Full Ebook of A Course in Quantum Many Body Theory 1St Edition Michele Fabrizio Online PDF All Chapter

A Course in Quantum Many Body
Theory 1st Edition Michele Fabrizio

Visit to download the full and correct content document:
https://ebookmeta.com/product/a-course-in-quantum-many-body-theory-1st-edition-mi
chele-fabrizio/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Quantum Scaling in Many Body Systems An Approach to

Quantum Phase Transitions 2nd Edition Mucio
Continentino
https://ebookmeta.com/product/quantum-scaling-in-many-body-
systems-an-approach-to-quantum-phase-transitions-2nd-edition-
mucio-continentino/
Quantum Mathematics II 1st Edition Michele Correggi
https://ebookmeta.com/product/quantum-mathematics-ii-1st-edition-
michele-correggi/
Back-Of-The-Envelope Quantum Mechanics: with Extensions

to Many-Body Systems and Integrable Pdes 2nd Edition
Maxim Olshanii
https://ebookmeta.com/product/back-of-the-envelope-quantum-
mechanics-with-extensions-to-many-body-systems-and-integrable-
pdes-2nd-edition-maxim-olshanii/
A Course in Stochastic Game Theory Eilon Solan
https://ebookmeta.com/product/a-course-in-stochastic-game-theory-
eilon-solan/
A First Course in Spectral Theory 1st Edition Milivoje
Luki■
https://ebookmeta.com/product/a-first-course-in-spectral-
theory-1st-edition-milivoje-lukic/
A First Course in Group Theory 1st Edition Bijan Davvaz
https://ebookmeta.com/product/a-first-course-in-group-theory-1st-
edition-bijan-davvaz/
Quantum Mechanics a graduate course Solutions manual

1st Edition Hora■iu N■stase
https://ebookmeta.com/product/quantum-mechanics-a-graduate-
course-solutions-manual-1st-edition-hora%c6%abiu-nastase/
Quantum Field Theory A Diagrammatic Approach 1st

Edition Ronald Kleiss
https://ebookmeta.com/product/quantum-field-theory-a-
diagrammatic-approach-1st-edition-ronald-kleiss/
Factorization Algebras in Quantum Field Theory 1st

Edition Kevin Costello
https://ebookmeta.com/product/factorization-algebras-in-quantum-
field-theory-1st-edition-kevin-costello/
Graduate Texts in Physics
Michele Fabrizio
A Course
in Quantum
Many-Body
Theory
From Conventional Fermi Liquids
to Strongly Correlated Systems
Series Editors
Kurt H. Becker, NYU Polytechnic School of Engineering, Brooklyn, NY, USA
Jean-Marc Di Meglio, Matière et Systèmes Complexes, Bâtiment Condorcet, Université
Paris Diderot, Paris, France
Sadri Hassani, Department of Physics, Illinois State University, Normal, IL, USA
Morten Hjorth-Jensen, Department of Physics, Blindern, University of Oslo, Oslo,
Norway
Bill Munro, NTT Basic Research Laboratories, Atsugi, Japan
Richard Needs, Cavendish Laboratory, University of Cambridge, Cambridge, UK
William T. Rhodes, Department of Computer and Electrical Engineering and Computer
Science, Florida Atlantic University, Boca Raton, FL, USA
Susan Scott, Australian National University, Acton, Australia
H. Eugene Stanley, Center for Polymer Studies, Physics Department, Boston
University, Boston, MA, USA
Martin Stutzmann, Walter Schottky Institute, Technical University of Munich,
Garching, Germany
Andreas Wipf, Institute of Theoretical Physics, Friedrich-Schiller-University Jena,
Jena, Germany
Graduate Texts in Physics publishes core learning/teaching material for graduate- and
advanced-level undergraduate courses on topics of current and emerging fields within
physics, both pure and applied. These textbooks serve students at the MS- or
PhD-level and their instructors as comprehensive sources of principles, definitions,
derivations, experiments and applications (as relevant) for their mastery and teaching,
respectively. International in scope and relevance, the textbooks correspond to course
syllabi sufficiently to serve as required reading. Their didactic style, comprehensive-
ness and coverage of fundamental material also make them suitable as introductions
or references for scientists entering, or requiring timely knowledge of, a research field.
Michele Fabrizio
A Course in Quantum
Many-Body Theory
From Conventional Fermi Liquids to
Strongly Correlated Systems
Michele Fabrizio
Department of Physics
International School for Advanced Studies
Trieste, Italy
ISSN 1868-4513 ISSN 1868-4521 (electronic)

ISBN 978-3-031-16304-3 ISBN 978-3-031-16305-0 (eBook)
https://doi.org/10.1007/978-3-031-16305-0
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature
Switzerland AG 2022
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse
of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or
the editors give a warranty, expressed or implied, with respect to the material contained herein or for any
errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.
Cover image: Olga_Kostrova
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
To my family, especially my wife Laura
Preface
This book collects the Lecture Notes of the Ph.D. course on Many-Body Theory
that I have been teaching for many years now. As such, the book is not intended
to be a compendium of all possible topics in Many-Body Theory, but just aims
to provide Ph.D. students with concepts and tools that I personally consider indis-
pensable to tackle modern issues of Many-Body Theory, most notably strongly
correlated electron systems. For that same reason, the book unavoidably reflects
my own limited research experience.
Besides a core part, Chaps. 1–4, where basic tools are presented in detail, there
are frequent excursions in topics of strongly correlated electron systems, among
which the two final chapters are on Luttinger Liquids and the Kondo Effect. In
addition, the microscopic derivation of Landau’s Fermi Liquid Theory in Chap. 5
includes conventional Fermi Liquids as a special case, but also less conventional
states, like, for instance, the Mott insulators of Sect. 5.6 that display the same
spin and thermal properties of normal metals, which might be relevant to strongly
correlated materials.
All chapters contain very few simple problems, whose solution is not even
shown. In addition, most chapters include sections with applications that are
simply more involved problems whose solution is explicitly derived.
Trieste, Italy Michele Fabrizio
Acknowledgements I am grateful to Prof. Claudio Castellani, whose master’s course on Quantum

Many-Body Theory first sparked my interest in the field, and to Profs. Philippe Noziéres and Erio
Tosatti, who nourished and mentored that interest.
vii
Contents
1 Second Quantization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Fock States and Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Fermionic Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Fermi Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2.2 Second Quantisation of Multiparticle Operators . . . . . . . . . 6
1.3 Bosonic Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3.1 Bose Fields and Multiparticle Operators . . . . . . . . . . . . . . . . 10
1.4 Canonical Transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.4.1 Canonical Transformations with Charge
Non-conserving Hamiltonians . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.4.2 Harmonic Oscillators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.5 Application: Electrons in a Box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.6 Application: Electron Lattice Models and Emergence
of Magnetism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.6.1 Hubbard Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.6.2 Mott Insulators and Heisenberg Models . . . . . . . . . . . . . . . . 26
1.7 Application: Spin-Wave Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
1.7.1 Classical Ground State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1.8 Beyond the Classical Limit: The Spin-Wave Approximation . . . . . 32
1.8.1 Hamiltonian of Quantum Fluctuations . . . . . . . . . . . . . . . . . . 34
1.8.2 Spin-Wave Dispersion and Goldstone Theorem . . . . . . . . . 36
1.8.3 Validity of the Approximation
and the Mermin-Wagner Theorem . . . . . . . . . . . . . . . . . . . . . . 37
1.8.4 Order from Disorder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2 Linear Response Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.1 Linear Response Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.2 Kramers-Kronig Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.2.1 Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.3 Fluctuation-Dissipation Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.4 Spectral Representation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
ix
x Contents
2.5 Power Dissipation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

2.5.1 Absorption/Emission Processes . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.5.2 Thermodynamic Susceptibilities . . . . . . . . . . . . . . . . . . . . . . . . 57
2.6 Application: Linear Response to an Electromagnetic Field . . . . . . 59
2.6.1 Quantisation of the Electromagnetic Field . . . . . . . . . . . . . . 61
2.6.2 System’s Sources for the Electromagnetic Field . . . . . . . . 63
2.6.3 Optical Constants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
2.6.4 Linear Response in the Longitudinal Case . . . . . . . . . . . . . . 67
2.6.5 Linear Response in the Transverse Case . . . . . . . . . . . . . . . . 74
2.6.6 Power Dissipated by the Electromagnetic Field . . . . . . . . . 76
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3 Hartree-Fock Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.1 Hartree-Fock Approximation for Fermions at Zero
Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
3.1.1 Alternative Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.2 Hartree-Fock Approximation for Fermions at Finite
Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.2.1 Saddle Point Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.3 Time-Dependent Hartree-Fock Approximation . . . . . . . . . . . . . . . . . . 96
3.3.1 Bosonization of the Low-Energy Particle-Hole
Excitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.4 Application: Antiferromagnetism in the Half-Filled
Hubbard Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
3.4.1 Spin-Wave Spectrum by Time-Dependent
Hartree-Fock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
4 Feynman Diagram Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.1 Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.1.1 Imaginary-Time Ordered Products . . . . . . . . . . . . . . . . . . . . . . 121
4.1.2 Matsubara Frequencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.1.3 Single-Particle Green’s Functions . . . . . . . . . . . . . . . . . . . . . . 127
4.2 Perturbation Expansion in Imaginary Time . . . . . . . . . . . . . . . . . . . . . . 133
4.2.1 Wick’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
4.3 Perturbation Theory for the Single-Particle Green’s
Function and Feynman Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
4.3.1 Diagram Technique in Momentum and Frequency
Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4.3.2 The Dyson Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4.3.3 Skeleton Diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.3.4 Physical Meaning of the Self-energy . . . . . . . . . . . . . . . . . . . 148
4.3.5 Emergence of Quasiparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
4.4 Other Kinds of Perturbations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
4.4.1 Scalar Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
4.4.2 Coupling to Bosonic Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Contents xi
4.5 Two-Particle Green’s Functions and Correlation Functions . . . . . . 161

4.5.1 Diagrammatic Representation of the Two-Particle
Green’s Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
4.5.2 Correlation Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
4.6 Coulomb Interaction and Proper and Improper Response
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.7 Irreducible Vertices and the Bethe-Salpeter Equations . . . . . . . . . . . 170
4.7.1 Particle-Hole Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
4.7.2 Particle-Particle Channel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
4.7.3 Self-energy and Irreducible Vertices . . . . . . . . . . . . . . . . . . . . 175
4.8 The Luttinger-Ward Functional . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
4.8.1 Thermodynamic Potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
4.9 Ward-Takahashi Identities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
4.9.1 Ward-Takahashi Identity for the Heat Density . . . . . . . . . . 187
4.10 Conserving Approximation Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
4.10.1 Conserving Hartree-Fock Approximation . . . . . . . . . . . . . . . 191
4.10.2 Conserving GW Approximation . . . . . . . . . . . . . . . . . . . . . . . . 193
4.11 Luttinger’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
4.11.1 Validity Conditions for Luttinger’s Theorem . . . . . . . . . . . . 197
4.11.2 Luttinger’s Theorem in Presence of Quasiparticles
and in Periodic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
5 Landau’s Fermi Liquid Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
5.1 Emergence of Quasiparticles Reexamined . . . . . . . . . . . . . . . . . . . . . . . 211
5.2 Manipulating the Bethe-Salpeter Equation . . . . . . . . . . . . . . . . . . . . . . 215
5.2.1 A Lengthy but Necessary Preliminary Calculation . . . . . . 216
5.2.2 Interaction Vertex and Density-Vertices . . . . . . . . . . . . . . . . . 221
5.3 Linear Response Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
5.3.1 Response Functions of Densities Associated
to Conserved Quantities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
5.4 Thermodynamic Susceptibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
5.4.1 Charge Compressibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
5.4.2 Spin Susceptibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
5.4.3 Specific Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
5.5 Current-Current Response Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
5.5.1 Thermal Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
5.5.2 Coulomb Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
5.6 Mott Insulators with a Luttinger Surface . . . . . . . . . . . . . . . . . . . . . . . . 236
5.7 Luttinger’s Theorem and Quasiparticle Distribution
Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
5.7.1 Oshikawa’s Topological Derivation of Luttinger’s
Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
xii Contents
5.8 Quasiparticle Hamiltonian and Landau-Boltzmann

Transport Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
5.8.1 Landau-Boltzmann Transport Equation
for Quasiparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
5.8.2 Transport Equation in Presence
of an Electromagnetic Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
5.9 Application: Transport Coefficients with Rotational
Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
6 Brief Introduction to Luttinger Liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
6.1 What Is Special in One Dimension? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
6.2 Interacting Spinless Fermions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
6.2.1 Bosonized Expression of the Non-interacting
Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
6.2.2 Bosonic Representation of the Fermi Fields . . . . . . . . . . . . 275
6.2.3 Operator Product Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
6.2.4 Non-interacting Green’s Functions
and Density-Density Response Functions . . . . . . . . . . . . . . . 280
6.2.5 Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
6.2.6 Interacting Green’s Functions and Correlation
Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
6.2.7 Umklapp Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
6.2.8 Behaviour Close to the K = 1/2 Marginal Case . . . . . . . . 293
6.3 Spin-1/2 Heisenberg Chain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
6.4 The One-Dimensional Hubbard Model . . . . . . . . . . . . . . . . . . . . . . . . . . 305
6.4.1 Luttinger Versus Fermi Liquids . . . . . . . . . . . . . . . . . . . . . . . . 309
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
7 Kondo Effect and the Physics of the Anderson Impurity Model . . . . . . 313
7.1 Brief Introduction to Scattering Theory . . . . . . . . . . . . . . . . . . . . . . . . . 314
7.1.1 General Analysis of the Phase-Shifts . . . . . . . . . . . . . . . . . . . 319
7.2 The Anderson Impurity Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
7.2.1 Non Interacting Impurity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
7.2.2 Hartree-Fock Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
7.3 From the Anderson Impurity Model to the Kondo Model . . . . . . . 325
7.3.1 The Emergence of Logarithmic Singularities
and the Kondo Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
7.3.2 Anderson’s Poor Man’s Scaling . . . . . . . . . . . . . . . . . . . . . . . . 330
7.4 Noziéres’s Local Fermi Liquid Theory . . . . . . . . . . . . . . . . . . . . . . . . . . 332
7.4.1 Ward-Takahashi Identity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
7.4.2 Luttinger’s Theorem and Thermodynamic
Susceptibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Second Quantization
1
The first difficulty encountered in solving a many-body problem is how to deal with a
many-body wavefunction. The reason is that a many-body wavefunction has to take
into account both the indistinguishability of the particles as well as their statistics,
whereas in first quantisation any operator, including the Hamiltonian, does not have
those properties. Therefore, it is desirable to have at disposal an alternative scheme
where the indistinguishability principle as well as the statistics of the particles are
already built in the expression of operators. That is actually the scope of second
quantisation.
1.1 Fock States and Space
Let us take a system of N particles, either fermions or bosons. The Hilbert space
spans a basis of N -body orthonormal wavefunctions which should satisfy both the
indistinguishability principle as well as the appropriate statistics of the particles. The
simplest way to construct this space is as follows.
We start by choosing an orthonormal basis of single-particle wavefunctions:
φa (x) , a = 1, 2, . . . .
Here x is a generalised coordinate that includes both the space coordinate r as well as,
e.g., the z-component, σ , of the spin, which is half-integer for fermions and integer
for bosons. The suffix a is a quantum label and, by definition,

d x φa∗ (x) φb (x) = δab , φa∗ (x) φa (y) = δ(x − y) , (1.1)
a
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 1

M. Fabrizio, A Course in Quantum Many-Body Theory, Graduate Texts in Physics,
https://doi.org/10.1007/978-3-031-16305-0_1
2 1 Second Quantization

where d x . . . means σ dr . . . , while δ(x − y) ≡ δ(r − r ) δσ σ with x =

(r, σ ) and y = (r , σ ). A generic N -body wavefunction with the appropriate sym-
metry properties can be constructed through the above single-particle states. Since
the particles are not distinguishable, we do not need to know which particle occupies
a specific state. Instead, what we need to know are just the occupation numbers n a ’s,
i.e., the number of particles occupying each single-particle state φa . That number is
either n a = 0, 1 for fermions, because of the Pauli principle, or an arbitrary integer
n a ≥ 0 for bosons. Apart from that, the occupation numbers should satisfy the trivial
particle-conservation constraint

na = N .
a
Since the occupation numbers are the only ingredients we need in order to build up
the N -body wavefunction, we can formally denote the latter as the ket
|n 1 , n 2 , . . . , (1.2)
which is called a Fock state, while the space spanned by the Fock states is called
Fock space. Within the Fock space, the state with no particles, the vacuum, will be
denoted by |0.
For instance, if the N fermions with coordinates xi , i = 1, . . . , N , occupy the
states a j , j = 1, . . . , N , with a1 < a2 < · · · < a N , namely n a = 1 for a ∈ {a j },
otherwise n a = 0, then the appropriate wavefunction is the Slater determinant

φa1 (x1 ) · · · φa1 (x N )

1 φa2 (x1 ) · · · φa2 (x N )
{n a } (x1 , . . . , x N ) = .. .. .. .. .. , (1.3)
N! . ... .

φa (x1 ) · · · φa (x N )
N N
which is therefore the first-quantisation expression of the Fock state |{n a } with
the same occupation numbers. The above wavefunction satisfies the condition of
being antisymmetric if two coordinates are interchanged, namely two columns in
the determinant. Analogously, it is antisymmetric by interchanging two rows, i.e.,
two quantum labels.
On the contrary, if we have N bosons with coordinates xi , i = 1, . . . , N , which
occupy the states a j , j = 1, . . . , M, M ≤ N and a1 < a2 < · · · < a M , with occu-
pation numbers n a j , then the appropriate wavefunction is the permanent

j n j!
{n a } (x1 , . . . , x N ) = φa1 (x p1 ) . . . φa1 (x pna )
N! p
1
(1.4)
φa2 (x pna ) . . . φa2 (x pna )
1 +1 1 +n a2
. . . φa M (x p N −na ) . . . φa M (x p N ) ,
M +1
1.2 Fermionic Operators 3
where the sum is over all non-equivalent permutations p’s of the N coordinates.1
Indeed, the wavefunction is even by interchanging two coordinates or two quantum
labels.
In conclusion, the space spanned by all possible Slater determinants built with
the same basis set of single-particle wavefunctions constitutes an appropriate Hilbert
(Fock) space for many-body fermionic wavefunctions. Analogously, the space
spanned by all possible permanents is an appropriate Hilbert space for many-body
bosonic wavefunctions.
In the following, we will introduce operators acting in the Fock space. We will
consider separately the fermionic and bosonic cases.
1.2 Fermionic Operators
Let us introduce the creation, ca† , and annihilation, ca , operators which add or remove,
respectively, one fermion in state a. The operator ca† ca first annihilates then creates
a particle in a, which can be done as many times as many particles occupy that state.
Therefore,
ca† ca |n a = n a |n a , (1.5)
so it acts like the occupation number operator ca† ca ≡ n̂ a . Since by the Pauli principle
n a = 0, 1, then
ca† ca | 0 = 0, ca† ca | n a = 1 =| n a = 1. (1.6)
Analogously, the operator ca ca† first creates then destroys a fermion in state a, which
it cannot do if a is occupied because of the Pauli principle, while it can if it is empty.
Therefore,
ca ca† | 0 =| 0, ca ca† | n a = 1 = 0. (1.7)
Thus, either a is empty or occupied, the following equation holds:
ca ca† + ca† ca | n a =| n a , n a = 0, 1 , (1.8)
which leads to the operator identity
ca ca† + ca† ca = ca , ca† = 1, (1.9)
where the symbol {. . . } means the anti-commutator. Moreover, since we cannot

create nor destroy two fermions in the same state, it also holds that
ca , ca = ca† , ca† = 0. (1.10)
1 Non-equivalent means for instance that φi (x)φi (y) is equivalent to φi (y)φi (x).
Equations (1.9) and (1.10) are the anti-commutation relations satisfied by the fermion
operators with the same quantum label. Going back to (1.6) and (1.7), we readily see
that all are satisfied if
ca† | 0 =| n a = 1, ca† | n a = 1 = 0, ca | 0 = 0, ca | n a = 1 =| 0,

(1.11)
also showing that ca† is the Hermitian conjugate of ca .
Let us now consider the action of the above operators on a Fock state. First, we
need to provide a prescription to build a Fock state by means of the creation operators.
We shall assume that, if
| n a = 1 = ca† | 0,
then by definition and for b = a
cb† | n a = 1 = cb† ca† | 0 ≡| n b = 1, n a = 1. (1.12)
Since the Slater determinant, hence the corresponding Fock state, is odd by inter-
changing two rows, then
| n b = 1, n a = 1 = − | n a = 1, n b = 1 ≡ −ca† cb† | 0. (1.13)
Comparing (1.13) with (1.12), we conclude that

ca† cb† = −cb† ca† ⇒ ca† , cb† = 0. (1.14)
The Hermitian conjugate thus implies that also

ca , cb = 0. (1.15)
We note that because of (1.10), both (1.14) and (1.15) remain valid even if a = b.
Finally, we need to extract the reciprocal properties of cb† and ca for a = b. We
assume the following result:
ca | n a = 1, n b = 1 = ca ca† cb† | 0 ≡| n b = 1 = cb† | 0. (1.16)
Since | 0 = ca ca† | 0, and ca† cb† = −cb† ca† , it follows that
−ca cb† ca† | 0 = cb† ca ca† | 0,
namely that, for a = b,

ca , cb† = 0. (1.17)
All the above anti-commutation relations can be cast in the simple formulas

ca , cb† = δab , ca† , cb† = 0, ca , cb = 0, ∀ a, b . (1.18)
If we extend our prescription (1.12) to more than two electrons, we find that any
Fock state has the simple expression
† n i
| n1, n2, . . . = ci | 0, (1.19)
i≥1
where the occupation numbers n i = 0, 1.
1.2.1 Fermi Fields
Till now we have defined fermionic operators for a given set of single-particle wave-
functions. Let us now introduce new operators which are independent of that choice.
We define annihilation and creation Fermi fields by

σ (r) ≡ (x) = φi (x) ci , (x)† = φi (x)∗ ci† . (1.20)
i i
They have the following anti-commutation relations:

(x), (y)† = φi (x) φ j (y)∗ ci , c†j
ij
(1.21)
= φi (x) φi (y)∗ = δ(x − y),
i
as well as

(x), (y) = (x)† , (y)† = 0 , (1.22)
which are indeed independent of the basis. If we change the basis via the unitary
transformation Û acting on the basis set

φi (x) = Ui,α φα (x),
α
with unitary Û , then

(x) = φi (x) ci = φα (x) Ui,α ci = φα (x) cα , (1.23)
i i,α α
showing that the proper transformation of the fermionic operators is

cα = Ui,α ci . (1.24)
i
Through the Fermi fields we can give a representation of the N -fermion Fock space
not through a basis of single-particle wavefunctions but directly in the space of the
coordinates. Specifically, such Fock space is spanned by the states
1
| x1 , . . . , x N ≡ √ † (x1 ) . . . † (x N ) | 0 , (1.25)
N!
which are indeed antisymmetric upon exchanging two coordinates. We note that,
through (1.1) and (1.20),

d x1 . . . d x N | x1 , . . . , x N x1 , . . . , x N |

1
= d x1 . . . d x N † (x1 ) . . . † (x N ) | 0 0 | (x N ) . . . (x1 )
N!
1 † (1.26)
= ca1 . . . ca†N | 0 0 | ca N . . . ca1
N ! a ,...,a
1 N

= ca†1 . . . ca†N | 0 0 | ca N . . . ca1 ≡ I N ,
a1 <a2 ···<a N
where I N is the identity in the N -fermion Hilbert space. In addition, it is straight-

forward to show that the Slater determinant identified by the occupation numbers
{n a } ≡ (n 1 , n 2 . . . ) can be simply written as
{n a } (x1 , . . . , x N ) = x1 , . . . , x N | n 1 , n 2 , . . .
1 (1.27)
=√ 0 | (x N ) . . . (x1 ) | n 1 , n 2 , . . . .
N!
1.2.2 Second Quantisation of Multiparticle Operators
Let us consider a generic single-particle operator in the first quantisation

N
V̂ = V (xi ) , (1.28)
i=1
where the sum runs over all N particles and V (xi ) is an operator acting both on space
coordinates and spins, namely
V (x) ≡ Vσ σ (r, p) , (1.29)
where p = −i∇ is the conjugate momentum of r.

Using (1.27) and the indistinguishability of the fermions, the matrix element
between two different Slater determinants can be shown to be
n 1 , n 2 , . . . | V̂ | n 1 , n 2 , . . .
N N
∗
≡ d xi {n } (x 1 , . . . , x N ) V (x i ) {n a } (x 1 , . . . , x N )
a
i=1 i=1

N
∗
=N d xi {n } (x 1 , . . . , x N ) V (x 1 ) {n a } (x 1 , . . . , x N ) (1.30)
a
i=1

N
N
= d xi V (x1 ) n 1 , n 2 , . . . | † (x1 ) † (x2 ) . . . † (x N ) | 0
N!
i=1
0 | (x N ) . . . (x2 ) (x1 ) | n 1 , n 2 , . . . .
We note that the integral over x2 , . . . , x N can be readily performed through (1.26)

N
d x2 . . . d x N † (x2 ) . . . † (x N ) | 0 0 | (x N ) . . . (x2 ) = I N −1 , (1.31)
N!
so that, dropping the identity operator I N −1 ,

n 1 , n 2 , . . . | V̂ | n 1 , n 2 , . . . = n 1 , n 2 , . . . | d x1 † (x1 ) V (x1 ) (x1 ) | n 1 , n 2 , . . . .
(1.32)
It thus follows that the second quantised expression of V̂ is simply

V̂ = d x † (x) V (x) (x) = Vab ca† cb , (1.33)
ab
where

Vab ≡ d x φa∗ (x) V (x) φb (x) = dr φa∗ (r, σ ) Vσ σ (r, p) φb (r, σ ).
σσ
(1.34)
As an example, let us consider the density operator that, in first quantisation, reads

N
ρ̂σ (r) = δ(r − ri ) δσ,σi ,
i=1
while in second quantisation is

ρ̂σ (r) = dr σ (r )† δ(r − r ) δσ,σ σ (r )
σ (1.35)
= σ (r) σ (r) = (x) (x),
† †
which also shows that (x)† is nothing but the operator which creates a particle at
x, while (x) destroys it.
Let us continue and consider a two-particle operator
1
Û = U (xi , x j ). (1.36)
2
i=j
We can simply repeat the previous steps and find that
n 1 , n 2 , . . . | Û | n 1 , n 2 , . . .
N
1 ∗
= d xi U (xi , x j ) {n (x 1 , . . . , x N ) {n a } (x 1 , . . . , x N )
2 a}
i=j i=1

N
N (N − 1) ∗
= d xi {n } (x 1 , . . . , x N ) U (x 1 , x 2 ) {n a } (x 1 , . . . , x N )
2 a
i=1

N
N (N − 1)
= d xi U (x1 , x2 )
N!
i=1
n 1 , n 2 , . . . | † (x1 ) † (x2 ) † (x3 ) . . . † (x N ) | 0
0 | (x N ) . . . (x3 ) (x2 ) (x1 ) | n 1 , n 2 , . . . ,
which, since,

N (N − 1)
d x3 . . . d x N † (x3 ) . . . † (x N ) | 0 0 | (x N ) . . . (x3 ) = I N −2 ,
N!
leads to the second quantised expression

1
Û = d x1 d x2 † (x1 ) † (x2 ) U (x1 , x2 ) (x2 ) (x1 )
2
1 (1.37)
= Uabcd ca† cb† cc cd ,
2
abcd
where

Uabcd = d x d y φa (x)∗ φb (y)∗ U (x, y) φc (y) φd (x) .
Analogously, a generic m-particle operator
1
Ûm = U (x1 , x2 , . . . , xm )
m!
i 1 =i 2 =··· =i m
1.3 Bosonic Operators 9
translates in second quantisation into

1
Ûm = d x1 . . . d xm † (x1 ) . . . † (xm ) U (x1 , . . . , xm ) (xm ) . . . (x1 ) .
m!
(1.38)
We conclude by emphasising the advantages of second quantisation with respect
to first quantisation. In the former, any multiparticle operator depends explicitly on
the particle coordinates. It is only the wavefunction that contains information about
the indistinguishability of the particles as well as their statistics. On the contrary,
in second quantisation, those properties are hidden in the definition of creation and
annihilation operators; hence, the multiparticle operators do not depend anymore on
the particle coordinates. Moreover, in second quantisation, we can also introduce
operators which have no first-quantisation counterpart. For instance, we can define
particle-non-conserving operators which connect subspaces with different numbers
of particles. For instance,

∗
ab ca† cb† + ab cb ca
ab
is a Hermitian operator that connects Fock states with particle numbers differing by
two. Those operators are useful when discussing superconductivity.
1.3 Bosonic Operators
As previously done for the fermionic case, we introduce the creation, da† , and its
Hermitian conjugate, the annihilation da , operators which, respectively, create and
destroy a boson in state a. Again the operator da† da counts how many times we can
destroy and create back a boson in state a; hence, it is just the occupation number
n a . However, since the Pauli principle does not hold for bosons, the operator da da†
first adds one boson in state a, next increases n a → n a + 1, and finally destroys one
boson in that same state. This latter process can be done n a + 1 times, being n a + 1
the actual occupation number once one more boson has been added. Therefore,
da† da |n a = n a | n a , da da† | n a = 1 + n a | n a ; (1.39)
Hence, the following commutation relation holds:

da , da† = da da† − da† da = 1, (1.40)
where [. . . ] denotes the commutator. Equation (1.39) is satisfied, e.g., by

√
da | n a = n a | n a − 1, da† | n a = n a + 1 | n a + 1 , (1.41)
which we assume hereafter. The permanent, contrary to the Slater determinant, is

invariant upon interchanging two quantum labels. As a consequence, bosonic oper-
ators corresponding to different states commute instead of anti-commuting as the
fermionic ones. Namely, for a = b,

da† , db = da† , db† = 0.
Therefore, in general,

da , db† = δab , da† , db† = da , db = 0. (1.42)
Moreover, through (1.41) and (1.42), we can write a generic Fock state as
d † ni
|n 1 , n 2 , . . . = i | 0 . (1.43)
i ni !
1.3.1 Bose Fields and Multiparticle Operators
The analogous role of the Fermi fields is now played by the Bose fields defined
through

(x) = φa (x) da , (x)† = φa (x)∗ da† , (1.44)
a a
which satisfy the commutation relations

(x), (y)† = δ(x − y) , (x), (y) = (x)† , (y)† = 0 . (1.45)
Exactly like in the fermionic case, if we define
1
| x1 , . . . , x N ≡ √ † (x1 ) . . . † (x N ) | 0 , (1.46)
N!
we find that

d x1 . . . d x N | x1 , . . . , x N x1 , . . . , x N |

1
= d x1 . . . d x N † (x1 ) . . . † (x N ) | 0 0 | (x N ) . . . (x1 ) (1.47)
N!
1 †
= d . . . da†N | 0 0 | da N . . . da1 ≡ I N ,
N ! a ,...,a a1
1 N
1.4 Canonical Transformations 11
as well as that the permanent with occupation numbers {n a } ≡ (n 1 , n 2 . . . ) is just
{n a } (x1 , . . . , x N ) = x1 , . . . , x N | n 1 , n 2 , . . .
1 (1.48)
=√ 0 | (x N ) . . . (x1 ) | n 1 , n 2 , . . . .
N!
These two equations imply that the second quantised expressions of the operators
are simply the same as in the fermionic case; for instance, a single-particle operator
is

V̂ = d x (x)† V (x) (x) , (1.49)
a two-particle one

1
Û = d x d y (x)† (y)† U (x, y) (y)(x) , (1.50)
2
and an m-particle operator

1
Ûm = d x1 . . . d xm (x1 )† . . . (xm )† U (x1 , . . . , xm ) (xm ) . . . (x1 ) .
m!
(1.51)
1.4 Canonical Transformations
In general, an interacting Hamiltonian that contains besides bilinear also quartic and
higher order terms in creation and annihilation operators cannot be diagonalised. On
the contrary, a bilinear Hamiltonian is diagonalisable by a canonical transformation
that preserves the commutation/anti-commutation properties of the operators.
Since the interaction is commonly analysed perturbatively starting from an appro-
priate non-interacting theory, it is useful to begin with bilinear Hamiltonians and
introduce the canonical transformations that diagonalise them.
The simplest bilinear Hamiltonian is the second quantised expression of a non-
interacting first-quantisation Hamiltonian, which has the general form

H= tab ca† cb , (1.52)
ab
both for fermions and bosons. Since H is Hermitian, then

∗
tab = tba .
If we define a column vector c, i.e., a spinor, with components ca , its Hermitian con-
jugate c† , a row vector, and the Hermitian matrix tˆ with elements tab , the Hamiltonian
can be shortly written as
H = c† tˆ c . (1.53)
Since tˆ is Hermitian, there exists a unitary transformation Û , i.e., satisfying Û † Û =

Û Û † = Iˆ, with Iˆ the identity matrix, such that
tˆ = Û † Ê Û ⇒ Ê = Û tˆ Û † , (1.54)
where Ê is diagonal with real elements a. Therefore, if we define a new spinor

operator d through
d ≡ Û c , (1.55)
with elements

da = Uab cb , (1.56)
b
then

H = d † Ê d = †
a da da (1.57)
α
is diagonal. We have now to check whether the above is a canonical transformation.
If the ca ’s are fermionic/bosonic operators, then

da , db† = †
Uan Umb cn , cm
†
= †
Uan Unb = δab ,
± ±
nm n
where {. . . }± stands for the anti-commutator (+) and commutator (-), respectively.
Therefore, also the da ’s are fermions/bosons; hence, Û is indeed canonical.
Once the Hamiltonian has been transformed into the diagonal form (1.57), the
problem is solved. Indeed any Fock state constructed through the new basis set with
operators da , namely a wavefunction | {n a } where each state a is occupied by n a
da -particles, is an eigenstate of the Hamiltonian

H | {n a } = a na | {n a }. (1.58)
a
It is important to remark that in the case of bosons the eigenvalues a must all be
greater or equal to zero; otherwise, the Hamiltonian would be ill-defined since the
ground state would correspond to putting an infinite number of bosons in the most
negative energy state, thus with infinitely negative total energy. The condition a ≥ 0
implies that tˆ must be semi positive definite.
An equivalent way to implement the canonical transformation (1.55) is through
a unitary operator
† φ̂
U = ei c c
≡ eiϕ , (1.59)
where φ̂ = φ̂ † is Hermitian and thus U † U = 1 unitary. We recall that, given two

operators A and B, then
(−i)n
e−i A B ei A = Cn , C0 = B , Cn = A, Cn−1 , (1.60)
n!
n≥0
which can be readily demonstrated by expanding the exponentials, so that, in the

specific case of U † c U

C0 = c , C1 = ϕ, c = −φ̂ c , Cn = (−1)n φ̂ n c , (1.61)
namely
in
U† c U = φ̂ n c = ei φ̂ c , U † c† U = c† e−i φ̂ , (1.62)
n!
n≥0
which coincides with (1.55) if ei φ̂ = Û , and thus e−i φ̂ = Û † . In this case
U † c U = Û c , U † c† U = c† Û † , (1.63)
which also imply

U c U † = Û † c , U c† U † = c† Û . (1.64)
In this formulation, the diagonalization of the Hamiltonian means the following.
Since the eigenstates and eigenvalues of a Hermitian operator do not change under a
unitary transformation, then we are allowed to study, instead of H , the transformed
Hamiltonian H̃ ≡ U H U † , which, through (1.64) and noting that Û tˆ Û † = Ê, can
be written as
H̃ ≡ U H U † = U c† tˆ c U † = U c† U † tˆ U c U † = c† Û tˆ Û † c = c† Ê c,
(1.65)
and therefore, it is diagonal.
Conversely, H = U † H̃ U , so that its eigenstates are simply U † | {n a }, where
| {n a } is a Fock state. Indeed

H U † | {n a } = U † H̃ | {n a } = a na U † | {n a } , (1.66)
a
which coincides with (1.58). In this language, the partition function

Z ≡ Tr e−β H = Tr U † U e−β H = Tr U e−β H U †
(1.67)
= Tr e−β U H U = Tr e−β H̃ ≡ Z̃
†
is evidently invariant under the unitary transformation, and the thermal average of
any operator A can be simply calculated through the diagonal Hamiltonian H̃ with
A transformed into U A U † ,
1 −β H 1 1
A = Tr e A = Tr U † U e−β H A = Tr e−β H̃ U A U † .
Z Z Z̃
(1.68)
1.4.1 Canonical Transformations with Charge Non-conserving

Hamiltonians
The above results show that a unitary transformation of the fermionic/bosonic oper-
ators ca preserves the anti-commutation/commutation relations. Indeed, if
ca → U † ca U , ca† → U † ca† U , (1.69)
with any unitary operator U , i.e., such that U † U = U U † = 1, then

U † ca U , U † cb† U = U † ca , cb† U = δab ,
± ±
(1.70)
U † ca U , U † cb U = U † ca , cb U = 0.
± ±
In second quantisation, we have the opportunity to introduce bilinear Hamiltonians

that does not conserve the number of particles:

H= tab ca† cb + ab ca cb + ab cb ca = c† tˆ c + cT ˆ c + H .c.
∗ † †
ab
(1.71)
and are relevant for a wide class of physical problems. Again tˆ = tˆ† , while ˆ is
antisymmetric for fermions and symmetric for bosons. Since H is bilinear in single-
particle operators, it can always be diagonalised by a proper unitary transformation
(1.69). However, if we write as before U = eiϕ , now the Hermitian operator ϕ must
be of the form

ϕ = c† φ̂ c + cT ψ̂ c + H .c. , (1.72)
with φ̂ = φ̂ † and ψ̂ antisymmetric/symmetric for fermions/bosons. In other words,

ϕ is still bilinear but does not conserve the number of particles. It is always possible
to find φ̂ and ψ̂ such that
U † H U = c† Ê c ≡ H̃ , (1.73)
that, given a Fock state | {n a }, then

with Ê diagonal with real elements a . It follows
U | {n a } is eigenstate of H with eigenvalue a a n a . As previously discussed, in
the case of bosons a ≥ 0, which poses constraints on tˆ and ˆ .
We just note that, while | {n a } is eigenstate of the number of particles with

eigenvalue N = a n a , the eigenstate U | {n a } = eiϕ | {n a } of the Hamiltonian

is not, since ϕ does not commute with the number of particle operator a ca† ca .
In the case of fermions, the diagonalization of the Hamiltonian (1.71) can still be
recast to that of a Hermitian matrix. Indeed, if we introduce the new spinors

c
ψ≡ , ψ † = c† , cT ,
c†T
the Hamiltonian (1.71) can be rewritten, dropping constant terms, as

tˆ/2 ˆ †
H =ψ †
ˆ −tˆT /2 ψ = ψ † Ĥ ψ ,
where now Ĥ is Hermitian and thus has real eigenvalues. We further note that, since
tˆ is Hermitian, then tˆT = tˆ∗ , while, since ˆ is antisymmetric, then ˆ † = − ˆ ∗ , so
that

tˆ/2 ˆ † tˆ/2 − ˆ ∗ 0 Iˆ ∗ 0 Iˆ
Ĥ = ˆ = ˆ =− ˆ Ĥ ≡ −σ̂1 Ĥ ∗ σ̂1 ,
−tˆT /2 −tˆ∗ /2 I 0 Iˆ 0
where σ̂1 is a generalised first Pauli matrix, thus σ̂1 Ĥ σ̂1 = − Ĥ ∗ . Therefore, if

uλ
χλ = : Ĥ χλ = λ χλ , λ ∈ R ,
vλ
so that Ĥ ∗ χλ∗ = λ χλ∗ , then
Ĥ σ̂1 χλ∗ = σ̂1 σ̂1 Ĥ σ̂1 χλ∗ = −σ̂1 Ĥ ∗ χλ∗ = −λ σ̂1 χλ∗ .
In other words,

v∗
σ̂1 χλ∗ = λ∗ ≡ χ−λ
uλ
is the eigenstate with opposite eigenvalue. Therefore, the unitary matrix

u λ1 u λ2 v∗λ1 v∗λ2
Û = ... ...
vλ 1 vλ 2 u∗λ1 u∗λ2
brings the Hamiltonian into the diagonal form

λ̂ 0
Û † Ĥ Û = ,
0 −λ̂
where λ̂ is diagonal with components λi , i = 1, 2, . . . , which also implies that the

canonical transformation

d
ψ = Û
d †T
allows to rewrite H in (1.71) as
λ̂ 0 d
H = ψ Ĥ ψ = d , d
† †
†T = λi 2di† di − 1 ,
0 −λ̂ d
i
namely in a diagonal form.
1.4.2 Harmonic Oscillators
When dealing with bosons, bilinear Hamiltonians very often reduce to coupled har-
monic oscillators, like, e.g., in the case of lattice vibrations in the harmonic approx-
imation. In those cases, it is more convenient to apply canonical transformations
directly to the conjugate variables rather than to the bosonic creation and annihila-
tion operators. For that, let us consider the prototypical example of a single harmonic
oscillator with Hamiltonian
2 2 g 2
H= p + x , (1.74)
2m 2

with x and p conjugate variables, i.e., x, p = i. We consider the transformation
√ 1
x= K X, p=√ P, (1.75)
K
which evidently preserves the commutation relations and therefore is canonical. In

terms of the new variables
2 Kg 2
H= P2 + X . (1.76)
2m K 2
If we fix K such that
2 2
= K g ⇒ K2 = , (1.77)
mK mg
we readily find that

g ω
H= P2 + X2 ≡ P2 + X2 . (1.78)
2 m 2
1.5 Application: Electrons in a Box 17
Hereafter, we denote the above expression as the ‘canonical form’ of the Hamiltonian.
Indeed, if we now introduce bosonic creation and annihilation operators through
1 i
X = √ a + a† , P = −√ a − a† , (1.79)
2 2
the Hamiltonian becomes

1
H = ω a† a + , (1.80)
2
i.e., the standard expression of a harmonic oscillator Hamiltonian with frequency ω.

We now consider the unitary operator

α
U = exp − i x p+ px ,
2
whose action on the conjugate variables can be readily found to be
U † x U = eα x , U † x U = e−α p ,
which yields the canonical transformation (1.75) if e2α = K . In other words,
ω
U† H U = p2 + x 2
2
directly brings the Hamiltonian in the canonical form. The eigenstates are therefore
U | n, where | n is the Fock state with n bosons.
1.5 Application: Electrons in a Box
Let us consider electrons in a square box of linear length L with periodic boundary
conditions. As a basis of single-particle wavefunctions, we use simple plane waves
and spin, namely the quantum label a = (k, σ ) with
2π
k= nx , n y , nz ,
L
being n i ’s integers, and σ =↑, ↓. Hence, the single-particle wavefunctions of the
basis set are
1
φa (x) = √ eik·r χσ .
L3
The annihilation and creation operators are defined as ck σ and ck† σ , respectively;
hence, the Fermi fields are
1 ik·r 1 −ik·r †
σ (r) = √ e ck σ , σ† (r) = √ e ck σ . (1.81)
L3 k L3 k
The kinetic energy in the first quantisation reads

N
2 2
Hkin = − ∇
2m i
i=1
and is diagonal in the spin. Since

1 −ik·r 2 2 2 k 2
dr e − ∇ eip·r = δk p ,
L3 2m 2m
then, on the plane-wave basis, the kinetic energy in the second quantisation becomes
2 k 2 † †
Hkin = c k σ ck σ ≡ k c k σ ck σ . (1.82)
2m
kσ kσ
The ground state with N , assumed to be even, electrons is the Fermi sea |F S obtained
by filling with spin-up and spin-down electrons the momentum states from k = 0 up
to the Fermi momentum k F defined through

N =2 ,
|k|≤k F
namely

|F S = ck† ↑ ck† ↓ | 0 ; (1.83)
k: |k|≤k F
hence, the occupation number in momentum space is
n k ↑ = n k ↓ = θ (k F − |k|) ,
where the θ -function is defined through θ (x) = 1 if x ≥ 0, otherwise θ (x) = 0.

1.5 Application: Electrons in a Box 19
Let us add an electron-electron interaction of the general form
1
Hint = U (ri − r j ) .
2
i=j
In second quantisation, it becomes

1
Hint = dx dy σ† (x)σ† (y) U (x − y) σ (y)σ (x)
2
σσ
1
= ck†1 σ ck†2 σ ck3 σ ck4 σ
2L 3
σ σ ki ,i=1,...,4

1
dx dy e−ik1 ·x e−ik2 ·y eik3 ·y eik4 ·x U (x − y) .
L3
We define the Fourier transform as

U (q) = dr e−iq·r U (r) ,
so that
1 iq·(x−y)
U (x − y) = e U (q) ,
L3 q
and the interaction is finally found to be
1
Hint = U (q) ck† σ cp+q
†
σ cp σ ck+q σ . (1.84)
2L 3 σσ k pq
We showed that the electron density for spin σ in second quantization is
ρσ (r) = σ† (r) σ (r) .
Its Fourier transform is

ρσ (q) = dr e−iq·r σ† (r) σ (r) = ck† σ ck+q σ .
k
Let us consider interacting electrons also in the presence of a single-particle potential,

provided, e.g., by the ions,

H pot = dx V (x) ρσ (x) ,
σ
so that the total Hamiltonian reads
H = Hkin + Hint + H pot .
Since the Hamiltonian conserves the total number of electrons, the electron density
summed over the spins, i.e., ρ = ρ↑ + ρ↓ must satisfy a continuity equation. Let us
define the Heisenberg evolution of the density through
ρ(r, t) = ei H t ρ(r) e−i H t .
Since the integral of the density over the whole volume is the total number of elec-
trons, N , which is conserved, it follows that

∂ρ(r, t) ∂ ∂N
dr = dr ρ(r, t) = ≡ 0.
∂t ∂t ∂t
This condition is satisfied if

∂ρ(r, t)
= −∇J(r, t),
∂t
where J(r, t) is the current density operator. In fact, the integral over the volume of
the left-hand side is also equal to minus the flux of the current out of the surface
of the sample. If the number of electrons is conserved, it means that the flux of the
current through the surface vanishes, which is the desired result. The equation
∂ρ(r, t)
+ ∇J(r, t) = 0 , (1.85)
∂t
is the continuity equation associated to the number of particles, which is a conserved
quantity. Similar equations can be derived for any conserved quantity.
1.6 Application: Electron Lattice Models and Emergence of

Magnetism
Let us consider the electron Hamiltonian in presence of the periodic potential pro-
vided by the ions in a lattice:

N
pi 2 1
H= + V (ri − R) + U (ri − r j )
2m 2 (1.86)
i=1 i R i=j
= Hkin + Hel−ion + Hel−el = H0 + Hel−el ,
1.6 Application: Electron Lattice Models and Emergence of Magnetism 21
where R are lattice vectors. We start by rewriting the Hamiltonian in the second
quantisation. For that purpose, we need to introduce a basis set of single-particle
wavefunctions. Since the Hamiltonian is spin independent, it is convenient to work
with factorised single-particle wavefunctions: φ(r, σ ) = φ(r) χσ . As a basis for the
space-dependent φ(r), we use Wannier orbitals φn,R (r) satisfying

dr φn,R1 (r)∗ φm,R2 (r) = δnm δR1 R2 , φn,R (r)∗ φn,R (r ) = δ(r − r ) ,
n,R
as well as
φn,R+R0 (r) = φn,R (r − R0 ) . (1.87)
†
Consequently, we associate to any wavefunction φn,R (r) χσ creation, cn,R,σ , and
annihilation, cn,R,σ , operators and introduce the Fermi fields
†
σ (r) = φn,R (r) cn,R,σ , σ† (r) = σ (r) .
n,R
Let us start by second quantization of the non-interacting part of the Hamiltonian,

H0 in (1.86):

2 ∇ 2
H0 = dr σ (r) −
†
+ V (r − R) σ (r)
σ
2m
R

= tRnm c† c
1 ,R2 n,R1 ,σ m,R2 ,σ
.
σ nm R1 ,R2
The matrix elements are

2 ∇ 2
tRnm
1 ,R2
= dr φn,R1 (r)∗ − + V (r − R) φm,R2 (r) (1.88)
2m
R
and satisfy
∗
tRnm
1 ,R2
= tRmn
2 ,R1
. (1.89)
By the property (1.87), it follows that

2 ∇ 2
∗
tRnm = dr φn,R1 +R0 (r) −
1 +R0 ,R2 +R0
+ V (r − R) φm,R2 +R0 (r)
2m
R

∗ 2 ∇ 2
= dr φn,R1 (r − R0 ) − + V (r − R) φm,R2 (r − R0 )
2m
R

∗ 2 ∇ 2
= dr φn,R1 (r) − + V (r + R0 − R) φm,R2 (r) = tRnm1 ,R2
,
2m
R
since

V (r + R0 − R) = V (r − R) .
R R
Let us now introduce the operators in the reciprocal lattice through the canonical
transformation
1 −ik·R
cn,k,σ = √ e cn,R,σ , (1.90)
V R
and its inverse
1 ik·R
cn,R,σ = √ e cn,k,σ , (1.91)
V k
where V is the number of lattice sites and k belongs to the reciprocal lattice, namely

eik·R = V δk0 , e−ik·R = V δR0 .
R k
One can check that the above transformation is indeed canonical:

1 −ik1 ·R1 ik2 ·R2
† †
cn,k1 ,σ1 , cm,k2 ,σ2
= e e c , c
n,R1 ,σ1 m,R2 ,σ2
V
R1 ,R2
1 −i(k1 −k2 )·R1
= δnm δσ1 σ2 e = δnm δσ1 σ2 δk1 k2 .
V
R1
Let us substitute (1.91) into (1.88):
1 −ik1 ·R1 ik2 ·R2 nm †

H0 = e e tR1 ,R2 cn,k c
1 ,σ m,k2 ,σ
.
V σ nm
R1 ,R2 k1 ,k2
We observe that
1
e−ik1 ·R1 eik2 ·R2 tRnm
1 ,R2
= e−ik1 ·R1 eik2 ·R2 tRnm
1 −R0 ,R2 −R0
V
R1 ,R2 R1 ,R2 ,R0
1
= e−ik1 ·(R1 +R0 ) eik2 ·(R2 +R0 ) tRnm
1 ,R2
V
R1 ,R2 ,R0
1 −i(k1 −k2 )·R0
= e−ik1 ·R1 eik2 ·R2 tRnm
1 ,R2
e
V
R1 ,R2 R0

= δk1 k2 e−ik1 ·(R1 −R2 ) tRnm
1 ,R2
= V δk1 k2 tknm
1
,
R1 ,R2
where we define
1 −ik·(R1 −R2 ) nm
tknm ≡ e tR1 ,R2 = e−ik·R tR,0
nm
. (1.92)
V
R1 ,R2 R
Since (1.89) holds, it also follows that

∗ ∗
tknm = eik·R tR,0
nm
= eik·R t0,R
mn
= eik·R t−R,0
mn
R R R
(1.93)
= e−ik·R tR,0
mn
= tkmn ,
R
namely the matrix tˆk , with elements tkmn , is Hermitian. The Hamiltonian is therefore
†
H0 = tknm cn,k,σ cm,k,σ . (1.94)
σ nm k
Since tˆk is Hermitian, we can write
tˆk = Û (k)† ˆk Û (k) → tknm = U † (k)ni i,k U (k)im ,
with Û (k)† Û (k) = Iˆ. Therefore, upon applying the canonical transformation

ci,k,σ = U (k)in cn,k,σ ,
n
the non-interacting Hamiltonian acquires a diagonal form

†
H0 = i,k ci,k,σ ci,k,σ . (1.95)
σ i k
The index i identifies the band, and i,k is the energy dispersion in the reciprocal
lattice: we have thus obtained the band structure.
We can formally write
1 −iR·k
ci,k,σ = √ e ci,R,σ ,
V R
thus introducing a new basis of Wannier functions. Since the Fermi field is invariant
upon the basis choice, then
1
σ (r) = φn,R (r) χσ cn,R,σ = √ φn,R (r) χσ eik·R cn,k,σ
R,n
V R,n,k
1
= √ φn,R (r) χσ eik·R U † (k)ni ci,k,σ
V R,n,i,k
1
= φn,R (r) χσ eik·(R−R ) U † (k)ni ci,R ,σ ≡ φi,R (r) χσ ci,R ,σ ,
V
R,R ,n,i,k R
thus implying the following expression of the new Wannier functions:

1 ik·(R−R ) †
φi,R (r) = e U (k)ni φn,R (r) . (1.96)
V
n,R,k
Going back to the Hamiltonian in the diagonal basis, we can also rewrite it as
†
H0 = i,k ci,k,σ ci,k,σ
σ i k
1 ik·(R1 −R2 ) †
= i,k e ci,R1 ,σ ci,R2 ,σ (1.97)
V σ
i k R1 ,R2
†
≡ tRi 1 ,R2 ci,R c
1 ,σ i,R2 ,σ
,
σ i R1 ,R2
namely like a tight-binding Hamiltonian diagonal in the band index. Once we know
the band structure, the ground state of the non-interacting Hamiltonian is simply
obtained by filling all the lowest bands with the available electrons. If the highest
occupied band is full, the model is a band insulator, otherwise is a metal. In particular,
since each band can accommodate 2V electrons, V of spin up and V of spin down,
a necessary condition for a band insulator is to have an even number of available
electrons per unit cell. This is not sufficient since the bands may overlap.
1.6.1 Hubbard Models
Hereafter, we consider the case of a putative metal, in which the highest occupied
bands are partly filled. In many physical situations, the Wannier orbitals φi,R (r) of
(1.96) for those bands, the conduction ones, are quite delocalised; hence, the lattice
vector label R looses its physical meaning. In those cases, although formally exact,
the tight-binding Hamiltonian (1.97) is of little use since the hopping matrix elements
tRi 1 ,R2 are very long ranged.
However, there is a wide class of materials, commonly called strongly correlated
ones, where the tight-binding formalism is meaningful. There, the valence bands
derive from d or f orbitals of transition metals, rare earth or actinides, and the Wan-
nier orbitals keep noticeable atomic character, thus leading to short-range hopping
elements tRi 1 ,R2 in (1.97). Let us consider just the above circumstance and write down
the left-over electron-electron interaction on the Wannier basis. We further assume
that there is only one conduction band well separated from lower and higher ones, so
that we can safely neglect interband transition processes due to interaction and just
project the latter onto the conduction band. For that reason, we hereafter drop the
band index i, so that the Wannier orbitals of the conduction band are simply denoted
as φR (r), and the tight-binding Hamiltonian projected on the same band reads
†
H0 = tR1 ,R2 cR c
1 σ R2 σ
. (1.98)
σ R1 ,R2
The interaction term is, correspondingly,
1 †
Hint = UR1 ,R2 ;R3 ,R4 cR c† c c
1 σ1 R2 σ2 R3 σ2 R4 σ1
, (1.99)
2σσ
1 2 R1 ,R2 ,R3 ,R4
where

UR1 ,R2 ;R3 ,R4 = drdr φR1 (r)∗ φR2 (r )∗ U (r − r ) φR3 (r ) φR4 (r) . (1.100)
Let us make use of our assumption of well-localised Wannier orbitals, namely that
φR (r) decays sufficiently fast with |r − R|. Within this assumption, the leading
matrix element is when all lattice sites are the same:
UR,R;R,R ≡ U . (1.101)
This term gives rise to an interaction
U † †
HU = cRσ cRσ cRσ cRσ = U n R↑ n R↓ , (1.102)
2 R σσ R
since cRσ cRσ = 0 by the Pauli principle, where the operator
†
n Rσ = cRσ cRσ ,
counts the number of spin-σ electrons at site R.

The interaction term (1.102), which simply describes an on-site Coulomb repul-
sion, plus the hopping (1.98) yield the so-called Hubbard model [1,2], which is the
prototype of strongly correlated lattice models.
Let us continue and consider in (1.100) two other cases: either (1) R1 = R4 ,
R2 = R3 with R2 nearest neighbour of R1 or (2) R1 = R3 , R2 = R4 still with R2
nearest neighbour of R1 . In case (1), we obtain

U1 n R n R ,
<RR >

where n R = σ n Rσ , < RR > stands for the sum over nearest neighbour sites,
and U1 = UR,R ;R ,R . In case (2), we instead find
† †
† †
U2 cRσ cR σ cRσ cR σ = −U2 cRσ cRσ cR σ cR σ ,
<RR > σ σ <RR > σ σ
where U2 = UR,R ;R,R . One can easily show that

† † 1
cRσ cRσ cR σ cR σ = n R n R + 4 SR · SR , (1.103)
2
σσ
where the spin-density operator
1 †
SR = cRα σ αβ cRβ ,
2
αβ
being σ = (σ1 , σ2 , σ3 ), with σi ’s the Pauli matrices. Therefore, upon defining V =

U1 − U2 /2 and Jex = −2U2 , the two nearest neighbour interaction terms lead to

Hn.n. = V n R n R + Jex SR · SR . (1.104)
<RR > <RR >
Since U1 > U2 > 0, the first term describes a nearest neighbour repulsion, while
the second a spin exchange which tends to align the spin ferromagnetically since
Jex < 0, so called direct exchange. Therefore, although we have started from a spin-
independent interaction, projecting onto the Wannier basis makes a spin interaction
emerge, thus showing in a simple way how magnetism raises out of the charge
Coulomb repulsion.
1.6.2 Mott Insulators and Heisenberg Models
Let us summarise the approximate Hubbard Hamiltonian which we have so far

derived, by further assuming a nearest neighbour hopping:
†

H = −t cRσ cR σ + H .c. + U n R↑ n R↓
σ <RR > R
(1.105)
+V nR n R + Jex SR · S
R ≡ H0 + Hint .
<RR > <RR >
By construction U > V > 0 and Jex < 0. We consider the case in which the number
of conduction electrons is equal to the number of sites N , i.e., density equal to half-
filling. In the absence of interaction, the hopping forms a band that can accommodate
2N electrons while there are just N of them: the band is therefore half-filled and the
system metallic.
Let us analyse the opposite case of a very large U t, V , |Jex |. In this case, it is
better to start from the configuration which minimises the Coulomb repulsion U and
treat what is left in perturbation theory. That lowest energy electronic configuration
is the one in which each site is singly occupied. Indeed, the energy cost in having
just an empty site and a doubly occupied one instead of two singly occupied sites is
given by
E(2) + E(0) − 2E(1) = U ,
and it is much larger than the hopping energy t. In this situation, the model describes
an insulator but of a particular kind. Namely, the insulating state is driven by the
strong correlation, while the conventional counting argument instead predicts a metal.
This correlation-induced insulator is called a Mott insulator after Sir Nevil Mott [3].
Therefore, as the strength of U increases with respect to the bandwidth, which is pro-
portional to t, an interaction-driven metal-to-insulator transition, commonly referred
to as a Mott transition, must occur at some critical Uc .
However, the configuration with one electron per site is hugely degenerate, since
the electron can have either spin up or down, thus a degeneracy 2 N that is going to be
split by the other terms in the Hamiltonian. The nearest neighbour interaction V is
not effective in splitting the degeneracy, contrary to the direct exchange that favours
a ferromagnetic ordering.
However, this is not the only source of spin correlations. Indeed, also the hop-
ping term is able to split the degeneracy within second order in perturbation theory.
Specifically, if P is the projector onto the degenerate ground state manifold with
energy E 0 = 0, at second order, the hopping Hamiltonian H0 generates an effective
spin exchange operator, called super-exchange.
1
Hs−ex = P H0 | n n | H0 P , (1.106)
n
E0 − En
where | n is an intermediate excited state with energy E n . Because of the projector,

the only allowed hopping processes are those where one electron hops to a nearest
neighbour site and then hops back to the initial site. It follows that the intermediate
states allowed in the sum are just those where one site is doubly occupied and a
nearest neighbour one empty, thus E n = U independent of n and
1 1
Hs−ex = − P H0 | n n | H0 P = − P H0 H0 P
U n U
2
t † † † †
=− cRσ cR σ c R c
σ Rσ + c c c c
R σ Rσ Rσ R σ (1.107)
U
σ σ <RR >
2t 2 † †
=− cRσ cR σ cR σ cRσ ,
U
σ σ <RR >
which, through (1.103) and since n R = 1, is simply
4t 2 1

Hs−ex = SR · SR − . (1.108)
U 4
<RR >
Therefore, the second-order perturbation theory in the hopping gives rise to an anti-
ferromagnetic spin super-exchange. In conclusion, the effective Hamiltonian at large
U describes localised spin-1/2 (each site occupied by one electron) coupled by the
overall spin exchange (dropping constant terms)

H H eis = J SR · SR , (1.109)
<RR >
where
4t 2 4t 2
J = Jex += −2U2 + .
U U
The spin exchange may be either ferromagnetic or antiferromagnetic depending on
the strength of the direct-exhange constant with respect to the super-exchange one.
The effective Hamiltonian (1.109) thus corresponds just to a spin-1/2 Heisenberg
model.
1.7 Application: Spin-Wave Theory
When the hopping involves further neighbour sites and more than a single band, as
it is commonly the case for magnetic ions, the corresponding Heisenberg model has
the general expression

H= JR,R SR · SR . (1.110)
RR
The Hamiltonian to be Hermitian requires that JR,R = JR ,R ∈ R, while transla-
tional symmetry that
JR,R = JR−R0 ,R −R0 , ∀ R0 . (1.111)
For instance, taking R0 = R or R0 = R, and using the property JR,R = JR ,R ,
JR,R = JR−R ,0 = J0,R −R = JR −R,0 . (1.112)
We can introduce Fourier transforms defined through

1 iq·R
S(q) = e−iq·R SR , SR = e S(q) , (1.113)
N q
R
where N is the number of lattice sites, and note that
1 −iq·R iq ·R 1 1 −iq·R iq ·R

Jq,q = e e JR,R = e e JR,R
N
N N
RR R0 RR
1 1 −iq·R iq ·R
= e e JR−R0 ,R −R0
N N
R0 RR
1 −i(q−q )·R0 1 −iq·R iq ·R
= e e e JR,R
N N
R0 RR
1 −iq·(R−R )
= δq,q e JR,R = δq,q e−iq·R JR,0 ≡ δq,q J (q) .
N
RR R
(1.114)
1.7 Application: Spin-Wave Theory 29
It follows that the Hamiltonian in Fourier space is simply
1
H= J (q) S(−q) · S(q) . (1.115)
N q
Moreover, since SR = S†R ,

S(q)† = eiq·R SR = S(−q) , (1.116)
R
and, since JR,R ∈ R,

J (q)∗ = eiq·R JR,0 = J (−q) = eiq·R JR−R,0−R = eiq·R J0,−R
R R R

−iq·R
= eiq·R
J−R,0 = e J R,0 = J (q) ,
R R
(1.117)
namely J (q) = J (−q) ∈ R. In conclusion,
1
H= J (q) S(q)† · S(q) . (1.118)
N q
1.7.1 Classical Ground State
Let us first assume that the spin SR is a classical vector satisfying
SR · SR = S 2 , ∀ R , (1.119)
so that the Hamiltonian becomes

1
H= J (q) S(q)∗ · S(q) . (1.120)
N q
The classical ground state is actually identified by a vector Q and a classical spin
configuration
S(Q) = S(−Q)∗ = 0 , S(q) = 0 , ∀ q = (Q, −Q) . (1.121)
To prove that, let us assume it is indeed the case, and first identify what Q is. Under
that assumption, and if Q = −Q, where the equality holds apart from a reciprocal
lattice vector G, then

1 ∗ −iQ·R
SR = S(Q) e iQ·R
+ S(Q) e . (1.122)
N
In order to verify (1.119), We have to impose that
S(Q) · S(Q) = S(−Q) · S(−Q) = 0 ,

N2 2 (1.123)
S(Q)∗ · S(Q) = S(−Q) · S(Q) = S ,
2
which can be satisfied by taking any two orthogonal real unit vectors, u1 and u2 ,
thus ui · u j = δi j , and writing
SN SN
S(Q) = u1 − i u2 , S(−Q) = S(Q)∗ = u1 + i u2 . (1.124)
2 2
With that choice
SR = S u1 cos Q · R + S u2 sin Q · R . (1.125)
If instead Q = −Q + G, i.e., 2Q = G, then S(Q) = S(−Q) ∈ R, so that, since
1
Q·R = G · R ≡ π n R (G) , (1.126)
2
with n R (G) integer, then
1 1
SR = S(Q) cos Q · R = S(Q) (−1)n R (G) . (1.127)
N N
Equation (1.119) is now satisfied if
S(Q) = N S u , u · u = 1. (1.128)
In both cases, 2Q = G and 2Q = G, the energy of the classical configuration is
1
E(Q) = J (Q) S(Q)∗ · S(Q) + J (−Q) S(−Q)∗ · S(−Q) = N S 2 J (Q) ,
N
(1.129)
which is minimised by Q such that
J (Q) = min J (q) . (1.130)

q
We note that the energy does not depend on the choice of u1 and u2 , if 2Q = G, or u
if 2Q = G, which reflects the invariance of the Hamiltonian under spin O(3) rotations
and, if 2Q = G, an additional U (1) symmetry related to S(Q) being complex.
Now suppose we choose a combination of two Q, Q1 and Q2 , assuming, just for
simplicity, that both satisfy 2Qi = G i . In this case,
1 1
SR = S(Q1 ) cos Q1 · R + S(Q2 ) cos Q2 · R . (1.131)
N N
1.7 Application: Spin-Wave Theory 31
The equation (1.119) is fulfilled if

S(Q1 ) = N S cos φ u1 , S(Q2 ) = N S sin φ u2 , ui · u j = δi j , (1.132)
in which case the energy of such configuration reads

E Q1 , Q2 , φ = N S 2 cos2 φ J Q1 + sin2 φ J Q2 . (1.133)
It is easy to realise that this energy is always higher than choosing the single Q that
minimises J (q), thus proving the claim (1.121).
We end by observing that the average over the Brillouin zone of J (q), i.e.,
1
J (q) ≡ JR,R = 0 , (1.134)
N q
is just the on-site exchange, which is zero. Therefore, if the average over the Brillouin
zone vanishes, then J (q) must have both positive and negative values; hence, the
minimum, i.e., J (Q1 ), is necessarily negative and so is the classical energy.
1.7.1.1 Symmetry Considerations

The lattice Hamiltonian is generally invariant under a space group S, i.e., under the
symmetry transformations g ∈ S. It follows that, if J (q) is minimised by Q1 , then
J g(Q1 ) = J (Q1 ) (1.135)
namely is minimum for the whole set of vectors obtained by applying any g ∈ S
to Q1 . In general g(Q1 ) for all g ∈ S generates a finite number of inequivalent Qi ,
i = 1, n ∗ , called the star of Q1 . We emphasise that the Hamiltonian (1.110) implicitly
assumes, besides spin O(3), also inversion symmetry, in which absence additional
terms would be present, as, e.g., SR ∧ SR · SR . It follows that S at least includes
inversion, Q1 → −Q1 , besides the identity.
If n ∗ > 1, the classical configuration choosing any of the Qi , e.g., Q1 , breaks the
space group symmetry. Specifically, such classical configuration will be invariant
under a subgroup S ⊆ S that contains all g ∈ S such that g (Q1 ) = Q1 , apart from
a reciprocal lattice vector.
If the star of Q contains just Q, which also implies that −Q ≡ Q apart from a
reciprocal lattice vector G, the classical configuration still seems to break translation
symmetry, which belongs to the space group, since
G·R
SR = N S u cos Q · R = N S u cos (1.136)
2
is not invariant under translation unless G = 0, namely Q = 0 = equal to the
point. However, if we combine the translation TR0 such that
G · (R + R0 ) G·R
cos = − cos , (1.137)
2 2
with time reversal, T (SR ) = −SR , we do recover full translational symmetry.
In conclusion, any classical configuration that minimises the energy is unavoid-

ably not invariant under spin O(3) symmetry, time reversal symmetry, and, eventually,
under the space group symmetry of the Hamiltonian, which entails the existence of
a whole manifold of different classical configurations with the same energy.
1.8 Beyond the Classical Limit: The Spin-Wave Approximation
In reality SR is not a classical vector, but a quantum operator whose components

satisfy the commutation relations

SiR , S jR = i δR,R i jk SkR , (1.138)
with i jk the antisymmetric tensor and taking = 1, which translates, for the Fourier
transformed operators, into

Si (q) , S j (q ) = i i jk Sk (q + q ) . (1.139)
Our aim here is to add the quantum mechanics implicit in the non-trivial commutation
relations under the assumption that the quantum fluctuations do not alter completely
the classical ground state. For that, we associate to the chosen classical configuration,
assuming for simplicity that J (q) is minimum for q = Q such that 2Q = G, the
vacuum | 0 of the quantum fluctuations. Specifically, if we choose the classical
state such that
Siclass (q) = N S δi,z δq,Q , (1.140)
namely taking u = (0, 0, 1), we define | 0 through
0 | Si (q) | 0 = Siclass (q) = N S δi,z δq,Q . (1.141)
We thus write
Si (q) = Siclass (q) + δSi (q) , (1.142)
where δSi (q) is an operator that describes the quantum fluctuation corrections, while
Siclass (q) just a c-number, and satisfies
0 | δSi (q) | 0 = 0 . (1.143)

√
We adopt a semiclassical approach, and thus assume that δ S(q) ∼ O( N ), unlike
Szclass (Q) ∼ N , and proceed consistently. For instance,

Sx (q) , S y (q ) = δSx (q) , δS y (q )
= i Sz (q + q ) = i δq+q ,Q N S + δSz (q + q ) i δq+q ,Q N S + · · · ,
(1.144)
1.8 Beyond the Classical Limit: The Spin-Wave Approximation 33
where . . . mean corrections subleading in N . It follows that, at leading order in N ,

the above commutation relation is fulfilled if
√ √
δSx (q) N S xq , δS y (q) N S pq−Q , (1.145)
† †
where xq = x−q and pq = p−q are conjugate variables satisfying

xq , pq† = i δq,q , (1.146)
and such that the expectation values on the vacuum,
0 | xq | 0 = 0 | pq | 0 = 0 . (1.147)
It follows that, if we write these conjugate variables in terms of bosonic operators,

i.e.,
1 †
i †

xq = √ bq + b−q , pq = − √ bq − b−q , (1.148)
2 2
then | 0 is their vacuum.
However, we have to enforce two more commutation relations, specifically

Sx (q) , Sz (q ) = −i S y (q + q ) , S y (q) , Sz (q ) = i Sx (q + q )
(1.149)
which imply that
√
Sx (q) , Sz (q ) = δSx (q) , δSz (q ) = N S xq , δSz (q )
√
= −i δS y (q + q ) = −i N S pq+q −Q ,
√ (1.150)
S y (q) , Sz (q ) = δS y (q) , δSz (q ) = N S pq−Q , δSz (q )
√
= i δSx (q + q ) = i N S xq+q ,
namely

xq , δSz (q ) = −i pq+q −Q , pq−Q , δSz (q ) = i xq+q . (1.151)
The solution for δSz (q) with the condition (1.143) can be readily shown to be
1 †
δSz (q) = − xk xk+q−Q + pk† pk+q−Q − δq,Q , (1.152)
2
k
since
1
0 | xq† xq | 0 = 0 | pq† pq | 0 = δq,q . (1.153)
2
In conclusion, at leading order in an expansion for large N , we can represent the spin
operators as
√
Sx (q) = N S xq ,
√
S y (q) = N S pq−Q ,
(1.154)
1 †
Sz (q) = N S δq,Q − xk xk+q−Q + pk† pk+q−Q − δq,Q .
2
k
We note that the expansion also looks like an expansion for large S. This is not surpris-
ing. Indeed, if we normalise the spin operators, SiR → siR /S, then the commutation
relations for the operators siR , which are O(1) in S, read
1
siR , s jR = i i jk skR −−−→ 0 , (1.155)
S S→∞
showing that the spin operators become classical vectors in the large S limit.2
1.8.1 Hamiltonian of Quantum Fluctuations
Now we substitute the expressions in (1.154) in the Hamiltonian
1
H= J (q) S(q)† · S(q) , (1.158)
N q
2 There are actually more rigorous ways to implement the mapping from spin operators to bosonic
annihilation and creation ones. However, we preferred the less rigorous approach presented, since
it is more flexible and can be adopted in more general cases than just spin models. However, for
completeness, we here mention one of those rigorous mappings, known as the Holstein-Primakoff
transformation.
Suppose that one of the equivalent classical configurations corresponds to the spin SR at site R
polarised along the z-direction. One can readily demonstrate that the following way of writing spin
operators in terms of bosonic one does preserve the spin commutation relations:
SRz = S − dR† dR , SR+ = 2S − dR† dR dR , SR− = dR† 2S − dR† dR , (1.156)
where the classical configuration is the vacuum of the bosonic operators dR and dR† . The square
roots in (1.156) assure that it is not possible to create more than 2S bosons at any given site, so that,
correctly, SRz = −S, . . . , S. The spin-wave theory is recovered by assuming that the number of
excited bosons is negligible with respect to S, namely that the actual quantum ground state is close
√
to the classical one. In that case 2S − dR† dR 2S, and thus
√ √
SRz = S − dR† dR , SR+ 2S dR , SR− = 2S dR† . (1.157)
Substituting those expressions in the Hamiltonian reproduces the spin-wave theory we discuss.
Another random document with
no related content on Scribd:
the first question of the sheikh’s was, as to the distance of our
country from India: and when told it was four months by sea, he said,
“What could induce you to go so far from home—to find it out, and
fight with the people?” We replied, “that we had plenty of ships, and
were great lovers of discovery; that the French and the Dutch had
been there before us; and we were always jealous of our neighbours
doing more than ourselves.” “And now it is all yours,” said he, “and
governed by your laws!” Our reply was, “that we only kept
possession of the part near the sea—that their own laws were in full
force—but that even Mussulmans often preferred the English laws to
their own.” “Wonderful!” said he, “and you went at first with only a
few ships, as friends?” “We are friends now,” said I, “and by trade
have not only made ourselves rich, but the natives also.” “By God!”
said the Marroquin, “they eat the whole country—they are no friends:
these are the words of truth.” We had then a few remarks (not good-
natured ones) as to the right of dictating to Algiers and the other
Barbary powers. Algiers we described as unfaithful to their word, and
little better than pirates.
Sketch by Major Denham. Engraved by E. Finden.
LANCERS OF THE SULTAN OF BEGHARMI.

Published Feb. 1826, by John Murray, London.
Aug. 11.—Soon after daylight, Karouash, with Hadgi Mustapha,

the chief of the Shouaas, and the sheikh’s two nephews, Hassein
and Kanemy, came to our huts. Hadgi Mustapha had been one of
the original four hundred who commenced the liberation of Bornou
from the Felatahs. They were attended by more than a dozen slaves,
bearing presents for us, for King George, and the consul at Tripoli. I
had applied for a lebida[54], after seeing those taken from the
Begharmis: the sheikh now sent a man, clothed in a yellow wadded
jacket, with a scarlet cap, and mounted on the horse taken from the
Begharmis, on which the sultan’s eldest son rode. He was one of the
finest horses I had seen; and covered with a scarlet cloth, also
wadded. “Every thing,” Hadgi Mustapha said, “except the man, is to
be taken to your great king.” He also brought me twelve very
beautiful tobes, of every manufacture, from Nyffé to Loggun and
Waday, four parrots, and a box of zibet. For the consul he also sent
six tobes, and a small box of zibet, worth thirty or forty dollars, with
two parrots.
August 13.—The long expected kafila arrived from Soudan, which
was a signal for our departure: they had been fifty days on the road
from Kano, in consequence of the waters; and had they been
delayed much longer, the season would have been so far advanced
as to have prevented the departure of all those merchants that had
many slaves: going, as they do, poor creatures, nearly naked, the
cold of Fezzan, in the winter season, kills them by hundreds. With
the Soudan kafila came Khalifa, a Moor and a fighi; he had been at
Saralo, as he called Sierra Leone, and desired to be brought to me,
as he knew English. He certainly knew enough to convince me of his
truth, when he asserted that he had met my countrymen. “Gun, cap,
and water!” he kept continually saying: and my Bornou friends were
not a little surprised when I told them it was the language of my
country. He spoke greatly of some person he called the Doctor[55],
whom he had seen at Bammakoo and Bunjalow, a good looking
man, with a red beard, and long projecting nose, with bad front teeth.
He gave away many things, wrote a great deal, and was much liked
by all the people. Two persons were with the Doctor, whom he
believed were French, and had come from Ender; one was called
Gentleman, and the other Fausta, or Forster; “but they held their
heads down, and did not talk to the people like the Doctor,” said
Khalifa. “The Doctor,” he said, “wanted to go to Sego, but the sultan
would not allow him to come to his country, and would not even look
at the presents he sent him, as he feared they had charms which
would kill him, either by the sight or smell. He, however, sent him
slaves, and horses, which he, the Doctor, also returned, saying that
he wanted nothing but to see the country and the rivers. The sultan
of Sego replied, he had heard that his (the doctor’s) king had water
all round his country, and he might go and look at that. Khalifa said
the Sego people were Kaffirs, and knew not God, therefore were
afraid of Christians; but the Moors knew them, and liked them. When
the present king of Sego’s father was alive, he, Khalifa, then a boy,
remembered Christians going to D’Jennie and Timboctoo, and
hearing that the Tuaricks killed them in their boat near Nyffé[56].”
On Monday, the 20th of the Mohamedan month del Khadi, and the
16th of August, we took our final leave of Kouka, and not without
many feelings of regret, so accustomed had we become, particularly
myself, to the people. In the morning I had taken leave of the sheikh
in his garden, when he had given me a letter to the king, and a list of
requests: he was all kindness, and said he had only one wish, which
was that I might find all my friends well, and once more return to
them. He gave me his hand at parting, which excited an involuntary
exclamation of astonishment from the six eunuchs and Karouash,
who were the only persons present.
I preceded the kafila for the following reason: I had, ever since my
return from Tangalia, determined to attempt the east side of the
Tchad, by Lari, previous to returning home. Many had been the
objections, many the reports of danger from the Waday people and
Amanook, who had now boldly forsaken the lake, and was
encamped at no great distance from Barca Gana, to whom he twice
paid a night visit, and had been beaten back. I, however, told the
sheikh I could take no present, or promise to the execution of any
commission, unless this duty was accomplished, or at least until I
had done my utmost, and that I would take care not to go into
danger. Bellal, my old companion, was once more appointed to
attend me, and we moved with two camels, lightly laden; for the
more train, always the more trouble and the more expense. All my
friends then in Kouka mounted to escort me from the town: the
women assembled outside the gate, and screamed an adieu; and I
am persuaded our regrets were mutual.
About midnight, while we were all asleep at Dowergoo, a
despatch came to say, that the skin of a camelopard had been
brought to Kouka, which the sheikh had procured for me. Columbus,
therefore, returned to prepare it for preservation, while we moved on
to N’Gortooah: he came up again in the evening, and reported, that
though small, it was a fine specimen. On Wednesday we slept at
Kaleeluwha, and on the 23d came once more on the Yeou, now a
considerable stream, full of water, and running towards the Tchad, at
the rate of three miles an hour[57]. My feelings on seeing this river for
a second time were very different to what they had been when I first
looked on its waters. We then had an escort of two hundred men,
and yet could not feel ourselves in perfect safety one hundred yards
from our tents. Now I had only one attendant—the people about me
were all natives, and I wandered about the banks of the river with
perfect freedom, and slept with my tent door open, in as great
security as I could have done in any part of England, had I been
obliged so to travel. Other feelings also obtruded themselves; I was
about to return home, to see once more dear friends, and a dear
country, after an absence of nearly three years, on a duty full of
perils and difficulties: two out of four of my companions had fallen
victims to climate and disease, while those who remained were
suffering, in no small degree, sickness and debility from the same
causes: I was in health, and notwithstanding the many very trying
situations, in which we had all been placed, some of them of great
vexation and distress, yet had we been eminently successful.
In the afternoon Bellal accompanied me down the river, about
nine miles, where, increasing in width to about one hundred yards, it
flows into the Tchad, with a strong and deep current of water. On its
banks are five considerable villages of Kanemboos, called Ittaquoi,
Belagana, Afaden, Yeougana, and Boso. At Belagana, the sheikh
has a large inclosure of huts, within a wall, where he generally has
from five hundred to eight hundred slaves of both sexes, under the
charge of four eunuchs, who are employed in preparing cotton, and
spinning the linen (gubbok), of which the tobes are made.
The manner of fishing in the Yeou, a very considerable source of
commerce to the inhabitants of its banks, must not be omitted: dried
fish from the Yeou is carried to all the towns to the south-west, quite
as far as the hills; and at this season they are usually taken in great
numbers. The Bornouese make very good nets of a twine spun from
a perennial plant called kalimboa: the implements for fishing are
ingenious, though simple: two large gourds are nicely balanced, and
then fixed on a large stem of bamboo, at the extreme ends; the
fisherman launches this on the river, and places himself astride
between the gourds, and thus he floats with the stream, and throws
his net. He has also floats of cane, and weights, of small leathern
bags of sand: he beats up against the stream, paddling with his
hands and feet, previous to drawing the net, which, as it rises from
the water, he lays before him as he sits; and with a sort of mace,
which he carries for the purpose, the fish are stunned by a single
blow. His drag finished, the fish are taken out, and thrown into the
gourds, which are open at the top, to receive the produce of his
labour. These wells being filled, he steers for the shore, unloads, and
again returns to the sport.
25th.—At Woodie I met Barca Gana, Ali Gana, and Tirab, with
their forces, on their return from Kanem: they had been out, some of
them five months, had made the complete tour of the lake, and were
in a sad plight, with scarcely thirty horses left, having literally fought
their way: Amanook had twice attacked them, and had sent off all his
flocks and women to Begharmi, but had not gone himself, and they
were so reduced as not to be able to do any thing to prevent him.
They were so badly off for every thing, that they were obliged to
come down on Kanem for supplies: the people refused them any
assistance, and after being half starved, they were obliged to make a
running fight of it, and get home as well as they could. The Kanem
people were all in a state of mutiny, and the Dugganah had gone off
towards Waday. This was sorry news; Bellal wanted to turn back. I
saw Barca Gana, who said, “It is dangerous, but I think you may go
on if you wish it. I would give you eighteen men, but you are better
without them: they expect the sheikh, and going with Bellal, wanting
but little, and paying for that, for the crops have failed them, you will
have little to fear; but you cannot go beyond the Bornou Kanemboos
with less than one hundred men.”
From a Sketch by Major Denham. Engraved by E. Finden.
MANNER OF FISHING IN THE RIVER YEOU.

Published June 1826, by John Murray, London.
With this advice I determined on going on, and after halting the
whole of Sunday, on the 27th I proceeded. Barca Gana had, on his
return, bivouacked in the wadey where once the Bahr-al-Ghazal ran
from the Tchad; the valley is now filled with trees and grass. This
was the fourth time Barca Gana had raised his tents in the same
place[58], as the sheikh had before informed me. This valley runs
between Kangara and N’Gussum, less than twenty miles from
Tangalia. We were overtaken by so dreadful a storm, that we halted,
and pitched the tent on a high sand-hill within five miles of N’Gygami.
Near this hill we had a beautiful view of the open lake, with several
floating islands, when the storm cleared away. The Biddomahs are
constantly landing hereabouts; and we met some poor people who
had been robbed of their whole flock of goats, with their daughter,
only the day before—indeed no single travellers can pass this road.
Towards evening, we saw their canoes in the offing; and below us, in
the low grounds, three Biddomahs making for the lake:—they saw
us, and quickened their pace. For safety we all slept outside the huts
of N’Gygami: this ground is the highest part of the borders of the
lake, and here deep water commences immediately off the shore,
while, in some parts, miles of marsh are to be waded through
previous to arriving at the lake. Tuesday, we made Lari, where we
were to find a Malem fighi, whom the sheikh had ordered to proceed
with us.
Aug. 29.—Moved from Lari. Here we found four men, with a chief
whom Barca Gana had left at Kuskoua, returning, as the people
would give them no provisions. It was near sunset when we reached
Zogany, thirty miles; the country was quite a flat, covered with a plant
resembling a heath which I had seen nowhere else; and in many
parts I observed incrustations of trona. This heath is called kanuskin:
the camels eat it; and in the neighbourhood of trona it is generally
found.
Aug. 30.—After a night of intolerable misery to us all, from flies
and mosquitoes, so bad as to knock up two of our blacks, we
mounted and advanced; and leaving our tents, for Bellal would not
carry them on, we proceeded to Garouah and Mabah;—which are
full of people, and though annually pillaged by Tuaricks and Tibboo
Arabs, yet still they will not quit their native soil. The character of the
country here, which is different to the south or west sides of the lake,
extends to Gala, where the land is again varied, and a little higher:
for many miles on this side we had one continued marsh and
swamp. I was at the northernmost part of the lake, and pursuing a
course first to the west, and then to the eastward of south, for five or
six miles, nearly up to the body of the horse in water, and with reeds
and high grass overtopping our heads, I at last got a sight of the
open lake. We disturbed hippopotami, buffaloes, enormous fish, and
innumerable hosts of insects. At the commencement of the water it
had a taste of trona, which, as we advanced, became gradually
sweeter: indeed, all the people say, when you ask if this water you
drink so strong of trona is the lake? La! la! inki kora kora kitchi
engobboo tilboo baco. (No! no! the water of the great lake is very
sweet, no trona).
Completely fatigued, we returned to the village of Chirgoa, near
which our tents were pitched: this was a most distressing day, and
we had been on our horses nearly thirteen hours. Garouah is twelve
miles from Zogany, and Mabah twenty. We were some way in
advance of the latter, but to Kuskoua I could not induce my guide to
venture; and so many proofs had I seen, not only of his bravery, but
his desire to satisfy my curiosity, that I could not doubt his fears were
just. Notwithstanding our fatigue, no rest could we obtain, and
another night was passed in a state of suffering and distress that
defies description: the buzz from the insects was like the singing of
birds; the men and horses groaned with anguish; we absolutely
could not eat our paste and fat, from the agony we experienced in
uncovering our heads. We at last hit upon an expedient that gave us
a little relief: as they came at intervals, in swarms, we thought they
might also be driven off in the same quantities; and we found, by
occasionally lighting a line of fires with wet grass, to windward of our
tents, that the smoke carried off millions, and left us a little at ease. I
do not think our animals could have borne such another night; their
legs and necks were covered with blood, and they could scarcely
stand, from the state of irritation in which they had been kept for so
many hours.
On Friday we returned to Lari by the lower road, where there are
frequent large detached pieces of water, strongly impregnated with
trona. On the road, to-day, we fell in with a tribe of the Biddomah,
who had, during the last three months, taken up their abode on the
sheikh’s land, and asked for what was instantly granted them,
permission to remain. Internal wars cause these fallings off of one
tribe from another, which the sheikh encourages: only one of their
chiefs could ever be induced to proceed so far as Kouka. The sheikh
takes no notice, and suffers them to do as they please: he sent them
tobes, and a fighi, and desired they would learn to pray: and they are
now beginning, as my guide said, to have the fear of God. They were
the most savage beings I had seen in the shape of men, except the
Musgowy; and we had sat some time under a tree before they would
come near us. The men, until they are married, wear their hair, and
collect as many beads and ornaments as they can, which they wear
round their necks; their hair is long and plaited, or twisted in knots;
they have ear-rings also: and this collection of beads and metal is
always given to the wife on their marriage. The upper part of the face
is very flat, and the eye sunk; they have large mouths, and long
necks; a sulky, reserved look about them, any thing but agreeable:
they have no style of salutation like other negroes, who greet
strangers over and over again, sitting down by them:—these stand
up, leaning on a spear, and look steadfastly at you without speaking.
I gave a little boy some white beads, which were directly tied round
his neck, I suppose as the commencement of his marriage portion.
They, however, at length, produced some sour milk; and some of
them came round my horse when I mounted, and nodded their
heads at me when I rode off, which I returned, much to their
amusement.
When we arrived at Lari, which was comparatively free from flies,
the horses lay down, and, stretching themselves out, fell asleep in a
way, and with an expression of enjoyment, I never saw animals do
before, and did not look for their nose-bags until after midnight. We
here found that one of Barca Gana’s people had the night before lost
his horse, which had been stolen by the Biddomahs we saw on the
look out.
Kanem, the most persecuted and unfortunate of negro countries,
was daily becoming more miserable; they were pillaged alternately
by the Fezzaneers, the Tuaricks, and the Waday people. Between
the latter and the sheikh they hung for protection, and from neither
could they obtain what they sought: the country was becoming
abandoned, and the villages deserted, part taking refuge in Waday,
and part in the sheikh’s dominions: the land communication between
Bornou and Kanem was too difficult and distant, either by the south
or north, for the sheikh to render them any effectual support. An
army almost every year went to Kanem for this purpose, but they
usually returned with the loss of horses, camels, and men, and were
seldom able much to annoy their enemies, the Wadays. This year his
expedition had been upon a larger scale, and his losses were
greater than on any former occasion.
In consequence of the waters, which fill the rivers at this season,
the ford across the Shary had become impassable, and they were
therefore obliged to return home through Kanem. Not the least
assistance would any of the towns give them, except Gala, and a
more wretched state I never saw men in: some of the chiefs on foot,
without horses; and those who were mounted, bestriding sorry
animals, with torn appointments: they all said, fighting without the
sheikh was useless, as he alone could lead them to victory.
Drawn by Major Denham. Engraved by E. Finden.
ENCAMPMENT NEAR WOODIE.

AND PART OF THE LAKE TCHAD.
Published Feb. 1826, by John Murray, London.
Sept. 3.—I had now been six days at Woodie, waiting the arrival
of my companions with our camels, and the kafila of merchants
whom we were to accompany to Fezzan. Woodie is no very pleasant
place of sojourn, as the Biddomah have a sort of agreement with the
kaid to be allowed to plunder all strangers and travellers, provided
the property of the inhabitants is respected. We were, however, told
to be on our guard, and not without reason. Our tents were pitched
near each other, and a look-out kept up the whole night,
notwithstanding which they paid us a visit, during a storm of thunder
and rain, and from the entrance of Bellal’s tent, only eight paces from
my own, stole both his horses. Although six or seven negroes were
sleeping quite close to them, they got completely off, and had an
hour’s start before even the loss was discovered. Bellal pursued
them, with about a dozen people, quite to the lake, tracing their
footsteps in the sand, which was not difficult after the rain; but finding
here that they had embarked, the pursuit was given up.
At length, however, on Tuesday the 14th, we had assembled our
kafila, and we moved on towards the desert: on the 22d of
September, in the afternoon, we halted half way to the well of Beere
Kashifery.
Sept. 23.—We made the well soon after mid-day; and fortunately
for us we brought some water with us, for the power of our friend
Mina Tahr here began to appear. This well was guarded, and we
were told, that until the sheikh Mina appeared, not a drop was to be
drawn. It required some exertion of patience and forbearance, in a
sultry oppressive day, with the thermometer at 110° in the tent, to be
obliged to drink muddy water from goat skins, when a well of the
best water between Kouka and Bilma was under our feet: but we
were inured to hardships and contradictions, and submitted, I hope,
like good Christians. Towards evening the Tahr appeared on the hills
to the north-west, attended by his troop: he seemed vastly glad to
see us; said “the well was ours—that our water-skins should be
filled, and camels watered, before any body, and for nothing; and
then,” said he, “sultan George the Great must be obliged to Mina
Tahr, the wandering chief of Gunda, and that will give more pleasure
to Tahr’s heart than payment: and who knows,” said he, “but when
sultan George hears this, he may send me a sword?”
Sept. 24.—In consequence of the number of camels to be
watered, and the large flocks of the Tibboo, it was not until the
evening that our animals could drink; and even then we were almost
obliged to take possession of the well by force. Our old Maraboot
was struck by a spear, as well as our servants; and it was not until
after I had mounted a horse and repaired to the well myself,
accompanied by the Tahr, that we could complete this most
important business of the day.
In the evening Tahr came for his present. I gave him a tobe from
Soudan, a red cap, and a turkodie: the tobes and cap he looked at,
and said, “Ah! this is very well for me; I am one, but my wives are
three:—what shall I do with one turkodie?” Tahr now began a
speech: he was greatly distressed that he had nothing to send sultan
George. “By the head of Mustapha!” said he, “I consider him as
much my master as the bashaw;—ay and more—for you say he sent
you to see me, which is more than the bashaw ever did. I can send
him a tiger-skin, and I will write him a letter—for Tahr’s enemies are
never quiet, and he has no time to kill ostriches now. The well Beere
Kashifery, whose waters are here like gold, and better than gold, and
all that Kashella Tahr and the Gunda Tibboos have to give, shall
always be, as long as he or his children govern, at the service of
sultan George Inglesi.” He now asked for water, and began washing
the ink from a paper which had previously been ornamented with a
charm, drinking the dirty water, and rubbing it over his head and
neck: when this was finished, he laid the paper in the sun; and I was
a little amused when I found, that it was on this dirty scrap that he
intended writing to king George.
Sept. 25.—In the afternoon of this day we left Beere Kashifery,
taking a very hearty leave of Tahr. In order to save my camels, who
had seventeen skins of water to carry in addition to their loads, as
we were to be four days on our road to Aghadem, I hired a maherhy
to carry two heavy boxes to Bilma, for seven dollars. The moon,
which was in its first quarter, assisted us after sunset, but we were
obliged to move on for at least two hours after she had sunk quietly
to rest. We halted for a few hours, at a spot called Geogo Balwy
(honey spot), and a little after midnight proceeded on our dreary way.
There was great difference of opinion as to our route, which,
probably, by following our own back bearings, we might have found
better ourselves: but we were not yet quite so bold in the desert as
on the main; and I insisted on the Mina Hamedee, the guide whom
the sheikh had given us, being allowed his own way, and my
confidence was not misplaced. By daylight on Sunday we discovered
the foot-marks of the camels and slaves of Hadge-Boo-Said, a
Fezzan merchant, who had absolutely been in the right track, but
had turned back, and gone southerly. Numbers now exclaimed that
we were wrong: there appeared, however, no sort of timidity about
our guide; he looked confident, and bore the abuse that was levelled
at him on all sides, with great coolness: and there was a sort of
conscious ability about him that determined me to rely completely on
his judgment alone. No man is ever afraid of doing what he knows
he can do well; and in most cases a sense of power gives
confidence: so it proved with my Tibboo. We continued travelling
another night and day in these most dreary wastes, with nothing but
the wide expanse of sand and sky to gaze upon.
On Monday the 27th of September, a little before noon, we
observed something in the distance, which had the appearance of a
body of men moving towards us; but from the effect of the mirage
assuming different shapes, and sometimes appearing twelve or
fifteen feet above the surface of the desert, the Arabs declared it to
be a Tuarick party on the forage, and all our followers loaded and
prepared for action. On their approach, however, we found to our
great joy that it was a kafila from Fezzan: they had been as much
alarmed at us as we at them, and were all formed, in front of their
camels several hundred yards, in extended order, as the Arabs
always fight: they gave us some Fezzan dates, which were a great
luxury; and some of the traders who were short of water exchanged
a jar of butter, worth at least two dollars, for every full skin they could
so purchase. They told us the road was perfectly safe, although their
fears of falling in with the Tuaricks had detained them seventy-two
days on the journey from Mourzuk.
It is scarcely possible to convey an idea of the sensations of all
parties on a meeting of this nature on the desert. The Arabs were
equally alive to these feelings as ourselves; and, in their usual wild
expressive manner, sang, for days after such a rencontre, ballads
descriptive of the event[59].
We halted at noon, at a place called Gassooma-foma. In the
afternoon we moved again; and the guides told me that the road was
so difficult that, until the moon fell, we should make the best of our
way, and then rest. On these occasions we pitched no tents, but laid
the boxes together; and, either with a little boiled kouskosou, or still
oftener without, soon forgot our fatigues in sleep. When we saw the
black ridge that extends along the wadey of Aghadem, the negroes,
female slaves, and followers, set up screams of joy, and began
dancing and singing with all their might. It was almost noon when we
got to the well, and several slaves, belonging to an old Shouaa who
was going to Hage, were speechless from want of water; yet they
ran several miles to reach the well, like things distracted, with their
mouths open, and eyes starting from their heads.
On the 2d of October we left Aghadem, and by the help of a
blessed moon we were enabled to travel until near midnight, without
losing our way. A very sharp storm of wind from the east obliged us
to halt; and we had scarcely time to shelter ourselves with the skins
and boxes, before it came on with most disagreeable force: this
detained us until daylight, when we rose from between the hillocks of
sand that had formed on each side of us in the night.
We had the satisfaction throughout our journey to find, that, young
as we were at desert travelling, yet we got on as well, if not better,
than our companions; and though children of the soil, they always
looked to us, instead of us to them, both for safety and protection, as
well as for the direction of the route. It was noon on Thursday the
7th, when we made Zow, an oasis situated under some high black
sandstone hills, where there is good water and ahgul in abundance
for the camels, who had scarcely broken their fast since leaving
Dibla. Zow is most appropriately named “difficult,” from the road
which leads to it—a frightful sandy waste of moveable sand hills,
exceeding fifty miles. Some little girls, and children of the kafila,
panting with thirst, augmented by fever and illness, were scarcely
able to creep along the deep sand: the whip shaken over the head
urged them on—for in justice it must be said, the Arabs use it but
rarely in any other way—and not to urge them on would be still more
cruel, for the resolution and courage of these poor things would
never carry them through; they would lie down, and if sleep once
overcame them, so as to be left behind, death would be inevitable.
Oct. 11.—We arrived at Bilma. Without the supply of dates, which
are procured here, kafilas would often suffer extreme hunger, so
scarce are provisions, and so difficult is the transport: all followers
from hence agree to have one meal per day of dates, and one of
flour and fat; while, previous to arriving at Bilma, they are obliged to
have two of flour and fat. Slaves of poor merchants will for twenty
days together be fed by a handful of dates, night and morning, and
they generally thrive well on this nourishing food.
Oct. 15.—We laid in a stock of dates for the next fourteen days,
and man and beast were nearly subsisted upon them: a camel-load
is worth from four to five dollars; they will, however, take camels’
flesh eagerly instead of money, or Soudan goods at one hundred per
cent. profit. Our tents were surrounded by daylight with women and
men; the former to sell us their commodities, and the latter to look
on.
Oct. 17.—We had another day of rest, and were pretty tranquil.
The women came in throngs to our tents, and were willing to sell us
corn and dates, for either dollars or Soudan tobes, at one hundred
and fifty per cent. profit: two lean goats they asked me four dollars
for; and for a sheep, six. A great deal of bustle was made about the
settlement of the dispute with the Mesurata Arabs, and the Tibboo:
“The Book” was to be referred to, but Hadge Mohammed Abedeen,
the brother of the kadi at Mourzuk, would not open the leaves until
the relations of the deceased swore to rest satisfied with his
decision. This preliminary being arranged on Monday morning, the
parties all assembled: the kadi, Hadge Ben Hamet, and Ben Taleb,
the chief merchants of our kafila, were present: they found, by the
Koran, that if any man lifts his hand higher than his shoulder, in a
menacing attitude, though he should not be armed, the adversary is
not to wait the falling of the blow, but may strike even to death. The
law was, of course, in favour of the Arab, as he proved the Tibboo’s
having his hand, armed with a spear, raised above his head, when
he shot him dead. On this being declared, the Arabs ran about,
throwing their guns over their heads, shouting and, what we should
call crowing, to such a degree, that I fully expected the Tibboos
would be aggravated to renew hostilities.
Oct. 25.—From hence we were to proceed by a different route to
that by which we went to Bornou: crossing, therefore, another part of
the range, we moved until night, and halted in what appeared to us a
beautiful oasis, under a ridge of dark sand hills. This spot of dingy
fertility extended several miles to the west, and afforded us water,
grass, and wood, for that and the two following days, which were to
be passed in deserts. A few miserable inhabitants had fixed
themselves here, for the sake of a small crop of dates, yielded by a
few palms: they were all anxious to exchange the produce of their
valley for a blue or a white shirt of the coarsest kind,—a luxury they
were the more in want of from possessing no other clothing. This is
by far the best road; soft sand gravel, instead of rough broken
stones; and the kafilas prefer it on account of the wells. The oasis is
called Seggedem. From hence, eight days’ distance, is a Tibboo
town, and by this road kafilas sometimes pass to Ghraat.
Oct. 26.—We left Seggedem after a blowing night, which either
overset the tents, or buried them several feet in the sand. Towards
evening we rested, and starting again at daylight, made the wells of
Izhya by noon next day.
From El Wahr to Meshroo are three very fatiguing days without
water, or a single vestige of verdure. We were not able to reach the
well, and halted short of the Beeban el Meshroo, the pass leading to
the well, nearly four miles. On Sunday, the 8th of November, we
arrived at the well,—watered our weary camels, and our more weary
men, and again pursued our route until night, when we pitched
westward of the well of Omhah; and after one more dreary day, at
night (Nov. 9th) we slept under the palm trees which surround
Tegerhy, the most miserable inhabited spot in Fezzan, nay, in the
world, I might almost say, and yet we hailed it with inexpressible joy,
after the pitiless deserts we had passed.
The fatigue and difficulty of a journey to Bornou is not to be
compared with a return to Fezzan: the nine days from Izhya to
Tegerhy, without either forage or wood, is distressing beyond
description, to both camels and men, at the end of such a journey as
this. The camels, already worn out by the heavy sand-hills, have the
stony desert to pass; the sharp points bruise their feet, and they
totter, and fall under their heavy loads: the people, too, suffer
severely from the scanty portion of provisions, mostly dates, that can
be brought on by these tired animals,—and altogether it is nine days
of great distress and difficulty. There is something about El Wahr
surpassing dreariness itself: the rocks are dark sandstone, of the
most gloomy and barren appearance; the wind whistles through the
narrow fissures, which disdain to afford nourishment even to a single
blade of wild grass; and as the traveller creeps under the lowering
crags, to take shelter for the night, stumbling at each step over the
skeleton of some starved human being, and searching for level spots
on the hard rock, on which to lay his wearied body, he may fancy
himself wandering in the wilds of desolation and despair.
On the day of our making El Wahr, and the two following days,
camels in numbers dropped and died, or were quickly killed, and the
meat brought on by the hungry slaves. Kafilas are obliged to rely on
the chance of Tibboos and Arabs from Mourzuk hearing of their
having passed the desert, and bringing them supplies; should these
fail, many poor creatures must fall a sacrifice for the salvation of the
rest. These bringers of supplies usually sell their dates and corn to
eager buyers, at about four times the price they could obtain for
them in Fezzan; besides which, the merchants gladly hire their
unburthened camels to quicken their passage to a better country.
A Tibboo trader, who was returning to his own country from
Fezzan, gave me a gratifying proof of the confidence he was willing
to place in the word of an Englishman. It was nearly night, and I was
in front of the camels: he had dates to sell, and mine were
expended, but I told him that my money was in my trunk, and that my
camels were too tired for me to unload them: “God bless you!” said
he, “why, I wish you would buy all I have, camels and all: I know who
the English are! Are they not almost Mislem, and people of one
word? Measure the dates, and go on:—pay the kaid at Mourzuk.”
We here voraciously bought up a few bad onions, to give a little
flavour to our insipid meal of flour and water; and soon after, the kaid
brought me a sheep, the only one in the town, which we cut up and
divided, so that we had a sumptuous meal about nine o’clock in the
evening.
On Sunday the 14th of November, by easy journeys we reached
Gatrone, which, before so miserable in our eyes, now really seemed
a little Paradise; and the food which the old hadge who governs
there sent us, of the same kind we before thought so unpalatable on
our outward voyage, now seemed delicious. I literally got up at
daylight to feast on a mess of hot broth and fresh bread, most highly
peppered, and made as good a meal as ever I did in my life.
At Gatrone, as well as at Tegerhy, our tents were pitched in a
palm grove, the trees shading us during the day from the sun-
beams, and at night from the easterly winds: the gentle moaning of
the breeze through its slowly-waving branches was to us a most
pleasing novelty; and the noble, nutritious, and productive palm,
seemed in our eyes fully to merit the beautiful lines of Abulfeda:
“The stately date, whose pliant head, crowned with pendent
clusters, languidly reclines like that of a beautiful woman overcome
with sleep.”—Abulfeda Descr. Egypt, a Michaelis, p. 6.
To do them justice, the Fezzan people seemed as glad we were
come back, as we were ourselves. “To go and come back from the
black country! Oh, wonderful!—you English have large hearts!—God
bless you!—the poor doctor to die too, so far from home!—Health to
your head! it was written he was to die, and you to come back.—God
is great!—and the young Rais Ali too! (Mr. Toole)—Ah! that was
written also:—but he was a nice man—so sweet spoken. Now you
are going home: well, good fortune attend you! How all your friends
will come out to meet you with fine clothes—and how much
gunpowder they will fire away!”
At the mosque of Sidi Bouchier the usual prayer was offered for
our safe arrival in our own country; and on the 21st of November,
Sunday, we made our entry into Mourzuk, and took possession of
our old habitation.
Nov. 21.—All welcomed our return: we had bowls of bazeen and
kouskosou night and morning, and visitors from daylight until long
after sunset, notwithstanding we had no tea, coffee, or sugar, to
regale them with, as on our former residence amongst them. The
new sultan, Sidi Hassein, who succeeded Mustapha, had only
arrived the day before us; and as he had entered in mourning on
account of the death of the bashaw’s wife, the Lilla Gibellia[60], no
rejoicings were allowed on the occasion: he however sent us two fat

Full Ebook of A Course in Quantum Many Body Theory 1St Edition Michele Fabrizio Online PDF All Chapter

Uploaded by

Document Information

Original Title

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

Full Ebook of A Course in Quantum Many Body Theory 1St Edition Michele Fabrizio Online PDF All Chapter

Uploaded by

Copyright:

Available Formats

A Course in Quantum Many Body

Theory 1st Edition Michele Fabrizio

Quantum Scaling in Many Body Systems An Approach to

Quantum Mathematics II 1st Edition Michele Correggi

Back-Of-The-Envelope Quantum Mechanics: with Extensions

A Course in Stochastic Game Theory Eilon Solan

A First Course in Group Theory 1st Edition Bijan Davvaz

Quantum Mechanics a graduate course Solutions manual

Quantum Field Theory A Diagrammatic Approach 1st

Factorization Algebras in Quantum Field Theory 1st

ISSN 1868-4513 ISSN 1868-4521 (electronic)

Cover image: Olga_Kostrova

Trieste, Italy Michele Fabrizio

Acknowledgements I am grateful to Prof. Claudio Castellani, whose master’s course on Quantum

2.5 Power Dissipation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

4.5 Two-Particle Green’s Functions and Correlation Functions . . . . . . 161

5.8 Quasiparticle Hamiltonian and Landau-Boltzmann

1.1 Fock States and Space

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2022 1

1.2 Fermionic Operators

ca† ca | 0 = 0, ca† ca | n a = 1 =| n a = 1. (1.6)

ca ca† + ca† ca | n a  =| n a , n a = 0, 1 , (1.8)

which leads to the operator identity

ca ca† + ca† ca = ca , ca† = 1, (1.9)

where the symbol {. . . } means the anti-commutator. Moreover, since we cannot

ca , ca = ca† , ca† = 0. (1.10)

ca† | 0 =| n a = 1, ca† | n a = 1 = 0, ca | 0 = 0, ca | n a = 1 =| 0,

cb† | n a = 1 = cb† ca† | 0 ≡| n b = 1, n a = 1. (1.12)

| n b = 1, n a = 1 = − | n a = 1, n b = 1 ≡ −ca† cb† | 0. (1.13)

Comparing (1.13) with (1.12), we conclude that

The Hermitian conjugate thus implies that also

ca | n a = 1, n b = 1 = ca ca† cb† | 0 ≡| n b = 1 = cb† | 0. (1.16)

−ca cb† ca† | 0 = cb† ca ca† | 0,

namely that, for a = b,

where the occupation numbers n i = 0, 1.

1.2.1 Fermi Fields

They have the following anti-commutation relations:

with unitary Û , then

showing that the proper transformation of the fermionic operators is

where I N is the identity in the N -fermion Hilbert space. In addition, it is straight-

1.2.2 Second Quantisation of Multiparticle Operators

Let us consider a generic single-particle operator in the first quantisation

V (x) ≡ Vσ σ (r, p) , (1.29)

where p = −i∇ is the conjugate momentum of r.

so that, dropping the identity operator I N −1 ,

while in second quantisation is

We can simply repeat the previous steps and find that

leads to the second quantised expression

Analogously, a generic m-particle operator

translates in second quantisation into

1.3 Bosonic Operators

da† da |n a  = n a | n a , da da† | n a  = 1 + n a | n a  ; (1.39)

Hence, the following commutation relation holds:

where [. . . ] denotes the commutator. Equation (1.39) is satisfied, e.g., by

which we assume hereafter. The permanent, contrary to the Slater determinant, is

1.3.1 Bose Fields and Multiparticle Operators

Exactly like in the fermionic case, if we define

as well as that the permanent with occupation numbers {n a } ≡ (n 1 , n 2 . . . ) is just

1.4 Canonical Transformations

both for fermions and bosons. Since H is Hermitian, then

Since tˆ is Hermitian, there exists a unitary transformation Û , i.e., satisfying Û † Û =

where Ê is diagonal with real elements a. Therefore, if we define a new spinor

where φ̂ = φ̂ † is Hermitian and thus U † U = 1 unitary. We recall that, given two

ca† ca | 0 = 0, ca† ca | n a = 1 =| n a = 1. (1.6)

ca ca† + ca† ca | n a =| n a , n a = 0, 1 , (1.8)

ca† | 0 =| n a = 1, ca† | n a = 1 = 0, ca | 0 = 0, ca | n a = 1 =| 0,

cb† | n a = 1 = cb† ca† | 0 ≡| n b = 1, n a = 1. (1.12)

| n b = 1, n a = 1 = − | n a = 1, n b = 1 ≡ −ca† cb† | 0. (1.13)

ca | n a = 1, n b = 1 = ca ca† cb† | 0 ≡| n b = 1 = cb† | 0. (1.16)

−ca cb† ca† | 0 = cb† ca ca† | 0,

da† da |n a = n a | n a , da da† | n a = 1 + n a | n a ; (1.39)

that, given a Fock state | {n a }, then

We just note that, while | {n a } is eigenstate of the number of particles with

ρσ (r) = σ† (r) σ (r) .

where | n is an intermediate excited state with energy E n . Because of the projector,