Chemical Physics 117 (1987) 415-420 415

North-Holland, Amsterdam

AB INITIO STUDY OF THE J(CC) INDIRECT NUCLEAR SPIN-SPIN COUPLING

CONSTANTS IN CYCLOBUTANE AND RELATED SYSTEMS

V. GALASSO
Dipartimento di Scienre Chimiche, Universitci di Trieste, I-34127 Trieste, Italy

Received 20 May 1987

Non-empirical calculations using the equations-of-motion approach, which incorporates the main portion of the electron
correlation effects, are reported for the carbon-carbon nuclear spin-spin coupling constants in cyclobutane, bicyclobutane,
tricyclobutane, cyclobutene, cyclobutyne, cyclobutadiene, bicyclobutene, methylenecyclopropane, and methylenecyclopropene.
The results provide an overall picture of the influences exerted on sign and magnitude of the J(CC) by progressive
condensation, unsaturation, and branching rearrangement of the cyclobutane frame.

1. Introduction

Highly strained microcycles possess unusual
chemical reactivity and physical properties in
comparison with the parent open-chain systems,
and their synthesis and characterization have
elicited a still-active experimental as well as theo-
retical interest (e.g., ref. [l]). Among the various
spectroscopic properties, the nJ(13C’3C) nuclear
spin-spin coupling constants deserve a particular
attention since they behave as efficient monitors
of alteration in the electronic structures of these
systems, extremely sensitive to substituents and/or
steric influences. An amazing example is apparent
in the ‘J(CC) coupling constant of the bridgehead
bond in bicyclobutane. In a previous paper we
have analyzed theoretically at the ab initio level
the dependence of the main features of ‘J(CC) on
changing the bridgehead heteroatom from B to 0
and from Si to S in the series of monoheteroana-
logs of cyclopropane and cyclopropene [2]. Here,
we report on a systematic theoretical study of the
nJ(CC) coupling constants for the most repre-
sentative homocyclic systems that can be formally
built from a butylic frame, cyclobutane (l), bi-
cyclobutane (2), tetrahedrane (formally tri-
cyclobutane) (3), cyclobutene (4), cyclobutyne (5),
cyclobutadiene (6), bicyclobutene (7), methylene-
cyclopropane (8), and methylenecyclopropene (9)
(see structures in scheme 1). Two peculiar features
of these systems are the wide ranges both of
‘J(C-C), which exceptionally bears the negative
sign for the bridgehead bond in bicyclobutane,
and of 2J(CC), which is found even larger than
‘J(C-C) in some derivatives of cyclobutene.
Non-empirical calculations were performed via

Scheme 1

0301-0104/87/$03.50 0 Elsevier Science Publishers B.V.

(North-Holland Physics Publishing Division)
416 V. Galasso / Spin -spin coupling constants in cyclobutane
the equations-of-motion method [3], which incor-
porates the main dynamical correlation effects
and was applied recently by the author to the ab
initio investigation of the J tensors of many mole-
cules [2,4]. -
2. Computational details
The “J(CC) nuclear spin-spin coupling con-
stants were determined at ab initio level by the
variational form of the equations-of-motion
(EOM) method described by Rowe [3],
(0 I[% H, Ox+]
IO) = 40 I [ Q, Ox+]
IO>,
which was solved according to the particular set of
approximations proposed by Shibuya and McKay
[5], by restricting the excitation operator 0:) which
generates an excited state 1A) from the correlated
ground-state (GS) IO), i.e. 0: 10) = 1A), to the
single particle-hole (lp-lh) pairs.
Full J tensors were calculated considering all
the four electron-nucleus spin perturbations of
Ramsey’s theory [6], i.e. the Fermi-contact (FC),
orbital-diamagnetic (OD), orbital paramagnetic
(OP), and spin-dipolar (SD) terms. The nondia-
magnetic contributions to J were evaluated as the
GS expectation values:
J(Xy)=KjC(ol[o~, IO)
OX+]
h basis sets and of 0: built from only lp-lh inter-
x (0 IKL %I low,
where Kj is a constant factor, the sum runs over
all the singlet or triplet excitation energies wA, and
Ox and OY are the relevant electron-nucleus
interaction operators. Taking into account the fact
that recent CI calculations [7] have shown the
minimal influence of electron correlation on the
OD component of J, this contribution was calcu-
lated by assuming the GS equal to the Hartree-
Fock determinant.
The split-valence-plus-polarization-functions
6-31G* basis set [8] was used for the carbon
atoms, while a simple ls-5G function [9] was used
for H. Note that the neglect of polarization func-
tions on hydrogens has only a modest influence on
the J(CC), less than 7% according to our results
on cyclobutadiene, and does not alter the signifi-
cance of the various trends examined here.
The experimentally determined geometries were
used, if known [lo], otherwise theoretical struc-
tures were assumed [ll-151; for cyclobutane the
fully planar conformation was assumed.
It is to be emphasized that the present un-
saturated molecules cannot be studied by the
standard ab initio CHF or SOS CI (restricted to
only monoexcited configurations) formalisms,
since they suffer from triplet instability [16] due to
strong electron correlation in the HF GS.
3. Results and discussion
The results of the ab initio EOM calculations
are presented in table 1 together with available
experimental data.
It must be noted that highly accurate reproduc-
tions of the J property require very large basis
sets, in particular for the FC term (e.g., ref. [26]).
(For acetylene, as an example, the theoretical
estimate of ‘J(C=C) is 217 Hz with the 6-31G*
basis set while it is 173 Hz with the 6-311G* *
basis set, in closer agreement with the experimen-
tal value of 171.5 Hz [27]). The quality of the
EOM results could be further improved by includ-
ing the 2p-2h components in the excitation, op-
erator O,+, but this step is computationally very
heavy and expensive. The present use of limited
actions restricts our interest to ‘analysis of the
main features of the J and not of the absolute
quality of the theoretical predictions. The primary
emphasis of this theoretical study is on the in-
fluences exerted on sign and magnitude of the
nJ(CC) by progressive condensation, unsaturation,
and branching rearrangement of the cyclobutane
frame. In the following paragraphs we discuss
these points with particular attention paid to
‘J(CC) and then we comment on 2J(CC).
3.1. Effects of progressive condensation
In agreement with expectation [17,22], upon
progressive condensation of the butylic frame a
lowering of = 10 Hz in *J(C-C) is forecast by the
EOM results on each step along the sequence
mono-bi-tricyclobutane. The ‘J(C-C) constant
 V. Gahso / Spin -spin coupling constants in eyclobutane 417

Table 1
“J(CC) isotropic indirect nuclear spin-spin coupling tensors in Hz

Compound ‘J(C=C)/‘J(C=C) ‘J(C-C) *J(CC)

1FC 37.73 - 6.06

OD 0.19 0.02
OP 0.77 - 0.08
SD 0.97 - 0.06
J 39.66 -6.18
exp 29.8 ‘) (-)8.1 b,

2 Cl--G G-C,
FC 28.66 - 7.00 - 7.64
OD 0.19 0.19 -0.04
OP - 0.89 - 0.46 -0.15
SD -0.13 - 0.50 -0.04
J 27.83 - 7.77 -7.87
exp. 21.0 =) -5.4 d, (-)16 =)
-17.49 6 ’

3FC 17.87
OD 0.18
OP -0.08
SD - 0.61
J 17.36
exp. 9.2 s)

4 cl-c4 G-C4
FC 81.59 43.86 35.39 - 8.60
OD 0.16 0.19 0.18 - 0.01
OP - 6.60 - 0.29 0.80 -0.11
SD 1.59 0.52 0.94 - 0.01
J 76.74 44.28 37.31 - 8.73
exp. 58.5 ‘) 32.13 i, 26.25 h’ (-)8.93 i)

5FC 140.77 46.20 44.33 4.31

OD 0.05 0.20 0.17 - 0.02
OP - 10.16 -0.16 - 2.01 -0.16
SD - 0.65 - 0.68 0.45 - 0.05
J 130.01 45.56 42.94 4.08

6FC 96.08 57.29 - 19.85

OD 0.15 0.18 -0.02
OP - 7.02 1.36 0.23
SD 3.12 1.92 - 1.18
J 92.33 60.75 - 20.82

7FC 40.92 20.95 - 6.93

OD 0.13 0.15 - 0.07
OP - 10.64 -2.15 - 0.69
SD 0.61 -0.16 0.37
J 31.02 18.79 - 7.32

8 Cl=c4 G-C* G-C,

FC 111.14 31.55 17.08 - 1.72
OD 0.17 0.20 0.15 - 0.11
OP -5.82 -1.04 0.24 -0.18
SD 1.91 -0.17 -0.03 0.10
J 107.40 30.54 17.44 - 1.91
95.2 j) 23.2 j)
(continued on next page)
418 V. Galasso / Spin -spin coupling constants in cyclobutane

Table 1 (continued)

Compound ‘J(C=c)/?I(C=C) ‘J(C-C) *J(CC)

9 CZ=G Cl=c4
FC 50.14 108.55 31.41 -1.38
OD 0.07 0.17 0.17 -0.14
OP - 5.50 -4.87 - 1.63 - 0.70
SD 1.03 1.27 - 0.53 0.17
J 45.74 105.12 29.42 - 2.05

‘) 1,1,3,3-tetramethylcyclobutane [17]. b, 1-methylcyclobutane [18].

‘) Ref. [17]. d, l-metbyl-3-phenyl-dicyciobutane-2,4-dicarboxylate ester [19].
e, 1-cyano-bicyclobutane 1201. u 2,2,4,4_tetra(trideuteromethyl)bicyclobutane [21].
@ tetra-t-butyhetrahedrane [22]. h, 1-methykyclobutene [23].
‘) Ref. [24]. j) Ref. [25].’

changes from a typical ethane-like value in mono-
cylobutane to a quite small value in tricyclobu-
tane, where it exhibits a magnitude similar to that
found in the single three-membered unit cyclopro-
pane of 12.4 Hz [28]. This trend reflects the paral-
lel decrease of s character in the C-C bond, as
shown by the total s overlap population according
to Mulliken population analysis.
The ‘J(C-C) for the bridgehead C-C bent bond
in bicyclobutane, which is referred to as a rather
unusual (nearly) pure p-p single bond [29], is
confirmed to be negative, in agreement with the
experimental sign determinations obtained on
some derivatives [19,21]. Direct comparison of the
EOM prediction with the spectroscopic data avail-
able for some derivatives is hampered by the
surprising flexibility of the bicyclobutane skeleton,
which upon substitution undergoes dramatic
changes in the dihedral angle between the cyclo-
propane rings and in the transannular C-C bond
length (as large as 0.10 A) [30]. Finally, it is to be
stressed that, in contrast to previous semi-em-
pirical CHF INDO results [31], the present ab
initiatio treatment shows that the dominant con-
tribution to the J of this unique bonding in bicy-
clobutane comes from the FC term and the non-
contact terms provide far smaller contributions.
The FC term is, exceptionally, negative as the
noncontact terms are.
3.2. Effects of progressive unsaturation
Upon passing from the single- to double- and
triple-bond rings, the corresponding ‘J(CC) in-
creases monotonically, as expected for a hybridi-
zational change from nominally sp3 to sp* and sp,
respectively, but this variation does not follow
closely that exhibited by the open-chain reference
series, ethane 34.6 Hz, ethylene 67.6 Hz, and
acetylene 171.5 Hz [27]. A strong deviation from
this pattern is, indeed, predicted on passing from
cyclobutene to cyclobutyne. This is, however, not
surprising since according to accurate ab initio
calculations [12] the C=C triple bond in cyclobu-
tyne is weak, as indicated by either bond distance
(1.26 A) or stretching vibrational frequency (=
1700 cm-‘), both significantly different from typi-
cal values for normal CS bonds. The “long” and
strained triple bond in cyclobutyne accounts also
for the marked difference noted between the theo-
retical prediction for its ‘J(M) (130 Hz) and the
experimental value of 166 Hz (quite similar to that
of acetylene) measured for cyclooctyne [32], where
the triple bond is reasonably normal, as reflected
also by its stretching frequency 2210 cm-’ and
bond distance 1.232 A [32].
IR [13] and PE [33] spectroscopic results have
provided a convincing argument in favour of the
rectangular geometry of cyclobutadiene. Its two
‘J(CC) are predicted by the EOM approach to be
significantly larger than not only those measured
for the unstrained 1,3-butadiene (68.80 and 53.70
Hz) [34] but also those calculated for the related
compound cyclobutene. Unfortunately, no experi-
mental values are yet available for comparison. It
is worthwhile noting that the average of the two
CC bond distances in the rectangular structure of
cyclobutadiene is quite similar to the CC distance
 K Galasso / Spin -spin
predicted ab initio for the quadratic form [13]. It
may be, therefore, tempting to look also at the
average of the corresponding ‘J(CC) values (83
Hz) as a theoretical estimate of the ‘J(CC) in the
quadratic structure. One must be cautious here,
however, since we do not know the consequences
brought about on the ‘J(CC) by the change in the
GS wavefunction, which varies from singlet (in the
rectangular ring) to triplet (square ring).
On passing from bicyclobutane to bicyclobu-
tene, the conversion of the transannular linkage
from single to double bond is manifested by the
occurrence of remarkably small ‘J(C=C). This
value, which turns out to be distinctly lower than
the one forecast for the single three-membered
unit cyclopropene 67.6 Hz [2], reflects the signifi-
cant biradical character of bent bicyclobutene [15].
3.3. Effects of branching rearrangement
The branching rearrangement of the cyclobu-
tane frame switches on a concerted action of ring
strain and bond polarity, which lead to a striking
alteration in the pattern of the “J(CC).
In particular, the conversion of cyclobutadiene,
the smallest [4n]annulene, into triafulvene, the
simplest cross-conjugated molecule, is accompa-
nied by ring contraction and, correspondingly, by
increase in ring strain, which is responsible for the
marked lowering predicted for the ‘J(CC) of the
endocyclic single and double bonds. The polar
exocyclic double bond of triafulvene exhibits,
instead, a coupling constant slightly larger than
that in the isomeric four-membered ring.
Similarly, upon isomerization of cyclobutene to
methylenecyclopropane, the double bond is
extruded out of the ring and becomes more “nor-
mal”. Thus, the ‘J(C=C) undergoes a significant
increase of the order of 30 Hz. The two endocyclic
‘J(C-C) get instead depressed to smaller values,
due to introduction of additional strain in the
ring.
Both the considered branching rearrangements
cause a sizeable reduction in the absolute value of
the geminal coupling constant.
coupling constants in cyclobutane 419
3.4. Geminal coupling constant
The theoretical geminal coupling constants
2J(CC) reproduce fairly well the experimental data
available for cyclobutane and cyclobutene. A like
value of = - 8 Hz is anticipated by the EOM
theory for bicyclobutane, while twice this value
(- 17 Hz) is predicted for cyclobutadiene. (Thus,
it appears that in bicyclobutane the 2J(CC) should
be quite similar to the ‘J(C-C) for the bridgehead
bond.) In order to rationalize qualitatively such a
pattern of rather large values of 2J(CC) in terms of
popular arguments, we could appeal to the dual-
path coupling mechanism (see refs. [18] and, e.g.,
[35]). In fact, 2J(CC) involves two equivalent
C-C-C paths in cyclobutane, two non-equivalent
paths, one across only single bonds and the other
across a single and a double bond, in cyclobutene,
and two equivalent paths across a C=C-C frame-
work in cyclobutadiene. However, within the EOM
formalism it is not possible to disentangle the
intimate interplay of these contributions.
The small negative values (= -2 Hz) calcu-
lated for the branched systems, triafulvene and
methylenecyclopropane, are consistent with the
< 1 Hz value measured for the related open-chain
reference compound 1,3-butadiene [34].
The 2J(CC) is predicted to be always negative
with the exception of cyclobutyne, where a posi-
tive value as large as 4 Hz is found, likely as a
consequence of the anomalous triple bond. Apart
from the special case of cyclobutyne, in the light
of our results we can conclude that in all cyclobu-
tanes 2J(CC) should be negative.
Finally, as to the specific role played by the
various spin perturbations, the information pro-
vided by the EOM results is that, irrespective of
the nature of the bonds between the carbons, the
“J(CC) couplings are all largely controlled by the
FC term. It is to be stressed, however, that in
bicyclobutene, where the FC term of ‘J(C=C) is
comparatively small, the OP term has the greatest
relative importance and it is in proportion a fourth
of the FC term and of contrary sign.
420 V. Galasso / Spin -spin coupling constants in cyclobutane

Acknowledgement 1151 B. Andes Hess Jr., W.D. A&n, D. Michalska, L.J. Schaad
and H.F. Schaefer III, J. Am. Chem. Sot. 109 (1987) 1615.

The author thanks Professor W. Liittke, Uni- WI J. C&k and J. PaIdus, J. Chem. Phys. 47 (1967) 3976.
P71 R.D. Bertrand, D.M. Grant, E.L. ABred, J.C. Hinshaw
versitgt Gijttingen (FRG), for helpful correspon- and A.B. Strong, J. Am. Chem. Sot. 94 (1972) 997.
dence. WI M. Klessinger, H. van Megen and K. Wilhelm, Chem. Ber.
115 (1982) 50.
1191 M. Pomeranz, R. Fink and G.A. Gray, J. Am. Chem. Sot.
98 (1976) 291.
References
WI M. Pomeranz and D.F. HiIIebrand, J. Am. Chem. Sot. 95
(1973) 5809.
[l] A. Greenberg and J.F. Liebman, Strained organic mole-
WI H. Finkelmeier and W. Lbttke, J. Am. Chem. Sot. 100
cules (Academic Press, New York, 1978).
(1978) 6261.
[2] G. Fronzoni and V. Galasso, J. Magn. Reson. 71 (1987)
WI T. Loerzer, R. Machinek, W. Liittke, L.H. Franz, K.-D.
229.
MaIsch and G. Maier, Angew. Chem. Intern. Ed. 22
[3] D.J. Rowe, Rev. Mod. Phys. 40 (1968) 153.
(1983) 878.
[4] V. Galasso, J. Chem. Phys. 82 (1985) 899;
v31 W. Ltittke, private communication.
V. Galasso and G. Fronzoni, J. Chem. Phys. 84 (1986)
~241 P.E. Hansen and J.J. Led, Org. Magn. Reson. 15 (1981)
3215; 85 (1986) 5200;
288.
G. Fronzoni and V. Galasso, Chem. Phys. 103 (1986) 29.
[251 H. Gunther and W. Herrig, Chem. Ber. 106 (1973) 3938.
[5] T. Shibuva and V. McKov. Phys. Rev. A 2 (1970) 2208;
WI M.F. Guest, V.R. Saunders and R.E. Gverill, Mol. Phys.
T. Shibuya, V. Rose and-V. McKay, J. Chem. Phys. 58
35 (1978) 427.
(1973) 500.
v71 R.M. Lynden-BeII and N. Sheppard, Proc. Roy. Sot. A
161N.F. Ramsey, Phys. Rev. 91 (1953) 303. 269 (1962) 385.
(71 R.E. OveriIl and M.F. Guest, Chem. Phys. Letters 98 WI J. Wardeiner, W. Ltittke, R. Bergholz and R. Machinek,
(1983) 229;
Angew. Chem. Intern. Ed. 21 (1982) 872.
R.E. OveriIl and V.R. Saunders, Chem. Phys. Letters 106
v91 M.D. Newton and J.M. Schuhnan, J. Am. Chem. Sot. 94
(1984) 197;
(1972) 767;
G. Fronzoni and V. Galasso, J. Mol. Struct. THEOCHEM
M. Eckert-Maksic, Z.B. Maksic and R. Gleiter, Theoret.
122 (1985) 327
Chim. Acta 66 (1981) 193.
PI P.C. Hariharan and J.A. Pople, Theoret. Chim. Acta 28
[301 P.G. Gassman, M.L. Greenlee, D.A. Dixon, S. Richtsmeier
(1973) 213.
and J. Zanos Gougoutas, J. Am. Chem. Sot. 105 (1983)
[91 W.J. Hehre, R.F. Stewart and J.A. Pople, J. Chem. Phys. 5865.
51 (1969) 2657.
[311 J.M. Schulman and M.D. Newton, J. Am. Chem. Sot. 96
WI B. Bak, J.J. Led, L. Nygaard and J. Rastrup-Andersen, J.
(1974) 6295;
Mol. Struct. 3 (1969) 369;
J.M. Schulman and T. Venanzi, Tetrahedron Letters (1976)
K.W. Cox, M.D. Harmony, G. Nelson and K.B. Wiberg,
1461.
J. Chem. Phys. 50 (1969) 1976;
[32] M. Traettenberg, W. Ltittke, R. Machinek, A. Krebs and
V.W. Laurie and W.M. Stighani, J. Am. Chem. Sot. 92
H.J. HohIt, J. Mol. Struct. 128 (1985) 217.
(1970) 1485;
[33] S. Masamune, F.A. Souto-Bachiher, T. Machiguchi and
T.D. Norden, S.W. Staley, W.H. Taylor and M.D.
J.E. Bertie, J. Am. Chem. Sot. 100 (1978) 4889.
Harmony, J. Am. Chem. Sot. 108 (1986) 7912.
[34] G. Becher, W. Ltittke and G. Schrumpf, Angew. Chem.
WI D. Cremer, J. Am. Chem. Sot. 99 (1977) 1307. Intern. Ed. 12 (1973) 339.
P21 G. Fitzgerald, P. Saxe and H.F. Schaefer III, J. Am. [35] M. Klessinger and M. Stocker, Org. Magn. Reson. 17
Chem. Sot. 106 (1983) 690.
(1981) 97;
1131 B. Andes Hess Jr., P. Carsky and L.J. Schaad, J. Am. J.-H. Cho, M. Klessinger, U. Tecklenborg and K. Wilhelm,
Chem. Sot. 105 (1983) 695.
Org. Magn. Reson. 23 (1985) 95.
[14] C.A. Scamehom, S.M. Hermiher and R.M. Pitzer, J. Chem.
Phys. 84 (1986) 833.