175

522 Index
Born–Oppenheimer Molecular Dynamics carboxy-substituted naphthols 265

spectroscopic 41 Car–Parrinello molecular dynamics 41,
Brønsted-type PT reactions 261 56
bulk water CBr3 COOH 217
OH (OD)-stretching mode 438 CCl3 COOH 217
temperature-dependent infrared CF3 COOH 217
spectra 439–441 chain length effect on the degree of
self-association of 1-alcohols
c 237–240
campestarenes 508, 510 charge assisted hydrogen bonds (CAHBs)
canonical tunneling 92 124, 131–134
carboxylic acid dimers ignoring Fermi chelate-ring breathing 100
resonances chelate-ring deformation modes 100
gaseous acrylic and propynoic acids chemically equivalent oxygen moieties
15, 17 91
gaseous and liquid acetic acid dimers 2-chlorobenzoic acid (2CBA) 221, 222,
14–16 224, 225
carboxylic acids account Fermi cis-1,2-dimethylcyclohexane 513
resonances classical hydrogen bond 129–131
combined crystalline oxindole acid classical view 442
dimers 31 Cl/F-H stretchings 39
crystalline adipic acid 16, 18 1-Color REMPI (1-C-REMPI) spectrum
crystalline furoic acid dimer with slow 330
mode Morse potential and Fermi condensed-phase acid dissociation
resonances 28–30 constants 89
crystalline H(D)-3-thiopheneacrylic continuum of hydrogen bonding 87–93,
acid 20, 24 97
crystalline polarized and unpolarized cooperative resonance-assisted hydrogen
glutaric acid 17, 20, 21 bonds 134–138
crystalline thiopheneacetic acid 17, corrphycene 146, 152–154, 169
20, 23 Cotton effect 510–512
H-bonded compounds, combined Coulomb interaction 263, 268, 285
crystalline oxindole acid dimers Coulomb potential 263
31 coupled-cluster theory 93, 94, 107
liquid formic acid mixing of monomer coupled reaction-diffusion reactive system
and dimer 27, 29 262
l.2-naphtylacetic acid (2-NA) crystals 4-cyano-2,2,6,6-tetramethyl-3,5-
20, 21, 25 heptanedione enol 197
monomer of (CH3 )2 O… HCl, electrical cyclic hydrogen bond chains
anharmonicity 34–36 in alleno-acetylenic cages 512
Morse potential 23–24, 26 amide-appended corannulenes 511
phosphinic acid dimer 31, 33, 34 G-quadruplexes 508
phthalic and terephthalic acid crystals open and closed form of the AAC 512
25, 27–29 4-pyrrolidinopyridines 513
6-carboxy-2-naphthol 264 tautomerism in campestarenes 510
Index 523
cyclic trimer of (HCl)3 305 dimethyl ether (DME) 39, 40, 42, 314
cyclochiral 513, 514 dimethylsulfoxide (DMSO) 130, 131,
cyclohexane 243, 482, 487, 503 135–138, 271, 273, 486, 506
cyclohexanol 243, 244 dipalmitoylphosphatidylcholine
cyclohexene-1,2-dicarboxylate (DOPC)-derived vesicles 496
monoanion 195 dipole moment function (DMF) 36,
cyclopropanes 482 253–256, 309, 385
cytosine crystal 51–52 2,3-dipyrrol-2-ylquinoxalines 190
Dirac constant 69
d dispersed fluorescence (DF) spectra 102,
delocalization of probability density 86 103
density functional theory (DFT) dispersed fluorescence (DF) spectroscopy
calculations 40, 68, 75, 146, 173, 98
175, 176, 186, 240, 241, 253, 453 dissociation energy 296
deuterium 74 of cationic H-bonded complexes
isotope effects on chemical shifts 325–328
176–177 of H-bonded complexes 296–297
isotope effects on 1 H chemical shifts SEP-REMPI technique 328–331
187 of the ionic complex 316
deuterium kinetic isotope effect (DKIE) DME-HCl complex 40, 46, 47
91, 97, 103–105 DME-H(D)Cl complexes 45–47
2,4-dialkoxy-meta-aminomethylbenzoic DME-HF complexes 39, 40, 44–48
acid (MAMBA) scaffold 486 donor-acceptor distance 85, 86, 88–92,
2,6-dibromophenol 249–251 94, 95, 104–107, 145, 152
dicarboxylic acid anion system 195 double boundary (DB) solution 267, 284
2,6-dichlorophenol 249–251 dynamical tunneling phenomena 92
difluorobenzene 299–300, 331 dynamic foldamers 489, 494
difluoromaleate monoanion 195 hydrogen-bonded 488–501
2,6-difluorophenol 249–251 reversible hydrogen-bond directionality
2,6-dihalogenated phenols 249, 251 in 501–508
2,6-dihalogenophenols 249, 257 dynamic view 442
dihydrogen bonds 124
diisopropylamine-formate system 189 e
β-diketone systems 197 electrolyte aqueous solutions 411–412
dimensionless oscillator strength 309, electron density 50–52, 89, 94, 126, 169,
310, 313 192, 193, 295, 413, 468, 501
dimer strong anharmonic coupling theory electronically excited HFF 102, 107
diabatic approximation for monomer Energy gap law 301
3, 5–6 energy splitting 67, 69, 97, 104
dimer involving damping Davydov EOM-CCSD method 414–416, 418
coupling and Fermi resonances equilibrium constant 271, 279, 280, 285,
12–13 307–309, 351, 353
Fermi resonances 6–8 equilibrium dissociation energy 294,
H-bonded centrosymmetric dimer 333–335
8–13 ethaline (ETH) 132
524 Index
excess spectroscopy 126–129 Franck–Condon type progressions 43

excited state intramolecular hydrogen Frank–Condon (FC) distribution 317
transfer in tropolone 71, 75–79 free energy of complexation 307
excited-state potential surfaces 106 freon cryosolvents 188
experimental binding energy 295, 305 Fukui functions 50
f g
far-ultraviolet (FUV) spectroscopy GAMESS 71
biomedical science 429
gas-phase proton affinities 89
characteristics and advantages of
gauche/anti conformations 487
410–411
Gibbs Free energy of H-bonded complex
development of 428
formation 307–314
of electrolyte aqueous solutions
G protein-coupled receptor (GPCR) 496
411–412
G-quadruplexes 508
Group I, II, and XIII cation 412
interfacial water adsorbed 416–418 ground-state potential surfaces 95
lithium ion (Li+) 424–427 ground-state properties of model systems
nylons 422–424 93–97
F-H stretching absorption band 46 ground-state TrOH and HFF 94
fifth, interstitial, non-tetrahedral or guanine crystal 51
mismatched water molecules 454 Gutmann-Beckett scale 360
first-order saddle point 67, 87
five unit cells 52, 53 h
fluorescence-based spectroscopic probes β-hairpin-mimetic trimeric MAMBA
98 foldamers 486
fluorescence excitation spectra of HFF halogen-substituted acetic acids
isotopologs 99 215–219, 228
2-fluorobenzoic acid (2FBA) 220, 221 H-bond acceptor (HBA) 273, 304, 314,
foldamers 480 321
Aib 490, 495, 499, 500 H-bond characteristics
β-Hairpin-mimetic trimeric MAMBA correlations between geometry 165,
486
166
dynamic (see dynamic foldamers)
corrphycene 152–154
helical aromatic oligoamide 484
hemiporphycene 150–152
hydrogen-bonded (see hydrogen-
inverted/confused porphyrin 162–164
bonded foldamers)
isoporphycene 154–156
hydrogen-bonded dynamic (see
hydrogen-bonded dynamic) neo-confused porphyrin 164–165
reversible hydrogen-bond directionality parameters 167–168
in dynamic (see reversible porphine 147–148
hydrogen-bond directionality in porphycene 148–150
dynamic) porphyrin-(2.2.0.0) 156, 157
fragment molecular-orbital- porphyrin-(3.1.0.0) 157–158
multicomponent molecular-orbital porphyrin-(4.0.0.0) 158–162
method 175 tautomerization mechanisms 168–169
Index 525
H-bond dissociation energy, multiphoton attractive or repulsive interactions 488

ionization (MPI) techniques 2,6-bis(N-imidazolidin-2-onyl)pyridine
316–324 oligomers 489
H-bond donor (HBD) 270, 273, 274, 304, BQPA 495, 496
308, 311, 314, 320 Cu(II)-BQPA binding site 496
H-bonded complex difluorobenzene 299 definition 488
H-bonding interactions, microwave determination of helical excess
spectroscopy 332–335 490–501
Heaviside step function 70 light-modulated screw-sense induction
helical aromatic oligoamide foldamers in phospholipid bilayers 498
484 screw-sense preference 490, 493
hemiporphycene 145, 146, 150–152, 157, screw-sense preference in
158, 167 DOPC-derived vesicles 497
Henri-Rousseau and Blaise model 3 synthetic molecular communication
1-hexanol 240–241, 243, 244, 257 489
homo-conjugated anions 356, 357, 359, Tendril perversions in Aib foldamers
370, 371
499
H2 S dimer 294
hydrogen-bonded foldamers
hydrogen bond 40, 85
accessible conformations of
charge assisted hydrogen bonds
ethylene-bridged triureas 487,
(CAHBs) 131–134
488
classical 129–131
ACHC-and ACPC-derived foldamers
continuum of 87–93, 97
480, 481
cooperative resonance-assisted
amide in 482
hydrogen bonds 134–138
β-sheet-mimetic oligourea foldamers
criteria of the existence of 124–125
487
definition 123, 293
exploiting repulsive interactions 485
excess spectroscopy 126–129
helical aromatic oligoamide foldamers
polarizability 348
potentials 173–175, 189, 197, 202 484
spectroscopic metrics for (see N,N-diarylformamide balances 480
spectroscopic metrics for hydrogen π-stacking interactions 484
bonding) supramolecular and biological systems
strength of 125–126 479
types 176, 385 VT NMR 483
by use of higher overtones 249 X-ray crystallography 482
of water 235 hydrogen-bonded systems
weak/moderate hydrogen bonds adenine and thymine crystals 50, 51
138–142 DME-H(D)Cl and DME-HF complexes
hydrogen-bond donors (HBD) 270, 273, 45–48
274, 455, 479–483, 501–504, 506, guanine and cytosine crystals 51–52
508 1-methyl-4-thiouracil 50
hydrogen-bonded dynamic foldamers 1-methyluracil 50
Aib helix 494, 495 molecular dynamics, methodology of
Aib nonapeptides 490 47–48
526 Index
hydrogen-bonded systems (contd.) 8-hydroxypyrene 1,3,6-trisulphonate

(CH3 )2 O-HCl and (CH3 )2 O-HF (HPTS) photoacid 262
complexes 42–45 8-hydroxyquinoline molecular assemblies
spectroscopic signature for ferroelectric 124
ice 52–55 3-hydroxyquinoline photoacid 284
uracil 49–50
hydrogen-bonded water molecules 418, i
453 icebergs 442, 450, 453, 470
hydrogen bond energy 193 infrared (IR) excitation
dissociation energy of H-bonded
definition 375–377
complexes 296–297
IR characterization of
Gibbs free energy of H-bonded complex
carbonyl stretching vibration 387
formation 307–314
low-frequency hydrogen bond
integrated absorbance 308, 309, 313
stretching frequency 385–386
intermolecular complexes 349, 360, 361,
proton donor deformational
375, 376
vibrations 385
intermolecular H-bond energy 294, 295
proton donor stretching band
intermolecular hydrogen bonds (IHB)
intensity 384–385
40, 42, 49, 86, 176, 177, 187–189,
proton donor stretching band shift
217, 251, 372
378, 380–384
internal complex salt-bridge 214
stretching vibrations’ force constants
intramolecular hydrogen bonds (IMHB)
386–387
93, 177
NMR characterization of 377–379
halogen-substituted acetic acids
hydrogen bond geometry
215–219
averaging of NMR parameters and ortho chloro-and/or fluoro-substituted
proton tautomerism 350–353 benzoic acids 219–225
description of 348–350 thiotropolone (TT) 225–228
IR hydrogen bond correlations intramolecular hydrogen transfer in
371–375 malonaldehyde 68, 71, 73–74, 79
NMR hydrogen bond correlations intramolecular vibrational relaxation
353–371 (IVR) process 228, 328
hydrogen-bonding motifs 87, 88, 92, 97, intrinsic reaction coordinates (IRC)
100 67–77, 79, 227
hydrogen rac-dimethylsuccinate 197 intrinsic reaction coordinate-vibrational
10-hydroxybenz[h]quinolines 181 adiabatic (IRC-VA) 72–74
9-hydroxy-9-fluorenecarboxylic acid intrinsic reaction coordinate-Wentzel–
(9-HFCA) 315 Kramers–Brillouin (IRC-WKB)
6-hydroxy-2-formylfulvene (HFF) 87, 93 67–69, 71–75, 77, 79
excited-state spectroscopy of 98–102 inverted/confused porphyrin 146, 162
ground-state spectroscopy of 102–105 ion cluster-CH3 CN complex 139
2-hydroxynaphaldehyde 199 ion clusters 139
8-hydroxypyrene 1,3,6- ionic liquids (ILs) 124, 189, 194, 383,
trisdimethylsulfonamide 411, 428
(HPTA) 271 IR hydrogen bond correlations
Index 527
carbonyl stretching vibration 375 linear distortion parameter 44

proton donor deformational vibrations lithium ion (Li+ ) 424–427
374 local mode perturbation theory (LMPT)
proton donor stretching vibration model 310
371–374 long-range isotope effects 182, 184–185,
IR-UV double resonance spectroscopy 191
314–316 low-barrier hydrogen bonding (LBHBing)
isodesmic reactions 215 85–109
isoporphycene 146, 154–156, 165, 169
isotope D/H effect 283–285 m
isotope effects magnetic anisotropy effects 378
calculation of 175–176 Makri–Miller method 68, 79
deuterium isotope effects 187 malonaldehyde 68, 71, 73–75, 78, 79
deuterium isotope effects on chemical malondialdehyde 71, 73, 74
shifts 176–177 mass analyzed threshold ionization
exchange in the solid state 192–193 (MATI) technique 323, 324
hydrogen bond energies 193 mass selected two-color multiphoton
hydrogen bond potentials 173–174 ionization technique 321, 322
hydrogen bond types 176 1-methyl-4-thiouracil 40, 41, 49–50
intermolecular hydrogen bonds 1-methyluracil 40, 41, 49–50
187–189 microheterogeneity in alcohol/alcohol
intramolecular hydrogen bonds and alcohol/alkane binary
177–185 mixtures 241–244
primary deuterium isotope effects microwave spectroscopy 213, 214, 224,
189–191 317, 321, 332–335
proteins 185–187 molar absorption coefficient 127, 128
protonation states 191 molecular dynamics (MD) simulations
solid state NMR 197–202 41, 47, 48, 356, 367, 375, 453, 454
solvent isotope effects and exchange Møller–Plesset perturbation theory 105
rates 192 Møller–Plesset second-order perturbation
tautomerism 194–197 theory (MP2) 72
isotope substitution 173, 175, 191, 194, multiphoton ionization (MPI) techniques
202, 287 316–324
multivariate curve resolution (MCR)
k analysis 238, 242, 243, 440, 441,
kinetic isotope effects (KIE) 86, 91, 97, 456, 463–465
108, 283–287 mutual diffusion coefficient 263, 268,
Kramers–Kronig transformation 418, 285
419, 440
n
l 1-naphthol…2-butyne complex 330, 331
laser-induced fluorescence (LIF) natural transition orbitals (NTOs) 413
spectroscopy 98–100, 102, 108 N-(3,5-dibromosalicylidene)-methylamine
level of theory 96, 97, 105, 133, 244, 309, 194
334, 355, 384, 386, 413, 425 2N68dS photoacid 266, 268
528 Index
neo-confused porphyrin 146, 164–165 N-(pyridoxylidene)-methylamine 194,

n-hexane 240–241, 243, 244, 249–251, 201, 202
257 N-(pyridoxylidene)tolylamine 201, 202
NH stretching frequency 147, 153, 160, N-triflylphosphoramide 492
166–168
NIR spectroscopy o
association of 1-hexanol in n-hexane OD(H)-stretching band 438
240–241 (CH3 )2 O-HCl complex 42–45
chain length effect on the degree of (CH3 )2 O-HF complex 42–45
self-association of 1-alcohols OH photoacid 263, 270–273
237–240 OH (OD)-stretching mode 438
characteristics 233–235 o-hydroxy Schiff bases 194
hydrogen bonding by use of higher o-hydroxythioacetophenones 181, 198
overtones 249–251 one-bond deuterium isotope effects on
15
hydrogen bonding of water 235–237 N chemical shifts 185, 191, 199
microheterogeneity in alcohol/alcohol one-water-bridge ESPT reaction 286
optothermal bolometric determination
and alcohol/alkane binary
297–299
mixtures 241–244
ortho chloro-and/or fluoro-substituted
overtones of 𝜈C≡N vibration 244–245
benzoic acids 219
pyrrole 252–256
oscillator strength 309–311, 313, 414,
weak hydrogen bond in poly(3-
423, 424
hydroxybutyrate) (PHB) 246–249
overtones of 𝜈C≡N vibration 244–245
nitromalonamide 197
N-methylacetamide (NMA) 135–138,
p
418–422
pentachlorophenol-4-methylpyridine
N-methylformamide (NMF) 135–138,
complex 202
418–421
p-fluorophenol 325–328
NMR hydrogen bond correlations phenol 249–251, 257, 272, 273, 320, 327,
FHF, FHN and FHO bonds 365–369 380, 479, 480
NHN bonds 363–365 phosphinic acids 31, 33, 189, 358–360,
OHN bonds 360–363 370, 371
OHO bonds-13 C and 31 P NMR chemical photoactive yellow protein 175, 370, 371
shifts 356–360 photoionization efficiency curve 325
OHO bonds-1 H chemical shifts phthalate monoanion 195
353–356 piroxicam 195
vicinal H/D isotope effects 369–371 point-to-point PT reactions 261
N,N-diarylformamides 479 poly(3-hydroxybutyrate) (PHB)
N,N’-disubstituted (thio)ureas 482 246–249, 257
noncovalent interactions 85–87, 92, 93, porphine 146–148, 164, 168
95, 96 porphycene 145, 146, 148–150, 155, 156,
nonlinear dielectric effect (NDE) 240, 165, 167–169, 175, 363, 364
257 porphyrin-(2.2.0.0) 146, 156
normal mode (NM) model 295, 310 porphyrin-(3.1.0.0) 146, 157–160
Nowick’s β-sheets 487 porphyrin-(4.0.0.0) 146, 158, 160–162
Index 529
8-position monosulfonated 2-naphthol resonant-2-photon ionization (R2PI)

2N8S photoacid 268 technique 328–330
potential-barrier tunneling 86 reverse spectroscopic problem 345–346
potential energy surface (PES) 47, 48, reversible geminate proton recombination
67, 68, 75, 89, 93, 101, 103, 105, GEM reactions 262
108, 214, 215, 224, 227–229, 296, reversible hydrogen-bond directionality in
303, 305, 306, 309, 310, 484, 488 dynamic foldamers
primary deuterium isotope effect 176, dual hydrogen-bond donor/acceptor
188–191, 202 monomers 502
proteins 134, 175, 176, 185–187, 193, ethylene-bridged oligourea foldamers
357, 362, 370, 429, 437, 479, 481, 505, 506
483, 497 in ethylene-bridged oligoureas 507
proton association 263, 264 Eyring-Polanyi equation 505
protonation reaction 262, 264, 266, 269, hydrogen-bond acceptor 501
288 hydrogen bond-donating/accepting
proton dissociation 262–264, 268, 269, ability 504
285 oligoureas 503, 504
proton transfer 85–87, 89–94, 96, 97, screw-sense preference 503
100, 103, 104, 106–108, 173, 176, reversible proton dissociation-
188, 196, 261, 274–282, 285–289, recombination reaction 262
346–349, 351–354, 356–358, 360, rotameric formamides 479
363–367, 369, 375–377
pseudo-diatomic model 333 s
pyridoyl benzoyl β-diketones 198 SAC-CI calculations 421
4-pyridoyl derivative 198 Sapporo-double zeta polarization (DZP)
pyrrole 252 basis function 72
monomer 253 Schiff-base macrocycles 508
pyridine complex 253, 254 secondary intrinsic isotope effects 177
4-pyrrolidinopyridines (PPY) 513, 514 self-modeling curve resolution (SMCR)
236
q semiclassical tunneling approach
quasi-lattice model 236, 256 69–71
SEP-REMPI technique 328–331
r short strong hydrogen bonds (SSHBs)
Reisler’s measurement 293 126, 377
remote protonation 261, 264, 265 single-boundary (SB) GEM model 268
resonance assisted hydrogen bonds solid state NMR 192, 197–202, 362
(RAHBs) 124, 134, 135, 137, 138, solvatomers 109, 350, 356, 375
176, 177, 190, 384 solvent isotope effects and exchange rates
resonance enhanced multi-photon 192
ionization (REMPI) detection SO3 -substituted 2-naphthol 266
scheme 300–306, 315, 319, 330 spectral density 13–14, 33
resonant-assisted hydrogen bonding spectral markers for proton transfer and
(RAHBing) 93, 95 H-bond length 346–348
530 Index
spectroscopic metrics for hydrogen thiophenoxyketenimines 183

bonding thiotropolone (TT) 225–228
continuum of hydrogen bonding β-thioxoketones 198
87–92 thymine 41, 50, 51, 184
excited-state properties of model time correlated single photon counting
systems 105–108 277
excited-state spectroscopy of time-dependent density functional theory
6-hydroxy-2-formylfulvene (TDDFT) 75, 413
98–102 time-independent Schrodinger equation
ground-state properties of model 251
systems 93–97 time-of-flight (TOF) mass spectrometer
ground-state spectroscopy of 323
6-hydroxy-2-formylfulvene time resolved fluorescence (TRF) 262
102–105 decay signal 268
relationship to tunnelling 92–93 measurements 277
spectroscopic signature time-resolved Stokes shift 271
ferroelectric ice 52–55
TOF-MD-WKB method 69–71, 75, 76, 79
of low-barrier hydrogen bonding in
trajectory on-the-fly molecular dynamics
neutral species 85–109
(TOF-MD) 67–79
stimulated emission pumping (SEP)
trans-2-aminocyclopentanecarboxylic
328, 329
acid (ACPC) 480, 481, 502
stretching force constant 332–335
trans-1,2-dimethylcyclohexane 513
strongly hydrogen-bonded (SHB) water
transient absorption spectroscopy 262
species 235–236
trihalogeno-substituted acetic acids 217,
sulfonamides 482
221
sulfonate substituted naphthols 265
tropolone (TrOH) 69, 71, 75–79, 93, 97,
t 105, 106, 109
tautomerism 194 two-bond deuterium isotope effects on 13 C
averaging of NMR parameters and chemical shifts 178–183, 203
proton 350–353 two-color photodissociation experiments
in campestarenes 510 319
isotope effects 194–197 two-color REMPI (2C-REMPI)
tautomerization mechanisms measurements 319
168–169 two-colour two-photon resonant
t-butyl alcohol 314 ionization techniques 321
tetraethylammonium hydrogen two-dimensional correlation spectroscopy
R-(+)-methylsuccinate 197 (2DCOS) 238, 239
tetraethylammonium (TEA) hydrogen two-dimensional hydrogen bond
succinate 197 potentials 173
2,3,5,6-tetrafluoroterephthalic acid 175
tetramethylammonium (TMA) hydrogen u
meso-dimethylsuccinate 197 Ubbelohde effect 89, 175, 356, 366
2,2,2-t-fluoroethanol 314 umbrella inversion of ammonia 71–73
thiodibenzoylmethane 198 uracil 40, 49, 50, 184
Index 531
v water spectroscopy
vacuum ultraviolet (VUV) region 409 bulk water 438–441
valence bond orders 350, 355, 361, 365 IR spectroscopy, water hydrogen
variable-temperature NMR (VT NMR) bonding, alcohol-water systems
483, 491, 505 455–470
velocity map imaging (VMI) 299–307 water near fully hydrophobic solutes
vibrational local mode (LM) theory 309 442–454
vibrationally-induced selective weak hydrogen bond in poly(3-
conformational isomerization hydroxybutyrate) (PHB) 246–249
216 weakly hydrogen-bonded (WHB) water
vibrational pre-dissociation 296 species 235
vibrationless zero-point level 91 weak/moderate hydrogen bonds
vibronically resolved fluorescence 138–142
techniques 108 weak organic Brønsted acids 261
vibronic transitions 101 Wentzel–Kramers–Brillouin (WKB)
vicinal H/D isotope effect 369 67–69, 227
w x
water…amine complexes 311 XH spectroscopic method 233
water near fully hydrophobic solutes
experimental procedure 443–446 z
clathrate formations 448, 449 zero electron kinetic energy (ZEKE)
temperature and pressure on the OD spectroscopy 315, 323, 325, 326
stretching band 446–448 zero-point averaged rotational constants
Monte Carlo and MD simulations 333
453–454 zero point energy (ZPE) 5, 7, 72–75, 79,
pure hydrophobic solutes in water 86, 90, 91, 96, 108, 147, 148, 174,
solution 449–453 227, 294–296, 376, 377
transmission high pressure cell 443 zero-point vibrational energy 72

Marek J. Wójcik, Yukihiro Ozaki - Spectroscopy and Computation of Hydrogen-Bonded Systems-Wiley-VCH (2023)

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Marek J. Wójcik, Yukihiro Ozaki - Spectroscopy and Computation of Hydrogen-Bonded Systems-Wiley-VCH (2023)

Uploaded by

Copyright:

Available Formats

Downloaded from https://onlinelibrary.wiley.com/doi/ by ibrahim ragab - Cochrane Germany , Wiley Online Library on [13/12/2022].

Edited by Marek J. Wójcik and Yukihiro Ozaki

Prof. em. Yukihiro Ozaki

© 2023 WILEY-VCH GmbH, Boschstraße 12,

All rights reserved (including those of

Print ISBN: 978-3-527-34972-2

Typesetting Straive, Chennai, India

1 Linear Response Theory Applications to IR Spectra of

2 Dynamic Interactions Shaping Vibrational Spectra of

3 Trajectory On-the-Fly Molecular Dynamics Approach to

4 Spectroscopic Signatures of Low-Barrier Hydrogen Bonding in

5 Hydrogen-Bonding Interactions Using Excess

6 Intramolecular Hydrogen Bonding in Porphyrin Isomers 145

6.2.7 Porphyrin-(3.1.0.0) (7) 157

7 Isotope Effects in Hydrogen Bond Research 173

8 Intramolecular Hydrogen Bonding: Shaping Conformers’

9 Hydrogen Bonding from Perspective of Overtones and

10 Direct Observation and Kinetic Mapping of Point-to-Point

10.3 Water Is Able to Donate and Accept an H-bond as Demonstrated by IR

11 Spectroscopic Determination of Hydrogen Bond

12 IR and NMR Spectral Diagnostics of Hydrogen Bond Energy

12.1.1 Solving the Reverse Spectroscopic Problem 345

13 ATR-Far-Ultraviolet Spectroscopy Holds Unique Advantages

14 Water–Hydrogen-Bond Network and Hydrophobic Effect 435

15 Hydrogen Bond Chains in Foldamers and Dynamic

Linear Response Theory Applications to IR Spectra of

1.2 Dimer Strong Anharmonic Coupling Theory

𝜔 (Q) = 𝜔∘ + bQ with b < 0 (1.1)

Blaise, Wójcik and Henri-

V=0 Maréchal and Witkowski V=0

V=0 V=0 V=0

Boulil et Maréchal and Witkowski Rösch-Ratner

Quantum model of direct

Sakun Commutation of stochastic Abramczyk

Memory Robertson and Yarwood Rotation

Slow modulation limit

The full Hamiltonian may be partitioned as follows:

Figure 1.3 Coordinates of single q

Table 1.1 Different sorts of Hamiltonians.

1.2.1.2 Introduction of Fermi Resonances

∣(m)〉 ⊗ ∣{1}〉 ⊗ ∣[0]〉

∣(m)〉 ⊗ ∣{0}〉 ⊗ ∣[2]〉

The different Hamiltonians are given as follows:

1.2.1.3 H-Bonded Centrosymmetric Dimer

∣(m)2〉 ⊗ ∣{1}2〉 ∣{0}1〉 ⊗ ∣(m)1〉

∣(m)2〉 ⊗ ∣(0)2〉 ⊗ ∣{1}1〉 ⊗ ∣(m)1〉

where Θ is a dimensionless parameter governing the linear dependence of the Davy-

Then the Davydov Hamiltonian becomes:

Now, the action of the parity operator on the symmetrized operators

Table 1.2 Symmetrized coordinates.

Symmetric coordinate Antisymmetric coordinate

Figure 1.7 Action of the C2 operator on coordinates and eigenstates. Source:

⎜||𝛽 (+) ⎟ = ⎜|{1} ⊗ || 𝛽 (+) {1} ⎟ (1.20)

As a consequence, the Hamiltonian of the dimer involving Davydov coupling,

1.2.2 The Spectral Density

in which 𝜁 is the electrical anharmonicity parameter.

At time t the u transition operator is obtained by the Heisenberg transformation

It may be shown that the ACF takes the form:

where 𝛾 ∘ is the relaxation parameter of the high-frequency mode. Besides, [G(t)]g

1.3 Comparison with Experiments

1900 2750 3600

1900 2750 3600

1900 2750 3600