0% found this document useful (0 votes)

294 views178 pages

Julia-Diaz B. Problems and Solutions in Many-Body Quantum Theory..Fermi Sea 2025

This guidebook offers a comprehensive collection of exercises and detailed solutions for quantum many-body theory, aimed at advanced undergraduate and graduate students. It covers a wide range of topics from nuclear physics to ultracold atoms and is designed to enhance understanding through a problems-based approach. The book serves as a practical resource for mastering the complexities of quantum many-body physics, filling a gap in existing literature that often lacks solutions to proposed problems.

Uploaded by

aarnaouty

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

294 views178 pages

Julia-Diaz B. Problems and Solutions in Many-Body Quantum Theory..Fermi Sea 2025

Uploaded by

aarnaouty

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 178

Problems and Solutions

in Many-Body Quantum
Theory
This practical guidebook provides a comprehensive set of exercises which illustrate
the most relevant concepts in a first course on quantum many-body theory as part of
advanced undergraduate or graduate courses. The problems come with detailed solu-
tions which can easily be followed either by professors or students. Quantum many-body
theory is relevant in several fields, from condensed matter to astrophysics. The problems
proposed in this book cover this variety of topics and are illustrated whenever possible
with state-of-the-art examples.
Key Features:
• Provides a problems-based approach to quantum many-body theory, unlike
existing textbooks.
• In-depth solutions to problems are presented, with the aim to maximize under-
standing and improve the teaching experience of the subject.
• The multidisciplinary nature of quantum many-body theory is explored in
problems that deal with nuclear physics to ultracold atoms and astrophysics.
Problems and Solutions
in Many-Body Quantum
Theory
Sailing the Fermi Sea

Bruno Juliá Díaz, Arnau Rios Huguet,

Héctor Briongos Merino, Javier Rozalén
Sarmiento and Artur Polls Martí
Designed cover image: Shutterstock

First edition published 2026

by CRC Press
2385 NW Executive Center Drive, Suite 320, Boca Raton FL 33431

and by CRC Press

4 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN

CRC Press is an imprint of Taylor & Francis Group, LLC

Reasonable efforts have been made to publish reliable data and information, but the author and pub-
lisher cannot assume responsibility for the validity of all materials or the consequences of their use.
The authors and publishers have attempted to trace the copyright holders of all material reproduced
in this publication and apologize to copyright holders if permission to publish in this form has not
been obtained. If any copyright material has not been acknowledged please write and let us know so
we may rectify in any future reprint.

Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced,
transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying, microfilming, and recording, or in any information
storage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, access www.copyright.
com or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA
01923, 978-750-8400. For works that are not available on CCC please contact mpkbookspermis-
[email protected]

Trademark notice: Product or corporate names may be trademarks or registered trademarks and are
used only for identification and explanation without intent to infringe.

ISBN: 978-1-032-04296-1 (hbk)

ISBN: 978-1-032-01355-8 (pbk)
ISBN: 978-1-003-19133-9 (ebk)

DOI: 10.1201/9781003191339

Typeset in Nimbus Roman

by KnowledgeWorks Global Ltd.

Publisher’s note: This book has been prepared from camera-ready copy provided by the authors.
i i

“output” — 2025/8/11 — 7:50 — page v — #7

i i

Dedication

To Artur Polls Martı́, in memoriam

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page vii — #9

i i

Contents

SECTION I Basics
Chapter 1 Symmetries ...................................................................................... 3
Problem 1.1: Symmetrization and antisymmetrization
operators for N = 2 .................................................... 3
Problem 1.2: Symmetrization and antisymmetrization
operators..................................................................... 5
Problem 1.3: Symmetrization in the space representation......7

Chapter 2 Second quantization....................................................................... 13

Problem 2.1: Fock space basis states are eigenvectors of
the particle number operator N̂ ................................ 13
Problem 2.2: The particle number operator N̂ commutes
with the Hamiltonian................................................ 14
Problem 2.3: Anticommutation relations of field operators .17
Problem 2.4: Particle density operator.................................. 18
Problem 2.5: The states generated by field operators are
eigenvectors of the particle number operator...........19
Problem 2.6: Heisenberg representation of creation
and destruction operators for a single-particle
Hamiltonian.............................................................. 20
Problem 2.7: A canonical linear transformation...................22
Problem 2.8: Spin operators in second quantization ............ 22

Chapter 3 Wick’s theorem.............................................................................. 27

Problem 3.1: Contraction of two creation and destruc-
tion operators ........................................................... 28
Problem 3.2: Matrix elements using the Wick theorem:
One-body operator ................................................... 29
Problem 3.3: Matrix elements using the Wick theorem:
Two-body operator...................................................31

Chapter 4 Green’s functions........................................................................... 37

Problem 4.1: The Green’s function of a free particle in 1D .38
Problem 4.2: Single-particle spectral function for a non-
interacting system. ...................................................40

vii

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page viii — #10

i i

viii Contents

Problem 4.3: Fourier transform of the single-particle

propagator. ............................................................... 42
Problem 4.4: Green’s function. Field operator definition. .... 44
Problem 4.5: Single-particle spectral functions for a
non-interacting system. ............................................45
Problem 4.6: Kinetic energy and Koltun sum-rule for the
free Fermi sea...........................................................47

SECTION II Applications

Chapter 5 The Fermi sea in one dimension....................................................51

Problem 5.1: Thermodynamic properties of the free
Fermi sea in 1D........................................................ 52
Problem 5.2: Two body distribution function of a free
Fermi sea in 1D........................................................ 54
Problem 5.3: Properties of the Slater function in 1D...........57
Problem 5.4: Trapped fermions ............................................58

Chapter 6 The Fermi sea in three dimensions at T = 0 .................................63

Problem 6.1: Energy and first derivatives .............................64
Problem 6.2: Symmetry energy of the free Fermi sea .......... 66
Problem 6.3: A simple lower bound to the kinetic energy ... 69
Problem 6.4: Thermodynamic properties of the rela-
tivistic three-dimensional free Fermi sea .................70

Chapter 7 Free Fermi sea at finite temperature .............................................. 75

Problem 7.1: Specific heat at the low temperature limit
of the free Fermi sea ................................................76
Problem 7.2: First steps: density and energy per particle .....78
Problem 7.3: High T limit of the free Fermi sea .................. 80
Problem 7.4: Low T expansion.............................................83
Problem 7.5: Partition function of the free Fermi system .... 88
Problem 7.6: Entropy of the non-interacting Fermi sea .......90

Chapter 8 Distribution functions ...................................................................93

Problem 8.1: Sequential conditions for the two-body
distribution function of the free Fermi sea...............94
Problem 8.2: Spin components of the two-body distri-
bution function .........................................................95
Problem 8.3: Spin of the free Fermi sea ............................... 98

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page ix — #11

i i

Contents ix

Chapter 9 Dynamic structure functions........................................................ 101

Problem 9.1: Dynamic structure function for the Fermi
sea at high momentum transfer ..............................102
Problem 9.2: Sum rules of the dynamic structure func-
tion for q ≥ 2 kF ..................................................... 103
Problem 9.3: Maximum and width of the dynamic struc-
ture function of the free Fermi sea for q ≥ 2 kF .... 106
Problem 9.4: Compton profile and Y-scaling for the free
Fermi sea................................................................ 107

Chapter 10 Hartree-Fock................................................................................ 111

Problem 10.1: Hartree-Fock for nuclear matter with a
simple Skyrme interaction ..................................... 112
Problem 10.2: Hartree-Fock for a system of N fermions
enclosed in a one dimensional box ........................ 116
Problem 10.3: Single-particle potential for a uniform
system of fermions in the Hartree-Fock approx-
imation ................................................................... 119

Chapter 11 Density matrices .......................................................................... 123

Problem 11.1: Derivatives of the one-body density matrix 123
Problem 11.2: Sequential condition.................................... 126
Problem 11.3: Properties of the two-body density matrix ..128
Problem 11.4: One body density matrix in a 2D non in-
teracting Bose-Hubbard gas...................................130

Chapter 12 Superfluidity ................................................................................135

Problem 12.1: The Bardeen-Cooper-Schrieffer (BCS)
ansatz......................................................................136
Problem 12.2: Particle number projection ..........................141

Chapter 13 Mixed statistics ............................................................................ 145

Problem 13.1: Mixed statistics: a system of fermions
and bosons.............................................................. 145

Appendix A Second-quantization formalism................................................... 151

Appendix B Wick’s theorem............................................................................ 155

References .............................................................................................................161

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page x — #12

i i

x Contents

Index...................................................................................................................... 163

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page xi — #13

i i

Preface
This book is the result of the teaching experience associated to a “Many-Body Quan-
tum Theory” module which has existed as part of the Physics curriculum at the Uni-
versity of Barcelona, in different shapes and forms, in the last few decades. The
module was mostly carried by Prof. Artur Polls for over 20 years, together with
Profs. Manuel Barranco, Muntsa Guilleumas, Bruno Juliá-Dı́az, and Àngels Ramos.
Prof. Polls was a pioneer of the field with a broad national and international reputa-
tion. He passed away in 2020, but this book remains as part of his legacy to a disci-
pline, quantum many-body physics, that he enjoyed. He also taught it with exquisite
clarity and, perhaps more importantly, set the foundations for several generations of
physicists working on it now.
One of us (Prof. Juliá-Dı́az) took over the module in 2020 and curated the last ver-
sion of the problem sets that are taught in conjunction with standard theory lectures.
Dr. Rios joined the project in 2021, but it wasn’t until the arrival of PhD students
Héctor Briongos Merino and Javier Rozalén Sarmiento that the final book write-up
really took off in 2024.
Just like any other University module, the choice of topics in the “Quantum
Many-Body Physics” course is biased by the lecturers that teach it. In this case, the
broad scope of the quantum many-body physics field is showcased in the problems
that verse on nuclear physics; ultracold atoms (including fermions and bosons); and
systems with low dimensionality.
A number of excellent books discussing quantum many-body physics have been
published over the years with a variety of perspectives [9, 3, 11, 7, 6, 5, 1, 12, 4].
These manuals, typically aimed at master’s and graduate levels, often incorporate
problems as work-at-home examples. Unfortunately, they rarely include the actual
solutions to the problems, potentially causing a gap in understanding. In some cases,
this can hamper the learning activity of students that cannot solve them for whatever
reason.
The aim of our book is precisely to fill this gap in the current literature. We
have prepared a manual of problems and solutions for master’s and graduate students
that want to deepen their practical understanding of quantum many-body physics.
Because many theory manuals are already available in the literature, our approach
has focused mostly on the solutions themselves. We provide a very cursory reminder
of theory essentials in each chapter, which focuses on specific content, but the heart
of our book is the solution to the proposed problems. We have also prepared two
appendices discussing briefly second quantization (Appendix A) and Wick’s theorem
(Appendix B), to furnish some further context to problems that required so.
In our minds, the target audience of this book are master’s or graduate students
trying to start their way into the beautiful, yet sometimes challenging, discipline that
is quantum many-body physics. All the problems in this collection have analytical
solutions. As many-body practitioners know, analytical solutions are the exception in

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page xii — #14

i i

xii Preface

this field, not the rule. Having said that, we also feel that in laying the basic ground-
work for the discipline, a solid understanding of some concepts cannot be achieved
without some of the analytical work we dissect here.
Some of the problems we propose are classics in the field: computing the pair
distribution function of the free Fermi gas or an explicit calculation of the Lindhard
function. Others, like those on Wick’s theorem, can be daunting to work on at first,
but provide a solid understanding that is of course also relevant for other disciplines,
such as quantum field theory. Finally, we give some examples that are not mathe-
matically cumbersome, but we hope that they yield some intuition of the physics at
play.
In most cases, we have provided a step-by-step solution to the problems, includ-
ing mathematical details that may have been skipped in the interest of brevity. We
hope that, removing the intermediate mathematical complexity, may render a more
physically-oriented learning experience. In doing so, we hope to convey Prof. Polls’
teaching style, which was very much one of being extremely careful in exposition;
elucidating all possible pitfalls and errors; and ultimately reaching the correct solu-
tion, providing along the way a clear physical intuition.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page xiii — #15

i i

Contributors

Bruno Julia´ Dı́az

Departament de Fı́sica Quàntica i
Astrofı́sica and Institut de Ciències Héctor Briongos Merino
del Cosmos (ICCUB), Universitat de Departament ´de Fı́sica Quàntica i
Barcelona Astrofı́sica and Institut de Ciències
Barcelona, Spain del Cosmos (ICCUB), Universitat de
Barcelona
Artur Polls Martı́ Barcelona, Spain
Departament de Fı́sica Quàntica i
Astrofı́sica and Institut de Ciències Javier Rozalén Sarmiento
del Cosmos (ICCUB), Universitat de Departament de Fı́sica Quàntica i
Barcelona Astrofı́sica and Institut de Ciències
Barcelona, Spain del Cosmos (ICCUB), Universitat de
Barcelona
Arnau Rios Huguet Barcelona, Spain
Departament de Fı́sica Quàntica i
Astrofı́sica and Institut de Ciències
del Cosmos (ICCUB), Universitat de
Barcelona and University of Surrey,
Guildford (UK)

xiii

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 1 — #17

i i

Section I

Basics

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 3 — #19

i i

Symmetries
1
Symmetry considerations lie at the very heart of quantum many-body theory. In par-
ticular, symmetries dictate and restrict the properties of many-body wavefunctions.
A many-body system with N particles can be described at zero temperature by a
many-body wavefunction Ψ(~x1 , · · · ,~xA ), where ~xi represents the coordinate of par-
ticle i. Note that this coordinate can be discrete, e.g. a lattice site in a spin system;
continuous, if, for instance, it describes the position of a particle; or a combination
of both, when we work with particles of definite position and spin, ~xi = (~xi , σi ).
Permutation symmetry dictates how the many-body wavefunction of a system
transforms when two particles are exchanged. Consider the permutation operator Pi j ,
which exchanges the coordinates of particles i and j in the many-body wavefunction,

Pi j Ψ(~x1 ,~x2 , . . . ,~xi , . . . ,~x j , . . . ,~xN ) = Ψ(~x1 ,~x2 , . . . ,~x j , . . . ,~xi , . . . ,~xN ) . (1.0.1)

Since applying the operator Pi j twice leads to the original wavefunction, we know
that Pi2j = 1. In consequence, the eigenvalues of this operator when acting on a wave-
function can either be +1 or −1.
The symmetrization postulate indicates that the only admissible many-body
wavefunctions are either fully symmetric or fully antisymmetric with respect to all
permutation operators. This is to be expected because we assume that the system
is formed of A indistinguishable particles. Permuting any one particle with another
cannot change the state of the system more than by a single phase. This phase, or
the sign associated to a permutation of coordinates, is one of the more fundamental
properties of a many-body wavefunction. It takes a value of +1 for bosonic systems,
or a value of −1 for fermionic systems.
In this chapter, we provide some insight on the symmetrization and antisym-
metrization operators, two key projectors for the two fundamental species of identi-
cal particles, and their effect on wavefunctions of quantum many-body systems.

PROBLEM 1.1
Symmetrization and antisymmetrization operators for N = 2

The symmetrization and antisymmetrization operators for 2 identical particles are

defined as:
1 1
Ŝ = 1 + P̂12 , Â = 1 − P̂12 . (1.1.1)
2 2
Here, P̂12 is the exchange operators of particles 1 and 2. Show explicitly that:

(a) Ŝ2 = Ŝ and Â2 = Â;

(b) Ŝ + Â = 1;

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 4 — #20

i i

4 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

(c) ŜÂ = ÂŜ = 0;

(d) P̂12 Ŝ = Ŝ, and P̂12 Â = −Â.

(a) We start with the symmetrization operator. We apply it twice and find, using
P12 P12 = 1 where 1 is the identity operator,
1 1 1 1
ŜŜ = 1 + P̂12 1 + P̂12 = 1 + P̂12 + P̂12 + 1 = 1 + P̂12 = Ŝ .
2 2 4 2
Proceeding similarly with the antisymmetrization operator:
11 1 1
ÂÂ = 1 − P̂12 1 − P̂12 = 1 − P̂12 − P̂12 + 1 = 1 − P̂12 = Â .
22 4 2
We have shown that both operators are idempotent, i.e., they fulfill the condition
Ô2 = Ô. This is a sufficient condition to say that both are projectors.

(b) The sum of the two operators is easily identified with the identity,
1 1 1
Ŝ + Â = 1 + P̂12 + 1 − P̂12 = 1 + P̂12 + 1 − P̂12 = 1 .
2 2 2
This property shows that operators Ŝ and Â form a set of operators that covers all
Hilbert space of 2 particles, and thus each state can be decomposed in a symmet-
ric and an antisymmetric part by means of these projectors.
(c) In contrast, the product of the two operators is null:
1 1 1
ŜÂ = 1 + P̂12 1 − P̂12 = 1 + P̂12 − P̂12 − 1 = 0 .
2 2 4

(d) We now demonstrate that a permutation of two particles leaves the symmetriza-
tion operator unchanged:
1 1
P̂12 Ŝ = P̂12 1 + P̂12 = P̂12 + 1 = Ŝ .
2 2

Exchanging two particles before the antisymmetrization operator instead brings

in a minus sign:
1 1
P̂12 Â = P̂12 1 − P̂12 = P̂12 − 1 = −Â .
2 2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 5 — #21

i i

Symmetries 5

PROBLEM 1.2
Symmetrization and antisymmetrization operators.

The definitions for the symmetrization

1
Ŝ = ∑ P̂ , (1.2.1)
N! p∈SN

and antisymmetrization operator

1
Â =
N! ∑ (−1)ε(p) P̂ , (1.2.2)
p∈SN

for N identical particles include the permutation operator, P̂, running over all per-
mutations of the symmetric group of N objects. ε(p) is the signature of a given
permutation, p. Using these definitions, show that:

(a) Ŝ2 = Ŝ and Â2 = Â;

(b) ŜÂ = ÂŜ = 0;

(c) P̂jk Ŝ = Ŝ; and

(d) P̂jk Â = −Â.

(a) We start by applying the symmetrization operator twice to get Ŝ2 . We use two
different dummy operators, P̂ and Q̂, for the corresponding sums and get:
! !
1 1 1 1
Ŝ2 = ŜŜ = ∑ P̂ N! ∑ Q̂ = N! N! ∑ ∑ P̂Q̂ .
N! P∈SN Q∈SN Q∈SN P∈SN

When P̂ runs over all SN , due to the closure property of groups, P̂Q̂ (where Q̂ is
a given element of SN ) runs over all SN too. In consequence, we can replace P̂Q̂
by a single permutation operator, R̂. The sum over Q̂ becomes redundant, so that
1 1 1 1
Ŝ2 = ∑ ∑ R̂ = N! N! N! ∑ R̂ = Ŝ .
N! N! Q∈SN R∈SN R∈SN

We follow a similar strategy to compute Â2 , following the same steps and intro-

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 6 — #22

i i

6 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

ducing a new R̂ operator:

! !
1 1
Â2 = ÂÂ = ∑ (−1)ε(Q̂) Q̂ ∑ (−1)ε(P) P̂
N! Q∈SN
N! P∈SN
1 1 1 1
= ∑ ∑ (−1)ε(Q)+ε(P) Q̂P̂ = N! N! ∑ ∑ (−1)ε(R) R̂
N! N! Q∈SN P∈SN Q∈SN R∈SN
1 1
= N! ∑ (−1)ε(R) R̂ = Â ,
N! N! R∈SN

where we have used the following property:

(−1)ε(Q)+ε(P) = (−1)ε(R) ,
which holds given that the parity of a given permutation is (−1)n with n the
smallest number of 2-particle transpositions needed to obtain the permutation.
We have just shown that both the symmetrization and the antisymmetrization
operators, Ŝ and Â, are idempotent, and this condition is sufficient to demonstrate
that they are projectors.
(b) To prove the equality ŜÂ = 0, we write explicitly the sums associated to each
operator in terms of the permutation operators,
1 1 1 1
ŜÂ = ∑ Q̂ ∑ (−1)ε(P) P̂ = N! N! ∑ (−1)ε(Q) ∑ (−1)ε(Q)+ε(P) Q̂P̂ .
N! N! Q∈SN P∈SN Q∈SN P∈SN

In this last line, we have used the trick (−1)ε(Q) (−1)ε(Q) = 1. We now proceed
by noting, again, that the product of two permutations, Q̂P̂, is a permutation R̂,
and find:
1 1 1
ŜÂ = ∑ (−1)ε(Q) N! ∑ (−1)ε(R) R̂ = N! ∑ (−1)ε(Q) Â
N! Q∈SN R∈SN Q∈SN
1
= Â ∑ (−1)ε(Q) = 0 .
N! Q∈SN

(c) We now consider the exchange of two particles, j and k, in front of the sym-
metrization operator. Using explicitly Eq. (1.2.1), we find:
1 1 1
P̂jk Ŝ = P̂jk ∑ P̂ = N! ∑ P̂jk P̂ = N! ∑ R̂ = Ŝ .
N! P∈S N P∈SN R∈SN

(d) A very similar derivation, but starting from Eq. (1.2.2), yields:
1 1
P̂jk Â = P̂jk ∑
N! P∈SN
(−1)ε(P) P̂ = ∑ (−1)ε(Pjk ) (−1)ε(Pjk ) (−1)ε(P) P̂jk P̂
N! P∈SN
1 1
= (−1)ε(Pjk ) ∑ (−1)ε(Pjk )+ε(P) P̂jk P̂ = − ∑ (−1)ε(R) R̂ = −Â ,
N! P∈SN N! R∈SN

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 7 — #23

i i

Symmetries 7

where we have taken into account that

(−1)ε(Pjk ) = −1 ,

and that
(−1)ε(Pjk )+ε(P) = (−1)ε(Pjk P) .

PROBLEM 1.3
Symmetrization in the space representation

Consider the properly symmetrized wavefunction of two spinless bosons, which

are distributed over two single-particle levels, 1 and 2,

1
Ψ(~r1 ,~r2 ) = √ [ψ1 (~r1 )ψ2 (~r2 ) + ψ1 (~r2 )ψ2 (~r1 )] . (1.3.1)
2

The two single-particle states are stationary and properly normalized. Assuming
that state 1 has positive parity and state 2 has negative parity:

(a) Determine the probability of finding a particle at position ~r, P(~r), if the posi-
tion of the other one is arbitrary.

(b) Find the probability of finding one particle in the upper half-space, z ≥ 0.

(d) Answer these questions for two fermions in the same states.

(e) If the two particles were distinguishable, we would know in which state each
particle sits. The wavefunction would be a simple product state, Ψ(~r1 ,~r2 ) =
ψ1 (~r1 )ψ2 (~r2 ). Answer the previous questions for this case.

This problem is inspired by problem 15.5 in Ref. [15].

(a) We start by finding the probability of particle 1 without knowing anything about
the position of particle 2. We marginalize over one of the positions by integrating

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 8 — #24

i i

8 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

over the whole volume:

Z ∞
P(~r) = d~r2 |Ψ(~r1 =~r,~r2 )|2
−∞
1 1
Z ∞ ∞ Z
= |ψ1 (~r)|2 d~r2 |ψ2 (~r2 )|2 + ψ1∗ (~r)ψ2 (~r) d~r2 ψ2∗ (~r2 )ψ1 (~r2 ) +
2 −∞ 2 −∞
| {z } | {z }
=1 =0
1 1
Z ∞ ∞ Z
+ ψ2∗ (~r)ψ1 (~r) d~r2 ψ1∗ (~r2 )ψ2 (~r2 ) + |ψ2 (~r)|2 d~r2 |ψ1 (~r2 )|2
2 −∞ 2 −∞
| {z } | {z }
=0 =1
1 1
= |ψ1 (~r)|2 + |ψ2 (~r)|2 . (1.3.2)
2 2
The integrals of the square wavefunction are assumed to be normalized. The
crossed terms are instead zero due to parity. To see this, we use the parity of each
state, ψ1 (−~r) = +ψ1 (~r) but ψ2 (−~r) = −ψ2 (~r), and look at the crossed terms
under a change of the integration variable~r2 → −~r2 :
Z ∞ Z ∞
d~r2 ψ2∗ (~r2 )ψ1 (~r2 ) = d~r2 ψ2∗ (−~r2 )ψ1 (−~r2 )
−∞ −∞
Z ∞
=− d~r2 ψ2∗ (~r2 )ψ1 (~r2 ) ≡ 0.
−∞

The result of Eq. 1.3.2 has an ”intuitive” explanation, in that, upon integrating
over the second particle, we find that the remaining particle is equally distributed
between either one of the single-particle wavefunctions.
(b) To determine the probability of finding one particle over the upper half-space, we
integrate the previous result for z ≥ 0,
1 1
Z Z Z
Wz≥0 = d~r P(~r) = d~r |ψ1 (~r)|2 + d~r |ψ2 (~r)|2 .
z≥0 2 z≥0 2 z≥0
The two wavefunctions have definite parity, so the integrals over the lower and
upper half-spaces are the same, as shown below,
Z Z ∞ Z ∞ Z ∞
d~r |ψi (~r)|2 = dx dy dz |ψi (x, y, z)|2
z≥0 −∞ −∞ 0
Z ∞ Z ∞ Z ∞
= dx dy dz |ψi (x, y, −z)|2
−∞ −∞ 0
Z ∞ Z ∞ Z 0
= dx dy dz|ψi (x, y, z)|2 = A.
−∞ −∞ −∞
We choose the variable A to denote this half-volume integral. Moreover, since
the wavefunctions are normalized over the whole volume,
Z ∞ Z ∞ Z ∞ Z ∞
d~r|ψi (~r)|2 = dx dy dz|ψi (x, y, z)|2
−∞ −∞ −∞ 0
Z ∞ Z ∞ Z 0
+ dx dy dz|ψi (x, y, z)|2 = 2A = 1,
−∞ −∞ −∞

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 9 — #25

i i

Symmetries 9

so clearly A = 12 . With this, we find that

1 1 1 1 1
× + × = .
Wz≥0 = =
2 2 2 2 2
In other words, there is a 50% chance of finding one particle in the upper half-
space.
(c) To determine the probability of finding two particles in the upper half-space,
z ≥ 0, we need to look at the two-body probability distribution. This is obtained
from the two-body wavefunction, integrated over the upper half-space of each
particle. This involves known integrals over squared wavefunctions in this region
as well as an unknown, possibly complex term involving the crossed integral of
the two states in the upper half plane, which we dub I:
Z Z
Wz1 ≥0,z2 ≥0 = d~r1 d~r2 |ψ(~r1 ,~r2 )|2
z1 ≥0 z2 ≥0
"Z
1
Z
2
= d~r1 |ψ1 (~r)| d~r2 |ψ2 (~r2 )|2
2 z1 ≥0 z2 ≥0
| {z }| {z }
=1/2 =1/2
Z Z
+ d~r1 ψ1∗ (~r1 )ψ2 (~r1 ) d~r2 ψ2∗ (~r2 )ψ1 (~r2 )
z1 ≥0 z2 ≥0
| {z }| {z }
=I =I ∗
Z Z
+ d~r1 ψ2∗ (~r1 )ψ1 (~r1 ) d~r2 ψ1∗ (~r2 )ψ2 (~r2 )
z1 ≥0 z2 ≥0
| {z }| {z }
=I ∗ =I
Z Z
#
+ d~r1 |ψ2 (~r)|2 d~r2 |ψ1 (~r2 )|2
z1 ≥0 z2 ≥0
| {z }| {z }
=1/2 =1/2
1
= + |I|2 .
4
With this, we find that the probability of finding two particles in the upper half-
space is either 25% or larger. Indeed, the probability is increased by a real and
positive factor, |I|2 , which depends on the problem under consideration.
(d) For fermions, the antisymmetrized wavefunction reads
1
Ψ(~r1 ,~r2 ) = √ [ψ1 (~r1 )ψ2 (~r2 ) − ψ1 (~r2 )ψ2 (~r1 )] , (1.3.3)
2
and the results for problems (a) and (b) are the same. In (a) the mixed terms,
which carry a minus sign for fermions, cancel, and so one obtains the same P(~r).
In contrast, for (c), the sign of the mixed terms is relevant, and opposite to the
bosonic case. One finds
1
Wz1 ≥0,z2 ≥0 = − |I|2 .
4

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 10 — #26

i i

10 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Distinguishable Bosons Fermions

1.0 1.0
0.8
0.6
0.5
Position, x1

0.4
0.2
0.0 0.0
−0.2
−0.4
−0.5
−0.6
−0.8
−1.0 −1.0
−1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0 −1.0 −0.5 0.0 0.5 1.0
Position, x2 Position, x2 Position, x2

Figure 1.1 Contour density plots for the two-body wavefunction Ψ(x1 , x2 ) from different
product states using the ground and first excited states of an infinite well with limits at x =
±1. The left panel shows the product of two distinguishable particle states. The central and
right panel correspond to the bosonic and fermionic cases. Note the very different structure of
these wavefunctions along the x1 = x2 line. Dashed (solid) lines represent negative (positive)
contours spaced in steps of 0.1.

In other words, the probability of finding 2 fermions on the upper half space is
less than 25%. In fact, this is a factor 2|I|2 smaller than in the bosonic case.
(e) The results in the distinguishable case are very different. To start with, the one-
particle probability depends on which particle we are looking into. The proba-
bility of finding particle 1 is obtained from the total two-body wavefunction by
marginalizing over particle 2:
Z ∞ Z ∞
P1 (~r) = d~r2 |Ψ(~r =~r1 ,~r2 )|2 = |ψ1 (~r)|2 d~r2 |ψ2 (~r2 )|2 = |ψ1 (~r)|2 ,
−∞ −∞
| {z }
=1

which yields the single-particle probability of particle 1. In contrast, for particle

2, the result is just the probability of particle 2,
Z ∞ Z ∞
P2 (~r) = d~r1 |Ψ(~r =~r1 ,~r2 )|2 = d~r1 |ψ1 (~r)|2 |ψ2 (~r2 )|2 = |ψ2 (~r)|2 .
−∞ −∞
| {z }
=1

In addition, the probability to find either particle in the upper half-space is just
1/2. The probability to find the two of them simultaneously in this same region
is 1/4.
As a practical example of the importance of the symmetrization of wavefunc-
tions, we show in Fig. 1.1 the result of three different symmetrizations of a
two-body wavefunction. We now have an infinite square well Hamiltonian, with
bounds placed at x = ±L. The single-particle wavefunctions for this case are:
q
 1 cos n π x if n odd,
2L
ψn (x) = q L (1.3.4)
 1 sin n π x if n even.
L 2L

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 11 — #27

i i

Symmetries 11

For plotting purposes, we work with L = 1. We use the convention where n = 1

represents the ground state here.
Two distinguishable particles have a simple product wavefunction. We choose in
which state each particle is placed, and we put particle 1 in the ground state, ψ1 ,
and particle 2, in the first excited state, ψ2 :
π
Ψ(x1 , x2 ) = ψ1 (x1 )ψ2 (x2 ) = cos x1 sin (πx2 ) . (1.3.5)
2
A contour plot corresponding to this two-body wavefunction is shown in the left
panel of Fig. 1.1. The wavefunction is only zero if x2 = 0.
For two bosons in the same two states, one finds the two-body wavefunction:
1
Ψ(x1 , x2 ) = √ [ψ1 (x1 )ψ2 (x2 ) + ψ1 (x2 )ψ2 (x1 )]
2
(1.3.6)
1 h π π i
= √ cos x1 sin (πx2 ) + sin (πx1 ) cos x2 .
2 2 2

The wavefunction only cancels when x1 = −x2 . This is shown in the central panel
of Fig. 1.1. We stress that two bosons could also sit on the same state, with a total
wavefunction Ψ(x1 , x2 ) = ψ1 (x1 )ψ1 (x2 ). This product state may have a lower
energy.
In contrast, for two fermions, the wavefunction has a − sign and thus
1
Ψ(x1 , x2 ) = √ [ψ1 (x1 )ψ2 (x2 ) − ψ1 (x2 )ψ2 (x1 )]
2
(1.3.7)
1 h π π i
= √ cos x1 sin (πx2 ) − sin (πx1 ) cos x2 .
2 2 2

Here, the nodal line occurs when x1 = x2 , as displayed in the right panel of
Fig. 1.1.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 13 — #29

i i

Second quantization
2
In non-relativistic quantum mechanics, each particle is characterized by a set of pa-
rameters. These are intrinsic to a given particle, and they are the same for a given
particle species. Think, for instance, of the mass, the charge, or the spin of an elec-
tron. All electrons have the same mass, charge and spin and, from these properties
alone, we cannot distinguish two different electrons. Based on these properties, we
can divide particles into classes depending on their intrinsic parameters. Particles
belonging to the same class are said to be indistinguishable.
The way this works is different in quantum than in classical mechanics. In clas-
sical mechanics, we can always distinguish particles depending on some of their
properties, like their positions or momenta. The positions of two identical quantum
mechanical particles, in contrast, cannot be distinguished. In quantum mechanics,
we face a different problem with a different meaning for “indistinguishability” than
in classical mechanics.
The natural language to describe indistinguishable particles in quantum mechan-
ics is the so-called second quantization formalism. Rather than working in a Hilbert
space with a fixed number of particles, H (N), in second quantization we let go of
the idea of a fixed number of particles, and form a Hilbert space that is large enough
to accommodate an undetermined number of particles, called the Fock space. The
mathematical tools used in this formalism also differ from those found in first quan-
tization. We refer the reader to Appendix A for a more detailed introduction to the
second quantization formalism.
In this chapter, we explore the basics of second quantization through different
problems. These exercises take us from the most basic properties of creation and
destruction operators, to the operators we can build with them, and to more advanced
constructions such as field and spin operators.

PROBLEM 2.1
Fock space basis states are eigenvectors of the particle number operator N̂

Using the anticommutation relations for the creation and destruction operators in
a system of identical fermions, show that:

(a) â†α |0i is an eigenstate of the particle number operator N̂ = ∑α â†α âα .

(b) â†β â†γ |0i is also an eigenstate of N̂.

In both cases, determine the corresponding eigenvalues.

(a) The trick to this problem is to consider first the state â†α |0i and then apply the

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 14 — #30

i i

14 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

number operator to it on the left. We are told these are fermions, so we can then
immediately use anticommutation relations to find:
" #
N̂ âα |0i = ∑ âβ âβ â†α |0i = ∑ â†β âβ â†α |0i
† †

β β | {z }
δαβ −â†α âβ

= ∑ â†β δαβ − â†α âβ |0i = â†α |0i ,

(b) For the state â†β â†α |0i, the situation is very similar. We apply the number operator
to the left of the state and employ the anticommutation relations twice:
" #
N̂(â†β â†α |0i) = ∑ â†γ âγ â†β â†α |0i = ∑ â†γ (δγβ − â†β âγ )â†α |0i
γ γ

= â†β â†α |0i − ∑ â†γ â†β (δγα − â†α âγ ) |0i

= â†β â†α |0i − â†α â†β |0i = 2 â†α â†β |0i .

Therefore, the state â†β â†α |0i is indeed an eigenvector of N̂ with eigenvalue 2.

PROBLEM 2.2
The particle number operator N̂ commutes with the Hamiltonian

Show that the particles number operator, N̂ = ∑α â†α âα , commutes with the Hamil-
tonian, Ĥ, of a system of fermions. To this end, consider the partition Ĥ = Ĥ0 + Ĥ1 ,
with
(a) A one-body operator term,

Ĥ0 = ∑ tαβ â†α âβ ,

αβ

where the matrix elements are defined as tαβ = hα|T̂ |β i.

(b) A two-body operator term,
1
Ĥ1 = ∑ vαβ γδ â†α â†β âδ âγ ,
2 αβ δγ

with matrix elements vαβ γδ = (αβ |V̂ |γδ ).

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 15 — #31

i i

Second quantization 15

(c) How does the result change for bosons?

(a) We look first at the commutator h ofi the one-body operator with the number oper-
†
ator, Ĥ0 , N̂ = ∑αβ tαβ âα âβ , N̂ . We compute each term within the sum sepa-
rately. We employ the following relation,
†
âα âβ , N̂ = â†α âβ N̂ − N̂ â†α âβ = â†α âβ N̂ − â†α N̂ âβ + â†α N̂ âβ − N̂ â†α âβ

= â†α âβ , N̂ + â†α , N̂ âβ ,

(2.2.1)

to break down the commutator of a product of creation and annihilation operators

into commutators of individual creation and annihilation operators with N̂. In
the next step, we try to express the commutation relation with the annihilation
operator in terms of the fundamental anticommutation relations for fermions,
" #
âβ , N̂ = âβ , ∑ âγ âγ = ∑ âβ , â†γ âγ
†

γ γ

âβ â†γ âγ + â†γ âβ âγ − â†γ âβ âγ − â†γ âγ âβ

=∑
γ

âβ , â†γ âγ − â†γ âβ , âγ = ∑ δβ γ âγ = âβ .

=∑ (2.2.2)
γ γ

We follow similar steps for the commutator with the corresponding creation op-
erator:
" #
[âα , N̂] = âα , ∑ âγ âγ = ∑ â†α , â†γ âγ
† † †

γ γ

â†α â†γ âγ + â†γ â†α âγ − â†γ â†α âγ − â†γ âγ â†α

=∑
γ

= ∑ â†α , â†γ âγ − ∑ â†γ â†α , âγ = − ∑ â†γ âγ , â†α = − ∑ â†γ δαγ

γ γ γ γ

= −â†α . (2.2.3)

Using these commutators in Eq. (2.2.1), we find

†
âα âβ , N̂ = â†α âβ , N̂ + â†α , N̂ âβ = â†α âβ − â†α âβ = 0.

The particle number operator thus commutes with any one-body Hamiltonian.
(b) For the commutator with the two-body part of the Hamiltonian, we proceed
in a similar manner. First, h we identify i the terms within the relevant sums,
Ĥ1 , N̂ = 21 ∑αβ δ γ vαβ γδ â†α â†β âδ âγ , N̂ . To evaluate this commutator, we add

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 16 — #32

i i

16 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

and remove terms to express it as commutators of individual creation and anni-

hilation operators with N̂:
h i
â†α â†β âδ âγ , N̂ = â†α â†β âδ âγ N̂ − N̂ â†α â†β âδ âγ

= â†α â†β âδ âγ N̂ − â†α â†β âδ N̂ âγ + â†α â†β âδ N̂ âγ − N̂ â†α â†β âδ âγ
= â†α â†β âδ âγ , N̂ + â†α â†β âδ N̂ âγ − â†α â†β N̂ âδ âγ + â†α â†β N̂ âδ âγ

− N̂ â†α â†β âδ âγ

= â†α â†β âδ âγ , N̂ + â†α â†β âδ , N̂ âγ + â†α â†β N̂ âδ âγ

− â†α N̂ â†β âδ âγ + â†α N̂ â†β âδ âγ − N̂ â†α â†β âδ âγ
h i
= â†α â†β âδ [âγ , N̂] + â†α â†β [âδ , N̂]âγ + â†α â†β , N̂ âδ âγ

+ â†α , N̂ â†β âδ âγ

= â†α â†β âδ âγ + â†α â†β âδ âγ − â†α â†β âδ âγ − â†α â†β âδ âγ = 0 .

Taking into account that the commutator is a linear operation, we can thus write

Ĥ, N̂ = Ĥ0 , N̂ + Ĥ1 , N̂ = 0 . (2.2.4)

This indicates that the two operators commute and hence can be diagonalized
separately. This has two significant consequences. First, N̂ is a constant of mo-
tion. Second, a common basis of many-body states can be built from the eigen-
states of both N̂ and Ĥ. These eigenstates can be labeled by the eigenvalues of
both N̂ and Ĥ providing eigenenergies EαN .

(c) We can repeat the same steps followed in the fermionic case, but this time we
want to express our commutators in terms of fundamental commutators (and not
anticommutators) of annihilation and creation operators. This essentially reduces
to recomputing the commutators in Eqs. (2.2.2) and (2.2.3):
" #
âβ , N̂ = âβ , ∑ âγ âγ = ∑ âβ , â†γ âγ = ∑ âβ , â†γ âγ + â†γ âβ , âγ
†

γ γ γ

= ∑ δβ γ âγ = âβ .
γ

This coincides with the results for fermions. Now, the commutator with the cre-
ation operator:
" #
†
âα , N̂ = âα , ∑ âγ âγ = ∑ â†α , â†γ âγ = ∑ â†α , â†α âγ + â†γ â†α , âγ
† †

γ γ γ

= −∑ â†γ δαγ = −â†α ,

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 17 — #33

i i

Second quantization 17

which also matches what was obtained for fermions. Now, given that the com-
mutators with the number operator N̂ for bosons are identical to their fermionic
counterparts, it is immediate from the previous proof that any bosonic
Hamilto-
nian Ĥ will also commute with the number of particles operator, Ĥ, N̂ = 0.

PROBLEM 2.3
Anticommutation relations of field operators

We define the field operators as

Ψ̂(~r) ≡ ∑ ϕ~k (~r)â~k , Ψ̂† (~r) ≡ ∑ ϕ~k∗ (~r)â~† , (2.3.1)

k
~k ~k

where
1 i~k·~r
ϕ~k (~r) = e (2.3.2)
Ω1/2
are single-particle momentum eigenfunctions, normalized to a volume Ω assum-
ing periodic boundary conditions. â~† (â~k ) are the creation (annihilation) operators
k
of single-particle states with a well-defined momentum ~k. Prove that the field op-
erators satisfy the following anticommutation relations:

Ψ̂(~r), Ψ̂(~r 0 ) = Ψ̂† (~r), Ψ̂† (~r 0 ) = 0;

(a)

(b) Ψ̂(~r), Ψ̂† (~r 0 ) = δ (~r −~r 0 ).

(a) To find the corresponding operators, we use the definition of the field operators
in terms of creation and annihilation operators:
( )
Ψ̂(~r), Ψ̂(~r ) = ∑ ϕ~k (~r)â~k , ∑ ϕ~k (~r )â~k 0 = ∑ ∑ ϕ~k (~r)ϕ~k 0 (~r 0 ) â~k , â~k 0 .
0 0

~k ~k 0 ~k ~k 0

Then, using the anticommutation relations of the creation and annihilation oper-
ators, â~k , â~k 0 = 0, we conclude that

Ψ̂(~r), Ψ̂(~r 0 ) = 0 .

Taking Hermitian conjugates, or working with a very similar argument, leads to

the same result for the conjugate operators, Ψ̂† (~r), Ψ̂† (~r 0 ) .
(b) To find the mixed anticommutator, we use similar mathematical machinery, start-
ing from
( )
n o
Ψ(~r), Ψ̂† (~r 0 ) = ∑ ϕ~k (~r)â~k , ∑ ϕ~k∗0 (~r)â~† = ∑ ∑ ϕ~k (~r)ϕ~k∗0 (~r 0 ) â~k , â~† 0 .

k k
~k ~k 0 ~k ~k 0

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 18 — #34

i i

18 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory
n o
Then, taking into account that â~k , â~† 0 = δ~k~k 0 one gets that
k

Ψ(~r), Ψ̂† (~r 0 ) = ∑ ∑ ϕ~k (~r)ϕ~k∗0 (~r 0 )δ~k~k 0 = ∑ ϕ~k (~r)ϕ~k∗ (~r 0 ) = δ (~r −~r 0 ) ,

~k ~k 0 ~k

due to the completeness relation of the wavefunctions ϕ~k (~r).

PROBLEM 2.4
Particle density operator

The density operator for a system of N particles in first quantization is expressed

as
N
ρ(~r) = ∑ δ (~r −~ri ) . (2.4.1)
i=1

(a) Show that the same operator in second quantization is ρ̂(~r) = Ψ̂† (~r)Ψ̂(~r),
where Ψ̂(~r) = ∑l ϕl (~r)âl is the field operator.

(b) Show that the particle number operator N̂ = ∑l â†l âl can be expressed as
N̂ = d~rρ̂(~r) . Note that â†l (âl ) are the creation (annihilation) operators cor-
R

responding to a complete single-particle basis.

(a) In second quantization, the expression for any one-body operator is given, see
Eq. (A.0.12), by
ρ̂(~r) = ∑ hl|δ (~r −~ri )|mi â†l âm ,
lm

where hl|δ (~r −~ri )|mi are the one-body matrix elements of the single-particle
component of the operator, which is the same independently of the value of i.
We employ a complete single-particle basis ϕl (~r), with associated creation (de-
struction) operators â†l (âl ). With this in mind, the previous expression can be
written as
Z
ρ̂(~r) = ∑ â†l âm d~ri ϕl∗ (~ri )δ (~r −~ri )ϕm (~ri ) = ∑ â†l âm ϕl∗ (~r)ϕm (~r)
lm lm
!
= ∑ ϕl∗ (~r)â†l ∑ ϕm (~r)âm = Ψ̂† (~r)Ψ̂(~r). (2.4.2)
l m

It is important to note that the integration is performed over the single-particle

coordinate ~ri , whereas the position ~r represents the point at which the density is
evaluated.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 19 — #35

i i

Second quantization 19

(b) We start from the integral and use the result of the previous section to find:
Z Z Z
!
† ∗ †
d~rρ̂(~r) = d~r Ψ̂ (~r)Ψ̂(~r) = d~r ∑ ϕl (~r)âl ∑ ϕm (~r)âm
l m
Z
= ∑ â†l âm d~rϕl∗ (~r)ϕm (~r) = ∑ â†l âm δlm = ∑ â†l âl ≡ N̂, (2.4.3)
lm lm l

where we have used the normalization of the single-particle wavefunctions.

PROBLEM 2.5
The states generated by field operators are eigenvectors of the particle num-
ber operator

Using the anticommutation relations of the field operators Ψ̂† (~r) and Ψ̂(~r) of a
system of identical fermions, show that

(a) Ψ̂† (~r) |0i is an eigenstate of the particle number operator.

(b) Ψ̂† (~r)Ψ̂† (~r 0 ) |0i is also an eigenstate of N̂.

In both cases, determine the corresponding eigenvalues.

(a) We can directly apply N̂ on Ψ̂† (~r 0 ) |0i and manipulate the expression to find:
Z Z
d~r Ψ̂† (~r)Ψ̂(~r) Ψ̂† (~r 0 ) |0i = d~r Ψ̂† (~r) δ (~r −~r 0 ) − Ψ̂† (~r 0 )Ψ̂(~r) |0i

= Ψ̂† (~r 0 ) |0i . (2.5.1)

Therefore, Ψ̂† (~r) |0i is an eigenvector of N̂ with eigenvalue 1.

(b) Again, we can directly apply N̂ onto the state and see that:
Z
d~r Ψ̂ (~r)Ψ̂(~r) Ψ̂† (~r 0 )Ψ̂† (~r 00 ) |0i
†

Z
d~r Ψ̂† (~r) −Ψ̂† (~r 0 )Ψ̂(~r) + δ (~r −~r 0 ) Ψ̂† (~r 00 ) |0i

=
Z Z
= d~r Ψ̂† (~r)δ (~r −~r 0 )Ψ̂† (~r 00 ) |0i − d~r Ψ̂† (~r)Ψ̂† (~r 0 )Ψ̂(~r)Ψ̂† (~r 00 ) |0i
Z
= Ψ̂† (~r 0 )Ψ̂† (~r 00 ) |0i − d~r Ψ̂† (~r)Ψ̂† (~r 0 ) δ (~r −~r 00 ) − Ψ̂† (~r 00 )Ψ̂(~r) |0i

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 20 — #36

i i

20 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Therefore, Ψ̂† (~r 0 )Ψ̂† (~r 00 ) |0i is an eigenvector of N̂ with eigenvalue 2.

PROBLEM 2.6
Heisenberg representation of creation and destruction operators for a single-
particle Hamiltonian

Consider a Hamiltonian which is diagonal in a given single-particle basis:

Ĥ0 = ∑ εα â†α âα . (2.6.1)

Find âα and â†α in the Heisenberg representation .

We start with the definition of a Heisenberg picture annihilation operator,

Ĥ0 t Ĥ0 t
âα,H (t) = ei h̄ âα e−i h̄ .

This definition can be expanded in powers of t in terms of nested commutators,

t i2 t2
âα,H (t) = âα + i[Ĥ0 , âα ] + [Ĥ0 , [Ĥ0 , âα ]] 2 + ... . (2.6.2)
h̄ 2! h̄
We evaluate the first commutator,
" #
Ĥ0 , âα = ∑ εγ â†γ âγ , âα = ∑ εγ [â†γ âγ , âα ] = ∑ εγ (â†γ âγ âα − âα â†γ âγ )

γ γ γ

=∑ εγ (â†γ âγ âα + â†γ âα âγ − â†γ âα âγ − âα â†γ âγ )
γ

= ∑ εγ â†γ âγ , âα − â†γ , âα aγ = − ∑ εγ δγα âα

γ γ

= −εα âα ,

employing the fermionic anticommutation relations. In other words, the first

commutator is just
Ĥ0 , âα = −εα âα .
The second commutator in Eq. (2.6.2) can easily be computed from this result,

[Ĥ0 , [Ĥ0 , âα ]] = [Ĥ0 , −εα âα ] = −εα [Ĥ0 , âα ] = εα2 âα .

In turn, the n−th order nested commutator is simply

[Ĥ0 , [Ĥ0 , . . . , [Ĥ0 , âα ]] . . . ] = (−1)n εαn âα .

| {z }
n times

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 21 — #37

i i

Second quantization 21

Taking into account this result, one can write the complete series expansion on t,

t i2 t 2 2 i3 3 t 3
âα,H (t) = âα − iεα âα + εα âα − ε âα + · · ·
h̄ 2! h̄2 3! α h̄3
i2

t
= âα 1 − iεα + εα2 − · · · , (2.6.3)
h̄ 2!
and therefore,
εα t
âα,H (t) = âα e−i h̄ . (2.6.4)
This gives the Heisenberg picture version of the destruction operator.
We could have obtained this result with an alternative procedure. One may write
the equation of motion of the destruction operator âα,H :
d h i
âα,H (t) = âα,H (t), Ĥ0 = eiĤ0 t/h̄ âα e−iĤ0 t/h̄ , Ĥ0

ih̄
dt " #
= eiĤ0 t/h̄ âα , ∑ εγ â†γ âγ e−iĤ0 t/h̄
γ
iĤ0 t/h̄
∑ εγ âα , â†γ âγ e−iĤ0t/h̄ .

=e
γ

Now we calculate the commutator

h i
âα , â†k âk = δkα âα ,

and, substituting in the previous expression, we have:

d
ih̄
dt
âα,H (t) = eiĤ0 t/h̄ ∑ εk δkα âα e−iĤ0t/h̄ = eiĤ0t/h̄ εα âα e−iĤ0t/h̄ = εα âα,H (t).
k

This indicates that the destruction operator in the Heisenberg picture fulfills the
following closed first-order differential equation:
d
ih̄ âα,H (t) = εα âα,H (t). (2.6.5)
dt
The solution of this equation is straightforward:

âα,H (t) = âα e−iεα t/h̄ , (2.6.6)

with the boundary condition for t = 0, âα,H (t = 0) = âα . This is the same result
obtained in Eq. (2.6.4).
We could follow either of these two methods to find the equivalent expression for
the creation operator â†α,H (t). This, however, can be obtained directly from the
Hermitian conjugate of Eq. (2.6.4), so we get

â†α,H (t) = â†α eiεα t/h̄ . (2.6.7)

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 22 — #38

i i

22 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 2.7
A canonical linear transformation

Show that a linear transformation of the creation and destruction operators,

defined by
dˆk† = ∑ Dlk â†l , dˆk0 = ∑ D∗lk0 âl , (2.7.1)
l l

is a canonical transformation , i.e. it preserves the anticommutation relations of

the creation and annihilation operators, as long as D is unitary.

We first compute the anticommutation relations of the new set of creation opera-
tors,
n o
dˆk† , dˆk†0 = ∑ Dnk â†n , ∑ D∗mk â†m = ∑ ∑ Dnk Dmk0 â†n , â†m = 0 . (2.7.2)

n m n m

This is the expected

result. We can do a similar calculation for the destruction
operators, to find dˆk , dˆk0 = 0. Note that these anticommutation relations are
preserved for any linear transformation, independently of whether it is unitary or
not.
We now compute the mixed anticommutator, and see that
n o
dˆk , dˆk†0 = ∑ D∗nk ân , ∑ Dmk0 â†m = ∑ D∗nk Dmk0 ân , â†m

n m nm
= ∑ D∗nk Dmk0 δnm = ∑ D∗nk Dnk0 = ∑(D† )kn (D)nk0
nm n n
= (D† D)kk0 = δkk0 . (2.7.3)

We have exploited the fact that D is unitary in the last line.

PROBLEM 2.8
Spin operators in second quantization

Consider a uniform system of fermions of spin S = 1/2. We build a Fock space

using a single-particle basis of plane waves, i.e. states with good linear momentum
~k and spin projections (↑, ↓) on the z-axis. We associate to these single-particle
states the corresponding creation (annihilation) operators â~† , â~† , â~k↑ , and â~k↓ .
k↑ k↓

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 23 — #39

i i

Second quantization 23

(a) Prove that the components of the total spin operator, ~S, are given by:

1 h † i
Ŝx = ∑ â~ â~k↓ + â~† â~k↑ ,
2 ~ k↑ k↓
k
−i h i
Ŝy = ∑ â~† â~k↓ − â~† â~k↑ ,
2 ~ k↑ k↓
k
1 h i
Ŝz = ∑ â~† â~k↑ − â~† â~k↓ . (2.8.1)
2 ~ k↑ k↓
k

(b) Calculate the expectation value of Ŝz , Ŝx , Ŝy on the unpolarized Fermi sea,

|ΨFS i = ∏ â~†k↑ ∏ â~†k↓ |0i . (2.8.2)

k≤kF k≤kF

|ΨPFS i = ∏ â~†k↑ |0i . (2.8.3)

k≤kF

(a) The operators Ŝx , Ŝy , and Ŝz are one-body operators. We shall employ
Eq. (A.0.12) to transform each operator into the second quantization represen-
tation. We start with Ŝz , which in second quantization reads

Ŝz = ∑ h~kσ |ŝz |~lσ 0 i â~† â~lσ 0 . (2.8.4)

kσ
~kσ ,~lσ 0

Here, ŝz = (1/2)σ̂z , with σ̂z the corresponding Pauli matrix. With this, the one-
body matrix elements read
1
h~kσ |ŝz |~lσ 0 i = h~k|~l i hσ |ŝz |σ 0 i = δ~k~l (σ̂ z )σ σ 0 , (2.8.5)
2
where (σ̂ k )σ σ 0 denote the matrix elements of the k−th Pauli matrix, such that
(σ̂ k )↑↑ = +1 and (σ̂ k )↓↓ = −1. With this, we find that the operator is expressed
as
1 h i
Ŝz = ∑ â~† â~k↑ − â~† â~k↓ . (2.8.6)
2 ~ k↑ k↓
k

We now proceed to find the second quantization expression for Ŝx . We start once
again from Eq. (A.0.12),

Ŝx = ∑ h~kσ |ŝx |~lσ 0 i â~† â~lσ 0 , (2.8.7)

kσ
~kσ ,~lσ 0

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 24 — #40

i i

24 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

and the one-body matrix elements are now related to the Pauli matrix σ̂x ,
1
h~kσ |ŝx |~lσ 0 i = h~k|~l i hσ |ŝx |σ 0 i = δ~k~l (σ̂ x )σ σ 0 . (2.8.8)
2
Using the fact that
(σ̂ x )σ σ 0 = δσ −σ 0 , (2.8.9)
we find that
1 h † i
Ŝx = ∑ â~ â~k↓ + â~† â~k↑ . (2.8.10)
2 ~ k↑ k↓
k

Finally, for Ŝy , the expression reads

Ŝy = ∑ h~kσ |ŝy |~lσ 0 i â~† â~lσ 0 . (2.8.11)

kσ
~kσ ,~lσ 0

In this case, the one-body matrix elements are related to the σ̂ y Pauli matrix,
1
h~kσ |ŝy |~lσ 0 i = h~k|~l i hσ |ŝy |σ 0 i = δ~k~l (σ̂ y )σ σ 0 . (2.8.12)
2
y y
Now, for this matrix we know that σ̂↓,↑ = −i and σ̂↑,↓ = i, which translates into
the following second quantization expression
−i h † i
Ŝy = ∑ â~ â~k↓ − â~† â~k↑ . (2.8.13)
2 ~ k↑ k↓
k

(b) We now proceed to find the expectation value of each total spin component in an
unpolarized free Fermi sea. We start with the Ŝz component:
* +
1 h † †
i
ΨFS Ŝz ΨFS = ΨFS ∑ â~ â~k↑ − â~ â~k↓ ΨFS
2 ~ k↑ k↓
k
N N
= − = 0. (2.8.14)
2 2

This result is a direct consequence of the fact that the operator â~† â~k↑ counts the
k↑
number of particles with spin up, which by construction in the unpolarized Fermi
sea is N/2. Similarly, the operator â~† â~k↓ counts the number of particles with spin
k↓
down, which by construction is also N/2. Overall, the expectation value is zero.
In other words, |ΨFS i is an eigenstate of Ŝz with a zero eigenvalue.
Let us analyze now the expectation value of Ŝx . We use the expression of
Eq. (2.8.10) and compute the expectation value to find
1 h † i
hΨFS |Ŝx |ΨFS i = hΨFS | ∑ â~ â~k↓ + â~† â~k↑ |ΨFS i = 0 . (2.8.15)
2 ~ k↑ k↓
k

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 25 — #41

i i

Second quantization 25

This result arises because, in the first term, we destroy a particle with momentum
~k and spin projection ↓ below the Fermi sea. If we try to create a particle with the
same momentum (below the Fermi sea) and opposite spin projection, the state is
already occupied by construction of the free Fermi sea. The Pauli principle does
not allow this operation. The same is true for the second operator, which destroys
a particle with up spin and tries to create one with a down component.
This very same argument applies to the Ŝy operator. Indeed, Eq. (2.8.13) is just
a different linear combination of the same operators involved in Ŝx , Eq. (2.8.10).
Consequently, the same arguments are valid and one easily proves that

hΨFS |Ŝy |ΨFS i = 0 . (2.8.16)

(c) We build the polarized Fermi, from a state built of spins pointing upwards,
|ΨPFS i = ∏k≤kF â~† |0i. The expectation value of Ŝz in this state thus reads
k↑

1 h † i N
hΨPFS |Ŝz |ΨPFS i = hΨPFS | ∑ â~ â~k↑ − â~† â~k↓ |ΨPFS i = . (2.8.17)
2 ~ k↑ k↓ 2
k

The operator counts only the particles with spin ↑, which are the ones present in
the spin-polarized system.
However, for Ŝx and Ŝy one can apply similar arguments to that given in the previ-
ous section in terms of the mixed products of creation and destruction operators.
With the combination a~† â~k↑ , for instance, one destroys a particle with spin ↑ and
k↓
creates one with spin ↓. The corresponding state is, however, orthogonal to the
spin-up state |ΨPFS i. The opposite term, a~† â~k↓ , tries to destroy a particle with
k↑
spin ↓, which does not exist in |ΨPFS i. All in all, we find that both expectation
values are zero. For instance,
1 h i
hΨPFS |Ŝx |ΨPFS i = hΨPFS | ∑ â~† â~k↓ + â~† â~k↑ |ΨPFS i = 0 . (2.8.18)
2 ~ k↑ k↓
k

The same arguments apply for Ŝy and we indeed get:

hΨPFS |Ŝy |ΨPFS i = 0 . (2.8.19)

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 27 — #43

i i

Wick’s theorem
3
In the context of many-body quantum mechanics, one often encounters the need to
evaluate expectation values of operators or their matrix elements. For instance, given
a Hamiltonian Ĥ, we typically want to evaluate its expectation value, which is the
energy of the system, E = hĤi. In second quantization, the matrix element of an op-
erator can always be expressed as a vacuum expectation value. This generally yields
an expression which is hard to evaluate. Fortunately, a powerful tool to simplify
these calculations exists: Wick’s theorem. This theorem provides a systematic way
to break down expressions involving products of creation and annihilation operators
into sums of simpler terms.
The statement of the theorem is as follows:

Â1 . . . Ân =N (Â1 . . . Ân )

+ ∑ N (Â1 . . . Âi . . . Â j . . . Ân )

(i j)

+ ∑ N (Â1 . . . Âi . . . Âk . . . Â j . . . Âl . . . Ân )

(i j)(kl)
..
.
+ sum over fully-contracted products .

Here, Âi denotes an arbitrary creation or destruction operator, and N is the normal-
ordering operator. This ordering is such that all creation operators are to the left of
all destruction operators. In addition, Wick contractions of two operators Âi , Â j are
indicated with Âi Â j .
In essence, Wick’s theorem separates the products of creation and destruction
operators into their normal-ordered part, plus all the possible normal-ordered con-
tractions (which may involve a phase in the fermionic case). When evaluated on
the vacuum, normal-ordered strings of operators are typically zero, and only fully-
contracted terms remain. We refer the reader to Appendix B for a more detailed
explanation of Wick’s theorem.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 28 — #44

i i

28 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 3.1
Contraction of two creation and destruction operators

Consider a basis of single-particle states (α, β , γ, ...) associated to a fermionic sys-

tem, where α destruction represents one or more quantum numbers necessary to
characterize the single-particle states. Consider now the creation and destruction
operators corresponding to this single-particle basis, that fulfill the standard an-
ticommutation relations. For two generic operators â and b̂, their contraction is
defined as:
âb̂ = âb̂ − N (âb̂) (3.1.1)

(a) Calculate the following contractions and their expectation values on the vac-
uum:
âα âβ , â†α â†β , â†α âβ , âα â†β . (3.1.2)

(b) What is the expectation value in the vacuum of a product of an odd number of
creation and destruction operators?

(a) We compute the contraction of the four terms one by one:

âα âβ = âα âβ − N (âα âβ ) = âα âβ − âα âβ = 0 , (3.1.3)

â†α â†β = â†α â†β − N (â†α â†β ) = â†α â†β − â†α â†β = 0 , (3.1.4)

â†α âβ = â†α âβ − N (â†α âβ ) = â†α âβ − â†α âβ = 0 , (3.1.5)

âα â†β = âα â†β − N (âα â†β ) = âα â†β + â†β âα = δαβ . (3.1.6)

The first two contractions are zero since the operators are of the same type, and
the normal ordering does not affect the order of such operators. The third term
is zero because the initial string is already normal-ordered. The only contraction
that is different from zero is the last one. Here, the operators were not in normal
order from the start. Upon ordering them, a − sign arises from the corresponding
fermionic permutation and the final result is a Kronecker delta of the two states.
Clearly, the vacuum expectation values of the first three contractions is zero. For
the fourth one, we find:

h0|âα â†β |0i = h0|δαβ |0i = δαβ . (3.1.7)

This expectation value is different from zero only when α = β .

(b) By the Wick theorem, products of creation and destruction operators are formed
of normal-ordered terms with no contractions; a pair of contractions; two pairs of

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 29 — #45

i i

Wick’s theorem 29

contractions; and so on until all operators are contracted in pairs. The expectation
value of any normal-ordered term is zero, because destruction operators are on
the right and, upon acting on the vacuum, they yield 0. The only term that can
be non-zero is the fully contracted one. However, if the number of operators is
odd, is not possible to have all operators contracted in pairs. There will always
be either a creation or a destruction operator left without a contraction. Upon
evaluating the expectation value on the vacuum, the result is necessarily zero.

PROBLEM 3.2
Matrix elements using the Wick theorem: One-body operator

Consider the one-body operator

T̂ = ∑ (α|tˆ|β )â†α âβ , (3.2.1)

α,β

where the single-particle states α are the basis used to build the Fock space and
the one-body operator in first quantization is

N
T̂ = ∑ tˆi . (3.2.2)
i=1

(a) c1 T̂ c1 ,

(b) c1 T̂ c2 ,

(a) We start with the diagonal matrix element,

c1 T̂ c1 = h0| â2 â1 ∑ (α | tˆ | β )â†α âβ â†1 â†2 |0i

α,β

= ∑ (α | tˆ | β ) h0|â2 â1 â†α âβ â†1 â†2 |0i . (3.2.3)

α,β

Now, we should recall the results of the previous problem,

âα âβ = 0, â†α â†β = 0, â†β âβ = 0, âα â†β = δαβ . (3.2.4)

The only non-zero result of the expectation value is the one obtained from con-
tracting all possible operators. We only consider contractions between different

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 30 — #46

i i

30 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

types of operators that are not normal ordered, which reduces to two possible
fully-contracted terms:

c1 T̂ c1 = ∑(α | tˆ | β ) h0|â2 â1 â†α âβ â†1 â†2 |0i

αβ

+ ∑(α | tˆ | β ) h0|â2 â1 â†α âβ â†1 â†2 |0i

αβ

= ∑(α | tˆ | β ) δα1 δβ 1 + δα2 δβ 2
αβ

= (1 | tˆ | 1) + (2 | tˆ | 2) . (3.2.5)

In other words, the expectation value is the sum of the single-particle expectation
values over the two single-particle states, 1 and 2, included in the |c1 i state.
(b) We now turn our attention to the calculation of a matrix element, which differs
from the expectation value in that the bra and ket states are different. In this case,
we find:

c1 T̂ c2 = ∑(α | tˆ | β ) h0|a2 a1 a†α aβ a†1 a†4 |0i

αβ

= ∑(α | tˆ | β ) h0|â2 â1 â†α aβ a†1 â†4 |0i

αβ

+ ∑(α | tˆ | β ) h0|â2 â1 â†α aβ â†1 a†4 |0i

αβ

= ∑(α | tˆ | β )δ2α δβ 4
αβ

= (2 | tˆ | 4) . (3.2.6)

In the end, the many-body matrix element is equal to one single-particle matrix
elements between the two different states, 2 and 4. Clearly, this matrix element
is non-zero because there is a contraction between the common state 1 included
in both |c1 i and |c2 i.
(c) We are finally asked to compute the matrix element c1 T̂ c3 . In this case, |c1 i
and |c3 i do not share any single-particle indices. Because the two states differ in
all single-particle states and the operator is of one-body nature, there is no way
to connect these two states with one another. As a consequence, the expectation
value is null,

c1 T̂ c3 = 0. (3.2.7)

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 31 — #47

i i

Wick’s theorem 31

PROBLEM 3.3
Matrix elements using the Wick theorem: Two-body operator

Consider the two-body operator

1
V̂ = ∑ (αβ | v̂ | γδ )â†α â†β âδ âγ ,
2 αβ
(3.3.1)
γδ

where
Z Z
(αβ | v̂ | γδ ) = d~r d~r0 ϕα∗ (~r)ϕβ∗ (~r0 )v(|~r −~r0 |)ϕγ (~r)ϕδ (~r0 ) (3.3.2)

are the two-body matrix elements expressed in real space, and the single-particle
states α are the basis used to build the Fock space.
Consider the Fock states:

Calculate the following expectation values:

(a) A V̂ A ,

(b) A V̂ B ,

(d) A V̂ D .

(a) We start with the diagonal matrix element,

We employ once again the results of Problem 3.1,

âα âβ = 0, â†α â†β = 0, â†β âβ = 0, âα â†β = δαβ . (3.3.4)

We also need to take into account the phases originated by the parity of the per-
mutations required to put the contracted operators together. With this, we find 12

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 32 — #48

i i

32 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

contractions that are not zero:

1
A V̂ A = ∑ (αβ | v̂ | γδ )
2 αβ γδ


× h0|â1 â2 â3 â†α â†β âδ âγ â†3 â†2 â†1 |0i + h0|â1 â2 â3 â†α â†β âδ âγ â†3 â†2 â†1 |0i


+ h0|â1 â2 â3 â†α â†β âδ âγ â†3 â†2 â†1 |0i + h0|â1 â2 â3 â†α â†β âδ âγ â†3 â†2 â†1 |0i

+ h0|â1 â2 â3 â†α â†β âδ âγ â†3 â†2 â†1 |0i + h0|â1 â2 â3 â†α â†β âδ âγ â†3 â†2 â†1 |0i


+ h0|â1 â2 â3 â†α â†β âδ âγ â†3 â†2 â†1 |0i + h0|â1 â2 â3 â†α â†β âδ âγ â†3 â†2 â†1 |0i .


(3.3.5)

The result of the contractions can be expressed in terms of Kronecker deltas,

1
A V̂ A = ∑ (αβ | v̂ | γδ )
2 αβ γδ

× −δα2 δβ 3 δγ3 δδ 2 + δα2 δβ 3 δγ2 δδ 3
+ δα3 δβ 2 δγ3 δδ 2 − δα3 δβ 2 δγ2 δδ 3
− δα1 δβ 3 δγ3 δδ 1 + δα1 δβ 3 δγ1 δδ 3
+ δα3 δβ 1 δγ3 δδ 1 − δα3 δβ 1 δγ1 δδ 3
− δα1 δβ 2 δγ2 δδ 1 + δα1 δβ 2 δγ1 δδ 2

+ δα2 δβ 1 δγ2 δδ 1 − δα2 δβ 1 δγ1 δδ 2 . (3.3.6)

Collapsing the sums over single-particle indices with the help of the Kronecker
functions, and using the fact that V̂ is a central potential such that (αβ | v̂ | γδ ) =

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 33 — #49

i i

Wick’s theorem 33

(β α | v̂ | δ γ), we get:

1
A V̂ A = [−(23 | v̂ | 32) + (23 | v̂ | 23)
2
+ (32 | v̂ | 32) − (32 | v̂ | 23)
− (13 | v̂ | 31) + (13 | v̂ | 13)
+ (31 | v̂ | 31) − (31 | v̂ | 13)
− (12 | v̂ | 21) + (12 | v̂ | 12)
+ (21 | v̂ | 21) − (21 | v̂ | 12)]
= (23 | v̂ | 23 − 32) + (13 | v̂ | 13 − 31) + (12 | v̂ | 12 − 21) . (3.3.7)

Each one of the terms on the r.h.s. has both a so-called direct (or Hartree) matrix
element, (αβ | v̂ | αβ ), and an exchange (or Fock) term, (αβ | v̂ | β α), where
the bra and ket states are exchanged. The result is the sum over the 3 possible
combinations of the 3 available states, 12, 13 and 23.
(b) Let us now calculate the matrix element A V̂ B . In analogy with the previous
case, we can write the matrix element in terms of fully contracted terms,
1
A V̂ B = ∑ (αβ | v̂ | γδ )
2 αβ γδ


× h0|â1 â2 â3 â†α â†β âδ âγ â†4 â†2 â†1 |0i + h0|â1 â2 â3 â†α â†β âδ âγ â†4 â†2 â†1 |0i


+ h0|â1 â2 â3 â†α â†β âδ âγ â†4 â†2 â†1 |0i + h0|â1 â2 â3 â†α â†β âδ âγ â†4 â†2 â†1 |0i

+ h0|â1 â2 â3 â†α â†β âδ âγ â†4 â†2 â†1 |0i + h0|â1 â2 â3 â†α â†β âδ âγ â†4 â†2 â†1 |0i


+ h0|â1 â2 â3 â†α â†β âδ âγ â†4 â†2 â†1 |0i + h0|â1 â2 â3 â†α â†β âδ âγ â†4 â†2 â†1 |0i .


(3.3.8)

Notice that now we have 8 terms instead of the 12 terms appearing in A V̂ A ,

because state |Bi has only 2 matching particles with state |Ai. Then, taking into
account the value of the contractions and the corresponding phases to put together

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 34 — #50

i i

34 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

the operators to be contracted, we obtain

1
A V̂ B = ∑ (αβ | v̂ | γδ )
2 αβ γδ

× δα2 δβ 3 δγ2 δδ 4 − δα2 δβ 3 δγ4 δδ 2
− δα3 δβ 2 δγ2 δδ 4 + δα3 δβ 2 δγ4 δδ 2
+ δα1 δβ 3 δγ1 δδ 4 − δα1 δβ 3 δγ4 δδ 1

− δα3 δβ 1 δγ1 δδ 4 + δα3 δβ 1 δγ4 δδ 1
1
= [(23 | v̂ | 24) − (23 | v̂ | 42) + (32 | v̂ | 42) − (32 | v̂ | 24)
2
+ (13 | v̂ | 14) − (31 | v̂ | 41) + (31 | v̂ | 41) − (31 | v̂ | 14)]
1
= [(23 | v̂ | 24 − 42) + (32 | v̂ | 42 − 24)
2
+ (13 | v̂ | 14 − 41) + (31 | v̂ | 41 − 14)]
= (23 | v̂ | 24 − 42) + (13 | v̂ | 14 − 41) . (3.3.9)

(c) The matrix element A V̂ C is somewhat simpler, because there is only one
common state (state 1), which reduces the number of fully contracted terms to 4:
1
A V̂ C = ∑ (αβ | v̂ | γδ )
2 αβ γδ


× h0|â1 â2 â3 â†α â†β âδ âγ â†5 â†4 â†1 |0i + h0|â1 â2 â3 â†α â†β âδ âγ â†5 â†4 â†1 |0i


+ h0|â1 â2 â3 â†α â†β âδ âγ â†5 â†4 â†1 |0i + h0|â1 â2 â3 â†α â†β âδ âγ â†5 â†4 â†1 |0i .


(3.3.10)
With the results of the contractions and the corresponding phases, the sums re-
duce to a single term,
1
A V̂ C = ∑ (αβ | v̂ | γδ ) δα2 δβ 3 δγ4 δδ 5 − δα2 δβ 3 δγ5 δδ 4
2 αβ γδ

− δα3 δβ 2 δγ4 δδ 5 + δα3 δβ 2 δγ5 δδ 4
1
= [(23 | v̂ | 45) − (23 | v̂ | 54) + (32 | v̂ | 54) − (32 | v̂ | 45)]
2
= (23 | v̂ | 45 − 54) . (3.3.11)

(d) Finally, the last matrix element is

A V̂ D = 0 , (3.3.12)

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 35 — #51

i i

Wick’s theorem 35

because states |Ai and |Di differ in two or more single-particle states. They can-
not be connected by a two-body operator, and hence the matrix element is neces-
sarily zero.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 37 — #53

i i

Green’s functions
4
Green’s functions are a cornerstone of quantum mechanics and many-body physics,
providing a systematic framework to describe the propagation of particles and exci-
tations in a system. They encode essential information about the evolution of a quan-
tum particle in space and time, serving as the foundation for understanding properties
such as energy levels, spectral functions and system responses to external perturba-
tions. Their broad applicability makes them invaluable for studying systems ranging
from simple, non-interacting particles, to highly correlated many-body systems.
This chapter introduces Green’s functions through a progression of problems
aimed at building both conceptual and practical skills. Starting with the Green
function of a free particle in one dimension, the readers will explore how these func-
tions capture the dynamics of non-interacting systems. Problems involving Fourier
transforms of propagators illustrate how Green’s functions connect the time and fre-
quency domains, highlighting their importance in quantum dynamics and spectral
analysis. This chapter also covers advanced topics, such as the definition of Green’s
functions in terms of field operators, and demonstrates their use in deriving key re-
sults like the kinetic energy and the Koltun sum rule for the free Fermi sea. By work-
ing through these problems, the readers will gain an understanding of the insights
Green’s functions provide into quantum mechanics and their pivotal role in modern
theoretical physics.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 38 — #54

i i

38 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 4.1
The Green’s function of a free particle in 1D

Calculate the propagator of a free particle ,

−i D Ĥt
E
G(α, β ;t) = α e−i h̄ β , (4.1.1)
h̄
in one dimension in both the coordinate and the momentum representations.

The Hamiltonian and the Schrödinger equation of a free particle read

p̂2 ∂ |Ψ(t)i
Ĥ = , ih̄ = Ĥ |Ψ(t)i , (4.1.2)
2m ∂t
while the time evolution operator is given by

Û(t) = e−iĤt/h̄ , (4.1.3)

such that
|Ψ(t)i = Û(t) |Ψ(t = 0)i . (4.1.4)
The eigenvectors of the Hamiltonian normalized to the continuum are given by

1 p2
hx|pi = √ eipx/h̄ , Ĥ |pi = |pi . (4.1.5)
2π h̄ 2m

In momentum space, the expression for the single-particle propagator reads

−i D Ĥt
E i i p2
G(p, p0 ;t) = p e−i h̄ p0 = − δ (p − p0 )e− h̄ 2m t , (4.1.6)
h̄ h̄
where we have taken into account that the momentum eigenstates, |pi, are the
eigenfunctions of Ĥ,
p2
Ĥ |pi = |pi , (4.1.7)
2m
and they are orthonormal, hp|p0 i = δ (p − p0 ). Note that Eq. (4.1.6) is indeed the
propagator in momentum space, and it is thus the answer to the question for the
momentum representation.
The definition of the propagator in the space representation is
−i D −i Ĥt 0 E
G(x, x0 ;t) = x e h̄ x . (4.1.8)
h̄

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 39 — #55

i i

Green’s functions 39
R
Using the closure relation in momentum space d p |pi hp| = 1 two times in this
equation, we find
i
Z Z D E
Ĥt
G(x, x0 ;t) = − dp d p0 hx|p0 i p e−i h̄ p0 hp|x0 i . (4.1.9)
h̄
We can now incorporate the expression of Eq. (4.1.6) in the integrals to find

i 1 p0 x i p2 px0
Z Z
G(x, x0 ;t) = − d p d p0 ei h̄ δ (p − p0 )e− h̄ 2m t e−i h̄
h̄ 2π h̄
i 1 p(x−x0 ) i p2
Z
=− d p ei h̄ e− h̄ 2m t .
h̄ 2π h̄
With this, we can identify,

i 1 p(x0 −x) i p2
Z
G(x, x0 ;t) = − d p e−i h̄ e− h̄ 2m t . (4.1.10)
h̄ 2π h̄
The next step is to perform this integral. The strategy is to complete the square
2 √
dφ e−φ = π. In fact, we
R∞
in the exponent, and take advantage of the result −∞
can write
"r √ #2
p2t p(x0 − x) it m 0 m 0
−i −i =− p+i√ (x − x) − (x − x)2 .
2mh̄ h̄ 2mh̄ 2it h̄ 2it h̄
(4.1.11)
Using this expression in the integral, one finds:
√ i2
p(x0 −x)
hq
1 i p2 1 √ m
Z ∞ Z ∞ it 0
− 2mh̄ p+i 2it h̄ (x −x)
m 0 2
d p e− h̄ 2m t e−i h̄ = dp e e− 2it h̄ (x −x) .
2π h̄ −∞ 2π h̄ −∞
(4.1.12)
Now, one can perform the change of variables,
r √ r
it m 0 2mh̄
ξ= p+i√ (x − x) , d p = dξ , (4.1.13)
2mh̄ 2it h̄ it

so the integral becomes

r
1 p2 p(x0 −x) 1 2mh̄ − m (x0 −x)2
Z ∞ Z ∞
− h̄i 2m t −i h̄ 2
dp e e = e 2it h̄ dξ e−ξ ,(4.1.14)
2π h̄ −∞ 2π h̄ it −∞

where we have pulled out the momentum-independent term in the exponent. Fi-
nally, one writes the propagator in real space in 1D:
r
i 1 2mh̄ − m (x0 −x)2 ∞
Z
0 2
G(x, x ;t) = − e 2it h̄ dξ eξ
h̄ 2π h̄ it −∞
r
i m i m (x0 −x)2
=− e 2t h̄ . (4.1.15)
h̄ 2πih̄t

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 40 — #56

i i

40 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

This is an oscillatory function in terms of the difference (x − x0 )2 .

In 3D, the Green’s function can be determined following the same procedure for
each dimension (x, y and z). Taking the product of the three terms, we find the
expression of the 3D propagator:
i m 3/2 i m (~r0 −~r)2
G(~r,~r0 ;t) = − e 2t h̄ . (4.1.16)
h̄ 2πih̄t

PROBLEM 4.2
Single-particle spectral function for a non-interacting system

Calculate the one-body Green’s function,

−i N h i
g(0) (α, β ;t,t 0 ) ≡ hΨ0 | T âαH (t)â†β (t 0 ) |ΨN
0i, (4.2.1)
h̄ H

for a non-interacting system described by the Hamiltonian

Ĥ0 = ∑ εα â†α âα , (4.2.2)

with all single-particle states occupied up to the Fermi level αF . The ground state
†
|ΨN
0 i = ∏i≤F âi |0i is an eigenstate of Ĥ0 ,

Ĥ0 |ΨN N
0 i = E0 |Ψ0 i , E0 = ∑ εα . (4.2.3)
α≤F

The time ordered single-particle Green’s function is defined by the expectation

value in Eq. (4.2.1), where | ΨN0 i is the ground state and all quantities are defined
in the Heisenberg representation. The time ordering operator is given by
h i
T âαH (t) â†β (t 0 ) = Θ(t − t 0 ) âαH (t)â†β (t 0 ) − Θ(t 0 − t) â†β (t 0 )âαH (t) , (4.2.4)
H H H

where Θ(t) is the step function, which is 0 for t < 0 and 1 for t > 0. Notice that
this definition includes a sign change when the two fermion operators are inter-
changed. The idea is that the time-ordering operator puts operators with earlier
time argument to the right, so they act first on the corresponding kets.
In the definition, | ΨN0 i is the normalized Heisenberg-representation ground state.
Finding this state is usually a problem of its own. However, full knowledge of the
state itself is not necessary to evaluate the Green’s function. The key point in the
present case is that this ground state has a known energy, E0 , as per Eq. (4.2.3).
In the Heisenberg picture, the single-particle operators âαH (t) and â†αH are given
by
Ĥt Ĥt
âαH (t) = ei h̄ âα e−i h̄ , (4.2.5)

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 41 — #57

i i

Green’s functions 41

and
Ĥt Ĥt
â†αH (t) = ei h̄ â†α e−i h̄ . (4.2.6)
Then, taking into account all these definitions, for any Ĥ we can write

−i

0 Ĥt Ĥt Ĥt 0 Ĥt 0
g(α, β ;t − t ) = Θ(t − t 0 ) hΨN0 | ei h̄ âα e−i h̄ ei h̄ â†β e−i h̄ |ΨN0 i
h̄
0 0

0 N i Ĥth̄ † −i Ĥth̄ i Ĥt −i Ĥt N
− Θ(t − t) hΨ0 | e âβ e e âα e
h̄ h̄ |Ψ0 i , (4.2.7)

which can be simplified as

−i n N 0 0
g(α, β ;t − t 0 ) = Θ(t − t 0 )eiE0 (t−t )/h̄ hΨN0 | âα e−iĤ(t−t )/h̄ â†β |ΨN0 i
h̄ o
N 0 0
− Θ(t 0 − t)eiE0 (t −t)h̄ hΨN0 | â†β e−iĤ(t −t)/h̄ âα |Ψn0 i . (4.2.8)

Using a closure relation for the eigenstates of the N + 1 and the N − 1 particle
systems, we can write

g(α, β ;t − t 0 ) =
−i

N N+1 0
Θ(t − t 0 ) ∑ ei(E0 −Em )(t−t )/h̄ hΨN0 |âα |ΨN+1 N+1 † N
m i hΨm |âβ |Ψ0 i
h̄ m
)
N N−1
)(t 0 −t)/h̄
− Θ(t 0 − t) ∑ ei(E0 −Em0 hΨN0 |â†β |ΨN−1 N−1 N
m0 i hΨm0 |âα |Ψ0 i . (4.2.9)
m0

Now, we should take into account that for the overlap to be nonzero, the state
|ΨN+1 N
m i must be constructed by just adding a particle to the state |Ψ0 i. To add a
particle, this must lie on a single-particle energy level above the Fermi level αF ,
which means that
N+1 † N+1 †
hΨN0 |âα |ΨN+1 N N
m i hΨm |âβ | Ψ0 i = δαβ Θ(α − αF ) hΨm |âα |Ψ0 i , (4.2.10)

Similarly, for the state |ΨN−1 N

m i a particle is subtracted to the state |Ψ0 i from a
single-particle energy level below the Fermi level

hΨN0 |â†β |ΨN−1 N−1 N N−1 N

m i hΨm |âα |Ψ0 i = δαβ Θ(αF − α) hΨm |âα |Ψ0 i . (4.2.11)

Notice that in our case, i.e. the free Hamiltonian, we have E0 − EmN+1 = −εα . The
excitation energy of the N + 1 particle state measured with respect to the ground
state of N particles is just the single-particle energy of the state added to the
system: εα . Similarly, for the non-interacting case, we know that E0N − EmN−1 =
εα .

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 42 — #58

i i

42 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

The hΨN−1 N N+1 † N

m |âα |Ψ0 i, hΨm |âα |Ψ0 i conditions permit to perform the summa-
tions and we get :

−iδαβ n 0
g(α, β ;t − t 0 ) = Θ(t − t 0 ) e−iεα (t−t )/h̄ Θ(α − αF )
h̄ o
0
−Θ(t 0 − t) eiεα (t −t)/h̄ Θ(αF − α) . (4.2.12)

PROBLEM 4.3
Fourier transform of the single-particle propagator

Calculate the Fourier transform of the single-particle propagator

Z ∞
0
g(0) (α, β ; E) = d(t − t 0 )eiE(t−t )/h̄ g(0) (α, β ;t − t 0 ) , (4.3.1)
−∞

where
iδαβ n 0
g(0) (α, β ;t − t 0 ) = − Θ(t − t 0 )e−iεα (t−t )/h̄ Θ(α − αF )
h̄ (4.3.2)
0
o
−Θ(t 0 − t)eiεα (t −t)/h̄ Θ(αF − α)

is the one-body Green’s function corresponding to the Hamiltonian

Ĥo = ∑ εα â†α âα (4.3.3)

with all single-particle states occupied up to the Fermi level αF , and the ground
state |φ0N i = ∏i≤F â†i |0i is an eigenstate of Ĥ0

Ĥ0 |φ0N i = E0 |φ0n i , E0 = ∑ εα (4.3.4)

α≤F

To calculate the Fourier transform, it is useful to consider the integral representa-

tion of the Θ-function:
0 0
1 e−iE (t−t )/h̄
Z ∞
Θ(t − t 0 ) = − dE 0 (4.3.5)
2πi −∞ E 0 + iη +

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 43 — #59

i i

Green’s functions 43

We start from the definition of the Fourier transform:

Z ∞
g(0) (α, β ; E) = d(τ)eiEτ/h̄ g(0) (α, β ; τ)
−∞
−iE 0 τ/h̄
( )
1 0 e
Z ∞ Z ∞
τ iEτ/h̄
=−i d e − dE 0 δαβ Θ(α − αF )e−iεα τ/h̄
−∞ h̄ 2πi −∞ E + iη +
iE 0 τ/h̄
( )
1 e
Z ∞ Z ∞
τ iEτ/h̄
+i d e − dE 0 0 δαβ Θ(α − αF )e−iεα τ/h̄ .
−∞ h̄ 2πi −∞ E + iη +
(4.3.6)
where τ = t − t 0 has been used.
The next step, is to change the order of integration,
1 1 ∞ τ i(E−E 0 −εα )τ/h̄
Z Z
g(0) (α, β ; E) = dE 0 δ Θ(α − αF ) d e
E 0 + iη + αβ 2π −∞ h̄
1 1 ∞ τ i(E+E 0 −εα )τ/h̄
Z Z
− dE 0 0 δαβ Θ(α F − α) d e .
E + iη + 2π −∞ h̄
(4.3.7)
The integration over d(τ/h̄) provides the δ -functions on the energy variables:

1
Z ∞ τ 0
d ei(E−E −εα )τ/h̄ = δ (E − E 0 − εα ) . (4.3.8)
2π −∞ h̄
1
Z ∞ τ 0
d ei(E+E −εα )τ/h̄ = δ (E + E 0 − εα ) . (4.3.9)
2π −∞ h̄
With the help of the δ -functions one can easily perform the integrations over the
energy variable:
1
Z
g(0) (α, β ; E) = dE 0 δ Θ(α − αF )δ (E − E 0 − εα )
E 0 + iη + αβ
1
Z
− dE 0 0 δ Θ(αF − α)δ (E + E 0 − εα )
E + iη + αβ

1 1
=δαβ Θ(α − αF ) − Θ(αF − α)
E − εα + iη + εα − E + iη +
Θ(α − αF ) Θ(αF − α)

=δαβ + . (4.3.10)
E − εα + iη + E − εα − iη +

Notice that one can also arrive at this expression by starting from the definition

1
g(0) (α, β ; E) = hΨN0 | aα a† |ΨN0 i
E − (Ĥ0 − EΨN ) + iη + β
0
1
+ hΨN0 | a†β aα |ΨN0 i . (4.3.11)
E − (EΨN − Ĥ0 ) − iη +
0

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 44 — #60

i i

44 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

In our case,
Ĥ0 a†β |ΨN0 i = (EΨN + εβ )a†β |ΨN0 i , (4.3.12)
0

and
Ĥ0 aβ |ΨN0 i = (EΨN − εβ )aβ |ΨN0 i . (4.3.13)
0

Finally,
"
(0) Θ(α − αF )
g (α, β ; E) =δαβ
E − (EΨN + εα − EΨN ) + iη +
0 0
#
Θ(αF − α)
+
E − (EΨN − (EΨN − εα )) − iη +
0 0
Θ(α − αF ) Θ(αF − α)

=δαβ + . (4.3.14)
E − εα + iη + E − εα − iη +

PROBLEM 4.4
Green’s function. Field operator definition.

Calculate the one-body Green’s function for the free Fermi sea, using the definition
in terms of field operators:

i h i
Gσ σ 0 (~rt,~r0 t 0 ) = − hφ0 | T Ψ̂H,σ (~r,t)Ψ̂†H,σ 0 (~r0 t 0 ) | φ0 i (4.4.1)
h̄
n o
where |φ0 i = ∏k<kF â†k |0i with k ≡ ~k, σ , is the ground state of the free Fermi
sea in the Fock space.

The field operators in the Heisenberg picture are given by

Ψ̂H (~r,t) = ∑ ϕm (~r) e−iEm t/h̄ âm , (4.4.2)

where âm is the annihilation operator associated to the single-particle state whose
wavefunction is
1 ~
ϕ(~r) = 1/2 eikm~r , (4.4.3)
Ω
with energy Em = h̄2 km
2 /2m. Introducing the definition of the time ordering op-

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 45 — #61

i i

Green’s functions 45

erator we can write:

i
G(~rt;~r0t 0 ) = − hφ0 |Ψ̂H (~r,t)Ψ̂†H (~r0 ,t 0 )|φ0 i θ (t − t 0 )
h̄
− hφ0 |Ψ̂†H (~r0 ,t 0 )Ψ̂H (~r,t)|φ0 iθ (t 0 − t)
i 0
=− ϕl (~r)ϕm∗ (~r0 ) e−i(El t−Em t )/h̄ hφ0 |âl â†m |φ0 i θ (t − t 0 )
h̄ ∑
lm
i 0
− (−) ∑ ϕl (~r)ϕm∗ (~r0 ) e−i(El t−Em t )/h̄ hφ0 |â†m âl |φ0 i θ (t 0 − t) .
h̄ lm
(4.4.4)
For t > t 0 we need to evaluate hφ0 |âl â†m | φ0 i = θ (m − kF ) δlm and for t < t 0 we
calculate hφ0 |â†m âl |φ0 i = θ (kF − l) δlm .
Then, the presence of the δ -functions allows us to perform part of the summa-
tions. Introducing also the step function in the summation indices we get
−i 0
G(~rt;~r0t 0 ) = ∑ ϕm (~r)ϕm∗ (~r0 ) e−iEm (t−t )/h̄ θ (t − t 0 )
h̄ m>k
F
(4.4.5)
i 0
+ ∑ ϕm (~r)ϕm∗ (~r0 )e−iEm (t−t )/h̄ θ (t 0 − t) .
h̄ m<k
F

Now introducing again the spin-indexes, transforming the summation over mo-
menta in an integral, and taking into account that the single-particle wavefunc-
tions are plane waves, we arrive at
1 −i
Z
0 0 ~ 0 0
Gσ σ 0 (~rt;~r t ) = δσ σ 0 d~k eik(~r−~r ) e−iEk (t−t )/h̄ θ (t − t 0 ) θ (k − kF )
(2π)3 h̄
Z
~ i~k(~r−~r0 ) −iEk (t−t 0 )/h̄ 0
− dk e e θ (t − t) θ (kF − k) ,
(4.4.6)
with Ek = h̄2 k2 /2m.

PROBLEM 4.5
Single-particle spectral functions for a non-interacting system

Consider the single-particle Green’s function of the free Fermi sea

Θ(k − kF ) Θ(kF − k)
g(0) (k, E) = + . (4.5.1)
h̄2 k2 2 2
E− 2m + iη E − h̄2mk − iη
(a) Calculate the single-particle spectral functions
(b) Calculate the occupation and the disoccupation of the single-particle states.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 46 — #62

i i

46 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

(a) To calculate the single-particle spectral function associated to g(0) (k, E) one
should identify the imaginary part of g(0) (k, E). To this end, one should take
into account that
1 A ∓ iη A η
= 2 2
= 2 2
∓i 2 (4.5.2)
A ± iη A +η A +η A + η2
and in the limit η → 0
1 1
lim = ∓ iπδ (A) , (4.5.3)
η→0 A ± iη A

where the delta function

appears following its
representation
with functional se-
ries δ (x) = limτ→∞ π1 τ 2 xτ2 +1 = limη→0 π1 x2 +η
η
2 .
The hole part of the spectral function is defined as:
1
Sh (k, E) = Im g(0) (k, E) , E < εF . (4.5.4)
π
Using the previous expression we have that

h̄2 k2

Sh (k, E) = δ E − Θ(kF − k), E < εF (4.5.5)
2m

and for the particle part, we have

h̄2 k2

1
S p (k, E) = − Im g(0) (k, E) = δ E − Θ(k − kF ) E > εF . (4.5.6)
π 2m

(b) In this case, the occupation coincides with the momentum distribution

h̄2 k2
Z εF Z εF
n(k) = Sh (k, E)dE = Θ(kF −k) δ E− dE = Θ(kF −k) (4.5.7)
−∞ −∞ 2m

The disoccupation will be given by :

Z ∞
d(k) = S p (k, E)dE = Θ(k − kF ) . (4.5.8)
εF

Notice that
n(k) + d(k) = 1 (4.5.9)
i.e., the occupation plus the disoccupation of each single-particle state is 1.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 47 — #63

i i

Green’s functions 47

PROBLEM 4.6
Kinetic energy and Koltun sum-rule for the free Fermi sea.

Using the hole part of the single-particle spectral function of the free Fermi
sea, !
h̄2 k2
Sh (k, E) = δ E− Θ(kF − k) , (4.6.1)
2m

(a) Calculate the expectation value of the kinetic energy per particle in the free
Fermi sea.

(b) Using the Koltun sum-rule, calculate the total energy (kinetic energy) of the
free Fermi sea.

(a) The expectation value of the kinetic energy per particle of a uniform system can
be expressed as
1 1 Ω
Z Z εF
hT̂ i = ν d~k h~k|tˆ|~ki Sh (k, E) dE , (4.6.2)
N N (2π)3 −∞

where Ω is the volume enclosing the system, N is the number of particles, ν is

the spin-isospin degeneracy and the one-body matrix element h~k|tˆ|~ki is given by

h̄2 k2
h~k|tˆ|~ki = , (4.6.3)
2m
and the hole part of the spectral function is

h̄2 k2

Sh (k, E) = δ E − Θ(kF − k) , (4.6.4)
2m
Then, Z εF
Sh (k, E) dE = Θ(kF − k) (4.6.5)
−∞
and therefore
1 1 ν h̄2 k2
Z
hT̂ i = d~k Θ(kF − k)
N ρ (2π)3 2m
h̄2 k4
Z kF Z 2π
1 ν
Z π
= dk dθ sin(θ ) dϕ (4.6.6)
ρ (2π)3 0 2m 0 0
1 ν 1 h̄2 kF5 3 h̄2 kF2
= 4π = .
ρ (2π)3 5 2m 5 2m

where ρ = N/Ω is the density of the system and kF3 = 6π 2 ρ/ν.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 48 — #64

i i

48 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

(b) The energy per particle in the ground state of a uniform system interacting
through a two-body potential can be expressed in terms of the hole part of the
spectral function by means of the Koltun sum-rule:
1 1 ν 1
Z Z εF
hĤi = d~k h~k|tˆ|~ki + E Sh (k, E)dE . (4.6.7)
N ρ (2π)3 2 −∞

Taking into account the previous calculations and the expression of Sh (k, E) for
the free Fermi sea, we have
Z εF 2 2
1 1 ν 1 h̄ k
Z
hĤiFS = d~k + E Sh (k, E)dE
N ρ (2π)3 2 −∞ 2m
(4.6.8)
1 ν 1 h̄2 k2 3 h̄2 kF2 1
Z
= ~
dk 2 = = 2 hT̂ i
ρ (2π)3 2 |~k|≤kF 2m 5 m N

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 49 — #65

i i

Section II

Applications

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 51 — #67

i i

The Fermi sea in one

5 dimension
Most of the visible matter around us is made of fermions, e.g. neutrons, protons and
electrons. Their properties are thus relevant to understand the behavior of matter in
different situations and at very different scales. Fermions, in contrast to bosons, are
described by antisymmetric wavefunctions, as explained in Chapter 1.3. This has
an important consequence already for non-interacting fermionic systems. Namely,
the way a set of non-interacting identical fermions builds its ground state is quite
peculiar as compared to bosons and gives rise to the so-called Fermi sea.
In general, fermionic gases appear in nature in three-dimensional configurations,
namely, building the electronic cloud of a nucleus, the electronic gas of a metal, or
even the interior of a neutron star. Understanding the properties of fermionic gases
provides a consistent explanation to phenomena in a wide range of scales appearing
in many fields of physics. A prominent example is the fact that a free Fermi sea ex-
erts a certain pressure. This stems directly from the Pauli exclusion principle, and
far from being an academic boutade, it has direct consequences in seemingly uncor-
related physical systems, such as the stability of white dwarves in an astrophysical
context or that of atomic nuclei in nuclear physics.
Up to very recently, we had at hand a variety of Fermi seas in nature that we could
study, e.g. different nuclei, metals, neutron stars, etc. This situation has changed re-
cently, with ultracold atomic gases experiments, where fermionic systems can be
prepared in an extremely controlled way [13], allowing also to extract thermody-
namic properties of custom-made fermionic gases [10].
Low dimensional Fermi gases are also of great interest today, again triggered
by the fast advances in experimental methods to produce one- and two-dimensional
degenerate quantum gases. The one dimensional case is particularly interesting, as
the quantum effects are in general enhanced [8].
In this chapter we present the main tools to compute the properties of one-
dimensional Fermi seas, exploring the equation of state, the pressure, etc. We also
discuss the correlations appearing between the fermions in the sea. These correla-
tions, which build from the Pauli principle, are fairly universal and thus important
for a variety of systems. Finally, we consider the case of trapped fermions, which
can in principle be produced in ultracold atomic gases.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 52 — #68

i i

52 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 5.1
Thermodynamic properties of the free Fermi sea in 1D

Consider a uniform system of non-interacting spin-1/2 fermions in one dimension

at zero temperature. You can consider the system as confined in a line of length L
with periodic boundary conditions.

(a) Find the relation between the density, ρ, and the Fermi momentum , kF .

(b) Calculate the total energy of the system as a function of ρ.

(c) Calculate the chemical potential, µ.

(d) Calculate the pressure exerted by these particles.

(e) Calculate the incompressibility, kT−1 .

(f) Calculate the speed of sound.

(a) The volume occupied in configuration space for each quantum state is given by
2π/L where L is the length of the box on which we will impose periodic boundary
conditions. Then,
L ν kF νL
Z
N= dk = 2kF , (5.1.1)
2π −kF 2π
where ν is the spin degeneracy, ν = 2 for spin 1/2. Then, we can write
N ν kF
ρ= = . (5.1.2)
L π
In one dimension, the density is linear with the Fermi momentum.
(b) To calculate the total energy, we must sum up the kinetic energy of all occupied
single-particle states:

h̄2 k2 L ν h̄2 kF3

Z kF
Lν
E= dk = 2 . (5.1.3)
2π −kF 2m 2π 2m 3
Now, if we identify N = L ν kF /π, we can write

1 h̄2 kF2
E =N , (5.1.4)
3 2m
and the energy per particle as

E 1 h̄2 kF2
e= = . (5.1.5)
N 3 2m

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 53 — #69

i i

The Fermi sea in one dimension 53

The total energy per particle is one-third of the Fermi energy. The total energy
per particle can also be expressed in terms of the density as:

1 h̄2 π 2 ρ 2
e(ρ) = . (5.1.6)
3 2m ν 2
The total energy per particle scales as the density squared.
(c) For the chemical potential we should calculate:

∂E ∂ (e(ρ)N) ∂ e(ρ)
µ= = = e(ρ) + ρ , (5.1.7)
∂N L ∂N L ∂ρ
that in our case is
1 h̄2 π 2 ρ 2 2 h̄2 π 2 ρ 2 h̄2 π 2 ρ 2
µ(ρ) = + = . (5.1.8)
3 2m ν 2 3 2m ν 2 2m ν 2
Now, if one recognizes kF2 = π 2 ρ 2 /ν 2 , then

h̄2 kF2
µ= . (5.1.9)
2m
The chemical potential is the energy of the last occupied single-particle state.
(d) The pressure is given by

∂E ∂ e(ρ) ∂ ρ ∂ e(ρ)
P=− = −N = ρ2 . (5.1.10)
∂L N ∂ρ ∂L ∂ρ
Then,
∂ e(ρ) 1 h̄2 2π 2 3
P = ρ2 = ρ . (5.1.11)
∂ρ 3 2m ν 2
(e) To calculate the incompressibility,

∂ P(ρ) ∂ e(ρ) ∂ 2 e(ρ)

kT−1 = ρ = 2ρ 2 + ρ3 , (5.1.12)
∂ρ ∂ρ ∂ ρ2
which in our case becomes,

h̄2 π 2
kT−1 = 2ρ 3 . (5.1.13)
2m ν 2

(f) Finally, the speed of sound:

kT−1 h̄2 π 2 ν 2 k2 h̄2 π 2 h̄2 kF2

c2s = = 2ρ 2 2 2 = 2 2F = = v2F . (5.1.14)
ρm 2m ν π 2m2 ν 2 m2
The speed of sound coincides with the speed associated to the Fermi momentum.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 54 — #70

i i

54 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 5.2
Two-body distribution function of a free Fermi sea in 1D

Consider a uniform system of non-interacting spin-1/2 fermions in one dimension

at zero temperature. One can take a box of length L with periodic boundary con-
ditions. The mean value per particle of a two-body operator V̂ = ∑i< j v(xi j ) in a
uniform system is given by

1
Z
φgs V̂ φgs = ρ dx12 v(x12 ) g(|x12 |) , (5.2.1)
2

where g(|x12 |) is the two-body distribution function that depends only on the dis-
tance between two particles, and x12 = x2 − x1 . The distribution function measures
the probability to find two particles at a given distance |x12 |, normalized such that
limx12 →∞ g(|x12 |) → 1 .
Calculate the two-body distribution function of a uniform system of non-
interacting fermions of spin 1/2.

In the case of the free Fermi sea, the wavefunction of the ground state is a Slater
determinant built with all single-particle states filled up to the Fermi level. In this
case, the expectation value of V̂ can be calculated as the sum of all two-body
matrix elements,
1
ΦFS V̂ ΦFS = ∑ hi j|v̂|i j − jii , (5.2.2)
2 ij

where the single-particle states are plane waves,

1
ϕ j (x) = eikx s j , (5.2.3)
L1/2
where kn = 2πn/L, x ∈ [−L/2, L/2], n = 0, ±1, ±2, . . . are the allowed momenta
when imposing periodic boundary conditions, and si represents the spin projec-
tion on the z-axis which can be up (↑) or down (↓). Now let’s perform the sum-
mation of the two-body matrix elements:

1 1 L2
Z kF Z kF
ΦFS V̂ ΦFS
= dk dk0
N 2 N (2π)2 −kF −kF
× ∑ hks1 , k0 s2 |v̂(x2 − x1 )|ks1 , k0 s2 − k0 s2 , ks1 i . (5.2.4)
s1 s2

We now introduce the single-particle wavefunctions in the two-body matrix ele-

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 55 — #71

i i

The Fermi sea in one dimension 55

ments and indicate the integration over the space coordinates:

ΦFS V̂ ΦFS 1 1 L2 kF kFZ Z

= 2
dk dk0
N 2 N (2π) −kF −kF
1
Z Z
0
× ∑ 2 dx1 dx2 e−ikx1 s1 (1)e−ik x2 s2 (2) v(x2 − x1 )
s1 s2 L
h 0 0
i
× eikx1 s1 (1)eik x2 s2 (2) − eik x1 s2 (1)eikx2 s1 (2)
Z kF Z kF
11 1
Z Z
= dk dk0 dx1 dx2 v(x2 − x1 )
2 N (2π)2 −kF −kF
" #
0
× ∑ 1 − ∑ δs1 s2 ei(k−k )(x2 −x1 )
s1 s2 s1 s2
Z kF Z kF
11 1
Z Z
= ν 2 dx1 dx2 v(x2 − x1 ) dk dk0
2 N (2π)2 −kF −kF

1 0
× 1 − eik(x2 −x1 ) e−ik (x2 −x1 ) , (5.2.5)
ν

where we have used that

∑ 1 = ν2 , ∑ δs1 s2 = ν (5.2.6)
s1 s2 s1 s2

and we have exchanged the order of integration of coordinates and momenta.

Now we evaluate the following integrals:
Z kF
ρ
dk = 2kF = (2π) , (5.2.7)
−kF ν

Z kF Z kF
dk eikx = dk (cos(kx) + i sin(kx)) (5.2.8)
−kF −kF
Z kF
sin kx kF 2 sin(kF x)
= dk cos(kx) = |−kF = ,
−kF x x

and introduce the definition of the Slater function in 1D:

Z kF
ν ν sin(kF x)
l1D (kF x) ≡ dk eikx = 2 . (5.2.9)
(2π)ρ −kF (2π)ρ x

Then we can write:

Z kF Z kF
ρ
dk e ikx
= dk e−ikx = (2π) l1D (kF x) . (5.2.10)
−kF −kF ν

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 56 — #72

i i

56 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Therefore,
2 ZZ

ΦFS V̂ ΦFS 11 1 2 2ρ
= ν (2π) dx1 dx2 v(x12 )
N 2 N (2π)2 ν2
l 2 (kF x12 )

× 1 − 1D
ν
l 2 (kF x12 )

11 2
ZZ
= ρ dx1 dx2 v(x12 ) 1 − 1D . (5.2.11)
2N ν

Since the integrand depends solely on the relative coordinate x12 , it is convenient
to perform the following coordinate transformation:
x2 + x1
x12 = x2 − x1 , xCM = ,
2
where xCM is the position of the center of mass of the system. One can easily
check that the Jacobian of this transformation is +1, and we thus have:

ΦFS V̂ ΦFS 11 2
ZZ 2 (k x )
l1D F 12
= ρ dxCM dx12 v(x12 ) 1 −
N 2N ν
11 2
Z 2 (k x )
l1D F 12
= ρ L dx12 v(x12 ) 1 −
2N ν
l 2 (kF x12 )

1
Z
= ρ dx12 v(x12 ) 1 − 1D ,
2 ν
R L/2
where we have used that −L/2 dxCM = L . We can now identify the two-body
distribution function as:
2 (k x )
l1D ν sin kF x
F 12
g(x12 ) = 1 − , l1D (kF x) = 2 . (5.2.12)
ν (2π)ρ x

Therefore, we can write:

ΦFS V̂ ΦFS 1
Z ∞
= ρ dx12 v(x12 ) g(|x12 |) . (5.2.13)
N 2 −∞

The distribution function depends only on the distance |x12 |. Notice also that
l1D (kF x) = l1D (−kF x), which is why we can write g(x12 ) = g(|x12 |). The proba-
bility of having two particles at the same place is given by:
1
g(0) = 1 − , (5.2.14)
ν
2 (0) = 1. The distribution functions go to 1 − O(1/x2 ) when x → ∞.
because l1D 12 12

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 57 — #73

i i

The Fermi sea in one dimension 57

PROBLEM 5.3
Properties of the Slater function in 1D. Sequential condition of the two-body
distribution function.

The two-body distribution function for the free Fermi sea in 1D with spin degen-
eracy ν = 2 reads:
l 2 (kF |x12 |)
g(x12 ) = 1 − 1D , (5.3.1)
ν
where Z kF
ν ν sin(kF x)
l1D (kF x) = dk eikx = 2 (5.3.2)
(2π)ρ −kF (2π)ρ x
is the Slater function in 1D and kF is the Fermi momentum.

(a) Check that Z ∞

ν
dx l1D (kF x) = . (5.3.3)
−∞ ρ

(b) Check that Z ∞

2 ν
dx l1D (kF x) = . (5.3.4)
−∞ ρ

(c) Check that g(x12 ) fulfills the sequential condition .

(a) We start with the normalization of the Slater function,

Z kF Z kF
1
Z ∞ Z ∞ Z ∞
ν ν
dx l1D (kF x) = dx eikx = dk dx eikx . (5.3.5)
−∞ −∞ (2π)ρ −kF ρ −kF 2π −∞

One can identify the delta function δ (x) in the last expression,
Z ∞ Z kF
ν ν
dx l1D (kF x) = dk δ (k) = . (5.3.6)
−∞ ρ −kF ρ

(b) We start by substituting the definition of the Slater function,

Z ∞ Z ∞ Z k Z k
ν ν F F 0
2
dx l1D (kF x) = dx dk eikx dk0 eik x
−∞ −∞ (2π)ρ (2π)ρ −kF −kF
ν2
Z −kF Z kF
1
Z ∞
0
= dk dk0 2
dx ei(k+k )x . (5.3.7)
kF −kF ρ (2π) 2π −∞

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 58 — #74

i i

58 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Now, one can identify the delta function δ (k + k0 ) in the last expression. Then,
ν2 ν2
Z ∞ Z kF Z kF Z kF
2
dx l1D (kF x) = 2
dk dk0 δ (k + k0 ) = 2
dk
−∞ ρ (2π) −kF −kF ρ (2π) −kF
ν2 ν
= 2kF = , (5.3.8)
ρ 2 (2π) ρ
where we have used Eq. (5.1.2) in the last equality. It is curious to notice that
both integrals give the same result.
(c) Finally, let us calculate the sequential condition,
l 2 (kF x12 )
Z ∞ Z ∞
ρ dx12 (g(x12 ) − 1) = ρ dx12 1 − 1D −1
−∞ −∞ ν
Z ∞ 2 (k x )
l1D ν
F 12
= −ρ dx12 = −ρ = −1 . (5.3.9)
−∞ ν ρν

PROBLEM 5.4
Trapped fermions

N non-interacting fermions are trapped in a one-dimensional harmonic oscillator

potential. The fermions are spin-polarized, with all spins oriented in a direction
perpendicular to the confinement.

(a) Calculate the total, kinetic and harmonic potential energies.

(b) Calculate the chemical potential.

(c) Derive an energy density functional in terms of the local density. Assume that,
locally, the system can be represented by a one-dimensional Fermi sea.

(d) Find the density profile by minimizing the energy functional with the condi-
tion that the number of particles is constant.

(e) Find the chemical potential associated with the Lagrange multiplier used to
impose a fixed number of particles.

(f) Calculate the total energy, harmonic potential energy and kinetic energy from
the density profile that minimizes the density functional and compare with the
exact values.

(a) The single-particle spectrum of a one-dimensional harmonic oscillator is

1
εi = + i h̄ω . (5.4.1)
2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 59 — #75

i i

The Fermi sea in one dimension 59

Then, the total energy of the ground state of the N fermions is obtained by sum-
ming the single-particle energies of the lowest filled single-particle levels:

E 1 3 5 1
= + + +···+ +N −1
h̄ω 2 2 2 2
1 1
= (1 + 3 + 5 + · · · + 2N − 1) = N 2 . (5.4.2)
2 2
Therefore, using the virial theorem we can say that
1 1
E = N 2 h̄ω , T = VHO = N 2 h̄ω . (5.4.3)
2 4

(b) The chemical potential at zero temperature is given as the difference of the
ground-state energies of the system of N particles and the system of N − 1 parti-
cles, which therefore coincides with the energy of the last occupied level, which
in our case is µ = εN = (N − 1/2)h̄ω. When the number of particles is large,
the energy and the number of particles can be considered as continuous variables
and the chemical potential can be approximately calculated as the derivative of
the total energy with respect to the number of particles:

dE
µ= = N h̄ω , (5.4.4)
dN
which is very similar to the exact value when N is large.

(c) To build the density functional, we should first recall which is the energy per unit
length in a free Fermi gas in 1D. The relation between the linear density and the
Fermi momentum is
Z kF
Lν νL ν kF
N= dk = 2kF →ρ = , (5.4.5)
2π −kF 2π π

the energy is
h̄2 k2 1 h̄2 kF2
Z kF
Lν
E= dk = N , (5.4.6)
2π −kF 2m 3 2m
and the energy density for ν = 1 is:

E h̄2 π 2 3
= ρ . (5.4.7)
L 2m 3
Notice that we have assumed that, locally, the system can be seen as a Fermi
sea (otherwise we could not use the above relation between ρ and kF ). Then, the
energy functional can be written as:
2 2
h̄ π 3 1
Z
E = dx ρ + mω 2 x2 ρ . (5.4.8)
2m 3 2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 60 — #76

i i

60 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

In the next step we want to write the functional in units of the harmonic oscillator,
i.e. the energy in units of h̄ω and the length in units of the harmonic oscillator
length: b = (h̄/(mω))1/2 . To this end, we transform the density as ρ̄(x1 ) = bρ(x)
where x = bx1 . Notice that this transformation keeps the density normalized to
the number of particles:
1
Z Z Z
dx1 ρ̄(x1 ) = dx1 bρ(bx1 ) =
dy b ρ(y) = N , (5.4.9)
b
where we have used the change y = x1 b, and taken into account that the original
density was normalized to the number of particles. Using the harmonic oscillator
units, the functional can be written as
2
π 1 1 2
Z
3
EHO = dx1 ρ̄(x1 ) + x1 ρ̄(x1 ) . (5.4.10)
3 2 2
(d) Now, we should minimize the functional with respect to the density profile but
keep the normalization of the density. To this end, we introduce a Lagrange
multiplier, and require the total functional to be stationary with respect to small
changes of ρ̄: R
δ (EHO − µ dx1 ρ̄(x1 ))
= 0, (5.4.11)
δ ρ̄
to get the following condition:
1 2 1
π ρ̄(x1 )2 + x12 − µ = 0 . (5.4.12)
2 2
The density should be always positive or null, so:
( q
1 √
2µ − x12 | x1 |≤ 2µ
ρ̄(x1 ) = π √ .
0 | x1 |> 2µ

(e) In the next step, we can determine the chemical potential by imposing that the
integration of the density should provide the total number of particles:
Z √2µ Z √2µ
1
q
√ dx1 ρ̄(x1 ) = √ dx1 2µ − x12 = N . (5.4.13)
− 2µ − 2µ π
To perform this integral we should recall that
r
1 p 1 a c
Z p
a + cx2 dx = x a + cx2 + √ arcsin x − , (5.4.14)
2 2 −c a
and therefore
Z √2µ q √2µ
11 2 x
√ dx1 ρ̄(x1 ) = x1 2µ − x1 + 2µ arcsin √
− 2µ π2 2µ −√2µ
1 h π π i
= µ − − = µ =N.
π 2 2
Therefore, µ = N in harmonic oscillator units.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 61 — #77

i i

The Fermi sea in one dimension 61

(f) Let us now proceed with the calculation of the harmonic potential energy:
Z √2µ Z √2µ
1 2 1 2 1
q
V̂HO = √ dx1 x1 ρ̄(x1 ) = dx 1 x 2µ − x12 . (5.4.15)
− 2µ 2 2 −√2µ 1
π
Now, we should take into account the following integral:
1 xa3 1 axu 1 a2
Z p
dx x2 a + cx2 = − − I1 , (5.4.16)
4 c 8 c 8 c
where
r
p 1 c
u = a + cx2 , I1 = √ arcsin x − , c < 0. (5.4.17)
−c a
In our case, a = 2µ and c = −1, and we have:
√
Z √2µ 2µ
1
q
√ dx1 x12 2µ − x12 = − x1 (2µ)3 √
− 2µ 4 − 2µ
√
2µ
1
q
+ 2µx1 2µ − x12 √
8 − 2µ
√
2µ
µ 2 2π

1 2 x1
+ (2µ) arcsin √ √
= ,
8 2µ − 2µ 4
where the first two contributions to the integral are zero. Finally,
1 1 µ 2 2π µ2
V̂HO = = , (5.4.18)
2π 4 4
and taking into account that µ = N, we have
N2
V̂HO = . (5.4.19)
4
Now the kinetic energy is:
Z √2µ Z √2µ
π2 1 11 3/2
T̂HO = √ dx1 ρ̄(x1 )3 = √ dx1 2µ − x12 . (5.4.20)
− 2µ 3 2 π6 − 2µ

One should take into account the following integral:

1 3 3
Z
u3 dx = xu3 + axu + a2 I1 , (5.4.21)
4 8 8
√
where u = a + cx2 , and therefore
√ √
Z √2µ 2µ 2µ
1 3
q
2 3/2 3/2
= x1 2µ − x12

√ dx1 2µ − x1 √
+ 2µx 1 2µ − x12 √
− 2µ 4 − 2µ 8 − 2µ
√2µ
3 2 x 3
+ (2µ) arcsin √ √
= µ 2π ,
8 2µ − 2µ 2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 62 — #78

i i

62 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

where the first two terms are identically zero. The kinetic energy is, finally:

1 3 2 µ2 N2
T̂HO = µ π= = . (5.4.22)
6π 2 4 4

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 63 — #79

i i

The Fermi sea in three

6 dimensions at T = 0
The concept of the Fermi sea is central to understanding the electronic properties
of many-body quantum systems. In one-dimensional systems, the Fermi sea is char-
acterized by a simpler topology, with electrons occupying states along a linear mo-
mentum axis. In contrast, in three dimensions, the free Fermi sea takes on a more
complex geometry, forming a spherical surface in momentum space. This difference
profoundly influences the density of states, interactions between particles, and the
emergent phenomena in these systems. Three-dimensional fermionic systems are
governed by a rich spectrum of quantum states and exhibit different collective be-
haviors.
At zero temperature (T = 0), the Fermi sea is in its simplest and most well-
defined form, with all states below the Fermi energy completely filled and those
above completely empty. This limit case is particularly valuable for theoretical and
computational studies, as it allows for analytic calculations to be performed with
relative ease compared to finite-temperature scenarios. Despite its simplicity, the T=0
approximation remains useful, providing insights into the ground-state properties of
quantum systems and serving as a foundation for understanding more complex finite-
temperature properties.
This chapter includes four problems to explore in detail properties of the three-
dimensional Fermi sea at zero temperature. The problems cover fundamental topics
such as energy, thermodynamic properties and relativistic modifications of the ther-
modynamic properties. These problems provide a framework for exploring the theo-
retical and practical aspects of the system, offering essential tools for understanding
both the classical and relativistic behavior of the Fermi sea.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 64 — #80

i i

64 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 6.1
Energy and first derivatives.

Consider a uniform system of non-interacting non-relativistic fermions with num-

ber density ρ usually known as a free Fermi sea.

(a) Calculate the relation between the density and the Fermi momentum.

(b) Calculate the energy per particle as a function of the density.

(c) Calculate the pressure and the chemical potential.

(d) Calculate the incompressibility and the speed of sound.

(e) Make numerical applications for the case of nuclear matter at ρ = 0.16 fm−3
taking h̄2 /2m = 20.7344 MeV fm2 .

(f) Do the same for liquid Helium-3 with ρ = 0.0164 Å−3 and h̄2 /2m = 8.03 K
Å2 .

(a) We assume that the fermions are enclosed in a cubic box of volume Ω and that the
single-particle states are plane waves normalized to this volume, characterized by
the momenta allowed by the periodic boundary conditions. The ground state is
then given by a Slater determinant where the lowest single-particle states are
filled up. Thus, we now start by calculating the Fermi momentum associated to
the density. The total number of particles should be related to the total number of
filled states,
4
Z
Ω Ω
N= 3
ν d~k = 3
ν πkF3 , (6.1.1)
(2π) ~
|k|≤kF (2π) 3
where ν is the spin/isospin degeneracy, i.e., ν = 4 for nuclear matter, ν = 2 for
neutron matter, ν = 2 for liquid 3 He, and ν = 2 for electrons. Therefore
2 1/3
νkF3 6π ρ
ρ = 2 ⇒ kF = (6.1.2)
6π ν
which establishes a relation between a macroscopic quantity, ρ, and a micro-
scopic one, the Fermi momentum kF .
(b) As the kinetic energy is a one-body operator, the expectation value in a Slater
determinant is the sum of the expectation values of the kinetic energy operator in
the occupied single-particle states,
E p̂2
e= = ∑ h~k ms | |~k ms i (6.1.3)
N ~ 2m
k,ms

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 65 — #81

i i

The Fermi sea in three dimensions at T = 0 65

where
p̂2 p̂2
h~k ms | |~k ms i = h~k| |~ki hms | ms i , (6.1.4)
2m 2m
with 2 2
p̂2 1 h̄ ∇ h̄2 k2
Z
−i~k~r ~
h~k| |~ki = d~r e − eik~r = . (6.1.5)
2m Ω ~r ∈ Ω 2m 2m
Then,
h̄2 k2 1 Ω h̄2 kF5
Z
Ω ν
e= d~k = ν 4π , (6.1.6)
(2π)3 N |~k|≤kF 2m N (2π)3 2m 5
and thus writing N in terms of Ω and kF results in,
2/3
3 h̄2 kF2 3 h̄2 6π 2 ρ

e(ρ) = = . (6.1.7)
5 2m 5 2m ν

(c) The first derivative of the total energy with respect to the volume, keeping the
number of particles constant, provides the pressure. It is convenient to perform
the derivative with respect to the density and to express the total energy as E =
Ne(ρ),

ρ 2 d e(ρ)

dE d e(ρ) dρ
P(ρ) = − =− = . (6.1.8)
dΩ N dρ dΩ N N dρ
Therefore,
2/3
h̄2 2 6π 2 ρ 5/3

P(ρ) = . (6.1.9)
2m 5 ν N
On the other hand, the chemical potential

dE P(ρ)
µ(ρ) = = e(ρ) + . (6.1.10)
dN ρ
Therefore,
2/3 2/3
h̄2 3 6π 2 h̄2 2 6π 2 h̄2 kF2

µ(ρ) = ρ 2/3 + ρ 2/3 = (6.1.11)
2m 5 ν 2m 5 ν 2m
i.e, the energy of the last occupied single-particle state.
(d) The compressibility modulus is the change in volume per unit volume when we
vary the pressure,
1 dΩ
κ =− . (6.1.12)
Ω dP N
Usually we calculate
2/3 5/3
h̄2 2 6π 2

−1 dP dP ρ
κ =− Ω= ρ= . (6.1.13)
dΩ N dρ 2m 3 ν N

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 66 — #82

i i

66 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Then, the speed of sound at T = 0 is related to the compressibility modulus by:

2/3
κ −1 h̄2 2 6π 2 h̄2 kF2 v2F

c2s = = ρ 2/3 = = . (6.1.14)
ρm 2m2 3 ν 3m2 3

Finally,
h̄ kF v
cs = √ = √F . (6.1.15)
m 3 3

(e) Let us now consider nuclear matter at saturation density ρ = 0.16 fm−3 . Then
kF = 1.33 fm−1 . Taking h̄2 /2m = 20.7344 MeV fm2 , we have hΦFS |T |ΦFS i =
22.01 MeV and the speed of sound can be expressed in terms of the speed of
light c
cs h̄ c kF
=√ = 0.16 (6.1.16)
c 3mc2
which is not relativistic.
(f) For the liquid 3 He, we have ρ = 0.0164 Å−3 and h̄2 /2m = 8.03 K Å2 , which
gives kF = 0.786 Å−1 , the average kinetic energy, hT i = 2.976 K and the speed
of sound cs /c = 3.2 10−7 .

PROBLEM 6.2
Symmetry energy of the free Fermi sea

Consider a uniform system of protons and neutrons, with total number density
density ρ = (Z + N)/Ω where Z is the number of protons, N is the number of
neutrons and Ω is the volume that encloses the system; and asymmetry ∆ = (ρn −
ρ p )/ρ where ρn and ρ p are the neutron and proton number densities, respectively.
The system is modeled by an asymmetric Fermi sea. Take h̄2 /m = 41.49 MeV
fm2 .

(a) Calculate the total energy as a function of the number density and the asym-
metry parameter ∆.

(b) Write the Taylor expansion of the energy for a given density in terms of the
asymmetry parameter.

(c) Identify the contribution to the asymmetry energy. Give the result at the em-
pirical saturation density of nuclear matter, ρ0 = 0.16 fm−3 .

(d) Calculate the density derivative of the asymmetry energy at ρ0 .

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 67 — #83

i i

The Fermi sea in three dimensions at T = 0 67

(a) We proceed to work out the different questions. For the first one we start by cal-
culating the Fermi momentum associated to the total density and the Fermi mo-
menta associated to the proton and neutron components. To describe the system
one can use the total density ρ = ρN + ρZ and the asymmetry: ∆ = (ρN − ρZ )/ρ.
At the same time, ρZ and ρn can be expressed in terms of the total density ρ and
the asymmetry ∆:
ρ ρ
ρZ = (1 + ∆), ρN = (1 − ∆) . (6.2.1)
2 2
Then we have a global Fermi momentum associated to the total density:
2 1/3 2 1/3
6π ρ 6π ρ
kF = = , (6.2.2)
ν 4
where we have considered the spin-isospin degeneracy, ν = 4. In a similar man-
ner, we define the Fermi momenta associated to protons and neutrons:
2 1/3 2 1/3
6π ρZ π ρ
kFZ = = 6 (1 − ∆) = kF (1 − ∆)1/3 , (6.2.3)
νs 2 2
where νs = 2 is the spin degeneracy, and
2 1/3 2 1/3
6π ρN π ρ
kFN = = 6 (1 + ∆) = kF (1 + ∆)1/3 . (6.2.4)
νs 2 2
The total kinetic energy is the sum of the kinetic energies of protons and neutrons:
h̄2 k2 h̄2 k2
Z Z
Ω Ω
E= νs d~k + ν
3 s
d~k , (6.2.5)
(2π)3 ~
|k|≤kF 2m (2π) |~k|≤kF 2m
where we have considered that the masses of protons and neutrons are the same,
mz = mN = m. Performing the integrals:
" #
Ω h̄2 (kFZ )5 h̄2 (kFN )5
E = 4π 2 +
(2π)3 10m 10m
Ω 1 5
h
5/3 5/3
i
= 4π 2k (1 + ∆) + (1 − ∆) . (6.2.6)
(2π)3 10m F
As we are interested in the kinetic energy per particle,
1
2kF5 (1 + ∆)5/3 + (1 − ∆)5/3
Ω

E (2π)3
4π 10m
e(kF , ∆) = = Ω 4
N
(2π)3 3
πkF3 4
3 h̄2 kF2 h i (6.2.7)
= (1 + ∆)5/3 + (1 − ∆)5/3
2 10m
3 h̄2 kF2 1 h i
= (1 + ∆)5/3 + (1 − ∆)5/3 ,
5 2m 2
where N = ρΩ. Notice that if ∆ = 0 we recover the kinetic energy per particle of
the symmetric free Fermi sea.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 68 — #84

i i

68 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

(b) Now we perform the Taylor expansion of (1 + ∆)5/3 + (1 − ∆)5/3 around ∆ = 0:

h
5/3 5/3
i 5 10 2 5 10 2
(1 + ∆) + (1 − ∆) ∼ 1+ ∆+ ∆ +···+1− + ∆ +···
3 18 3 18
10 2
∼ 2+ ∆ +··· (6.2.8)
9
Notice that by symmetry, all terms with odd powers, in particular the linear term,
in ∆ disappear. Therefore, up to second order,

3 h̄2 kF2

5
e(kF , ∆) = 1 + ∆2 + · · · . (6.2.9)
5 2m 9

(c) By definition, the symmetry energy is associated with the coefficient of the term
∆2 or equivalently, by looking at the Taylor expansion, with the one-half of the
second derivative with respect to ∆ of e(ρ, ∆):

3 h̄2 kF2 3 h̄2 kF2 5 2

e(ρ, ∆) ∼ + ∆ +··· , (6.2.10)
5 2m 5 2m 9
and therefore
1 h̄2 kF2
aI = . (6.2.11)
3 2m
For the case of nuclear matter, with kF = 1.36 fm−1 which corresponds to the
nuclear saturation density, and taking h̄2 /m = 41.46 MeV fm2 , we have that aI =
12.78 MeV. Notice that a commonly accepted value for nuclear matter is 30 MeV,
which means that correlations will have an important role in the evaluation of this
quantity and that the free Fermi gas model is rather limited in this case. If only
the second power of ∆ would be present in the expansion, it would be possible to
calculate the symmetry energy as aI ∼ e(ρ, ∆ = 1) − e(ρ, ∆ = 0). It is therefore
interesting to evaluate the difference between the two procedures:

3 h̄2 kF 2/3
ãI = e(ρ, ∆ = 1) − e(ρ, ∆ = 0) = (2 − 1) (6.2.12)
5 2m
and the relative error :
5
aI − ãI 9 − (22/3 − 1)
= 5
= −0.06 . (6.2.13)
aI 9

(d) Also of interest is to provide the behavior of the density derivative of aI . To this
end, it is useful to express the symmetry energy in terms of the density instead
than the Fermi momentum:
2/3 2/3
3 h̄2 6π 2 ρ 3 h̄2 6π 2 ρ

5 2
e(ρ, ∆) = + ∆ +··· (6.2.14)
5 2m ν 5 2m ν 9

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 69 — #85

i i

The Fermi sea in three dimensions at T = 0 69

Therefore,
2/3
1 h̄2 6π 2 ρ

5
aI = (6.2.15)
3 2m ν 9
and
2/3
daI (ρ) 1 h̄2 6π 2

2 −1/3 5 2 aI
= ρ = . (6.2.16)
ρ 3 2m ν 3 9 3 ρ 1/3
The value at ρ = 0.16 fm−3 ,

daI
= 15.70 MeV fm3 (6.2.17)
dρ

PROBLEM 6.3
The energy of the free Fermi sea as a lower bound of the kinetic energy

Consider a non-relativistic uniform system of fermions characterized by a density

ρ. Show that the energy of the free Fermi sea associated to that system is a lower
bound to the expectation value of the kinetic energy of the system.

The kinetic energy operator

h̄2 ∇2i
T̂ = − ∑ (6.3.1)
i 2m

is the Hamiltonian, Ĥ0 ≡ T̂ , of the free Fermi sea and its ground state, φ0 is the
Slater determinant built with plane waves filling up all momenta up to the Fermi
level. The energy per particle of the ground state is the well known:
2/3
3 h̄2 kF2 3 h̄2 6π 2 ρ

Ĥ0 |φ0 i = E0 |φo i , E0 = = (6.3.2)
5 2m 5 2m ν

If we consider now the ground state |Ψ0 i of another Hamiltonian, Ĥ = T̂ + V̂ , and

calculate the expectation value of Ĥ0 ≡ T̂ in this wavefunction, the variational
principle tells us that

hΨ0 |T̂ |Ψ0 i = hΨ0 |Ĥ0 |Ψ0 i = Evar ≥ E0 (6.3.3)

and therefore, E0 is a lower bound of the expectation value of the kinetic energy
in the ground state of Ĥ.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 70 — #86

i i

70 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 6.4
Thermodynamic properties of the relativistic three-dimensional free Fermi
sea.

Consider a uniform system of non-interacting spin-1/2 fermions in three dimen-

sions at zero temperature. Use the relativistic dispersion relation between momen-
tum and energy:
p p
ε(k) = m2 c4 + h̄2 k2 c2 = mc2 1 + x2 , (6.4.1)

where
h̄ck
x= . (6.4.2)
mc2
This parameter measures the relativistic effects. For x 1 , relativistic effects are
small.

(a) Calculate the energy density.

(b) Calculate the chemical potential.

(c) Calculate the pressure.

(a) To calculate the energy density, one should sum up all the single-particle energies
of the occupied states. To this end,
E 1
Z
εΩ = =ν d~k Θ(kF − k) ε(k) (6.4.3)
Ω (2π)3

Now, using the dispersion relation we can write:

1
Z q
εΩ =ν mc2
d~k Θ(kF − k) 1 + x(k)2
(2π)3
1
Z q
2 2
=ν mc 4π dk k 1 + x(k)2
(2π)3 k≤kF
1 m2 c4 mc2 2
Z xF p
=ν mc 4π x2 1 + x2 dx
(2π)3 h̄2 c2 h̄c 0
4
mc2
Z xF p
1
= ν x2 1 + x2 dx = ε0 ε̄(xF ) , (6.4.4)
(h̄c)3 2π 2 0

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 71 — #87

i i

The Fermi sea in three dimensions at T = 0 71

where ε0 = (mc2 )4 /(h̄c)3 , and

Z xF p
ν
ε̄(xF ) = x2 1 + x2 dx
2π 2 0
q q
ν 2 2 2
= xF (1 + 2xF ) 1 + xF − ln xF + xF + 1
16π 2
q
ν 2 2 −1
= xF (1 + 2xF ) 1 + xF − sinh (xF ) . (6.4.5)
16π 2

The non-relativistic limit is recovered for xF 1 which should coincide with

E 3 h̄2 kF2 ν 3 h̄2 ν 5

= ρmc2 + ρ = 2 mc2 kF3 + k , (6.4.6)
Ω 5 2m 6π 5 2m 6π 2 F
In fact, doing the series expansion of ε̄(xF ) around xF = 0 one gets:

ν 8 3 4 5
ε̄(xF ) ∼ x + x + · · · , (6.4.7)
16π 2 3 F 5 F

which can be expressed as

(mc2 )4 ν 8 (h̄c)3 3 4 (h̄c)5 5

εΩ =ε0 ε̄(xF ) ∼ k + k +···
(h̄c)3 16π 2 3 (mc2 )3 F 5 (mc2 )5 F
ν 3 h̄2 kF2 3 ν 3 3 h̄2 kF2
=mc2 2
kF + 2
kF = mc2 ρ + ρ. (6.4.8)
6π 2m 5 6π 5 2m

(b) The chemical potential is given by

∂E ∂ (εΩ Ω) ∂ εΩ (ρ) ∂ ρ ∂ εΩ (ρ)
µ= = =Ω = , (6.4.9)
∂N Ω ∂N Ω ∂ρ ∂N ∂ρ
which in terms of xF is expressed as

∂ ε̄(xF ) ∂ xF
µ = ε0 (6.4.10)
∂ xF ∂ ρ

where
1/3
ν kF3 6π 2 ρ h̄c 1 6π 2 1

∂ xF
ρ= , kF = , = . (6.4.11)
6π 2 ν ∂ρ mc2 3 ν kF2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 72 — #88

i i

72 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Then,

∂ ε̄(xF )
q q
ν ∂ 3 2 − ln x + 2
= (xF + 2xF ) 1 + xF F 1 + xF
∂ xF 16π 2 ∂ xF

2xF
q
ν 
= 2
(1 + 6x F ) 1 + xF2 + (xF + 2xF3 ) q (6.4.12)
16π 2 1 + xF2
 
1 1 + q2xF 
− q
xF + 1 + xF 2 2 1 + xF2
q
ν
= 2 xF2 1 + xF2 . (6.4.13)
2π
Then,

∂ ε̄(xF ) ∂ xF (mc2 )4 ν 2 h̄c 1 6π 2 1

q
µ = ε0 = x 1 + xF2
∂ xF ∂ ρ (h̄c)3 2π 2 F mc2 3 ν kF2
q
= mc2 1 + xF2 = ε(xF ) , (6.4.14)

which is nothing other than the energy of the highest occupied level.
The non-relativistic limit is obtained when xF 1, then

xF2 h̄2 kF2

q
2 2
2
µ(xF 1) = lim mc 1 + xF ∼ mc 1 + = mc2 + (6.4.15)
xF →0 2 2m

(c) The pressure is given by

∂E ∂ (ΩεΩ (ρ))
P=− =− =
∂Ω N ∂Ω N

∂ εΩ ∂ρ ∂ εΩ
− εΩ (ρ) − Ω = −εΩ + ρ (6.4.16)
∂ρ ∂Ω N ∂ρ

which can be expressed as

P = −εΩ + µρ . (6.4.17)
Now, taking into account that, εΩ (ρ) = ε0 ε̄(xF ), we can write

∂ εΩ ∂ ε̄(xF ) ∂ xF
ρµ = ρ = ρ ε0
∂ρ ∂ xF ∂ ρ
ν 4 3 ∂ ε̄(xF ) h̄c 1 6π 2 1
= πk F ε 0
(2π)3 3 ∂ xF mc2 3 ν kF2
h̄c 1 ∂ ε̄(xF ) 1 ∂ ε̄(xF )
= kF 2 ε0 = xF ε0 . (6.4.18)
mc 3 ∂ xF 3 ∂ xF

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 73 — #89

i i

The Fermi sea in three dimensions at T = 0 73

Then,
P = P0 P̄(xF ) (6.4.19)
with
1
P̄(xF ) = xF ε̄ 0 (xF ) − ε̄(xF ) (6.4.20)
3
and P0 = ε0 , and finally

∂ ε̄(xF )
q
ν
ε̄ 0 (xF ) = = 2 xF2 1 + xF2 . (6.4.21)
∂ xF 2π

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 75 — #91

i i

Free Fermi sea at finite

7 temperature
At finite temperatures, the sharp distinction between occupied and unoccupied states
near the Fermi surface becomes blurred, as the filling of the states is ruled by the
Fermi distribution . This leads to significant changes in the physical properties of the
system. Analyzing these effects is crucial for understanding phenomena like electri-
cal and thermal conductivity, as well as the response of materials to external pertur-
bations. Moreover, working at nonzero temperatures allows us to study transitions
between quantum and classical regimes, shedding light on the statistical mechanics
that govern systems from condensed matter to nuclear physics and astrophysics.
This chapter includes problems that explore the effects of temperature on the
Fermi sea, bridging the gap between the zero-temperature idealization and realistic
finite-temperature scenarios. Important tools are included to compute thermal aver-
ages, some thermodynamic properties, and derive temperature-dependent corrections
in the high and low temperature limits, in systems where thermal excitations play a
crucial role in shaping their behavior.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 76 — #92

i i

76 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 7.1
Specific heat at the low temperature limit of the free Fermi sea.

At low temperature, i.e. in the degenerate limit, the specific heat per particle at
constant volume Ω of the free Fermi sea can be expressed as

π 2T
CΩ (ρ, T ) = , (7.1.1)
2εF

where εF = h̄2 kF2 /2m. Use this expression to calculate the low temperature behav-
ior of:

(a) The energy per particle of the free Fermi sea.

(b) The entropy per particle.

(c) The free energy per particle.

(d) The pressure.

(e) The chemical potential.

(a) The increment of energy between two temperatures at a given density can be
expressed as
Z T
e(ρ, T ) = e(ρ, T = 0) + CΩ (ρ, T 0 )dT 0 , (7.1.2)
0
which in our case is:
Z T 2 2
3 π 0 3 0 1 T
e(ρ, T ) = εF + T dT = εF + π 2 ε f , (7.1.3)
5 0 2εF 5 4 εF
where εF = h̄2 kF2 /2m is the Fermi energy of the free Fermi sea at T = 0. There-
fore, the energy increases quadratically with temperature as
2 !
3 5 T
e(ρ, T ) = εF 1 + π 2 . (7.1.4)
5 12 εF

(b) By definition,
CΩ (ρ, T 0 )
Z T Z T 2 0
π T 1
s= dT 0 = dT 0 . (7.1.5)
0 T0 0 2εF T 0
Performing the integral, we get:
π2 π2
Z T
s= dT 0 = T, (7.1.6)
2ε f 0 2εF

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 77 — #93

i i

Free Fermi sea at finite temperature 77

which is linear in T .
(c) The free energy per particle is defined as f (ρ, T ) = e(ρ, T ) − T s(ρ, T ), which in
our case is

π2 T 2 π2 2 3 1 π2 2

3
f (ρ, T ) = εF + εF − T = εF − T , (7.1.7)
5 4 εF 2εF 5 εF 4

which can be written as

2 !
5π 2

3 T
f (ρ, T ) = εF 1− . (7.1.8)
5 12 εF

At very low temperatures, the free energy decreases in respect to T = 0 due to

the fact that the increase of entropy dominates over the increase of energy.
(d) The pressure is related to the free energy by

∂ f (ρ, T )
P(ρ, T ) = ρ 2 . (7.1.9)
∂ρ
To perform this derivative it is useful to have εF expressed in terms of the density,
which by Eq. (6.1.2) can be put as

h̄2
εF = (6π 2 ν −1 )2/3 ρ 2/3 , (7.1.10)
2m
where ν is the spin-isospin degeneracy. Then,
2 !!
5π 2

∂ 3 T
P(ρ, T ) =ρ 2 εF 1−
∂ρ 5 12 εF
2 !
h̄2 5π 2

23 2 T
=ρ (6π 2 ν −1 )2/3 ρ −1/3 1 −
5 2m 3 12 εF
3 5π 2 2 2 h̄2 2
+ ρ 2 εF T 3 (6π 2 ν −1 )2/3 ρ −1/3
5 12 εF 2m 3
2

2 T 2 4
= ρεF + ρ π 2 − + ,
5 εF 12 12

which finally is expressed as

π2 T 2
P(ρ, T ) = P(ρ, T = 0) + ρ , (7.1.11)
6 εF
where we have put in evidence the value of the pressure at T = 0,
2
P(ρ, T = 0) = ρεF . (7.1.12)
5

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 78 — #94

i i

78 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Finally,
2 !
5π 2

T
P(ρ, T ) = P(ρ, T = 0) 1 + . (7.1.13)
12 εF
The pressure increases quadratically with the temperature.
(e) The chemical potential is related to the free energy and the pressure by

P(ρ, T )
µ(ρ, T ) = f (ρ, T ) + . (7.1.14)
ρ
Taking into account the expressions of the free-energy and the pressure we get
! !
3 5π 2 T 2 2 5π 2 T 2
µ(ρ, T ) = εF 1 − + εF 1 + , (7.1.15)
5 12 εF 5 12 εF

which finally can be written as

2 2 !
π2

2 T 1 1 T
µ(ρ, T ) = εF − π εF − = εF 1− . (7.1.16)
εF 4 6 12 εF

The chemical potential decreases quadratically with temperature in respect to

the value at T = 0 because the decrease of the free energy dominates over the
increase of the pressure term.

PROBLEM 7.2
Free Fermi sea at finite temperature

Consider a uniform free Fermi sea of non-interacting, non-relativistic fermions of

density ρ at temperature T . The occupation of a given momentum ~k is given by
ν
n(~k) = ε(k)−µ
, (7.2.1)
1+e T

where ε(k) = h̄2 k2 /2m, ν is the spin-isospin degeneracy and µ is the chemical
potential.

(a) Give an expression of the density in terms of the Fermi integrals defined as

xν dx
Z ∞
Jν (η) = . (7.2.2)
0 1 + ex−η

(b) Obtain the energy per particle in terms of the Fermi integrals.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 79 — #95

i i

Free Fermi sea at finite temperature 79

(a) The total number of particles can be expressed as the sum of all occupations,
ν
N=∑ ε(k)−µ
. (7.2.3)
~k 1+e T

To perform the summation it is useful to transform the sum into an integral by

means of Z
Ω
∑ −→ d~k , (7.2.4)
~
(2π)3
k

where Ω is the volume of the enclosing box, or equivalently,

Z
Ω
∑ −→ (2π h̄)3 d~p . (7.2.5)
~p

Then, Z
Ω ν
N= d~p . (7.2.6)
(2π h̄)3 1+e
ε(p)−µ
T

Therefore,
dN 4π νΩp2
= 3
, (7.2.7)
dp (2π h̄) 1 + e ε(p)−µ
T

where ε(p) = p2 /2m, and d p = (2m)1/2 dε/(2ε 1/2 ). Then the density of states
per unit energy can be expressed as

dN 4πνΩ 2mε(2m)1/2 1
= , (7.2.8)
dε (2π h̄)3 1 + e ε−µ
T 2ε 1/2

and the total number of particles:

νΩm3/2 ε 1/2
Z ∞ Z ∞
N= dNε = dε . (7.2.9)
21/2 π 2 h̄3
ε−µ
0 0 1+e T

Now, it is convenient to do the following change of variables: z = ε/T , and dε =

dzT to get
νΩm3/2 T 3/2 ∞ z1/2 dz
Z
N= . (7.2.10)
21/2 π 2 h̄3 0 1 + e(z−µ/T )
Finally,
mT 3/2 ∞ z1/2
Z
ν
ρ = 1/2 2 , (7.2.11)
2 π h̄2 0 1+e
z−η

with η = µ/T , which can be expressed in terms of the Fermi integrals as

3/2
ν mT
ρ= J1/2 (η) . (7.2.12)
21/2 π 2 h̄2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 80 — #96

i i

80 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

(b) For the total energy we must sum the energies of each state times their occupa-
tions:
νΩm3/2 ∞ ε 3/2
Z ∞ Z
E= ε dNε = 1/2 2 3 dε . (7.2.13)
0 2 π h̄ 0 1 + e(ε−µ)/T
Performing the same change of variable as before,
νΩm3/2 T 5/2 z3/2
Z ∞
E(η) = dz . (7.2.14)
21/2 π 2 h̄3 0 1 + ez−η
The energy density will be:
E(η) νm3/2 T 5/2 z3/2 νm3/2 T 5/2
Z ∞
eΩ = = 1/2 2 3 dz = J3/2 (η) , (7.2.15)
Ω 2 π h̄ 0 1 + ez−η 21/2 π 2 h̄3
and the energy per particle,
E(η) 1 νm3/2 T 5/2
e= = J (η) . (7.2.16)
N ρ 21/2 π 2 h̄3 3/2
Usually we give ρ and T ; then we determine η by inverting the relation between
ρ and η, which permits us to obtain the chemical potential, µ = ηT , and after-
wards we calculate the energy per particle e(ρ, T ).

PROBLEM 7.3
High T limit of the free Fermi sea

Consider a uniform free Fermi sea of non-interacting, non-relativistic fermions of

density ρ at temperature T . For a given density, when T becomes very large or
for a given temperature, when ρ becomes very small, then z = eµ/T is such that
0 < z 1. In this limit, the function

1 xn−1
Z ∞
fn (z) = dx (7.3.1)
Γ(n) 0 1 + ex z−1

can be expanded as

1 ∞
z2 z3
Z ∞
fn (z) = xn−1 ∑ (−1)l−1 (ze−x )l dx = z − + +... (7.3.2)
Γ(n) 0 l=1 2n 3n

(a) Using the first term of this expansion find the chemical potential from the
expression
mT 3/2

ν
ρ = 1/2 2 J1/2 (η) , (7.3.3)
2 π h̄2
where η = µ/T , and

x1/2
Z ∞
J1/2 (η) = dx . (7.3.4)
0 1 + ex e−η

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 81 — #97

i i

Free Fermi sea at finite temperature 81

(b) Calculate the energy per particle in this limit.

(c) Calculate the pressure in this limit.

(a) The density and the chemical potential at a given temperature are related by
3/2
ν mT
ρ= J1/2 (η) , (7.3.5)
21/2 π 2 h̄2

where η = µ/T . Using the expansion of f3/2 (z) we can write,

3/2 Z 3/2
x3/2−1

ν mT ∞ ν mT
ρ= −η
dx ≈ 1/2 2 Γ(3/2)z , (7.3.6)
21/2 π 2 h̄2 0
x
1+e e 2 π h̄2

where z = eη and Γ(3/2) = π 1/2 /2. Therefore,

" 3/2 #
ρ 2π h̄2
η = ln . (7.3.7)
ν mT

Taking into account that η = µ/T , we have

" 3/2 #
ρ 2π h̄2
µ = T ln , (7.3.8)
ν mT

which coincides with the chemical potential of a classical ideal gas. Notice that
µ → −∞ when ρ → 0 for a given T , or when T → ∞ for a given ρ.
(b) The energy per particle can be expressed as:

1 νm3/2 T 5/2 x3/2

Z ∞
e= dx . (7.3.9)
ρ 21/2 π 2 h̄3 0 1 + ex e−η

Performing the expansion of J3/2 (z), when z = eη → 0,

x3/2
Z ∞ 5/2−1
1 x
Z ∞
J3/2 (z) = x −1
= Γ(5/2) dx ≈ Γ(5/2)z
0 1+e z Γ(5/2) 0 1 + ex z−1
=Γ(5/2)eη .

Then,
1 νm3/2 T 5/2
e= Γ(5/2)eη . (7.3.10)
ρ 21/2 π 2 h̄3

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 82 — #98

i i

82 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Taking into account that in this limit,

" 3/2 #
ρ 2π h̄2
η = ln ,
ν mT

one gets,
3/2
1 νm3/2 T 5/2 2π h̄2

ρ
e∼ Γ(5/2) , (7.3.11)
ρ 21/2 π 2 h̄3 ν mT
with Γ(5/2) = π 1/2 3/4 , and therefore
3
e∼ T, (7.3.12)
2
which is the energy per particle of an ideal classical gas, which depends linearly
on temperature. Notice that the expression is in agreement with the equipartition
theorem that assigns 1/2T for each degree of freedom and therefore in 3D one
gets 3T /2.
(c) In the grand canonical ensemble,
PΩ
= ∑ ln(1 + ze−β ε ) , (7.3.13)
T ε

where z = eµ/T and β = 1/T . To perform the summation over all the energy
states it is convenient to transform the summation into an integral:
1
Z ∞
Ων4π
∑→ (2m)3/2 ε 1/2 dε . (7.3.14)
ε (2π h̄)3 2 0

Therefore
1
Z ∞
Ων4π
∑ ln(1 + ze−β ε ) = (2π h̄)3 (2m)3/2 2 0
ε 1/2 ln(1 + ze−β ε )dε . (7.3.15)
ε

Performing the change of variables x = ε/T , dε = T dx we have:

1
Z ∞
Ων4π
∑ ln(1 + ze−β ε ) = (2π h̄)3
(2mT )3/2
2 0
ln(1 + ze−x )x1/2 dx . (7.3.16)
ε

Integrating by parts, with ds = x1/2 dx then s = 2/3x3/2 and u = ln(1 + ze−x )

which implies du = −ze−x /(1 + ze−x ) = −1/(1 + z−1 ex ) we get

12 x3/2
Z ∞
Ων4π
∑ ln(1 + ze−β ε ) = (2π h̄)3
(2mT )3/2
23 0 1 + z−1 ex
dx , (7.3.17)
ε

which can be written as:

1 1 1/2 1 x3/2
Z ∞
Ων4π
(2mT )3/2 π dx . (7.3.18)
(2π h̄)3 22 (3/2)(1/2)π 1/2 0 1 + z−1 ex

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 83 — #99

i i

Free Fermi sea at finite temperature 83

Taking into account that Γ(5/2) = (3/2)(1/2)π 1/2 we can identify the function
f5/2 (z) in the previous expression and we obtain:

PΩ (2πmT )3/2
= ∑ ln(1 + ze−β ε ) = Ων f (z) . (7.3.19)
T ε (2π h̄)3 5/2
If we eliminate the volume from both sides we have
P (2πmT )3/2
=ν f (z) . (7.3.20)
T (2π h̄)3 5/2
In the limit z → 0 we have that f5/2 (z) → z and therefore,

P (2πmT )3/2 µ/T

=ν e , (7.3.21)
T (2π h̄)3
Taking into account that in this limit,
3/2 ! 3/2
ρ 2π h̄2 ρ 2π h̄2

µ/T
µ = T ln =⇒ e = , (7.3.22)
ν mT ν mT
one obtains
P = ρT (7.3.23)
which is the pressure as function of ρ and T for a classical ideal gas. The pressure
is linear in temperature and the proportionality factor is just the density.

PROBLEM 7.4
Low T expansion

For a given density, at low temperatures η = µ/T 0. The Fermi integrals

xν
Z ∞
Jν (η) = dx ,
0 1 + ex−η
can be expanded for η 0 as

π2 7π 4 d3φ

φ (x) dφ
Z ∞ Z η
dx = φ (x) dx + + +···
0 1 + ex−η 0 6 dx x=η 360 dx3 x=η

(a) Apply the previous expression to derive an expansion for the Fermi integrals
J1/2 (η) and J3/2 (η) when η 0.

(b) Using these results find a low-temperature expansion for the chemical poten-
tial, by inverting the relation

mT 3/2 ∞ x1/2
Z
ν
ρ = 1/2 2 x−η
dx , (7.4.1)
2 π h̄2 0 1+e

where η = µ/T .

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 84 — #100

i i

84 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

(c) Using the expansion of µ derive a low-temperature expansion for the energy
per particle,
3/2 Z
x3/2

1 ν mT ∞
e= T dx . (7.4.2)
ρ 21/2 π 2 h̄2 0 1 + ex−η

(d) Derive the low-temperature expansion of the entropy per particle by using that
the entropy can be written as

5 e(ρ, T )
s(ρ, T ) = −η . (7.4.3)
3 T

(e) Find also the low-temperature expansion of the free-energy per particle

f (ρ, T ) = e(ρ, T ) − T s(ρ, T ) . (7.4.4)

(f) Derive the low-temperature expansion for the pressure by using

∂ f (ρ, T )
P(ρ, T ) = ρ 2 . (7.4.5)
∂ρ

(g) Finally, check that the low-temperature expansion of µ(ρ, T ) given by

P(ρ, T )
µ(ρ, T ) = f (ρ, T ) + (7.4.6)
ρ

coincides with the low temperature expansion of µ(ρ, T ) previously derived

by inverting the relation between the density and the chemical potential.

(a) For a given density, at very low temperatures, η = µ/T 0; then, the Fermi
integrals admit the following expansion:
π 2 dφ 7π 4 d 3 φ

φ (x)
Z ∞ Z η
x−η
dx = φ (x) dx + + +···
0 1+e 0 6 dx x=η 360 dx x=η
(7.4.7)
In particular,
x1/2 π 2 1 −1/2
Z ∞ Z η
J1/2 (η) = dx = x1/2 dx + x +···
0 1 + ex−η 0 6 2 x=η
2 η
π2 1 2 π2 1
= x3/2 + 1/2
+ · · · = η 3/2 + +···
3 0 12 x x=η 3 12 η 1/2
!
π2 T 2

2 µ 3/2
= 1+ +··· .
3 T 8 µ

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 85 — #101

i i

Free Fermi sea at finite temperature 85

In a similar way,

x3/2 π 2 3 1/2
Z ∞ Z η
J3/2 (η) = = x x3/2 dx +
+···
0 1 + ex−η
0 6 2 x=η
!
2 5/2 η π 2 3 1/2 5 π2 3 T 2

2 µ 5/2
= x + η +··· = 1+ +···
5 0 6 2 5 T 2 6 2 µ
!
5π 2 T 2

2 µ 5/2
= 1+ +··· .
5 T 8 µ

(b) At the quadratic order in T /µ 1,

3/2 2 !
π2

ν m 2 3/2 T
ρ' µ 1+ +··· .
21/2 π 2 h̄2 3 8 µ

We should isolate µ, and find the quadratic correction in temperature to the zero-
temperature chemical potential εF (T = 0) of the free Fermi sea,
2/3
h̄2 6π 2 ρ

εF (T = 0) = . (7.4.8)
2m ν
Then,
2/3 2 !2/3
h̄2 6π 2 ρ π2

T
' µ 1+ . (7.4.9)
2m ν 8 µ

Using the expansion (1 + x)2/3 ' 1 + 2x/3 valid when x → 0, we get

2/3 !
h̄2 6π 2 ρ π2 T 2

' µ 1+ . (7.4.10)
2m ν 12 µ

We solve this equation, iteratively

2 !
1 π2 T
µ ' εF 2 ' εF 1− , (7.4.11)
2 T 12 µ
1 + π12 µ

valid when T /µ 1. Now we substitute again µ by εF , and get the expansion

up to T 2
!
2 π
2 π2 T 2
µ ' εF − T = εF 1 − . (7.4.12)
12εF 12 εF

(c) The energy per particle is given by

3/2 Z
x3/2

1 ν mT ∞
e= T dx , (7.4.13)
ρ 21/2 π 2 h̄2 0 1 + ex−η

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 86 — #102

i i

86 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

which for η 0 is expanded as

3/2 µ 5/2 !
25 T 2
1 ν m 2 π
e' T 5/2 1+ . (7.4.14)
ρ 21/2 π 2 h̄2 5 T 8 µ

Then using the expansion of µ up to second order in T /µ,

!
π2 T 2
µ ' εF 1 − , (7.4.15)
12 εF

we obtain
3/2 2 !5/2
π2

1 ν m 2 5/2 T
e ' 1/2 2 εF 1−
ρ2 π h̄2 5 12 εF
 
2
2
 5π T 1 
× 1 + .
 
8 εF
2 2 

π2
 T

1 − 12 εF

Now, keeping quadratic terms in T /εF , we can compute the Taylor expansion of
the T −dependent factors:
!5/2
π2 T 2 5π 2 T 2

1− ' 1− ,
12 εF 24 εF

2 2
5π 2 5π 2

T T 1
1+ '1 + 2 2
8 8

εF εF 2
1 − π12 εTF
!
5π 2 T 2 π2 T 2

'1 + 1+
8 εF 6 εF
2
5π 2 T

'1 + .
8 εF

Now we need to get ρ to substitute it in the expression for e. We can obtain its
expression using the relation between the Fermi momentum and the energy at
T = 0 and substituting the degeneracy ν = 2:
2/3
h̄2 6π 2 ρ

2 1 m 2 3/2
εF (T = 0) = =⇒ ρ = εF .
2m ν 3 21/2 π 2 h̄
Bringing everything together,
! ! 2 !
5π 2 T 2 5π 2 T 2 5π 2

3 3 T
e ' εF 1 − 1+ ' εF 1+ .
5 24 εF 8 εF 5 12 εF

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 87 — #103

i i

Free Fermi sea at finite temperature 87

At low temperature, the energy per particle increases quadratically with T ,

e(ρ, T ) ' e(ρ, T = 0) + aT 2 , (7.4.16)

with a = π 2 /(4εF ).
(d) The entropy can be expressed as
5 e(ρ, T )
s(ρ, T ) = −η . (7.4.17)
3 T
Now using the expansion of each quantity we can write
" !# !
51 3 5π 2 T 2 εF π2 T 2
s= εF 1 + − 1−
3T 5 12 εF T 12 εF
εF 5π 2 T εF π 2 T 1 T
= + − + = π2 ,
T 12 εF T 12 εF 2 εF
which is linear in T .
(e) The free energy is expressed as

f (ρ, T ) = e(ρ, T ) − T s(ρ, T ) . (7.4.18)

Then using the expansion of each of term we have:

!
3 5π 2 T 2 π2 T
f (ρ, T ) ' εF 1 + −T
5 12 εF 2 εF
!
3 π2 T 2 3 5π 2 T 2
= εF − = εF 1 − ,
5 4 εF 5 12 εF

which decreases quadratically with T . Actually it can be expressed as f (ρ, T ) '

e(ρ, T = 0) − aT 2 where a is the same factor that appears in the low temperature
expansion of e(ρ, T ), i.e. a = π 2 /(4εF ) .
(f) Doing a similar process for the pressure
∂ f (ρ, T )
P(ρ, T ) = ρ 2 , (7.4.19)
∂ρ
one gets

∂ εF π 2 2 2 ∂ π2 T 2

23 1 2
P(ρ, T ) =ρ − T ρ = ρεF + ρ
5 ∂ρ 4 ∂ ρ εF 5 6 εF
2
2 !
2 5π T
= ρεF 1 + .
5 12 εF

The pressure increases quadratically with T .

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 88 — #104

i i

88 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

(g) Finally, the chemical potential can be also derived as

P(ρ, T )
µ(ρ, T ) = f (ρ, T ) + .
ρ

Then, introducing the expansion of each quantity we have

!
3 5π 2 T 2 2 π2 T 2
µ(ρ, T ) ' εF 1 − + εF +
5 12 εF 5 6 εF
!
π2 T 2

=εF 1 − ,
12 εF

which coincides with the expression of µ obtained from the normalization of the
density.

PROBLEM 7.5
Partition function of the free Fermi system

Consider a free Fermi system described by the Hamiltonian

Ĥ0 = ∑ εk ĉ†k ĉk . (7.5.1)

The Fock states associated to this Hamiltonian are described as

|φi i = |ni1 , ni2 , . . . nik , . . .i , (7.5.2)

with nik = 1 or 0, depending on whether the single-particle state k is occupied or

not, respectively.
These states are eigenstates of Ĥ0 , Ĥ0 |φi0 i = E0i |φi0 i with E0i = ∑k nik εk . No-
tice that we are working in the grand canonical ensemble and ∑k nik = N i is the
number of particles in each state. The number of particles is not fixed. The distri-
bution operator is defined as

ρ̂0 = e−β (Ĥ0 −µ N̂) , (7.5.3)

where β = 1/T , i.e., the inverse of the temperature, and µ is the chemical potential.

(a) Calculate the grand partition function Z0 = Tr(ρ̂0 ).

(b) Calculate the thermal average of the occupation number operator, ĉ†k ĉk , as a
function of µ and T in the grand canonical ensemble.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 89 — #105

i i

Free Fermi sea at finite temperature 89

(a) The matrix elements of ρ̂0 in the Fock states are:

With the matrix elements, one can calculate the trace:

Z0 = Tr(ρ̂0 ) = ∑ ρ0,i . (7.5.6)

This summation is equivalent to:

i
∑ ρ0,i = ∑ ∏ e−β (εk −µ)nk .
i i k

In
i this summation, the first term i = 1 corresponds to the vacuum:
n1 , . . . , nik , . . . = {0, 0, 0, . . . }; the next terms i = 2, 3, . . . correspond to the
states with one particle: {1, 0, 0, . . . } , {0, 1, 0, . . . } , {0, 0, 1, . . . } , . . . ; then, the
states with two particles, and so on.
Finally, the summation is given by:

Z0 = ∑ ρ0,i = 1 + e−β (ε1 −µ) + e−β (ε2 −µ) + · · ·

i
+ e−β (ε1 −µ) e−β (ε2 −µ) + e−β (ε1 −µ) e−β (ε3 −µ) + · · ·

= ∏ 1 + e−β (εk −µ) .
k

(b) Having Z0 , one can calculate the thermal average of any operator, in particular of
ĉ†k ĉk , as
D E Tr(ĉ†k ĉk ρ̂0 )
ĉ†k ĉk = . (7.5.7)
0 Tr(ρ̂0 )
In the free system, ρ̂0 is diagonal in the Fock basis, and ĉ†k ĉk is also diagonal.
Therefore D E D E hφ | ρ̂ | φ i
i 0 i
ĉ†k ĉk = ∑ φi ĉ†k ĉk φi (7.5.8)
0 i Z0
and
i i
ρ0,i hφi | ρ̂0 | φi i ∏l e−β (εl −µ)nl e−β (εl −µ)nl
= = = ∏ (1 + e−β (εl −µ) ) . (7.5.9)
Z0 Z0 ∏l (1 + e−β (εl −µ) ) l

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 90 — #106

i i

90 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Then,
i
D E D E e−β (εl −µ)nl
ĉ†k ĉk = ∑ ni1 , . . . nik , . . . c†k ck ni1 , . . . , nik , . . . ∏ . (7.5.10)
0 i l (1 + e−β (εl −µ) )

In the summation, we have contribution only from the states with nk = 1, and all
of them contain the factor
e−β (εk −µ)
. (7.5.11)
1 + e−β (εk −µ)
Therefore, we can write
i
D E e−β (εk −µ) e−β (εk −µ)nl
ĉ†k ĉk = ∑ ∏ . (7.5.12)
0 1 + e−β (εk −µ) i l6=k 1 + e−β (εl −µ)

Now,
h i
−β (εl −µ)nil 1 + e−β (εk −µ)
∑i ∏l6=k e ∏ l6=k
= = 1. (7.5.13)
−β −µ) ∏l6=k 1 + e−β (εk −µ)

∏l6=k 1 + e (εk

Finally,
D e−β (εk −µ)
E 1
ĉ†k ĉk = = . (7.5.14)
0 1 + e−β (εk −µ) 1 + eβ (εk −µ)
In the case of the free Fermi sea, εk = h̄2 k2 /(2m), where ~k is the momentum of
the single-particle state. Usually, for a given ρ and β = 1/T , one determines µ
by inverting the relation
1
Z
ν
ρ= d~k . (7.5.15)
(2π)3 1 + eβ (εk −µ)
Alternatively, for a given µ, we can calculate ρ.

PROBLEM 7.6
Entropy of the non-interacting Fermi sea

Consider a free Fermi system described by the Hamiltonian

Ĥ = ∑ εk ĉ†k ĉk . (7.6.1)

The grand canonical potential for the system is given by

Ω = −T ln Z , (7.6.2)

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 91 — #107

i i

Free Fermi sea at finite temperature 91

where Z is the partition function defined by:

Z(µ,V, T ) = Tr e−β (Ĥ−µ N̂) , (7.6.3)

where µ is the chemical potential, N̂ = ∑~k â~† â~k is the number operator, β = 1/T
k
and V is the volume.
Show that the entropy, which is defined like

∂Ω
S=− , (7.6.4)
∂ T µ,V

can be written as

S = − ∑ { f (ε) ln f (ε) + (1 − f (ε)) ln(1 − f (ε))} , (7.6.5)

where
1
f (ε) = . (7.6.6)
1 + e(ε−µ)/T

The partition function for the free Fermi sea can be expressed as

Z = ∏ 1 + e−(ε(k)−µ)/T . (7.6.7)
~k

Therefore the grand canonical potential is written as

h i
Ω = −T ∑ ln 1 + e−(ε(k)−µ)/T , (7.6.8)
~k

and the entropy S = − (∂ Ω/∂ T )µ,V has several contributions,

( )
h
−(ε(k)−µ)/T
i e−(ε(k)−µ)/T (ε(k) − µ)
S = ∑ ln 1 + e +T , (7.6.9)
~ 1 + e−(ε(k)−µ)/T T2
k

which can be transformed to

( " # )
e(ε(k)−µ)/T h −(ε(k)−µ)/T
i ε(k) − µ 1
S = ∑ ln (ε(k)−µ)/T 1 + e + ,
~k e T 1 + e(ε(k)−µ)/T
(7.6.10)
then
h i h i ε(k) − µ 1

S = ∑ ln e(ε(k)−µ)/T + 1 − ln e(ε(k)−µ)/T + .
~k
T 1 + e(ε(k)−µ)/T
(7.6.11)

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 92 — #108

i i

92 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

From now on we will not specify the argument in ε(k), but just write ε. Then we
have
h i ε −µ 1

S = − ∑ − ln 1 + e(ε−µ)/T + 1− . (7.6.12)
~
T 1 + e(ε−µ)/T
k

But,
ε −µ

1
(ε−µ)/T

= ln 1 − + ln 1 + e . (7.6.13)
T 1 + e(ε−µ)/T
Then,
i
h 1
S = − ∑ − ln 1 + e(ε−µ)/T + ln 1 − + ln 1 + e(ε−µ)/T
~k 1 + e(ε−µ)/T

1
× 1− ,
1 + e(ε−µ)/T
which can be grouped as

(ε−µ)/T 1
S = − ∑ ln e + 1 −1 + 1 −
~k 1 + e(ε−µ)/T

1 1
+ 1− ln 1 −
1 + e(ε−µ)/T 1 + e(ε−µ)/T
and written as

1 1
S =−∑ ln
~k 1 + e(ε−µ)/T 1 + e(ε−µ)/T

1 1
+ 1− ln 1 − .
1 + e(ε−µ)/T 1 + e(ε−µ)/T
Finally, if we introduce the Fermi thermal occupation function
1
f (ε) = , (7.6.14)
1 + e(ε−µ)/T
and taking into account the spin degeneracy ν, we have
S = −ν ∑ { f (ε) ln f (ε) + (1 − f (ε)) ln(1 − f (ε))} . (7.6.15)
~k

If we transform the discrete sum to the continuum we have:

V
Z
∑ = (2π)3 d~k . (7.6.16)
~k

Finally, we get:
Z
νV
S=− d~k { f (ε) ln f (ε) + (1 − f (ε)) ln(1 − f (ε))} . (7.6.17)
(2π)3

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 93 — #109

i i

Distribution functions
8
Correlation is a crucial concept in science. It is deeply related to our understanding
of the interrelation between different parties or physical systems, and thus to the
casual connection between phenomena. In physics, correlations appear, usually but
not only, as a consequence of the interaction between different parts of a system.
In usual statistics two variables are said to be correlated if their outcomes depend
linearly on each other. In our case, more physics motivated, two systems are cor-
related if their outcomes are not independent of each other. Interestingly, in many-
body quantum physics, there are non-trivial correlations that appear already in non-
interacting fermionic systems. This is due to the Pauli exclusion principle, that builds
a strong correlation on the measurement of, for instance, the positions of two identi-
cal fermions.
To quantify the number and properties of the correlations we can use so-called
distribution functions . In this chapter we present three exercises about the distribu-
tion functions of the Fermi sea.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 94 — #110

i i

94 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 8.1
Sequential condition for the two-body distribution function of the free Fermi
sea

Consider the two-body distribution function , g(r), of a free Fermi sea, given by
the expression

l(kF r)2
g(r) = 1 − , (8.1.1)
ν
where ν = 2(4) is the spin(-isospin) degeneracy and the Slater function l(kF r) is
expressed in terms of Bessel functions

3 j1 (kF r)
Z
ν ~
l(kF r) = d~k eik~r = .
(2π)3 ρ |~k|<kF kF r

Here, the Bessel function j1 (kF r) is given by the expression

sin(kF r) cos(kF r)
j1 (kF r) = − .
(kF r)2 kF r

Show that the sequential condition

Z
ρ d~r (g(r) − 1) = −1 (8.1.2)

is fulfilled.

(a) The two-body distribution function of the Fermi sea reads

l(kF r)2
g(r) = 1 − , (8.1.3)
ν
For the purpose of this exercise, it is convenient to learn how to perform two
radial integrals over l(kF r) and l 2 (kF r). Starting with the one for l(kF r), we find:
Z Z Z
ν ~
d~r l(kF r) = 3
d~r d~k eik~r
(2π) ρ ~
|k|<kF
1
Z Z Z
ν i~k~r ν ν
= d~k d~r e = d~k δ (~k) = .
ρ |~k|<kF (2π)3 ρ |~k|<kF ρ

Similarly, the integral for l 2 (kF r) reads:

Z Z Z Z
ν ν ~ 0 i~k0~r
2
d~r l (kF r) = d~r d~k eik~r ~
dk e .
(2π) (2π)3
3
|~k|<kF |~k0 |<kF

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 95 — #111

i i

Distribution functions 95

To perform this integral, we change the order of integration,

ν2 1
Z Z Z
~ ~0
d~k d~k0 d~r ei(k+k )~r
|~k|<kF |~k0 |<kF ρ (2π) (2π)3
2 3

ν2 ν2
Z Z Z
ν
= d~k d~k0 2 3
δ (~k +~k0 ) = 2 d~k = .
|~k|<kF |~k0 |<kF ρ (2π) ρ (2π)3 |~k|<kF ρ
With this expression and, using Eq. (8.1.1), we find
l(kF r)2
Z Z
ρν
ρ d~r (g(r) − 1) = ρ d~r − =− = −1 .
ν νρ

PROBLEM 8.2
Spin components of the two-body distribution function

The expectation value of a two-body operator

V̂ = ∑ v(ri j )Ôi j (8.2.1)

i< j

is given by

< V̂ > 1
Z
d~r v(r) Tr Ô12 G (~r1~σ1 ,~r2~σ2 )

= ρ (8.2.2)
N 2
where Tr indicates the trace in the spin-space and
N(N − 1)
Z
G (~r1~σ1 ;~r2~σ2 ) = Ψ∗ (x1 , . . . , xN )Ψ(x1 , . . . , xN )dx3 . . . dxN (8.2.3)
ρ2
where integration over xi implies both the integration over the spatial coordi-
nates and a summation over the spin variables ~σi . Then the distribution function
G (~r1~σ1 ;~r2~σ2 ) can be expressed in several ways:
1
G (~r1~σ1 ;~r2~σ2 ) = [gc (r12 ) + (~σ1~σ2 )gσ1 σ2 (r12 )]
ν2
1
= 2 P̂s gs (r12 ) + P̂T gT (r12 )
ν
1
= 2 g0 (r12 ) + P̂12 gM (r12 ) , (8.2.4)
ν
where P̂s and P̂t are the singlet and triplet spin projection operators, and P̂12 is the
spin exchange operator.

(a) Find out the relation between the different spin components of the distribution
function.

(b) Calculate explicitly the different components for the free Fermi sea, with a
Fermi momentum kF and spin degeneracy two.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 96 — #112

i i

96 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

(a) To obtain the different component form G (~r1~σ1 ;~r2~σ2 ) we should calculate some
traces in the spin space. For instance,
gc (r12 ) = Tr [G (~r1~σ1 ;~r2~σ2 )] , (8.2.5)
and
1
gσ1 σ2 = Tr [(~σ1~σ2 )G (~r1~σ1 ;~r2~σ2 )] . (8.2.6)
3
In fact, if we take
1
G (~r1~σ1 ;~r2~σ2 ) = [gc (r12 ) + (~σ1~σ2 )gσ1 σ2 (r12 )] , (8.2.7)
ν2
and taking into account that Tr(I) = ν 2 and Tr(~σ1~σ2 ) = 0 we easily obtain the
first relation. Next, we consider
1
Tr [(~σ1~σ2 )G (~r1~σ1 ;~r2~σ2 )] = Tr [(~σ1~σ2 )gc (r12 ) + (3 − 2(~σ1~σ2 ))gσ1 σ2 (r12 )] .
ν2
(8.2.8)
Therefore,
3ν 2
Tr[(~σ1~σ2 )G (~r1 , ~σ1 ;~r2~σ2 )] =
gσ σ (r12 ) . (8.2.9)
ν2 1 2
Then if we want to find the relation between two representations we should just
use these relations but with the other representation. For instance, we know that
gc (r12 ) = Tr[G (~r1~σ1 ,~r2~σ2 )] , (8.2.10)
but now we express G (~r1~σ1 ,~r2~σ2 )in terms of g0 (r12 ) and gM (r12 ),

1 1 1
gc (r12 ) = 2 Tr g0 (r12 ) + (1 + ~σ1~σ2 )gM (r12 ) = g0 (r12 ) + gM (r12 ) .
ν 2 2
(8.2.11)
On the other hand,
1
g~σ1~σ2 (r12 ) = Tr [(~σ1~σ2 )G (~r1~σ1 ,~r2~σ2 )] . (8.2.12)
3

Now, we use that P̂12 = 12 (1 + ~σ1~σ2 ) to transform

~σ1~σ2 3 − 2(~σ1~σ2 )
~σ1~σ2 P̂12 = + , (8.2.13)
2 2
and
3
Tr[(~σ1~σ2 )P̂12 ] = ν 2 . (8.2.14)
2
Therefore,
1 1
g~σ1~σ2 (r12 ) = 2
Tr (~σ1~σ2 ) g0 (r12 ) + P̂12 gM (r12 ) (8.2.15)
3ν
1 1 3 1
= 2 ν 2 gM (r12 ) = gM (r12 ) . (8.2.16)
3ν 2 2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 97 — #113

i i

Distribution functions 97

Therefore
1
gc (r12 ) = g0 (r12 ) + gM (r12 ) ,
2
1
g~σ1~σ2 (r12 ) = gM (r12 ) , (8.2.17)
2
and

g0 (r12 ) = gc (r12 ) − g~σ1~σ2 (r12 ) ,

gM (r12 ) = 2 g~σ1~σ2 (r12 ) . (8.2.18)

Now we establish the relations with gS (r12 ) and gT (r12 ),

1
gc (r12 ) = Tr [G (~r1~σ1 ,~r2~σ2 ))] = Tr[P̂S ]gS (r12 ) + Tr[P̂T ]gT (r12 ) . (8.2.19)
ν2
Now we use that Tr(P̂s ) = ν 2 /4 and that Tr(P̂T ) = 3/4ν 2 , to get that
1 3
gc (r12 ) = gS (r12 ) + gT (r12 ) , (8.2.20)
4 4
while
1
g~σ1~σ2 (r12 ) = Tr[(~σ1~σ2 )G (~r1~σ1 ,~r2~σ2 ))]
3
1 1
= 2
Tr (~σ1~σ2 )P̂S gS (r12 ) + (~σ1~σ2 )P̂T gT (r12 ) . (8.2.21)
3ν
Now we use P̂S = 14 (1 − ~σ1~σ2 ) and P̂T = 34 (1 + ~σ1~σ2 ) to simplify,
3 3
(~σ1~σ2 )P̂S = (~σ1~σ2 ) − 1) , (~σ1~σ2 )P̂T = (1 − (~σ1~σ2 ) . (8.2.22)
4 4
Performing the traces we get
1 1
g~σ1~σ2 (r12 ) = − gS (r12 ) + gT (r12 ) (8.2.23)
4 4
and we can also express the inverted relation:

gT (r12 ) = gc (r12 ) + g~σ1~σ2 (r12 ), gS (r12 ) = gc (r12 ) − 3 g~σ1~σ2 (r12 ) (8.2.24)

Finally we can also give the relations with g0 (r12 ) and gM (r12 ), which would be
obtained in a similar way:

gT (r12 ) = g0 (r12 ) + gM (r12 ), gS (r12 ) = g0 (r12 ) − gM (r12 ) . (8.2.25)

(b) In particular for the unpolarized free Fermi sea, we can start from
1
GFS (~r1~σ1 ,~r2~σ2 ) = [g0 (r12 ) + P12 gM (r12 )] , (8.2.26)
ν2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 98 — #114

i i

98 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

with ν = 2, g0 (r12 ) = 1 and gM (r12 ) = −l 2 (kF r), where l 2 (kF r) is the square of
the Slater function and kF is the Fermi momentum. Then we have
l 2 (kF r) 1
gc (r12 ) = 1 − , g~σ1~σ2 (r12 ) = − l 2 (kF r) , (8.2.27)
ν ν
and
gT (r12 ) = 1 − l 2 (kF r) , gS (r12 ) = 1 + l 2 (kF r) . (8.2.28)

PROBLEM 8.3
Spin of the free Fermi sea

Taking into account that the total spin of a system is

~S = ∑~si = 1/2 ∑ ~σi (8.3.1)

i i

calculate ~S2 for the free Fermi sea.

(a) The total spin operator in units of h̄ for an N particle system is

~S = ∑~si = 1 ∑ ~σi , (8.3.2)

i 2 i
then
~S2 = 1 ∑ ~σi2 + 1 2 ∑ ~σi~σ j , (8.3.3)
4 i 4 i< j

To evaluate the two-body part, we will use the two-body distribution function of
the free Fermi sea:
1 1 1 1
Z
hφFS | 2 ∑ ~σi~σ j |φFS i = 2 ρ d~r Tr [(~σ1~σ2 )G (~r1~σ1 ,~r2~σ2 )] , (8.3.5)
N 4 i< j 4 2

where ρ is the density of the free Fermi sea and Tr indicates the trace operation
in the spin space of a two-particle system. The distribution function in given by:
1
G (~r1~σ1 ,~r2~σ2 ) =

2
gc (r12 ) + (~σ1~σ2 ) g~σ1~σ2 (r12 ) , (8.3.6)
ν

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 99 — #115

i i

Distribution functions 99

where ν = 2 is the spin degeneracy. Now we can use that

(~σ1~σ2 )2 = 3 − 2(~σ1~σ2 ) , (8.3.7)

and that

Tr(~σ1~σ2 ) = 0, Tr[(~σ1~σ2 )2 ] = Tr(3) − 2Tr(~σ1~σ2 ) = 3ν 2 = 12 , (8.3.8)

and that for the free Fermi sea

l(kF r)2 ˆ l(kF r)2

1
G (~r1~σ1 ,~r2~σ2 ) = 2 (1 − )I − (~σ1~σ2 ) , (8.3.9)
ν ν ν

where l(kF r) is the Slater function. Then,

1 1 1 1 1
Z
hφFS | 2 ∑ ~σi~σ j |φFS i = 2 2 ρ d 3 r Tr((~σ1~σ2 )(~σ1~σ2 ))g~σ1~σ2 (r)
N 4 i< j 4 2ν
1 1 1
Z
= 2 ρ 2 ν 2 3 d~r g~σ1~σ2 (r)
4 2 ν
3
Z
= d~r g~σ1~σ2 (r) . (8.3.10)
4
For the free Fermi sea we have:
l(kF r)2
g~σ1~σ2 (r) = − , (8.3.11)
ν
and recall from Prob 8.1
l(kF r)2
Z
−ρ d3r = −1 . (8.3.12)
ν
Therefore,

l(kF r)2

1 1 3 3
Z
3
hφFS | 2 ∑ ~σi~σ j |φFS i = −ρ d r =− . (8.3.13)
N 4 i< j 4 ν 4

On the other hand,

1 1 3N 3
hφFS | ∑(~σi )2 |φFS i = hφFS | |φFS i = . (8.3.14)
N i N 4 4

Finally,
1 3 3
hφFS | ~S2 |φFS i = − = 0 . (8.3.15)
N 4 4

Now, if we consider that ~S2 = Sx2 + Sy2 + Sz2 , where each contribution is positive
defined, we can conclude that | φFS i is an eigenstate of ~S2 with zero eigenvalue.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 101 — #117

i i

Dynamic structure functions

9
The dynamic structure function, closely tied to the linear response function, describes
the system’s reaction to external forces and encodes information about its excitations
and correlations. It is thus an appropriate tool to study phenomena such as collective
modes, particle-hole excitations, and energy dissipation. The linear response func-
tion quantifies the proportional relationship between an applied external perturba-
tion and the induced response in the system, making it a cornerstone of many-body
physics. Problems in this section also explore key concepts such as sum rules, which
are integral constraints that provide insights into the conservation laws and spectral
properties of the system.
Together, these topics equip readers with essential tools for analyzing the micro-
scopic dynamics using the dynamic structure function as a starting point.

101

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 102 — #118

i i

102 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 9.1
Dynamic structure function for the Fermi sea at high momentum transfer

The dynamic structure function of the free Fermi sea can be expressed as
!!
h̄2 (~k +~q)2 h̄2 k2
Z
ν
S(q, E) = d~k Θ(kF − k) Θ(|~k +~q| − kF ) δ E− − ,
(2π)3 ρ 2m 2m

where q and E are the momentum and energy transferred to the system.

(a) Demonstrate that for q > 2kF the impulse approximation is exact.

(b) Calculate the dynamic structure function for q > 2kF .

(a) For q > 2kF and k < kF , the function Θ(|~k +~q | −kF ) is superfluous in the sense
that for q > 2kF and k < kF , |~k +~q| > kF . Therefore, the dynamic structure func-
tion reduces to:
!!
2~ 2 2 2
h̄ (k +~
q) h̄ k
Z
ν
S(q, E) = d~k Θ(kF − k)δ E − −
(2π)3 ρ 2m 2m
!!
2~ 2 2 2
~k n(k)δ E − h̄ (k +~q) − h̄ k
Z
ν
= d ,
(2π)3 ρ 2m 2m
where n(k) = Θ(kF − k) is the momentum distribution.This last expression coin-
cides with the impulse approximation.
(b) For q > 2kF , we have
!!
h̄2 (~k +~q)2 h̄2 k2
Z
ν
S(q, E) = d~k Θ(kF − k) δ E− − .
(2π)3 ρ 2m 2m
Switching to spherical coordinates will allow us to ignore the Θ function and to
compute the sum (~k +~q)2 :
Z Z kF Z π
d~k −→ 2π dk k2 dθ sin θ , (~k +~q)2 = k2 + q2 + 2kq cos θ .
R3 0 0
If we now use the change of variables x = cos θ and basic properties of the Dirac
δ function:
2
2h̄ kqx h̄2 q2
Z kF Z 1
ν 2
S(q, E) = dk k dxδ E − +
(2π)2 ρ 0 −1 2m 2m
Z kF Z 1
ν 2 m Em q
= dk k dx δ − − x .
(2π)2 ρ 0 −1 h̄2 kq h̄2 kq 2k

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 103 — #119

i i

Dynamic structure functions 103

From the δ function we realize that we have contribution only if x =

Em/(h̄2 kq) − q/(2k) happens to be such that −1 ≤ x ≤ 1. Therefore we will
have contribution different than zero, only if
Em q
−1 ≤ 2
− ≤ 1,
h̄ kq 2k
which can be easily manipulated to get
h̄2 kq h̄2 q2 h̄2 kq h̄2 q2
− + ≤E ≤ + .
m 2m m 2m
For each q and E, even if E fulfills the previous condition, the integral over x
imposes also a condition in the lower limit integration on k, as we require that
|x| < 1. Therefore,

Em q 1 Em q Em q
2
− <1⇒ 2
− <1⇒k> 2 − ,
h̄ kq 2k k h̄ q 2 h̄ q 2
and we can write
Z kF kF
ν m ν m 2
S(q, E) = dkk2 2
= 2
k ,
(2π)2 ρ q
2
E −1
2
h̄ kq (2π) ρ h̄ 2q q E −1
h̄2 q2 2 h̄2 q2
2m 2m

and finally " #

ν m kF2 1 mE q 2

S(q, E) = 2 2 − − .
4π ρ h̄ q 2 2 h̄2 q 2

PROBLEM 9.2
Sum rules of the dynamic structure function for q ≥ 2 kF

Use the definition of the dynamical structure function of the free Fermi sea for a
momentum transferred q ≥ 2 kF :

!!
h̄2 (~k +~q)2 h̄2 k2
Z
ν
S(q, E) = d~k θ (kF − k) δ E− − ,
(2π)3 ρ 2m 2m

where ν is the spin-isospin degeneracy and kF is the Fermi momentum, to

(a) Calculate the m0 and m1 energy-weighted sum rules.

(b) Calculate m0 and m1 using the explicit expression for q ≥ 2kF ,

" #
ν m 1 kF2 1 mE q 2

S(q, E) = 2 2 − − ,
4π ρ h̄ q 2 2 h̄2 q 2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 104 — #120

i i

104 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

when h̄2 q2 /2m − h̄2 kF q/2m ≤ E ≤ h̄2 q2 /2m + h̄2 kF q/2m and zero other-
wise.

(a) The m0 sum rule, Z ∞

m0 (q) = dE S(q, E) ,
0
using the definition given above, we have:
Z ∞ Z ∞ Z
ν
dE S(q, E) = dE d~k θ (kF − k)
0 0 (2π)3 ρ
!!
h̄2 (~k +~q)2 h̄2 k2
×δ E− − .
2m 2m

Changing the order of integration, i.e., integrating first over E, we can take ad-
vantage of the fact that:
!!
h̄2 (~k +~q)2 h̄2 k2
Z ∞
dE δ E − − = 1.
0 2m 2m

Therefore,
4 3
Z
ν ν
m0 (q) = d~k θ (kF − k) = πk = 1 ,
(2π)3 ρ (2π) ρ 3 F
3

as it should be, because

Z ∞
dE S(q, E) = S(q) ,
0

and for the free Fermi sea, S(q) = 1 when q ≥ 2kF .

On the other hand,
Z ∞ Z ∞
ν
m1 (q) = dE E S(q, E) = dE E
0 0 (2π)3 ρ
!!
h̄2 (~k +~q)2 h̄2 k2
Z
× d~k θ (kF − k)δ E− − .
2m 2m

Again, one should perform first the integral over the energy:
!!
h̄2 (~k +~q)2 h̄2 k2 h̄2~k ·~q h̄2 q2
Z ∞
dE E δ E − − = + .
0 2m 2m m 2m

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 105 — #121

i i

Dynamic structure functions 105

Therefore,
!
h̄2~k ·~q h̄2 q2
Z
ν
m1 (q) = d~k θ (kF − k) + .
(2π)3 ρ m 2m

Due to the angular integration, the term with h̄2~k ·~q/m gives zero contribution.
Then
h̄2 q2 h̄2 q2
Z
ν
m1 (q) = 3
d~k θ (kF − k) = ,
2m (2π) ρ 2m
as it should be. Actually this result is true also for an interacting system, at any
q.
(b) We now make the calculation using the explicit expression of the response for
q ≥ 2kF . Using the limits for E obtained in Exercise 9.1:
h̄2 q2 h̄2 kF q
" #
2m + m ν m 1 kF2 1 mE q 2
Z
m0 (q) = h̄2 q2 h̄2 kF q
dE − − .
2m − m
4π 2 ρ h̄2 q 2 2 h̄2 q 2

Now the following change of variable is convenient: E 0 = E − h̄2 q2 /2m, then

h̄2 kF q
" #
ν m 1 kF2 1 mE q 2
Z
m 0
m0 (q) = h̄2 k q dE − −
− mF 4π 2 ρ h̄2 q 2 2 h̄2 q 2
Z h̄2 kF q 2
1 m2

ν m1 m 0 kF 0 2
= 2 2 dE − (E )
4π ρ h̄ q − h̄2mkF q 2 2 h̄4 q2
!
ν m 1 kF2 2h̄2 kF q 1 m2 2 h̄6 kF3 q3
= 2 2 − 4 2
4π ρ h̄ q 2 m 2 h̄ q 3 m3
ν m 1 2 3 h̄2 q
= k = 1.
4π 2 ρ h̄2 q 3 F m

Similarly for m1 (q),

h̄2 q2 h̄2 kF q
" #
2m + m ν m 1 kF2 1 mE q 2
Z
m1 (q) = h̄2 q2 h̄2 kF q
dE E − − ,
2m − m
4π 2 ρ h̄2 q 2 2 h̄2 q 2

and performing again the same change of variables, E 0 = E − h̄2 q2 /2m, we get:
h̄2 kF q
h̄2 q2

ν m1
Z
m 0 0
m1 (q) = 2 2 h̄2 k q
dE E +
4π ρ h̄ q − mF 2m
2 2 2 !
kF2 1

m h̄ q q
× − E0 + − ,
2 2 h̄2 2q 2m 2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 106 — #122

i i

106 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

and then
h̄2 kF q 2 !
h̄2 q2 kF2 1 mE 0

ν m1
Z
m 0
0
m1 (q) = 2 2 h̄2 kF q
dE E + − .
4π ρ h̄ q − m
2m 2 2 h̄2 q
2 2
The integral multiplying the term h̄2mq is 1, because it corresponds to the normal-
ization integral calculated above for the m0 sum rule, and the contribution of the
terms with E 0 is zero, as it is an integral of an odd function over a symmetric
interval with respect to the origin:

Z h̄2 kF qm
kF2 1 m2

h̄2 kF q
dE 0 E 0 − 4 2 (E 0 )2 = 0.
m
2 2 h̄ q
Therefore,
h̄2 q2
m1 (q) = .
2m

PROBLEM 9.3
Maximum and width of the dynamic structure function of the free Fermi sea
for q ≥ 2 kF

The dynamic structure function of the free Fermi sea for a momentum transferred
q ≥ 2 kF is given by:

!!
h̄2 (~k +~q)2 h̄2 k2
Z
ν
S(q, E) = d~k θ (kF − k) δ E− − ,
(2π)3 ρ 2m 2m

where ν is the spin-isospin degeneracy and kF is the Fermi momentum. Perform-

ing the angular integration:
Z kF
ν m
S(q, E) = 2π k dk ,
(2π)3 ρ h̄2 q q E
2 | h̄2 q2 −1|
2m

we finally get:
2 !
kF2

ν m1 1 mE q
S(q, E) = 2 2 − −
4π ρ h̄ q 2 2 h̄2 q 2

when h̄2 q2 /2m − h̄2 kF q/2m ≤ E ≤ h̄2 q2 /2m + h̄2 kF q/2m and zero otherwise.
(a) Find the value of E that maximizes S(q, E) for a given q.
(b) Find the value of S(q, E) at the maximum.
(c) Find the width of S(q, E), i.e. the energy interval at which S(q, E) 6= 0.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 107 — #123

i i

Dynamic structure functions 107

(a) To find the maximum for a given q we can, by inspection of the definition of
S(q, E), see that S(q, E) will be maximum when E = h̄2 q2 /2m, because in this
case the lowest limit of the integral will be zero, and that will define the maximum
value of the integral.
Another way to find the maximum is to take the analytic expression of S(q, E)
and find the E for which the derivative of S(q, E) with respect to E is zero,

h̄2 q2

∂ S(q, E) mE q
= 0 =⇒ 2 2 − = 0 =⇒ E = .
∂E h̄ q 2 2m
One can also check that the second derivative with respect to the energy is nega-
tive and therefore this stationary point corresponds to a maximum.
This maximum is known as the quasi-elastic peak, which means that all the en-
ergy associated to the momentum q has been transferred to a particle of momen-
tum zero in the free Fermi sea as if the collision was elastic.
(b) The value at the maximum is given by:
ν mkF2 1 3 m
S(q, Emax ) = = .
4π 2 ρ h̄2 2 q 4 h̄2 qkF
Notice that S(q, E) has dimensions of the inverse of energy and decreases in-
versely proportional to q.
(c) Taking the definition of S(q, E) and looking where it is not zero, we immediately
see that the width is given by
2h̄2 kF q
.
m
Notice that the width increases linearly with q, and the value at the maximum
decreases inversely proportional to q, in such a way that the total surface below
S(q, E) for q > 2kF is equal to 1, independently of the value of q.

PROBLEM 9.4
Compton profile and Y −scaling for the free Fermi sea

The Compton profile of the free Fermi sea for a momentum transfer q ≥ 2 kF is
given by:
Z ∞
J(Y ) = 2π dq q n̄(q) ,
|Y |

where Y = mE/h̄2 q − q/2 and

ν
n̄(q) = n(q) ,
(2π)3 ρ

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 108 — #124

i i

108 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

with n(q) = θ (kF − q). Then n̄(q) is normalized to unity:

Z
d~q n̄(q) = 1 .

Actually, the definition of the Compton profile exploits the fact that S(q, E) for
large q (q ≥ 2kF for the free Fermi sea) depends only on Y , i.e., scales on this
variable. In fact,
m
J(Y ) = 2 S(q, E) .
h̄ q

(a) Calculate J(Y ).

(b) Calculate m0 .
Z kF
m0 = J(Y )dY .
−kF

(c) Calculate m1 and m2

Z kF Z kF
2m
m1 = Y J(Y ) , m2 = Y 2 J(Y )dY = htˆi ,
−kF −kF 3h̄2

where htˆi is the expectation value of the kinetic energy per particle in the free
Fermi sea.

(a) The Compton profile is defined as

Z ∞
J(Y ) = 2π dqq n̄(q) .
|Y |

Therefore,
Z ∞ Z kF
ν ν
J(Y ) =2π θ (kF − q)qdq = 2π q dq
(2π)3 ρ |Y | (2π)3ρ
|Y |
kF
ν 1 2 ν 1 2
kF −Y 2 ,

=2π 3
q = 2
(2π) ρ 2 |Y | (2π) ρ 2

which is a parabolic function of Y defined for |Y | < kF . J(Y ) is an even function

of Y , whose maximum is located at Y = 0, and whose value at the maximum is
J(0) = 3/(4kF ).

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 109 — #125

i i

Dynamic structure functions 109

(b) We can calculate m0 as:

Z kF Z kF
ν 1
m0 = J(Y )dY = 2ρ 2
(kF2 −Y 2 )dY
−kF (2π) −kF
" kF
#
Y3

ν 1 2 kF ν 1 3 2 3
= kF Y −k − = 2k F − k
2
(2π) ρ 2 F 3 −kF (2π)2 ρ 2 3 F
ν 4π 3
= k = 1.
(2π)3 ρ 3 F

(c) Let us calculate m1 :

Z kF
m1 = Y J(Y )dY = 0
−kF

because it is the integral of an odd function over a symmetric interval with respect
to the origin.
On the other hand, for m2 we have:

1 kF 2 2
Z ∞ Z
ν
m2 = Y 2 J(Y )dY = 2
Y (kF −Y 2 )dY
−∞ (2π) ρ 2 −kF
" #
1 2 1 3 kF 1 5 kF 1 2kF5 2kF5

ν ν
= k Y − Y = −
(2π)2 ρ 2 F 3 −kF 5 −kF (2π)2 ρ 2 3 5
ν 4πkF3 kF2 2m 3 h̄2 kF2 2m
= 3
= 2
= 2 htˆi ,
(2π) ρ 3 5 3h̄ 5 2m 3h̄
where
3 h̄2 kF2
htˆi =
5 2m
is the average kinetic energy of the free Fermi sea.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 111 — #127

i i

Hartree-Fock
10
This chapter focuses on problems related to the Hartree-Fock method, an important
method used in quantum many-body theory for many years. The Hartree-Fock ap-
proach approximates the many-electron wavefunction as a single Slater determinant,
capturing the effects of exchange interactions between electrons while neglecting ex-
plicit electron correlation. The Hartree-Fock method also plays a foundational role
in the development and application of Density Functional Theory, a self-consistent
method that recasts the problem in terms of electron density, making it computation-
ally efficient for large systems. By solving the equations iteratively, one can obtain
self-consistent solutions for the electronic structure of atoms, molecules, and solids.
With these problems the reader will gain a thorough understanding of the core
concepts, techniques, and limitations of this fundamental quantum mechanical ap-
proach.

111

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 112 — #128

i i

112 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 10.1
Hartree-Fock for nuclear matter with a simple Skyrme interaction

Consider uniform symmetric nuclear matter at a given density. The nucleon-

nucleon interaction is described by means of an effective contact interaction

1
v(ri j ) = t0 + t3 ρ α δ (~ri −~r j ) (10.1.1)
6

with t0 = −1794 MeV fm3 , t3 = 12817 MeV fm3+α and α = 1/3. Remember that
for nucleons h̄2 /m = 41.4687 MeV fm2 .
This density dependent effective interaction has been constructed to produce
a good saturation density and binding energy.

(a) Calculate the diagonal two-body matrix elements taking plane waves normal-
ized to volume as single-particle states.

(b) Calculate the energy in the Hartree-Fock approximation as a function of den-

sity for nuclear matter.

(d) Calculate the speed of sound in nuclear matter.

(a) The direct part of the diagonal two-body matrix element:

D
~k1 ms mτ ~k2 ms mτ 1 α ~k1 ms mτ ~k2 ms mτ
E
1 1 2 2 t0 + t 3 ρ δ (~
r 1 −~r2 ) 1 1 2 2
6

1 1
Z Z
~ ~ ~ ~
= t0 + t3 ρ α d~r1 d~r2 e−ik1~r1 e−ik2~r2 δ (~r1 −~r2 ) eik1~r1 eik2~r2
6 Ω2
hms1 ms2 |ms1 ms2 i hmτ1 mτ2 |mτ1 mτ2 i
Z
1 1 α 1 1 α
= 2 t0 + t3 ρ d~r1 = t0 + t3 ρ ,
Ω 6 Ω 6

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 113 — #129

i i

Hartree-Fock 113

while the exchange term reads:

D
~k1 ms mτ ~k2 ms mτ 1 α E
1 1 2 2 t0 + t 3 ρ δ (~r1 −~r2 ) ~k2 ms2 mτ2~k1 ms1 mτ1
6

1 1
Z
~ ~ ~ ~
= t0 + t3 ρ α d~r1 d~r2 e−ik1~r1 e−ik2~r2 δ (~r1 −~r2 )eik2~r1 eik1~r2
6 Ω2
hms1 ms2 |ms2 ms1 i hmτ1 mτ2 |mτ2 mτ1 i

1 1
Z
~ ~
= t0 + t3 ρ α δm m δm m d~r1 d~r2 ei(k2 −k1 )(~r1 −~r2 ) δ (~r1 −~r2 )
6 Ω2 s1 s2 τ1 τ2

1 1
= t0 + t3 ρ α δms1 ms2 δmτ1 mτ2 .
Ω 6

Finally the antisymmetrized two-body matrix element is:

D E
~k1 ms mτ ,~k2 ms mτ v(r12 ) ~k1 ms mτ ,~k2 ms mτ −~k2 ms mτ ,~k1 ms mτ
1 1 2 2 1 1 2 2 2 2 1 1

1 t3
= t0 + ρ α 1 − δms1 ms2 δmτ1 mτ2 .
Ω 6

(b) Now, to evaluate the expectation value of the interaction in the free Fermi sea we
should sum up all the two-body matrix elements:
hΦFS |V |ΦFS i 1 1
Z Z
Ω Ω
= ∑ d~k1 d~k2
N N 2 ms1 ms2 (2π) (2π)3
3
|~k1 |≤kF |~k2 |≤kF
mτ1 mτ2
1 t3
t0 + ρ α 1 − δms1 ms2 δmτ1 mτ2 .
Ω 6
Performing the integrals and the summations we arrive to:
t
4 3 2

hΦFS |V |ΦFS i 1 1 Ω t0 + 63 ρ α

= πk [Tr(1) − Tr(Pσ Pτ )]
N 2 N (2π)3 (2π)3 3 F
4 3 2

1Ω 1 t3 α 2 1
= πk t0 + ρ ν 1−
2 N (2π)6 3 F 6 ν

1 t3 α 1 1 t3 α 3
= ρ t0 + ρ 1− = ρ t0 + ρ ,
2 6 ν 2 6 4
where ν is the spin-isospin degeneracy ν = 4. Adding the average kinetic energy,
we finally can write the energy per particle as
2/3
h̄2 3 3π 2

1 t3 3
e(ρ) = ρ 2/3 + ρ t0 + ρ α . (10.1.2)
2m 5 2 2 6 4

The energy has a minimum at the saturation density, ρ0 = 0.16 fm−3 and e0 =
−15.95 MeV. By definition, at ρ0 the pressure is zero, and therefore the chemical

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 114 — #130

i i

114 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

e(ρ) µ(ρ)
0.0 0.0

−2.5 −2.5

Chemical potential µ(ρ) [MeV]

Energy per particle e(ρ) [MeV]

−5.0 −5.0

−7.5 −7.5

−10.0 −10.0

−12.5 −12.5

−15.0 −15.0

−17.5 −17.5

−20.0 −20.0

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35

Number density ρ [fm−3 ]

Figure 10.1 Energy per particle and chemical potential of nuclear matter as a function of
the density. The black point marks the energy minimum at saturation density, and the gray one
the spinodal point.

potential,
P(ρ)
µ(ρ) = e(ρ) + , (10.1.3)
ρ
at ρ0 coincides with the binding energy. Due to the contribution of the pressure,
µ(ρ) grows faster, for ρ > ρ0 than e(ρ). On the other hand, for ρ < ρ0 , as P(ρ) <
0, µ(ρ) lies below e(ρ). At ρ = 0.354 fm−3 the binding energy is zero.
(c) The chemical potential is

∂E ∂ (Ne(ρ)) ∂e ∂ρ ∂ e(ρ)
µ(ρ) = = = e(ρ) + N = e(ρ) + ρ ,
∂N Ω ∂N Ω ∂ ρ ∂ N Ω ∂ρ
(10.1.4)
and therefore
2/3
h̄2 3 3π 2

1 t3 3
µ(ρ) = ρ 2/3 + ρ t0 + ρ α
2m 5 2 2 6 4
2 2 2/3
h̄ 3 3π 2 2/3 3 1 t3 3
+ ρ + ρt0 + (α + 1)ρ α+1
2m 5 2 3 8 2 64
2/3
h̄2 3π 2

3 t3 3
= ρ 2/3 + ρt0 + (α + 2)ρ α+1 ,
2m 2 4 68

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 115 — #131

i i

Hartree-Fock 115

Presure P(ρ) [MeV fm−3 ] 4

−1
0.00 0.05 0.10 0.15 0.20 0.25
−
Number density ρ [fm ]3

Figure 10.2 Pressure as a function of density. The black point marks the saturation density,
and the grey one the spinodal point.

and the pressure is

∂E de(ρ)
P(ρ) = − = ρ2 = ρ[µ(ρ) − e(ρ)] . (10.1.5)
∂Ω N dρ
Therefore,
2/3
h̄2 2 3π 2

3 3
P(ρ) = ρ 5/3 + ρ 2t0 + (α + 1)ρ α+2t3 . (10.1.6)
2m 5 2 8 48

The pressure is zero at the saturation density and has a minimum at the spinodal
point, ρs = 0.102 fm−3 . Below this density, the system is mechanically unstable
and the speed of sound is an imaginary number.
(d) The speed of sound can be calculated as
1/2
1 h̄2 c2 −1

cs (ρ) = k , (10.1.7)
h̄ mc2 ρ
where c is the speed of light and k−1 is the incompressibility

−1 dP(ρ)
k (ρ) = ρ , (10.1.8)
dρ
and therefore,
2/3
h̄2 2 3π 2

−1 6 3(α + 1)(α + 2) α+2
k (ρ) = ρ 5/3 + ρ 2t0 + ρ t3 . (10.1.9)
2m 3 2 8 48

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 116 — #132

i i

116 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Finally, the speed of sound in units of the speed of light is:

1/2
k−1

cs
= . (10.1.10)
c mc2 ρ

PROBLEM 10.2
Hartree-Fock for a system of N fermions enclosed in a one-dimensional box

Consider a system of N fermions of mass m and spin 1/2, all with the same spin
orientation, in a one-dimensional box. The single-particle Hamiltonian of the sys-
tem is
N n
h̄2 N
Ĥ0 = ∑ ĥ0 (i) = ∑ − ∇2i + ∑ U(xi ) , (10.2.1)
i=1 i=1 2m i=1

where U(x) is a square well potential with infinite walls: U(x) = 0 if 0 < x <
a and U(x) = ∞ if x > a or x < 0. The single-particle eigenfunctions and their
eigenenergies are:
1/2
h̄2 π 2 k2

2 πkx
φk = sin , εk = , k = 1, 2, 3, ... . (10.2.2)
a a 2ma2

Suppose that there is a two-body interaction of the type V̂ = ∑i< j v(xi , x j ) with
v(xi , x j ) = C, where C is a constant measured in units of energy. Then the total
Hamiltonian is Ĥ = Ĥ0 + V̂ .

(a) Calculate the two-body matrix element

hi j| v |kl − lki

where |kl − lki = |kli − |lki .

(b) For this Hamiltonian, the single-particle wavefunctions solution of the

Hartree-Fock equations are the single-particle wavefunctions defined above.
Calculate the single-particle Hartree-Fock energies:

ε(i) = i|ĥ0 |i +UHF (i) , (10.2.3)

N
UHF (i) = ∑ hi j| v |i j − jii . (10.2.4)
j=1

(c) Calculate the total energy by using

N N
1
EHF = ∑ εHF (i) − 2 ∑ UHF (i) . (10.2.5)
i=1 i=1

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 117 — #133

i i

Hartree-Fock 117

(d) Check that this energy coincides with the expectation value , hΨHF | Ĥ |ΨHF i,
where |ΨHF i is the Slater determinant built with the lowest N single-particle
wavefunctions.

(e) Check that Ĥ in second quantization is written as

1 1
Ĥ = ∑ εi â†i âi + CN̂ 2 − CN̂ , (10.2.6)
i 2 2

where â†i (âi ) are the creation (annihilation) operators associated to the single-
particle basis φi (x), and N̂ = ∑i â†i âi is the number of particles operator.

†
(f) Demonstrate that the state |Ψi = ∏N i=1 âi |0i is an eigenstate of Ĥ.

(a) Taking into account the orthonormality of the single-particle states and the fact
that the interaction is a constant, we have

hi j| v |kl − lki = C hi j|kl − lki = C(δik δ jl − δil δ jk ) . (10.2.7)

(b) The HF single-particle energies are given by

εHF (i) = hi| ĥo |ii +UHF (i) , (10.2.8)

where
h̄2 π 2 i2
hi| ĥo |ii = (10.2.9)
2ma2
and

UHF (i) = ∑ hi j| v |i j − jii = C ∑ (1 − δi j ) = C(N − 1) , (10.2.10)

j≤N j≤N

which is a reasonable result, as UHF (i) takes into account the interaction of the
state i with the rest of occupied states.
(c) The total HF energy is:
N
1 N
EHF = ∑ εHF (i) − ∑ UHF (i) . (10.2.11)
i=1 2 i=1

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 118 — #134

i i

118 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Taking into account all the previous results we have:

h̄2 π 2 N 2 N 1 N
EHF = ∑ i + ∑ C(N − 1) − ∑ C(N − 1)
2ma2 i=1 i=1 2 i=1
h̄2 π 2 N 2 1
=
2ma2 i=1∑ i +CN(N − 1) − 2 CN(N − 1)
h̄2 π 2 N 2 1
=
2ma2 i=1∑ i + 2 CN(N − 1) .

(d) The Hartree-Fock energy can be also calculated as the expectation value of Ĥ in
the HF wavefunction
N
h̄2 π 2 2 N
EHF = hΨHF | Ĥ |ΨHF i = ∑ 2
i + ∑ hi j| v |i j − jii , (10.2.12)
i=1 2ma i< j

which corresponds to the evaluation of a one-body operator Ĥ0 and a two-body

operator V̂ , in the Slater determinant built with the first N single-particle solutions
of the HF equations. Then
N
h̄2 π 2 2 N(N − 1)
EHF = ∑ 2
i +C . (10.2.13)
i=1 2ma 2

(e) The Hamiltonian in the second quantization for this specific case, in which the
interaction is a constant, is written as
1
Ĥ = εi â†i âi + hi j|v|i jiâ†i â†j â j âi , (10.2.14)
2∑ij

where
h̄2 π 2 i2
εi = , hi j|v|i ji = C . (10.2.15)
2ma2
On the other hand,
â†i â†j â j âi = −â†i âi δi j + â†i âi â†j â j . (10.2.16)
Therefore,
1 1
∑ hi j| v |i ji â†i â†j â j âi = C ∑(−â†i âi δi j + â†i âi â†j â j )
2 ij 2 ij
1 1
= − C ∑ â†i âi + C(∑ â†i âi )(∑ â†j â j )
2 i 2 i j
1 1
= − CN̂ + CN̂ 2 .
2 2
Therefore
1 1
Ĥ = ∑ εi â†i âi + N̂ 2 − CN̂ . (10.2.17)
i 2 2

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 119 — #135

i i

Hartree-Fock 119

(f) In this case, for a constant interaction, the HF-ground state is an eigenstate. In
fact,
! ! ! ! !
N N N N
∑ εi â†i âi ∏ â†j |0i = ∑ εi ∏ â†j |0i = ∑ εi |ΨHF i ,
i j=1 i=1 j=1 i=1
(10.2.18)
and the interaction term is

1 2 1 1 1
CN̂ − CN̂ |ΨHF i = C(N 2 − N) |ΨHF i = CN(N − 1) |ΨHF i .
2 2 2 2
(10.2.19)
Therefore, we can conclude that
!
h̄2 π 2 N 2 1
Ĥ |ΨHF i = ∑ i + 2 CN(N − 1) |ΨHF i . (10.2.20)
2ma2 i=1

PROBLEM 10.3
Single-particle potential for a uniform system of fermions in the Hartree-Fock
approximation

Consider uniform system of fermions of spin 1/2, with the same number of
fermions with spin up and spin down, interacting through an interaction depending
only on the distance:
V = ∑ v(ri j ) . (10.3.1)
i< j

(a) Show that the single-particle potential can be written in the following way:

l(kF r) i~k~r
Z
U(k) = ρ d~r v(r) 1 − e , (10.3.2)
ν

where l(kF r) is the so-called Slater function and ν is the spin degeneracy.

(b) Demonstrate that the expectation value hΦFS |V |ΦFS i can be calculated as

hΦFS |V |ΦFS i 1 ν 1
Z
= d~k U(k) . (10.3.3)
N 2 ρ (2π)3 |~k|≤kF

(a) The single-particle potential for k is the sum of the interactions of the parti-
cle with momentum ~k with all the other particles. For a uniform system and an

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 120 — #136

i i

120 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

isotropic interaction, the potential depends only on the modulus of ~k.

1
U(k) = ∑ h~kmσ~k0 m0σ |v(r12 )|~kmσ~k0 m0σ −~k0 m0σ~kmσ i
ν~
km0σ mσ
1 Ω 1
Z Z
= d~k0 ∑ 2 d~r1 d~r2 v(r12 )
ν (2π)3 |~k0 |≤kF m0σ mσ
Ω
h 0
i
~ ~ ~ ~0 ~0 ~
e−ik~r1 e−ik r2 mσ m0σ eik~r1 eik r2 mσ m0σ − eik ~r1 eikr2 m0σ mσ
1 Ω 1
Z Z
= d~k0 d~r1 d~r2 v(r12 )
ν (2π)3 Ω2 |~k0 |≤kF
0

Tr(I) − Tr(Pσ ) eik̃(r̃2 −r̃1 ) eik̃ (r̃1 −r̃2 ) .

Now, one should take into account the following integrals:

4
Z
ρ
d~k = πkF3 = (2π)3 (10.3.4)
|~k|≤kF 3 ν

and
(2π)3 ρ
Z
~
d~k eik~r = l(kF r) . (10.3.5)
|~k|≤kF ν
Taking into account these integrals, one can integrate first over momenta:

1 Ω ν2

1 i~k0 (~r1 −~r2 ) i~k(~r2 −~r1 )
Z Z
~ 0
U(k) = d~r1 d~r2 v(r12 ) dk 1 − e e ,
ν (2π)3 Ω2 |~k0 |≤kF ν
(10.3.6)
obtaining

(2π)3 ρ (2π)3 ρ

1 Ω 1 2
Z
i~k(~r2 −~r1 )
U(k) = ν d~r d~
1 2 r v(r12 ) − l(k r
F 12 )e
ν (2π)3 Ω2 ν ν2
(10.3.7)
and finally
l(kF r) i~k~r
Z
U(k) = ρ d~r v(r) 1 − e (10.3.8)
ν
that will depend only on the modulus of ~k.

(b) Now, integrating the single-particle potential we recover the expectation value of
the interaction. In fact,
hΦFS |V |ΦFS i 1 Ω
Z
= ν d~kU(k)
N 2 (2π)3 N |~k|≤kF

1 1 1 ~k ρ d~r v(r) 1 − l(kF r) ei~k~r ,
Z Z
= ν d
2 (2π)3 ρ |~k|≤kF ν

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 121 — #137

i i

Hartree-Fock 121

changing the order of integration:

hΦFS |V |ΦFS i 1 ν

l(kF r) i~k~r
Z Z
= d~r v(r) ~
dk 1 − e
N 2 (2π)3 |~k|≤kF ν
(2π)3 ρ l(kF r) (2π)3 ρ

1 ν
Z
= d~r v(r) − l(kF r)
2 (2π)3 ν ν ν
l(kF r)2

1 1
Z Z
= ρ d~r v(r) 1 − = ρ d~r v(r)gFS (r) ,
2 ν 2

where gFS (r) is the two-body distribution function associated to the free Fermi
sea (see Problem 8.2).

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 123 — #139

i i

Density matrices
11
This chapter delves into the concept of density matrices, a fundamental tool in quan-
tum mechanics for describing the state of a many-body system. Density matrices
provide a compact and powerful representation of quantum states, encompassing
both pure and mixed states. The study of one-body and two-body density matrices
is particularly important for understanding the behavior of many-body systems, es-
pecially in contexts like quantum gases, solid-state physics, and quantum chemistry.
These matrices encapsulate information about particle distributions and correlations,
enabling the derivation of key properties such as energy, momentum, and entropy.
By working through the problems in this chapter, readers will gain a deeper un-
derstanding of how density matrices connect the microscopic wavefunction to exper-
imentally accessible quantities, providing a versatile framework for studying non-
interacting and interacting systems alike.

PROBLEM 11.1
Derivatives of the one-body density matrix at r = 0.

One way to calculate the one-body density matrix of a homogeneous system is

through the momentum distribution of the system:
Z
ν ~
ρ(r) = d~k n(k) eik·~r ,
(2π)3

where n(k) is the momentum distribution, independent of the spin orientation, and
normalized such that
Z
ν
ρ0 = d~k n(k) ,
(2π)3

where ρ0 is the density of the homogeneous system. Notice that the diagonal of
the density matrix fulfills that ρ(r = 0) = ρ0 .

(a) Do the series expansion of ρ(r) around r = 0 and show that all odd derivatives
of ρ(r) at the origin are zero, and that

d 2 ρ(r) ν 1
Z
2
=− d~k k2 n(k) .
dr (2π)3 3

(b) Show that the kinetic energy per particle can be related to the second derivative
of the one-body density matrix at the origin:

T̂ 3 h̄2 1 d 2 ρ(r)
=− .
N 2 m ρ0 dr2 r=0

123

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 124 — #140

i i

124 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

(c) Consider a free Fermi sea of fermions of spin 1/2, whose one-body density
matrix is given by ρ(r) = ρ0 l(kF r), where kF is the Fermi momentum and
l(kF r) is the Slater function , defined as

3 j1 (z)
l(z) = ,
z

where j1 (z) is the spherical Bessel function of order one:

sin(z) − z cos(z)
j1 (z) = .
z2

Do the series expansion of ρ(r) around r = 0 and check the previous properties
for a free Fermi sea with spin degeneracy 2 and n(k) = θ (kF − k).

(a) To perform the series expansion of ρ(r) we start from its definition, as the Fourier
transform of the momentum distribution. Notice that here we consider that the
system is isotropic and therefore n(k) depends only on the modulus of ~k. On the
other hand both spin orientations, up and down with respect to the spin quantiza-
tion axis, have the same n(k). Then ρ(r) is given by
Z
ν ~
ρ(r) = d~k n(k) eik·~r .
(2π)3

Now we perform the angular integrations and check that actually ρ(r) depends
only on the modulus of~r:

ν
Z∞ 2π Z 1 Z
~
ρ(r) = 3
dk k2 dφ dxe−ik·~r n(k)
(2π) 0 0 −1
Z ∞ Z 1
ν 2
= 2π dk k n(k) dx (cos(krx) + i sin(krx))
(2π)3 0 −1
1
sin(krx)
Z ∞
ν
= 2π dk k2 n(k)
(2π)3 0 kr −1
sin(kr)
Z
ν
= 4π dk k2 n(k) .
(2π)3 kr

Now we take into account the Taylor expansion of sin(kr)/kr,

sin(kr) (kr)2 (kr)4

' 1− + −··· .
kr 3! 5!

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 125 — #141

i i

Density matrices 125

Therefore,
(kr)2 (kr)4
Z
ν
ρ(r) = 4π dkk2 n(k) 1 − + −···
(2π)3 3! 5!
Z Z 2 Z 4
ν 2 4 r 6 r
= 4π dkk n(k) − dkk n(k) + dkk n(k) +···
(2π)3 3! 5!
r2
4
r
Z Z Z
ν ~ ν ~ 2 ν ~ 6
= d kn(k) − d kk n(k) + d kk n(k)
(2π)3 (2π)3 3! (2π)3 5!
+··· .

Then, we should compare the previous expansion with the Taylor expansion of
ρ(r):

dρ 1 d2ρ 1 d3ρ
ρ(r) = ρ(0) + r + r2 + r3 + · · · ,
dr r=0 2! dr2 r=0 3! dr3 r=0

and by equating the coefficients we can identify the derivatives of ρ(r). The first
thing to realize is that all odd derivatives of ρ(r) for r = 0 are zero. The second
thing we can conclude is :
Z
ν
ρ(0) = 4π dk k2 n(k) = ρ0 ,
(2π)3
and for the second derivative:
1 d2ρ

1
Z
ν ~k k2 n(k) .
=− d
2 dr2 r=0 3! (2π)3

Therefore,
d2ρ 1 ν
Z
=− d~k k2 n(k) .
dr2 r=0 3 (2π)3
(b) Then this second derivative could be related with the kinetic energy per particle:

T̂ ν 1 h̄2 k2 3 h̄2 1 d 2 ρ
Z
= d~k n(k) = − ,
N (2π)3 ρ0 2m 2 m ρ0 dr2 r=0

which is valid for any homogeneous Fermi system.

ρ(r) = ρ0 l(kF r) .

The Slater function, l(kF r), can be developed in powers of kF r by using the
expansion of the spherical Bessel function j1 (z),
!
1 2 1 4
z 2z 4z z z3 z5
j1 (z) = 1− + −··· = − + −··· .
3 15 257 3 30 840

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 126 — #142

i i

126 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Doing this we obtain:

kF2 2 kF4 4

ρ(r) = ρ0 1 − r + r −··· ,
10 280
and we can immediately identify

dρ(r) d2ρ kF2

ρ(0) = ρ0 , = 0, = −ρ0 .
dr r=0 d2r r=0 5
Therefore,

T̂ ν 1 h̄2 k2 3 h̄2 1 d 2 ρ 3 h̄2 kF2

Z
= d~k n(k) = − = ,
N (2π)3 ρ0 2m 2 m ρ0 dr2 r=0 5 2m
which is nothing other than the kinetic energy per particle of the free Fermi sea.

PROBLEM 11.2
Sequential condition of the three-body distribution function.

The three-body distribution function of a many-body system is defined as,

N(N − 1)(N − 2)
Z
g3 (~r1 ,~r2 ,~r3 ) = d~r4 . . . d~rN Ψ∗ (~r1 , . . . ,~rN )Ψ(~r1 , . . . ,~rN ) ,
ρ3

where Ψ(~r1 , . . . ,~rN ) is a unit-normalized pure state. For an infinite, homogeneous

system, g3 depends only on the distances between the three particles r12 , r13 , r23 .
Usually one writes g3 (~r12 ,~r13 ).
In the case of the free Fermi sea we have,

1 2 1 1
g3 (~r12 ,~r13 ) =1 − l (kF r12 ) − l 2 (kF r13 ) − l 2 (kF r23 )
ν ν ν
2
+ l(kF r12 )l(kF r13 )l(kF r23 ) ,
ν
where ν is the spin degeneracy.

(a) Derive the sequential condition

Z
ρ d~r3 g3 (~r12 ,~r13 ) = (N − 2) g2 (r12 ) ,

where g2 (r12 ) is the two-body distribution function .

(b) Check the validity of this relation for the free Fermi sea.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 127 — #143

i i

Density matrices 127

(a) To derive the sequential condition we should use the definition of the n−body
distribution functions. Actually this property is valid both for Bose and Fermi
systems.

N(N − 1)(N − 2)
Z Z
ρ g3 (~r12 ,~r13 )d~r3 =ρ d~r3 d~r4 . . . d~rN Ψ∗ (~r1 , . . . ,~rN )Ψ(~r1 , . . . ,~rN )
ρ3
N(N − 1)
Z
=(N − 2) d~r3 d~r4 . . . d~rN Ψ∗ (~r1 , . . . ,~rN )Ψ(~r1 , . . . ,~rN )
ρ2
=(N − 2)g2 (~r1 ,~r2 ) = (N − 2)g2 (r12 ) .

(b) To check the validity of the sequential condition for the g3 (~r12 ,~r13 ) of the free
Fermi sea we need several properties of the Slater functions,
Z
ν
d~r l 2 (kF r) = ,
ρ

that was derived in Problem 8.1.

We still need another relation,
Z Z
ρ d~r3 l(kF r13 ) l(kF r23 ) = ρ d~r3 l(kF r13 ) l(kF r32 ) = ν l(kF r12 ) .

To derive this relation one should use the explicit definition of the Slater function,
Z Z Z
ν ~
ρ d~r3 l(kF r13 ) l(kF r32 ) =ρ d~r3 d~k eik·(~r3 −~r1 )
(2π)3 ρ |~k|≤kF
Z
ν ~0
× d~k0 eik ·(~r2 −~r3 ) .
(2π)3 ρ |~k0 |≤kF

Now we change the order of integration,

(ν)2 −i~k·~r1 i~k0 ·~r2

Z Z Z
ρ d~r3 l(kF r13 )l(kF r32 ) =ρ d~k d~k0 e e
|~k|≤kF |~k0 |≤kF ρ 2 (2π)3
1
Z
~ ~0
× d~r3 ei~r3 ·(k−k )
(2π)3
(ν)2
Z Z
~ ~0
=ρ d~k d~k0 e−ik·~r1 eik ·~r2 δ (~k −~k0 )
|~k|≤kF |~k0 |≤kF ρ 2 (2π)3
(ν)2
Z
~
=ρ eik·(~r2 −~r1 ) = ν l(kF r12 ) .
|~k|≤kF ρ 2 (2π)3

Now we can do the explicit calculation for the three-body distribution function

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 128 — #144

i i

128 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

of the free Fermi sea gFS

3 (~r12 ,~r13 ),

1 1 1
Z
ρ d~r3 1 − l 2 (kF r12 ) − l 2 (kF r13 ) − l 2 (kF r23 )
ν ν ν

2
+ l(kF r12 )l(kF r13 )l(kF r23 )
ν

1 2
Z Z
ρ ρ
=N 1 − l (kF r12 ) − d~r3 l 2 (kF r13 ) − d~r3 l 2 (kF r23 )
ν ν ν
2
Z
+ 2 l(kF r12 )ρ l(kF r13 )l(kF r23 )d~r3
ν
ρν ρν 2 l 2 (kF r)
=NgFS2 (r12 ) − − + 2 νl(kF r12 )l(kF r12 ) = NgFS 2 (r12 ) − 2 + 2
νρ νρ ν ν
=NgFS FS FS
2 (r12 ) − 2 g2 (r12 ) = (N − 2)g2 (r12 ) .

PROBLEM 11.3
Properties of the two-body density matrix.

The two-body density matrix of an N-body system is defined as

Z
ρ2 (~r1 ,~r2 ; ~r1 0 , ~r2 0 ) = N(N − 1) d~r3 ...d~rN Ψ∗ (~r1 , ...,~rN ) Ψ(~r1 0 ,~r2 0 ,~r3 , ...,~rN ) ,

where Ψ(~r1 , . . . ,~rN ) is a unit-normalized pure state, and where we have omitted
the spin degrees of freedom. It is also common to define the semidiagonal:
ρ2sd (~r1 ,~r2 ; ~r1 0 ) = ρ2 (~r1 ,~r2 ; ~r1 0 ,~r2 ) .
In the particular case of the homogeneous free Fermi sea, the two-body density
matrix is given by

l(kF r120 ) l(kF r210 )
ρ2 (~r1 ,~r2 ; ~r1 0 ,~r2 0 ) = ρ 2 l(kF r110 ) l(kF r220 ) − ,
ν
where l(z) is the Slater function, ν is the spin degeneracy and rab0 = |~ra − ~rb 0 |.
Notice that the spin degrees of freedom have been summed up.
Demonstrate the following properties:
(a)
ρ2 (~r1 ,~r2 ;~r1 ,~r2 ) = ρ 2 g(r12 ) .
(b) Z
ρ2sd (~r1 ,~r2 ; ~r1 0 ) d~r2 = (N − 1) ρ1 (~r1 ,~r1 0 ) ,
which in the case of a homogeneous system depends only on r110 , i.e.,
ρ1 (~r1 , ~r1 0 ) = ρ1 (r110 ) .

In both cases, check the properties for the free Fermi sea.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 129 — #145

i i

Density matrices 129

(a) Taking into account the definition of g(r12 ) for a homogeneous system,

N(N − 1)
Z
g(r12 ) = d~r3 . . . d~rN Ψ∗ (~r1 , . . .~rN ) Ψ(~r1 , . . .~rN ) ,
ρ2
it is immediate to check that
Z
ρ2 (~r1 ,~r2 ;~r1 ,~r2 ) = N(N − 1) d~r3 ...d~rN Ψ∗ (~r1 , ...~rN ) Ψ(~r1 , ...~rN ) = ρ 2 g(r12 ) .

For the free Fermi sea we have:

2 1
ρ2 (~r1 ,~r2 ;~r1 ,~r2 ) =ρ l(0)l(0) − l(kF r12 )l(kF r12 )
ν
2

2 l (kF r12 )
=ρ 1 − = ρ 2 gFS (r12 ) .
ν

(b) For the second property,

Z Z Z
ρ2 (~r1 ,~r2 ;~r1 0 ,~r2 )d~r2 =N(N − 1) d~r2 d~r3 . . . d~rN Ψ∗ (~r1 , . . .~rN )Ψ(~r1 0 , . . .~rN )
=(N − 1)ρ1 (~r1 ,~r1 0 ) ,

where Z
ρ1 (~r1 ,~r1 0 ) = N d~r2 . . . d~rN Ψ∗ (~r1 , . . .~rN )Ψ(~r1 0 , . . . ,~rN ) ,

which in the case of a homogeneous system depends only on r110 = |~r1 − ~r1 0 |.
In the case of the free Fermi sea,

1
sd
ρ2,FS (~r1 ,~r2 ;~r1 0 ) = ρ 2 l(kF r110 ) − l(kF r12 ) l(kF r10 2 ) .
ν

The first term does not depend on~r2 and therefore can be immediately integrated,
Z
ρ2 l(kF r110 ) d~r2 = ρ 2 Ω l(kF r110 ) = N ρ l(kF r110 ) ,

while for the second term we can use that

Z
ρ2 d~r2 l(kF r12 ) l(kF r10 2 ) d~r2 = ρνl(kF r110 ) ,

and therefore,
l(kF r110 )
Z
sd
ρ2,FS (~r1 ,~r2 ;~r1 0 ) d~r2 =N ρ l(kF r110 ) − ρ ν
ν
=(N − 1) ρ l(kF r110 ) = (N − 1) ρ1,FS (r110 ) .

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 130 — #146

i i

130 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 11.4
One-body density matrix in a 2D non-interacting Bose-Hubbard gas

The non-interacting Bose-Hubbard Hamiltonian is defined as

Ĥ = −J ∑ â†j âk + â†k â j ,

h j,ki

where âk (â†k ) is the annihilation (creation) operator for site k and h j, ki represents
that the sum is only performed over neighboring indices. For a 2D system with
square lattice of spacing d and size Ns × Ns , with periodic boundary conditions
(âNs +1 = â1 ), filled with N bosons:

(a) Find the normalization coefficient for the following transformation:

â j = C(Ns ) ∑ ∑ ei~r j ·~q b̂qx qy ,

qx qy

where ~r j = mx d~ux + my d~uy is the position of site j (mx , my = 1, 2, . . . , Ns ),

for b̂qx qy to be the annihilation operator of a particle with momentum ~q =
qx~ux + qy~uy which fulfills the following commutation relations:
h i
b̂qx qy , b̂†q0 q0 = δqx q0x δqy q0y . (11.4.1)
x y

(b) Diagonalize the Hamiltonian using the previous transformation.

1D E
ρ jk = Ψ0 â†j âk Ψ0 .
N

(a) Let us write the commutation relations in the positions space to check what is

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 131 — #147

i i

Density matrices 131

needed for C(Ns ) to fulfill.


h i 0
â j , â† =|C(Ns )|2 ∑ ∑ ei~r j ·~q b̂qx qy ∑ ∑ e−i~rk ·~q b̂†0
k qx q0y
qx qy q0x q0y

·~q0
− ∑ ∑ e−i~rk b̂†q0 q0 ∑ ∑ ei~r j ·~q b̂qx qy 
x y
q0x q0y qx qy
0

=|C(Ns )|2 ∑ ∑ ∑ ∑ ei (~r j ·~q−~rk ·~q ) b̂qx qy b̂†q0 q0 − b̂†q0 q0 b̂qx qy
x y x y
qx qy q0x q0y
0
h i
=|C(Ns )|2 ∑ ∑ ∑ ∑ ei (~r j ·~q−~rk ·~q ) b̂qx qy , b̂†q0 q0 .
x y
qx qy q0x q0y

Now, using the relation in Eq. (11.4.1) we obtain:

h i 0
â j , â†k =|C(Ns )|2 ∑ ∑ ∑ ∑ ei (~r j ·~q−~rk ·~q ) δqx q0x δqy q0y
qx qy q0x q0y

=|C(Ns )|2 ∑ ∑ ei (~r j −~rk )~q = |C(Ns )|2 ∑ ∑ ei (r jx −rkx )qx ei (r jy −rky )qy .
qx qy qx qy

Since we know that the momentum values in the first Brillouin zone are of the
form
2πkx 2πky
~q = ~ux + ~uy , kx , ky = −bNs /2c, . . . , −1, 0, 1, . . . , dNs /2e
dNs dNs
we are able to perform the summations using the following property of complex
sums: ∑Mk=0 e
i2πk/M = Mδ . Then, ignoring a global phase that will not affect the
k
system, we can define the normalization value:
h i 1 1
â j , â†k = |C(Ns )|2 Ns2 δ jk =⇒ |C(Ns )|2 = 2 =⇒ C(Ns ) = .
Ns Ns

(b) First of all, let us write the Hamiltonian using the fact that the sum is only per-
formed over first neighbors and we have a square lattice:
Ns Ns
Ĥ = −J ∑ ∑ â†mx +1,my âmx ,my + â†mx ,my +1 âmx ,my + h. c. ,
mx =1 my =1

where h.c. stands for the Hermitian conjugate of the first terms and the creation
and annihilation operators now are labeled following the spatial integer coeffi-
cients mx , my that identify each lattice point.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 132 — #148

i i

132 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

To diagonalize the Hamiltonian, we substitute the annihilation and creation op-

erators by their momentum series:
1 2π 0 0 2π
Ĥ = − J ei Ns (mx (kx −kx )+my (ky −ky )) e−i Ns kx b̂†kx ky b̂kx0 ky0
Ns2 m∑
x ,my k
∑
0 0
x ,kx ,ky ,ky

i 2π (mx (kx0 −kx )+my (ky0 −ky )) 2π

+e Ns e−i Ns ky b̂†kx ky b̂kx0 ky0
1 2π 0 0 2π 0
−J ei Ns (mx (kx −kx )+my (ky −ky )) ei Ns kx b̂†kx ky b̂kx0 ky0
Ns2 m∑
x ,my k
∑
0 0
x ,kx ,ky ,ky
2π 0 0 2π 0

+ei Ns (mx (kx −kx )+my (ky −ky )) ei Ns ky b̂†kx ky b̂kx0 ky0
J 2π 0 0
2π 2π
= − 2 ∑ ei Ns (mx (kx −kx )+my (ky −ky )) e−i Ns kx + e−i Ns ky
Ns mx ,my
kx ,kx0 ,ky ,ky0
2π 0 2π 0

+ei Ns kx + ei Ns ky b̂†kx ky b̂kx0 ky0 .

Now, performing the summation over the spatial indices mx and my ,

J 2π 2π 2π 0 2π 0

Ĥ = − ∑ Ns2 δkx kx0 δky ky0 e−i Ns kx + e−i Ns ky + ei Ns kx + ei Ns ky b̂†kx ky b̂kx0 ky0
Ns2 k 0 0
x ,kx ,ky ,ky

2πkx 2πky
= − J ∑ 2 cos + cos b̂†kx ky b̂kx ky
kx ,ky Ns Ns

2πkx 2πky
= − J ∑ 2 cos + cos m̂kx ky ,
kx ,ky Ns Ns

where m̂kx ky is the number operator in momentum space. Then this is an already
diagonal Hamiltonian, whose ground state for a system with N bosons is the state
with all bosons in the state kx = 0, ky = 0:

1 † N
|Ψ0 i = √ b̂ |vaci .
N! 00

(c) First, note the following relations:

† †
â . . . â}† â
| â {z . . . â} = n̂ (n̂ − 1) (n̂ − 2) · · · (n̂ − (l − 1)) ,
| â{z (11.4.2)
l times l times
† † †
â . . . â} â
| â{z . . . â} = (n̂ + 1) (n̂ + 2) · · · (n̂ + l) .
| â {z (11.4.3)
l times l times

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 133 — #149

i i

Density matrices 133

Now we compute the one-body density matrix elements ρ jk :

1 N N
ρ jk = hvac| b̂00 a†j ak b̂†00 |vaci
NN!
1 N
i~rk ·~q0
N −i~r j ·~q †
= hvac| b̂ 00 ∑ ∑ e b̂q q ∑ ∑ e b̂ q0 q0 b̂†00 |vaci
NN!Ns2 qx qy
x y
q0x q0y
x y

 
1 N
hvac| b̂00 b̂†00 + ∑ e−i~r j ·~q b̂†qx qy 

=

NN!Ns2 qx ,qy
qx ,qy 6=0,0
 
 0
 N
× b̂00 + ei~rk ·~q b̂q0x q0y  b̂†00 |vaci .
 
∑
0 0
qx ,qy
 
q0x ,q0y 6=0,0

Taking into account that the terms under the sum commute with the operators that
act on the vacuum state, it is clear that they do not contribute to this computation.
Then, the expression reduces to
1 N † N
ρ jk = hvac| b̂ 00 b̂00 b̂ 00 b̂†00 |vaci
NN!Ns2
1 N+1 † N+1 N † N
= hvac| b̂ 00 b̂ 00 − b̂ 00 b̂00 |vaci ,
NN!Ns2

and applying relation Eq. (11.4.3) we have:

1
ρ jk = hvac| (m̂00 + 1) · · · (m̂00 + N + 1) − (m̂00 + 1) · · · (m̂00 + N) |vaci
NN!Ns2
(N + 1)! − N! N!(N + 1 − 1) 1
= 2
= 2
= 2.
NN!Ns NN!Ns Ns

The general expression is found to be independent of the indices j, k. This is a

feature that highlights the non-decaying correlations of the ground state between
different sites.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 135 — #151

i i

Superfluidity
12
Superfluidity is one of the most striking manifestations of quantum mechanics at
macroscopic scales, characterized by frictionless flow and the emergence of collec-
tive quantum phenomena. The theoretical foundation of superfluidity in fermionic
systems was revolutionized by the Bardeen-Cooper-Schrieffer (BCS) theory , de-
veloped in 1957 to explain superconductivity in metals. This groundbreaking work
provided a microscopic description of how attractive interactions between electrons
near the Fermi surface can lead to the formation of Cooper pairs—correlated pairs
of fermions that condense into a coherent quantum state. The BCS theory not only
explained superconductivity but also laid the groundwork for understanding super-
fluidity in systems like liquid helium-3 and ultracold fermionic gases, becoming a
cornerstone of condensed matter physics.
This chapter explores the BCS state through a series of problems designed to
introduce the reader to the BCS theory. The BCS wavefunction captures the essence
of superfluidity by describing a macroscopic quantum state with off-diagonal long-
range order. Readers will engage with the mathematical structure of the BCS state
and calculate some properties. Through this exploration, the chapter underscores the
profound impact of the BCS framework on both theoretical physics and technologi-
cal advancements.

135

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 136 — #152

i i

136 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

PROBLEM 12.1
The Bardeen-Cooper-Schrieffer (BCS) ansatz

The celebrated BCS ansatz reads

|BCSi = ∏ uk + vk â†k â†−k |vaci , uk , vk ∈ C ,
k>0

where k denotes all quantum numbers necessary to describe a state, and the minus
sign in front means time-reversed. The notation k > 0 in the product refers to the
positive eigenvalue of time reversal.

(a) Derive the normalization equations for uk and vk , and show that |uk |2 + |vk |2 =
1 ∀k is a valid choice. Use this condition for the next sections.

(b) Compute the average particle number of a BCS state in terms of the constants
uk , vk .
2
(c) Compute the fluctuations in the particle number operator , ∆N̂ .

(a) We can start by imposing that the BCS state have unit norm,
1 = hBCS|BCSi .
Then, using the definition of the BCS state:

1 = hvac| ∏ (uk0 + vk0 â−k0 âk0 ) ∏ uk + vk â†k â†−k |vaci .
k0 >0 k>0

We can now bring together the pairs of factors labeled the same, i.e., k0 = k,
and write the equation above as a single product over k. To see this, notice that
factors within each product commute. Furthermore, the k-th factor in the right
product also commutes with all factors in the left product except for the one with
k0 = k, and therefore we can bring these two factors together, respecting their
initial order:

1 = hvac | ∏ (uk + vk â−k âk ) uk + vk â†k â†−k | vaci
k>0
 

= hvac| ∏ |uk |2 + uk vk â†k â†−k + vk uk â−k âk +|vk |2 â−k âk â†k â†−k  |vaci .
 
k>0 | {z } | {z } | {z }
(I) (II) (III)

Here, the term (I) is zero because h0| â†k = 0 ∀k. The term (II) is also null, since
âk |0i = 0 ∀k. The braced factor (III) is the identity. Then,

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 137 — #153

i i

Superfluidity 137

1 = ∏ |uk |2 + |vk |2 .

k>0

One way to fulfill this condition is to have:

|uk |2 + |vk |2 = 1 ∀k .

(b) We start by inserting the definition of the particle number operator N̂ in the ex-
pectation value:

Now we cannot bring equally-labeled terms together, since the factors â†q âq will
not commute with the factors in the products when k, k0 = q. What we can then
do is the following:

hN̂iBCS = ∑ hvac| ∏ (uk0 + vk0 â−k0 âk0 ) (uq + vq â−q âq ) â†q âq uq + vq â†q â†−q
q k0 6=q

× ∏ uk + vk â†k â†−k |vaci
k6=q
"

= ∑ hvac | ∏ |uq |2 (uk0 + vk0 â−k0 âk0 ) â†q âq ∏ uk + vk â†k â†−k | vaci
q k0 6=q k6=q

+ hvac | ∏ uq vq (uk0 + vk0 â−k0 âk0 ) â†q âq â†q â†−q u + v â† †
∏ k k k −k | vaci
â
k0 6=q k6=q

+ hvac | ∏ vq uq (uk0 + vk0 â−k0 âk0 ) â−q âq â†q âq ∏ uk + vk â†k â†−k | vaci
k0 6=q k6=q

+ hvac| ∏ |vq | 2
(uk0 + vk0 â−k0 âk0 ) â−q âq â†q âq â†q â†−q
k0 6=q
#

× ∏ uk + vk â†k â†−k |vaci .
k6=q

The first term is immediately zero, since the operators âq commute with the
whole product to their right, and they will then annihilate the kets |0i. With simi-
lar arguments, one can see that the same happens for the next two terms, leaving
us only with the fourth term, which we can rewrite as a single product. Ignor-
ing the crossed terms (those with factors ūk vk and v̄k uk ) because they also vanish

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 138 — #154

i i

138 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

following the arguments of the previous question, the result is:

= ∑ |vq |2 ∏ |uk |2 hvac |â−q âq â†q âq â†q â†−q | vaci
q k6=q


+ |vk |2 hvac |â−k âk â−q âq â†q âq â†q â†−q â†k â†−k | vaci


|vq |2
= ∑ |vq |2 ∏ |uk |2 + |vk |2 = ∑ = ∑ |vq |2 ,

2 2
q k6=q q |uq | + |vq | q

where we have used Wick’s theorem, and the normalization condition |uk |2 +
|vk |2 = 1 ∀k has been used in the last step. It is interesting to note that the average
number of particles in a BCS ansatz depends only on the {vk }k .1
(c) The fluctuation in the number of particles, ∆N̂, is defined as the variance of N̂,
2
∆N̂ = N̂ 2 − N̂ . The second term is trivially
2
N̂ = ∑ |vq |2 |vq0 |2 ,
q,q0

whereas the first term requires some more work. By definition,

N̂ 2 = ∑ â†q0 âq0 â†q âq ,

q,q0

and therefore we can write:

† † †
N̂ 2 BCS
= ∑ hvac | ∏ (uk
†
0 + vk0 â−k0 âk0 ) â 0 âq0 âq âq ∏ uk + vk â a
q k −k | vaci .
q,q0 k0 >0 k>0

Here, in order to be able to use the trick of pulling out the factors with k, k0 = q, q0 ,
we must first differentiate the cases when q0 = q and q0 6= q:
!
∑0 = ∑ [q0 = q] + ∑ .
q,q q q0 6=q

1 In the literature, this result is often expressed as a sum over q > 0, and with an extra factor of 2 to

exploit the time-reversal symmetry of the BCS ansatz.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 139 — #155

i i

Superfluidity 139

Let us start with the terms where q0 = q:

∑ hvac | ∏ (uk0 + vk0 â−k0 âk0 ) â†q âq â†q âq ∏ uk + vk â†k a†−k | vaci
q 0 k >0 k>0

= ∑ hvac| ∏ (uk0 + vk0 â−k0 âk0 ) (uq + vq â−q âq ) â†q âq â†q âq uq + vq â†q a†−q
q k0 6=q

× ∏ uk + vk â†k a†−k |vaci
k6=q

= ∑ |vq |2 ∏ |uk |2 hvac |â−q âq â†q âq â†q âq â†q â†−q | vaci
q k6=q


+ |vk |2 hvac |â−k âk â−q âq â†q âq â†q âq â†q â†−q â†k â†−k | vaci



|vq |2
= ∑ |vq |2 ∏ |uk |2 + |vk |2 = ∑

.
q k6=q q |uq |2 + |vq |2

Let us now compute the terms where q0 6= q:

∑∑ hvac | ∏ (uk0 + vk0 â−k0 âk0 ) â†q0 âq0 â†q âq ∏ uk + vk â†k â†−k | vaci
0 q q 6=q0 k >0 k>0

=∑ ∑ hvac| ∏ (uk0 + vk0 â−k0 âk0 ) (uq + vq â−q âq ) â†q0 âq0 â†q âq
0
q q 6=q 0 k 6=q

× uq + vq â†q â†−q ∏ uk + vk â†k â†−k |vaci
k6=q

=∑ ∑ hvac| ∏ |vq |2 (uk0 + vk0 â−k0 âk0 ) â−q âq â†q0 âq0 â†q âq â†q â†−q
0
q q 6=q 0 k 6=q

× ∏ uk + vk â†k â†−k |vaci .
k6=q

Here, notice that for every q we have q0 6= q, and also the products do not con-
tain factors labelled by q. This means that the factors âq , â−q and their adjoints
must be contracted with themselves, and there is only one such allowed Wick

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 140 — #156

i i

140 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

contraction. This leaves us with:

=∑ ∑ |vq |2 |vq0 |2 hvac| ∏ (uk + vk â−k âk ) â−q0 âq0 â†q0 âq0 â†q0 â†−q0
0
q q 6=q 0 k6=q,q

× uk + vk â†k â†−k |vaci
|vq |2 |vq0 |2
|vq |2 |vq0 |2 ∏ |uk |2 + |vk |2 = ∑

=∑ ∑ ∑ ,
0
q q 6=q 0 k6=q,q q 0 |u |2 + |vq |2 |uq0 |2 + |vq0 |2
q 6=q q

where we have used contraction lines to indicate the only possible contraction
of the operators âq0 , â−q0 and their adjoints. Bringing both contributions together
(q0 = q and q0 6= q) yields:
!
2 |vq |2 |vq |2 |vq0 |2
N̂ BCS = ∑ + .
q |uq |2 + |vq |2 q∑ 2 2 2
0 6=q |uq | + |vq | |uq0 | + |vq0 |
2

Finally,

2 |vq |2 |vq |2 |vq0 |2

∆N̂ = N̂ 2 − N̂ =∑ + ∑
q |uq |2 + |vq |2 q0 6=q |uq |2 + |vq |2 |uq0 |2 + |vq0 |2
!
|vq |2 |vq0 |2
−∑
q0
|uq |2 + |vq |2 |uq0 |2 + |vq0 |2

|vq |2 |vq |2 |vq0 |2

=∑ +
q |uq |2 + |vq |2 q∑ 2 2 2
0 6=q |uq | + |vq | |uq0 | + |vq0 |
2
!
|vq |2 |vq0 |2 |vq |4
− ∑ −
|u |2 + |vq |2 |uq0 |2 + |vq0 |2 (|uq |2 + |vq |2 )2
q0 6=q q

|vq |2 |vq |2

=∑ 2 2
1− = ∑ |vq |2 |uq |2 ,
q |uq | + |vq | |uq |2 + |vq |2 q

where we have used |uk |2 + |vk |2 = 1 ∀k in the last step.

Notice that we could have used the time-reversal symmetry of the BCS ansatz in
our favor: since particles are created in pairs with quantum numbers (k, −k), the
number of particles in a state k will be the same as that in the state −k. Therefore,

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 141 — #157

i i

Superfluidity 141

we could run the sum exclusively over states of positive time-reversal parity and
account for the missing factor of 2,

N̂ = 2 ∑ â†q âq ,
q>0

and also
N̂ 2 = 4 ∑0 â†q0 âq0 â†q âq ,
q,q >0

which are the usual expressions found in the literature.

PROBLEM 12.2
Particle number projection

In a previous problem we have seen the BCS ansatz which is manifestly not an
eigenstate of the particle number operator, N̂:

|BCSi = ∏ uk + vk â†k â†−k |vaci , uk , vk ∈ C ,
k>0

where k denotes all quantum numbers necessary to describe a state, and the minus
sign in front means time-reversed. In practice, however, these states are still used as
ansätze for many-body wavefunctions of definite particle number. This is possible
because we can project onto eigenstates of N̂.

(a) Rewrite the BCS wavefunction as a sum over the eigenstates of N̂.

(b) The particle number projection operator reads

Z 2π
1
P̂N0 = dϕeiϕ (N̂−N0 ) ,
2π 0

where N0 is the desired eigenstate of N̂. Prove that this is indeed a projector
and that it actually works as intended, that is, that it projects any state into its
N0 component.

(a) First, we have to express the product of sums appearing in the BCS ansatz as a
sum of products. To do that, we can consider the general product of binary sums
N
∏ (ak + bk ) .
k=1

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 142 — #158

i i

142 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

In the cases N = 2 and N = 3 we can explicitly write:

2
∏ (ak + bk ) =a1 a2 (2 a’s, 0 b’s)
k=1
+ a1 b2 + a2 b1 (1 a, 1 b)
+ b1 b2 (0 a’s, 2 b’s)
3
∏ (ak + bk ) =a1 a2 a3 (3 a’s, 0 b’s)
k=1
+ a1 a2 b3 + a1 a3 b2 + a2 a3 b1 (2 a’s, 1 b)
+ a3 b1 b2 + a1 b2 b3 + a2 b1 b3 (1 a, 2 b’s)
+ b1 b2 b3 (0 a’s, 3 b’s) .
Every term has N factors, and the number of a’s is complementary to the number
of b’s, each ranging from 0 to N. Furthermore, the subindices of the factors can-
not repeat within the same term. We can express this fact as follows. Consider
the set ΩM ≡ {0, 1, 2, . . . , M}. Then:
!
M
∏ (ak + bk ) = ∑ ∏ ak ∏ bk , (12.2.1)
k=1 S⊆ΩM k∈Sc k∈S

where the complement of S is taken with respect to ΩM . In the case of the BCS
ansatz, M denotes the number of single-particle states available in the system.
Using Eq. (12.2.1) we can now rewrite the BCS ansatz for a system with M
available states as:
!
|BCSi = ∑ ∏c uk ∏ vk â†k â†−k |vaci
S⊆ΩM k∈S k∈S

= ∑ ∏ uk ∏ vk |k − ki
S⊆ΩM k∈Sc k∈S

= ∑ ∏ uk ∏ vk |N = 2|S|i ,
S⊆ΩM k∈Sc k∈S

where |S| is the cardinal number of the subset S. What we want is to have a sum
over N. To achieve this, we can write the sum over subsets as a sum over the
possible cardinal numbers of these subsets, and for each cardinal number, a sum
over the subsets with this cardinality:
|BCSi = ∑ ∑ ∏ uk ∏ vk |N = 2Ci .
C=0,1,...,M S⊆ΩM s.t. |S|=C k∈Sc k∈S

Finally, we can rewrite the above as:

!
|BCSi = ∑ ∑ ∏ uk ∏ vk |Ni
N=0,2,...,2M S⊆ΩM s.t. 2|S|=C k∈Sc k∈S

≡ ∑ aN |Ni ,
N=0,2,...,2M

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 143 — #159

i i

Superfluidity 143

or, more correctly, M

|BCSi = aN |Ni ,
N=0,2,...,2M

since, by construction of the Fock space, states of different particle numbers can-
not be entangled. Similarly, the product symbols used above should actually be
tensor products, since they relate states in the same Hilbert space (i.e., with an
equal number of particles).
(b) A projector P̂ in quantum mechanics has to be (I) idempotent (P̂2 = P̂), and (II)
self-adjoint (P̂† = P̂). Let us start by checking the second property.

1 2π
Z 2π
1
† Z
dϕe−iϕ (N̂ −N0 ) = dϕe−iϕ (N̂−N0 )
†
P̂N0 =
2π 0 2π 0
1 2π
Z
= dϕeiϕ (N̂−N0 ) = P̂N0 ,
2π 0

where we have used the fact that N̂ † = N̂, and also the fact that, in the basis
where N̂ is diagonal, P̂N0 = δN,N0 = δN0 ,N . Using this identification of P̂N0 with
the Kronecker delta, proving the first property, P̂2 = P̂, is immediate.
We should now check that it actually projects a state onto a subspace of states
with a definite number of particles. To do this, consider a general state which is
a superposition of states with different particle numbers,
∞
|ψi = ∑ aN |Ni .
N=0

Then, applying P̂N0 onto this state,

∞ Z 2π
1
N0
P̂ |ψi = ∑ aN dϕeiϕ (N̂−N0 ) |Ni .
N=0 2π 0

Since N̂ is diagonal in the basis of particle number eigenstates, N̂ |Ni = N |Ni,

we can immediately identify a Kronecker delta in the equation above:
∞ Z 2π
1
P̂N0 |ψi = ∑ aN dϕeiϕ(N−N0 ) |Ni = aN0 |N0 i .
N=0 2π 0
| {z }
δN,N0

These projectors are especially useful in many-body calculations where the

wavefunction is written as a superposition of definite particle number states, such
as the BCS wavefunctions (first part of this exercise). This is because, for most
physical systems, we want to work with a definite number of particles, and there-
fore we need to project onto subspaces with the concrete number of particles that
we want.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 145 — #161

i i

Mixed statistics
13
In Chapter 2 we saw that the Fock space is defined as:
∞
F= H (N) ,
M

N=0

where H (N) are Hilbert spaces with N particles. This construction is valid when
dealing with a single species of particles. There are times, however, when we have
to deal with systems of different types of particles, such as fermions and bosons. In
this case, the construction of the Fock space of bosons and fermions is:
" # " #
∞ ∞
F= Hb (N) H f (N) ,
M O M

N=0 N=0

where the subscripts b and f are for bosonic and fermionic Hilbert spaces, respec-
tively. This tells us that we can have mixed (entangled) states containing both bosons
and fermions, for instance. The same construction can be applied for systems with
more than two distinct particle species. In this chapter we provide a working example
of a system containing a mixture of one fermion and two bosons.

PROBLEM 13.1
Mixed statistics: a system of fermions and bosons

Consider a system consisting of one fully-polarized fermion and two spin-0

bosons interacting via the following Hamiltonian:

Ĥ = T̂ + V̂0 + V̂BF ,

where T̂ and V̂0 contain only one-body operators, and V̂BF is a two-body operator
which describes the interaction between the fermion and one boson.

(a) Write the general expression of Ĥ in terms of fermionic and bosonic operators.

(b) Consider V̂BF = 0 for now. We will suppose that there are only two single-
particle states available, both for fermions and for bosons, with associated
single-particle energies e0, f ,b , e1, f ,b fulfilling e0, f ,b < e1, f ,b and e1, f < e1,b .
Write the ground state of Ĥ in this case and compute its energy, E0 .

(c) Consider now V̂BF 6= 0. Assuming that the interaction allows the hopping of
the fermion to the excited state only when there is a boson on the lowest single-
particle level, the only non-zero two-body matrix elements are vi jαβ = v0010 =

145

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 146 — #162

i i

146 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

v0001 ∈ R. Write Ĥ using the basis where Ĥ0 = T̂ + V̂0 is diagonal, i.e., the
single-particle states

{|10; 20i , |01; 20i , |10; 11i , |01; 11i , |10; 02i , |01; 02i} ,

with energies
{E0 , E1 , E2 , E3 , E4 , E5 } ,
respectively, fulfilling the condition:

E0 < E1 < · · · < E5 .

Find the ground state. (Advice: set E0 = 0 to simplify the calculations). We

(a) Bosons and fermions are not identical particles (i.e., they are distinguishable),
and our Hamiltonian must reflect this. To start with, the single-particle states
of fermions and bosons need not be identical, and therefore we need to choose
different single-particle indices. We will choose Latin indices for bosons and
Greek indices for fermions, and we will use b̂, b̂† for bosonic operators, and ĉ, ĉ†
for fermionic operators. The one-body operators are:
f
T̂ = ∑ tibj b̂†i b̂ j + ∑ tαβ ĉ†α ĉβ ,
ij αβ
f
V̂0 = ∑ vbij b̂†i b̂ j + ∑ vαβ ĉ†α ĉβ ,
ij αβ

where we have used the superindex b to refer to bosonic matrix elements, and f
for the fermionic ones. The two-body potential reads

V̂BF = ∑ viα jβ b̂†i ĉ†α b̂ j ĉβ = ∑ vi jαβ b̂†i b̂ j ĉ†α ĉβ .

iα jβ i jαβ

Here we have used the fact that the canonical operators

i h of distinct particle species
h †
i h
† † †
i
always commute, b̂i , ĉ j = b̂i , ĉα = b̂i , ĉα = b̂i , ĉα = 0. Because of that,
notice that, contrary to the purely fermionic case, the ordering of the destruction
operators need not be reversed with respect to that of the indices (see chapter 2).
(b) We have that, both for fermions and for bosons, the single-particle energies of
the (single-particle) state |0i are lower than that of the state |1i. Besides, we have
that Ĥ0 has only one-body operators (no inter-particle interaction), and therefore
there is no mixing of single-particle states. With this information, the ground

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 147 — #163

i i

Mixed statistics 147

Figure 13.1 Left: potential energy well with two energy levels and two bosons in the single-
particle bosonic ground state. Right: similar to the left but for a single fermion. The tensor
product of the two wells represents the lowest energy level of the many-body state. Both the
bosonic and fermionic ground-state energies are set to 0, for convenience.

state must be:

1
|GSi = |0i f ⊗ (|0ib )2 := |10; 20i = √ ĉ†0 b̂†0 b̂†0 |vaci .
2
√
The factor 1/ 2 comes from the totally-symmetric character of the bosonic part
of the state. This information can be summarized in Fig. 13.1.
We will compute E0 by computing the following expectation value.
f f
hĤi = h10; 20| ∑(tibj + vbij )b̂†i b̂ j + ∑(tαβ + vαβ )ĉ†α ĉβ |10; 20i .
ij αβ

Using the rules of how creation and annihilation operators act on Fock states, the
result is:
√ f f
√
hĤi = ∑(tibj + vbij )( 2)2 δ j0 δi0 + ∑(tαβ + vαβ )( 1)2 δβ 0 δα0
ij αβ
b f f
=2(t00 + vb00 ) + (t00 + v00 ) = 2e0,b + e0, f .

As expected for a non-interacting system, the total energy is the sum of the single-
particle energies, so it could have been guessed without any computation.
(c) Let us start by computing the matrix elements of V̂BF . First, looking at the term
with vi jαβ = v0010 , the only non-zero matrix elements are:
1
h01; 20|V̂BF |10; 20i = v0010 hvac |b̂0 b̂0 b̂†0 b̂0 b̂†0 b̂†0 | vaci hvac |ĉ1 ĉ†1 ĉ0 ĉ†0 | vaci
2
1
= v0010 4 hvac |b̂0 b̂0 b̂†0 b̂0 b̂†0 b̂†0 | vaci hvac |ĉ1 ĉ†1 ĉ0 ĉ†0 | vaci
2
=2v0010 ,
which is the associated energy for the transition of the fermion between the
ground state and the excited state when there are two bosons in their single-
particle state.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 148 — #164

i i

148 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

=v0010 hvac |b̂1 b̂0 b̂†0 b̂0 b̂†0 b̂†1 | vaci hvac |ĉ1 ĉ†1 ĉ0 ĉ†0 | vaci
=v0010 ,

which is the associated energy for the transition of the fermion between the
ground state and the excited state when there is one boson in the single-particle
state. The value is half of the other element because there are fewer bosons al-
lowing the transition.
The matrix elements stemming from vi jαβ = v0001 are necessarily the adjoints of
those with vi jαβ = v0010 , as Ĥ must be Hermitian.
To diagonalize Ĥ it is convenient to choose the following basis ordering:

{|10; 02i , |01; 02i , |10; 20i , |01; 20i , |10; 11i , |01; 11i} .

Now, Ĥ is block-diagonal,
 
E4 0 0 0 0 0
 0 E5 0 0 0 0 
 
0 0 E0 2v0010 0 0 
Ĥ = 
0
,
 0 2v0010 E1 0 0 

0 0 0 0 E2 v0010 
0 0 0 0 v0010 E3

and we can observe that the interaction only couples 1 fermion transition, and
that the first two states are not affected by this interaction.
To diagonalise this matrix, we can use the fact that the determinant of a block-
diagonal matrix is the product of the determinants of the blocks:

E −λ 2v0010 E2 − λ v0010
det Ĥ − λ Iˆ =(E4 − λ )(E5 − λ ) 0

.
2v0010 E1 − λ v0010 E3 − λ

The 2 × 2 determinants are easily computed:

E0 − λ 2v0010
=(E0 − λ )(E1 − λ ) − 4v20010
2v0010 E1 − λ
=λ 2 − λ (E0 + E1 ) − 4v20010 + E0 E1 ,

E2 − λ v0010
=(E2 − λ )(E3 − λ ) − v20010
v0010 E3 − λ
=λ 2 − λ (E2 + E3 ) − v20010 + E2 E3 .

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 149 — #165

i i

Mixed statistics 149

The solutions of setting these polynomials to zero are:

1
q
2 2
λ± = E0 + E1 ± (E0 + E1 ) + (4v0010 ) − 4E0 E1
2

and
1
q
λ±0 = E2 + E3 ± (E2 + E3 )2 + v20010 − 4E2 E3
2
respectively. Then, the overall expression for det Ĥ − λ Iˆ is:

det Ĥ − λ Iˆ = − (E4 − λ ) (E5 − λ ) (λ+ − λ ) (λ− − λ ) λ+0 − λ λ−0 − λ .

Since we have E0 < E1 < · · · < E5 and we are looking for the ground state, from
the equation above, we deduce that the ground state must either have energy
E4 , λ− or λ−0 . If we rewrite E4 , λ− and λ−0 as a function of the single-particle
energies ekf ,b = tkf ,b + vkf ,b , setting e0f = eb0 = 0:

E4 =2eb1 ,

1
q
f f 2 2
λ− = e − (e1 ) + (4v0010 ) ,
2 1

1
q
0 f b f 2 2
λ− = e + 2e1 − (e1 ) + v0010 .
2 1

We see that λ− < h1f < 2eb1 = E4 , and λ−0 − λ− ≥ eb1 > 0. Therefore, the ground
state will have energy λ− .
Since the Hamiltonian is block-diagonal, what could have been a system of 6
coupled equations is now a set of 4 independent systems (one for each block).
We now need to find the eigenvector with this eigenvalue, for which we will work
only in the subspace associated to this eigenvalue, i.e., the subspace spanned by
{|10; 20i , |01; 20i}.

γ γE0 + 2δ v0010 γ
Ĥ = = λ− .
δ 2γv0010 + δ E1 δ
Using again the single-particle energies,
)
2δ v0010 = λ− γ
.
2γv0010 + δ e1f = λ− δ

If we now write δ (γ) using the first equation and insert it into the second equa-
tion, we obtain:
λ−2

f λ−
γ 2v0010 + e1 − = 0.
2v0010 2v0010

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 150 — #166

i i

150 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

The solution γ = 0 is not admissible as it leads to a null 8−dimensional vector,

which is the trivial solution. However, we can check that the condition in paren-
theses is always satisfied, and therefore there is no constraint on γ. The ground
state is therefore:

λ−
|GSi = |λ− i = γ |10; 20i + |01; 20i ,
2v0010

up to a normalization factor.
This result is somewhat expected: the ground state of the interacting system is
a superposition of the ground state of the non-interacting system and the state
with an excited fermion, since the interaction allows the fermionic |0i f → |1i f
single-particle transition (and the opposite one). In the limit v0010 → ±∞, it is
easy to see that |GSi = γ (|10; 20i ∓ |01; 20i). Physically, this means that if we
increase the interaction strength, the lowest-energy configuration will be an equal
superposition of |10; 20i and |01; 20i.
Finally, the limit v0010 → 0 is also interesting. One can see that the ground state
is then |GSi = |10; 20i. This was expected, since having no interaction should
return the same result as in the previous section, where V̂BF = 0 from the start.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 151 — #167

i i

Second-quantization
A formalism
The natural language to describe indistinguishable particles in quantum mechanics
is the so-called second quantization approach [9, 11, 3, 6, 5, 1]. Rather than working
in a Hilbert space with a fixed number of particles, H (N), in second quantization
we let go of the idea of a fixed number of particles, and form a Hilbert space that is
large enough to accommodate an undetermined number of particles. This is achieved
by employing the so-called Fock space,
∞
F= H (N) .
M

N=0

This Fock space contains a vacuum state, |0i, with no particles; a complete set of
single-particle states, |αi, with α = 1, 2, 3, . . . ; a complete set of two-particle states,
|αβ i; a complete set of three-particle states, |αβ γi; and so on. These complete sets
of many-particle states contain only vectors of the correct permutation symmetry,
which can either be fermionic or bosonic.
The key tools to exploit the extension into Fock space are the creation, â† , and de-
struction, â, operators. These act as mappings between many-particle Hilbert spaces
of different particle numbers,

â†α : H (N − 1) → H (N) ,
âα : H (N) → H (N − 1) .

Let us first discuss the creation operator, and we shall derive the properties of the
annihilation operator from it. We also restrict the discussion to fermions in the first
instance, although the extension to bosons is relatively straightforward and discussed
in detail in several literature sources [2, 15, 7]. Acting on top of the vacuum state,
with no particles, the operator â†α creates a particle in the state α,

â†α |0i = |αi . (A.0.1)

Acting on top of a one-particle state |β i, the same operator creates a two-particle

state,

â†α |β i = |αβ i . (A.0.2)

It is useful to demand this state to be antisymmetric from the outset. For this to occur,
we must have

â†α â†β |0i = |αβ i = − |β αi = −â†β â†α |0i .

151

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 152 — #168

i i

152 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

We can impose this condition through the anticommutation relation

â†α â†β + â†β â†α = 0 ⇒ {â†α , â†β } = 0 . (A.0.3)

Taking adjoints of this relation, one can easily prove that

{âα , âβ } = 0 . (A.0.4)

Adding a third particle to a two-body state immediately yields an anti-

symmetrized three-body state. It is important to note the ordering of the states on
the antisymmetrized ket. Every time we create a state, we do it on the left-hand side
of the ket:

â†α |β γi = |αβ γi . (A.0.5)

Some additional algebra immediately shows that fermionic creation and annihilation
operators also fulfill a fundamental anticommutation relation,

{âα , â†β } = δαβ . (A.0.6)

For bosons, the relevant relations between creation and destruction operators are
the following commutation relations:

[âα , â†β ] = δαβ , (A.0.7)

[âα , âβ ] = 0 , (A.0.8)
[â†α , â†β ] = 0 . (A.0.9)

We have seen that the creation and annihilation operators allow us to define prac-
tically many-body states. This in itself would not be useful if we could not also
transform operators into their Fock space counterparts. Let us first consider opera-
tors that act over one-body properties. Think, for instance, about the kinetic energy
of an N−body system of identical fermions (and hence with the same mass m),
N N
p̂2i
T̂ = ∑ = ∑ tˆi . (A.0.10)
i=1 2m i=1

The single-particle operators tî act on one particle at a time. Similar expressions hold
for other one-body properties like the total momentum, or the total external potential
energy of a system of N particles. We use tî as a generic one-body operator in the
following. Since it acts on individual particles, tî has the matrix elements

tαβ = (α|tˆi |β ) . (A.0.11)

A key point of this expression is that it does not depend on the particle index i. The
single-particle states |αi are predefined, and the operator tˆi is exactly the same for
each particle— so the corresponding single-particle matrix elements do not know to
which particle they need to be applied to.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 153 — #169

i i

Second-quantization formalism 153

One can construct a second quantization version of the very same operator. For
operators of a one-body nature, the expression reads:

T̂ = ∑ tαβ â†α âβ . (A.0.12)

αβ

Compared to the original expression, this operator does not explicitly depend on the
number of particles since it acts and lives in Fock space. It is only when T̂ is applied
to an antisymmetric N−body ket that the dependence over N is restored.
One-body operators are not all there are, though. In an interacting system, a key
two-body operator is the interaction operator, which provides the potential between
particles i and j,
1
V̂ = ∑ V̂i j = V̂i j . (A.0.13)
ij 2 i6∑
=j

The second-quantized version of this operator is given by the expression,

1
V̂ = ∑ vαβ γδ â†α â†β âδ âγ ,
2 αβ
(A.0.14)
δγ

where it is important to stress that the order of the last two destruction operators is
inverted with respect to the matrix element,

vαβ γδ = αβ |V̂i j |γδ . (A.0.15)

Here and in Eq. (A.0.11) we employ a parenthesis notation |·) in the matrix elements
obtained from specific single-particle states, without considering explicitly antisym-
metry. Fully symmetric or antisymmetric kets, like those in Eq. (A.0.2), are instead
represented by |·i.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 155 — #171

i i

Wick’s theorem
B
In the context of quantum many-body physics, one often encounters the need to
evaluate expectation values of operators and their matrix elements. In second quan-
tization, we denote the matrix element O~α ,~β of a generic N-body operator Ô as:
D E
O~α ,~β = ~α Ô ~β = α1 . . . αN Ô β1 . . . βN .

For the sake of this example, we introduce the vector notation |~α i = |α1 . . . αN i. Let
us now assume that Ô is of the form â†γ âδ for some indices γ, δ . The matrix element
then reads D E
O~α ,~β = 0 âαN . . . âα1 â†γ âδ â†β . . . â†β 0 .
1 N

We now have an expectation value on the vacuum of a product of creation and an-
nihilation operators. In general, in second quantization, we can always express any
matrix element, expectation value, or overlap between two states as a vacuum expec-
tation value of a product of operators.
In order to carry out actual calculations, one generally wishes to rearrange such
products of operators so that all creation operators lie to the left of all annihilation
operators. Whenever a product is thusly arranged, we shall say that it is in normal or-
der. Why is this a good idea? Precisely because if a product Â1 Â2 . . . ÂN is in normal
order form, then:
0 Â1 Â2 . . . ÂN 0 = 0 ,
where we have used upper-case letters to denote fully general operators, i.e., that can
be creation or annihilation operators. Let us see how to write any product of operators
as a sum of normal-ordered terms. Take as an example the following product of 3
operators:

âα â†β â†γ = −â†β âα + δαβ â†γ = δαβ â†γ − â†β −â†γ âα + δαβ

= â†β â†γ âα − δαβ â†γ − δαγ â†β .

The right-hand side in this last line is formed by a term, â†β â†γ âα , which is in normal
order, minus two additional terms involving Kronecker δ ’s on single-particle indices.
The process to arrive at normal order products of creation and annihilation opera-
tors becomes quite tedious as the number of operators in the product increases. Gian
Carlo Wick (1909 - 1992) suggested an easier way to evaluate such products in his
1950 article entitled Evaluation of the Collision Matrix [14]. Before we announce
this theorem, we need to introduce some definitions.
Definition (normal ordering operator). Consider the product of operators Π
that contains both creation and annihilation operators. The normal ordering operator

155

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 156 — #172

i i

156 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

N acting on Π is defined as:

N (Π) ≡ (±1)P Π(creation × annihilation) , (B.0.1)

where the + sign corresponds to the case where all operators in Π are bosonic, and
the − sign, to the case where they are all fermionic. P is the sign of the permutation
that takes all creation operators to the left of all annihilation operators. One can think
of the above equation as rearranging the original product Π in normal order by using
the fermionic (bosonic) anticommutation (commutation) relations, and then ignoring
the terms with Kronecker deltas.
Notice that the normal ordering operator outcome is unique up to the reordering
of creation (or annihilation) operators within themselves. In fact, let us assume that
a product is already in normal order, and now we wish to exchange two creation
(or annihilation) operators. If these operators are bosonic, they will commute, and
therefore we end up with a reordering of the same product. If they are fermionic,
since we are effectively ignoring the Kronecker deltas, we are introducing an extra
minus sign at most. However, this will increase the parity of the overall permutation
by one, and therefore the overall factor (−1)P will counteract the change of sign,
leaving us with the same value for the reordering of the initial product.
Definition (contraction). We define the contraction of two operators Â, B̂ as:

ÂB̂ ≡ ÂB̂ − N (ÂB̂) . (B.0.2)

Because N (ÂB̂) is only a rearrangement of ÂB̂, the contraction of two operators is

actually an (anti)commutation relation, and therefore a c−number.
The usefulness of this definition lies partly in the fact that, since the vacuum
expectation of a normal-ordered product vanishes, and since a contraction is just a
c−number, we have:
ÂB̂ = h0|ÂB̂|0i = 0 ÂB̂ 0 .
Definition. We define the contraction of two operators Â, B̂ which are separated,
within a product, by a chain of n operators, as:

ÂÂ1 . . . Ân B̂ ≡ (±1)n ÂB̂Â1 . . . Ân , (B.0.3)

where the + sign corresponds to the bosonic case, and the − sign, to the fermionic
case. n can also be seen as the number of operator transpositions which takes the
contracted pair to the left of the product.
Let us now see how these definitions can help us in rearranging operator products.
We start by looking at some simple products of fermionic operators and manually
expressing them in normal order. The product of two canonical operators is

âα â†β = −â†β âα + δαβ ,

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 157 — #173

i i

Wick’s theorem 157

and by the definition of a contraction, Eq. (B.0.2), the following must also be true:

âα â†β = N (âα â†β ) + âα â†β .

Let us now look into products with three elements:

âα â†β âγ = −â†β âα + δαβ âγ = −â†β âα âγ + δαβ âγ , (B.0.4)
âα âβ â†γ = âα −â†γ âβ + δβ γ = δβ γ âα − −â†γ âα + δαγ âβ

= â†γ âα âβ + δβ γ âα − δαγ âβ , (B.0.5)

âα â†β â†γ = â†β â†γ âα − δαβ â†γ − δαγ â†β , (B.0.6)
â†α âβ â†γ = â†α −â†γ âβ + δβ δ = −â†α â†γ âβ + δβ γ â†α .

(B.0.7)

Finally, we provide an example of a 4-operator product:

âα âβ â†γ â†δ =âα −â†γ âβ + δβ γ â†δ = δβ γ âα â†δ − −â†γ âα + δαγ âβ â†δ

=δβ γ âα â†δ − δαγ âβ â†δ + â†γ âα −â†δ âβ + δβ δ

=δβ γ −â†δ âα + δαδ − δαγ −â†δ âβ + δβ δ + δβ δ â†γ âα

− â†γ −â†δ âα + δαδ âβ

=δβ γ δαδ − δαγ δβ δ − δβ γ â†δ âα + δαγ â†δ âβ + δβ δ â†γ âα − δαδ â†γ âβ
+ â†γ â†δ âγ âβ . (B.0.8)

There is a clear pattern in all the above expressions. A normal-ordered product

always appears as one of the terms in the r.h.s., and then we have a sum of terms
containing deltas and normal-ordered products with fewer operators. First, we can
think of the Kronecker deltas as the contraction of two operators (see Problem 3.1).
Second, by the definition of normal ordering it is immediate that we can freely take
any multiplicative factor in and out of the normal ordering operation, so we can
actually express all r.h.s. terms as normal-ordered products with contractions. Take
for instance the last term in Eq. (B.0.4):

δαβ âγ = âα â†β N (âγ ) = N (âα â†β âγ ) .

This can be expressed as a normal-ordered and contracted term. The same applies to
all the other terms on the r.h.s. of Eqs. (B.0.4)-(B.0.8) that have Kronecker deltas. In
some cases, however, we may need to be careful with the fermionic sign. Consider
the third term in Eq. (B.0.5),

δαγ âβ = âα â†γ N (âβ ) = −N (âα âβ â†γ ) ,

where we require a − sign by virtue of Eq. (B.0.3).

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 158 — #174

i i

158 Sailing the Fermi sea: Problems and Solutions in Many-Body Quantum Theory

Taking these ideas into consideration, we can re-express all the products above in
a way that involves only sums (no negative signs) of normal-ordered and contracted
terms:

âα â†β =N (âα â†β ) + N (âα â†β )

âα â†β âγ =N (âα â†β âγ ) + N (âα â†β âγ )

âα âβ â†γ =N (âα âβ â†γ ) + N (âα âβ â†γ ) + N (âα âβ â†γ )

âα â†β â†γ =N (âα â†β â†γ ) + N (âα â†β â†γ ) + N (âα â†β â†γ )

â†α âβ â†γ =N (â†α âβ â†γ ) + N (â†α âβ â†γ )

âα âβ â†γ â†δ =N (âα âβ â†γ â†δ )

+ N (âα âβ â†γ â†δ ) + N (âα âβ â†γ â†δ ) + N (âα âβ â†γ â†δ ) + N (âα âβ â†γ â†δ )

+ N (âα âβ â†γ â†δ ) + N (âα âβ â†γ â†δ ) .

Notice that, in each of these expressions, we could have added normal-ordered terms
involving contractions of the sort: âα âβ , â†α â†β , â†α âβ , as they are all null (see Prob-
lem 3.1). This, in turn, is an indication that we might be able to express any product
of operators as a sum of normal-ordered products with different numbers of contrac-
tions. This result, obtained here with rather hand-wavy arguments, can be rigorously
proven to be true. This is in fact the well-known Wick’s Theorem:
Theorem (Wick’s Theorem):

Â1 . . . Ân =N (Â1 . . . Ân )

+ ∑ N (Â1 . . . Âi . . . Â j . . . Ân )

(i j)

+ ∑ N (Â1 . . . Âi . . . Âk . . . Â j . . . Âl . . . Ân )

(i j)(kl)
..
.
+ sum over fully-contracted products . (B.0.9)

The first sum runs over all single contractions; the second sum runs over all combina-
tions of double contractions; the last term runs over all full contractions, i.e., where
all operators are contracted. The proof of this theorem is somewhat long and will not
be given here. The interested reader can find it in G. C. Wick’s 1950 paper [14] and
other books [3, 5, 7, 9].
The theorem is of particular use in the following situations: whenever we have
a vacuum expectation value, it is clear that all normal-ordered terms will cancel. By
Wick’s theorem, the only terms that remain are the fully-contracted contributions

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 159 — #175

i i

Wick’s theorem 159

from the last line, such that:

0 Â1 . . . ÂN 0 = ∑ (±1)P × fully-contracted products , (B.0.10)

all pair combinations

where P is the number of transpositions needed to take all contracted pairs together.
When writing vacuum expectation values of fully-contracted products, it is common
practice to omit the normal-ordering symbol. In fact, the result of performing a full
contraction is a c-number, so the normal-ordering does not play a role anyway.
There is one last result which greatly simplifies the calculation of any expectation
value using Wick’s theorem:
Theorem (Sign Rule and Line Crossings): the factor (−1)P appearing in
Eq. (B.0.10) equals the number of contraction line crossings.
This allows us to directly compute the fermionic phase associated to normal or-
dering from the contractions themselves.

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 161 — #177

i i

References
1. A. Altland and B. D. Simons. Condensed Matter Field Theory. Cambridge University
Press, 2 edition, 2010.

2. L. E. Ballentine. Quantum Mechanics: A Modern Development. World Scientific, 1998.

3. J.-P. Blaizot and G. Ripka. Quantum theory of finite systems. MIT Press, Cambridge,
Massachusetts, 1986.

4. P. Coleman. Intoduction to Many-Body Physics. Cambridge University Press, 2015.

5. W. H. Dickhoff and D. van Neck. Many-body theory exposed!: propagator description

of quantum mechanics in many-body systems. World Scientific, first edition, 2005.

6. A. L. Fetter and J. D. Walecka. Quantum Theory of Many-Particle Systems. Dover, NY,

2003.

7. E. K. U. Gross, E. Runge, and O. Heinonen. Many-Particle Theory. Taylor & Francis,

1991.

8. X.-W. Guan, M. T. Batchelor, and C. Lee. Fermi Gases in One Dimension: From Bethe
Ansatz to Experiments. Rev. Mod. Phys., 85:1633–1691, Nov 2013.

9. G. D. Mahan. Many-particle Physics. Plenum Press, NY, Second edition, 1990.

10. N. Navon, S. Nascimbène, F. Chevy, and C. Salomon. The Equation of State of a Low-
Temperature Fermi Gas with Tunable Interactions. Science, 328(5979):729–732, 2010.

11. J. W. Negele and H. Orland. Quantum Many-Particle Systems. Addison-Wesley, 1987.

12. Gianluca Stefanucci and Robert van Leeuwen. Nonequilibrium Many-Body Theory of
Quantum Systems: A Modern Introduction. Cambridge University Press, first edition,
2013.

13. A. N. Wenz, G. Zürn, S. Murmann, I. Brouzos, T. Lompe, and S. Jochim. From Few
to Many: Observing the Formation of a Fermi Sea One Atom at a Time. Science,
342(6157):457–460, 2013.

14. G. C. Wick. The Evaluation of the Collision Matrix. Phys. Rev., 80:268–272, Oct 1950.

15. V. Zelevinsky. Quantum Physics: Volume 2. Wiley, 2010.

161

i i

i i
i i

“output” — 2025/8/11 — 7:50 — page 163 — #179

i i

Index
antisymmetrization operator, 3, 5 spin operator, 23
symmetrization operator, 3, 5
BCS ansatz, 136, 141 symmetrization postulate, 3
BCS theory, 135
three-body distribution function, 126
canonical transformation, 22 two-body density matrix, 128
closure relation, 41 two-body distribution function, 54, 57,
Compton profile, 107 94, 126
creation operator, 151 two-body operator, 14, 31
density operator, 18 wavefunction, 3
destruction operator, 151 Wick’s theorem, 27
distribution function, 93
dynamic structure function, 102

Fermi distribution, 75
Fermi energy, 40, 63
Fermi momentum, 52
field operator, 17, 19
Fourier transform, 42
free particle propagator, 38

Green’s function, 37

Hartree-Fock energy, 112, 116

Hartree-Fock method, 111
Heisenberg representation, 20

Koltun sum-rule, 47

linear response function, 101

one-body density matrix, 123

one-body operator, 14, 29
operator contraction, 156

particle number operator, 14, 18, 136

permutation symmetry, 3

second quantization, 13
sequential condition, 57, 94, 126
Slater determinant, 111
Slater function, 57, 94, 124
spectral function, 40, 45, 47

163

i i

Entan Interpret
No ratings yet
Entan Interpret
16 pages
Delta Function Normalization in QM
No ratings yet
Delta Function Normalization in QM
11 pages
M.Phys. C6 Theory Option/M.Math - Phys. Advanced Quantum Theory Lecture Notes For 2020-2021
No ratings yet
M.Phys. C6 Theory Option/M.Math - Phys. Advanced Quantum Theory Lecture Notes For 2020-2021
123 pages
High Energy Theory List
No ratings yet
High Energy Theory List
42 pages
(Fundamental Theories of Physics) V. Kulish - Hierarchical Methods - Hierarchy and Hierarchical Asymptotic Methods in Electrodynamics. Volume 1 - Kluwer Academic Publ (2002)
No ratings yet
(Fundamental Theories of Physics) V. Kulish - Hierarchical Methods - Hierarchy and Hierarchical Asymptotic Methods in Electrodynamics. Volume 1 - Kluwer Academic Publ (2002)
385 pages
Advanced Quantum Physics Insights
No ratings yet
Advanced Quantum Physics Insights
372 pages
String Theory Lecture Notes Overview
No ratings yet
String Theory Lecture Notes Overview
154 pages
Janis-Newman Algorithm: Generating Rotating and NUT Charged Black Holes
No ratings yet
Janis-Newman Algorithm: Generating Rotating and NUT Charged Black Holes
68 pages
Swanson M. Classical Field Theory and The Stress-Energy Tensor 2ed 2022
No ratings yet
Swanson M. Classical Field Theory and The Stress-Energy Tensor 2ed 2022
598 pages
Indirect Exchange Magnetic Electrons : of Nuclear Moments by Conduction
No ratings yet
Indirect Exchange Magnetic Electrons : of Nuclear Moments by Conduction
4 pages
Fate of the False Vacuum Theory
100% (1)
Fate of the False Vacuum Theory
8 pages
CEE 471, Fall 2019: HW1 Solutions
No ratings yet
CEE 471, Fall 2019: HW1 Solutions
8 pages
QFT Lecture Notes
No ratings yet
QFT Lecture Notes
205 pages
2 Regular Surfaces: E X: E R
No ratings yet
2 Regular Surfaces: E X: E R
37 pages
Rose RelativisticElectronTheory Outtakes
No ratings yet
Rose RelativisticElectronTheory Outtakes
43 pages
Quantum Corrections in False Vacuum
No ratings yet
Quantum Corrections in False Vacuum
7 pages
ElectromagneticFieldTheory ZahnSolutionsManual PDF
No ratings yet
ElectromagneticFieldTheory ZahnSolutionsManual PDF
302 pages
Vector Mechanics For Engineers-STATICS
67% (3)
Vector Mechanics For Engineers-STATICS
15 pages
Quantum Field Theory Basics
No ratings yet
Quantum Field Theory Basics
63 pages
Complex Number Theory
No ratings yet
Complex Number Theory
33 pages
The Adiabatic Motion of Charged Particles (Interscience Tracts On Physics Astronomical) (Theodore G. Northrop)
No ratings yet
The Adiabatic Motion of Charged Particles (Interscience Tracts On Physics Astronomical) (Theodore G. Northrop)
136 pages
Wigner D-Matrix
100% (1)
Wigner D-Matrix
8 pages
A Review On Cartans Structure Equations For Certa
No ratings yet
A Review On Cartans Structure Equations For Certa
7 pages
Lorentz and Poincaré Group Representations
No ratings yet
Lorentz and Poincaré Group Representations
11 pages
Einstein's Vierbein Field Theory of Curved Space: Jeffrey Yepez
No ratings yet
Einstein's Vierbein Field Theory of Curved Space: Jeffrey Yepez
27 pages
The Philosophy of Physics - Max Planck, W.H. Johnston
No ratings yet
The Philosophy of Physics - Max Planck, W.H. Johnston
131 pages
Relativistic Quantum Physics - From Advanced Quantum Mechanics To Introductory Quantum Field Theory
No ratings yet
Relativistic Quantum Physics - From Advanced Quantum Mechanics To Introductory Quantum Field Theory
54 pages
Introduction To General Relativity Solutions 1-5
No ratings yet
Introduction To General Relativity Solutions 1-5
5 pages
Asher Peres
No ratings yet
Asher Peres
5 pages
An Introduction To Relativistic Processes and The Standard Model of Electroweak Interactions - Carlo M. Becchi, Giovanni Ridolfi (2006 Springer, 140p)
No ratings yet
An Introduction To Relativistic Processes and The Standard Model of Electroweak Interactions - Carlo M. Becchi, Giovanni Ridolfi (2006 Springer, 140p)
140 pages
Quantum Field Theory Ryder
100% (1)
Quantum Field Theory Ryder
27 pages
Frontiers in Modern Condensed Matter Physics
No ratings yet
Frontiers in Modern Condensed Matter Physics
19 pages
Solutions To Feynman-Hibbs Classical Action Problems
No ratings yet
Solutions To Feynman-Hibbs Classical Action Problems
15 pages
Pseudoparticle Solutions of The Yang-Mills Equations
100% (1)
Pseudoparticle Solutions of The Yang-Mills Equations
3 pages
To Orbital Perturbations: James M. Longuski Felix R. Hoots George E. Pollock IV
No ratings yet
To Orbital Perturbations: James M. Longuski Felix R. Hoots George E. Pollock IV
349 pages
Spontaneous Breakdown of Parity in A Class of Gauge Theories
No ratings yet
Spontaneous Breakdown of Parity in A Class of Gauge Theories
31 pages
Renormalization Made Easy, Baez
No ratings yet
Renormalization Made Easy, Baez
11 pages
General Understanding GTR
100% (1)
General Understanding GTR
35 pages
Bohigas 1984
No ratings yet
Bohigas 1984
4 pages
Lecture Notes QFT Temp 3
No ratings yet
Lecture Notes QFT Temp 3
77 pages
QM Postulates
No ratings yet
QM Postulates
11 pages
A Treatise On Octonions: A Development of Clifford's Bi-Quaternions
No ratings yet
A Treatise On Octonions: A Development of Clifford's Bi-Quaternions
278 pages
CMT1
No ratings yet
CMT1
55 pages
Itzhak Bars - 2t-Physics and The Standard Model of Particles and Forces
100% (1)
Itzhak Bars - 2t-Physics and The Standard Model of Particles and Forces
23 pages
Frolov Blackhole
100% (2)
Frolov Blackhole
788 pages
Introduction To Quantum-State Estimation
100% (1)
Introduction To Quantum-State Estimation
396 pages
Superstrings: Theory of Everything
No ratings yet
Superstrings: Theory of Everything
242 pages
String Theory Part I Solutions
No ratings yet
String Theory Part I Solutions
202 pages
(Oxford Graduate Texts) Jean Zinn-Justin - Path Integrals in Quantum Mechanics (2010, Oxford University Press) - Libgen - Li
No ratings yet
(Oxford Graduate Texts) Jean Zinn-Justin - Path Integrals in Quantum Mechanics (2010, Oxford University Press) - Libgen - Li
335 pages
Relativistic Quantum Chemistry Intro
No ratings yet
Relativistic Quantum Chemistry Intro
111 pages
(FREE PDF Sample) A Course in Field Theory 1st Edition Pierre Van Baal Ebooks
No ratings yet
(FREE PDF Sample) A Course in Field Theory 1st Edition Pierre Van Baal Ebooks
67 pages
Functional Methods in Quantum Field Theory and Statistical Physics First Edition Vasiliev Instant Download
No ratings yet
Functional Methods in Quantum Field Theory and Statistical Physics First Edition Vasiliev Instant Download
164 pages
Henrik Bruus, Karsten Flensberg Many-Body Quantu
No ratings yet
Henrik Bruus, Karsten Flensberg Many-Body Quantu
352 pages
Many-Body Quantum Theory in Condensed Matter Physics Henrik Bruus and Karsten Flensberg
100% (1)
Many-Body Quantum Theory in Condensed Matter Physics Henrik Bruus and Karsten Flensberg
352 pages
Many-Body Quantum Theory in - Condensed Matter Physics - Henrik Bruus and Karsten Flensberg
No ratings yet
Many-Body Quantum Theory in - Condensed Matter Physics - Henrik Bruus and Karsten Flensberg
352 pages
Taiz & Zaiger. Plant Physiol
No ratings yet
Taiz & Zaiger. Plant Physiol
16 pages
Induction Motor Dynamics Study
No ratings yet
Induction Motor Dynamics Study
27 pages
Marketing Plan For HP PDF
No ratings yet
Marketing Plan For HP PDF
14 pages
Vial Integrity
No ratings yet
Vial Integrity
87 pages
PHY 104 Lecture Schedule
No ratings yet
PHY 104 Lecture Schedule
2 pages
Address: Plot 71 Stadium Road and ISPON Secretariat, Igwuruta, R/S
No ratings yet
Address: Plot 71 Stadium Road and ISPON Secretariat, Igwuruta, R/S
4 pages
Business Strategy Insights
No ratings yet
Business Strategy Insights
3 pages
News Release 1
No ratings yet
News Release 1
3 pages
Electronic Diagrams and Schematics Guide
100% (1)
Electronic Diagrams and Schematics Guide
29 pages
C11000 Copper Specification Sheet
No ratings yet
C11000 Copper Specification Sheet
8 pages
Modern Report Powerpoint Template: Free Template Site2Max - Pro
No ratings yet
Modern Report Powerpoint Template: Free Template Site2Max - Pro
19 pages
Curriculum Vitae of Nishant Shah
No ratings yet
Curriculum Vitae of Nishant Shah
3 pages
Requisition of Erector Mechanical
No ratings yet
Requisition of Erector Mechanical
9 pages
ISO 4136 - 2022 (2022) - Libgen - Li
100% (1)
ISO 4136 - 2022 (2022) - Libgen - Li
16 pages
Lecturette Topics SSB 2025
No ratings yet
Lecturette Topics SSB 2025
3 pages
Whitetopping Guidelines for Indian Roads
No ratings yet
Whitetopping Guidelines for Indian Roads
48 pages
Dr. Venu - Management Accounting
No ratings yet
Dr. Venu - Management Accounting
2 pages
Important Telephone Nos
No ratings yet
Important Telephone Nos
1 page
Nerval's Thesis Illuminés, Eccentricity, and The Evolution of Madness
No ratings yet
Nerval's Thesis Illuminés, Eccentricity, and The Evolution of Madness
303 pages
Understanding Program Quality Assessment
No ratings yet
Understanding Program Quality Assessment
21 pages
Deep Learning UNIT-II Part1
No ratings yet
Deep Learning UNIT-II Part1
48 pages
MBARARA DDP III Vol I
No ratings yet
MBARARA DDP III Vol I
173 pages
Tecumseh Engine Valve Specs Guide
No ratings yet
Tecumseh Engine Valve Specs Guide
1 page
STGB30H60DFB, STGP30H60DFB: Trench Gate Field-Stop 600 V, 30 A High Speed HB Series IGBT
No ratings yet
STGB30H60DFB, STGP30H60DFB: Trench Gate Field-Stop 600 V, 30 A High Speed HB Series IGBT
21 pages
Thesis 2.1 Custom PHP
100% (3)
Thesis 2.1 Custom PHP
6 pages
000cor CT Tissue Sim Dose PMB
No ratings yet
000cor CT Tissue Sim Dose PMB
21 pages
Indian Monsoon
No ratings yet
Indian Monsoon
14 pages
Forms of Energy Song
No ratings yet
Forms of Energy Song
2 pages
Finite Automata and Language Theory
No ratings yet
Finite Automata and Language Theory
16 pages
Present Modals: Paper Itermediate Structure
No ratings yet
Present Modals: Paper Itermediate Structure
10 pages