Notes 2003

Many-particle physics
Lecture notes for P654, Cornell University, spring 2003.

c
Piet Brouwer, 2004
0
1 Preliminaries
1.1 Thermal average
In the literature, two types of theories are constructed to describe a system
of many particles: Theories of the many-particle ground state, and theories
that address a thermal average at temperature T . In the book by Doniach
and Sondheimer, both approaches are described. Here, we’ll limit ourselves
to the finite-temperature theory. In all cases of interest to us, results at
zero temperature can be obtained as the zero temperature limit of the finite-
temperature theory.
The thermal average of an operator A is defined as
hAi = tr Aρ, (1.1)
where ρ is the density matrix. It describes the thermal distribution over the
different eigenstates of the system. The symbol tr denotes the “trace” of an
operator, the summation over the expectation values of the operator over an
orthonormal basis set, X
tr Aρ = hn|Aρ|ni. (1.2)
n
The basis set {|ni} can be the collection of many-particle eigenstates of the
Hamiltonian H, or any other orthonormal basis set.
In the canonical ensemble of statistical thermodynamics, the trace is taken
over all states with N particles and one has
1 −H/T
ρ= e , Z = tr e−H/T . (1.3)
Z
Using the basis of N-particle eigenstates |ni of the Hamiltonian H, the ther-
mal average hAi can then be written as
P −En /T
e hn|A|ni
hAi = n P −En /T . (1.4)
ne
In the grand-canonical ensemble, the trace is taken over all states, irre-
spective of particle number, and one has
1 −(H−µN )/T
ρ= e , Z = tr e−(H−µN )/T . (1.5)
Z
1
1.2 Schrödinger, Heisenberg, and interaction picture
There exist formally different but mathematically equivalent ways to for-
mulate the quantum theoretical dynamics. These different formulations are
called “pictures”. The idea behind the different pictures is that the physi-
cal content of the quantum theoretical formalism is contained in expectation
values only. For an operator A, the expectation value in the quantum state
|ψi is
A = hψ|A|ψi. (1.6)
The time-dependence of the expectation value of A can be encoded through
the time-dependence of the quantum state ψ, or through the time-dependence
of the operator A, or through both.1
In the Schrödinger picture, all time-dependence is encoded in the quantum
state |ψ(t)i; operators are time independent, except for a possible explicit
time dependence.2 Probably, this is the picture of quantum mechanics that
you are best used to. The time-evolution of the quantum state |ψ(t)i is
governed by a unitary evolution operator U(t, t0 ) that relates the quantum
state at time t to the quantum state at a reference time t0 ,
|ψ(t)i = U(t, t0 )|ψ(t0 )i. (1.7)
Quantum states at two arbitrary times t and t′ are related by the evolution
operator U(t, t′ ) = U(t, t0 )U † (t′ , t0 ). The evolution operator U satisfies the
group properties U(t, t) = 1̂ and U(t, t′ ) = U(t, t′′ )U(t′′ , t′ ) for any three times
t, t′ , and t′′ , and the unitarity condition
U(t, t′ ) = U † (t′ , t). (1.8)
The time-dependence of the evolution operator U(t, t0 ) is given by the Schrö-

dinger equation
∂U(t, t0 )
i~ = HU(t, t0 ). (1.9)
∂t
1
You can find a detailed discussion of the pictures in a textbook on quantum mechanics,
e.g., chapter 8 of Quantum Mechanics, by A. Messiah, North-Holland (1961).
2
An example of an operator with an explicit time dependence is the velocity operator
in the presence of a time-dependent vector potential,
1 ~ − eA(t)
~

~v = −i~∇ .
m
2
For a time-independent Hamiltonian H, the evolution operator U(t, t0 ) is the
well-known exponential function of H,
U(t, t′ ) = exp[−iH(t − t′ )/~]. (1.10)
If the Hamiltonian H is time dependent, there is no such simple result as Eq.

(1.10) above, and one has to solve the Schrödinger equation (1.9).
In the Heisenberg picture, the time-dependence is encoded in operators
A(t), whereas the quantum state |ψi is time independent. The Heisenberg
picture operator A(t) is related to the Schrödinger picture operator A as
A(t) = U(t0 , t)AU(t, t0 ), (1.11)
where t0 is a reference time. The Heisenberg picture quantum state |ψi has
no dynamics and is equal to the Schrödinger picture quantum state |ψ(t0 )i
at the reference time t0 . The time evolution of A(t) then follows from Eq.
(1.9) above,
dA(t) i ∂A(t)
= [H(t), A(t)]− + (1.12)
dt ~ ∂t
where H(t) = U(t0 , t)H(t0 )U(t, t0 ) is the Heisenberg representation of the
Hamiltonian and the last term is the Heisenberg picture representation of an
eventual explicit time dependence of the observable A. The square brackets
[·, ·]− denote a commutator. The Heisenberg picture Hamiltonian H(t) also
satisfies the evolution equation (1.12),
dH(t) i ∂H ∂H(t)
= [H(t), H(t)]− + = .
dt ~ ∂t ∂t
For a time-independent Hamiltonian, the solution of Eq. (1.12) takes the
familiar form
A(t) = eiH(t−t0 )/~Ae−iH(t−t0 )/~. (1.13)
You can easily verify that the Heisenberg and Schrödinger pictures are
equivalent. Indeed, in both pictures, the expectation values (A)t of the ob-
servable A at time t is related to the operators A(t0 ) and quantum state
|ψ(t0 )i at the reference time t0 by the same equation,
(A)t = hψ(t)|A|ψ(t)i
= hψ(t0 )|U(t0 , t)AU(t, t0 )|ψ(t0 )i
= hψ|A(t)|ψi. (1.14)
3
The interaction picture is a mixture of the Heisenberg and Schrödinger
pictures: both the quantum state |ψ(t)i and the operator A(t) are time
dependent. The interaction picture is usually used if the Hamiltonian is
separated into a time-independent unperturbed part H0 and a possibly time-
dependent perturbation H1 ,
H(t) = H0 + H1 . (1.15)
In the interaction picture, the operators evolve according to the unperturbed

Hamiltonian H0 ,
A(t) = eiH0 (t−t0 )/~A(t0 )e−iH0 (t−t0 )/~, (1.16)
whereas the quantum state |ψi evolves according to a modified evolution

operator UI (t, t0 ),
UI (t, t0 ) = eiH0 (t−t0 )/~U(t, t0 ). (1.17)
The evolution operator that relates interaction picture quantum states at
two arbitrary times t and t′ is
′
UI (t, t′ ) = eiH0 (t−t0 )/~U(t, t′ )e−iH0 (t −t0 )/~. (1.18)
The interaction picture evolution operator also satisfies the group and uni-
tarity properties.
You easily verify that this assignment leads to the same time-dependent
expectation value (1.14) as the Schrödinger and Heisenberg pictures.
A differential equation for the time dependence of the operator A(t) is
readily obtained from the definition (1.16),
dA(t) i ∂A(t)
= [H0 , A(t)]− + . (1.19)
dt ~ ∂t
The time dependence of the interaction picture evolution operator is governed
by the interaction-picture representation of the perturbation H1 only,
∂UI (t, t0 )
i~ = H1 (t)UI (t, t0 ), (1.20)
∂t
where
H1 (t) = eiH0 (t−t0 )/~H1 e−iH0 (t−t0 ) . (1.21)
4
Note that, in the interaction picture, the perturbation H1 acquires an ad-
ditional time dependence from the evolution according to the unperturbed
Hamiltonian H0 .
In these notes we’ll use all three pictures. Mostly, the context provides
sufficient information to find out which picture is used. As a general rule,
operators without time argument are operators in the Schrödinger picture,
whereas operators with time argument are in the Heisenberg and interaction
pictures. We’ll use this convention even if the Schrödinger picture operators
have an explicit time dependence, cf. Eq. (1.15) above.
1.3 imaginary time

When performing a calculation that involves a thermal average, it is often
convenient to introduce operators that are a function of the imaginary time
τ.
Imaginary times have no direct physical meaning. Real times do, e.g., for
time-dependent correlation functions. We use imaginary time for technical
convenience only. Later in this course, when we develop the diagrammatic
perturbation theory, it will become clear why. At this moment we can say
the following: A thermal average brings about a negative exponents e−H/T ,
whereas the real time dependence brings about imaginary exponents eiHt . For
calculations, it would have been much easier if all exponents were either real
or imaginary. To treat both real and imaginary exponents at the same time is
rather awkward. Therefore, we’ll opt to consider operators that depend on an
imaginary time argument, so that the imaginary exponent eiHt is replaced by
eHτ . If only imaginary times are used, only real exponents occur. This leads
to much simpler calculations, as we’ll see soon. The drawback of the change
from real times to imaginary times is that, at the end of the calculation,
one has to perform an analytical continuation from imaginary times to real
times, which requires considerable mathematical care.
In the Schrödinger picture, the imaginary time variable appears in the
evolution operator U(τ, τ ′ ), which now reads
′ )/~ ∂
U(τ, τ ′ ) = e−H(τ −τ , −~ U(τ, τ ′ )HU(τ, τ ′ ). (1.22)
∂τ
Since a thermal average necessarily involves a time-independent Hamiltonian,
we can give an explicit expression for the evolution operator. Note that
5
0
t
1/T
τ
Figure 1: Relation between real time t and imaginary time τ .
the evolution operator (1.22) satisfies the group properties U(τ, τ ) = 1 and
U(τ, τ ′ )U(τ ′ , τ ′′ ) = U(τ, τ ′′ ), but that it is no longer unitary.
In the Heisenberg picture, the imaginary time variable appears through
the relation
∂A(τ ) 1
A(τ ) = eHτ /~Ae−Hτ /~, = [H, A]− , (1.23)
∂τ ~
where A is the corresponding real-time operator. Again, H is assumed to be
time independent.
In the interaction picture, the time dependence of the operators is given
by
∂A(τ ) 1
A(τ ) = eH0 τ /~Ae−H0 τ /~, = [H0 , A]− , (1.24)
∂τ ~
whereas the time dependence of the quantum state |ψ(τ )i is given by the
interaction picture evolution operator
′ ′ ∂
UI (τ, τ ′ ) = eH0 τ e−H(τ −τ ) e−H0 τ , −~ UI (τ, τ ′ ) = H1 (τ )U(τ, τ ′ ). (1.25)
∂τ
Here H1 (τ ) = eH0 τ H1 e−H0 τ is the perturbation in the interaction picture.
The relation between the real time t and the imaginary time τ is illus-
trated in Fig. 1.
6
1.4 Exercises
Exercise 1.1: Evolution equation for expectation value
Show that the time dependence of the expectation value of an observable A

satisfies the differential equation
d i
A t= [H, A]− ,
dt ~ t
if the operator A has no explicit time dependence.
Exercise 1.2: Free particle, Schrödinger picture.
For a free particle with mass m, express the expectation values (xα )t and
(pα )t of the components of the position and momentum at time t in terms
of the corresponding expectation values at time 0 (α = x, y, z). Use the
Schrödinger picture.
Exercise 1.3: Free particle, Heisenberg picture
Derive the Heisenberg picture evolution equations for the position and mo-
mentum operators of a free particle. Use your answer to express the expec-
tation values (xα )t and (pα )t of the components of the position and momen-
tum at time t in terms of the corresponding expectation values at time 0
(α = x, y, z).
Exercise 1.4: Harmonic oscillator
Derive and solve the Heisenberg picture evolution equations for the position
and momentum operators x and p of a one-dimensional harmonic oscillator
with mass m and frequency ω0 . Do the same for the corresponding creation
and annihilation operators
a† = (ω0 m/2~)1/2 x − i(2~ω0 m)−1/2 p

a = (ω0 m/2~)1/2 x + i(2~ω0 m)−1/2 p. (1.26)
7
Verify that the Hamiltonian is time-independent.
Exercise 1.5: Imaginary time
Derive and solve the imaginary-time Heisenberg picture evolution equations

for the position and momentum operators of a one-dimensional harmonic
oscillator with mass m and frequency ω0 . Do the same for the corresponding
creation and annihilation operators, see Eq. (1.26).
8
2 General formalism; equilibrium theory
Throughout this course we’ll make use of “Green functions”. There are
many ways one can define Green functions, and there are important relations
between the different definitions. Quite often, a certain application calls for
one Green function, whereas it is easier to calculate a different Green function
for the same system.
We now present the various definitions of Green functions in the general
case and the general relations between the Green functions. In the later
chapters we discuss various physical applications that call for the use of
Green functions.
We consider Green functions (or “correlation functions”) defined for two
operators A and B, which do not need to be hermitian. In the applications,
we’ll encounter cases where the operators A and B are fermion or boson
creation or annihilation operators, or displacements of atoms in a lattice,
or current or charge densities. We say that the operators A and B satisfy
commutation relations if they describe bosons or if they describe fermions
and they are even functions of fermion creation and annihilation operators.
We’ll refer to this case as the “boson” case, and use a − sign in the formulas
below. We say that the operators A and B satisfy anticommutation relations
if they describe fermions and they are odd functions of in fermion creation
and annihilation operators. We’ll call this case the “fermion case”, and use
a + sign in the formulas of the next sections.
2.1 Temperature Green functions

The temperature Green function GA;B (τ1 ; τ2 ) for the operators A and B is
defined as3
GA;B (τ1 ; τ2 ) = −hTτ [A(τ1 )B(τ2 )]i. (2.1)
Here, the brackets denote a thermal average and the symbol Tτ denotes time-
ordering,
Tτ [A(τ1 )B(τ2 )] = θ(τ1 − τ2 )A(τ1 )B(τ2 ) − (±1)θ(τ2 − τ1 )B(τ2 )A(τ1 ). (2.2)
As discussed previously, the + sign applies if the operators A and B satisfy

fermion anticommutation relations, i.e., if they are of odd degree in fermion
3
DS denote the phonon temperature Green function as G. Their definition differs a
minus-sign with respect to that of Eq. (2.1).
9
creation/annihilation operators. The − sign applies if A and B are boson
operators or if they are of even degree in fermion creation/annihilation op-
erators. The temperature Green function is defined on the interval −~/T <
τ1 , τ2 < ~/T . In order to extend the definition to the entire imaginary time
axis −∞ < τ1 , τ2 < ∞ one makes use of the facts that (i) G depends on
the difference τ = τ1 − τ2 only and (ii) G is periodic with period ~/T ,
G(τ ) = G(τ + ~/T ) for bosons, and antiperiodic with the same period,
G(τ ) = −G(τ + ~/T ) for fermions. With these properties, G can be extended
to the entire imaginary time axis.
If the operators A and B and the Hamiltonian H are all symmetric,
the temperature Green function is symmetric or antisymmetric in the time
argument, for the boson and fermion cases, respectively,
GA;B (τ ) = −(±1)GA;B (−τ ) if A, B, H symmetric. (2.3)
Using the periodicity of G, one may write G as a Fourier series,4

X
GA;B (τ ) = T e−iωn τ GA;B (iωn ), (2.4)
n
with Matsubara frequencies ωn = 2πn/T , n integer, for bosons and ωn =

(2n + 1)π/T for fermions. The inverse relation is
Z ~/T
Giα;jβ (iωn ) = dτ eiωn τ Giα;jβ (τ ). (2.5)
0
2.2 Green functions with real time arguments

In addition to the temperature Green functions, Green functions with real
time arguments are used. The retarded, advanced, and time-ordered Green
functions are defined as
GR ′ ′ ′
A;B (t; t ) = −iθ(t − t )h[A(t), B(t )]± i, (2.6)
GA ′
A;B (t; t )
′
= iθ(t − t)h[A(t), B(t )]± i, ′
(2.7)
GA;B (t; t′ ) = −ihTt [A(t)B(t′ )]i
= −iθ(t − t′ )hA(t)B(t′ )i ± iθ(t′ − t)hB(t′ )A(t)i, (2.8)
4
DS define the Fourier transform without the prefactor T .
10
respectively. Here, [·, ·]+ is the anticommutator, [A, B]+ = AB + BA. One
verifies that
GA ′ R ′ ∗
B † ;A† (t ; t) = GA;B (t; t ) . (2.9)
By definition, GR is zero for t − t′ < 0 and GA is zero for t − t′ > 0. Further,
by virtue of their definitions, one has the relation
GR ′ A ′
A;B (t; t ) = −(±1)GB;A (t ; t), (2.10)
where the + sign applies to fermions and the − sign to bosons.

Since the averages h. . .i are taken in thermal equilibrium, the Green func-
tions depend on the time difference t − t′ only. Then, Fourier transforms of
these Green functions are defined as
Z ∞
R
GA;B (ω) = dteiωt GR
A;B (t), (2.11)
0
Z 0
A
GA;B (ω) = dteiωt GA
A;B (t), (2.12)
−∞
Z ∞
GA;B (ω) = dteiωt GA;B (t). (2.13)
−∞
The retarded Green function GR (ω) is analytic for Im ω > 0, while the ad-
vanced Green function GA (ω) is analytic for Im ω < 0.
The inverse Fourier transforms are
Z ∞
R 1
GA;B (t) = dωe−iωt GR
A;B (ω), (2.14)
2π −∞
Z ∞
A 1
GA;B (t) = dωe−iωt GA
A;B (ω), (2.15)
2π −∞
Z ∞
1
GA;B (t) = dωe−iωt GA;B (ω). (2.16)
2π −∞
Note that the fact that GR (ω) is analytic for Im ω > 0 implies that GR (t) = 0
for t < 0, and, similarly, that the fact that GA (ω) is analytic for Im ω < 0
implies that GR (t) = 0 for t > 0.
For some applications it is important to consider Green functions without
time ordering. These are the “greater” and “lesser” Green functions
G> ′ ′
A;B (t, t ) = −ihA(t)B(t )i, (2.17)
G< ′ ′
A;B (t, t ) = ±ihB(t )A(t)i, (2.18)
11
where the + sign is for fermions and the − sign for bosons. The retarded,
advanced, and time-ordered Green functions can be written in terms of these
two functions as
GR ′ ′ > ′ < ′
A;B (t, t ) = θ(t − t )[GA;B (t, t ) − GA;B (t, t )],
GA ′ ′ < ′ > ′
A;B (t, t ) = θ(t − t)[GA;B (t, t ) − GA;B (t, t )], (2.19)
GA;B (t, t′ ) = θ(t − t′ )G> ′ ′ < ′
A;B (t, t ) + θ(t − t)GA;B (t, t ).
Inversely, the greater and lesser Green functions can be expressed as
G> ′ ′ A ′
A;B (t, t ) = GA;B (t, t ) − GA;B (t, t )
G< ′ ′ R ′
A;B (t, t ) = GA;B (t, t ) − GA;B (t, t ). (2.20)
Since averages are taken in thermal equilibrium, the Green functions G> and
G< depend on the time difference t − t′ only. The Fourier transform of the
greater and lesser Green functions is defined as
Z ∞
>
GA;B (ω) = eiωt G>
A;B (t), (2.21)
−∞
Z ∞
<
GA;B (ω) = eiωt G<
A;B (t). (2.22)
−∞
2.3 Relations between Green functions

In fact, all six Green functions defined in the previous two sections are related.
Before we explain those relations, it is helpful to define a “real part” and
“imaginary part” of a Green function.
In the time representation, we define the “real part” ℜG of the retarded,
advanced, and time-ordered Green functions as
1 R
ℜGR ′
GA;B (t; t′ ) + GR ′ ∗

A;B (t; t ) = B † ;A† (t ; t) , (2.23)
2
1 A
ℜGA ′
GA;B (t; t′ ) + GA ′ ∗

A;B (t; t ) = B † ;A† (t ; t) , (2.24)
2
1
ℜGA;B (t; t′ ) = GA;B (t; t′ ) + GB† ;A† (t′ ; t)∗ .

(2.25)
2
Similarly, we define the “imaginary part” ℑG as
1
ℑGR ′
GR ′ R ′ ∗

A;B (t; t ) = A;B (t; t ) − GB † ;A† (t ; t) , (2.26)
2i
12
1
ℑGA ′
GA ′ A ′ ∗

A;B (t; t ) = A;B (t; t ) − GB † ;A† (t ; t) , (2.27)
2i
1
ℑGA;B (t; t′ ) = GA;B (t; t′ ) − GB† ;A† (t′ ; t)∗ .

(2.28)
2i
These “real” and “imaginary” parts of Green functions are different from the
standard real and imaginary parts Re G and Im G. In particular, ℜG and
ℑG are not necessarily real numbers. The quantities ℜG and ℑG are defined
with respect to a “hermitian conjugation” that consists of ordinary complex
conjugation, interchange and hermitian conjugation of the operators A and
B, and of the time arguments. The assignments ℜ and ℑ are invariant under
this hermitian conjugation and are preserved under Fourier transform of the
time variable t,
1
GA;B (ω) + GB† ;A† (ω)∗ ,

ℜGA;B (ω) = (2.29)
2
1
GA;B (ω) − GB† ;A† (ω)∗ .

ℑGA;B (ω) = (2.30)
2i
with similar expressions for the retarded and advanced Green functions.
By virtue of Eq. (2.9), the retarded and advanced Green functions are
hermitian conjugates, hence
ℜGR (ω) = ℜGA (ω), (2.31)

ℑGR (ω) = −ℑGA (ω). (2.32)
Similarly, by employing their definitions and with repeated use of the relation
hA(t)B(t′ )i∗ = hB † (t′ )A† (t)i, (2.33)
one finds a relation between the real parts of the time-ordered and the re-
tarded or advanced Green functions,
ℜGR (ω) = ℜG(ω), (2.34)

ℜGA (ω) = ℜG(ω). (2.35)
Shifting the t-integration from the real axis to the line t − i/T , one derives
the general relation
Z Z
iωt
dte hA(t)B(0)i = e ~ω/T
dteiωt hB(0)A(t)i. (2.36)
13
Using Eq. (2.36), together with Eq. (2.33), one finds a relation for the imag-
inary parts of the time-ordered and retarded or advanced Green functions.
This relation follows from the fact that the imaginary parts of the retarded,
advanced, and time-ordered Green functions do not involve theta functions.
Indeed, for the imaginary part of the retarded and advanced functions one
finds
ℑGR A
A;B (ω) = −ℑGA;B (ω)
Z ∞
1
= − dteiωt hA(t)B(0) ± B(0)A(t)i
2 −∞
1 ∞
Z
−~ω/T
= − 1±e dteiωt hA(t)B(0)i, (2.37)
2 −∞
whereas for the time-ordered Green function one has

1 ∞
Z
ℑGA;B (ω) = − dteiωt hA(t)B(0) − (±1)B(0)A(t)i
2 −∞
1 ∞
Z
−~ω/T
= − 1 − (±1)e dteiωt hA(t)B(0)i. (2.38)
2 −∞
Hence, we find
1 + (±1)e−~ω/T
ℑGR (ω) = ℑG(ω), (2.39)
1 − (±1)e−~ω/T
1 + (±1)e−~ω/T
ℑGA (ω) = − ℑG(ω), (2.40)
1 − (±1)e−~ω/T
where the + sign is for fermion operators and the − sign for boson operators.
With the help of the Fourier representation of the step function,
Z ∞ ′
1 eiω t
θ(t) = dω ′ ′ , (2.41)
2πi −∞ ω − iη
where η is a positive infinitesimal, and of the second line of Eq. (2.37), one
shows that
1 1
Z
R
GAB (ω) = − dω ′ ′
2π ω − ω − iη
Z ∞
′
× dteiω t hA(t)B(0) ± B(0)A(t)i
−∞
1 ∞ ℑGR ′
A;B (ω )
Z
′
= dω ′ . (2.42)
π −∞ ω − ω − iη
14
Similarly, one finds
1 ∞ ℑGA ′
A;B (ω )
Z
GA
A;B (ω) = − dω ′ . (2.43)
π −∞ ω ′ − ω + iη
The numerator of the fractions in Eqs. (2.42) and (2.43) is known as the
“spectral density”,
AA;B (ω) = −2ℑGR (ω) = 2ℑGA (ω). (2.44)
Other definitions of the spectral density can be found in the literature,5 but
they all have in common that the spectral density is the imaginary part of a
Green function. The spectral density satisfies the normalization condition
Z ∞
1
dωAA;B (ω) = h[A, B]± i. (2.45)
2π −∞
Using Eq. (2.20) and the results derived above, we can write the Fourier
transforms G> (ω) and G< (ω) of the greater and lesser Green functions as
iAA;B (ω)
G>
A;B (ω) = − , (2.46)
1 ± e−~ω/T
iAA;B (ω)
G<
A;B (ω) = . (2.47)
1 ± e~ω/T
We conclude that the greater and lesser Green functions are purely imaginary.
Finally, by shifting integration contours as shown in Fig. 2, one can write
the Fourier transform of the temperature Green function for ωn > 0 as
Z ~/T
GA;B (iωn ) = − dτ eiωn τ hA(τ )B(0)i
Z0 ∞
= −i dthA(t)B(0)ie−ωn t
Z0 ∞
+i dthA(t − i/T )B(0)ie−ωnt+iωn /T
0
5
DS define two spectral densities J1 (ω) and J2 (ω) for the boson case, with
J1 (ω) = iG> <

A;B (ω), J2 (ω) = −iGA;B (ω),
see Eq. (2.46) below.
15
0 1/T τ
Figure 2: Integration contours for derivation of Eq. (2.48).

Z ∞
= −i dthA(t)B(0)ie−ωn t
Z0 ∞
+i dthB(0)A(t)ie−ωn t+iωn /T
0
R
= GA;B (iωn ). (2.48)
Here, we made use of the fact that exp(iωn /T ) = −(±1). In order to find
the temperature Green function for negative Matsubara frequencies, one can
make use of the relation
GA;B (iωn ) = GB† ;A† (−iωn )∗ . (2.49)
Writing the retarded Green function in terms of the spectral density using
Eqs. (2.42) and (2.44), one thus obtains the general expression
1 AA;B (ω ′)
Z
GA;B (iωn ) = − dω ′ ′ , (2.50)
2π ω − iωn
which is valid for positive and for negative Matsubara frequencies.
In many applications is it difficult, if not impossible, to calculate Green
functions exactly. Instead, we have to rely on perturbation theory. Whereas
the retarded, advanced, greater, and lesser Green functions are the ones
that appear in most physical applications, good diagrammatic methods to
calculate Green functions in perturbation theory exist for temperature and
time-ordered Green functions only. The above relations allows one to relate
retarded, advanced, and temperature Green functions to the temperature
and time-ordered Green functions. Hence, these relations are a crucial link
between what can be calculated easily and what is desired to be calculated.
16
2.4 Exercises
Exercise 2.1: Lehmann representation
Explicit representations for Green functions can be obtained using the set of
many-particle eigenstates {|ni} of the Hamiltonian H, as a basis set. For
example, the greater Green function in the canonical ensemble can be written
as
i
G> ′ ′ −H/T
A;B (t, t ) = − tr A(t)B(t )e
Z
i X −En /T
= − e hn|A(t)B(t′ )|ni, (2.51)
Z n
P −En /T
where Z = ne is the canonical partition function. (In the grand
canonical ensemble, similar expressions are obtained.) Inserting a complete
basis set and using X
|n′ ihn′ | = 1,
n′
we write Eq. (2.51) as

i X −En /T +i(t−t′ )(En −En′ )
G> ′
A;B (t, t ) = − e hn|A|n′ ihn′ |B|ni. (2.52)
Z n,n′
Performing the Fourier transform to time, one finds

2πi X −En /T
G>
A;B (ω) = − e δ(En − En′ + ω)hn|A|n′ihn′ |B|ni. (2.53)
Z n,n′
With the Lehmann representation, the general relations we derived in Sec.

2.3 can be verified explicitly.
(a) Verify that G>A;B is purely imaginary (with respect to hermitian conju-
gation, as defined in Sec. 2.3).
(b) Derive Lehmann representations of the retarded, advanced, lesser, time-

ordered, and temperature Green functions.
(c) Derive the Lehmann representation of the spectral density AA;B (ω).
17
(d) Verify the relations between the different types of Green functions that
were derived in Sec. 2.3 using the Lehmann representations derived in
(b) and (c).
Exercise 2.2: Spectral density
In this exercise we consider the spectral density AA;B (ω) for the case A = c,
B = c† that the operators A and B are fermion or boson annihilation and
creation operators, respectively. In this case, the spectral density Ac;c† (ω)
can be viewed as the energy resolution of a particle created by the creation
operator c† . To show that this is a plausible idea, you are asked to prove
some general relations for the spectral density Ac;c† (ω). Consider the cases
of fermions and bosons separately.
(a) The spectral density Ac;c† (ω) is normalized,
Z ∞
1
dωAc;c† (ω) = 1. (2.54)
2π −∞
(b) For fermions, the spectral density Ac;c† (ω) ≥ 0. For bosons, Ac;c† (ω) ≥
0 if ω > 0 and Ac;c† (ω) ≤ 0 if ω < 0. (Hint: use the Lehmann
representation of the spectral density.)
(c) The average occupation n̄ = hc† ci is
Z ∞
1 1
n̄ = dωAc;c† (ω) . (2.55)
2π −∞ 1 ± e~ω/T
(d) Now consider a Hamiltonian that is quadratic in creation and annihi-

lation operators, X
H= εn c†n cn .
n
For this Hamiltonian, calculate the spectral density A(n, ω) ≡ Acn ;c†n (ω).
18
Exercise 2.3: Harmonic oscillator
For a one-dimensional harmonic oscillator with mass m and frequency ω0 ,

calculate the retarded, advanced, greater, lesser, time-ordered, and temper-
ature Green functions for the following choices of the operators A and B:
(a) Both A and B are equal to the position operator x.
(b) The operator A is equal to the position operator x, while B is equal to

the momentum operator p.
(c) The operator A is equal to the annihilation operator a, while B is equal

to the creation operator a† , see Eq. (1.26).
19
20
3 Introduction: phonons
3.1 Normal modes
We consider the dynamics of the atoms of a lattice. The atoms interact via a
potential V({~xi }) that depends on the location ~xi of all atoms i = 1, . . . , N.
Each atom has an equilibrium position ~ri . For small displacements ~ui = ~xi −~ri
from equilibrium, the Hamiltonian for the lattice atoms can be expanded in a
series in the displacements ~ui . The first order term in the expansion vanishes,
since the potential V is a minimum at the equilibrium positions of the atoms.
If we truncate the expansion at the second order, we have
N
X p~2i 1 XX
H= + uiα ujβ Vα;β (~ri , ~rj ), (3.1)
i=1
2M 2 i,j αβ
where the numbers Vα;β (~ri , ~rj ) are derivatives of the potential V at the equi-
librium positions ~ri , ~rj ,
Vα;β (~ri , ~rj ) = ∂xiα ∂xjβ V({~ri }). (3.2)
The displacement and momentum operators ~ui and p~j have canonical com-
mutation relations,
[piα , pjβ ]− = [uiα , ujβ ]− = 0 (3.3)
[piα , ujβ ]− = −i~δαβ δij . (3.4)
For an infinite lattice or for a lattice with periodic boundary conditions, the
real numbers Vα;β (~ri , ~rj )) depend on the difference ~ri − ~rj only. Inversion
symmetry implies Vα;β (~ri − ~rj ) = Vα;β (~rj − ~ri ), and, since derivatives can be
taken in any order, Vα;β (~ri − ~rj ) = Vβ;α (~ri − ~rj ).
The approximation (3.1) in which the Hamiltonian of the lattice atoms
is truncated after the quadratic term in the displacements is known as the
“harmonic approximation”.
Doniach and Sondheimer study the case that V is a sum of pair potentials,
X
V({~xi }) = U(~xi − ~xj ). (3.5)
i6=j
In that case, one has the expressions

Viα;jβ = −∂α ∂β U(~ri − ~rj ) if i 6= j, (3.6)
X
Viα;iβ = ∂α ∂β U(~ri − ~rj ). (3.7)
j6=i
21
The Hamiltonian (3.1) can be brought to diagonal form by a suitable
basis transformation. We construct this basis transformation in two steps.
First, we perform a Fourier transform with respect to the lattice coordinate
~ri and define
1 X i~k·~rj
p~~k = √ e p~j (3.8)
N j
1 X −i~k·~rj
~u~k = √ e ~uj . (3.9)
N j
Here and below, the wavevector ~k is in the first Brillouin zone of the reciprocal
lattice. After Fourier transform, the operators ~p~k and ~u~k still obey canonical
commutation relations,
h i h i
p~kα , p~k′β = u~kα , u~k′β = 0 (3.10)
− −
h i
p~kα , u~k′β = −i~δαβ δ~k~k′ . (3.11)
−
The momenta and displacements corresponding to opposite wavevectors ~k

and −~k are related by hermitian conjugation, ~p~k = p~†−~k and ~u~k = ~u†−~k . Note
that the total number of degrees of freedom has not changed upon Fourier
transform.
In terms of these new variables, the Hamiltonian reads
XX 1 1

H= p~kα pβ,−~k + Vαβ,~k u~kα uβ,−~k , (3.12)
α,β
2M 2
~k
where
~
X
Vαβ,~k = eik·~r Vα;β (~r). (3.13)
~
r
The matrix 3 × 3 matrix V~k is symmetric because Vα;β (~r) is symmetric and
real because Vα;β (~r) = Vα;β (−~r). Moreover, Vαβ,~k = Vα;β (−~k). Since the
matrix V~k is real and symmetric, it is diagonalized by three real and orthog-
λ
onal eigenvectors. We denote the eigenvectors as v~kα , λ = 1, 2, 3, and the
2
corresponding eigenvalues as Mω~kλ . Then we can introduce normal modes
X
p~λk = vαλ p~kα , (3.14)
α
X
u~λk = vαλ u~kα , (3.15)
α
22
that bring H into diagonal form,
X 1 1

λ λ 2 λ λ
H= p p + Mω~kλ u~k u−~k . (3.16)
2M ~k −~k 2
~k,λ
One verifies that the normal mode operators p~λk and u~λk still obey canonical
commutation relations.
Finally, we write the Hamiltonian in terms of phonon creation and anni-
hilation operators. These are defined as
s s
Mω ~ 1
b~λk = u~λk kλ
+ ipλ−~k , (3.17)
2~ 2Mω~kλ ~
s s
Mω ~kλ 1
b~kλ† = uλ−~k − ip~λk , (3.18)
2~ 2Mω~kλ ~
and obey the usual commutation rules for creation and annihilation opera-
tors,
h ′
i h ′
i
b~λk , b~kλ′ = b~kλ† , b~λk′ †
− −
= 0,
h ′
i
b~λk , b~λk′ † = δ~k~k′ δλλ′ . (3.19)
−
In terms of the new operators b~kλ† and b~λk , the Hamiltonian reads

X 1
H= ~ω~kλ b~kλ† b~λk + . (3.20)
2
~kλ
Since the operators b~kλ† and b~λk satisfy creation and annihilation operators
for bosons, the Hamiltonian (3.20) is interpreted as describing “phonons”,
quantized lattice vibrations with boson statistics. With or without such
interpretation, what is important for us is that the Hamiltonian (3.20) is
diagonal, so that any average of displacements and momenta of lattice atoms
can be calculated straightforwardly once it is expressed in terms of the op-
erators b~kλ† and b~λk . For this, it is important to have an expression for the
23
inverse of the normal-mode transformation we just derived,
s
1 X i~k·~rj λ ~ λ
ujα = √ e vα b~k + bλ†
−~k
, (3.21)
N ~ 2Mω~kλ
kλ
s
1 X ~ M~ω~kλ λ†
pjα = √ e−ik·~rj vαλ ib~k − ibλ−~k . (3.22)
N 2
~kλ
3.2 Phonon Green functions

For phonons we define Green functions with respect to the displacement
operators uiα . Following the formalism outlined in Sec. 2, we introduce the
phonon temperature Green function6
Diα;jβ (τ1 ; τ2 ) = −hTτ [uiα (τ1 )ujβ (τ2 )]i, (3.23)
where the brackets denote a thermal average and the symbol Tτ denotes
time-ordering,
Tτ [uiα (τ1 )ujβ (τ2 )] = θ(τ1 −τ2 )uiα(τ1 )ujβ (τ2 )+θ(τ2 −τ1 )ujβ (τ2 )uiα (τ1 ). (3.24)
Since the displacement operators u obey commutation relations (phonons are

bosons), we choose the boson variant of the expressions in Sec. 2.
For phonons, D is real, since u and H are real. According to Eq. (2.3),
this implies
Diα;jβ (τ ) = Diα;jβ (−τ ). (3.25)
Using the periodicity of D, one may write D as a Fourier series,7
X
Diα;jβ (τ ) = T e−iωn τ Diα;jβ (iωn ), (3.26)
n
where the ωn = 2πn/T , n integer, are the boson Matsubara frequencies.

The retarded, advanced, and time-ordered phonon Green functions are
defined as
R
Diα;jβ (t − t′ ) = −iθ(t − t′ )h[uiα (t), ujβ (t′ )]− i. (3.27)
6
DS denote the phonon temperature Green function as G. Their definition differs a
minus-sign with respect to that of Eq. (3.23).
7
DS define the Fourier transform without the prefactor T .
24
A
Diα;jβ (t − t′ ) = iθ(t′ − t)h[uiα (t), ujβ (t′ )]− i. (3.28)
Diα;jβ (t − t′ ) = −ihTt [uiα (t)ujβ (t′ )]i
= −iθ(t − t′ )huiα(t)ujβ (t′ )i
− iθ(t′ − t)hujβ (t′ )uiα (t)i, (3.29)
and their Fourier transforms are as given in Sec. 2. Finally, the phonon lesser
and greater Green functions are defined as
>
Diα;jβ (t − t′ ) = −ihuiα (t)ujβ (t′ )i, (3.30)
<
Diα;jβ (t − t′ ) = ihujβ (t′ )uiα (t)i. (3.31)
3.3 Calculation of phonon Green functions

3.3.1 Calculation using diagonal form of the Hamiltonian
The simplest method to calculate the phonon Green functions is to use the
diagonal form (3.16) of the phonon Hamiltonian. The time-dependence of
the phonon annihilation and creation operators is
b~λk (t) = e−iω~kλ t b~λk (0), b~kλ† (t) = eiω~kλ t b~kλ† (0). (3.32)
In thermal equilibrium, averages involving the phonon annihilation and cre-
ation operators are particularly simple:
′
hb~λk (t)b~kλ′ (0)i = 0, (3.33)
′
hb~kλ† (t)b~λk′ † (0)i = 0, (3.34)
′
hb~kλ† (t)b~kλ′ (0)i = δ~k~k′ δλλ′ eiω~kλ t hn~kλ i, (3.35)
′
hb~λk (t)b~λk′ † (0)i = δ~k~k′ δλλ′ e−iω~kλ t hn~kλ + 1i, (3.36)
where
1
hn~kλ i = ~ω~kλ /T
(3.37)
e −1
is the phonon distribution function. We can use these averages, together
with the variable transformation (3.21) and find for the time-ordered Green
function for t > 0
Diα;jβ (t > 0) = −ihuiα (t)ujβ (0)i
X ~

~
= −i vαλ vβλ eik·(~ri −~rj )
2MNω~kλ
~kλ
× (hb~λk (t)b~kλ† (0) + bλ†

−~k
(t)bλ−~k (0)i), (3.38)
25
where we omitted terms that give zero after thermal averaging. Performing
the thermal average in Eq. (3.38), and repeating the same analysis for t < 0,
we find
X ~

~
Diα;jβ (t) = −i vαλ vβλ eik·(~ri −~rj )
2MNω~kλ
~kλ
× hn~kλ ieiω~kλ |t| + hn~kλ + 1ie−iω~kλ |t| .

(3.39)
Similarly, for the retarded and advanced Green functions, one finds
X ~

~
R
Diα;jβ (t) = −θ(t) vαλ vβλ eik·(~ri −~rj ) sin(ω~kλ t), (3.40)
MNω~kλ
~kλ

A
X ~ ~
Diα;jβ (t) = θ(−t) vαλ vβλ eik·(~ri −~rj ) sin(ω~kλ t). (3.41)
MNω~kλ
~kλ
Performing a Fourier transform to time, we have

X ~

~
R
Diα;jβ (ω) = vαλ vβλ eik·(~ri −~rj )
2MNω~kλ
~kλ

1 1
× − . (3.42)
ω − ω~kλ + iη ω + ω~kλ + iη
The spectral density can now be found by taking the imaginary part of
R
D (ω),
X ~

~
Aiα;jβ (ω) = 2π vαλ vβλ eik·(~ri −~rj )
2MNω~kλ
~kλ

× δ(ω − ω~kλ ) − δ(ω + ω~kλ ) . (3.43)
The temperature Green function can be found using Eq. (2.50) or by
direct calculation using the imaginary-time dependence of the creation and
annihilation operators. Here we use the latter method. Starting point is the
general expression for the dependence of the phonon creation and annihilation
operators on the imaginary time τ . For 0 ≤ τ ≤ ~/T one has
′
hb~λk (τ )b~kλ′ (0)i = 0, (3.44)
λ′ †
hb~kλ† (τ )b~k′ (0)i = 0, (3.45)
′
hb~kλ† (τ )b~kλ′ (0)i = δ~k~k′ δλλ′ eω~kλ τ hn~kλ i, (3.46)
λ′ †
hb~λk (τ )b~k′ (0)i = δ~k~k′ δλλ′ e−ω~kλ τ hn~kλ + 1i, (3.47)
26
Calculation of the temperature Green function then yields, for 0 ≤ τ ≤ ~/T ,
Diα;jβ (τ ) = −huiα (τ )ujβ (0)i

X ~

~
= − vαλ vβλ eik·(~ri −~rj )
2MNω~kλ
~kλ
eω~kλ τ e−ω~kλ τ

× + . (3.48)
e~ω~kλ /T − 1 1 − e−~ω~kλ /T
It is easily verified that Diα;jβ (0) = Diα;jβ (~/T ), so that this result can be
extended periodically to the entire imaginary time axis. Fourier transform
to the imaginary time τ yields
X ~ ~ 1
Diα;jβ (iΩn ) = − vαλ vβλ eik·(~ri −~rj ) 2 . (3.49)
MN ω~kλ + Ω2n
~kλ
3.3.2 Calculation using the equation of motion

A different method to calculate the Green function is to use the “equation
of motion”. As an illustration of this method, we repeat the calculation of
the phonon temperature Green function below.
For the Heisenberg operators uiα (τ ) and piα (τ ) the equations of motion
are
1 i
∂τ uiα (τ ) = [H, uiα ]− = − piα , (3.50)
~ M
1 X
∂τ piα (τ ) = [H, piα ]− = i Vα;β (~ri − ~rj )ujβ . (3.51)
~ jβ
Then, using the definition of the temperature Green function
Diα;jβ (τ ) = −θ(τ )huiα (τ )ujβ (0)i − θ(−τ )hujβ (0)uiα (τ )i, (3.52)
we find
∂τ Diα;jβ (τ ) = −δ(τ )huiα (τ )ujβ (0) − ujβ (0)uiα (τ )i

i
+ θ(τ )hpiα (τ )ujβ (0)i
M
i
+ θ(−τ )hujβ (0)piα (τ )i. (3.53)
M
27
The first term on the r.h.s. of Eq. (3.53), multiplying the delta function δ(τ ),
is zero. Taking one more derivative of the remaining two terms, we find
i
∂τ2 Diα;jβ (τ ) = δ(τ )hpiα (τ )ujβ (0) − ujβ (0)piα (τ )i
M
i
+ θ(τ )h∂τ piα (τ )ujβ (0)i
M
i
+ θ(−τ )hujβ (0)∂τ piα (τ )i
M
~ 1 X
= δ(τ )δij δαβ + Vα;γ (~ri − ~rl )Dlγ;jβ (τ ). (3.54)
M M lγ
Rearranging terms, we find that D(τ ) satisfies an inhomogeneous differential

equation8
X 1

~
2
∂τ δil δαγ − Vα;γ (~ri − ~rl ) Dlγ;jβ (τ ) = δ(τ )δij δαβ . (3.55)
lγ
M M
This equation can be solved by Fourier transform to the imaginary time τ ,

X 1

~
2
− Ωn δli δγα + Vα;γ (~ri − ~rl ) Dlγ;jβ (iΩn ) = δij δαβ . (3.56)
lγ
M M
The remaining problem of dealing with the lattice indices and the coordi-
nate directions can be tackled by a further Fourier transform. Based on our
previous analysis of the normal modes, we can make use of the equation
1 1 X λ λ i~k·(~ri −~rl ) 2
Ω2n + Vα;γ (~ri − ~rl ) = vα vγ e 2
(Ωn + ω~kλ ) (3.57)
M N
~kλ
and the completeness relation

1 X λ λ′ i~rl ·(~k−~k′ ) ′
vγ vγ e = δ λλ δ~k~k′ , (3.58)
N lγ
to find
~ X λ λ i~k·(~ri −~rj ) 1
Diα;jβ (iΩn ) = − v~kα vβ~k e 2 2
, (3.59)
MN Ωn + ω~kλ
~kλ
in agreement with Eq. (3.49) above.

8
Note that DS have a minus sign on the right hand side, because definition of the
temperature Green function differs a minus sign from ours.
28
3.4 Neutron scattering cross-section
As an application of the use of Green functions, Doniach and Sondheimer
consider inelastic neutron scattering. An important quantity to calculate in
a scattering experiment is the differential cross section. It tells us the amount
of scattered neutrons that emerge from the sample at a given energy ~ω and
solid angle Ω.
In the first Born approximation, the differential scattering cross section
per unit solid angle and per unit energy is9
d2 σ X k ′ M 2 D E
= |h~k ′, f |H ′|~k, ii|2 δ(ω + Ei − Ef ), (3.60)
dΩdω f
k 2π
where ~k is the wavevector of the incoming neutron, ~k ′ is the wavevector of

the outgoing neutron, M is the reduced mass of the neutron, i and f refer
to initial state and final state, respectively, H ′ is the interaction between
the neutron and the target, and the outer brackets h. . .i indicate a thermal
average over the initial states i of the target.
Doniach and Sondheimer describe how Eq. (3.60) can be rewritten as
2 ∞
d2 σ k′

M
Z
2
= |Vq~| dteiωt F (~q, t), (3.61)
dΩdω 2πk 2π −∞
where ~q = ~k − ~k ′ and the correlation function F (~q, t) is defined as

X
ei~q·(~rl −~rj ) e−i~q·~uj (t) ei~q·~ul(0) .

F (~q, t) = (3.62)
j,l
In Eq. (3.62), the summation is over the lattice sites j and l and the opera-
tors ~uj (t), ~ul (0) are the corresponding displacements of the lattice atoms in
the Heisenberg picture. The equilibrium positions of the lattice atoms are
denoted ~rj and ~rl . In Eq. (3.61), Vq~ is the Fourier transform of the potential
that describes how a neutron scatters off a single atom. The combined effect
of many atoms in a vibrating lattice is described by the correlation function
F (~q, t).
9
This result is derived in standard texts on quantum mechanics, e.g., Quantum Me-
chanics, by A. Messiah, North-Holland (1961).
29
For small ~q, we now write
D E
ul (0))+ 12 [~
X
i~
q ·(~
rl −~
rj ) −i~
q ·(~
uj (t)−~ q ·~
uj (t),~
q ·~
ul (0)]−
F (~q, t) = e e , (3.63)
j,l
where the expression in the exponent is correct up to corrections of order q 3 .

In order to perform the thermal average, we use the cumulant expansion,10
and obtain
X 1 2 1 2
F (~q, t) = ei~q·(~rl −~rj )− 2 h(~q·~uj (t)) i− 2 h(~q·~ul (0)) i+h(~q·~uj (t))(~q·~ul (0))i , (3.64)
j,l
again up to corrections of order q 3 in the exponent.11

The second and third terms in the exponent do not depend on j, l, or t;
they form an overall prefactor known as the Debije-Waller factor. Writing
the Debije-Waller factor as e−2W , we may calculate W from the time-ordered
phonon Green function
iX
W = qα qβ lim Djα;jβ (t). (3.65)
2 αβ t↓0
Substituting the exact result (3.39) for the time-ordered Green function, we
obtain
XX ~

W = qα qβ vαλ vβλ coth(~ω~kλ /2T ). (3.66)
αβ
4MNω ~kλ
~kλ
We further expand F in the fourth term of the exponent, and separate F

into an elastic and inelastic contribution,
F (~q, t) = F el (~q, t) + F inel (~q, t) (3.67)

10
According to the cumulant expansion, any average of the form hexp(A)i can be cal-
culated as
heA i = exp hAi + 12 var A + . . . ,

where var A = hA2 i− hAi2 is the variance (i.e., the second cumulant) and the dots indicate
higher cumulants. If the probability distribution of A is Gaussian, the first two terms in
the cumulant expansion are sufficient.
11
In fact, Eqs. (3.63) and (3.64) are exact because the Hamiltonian is quadratic in the
displacements ~ui , i = 1, . . . , N .
30
with
X
F el (~q, t) = e−2W ei~q·(~rl −~rj ) (3.68)
j,l
X X
inel −2W
F (~q, t) = e ei~q·(~rl −~rj ) qα qβ hujα(t)ulβ (0)i. (3.69)
j,l αβ
Fourier transform of F el gives a term proportional to δ(ω) in the scattering

cross section, corresponding to elastic scattering from the lattice without the
excitation of phonons. Fourier transform of F inel gives
Z ∞ X X
dteiωt F inel (~q, t) = ie−2W ei~q·(~rl −~rj ) >
qα qβ Djα;lβ (ω), (3.70)
−∞ j,l αβ
>
where Djα;lβ (ω) is the phonon greater Green function, see Eq. (3.31) above.
>
Using Eqs. (2.46) together with the exact result (3.43) to calculate Djα;lβ (ω),
we find
Z ∞ XX π~eω/2T
dteiωt F inel (~q, t) = Ne−2W qα qβ vαλ vβλ
−∞ 2Mω sinh(ω/2T )
~kλ αβ

× δ(ω − ω~kλ ) − δ(ω + ω~kλ ) . (3.71)
3.5 Calculation of the phonon Green function using

diagrammatic perturbation theory
In the harmonic approximation, the phonon problem is exactly solvable.
However, most cases of interest are not exactly solvable. In order to demon-
strate the techniques used for a more general calculation, Doniach and Sond-
heimer return to the calculation of the phonon Green function.
In this section, we put ~ = 1, restoring ~ in the final formulas only.
Following Doniach and Sondheimer, we ignore the vector nature of the
displacement ~ui , i = 1, . . . , N, and consider a scalar displacement ui instead,
with the Hamiltonian
X p2 1X
i
H= + uiuj V (~ri − ~rj ). (3.72)
i
2M 2 i,j
We separate the potential energy in diagonal and off-diagonal parts,

H = H0 + H1 (3.73)
31
with
X p2 1

i 2 2
H0 = + Mω0 ui , (3.74)
i
2M 2
1X
H1 = ui uj V ′ (~ri − ~rj ), (3.75)
2 ij
where we abbreviated Mω02 = V (0) and

′ V (~r) if ~r 6= 0,
V (~r) = (3.76)
0 if ~r = 0.
Below, we’ll consider H1 as a “perturbation”, although there is no param-

eter that guarantees that this term is small. In the absence of the “perturba-
tion”, the calculation of the temperature Green function is straightforward,
eω0 τ e−ω0 τ

0 ~δij
Dij (τ ) = − + . (3.77)
2Mω0 e~ω0 /T − 1 1 − e−~ω0 /T
The purpose of the calculation is to develop a systematic way to include

the perturbation H1 , without resorting to the exact solution of the previous
sections.
In order to incorporate the “perturbation” H1 we switch to an imaginary-
time version of the so-called “interaction picture” of quantum mechanics.
The standard “interaction picture” is a mixture of the Heisenberg and Schrö-
dinger pictures: the time-dependence of the operators is given by the unper-
turbed Hamiltonian H0 , while the remaining time dependence is encoded in
the wavefunction. For imaginary times, this procedure is implemented with
the introduction of an imaginary time “evolution operator” for the quantum
state,
′ ′
U(τ, τ ′ ) = eτ H0 e−(τ −τ )(H0 +H1 ) e−τ H0 . (3.78)
The middle factor in Eq. (3.78) represents the exact time evolution with the
total Hamiltonian H = H0 + H1 , whereas the right and left factors “offset”
for the time-evolution with the unperturbed Hamiltonian H0 which will be
encoded in the operators.
The operator U satisfies the evolution equation
∂U
= −H1 (τ )U(τ, τ ′ ), (3.79)
∂τ
32
where the time-dependence of H1 is defined with respect to the unperturbed
Hamiltonian H0 only,
H1 (τ ) = eH0 τ H1 e−H0 τ . (3.80)
The evolution operator U(τ, τ ′ ) has the group property
U(τ, τ ′ )U(τ ′ , τ ′′ ) = U(τ, τ ′′ ) (3.81)
and U(τ, τ ) = 1. However, note that U(τ, τ ′ ) is not a unitary operator, unlike
the evolution operator in a real-time formalism.
The solution of the evolution equation (3.79) can be written as
Z τ
′
U(τ, τ ) = 1 − dτ1 H1 (τ1 )
τ′
Z τ Z τ1
+ dτ1 dτ2 H1 (τ1 )H1 (τ2 ) + . . . . (3.82)
τ′ τ′
It is important to note that in Eq. (3.82) the time arguments of the pertur-
bation H1 always appear in descending order if τ > τ ′ . Hence, formally, Eq.
(3.82) may be written as
Z τ
′ ′′ ′′
U(τ, τ ) = Tτ exp − H1 (τ )dτ , (3.83)
τ′
where Tτ denotes τ -ordering.

With the help of the evolution operator U(τ, τ ′ ) a thermal average with
respect to the full Hamiltonian H = H0 + H1 can be expressed in terms
of a thermal average with respect to the unperturbed Hamiltonian H0 only.
That is a useful transformation if the Hamiltonian H0 is simple, whereas
the full Hamiltonian H0 + H1 is complicated. For example, for τ > τ ′ the
temperature Green function Dij (τ ) can be written as
′ ′
′ tr e−(H0 +H1 )/T e(H0 +H1 )τ ui e−(H0 +H1 )τ e(H0 +H1 )τ uj e−(H0 +H1 )τ
Dij (τ, τ ) =
tr e−(H0 +H1 )/T
−H0 /T
tr e U(1/T, τ )ui (τ )U(τ, τ ′ )uj (τ ′ )U(τ ′ , 0)
= , (3.84)
tr e−H0 /T U(1/T, 0)
where the operator ui(τ ) is defined as
ui (τ ) = eH0 τ ui e−H0 τ , (3.85)
33
in analogy to Eq. (3.80). Similarly, for τ < τ ′ we find
tr e−H0 /T U(1/T, τ ′ )uj (τ ′ )U(τ ′ , τ )ui (τ )U(τ, 0)
Dij (τ, τ ′ ) = . (3.86)
tr e−H0 /T U(1/T, 0)
We can summarize Eqs. (3.84) and (3.86) with the concise expression
hTτ [U(1/T, 0)ui (τ )uj (τ ′ )]i0
Dij (τ, τ ′ ) = , (3.87)
hTτ U(1/T, 0)i0
where Tτ [. . .] denotes τ -ordering and the brackets h. . .i0 denote a thermal
average with respect to the unperturbed Hamiltonian H0 .
The reason why it is favorable to express the temperature Green function
in terms of an thermal average with respect to the unperturbed Hamiltonian
H0 is that such an average can be done with the help of Wick’s theorem.
Wick’s theorem applies to a time-ordered average with respect to a Hamilto-
nian that is quadratic in creation/annihilation operators. According to the
Wick theorem, the average of any product of creation/annihilation operators
can be found by forming all pairs of operators and replacing these by their
average. By definition, the average of a pair of operators is the unperturbed
temperature Green function D 0 . Hence, for the case of the phonon Green
function, Wick’s theorem states that
n X
1
hTτ ui1 (τ1 ) . . . ui2n (τ2n )i0 = − Di0P (1) ,iP (2) (τP (1) , τP (2) ) × . . .
2 P
× Di0P (2n−1) ,iP (2n) (τP (2n−1) , τP (2n) ), (3.88)
where the summation is over all permutations P of the numbers 1, 2, . . . , 2n.
The factor (−1)n accounts for the minus sign in the definition of Di,j ; the
factor 2−n accounts for the possible exchanges 1 ↔ 2, 3 ↔ 4, . . . , (2n − 1) ↔
2n. For a proof of the Wick theorem, see the book Methods of quantum field
theory in statistical physics, by A. A. Abrikosov, L. P. Gorkov, and I. E.
Dzyaloshinski, Dover (1963).
For the first terms in a series expansion of the numerator in Eq. (3.87)
one thus finds
− hTτ [U(1/T, 0)ui (τ )uj (τ ′ )]i0 =
Z 1/T
0 ′ 1X ′
= Dij (τ, τ ) + V (~rl − ~rk ) dτ1 Dij0 (τ, τ ′ )Dkl 0
(τ1 , τ1 )
2 kl 0
0
(τ, τ1 )Dlj0 (τ1 , τ ′ ) + Dil0 (τ, τ1 )Dkj
0
(τ1 , τ ′ ) + . . . .

+ Dik (3.89)
34
numerator:
τ’j τi
0th order
τ’j τi τ1k τ1k

τ’j τi τ’j τi
1st order τ1k +
V’
kl +
V’
kl
V’
kl
τ1l
τ1l τ1l
denominator: τ1k
V’
kl
1st order
τ1l
Figure 3: Diagrammatic representation of Eqs. (3.89) and (3.90).
Similarly, the denominator gives

1/T
1X ′
Z
0
hTτ U(1/T, 0)i0 = 1+ V (~rl − ~rk ) dτ1 Dkl (τ1 , τ1 ) + . . . .
2 kl 0
(3.90)
It is possible to represent these terms diagrammatically, see Fig. 3. With the
help of such a diagrammatic representation, higher-order terms can be easily
classified.
For the terms contributing to the numerator to second order, we dis-
criminate two types of contributions: disconnected diagrams (the left-most
one) and connected diagrams (the middle and right diagrams in Fig. 3). In
fact, one can easily convince oneself that the disconnected diagrams in the
perturbative expansion of the numerator cancel against the diagrammatic
expansion of the numerator. Thus, for the calculation of D we can ignore
the denominator if we keep connected diagrams only.
For this calculation, the two connected diagrams that contribute to first
order are equal, because Vij = Vji. Moreover, the unperturbed Green func-
tion is diagonal in the site indices i and j, cf. Eq. (3.77) above. With this
in mind, and with a slight change of diagrammatic notation, we write the
diagrammatic expansion of the phonon Green function as in Fig. 4. Note
that for the higher order diagrams (second order and higher), the combina-
torial factors arising from the ways in which one can assign the indices in
the diagram cancel the factorials in the Taylor expansion of the exponent
35
τ’ τ1 τ τ’ τ1 τ2 τ
τ’ τ j i j k i
j=i + + + ...
V’
ij V’
kj
V’ik
Figure 4: Diagrammatic representation the diagrams that contribute to the

phonon Green function.
and the prefactor 1/2 in the definition of the perturbation H1 , cf. Eq. (3.75).
Hence, every order has unit weight.
Summing over all orders, we then find
∞
X X Z 1/T Z 1/T
′
Dij (τ, τ ) = Dii0 (τ, τ ′ ) δij + dτ1 . . . dτn Dii0 (τ, τ1 )
n=1 l1 ,...,ln−1 0 0
× V ′ (~ri − ~rl1 )Dl01 l1 (τ1 , τ2 ) . . . V ′ (~rln−1 − ~rj )Djj

0
(τn , τ ′ ).
(3.91)
We note that Eq. (3.91) contains a convolution of unperturbed Green func-

tions. Taking the Fourier transforms with respect to imaginary time and
position,
1 1/T
Z X ~
~
D(iΩn , k) = dτ eiΩn τ e−ik·(~ri −~rj ) Dij (τ ), (3.92)
N 0 i,j
all convolutions transform into simple products,

∞
D(iΩn , ~k) = D 0 (iΩn , ~k)n+1 (V~k′ )n
X
n=0
D 0 (iΩn , ~k)
= , (3.93)
1 − D 0 (iΩn , ~k)V ′ ~k
where V~k′ is the Fourier transform of the perturbation,
~
X
V~k′ = eik·~r V (~r)
~
r6=0
= V~k − Mω02
= Mω~k2 − Mω02 . (3.94)
36
Substitution of
1
D 0 (iΩn , ~k) = − , (3.95)
M(Ω2n+ ω02)
yields
1 1
D(iΩn , ~k) = − 2 ′
=− . (3.96)
M(Ω2n + ω0 ) + V~k M(Ωn + ω~k2 )
2
This is the same result as that obtained in Eq. (3.49).

The calculation we just performed is a good example of the use of dia-
grammatic perturbation theory. We collected all parts of the Hamiltonian
that we knew how to deal with in the “unperturbed Hamiltonian” H0 and
treated the remaining terms in the Hamiltonian as a perturbation. We then
expanded the Green function in the perturbation using the diagrammatic
representation as a tool to organize the expansion. This problem was spe-
cial, because we could sum the perturbation theory to all orders. Quite often,
that is not the case, and we can only include a subset of diagrams.
3.6 Diagrammatic calculation of the phonon Free en-

ergy
As another application of the diagrammatic technique, we calculate the free
energy of the lattice vibrations. Of course, this problem can be solved exactly
using the normal mode representation. Here, we’ll show how one can obtain
the same answer using the perturbation methods of the previous section.
As before, we ignore the vector nature of the displacement and momen-
tum operators and separate the Hamiltonian into an unperturbed part and
a perturbation, see Eq. (3.73). Calculation of the free energy for the unper-
turbed part is simple. The corresponding partition function is that for N
harmonic oscillators of frequency ω0 ,
N
1
Z0 = , (3.97)
2 sinh(~ω0/2T )
so that the corresponding Free energy reads
F0 = −T ln Z0 = NT ln[2 sinh(~ω0 /2T )]. (3.98)
Using the definition (3.78) of the imaginary-time evolution operator, the

partition function Z for the total Hamiltonian H = H0 +H1 can be expressed
37
j=i j=i k=l i
τ
1 2 τ’ + ...
1+ τ + τ τ’ + 1! j
V’ 2! V’ V’ 2!
ii ii kk V’ij V’
ij
Figure 5: Diagrammatic representation the diagrams that contribute to the

phonon partition function.
as
Z tr e−H0 /T U(1/T, 0)
= = hU(1/T, 0)i0, (3.99)
Z0 tr e−H0 /T
where the brackets h. . .i denote a thermal average with respect to the unper-
turbed Hamiltonian H0 .
We now expand the evolution operator U(1/T, 0) in powers of the per-
turbation H1 in order to calculate the free energy difference
∆F = F − F0 = −T ln(Z/Z0) = −T lnhU(1/T, 0)i0 . (3.100)
A diagrammatic expansion of hU(1/T, 0)i0 can be obtained using the rules
derived in the previous section. In the first few orders, we find the diagrams
shown in Fig. 5. As before, we distinguish two types of diagrams: connected
diagrams (all but the third diagram in Fig. 5) and disconnected diagrams
(the third diagram in Fig. 5). Upon taking the logarithm, the disconnected
diagrams cancel, and one is left with the connected diagrams only.
The combinatorial factor for the connected diagrams is slightly more com-
plicated than for the calculation of the phonon Green function. You can verify
that, for the nth order diagram, there are 2n−1 (n − 1)! connected diagrams.
Combining this combinatorial factor with the factor (−1)n /n! from the Tay-
lor expansion of the exponent and the prefactor 1/2 in the perturbation H1 ,
we find that the free energy is
∞ Z 1/T
X 1 X
∆F = −T dτ1 . . . dτn Di01 ,i1 (τ1 − τ2 )V ′ (~ri1 − ~ri2 )
n=1
2n i ,...,i 0
1 n
× Di02 ,i2 (τ2 − τ3 ) . . . Di0n ,in (τn − τ1 )V ′ (~rin − ~ri1 ). (3.101)

In order to make progress, we change to Matsubara frequencies, and perform
a Fourier transform with respect to the lattice coordinate ~ri . We then find
∞
XXX 1 n
∆F = −T V~k′ D~k0 (iωm )
m n=1
2n
~k
38
C1 C1
C2 C2
−ω −ω0 ω0 ω
k k
Figure 6: Integration contours for the evaluation of the Matsubara sum of Eq.
(3.102).
!
2
T XX ωm + ω~k2
= ln 2 + ω2
. (3.102)
2 m
ωm 0
~k
At zero temperature, one can replace the summation over Matsubara

frequencies by an integration. Since the zero temperature free energy equals
the ground state energy EG , we find the result
!
ω 2 + ω~k2
Z
~ X
∆EG = dω ln
4π ω 2 + ω02
~k
~X
= (ω~k − ω0 ). (3.103)
2
~k
For finite temperatures one has to perform the summation over Matsubara
frequencies. Hereto we represent the summation as an integration in the
complex plane, as shown in Fig. 6,
!
2 2
−z + ω
Z
~ X ~k
∆F = dz coth(z~/2T ) ln . (3.104)
8πi C1 −z + ω02
2
~k
To see that Eq. (3.104) is correct, note that the function coth z has poles at
the points z = 2πmiT /~, with m integer. Performing the integration along
the contour C1 picks out precisely these poles. The integrand is analytic
everywhere else in the complex plane, except for branch cuts on the real axis,
for ω02 < z 2 < ω~k2 , see Fig. 6. Hence we can deform the integration contour
C1 into a contour C2 that is around the branch cuts instead of around the
39
Matsubara poles. The real part of the logarithm is continuous at the branch
cuts, so that it does not contribute to the integral. The imaginary part jumps
by 2π at the branch cuts. Adding contributions from both branch cuts, we
find
~ X ω~k
Z
∆F = dω coth(~ω/2T )
2 ω0
~k
X sinh(~ω~ /2T )
k
= T ln . (3.105)
sinh(~ω0 /2T )
~k
Combining the result (3.105) for ∆F with that for the unperturbed free
energy F0 , Eq. (3.98), we find
X
F =T ln sinh(~ω~k /2T ) . (3.106)
~k
This is the same answer as one would have obtained from the normal-mode
analysis.
40
3.7 Exercises
Exercise 3.1: Jellium model for phonons
Sometimes, one wants a simplified description of phonons without reference

to the lattice. This may be the case, e.g., if one is interested in long-
wavelength phenomena only, or if one only needs an order-of-magnitude es-
timate of the phonon contribution to a certain physical observable. Such
“structureless” models that do not refer to the underlying crystal lattice are
called “jellium models”.
We start with a description of lattice ions by means of their charge den-
sity ρion (~r), which is taken to be a continuous function of the coordinate
~r. We first treat the electron gas as a negatively charged static background
that neutralizes the positive charge of the ions. Then any deviation from
equilibrium δρion = ρion − ρ0ion is associated with an electric field
~ · E = 1 δρion ,
∇ (3.107)
ǫ0
where ǫ0 is the electric permittivity of vacuum. The electric field, in turn
exerts a force field on the ions,
f~ = ρ0ion E.
~ (3.108)
(a) Show that the ion charge density satisfies the equation
∂2 Zeρ0ion
M δρion + δρion = 0, (3.109)
∂t2 ǫ0
where M is the ionic mass and Ze is the charge on one ion.
(b) Describe the possible excitations of the ion “jellium”. These are not
phonons. Why not?
The physical cause that the jellium excitations we found are not phonons
is the long range of the Coulomb interaction. In fact, those excitations are
known as “plasmon” excitations. They do not exist for a real ion lattice, but
they are relevant for electrons.
In a real material, the long range Coulomb interaction between the ions is
screened by the electrons. Since electrons are much lighter than the phonons,
41
they react quickly to any change in the ion density. In fact, to a good ap-
proximation we can assume that the electrons follow the ion density instanta-
neously, maintaining charge neutrality throughout the material at all times.
As a consequence, the energy price for a perturbation δρion of the ion
density is not the Coulomb energy, but the kinetic energy cost of increasing
or decreasing the electron density: If the ionic charge density is changed by
an amount δρion , the electronic charge density needs to be changed by −δρion
in order to maintain charge neutrality. The energy cost associated with the
change of electron density leads to a restoring force for the ions.
(b) The easiest way to estimate that restoring force, is to view it as an

“electron pressure”. Show that, for a free electron gas, the ground
state energy for N electrons confined to a volume V is
5/3
~2 (3π 2 N)
Ee = . (3.110)
10π 2 mV 2/3
(c) Use your answer to (b) to argue that the electronic pressure equals
~2 2
5/3
pe = 3π ρe . (3.111)
15π 2m
(d) The ionic motion is driven by the gradient of pe . Show that the change
δρion obeys the wave equation
∂2 2 ~ 2 δρion = 0.
M δρion − ZεF ∇ (3.112)
∂t2 3
(e) From here, derive the dispersion relation for phonons in the jellium
model. What is the phonon velocity? This expression for the phonon
disperson is known as the “Bohm-Staver law”.
(f) In the jellium model, you find only one phonon mode per wavevector.
For a real lattice, there are three phonon modes per wavevector: one
longitudinal mode and two transverse modes. Is the single mode you
find in the jellium model a longitudinal mode or a transverse mode?
What happened to the other two modes of the lattice model?
42
Exercise 3.2: Free energy
In this exercise we consider an alternative method to calculate the phonon

free energy using diagrammatic perturbation theory. This alternative method
uses the diagrammatic rules for Green functions, and, thus, avoid the ap-
pearance of the nontrivial combinatorial factors that appeared in the direct
calculation of the free energy in Sec. 3.6.
In this calculation we consider a Hamiltonian H(λ) in which the pertur-
bation H1 is multiplied by a coupling constant λ,
H(λ) = H0 + λH1 . (3.113)
Hence, H(0) corresponds to the unperturbed Hamiltonian H0 whereas H(1)

corresponds to the full Hamiltonian H0 + H1 . For this Hamiltonian, the Free
energy becomes a function of λ
F (λ) = −T ln Z(λ), Z(λ) = tr e−(H0 +λH1 )/T . (3.114)
Taking a derivative to λ, we find

∂F (λ) tr H1 e−(H0 +λH1 )/T 1
= −(H +λH )/T
= hλH1 iλ (3.115)
∂λ tr e 0 1 λ
where the brackets h. . .iλ denote a thermal average with respect to the Hamil-
tonian H(λ).
Integrating Eq. (3.115) from λ = 0 to λ = 1 we find the free energy
difference ∆F between the full Hamiltonian H0 + H1 and the unperturbed
Hamiltonian H0 , Z 1
1
∆F = dλ hλH1 iλ . (3.116)
0 λ
Hence, if we know the average hλH1 iλ for all λ, we can find the change in
the free energy.
The average hλH1 iλ can be expressed in terms of the phonon temperature
Green function for the Hamiltonian H(λ),
λX ′ (λ)
hλH1 iλ = − V (~ri − ~rj ) lim Dij (τ ), (3.117)
2 ij τ ↓0
where
(λ)
Dij (τ ) = −hTτ ui(τ )uj (0)iλ . (3.118)
43
(a) Verify Eq. (3.117).
(b) Calculate ∆F using diagrammatic perturbation theory and Eq. (3.116).
(c) Verify that your result is the same as that of the direct perturbation
expansion of ∆F , order by order in the perturbation theory in H1 . How
are the combinatorial factors taken into account in the method you use
here?
The trick with the integration over the coupling constant can be applied
to other systems as well.
44
4 Scattering from a single impurity
In this section, we consider non-interacting electrons in the presence of an
impurity. This is another problem for which an exact solution is possible.
In the next two subsections we first describe this exact solution. In the
last subsection we then show how the same results can be obtained using
the many-electron Green function formalism and diagrammatic perturbation
theory.
In first quantization notation, the Hamiltonian we consider is
1 2
H=− ∂ + U(~r), (4.1)
2m ~r
where U(~r) is the impurity potential. In second quantization notation, the
Hamiltonian is written in terms of creation and annihilation operators ψσ† (~r)
and ψσ (~r) of an electron at position ~r and with spin σ,
1 X
Z XZ
† 2
H=− d~rψσ (~r)∂~r ψσ (~r) + d~rψσ† (~r)U(~r)ψσ (~r). (4.2)
2m σ σ
In the Heisenberg picture, the operators ψ † and ψ are time dependent, and
their time-dependence is given by
i
∂t ψσ (~r, t) = [H, ψσ (~r, t)]−. (4.3)
~
Using the anticommutation relations for the creation and annihilation oper-
ators ψσ† (~r) and ψσ (~r), the time-evolution can be rewritten in terms of the
first-quantized Hamiltonian in a form that is quite similar to the Schrödinger
equation,
i
∂t ψσ (~r, t) = − Hψσ (~r, t). (4.4)
~
4.1 Scattering theory and Friedel Sum Rule

4.1.1 General formalism
We first consider the problem of one electron in the presence of an impurity
with scattering potential U(~r). There are no other electrons present. We
drop reference to the spin index σ.
45
For the one-electron problem, the expectation values h. . .i in the Green
function definitions are taken with respect to the vacuum |0i. The temper-
ature Green function has no meaning in that case. In the absence of other
electrons, the retarded and time-ordered Green functions coincide,
GR (~r, ~r′ ; t) = −iθ(t)h0|[ψ(~r, t), ψ † (~r′ , 0)]− |0i
= −iθ(t)h0|Tt [ψ(~r, t)ψ † (~r′, 0)]|0i
= G(~r, ~r′ ; t). (4.5)
Using the equation of motion for the operator ψ(~r, t), one can derive an
equation of motion for the retarded Green function,
∂t GR (~r, ~r′ ; t) = −iδ(t)h0|ψ(~r, t)ψ † (~r′ , 0)|0i − iθ(t)h0|∂t ψ(~r, t)ψ † (~r′, 0)|0i
i
= −iδ(t)δ(~r − ~r′ ) − HGR (~r, ~r′; t). (4.6)
~
where H is the first-quantization form of the single-particle Hamiltonian
operating on the first argument of the Green function, see Eq. (4.4) above.
Performing a Fourier transform, one finds
(ω − ~−1 H)GR (~r, ~r′; ω) = δ(~r − ~r′ ). (4.7)
A differential equation like Eq. (4.7) requires the specification of boundary
conditions as |~r−~r′ | → ∞. Since the retarded Green function is zero for t < 0,
we require outgoing wave boundary conditions as |~r − ~r′ | → ∞. Similarly,
the advanced Green function GA the solution of Eq. (4.7) with ingoing wave
boundary conditions.
The advanced and retarded Green functions can be expressed in terms of
the eigenvalues εµ and eigenfunctions φµ (~r) of H,
X φµ (~r)φ∗µ (~r′ )
GR (~r, ~r′ ; ω) = , (4.8)
µ
ω − εµ + iη
X φµ (~r)φ∗µ (~r′ )
A ′
G (~r, ~r ; ω) = , (4.9)
µ
ω − εµ − iη
where η is a positive infinitesimal. One verifies that both the retarded and the
advanced Green functions are analytical in the upper and lower half planes
of the complex plane, respectively, and that they satisfy the relation
∂
tr G(ε) = − ln det G(ε). (4.10)
∂ε
46
In the absence of an impurity potential, the Hamiltonian is
~2 2
H0 = − ∂ . (4.11)
2m ~r
In the basis of plane wave states |~ki, the retarded and advanced Green func-
tions are
δ~k,~k′ δ~k,~k′
G~0R
k,~k ′
(ω) = , G~0A
k,~k ′ (ω) = . (4.12)
ω + iη − ε~k ω − iη − ε~k
In coordinate representation one has
~′ ′
0 ′ 1 X eik ·(~r−~r )
G (~r, ~r ; ω) = , (4.13)
V ′ ω ± iη − ε~k′
~k
where η is a positive infinitesimal and the + (−) sign applies to the retarded
(advanced) Green function. In order to further simplify this result, we define
k through εk = ω and linearize the spectrum,
ε~k′ = ω + v(|~k ′| − k), (4.14)
where v is the velocity. Then, performing the Fourier transform, one finds
Z 2π Z 1 Z ∞ ′ ′
0 ′ 1 ′2 ′ eik |~r−~r | cos θ
G (~r, ~r ; ω) = dφ d cos θ k dk
(2π)3 0 −1 0 ±iη − v(k ′ − k)
Z ∞ ik ′ |~ r′ |
r−~ ′ ′
1 ′ ′e − e−ik |~r−~r |
= k dk
(2π)2 i|~r − ~r′ | 0 ±iη − v(k ′ − k)
Z ∞ ′ ′
k ei(k+x/v)|~r−~r | − e−i(k+x/v)|~r−~r |
≈ dx
4π 2 i|~r − ~r′ |v −∞ ±iη − x
±ik|~ r′ |
r−~
m e
= − . (4.15)
2π~2 |~r − ~r′ |
This result can be derived without approximations for a quadratic disper-

sion relation ε~k = ~2 |~k|2 /2m. The derivation given here shows that the
result (4.15) is not restricted to a parabolic dispersion relation. For non-
′
infinitesimal η, the final result needs to be multiplied by e−η|~r−~r |/v .
Let us now add an impurity to the system, which is described by a the
potential U(~r). We want to write a solution |ψ~k i of the total Hamiltonian
47
H = H0 + U with incoming boundary condition corresponding to the unper-
turbed state |~ki. This solution can be written as12
|ψ~k i = |ki + G0R (ε~k )U|ψ~k i
= |ki + G0R (ε~ )U|~ki + G0R (ε~ )UG0R (ε~ )U|~ki + . . . ,
k k k
where the symbol U should be interpreted as the quantum operator cor-

responding to the potential U(~r). The series (4.16) is known as the Born
series. Truncating the series after the nth order in U yields the nth Born
approximation. Defining the T -matrix as
T (ω + ) = U + UG0R (ω)U + . . .
= U[1 − G0R (ω)U]−1
= U + UG0R (ω)T (ω + ), (4.16)
where ω + = ω + iη, η being a positive infinitesimal, we can rewrite this as
X T~ ′~ (ε~ + iη)
|ψ~k i = |ki + kk k
|k ′ i, (4.17)
ε~k − ε~k ′ + iη
′ ~k
with T~k′~k (ω) = h~k ′ |T (ω)|~ki. The T -matrix is related to the retarded Green
function GR of the total Hamiltonian H0 + U as
GR (ω) = G0R (ω) + G0R (ω)T (ω + )G0R (ω). (4.18)
Note that in this notation, where we have a summation over discrete set of
wavevectors, the T -matrix is of order 1/V . Using Eq. (4.15) one then finds
that the asymptotic form of the wavefunction of the scattered wave (4.17) is
i~k ′ ·~
r
i~k~ e mV
e r
+ f (~k ′ , ~k) , f (~k, ~k ′ ) = − T~~ ′ (ε~ + iη), (4.19)
r 2π~2 kk k
where the wavevector ~k ′ is defined as kr̂, r̂ being the unit vector in the
direction of ~r.
With this result we can now write expressions for two useful quantities
from collision theory: the differential cross section dσ/dΩ, where dΩ denotes
an element of solid angle, which is
dσ ~ ′ ~ 2 m2 V 2
= |f (k , k)| = 2 4 |T~k′ ,~k (ε~k + iη)|2 , (4.20)
dΩ 4π ~
12
You can find a good discussion of stationary scattering theory in section 37 of Quantum
Mechanics, by L. I. Schiff, McGraw-Hill (1968).
48
and the transition rate
2π
W~k′~k =|T~k′~k (ε~k + iη)|2 δ(ε~k′ − ε~k ). (4.21)
~
The transition rate gives the probability per unit time that a particle is
scattered out of plane wave state ~k into plane wave state ~k ′ . To lowest order
in the impurity potential U, Eq. (4.21) is nothing but the Fermi golden rule.
The total rate for scattering out of the state |~ki is found by summation of
Eq. (4.21) over all outgoing momenta,
1 2π X
= |T~k′~k (ε~k + iη)|2 δ(ε~k′ − ε~k ). (4.22)
τ~k ~ ′
~k
Taking the hermitian conjugate of Eq. (4.16) and substituting U = U † =

T − T † G0A U in the second term of Eq. (4.16), one finds
†
T = U + T † G0R T − T † G0A UG0R T . (4.23)

The first and third term on the right hand side are hermitian, so that
ℑT = T † ℑG0R T . (4.24)
Using Eq. (4.12), this result can be written as
1 X
T~k~k′ (ω + ) − T~k′~k (ω + )∗ = −π T~k∗1~k (ω + )T~k1~k′ (ω + )δ(ω − ε~k1 ).

(4.25)
2i
~k1
This result is known as the “generalized optical theorem”. The “standard”

optical theorem is obtained setting ~k ′ = ~k,
X
Im T~k~k (ω + ) = −π |T~k1~k (ω + )|2 δ(ω − ε~k1 ). (4.26)
~k1
The density of levels dN(ω)/dω is expressed in terms of the Green func-

tion as
dN(ω) X 1
= δ(ω − εµ ) = − Im tr GR (ω). (4.27)
dω µ
π
The impurity density of states (dN/dω)imp is defined as the difference between
the density of states with and without impurity. Using Eq. (4.10) above, one
can show that
dN 1 ∂δ(ω)
= , (4.28)
dω imp π ∂ω
49
where δ(ω) = arg det T (ω) is the so-called “total scattering phase shift”.
Zero eigenvalues of T , which correspond to states that do not scatter at all,
should be left out from the calculation of the total scattering phase shift.
Equation (4.28) is known as the “Friedel sum rule”. At zero temperature,
as a result of the introduction of the impurity, the number of electrons in a
many-electron system is changed by the amount
Z εF
dN δ(εF )
Nimp = dω = , (4.29)
−∞ dω imp π
where we took δ(−∞) = 0 and εF is the Fermi energy.
4.1.2 Example of an s-wave scatterer

Following Doniach and Sondheimer, we study the case U(~r) = uδ(~r) of a
delta-function potential at the origin in detail. Fourier transforming to the
basis of plane-wave states and using second-quantization notation, the po-
tential U is written as
uX †
U= ψ~k+~qψ~k . (4.30)
V
~k,~
q
Then the equation for the T -matrix becomes

u 1
T~k,~k′ (ω + iη) = , (4.31)
V 1 − uG0R (ω)
where we abbreviated
1 X 0R 1 X 1
G0R (ω) = G~k~k (ω) = . (4.32)
V V ω + iη − ε~k′
~k ~k
(A similar definition can be made for G0A (ω).) Note that the T -matrix is
isotropic: For a delta-function scatterer, scattering is equally probable in all
directions. Since T is isotropic, only one of the eigenvalues of T is nonzero.
The simplest way to calculate the corresponding phase shift δ is to note that
T (ω + iη) = T (ω − iη)∗ , so that
T (ω + iη) 1 − uG0 (ω − iη)

e2iδ(ω) = = . (4.33)
T (ω − iη) 1 − uG0 (ω + iη)
50
The real part of G0 (ω + iη) depends weakly on ω. Formally, for a parabolic
potential and a delta-function scatterer, Re G0 (ω + iη) is ultraviolet diver-
gent. The high-energy part of the energy integration can be cut off by in-
troducing a finite width of the conduction band, or by noting that for high
energies the approximation of a delta function scatterer breaks down. The
imaginary part of G0 (ω + iη) is related to the density of states ν0 (per unit
volume) at the energy ω in the absence of the impurity,
k2
Im G0 (ω + iη) = −πν0 (ω) = − , (4.34)
2π~v
Here v = ~−1 ∂ε/∂k is the velocity at wavevector ~k.
With the help of Eq. (4.34), we can express T in terms of δ and ν0 ,
1 ±iδ(ω)
T~k~k′ (ω ± iη) = − e sin δ(ω)
πν0 V
2π~v
= − 2 e±iδ(ω) sin δ(ω). (4.35)
k V
Hence, in this case we find for the differential cross section
dσ m2 v 2
= 4 2 sin2 δ(ω), (4.36)
dΩ k ~
and for the total scattering cross section (the integral of dσ/dΩ over the full
solid angle)
4πm2 v 2
σ= sin2 δ(ω). (4.37)
k 4 ~2
4.2 Alternative derivation of the Friedel Sum Rule

The Friedel sum rule relates the change ∆N in the total number of particles
as the result of the presence of an impurity to the phase shifts that particles
experience while scattering off the impurity. In terms of the T -matrix, the
Friedel sum-rule reads
δtot (εF )
∆N = , (4.38)
π
where the total phase shift δtot is defined as13
δtot = arg det T (ε + i0).
13
Strictly speaking, Eq. (4.38) and the derivation given below apply to potential scat-
tering only. One may, however, extend the Friedel sum rule to the case of scattering from
a dynamical impurity. See, e.g., chapter 5 of The Kondo problem to heavy fermions, by
A. C. Hewson.
51
You have seen a formal derivation of the Friedel sum rule in the previous
section. Here, we will consider a more elementary derivation of the Friedel
sum rule for the case of a spherically symmetric scatterer.
If a flux of noninteracting particles is elastically scattered on a spherically
symmetric potential at ~r = 0, the wave function at large values of r will be
of the form
eikr
Ψ ≈ eikz + f (θ) , (4.39)
r
where z is the direction of the incoming flux. The first term represents the
incoming flux (with wavenumber k), the second one the scattered wave. The
T -matrix is related to the function f (θ) as f (θ) = −(m/2π~2 )T~k~k′ , where
|~k| = |~k ′ | = k and θ is the angle between ~k and ~k ′ , see Eq. (4.19) above.
Since the scattered wave in Eq. (4.39) has axial symmetry about the
z-axis, it may be expanded in Legendre polynomials,
∞
X
Ψ= aℓ Pℓ (cos θ)Rkℓ (r) (4.40)
ℓ=0
The constants aℓ are just expansion coefficients, and the radial functions Rkℓ
satisfy the “radial Schrödinger equation”

−2 ∂ 2 ∂Rkℓ 2 ℓ(ℓ + 1) 2m
r r + k − − 2 U(r) Rkℓ = 0,
∂r ∂r r2 ~
where the energy eigenvalues are written as ~2 k 2 /2m, the energy of a free
particle, and U(r) is the scattering potential.
At distances sufficiently large that U(r) ≈ 0, the radial functions Rkℓ will
approach the asymptotic expansion
2
Rkℓ ≈ sin(kr − ℓπ/2 + δℓ ) (4.41)
r
Both the factor of 2 and the factor ℓπ/2 are introduced for convenience, so
that the asymptotic behavior of Rkℓ reduces to the solutions for Rkℓ when
U(r) vanishes identically, by putting δℓ = 0. The δℓ are the so-called scat-
tering phases. They depend on the scattering potential U(r) and approach
zero when U(r) vanishes.
The change in the number of states near the Fermi surface as a result
of the scattering potential generally depends on the δℓ ’s. To see this, we
52
first have to connect the general δℓ ’s in the expression (4.41) for the radial
functions in the expansion of the total wavefunction (4.40) to the function
f (θ) defined in (4.39). Apart from an overall normalization, the coefficients
aℓ in (4.40) are fixed by the requirement that the function Ψ − eikz is of the
form f (θ)eikr /r, i.e. has only terms of the form of an outgoing wave, eikr /r.
This requirement leads to the identification
∞
1 X
(2ℓ + 1) e2iδℓ − 1 Pℓ (cos θ).

f (θ) = (4.42)
2ik ℓ=0
You may derive Eq. (4.42) using the fact that for large r the plane wave eikz
can be expanded in terms of Legendre polynomials as
∞
ikz ikr cos θ 1 X
(2ℓ + 1)Pℓ (cos θ) (−1)ℓ+1 e−ikr + eikr .

e =e =
2ikr
ℓ=0
Let us now consider the change in the number of states in the presence of
the impurity. Hereto we imagine a large sphere of radius L about the origin,
and impose that the wavefunction vanishes on this sphere. In the absence of
the potential (δℓ = 0), Eq. (4.41) then yields
ℓπ ℓπ nπ
kL − = nπ =⇒ knℓ = + (4.43)
2 2L L
′ ′
In the presence of the scattering potential, the wavenumbers knℓ satisfy knℓ =
ℓπ/(2L) + nπ/L − δℓ /L, so we get for the change in wavenumber
′ δℓ
knℓ − knℓ = − (4.44)
L
Let us first consider the special case in which only δ0 is nonzero. The spacing
of successive values of kn0 is π/L [see Eq. (4.43)]; if we now consider a given
value of k, say k = kF , and imagine turning on the potential, then each
allowed k value shifts, according to Eq. (4.44), by an amount −δ0 (kF )/L.
Then the number of states k with k < kF increases by an amount
δ0 L δ0
Nimp =· = .
L π π
Now you are invited to prove the general Friedel sum rule (4.38), which reads,
in terms of the phase shifts δℓ ,
X
Nimp = π −1 (2ℓ + 1)δℓ (kF ).
ℓ
53
Using these results, one can also show that, at zero temperature, the density
of electrons far from a small spherically symmetric impurity shows oscillations
1
nimp (r) ≈ − cos[2kF r + δ0 (kF )] sin[δ0 (kF )], (4.45)
4π 2 r 3
where δ0 is a scattering phase shift evaluated at the Fermi energy.
4.3 Many-particle formulation

In this section, we’ll use the many-particle formalism to describe the electron
gas in the presence of an impurity.
We are interested in the electron density n(~r) of a gas of noninteracting
electrons in the presence of a potential U(~r). The electron density can be
expressed in terms of the many-electron temperature Green function, using
n(~r) = hψ † (~r)ψ(~r)i
= lim G(~r, ~r; τ ), (4.46)
τ ↑0
where the temperature Green function is defined with respect to the electron
creation and annihilation operators,
G(~r, ~r′; τ ) = −hTτ [ψ(~r, τ )ψ † (~r′ , 0)]i. (4.47)
We now calculate the many-electron temperature Green function using

diagrammatic perturbation theory. It is advantageous to change to the mo-
mentum representation,
1 1 X
Z
−i~k·~
r ~
ψ~k = 1/2
d~
r ψ(~
r )e , ψ(~
r ) = 1/2
ψ~k eik·~r ,
V V
~k
so that the temperature Green function in momentum representation reads
G~k;~k′ (τ ) = −hTτ [ψ~k (τ )ψ~k† ′ (0)]i

1
Z
~ ~′ ′
= d~rd~r′ G(~r, ~r′ ; τ )e−i(k·~r−k ·~r ) . (4.48)
V
The Hamiltonian is separated into an “unperturbed” part H0 and the “per-
turbation” H1 ,
H = H0 + H1 , (4.49)
54
with
X
H0 = (ε~k − µ)ψ~k† ψ~k , (4.50)
~k
XX †
H1 = Uq~ψ~k+~ ψ~ ,
q k
(4.51)
~k q~
where Uq~ is the Fourier transform of the scattering potential,

1
Z
Uq~ = d~rU(~r)e−i~q·~r . (4.52)
V
Here, we have included the chemical potential into the unperturbed Hamil-
tonian H0 .
As in the phonon case, it is not clear a priori that the perturbation H1
is small. Still, in the method of diagrammatic perturbation theory, it is
assumed that an expansion in H1 is possible, and that resummation of that
expansion gives the correct results. Since we are going to expand in H1 , it
is useful to calculate the temperature Green function for the “unperturbed”
Hamiltonian H0 ,
0 e−(ε~k −µ)τ e−(ε~k −µ)τ

G~k;~k ′ (τ ) = −δ~k,~k ′ θ(τ ) θ(τ ) + δ~k;~k ′ θ(−τ ) . (4.53)
1 + e−(ε~k −µ)/T 1 + e(ε~k −µ)/T
This result is obtained directly from the second-quantized form of the Hamil-
tonian H0 , using the Fermi-Dirac distribution function and the fact that the
imaginary-time dependence of the operators ψ~k and ψ~k† is
ψ(~k; τ ) = e−(ε~k −µ)τ ψ(~k; 0), ψ † (~k; τ ) = e(ε~k −µ)τ ψ † (~k; 0), (4.54)
for −~/T < τ < ~/T . The Fourier transform of the temperature Green
function is
0
δ~k,~k′
G~k;~k ′ (iω n ) = . (4.55)
iωn − (ε~k − µ)
We now calculate the Green function in the presence of the impurity using
diagrammatic perturbation theory. The calculation is done along the lines of
that of Sec. 3.5. We introduce the “interaction picture” evolution operator
′ ′
U(τ, τ ′ ) = eτ H0 e−(τ −τ )(H0 +H1 ) e−τ H0
Z τ
= Tτ exp − dτ1 H1 (τ1 ) , (4.56)
τ′
55
and use time-dependent operators for which the time-dependence stems from
H0 only,
H1 (τ ) = eτ H0 H1 e−τ H0 . (4.57)
Following the same steps as in Sec. 3.5, we then write the temperature Green
function as
hTτ [U(1/T, 0)ψ~k (τ )ψ~k† ′ (0)]i0
G~k;~k′ (τ ) = . (4.58)
hTτ U(1/T, 0)i0
The brackets h. . .i0 indicate an thermal average with respect to the unper-
turbed Hamiltonian H0 only.
The next step is to expand the evolution operator in powers of H1 and
calculate averages using the Wick theorem. In the language of the general
formulation of Sec. 2, the Wick theorem allows one to calculate an average
of the form hA1 (τ1 ) . . . A2n (τ2n )i0 , where all operators Ai , i = 1, . . . , 2n are
linear in fermion creation and annihilation operators and the average is taken
with respect to a Hamiltonian H0 that is quadratic in fermion creation and
annihilation operators. Clearly, every order in the perturbation expansion of
Eq. (4.58) satisfies these criteria. The Wick theorem states that
n X
1
hA1 (τ1 ) . . . A2n (τ2n )i0 = − (−1)P GA0 P (1) ,AP (2) (τP (1) , τP (2) ) × . . .
2 P
× GA0 P (2n−1) ,AP (2n) (τP (2n−1) , τP (2n) ), (4.59)
where the sum is over all permutations P of the numbers 1, 2, . . . , 2n and

(−1)P is the sign of the permutation. In our case, only Green functions
with one creation operator and one annihilation operator are nonzero. That
simplifies the set of allowed permutations, and we can write the Wick theorem
as
hψ~k† (τ1 ) . . . ψ~k† (τn )ψ~kn′ (τn′ ) . . . ψ~k′ (τ1′ )i0

1 n 1
X
= (−1) G~k1 ,~k′ (τ1 , τP (1) ) . . . G~k0n ,~k′
P 0 ′
(τn , τP′ (n) ). (4.60)
P (1) P (n)
P
With the help of the Wick theorem, we can now calculate the average
(4.58) to every order in H1 by drawing all possible diagrams. As in the case
of phonons, all disconnected diagrams cancel between numerator and denom-
inator, and one is left with connected diagrams only. Also, the combinatorial
56
τ τ1 τ’ τ τ1 τ2 τ’
τ τ’ = τ τ’ + k k’ + k k1 k’ + ...
k k’ k=k’
Figure 7: Diagrammatic representation of Eq. (4.61). The double arrow cor-

responds to the full Green function G~k,~k′ (τ, τ ′ ), single arrows correspond to the
unperturbed Green function G~0 ~ ′ (τ, τ ′ ), and the dashed lines with the stars cor-
k,k
respond to the impurity potential Uq~ . (The momentum argument of the impurity
potential is the difference in electron momenta at the scattering vertex.)
factor cancels the factorial in the Taylor expansion of the exponent in the
evolution operator U(1/T, 0), so that one sums over topologically different
diagrams only. We thus find
Z 1/T
0
G~k;~k′ (τ ) = G~k;~k (τ )δ~k,~k + dτ1 U(~k − ~k ′ )G~k;
0 0
~k (τ − τ1 )G~k ′ ;~k ′ (τ1 )
0
Z 1/T Z 1/T
~ ~
X
0
+ dτ1 dτ2 G~k;~k (τ − τ1 )U(k − k1 )
0 0 ~k1
× G~k01 ;~k1 (τ1 − τ2 )U(~k1 − ~k ′ )G~k0′ ;~k′ (τ2 ) + . . . . (4.61)
Diagrammatically this result may be represented as in Fig. 7. The double

arrow represents the full Green function, while the single arrows correspond
to the Green functions of the free electron Hamiltonian H0 . The dashed lines
with the stars correspond to scattering from the impurity potential U. The
series can be written as a self-consistent equation for G~k,~k′ (τ ),
XZ 1/T
G~k;~k′ (τ ) = 0
G~k;~k (τ )δ~k,~k ′ + 0
dτ1 G~k; ~ ~
~k (τ − τ1 )U(k − k1 )G~k1 ;~k ′ (τ1 ). (4.62)
~k1 0
This equation is called the “Dyson” equation. It is represented diagrammat-

ically in Fig. 8. The convolution with respect to the imaginary time variable
τ can be handled by Fourier transformation,
U(~k − ~k1 )G~k;

X
0 0
G~k;~k′ (iωn ) = G~k;~k (iω )δ
n ~k,~k ′ + ~k (iωn )G~k1 ;~k ′ (iωn ). (4.63)
~k1
The summation over the momentum ~k1 is more complicated. It can not
be carried out exactly for an arbitrary impurity potential. As a simple
57
τ τ1 τ’
τ τ’ = τ τ’ + k k1 k’
k k’ k=k’
Figure 8: Diagrammatic representation of the Dyson equation for impurity scat-

tering, Eq. (4.62).
ωn ωn
ωn = ωn + k k1 k’
k k’ k=k’
Figure 9: Diagrammatic representation of the Matsubara frequency representa-

tion of the Dyson equation for impurity scattering, Eq. (4.63). Notice that the
Matsubara frequency ωn is conserved at every scattering event.
example, we return to the case of a point scatterer, U(~r) = uδ(~r). In this

case, U(~k −~k1 ) = u/V does not depend on ~k and ~k1 , see Eq. (4.30). Summing
Eq. (4.63) over ~k, we find
  
X u X 0 X
G~k;~k′ (iωn ) = G~k0′ ;~k′ (iωn ) +  G~k;~k (iωn )  G~k;~k′ (iωn ) , (4.64)
V
~k ~k ~k
from which we derive

X G~k0′ ;~k′ (iωn )
G~k;~k′ (iωn ) = , (4.65)
1 − Vu ~k G~k;
P 0
~k ~k (iωn )
and, hence,
0 u 0 1
G~k;~k′ (iωn ) = G~k;~k (iωn )δ~k,~k ′ + G~k;~k (iωn ) G~k0′ ;~k′ (iωn ), (4.66)
V 0
1 − u G (iωn )
1 X 0
G 0 (iωn ) = G~k;~k (iωn ). (4.67)
V
~k
What remains is to perform the inverse Fourier transform,

X
G~k;~k′ (τ ) = T e−iωn τ G~k;~k′ (iωn ). (4.68)
n
58
C1 C1
C2
C2
Figure 10: Integration contours for the derivation of Eq. (4.69) and (4.70).
This is done with the help of the following trick. The summation in Eq.
(4.68) is written as an integration over the contour C1 , see Fig. 10,
1
Z
G~k;~k′ (τ ) = dze−zτ [tanh(z/2T ) − 1] G~k;~k′ (iωn → z). (4.69)
4πi C1
We are interested in the case of negative τ .14 Then we make use of the fact
that G is analytic for all frequencies away from the real and imaginary axes,
so that we may deform the contour C1 into the contour C2 , see Fig. 10. Note
that there is no contribution from the segments near infinity by virtue of
the factor [tanh(z/2T ) − 1] for Re z > 0 and by virtue of the factor e−zτ for
Re z < 0. We thus obtain
Z ∞
1 X
G~k;~k′ (τ ) = (±) dωe−ωτ [tanh(ω/2T ) − 1]G~k;~k′ (iωn → ω ± iη),
4πi ± −∞
Z ∞
e−(ε~k −µ)τ e−ωτ

u X ±1 1
= (ε~k −µ)/T
− dω ω/T ±
1+e V ± 2πi −∞ 1+e ω − (ε~k − µ)
1 1
× , (4.70)
1− u G 0 (ω ± ) ω± − (ε~k′ − µ)
where ω ± = ω ± iη and η is a positive infinitesimal. The first term in Eq.

(4.70) is the temperature Green function in the absence of the potential,
cf. Eq. (4.53). The second term is the correction that arises as a result of
scattering from the impurity at ~r = 0.
14
If we would have wanted to calculate the temperature Green function for positive τ ,
we would have used Eq. (4.69) with a factor [tanh(z/2T ) + 1] instead of [tanh(z/2T ) − 1].
59
We first use this result to calculate the impurity contribution to the total
electron density
X
0
Nimp = lim G~k;~k (τ ) − G~k;~k (τ )
τ ↑0
~k
u X ±1 X ∞

1 1 1
Z
= − dω ω/T ± 2
V ± 2πi −∞ 1+e (ω − (ε~k − µ)) 1 − uG 0 (ω ± )
~k
X ±1 Z ∞ 1 ∂ h i
= − dω 0 ±
log 1 − uG (ω ) . (4.71)
±
2πi −∞ 1 + eω/T ∂ω
This result can be rewritten in terms of the T matrix and the scattering
phase shifts. Using the relation (4.33) between Ḡ, the T -matrix, and the
scattering phase shift, we recover the Friedel sum rule,

1 ∂ 1
Z
Nimp = dω − δ0 (ω + µ). (4.72)
π ∂ω 1 + eω/T
At zero temperature, Eq. (4.72) simplifies to Eq. (4.29) above.

In order to find how the density change depends on the distance r to the
impurity site, we have to change variables from the momenta ~k and ~k ′ to the
positions ~r and ~r′. Hereto, we start again from the second term in Eq. (4.70)
and note that the ~k-dependence is through the factors 1/(ω + −(ε~k −µ)) only.
The corresponding Fourier transform has been considered in Sec. 4.1.1 and
the result for the asymptotic behavior as r → ∞ is
X ±1 Z ∞ m2 v
nimp (~r) = lim dω
τ ↑0
±
2πi −∞ 2πr 2k 2 ~3
sin δ(ω + µ)e−ωτ ±i[δ(ω+µ)+2(kF +ω/~vF )r]
× e .
1 + eω/T
In order to perform the ω-integration, we expand the phase shift δ(ω + µ)
around ω = 0 and keep terms that contribute to leading order in 1/r only.
The result is

sin δ(µ) ∂ 1
Z
nimp (~r) = − dω − cos[δ(µ) + 2(kF + ω/~vF )r].
4π 2 r 3 ∂ω 1 + eω/T
(4.73)
60
At zero temperature, this simplifies to Eq. (4.45). These density oscillations
are known as “Friedel oscillations”. At zero temperature, Friedel oscillations
fall off algebraically, proportional to 1/r 3 . At finite temperatures, the main
cause of decay are the thermal fluctuations of the phase shift 2ωr/~vF , which
are important for T & ~vF /r.
The final results (4.72) and (4.73) are to be multiplied by two if spin
degeneracy is taken into account.
61
4.4 Exercises
Exercise 4.1: Scattering from a single impurity in one dimension
In this exercise, we consider scattering from a delta-function impurity in one

dimension. The impurity has potential U(x) = uδ(x) and is located at the
origin.
It is our aim to calculate the full Green function G(x, x′ ; iωn ) in coordinate
representation. In order to calculate G(x, x′ ; iωn ), one may either use the
Green function in momentum space and Fourier transform back to coordinate
space, as we did in Sec. 4.3, or derive the Dyson equation in coordinate
representation. Following the latter method, you can verify that one finds
G(x, x′ ; iωn ) = G 0 (x, x′ ; iωn ) (4.74)

0 u
+ G (x, 0; iωn ) 0
G 0 (0, x′ ; iωn ).
1 − G (0, 0; iωn )u
(a) Calculate the temperture Green function in coordinate representation.
(b) Use your answer to (a) to calculate the retarded Green function. If
x′ < 0, show that it can be written as
GR (x, x′ ; ω) = tG0R (x, x′ ; ω)θ(x)

′
+ [1 + reiφ(x,x ) ]G0R (x, x′ , ω)θ(−x). (4.75)
The complex numbers t and r are called “transmission” and “reflection”

amplitude.
(c) Explain the origin of the phase shift φ(x, x′ ) in Eq. (4.75).
Exercise 4.2: Anderson model for impurity scattering
In some applications, a simple scattering potential is not a good description

for an impurity. This is the case, for example, if the impurity introduces an
extra localized state for electrons in an otherwise perfect lattice. In this case,
one can use the following Hamiltonian to describe the impurity,
X †
X †
H = ε~kσ ψ~kσ ψ~kσ + εimp ψiσ ψiσ
~k,σ σ
62
X † †
+ (V~k ψ~kσ ψiσ + V~k∗ ψiσ ψ~kσ ). (4.76)
~kσ
Here εimp is the energy of the extra localized state. The second line of Eq.
(4.76) represents “hybridization” of the impurity state with the Bloch states
|~ki of the lattice. The matrix element V~k describes the overlap of the impurity
state with the state |~ki. For a strongly localized impurity orbital at the origin,
one can neglect the ~k-dependence of the matrix elements V~k .
For this system, we define Green functions with respect to the wavevec-
tors of the Bloch states and with respect to the impurity site. Hence, the
temperatue Green function is denoted as G~k′ ,~k (iωn ), Gi,~k (iωn ), G~k′ ,i (iωn ), and
Gi,i (iωn ), etc.
(a) Show that the temperature Green function satisfies the equations
X
1 = (iωn − εimp)Gi,i (iωn ) − V~k∗′ G~k′ ,i (iωn ),
~k ′
0 = (iωn − ε~k′ )G~k′ ,i (iωn ) − V~k′ Gi,i (iωn ),

X
0 = (iωn − εimp)Gi,~k (iωn ) − V~k∗′ G~k′ ,~k (iωn ),
~k ′
δ~k′~k = (iωn − ε~k′ )G~k′ ,~k (iωn ) − V~k′ Gi,~k (iωn ).
Solve these equations and analytically continue your answer to find the
retarded Green function GR .
(b) Show that the T -matrix is given by
T~k′~k (ε) = V~k′ Gi,i (ε)V~k∗ .
Calculate the total scattering phase shift δ(ε), and the impurity density
of states νimp (ε).
(c) Evaluate your expressions for the special case of a “flat conduction
band”: the conduction electron states have a constant density of states
ν0 for energies −D < ε < D, where D is the bandwidth. For the hy-
bridization matrix elements, you may set |V~k |2 = V 2 /Ns , where Ns is
the total number of lattice sites in the sample. (For a lattice, the num-
ber of available wavevectors also equals Ns .) The conduction electron
density of states ν0 is defined as density of states per site.
63
Exercise 4.3: Resonant tunneling
In this exercise we consider a “quantum well”, which is formed between two

tunnel barriers in a three-dimensional metal or semiconductor structure. The
tunnel barriers have a potential U(~r) = u(δ(x − a) + δ(x + a)).
Since the system is translation invariant in the y and z directions, but
not in the x direction, we use a mixed notation, in which Green functions
are Fourier transformed with respect to y and z, but not with respect to x.
The Green function is diagonal in the momenta ky and kz .
(a) Following the same method as in exercise 4.1, calculate the retarded
Green function GR ′
ky ,kz ;ky ,kz (x, x ; ω). Express your answer in the same
form as Eq. (4.75) and calculate the transmission and reflection ampli-
tudes tky ,kz and rky ,kz .
(b) When is the transmission is unity? Find a simple physical argument

for this relation. (Hint: use your answer to Ex. 4.1.)
64
5 Many impurities
We have seen that it was a difficult task to calculate the Green function for
a single impurity. In the presence of many impurities, the calculation of the
Green function is as good as impossible.
One might ask whether that is a big problem. Do we really want to
calculate the Green function in the presence of many impurities? The answer
is no, and the reason is very simple: Nobody knows the exact locations and
potentials of all impurities in a realistic sample! Of course, if we don’t know
the exact potential U(~r), it does not make sense to try to calculate the exact
Green function in the presence of that potential.
The information one usually does know is the concentration of impurities,
i.e., the number of impurities per unit volume. The impurity concentration
can also be characterized by the mean time τ between collisions. Because
that is all information one has, one considers an ensemble of different sam-
ples, all with the same impurity concentration, but with different locations of
the impurities. Such samples are called “macroscopically equivalent but mi-
croscopically different”. We then calculate quantities that are averaged over
this ensemble. Such an average is called a disorder average, to distinguish it
from the thermal average we have been considering so far.
There is an important difference between disorder averages and thermal
averages. In a thermal average, one averages over time-dependent fluctu-
ations. According to the ergodic hypothesis, if one waits long enough, a
sample will access all accessible states. Hence, the thermal average equals
the time average. In a disorder average, one averages over impurity configu-
rations. For a given sample, the impurities do not move. Hence, in order to
obtain a disorder average experimentally one has to, somehow, involve many
samples that have the same impurity concentration but different impurity
locations.15
One way of obtaining a disorder average experimentally is “thermal cy-
cling”: to repeatedly heat up and cool the same sample. The reason why
this works is that, at sufficiently high temperatures, impurities are somewhat
mobile. Hence, after heating up and cooling down the sample once, a differ-
ent static impurity configuration is obtained. An example of this procedure
is shown in Fig. 11. Here, the conductance of a metal wire is measured as
15
Some observables may be “self-averaging”: all samples in the ensemble will behave in
the same way, irrespective of the microscopic details of impurity configuration. In the case
of a self-averaging observable, it is sufficient to measure one sample only.
65
Figure 11: Reproducible magnetoconductance curves measured at T = 45 mK.
Different curves are measured for different cool downs of the same GaAs wire.
Figure is taken from D. Mailly and M. Sanquer, J. Phys. (France) I 2, 357 (1992).
Figure 12: Average (a) and variance (b) of the magnetoconductance curves of
Fig. 11, together with theoretical predictions. The average and variance were
taken over 50 different disorder configurations. The figure is taken from D. Mailly
and M. Sanquer, J. Phys. (France) I 2, 357 (1992).
a function of the magnetic field. The different curves correspond to differ-

ent cool downs of the same wire. Within one cool down, the conductance
curves are fully reproducible, as one would expect for a static impurity con-
figuration. There are small differences between cool downs. Indeed, different
cool downs should correspond to slightly different impurity configurations,
for which one expect slightly different Green functions, and, hence, a slightly
different conductance. From the experimental data of Fig. 11, one can calcu-
late the average conductance for this “ensemble” of wires, the variance, and
so on. The average and variance can then be compared to a theory. You can
see plots of the average conductance and of the variance in Fig. 12.
66
5.1 Disorder average
We look at an impurity potential
Nimp
X
U(~r) = u(~r − ~rj ), (5.1)
j=1
where ~rj is the location of the jth impurity. The Fourier transform of the
impurity potential is
1
Z
Uq~ = d~rU(~r)e−i~q·~r
V
Nimp
X
= uq~ e−i~q·~rj , (5.2)
j=1
where uq~ is the Fourier transform of the scattering potential for a single
impurity. For delta-function impurities with u(~r) = uδ(~r) one has uq~ = u/V ,
see Eq. (4.30).
In performing the disorder average for a Green function, we have to aver-
age products of the disorder potential Uq~ over the positions of the impurities.
The corresponding averages are
hUq~i = uq~Nimp δq~,0 ,
hUq~1 Uq~2 i = uq~1 uq~2 [Nimp (Nimp − 1)δq~1 ,0 δq~2 ,0 + Nimp δq~1 +~q2 ,0 ],
hUq~1 Uq~2 Uq~3 i = uq~1 uq~2 uq~3 [Nimp (Nimp − 1)(Nimp − 2)δq~1 ,0 δq~2 ,0 δq~3 ,0
+ Nimp(Nimp − 1)(δq~1 ,0 δq~2 +~q3 ,0 + δq~2 ,0 δq~1 +~q3 ,0 + δq~3 ,0 δq~1 +~q2 ,0 )
+ Nimpδq~1 +~q2 +~q3 ,0 ], (5.3)
and so on. We’ll consider the case when the sample is large and there are
many impurities. In that case, we can neglect the difference between Nimp −1
or Nimp − 2 and Nimp in the above formulas. Studying the case of many
impurities does not mean that one can neglect, e.g., the term Nimp δq~1 +~q2 ,0
2
in the second line of Eq. (5.3) with respect to the term Nimp δq~1 ,0 δq~2 ,0 . The
reason is that the former term has one more summation over the momenta,
which may bring in extra factors volume, etc.
The disorder average is taken after the thermal average. One can for-
mulate a set of diagrammatic rules for the impurity average that should be
implemented on top of the diagrammatic rules for the thermal average. These
rules are
67
ωn = + + +
k k
+ + +
+ + + ...
Figure 13: Diagrammatic expansion of the disorder averaged single-particle Green

function hG~k,~k′ (iωn )i.
• Momentum has to be conserved at an impurity vertex (a star in the

diagram),
• Every impurity vertex (represented by a star in the diagram) con-

tributed a factor Nimp .
• An arbitrary number of dashed lines can meet at a single impurity

vertex. Every dashed impurity line carrying a momentum ~q contributes
a factor uq~.
5.2 Disorder average of single-particle Green function

We now use the diagrammatic rules derived in the previous section to calcu-
late the disorder average of the single-particle Green function,
hG~k,~k′ (τ )i = −hTτ ψ~k (τ )ψ~k† ′ (0)i. (5.4)
Our first observation is that the disorder average restores translation in-
variance. Hence, hG~k,~k′ (τ )i is nonzero only if ~k = ~k ′ .
A diagrammatic expansion for hG~k,~k′ (iωn )i is shown in Fig. 13. In this
figure, all labels for the imaginary frequency and wavevector have been sup-
pressed. You can include them in accordance with the diagrammatic rules
of the previous section. The diagrammatic expansion of Fig. 13 looks rather
wild. One can introduce some order in this expansion with the concept of
“irreducible diagrams”. A diagram is called “irreducible” if it cannot be cut
into two pieces by cutting a single internal fermion line. For example, the
68
ωn
1111
0000
0000
1111
0000
1111
0000
1111 = + + + + ...
0000
1111
0000
1111
0000
1111
0000
1111
k
Figure 14: Diagrammatic expansion of the self energy Σ~k for the disorder aver-
aged single-particle Green function hG~k,~k′ (iωn )i. The self energy is denoted by the
shadded bubble at the left hand side of the figure.
ωn ωn ωn
ωn ωn ωn 0000
1111
1111
0000
0000
1111
ωn ωn 0000
1111
1111
0000
0000
1111
ωn 0000
1111
1111
0000
0000
1111
ωn
= + 0000
1111
0000
1111
+ 0000
1111
0000
1111
0000
1111
0000
1111
+ ...
k 0000
1111 k k 0000
1111 k 0000
1111 k
k k k 0000
1111 0000
1111 0000
1111
0000
1111 0000
1111 0000
1111
k k k
ωn
ωn ωn 1111
0000
0000
1111
0000
1111
ωn
= + 0000
1111
0000
1111
k 0000
1111
k 0000
1111
0000
1111 k
k
Figure 15: Diagrammatic representation of the Dyson equation for the disorder
averaged Green function hG~k,~k′ (iωn )i.
first, second, third, fifth, and eigth diagram in the expansion of Fig. 13 are
irreducible. The other diagrams in Fig. 13 are “reducible”. The sum of all
irreducible diagrams, without the external fermion lines, is called the “self
energy” Σ~k . Diagrammatically, the self energy is given by the expansion of
Fig. 14. In terms of the self energy, the disorder averaged Green function is
seen to obey a Dyson equation, see Fig. 15,
0 0
hG~k,~k (iωn )i = G~k,~k (iωn ) + G~k,~k (iωn )Σ~k (iωn )hG~k,~k (iωn )i. (5.5)
The solution of the Dyson equation is
0
G~k,~k (iωn ) 1
hG~k,~k (iωn )i = 0
= . (5.6)
1− G~k,~k (iωn )Σ~k (iωn ) iωn − ε~k + µ − Σ~k (iωn )
With this result, we have reduced the calculation of the disorder averaged
single-particle Green function to that of the self energy Σ~k (iωn ).
The reason why the quantity Σ~k is called “self energy” is clear from Eq.
(5.6): the self energy provides a shift of the energy ε~k that appears in the
denominator of the single-particle Green function. The shift is caused by the
average effect of scattering from all impurities.
69
ωn
1111
0000
0000
1111
0000
1111
0000
1111 = + + + ...
0000
1111
0000
1111
0000
1111
0000
1111
k
Figure 16: Diagrammatic expansion of the Born approximatino for the self energy
Σ~k for the disorder averaged single-particle Green function hG~k,~k′ (iωn )i.
Still, the diagrammatic expansion for the self energy is too complicated
to be solved exactly. In the limit of a low impurity concentration, one can
restrict the diagrammatic expansion to those diagrams that have one im-
purity vertex. Indeed, according to the diagrammatic rules of the previous
section, every impurity vertex contributes a factor Nimp, so that diagrams
with two or more impurity vertices contribute to higher order in the impu-
rity concentration. (See also exercise 5.1).) The approximation in which one
considers diagrams with one impurity vertex is referred to as the “Born ap-
proximation”. One refers to the “nth order Born approximation” as the Born
approximation where only self energy diagrams with n + 1 or less impurity
lines are kept.
Diagrammatically, the Born approximation corresponds to the expansion
shown in Fig. 16. For the self energy this implies

X
0
Σ~k (iωn ) = Nimp u0 + u−~qG~k+~
q
(iωn )uq~
q~

X
0 0
+ u−~q1 G~k+~
q1
(iωn )uq~1−~q2 G~k+~
q2
(iωn )uq~2 + . . . . (5.7)
q~1 ,~
q2
If we compare this with the expression for the T -matrix of a single impurity,
Eq. (4.16), we arrive at the important conclusion
Σ~k (iωn ) = Nimp T~k,~k (iωn ). (5.8)
In calculations, we’ll be interested in the analytical continuation of the

self energy to real frequencies ω ± iη. After analytical continuation, the real
70
part of the self energy is nothing but a shift of the pole of the single-particle
Green function. It can be absorbed in a redefinition of the chemical potential
or in a redefinition of the energy ε~k , and will be neglected. The imaginary
part of the self energy is more important. It can be expressed in terms of the
scattering time τ~k using the optical theorem, Eq. (4.26),
Im Σ~k (ω + iη) = −Im Σ~k (ω − iη)
X
= −Nimp π |T~k′~k (ω)|2δ(ω − ε~k1 )
~k ′
1
= − , (5.9)
2τ~k
where 1/τ~k is the rate for scattering out of the momentum eigenstate |~ki.
The scattering rate is Nimp times the scattering rate due to a single impurity
only, see Eq. (4.22). (It is only in the limit of a low impurity concentration
that the scattering rate is proportional to the number of impurities.)
The imaginary part of the self energy gives the particle in a momentum
eigenstate |~ki a finite lifetime, as you would expect for impurity scattering.
You can see this explicitly if you use the result (5.6) for the disorder averaged
temperature Green function to calculate the retarded Green function,
hG~R
k,~k
(t)i = G~0R
k~k
(t)e−t/2τ~k . (5.10)
5.3 Gaussian white noise

In the remainder of this course, we’ll use a simpler model for the impurity
potential than Eq. (5.1) above. In this model, both the positions and the
potentials of the impurities are random. The simplest way to characterize
the impurity potential U(~r) is to say that it is a random function, with a
Gaussian distribution of zero average,
hU(~r)i = 0, (5.11)
and Gaussian delta-function correlations,
1
hU(~r)U(~r′ )i = δ(~r − ~r′ ), (5.12)
2πντ
where τ is the mean free scattering time (see Sec. 5.2) and ν is the density
of states at the Fermi level per unit volume,
k2
ν= . (5.13)
2π 2 ~vF
71
ωn ωn
1111
0000
0000
1111
0000
1111
0000
1111
k−q
0000
1111 =
0000
1111
0000
1111
0000
1111
k
Figure 17: Self energy diagram for Gaussian white noise potential.
Higher moments of the impurity potential can be constructed using the rules
for Gaussian averages. This random potential U(~r) is called “white noise”
because it is delta-function correlated in space.
Performing the Fourier transform,
1
Z
Uq~ = d~rU(~r)e−i~q·~r ,
V
we find that the potential Uq~ is still random and Gaussian. Its first two
moments are given by
hUq~i = 0
1
hUq~Uq~′ i = δq~+~q′ ,0 . (5.14)
2πντ V
For the diagrammatic rules for impurity average this means that only those
impurity vertices with two dashed lines contribute to the average.
To leading order in the impurity concentration, we then find only one
diagram for the self energy, see Fig. 17. Hence, the self energy is given by
X 1
Σ~k = G~k−~q,~k−~q(iωn )
2πντ V
q~
1 d~q 1
Z
= 3
2πντ (2π) iωn − ε~k−~q + µ
i
= − sign (ωn ). (5.15)
2τ
This is the same answer as we obtained previously in Sec. 5.2. In fact, the
normalization (5.12) was chosen precisely to get this answer.
72
5.4 Electrical conductivity from Boltzmann equation
Impurities in a conductor form an obstacle for the flow of electrical current.
In the next chapter, we will present a rigorous calculation of the electrical
conductivity in the presence of many impurities. Here, we’ll give a simple,
more phenomenological calculation of the conductivity that makes use of the
Boltzmann equation.
The central object of the Boltzmann equation is a distribution function
f~k for electrons with momentum ~~k. In equilibrium, the distribution function
f~k is the Fermi function f 0 ,
1
f~k0 = . (5.16)
1 + e(ε~k −µ)/T
The presence of an electric field causes the flow of electrical current, i.e., it
changes the distribution f .
The conductivity σ is defined as the proportionality constant between the
~ for small electric fields,
current density ~j and the electric field E
~
~j = σ E. (5.17)
Since we are only interested in the response to small electric fields, we only
consider deviations from f 0 to linear order in the electric field,
f~k = f~k0 + f~k1 . (5.18)
The Boltzmann equation provides an expression for the time derivative
df /dt as a result of the electric field E~ and impurity scattering,
!
df~k ∂~k ~ ~ f~ = I~ .
= ·∇ k k k (5.19)
dt ∂t
Here I is the so-called “collision integral”, which describes changes in f~k as a

result of impurity scattering. The derivative ∂~k/∂t is nothing but the force
exerted by the electric field,
∂~k ~
~ = −eE. (5.20)
∂t
The collision integral I can be expressed as
X
I~k = −Nimp W~k~k′ δ(ε~k − ε~k′ ) f~k (1 − f~k′ ) − f~k′ (1 − f~k ) , (5.21)
~k
73
where W~k~k′ = W~k′~k is the transition probability from a state with wavevector
~k to a state with wavevector ~k ′ as a result of scattering from a single impurity,
or vice versa, see Eq. (4.21). The first term in Eq. (5.21) represents scattering
out of the state |~ki, the second term represents scattering into the state |~ki.
After some algebra, one can rewrite Eq. (5.21) into a simple form,
f~k1
I~k = − , (5.22)
~τ~k,tr
In general, this simple form is known as the “relaxation time approximation”,

although it is exact for the problem we consider here. The time τ~k,tr is known
as the “transport mean free time”. It is expressed in terms of the transition
probabilities as
1 X
= Nimp W~k~k′ (1 − cos θ′ )δ(ε~k − ε~k′ ), (5.23)
τ~k,tr
~k ′
where θ′ is the angle between ~k and ~k ′ .16 For the white noise potential, the
transition rate W~k~k′ does not depend on the angle between ~k and ~k ′ , so that
one has τtr = τ .
Combining Eq. (5.22) with the Boltzmann equation (5.19), and, again,
~ we find
linearizing in the electric field E,
f~k1 = eτ~k,tr E ~ ~ f 0 (~k).

~ ·∇ (5.24)
k
We can also express ~j in terms of the distribution function f and find
~j = − 2e
X ~~k
f~k
V m
~k
2e X ~
= − eτ~k,tr E ~ ~ f~0 ~k
~ ·∇
k k
V m
~k
2
2e X ~ ∂f 0
~k
= − ~v~k E · ~v~k τ~ , (5.25)
V ∂ε~k k,tr
~k
16
You can find a derivation of Eq. (5.23) in the lecture notes for P636, Advanced Solid
State Physics.
74
where we introduced the velocity ~v~k = ~~k/m = ~−1 ∂ε~k /∂~k. We then perform
the integration over ~k by first averaging over the angles and find
2e2 X 2 ∂f
σ=− v~k τ~k,tr ,
3V ∂ε~k
~k
which we can rewrite in terms of the density n of electrons and for low
temperatures T as
ne2 τtr (kF )

σ= . (5.26)
m
This is nothing but the well-known Drude formula for the conductivity.
75
ωn ωn ωn ωn ωn ωn
k−q−q’
k−q−q’
k−q’
k−q
k−q
k−q
Figure 18: Diagrams that were omitted in the calculation of the self energy.
5.5 Exercises
Exercise 5.1: Corrections to the self energy
For the Gaussian white noise potential, we neglected some self-energy dia-
grams. The simplest diagrams we left out are shown in Fig. 18. Calculate
these contributions to the self energy and show that they can be omitted if
disorder if the impurity concentration is small.
Exercise 5.2: Boltzmann Equation
Derive Eq. (5.22).
Exercise 5.3: Self-consistent Born approximation
In the self-consistent Born approximation, one substitutes the unperturbed

Green function G 0 with the disorder averaged full Green function G in the
calculation of the self energy. Since G depends on Σ, the result of this sub-
stitution is a self-consistent equation for the self energy Σ.
(a) Draw diagrams representing the equation for the disorder averaged
Green function G and the self energy Σ in the self-consistent Born
approximation.
(b) Write down a self-consistency equation for the self energy in terms of
the T -matrix. Hint: since the real part of Σ can be absorbed into the
chemical potential, it is the imaginary part only that plays a role.
76
(c) Discuss the solution if the T -matrix has only a weak energy dependence
for energies within a distance ≪ εF from the Fermi level.
(d) Discuss the self-consistent Born approximation for the Gaussian white
noise potential.
77
78
6 Linear Response
6.1 Kubo formula
Quite often, instead of calculating the expectation value of an operator in
thermal equilibrium, one wants to see how an observable responds to a per-
turbation of the system. For example, an electric field generates a current,
or an external charge in the proximity of a metal generates a polarization
cloud.
If one is only interested in the linear response, the change in an observable
A to first order in the perturbation, the equilibrium formalism outlined in
the previous sections can still be used.
We consider the time-dependent Hamiltonian
H = H0 + H1 , (6.1)
where the unperturbed Hamiltonian H is time-independent and the pertur-

bation is H1 . We assume that the perturbation is switched on slowly, so that
H1 → 0 for t → −∞. We are still interested in the average of the operator
A, which now depends on time as a result of the perturbation H1 . Using the
Schrödinger picture, we have
hAi = tr Aρ(t), (6.2)
where ρ(t) is the now time-dependent density matrix. In the presence of the
perturbation H1 , ρ(t) is different from the density matrix ρ0 in the absence of
the perturbation, which is given by the standard expressions for the canonical
or grand canonical ensembles,
−1 −H/T
Z e canonical,
ρ0 = −1 −(H−µN )/T (6.3)
Z e grand canonical,
see Sec. 1.1 The time-dependence of ρ is a consequence of the time-dependence

of the total Hamiltonian H. It is described by the equation
∂ i
ρ(t) = − [H, ρ]− . (6.4)
∂t ~
Note that the sign in Eq. (6.4) differs from that of the Heisenberg picture
evolution equation (1.12) for the operator A. In principle, the density matrix
79
ρ(t) can be solved from Eq. (6.4) using the boundary condition ρ → ρ0 if
t → −∞.
In order to find the linear response to the perturbation H1 , we expand
the density matrix in a power series in H1 and truncate the series after first
order,
ρ(t) = ρ0 + ρ1 (t) + . . . (6.5)
Truncating the evolution equation (6.4) for ρ(t) at first order, we find
∂ i i
ρ1 (t) = − [H0 , ρ1 (t)]− − [H1 , ρ0 ]− , (6.6)
∂t ~ ~
with the boundary condition ρ1 (t) → 0 if t → −∞. Equation (6.6) is a linear
equation, so that it can be solved by standard integration. Its solution is
Z t
−iH0 t/~ ′ iH0 t′ −iH0 t′ /~
ρ1 (t) = −ie dt e [H1 , ρ0 ]− e eiH0 t/~. (6.7)
−∞
Substituting this result into Eq. (6.2), we find the average hAi to first order
in the perturbation H1 ,
hAi = tr Aρ0 + tr Aρ1 (t) (6.8)

Z t
′ ′
= hAi0 − itr Ae−iH0 t/~ dt′ eiH0 t /~[H1 , ρ0 ]− e−iH0 t /~ eiH0 t/~,
−∞
where the brackets h. . .i0 indicate a thermal average with respect to the
unperturbed Hamiltonian H0 . Switching to the Heisenberg picture,
A(t) = eiH0 t/~Ae−iH0 t/~, H1 (t) = eiH0 t/~H1 e−iH0 t/~, (6.9)
and using the cyclic property of the trace, this can be rewritten as
Z t
hAi − hAi0 = −i dt′ h[A(t), H1 (t′ )]− i0
−∞
Z t
= dt′ GR ′
A;H1 (t; t ). (6.10)
∞
This result is known as the Kubo formula. Since the right hand side of Eq.
(6.10) is a retarded Green function, defined with respect to a thermal average
with the equilibrium Hamiltonian H0 , Eq. (6.10) can be calculated with the
methods described in the previous sections.
80
Quite often, we are interested in the response to a perturbation that is
known in the frequency domain,
1
Z
H1 = dωe−i(ω+iη)t H1 (ω). (6.11)
2π
Here η is a positive infinitesimal and the factor exp(ηt) has been added to
ensure that H1 → 0 if t → −∞. Then Eq. (6.10) gives
Z
hA(ω)i − hA(ω)i0 = dteiωt (hAi − hAi0 )
= GR
A;H1 (ω + iη). (6.12)
Although the Kubo formula involves an integration over the real time t′ ,
for actual calculations, we can still use the imaginary time formalism and
calculate the temperature Green function GA;H1 (τ ). The response hA(ω)i is
then found by Fourier transform of G to the Matsubara frequency domain,
followed by analytical continuation iωn → ω.
6.2 Kubo formula for dielectric function

Consider a metal or other dielectric with an added charge. One example of
such a charge is an ion with a different valence in a metal. We’ll call this
charge the “external charge” and denote the corresponding charge density
by ρext .
~ which is
The external charge determines the “electrical displacement” D,
the solution of the Maxwell equation
~ ·D
ǫ0 ∇ ~ = 4πρext . (6.13)
Here, ǫ0 is the permittivity of vacuum. In a dielectric, the external charge

will induce a screening charge ρind . The total charge, screening charge plus
~
external charge, determines the electric field E,
~ ·E
ǫ0 ∇ ~ = 4π(ρext + ρind ). (6.14)
In the language of “external” and “induced” quantities, you may consider

~ as the “total electric field”, and D
E ~ as the “external electric field”. The
~ ~
difference E − D then corresponds to the “induced electric field”.
81
In general, the electric field, the electric displacement, and their difference
(or, more precisely, the longitudinal parts of these fields) can be written
as gradients of electric potentials φtot , φext , and φind , respectively. These
potentials are expressed in terms of the corresponding charges as
Z
φ(~r) = d~r′ VC (~r − ~r′)ρ(~r′ ), (6.15)
where VC (~r − ~r′ ) = 1/ǫ0 |~r − ~r′ | is the Coulomb interaction.

For a sufficiently small external charge, the induced and total potentials
are proportional to the external potential. This is the regime of “linear
screening”. The proportionality constant, which may be a non-local function
in space and time, is called the inverse “relative permittivity” of the dielectric,
or the dielectric response function,
Z Z
′
φtot (~r, t) = d~r dt′ ǫ−1 (~r, t; ~r′, t′ )φext (~r′ , t′ ). (6.16)
Since ǫ is defined as the linear response to the external charge, we can use
the Kubo formula to calculate ǫ. We make use of the fact that the external
charge causes a change of the Hamiltonian given by
Z
H1 = d~rρe (~r)φext (~r, t), (6.17)
where ρe is the electronic charge density operator,

X
ρe = −e ψσ† (~r)ψσ (~r). (6.18)
σ
Then, according to the Kubo formula, we find that the induced charge density
is Z Z
′
ρind (~r, t) = d~r dt′ GR r, t; ~r′ , t′ )φext (~r′ , t′ ).
ρe ,ρe (~ (6.19)
As in any use of the Kubo formula, we assume that the external potential
φext was switched on at some time before t. The quantity
χR r , t; ~r′ , t′ ) = GR
e (~ r, t; ~r′ , t′ )
ρe ,ρe (~ (6.20)
is known as the “polarizability function”.
82
From Eq. (6.19) we then conclude that
ǫ−1 (~r, t; ~r′, t′ ) = δ(~r − ~r′ ) δ(t − t′ )

Z
+ d~r′′ VC (~r − ~r′′ )χR (~r′′ , t; ~r′ , t′ ). (6.21)
In a translationally invariant medium, the dielectric response function

ǫ−1 (~r, t; ~r′ , t′ ) depends on the differences ~r − ~r′ and t − t′ only. Then it is
advantageous to perform a Fourier transform to position and time. Defining
1
Z Z
−1 ′ ′
ǫ (~q, ω) = d~rd~r dte−i~q·(~r−~r )+iωt ǫ−1 (~r, t; ~r′, 0), (6.22)
V
Z Z
φ(~q, ω) = d~r dteiωt−i~q·~r φ(~r, t), (6.23)
we find
φtot (~q, ω) = ǫ−1 (~q, ω)φext (~q, ω). (6.24)
Fourier transforming Eq. (6.21), we thus find
ǫ−1 (~q, ω) = 1 + VC (~q)χR

e (~
q , ω), (6.25)
where VC (~q) = 4π/ǫ0 q 2 is the Fourier transform of the Coulomb potential.

As an application, we now calculate the dielectric response function for a
noninteracting electron gas
1
Z
R ′
χe (~q; t) = d~rd~r′ χR r, t; ~r′ , 0)e−i~q·(~r−~r )
e (~
V
1
= −iθ(t) h[ρe,~q(t), ρe,−~q(0)]− i, (6.26)
V
where the Fourier transform of the electron charge density operator follows
from Eq. (6.18),
XZ
ρe (~q) = −e d~re−i~q·~r ψσ† (~r)ψσ (~r)
σ
X †
= −e ψ~k,σ ψ~k+~q,σ . (6.27)
~k,σ
83
†
Inserting the proper time dependence of the operators ψ~k+~q,σ and ψ~k,σ , we
find
1 XX †
χR
e (~
q ; t) = −ie2
θ(t) h[ψ~k,σ ψ~k+~q,σ , ψ~k† ′ ,σ′ ψ~k′ −~q,σ′ ]− i
V ′ ′ ~kσ ~k σ
i(ε~k −ε~k+~q )t/~
×e . (6.28)
The thermal average of the commutator is easily found to give [nF (ε~k −
µ) − nF (ε~k+~q − µ)]δ~k,~k′−~qδσσ′ , where nF (ε) = 1/(1 + exp(ε/T )) is the Fermi
distribution function. Finally, performing a Fourier transform to time, we
find the result
2e2 X nF (ε~k − µ) − nF (ε~k+~q − µ)
χR
e (~
q , ω) = , (6.29)
V ε~k − ε~k+~q + ω + iη
~k
where η is a positive infinitesimal. Note that χR e satisfies the general integral

relation (2.42) that expresses the full retarded correlation function in terms
of its imaginary part only.
An alternative method to calculate χR e is to use the imaginary time for-
malism. One first calculates
e2 X X †
χe (~q, τ ) = − Tτ hψ~k,σ (τ + η)ψ~k+~q,σ (τ )ψ~k† ′ ,σ′ (η)ψ~k′ −~q,σ′ (0)i
V
~kσ ~k ′ σ′
2 XXh
e
= G~k+~q,σ;~k′ ,σ′ (τ )G~k′ −~q,σ′ ;~k,σ (−τ )
V
~kσ ~k ′ σ′
i
− G~k+~q,σ;~k,σ (−η)G~k′ −~q′ ,σ′ ;~k′ ,σ′ (−η) . (6.30)
The positive infinitesimal η has been added to ensure the placement of cre-
ation operators in front of the annihilation operators at the same imaginary
time in the time-ordered product. The second term does not depend on τ .
After Fourier transform with respect to τ it will give a contribution pro-
portional to δΩn ,0 , which can be discarded for the analytical continuation
iΩn → ω + iη we need to do at the end of the calculation. Fourier transform
of the first term and substitution of the exact results for the electron Green
functions gives
2T e2 X 1
χe (~q, iΩn ) = .
V [iωm − (ε~k − µ)][i(ωm + Ωn ) − (ε~k+~q − µ)]
~k,m
(6.31)
84
C1 C1
C2
εk−µ
εk+q−µ− i Ωn
C2
Figure 19: Integration contours for the calculation of the polarizability of the
noninteracting electron gas.
Here Ωn is a bosonic Matsubara frequency (recall that ρq~ is quadratic in

fermion creation and annihilation operators) and ωm is a fermionic Matsubara
frequency. In order to perform the summation over ωm , the summation is
written as an integral over a suitably chosen contour in the complex plane,
see Fig. 19. The integrand is a function of the complex variable z and has
poles at the fermionic Matsubara frequencies z = iωm , at z = ε~k − µ, and at
z = ε~k+~q − µ − iΩn , hence
e2 X tanh(z/2T )
Z
χe (~q, iΩn ) = dz
2πiV C1 [z − (ε~k − µ)][z + iΩn − (ε~k+~q − µ)]
~k
e2 X tanh[(ε~k+~q − µ)/2T ] − tanh[(ε~k − µ)/2T ]

= . (6.32)
V ε~k − ε~k+~q + iΩn
~k
Now the analytical continuation Ωn → ω + iη is easily done and one recovers

the result (6.29) upon substituting tanh(ε/2T ) = 1 − 2nF (ε).
Let us now consider the dielectric response at finite frequency in more
detail. The polarizability describes how the induced charge varies with the
time-dependent external charge. A time-dependent induced charge requires
a time-dependent current density ~j, which, by the continuity relation, is
proportional to the time derivative of the induced charge. The real part of the
polarizability describes that part of the induced current that is out of phase
with respect to the external electric field. For such currents, the time-average
~ is zero, so that no energy is dissipated in one cycle. Hence, Re χe
of ~j · E
describes the dissipationless response to an external potential. Similarly, the
imaginary part of the polarizability describes dissipation: current and field
85
11111111111111111111
00000000000000000000
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
ω/ε F 4 00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
2 00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
0 00000000000000000000
11111111111111111111
0000000000
1111111111
0 1 2 3 4
q/k F
Figure 20: Support of the imaginary part of the polarizability χR

e (~
q , ω) for a
noninteracting electron gas.
~ is nonzero.
are in phase, so that the time average of ~j · E
For the noninteracting electron gas, the dissipative response is given by
the imaginary part of Eq. (6.29) above,
πe2 X
Im χR q , ω) = −
e (~ [nF (ε~k − µ) − nF (ε~k+~q − µ)]
V
~kσ
× δ(ε~k − ε~k+~q + ~ω). (6.33)
Hence, electromagnetic energy can be dissipated by the excitation of particle-

hole pairs of energy ~ω and momentum ~q. Let us investigate Eq. (6.33) in
more detail for zero temperature and ω > 0.17 Then the difference nF (ε~k −
µ) − nF (ε~k+~q − µ) is one if ~k is inside the Fermi sphere and ~k + ~q is outside
the Fermi sphere, and zero otherwise. Taking a parabolic dispersion relation,
ε~k = ~2 k 2 /2m, the delta function then implies ω = q 2 /2m~ + ~k · ~q/m~. In
general, one has |~k · ~q| ≤ mvq/~, where v = ~k/m is the velocity for an
electron with wavenumber ~k. Since ~k is inside the Fermi sphere, one has
v ≤ vF , so that the momentum and frequency range where Im χR e is nonzero
is given by
q2 q2
− vF q < ω < + vF q. (6.34)
2m~ 2m~
The support of χR e is shown in Fig. 20.
17
The case ω < 0 follows from the relation χR q , ω) = χR
e (~ q , −ω).
e (−~
86
6.3 Kubo formula for electrical conductivity
The most common transport measurement for a conducting material is that
of the conductivity, the current that flows in response to an external electric
field.
For a normal metal, this response takes the form
Z Z X
′
jeα (~r, t) = dt d~r′ σαβ (~r, t; ~r′ , t′ )Eβ (~r′ , t′ ). (6.35)
β
Here σ is the conductivity. Note that the conductivity is a tensor and that
it describes a response to the electric field that is non-local in time and in
space.
The electric field is related to the electric potential φ and the vector
potential A,~
~ = −∇φ
E ~ − ∂ A. ~ (6.36)
∂t
This relation holds both for the external electric field and for the total electric
field. If the current is carried by electrons, the electrical current operator is a
sum of a “paramagnetic” contribution ~j p and a “diagmagnetic” contribution
~j d ,
~je (~r) = ~jep (~r) + ~jed (~r), (6.37)
where
~jep (~r) = − ~e
X
ψσ† (~r)(∂~r ψσ (~r)) − (∂~r ψσ† (~r))ψσ (~r) ,

(6.38)
2mi σ
~jed (~r) = e A(~
~ r )ρe (~r). (6.39)
m
Here ρe is the electronic charge density, see Eq. (6.18) above. To linear order
in the electric field, the electron’s Hamiltonian is H0 + H1 , where H0 is the
Hamiltonian in the absence of the electric field and
Z Z
H1 = d~rρe (~r)φext (~r, t) − d~r~je (~r)Aext (~r, t). (6.40)
Now we’ll apply the general Kubo formula to find an expression for the
electrical current density ~j in the presence of the electric field. We’ll choose
a gauge in which φext = 0. This would be the natural choice if, e.g., the
87
E(t)
Φ(t)
Figure 21: The gauge with zero scalar potential φ and a time-dependent vector po-
~ is the natural choice if the electric field is generated by a time-dependent
tential A
flux through the sample.
sample is a cylinder and the electric field is generated by a time-dependent

flux through the sample, see Fig. 21.
Fourier transforming Eq. (6.35), we find
Z X
jeα (~r, ω) = d~r′ σαβ (~r, ~r′ ; ω)Eβ (~r′ , ω), α = 1, 2, 3, (6.41)
β
where, in frequency representation, the electric field is related to the vector

potential as
1 ~
Z
~
Aext (~r, ω) = Eext (~r, ω) = dteiωt Aext (~r, t) (6.42)
iω
and Z
′
σαβ (~r, ~r ; ω) = dteiωt σαβ (~r, t; ~r′ , 0). (6.43)
Since we are interested in linear response, we only need to know H1 to

~ so that we can replace the electrical current density ~je in
linear order in A,
Eq. (6.40) by the paramagnetic current density ~jep .18 Then the Kubo formula
18 ~ 0 due to a magnetic field, we
If the Hamiltonian H0 contains a vector potential A
~ ~ p ~
should replace je by je + (e/m)A0 ρe .
88
gives
i R eρe (~r)
σαβ (~r, ~r′; ω) = Παβ (~r, ~r′; ω) + δ(~r − ~r′ )δαβ . (6.44)
ω iωm
Here the function ΠR
αβ is the retarded current-current correlation function,
ΠR r , ~r′ ; t) = GR
αβ (~ r, ~r′ ; t) = −iθ(t)h[jeα (~r, t); jeβ (~r′ , 0)]− .i
jeα ,jeβ (~ (6.45)
For a translationally invariant system, the conductivity depends on the

difference ~r −~r′ only. Performing Fourier transforms for the conductivity and
for the current density,
1
Z
′
σαβ (~q, ω) = d~rd~r′ e−~q·(~r−~r ) σαβ (~r, ~r′ ; ω), (6.46)
V
Z
~je,~q = d~re−i~q·~r~je (~r), (6.47)
we can write the Kubo formula for the conductivity as

i R eρe
σαβ (~q, ω) = Παβ (~q, ω) + δαβ . (6.48)
ω iωm
Here ρe is the electron charge density averaged over the entire sample (i.e.,
the ~q = 0 component of the charge density divided by the volume), and
1
ΠR q , t) = −iθ(t) h[jeα,~q(t), jeβ,−~q(0)]− i.
αβ (~ (6.49)
V
In general, the conductivity is defined as the response to the actual elec-
tric field in the sample, not as the response to the external field. If the electric
fields are time dependent, the currents are time dependent and, thus, gener-
ate their own magnetic and electric fields. In addition, as we discussed in the
previous section, an applied electric field may lead to a build-up of charge,
which can screen the field. What we calculated above is the response to the
external electric field. This procedure gives the correct result only if we can
neglect the induced electric field. This is the case if, on the one hand, the
frequency of the signal is sufficiently low that no magnetic fields are induced,
whereas, on the other hand, the frequency is large enough compared to the
size of the sample or the wavelength of the electric field that no screening
charges are built up. In practice, both conditions are met if the electric field
89
is perpendicular to the wavevector ~q or if the wavevector ~q is taken to zero
before the frequency ω is made small.
Let us now calculate the conductivity of a noninteracting electron gas.
If there are no impurities, the conductivity is infinite (see exercise 6.3). So,
in order to do a meaningful calculation, we consider the conductivity of an
electron gas with impurities.
The conductivity follows from the current-current correlation function
(6.49). For our calculation, we consider the imaginary-time version of that
correlation function,
1
Πα,β (~q, τ − τ ′ ) = − hTτ jq~,α (τ )j−~q,α (τ ′ )i, (6.50)
V
and perform an analytical continuation to real frequencies at the end of the
calculation. The current density operator ~jq~ is expressed in terms of the
creation and annihilation operators for electrons in momentum eigenstates
|~ki as
~jq~ = − e (2~k + ~q)c~†k,σ c~k+~q,σ .
X
(6.51)
2m
~kσ
Since we are dealing with a non-interacting electron gas, we can perform the
thermal average in Eq. (6.50) using Wick’s theorem, with the result
e2 X X
Πα;β (τ, τ ′ ) = G~k+~q,σ;~k′ ,σ′ (τ ′ − τ )G~k′ −~q,σ′ ;~k,σ (τ − τ ′ )
4m2 V ′ ′~kσ ~k σ
× (2kα + qα )(2kβ′ − qβ ), (6.52)
plus a term that does not depend on τ − τ ′ and can be discarded. It is im-
portant to note that the Green function in Eq. (6.52) is the full temperature
Green function, calculated in the presence of the exact impurity potential.
Fourier transforming Eq. (6.52) and changing variables ~k ′ → ~k ′ + ~q, we find
e2 T X
Πα;β (iΩn ) = G~k+~q,~k′ +~q(iωm + iΩn )G~k′ ;~k (iωm )
2m2 V ′
~k,~k ,m
× (2kα + qα )(2kβ′ + qβ ), (6.53)
where we summed over the spin index.

In order to calculate the full temperature Green function, we use the
perturbative expansion of G in powers of the impurity potential, which is
90
ωm ωm ωm ωm ωm
k k k’ k k" k’
+ + + ...
k+q k+q k’+q k+q k"+q k’+q
ωm+Ω n ωm+Ω n ωm+Ω n ωm+Ω n ωm+Ω n
Figure 22: Diagrammatic representation of all diagrams without crossed lines that
contribute to the conductivity. The double lines represent the disorder averaged
Green function.
shown in Fig. 7. Here, our aim is to calculate a disorder average. It is

important to realize that, here, we calculate the impurity average of the
product of two Green functions. That is not the same as the product of the
averages!
In order to keep the calculations simple, we’ll restrict our attention to the
case of a Gaussian white noise random potential. Then, diagrammatically,
the impurity average means that we connect all pairs of impurity lines, as
in Fig. 17. Still, it is an impossible task to sum all diagrams. We can
further simplify to limit ourselves to those diagrams in which no impurity
lines cross. This is a sensible approximation, since you have seen in exercise
5.1 that diagrams in which impurity lines cross are a factor kF l smaller than
diagrams without crossed impurity lines.
The diagrams without crossed lines can be summed up as in Fig. 22.
The building blocks are the impurity averaged Green functions, which are
represented by double arrows. The open circles at the two ends are called
“current” vertices. They represent the factors (2~k + ~q)/2m and (2~k ′ + ~q)/2m
that appear in the calculation of the current-current correlation function.
These factors are referred to as the “bare current vertex”, and they are
denoted by the “bare vertex function” Γ0α (~k, iωm ; ~k + ~q, iωm + iΩn ),
e(2kα + qα )
Γ0α (~k, iωm ; ~k + ~q, iωm + iΩn ) = − . (6.54)
2m
The middle part of the diagram is a “ladder” which has the impurity averaged
Green functions and impurity lines as building blocks.
All but the rightmost part of the ladder diagram can be written as a solid
dot, see Fig. 23. This part of the diagram is known as a vertex correction.
It is described by the vertex function ~Γ(~k, iωm ; ~k + ~q, iωm + iΩn ). A self-
consistent equation for the vertex function can be found from the second line
of Fig. 23,
Γα (~k, iωm ; ~k + ~q, iωm + iΩn )
91
ωm ωm
ωm k ωm k k’ k k’ k" k
= + + + ...
k+q k+q
ωm+Ω n ωm+Ω n k’+q
k+q ω +Ω
k’+q k"+q
k+q
ωm+Ω n m n
ωm
ωm k k’ k
= +
k+q
ωm+Ω n k’+q
k+q
ωm+Ω n
Figure 23: Diagrammatic representation of vertex correction Γ necessary for the

calculation of the current-current correlator. The double lines represent the disor-
der averaged Green function.
= Γ0α (~k, iωm ; ~k + ~q, iωm + iΩn )

1 X
+ hG~k′ ,~k′ (iωm )ihG~k′+~q,~k′ +~q(iωm + iΩn )i
2πντ V
~k ′
× Γ(~k ′ , iωm ; ~k ′ + ~q, iωm + iΩn ), (6.55)

where Γ0 is the bare vertex function, see Eq. (6.54) above. In order to
calculate the vertex function Γ, note that the left hand side of Eq. (6.55)
contains the quantity
hG~k′ ,~k′ (iωm )ihG~k′+~q,~k′+~q(iωm + iΩn )iΓ(~k ′ , iωm ; ~k ′ + ~q, iωm + iΩn )
X
~k ′
only. Hence, a self-consistent equation for the vertex function is found upon
multiplication of Eq. (6.55) with hG~k,~k (iωm )ihG~k+~q,~k+~q(iωm + iΩn )i and sum-
mation over ~k,
X 2kα + qα
−e hG~k,~k (iωm )ihG~k+~q,~k+~q(iωm + iΩn )i
2m
~k
 
=  hG~ ′ ~ ′ (iωm )ihG~ ′ ~ ′ (iωm + iΩn )iΓ(~k ′ , iωm ; ~k ′ + ~q, iωm + iΩn )
X
k ,k k +~
q ,k +~
q
~k ′
 
1 X
× 1 − hG~k,~k (iωm )ihG~k+~q,~k+~q(iωm + iΩn )i , (6.56)
2πντ V
~k
92
ωm
k
k+q
ωm+Ω n
Figure 24: Diagrammatic representation of the current-current correlator in terms

of the vertex function Γ (solid dot), the bare vertex function Γ0 (open dot) and
the disorder averaged Green function (double lines).
from which we conclude
Γα (~k, iωm ; ~k + ~q, iωm + iΩn )

e(2kα + qα )
= −
2m
e 1 P~ ′ (2k ′ + qα )hG~ ′ ~ ′ (iωm )ihG~ ′ ~ ′ (iωm + iΩn )i
V k α k ,k k +~q ,k +~
q
− .
2m 2πντ − V1 ~k′ hG~k′,~k′ (iωm )ihG~k′ +~q,~k′ +~q(iωm + iΩn )i
P
(6.57)
We are interested in the current-current correlator in the limit ~q → 0. The

numerator of the last term of Eq. (6.57) goes to zero if ~q → 0. Hence, we
conclude that, for our choice of the random potential, there is no vertex
correction.19 (That is one of the reasons why we have chosen this model for
the random potential.)
Now the current-current correlation function is easily expressed in terms
of the vertex function Γ and the bare vertex function Γ0 , see Fig. 24,
2T X ~
Πα,β (iΩn ) = Γ(k, iωm ; ~k + ~q, iωm + iΩn )Γ0 (~k + ~q, iωm + iΩn )
V
~k,m
× G~k,~k (iωm )G~k+~q,~k+~q(iωm + iΩn ). (6.58)
It remains to perform the summations over ~k and m and do the analytical

continuation iΩn → ω + iη. We first do the summation over m. As before,
19
For a different choice of the impurity potential Γ and Γ0 are different. This “vertex
correction” amounts to the replacement of the mean free time τ with the “transport mean
free time” τtr , see Sec. 5.4.
93
C1 C1
C2
C2
C2
C2
Figure 25: Integration contours for the calculation of the current-current correla-
tor.
this is done by representing the Matsubara summation as an integral in the

complex plane,
1
Z
Γ(~k, z; ~k + ~q, z + iΩn )Γ0 (~k + ~q, z + iΩn )
X
Πα,β (iΩn ) = dz
2πiV C1
~k
× G~k,~k (z)G~k+~q,~k+~q(z + iΩn ) tanh(z/2T ). (6.59)
The integrand has singularities at the real axis and at the axis z = x − iΩn ,
x real. Hence, we deform the contour as shown in Fig. 25.
1
Z XX
Πα,β (iΩn ) = dξ tanh(ξ/2T ) (±1)
2πiV ±
~k
h
× Γ(~k, ξ ± ; ~k + ~q, ξ + iΩn )Γ0 (~k + ~q, ξ + iΩn , ~k, ξ ± )
× G~k,~k (ξ ± )G~k+~q,~k+~q(ξ + iΩn )
+ Γ(~k, ξ − iΩn ; ~k + ~q, ξ ± )Γ0 (~k + ~q, ξ ± , ~k, ξ − iΩn )
i
× G~k,~k (ξ − iΩn )G~k+~q,~k+~q(ξ ± ) , (6.60)
where we used tanh(z/2T ) = tanh[(z + iΩn )/2T ] and ξ ± = ξ ± iη. It is not

necessary to add an infinitesimal ±iη to ξ ±iΩn , since iΩn is imaginary itself.
At this stage of the calculation we can perform the analytical continuation
94
iΩn → ω + iη,
1
Z X
Πα,β (ω + iη) = dξ tanh(ξ/2T )
2πiV
~k
h
× ΓRR (~k, ξ; ~k + ~q, ξ + ω)Γ0 RR (~k + ~q, ξ + ω; ~k, ξ)G~R
k,~k
(ξ)G~R q ,~k+~
k+~ q
(ξ + ω)
− ΓAR (~k, ξ; ~k + ~q, ξ + ω)Γ0 RA (~k + ~q, ξ + ω; ~k, ξ)G~A
k,~k
(ξ)G~R q,~k+~
k+~ q
(ξ + ω)
+ ΓAR (~k, ξ − ω; ~k + ~q, ξ)Γ0 RA (~k + ~q, ξ; ~k, ξ − ω)G~A
k,~k
(ξ − ω)G~R q ,~k+~
k+~ q
(ξ)
i
−Γ AA ~ ~
(k, ξ − ω; k + ~q, ξ)Γ 0 AA ~ ~ A A
(k + ~q, ξ; k, ξ − ω)G~k,~k (ξ − ω)G~k+~q,~k+~q(ξ) .
(6.61)
Sofar our result has been quite general. Equation (6.61) is valid for any
form of the vertex correction and even holds in the presence of electron-
electron interactions (which will be discussed in a later chapter). We now
specialize to the limit ~q → 0, which is needed to find the dc conductivity.
As explained above, we should take the limit ~q → 0 before we take the limit
ω → 0. For zero wavevector ~q there is no vertex renormalization, and hence
e2 X kα kβ
Z
Πα,β (ω + iη) = dξ tanh[(ξ − µ)/2T ]
2πiV m2
~
k
1
×
(ξ − ε~k + i/2τ )(ξ + ω − ε~k + i/2τ )
1
−
(ξ − ε~k − i/2τ )(ξ + ω − ε~k + i/2τ )
1
+
(ξ − ω − ε~k − i/2τ )(ξ − ε~k + i/2τ )

1
− , (6.62)
(ξ − ω − ε~k − i/2τ )(ξ − ε~k − i/2τ )
where we shifted the ξ-integration by the chemical potential µ. We first
perform a partial integration to ξ,
e2 1
Z
Πα,β (ω + iη) = dξ 2
4πiV T cosh [(ξ − µ)/2T ]
X kα kβ 1 (ξ − ε~k )2 − (ω + i/2τ )2

1
× − ln .
m2 ω ω + i/τ (ξ − ε~k )2 + 1/4τ 2
~k
95
With this manipulation the summation over the wavevector ~k is convergent.
The summation over ~k is separated into an angular average over the direction
of ~k and an integration over the magnitude. The latter integration can be
rewritten as an integral over the energy ε~k and yields
2e2 νkF2 δα,β

Πα,β (ω + iη) = − . (6.63)
3(1 − iωτ )m2
We use the relation ν = mkF /2π 2 for the density of states per spin direction
and per unit volume, and the relation n = kF3 /3π 2 = (−ρe /e) for the total
particle density to rewrite this as20
ρe eδα,β
ΠR
α,β (ω) = . (6.65)
m(1 − iωτ )
20
You can verify this result in the limit ω → 0 without expanding around the Fermi
surface. Starting point is Eq. (6.62), which we rewrite as
e2 X kα kβ
Z
ΠRα,β (0) = dξ tanh[(ξ − µ)/2T ]Im hG~R
k,~
(ξ)i2 .
πV m2 k
~
k
We make use of the fact that the integral of hGR (ξ)i2 vanishes and perform the angular
average over the wavevector ~k,
4e2 X k 2 1
Z
R
Πα,β (0) = − dξ Im hG~R
k,~
(ξ)i2 δαβ .
3πV m 2m 1 + e(ξ−µ)/T k
~
k
Using the explicit form of Im hG~R

k,~
k
(ξ)i one verifies that
1 ∂
Im hG~R
k,~
(ξ)i2 = − hA~ ~ (ξ)i. (6.64)
k 2 ∂ε~k k,k
Performing a partial integration with respect to ε~k one thus finds
2e2 X 1 ∂
Z
ΠR
α,β (0) = ε~k dξ hA~ ~ (ξ)iδαβ
3πV m 1 + e(ξ−µ)/T ∂ε~k k,k
~
k
2e2 X dξ 1
Z
= − (ξ−µ)/T
hA~k,~k (ξ)iδαβ
Vm 2π 1 + e
~
k
ρe e
= δαβ .
m
Note that the partial integration gives a factor 3/2 since the density of states is proportional
1/2
to ε~ . This is the desired result.
k
96
Substituting this in Eq. (6.48), we find that the conductivity at wavevector
~q = 0 and frequency ω is given by
e(−ρe )τ
σαβ (0, ω) = . (6.66)
m(1 − iωτ )
In this limit ω → 0 this reproduces the well-known Drude formula.

It may be somewhat disappointing to recover the Drude formula after so
much work. However, we could not have expected to find anything else! In
fact, this whole calculation was not for nothing. First of all, we have found
a firm microscopic derivation of the Drude formula. But more importantly,
with this formalism we are ready to study corrections to the Drude formula.
One of these corrections, the weak localization correction, will be studied at
length in a later chapter.
97
6.4 Exercises
Exercise 6.1: Tunneling density of states
In this exercise we consider two pieces of metal that are weakly coupled by a
tunnel barrier, see Fig. 26. The two pieces of metal are labeled A and B. In
the tunnelling Hamiltonian formalism, the total Hamiltonian H is written as
H = HA + HB + HAB , (6.67)
where HA and HB operate on states of each system separately, with creation

and annihilation operators c†Aµ , cAµ , and c†Bν , cBν for the single-electron states
in A and B, and HAB is a tunneling Hamiltonian HAB that describes the
coupling between the two systems,
X
HAB = (Tµν c†Aµ cBν + Tνµ ∗ †
cBν cAµ ), (6.68)
νµ
where, in principle, the indices µ and ν refer to any choice of an orthogo-

nal basis for the single-electron states in the two systems A and B. The
tunnel matrix elements Tµν are complex number numbers that depend on
the extension of the single-electron wavefunctions in A and B into the in-
sulating region between them. In general, the tunnel matrix elements are
exponentially small in the separation between A and B.
The current through the tunnel barrier follows from the rate of change of
the charge QA in metal A (or metal B) as
∂QA X †
I= = i[H, QA ]− , QA = −e cA,µ cA,µ . (6.69)
∂t µ
Substituting Eqs. (6.67) and (6.68) for H, one finds that only the tunneling
Hamiltonian contributes to the commutator, and that
X
I = ie (Tµν c†Aµ cBν − Tνµ
∗ †
cBν cAµ ). (6.70)
µν
The current through the tunnel barrier is linear in the tunneling matrix
elements Tνµ . In order to calculate I to lowest (second) order in the tunneling
matrix elements, it is sufficient to calculate the (change of the) current to
98
111111111
000000000
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
000
111
000000000
1111111110000000000
1111111111
000
111
000000000
1111111110000000000
1111111111
000
111
000000000
111111111
111
000000000
111111111 I
0000000000
1111111111
000
0000000000
1111111111
000
111
000000000
1111111110000000000
1111111111
000
111
000000000
1111111110000000000
1111111111
000
111
000000000
1111111110000000000
1111111111
000
111
000000000
1111111110000000000
1111111111
000
111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
111111111
A
000000000
1111111110000000000
1111111111
0000000000
1111111111 B
000000000
1111111110000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
V
Figure 26: Setup for tunnel spectroscopy. Two metals, labeled A and B, are
separated by an insulating material, e.g., an oxide or simply vacuum. A battery
maintains a finite difference between the chemical potentials of A and B.
linear response in the tunneling matrix elements. The Kubo formula then
gives Z ∞
I(t) = dt′ CI,H
R
AB
(t, t′ ), (6.71)
−∞
R
where the correlation function CI,H AB
is defined as
R
CI,H AB
(t, t′ ) = −iθ(t − t′ )h[I(t), HAB (t′ )]− i. (6.72)
Here, the thermal equilibrium average h. . .i and the time-dependence of the

tunneling Hamiltonian HAB are taken with respect to the Hamiltonian HA +
HB only, i.e., without inclusion of the tunneling Hamiltonian HAB .
In the average (6.72), the two systems are completely separated. Hence,
in thermal equilibrium, their chemical potentials µA and µB need not be
equal. In fact, it is only if µB − µA = eV 6= 0 that a nonzero current flows
between the reservoirs. Even in the presence of the tunneling current, a quasi
equilibrium in which µB − µA = eV 6= 0 can be maintained with the help of
a battery, as long as the contacts between the battery and the metals A and
B have a much smaller resistance than the tunnel barrier between A and B.
(a) Calculate the correlator CI,HAB and show that the tunneling current
99
can be written
dω X
Z
I = −e |Tµν |2 AA,µ (ω)AB,ν (ω + eV ) (nF (ω + eV ) − nF (ω)) ,
2π µν
(6.73)
where nF (ω) = 1/(1 + exp((ω − µ)/T )) is the Fermi distribution func-
tion. Be sure that your derivation makes no specific assumptions about
the Hamiltonians HA and HB ! Equation (6.73) is valid both with and
without electron-electron interactions.
(b) If metal B has a spectral density that is more or less constant,
X
|Tµν |2 AB,ν (ω) ≈ const., (6.74)
ν
show that the tunneling current can be used as a direct measurement of

the spectral density of metal A. This method of measuring the spectral
density is known as “tunnel spectroscopy”.
Exercise 6.2: Polarizability function
In this exercise, we consider the polarizability function for the non-interacting

electron gas at zero temperature
2e2 X θ(µ − ε~k+~q) − θ(µ − ε~k )
χ0e (~q, ω) = − (6.75)
V ω − ε~k+~q + ε~k + iη
~k
" #
2e2 X 1 1
= − .
V ω − ε~k+~q + ε~k + iη ω − ε~k−~q + ε~k + iη
~k<kF
(a) Perform the summation over ~k to show that the real part of χe is
Re χe (~q, ω) = −2νe2 (f (x, x0 ) + f (x, −x0 )) , (6.76)
where ν is the density of states at the Fermi level per spin direction
and per unit volume, x = q/2kF , x0 = ω/4εF , and
1 (1 − (x0 /x − x)2 ) x + x2 − x0

f (x, x0 ) = + ln .
4 8x x − x2 + x0
100
(b) The expressions for Im χe are somewhat more complicated. Show that
if q < 2kF and ω ≥ 0 one finds

2  4x0 /x if 0 < x0 < x − x2 ,
πνe
Im χe (~q, ω) = − (1 − (x0 /x − x) ) if x − x2 < x0 < x + x2 ,
2
8x 
0 if x0 > x + x2 .
(6.77)
The dimensionless quantities x and x0 were defined above. On the
other hand, if q > 2kF , one finds
 2 2 2
2  (1 − (x0 /x − x) ) if x − x < x0 < x + x,
πνe 2
Im χe (~q, ω) = − 0 if 0 < x0 < x − x
8x 
or if x0 > x + x2 .
(6.78)
Exercise 6.3: Conductivity without impurities
Calculate the retarded current density correlation function ΠR α;β (ω) for a non-
interacting electron gas without impurities. Is the result the same as the limit
τ → ∞ of Eq. (6.65)? What does your result imply for the relation between
current density and electric field?
Exercise 6.4: Kubo formula for conductance
The response of a bulk sample to an applied electric field is characterized

by the conductivity. However, for a sample of finite size, one measures the
conductance, the coefficient of proportionality between the total current I
flowing through a cross section of the sample and the total voltage drop V
over the sample.
In order to derive a relation between the total current I and the voltage
drop V , we introduce a coordinate system (ξ, ξ⊥), such that ξ is parallel with
the electric field lines and ξ⊥ is parallel with the equipotential lines, see Fig.
27. The current can then be written as an integral of the current density,
Z
I(ξ) = dξ⊥ ξˆ · ~j(ξ, ξ⊥ ), (6.79)
101
111
000 111
000
000
111 000
111
000
111 000
111
000
111
000
111 000
111
000
111
000
111 000
111
000
111
000
111 000
111
000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111
000
111 000
111
000
111
000
111 ξ 000
111
000
111 000
111
000
111
000
111 000
111
000
111
000
111 ξ 000
111
000
111 000
111
000
111 000
111
000
111
000
111
equipotential lines 000
111
000
111
000
111 electric field lines 000
111
000
111 000
111
I
Figure 27: Schematic of a conductance measurement. In the derivation of the

Kubo formula for the conductance one uses a coordinate system where the coor-
dinate ξ points along the electric field and the coordinate ξ⊥ points along equipo-
tential lines.
where ξˆ is the unit vector in the ξ-direction. We are interested in the dc

conductance only. Then, current conservation implies that I(ξ) does not
depend on the choice of the cross section, i.e., I(ξ) does not depend on ξ.
(a) Use the Kubo formula for the dc conductivity to express I(ξ) in terms
~ ′ , ξ ′ ).
of an integral over the electric field E(ξ ⊥
(b) Use current conservation to show that the conductance G can be writ-
ten as
i R
G = lim Re CII (ω), (6.80)
ω→0 ω
R
where CII is the retarded current-current correlation function,
R
CII (t) = −iθ(t)h[I(t), I(0)]− i. (6.81)
The current can be calcalated among an arbitrary cross section along

the sample.
(c) Discuss when the relation G = σA/L is true for a sample of cross
section A and length L.
102
Exercise 6.5: Vertex renormalization
In this exercise, we address the vertex function Γ. First, we consider the

numerator of the last term in Eq. (6.57). We perform the analytical contin-
uation iωm → ξ ± iη and iΩn → ω + iη.
(a) For the + sign, show that the numerator of the last term in Eq. (6.57)
is zero.
(b) For the − sign, show that the numerator of the last term in Eq. (6.57)
can be written as
1 X 2~k + ~q
V (ξ − ε~k − i/2τ )(ξ + ω − ε~k+~q + i/2τ )
~k

2πν q̂ 2kl − i i + ωτ ωτ + ql + i i ωτ + ql + i
= 1− ln + ln
vF ql 2ql ωτ − ql + i 2 ωτ − ql + i

4πνklq̂ i + ωτ ωτ + ql + i
≈ 1− ln . (6.82)
vF ql 2ql ωτ − ql + i
where k = kF + ξ/vF , ν is the density of states per spin direction and
per unit volume, and l = vF τ is the mean free path.
Now let us consider the denominator of the same term. Again, we consider
the analytical continuations iωm → ξ ± iη and iΩn → ω + iη.
(c) For the + sign, show that the denominator of the last term in Eq. (6.57)
equals 2πντ .
(d) For the − sign, show that the denominator of the last term in Eq. (6.57)
can be written as
1 X 1
2πντ −
V (ξ − ε~k − i/2τ )(ξ + ω − ε~k+~q + i/2τ )
~
k
i ωτ + ql + i
= 2πντ 1 + ln . (6.83)
2ql ωτ − ql + i
103
104
7 The interacting electron gas: Hartree-Fock
and random phase approximations
Sofar we have discussed electrons mainly as if they were non-interacting
particles. This is a huge simplification of reality that turns out to work
remarkably well. In this chapter we include electron-electron interactions into
our theory and uncover part of the picture why it is that the approximation
of noninteracting electrons is as good as it is.
In our theoretical description, we’ll consider a general electron-electron
interaction Hamiltonian of the form
H = H0 + H1 , (7.1)
where H0 is the “unperturbed” Hamiltonian, which we assume to be quadratic

in fermion or boson creation and annihilation operators, whereas H1 is the
perturbation. The Hamiltonian H1 will include the effect of electron-electron
interactions, which, in the most general case, read
1 X
H1 = Vνµ,ν ′ µ′ ψν† ψµ† ψµ′ ψν ′ . (7.2)
2 ′ ′
νν ,µµ
Here the indices µ, ν, µ′ , and ν ′ label a complete set of single-electron states

and Vνµ,ν ′ µ′ is the corresponding matrix element. For electrons, the fermion
statistics is taken care by the anticommutation relations of the creation and
annihilation operators ψ † and ψ. Depending on the problem of interest, the
effects of impurity scattering may be included in H0 or as an additional term
in H1 .
Of course, the true microscopic electron-electron interaction is the Coulomb
interaction. Using a representation in real space, the Coulomb interaction
Hamiltonian reads
1X e2
Z Z
H1 = d~r1 d~r2 ψ † (~r1 )ψσ† 2 (~r2 )ψσ2 (~r2 )ψσ1 (~r1 ). (7.3)
2 σ ,σ 4πǫ0 |~r1 − ~r2 | σ1
1 2
In momentum space, the Coulomb interaction reads

1 X XX
H1 = Vq~ψ~k† +~q,σ ψ~k† −~q,σ ψ~k2 ,σ2 ψ~k1 ,σ1 , (7.4)
2V σ ,σ 1 1 2 2
1 2 ~k1 ,~k2 q~
105
where
e2 −i~q·~r e2
Z
Vq~ = ed~r = . (7.5)
4πǫ0 r ǫ0 q 2
One important aspect of the Coulomb interaction is its long range. As
a matter of fact, the energy required to charge a system is infinite! That
is the reason why all objects are electrically neutral. In most cases, charge
neutrality is maintained by the positive charge density of the ion lattice, but
charge neutrality can also be maintained by the charge on objects that are
near the sample of interest, but that are not a part of it, such as nearby
pieces of metal or condensator plates.
The goal of this chapter is to include the Coulomb interaction into the
diagrammatic theory. However, first we briefly review other approaches to
treat interactions between electrons.
7.1 Hartree Fock approximation

In the Hartree Fock approximation, the electron-electron is replaced by an
effective potential, which is then determined self-consistently. Replacing the
interaction by a potential is a very useful apprximation, because potentials
can be dealt with on the single-particle level.
If the interaction is replaced by an effective potential, the Hamiltonian is
quadratic in electron creation and annihilation operators. This means that
Wick’s theorem holds for this Hamiltonian. The principle of the Hartree
Fock approximation is that the interaction Hamiltonian is replaced with a
Hamiltonian that is quadratic in electron creation and annihilation operators
in a way that respects Wick’s theorem,21
1X
H1HF = Vνµ,ν ′ µ′ ψν† ψν ′ hψµ† ψµ′ i + ψµ† ψµ′ hψν† ψν ′ i − hψµ† ψµ′ ihψν† ψν ′ i

2
1X
Vνµ,ν ′ µ′ ψν† ψµ′ hψµ† ψν ′ i + ψµ† ψν ′ hψν† ψµ′ i − hψµ† ψν ′ ihψν† ψµ′ i .

−
2
(7.6)
Here, the first line corresponds to what is known as the “Hartree Hamilto-
nian”, whereas the second line is the “Fock Hamiltonian”. If the labels µ and
ν refer to different particles, the Fock contribution is zero; for indistinguish-
able particles, however, the Fock contribution is important. The averages
21
You verify that any average of the Hamiltonian H1HF is the same as that of the original
interaction Hamiltonian H1 if the Wick theorem were to be used.
106
hψµ† ψµ′ i are to be regarded as parameters in the Hartree-Fock Hamiltonian
that should be calculated self-consistently at the end of the calculation.
In a non-magnetic translationally invariant system, any average of the
form hψ~kσ†
ψ~k′ σ′ i should be nonzero only if ~k = ~k ′ and σ = σ ′ . Denoting
†
n~kσ = hψ~kσ ψ~kσ i, (7.7)
the Hartree-Fock Hamiltonian in a translationally invariant system can be

written X
H HF = H0 + H1HF = ε~HF ψ† ψ ,
kσ ~kσ ~kσ
(7.8)
~kσ
where the Hartree Fock energy is given by

X
ε~HF
kσ
= ε~k + (V0 − δσσ′ V~k−~k′ )n~k′ σ′ . (7.9)
~k ′ σ
Note that the Hartree-Fock energies ε~HFkσ

depend on the occupation numbers
n~k′ σ′ which, in turn, depend self-consistently on the Hartree-Fock energies
ε~HF
kσ
.
In the next sections we return to the Hartree-Fock approximation and
give it a diagrammatic interpretation.
7.2 Diagrammatic Technique

The derivation of a diagrammatic technique for interacting electrons proceeds
in exactly the same way as in Sec. 4. We present the rules for the specific
case of the calculation of a single-particle Green function. Their extension
to more complicated correlation functions is straightforward.
The single particle Green function can be expressed as a thermal average
over the non-interacting Hamiltonian H0 ,
R 1/T
H1 (τ ′ )dτ ′
hTτ e− 0 ψσ1 (~r1 , τ1 )ψσ† 2 (~r2 , τ2 )i0
Gσ1 ,σ2 (~r1 , τ1 ; ~r2 , τ2 ) = − R 1/T
′ ′
. (7.10)
hTτ e− 0 H1 (τ )dτ i0
Here operators are represented in the interaction picture. As before, the

Green function is calculated in a series in H1 . Each term in the series can
be represented by a Feynman diagram. The diagrammatic representation
of the interaction (7.2) is as in figure 28: it consists of four fermion lines
107
ν’ ν
µ µ’
Figure 28: Diagrammatic representation of an interaction vertex
connected by a dotted line.22 Each interaction line carries the weight −Vµν;µ′ ν ′
appropriate to the representation that is used, the minus sign arising from the
minus sign in the exponent of the evolution operator in Eq. (7.10). For the
Coulomb interaction in coordinate representation, the interaction involves
one particle at coordinate ~r and one particle at coordinate ~r′ and has weight is
−e2 /4πǫ0 |~r − ~r′ |. For the Coulomb interaction in momentum representation,
the Coulomb interaction scatters electrons in momentum states ~k and ~k ′ to
momentum states ~k + ~q and ~k ′ − ~q, respectively, and has weight −Vq~/V . The
factor 1/V follows from the normalization of the Fourier transform of the
Coulomb interaction, see Eq. (7.4).
The diagrammatic rules for a perturbation H1 that includes interactions
are the same as those for the case of impurity scattering we discussed in
chapter 4. There are two additional rules, specific for the case of interactions,
1. The interaction Hamiltonian H1 involves four creation and annihila-

tions opertors at the same time. In order to ensure that creation op-
erators are always kept in front of annihilation operators, one adds a
positive infinitesimal η to their imaginary times,
1 X
H1 (τ ) = Vνµ,ν ′ µ′ ψν† (τ + η)ψµ† (τ + η)ψµ′ (τ )ψν ′ (τ ). (7.11)
2 νν ′ ,µµ′
With this convention, the time ordering operator Tτ guarantees that

ψ † (τ ) always appears in front of ψ(τ ).
2. With interactions, connected diagrams may include closed particle loops,

see, e.g., Fig. 29. For fermions, each loop corresponds to an odd per-
22
In the literature, the dotted line is often replaced by a wiggly line.
108
k k
k1
= + + +
k1 k k
k k’ k
+ + + +
+ + +
+ + + ...
Figure 29: Diagrammatic representation of all diagrams for the single particle
Green function, up to second order in the Coulomb interaction. The Coulomb
interaction lines are drawn dotted.
mutation of the creation and annihilation operators in the perturbation

expansion, and hence carries a weight −1.
As before, all internal times are integrated over, all internal spin indices
are summed over, and all internal coordinates or momenta are integrated
over or summed over, depending on the representation that is used in the
calculation. You verify that combinatorial factors conspire in such a way that
(i) all disconnected diagrams cancel between numerator and denominator and
(ii) one has to sum over topologically different diagrams only, with the same
weight for each diagram, up to a factor −1 for each fermion loop. You can
find a detailed proof of these rules in the book by Abrikosov, Gorkov, and
Dzyaloshinskii.
All diagrams for the single particle Green function, up to second order
in the Coulomb interaction, are shown in Fig. 29. The second, fourth, fifth,
eleventh, twelfth, and thirteenth diagram in the figure have an odd number
of fermion loops and, hence, receive an extra factor −1. The first three terms
in the expansion of the single-particle Green function read
Gσ,σ′ (~r, τ ; ~r′ , τ ′ ) = Gσ,σ

0
(~r, τ ; ~r′ , τ ′ )δσ,σ′
Z 1/T XZ (−e2 )
Z
0
− dτ1 d~r1 d~r2 Gσ,σ (~r, τ ; ~r1 , τ1 )
0 σ1
4πǫ0 |~r1 − ~r2 |
109
× Gσ01 ,σ1 (~r2 , τ1 ; ~r2 , τ1 + η)Gσ,σ
0
(~r1 , τ1 ; ~r′ , τ ′ )δσ,σ′
Z 1/T
(−e2 )
Z Z
0
+ dτ1 d~r1 d~r2 Gσ,σ (~r, τ ; ~r1 , τ1 )
0 4πǫ0 |~r1 − ~r2 |
0 0 ′ ′
× Gσ,σ (~r1 , τ1 ; ~r2 , τ1 + η)Gσ,σ (~r2 , τ1 ; ~r , τ )δσ,σ′ , (7.12)
if the coordinate representation is used, whereas in momentum representation

the first three diagrams read
G~kσ;~k′ σ′ (τ ; τ ′ ) = G~kσ;
0 ′
~kσ (τ ; τ )δσ,σ′ δ~k,~k ′
1 1/T
Z XX
0
− dτ1 G~kσ,~kσ (τ ; τ1 )(−V0 )
V 0 σ 1 ~k1
× G~k01 σ1 ;~k1 σ1 (τ1 ; τ1 + η)G~kσ;

0 ′
~kσ (τ1 ; τ )δσ,σ′ δ~k,~k ′
1 1/T
Z X
0
+ dτ1 G~kσ,~kσ (τ ; τ1 )(−V~k−~k1 )
V 0
~k1
0 0 ′
× G~k1 σ,~k1 σ (τ1 ; τ1 + η)G~kσ;~kσ (τ1 ; τ )δσ,σ′ δ~k,~k ′ . (7.13)
Note that the interaction matrix element that appears in the integrations
depends on the distance ~r1 −~r2 only (if the coordinate representation is used)
or on the transferred momentum ~k − ~k1 (if the momentum representation is
used). Also note that the Coulomb interaction does not flip spins, so that
spin needs to be conserved separately at each end of an interaction line.
In order to deal with the integrations over imaginary time, it is advanta-
geous to make a Fourier transform to imaginary time, i.e., to write
′
X
Gσ,σ′ (~r, τ ; ~r′ , τ ′ ) = T Gσ,σ (~r, ~r′ ; iωn )e−iωn (τ −τ ) ,
n
′
X
′
G~kσ;~k′ σ′ (τ ; τ ) = T G~kσ;~k′ σ′ (iωn )e−iωn (τ −τ ) , (7.14)
n
where the ωn are the fermionic of bosonic Matsubara frequencies. Each in-
ternal integration over imaginary time for each interaction vertex gives 1/T
times a Kronecker delta for the Matsubara frequencies. Since an interaction
vertex carries only one internal time, this implies that only the sum of the
four Matsubara frequencies is conserved at each interaction line, not the Mat-
subara frequencies of the individual fermions or bosons participating in the
110
interaction.23 In other words, at interaction vertices a Matsubara frequency
difference can be “transferred” between particles. Note that a “transferred”
Matsubara frequency difference is of the boson type if it is transferred be-
tween bosons and bosons or fermions and fermions, and of fermion type if it
is transferred between a boson and a fermion. Since the Coulomb interaction
is instantaneous, the weight of the Coulomb interaction lines do not depend
on the Matsubara frequency that is transferred at the interaction line. All
remaining intermediate Matsubara frequencies that are not fixed by the con-
servation rules are summed over; normalization requires an extra factor T for
each summation over Matsubara frequencies, in accordance with Eq. (7.14).
To illustrate the use of Matsubara frequencies, we rewrite Eq. (7.13) using
Matsubara frequencies,
G~kσ;~k′ σ′ (iωn ) = G~0kσ;~kσ (iωn )δσ,σ′ δ~k,~k′

T XXX 0
− G~kσ,~kσ (iωn )(−V0 )G~0k1 σ1 ;~k1 σ1 (iωm )eiηωm G~0kσ;~kσ (iωn )δσ,σ′ δ~k,~k′
V σ m
1 ~k1
T XX 0
+ G~kσ,~kσ (iωn )(−V~k−~k1 )G~0k1 σ,~k1 σ (iωm )eiηωm G~0kσ;~kσ (iωn )δσ,σ′ δ~k,~k′ .
V m ~k1
(7.15)
The factors exp(iηωm ) were added to account for the infinitesimal time dif-
ference in the intermediate Green functions, cf. Eqs. (7.13) and (7.14).
Looking at all the indices and summations that appear in Eq. (7.15), it
becomes useful to group all these indices together into one “four-vector” that
includes the momentum, the spin, and the Matsubara frequency,
k = (~k, σ, iωn ).
Then the symbol

X T XXX
=
k
V σ n
~k
will denote summation over the momentum ~k, the spin σ, and the Matsubara
frequency ωn and includes the appropriate normalization factors.
23
The statement that the sum of the four Matsubara frequencies is conserved means
that the sum of the Matsubara frequencies of the two outgoing particles equals the sum
of the Matsubara frequences of the incoming particles at the interaction vertex.
111
1111
0000
0000
1111
0000
1111
0000
1111
0000
1111 =
0000
1111
0000
1111
+ + +
0000
1111
+ + + + + ...
Figure 30: Diagrammatic expansion of the self energy, up to second order in the
Coulomb interaction.
7.3 Self energy

We saw that the concept of a “self energy” was very useful in order to organize
the summation of a diagrammatic perturbation series for a disorder-averaged
Green function. Here, we shall see that the same is true for a thermal average
of interacting electrons.
Again we call a diagram “irreducible” if it cannot be cut in two by removal
of a single fermion line. In Fig. 29, the seventh, tenth, twelfth, and thirteenth
diagrams are not irreducible. The self-energy is defined as the sum over all
all irreducible diagrams without the two external fermion lines, see Fig. 30.
With this definition of the self energy, the single-particle Green function
obeys the Dyson equation
X
Gk,k′ = G0k,k′ + G0k,k1 Σk1 ,k2 Gk2 ,k′ . (7.16)
k1 ,k2
Note that, in contrast to the case of the disorder average we studied before,
the self energy Σk,k′ does not need to be diagonal in the four-vectors k, k ′ .
A perturbative expansion for the self energy is shown in Fig. 29. The first
two diagrams of Fig. 29 have our special attention: The interaction line can
be seen as an effective “impurity line”. This is particularly clear for the first
diagram, which is known as the “Hartree” diagram. For the second diagram,
which is known as the “Fock diagram”, you should consider the fermion line
as being part of the “effective impurity”.
At first sight, the Hartree and Fock diagrams appear to be rather different.
However, they are related by simple exchange of particles participating in the
interaction! To understand this, look at the effect of exchange on the basic
112
ν ν’ ν ν’
µ µ’ µ µ’
Figure 31: Interaction vertex and exchanged interaction vertex.
interaction vertex, see Fig. 31. It is precisely this exchange operation that
maps the Hartree diagram into the Fock diagram and vice versa.
Let us calculate what these two contributions to the self energy are. Cal-
culation of the Hartree diagram gives
T XXX
Σ~H
kσ
(iωn ) = − (−V0 )G~k0′ σ′ (iωm )eiηωm . (7.17)
V ′ σ′ m
~k
The factor exp(iηωm ) arose because we take the creation operator in the
interaction Hamiltonian at an infinitesimally larger imaginary time thatn
the annihilation operator. Performing the summation over the Matsubara
frequency ωm and the momentum ~k, we find
X
ΣH = V0 n~0kσ = V0 n, (7.18)
~kσ
where n is the particle density. Similarly, calculation of the Fock diagram

gives
T XXX
Σ~Fkσ (iωn ) = (−V~k−~k′ δσσ′ )G~k′ σ′ (iωm )eiηωm
V ′ ′ m
~k σ
1 X
= − V~k−~k′ n~0k′ σ . (7.19)
V
~k ′
Here, there is no summation over spin, because spin is conserved throughout

the diagram.
Looking at these results, we first note that the Hartree and Fock correc-
tions to the self energy are real: they correspond to a shift of the particle’s
energy, but the particle’s lifetime remains infinite. The results (7.18) and
(7.19) are almost identical to the correction to the single particle-energy
in the Hartree-Fock approximation. The only difference is the absence of
113
1111
0000
0000
1111
0000
1111
0000
1111 =
0000
1111 +
0000
1111
0000
1111
0000
1111
Figure 32: Diagrammatric representation of the Hartree-Fock approximation. The

self energy is expressed in terms of the full single-particle Green function, which,
in turn, depends on the self energy as in Fig. 17.
self-consistency in Eqs. (7.18) and (7.19): in those equations, the occupa-

tion numbers n~kσ are calculated in the non-interacting ground state. Self-
consistency can be achieved if the single-particle Green function G 0 in the
Hartree and Fock diagrams is replaced by the full Green function G, see Fig.
32. In that case one recovers precisely the Hartree-Fock equation (7.9). You
easily verify that, because of the self-consistency condition, the Hartree-Fock
approximation includes all diagrams of Fig. 30 with the exception of the
fourth and the eigth diagrams of that figure.
Note that, formally, the Hartree self energy is divergent. In practice, the
negative charge of the electrons is balanced by the positive charge of the lat-
tice ions, and there is no contribution from the ~q = 0 mode of the interaction.
Hence, for a translationally invariant system, the Hartree contribution to the
self energy can be neglected.
7.4 RPA approximation and screening

Of course, one can continue and calculate more and more diagrams that
contribute to the self energy. As you can see from Fig. 30, this number of
diagrams grows rather fast with the order of the perturbation theory. Hence,
it would be useful if we had a method to pick those diagrams that have the
largest contributions.
In defining the relative importance of interactions, one introduces the
so-called “gas parameter” rs , which is a dimensionless parameter measuring
the density of the electron gas. The parameter rs is defined in terms of the
electron number density n as
a0 rs = (4πn/3)−1/3 . (7.20)
A sphere of radius a0 rs contains precisely one electron on average. Using the
114
relation n = kF3 /3π 2 , we can rewrite this as
rs = (9π/4)1/3 (a0 kF )−1 . (7.21)
A low gas parameter corresponds to a high density of electrons.

The importance of the electron density for a perturbation expansion in
the Coulomb interaction follows from the observation that the gas parameter
is proportional to the ratio of typical Coulomb interaction for neighboring
electrons, which is of order e2 /ǫ0 kF , and the electronic kinetic energy, which
is ~2 kF2 /2m. Hence, at high electron densities the main contribution to the
electron’s energy is kinetic.
For realistic metals, rs is between 1.5 and 6. That is not really small.
Hence, our small-rs expansion is an idealized theory, and we should expect
that some of our quantitative conclusions cannot be trusted.
In order to see which diagrams are important in the limit of high density,
we need two observations.
• First, we consider a diagram contribution to the self energy to nth

order in the Coulomb interaction. In order to estimate its dependence
on the gas parameter rs , we measure energy and temperature in units
of the Fermi energy, which is ∝ kF2 . Then, noting that every nth order
diagram has n integrations over internal momenta (there are 2n − 1
fermion lines and n − 1 constraints from the interaction lines) and that
it contains 2n − 1 fermion Green functions, we find that
Σ(n) /εF ∝ rsn . (7.22)
(The summations over internal Matsubara frequencies do not contribute

to this estimate, since, for each internal Matsubara frequency, the dia-
gram carries a factor T .)
• Second, we note that the Coulomb potential is singular: for small trans-
ferred momentum ~q, the Coulomb potential diverges Vq ∝ q −2 . The role
of these divergencies is stronger if the same momentum ~q is transferred
more than once, i.e., if there is more than one interaction vertex with
precisely the same momentum transfer. Hence, we conclude that those
diagrams with the maximal number of identical transferred momenta
have the largest contribution.
115
1111
0000
0000
1111
0000
1111
0000
1111
0000
1111 = + + + ...
0000
1111
0000
1111
0000
1111
Figure 33: Diagrammatric representation of the random phase approximation for

the self energy in a translationally invariant system.
With these two observations, we can identify the leading diagrams in the limit
of high electron density. Namely, for each set of diagrams with the same
largest number of identical transferred momenta, the diagram of minimal
order in the interaction potential dominates for small rs . Such diagrams are
precisely the diagrams shown in Fig. 33. The approximation in which one
keeps those diagrams only is known as the “random phase approximation”.
Note that the random phase approximation does not treat the direct
interaction and the exchange interaction on the same footing: performing
the same transformation as in Fig. 31 on the diagrams of Fig. 32 leads to
different contributions to the self energy. The reason that these contributions
are left out in the present calculation is that they are less singular. In the
next chapter we consider the effects of a short-range interaction, for which it
is essential that exchange and direct interactions are treated equally.
Before we sum the entire series for the self energy, we note that the
diagrams can be interpreted as an effective Coulomb interaction. In fact,
what the series of Fig. 33 represents is that the Coulomb interaction polarizes
the electron gas, and that the total interaction is found as the sum of the
direct interaction and the interaction with the polarization cloud. This is
nothing but “screening” of the Coulomb interaction! Of course, the random
phase approximation is only a subset of a more complete series of diagrams
that represents the modification of the Coulomb interaction by the electron
gas.24 However, as we discussed above, the random phase approximation
24
In a complete description of the effective Coulomb interaction, one introduces the
concept of an “irreducible” polarization propagator χirr , which is any part of the diagram
for the effective Coulomb interaction that cannot be cut into two parts by cutting a single
internal interaction line. Then the effective interaction can be written as
Vq~
Vq~eff (iΩn ) = .
1 + Vq~ χirr (~q, iΩn )/e2
The random phase approximation amounts to keeping only the zeroth-order in interaction
diagram for χirr , see Eq. (7.24) below.
116
= + + + ...
= +
Figure 34: Diagrammatric representation of the random phase approximation for

the self energy in a translationally invariant system.
describes the dominant effect for an electron gas with a high density.
A diagrammatic expression for the RPA screened Coulomb interaction is
shown in Fig. 34. The screened interaction is represented by a double dotted
line. The central object in this expansion the “pair bubble”, which is nothing
but the retarded polarizability function of the non-interacting electrons gas
χ0e~q(Ωn ), see Eq. (6.29),
2T e2 X X 0
χ0e~q(iΩn ) = G~k+~q(iωm + iΩn )G~k0 (iωm ). (7.23)
V m ~k
(We added the superscript “0” to indicate that χ0e is the polarizability func-
tion of non-interacting electrons. Later we will calculate the full polarizability
for interacting electrons.) Solving for the effective Coulomb interaction, we
find
Vq~ e2
Vq~RPA (iΩn ) = = . (7.24)
1 − Vq~e−2 χ0Rq (iΩn )
e~ ǫ0 q 2 − χ0R
e~q (iΩn )
For small momentum and energy exchange, we can replace the polariz-
ability χ0 by its value at ~q = 0. Since χ0R
e0 (0) is negative (the signs of induced
charge and external charge are opposite), we write
ǫ0 ks2 = −χ0R 2
q →0 (iΩn → 0) = 2νe ,
e,~ (7.25)
where ν = mkF /2π 2 is the density of states at the Fermi level, per spin
direction and per unit volume. Equation (7.25) can be rewritten in terms of
the Bohr radius a0 or the gas parameter rs ,
2/3
4 kF 16
ks2 = = kF2 rs . (7.26)
π a0 3π 2
117
Then the effective Coulomb interaction in the limit of small ~q and ω reads
e2
Vq~RPA (iωn ) = if ~q, ωn → 0. (7.27)
ǫ0 (q 2 + ks2 )
The quantity 1/ks is known as the “Thomas-Fermi screening length”. The
singularity of the Coulomb interaction at large wavelengths is cut off by
the screening of the interaction by the electron gas. The Thomas-Fermi
screening potential is found when we use the limiting form of Eq. (7.27) for
all momenta and frequencies. In that case, the screened Coulomb interaction
is instantaneous and has the spatial dependence of the Yukawa potential,
e2 −ks r
V TF (~r) = e . (7.28)
4πǫ0 r
The true RPA screened interaction we derived above, however, has a different
from on short length scales and, more importantly, is retarded (i.e., it depends
on frequency).
Returning to the self energy: if we sum the entire RPA series for the self
energy, we find
RPA iηωm
RPA T X X V~k−~k′ (iωn − iωn )e
Σ~k (iωn ) = − . (7.29)
V ′ m (iωm − ε~k′ + µ)
~k
7.5 Dielectric response

Sofar we have studied the behavior of the electron gas in thermodynamic
equilibrium. Let us now apply our results to study the response of the elec-
tron gas to a perturbation.
The response function we study first is the polarizability function χe ,
which was defined in Sec. 6.2. The polarizability function expresses the
induced charge density as a function of the external potential,
ρind (~q, ω) = χR
e (~
q , ω)φext (~q, ω). (7.30)
The polarizability function could be calculated as the retarded charge density
autocorrelation function, which is obtained as the analytical continuation of
the temperature charge density autocorrelation function
1
χR q, τ ) = − hTτ ρe,~q(τ )ρe,−~q(0)i
e (~ (7.31)
V
e2 X X †
= − hTτ ψ~k,σ (τ + η)ψ~k+~q,σ (τ )ψ~k† ′ +~q,σ′ (η)ψ~k′ ,σ′ (0)i.
V
~kσ ~k σ
′ ′
118
111
000
k 000
111 k’
−χe = 000
111
000
111 = + + +
000
111
000
111
k+q 000
111 k’+q
+ + + +
+ + + +
+ + ...
Figure 35: Diagrammatic expansion of the polarizability function χe . The figure

contains all diagrams to first order in the interaction and a subset of the diagrams
to second order in the interaction.
Diagrammatically, the polarizabibility function is obtained as in Fig. 35.

We can organize the diagrammatic expansion as we did previously when
we looked at the single-particle Green function. A polarization diagram is
called “reducible” if the two end points can be separated by cutting a single
interaction line. For example, of the diagrams in Fig. 35, the seventh, ninth,
tenth, eleventh, and thirteenth diagrams are reducible. All other diagrams
are irreducible in this sense. Defining the irredubile polarizability function as
(minus) the sum over all irreducible diagrams, we can write the polarizability
function as
χirr
e (~q, iΩn )
χe (~q, iΩn ) = . (7.32)
1 − Vq~χe (~q, iΩn )/e2
irr
The diagrammatic series for χirre is as in Fig. 36.

In the RPA approximation, one only keeps the zeroth order diagram for
irr
χe , which is the polarizability function for the noninteracting electron gas.
After analytical continuation to real frequencies, one finds
χ0R
e (~ q , ω)
χRPA
e (~q, ω) = . (7.33)
1 − Vq~χ0e (~q, iΩn )/e2
You may find that the calculation of the polarizability function is very
similar to that of the screened Coulomb interaction. That is no surprise.
119
111
000
k 000
111 k’
000
111
000
111
− χ irr
e =
000
111
000
111
000
111 = + + +
000
111
000
111
000
111
000
111
k+q 000
111 k’+q
+ + + + + ...
Figure 36: Diagrammatic expansion of the irreducible part χirr

e of the polarizabil-
ity function. The figure contains all diagrams to first order in the interaction and
a subset of the diagrams to second order in the interaction.
On the one hand, according to Eq. (7.30) above, the polarizabilitly function
describes what is the screening charge induced by an external potential. On
the other hand, the screened Coulomb potential describes how the “internal”
potential of a charged particle is modified by the charge density this particle
induces. As you see, one has solved the same problem twice!
We can now use our result for the polarizability function to discuss the
dielectric response function
ǫ−1 (~q, ω) = 1 + Vq~χR

e (~
q , ω). (7.34)
Substituting the RPA result (7.33), we find

−1
ǫ−1 (~q, ω) = 1 − χ0R

e (~
q , ω)Vq~ . (7.35)
The polarizability function of the non-interacting electron gas was calculated

in Ex. 6.2. In the static limit, ω → 0, the result is

0 kF q 2kF + q
χe (~q, 0) = −ν 1 + − ln
, (7.36)
q 4kF 2kF − q
where ν = mkF /2π 2 is the density of states at the Fermi level. Hence,
e2 kF m 2kF + q −1

−1 kF q
ǫ (~q, 0) = 1 + 1+ − ln . (7.37)
2ǫ0 q 2 π 2 q 4kF 2kF − q
The long wavelengths limit of Eq. (7.37) is the Thomas-Fermi dielectric func-
tion,
k2
ǫ(~q, 0) = 1 + 2s , (7.38)
q
where ks2 = me2 kF /π 2 ǫ0 is the inverse Thomas-Fermi screening length.
120
RPA
Uq = = + + + ...
RPA
Vq = = + + + ...
Figure 37: Diagrammatic representations of impurity potential and screened

Coulomb interaction in the RPA approximation.
An important feature of the RPA result for the static dielectric response
function is that it is singular at q = 2kF . This has a number of physical
consequences, that are discussed in the book by Doniach and Sondheimer.
For large frequencies ω ≫ qvF , the polarizability function χ0 can be
expanded as
kF3 q 2

R 3 qvF 2
χ0 (~q, ω) = 1+
3π 2 mω 2 5 ω
2

nq 3 qvF 2

= 1+ , (7.39)
mω 2 5 ω
where n = kF3 /3π is the density of electrons. Substituting this into the
dielectric response function gives
−1
ne2

−1 3 qvF 2
ε (~q, ω) = 1 − 1+ . (7.40)
mǫ0 ω 2 5 ω
Let us now consider the effect of impurities in the metal. The impuri-
ties are charged objects — otherwise they would hardly interact with the
electrons. However, in our previous treatment, we have modeled impurites
with a short-range potential. Now we see why that was correct: the impurity
potential is screened by the electron gas, so that
Uq~ → Uq~ + Uq~ χR

e (~
q, 0)Vq~. (7.41)
Diagrammatically, the screening of the impurity potential and of the Coulomb

interaction in the RPA approximation can be represented in striking similar-
ity, see Fig. 37.
121
7.6 Plasmon modes
For some frequency/wavelength combinations, the dielectric response func-
tion can vanish. Then, there is an induced charge density without an external
one! Such a situation is referred to as a “collective mode” of the electron
gas. It is a motion that can sustain itself without external input.
In priciple, the dielectric response function is complex. In those cases,
the occurence of zeroes is unlikely. However, there are regions of (q, ω) space
where ǫ is real. In such regions, long-lived collective excitations can be ex-
pected. The support of the imaginary part of the polarizability function
of the noninteracting electron gas (which equals the support of Im ǫ−1 ) was
shown in Fig. 20. Based on that figure, we expect collective excitations in the
region of finite frequencies and long wavelengths. In this parameter regime,
the dielectric response function is given by Eq. (7.40) above. We see that,
indeed, there are zeroes, given by the dispersion relation
3q 2 vF2 2 ne2
ω(~q) = ωp + , ωp = . (7.42)
10ωp mǫ0
The frequency ωp is known as the “plasma frequency” and the correspond-

ing collective modes are known as “plasma oscillations” or “plasmons”. The
plasmon modes are long-lived as long as the dielectric response function re-
mains real. At the point when the dispersion curve (7.42) enters that region
of (q, ω) space where ǫ−1 is complex, the zeroes of ǫ at real frequency ω
and wavevector q cease to exist. In this case, one can still find zeroes of
ǫ at complex frequencies. These correspond to damped collective modes.
The damping that arises from the fact that ǫ is complex is called “Landau
damping”.
The plasmon dispersion relation is shown schematically in Fig. 38, to-
gether with the support of Im ǫ(q, ω).
7.7 Free energy

With all the diagrammatic rules we derived above, the easiest way to calculate
the free energy of the interacting electron gas is to use Eq. (3.116) of exercise
3.2, Z 1
1
F − F0 = dλ hλH1 iλ , (7.43)
0 λ
122
11111111111111111111
00000000000000000000
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
ω/ε F 4 00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
2 00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
00000000000000000000
11111111111111111111
0000000000
1111111111
0 00000000000000000000
11111111111111111111
0000000000
1111111111
0 1 2 3 4
q/k F
Figure 38: Plasmon dispersion relation (thick curve) together with the support
of Im ǫ. Plasmon modes are strongly damped (Landau damping) if Im ǫ 6= 0.
where F0 is the free energy of the noninteracting electron gas and the pa-
rameter λ multiplies the Coulomb interaction. Rewriting H1 as
1 X X 1 X
H1 = Vq~ψ~k† +~q,σ ψ~k1 ,σ1 ψ~k† −~q,σ ψ~k2 ,σ2 − n Vq~, (7.44)
2V σ ,σ 1 1 2 2 2
1 2 ~k1 ,~k2 ,~
q q~
we can express the average hH1 i as

1 X 1 X
hH1 i = − 2 lim Vq~χe (~q, τ ) − n Vq~,
2e τ ↑0 2
q~ q~
T X X −iΩn η 1 X
= − 2 e Vq~χe (~q, iΩn ) − n Vq~, (7.45)
2e n
2
q~ q~
where χe is the polarizability function. Within the random phase approx-

imation we can use Eq. (7.33) for χe , replace Vq~ by λVq~ and perform the
integration to λ. The result is
T X X iΩn η 1 X
F − F0 = e ln 1 − Vq~χ0e (~q, iΩn )/e2 − n Vq~. (7.46)
2 n
2
q~ q~
We can perform the summation over n writing F − F0 as a contour integral

in the complex plane and shifting the contours to the real axis,
X Z ∞ dξ 1 X
F − F0 = coth(ξ/2T )Im ln (ǫ(~q, ξ)) − n Vq~. (7.47)
−∞ 4π q~
2
q~
Here we used Eq. (7.35) to express the argument of the logarithm in terms
of the RPA dielectric response function.
123
Evaluation of this expression for the Free energy is not entirely straight-
forward. Some insight into the physics of Eq. (7.47) is obtained if we look
at the contribution from the plasmons. Using Eq. (7.40) for the dielectric
response function ǫ(~q, ω) near the plasmon branch ω ≈ ωp (~q) ≫ qvF at low
wavenumber, we find

 0 if |ω| > ωp (~q),
Im ln ǫ(~q, ω) = π if qvF ≪ ω < ωp (~q), (7.48)
−π if −ωp (~q) < ω ≪ −qvF .

We then see that each plasmon mode with vF q ≪ ωp (~q) has a contribution
1 ωp (~q)
Z
δFq~ = dξ coth(ξ/2T )
2
= T ln(sinh(ωp /2T )) − const, (7.49)
where the additional constant reflects the contribution from the integration
close to ξ = 0 where the approximation (7.48) does not hold. The first term
on the r.h.s. of Eq. (7.49) represents the free energy of a plasmon mode at
wavevector ~q, while the additional constant represents the contribution from
low energy particle-hole excitations to the free energy.
Doniach and Sondheimer report a result for the ground state energy as
an expansion for small gas parameter rs ,

0.916
E − E0 = − + 0.0622 ln rs − 0.094 Ry. (7.50)
rs
In the same units, the ground state energy of the non-interacting electron
gas is (2.21/rs2) Ry. The unit Ry, for Rydberg, corresponds to e2 /2a0 ≈ 13.6
eV. Higher-order terms in the small-rs expansion of the ground state have
been calculated. However, for those terms, one needs to go beyond the RPA
approximations.
124
7.8 Exercises
Exercise 7.1: Coulomb interactions in two dimensions
In certain materials, such as a quantum well at the interface of the semicon-

ductors GaAs and GaAlAs, the electrons are confined to two spatial dimen-
sions only.
For such a system, the non-interacting part of the Hamiltonian can be
written X †
H0 = ε~kσ ψ~kσ ψ~kσ , (7.51)
~kσ
where the summation over wavevectors ~k is restricted to wavevectors in the

†
plane of the two-dimensional electron gas. The creation operator ψ~kσ creates
an electron with spin σ and orbital wavefunction
1 i~k·~r
ψ~k (~r, z) = e ζ(z), (7.52)
A1/2
where ~r is the coordinate in the plane of the two-dimensional electron gas,
and z is the transverse coordinate, and A is the area of the two-dimensional
electron gas. For most practical purposes, the function ζ(z) can be approxi-
mated by the delta function δ(z).
The interaction between the electrons in the two-dimensional electron gas
is the three-dimensional Coulomb interactions,
1
V (~r) = . (7.53)
4πǫ0 ǫr |~r|
The relative permittivity ǫr appears because the two-dimensional electron gas

is not surrounded by vacuum, but by a semiconducting material. For this
system, write down the interaction Hamiltonian for the Coulomb interaction
using the momentum representation. Find an explicit expression for the
Coulomb-interaction matrix element Vq~ in two dimensions.
Exercise 7.2: Four-vertex formalism
Abrikosov, Gorkov, and Dzyaloshinskii discuss a different diagrammatic for-

malism to describe electron-electron interactions. The diagrams for the direct
125
ν ν’ ν ν’ ν ν’
µ µ’ µ µ’ µ µ’
Figure 39: Diagrammatic representation of both the direct and exchange parts of
the electron-electron interaction, following the notation of the book by Abrikosov,
Gorkov, and Dzyaloshinskii.
and the exchange interaction are combined into one diagram, as shown in Fig.
39.
Discuss the diagrams contributing to the self-energy to second order in
the interaction using this formalism. Also discuss the Hartree-Fock approxi-
mation. Does the simplest four-vertex formulation of an effective interaction
agree with the RPA approximation? Why (not)?
Exercise 7.3: Plasma oscillations
It is possible to find the plasma frequency ωp from simple classical consid-

erations. Consider an electron gas in a rectangular box of length Lx and
cross section A. Since the mass of the ions is much larger than the mass of
the electrons, the ions can be treated as an inert positive background charge
that compensates the charge of the electrons. The electron gas can be set in
oscillating motion by translating it a distance δx in the x-direction, leaving
the ion background fixed. Show that the frequency of this oscillation is pre-
cisely the plasma frequency ωp . Can you generalize this argument to plasma
oscillations at finite wavevector ~q?
Exercise 7.4: Plasma oscillations in two dimensions
In this exercise we return to the case of a two-dimensional electron gas that

exists at a quantum well in a semiconductor heterostructure.
(a) Calculate the dielectric response function ǫ(~q, ω) for a two-dimensional

electron gas in the RPA approximation.
(b) From the zeroes of the dielectric response function, calculate the dis-
126
persion relation of plasma oscillations in the two-dimensional electron
gas. Discuss for what wavenumbers plasma oscillations are damped by
Landau damping (excitation of particle/hole pairs).
(c) Calculate the screened Coulomb interaction in a two-dimensional elec-

tron gas in the RPA approximation.
(d) Show that, at small wavevectors q ≪ kF and low frequencies ~ω ≪ εF

the screened interaction is given by
e2
Vq~RPA = , (7.54)
2ǫ0 ǫr (q + ks )
where ks is the two-dimensional Thomas-Fermi screening wavenumber.

Calculate ks for the case of a GaAs/GaAlAs heterostructure, for which
the effective electron mass is m∗ = 0.067m, relative permittivity ǫr =
13, and electron density n = 1015 m−2 . How does it compare to the
Fermi wavelength for this system?
127
128
8 Magnetism
In this chapter we study the magnetic excitations in a solid. We start with
a calculation of the magnetic susceptibility χ, which is the coefficient of pro-
portionality between the spin polarization induced by a magnetic field and
the magnetic field ~h. The susceptibility not only tells us how easy it is to
magnetize a solid, it also points to the onset of a magnetic phase. This
happens at the point where the susceptibility diverges. A divergent suscepti-
bility signals an instability: a spin polarization is formed for arbitrarily weak
magnetic field.
The magnetic instability is governed by a competition between two ef-
fects. On the one hand, the kinetic energy of the electrons is minimal if
there is no net spin polarization, because then all energy levels ε~k can be
doubly occupied. On the other hand, the short-range part of the electron-
electron interaction favors a magnetic state. This is because electrons with
the same spin have a strongly reduced repulsion since the Pauli principle
already guarantees that they are spatially separated.
We first consider the susceptibility for a gas of interacting electrons with
a parabolic dispersion (“free electrons”). This is the example that we studied
in detail in the previous chapters. A calculation within the RPA approxima-
tion shows that electron-electron interactions enhance χ, but also that they
do not lead to a divergence. In other words, for free electrons, the competi-
tion between kinetic energy and interaction energy always favors the kinetic
energy, and, hence, forbids a spontaneous spin polarization.
The picture is different once the ionic lattice is taken into account. Band
structure effects turn out to be crucial for the occurence of a ferromagnetic
instability. In a narrow band, the kinetic energy cost for polarizing elec-
tron spins is strongly reduced, and a magnetic instability becomes possible.
Depending on the material one looks at, one may have a ferromagnetic insta-
bility (uniform magnetization) or a spin-wave material (nonuniform magne-
tization). One example of a spin-wave material is an antiferromagnet, where
the spin polarization has opposite direction on neighboring lattice sites.
8.1 Magnetic susceptibility

In order to make the above picture more quantitative, we consider an electron
gas in the presence of a magnetic field ~h. We are interested in the spin density
129
~s(~r),
1X †
sα (~r) = ψ (~r)(σα )σ,σ′ ψσ′ (~r), α = x, y, z, (8.1)
2 ′ σ
σ,σ
where σα is the Pauli matrix. The response of a magnetic material to

an applied magnetic field is characterized by the magnetic susceptibility
χ(~r, t; ~r′, t′ ), which relates the change in spin density at position at position
~r and time t to the magnetic field at position ~r′ and time t′ ,
XZ Z
′
δsα (~r, t) = d~r dt′ χαβ (~r, t; ~r′ , t′ )hβ (~r′, t′ ). (8.2)
β
Note that χαβ is a tensor: in principle, a magnetic field ~h can cause a spin
density in a different direction than the direction of ~h. The magnetic field ~h
enters into the Hamiltonian through the Zeeman coupling,
Z
~
H1 = − µB g d~r~h(~r, t) · ~s(~r), (8.3)
2
where µB = e~/2mc is the Bohr magneton and g = 2 the electron g factor.
According to the Kubo formula, the spin susceptibility is equal to the
retarded spin-spin correlation function,
µB g
χαβ (~r, t; ~r′, t′ ) = i θ(t − t′ )h[sα (~r, t), sβ (~r′ , t′ )]− i. (8.4)
2
Note the absence of a minus sign on the r.h.s. of Eq. (8.4) because of the
minus sign in front of the magnetic field in the expression (8.3) for the Zeeman
energy. We may perform a Fourier transform to find the susceptibility in
frequency representation,
XZ
δsα (~r, ω) = d~r′χαβ (~r, ~r′; ω)hβ (~r′, ω), (8.5)
β
where
µB g ∞
Z
′
χαβ (~r, ~r ; ω) = i dteiωt h[sα (~r, t), sβ (~r′, 0)]− i. (8.6)
2 0
For a spatially homogeneous system, we may perform one further Fourier
transform to the coordinate ~r, which gives a susceptibility χαβ (~q, ω),
1
Z
′
χαβ (~q, ω) = d~rd~r′ e−i~q·(~r−~r ) χαβ (~r, ~r′; ω). (8.7)
V
130
For the calculation of χαβ (~q, ω) it is useful to write the spin density in
terms of the plane wave basis,
µB g
χαβ (~q; t, t′ ) = iθ(t − t′ ) h[sq~,α (t), s−~q,β (t′ )]− i, (8.8)
2V
where
1 X †
sq~,α = ψ~k,σ (σα )σ,σ′ ψ~k+~q,σ′ . (8.9)
2 ′~k,σ,σ
In practice, the x and y labels of the tensor χαβ are often replaced by
labels “+” and “−”, referring to the linear combinations
s± = s1 ± is2 . (8.10)
These combinations are expressed in terms of raising and lowering operators

for the spin, which are much easier to calculate. In particular, we have the
susceptibility
µB g
χ−+ (~r, t; ~r′, t′ ) = i θ(t − t′ )h[s− (~r, t), s+ (~r′ , t′ )]− i. (8.11)
2
The susceptibility χ−+ is known as the “transverse” susceptibility, while χzz
is known as the “longitudinal” susceptibility. For an isotropic system, the
tensor χαβ is proportional to the unit tensor. In that case one has
1
χαβ = δαβ χ−+ , α, β = x, y, z. (8.12)
2
A calculation of the susceptibility for non-interacting electrons is very
similar to the calculation of the polarizability that we discussed in Sec. 6.2.
One finds
µB g X †
χ−+ (~q, t) = i θ(t) h[ψ~k,↓ (t)ψ~k+~q,↑ (t), ψ~k† ′ ,↑ (0)ψ~k′ −~q,↓ (0)]− i.(8.13)
2V ′ ~k,~k ,~
q
Diagrammatically, the susceptibility for non-interacting electrons can be rep-

resented by the diagram shown in Fig. 40. The white dots are spin vertices.
For the transverse susceptibility χ−+ , they enforce that the ingoing and out-
going spins have different sign, see Fig. 40.
131
k k
χ −+=
k+q k+q
Figure 40: Diagrammatic representation of the transverse spin susceptibility χ−+

for non-interacting electrons.
As a matter of fact, we can find χ−+ without doing a calculation if we

compare Eqs. (8.13) and the expression for the polarizability of the non-
interacting electron gas, Eq. (6.29). We then immediately conclude
µB g X nF (ε~k − µ) − nF (ε~k+~q − µ)
χ0−+ (~q, ω) = − , (8.14)
2V ε~k − ε~k+~q + ω + iη
~k
where nF is the Fermi function and η is a positive infinitisimal.

In the limit ~q → 0, ω → 0 of a static and uniform magnetic field, one
has χ−+ → νµB g/2, where ν is the density of states per unit volume and per
spin direction. This result is known as the “Pauli susceptibility”. The Pauli
susceptibility accounts for the fact that it costs a finite kinetic energy to
have a finite spin density. The value of the Pauli susceptibility then follows
from the competition of the magnetic energy gain (proportional to the spin
density and the magnetic field) and the kinetic energy loss (proportional to
the square of the spin density).
The imaginary part of the susceptibility χ0 (~q, ω) characterizes energy
dissipation through magnetic excitations. The support of the imaginary part
of χ0 is the same as that of the imaginary part of the polarizability of the
non-interacting electron gas, see Fig. 20).
For the interacting electron gas, we can use the diagrammatic theory to
compute the susceptibility. Some diagrams are shown in Fig. 41. They can
be rearranged in three steps. First, we incorporate all interaction lines that
connect the upper or lower fermion line to itself into single-particle Green
functions. Second, we replace the interaction by the screened interaction by
inserting any number of “bubbles” into the interaction. Below, the single-
electron Green function will be calcualted using the Hartree-Fock approxi-
mation, the effective interaction in the RPA. Third, we write the diagram in
terms of a “vertex correction” as in Sec. 6.3, see Fig. 41. The renormalized
vertex, which is indicated by a black dot, then contains all contributions from
interaction lines that connect the upper and lower fermion lines.
132
111
000
k 000
111
000
111 k
000
111
χ −+= 000
111
000
111
000
111 = + + +
000
111
000
111
000
111
k+q 000
111 k+q
+ + σ + + + ...
σ
ωm
k
=
k+q
ωm+Ω n
Figure 41: Diagrammatic representation of the contributions to the transverse

spin susceptibility χ−+ for interacting electrons, up to second order in the inter-
action. The arrows indicate the spin direction, while “σ” denotes an additional
summation over spin. Note that spin is conserved throughout the upper and lower
fermion lines.
In the spirit of the RPA approximation, we keep only “ladder” diagrams

for the vertex correction, see Fig. 42. In this approximation, the calculation
has become very similar to that of the current-current correlation function
of Sec. 6.3. Denoting the renormalized vertex by Γ(~k, iωm ; ~k + ~q, iωm + iΩn )
and the screened interaction by Vq~eff (iΩn ), we have
T XX
Γ(~k, iωm ; ~k + ~q, iωm + iΩn ) = 1 − G~ ′ ~ ′ (iωp )
V ′ p k ↑,k ↑
~k
eff
× G~k′ +~q↓,~k′ +~q↓ (iωp + iΩn ) − V~k− ~k ′ (iωm − iωp )
× Γ(~k ′ , iωp ; ~k ′ + ~q, iωp + iΩn ). (8.15)
Note that the effective interaction depends on the Matsubara frequency dif-
ference iωm − iωp , in contrast to the bare interaction which is frequency
independent. Then, in terms of the renormalized vertex, the spin suscepti-
bility is calculated as
T µB g X X ~
χRPA q , iΩn ) = −
−+ (~ Γ(k, iωm ; ~k + ~q, iωm + iΩn )
2V m ~k
× G~k↑,~k↑ (iωm )G~k+~q↓,~k+~q↓ (iωm + iΩn ). (8.16)
The analytical continuation iΩn → ω + iη is done as in Sec. 6.3. The result
133
ωp ωp
ωm k ωm k k’ k k’ k" k
= + + + ...
k+q k+q
ωm+Ω n ωm+Ω n k’+q
k+q ω +Ω
k’+q k"+q
k+q
ωp +Ωn m n
ωp
ωm k k’ k
= +
k+q
ωm+Ω n k’+q
k+q
ωp +Ωn
Figure 42: Diagrammatic representation of the RPA approximation for the trans-
verse spin susceptibility χ−+ .
is
µB g
Z X
χ−+ (~q, ω + iη) = − dξ tanh(ξ/2T )
8πiV
~k
h
× ΓRR (~k, ξ; ~k + ~q, ξ + ω)G~Rk↑,~k↑
(ξ)G~R q ↓,~k+~
k+~ q↓
(ξ + ω)
− ΓAR (~k, ξ; ~k + ~q, ξ + ω)G~A
k↑,~k↑
(ξ)G~R q↓,~k+~
k+~ q↓
(ξ + ω)
+ ΓAR (~k, ξ − ω; ~k + ~q, ξ)G~A
k↑,~k↑
(ξ − ω)G~R q↓,~k+~
k+~ q↓
(ξ)
i
−Γ AA ~ ~ A A
(k, ξ − ω; k + ~q, ξ)G~k↑,~k↑ (ξ − ω)G~k+~q↓,~k+~q↓ (ξ) .
(8.17)
For the full frequency and momentum dependent screened interaction,

the RPA equation for the renormalized vertex cannot be solved in closed
form. Here, we’ll adopt the Thomas-Fermi approximation and replace the
screened interaction by the interaction at ~q = 0, ω = 0,
e2
Vq~TF (iΩn ) ≈ , (8.18)
ǫ0 ks2
where ks2 = 2νe2 /ǫ0 is the Thomas-Fermi wavenumber. This approxima-

tion corresponds to a delta-function interaction in real space, V TF (~r) ≈
(2ν)−1 δ(~r). With this approximation, calculation of the renormalized ver-
tex is straightforward, and we find the simple result
Γ(~k, iωm ; ~k + ~q, iωm + iΩn )
134
 −1
T X X
= 1 + G~k′ ↑,~k′ ↑ (iωp )G~k′ +~q↓,~k′+~q↓ (iωp + iΩn ) . (8.19)
2νV p ~k ′
The summation in the denominator has the same expression as the suscepti-
bility for non-interacting electrons, the only difference being that the single-
electron Green functions have been replaced by their values in the interacting
electron liquid. In other words, this is the susceptibility for the interacting
electron liquid without vertex renormalization. Using the Hartree-Fock ap-
proximation for the single-particle Green function, we write
µB gT X X
χ0,HF
−+ (~
q , iΩn ) = − G~k↑,~k↑ (iωm )G~k+~q↓,~k+~q↓ (iωm + iΩn )
2V m~k
HF HF
µB g X nF (ε~k,↑ − µ) − nF (ε~k+~q,↓ − µ)
= − , (8.20)
2V ε~HF
k,↓
− ε~HF
k+~
q ,↑
+ iΩn
~k
so that we can express the full transverse susceptibility in the Thomas-Fermi

approximation as
χ0,HF
−+ (~q , ω)
χTF
−+ (~
q, ω) = 1 0,HF
. (8.21)
1 − νµB g χ−+ (~q, ω)
For a uniform static magnetic field, one has χ0,HF

−+ (0, 0) = νµB g/2, where
the density of states ν should be interpreted as the “density of poles” of the
full Green function, i.e., the density of Hartree-Fock energy levels. Since this
is the same density of states as the one that occurs in the screened interaction,
one has
χTF
−+ (0, 0) = νµB g. (8.22)
We conclude that, according to the Thomas-Fermi approximation, in an
interacting electron gas the electron-electron interactions increase the spin
susceptibility by a factor two compared to the case of the non-interacting
electron gas.25
25
The Thomas-Fermi approximation is not quantitatively correct. A more correct treat-
ment of the RPA approximation which takes into account the momentum dependence of
the effective interaction gives
νµB g
χRPA
−+ = ,
2 − (2/3π ) rs ln(1 + (3π 2 /2)2/3 /rs )
2 2/3
see Ex. 8.1.
135
The enhancement of the susceptibility by the interactions follows from
the fact that the repulsive interaction between the electrons is lowered if
a finite spin density is created. After all, the Pauli principle forbids that
electrons of the same spin are at the same location, so that they have a
smaller electrostatic repulsion. The interaction energy gain is proportional
to the square of the spin density, so that the net energy cost for polarizing
the system is reduced with respect to the case of non-interacting electrons.
If the interaction energy gain is larger than the kinetic energy cost, a
spin polarization will form spontaneously. This instability is known as the
“Stoner instability”. We see that for the standard case of a parabolic band
the Stoner instability does not occur. The situation becomes different once
the precise band structure of the material is taken into account. The next
section discusses a simple model for this.
8.2 Hubbard model

The approximations of the previous subsection are appropriate for the stan-
dard parabolic band of free electrons. In real materials, the electrons occupy
bands, and the band structure needs to be taken into account for a calculation
of the magnetic susceptibility. A simple model to take the band structure
into account is the “Hubbard model”.
For the construction of the Hubbard model, we consider electrons in an
energy band with Bloch energies ε~k described by the Hamiltonian
X
H0 = ε~k ψ~k† ψ~k . (8.23)
~k
For such a band, wavefunctions can be represented in the basis of Wannier

wavefunctions at lattice site ~ri ,
X ~
φi (~r) = N −1/2 eik·~ri ψ~k (~r), (8.24)
~k
where the summation is restricted to wavevectors in the first Brillouin zone

and N is the number of lattice sites. The Wannier wavefunctions at different
lattice sites are orthogonal.
The Coulomb interaction can be written in terms of the basis of Wannier
wavefunctions. The corresponding matrix elements are
Z
Vij,i′j ′ = d~r1 d~r2 φ∗i (~r1 )φ∗j (~r2 )V eff (~r1 − ~r2 )φj ′ (~r2 )φi′ (~r1 ), (8.25)
136
where V eff is the screened Coulomb interaction. The approximation of the
Hubbard model consists in neglecting all interaction matrix elements except
for the matrix element U = Viiii . The justification for this approximation is
that the screened interaction is short range, while the Wannier wavefunctions
are mainly localized at one lattice site. This approximation leads to the
interaction Hamiltonian
X
H1 = U/2 niσ niσ′ , (8.26)
iσσ′
†
where niσ = ψiσ ψiσ is the Wannier number operator and
X † ~
ψiσ = N −1/2 ψ~kσ eik·~ri (8.27)
~k
is the Wannier annihilation operator.

The interaction Hamiltonian can be further simplified if we use n2iσ = niσ ,
so that X X
H1 = U ni↑ ni↓ + (U/2) (ni↑ + ni↓ ). (8.28)
i i
The second term is nothing but a shift of the chemical potential and will be
neglected. Fourier transforming the first term in Eq. (8.28) we then arrive at
U X †
H1 = ψ ψ~ ψ † ψ~ ′ . (8.29)
N ′ ~k,↑ k+~q,↑ ~k′ ,↓ k −~q,↓
~k,~k ,~
q
Let us now calculate the spin susceptibility for the Hubbard model. The
calculation proceeds along the same lines as that of the previous section. The
only differences are that, for the Hubbard model, the wavevector summation
is restricted to the first Brillouin zone, and that the strength of the interac-
tion is set independently by the interaction constant U. The result for the
transverse susceptibility is
χ0,HF
−+ (~ q , ω)
χ−+ (~q, ω) = 2U 0,HF
, (8.30)
1 − µB g χ−+ (~q, ω)
where
HF HF
χ0,HF
−+ (~q, ω) = . (8.31)
2N ε~HF
k+~q ,↓
− ε~
HF
k,↑
+ ω + iη
~k
137
For the Hubbard model, the Hartree-Fock energy levels have a very simple
connection to the bare energy levels ε~k ,
1 X
ε~HF
kσ
= ε~k + U n̄−σ , n̄σ = nF (ε~HF
kσ
)i. (8.32)
N
~k
In the absence of a spontaneous magnetization, the levels of electrons with

spin up and spin down experience the same uniform shift, which can be
incorporated into the chemical potential. In particular, this implies that the
support of the imaginary part of χ0,HF −+ and, hence, of the imaginary part of
χTF
−+ , is the same as that of χ0
−+, which is shown in Fig. 20. In the presence
of a spontaneous magnezation, the Hartree-Fock energy levels for electrons
with spin up and spin down experience different shifts. We return to this
case in Sec. 8.5.
As before, the limit of large wavelength and low frequency has χ0,HF−+ (0, 0) =
νµB g/2, where ν is the density of states per spin direction and per lattice
site. Hence
νµB g/2
χTF
−+ (0, 0) = . (8.33)
1 − Uν
We can use these results to find the interaction correction to the Free
energy for the Hubbard model. Repeating the analysis of Sec. 7.7, we find
the result
Z ∞
dξ X
F − F0 = coth(ξ/2T )Im ln(1 − 2Uχ0,HF (~q, ξ)/µB g). (8.34)
−∞ 4π
q~
This result may be analyzed using the detailed expressions for χ0,HF of Ex.
6.2. For low temperatures, the main contribution comes from frequencies
around ω = 0, and one finds that the specific heat is enhanced close to the
Stoner instability,
CV ∝ −T ln(1 − Uν). (8.35)
This is to be compared to the usual specific heat of the noninteracting electron
gas, which is proportional to νT .
8.3 The Stoner instability

As long as the interaction constant U is small enough, the susceptibility
remains finite. A finite susceptibility implies that it requires a finite value of
138
the magnetic field ~h to create a spin polarization. However, if the interaction
strength increases, the susceptibility increases. If the static susceptibility
χ(~q, 0) diverges, an instability occurs, and a spontaneous magnetization will
form.
From our general result (8.30) for the Hubbard model we find that the
instability criterion is26
U = Uc (~q) = 1/χ0,HF(~q, 0). (8.36)
This is the Stoner criterion.

For zero wavevector ~q = 0, the Stoner criterion describes the ferromag-
netic instability, which occurs at critical interaction strength Uν = 1. For
a parabolic band, the instability occurs first at ~q = 0, since χ0,HF (~q, 0) is a
monotonously decreasing function of |~q|, cf. Ex. 6.2. However, in the presence
of a more complicated band structure, instabilities at finite wavenumbers are
also possible. An instability at finite ~q leads to the formation of a so-called
“spin density wave” state, as was first pointed out by Overhauser [Phys. Rev.
128, 1437 (1962)]. As an example, we consider the band
X
ε~k = −2t cos kα a, (8.37)
α=x,y,z
which corresponds to electrons on a simple cubic lattice with lattice constant

a and hopping amplitude t, see Ex. 8.7. For this band, there is a competition
between the ferromagnetic instability at wavevector ~q = 0 and an antiferro-
magnetic instability at wavevector ~q = ~qA = (π/a)(1, 1, 1) at the far corner
of the first Brillouin zone. You verify that one has
Z µ
0,HF ν(ξ)
χ (~qA , 0) = − dξ . (8.38)
−6t ξ
For chemical potentials close to the band edges ±6t, the instability occurs at
wavevector ~q = 0, while for a chemical potential µ close to the band center at
ε = 0, the instability occurs at ~q = ~qA . For this simple cubic band structure,
26
For finite frequencies, the denominator of Eq. (8.30) is generally complex, so that one
can rule out the existence of zeroes. Physically, one expects the instability to happen at
zero frequency, since an instability at nonzero frequency correspons to a time-dependent
polarization, which is damped. Response at nonzero frequencies is studied in a later
section.
139
one has χ0,HF (~qA , 0) → ∞ as µ → 0, so that the antiferromagnetic instability
occurs already for arbitrarily small repulsive interaction!
Note that for interaction strength U > Uc , our expression (8.30) for the
spin susceptibility in the Thomas Fermi approximation is, again, finite. How-
ever, this result is unphysical, since, for U > Uc the ground state has a spon-
taneous spin polarization (or a spin-density wave state). Hence, we conclude
that Eq. (8.30) is valid for all U < minq~ Uc (~q), whereas a state with sponta-
neous magnetization is formed for U > Uc . A description of the ferromagnetic
state is given in Sec. 8.5 below.
8.4 Neutron scattering

Neutrons have spin 1/2. The neutron spins interact with the spins of elec-
trons in a solid through the magnetic dipole interaction. Hence, the neutron
scattering cross section is sensitive to the magnetic structure of a solid.
In order to describe the scattering of slow neutrons, we return to the first
Born approximation for the neutron scattering cross section, but take the
neutron spin into account explicitly,
2 D
d2 σ 1 X k′ M E
= |h~k ′σf , f |H ′|~kσi , ii|2 δ(ω + Ei − Ef ), (8.39)
dΩdω 2 f,σ ,σ k 2π
f i
where σf and σi denote the outgoing and incoming spin states of the neutron.
The extra factor 1/2 is added since we average over the possible incoming
spin states. For the interaction between an electron and a neutron we take
the first quantization form27
H ′ = V0 (~rn − ~r) + V1 (~rn − ~r)~sn · ~s, (8.40)
where ~rn and ~sn are the position and spin of the neutron, while ~r and ~s are
the position and spin of the electron. Using the second quantization language
for the electrons, H ′ can be rewritten as
Z Z
′
H = d~rV0 (~rn − ~r)n(~r) + ~sn · d~rV1 (~rn − ~r)~s(~r), (8.41)
27
For more details: see L.D. Landau and E.M. Lifschitz, Quantum mechanics, non-
relativistic theory, Addison and Wesley (1958).
140
where n(~r) and ~s(~r) are the second quantization operators for the electron
number density and spin density. Writing ~q = ~k − ~k ′ , we then find
Z
~ ′ ′~
hk σf , f |H |kσi , ii = d~rn d~re−i~q·~rn V0 (~rn − ~r)δσf ,σi hf |n(~r)|ii
1X
Z
+ d~rn d~re−i~q·~rn V1 (~rn − ~r)(σα )σf σi hf |s(~r)|ii
2 α
V1~q X
= V0~qhf |ρq~|iiδσf ,σi + (σα )σf σi hf |sq~α |ii, (8.42)
2 α
where V0~q and V1~q are the Fourier transforms of V0 and V1 , respectively. Next,
we write the delta function in Eq. (8.39) as an integration over time, absorb
the factors exp(iEi t) and exp(−iEf t) into a time-shift of the operators n and
~s, and perform the summation over final states. The result is
2 ′
d2 σ M k
= (8.43)
dΩdω 2π 2πk
|V1~q|2
Z
iωt 2
× dte |V0~q| hn−~q(0)nq~(t)i + h~s−~q(0) · ~sq~(t)i .
4
For an isotropic system, the spin-spin correlation function appearing in

the spin-dependent part of the inelastic neutron scattering cross section is
an example of a “greater Green function”, see Sec. 2.3. It can be written in
terms of the spin raising and lowering operators s+ and s− ,
3
h~s−~q(0) · ~sq~(t)i = hs−,−~q(0)s+,~q(t)i. (8.44)
2
Using the general relation (2.46) between Green functions, the Fourier trans-
form with respect to time can be written in terms of the imaginary part of
the retarded spin density correlation function,
3 Im χR−+ (~
q , ω)
Z
dteiωt h~s−~q(0) · ~sq~(t)i = . (8.45)
2µB g 1 − e−~ω/T
The minus-sign in the denominator appears because χ is a correlation func-
tion of spin densities, which are even in fermion creation and annihilation
operators.
141
The approach to the instability point leads to a strong enhancement of
the neutron scattering cross section. For small ~q and ω and a parabolic band,
one has (cf. Ex. 6.2)
πµB gω/8qvF
Im χR
−+ (~
q , ω) = . (8.46)
(1 − Uν)2 + (Uπω/4qvF )2
Thus, as the ferromagnetic instability is approached, inelastic neutron scat-

tering is strongly enhanced for low frequences. The physical origin of the
enhanced neutron scattering rate at low temperatures is the slowing down of
spin relaxation at the approach of the Stoner instability.
8.5 The ferromagnetic state

Beyond the instability point a spontaneous magnetization forms. In princi-
ple, the magnetization can point in any direction, although, for each sample,
a specific direction will be selected. Here, we assume that the magnetization
points in the positive z-direction.
The main difference between the ferromagnetic state and the paramag-
netic state is that, in a ferromagnet, the Hartree-Fock energy levels are dif-
ferent for up spins and down spins, see Eq. (8.32). For the Hubbard model,
the difference between levels for spin up and spin down does not depend on
the wavevector,
∆ = ε~HF
k↓
− ε~HF
k↑
= U(n̄↑ − n̄↓ ). (8.47)
The existence of the gap ∆ implies that electrons with spin up have a larger
Fermi momentum than electrons with spin down, see Fig. 43, which, in turn
reflects the fact that there is a net spin polarization. The electrons with
spin parallel to the magnetization direction are referred to as “majority elec-
trons”, whereas those with opposite to the magnetization direction are called
“minority electrons”.
One can calculate ∆ self-consistently from the requirement that the total
number of particles does not change upon the formation of the magnetic
moment. Together with Eq. (8.32), one can then solve for the energy gap ∆.
Nontrivial solutions exist for Uν > 1 only.
The difference between Hartree-Fock energy levels for different spin di-
rections leads to a modification of the expression for the transverse spin
142
kF
kF
Figure 43: Support of the imaginary part of χ−+ for a ferromagnet and spin
wave dispersion.
susceptibility,
HF HF
χ0,HF
−+ (~q , ω) =− . (8.48)
2N ε~k − ε~k+~q + ∆ + ω + iη
~k
The support of the imaginary part of χ0,HF and, hence, of χ, is affected

by the appearance of the gap ∆: The imaginary part of χ−+ is nonzero
if there are excitations of pairs of a particle and a hole of opposite spin.
Such excitations are called “Stoner excitations”. The existence of the gap ∆
implies that Stoner excitations at ~q = 0 require an energy ∆. Zero frequency
Stoner excitations require q to be at least equal to the difference between
the Fermi momenta of electrons with spin up and spin down. Repeating the
arguments of Sec. 6.2, we can find an expression for the support of Im χ−+ for
a ferromagnet, somewhat unrealistic case of a parabolic band. For positive
ω one finds that, at zero temperature, Im χ−+ is nonzero for
∆ q2 ∆ q2
−vF q + + < ω < vF q + + . (8.49)
~ 2m~ ~ 2m~
This support is shown in Fig. 44.
Substitution into Eq. (8.30) then yields for ~q → 0
gµ ∆
B
χ−+ (~q → 0, ω) = . (8.50)
2 Uω
143
111111111111111111111
000000000000000000000
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
ω/ε F 4 000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
Stoner
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
excitations
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
2 000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
spin waves
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
∆ 000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
0 000000000000000000000
111111111111111111111
000000000000000000000
111111111111111111111
0 1 2 3 4
q/k F
Figure 44: Support of the imaginary part of χ−+ for a ferromagnet and spin
wave dispersion.
The pole of χ as ω → 0 corresponds to a uniform rotation of the magne-

tization. This collective excitation has zero frequency because no energy is
required for such a uniform rotation.28 At finite wavevector ~q, the transverse
spin susceptibility χ−+ has a pole at frequencies ω = Dq 2 , where D is a pro-
portionality constant. Such an excitation is known as a “spin wave”. Spin
waves are virtually undamped for small wavenumbers. At larger wavenum-
bers, the spin wave branch enters into the region of the (q, ω) plane where
χ−+ is complex. Then, spin waves can decay by excitation of “Stoner exci-
tations”, a pair of a particle and a hole of opposite spin, see Fig. 44.
The fact that spin waves exist at low frequencies if ~q → 0 is guaranteed
by the rotation symmetry of the ferromagnetic state. A uniform rotation
of all spins does not cost energy, and, hence, a long-wavelength spin wave
has energy proportional to q 2 . Quite generally, low-lying collective modes
whose existence is guaranteed by a continuous symmetry of the problem are
referred to as “Goldstone modes”.
8.6 The transition to a magnetic insulator

In Sec. 8.2 we considered the interaction term in the Hubbard model as a
perturbation. Without interactions, the electrons are delocalized. In our
treatment, the interaction does not change that; the main effect of interac-
tions is to change the spin susceptibility, and, eventually, beyond the Stoner
28
Collective modes that do not require energy are known as Goldstone modes.
144
instability, to create a gap ∆ between the self-consistent energy levels of
electrons with spin up and spin down.
This perturbative picture should remain correct as long as the interaction
energy is small compared to the band width. In this section, the opposite case
is investigated, when the band width is small compared to the interaction
strength U. Doniach and Sondheimer call this regime the “atomic limit” of
the Hubbard model.
In the atomic limit, the coordinate representation is preferred over the
momentum representation. Fourier transforming the non-interacting part of
the Hamiltonian, Eq. (8.23) above, we find
X †
H0 = t(~ri − ~rj )ψiσ ψjσ , (8.51)
ij,σ
where ψiσ is the annihilation operator for an electron with spin σ in Wannier
wavefunction φi (~r). The interaction Hamiltonian is
X
H1 = U ni↑ ni↓ , (8.52)
i
which is the first term on the r.h.s. of Eq. (8.28) above. The indices “0” and
“1” refer to the non-interacting and interacting parts of the Hamiltonian,
respectively. As we discussed above, in the atomic limit the Hamiltonian H1
is, in fact, dominant.
The problem of studying the full Hamiltonian H0 + H1 has proven ex-
tremely difficult. Whereas an exact solution has been obtained in one di-
mension, in two and three dimensions no more than a few limiting are fully
understood. In order to make progress, we now make use of the fact that the
interaction U is much larger than the band width. Hereto, we write
H0 = H00 + H01 , (8.53)
where X X
† †
H00 = t(0)ψiσ ψiσ , H01 = t(~ri − ~rj )ψiσ ψjσ , (8.54)
i,σ i6=j,σ
and consider H01 as a perturbation. (Note that H00 cannot be considered a

perturbation, since t(0) determines the position of the band, not the band
width.)
145
Below, we are interested in the retarded single-particle Green function,
†
Gij,σ (t) = −iθ(t)h[ψiσ (t), ψjσ (0)]+ i. (8.55)
Recall that the poles of the single-particle Green function contain information
about the particle-like excitations of the system. Since the “unperturbed”
Hamiltonian H00 + H1 is not quadratic in the fermion creation/annihilation
operators, the diagrammatic perturbation theory we used in the previous
chapters cannot be employed to calculate Gij,σ . Instead, we have to calculate
Green functions using the equation of motion approach. An illustration of the
equation of motion approach for the calculation of phonon Green functions
was given in Sec. 3.3.2 above.
Let us first calculate the single-particle Green function in the absence of
the perturbation H01 . In that case, one has
∂Gij,σ (t) † †
= −iδ(t)h[ψiσ (t), ψjσ (0)]+ i − iθ(t)h[∂ψiσ (t)/∂t, ψjσ (0)]+ i. (8.56)
∂t
Using the equation of motion for the annihilation operator ψiσ (t),
∂ψiσ (t)
= −it(0)ψiσ (t) − iUni,−σ ψiσ (t), (8.57)
∂t
we find

∂
−i + t(0) Gij,σ (t) = −δ(t)δij (8.58)
∂t
†
− (−i)Uθ(t)h[ni,−σ (t)ψiσ (t), ψjσ (0)]+ i.
Repeating the same procedure for the Green function that appeared at the
r.h.s. of Eq. (8.58), we find
∂ h †
i
−iθ(t)h[ni,−σ (t)ψiσ (t), ψjσ (0)]+ i =
∂t
† †
= −iδ(t)h[ni,−σ (t)ψiσ (t), ψjσ (0)]+ i − t(0)θ(t)h[ni,−σ (t)ψiσ (t), ψjσ (0)]+ i
†
− Uθ(t)h[ni,−σ (t)2 ψiσ (t), ψjσ (0)]+ i
h i
†
= −iδ(t)δij hni,−σ i − i(t(0) + U) −iθ(t)h[ni,−σ (t)ψiσ (t), ψjσ (0)]+ i .
After Fourier transforming and substitution into Eq. (8.58), we find

(1 − hni,−σ iδij hni,−σ iδij
Gij,σ (ε) = + . (8.59)
ε − t(0) + iη ε − t(0) − U + iη
146
In a paramagnetic state, the expectation value hni,−σ i = n/2, where n is the
electron number density.
The result (8.59) has a very simple interpretation: If an electron of spin
−σ is present at a site, an added electron with spin σ has energy t(0) + U,
so that its retarded Green function is Gii,σ (ε) = 1/(ε − t(0) − U + iη). If no
electron of spin −σ is present, the Green function is Gii,σ (ε) = 1/(ε − t(0) +
iη). The probability that an electron of spin −σ is present is hni,−σ i, hence
Eq. (8.59).
Once the hopping term H01 is included, the equation of motion for the
Green function can no longer be truncated. However, following Hubbard,
one can make the following “guess” for the result. It should be pointed out,
however, that this guess has no formal justification; nevertheless, as you’ll
see below, it conveys a compelling picture of how the system behaves. The
single-electron Green function (8.59) can be interpreted as the inverse of
ε − Σi , where Σi is a measure of the energy of an electron at site i, including
the interaction modifications. Without interactions, the Green function of
the electron would be G~k,σ = 1/(ε − ε~k + iη), where
X ~
ε~k = t(0) + e−ik·(~ri −~rj ) t(~ri − ~rj ). (8.60)
i6=j
Note that we switched to the momentum representation. With interactions,

we replace ε − t(0) by (Gii,σ (ε))−1. Hence, we find
1
G~H = . (8.61)
kσ (Giiσ (ε))−1 − (ε~k − t(0)) + iη
Let us now consider the positions of the poles of the Green function G~H
kσ
.
They are at the solutions of the equation
(ε − t(0))(ε − t(0) − U) = (ε~k − t(0))(ε − t(0) − U(1 − n/2)). (8.62)
You verify that for n → 0 or U → 0 one recovers the Bloch band ε = ε~k . You
also verify that for the limit of very small bandwidth max |ε~k − t(0)| ≪ U
the solutions approach the on-site energies t(0) and t(0) + U of the Hubbard
model in the atomic limit. The effect of the inter-site hopping is to change
the N-fold degenerate solutions ε = t(0) and ε = t(0) + U into narrow bands
around t(0) and t(0) + U. For max |ε~k − t(0)| ≪ U these bands are given by
ε = t(0)(n/2) + ε~k (1 − n/2), ε = t(0)(1 − n/2) + U + ε~k (n/2). (8.63)
147
ε
t 0+U
k
t0
Figure 45: Schematic picture of the two energy bands for the Hubbard model for
the case that the interaction U is much larger than the bandwidth.
A schematic picture of these bands is shown in Fig. 45.

In general, the two bands may overlap. An exception is the case that n
is close to 1. The case n = 1 is known as “half filling”, since it amounts to
half of the maximum possible number of electrons in the system. At n = 1,
the two bands are always separated, although one may question whether this
result is realistic for small U. Since the lowest band is fully filled at n = 1,
there is an excitation gap, and the system becomes an insulator. In this case,
the excitation gap is referred to as “Coulomb gap” and the insulating state
is known as a “Mott insulator”.
8.7 Heisenberg model

The Heisenberg model gives a description for an insulating magnet. The
magnetization arises from the alignment of local magnetic moments, not
from the shift of Fermi levels for “majority” and “minority” electrons.29
The Heisenberg model is described by the Hamiltonian
X
HHeis = J~si · ~sj
ij nn
X
= J(sxi sxj + syi syj + szi szj ) (8.64)
ij nn
X 1 1 − +

+ − z z
= J si sj + si sj + si sj . (8.65)
ij nn
2 2
29
Parts of this section are based on lecture notes by W. van Saarloos, Leiden, 1996.
148
In writing (8.65), we consider a lattice whose sites are labeled with the index
i, the summation is over nearest neighbor pairs i and j on this lattice only,
and ~si denotes the spin 1/2 operator at site i. In principle, one can include
next-nearest neighbor interactions, etc., in the Hamiltonian (8.65). The spin
operators commute for different sites, and obey the usual spin commutation
relations on the same sites. Thus, in terms of the spin raising and lowering
operators,
[s− z − + z +
i , sj ] = δij si , [si , sj ] = −δij si [s+ − z
i , sj ] = 2δij si (8.66)
[sxi , syj ] = iδij szi , and cyclic. (8.67)
We will take the quantum Heisenberg model as a starting point here; the
Heisenberg model with antiferromagnetic coupling (J > 0) can be derived
from the Hubbard model at half filling, with nearest-neighbor hopping and
large positive U, see exercise 8.7.
In the literature, sometimes an anisotropic version of the Heisenberg
model (8.65) is studied,
X
anis
J⊥ (sxi sxj + syi syj ) + Jz szi szj

HHeis =
ij nn
X J⊥
= (s+ −
i sj + s− +
i sj ) + Jz szi szj . (8.68)
ij nn
2
Clearly, in this case the model is not invariant under arbitrary rotations of
the spins, but is invariant for rotations of the spins about the z-axis.30
Despite its simplicity, the Heisenberg model is difficult to study analyti-
cally. One reason that is of interest to us, is that Wick’s theorem does not
hold for the Heisenberg model. Hence, we cannot rely on the diagrammatic
perturbation theory.
An important question of interest for the Heisenberg model is when and
how a magnetic phase is formed. This question is in the realm of the theory
30 anis
In passing, we note that for J⊥ = 0, HHeis reduces to the celebrated Ising model, which
plays a central role in the theory of critical phenomena. Unlike the Heisenberg model, the
Ising model does not have a continuous symmetry, and, as a result, the Ising model has
no Goldstone modes. On a bipartite lattice, the antiferromagnetic Ising model at zero
magnetic field is equivalent to the ferromagnetic Ising model. This result does not extend
to the Heisenberg model; the ferromagnetic and antiferromagnetic Heisenberg models are
fundamentally different.
149
of phase transitions and critical phenomena, and we will not discuss it here.
Instead, we will assume that a low-temperature magnetic phase exists, and
aim at a description of the possible excitations of the system in the low-
temperature magnetic phase and the high-temperature paramagnetic phase.
We will carry out this program in detail for the ferromagnetic Heisenberg
model, for which the low-temperature magnetic phase is the fully polarized
phase. For the antiferromagnetic Heisenberg model, this question is much
more difficult, as the precise nature of the antiferromagnetic phase is not
known.
Let us now turn to the isotropic ferromagnetic (J < 0) Heisenberg model
(8.65) in absence of a field to introduce the concept of spin-waves in more
detail. The case of the ferromagnetic Heisenberg model is particularly simple
— but somewhat a-typical — in that the order parameter associated with
the broken symmetry, the total magnetization, is also a conserved quantity.
(This is sometimes referred to as an exact spontaneously broken symmetry.)
Since the total spin S~ commutes with HHeis , we can diagonalize the Hamil-
tonian in each subspace of eigenvalues of the operator S z . Let |+i denote
the state with all the spins pointing up, so that S z |+i = (N/2)|+i, where
N is the total number of spins. The state |+i is the ground state of HHeis if
J < 0, although it is not the only possible ground state. In fact, all states
obtained from |+i by multiple operation of the total spin lowering operator,
X
S− = s−
i (8.69)
i
is a ground state of the Heisenberg Hamiltonian as well. For a hypercubic

lattice in d dimensions, the ground state energy is E0 = dNJ/4.
Since the ground state of the ferromagnetic Heisenberg model is so simple,
some low-lying excitations, the spin waves — can be determined exactly. A
spin-wave state with wavevector ~q is created by the operator Sq~− operating
on the fully polarized state |+i, with
X
Sq~− = N −1/2 ei~q·~ri s−
i . (8.70)
i
You verify that a state with a single spin wave is an exact eigenstate of the
Heisenberg Hamiltonian HHeis , at energy E0 + ~ωq~, where E0 is the ground
state energy and
|J| X ~
~ωq~ = (1 − ei~q·δ ), (8.71)
2
~
δ
150
u e u e u
u: A
e u e u e
e: B
u e u e u
Figure 46: An example of a bipartite lattice. Every neighbor of the A sublattice

is a site on the B sublattice, and vice versa.
where the vector ~δ is summed over all nearest-neighbor directions in the

lattice. For small ~q, the energy ωq~ increases proportional to q 2 . Note that,
for ~q → 0, the state Sq~− |+i = |~qi approaches the uniformly rotated state
S − |+i continuously, and therefore we expect the excitation energy ~ωq~ of
this mode to vanish continuously as ~q → 0. This again illustrates that the
appearance of such “low lying modes” or “Goldstone modes” for ~q small is a
general phenomenon when a continuous symmetry is broken.
The spin-wave state |~qi is a collective mode; it describes a delocalized
state with many spin excitations involved. For small ~q, it describes a slow
rotation of the spins about the z-axis. To make this more explicity, one can
calculate the expectation value of the transverse spin correlation in the state
|~qi,
1
h~q|sxi sxj + syi syj |~qi = cos[~q · (~i − ~j)]. (8.72)
N
Spin waves do interact: the two-spin wave state Sq~−1 Sq~−2 |+i is not an exact
eigenstate of the Hamitonian (8.65). On the other hand, spin-wave interac-
tions are small if the number of excited spin wave modes is small. Then,
the spin waves can be considered independent boson modes, their distribu-
tion being given by the Bose-Einstein distribution function. The picture of
independent spin wave modes leads to the Bloch T 3/2 law for the tempera-
ture dependence of the low-temperature magnetization in three dimensions,
hSz (T )i − hSz (T = 0)i ∝ T 3/2 .31 For one and two dimensions, the density of
states of low-energy spin waves is higher than in three dimensions; there is
no regime of a small number of excited spin waves, leading to a breakdown
of the ferromagnetic state in one and two dimensions.
31
See, e.g., Solid State Physics, by N. W. Ashcroft and N. D. Mermin, Saunders (1976),
and the discussion below.
151
Within the Green function approach, these conclusions can be arrived at
if we look at the retarded transverse spin-spin correlation function,
R
Rij (t) = −iθ(t)h[s− +
i (t), sj (0)]− i, (8.73)
which is the Heisenberg-model equivalent of the transverse spin susceptibility

χ−+ we studied previously. (Note that the correlation function contains a
commutator of spin operators, since spin operators satisfy commutation re-
R
lations.) The function Rij (t) describes the response of the spin polarization
to a magnetic field, where both the response and the field are perpendicular
to the z axis, which is presumed to be the direction of spontaneous magne-
tization. The poles of the correlation function corresponds to the collective
modes. Further, the spin-spin correlation function is related to the average
magnetization through the relation
R
lim Rij (t) = 2ihsz iδij . (8.74)
t↓0
In order to calculate the transverse spin-spin correlation function, we

consider its equation of motion
∂ R X
Rij (t) = 2iδ(t)hszi i − θ(t) hJli[s− z z − +
l (t)si (t) − sl (t)si (t), sj (0)]− i, (8.75)
∂t l
where we defined Jli = J if l and i are nearest neighbor sites and Jli = 0
otherwise. In order to close this equation of motion, we replace the operator
szi (t) by the average hsz i ≡ hszi i in the ferromagnetic state,32
∂ R X
Rij (t) = 2iδ(t)hszi i − ihsz i Jli (RljR (t) − Rij
R
(t)). (8.76)
∂t l
The justification for this approximation is that, in the ferromagnetic state,

fluctuations of szi around its average are small. Moreover, since Eq. (8.76)
contains a sum over nearest neighbors, only the sum of szi over nearest neigh-
bors is important. In high dimensions, the summation over nearest neighbors
leads to an even further reduction of fluctuations, and one expects the ap-
proximation szi ≈ hsz i to continue to hold outside the almost fully polarized
32
This approximation is the equivalent of the Weiss molecular field approximation, see,
e.g, Statistical and Thermal Physics, F. Reif, Mc Graw Hill, 1965.
152
ferromagnetic state. Solving Eq. (8.76) by means of a Fourier transform, we
find
2hsz i
R~k (ω) = − , (8.77)
ω + ω~k + iη
where
~~
X
ω~k = hsz iJ (1 − e−ik·δ ), (8.78)
~
δ
the summation being over all nearest-neighbor vectors ~δ. Note that this
solution satisfies the relation (8.74).
For zero temperature, hsz i = s, and we recover the spin-wave dispersion
relation (8.71). For finite temperatures, the spin-wave frequencies depend
on the average spin polarization hsz i, which, in turn, is a function of the
temperature T . In order to determine hsz i, we use the “greater” correlation
function,
>
Rij (t) = −ihs− +
i (t)sj (0)i, (8.79)
for which one has
Rii> (0) = −is(s + 1) + ihsz (sz + 1)i. (8.80)
In keeping with the approximations we made previously, we neglect fluctua-
tions of sz and set
Rii> (0) ≈ −is(s + 1) + ihsz i(hsz i + 1). (8.81)
>
We calculate Rij (t) from the retarded correlation function (8.77) using the
general relation (2.46),
Im R~kR (ω)
R~k> (ω) = 2i
1 − e−~ω/T
hsz iδ(ω + ω~k )
= 4πi . (8.82)
1 − e−~ωT
Performing the inverse Fourier transform, we then find
2i X 1
Rii> (0) = − ~ω /T
, (8.83)
N e k −1
~
~ k
from which we derive the self-consistency equation

2 X 1
s(s + 1) − hsz i(hsz i + 1) = hsz i ~ω /T
. (8.84)
N e ~k − 1
~ k
153
In the limit T → 0 the r.h.s. vanishes in three dimensions and above, recov-
ering the solution hsz i = s at T = 0. For one and two spatial dimensions,
the r.h.s. does not vanish as T → 0, implying the breakdown of the ferro-
magnetic state by spin wave generation, as we discussed previously. For low
temperatures and three dimensions, one find hsz (T = 0)i − hsz (T )i ∝ T 3/2 ,
which is the Bloch T 3/2 law mentioned previously.
The self-consistent relation could be used to find the Curie tempera-
ture, the highest temperature at which the spontaneous magnetization exists.
However, close to the Curie temperature, spin fluctuations are abundant, and
the approximation szi → hsz i we made in the derivation of Eq. (8.84) is not
justified in systems with low dimensionality.
We now turn to the antiferromagnetic Heisenberg model J > 0. Clas-
sically, an antiferromagnetic state is possible only if the lattice is bipartite:
it can be separated in sublattices A and B such that all nearest neighbors
of a site in lattice A belong to lattice B and vice versa, see Fig. 46. If the
lattice is not bipartite, the antiferromagnet is “frustrated”, and a completely
different physical picture arises, which we do not discuss here. The classical
antiferromagnetic ground state is a state in which spins on one sublattice
point “up”, whereas spins on the other sublattice point “down”. This state
is known as the “Néel state”. Intuitively, we expect that the Néel state is
the ground state of the quantum antiferromagnetic Heisenberg model as well.
However, the Néel state is not an exact eigenstate of HHeis !33 Nevertheless
the true ground state has strong antiferromagnetic correlations reminiscent
of this state.
Even though the classical Néel state is not the exact ground state, it does
suggest a definition of the antiferromagnetic “order parameter”, which takes
the role of the magnetization for the ferromagnetic Heisenberg model. Let
us consider a hypercubic lattice with lattice spacing a in d dimensions and
let a reference spin on sublattice A point up; then the so called “staggered
magnetization” is defined as
z
AF hsi i if ~ri in sublattice A,
Mi = z . (8.85)
−hsi i if ~ri in sublattice B
In the classical Néel state, this order parameter is 1/2 for all sites.
This choice of the order parameter may be justified for a Heisenberg model
with spin s ≫ 1/2. The larger the spin s is, the more “classical” the behavior
33 anis
The classical Néel state is the ground state of HHeis with Jz > 0 and J⊥ = 0, but not
of the model with J⊥ 6= 0.
154
of the model becomes, as the commutation relations (and hence fluctuations)
become less important. In the classical limit, we can represent the spins by
vectors on a sphere of radius s. In this s → ∞ classical limit, the ground
state of the antiferromagnetic Heisenberg model consists of vectors pointing
in opposite directions on the two sublattices. In this limit, (8.85) is a good
order parameter, whose value approaches s for s large.
In recent years, there has been a lot of research on the properties of quan-
tum antiferromagnets, in particular in d = 2. The upsurge of interest in this
problem was stimulated to a large extent by the discovery of high temper-
ature superconductors, which are obtained by doping compounds which are
insulating quantum antiferromagnets. The consensus, basis on many large
scale numerical simulations, is that the ground state of the two dimensional
Heisenberg model does have antiferromagnetic order, but that the order pa-
rameter (8.85) is reduced from the classical Néel value 1/2 to about 0.34.
In the ferromagnetic case, the order parameter was a conserved quantity
and as a result a single spin wave state |~qi was an exact eigenstate of the
Hamiltonian. In the present case, the staggered order parameter (8.85) is
not a conserved quantity, and the spin wave analysis can not be performed
exactly (after all, the ground state is not known exactly either!), so we have
to resort to an approximate treatment. In order to understand the nature
of the approximation better, it will turn out to be useful to look at the
Heisenberg model for the case of large spin s. Since we expect the ground
state to get closer to the classical Néel state for large s, increasing s gives us
a way to get a controlled approximation.
As in the case of a ferromagnet, to analyze spin waves we need to consider
states where one spin is flipped. Since the ground state is not explicitly
known, it is more convenient to use (in the Heisenberg picture) the equation
of motion for the spin flip operator. Using the commutation relations (8.66)
(which are also valid for arbitrary spin s), we get
∂s~−
i 1 − J X − z − z

= [s~i , HHeis ] = s~i s~i+~δ − s~i+~δ s~i . (8.86)
∂t i~ i~
~
δ
As before, the sum over ~δ is a sum over the nearest neighbor vectors on the
quadratic (d = 2) or cubic (d = 3) lattice.
For large s, spin waves consist of minor distortions of the Néel state,
with the z-component of the spin slightly reduced from its maximal value
+s on the A sublattice, and slightly increased from its minimal value −s
155
on the B sublattice. To a good approximation, we can then write in (8.86)
s~zi |ground statei ≈ ±s|ground statei depending on whether ~ri is on the A or
B sublattice. Using the notation
|~iiA = s~−
i
|ground statei for ~ri in sublattice A, (8.87)
|~iiB = s~+
i
|ground statei for ~ri in sublattice B, (8.88)
we then get from (8.86)

 
∂ ~ J 
|~i + ~δiB  ,
X
|iiA = −2ds|~iiA − s
∂t i~
~
δ
 
∂ ~ J 
|~i + ~δiA  .
X
|iiB = 2ds|~iiB + s (8.89)
∂t i~
~
δ
Since the two sublattices are distinct, we now have two coupled equations,
instead of one in the case of a ferromagnet. The equations for the spin waves
can be solved by introducing Fourier transforms,
−1/2
X ~ −1/2
X ~
|~qiA = NA ei~q·i|~iiA , |~qiB = NB ei~q·i|~iiB . (8.90)
~i∈A ~i∈B
Writing the temporal evolution of these modes as e−iω~qt , we then get from
(8.89) the dispersion relation

~ωq~ + 2dsJ 2dsJγq~ 1 X i~q·~δ

−2dsJγq~ ~ωq~ − 2dsJ = 0, γ q~ = e . (8.91)
2d
~
δ
Solution of Eq. (8.91) yields

q
~ωq~ = 2dsJ 1 − γq~2 . (8.92)
By expanding γq~ for small q, we find that long wavelength antiferromagnetic

spin waves have a linear dispersion
√
ωq ≈ 2sJq d, q ≪ 1. (8.93)
This linear dispersion is reminiscent of that of acoustic phonons in a crystal.

The difference between the quadratic dispersion of ferromagnetic spin waves
156
and the linear dispersion of the antiferromagnetic spin waves — the Gold-
stone modes of the antiferromagnet — can be traced to the fact that in the
first case, the order parameter itself is conserved. This is rather exceptional,
and indeed Goldstone modes usually have a linear dispersion.
Our earlier discussion implies that (8.93) is a good approximation
√ for
large s. Although the estimate of the spin wave velocity c = sJ 2d/~ may
not be that accurate for spin 1/2, the linearity of the antiferromagnetic spin
wave dispersion is a robust feature that is not affected by the approximations
made.
In closing this section, we note that the present analysis may be extended
to a systematic 1/s expansion for quantities like the spin wave velocity and
the order parameter. Quantum mechanical fluctuations are more and more
suppressed as s increases. Intuitively, this becomes clear from the fact that
for s = 1/2, the operators s+ −
i or si lead to a local reversion of the spin, and
hence tends to locally make the order parameter of the wrong sign. For large
s, these operators just reduce the z-component of the spin slightly — they
do not lead to a complete spin reversal.
157
8.8 Exercises
Exercise 8.1: Spin susceptibility
The Thomas-Fermi approximation (8.22) for the spin susceptibility in an

interacting electron gas is not correct. It neglects the structure of the inter-
action on the scale of the Fermi wavelength. A better approximation is made
when the full Thomas-Fermi screened interaction is used,
e2
Vq~RPA (iωn ) = if ~q, ωn → 0,
ǫ0 (q 2 + ks2 )
see Eq. (7.27). Hence, the momentum-dependence of the screened interaction

is taken into account, but retardation effects are neglected.
(a) Argue that, in the limit ~q → 0, the renormalized vertex Γ(~k, iωm ; ~k +
~q, iωm + iΩn ) of Eq. (8.15) does not depend on the direction of the
wavevector ~k or on the energy ωm .
(b) Since Γ(~k, iωm ; ~k + ~q, iωm + iΩn ) of Eq. (8.15) does not depend on
the direction of the wavevector ~k, the effective interaction V~k−vk
eff
′
in
~
the vertex equation (8.15) may be integrated over the directions of k ′ .
Show that the integrated interaction is
e2 ks2 + (k + k ′ )2
V~k,eff~k′ = ln . (8.94)
4ǫ0 kk ′ ks2 + (k − k ′ )2
(c) Argue that the renormalized vertex is determined by the effective in-
teraction at the fermi level. From here, derive the result
νµB g
χRPA
−+ = . (8.95)
2 − (2/3π 2 )2/3 rs ln(1 + (3π 2 /2)2/3 /rs )
(d) Can you explain this result qualitatively in the limits rs ≪ 1− and
rs ≫ 1?
158
Exercise 8.2: Specific heat
Verify Eq. (8.35) for the low-temperature specific heat of the Hubbard model
close to the Stoner instability. For your verification is is sufficient if you find
the order of magnitude of the proportionality constant in Eq. (8.35). You do
not need to find precise numerical coefficients.
Exercise 8.3: Ferromagnet with parabolic dispersion relation
In this exercise we consider the Hubbard model, with a parabolic dispersion

of the band energies ε~k . This model is somewhat unrealistic, but it allows us
to calculate many quantities in closed form.
(a) Using the Hartree-Fock energy levels, show that the spin polarization
n̄↑ − n̄↓
ζ= (8.96)
n̄↑ + n̄↓
obeys the equation

n 1/3 2mζ
U = (1 + ζ)2/3 − (1 − ζ)2/3 . (8.97)
9π 4 ~2
Here n = n̄↑ + n̄↓ is the total electron density. Find a relation between
ζ and the energy gap ∆.
(b) Analyze Eq. (8.96) and find for what values of the Hubbard interaction
strength U a spontaneous spin polarization is formed.
(c) If the Hubbard interaction strength U is further increased, a full polar-

ization is achieved, ζ = ±1. This situation is known as a half metal.
(d) Calculate the transverse spin susceptibility χ−+ for low frequencies and
long wavelengths. What is the spin-wave dispersion that you find from
your answer?
159
Σ P= + + + ...
Figure 47: Singular contribution to the electron self energy at the Stoner insta-
bility.
Exercise 8.4: Self-energy correction
Close to the Stoner instability point there are important corrections to the
self energy that are not included in the Hartree-Fock approximation. One of
the leading corrections is shown in Fig. 47.
(a) Calculate the contribution to the self energy shown in Fig. 47 for the
Hubbard model.
(b) Discuss how the self-energy correction you calculated under (a) affects
the electron effective mass m∗ and the density of states at the Fermi
level.
Exercise 8.5: Coulomb gap
In this exercise, we consider the ansatz (8.59) for the single particle Green
function of the Hubbard model. The Green function not only contains infor-
mation about the possible single-particle excitations through its poles, but
also through the spectral weight of each pole.
(a) Find the spectral weights for the case of large U, U ≫ max |ε~k − t(0)|,
for which you can use Eqs. (8.63) for the poles of the Green function.
(b) Do the same for the case of small U, U ≪ max |ε~k − t(0)|. Formulate
your answer for the case n ≪ 1 and n = 1 separately.
160
Exercise 8.6: Bloch T 3/2 law
Derive the Bloch T 3/2 law for the magnetization of a three-dimensional fer-
romagnet from Eq. (8.84).
Exercise 8.7: Antiferromagnetic Heisenberg model from large-U Hubbard

model at half filling
In this exercise we consider the Hubbard model with nearest neighbor hop-
ping, X † X
†
HHub = −t ψiσ ψjσ + ψjσ ψiσ + U ni↑ ni↓ . (8.98)
ij nn,σ i
(a) When U = 0, the Hubbard model describes non-interacting electrons,

and we can explicitly determine the single electron energies ε~k . Show
that these are given by Eq. (8.37) above.
Because of the Pauli principle, there can not be more than two electrons per
site. Let us from now on consider half filling, i.e. the case where there is
one electron per site. Moreover, we consider the limit of large positive U
(U ≫ t). Then it is energetically very unfavorable to have two electrons on
the same site — in other words the low energy sector of the Hilbert space is
the one where all sites are occupied by one electron. Our strategy now will
be to project onto this part of Hilbert space of singly occupied sites.
What will the interactions look like? Well, the operator ni↑ ni↑ only counts
whether a site is doubly occupied, therefore it does not depend on the abso-
lute spin direction. Therefore, the effective spin interaction will only depend
on the relative spin orientation, in other words, it will be of the Heisenberg
type ~si · ~sj .
Furthermore the interaction will be antiferromagnetic. Physically, this
can be seen as follows. Virtual excitations to states in which one site is
doubly occupied involve the hop of an electron to a neighboring site and
back (see the figure); in the intermediate state, there is an extra energy U
associated with the double occupancy.
161
6
s s 6
s s 6
s s s6 s6
? ? ?
(a) (b) (c)

initial configuration virtual state final state hopping blocked
→ hop t → hop t
(b) Argue that, with these considerations, the large-U Hubbard model at
half filling reduces to an antiferromagnetic Heisenberg model with in-
teraction J = 2t2 /U.
The range over which the antiferromagnetic phase exists as a function

of the ratio U/t and doping of the Hubbard model, is still under active
investigation.
162
9 Superconductivity
Just like the presence of the electron gas modifies the electron-electron inter-
action, and leads to an effective, or “screened” interaction, the presence of
the ionic lattice and the interaction between electrons and phonons causes a
modification of the electron-electron interaction. This phonon-modified in-
teraction is attractive, and is the root cause for superconductivity. In this
chapter we will briefly review the origin of the attractive interaction, the
instability that leads to the superconducting state, and the fundaments of a
Green function description of superconductors.
9.1 Phonon-mediated electron-electron interaction

For a discussion of the phonon-mediated electron-electron interaction we first
need to describe the interaction between electrons and phonons.
Starting point is the potential the electrons feel from the lattice ions,
Z N
X
Hlattice = d~rn(~r) Vion (~r − ~xj ), (9.1)
j=1
where the vector ~xj runs over the positions of all ions in the lattice and n
is the electron number density. Expanding Hlattice around the equilibrium
positions ~xj = ~rj , j = 1, . . . , N, we have
Z N Z N
~ ~r Vion (~r − ~rj ). (9.2)
X X
Hlattice = d~rn(~r) Vion (~r − ~rj ) − d~rn(~r) ~uj · ∇
j=1 j=1
Here ~uj = ~xj − ~rj is the displacement of the jth lattice ion. The first term
represents a static potential and gives rise to the electronic band structure.
The second term depends on the actual ionic positions. It is the second term
that leads to the electron-phonon interaction.
Fourier transforming the second term of Eq. (9.2), we find
s !
i XX N~ λ †
Hep = − ~v · ~qVq~ie ψ~k+~ ψ~ (bλ + bλ†
q ), (9.3)
V 2Mωq~λ q~ q ,σ k,σ q~ −~
~kσ q~λ
where ψ~kσ is the annihilation operator for an electron with wavevector ~k

and spin σ, bqλ~ the annihilation operator for a phonon with momentum ~q
163
and polarization ~vq~λ . In the derivation of Eq. (9.3) we used Eq. (3.21) to
express the displacement ~uj in terms of the phonon creation and annihilation
operators bqλ† λ
~ and bq~ .
At low temperatures, the phase space available for “umklapp processes”,
corresponding to a momentum transfer ~q outside the first Brillouin zone, is
small. Moreover, the weight of umklapp processes is suppressed because of
the q-dependence of the potential Vq~ie for large q. Hence, we will restrict the
summation over ~q to the first Brillouin zone.
In an isotropic material, the polarization vector ~vq~λ is either perpendicular
or parallel to ~q. The corresponding phonon states are labeled “transverse”
and “longitudinal”, respectively. From Eq. (9.3), we conclude that only lon-
gitudinal phonons interact with electrons. If we want to simplify Eq. (9.3) to
the case that we keep longitudinal phonons only, we must remind ourselves
that in our original formulation of the phonon modes, we have chosen the
same polarization vectors for opposite wavevectors, ~vq~λ = ~v−~ λ
q . Hence, if the
longitudinal polarization ~vq~ points along ~q, ~v−~q points opposite to −~q. We re-
solve this problem by defining an arbitrary function sign (~q), which can take
the values 1 or −1 and which is such that sign (~q) = −sign (−~q). Keeping
longitudinal phonons only, the expression for the electron-phonon interaction
then simplifies to
s !
i XX N~ †
Hep = − Vq~ie q sign (~q) ψ~k+~ ψ~k,σ (bq~ + b†−~q). (9.4)
V 2Mωq~ q ,σ
~kσ q~
where the creation and annihilation operators b† and b refer to longitudinal

phonons only.
The calculation of Green functions for a system of electrons and phonons
proceeds in a way that is very similar to what we have seen in the previous
chapters. Since our discussion of phonon Green functions there focused on
the Green functions for the displacements rather than the phonon creation
and annihilation operators, we rewrite the electron-phonon Hamiltonian in
the form
1 XX †
Hep = gq~ψ~k+~ ψ~ u ,
q,σ k,σ q~
(9.5)
V
~kσ q~
where uq~ is a longitudinal displacement,

s
~q ~
uq~ = (bq~ + bq†~), (9.6)
q 2Mωq~
164
and gq~ describes the strength of the electron-phonon coupling,
gq~ = −iN 1/2 qVq~ie sign (~q). (9.7)
The properties of lattice vibrations were discussed in chapter 3 we dis-

cussed lattice dynamics. Lattice ions perform vibrations around their equi-
librium positions, the frequency of the vibrations being determined by the
second derivative of the interaction between ions. However, this interaction
is an “effective interaction”, mediated by the electrons. Without electrons,
the potential between lattice ions would be the Coulomb potential. As we
have seen in the electronic case, the long-range Coulomb potential does not
allow for phonon-like low-frequency collective modes. Instead, the collective
modes allowed by the Coulomb interaction are plasma modes, which have a
finite frequency even for zero wavevector.
A description of the interaction between electrons and the lattice and the
lattice dynamics itself is not straightforward. After all, in such a theory the
electrons interact with lattice vibrations, that themselves exist by virtue of
the existence of the electrons. A theory of phonon-mediated electron-electron
interactions in which the phonon dispersion is, in turn, governed by the
electron-mediated interaction between the lattice ions is a risky enterprise.
In order to circumvent such a “chicken and egg” problem, it is preferable
to start from first principles, and use a description in which the only interac-
tion is the Coulomb interaction, just like we did for the case of the electron
gas. As a complete “ab-initio” treatment is impossible here — for that we’d
have to find the lattice structure from first principles —, we use a simplified
model for the ionic lattice. That simplified model is the “jellium model” of
Ex. 3.1. In the jellium model, the lattice is replaced by a positively charged
ion “fluid” (or “jellium”). In such a model, the only difference between the
ions and the electrons is the different ratios of mass density to charge density:
the ion jellium has the same average charge density as the electron gas, but
its mass density is a factor ∼ 105 higher.
In our theory we’ll consider the electron-phonon interaction as a pertur-
bation. Without electron-phonon interaction, the phonons are the quantized
oscillations of the ion jellium without electron screening. Such oscillations,
which are the equivalent of plasma oscillations in the interacting electron gas,
have frequency ΩE
Z 2 e2 N Ze2 n
Ω2E = = , (9.8)
ǫ0 MV ǫ0 M
165
k q
k+q
Figure 48: Diagrammatic representation of phonon line and of electron-phonon
interaction vertex.
where n is the electron density. The frequency does not depend on the
wavelength of the phonons, as in the Einstein model of lattice vibrations.
Also, in the jellium one has longitudinal phonons only; transverse phonons,
which do not correspond to a change in ion density, can not be described in
a jellium model. Thus, the phonon Hamiltonian in the jellium model is
X
Hphonon = ~ΩE (bq†~bq~ + 1/2). (9.9)
q~
The relevant phonon Green function is defined with respect to the displace-
ment operators uq~. It was calculated in Chapter 3,
1
Dq~(τ ) = −hTτ uq~(τ )u−~q(0)i, Dq~0 (iΩn ) = − . (9.10)
M(Ω2n + Ω2E )
The superscript “0” indicates that this phonon Green function is calculated
without taking the electron gas into account.
In this first-principle jellium model, the interaction potential between
electrons and ions is the Coulomb interaction. We thus set Vq~ie = −Ze2 /ǫ0 q 2
in Eq. (9.7), s
Ze2 i 1/2 ΩE Me2
gq~0 = N =i . (9.11)
ǫ0 q ǫ0 q 2
Diagrammatic rules for a combined electron-phonon system are quite sim-
ilar to those of the previous chapters for electronic systems. We denote a
correlation function of displacements with a “springy” line, see Fig. 48. The
electron-phonon interaction is a vertex where a phonon line meets an incom-
ing and outgoing electron line. Energy and momentum are conserved at the
electron-phonon vertex. Each phonon line is represented by minus a phonon
Green function, −D 0 ; each electron-phonon vertex has weight gq~0 . Note that
0 0∗
g−~ q = gq~ .
166
111
000
000
111
000
111
000
111
000
111
= + 000
111
000
111
000
111
000
111
000
111
000
111
000
111
−D = −D 0 (−D0 ) (− χ e) (−D)
Figure 49: Diagrams for the effective phonon Green function D.
The first effect of the electron-phonon interaction is to modify the phonon

Green function. Diagrams for the phonon Green function in the random
phase approximation are shown in Fig. 49. In this diagram you recognize the
electron polarizability function χ, which we calculate in the RPA approxi-
mation as well, see Eq. (7.33). We find
Dq~0 (iΩn )
Dq~RPA (iΩn ) =
1 − χqRPA
~ (iΩn )Dq~0 (iΩn )|gq~0|2 /V e2
1
= − , (9.12)
M(Ωn + ωq~2 )
2
where
Ω2E Ze2 n
ωq~2 = = . (9.13)
ǫqRPA
~ (iΩn ) ǫRPA ǫ0 M
Here we wrote the polarizability function in terms of the dielectric response
function ǫRPA = (1 + Vq~χRPA /e2 )−1 , where Vq~ = e2 /ǫ0 q 2 denotes the electron-
electron interaction. Substituting the Thomas-Fermi approximation (7.38)
of the dielectric response function, we find, for small q,
r
Zm
ωq~ = vF q , (9.14)
3M
which is the Bohm-Staver expression for the phonon dispersion, see Ex. 3.1.
We thus arrive at the satisfactory conclusion that, once the electron-phonon
interaction is taken into account, the phonon frequencies are renormalized
from Ω to ωq~, where ωq~ is proportional to q for large wavelengths. You had
reached the same conclusion in Ex. 3.1, based on macroscopic arguments.
Next, let us consider the electron-phonon interaction. This interaction is
also modified by the presence of the electron gas. Instead of exciting a phonon
167
1111
0000
0000
1111
0000
1111
0000
1111
0000
1111
= + 0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
g = g0 + V (−χ e) g0
Figure 50: Diagrams for the effective electron-phonon interaction gq~.
directly, as is described by the electron-phonon Hamiltonian (9.5), an electron

may excite an electron-hole pair, which, in turn excites a phonon. Again
using the random phase approximation, such processes can be described by
the diagrams listed in Fig. 50. We thus find
gq~RPA (iΩn ) = gq~0 (1 + Vq~χqRPA

~ /e2 )
gq~0
= . (9.15)
ǫRPA (~q, iΩn )
Let us now turn to the electron-electron interactions. A diagrammatic

expression for the effective electron-electron interaction is shown in Fig. 51.
The left diagram in Fig. 51 corresponds to interactions that do not involve the
excitations of phonons. This is the RPA effective interaction we studied in
Sec. 7.4. The right diagram contains all processes where a phonon is excited.
Notice that the electron-phonon vertices are renormalized vertices of Eq.
(9.15) and that the phonon propagator is the full renormalized propagator of
Eq. (9.12). Combining everything, we find that the effective electron-electron
interaction is
Vq~
Vq~eff (iΩn ) = RPA
+ gq~RPA (iΩn )g−~
RPA RPA
q (iΩn )Dq~ (iΩn )
ǫq~ (iΩn )
!
Vq~ ωq~2
= RPA 1− 2 , (9.16)
ǫq~ (iΩn ) Ωn + ωq~2
where we have made use of the identity
gq~0g−~
0 2 RPA
q = Mωq~ εq~ (iΩn )Vq~. (9.17)
168
= +
V eff = V RPA + g D g
Figure 51: Diagrams for the effective electron-electron interaction Vq~eff .
Transforming to real frequencies, we finally obtain the phonon-mediated

effective electron-electron interaction
Vq~ ω2
Vq~eff (ω) = . (9.18)
ǫqRPA
~ (ω) ω 2 − ωq~2
Using the Thomas-Fermi approximation for the dielectric response function,
this simplifies to
e2 ω 2
Vq~eff, TF
(ω) = . (9.19)
ǫ0 (q 2 + ks2 )(ω 2 − ωq~2 )
In contrast to the RPA screened interaction, which hardly depended on fre-
quency as long as ω ≪ ǫF , the inclusion of electron-phonon interactions gives
rise to a frequency dependence on a much smaller energy scale: V eff depends
on frequency on the scale ωq~, as sketched in Fig. 52. For large frequencies
ωq~ ≪ ω the effect of the lattice is very small; V eff is very close to V RPA . The
effect of the lattice is most dramatic, however, for small frequencies, ω ≪ ωq~,
where the effective electron-electron interaction becomes attractive, rather
than repulsive.
What is the effect of the attractive interaction? At first sight, one is in-
clined to believe that the effect is small. After all, the interaction is attractive
only in a very small frequency range, ω . ωD , where the Debije frequency
ωD is the typical phonon frequency. Nevertheless, at low temperatures only
electrons close to the Fermi level play a role. For these electrons interactions
with small transferred energy are most important. It is precisely for those
processes that the interaction is attractive.
9.2 Cooper instability

What is the effect of an attractive interaction? In order to investigate the
effect of a weak attractive interaction, we consider a cartoon version of the
169
Vqeff(ω)
VqRPA
ωq ω
Figure 52: Sketch of the frequency dependence of the effective electron-electron

interaction.
effective interaction we derived in the previous section.

The cartoon version, used since the original formulation of the theory of
superconductivity by Bardeen, Cooper, and Schrieffer, reflects the fact that
the attractive interaction exists for small energy transfers only, and combines
it with our knowledge that only electrons near the Fermi surface play a role.
With this motivation, we look at the interaction Hamiltonian
λ X X
H1 = − ψ~k† ψ~k† ψ~k3 σ2 ψ~k4 σ1
2V 1 σ1 2 σ2
~k1 ,~k2 ,~k3 ,~k4 σ1 ,σ2
× δ~k1 +~k2 ,~k3 +~k4 θ(~k1 )θ(~k2 )θ(~k3 )θ(~k4 ), (9.20)
where the function θ(~k) is nonzero only for |ε~k − µ| < ωD and λ is a posi-
tive constant. Note that the interaction (9.20) is not the result of a formal
manipulation starting from the effective phonon-mediated electron-electron
interaction. Instead, it is merely a cartoon that is used because of its techni-
cal simplicity. In some applications, it is favorably to rewrite the interaction
(9.20) in coordinate space. The interaction is local on lenght scales large
compared to vF /ωD ,
λX
Z
H1 = − d~rψσ† 1 (~r)ψσ† 2 (~r)ψσ2 (~r)ψσ1 (~r). (9.21)
2 σ ,σ
1 2
If the coordinate representation (9.21) is used, care must be taken that it

is, in fact, the Fourier transformation of Eq. (9.20), and, thus, requires a
170
ωm ωm ωm ωp ωm −ω p+Ω n
k k’ k k’ k k’ k k’
R= − + −
−k+q −k’+q −k+q −k’+q −k+q −k’+q −k+q −k’+q
−ωm+Ω n −ωm+Ω n −ωm+Ω n −ω p+Ω n −ωm+Ω n ωp
ωm ωp ωm ωp
k k’’ k’ k k’’ k’
+ − + ...
−k+q −k’’+q −k’+q −k+q −k’’+q −k’+q
−ωm+Ω n −ω p+Ω n −ωm+Ω n −ω p+Ω n
Figure 53: Ladder diagrams contributing to the Cooper correlator R of Eq. (9.22).
momentum cut-off at momenta of order ωD /vF .34

For reasons that will become clear soon, we look at the correlator
σ1 σ2 ,σ3 σ4 †
R~k,~k ′ ,~
q
(τ ) = −hTτ ψ−~k+~q,σ1 (τ )ψ~k,σ2 (τ )ψ− ~k ′ +~
q ,σ
(0)ψ~k† ′ ,σ (0)i. (9.22)
3 4
We calculate R using diagrammatic perturbation theory, taking the effective

electron-electron interaction we calculated in the previous section. The di-
agrams contribution to R are very similar to the diagrams that contributed
to the spin susceptibility in chapter 8. They are shown in Fig. 53. Note that
there are contributions from the direct interaction and from the exchange
interaction, which come with an extra minus sign. Summing the geometric
series in Fig. 53, we find
σ1 σ2 ,σ3 σ4
R~k,~k ′ ,~
q
(iΩn ) = (δ~k,~k′ δσ1 σ3 δσ2 σ4 − δ~k,~q−~k′ δσ1 σ4 δσ2 σ3 )
!
X
× T G−~k+~q(iωm + iΩn )G~k (−iωm )
m
1
+ λ̃(δσ1 σ3 δσ2 σ4 − δσ1 σ4 δσ2 σ3 )
V
34
The exact Fourier transform of Eq. (9.20) has the form
λ X
Z Z
H1 = d~r d~r1 d~r2 d~r3 d~r4 θ(~r − ~r1 )θ(~r − ~r2 )θ(~r − ~r3 )θ(~r − ~r4 )
2 σ ,σ
1 2
× ψσ† 1 (~r1 )ψσ† 2 (~r2 )ψσ2 (~r3 )ψσ1 (~r4 ),
where
1 X i~k·~r ~
θ(~r) = e θ(k).
V
~
k
171
!
X
× T G−~k+~q(−iωm + iΩn )G~k (iωm )
m
!
X
× T G−~k′ +~q(−iωp + iΩn )G~k′ (iωp ) . (9.23)
p
Here λ̃ is a renormalized interaction,

 −1
T X
′
X
λ̃ = λ 1 + λ G~k′′ ,~k′′ (iωm )G−~k′′ +~q,−~k′′+~q(−iωm + iΩn ) , (9.24)
V m
′′~k
where the symbol ~′k denotes ~k θ(~k)θ(−~k + ~q). Performing the Matsubara
P P
summations, we have
X
T G−~k+~q(−iωm + iΩn )G~k (iωm )
m
X 1
= T
m
[−iωm + iΩn − (ε−~k+~q − µ][iωm − (ε~k − µ)]
1 tanh[(ε−~k+~q − µ)/2T ] + tanh[(ε~k − µ)/2T ]
= − . (9.25)
2 ε~k + ε−~k+~q − 2µ − iΩn
We perform the analytical continuation iΩn → ω + iη and consider the

limit ~q → 0, ω → 0. Substituting Eq. (9.25) into the expression (9.24) for
the effective interaction strength, replacing the momentum summation by an
integration over energy, and integrating by parts, we find
Z ~ωD
1 X ′ tanh[(ε~k − µ)/2T ] tanh(ξ/2T )
= ν dξ (9.26)
2V ε~k − µ −~ωD 2ξ
~k
Z ~ωD /2T
~ωD ~ωD ln x
= ν tanh ln − dx .
2T 2T 0 cosh2 (x)
The second term is convergent and approaches the value −0.81 as T →

0, whereas the first term diverges logarithmically. We thus see that the
denominator in Eq. (9.24) diverges if
1 = λν(ln(~ωD /2T ) + 0.81), or T = Tc = 1.14~ωD e−1/λν . (9.27)
172
It is important to realize that this divergence happens for an arbitrarily weak
attractive interaction, provided the temperature is sufficiently low.
Of course, a divergence of λ̃ corresponds to a divergence of the correlation
function R. The reason why we studied the correlation function R was pre-
cisely because of this divergence: a divergent correlation function is a signal
of an instability. To explain this in more detail, let us describe the system
using the grand canonical ensemble and calculate the response to a fictitious
perturbation of the Hamiltonian,
Z
H1 = d~r(∆(~r)ψ↑† (~r)ψ↓† (~r) + ∆∗ (~r)ψ↓ (~r)ψ↑ (~r)). (9.28)
It is important that we use a grand canonical formulation, since the pertur-

bation H1 does not conserve particle number. The response we look for is of
the quantity
F ∗ (~r) = hψ↑† (~r)ψ↓† (~r)i. (9.29)
The response of F ∗ to the perturbing field ∆(~r) is easily found using the
Kubo formula,
Z Z
∗ ′
F (~r, t) = dt d~r′RR (~r − ~r′ ; t − t′ )∆∗ (~r′ , t′ ), (9.30)
where
RR (~r − ~r′ , t − t′ ) = −iθ(t − t′ )h[ψ↑† (~r, t)ψ↓† (~r, t), ψ↓ (~r′, t′ )ψ↑ (~r′ , t′ )]− . (9.31)
Here we omitted a contribution to the response that involves a commutator of

four creation operators, which is nonzero in the non-interacting ground state.
Fourier transforming and switching from the retarded correlation function to
the temperature correlation function, we find
1 X
Rq~(τ ) = − hTτ ψ~k+~q↓ (τ )ψ−~k↑ (τ )ψ~k′ +~q↑ (0)ψ−~k′ (0)i
V
~k,~k ′
X ↓↑↑↓
= R~k~k′q~ (τ ), (9.32)
~k,~k ′
where R~k↓↑↑↓
~k ′ q~ is the correlator we defined previously in Eq. (9.22).
At high temperatures T ≫ Tc , the response function Rq~ is finite, indicat-
ing a finite response to the perturbation ∆(~r). Lowering the temperature, we
173
see that the response to the fictitious field ∆(~r) diverges upon approaching
the critical temperature Tc . The divergence signals a phase transition to a
state in which F ∗ (~r) acquires a nonzero value spontaneously.35 This is the
superconducting state. This important observation is the basis for the Green
function description of the superconducting state of the next section.
The fact that the quantity F ∗ becomes nonzero in the superconducting
phase has two consequences: First, the absolute value |F ∗ | becomes nonzero,
and, second, F ∗ acquires a phase. Upon rethinking this, the second statement
may be more striking than the first one. After all, what determines the phase
of F ∗ ? For this, it is useful to compare the normal-metal-superconductor
phase transition to the normal-metal–ferromagnet transition we studied in
the previous chapter. Upon passing the Stoner instability, the spin polar-
ization acquired a finite magnitude, and a certain direction, although the
Hamiltonian did not contain any preferred direction. The spin-rotational
symmetry is broken spontaneously upon entering the ferromagnetic phase.
In principle, the direction of the magnetization could be fixed by the ad-
dition of an infinitesimal magnetic field to the Hamiltonian. In the same
way, the superconducting phase is a “broken symmetry phase”. The quan-
tity F ∗ acquires a phase, whereas the Hamiltonian does not contain a term
that determines the phase of F ∗ . As in the case of the ferromagnet, the
symmetry may be broken by a small perturbation, which has the form of the
perturbation of Eq. (9.28).
9.3 Green’s function description of a superconductor

In this section, we will derive an equation of motion for the superconductor
Green functions. Hereto, we need the equation of motion for the creation
and annihilation operators. Using the coordinate representation (9.21) of
the interaction Hamiltonian, we find
" #
∂ X †
ψσ (~r) = −(H0 − µ)ψσ (~r) + λ ψσ′ (~r)ψσ′ (~r) ψσ (~r)
∂τ σ ′
" #
∂ † X †
ψ (~r) = (H0∗ − µ)ψσ (~r) − λψσ† (~r) ψσ′ (~r)ψσ′ (~r) . (9.33)
∂τ σ ′ σ
35
If one would insist on a canonical formulation, the quantity that turns nonzero at the
normal-metal–superconductor transition is hN + 2|ψ~† ψ~† |N i.
k↑ k↓
174
Here H0∗ is the complex conjugate of the Hamiltonian H0 . Eventually, the
repulsive part of the interaction at higher frequencies can be included into
H0 as a self-consistent Hartree-Fock potential.
Slightly generalizing the arguments of the preceding section, we charac-
terize the superconducting state by means of the standard Green function
Gσσ′ (~r, τ ; ~r′τ ′ ) and by the so-called “anomalous Green function”
Fσσ′ (~r, τ ; ~r′ , τ ′ ) = hTτ ψσ (~r, τ )ψσ′ (~r′ , τ ′ )i,

+
Fσσ r , τ ; ~r′ , τ ′ ) = hTτ ψσ† (~r, τ )ψσ† ′ (~r′ , τ ′ )i,
′ (~ (9.34)
where H0 is the electron Hamiltonian without interactions. Note that the

absence of a minus sign in the definition of the anomalous Green functions
F and F + . The anomalous Green functions F and F + are conjugates in the
sense of Ch. 2,
+
Fσσ r, τ ; ~r′ , τ ′ ) = [Fσ′ σ (~r′ , τ ′ ; ~r, τ )]∗ .
′ (~ (9.35)
Writing down the equation of motion for G, we the find
∂
− Gσσ′ (~r, ~r′ ; τ ) = δ(~r − ~r′ )δ(τ )δσσ′ + (H0 − µ) Gσσ′ (~r, ~r′; τ ) (9.36)
∂τ
X D E
+ λ Tτ ψσ† ′′ (~r, τ )ψσ′′ (~r, τ )ψσ (~r, τ )ψσ† ′ (~r′ , 0) .
σ′′
The imaginary-time arguments of the creation and annihilation operators in

the last term on the r.h.s. of Eq. (9.37) should be increased by an infinitesimal
amount in order to ensure the correct order under the time-ordering operation
Tτ . Anticipating a description in terms of a self-consistent field (as in the
Hartree-Fock approximation we discussed in Ch. 7), the average of a product
of four creation and annihilation operators on the right hand side of Eq.
(9.36) can be written as a product of two pair averages using Wick’s theorem.
Application of Wick’s theorem results in three terms. Two of those are the
Hartree and Fock terms we encountered before in Chapter 7. These can
be absorbed into the Hartree and Fock self consistent potentials, which is
included in the Hamiltonian H0 . The third term involves the anomalous
Green function F ,
D E
Tτ ψσ† ′′ (~r, τ )ψσ′′ (~r, τ )ψσ (~r, τ )ψσ† ′ (~r′ , 0) → Fσ′′ σ (~r, τ + η; ~r, τ )
× Fσ+′′ σ′ (~r, τ ; ~r′ , 0),
175
where η is a positive infinitesimal. Defining the field ∆σσ′ (~r) as36
∆σσ′ (~r) = −∆σ′ σ (~r) = λFσσ′ (~r, τ + η; ~r, τ ), (9.37)
we arrive at the equation of motion

∂
− − (H0 − µ) Gσσ′ (~r, τ ; ~r′ , τ ′ )
∂τ
X
+ ∆σσ′′ (~r)Fσ+′′ σ′ (~r, τ ; ~r′ , τ ′ ) = δ(~r − ~r′ )δ(τ − τ ′ ). (9.38)
σ′′
Note that ∆(~r) is antisymmetric in the spin indices. This reflects the fact
that the superconducting instability exists for pairs of electrons with zero
spin only. Similarly, one derives an equation of motion for the anomalous
Green function,

∂ ∗ +
− (H0 − µ) Fσσ ′ (~r , τ ; ~r′ , τ ′ )
∂τ
X
− ∆∗σ′′ σ (~r)Gσ′′ σ′ (~r, τ ; ~r′ , τ ′ ) = 0, (9.39)
σ ′′

∂
− − (H0 − µ) Fσσ′ (~r, τ ; ~r′ , τ ′ )
∂τ
X
− ∆σσ′′ (~r)Gσ′ σ′′ (~r′ , τ ′ ; ~r, τ ) = 0, (9.40)
σ′′
where
∆∗σ′ σ (~r) = λFσσ
+
′ (~
r , τ + η; ~r, τ ). (9.41)
These equations are known as Gorkov’s equations. Note that Eq. (9.39)
contains the complex conjugate of the Hamiltonian H0 .
It is the presence of the field ∆(~r) that makes the Gorkov equations dif-
ferent from the equations for the Green function in a normal metal. The field
∆ is related to the energy gap for quasiparticle excitations in a superconduc-
tor (see below). It also serves as an “order parameter” that distinguishes
36
Strictly speaking, one would need to take the finite range of the attractive interaction
into account. In accordance with the momentum cut off at |ε~k − µ| ≥ ~ωD , the interaction
(9.21) has a range of order vF /ωD . However, as we’ll see below, the anomalous Green
function F(~r, ~r′ ) is smooth on the scale vF /ωD , its spatial variations occurring on the
much larger scale vF /Tc only. This implies that the approximation made in setting both
coordinates equal in the argument of F is justified.
176
the superconducting state from the normal state. In that sense, it plays the
same role as the magnetization in the normal-metal–ferromagnet transition.
Solving these equations by means of a Fourier transform, we find
(iωn + (ε~k − µ))δ~k~k′ δσσ′

G~kσ,~k′ σ′ (iωn ) = −
ωn2 + (ε~k − µ)2 + |∆↑↓ |2
+ ∆∗σ′ σ
F~kσ,~k ′ σ′ (iω n ) = . (9.42)
ωn2 + (ε~k − µ)2 + |∆↑↓ |2
Self-consistency is achieved upon substitution of the solution for F into the

definition of ∆,
T XX 1
1=λ . (9.43)
V n 2
ωn + (ε~k − µ)2 + |∆↑↓ |2
~k
Performing the summation over Matsubara frequencies, replacing the sum-

mation over momenta ~k by an integration over energies ε~k and remember-
ing that only momenta with |ε~k − µ| < ~ωD play a role, we find the self-
consistency condition
Z ~ωD p
tanh[ ξ 2 + |∆↑↓ |2 /2T ]
1 = λν dξ p . (9.44)
0 ξ 2 + |∆↑↓ |2
This is the same equation as the “gap equation” from standard BCS theory.
The highest temperature with a nontrivial solution of the self-consistency
equation is the point where the normal-metal–superconductor phase transi-
tion occurs. Since ∆↑↓ = 0 at T = Tc , we find that the critical temperature
satisfies the equation
Z ~ωD
tanh(ξ/2Tc )
1 = λν dξ . (9.45)
0 ξ
This is precisely the same equation as we found for the divergence of the
correlator R in the previous section, see Eq. (9.26) above. Hence
Tc = 1.14~ωD e−1/λν . (9.46)
Quasiparticle excitations of the superconductor correspond to poles of

the retarded Green functions G and F . Analytical continuation of Eq. (9.42)
177
gives
(ω + ε~k − µ)δ~k~k′ δσσ′

G~R
kσ,~k ′ σ′
(ω) = −
(ε~k − µ)2 + |∆|2 − (ω + iη)2
+R ∆∗σ′ σ
F~kσ,~k σ (ω) = . (9.47)
′ ′
(ε~k − µ)2 + |∆|2 − (ω + iη)2
We thus conclude that the quasiparticle energies are

q
ξ~k = ± (ε~k − µ)2 + |∆|2 . (9.48)
From here we conclude that the excitation spetrum has a gap of size |∆|.
We close this section with a few remarks on the role of a vector poten-
~ r ) and gauge invariance. A vector potential can be included in the
tial A(~
Hamiltonian H0 by the replacement
~ →∇
∇ ~ − ieA. (9.49)
In the complex conjugate Hamiltonian H0∗ this corresponds to the replace-

ment ∇~ →∇ ~ + ieA.
~ Gauge invariance requires that the theory is invariant
under transformations
A~→A ~ + ∇χ,
~ (9.50)
where χ is an arbitrary function of the position ~r. Green functions are not
gauge-invariant objects. Under the gauge transformation (9.50), the Green
functions transform as
′
G(~r, ~r′) → G(~r, ~r′)eie(χ(~r)−χ(~r )) ,
′
F (~r, ~r′) → F (~r, ~r′ )eie(χ(~r)+χ(~r )) , (9.51)
′
F + (~r, ~r′) → F + (~r, ~r′ )e−ie(χ(~r)+χ(~r )) .
In particular, this implies that the field ∆(~r) transforms as
∆(~r) → ∆(~r)e2ieχ(~r) . (9.52)
The Gorkov equations serve as the basis for further investigations of the
superconducting state. Results obtained from the Gorkov equations agree
with those obtained from the mean-field BCS theory. For reviews, we refer
178
to the book by Tinkham.37 Here we forego a discussion of thermodynamic
properties and the Josephson effect, and limit ourselves to a discussion of the
response of superconductors to electromagnetic radiation.
9.4 A superconductor in a weak electromagnetic field

Here we calculate the current density in a superconductor in response to an
external vector potential A.~
Starting point of the calculation is the Kubo formula. Repeating results
of Sec. 6.3, we have
X eρe
je,~qα (ω) = − ΠRαβ (~
q , ω)Aβ,~q(ω) + Aα,~q (ω), (9.53)
m
β
where ΠR is the retarded current density autocorrelation function,

1
Παβ (~q, τ ) = − hTτ je,~qα (τ )je,−~qβ (0)i, (9.54)
V
and
~je,~q = − e (2~k + ~q)ψ~k,σ
X †
ψ~k+~q,σ . (9.55)
2m
~kσ
Equation (9.53) describes the linear response of the current density to a

vector potential A.~ We used this relation in Sec. 6.3 to find the conductivity
of a disordered normal metal, the coefficient of proportionality between the
~
current density and the electric field E(ω) ~
= iω A(ω). In the normal metal,
~
the choice of the gauge for the vector potential A was not relevant. This
is different for the case of a superconductor. The reason is that the vector
potential A ~ enters in the equation for the single-particle Green functions G
and F , and, hence, affects the value of the “order parameter” ∆. To linear
~ ∆ can only depend on ∇
order in A, ~ · A.
~ Hence, if we choose the gauge such
~ ~
that ∇ · A = 0, ∆ is unchanged to linear order in the vector potential. This
choice of the gauge is known as the “London gauge”.
37
See Introduction to Superconductivity, M. Tinkham (Mc Graw Hill College Div.,
1995). Other useful references are Superconductivity of Metals and Alloys, P.G. de
Gennes (Perseus, 1999) and Methods of Quantum Field Theory in Statistical Physics,
A.A. Abrikosov, L.P. Gorkov, and I.E. Dzyaloshinski (Dover 1977).
179
Let us now calculate the current density autocorrelation function for the
case of a clean superconductor. Using the Matsubara frequency language,
we find
e2 T X
Παβ (~q, iΩn ) = (2kα + qα )(2kβ + qβ )
2m2 V
~k,m
h
× G~k+~q,~k+~q(iωm + iΩn )G~k;~k (iωm )
i
+ F~k+~q,~k+~q(iωm + iΩn )F~k,+~k (iωm ) . (9.56)
Changing variables to ~k± = ~k ± (~q/2) and ωm± = ωm ± (Ωn /2), we rewrite

this as
2e2 T X h
Παβ (~q, iΩn ) = k α β G~k+ ,~k+ (iωm+ )G~k− ,~k− (iωm− )
k
m2 V
~k,m
i
+ F~k+ ,~k+ (iωm+ )F~k+ ,~k (iωm− )
− −
2
2e T X
= kα kβ (9.57)
m2 V
~k,m
(iωm+ + ε~k+ − µ)(iωm− + ε~k− − µ) + |∆|2

× 2 2
,
(ωm+ + (ε~k+ − µ)2 + |∆|2 )(ωm− + (ε~k− − µ)2 + |∆|2 )
where we substituted the solution (9.42) for the Green functions G and F .
Next, we replace the summation over ~k by an integration over ξ = ε~k − µ
and over the angle θ between ~q and ~k.
Next we introduce polar coordinates for the vector ~k. We choose the di-
rection of ~q as the polar axis. The average over the azimuthal angle φ can
be performed directly. Noting that ~q and A ~ are orthogonal in the London
gauge, upon performing the average over φ we find that Π is diagonal, and
proportional to (1/2) sin2 θ, where θ is the angle between ~q and ~k. Further,
since the main contribution comes from momenta close to the Fermi momen-
tum, we set the magnitude of the factors kα and kβ equal to kF . Finally,
using the equality νkF2 /m = 3n/2, where n is the electron density and ν the
normal-metal density of states at the Fermi level (per spin direction and per
unit volume), and replacing the summation over ~k by an integration over
180
ξ = ε~k − µ and an integration over θ, we find
Z π
3T e2 n XZ
Παβ (~q, iΩn ) = δαβ dξ dθ sin3 θ
4m m 0
(iωm+ + ξ+ )(iωm− + ξ− ) + |∆|2

× 2 2 2 2
,
(ωm+ + ξ+ + |∆|2 )(ωm− + ξ− + |∆|2 )
where ξ± = ξ ± (vF q/2) cos θ.

Because of the approximations involved, the summation over Matsubara
frequencies should be performed before the integration over ξ. Performing
the integration over ξ first gives a different answer. The answer depends on
the order of the integrations because the integral (9.58) is formally divergent.
Our formulas for the ξ-dependence of the single-electron Green functions are
valid close to the Fermi level only. After the summation over the Matsubara
frequencies, the ξ integration is, in fact, convergent, the main contribution
coming from momenta near the Fermi level.
For the case of the superconductor, there is a trick that allows us to do
the ξ integration first, nevertheless. The trick is to look at the difference
between current density correlation functions for the normal metal and for
the superconductor. For this difference, the integrations and summations
over ξ and ωm are convergent, so that the order of integrations is no longer
important. Hence, if we calculate the difference between current-current
correlation functions for the normal metal and superconducting states, we can
perform the ξ integration before the summation over Matsubara frequencies.
The current density autocorrelation function for the normal metal is
Z π
N 3T e2 n XZ 1
Παβ (~q, iΩn ) = δαβ dξ dθ sin3 θ
4m m 0 (iωm+ − ξ+ )(iωm− − ξ− )
eρe
= , (9.58)
m
see Ex. 6.3. Integrating with respect to ξ we then obtain
eρe 3πT eρe XZ π
Παβ (~q, iΩn ) = − δαβ dθ sin3 θ
m 4m m 0
" p p
2
(iωm+ + i ωm+ + |∆|2 )(iωm− − qvF cos θ + i ωm+ 2
+ |∆|2 ) + |∆|2
× 2
p
2
p
2
(ωm− + |∆|2 + (qvF cos θ − i ωm+ + |∆|2 )2 ) ωm+ + |∆|2
181
C1 C1
C2 −∆ ∆
C2
C2
−∆− i Ω n ∆− iΩ n C2
Figure 54: Integration contour for the calculation of Eq. (9.61).

p p #
2
(iωm− + i ωm− + |∆|2 )(iωm+ − qvF cos θ + i ωm− 2
+ |∆|2 ) + |∆|2
+ 2
p
2
p
2
.
(ωm+ + |∆|2 + (qvF cos θ − i ωm− + |∆|2 )2 ) ωm− + |∆|2
(9.59)
Next, we perform the summation over the Matsubara frequency ωm . We
undo the shift of variables for the Matsubara frequencies, i.e., we replace
ωm+ → ωm + Ωn and ωm− → ωm . We then represent the summation over
Matsubara frequencies as an integral in the complex plane and deform the
integration contour as shown in Fig. 54. The integrand has a square root
branch cut. In identifying the correct sign of the square root, one needs the
relations
p
p
2 2 |∆|2 − ωp 2 if |ω| < |∆|,
|∆| − (ω ± iη) = 2 2
(9.60)
∓i sign(ω) ω − |∆| if |ω| ≥ |∆|.
After the integration contours have been deformed as in Fig. 54, we can
perform the analytical continuation iΩn → ω + iη.
Obtaining the final equations for Παβ (~q, ω), which contain one integration
over a real energy variable and one integration over the angle θ is straightfor-
ward, but tedious. Below, we’ll restrict ourselves to static response ω → 0.
In this case, the integration above the lower branch cut cancels the integra-
tion below the upper branch cut in Fig. 54. What remains is the integration
above the upper branch cut and the integration below the lower branch cut,
Z π Z ∞
R eρe 3πeρe 3
Παβ (~q, 0) = − δαβ dθ sin θ dζ tanh(ζ/2T )
m 8πm 0 ∆
X |∆|2
× p .
± ζ 2 − |∆|2 (|∆|2 − (ζ ± iη)2 + (vF2 q 2 /4) cos2 θ)
182
(9.61)
The easiest way to evaluate this integral is to rewrite it as a sum over Mat-
subara frequencies,
eρe 3πeρe T XZ π
R
Παβ (~q, 0) = δαβ − δαβ dθ sin3 θ
m 4m m 0
|∆|2
× p . (9.62)
2 + |∆|2 + (v 2 q 2 /4) cos2 θ)
(ωm 2 + |∆|2
ωm
F
Evalution of this expression is further simplified in the limits qvF ≪ max(|∆|, T )

and qvF ≫ max(|∆|, T ).
If qvF ≪ max(|∆|, T ), we can neglect the q-dependence of the summand
and find

R eρe ns (T )
Παβ (~q, 0) = δαβ 1 − , (9.63)
m n
where ns is the so-called “superconducting density”,
nπT |∆|2 X 1
ns = . (9.64)
2 m
(ωm + |∆|2 )3/2
2
For zero temperature the superconducting density is equal to the electron

density n. The superconducting density approaches zero as T approaches Tc .
If qvF ≫ max(|∆|, T ), the main contribution to the integral over the angle
θ comes from angles close to θ = π/2. For those angles we may set sin θ = 1.
Replacing the θ-integration by an integration over cos θ and extending the
integration interval to the entire real axis, we find
eρe 3π 2 eρe T X |∆|2

ΠR
αβ (~
q , 0) = δαβ − δαβ 2 + |∆|2
m 2mqvF m
ωm
eρe 3π 2 eρe |∆| |∆|
= δαβ − δαβ tanh . (9.65)
m 4mqvF 2T
Unlike in the case of a normal metal, we find that for a superconductor
~ itself,
the current density ~j(~q, ω) is proportional to the vector potential A
and not to its time-derivative. This dependence is the cauase of the Meissner
effect, the phenomenon that no magnetic field can exist in a superconductor,
183
except for a thin layer of thickness δ close to the superconductor surface. For
a superconductor for which vF /δ ≪ max(T, ∆), the Meissner effect can be
described with the help of Eq. (9.63) above,
2
~j = − ns e A.
~ (9.66)
m
Combining Eq. (9.66) with the Maxwell equation ∇2 A ~ = −~j for the London
~ ·A~ = 0, we find δ =
p
gauge ∇ m/ns e2 . Equation (9.66) is known as
the London equation, and δ is known as the London penetration depth.
A superconductor for which δ ≫ vF / max(T, ∆) is said to be “of London
type”. For a superconductor for which δ ≪ vF / max(T, ∆), Eq. (9.65) has
to be used instead of Eq. (9.63). In this case, the mathematical solution is
more complicated, but the qualitative conclusion, no magnetic field inside
the superconductor, except for a surface layer of thickness δ, remains true.
Superconductors for which δ ≪ vF / max(T, ∆) are said to be “of Pippard
type”. Equation (9.65) is known as the Pippard equation.
Most pure superconductors at T ≪ Tc are of Pippard type. Close to Tc ,
δ increases, and a crossover to the London type of behavior is seen.
These calculations can be repeated for the case of a disordered supercon-
ductor with elastic mean free time τ . The calculation proceeds along the
lines of Sec. 6.3 and has the following results: First, the addition of impuri-
ties does not affect the magnitude of the superconducting order parameter
∆. Second, the effect of impurities is small if qvF τ ≫ 1, whereas, in the
opposite limit, qvF τ ≪ 1, one recovers the London equation (9.66), with a
reduced superconducting density
nπT |∆|2 X 1
ns = p . (9.67)
2 2 2
m (ωm + |∆| )(
2 + |∆|2 + 1/2τ )
ωm
As a result of the reduced superconducting density, the London penetration

depth δ is strongly enhanced. As a result, for sufficiently strong disorder,
disordered superconductors are always in the London limit.
184
Figure 55: Diagram contribution to the renormalization of electron-phonon vertex
by phonon scattering.
9.5 Exercises
Exercise 9.1: Migdal’s theorem
In principle, the electron-phonon vertex can also be renormalized by the

repeated exchange of phonons, as, e.g., in the diagram of Fig. 55. (Note that
the phonon line exchanged between the electrons is a renormalized phonon
line.) Show that such diagrams are smaller pthan the diagrams without the
exchange of a phonon by a large factor ∼ M/m, where M and m are ion
and electron mass, respectively.
Exercise 9.2: Thermodynamic properties
One can calculate the effect of the superconductivity on the free energy F
using the general relation between F and the interaction Hamiltonian H1 of
Eq. (9.20).
Z λ
1
F − F0 = dλ′ ′ hH1 (λ′ )i. (9.68)
0 λ
(a) Keeping contributions that are specific for a superconductor only, show
that the superconductivity contribution δF to the free energy reads
Z λ ′
dλ
δF = ′2
|∆(λ′ )|2 , (9.69)
0 λ
where ∆(λ′ ) is the superconducting order parameter at interaction
strength λ′ .
185
(b) One can rewrite the gap equation (9.44) as
" ∞
#
2ωD X
n+1
1 = λν ln −2 (−1) K0 (n|∆|/T ) . (9.70)
|∆| n=1
Using this result, together with Eq. (9.69), show that

" ∞
#
2 Z n|∆|/T
|∆|2 X T
δF = −ν −2 (−1)n+1 2 K1 (x)x2 dx , (9.71)
2 n=1
n 0
where K1 is a Bessel function.
(c) Show that, at temperatures T ≪ Tc , the heat capacity of a supercon-

ductor is r
2π∆50 −∆0 /T
C = 2ν e , (9.72)
T3
where ∆0 is the magnitude of the order parameter at zero temperature.
186
10 Diffusion modes
10.1 polarizability function for a dirty metal
Let us return to our calculation of the polarizability function of the non-
interacting electron gas. In Sec. 6.2 we looked at the response of a clean
system. What about a disordered metal?
First, at high frequencies or small wavelengths, we do not expect that the
disorder will play an important role. However, at low frequencies or large
wavelengths, impurity scattering becomes important and should affect the
way the electron density responds to a change of the potential.
Repeating the calculations of Sec. 6.2, we find that the polarizability
function χe is equal to
2T e2 X
χe (~q, iΩn ) = G~k+~q,~k′ +~q(iωm + iΩn )G~k′ ,~k (iωm ), (10.1)
V
~k,~k ,m
′
where the Green function G is calculated in the presence of the exact impu-
rity potential. We will be interested in the impurity averaged polarizability
function hχe i. For the disorder, we use the model of Gaussian white noise,
see Sec. 5.3. Evaluating the average, we limit ourselves to the ladder dia-
grams shown in Fig. 56. We rewrite the diagrams in terms of a propagator
D shown in Fig. 57,
D(~k, iωm ; ~k + ~q, iωm + iΩn )

 −1
1 X
= 1 − hG~ ′ (iωm + iΩn )ihG~k′ (iωm )i . (10.2)
2πντ V ′ k +~q
~k
We can replace the summation over ~k ′ by an integration over the energy

ξ = ε~k′ −µ and the angle θ between ~k ′ and ~q. We are interested in wavevectors
q ≪ kF , so that we can expand
ε~k+~q = ε~k + vF q cos θ. (10.3)
Since the main contribution of the ξ-integration comes from a window of

width ∼ ~/τ around the Fermi energy, we can assume a constant density
of states ν for the entire integration range. The result of the ξ-integration
depends on the position of the Matsubara frequencies iωm + iΩn and iωm
187
000
111
ωm 111
000
000
111
000
111
ωm ωm ωm ωm ωm
k 000
111 k k k’ k k" k’
000
111
000
111
000
111
000
111 = + + + ...
k+q111
000
000
111
000
111 k+q k+q k’+q k+q k"+q k’+q
000
111
ωm+Ω n 111
000
000
111 ωm+Ω n ωm+Ω n ωm+Ω n ωm+Ω n ωm+Ω n
Figure 56: Diagrammatic representation of ladder series for the polarizability

function is a disordered non-interacting electron gas.
with respect to the real axis. If they are on the same side of the real axis, the
integration over the energy ξ vanishes, and one finds D = 1. If ωm + Ωn > 0
whereas ωm < 0, the ξ-integration is finite, and one has
1 X
hG~ ′ (iωm + iΩn )ihG~k′ (iωm )i
2πντ V ′ k +~q
~k
Z 1
1 1
= d cos θ . (10.4)
2 −1 1 + Ωn τ + ivF qτ cos θ
For low frequencies, iΩn → ω ≪ 1τ and for long wavelengths, we can expand
the fraction in the integrand and perform the θ-integration. The result is
1 X 1
hG~k′ +~q(iωm + iΩn )ihG~k′ (iωm )i = 1 − Ωn τ − vF2 q 2 τ 2 .
2πντ V ′ 3
~k
Similarly, for ωm + Ωn < 0 and ωm > 0, one finds

1 X 1
hG~k′ +~q(iωm + iΩn )ihG~k′ (iωm )i = 1 + Ωn τ − vF2 q 2 τ 2 .
2πντ V ′ 3
~k
We thus conclude that for |Ωn |τ ≪ 1, qvF τ ≪ 1, one has
D(~k, iωm ; ~k + ~q; iωm + iΩn )



 1 if ωm + Ωn > 0, ωm > 0,
2
1/[τ (Dq + Ωn )] if ωm + Ωn > 0, ωm < 0,

= (10.5)
 1/[τ (Dq 2 − Ωn )] if
 ωm + Ωn < 0, ωm > 0,
1 if ωm + Ωn < 0, ωm < 0,

where we introduced the “diffusion coefficient” D = vF2 τ /3. Because of the

appearance of the factor Dq 2 ± Ωn in the denominator, which is the kernel of
the diffusion equation, the propagator D is referred to as the “diffuson” or
188
111
000
000
111
000
111
000
111
000
111
000
111
000
111
000
111 = + + + ...
000
111
000
111 1
000
111
000
111
000
111
000
111
000
111
Figure 57: Diagrammatic representation of the diffusion propagator.
“diffusion propagator”. Performing the analytical continuation iωm → ε±iη,

iωm + iΩn → ε + ω ± iη, we refer to the four cases listed in Eq. (10.5)
as retarded-retarded, retarded-advanced, advanced-retarded, and advanced-
advanced,
D RR (~k, ε; ~k + ~q; ε + ω) = 1,
D AR (~k, ε; ~k + ~q; ε + ω) = 1/[τ (Dq 2 − iω)],
D RA (~k, ε; ~k + ~q; ε + ω) = 1/[τ (Dq 2 + iω)], (10.6)
D AA
(~k, ε; ~k + ~q; ε + ω) = 1.
In terms of the diffusion propagator, the polarizability reads

2T e2 X
χe (~q, iΩn ) = hG~k+~q,~k+~q(iωm + iΩn )ihG~k,~k (iωm )i
V
~k,m
× D(~k, iωm ; ~k + ~q, iωm + iΩn ). (10.7)
The calculation of χe differs in an essential way from the calculation of the

diffusion propagator, since Eq. (10.7) involves both a summation over Mat-
subara frequencies ωm and a summation over wavevectors ~k. The summation
over ωm , which needs to be performed first, can be done as in Sec. 6.3. Re-
placing the summation over Matsubara frequencies by an integration in the
complex plane and deforming the contours as in Fig. 25, and performing the
analytical continuation iΩn → ω + iη, we find
e2
Z X
R
χe (~q, ω) = dξ tanh(ξ/2T )
2πiV
~k
h
× hG~R k,~k
(ξ)ihG~R q ,~k+~
k+~ q
(ξ + ω)iD RR(~k, ξ; ~k + ~q, ξ + ω)
− hG~A
k,~k
(ξ)ihG~R q ,~k+~
k+~ q
(ξ + ω)iD AR(~k, ξ; ~k + ~q, ξ + ω)
189
+ hG~A
k,~k
(ξ − ω)ihG~R q,~k+~
k+~ q
(ξ)iD AR (~k, ξ − ω; ~k + ~q, ξ)
i
AA ~ ~
− hG~A
k,~k
(ξ − ω)ihG A
~k+~
q ,~k+~
q
(ξ)iD ( k, ξ − ω; k + ~
q , ξ) .
For the first and fourth terms between the brackets, the diffuson propagator
D RR = D AA = 1. For the second and third term, we write D AR = 1+(D AR −
1). Then there is a contribution from all four terms between brackets, which
is identical to that calculated in Eq. (6.63), and a contribution from the
remaining part of the second and third term between brackets. Adding both
terms and making use of the limits qvF τ ≪ 1, ωτ ≪ 1 and εF τ ≫ 1, we find
e2

ξ ξ+ω
Z
R 2
χe (~q, ω) = −2e ν − dξ tanh − tanh
2πiV τ (Dq 2 − iω) 2T 2T
X
× hG~A
k,~k
(ξ)ihG~R q ,~k+~
k+~ q
(ξ + ω)i
~k
2 2e2 νiω
= −2e ν −
Dq 2 − iω
2e νDq 2
2
= − 2 . (10.8)
Dq − iω
To see what this result means, it is useful to perform a Fourier transform
to time t and coordinate ~r. For the density response to a sudden increase of
the potential δφ(~r, t) = δ(~r − ~r′ )θ(t)δφ we find from Eq. (10.8)
π 3/2
2 ′ −|~ r ′ |2 /4Dt
r−~
δρe (~r, t) = −2e ν δ(~r − ~r ) − e δφ. (10.9)
Dt
This result has a very simple interpretation: the change in the potential at
~r = ~r′ creates a density profile that spreads diffusively. This result, which was
obtained by summation of a set of diagrams, thus reproduces what we know
based on the picture of electron diffusion. Note that the above calculation
is quite similar to that of the Drude formula. So it is no surprise that we
recovered results from diffusion theory in both cases.38
The fact that dynamics in a dirty metal is diffusive, rather than ballistic,
for time scales longer than the mean free time τ and length scales longer than
38
Note that the Drude conductivity is a function of the diffusion constant as well,
σαβ (ω) = 2e2 νDδαβ .
190
φ (r,t)
electron motion
to screen potential
Figure 58: The electrons move periodically in order to screen an external time-
dependent potential. It depends on the frequency of the potential whether the
electron gas can sustain this periodic motion in the presence of disorder.
the mean free path l = vF τ , has important consequences for the screening
of the Coulomb interaction: since it is more difficult to move charge around,
the electron gas has a reduced capacity to screen time-dependent potentials.
The relevant frequency for the time dependence is not the ballistic frequency
vF q (as it is in the response of a clean electron gas), but the diffusive one,
ω ∼ Dq 2 , where 1/q ≫ l is the length scale that the system is probed at.
Qualitatively, the effect of disorder on the screened Coulomb interaction can
be seen from the following picture, see Fig. 58. In order to screen a time-
dependent potential with frequency ω and wavevector q, the electrons need
to move a distance ∼ 1/q in a time ∼ 1/ω. As long as ω ≪ Dq 2 the electron
gas can accomodate such a displacement, and Thomas-Fermi screening theory
applies. If, however, ω ≫ Dq 2 the electron gas cannot follow the potential,
and screening is absent.
The reduced screening capability is seen quantitatively in the RPA cal-
culation of the dielectric function of the interacting electron gas, which now
becomes
k2 Dq 2
ǫRPA (~q, iΩn ) = 1 + 2s , (10.10)
q |Ωn | + Dq 2
where ks2 = 2νe2 /ǫ0 is the Thomas-Fermi wavenumber. Similarly, the effec-
tive electron-electron interaction becomes
e2
Vq~RPA (iΩn ) =
ǫ0 q 2 ǫRPA (~q, iΩn )
−1
e2 Dq 2

2 2
= q + ks . (10.11)
ǫ0 |Ωn | + Dq 2
For |ω| ≪ Dq 2 the effect of the disorder is negligible, whereas for ω ≫ Dq 2

the Coulomb interaction is essentially unscreened. This is different from the
191
111111111
0000000000000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
000000000
1111111110000000000
1111111111
000
111
000000000
1111111110000000000
1111111111
000
111
000000000
1111111110000000000
1111111111
I000
111
000000000
1111111110000000000
1111111111
000
111
000000000
1111111110000000000
1111111111
000
111
disordered 000000000
1111111110000000000
1111111111
000
111
000000000
1111111110000000000
1111111111 reference
000
111
000000000
1111111110000000000
1111111111
000
111
000000000
1111111110000000000
1111111111
000
111
conductor 000000000
1111111110000000000
1111111111 electrode
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
1111111110000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
000000000
111111111
000000000
1111111110000000000
1111111111
0000000000
1111111111
V
Figure 59: Tunnel spectroscopy setup for a measurement of the spectral density
of a disordered conductor.
case of a clean metal, where Thomas-Fermi screening breaks down at the

much higher frequency ω ∼ vF q.
One example where the modified screening capability of electrons plays
a role is the so-called “zero bias anomaly” in the tunneling density of states.
The density of states can be measured by a tunneling experiment, in which
electrons from a reference conductor with known (and constant) spectral
density tunnel into the sample, see Fig. 59. Indeed, as shown in Ex. 6.1, the
current I through a tunneling barrier contacting the sample at position ~r is
proportional to the spectral density,
Z ∞
∂I(~r, eV ) 1 X 1
∝ dξA(~r, ξ) 2 . (10.12)
∂V 4T −∞ ±
cosh [(ξ ± eV )/2T ]
(At zero temperature, this simplifies to ∂I/∂V ∝ A(~r, eV ).)

For a non-interacting electron gas, the spectral density is essentially en-
ergy independent, hence the tunnel current is proportional to the applied
bias voltage V and independent of the temperature T . For the disordered
interacting electron gas, the situation is more complicated. When an electron
tunnels into the sample, it charges the sample locally. For a cloud that is
bigger than the mean free path l, dispersion of the charge cloud is a slow pro-
cess, which is determined by the diffusive motion of the electrons. Until the
energy of the charge cloud, which is a sum of kinetic and the Coulomb energy
contributions, has dropped below the bias voltage eV or the temperature T ,
192
111
000
000
111 111
000
000
111 111
000
000
111 111
000
000
111
000
111 000
111 000
111 000
111
000
111 000
111 000
111 000
111
000
111 000
111 000
111 000
111
000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111
000
111 000
111 000
111 000
111
000
111 000
111 000
111 000
111
000
111 000
111 000
111 000
111
000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111
000
111 000
111 000
111 000
111
000
111 000
111 000
111 000
111
Figure 60: Lowest order corrections to the single-particle Green function in a

disordered interacting electron gas. Left: Hartree diagram, right: Fock diagram.
The shaded blocks represent the diffusion propagator, see Fig. 57.
the electron that tunnels into the metal has to exist in a virtual state. The
prolonged existence in a virtual state suppresses the tunneling current or, in
other words, the density of states. At low voltages and low temperatuers,
the suppression should be maximal, the tunneling density of states being an
increasing function of V and T for small voltages and low temperatures.
A non-perturbative semiclassical theory of the zero bias anomaly has
been formulated by Shytov and Levitov [JETP Lett. 66, 214 (1997)]. Here
we present an argument based on perturbation theory in the electron-electron
interaction, following the original calculation of Altshuler and Aronov [Sov.
Phys. JETP 50, 968 (1979)] and Altshuler, Aronov, and Lee [Phys. Rev.
Lett. 44, 1288 (1980)]. To lowest order in the interaction, this correction
is given by Hartree and Fock diagrams. The diffusive electron motion is
reflected in the presence of the diffusion ladders around the interaction line,
which, in turn, represents the effective interaction of Eq. (10.11) above. The
two relevant diagrams for the correction to the single-particle Green function
are shown in Fig. 60.
We consider the Fock diagram first. The correction to the single-electron
Green function shown in Fig. 60 reads
T XX 0
δhG~k,~k (iωn )i = hG~k,~k (iωn )i2 D(~k, iωn ; ~k + ~q, iωm )2
V mq~
0
× (−Vq~(iωn − iωm ))hG~k+~
q ,~k+~
q
(iωm )i. (10.13)
The correction to the spectral density is found as
δA(~r, ω) = −4Im δGR (~r, ~r; ω)
193
C2
C1
C1
Figure 61: Integration contour for the calculation of the zero-bias anomaly to the
tunneling density of states.
4T XX
0 2 ~ ~ q , iωm )2
= − Im hG~k,~k (iωn )i D(k, iωn ; k + ~
V2
q m
~k,~

0
× (−Vq~(iωn − iωm ))hG~k+~ (iω )i . (10.14)

q ,~k+~
q m
iωn →ω+iη
where the extra factor two accounts for spin degeneracy.

Since the diffusion propagator is divergent for small momentum transfers,
the main contribution to the integration in Eq. (10.13) comes from small ~q.
For these momenta, we can use the screened interaction of Eq. (10.11).
A singular contribution to the density of states can come from the pole
of the diffusion propagator at small momentum and frequency. For the case
ωn > 0, this pole exists for ωm < 0 only. Replacing the summation over ωm
by an integration in the complex plane, deforming the integration contour as
in Fig. 61, and performing the analytical continuation iωn → ω + iη we find
1 X ∞ 1
Z
0R 2
δA(~r, ω) = Im 2 dξ tanh(ξ/2T )hG~k, ~k (ω)i
V πi −∞ (Dq 2 + iξ − iω)2 τ 2
~k,~
q
0A
× (−Vq~ (ω − ξ + iη))hG~k+~q,~k+~
q
(ξ)i. (10.15)
Since the main dependence on ω and ~q comes from the pole of the diffusion
operator, we set ~q = 0 and ω = ξ in the arguments of the Green functions
in Eq. (10.15). Performing the summation over ~k and shifting integration
variables ξ → ω − ξ we then find
2ν X ∞

ξ − ω Vq~(ξ + iη)
Z
δA(~r, ω) = −Im dξ tanh . (10.16)
V −∞ 2T (Dq 2 − iξ)2
q~
194
The summation over ~q and integration over ξ are, in fact, divergent. The
divergence is not physical, since our approximations break down when vF qτ
and ωτ become of order unity. In order to arrive at a meaningful result that
describes the temperature and energy dependence of the density of states, we
substract δA in the limit ω → 0, T → 0 (limits taken in this order). In Eq.
(10.16) this amounts to the replacement of tanh[(ξ − ω)/2T ] by tanh[(ξ −
ω)/2T ] − sign (ξ). For the interaction, we substitute the screened Coulomb
interaction in the presence of disorder, Eq. (10.11) above. First performing
the remaining summation over ~q, one finds
Z ∞
1 ξ−ω
δA(~r, ω) = −Im dξ tanh − sign (ξ)
4πD 3/2 −∞ 2T
" #
1 1
× √ −p . (10.17)
−iξ Dks2 − iξ
We omit the second term between the square brackets, which gives a vanish-
ing contribution to the spectral density if ω, T ≪ Dks2 . For the remaining
integration, we find
T 1/2
δA(~r, ω) = − √ f (ω/T ), (10.18)
π 2(~D)3/2
where ∞
dy sinh y
Z
f (x) = −1 . (10.19)
0 2y 1/2 cosh y + cosh x
Asymptotically, the function f (x) behaves as
√ π2
f (x) ≈ − x − x−3/2 if x ≫ 1,
24
f (x) ≈ −1.072 if x ≪ 1. (10.20)
You verify that, indeed, δA → 0 if we take the limit ω → 0, followed by
T → 0.
For two dimensional samples and one dimensional samples, the calculation
of the density of states is very similar. The first difference is that the Coulomb
interaction has a different form in one and two dimensions,
 h i−1
 (e2 /2ǫ0 ) q + (e2 /ǫ0 )ν Dq2 2 if d = 2,
|Ωn |+Dq
Vq~(iΩn ) = h i−1 (10.21)
 (e2 /ǫ ) − 4π + 2(e2 /ǫ )ν Dq2 if d = 1.
0 ln(q 2 a2 ) 0 |Ωn |+Dq 2
195
Here ν is the density of states per spin direction and per unit area (in two
dimensions) or unit length (in one dimension), whereas a is the thickness
or radius of the sample. In quasi two-dimensional samples, ν = ν3 a, where
ν3 is the three-dimensional density of states. With this modification, the
calculation proceeds as in the three dimensional case.
In two dimensions, the final integration over ξ and summation over ~q is
logarithmically divergent. Performing the summation over ~q first and trun-
cating the ξ-integration at ξ ∼ 1/τ , we find
Z ∞
1 dξ ξ ξ+ω ξ−ω
δA(~r, ω) = ln tanh + tanh
4πD 0 ξ De2 ν/ǫ0 2T 2T

1 max(ω, T )
≈ − ln 2 2 ln [τ max(ω, T )] . (10.22)
4πD D (e /ǫ0 )ν 2 τ
Similarly, in the one-dimensional case one has
−1/2 Z ∞
Dks2

aks dy sinh y
δA(~r, ω) = − √ ln 3/2
,
2 2πT D max(ω, T ) 0 y cosh y + cosh(ω/T )
(10.23)
where ks is the three-dimensional Thomas-Fermi wavenumber.
The contribution from the Hartree diagram is similar. The transferred
momentum ~q in interaction factor of the Hartree diagram does not have to
be small for the momentum in the diffusion propagators to be small. Since
the Hartree diagram involves large momentum exchanges in the interaction
matrix element, the precise magnitude depends on the details of the short-
range part of the Coulomb interaction. In that case we cannot use the simple
screened form (10.11) above, but need information about the microscopic
details of the sample. For the discussion here, we limit ourselves to the
statement that the energy and temperature dependence of δA is still the same
as that of Eq. (10.18), whereas the prefactor depends on sample details. For
details of this calculation we refer to the article Electron-electron interaction
in disordered conductors, by B.L. Altshuler and A.G. Aronov, in the book
Electron-electron interactions in disordered systems, edited by A.L. Efros and
M. Pollak (North-Holland, 1985).
10.2 Weak localization

Our calculation of the conductivity in Sec. 6.3 was done to leading order in
the impurity concentration. The leading contribution consisted of diagrams
196
1111
0000
ωm ωm ωm ωm 0000
1111
0000
1111
0000
1111
k k k’ 0000
1111
1 0000
1111
0000
1111
+ + ... = 0000
1111
0000
1111
k+q k+q k’+q 2πντ 0000
1111
0000
1111
0000
1111
0000
1111
ωm+Ω n ωm+Ω n ωm+Ω n ωm+Ω n 0000
1111
0000
1111
k k k’
0000
1111
1111
0000
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111 = 2πντ + + ...
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
q−k q−k q−k’
Figure 62: Diagrammatic representation of maximally crossed diagrams that con-

tribute to the conductivity (top). The double lines represent the disorder averaged
Green function; the solid dots refer to the renormalized current vertex, see Fig.
23. The maximally crossed diagrams can be represented in terms of the cooperon
propagator (right and below).
without crossed impurity lines. Diagrams that have impurity lines are a fac-
tor 1/ǫF τ smaller. However, as we’ll show here, such diagrams can have a
singular temperature dependence in one and two dimensions, actually diverg-
ing in the limit of zero temperature. The singularity arises from the diffusive
electron motion we considered in the previous section.
The leading correction in powers of 1/ǫF τ is given by the so-called “max-
imally crossed” diagrams.39 The maximally crossed diagrams are shown in
Fig. 62. They can be represented in terms of the renormalized current vertex
(see Sec. 6.3) and the so-called cooperon propagator. The cooperon is a series
of ladder diagrams, similar to the diffusion propegator we encountered in the
previous section. However, unlike for the diffuson, where the difference of the
momenta is constant throughout the ladder, for the cooperon the sum of the
momenta is constant throughout the ladder, see Fig. 62. The cooperon prop-
agator derives its name from its close resemblance to the electron-electron
correlation function that describes the superconducting instability.
39
Don’t be confused by the word “maximally crossed”. These diagrams, in fact, have the
smallest number of crossed impurity lines of all the diagrams that remain after calculation
of the Drude conductivity.
197
The cooperon propagator reads
1
P
2πντ V ~k hGq~−~k,~q−~k (iωm + iΩn )ihG~k,~k (iωm )i
C(~q; iωm , iωm + iΩn ) = 1
P .
1 − 2πντ V ~k hGq~−~k,~
q−~k (iωm + iΩn )ihG~k,~k (iωm )i
(10.24)
Performing the momentum summations, we find for qvF τ ≪ 1 and |Ωn |τ ≪ 1

1
C(~q; iωm , iωm + iΩn ) = θ[(−ωm )(ωm + Ωn )].
τ (Dq 2 + |Ωn |)
(10.25)
For the current-current correlator, the diagrams of Fig. 62 imply a correction

2e2 T X ′
X
δΠαβ (~q → 0, iΩn ) = k α k β hG~k,~k (iωm )ihG~k,~k (iωm + iΩn )
m2 V 2 ′ m
~k,~k
~
× C(~k + ~k ′ ; iωm , iωm + iΩn )
2πντ
× hG~k′ ,~k′ (iωm )ihG~k′ ,~k′ (iωm + iΩn )i, (10.26)
The summation over m and the integrations over ~k and ~k ′ can be done as in
the previous sections. Alternatively, we note that the Green functions have a
very weak dependence on ωm for |ωm | < 1/τ , so that we may set iωn → −iη in
the arguments of the Green functions. Then summation over m gives nothing
but a factor n = Ωn /2πT . Now the analytical continuation iΩn → ω + iη is
possible, followed by division by ω and the limit ω → 0. Also, if we write
~k ′ = −~k + Q~ and note that the mean contribution of the summation over
~ comes from Q
Q ~ close to 0, we can ignore the Q ~ dependence of the single-
electron Green functions in Eq. (10.26). The remaining integration over ~k
then gives
1 X
kα (−kβ )hG~R
k,~k
(0)i2 hG~A
k,~k
(0)i2 = 4πντ 3 kF2 /d, (10.27)
V
~k
where d is the dimensionality of the conductor. Combining everything, we

find
2τ e2 kF2 X iω
δΠR
αβ (~
q → 0, ω) = . (10.28)
πdm2 V DQ2
~
Q
198
Hence, using the general relation D = vF2 τ /d, we find that the correction to
the conductivity reads

1 R

δσαβ = − δΠ (~q → 0, ω)
iω ω→0
2 2 X
2τ e kF 1
= − 2
πdm V DQ2
~
Q
2e2 X 1
= − . (10.29)
πV Q2
~
Q
In three dimensions, the summation over Q~ is divergent for large Q. The

reason for the divergence is that the approximations we made cease to be
valid for Q larger than the inverse mean free path l = vF τ . Truncating the
summation for Ql ∼ 1, we find
e2
δσ ∼ − . (10.30)
l
This correction to the conductivity is a factor kF l smaller than the Drude
conductrivity. Hence, for three dimensions, we conclude that the strategy of
Sec. 6.3, where we kept diagrams without crossed lines only, is valid.
For two dimensions, the summation over Q ~ is divergent both for large Q
and for small Q. The divergence at large Q is removed by truncating the
integration at Q ∼ 1/l. The divergence at small Q, however, has an entirely
different origin. Small Q correspond to large length scales, and on large
length scales our description in terms of non-interacting electrons in a static
and unbounded environment breaks down. In a real sample, electrons exit
the sample through the contacts, or they loose their phase memory because of
scattering from photons or phonons, or because of electron-electron interac-
tions. That means that our description, in which such processes are not taken
into account, breaks down at long time scales or long length scales. Here we
do not attempt to formulate a full microscopic theory that includes the con-
tacts and decohering processes. Instead, we use a phenomenological solution.
We cut off the momentum summation from below at Q ∼ max(1/L, 1/Lφ ),
where L is the system size and Lφ is the length scale at which the elec-
trons loose their phase memory. Alternative, we keep the full momentum
summation but replace the frequency −iω in the Cooperon propagator by
199
max(1/τesc , 1/τφ ), where τesc = L2 /D is the time it takes to exit the sample
through the contacts and τφ = L2φ /D is the dephasing time. We then find40
−(e2 /π 2 ) ln[min(L, Lφ )/l], if d = 2,

δσ = (10.31)
−(e2 /π)[min(L, Lφ ) − l], if d = 1.
In three dimensions, the fact that we need to add a lower cut-off to the
momentum integration has a minor effect. It amounts to the replacement of
Eq. (10.30) by
e2 1

1
δσ = − 2 − . (10.32)
2π l Lφ
These results are very important for our understanding of electron trans-
port through disordered conductors. On the one hand, you verify that δσ
is, at first sight, a factor kF l smaller than the Drude conductivity. However,
δσ has an additional dependence on system size or temperature that leads
to a divergence at large system sizes and low temperatures. (The dephasing
length Lφ → ∞ if T → 0.) Hence, we see that these small corrections be-
come large in large samples at low temperatures, and that the conductivity
becomes significantly smaller than the Drude conductivity. In one and two
dimensions, the correction δσ we calculated is the first signature of the phe-
nomonon of “Anderson localization”, the statement that all “metals” in one
and two dimensions become insulators at sufficiently low temperatures. For
that reason, the correction δσ is called “weak localization correction”.
The weak localization correction has a simple physical explanation. In a
semiclassical picture, the electrons are described as wave packets that move
along well defined trajectories, acquiring a phase as they move along. Hence,
the electron propagation along a certain path is described by an amplitude,
i.e., by an intensity and a phase. The probability for an electron to go
from one location in the sample to another location is proportional to the
square of the amplitudes for all trajectories linking start and finish. Without
quantum phase coherence, one would add intensities of all trajectories, not
amplitudes, i.e., one squares the amplitudes first and then adds them. The
difference between the two descriptions — add squared amplitudes or square
added amplitudes — is “interference”. Interference can be constructive or
destructive. In most cases, the interference correction contains a sum over
40
In obtaining the numerical factor for d = 1 in Eq. (10.31) and for d = 3 in Eq. (10.32)
below, we replaced ω by i/τφ in the diffuson propagator and used τφ = L2 /D to express
τφ in terms of Lφ .
200
Figure 63: Constructive interference of two time-reversed trajectories returning
to the point of departure.
many terms with different phases, and averages to zero. However there is one
exception: interference of trajectories that return to the point of departure.
In this case, time-reversed trajectories acquire the same phase, and, hence,
interfere constructively, see Fig. 63. Thus, the probability to return to the
point of departure is enhanced by quantum interference. This also implies
that the rate at which particles diffuse through the metal is decreased, which,
in turn, leads to a reduction of the conductivity. This reduction is precisely
what we calculated above!
The interference between time-reversed trajectories is described by the
Cooperon propagator. Recall that the Cooperon propagator was important
precisely for the case of opposite momenta ~k and ~k ′ .
In this picture, the cut-offs needed for the calculation of the weak lo-
calization correction are more transparent. The large momentum cut-off
corresponds to the break down of diffusion theory at length scales below the
mean free path. Indeed, since one needs impurity scattering in order to re-
turn to the point of departure, there is a minimal length for such a trajectory.
Also, the interference correction exists only if the electrons keep their phase
memory along the trajectory. This explains the cut-off at low momenta.
In a magnetic field B, time-reversed trajectories acquire a phase difference
equal to 2π times the enclosed flux Φ divided by the flux quantum Φ0 = hc/e.
Since the typical area swept out for a trajectory that ventures a distance L
away from the point of departure is of order L2 , trajectories with L2 B/Φ0 & 1
201
do not contribute to the interference. Hence, in a magnetic field, we should
impose a lower momentum cutoff at Q ∼ 1/lm , where lm = (B/Φ0 )−1/2 is the
so-called magnetic length. If lm is smaller than the dephasing length Lφ and
the system size, the dephasing length Lφ should be replaced by LB in Eqs.
(10.31) and (10.32). Hence, for large magnetic fields, the weak localization
correction to the conductivity is suppressed.
10.3 Formulation in real space

In Sec. 10.1 we saw that the diffuson is equal to the pole of the diffusion
equation. We arrived at this conclusion from an analysis in momentum
represententain, which, in principle, is valid for a bulk system only. For
finite-sized systems or for inhomogeneous sytems, a description in real space
is more useful. In this section, we rederive the results of Sec. 10.1 using the
coordinate representation.
Before embarking on a diagrammatic calculation in coordinate represen-
tation, let us recall the expressions for the impurity averaged Green function,
m r ′ |−|ω1 ||~ r ′ |/vF −|~ r ′ |/2l
hG(~r, ~r′, iω1 )i = − e±ikF |~
r −~ r−~ r −~
, (10.33)
2π~2 |~r − ~r′ |
where the + sign is for ω1 > 0 and the − sign is for ω1 < 0, and the averaging
rule for the Gaussian white noise potential,
~
hU(~r)U(~r′ )i = δ(~r − ~r′ ), (10.34)
2πντ
where τ is the elastic mean free time and ν is the density of states per spin
direction. Equation (10.33) is valid in three dimensions; similar results can be
obtained in one and two dimensions. Throughout this discussion we assume
weak disorder, kF l ≫ 1, where l = vF /τ is the elastic mean free path.
Starting point is the diagrammatic expression for the diffuson operator
in Matsubara representation D(~r, ~r′ ; iω1 , iω2 ), which we repeat in Fig. 64.
Writing down the Dyson equation in real space, one finds
Z
′ ′ ~
D(~r, ~r ; iω1 ; iω2 ) = δ(~r − ~r ) + d~r′′ hG(~r, ~r′′ ; iω1 )i
2πντ
× hG(~r′′ , ~r; iω2 )iD(~r′′, ~r′ ; iω1 ; iω2 ). (10.35)
In coordinate representation, the disorder-averaged Green functions are short

ranged, see Eq. (10.33). Hence, anticipating that D will be a slowly varying
202
iω2
111
000 111
000
000
111 000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111
000
111
000
111
= 2πντ + + ... = 2πντ + 000
111
000
111
000
111
000
111 h h 000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111 000
111
iω1
Figure 64: Defining equation for the diffusion propagator.
function of ~r and ~r′ , we replace the product of single-electron Green functions

by a delta function. If ω1 ω2 > 0 (both Matsubara frequencies on the same
side of the real axis), we find
~ i
hG(~r, ~r′′; iω1 )ihG(~r′′, ~r; iω2 )i = sign (ω1 ) δ(~r − ~r′′ ), (10.36)
2πντ 2kF l
whereas
~
hG(~r, ~r′′ ; iω1 )ihG(~r′′ , ~r; iω2 )i = (1 − |ω1 − ω2 |τ )δ(~r − ~r′′ ), (10.37)
2πντ
if ω1 ω2 < 0 (iω1 and iω2 on different sides of the real axis). When iω1 and
iω2 are both on the same side of the real axis, the second term on the r.h.s.
of Eq. (10.35) can be neglected with respect to the l.h.s. of Eq. (10.35). We
thus find
D(~r, ~r′; iω1 , iω2 ) = δ(~r − ~r′ ) if ω1 ω2 > 0. (10.38)
If ω1 ω2 < 0, however, the l.h.s. and r.h.s. of Eq. (10.35) almost cancel, leaving
a contribution proportional to |ω1 −ω2 |τ only. In this case, approximating the
product of single-electron Green functions by a delta function is not correct,
and we need to take into account the spatial dependence of the diffuson
propagator,
~
hG(~r, ~r′′ ; iω1 )ihG(~r′′ , ~r; iω2 )i = (1 − |ω1 − ω2 |τ + Dτ ∇~r2 )δ(~r − ~r′′ ),
2πντ
(10.39)
where D = (1/3)vF2 τ is the diffusion constant.41 Hence, if ω1 ω2 < 0, the
diffuson propagator satisfies the diffusion equation
1
−D∇2 + |ω1 − ω2 | D(~r, ~r′ ; iω1 , iω2 ) = δ(~r − ~r′ ).

(10.40)
τ
41
Equation (10.39) is derived “under the integral sign”, considering the integral of
hG(~r, ~r′′ ; iω1 )ihG(~r′′ , ~r; iω2 )i times an arbitrary function f (~r′′ ) over ~r′′ . Expanding the
function f in a Taylor series around ~r′′ = ~r then gives the desired result.
203
11111111111111111
00000000000000000
00000000000000000
11111111111111111
00000000000000000
11111111111111111
insulator
00000000000000000
11111111111111111
00000000000000000
11111111111111111
00000000000000000
11111111111111111
00000000000000000
11111111111111111r
00000000000000000
11111111111111111
r
00000000000000000
11111111111111111
r
00000000000000000
11111111111111111
00000000000000000
11111111111111111
00000000000000000
11111111111111111 conducting
00000000000000000
11111111111111111
00000000000000000
11111111111111111 sample
00000000000000000
11111111111111111
00000000000000000
11111111111111111
Figure 65: Relative positions of a point ~r close to the sample boundary and its
mirror image ~rr . For the derivation of the boundary conditions for the diffusion
propagator, the point ~r is chosen well within a mean free path from the boundary,
but much further away than a Fermi wavelength.
For a bulk system, Fourier transform of Eq. (10.40) gives D(~q; iω1 , iω2 ) =
[τ (Dq 2 + |ω1 − ω2 |)]−1 , in agreement with the results of Sec. 10.1.
The boundary conditions for the diffuson propagator can be obtained in
a similar way. If we fix the point ~r′ to be inside the sample, the bound-
ary conditions are found by taking the correct form of the disorder averaged
single-electron Green functions at the sample boundary. Since the wavefunc-
tions have to vanish at the sample boundary, you verify that one has
hG(~r, ~r′′, iω1 )i = hG(~r, ~r′′ , iω1 )ibulk − hG(~rr , ~r′′ , iω1 )ibulk , (10.41)
where ~rr is the mirror image of ~r in the sample boundary, see Fig. 65, and
the subscript “bulk” indicates the expression for the average Green function
deep inside the sample, see Eq. (10.33) above. Since we are interested in
spatial variations of the diffusion propagator on length scales of the mean
free path l and beyond, we take the point ~r such that its distance from the
boundary is much smaller than the mean free path l but much larger than
the Fermi wavelength λF = 2π/kF . Then, using Eq. (10.33), we find
Z
~ ~ ~r )δ(~r − ~r′′ ),
d~r′′hG(~r, ~r′′ ; iω1 )ihG(~r′′, ~r; iω2 )i = (1 + DvF−1 n̂ · ∇
2πντ
(10.42)
where n̂ is the outward oriented unit vector normal to the sample boundary.
Hence, using Eq. (10.35), we conclude that, at the sample boundary, the dif-
fusion propagator D(~r, ~r′; iω1 , iω2 ) with ω2 ω1 < 0 must satisfy the boundary
204
condition
~ ~r D(~r, ~r′ ; iω1 , iω2 ) = 0.
n̂ · ∇ (10.43)
Similarly, at points where the sample is connected to a perfect conductor,

one has the boundary condition
D(~r, ~r′ ; iω1 , iω2 ) = 0. (10.44)
The boundary condition (10.43) describes the fact that, at a boundary with
an insulator, the current density is zero, whereas the boundary condition
(10.44) describes the fact that, at a boundary with a perfect conductor, the
particle accumulation must vanish.
In the presence of a vector potential A ~ that varies on a scale that is
slow compared to the mean free path, the Green function (10.33) acquires
an extra factor exp[i(eA/c~) ~ · (~r − ~r′ )]. However, upon multiplication of
′′ ′′
hG(~r, ~r ; iω1 )i and hG(~r , ~r; iω2 )i on the r.h.s. of Eq. (10.35), this extra factor
disappears from the equation for the diffuson. Similarly, an eventual slowly
varying potential V (~r) in the Hamiltonian, which causes a small change
kF → kF − V /~vF in the wavevector appearing in the exponent of the single-
electron Green function of Eq. (10.33), also drops out from the equation for
the diffusion propagator.
As a slight generalization of the previous discussion, we may discuss a dif-
fusion propagator for the case where the upper and lower lines in the diagram
of Fig. 64 correspond to Hamiltonians with the same disorder configuration,
but with a different magnetic field or with a different value of a slowly varying
background potential V (~r). In this case, the phase vectors from the vector
potentials and from the orbital motion do not cancel exactly, and one arrives
at the following equation for the diffuson if ω1 ω2 < 0,
2
~ ie ~ ~
−D ∇~r + (A1 − A2 ) D(~r, ~r′ ; iω1 , iω2 )
~c
1
± (ω1 − ω2 + iV1 − iV2 ) D(~r, ~r′ ; iω1 , iω2 ) = δ(~r − ~r′ ),
τ
(10.45)
where the + sign is taken for the case ω1 − ω2 > 0 and the − sign is taken
if ω1 − ω2 < 0. Similarly, in the boundary condition (10.43), the deriviative
~ ~r is replaced by the covariant derivative ∇
∇ ~ ~r + (ie/~c)(A
~1 − A
~ 2 ).
205
The same considerations apply to the Cooperon. One finds that the
Cooperon propagator is zero if ω1 ω2 > 0, whereas it satisfies the differential
equation
2
~ ie ~ ~
−D ∇~r + (A1 + A2 ) C(~r, ~r′; iω1 , iω2 )
~c
1
± (ω1 − ω2 + iV1 − iV2 ) C(~r, ~r′; iω1 , iω2 ) = δ(~r − ~r′ ),
τ
(10.46)
if ω1 ω2 < 0. The boundary conditions are the same as those for the diffuson
~ ~r
propagator. In the presence of a magnetic field, the differential operator ∇
~ ~r + (ie/~c)(A
is replaced by the covariant derivative ∇ ~1 + A~ 2 ). Note that,
unlike the diffuson, the cooperon propagator depends on the magnetic field,
even if A~1 = A
~ 2.
10.4 Random matrix theory

Now let us look at an isolated piece of a disordered conductor of finite size.
We are interested in the density of states in the conductor. Since the sample
has a finite size, the density of states will be a sum of Dirac delta functions,
X
N (ω) = 2 δ(ω − εµ ), (10.47)
µ
where µ labels the eigenvalues of the single-electron Hamiltonian (without

spin) and the factor two accounts for spin degeneracy.
The positions of the energy levels εµ and, hence, the density of states,
will depend on the precise impurity configuration. As before, we’ll take a
statistical approach, and calculate the probability distribution of the density
of states.
Calculation of the average is straightforward, hN i = 2νV , where V is
the volume of the sample. The more interesting question is that of the
fluctuations of the density of states, which is described by the function
R(ω1 − ω2 ) = hN (ω1)N (ω2 )i − hN (ω1)ihN (ω2)i. (10.48)
In order to calculate R, we first express the density of states in terms of the
electron Green functions,
1
Z
N (ω) = d~r[GA (~r, ~r; ω) − GR (~r, ~r; ω)]. (10.49)
πi
206
iω2 iω2
111
000 111
000
000
111 000
111
000
111 000
111
r’ 000
111 r’ 000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111 000
111
000
111 000
111
r’ 000
111
000
111
000
111
r’ 000
111
000
111
000
111
iω1 iω1
r r r r
1 2 1 2
iω1 iω1
000
111 000
111
000
111 000
111
r’’111
000
000
111
000
111
000
111
r’’111
000
000
111
000
111
000
111
111
000 111
000
000
111 000
111
000
111 000
111
000
111 000
111
000
111 000
111
r’’ 000
111 r’’ 000
111
000
111 000
111
000
111 000
111
000
111 000
111
iω2 iω2
Figure 66: Diagrams contributing to the fluctuations of the density of states in a

disordered metal grain.
With this expression, R(ω1 − ω2 ) can be calculated with the help of diagram-
matic perturbation theory. The relevant diagrams for R(ω1 − ω2 ) are shown
in Fig. 66. (Note that the diagrams without any lines connecting the inner
and outer electron lines contribute to the average of the density of states and
not to the fluctuations.)
After substitution of Eq. (10.49) into Eq. (10.48), we distinguish four
contributions to the density of states fluctuations. Performing the same
calculation as the one leading to Eqs. (10.36) and (10.37), one finds that the
contributions involving two retarded Green functions or two advanced Green
functions are smaller than the contributions involving one retarded and one
advanced Green function by a factor kF l ≫ 1. Hence, we conclude
2
Z
R(ω1 − ω2 ) = 2 Re d~r1 d~r2 hGR (~r1 , ~r1 , ω1 )GA (~r2 , ~r2 , ω2 )i. (10.50)
π
Making use of the auxiliary results
GR (~r1 , ~r′ , ω1 )GR (~r′′ , ~r1 , ω1 )GA (~r′, ~r′′ , ω2 ) = −2πiντ 2 δ(~r′ − ~r1 )δ(~r′′ − ~r1 )
GR (~r1 , ~r′ , ω1 )GR (~r′′ , ~r1 , ω2 )GA (~r′′ , ~r′, ω2 ) = −2πiντ 2 δ(~r′ − ~r1 )δ(~r′′ − ~r1 ),
(10.51)
we find that, according to the diagrams of Fig. 66,
2τ 2
Z
R(ω1 − ω2 ) = Re d~rd~r′[D RA (~r, ~r′ ; ω1 , ω2 )D RA (~r′ , ~r; ω1 , ω2 )
π2
207
+ C RA (~r, ~r′ ; ω1 , ω2)C RA (~r′ , ~r; ω1 , ω2 )]. (10.52)
We now calculate the density of states correlator at energy ω1 and vector
~ 1 and at energy ω2 and vector potential A
potential A ~ 2 . Hereto, we write the
diffuson and cooperon propagators as a sum over eigenmodes of the diffusion
equation,
2
~ ie ~ ~
−D ∇~r + (A1 − (±)A2 ) φm± (~r) = γm± φm± (~r), (10.53)
~c
where the φm are properly normalized eigenfunctions and the γm are the
corresponding eigenvalues,
1 X φm+ (~r)φm+ (~r′ )∗
D RA (~r, ~r′ ; ω1 , ω2) = , (10.54)
τ m γm+ − i(ω1 − ω2 )
1 X φm− (~r)φm− (~r′ )∗
C RA (~r, ~r′ ; ω1 , ω2) = . (10.55)
τ m γm− − i(ω1 − ω2 )
Substituting this into Eq. (10.52) we find

2 X 1 1

R(ω1 − ω2 ) = 2 Re + .
π m
(γm+ − i(ω1 − ω2 ))2 (γm− − i(ω1 − ω2 ))2
(10.56)
This result, which was first obtained by Altshuler and Shklovskii [Sov.
Phys. JETP 64, 127 (1986)], is a key result in the study of spectral statistics
in small disordered conductors. Let us look at it in more detail. First, we
consider the case in which there is no magnetic field. Then, the cooperon and
diffuson contributions are equal. Finding an exact expression for R(ω1 − ω2 )
requires solving the diffusion equation, which may be difficult in a sample
with an irregular shape. However, we can make three general statements:
First, all γm are non-negative. Second the smallest eigenvalue of the diffusion
equation is γ0 = 0. Third, the second smallest eigenvalue γ1 is of order D/L2 ,
where L is the sample size. (The length L is the longest dimension of the
sample for a sample with an anisotropic shape.) The smallest eigenvalue γ1
in the absence of a magnetic field is known as the “Thouless energy”, ETh .
Hence, if |ω1 − ω2 | ≪ ETh the sum in Eq. (10.56) is dominated by the m = 0
term, and one has
4
R(ω1 − ω2 ) = − if no magnetic field. (10.57)
π 2 (ω 1 − ω2 )
2
208
Notice that this result is completely universal: it does not depend on sample
size, sample shape, or disorder concentration.
If, on the other hand, there is a finite magnetic field, the lowest eigenvalue
γ0− for the cooperon propagator is nonzero. For small magnetic fields, one
can find γ0− using perturbation theory,
2
4De2 Φe ETh
Z
γ0− = 2 2 ~ 2
d~r|A| ∼ , (10.58)
~c hc δ
where δ = 1/νV is the mean spacing between energy levels and the vector
potential A ~ is taken in the London gauge (∇ ~ ·A~ = 0 inside the sample and
~A
n̂ · ∇ ~ = 0 on the sample boundary, where n̂ is the unit vector perpendicular
to the sample boundary). Equation (10.58) is valid as long as γ0− ≪ ETh .
Hence, for magnetic fields for which γ0− ≫ |ω1 − ω2 |, the cooperon contribu-
tion to the density of states fluctuations is suppressed, and one finds
2
R(ω1 − ω2 ) = − with magnetic field. (10.59)
π 2 (ω 1 − ω2 )
2
The results (10.57) and (10.59) are divergent if ω1 − ω2 → 0. This diver-

gence appears because of a level’s “self correlation”: Looking at the defini-
tions (10.47) and (10.48), one sees that the correlator R(ω1 − ω2 ) contains
the singular term
X
Rself (ω1 − ω2 ) = 4 [hδ(ω1 − εµ )δ(ω2 − εµ )i − hδ(ω1 − εµ )ihδ(ω2 − εµ )i]
µ
4 4
= δ(ω1 − ω2 ) − 2 . (10.60)
δ δ
The origin of the divergence in our diagrammatic calculation is that the
diagrammatic perturbation theory breaks down when the energy difference
ω1 − ω2 becomes comparable to the spacing δ = V /ν between energy levels
in the sample and thus fails to resolve the delta-function correlations of Eq.
(10.60). One way to cure this problem is to give all levels a finite width
γ, which regularizes the divergence arising from the self correlations. This
amounts to replacing Eqs. (10.57) and (10.59) by
4 1
R(ω) = 2
Re
π β (γ − iω)2
4 γ2 − ω2
= , (10.61)
π 2 β (γ 2 + ω 2)2
209
where β = 2 with a magnetic field and β = 1 without a field. Using field-
theoretic methods, Efetov has been able to calculate the exact correlation
function R(ω) down to ω = 0 [Adv. Phys. 32, 53 (1983)]. Without magnetic
field he found
Z ∞
4 sin2 (ωπ/δ) 4 ∂

sin(ωπ/δ) sin(ωx)
R(ω) = − 2 − 2 dx for ω 6= 0.
π ω2 π ∂ω ω δ/π x
(10.62)
With a magnetic field, only the first term on the r.h.s. of Eq. (10.62) is kept.
Note that Eq. (10.62) simplifies to our results (10.57) and (10.59) if |ω| ≫ δ.
One striking consequence of the correlator (10.57) and (10.59) is that
the spectrum is rigid. This means that the fluctuations of the number of
levels N(ω2 , ω1 ) in a certain energy interval ω1 < ω < ω2 increases slower
with |ω1 − ω2 | than |ω1 − ω2 |1/2 . (Fluctuations proportional to |ω1 − ω2 |1/2
are expected if the levels were distributed as uncorrelated random numbers.)
Indeed, one finds
Z ω2
var N(ω2 , ω1 ) = dωdω ′R(ω − ω ′ )
ω1
4 (ω1 − ω2 )2
∼ ln , (10.63)
π2β δ2
where we used the regularized result (10.61) with γ ∼ δ to cut off the diver-
gence of R(ω) at small ω.
The universality of the spectral fluctuations for energies below the Thou-
less energy is more general than the case of a disordered conductor of arbi-
trary shape we considered here. In fact, one finds the same spectral fluctu-
ations for the energy levels in a ballistic conductor with an irregular shape,
for the resonances in a heavy nucleus, and for the eigenvalues of a hermitian
matrix with randomly chosen elements.42 In these cases, the only relevant
input is the average spacing between levels δ and the symmetry of the sys-
tem: without a magnetic field, time-reversal symmetry is present, whereas
with a magnetic field, time-reversal symmetry is broken. For hermitian ma-
trices, the presence or absence of time-reversal symmetry corresponds to real
42
For a ballistic conductor, the Thouless energy is ETh ∼ vF /L, whereas for a random
matrix of size N one has ETh ∼ N δ. For a ballistic conductor, R(ω1 − ω2 ) for |ω1 − ω2 | ≪
vF /L does not depend on the precise sample shape as long as the shape is irregular,
whereas for a random matrix R(ω1 − ω2 ) does not depend on the distribution of the
matrix elements if |ω1 − ω2 | ≪ N δ. For a ballistic sample, the universality of spectral
statistics has the status of a conjecture, although there is extensive numerical evidence.
210
or complex matrix elements. Mathematically, random matrices are much
easier to deal with than disordered conductors. The study of eigenvalues
and eigenvectors of random matrices has grown into a field of its own, called
“random matrix theory”. In fact, the label “random matrix theory” is also
used to describe the universal aspect of spectral statistics of disordered met-
als and heavy nuclei within an energy window of width ≪ ETh . You can find
more about the mathematical aspects of the theory of random matrices in
the book Random matrices by M.L. Mehta (Academic, New York, 1991).
211
212
11 Beyond linear response
11.1 formalism
In Chapter 6 we discussed the response of a physical observable to a time-
dependent perturbation to linear order in the perturbation. The central
result of that chapter was the Kubo formula, which relates the expectation
value of an observable A at time t to the strength of the perturbation at
previous times t′ < t. As we saw in Ch. 6, the linear response can be
calculated in terms of an equilibrium correlation function.
If one wants to go beyond linear response, knowledge of equilibrium cor-
relation functions only is not sufficient. However, there is a very powerful
extension of the equilibrium formalism to the case of nonlinear response.
We’ll discuss that extension in this section.
The quantity of interest is, as always, a Green function. We start from
the greater and lesser Green functions, since all other Green functions can
be built from these, see Eq. (2.19),
G> ′ ′
A;B (t, t ) = −ihAH (t)BH (t )i, (11.1)
G< ′ ′
A;B (t, t ) = ±ihBH (t )AH (t)i, (11.2)
where the + sign is for fermions and the − sign for bosons and the operators
are taken in the Heisenberg picture (which is, for now, indicated explicitly
by the subscript H).
We write the Hamiltonian as
H = H0 + H1 , (11.3)
where the perturbation H1 includes all interactions between the particles and
time-dependent external fields and the perturbation H1 is switched on slowly
after a time t0 ≪ t, t′ . Since the system is not in thermal equilibrium at the
times t and t′ , we cannot use a direct thermal average to calculate the Green
functions G< and G> . However, it is known that the system is in thermal
equilibrium at times before the time t0 when H1 was switched on. Using the
evolution operator, we can express the operators at times t and t′ in terms
of the corresponding operators at times before t0 ,
AH (t) = U(t0 , t)A(t)U(t, t0 ). (11.4)
Here we have used the interaction picture for the operator A(t) and the
evolution operator U. In the interaction picture, the time evolution with
213
t0 t
Figure 67: Contour used to calculate the operator A(t) in the interaction picture
the Hamiltonian H0 is included in the operators, e.g., A(t) = exp(iH0 (t −

t0 )/~)A exp(−iH0 (t−t0 )/~), and excluded from the evolution operator U(t, t0 ).
The evolution operator U(t, t0 ) can be written as a time-ordered integral over
the perturbation H1 (t) (where H1 (t) is taken in the interaction picture as
well), Z t
†
U(t, t0 ) = U(t0 , t) = Tt exp −i dt′ H1 (t′ ) . (11.5)
t0
We now rewrite Eq. (11.4) as

i t dt′ H1 (t′ ) −i t dt′ H1 (t′ )
h R i h R i
AH (t) = Tt e t0 A(t) Tt e t0
−i t dt′ H1 (t′ )
h R t0 ′ ′
i h R i
= Tt e−i t dt H1 (t ) A(t) Tt e t0
dt′ H1 (t′ )
R
= Tc e−i c A(t), (11.6)
where “c” indicates integration along a contour that starts at t0 , goes through
t, and returns to t0 , see Fig. 67, and Tc implies time-ordering along the
contour.
Similarly, for a correlation function that invovles the product A(t)B(t′ )
one can use a representation in terms of a contour c that starts at t0 , goes
through both times t and t′ , and returns to t0 ,
R
hTc e−i c dt1 H1 (t1 ) A(t)B(t′ )i0
G(t, t′ ) = −i R . (11.7)
hTc e−i c dt1 H1 (t1 ) i0
Here h. . .i0 indicates a thermal average with respect to the Hamiltonian H0 .43
The arguments t and t′ refer to a point on the contour c, instead of a time on
the time axis. The contour c can be chosen arbitrariliy, as long as it starts
and ends in t0 and passes through the real times t and t′ . Direct application
of Eq. (11.4) for A(t) and a similar result for B(t′ ) would lead to the contour
of Fig. 68a if one were to calculate G< and the contour of Fig. 68b if one
43
Note that we assume that both the interactions and the perturbation are switched on
after the time t0 , so that H0 is the full Hamiltonian at time t0 .
214
t0 t t0 t’
t’ (a) t (b)
Figure 68: Contours used to calculate the greater and lesser Green functions G>
(b) and G< (a).
wants to calculate G> . The contour may be deformed, and parts that are
traversed in both directions may be canceled. You verify that one obtains
the Kubo formula if Eq. (11.7) is expanded to first order in H1 .
A convenient choice is the Keldysh contour, which starts at t0 , goes to
+∞ and then returns to t0 , see Fig. 69. The arguments t and t′ can be
chosen on the “upper” and “lower” branches of the contour, cf. Fig. 69. If
both arguments are on the upper branch, one has
G(t, t′ ) = −ihTt A(t)B(t′ )i, t, t′ upper branch, (11.8)
where Tt indicates the standard time ordering. If both arguments are on the
lower branch, G(t, t′ ) is equal to an “anti-time-ordered” Green function,
G(t, t′ ) = −iθ(t′ − t)hA(t)B(t′ )i ± iθ(t − t′ )hB(t′ )A(t)i
≡ −ihT̃t A(t)B(t′ )i, t, t′ lower branch. (11.9)
Finally, if t is on the upper branch and t′ is on the lower branch, or vice
versa, one has
G(t, t′ ) = G< (t, t′ ) if t upper branch, t′ lower branch,
G(t, t′ ) = G> (t, t′ ) if t lower branch, t′ upper branch. (11.10)
The Green function G(t, t′ ) can be represented as a 2 × 2 matrix, where the
matrix index indicates what branch of the Keldysh contour is referred to,
G11 (t, t′ ) G12 (t, t′ )

′
G(t, t ) = , (11.11)
G21 (t, t′ ) G22 (t, t′ )
where
G11 (t, t′ ) = −ihTt A(t)B(t′ )i,
G12 (t, t′ ) = G< (t, t′ ),
G21 (t, t′ ) = G> (t, t′ ),
G22 (t, t′ ) = −ihT̃t A(t)B(t′ )i. (11.12)
215
t0 t t’
Figure 69: Keldysh contour. The arguments t and t′ can be taken on each branch
of the contour.
In the literature, one usually uses a different representation of the matrix

Green function (11.11),
G = Lτ3 GL† , (11.13)
where
1 0 1 1 −1
τ3 = , L= √ . (11.14)
0 −1 2 1 1
Using Eq. (11.12) you quickly verify that
R
G (t, t′ ) GK (t, t′ )

G= , (11.15)
0 GA (t, t′ )
where GR (t, t′ ) and GA (t, t′ ) are the standard retarded and advanced Green
functions (but now calculated outside equilibrium) and
GK (t, t′ ) = G> (t, t′ ) + G< (t, t′ ) (11.16)
is the so-called Keldysh Green function. Note that the matrix form (11.15)
is preserved under matrix multiplication.
A perturbation expansion in powers of H1 can be done with the help
of Wick’s theorem for the contour-ordered products. One difference with
the imaginary-time version of the diagrammatic theory we used before is
that electron-electron interaction matrix elements and phonon Green func-
tion come with a factor i instead of a minus sign. Disconnected diagrams of
the numerator of Eq. (11.7) cancel the contribution from the denominator.
Integrations over the Keldysh contour may be replaced by integrations over
the real axis,
Z Z ∞ Z ∞
dt . . . → dt[ upper branch ] − dt[ lower branch].
c t0 t0
In the matrix notation of Eq. (11.11), this amounts to matrix multiplica-

tion at each vertex (impurity vertex, phonon vertex, or interaction vertex),
216
(a) (b)
Figure 70: First-order diagrams for impurity scattering and phonon scattering.
combined with insertion of a minus sign at the lower branch argument. For
example, the diagram of Fig. 70a for the single-electron Green function with
one impurity scattering is evaluated as
Z Z
(1)
X
′ ′ ′′
Gij (~r, t; ~r , t ) = d~r dt′′ Gik (~r, t; ~r′′ , t′′ )U(~r′′ )σk Gkj (~r, t′′ , ~r′, t′ ),
k
where σ1 = 1 and σ2 = −1. Similarly, the phonon diagram of Fig. 70b reads
Z Z
(1)
X
′ ′ 2
Gij (~r, t; ~r , t ) = g d~r1 d~r2 dt1 dt2 Gik (~r, t; ~r1 , t′′ )σk
kl
× (iDkl (~r1 , t1 ; ~r2 , t2 ))Gkl (~r1 , t1 ; ~r2 , t2 )σl Glj (~r2 , t2 ; ~r′ , t′ )
where Dkl (~r, t; ~r′ , t′ ) is the matrix phonon Green function and the electron-
phonon coupling constant gq~ in the electron-phonon Hamiltonian (9.5) is
chosen independent of ~q.
Upon transformation to the matrix representation of Eq. (11.15) those
diagrammatic rules are somewhat modified. Performing the transforma-
tion (11.13), one finds that an impurity vertex carries no sign, whereas the
electron-phonon or Coulomb vertices are modified as in Fig. 71, where
1 1
γij1 = γ̃ij2 = √ δij , γij2 = γ̃ij1 = √ (τ1 )ij , (11.17)
2 2
where τ1 is the Pauli matrix. Note that, unlike in the equilibrium diagrams
we considered previously and unlike in the matrix language of Eq. (11.11),
phonon and interaction lines need to have a direction. However, the results
do not depend on the direction of the assigned direction and the direction of
the phonon or interaction lines has no physical meaning.
For future use, we now write down the Dyson equation for the single-
electron Green function. As before, calling all diagrams that can be cut in
217
i i
k γ
k ijk
j absorption j
i i
k ~
γ
k ijk
j emission j
Figure 71: Diagrammatic rules for electron-phonon and Coulomb vertices.
two by deleting a single fermion line “reducible”, we can write

Z Z X
′ ′ ′ ′
Gij (~r, t; ~r , t ) = (G0 )ij (~r, t; ~r , t ) + dt1 dt2 d~r1 d~r2 (G0 )ik (~r, t; ~r1 , t1 )
kl
′ ′
× Σkl (~r1 , t1 ; ~r2 , t2 )Glj (~r2 , t2 ; ~r , t ), (11.18)
see Fig. 72. Here G0 is matrix Green function for the Hamiltonian H0 and Σ
is the self energy, the sum of all “irreducible” diagrams (excluding external
fermion lines, but including the matrices γ and γ̃ if necessary). The self
energy has matrix structure
R
Σ (~r, t; ~r′ , t′ ) ΣK (~r, t; ~r′, t′ )

Σ= . (11.19)
0 ΣA (~r, t; ~r′, t′ )
Using the operator language for the dependence on the coordinates ~r, ~r′ and
the times t and t′ and matrix representation, we can rewrite the solution of
the Dyson equation as −1
G = G−1

0 −Σ . (11.20)
In the operator language, the inverse Green function G−1
0 is equal to

∂
G−1
0 = i − H0 1, (11.21)
∂t
where H0 is the unperturbed Hamiltonian in first quantization form and 1 is
the 2 × 2 unit matrix. Note that adding an imaginary infinitesimal ±iη to
218
1111
0000 1111
0000 1111
0000
0000
1111 0000
1111 0000
1111
= + 0000
1111 + 0000
1111 0000
1111 + ...
0000
1111 0000
1111 0000
1111
0000
1111
0000
1111 0000
1111
0000
1111 0000
1111
0000
1111
0000
1111 0000
1111 0000
1111
0000
1111 0000
1111 0000
1111
0000
1111
1111
0000
= + 0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
0000
1111
Figure 72: Diagrammatic representation of the Dyson equation for the matrix
Green function G. The shaded circle represents the self energy Σ and the single
line represents the unperturbed Green function G0 .
G−1
0 does not make a difference here, hence there is no need to distinguish
retarded and advanced components. Alternatively, using operator language
for the dependence on ~r and t, the Dyson equation (11.20) can be rewritten
as
(G−1 −1
0 − Σ)G = 1 = G(G0 − Σ). (11.22)
In order to be a useful starting point for calculations, Eq. (11.20) should
be supplemented with a further theory or approximation scheme for the self
energy
As all single-electron Green functions, the matrix Green function G can
be chosen to depend on two positions ~r and ~r′ , on two momenta ~k and ~k ′ ,
or on the quantum numbers µ and µ′ of any other complete set of single-
electron states. The above diagrammatic rules do not depend on the choice
of the representation as long as the first and second arguments of the Green
functions are not mixed. For some applications, it is favorable to use a
“mixed” representation in which the Green function G is chosen to depend
on sum and difference coordinates, followed by a Fourier transform to the
difference coordinate,
Z Z
~ ~
~
G(R, T ; k, ω) = ~ + ~r/2, T + t/2; R
d~r dte−ik·~r+iωt G(R ~ − ~r/2, T − t/2).
(11.23)
~ and T as “center coordinate” and “center time”, and to the ~k
We’ll refer to R
and ω as momentum and frequency. The “mixed representation” or “Wigner
representation” is particularly useful in a semiclassical approximation, when
electrons are assigned both a momentum and a position. However, in the
mixed representation the convolution of Green functions is changed. Using
Eq. (11.23) and its inverse to express the operator product [G1 G2 ] in the
219
Wigner representation in terms of the Wigner representation of the two fac-
tors G1 and G2, one finds that the convolution of Green functions G1 and
G2 is given by

i ∂1 ∂2 ∂ ∂ ∂ ∂ ∂ ∂
− ∂ 1T ∂ 2ω − 2~ 1~ + ∂ 2T ∂ 1ω
~ T ; ~k, ω) = e
[G1 G2 ](R, 2 ~ ∂ ~
∂1 R 2k 1 2 ∂2 R ∂1 k 2 1 ~ T ; ~k, ω)
G1 (R,
~ T ; ~k, ω),
× G2 (R, (11.24)
where ∂1 refers to a derivative with respect to an argument of the first Green

function and ∂2 refers to a derivative with respect to an argument of the
second Green function.
In general, the advanced and retarded Green function contain the infor-
mation on the available states (appropriately modified for the nonequilibrium
state of the system at times t and t′ ), whereas the Keldysh Green function
contains information on the (nonequilibrium) occupation of these states. In-
deed, in equilibrium, the greater and lesser Green functions are related to the
spectral density via Eqs. (2.46) and (2.47). This relation implies a relation
between the Keldysh component and the advanced/retarded components of
the matrix Green function G,
GK R A

0 = tanh(ω/2T ) G0 − G0 . (11.25)
A similar relation holds for the self energy in equilibrium. (Note that, in
equilibrium, the Green functions do not depend on the “center time” T , but
they may depend on both the center coordinate R ~ and the momentum ~k.)
In the mixed representation, the unperturbed Hamiltonian H0 takes a
particularly simple form,
~
H0 = ε~k + U(R), (11.26)
where U is the potential.
11.2 Boltzmann equation

The Boltzmann equation, which was used in Sec. 5.4 as an alternative method
to derive the Drude formula for the conductivity of a normal metal, is an ex-
ample of a “kinetic equation”. In general, a “kinetic equation” is an equation
that describes the time-dependence of the occupation of quantum mechan-
ical states. A time-evolution equation of the Keldysh Green function is an
example of a quantum kinetic equation, since the Keldysh Green function
contains information about the quantum states and their occupation.
220
Physical observables are expressed in terms of the Green functions. Ex-
amples are the electron charge density
ρe (~r, t) = −eG< (~r, t; ~r, t), (11.27)
or the current density

~je (~r, t) = − ~e ~ ~r − ∇
~ ~r′ )G (~r, t; ~r , t)
< ′

(∇
2im ′
~
r →~
r
2
e ~
− A(~ r)G< (~r, t; ~r, t), (11.28)
m
where A~ is the vector potential and a summation over spin degrees of freedom
is implied. In linear response, one can find the charge and current densities
from the Kubo formula. Beyond linear response, one needs a solution of the
Dyson equation for the matrix Green function G. Here we’ll describe how
such a solution can be shown to be equivalent to a solution of the Boltzmann
equation.
Starting point of our discussion is the Dyson equation (11.22) in which
we subtract the far left and far right sides. In operator language, one thus
finds
[G−1
0 − Σ, G]− = 0, (11.29)
where [·, ·]± denotes the commutator (− sign) or anticommutator (+ sign).
The Keldysh component of this equation reads
G−1 K K −1 R K K A K A R K
0 G − G G0 − Σ G + G Σ − Σ G + G Σ = 0. (11.30)
We rewrite this result in terms of commutators and anticommutators,

i K i
[G−1 K K K
0 − ℜΣ, G ]− − [Σ , ℜG]− = [Σ , A]+ − [Γ, G ]+ , (11.31)
2 2
where A = i(GR − GA ) is the spectral weight of the full Green function and
Γ = i(ΣR − ΣA ),
1 R
ℜΣ = (Σ + ΣA ),
2
1 R
ℜG = (G + GA ). (11.32)
2
221
The definition of “real” and “imaginary” parts is the same one as used in
Chapter 2. Note that in equilibrium one has GK = −i tanh(ω/2T )A and
ΣK = −i tanh(ω/2T )Γ, so that the right hand side of Eq. (11.31) vanishes.
We can make further progress if we assume that the Green functions in-
volved are slowly varying functions of their arguments. Then, in taking the
Green function products according to the rule (11.24) we can expand the
exponential and truncate the expansion after first order. For an anticommu-
tator, the derivatives cancel, and one finds
[G1 , G2 ]+ = 2G1 G2 + higher order gradient terms. (11.33)
Similarly, one finds that a commutator becomes equal to a Poisson bracket,
−i[G1 , G2 ]− = [G1 , G2 ]Poisson (11.34)

∂1 ∂2 ∂1 ∂2 ∂2 ∂1 ∂2 ∂1
= − − + G1 G2 .
~ ~ ~ ~
∂1 R ∂2 k ∂1 T ∂2 ω ∂2 R ∂1 k ∂2 T ∂1 ω
This approximation is known as the “gradient expansion”.
Without further knowledge of the system under consideration this is as
far as one can go. We’ll now specialize to the case of a system with impurity
scattering but without interactions between the electrons. Taking the case
in which a slowly varying external potential Uext may be present, we have,
in the Wigner representation,
G−1 ~
0 = ω − (ε~k − µ) − Uext (R, T ). (11.35)
Using the self-consistent Born approximation for the self energy, we find,
in the Wigner representation,
~ T ; ω, ~k) = Nimp ~ T ; ω, ~k ′),

X
Σ(R, |u~k−~k′ |2 G(R, (11.36)
V ′ ~k
where u~k is the Fourier transform of the impurity potential of a single impu-
rity and Nimp is the number of impurities. In the dilute impurity limit and
for slowly varying external potentials, we can use the gradient expansion,
neglect the R~ and T dependence of the self-energy, and drop the self energy
from the l.h.s. of Eq. (11.31),
[G−1 K K K
0 , G ]Poisson = Σ A − ΓG . (11.37)
222
Also, in the dilute impurity limit and for slowly varying external potentials,
the spectral density A approaches a delta function,
A = 2πδ[ω − (ε~k − µ) − U]. (11.38)
As a result of the delta-function character of A, the Keldysh Green function
GK is also strongly peaked as a function of ω. Hence, we write
GK = −2πiδ[ω − (ε~k − µ) − U](1 − 2f~k ), (11.39)
where
1 1
Z
~
f~k (R, T ) = − dωGK (R, ~ T ; ω, ~k). (11.40)
2 4πi
Then, writing the Poisson bracket explicitly in Eq. (11.37), integrating over
ω, and performing one partial integration, we find

∂ ~ ~ −∇ ~ ~ Uext · ∇ ~ ~ f~ (R, ~ T)
+ ~v~k · ∇ R R k k
∂T
Nimp X
= −2π |u~k−~k′ |2 δ(ε~k − ε~k′ )(f~k − f~k′ ), (11.41)
V
~k
′
where ~v~k = ∇~ ~ ε~ .
k k
Equation (11.41) is the Boltzmann equation. In Sec. 5.4 you used it to
derive the Drude formula for the conductivity of a normal metal. Hence, the
present discussion, together with that of Sec. 5.4 can be seen as an alternative
derivation of that result.
The function f~k that appears in the Boltzmann equation plays the role of
an “occupation factor”. Note that it contains an integration over frequency ω,
so that f~k is related to an equal-time Green function. Using the relationship
G< = (GK + iA)/2, together with Eqs. (11.38) and (11.39) and Eqs. (11.27)
and (11.28), one quickly recovers the well-known formulas for charge density
and current density,
~ T) = − e ~ T ),
X
ρe (R, f~k (R, (11.42)
V
~k
~ T) = −
~je (R, e X
~ T ).
~v~k f~k (R, (11.43)
V
~k
Also note that, by Eqs. (11.25), (11.38) and (11.40), one has f~k = [1 +
~ and T .
exp((ε~k − µ)/T )]−1 in equilibrium, indepenent of R
223
The Boltzmann equation is only a simple example of a kinetic equa-
tion. Generalizations of the Boltzmann equation have been derived in the
Keldysh formalism for electron-phonon interactions, for weak localization
and electron-electron corrections to the conductivity of a normal metal, and
for quasiparticle dynamics in superconductors. For more details, you are re-
ferred to the review “Quantum field-theoretical methods in transport theory
of metals”, by J. Rammer and H. Smith, Rev. Mod. Phys. 58, 232 (1986).
224
12 Electrons in one dimension
12.1 Why is one dimension different?
With the exception of a short discussion in chapter 10 we have studied elec-
trons in three dimensions. We encountered two possible excitations of an
electron liquid in three dimensions: particle-hole pairs and plasma modes.
These excitations appeared as the support of the imaginary part of the po-
larizability and susceptibility at finite frequency ω and finite wavevector ~q.
The plasma modes have a high threshold frequency for excitation, leaving
the particle-hole pairs as the main type of excitations at long wavelengths
and low frequencies. This observation, together with the observation from
Fermi liquid theory that particle-hole pairs are long-lived excitations if their
energy is low, is the main justification of the use of the non-interacting elec-
tron picture as the starting point for our description of the three-dimensional
electron liquid.
In one dimension, the situation is completely different. Here, with “one di-
mension” we mean a truly one dimensional system, i.e., a quantum wire with
a width a so small that only one transverse mode is populated at the Fermi
energy. In practice, this means a ∼ λF , where λF is the Fermi wavelength.44
In order to see why the one-dimensional electron liquid is so different, we re-
examine the polarizability and spin susceptibility in a one-dimensional wire.
Before we look at the polarizability, we need to know the Coulomb inter-
action in one dimension. For a wire of thickness a in vacuum, the Coulomb
interaction is
e2 e−iqx
Z
Vq~ = dx
4πǫ0 |x|
2
e 1
≈ ln , (12.1)
2πǫ0 qa
which is valid for qa ≪ 1. We have cut off the integral at x ∼ a, since for
x . a a three-dimensional form of the Coulomb interaction should be used.
If interactions in the one-dimensional wire are screened by a nearby metal at
distance b, the interaction Vq~ saturates at q ∼ 1/b.
44
This use of the word “one dimensional” is different from that of Chapter 10, where
“one dimensional” referred to a system that is much longer than it is wide, but that does
not need to be as narrow as λF . In order to distinguish the two cases, the latter case is
sometimes referred to as “quasi one dimensional”.
225
In the random phase approximation, the polarizability reads
χ0R
e (~q, ω)
χe (~q, ω)RPA = , (12.2)
1 − Vq~χ0R
e (~q, ω)/e2
where Vq~ is given by Eq. (12.1) above and
e2 X tanh[(ε~k+~q − µ)/2T ] − tanh[(ε~k − µ)/2T ]

χ0R
e (~
q , ω) = (12.3)
V ε~k − ε~k+~q + ω + iη
~k
is the polarizability of the one-dimensional non-interacting electron gas. In

one dimension, the ~k summation in Eq. (12.3) is easily done: the momentum
~k is represented by a real number k, and one has εk+q = εk + qvF sign k.
Then, replacing the summation over k by an integration over εk , one has
e2 ν

0R 2qvF 2qvF
χe (~q, ω) = −
2 ω + iη − qvF ω + iη + qvF
2 2
2e q vF
= , (12.4)
π[(ω + iη)2 − q 2 vF2 ]
where we used the fact that, in one dimension, the density of states ν is
related to the Fermi velocity vF as ν = 1/πvF . Substituting this into Eq.
(12.2), one finds
2e2 q 2 vF
χe (~q, ω)RPA = . (12.5)
π(ω + iη)2 − πq 2 vF2 − (e2 /πǫ0 ) ln(1/qa)q 2 vF
The one-dimensional electron liquid can dissipate energy only if the imag-
inary part of χe (~q, ω) is nonzero. Inspection of Eq. (12.5) shows that that is
the case for
ω = qvc , (12.6)
where the velocity vc is
s
e2 1
vc = vF 1+ ln . (12.7)
π 2 vF ǫ0 qa
In the presence of a metal gate at distance b from the wire that screenes the
Coulomb interaction inside the wire, the momentum q in Eq. (12.7) should
be replaced by 1/b.
226
111111111111111111111
000000000000000000000
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
particle−hole
000000000000000000000
111111111111111111111
00000000000
11111111111
4 000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
plasmon excitations
000000000000000000000
111111111111111111111
00000000000
11111111111
ω/ε F 000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
modes 00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
2 000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
spin modes
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
000000000000000000000
111111111111111111111
00000000000
11111111111
0 000000000000000000000
111111111111111111111
00000000000
11111111111
0 1 2 3 4
q/k F
Figure 73: Support of the polarizability χe and the transverse spin susceptibility
χ−+ for a one-dimensional interacting electron liquid in the RPA approximation.
The result (12.6) describes the one-dimensional version of plasma oscil-

lations. Unlike in three dimensions, where the plasma frequency remained
finite (and large) if the wavevector q → 0, in one dimension ω → 0 if q → 0.
However, the plasmon speed is larger than the Fermi velocity, since the plas-
mon propagation speed is enhanced by the repulsive Coulomb interactions.
Hence, in one dimension, plasmons are relevant perturbations in the low
frequency, long wavelength limit.
However, this is not the entire story. The result (12.5) differs in one
more aspect from its three dimensional counterpart: There is no particle-
hole continuum for excitations, not even a single particle-hole branch! Note
that at ω = ±qvF the polarizability of Eq. (12.5) is well behaved and has no
imaginary part. Hence, we conclude that the only charged excitations in a
one dimensional electron gas are the plasmon collective modes.
The situation is different if the wavevector q is of order kF , or if the
temperature or frequency are of order εF . In that case, one cannot ap-
proximate the spectrum by a linear dispersion. Without interactions, the
polarizibility function χ0R
e (q, ω) becomes a truly complex function of q and
ω for qvF − q /2m < ω < qvF + q 2 /2m, just like in three dimensions, and,
2
hence, the full polarizability χRPA

e becomes complex as well for that wavevec-
tor and frequency range. The support of Im χe for the entire frequency range
is shown in Fig. 73.
If we want to investigate spin excitations, we should look at the spin sus-
227
ceptibility. Taking the Hubbard-model result for the transverse susceptibility
χ−+ (~q, ω),
R χ0R
−+ (~
q , ω)
χ−+ (~q, ω) = , (12.8)
1 − (2U/µB g)χ0R −+ (~
q , ω)
where
µG g X tanh[(ε~k+~q − µ)/2T ] − tanh[(ε~k − µ)/2T ]
χ0R q , ω) = −
−+ (~
4V ε~k − ε~k+~q + ω + iη
~k
µB gq 2 vF
= − (12.9)
2π[(ω + iη)2 − q 2 vF2 ]
is the transverse susceptibility of the non-interacting electron gas, we find

that the imaginary part of χ−+ (~q, ω) is nonzero if and only if
ω = vs q, (12.10)
where p
vs = vF 1 − U/πvF . (12.11)
The spin excitations are collective excitations as well, moving at a speed
that is below the Fermi velocity. The reason why vs is smaller than vF is
that the propagating of spin excitations is slowed down by the ferromagnetic
exchange interaction.45 At the Stoner instability, Uν = 1, the propagation
speed has come to zero, and large-scale ferromagnetic fluctuations become
possible. Note that, again, χ−+ is purely real at ω = vF q.
The phenomenon that, in general, vs is different from vc is known as
“spin-charge” separation. In a one-dimensional electron liquid, excitations
with spin move at a different speed than excitations with charge. In view of
what we just discussed, this is not a big surprise. Since, in one dimension, the
only possible excitations are collective modes, there is no reason to expect
that spin modes and charge modes have the same velocity. You may recall
the calculations of the zero sound velocity in a Fermi liquid, where spin and
density modes depend on different Fermi-liquid constants and, hence, have
different propagation velocities.
45
In three dimensions, the fact that vs < vF implies that spin excitations are strongly
damped because they can decay into particle-hole pairs (Landau damping). In one dimen-
sion, there are no particle-hole excitations, so that spin excitations are long-lived in spite
of the fact that their propagation speed is below the Fermi veloicty.
228
Quantitatively, our observations have been based on the random phase
approximation. Going beyond the random phase approximation will most
certainly change our estimates for the velocities vc and vs of collective charge
and spin excitations in the one-dimensional electron system, but not the qual-
ititative conclusion that the collective modes are the only long-wavelength
excitations of the one-dimensional electron system. We’ll see this in Sec.
12.3, where we start from a microscopic model for electrons in one dimen-
sion and show rigorously that all dynamics is collective. Another way to see
this is to note the similarity between electrons in one dimension and ions in
a lattice. The similarity arises because, in one dimension, electrons cannot
pass each other. In that sense, a description in terms of a “solid” would
be more appropriate than a description in terms of a “liquid”.46 Indeed,
all long-wavelength dynamics of lattice ions is collective; it is only at short
wavelengths that the properties of individual ions become important.
We conclude that, for long wavelengths and low frequencies, all excita-
tions of the one-dimensional metal are collective modes; there are no dissi-
pative modes at |ω/q| = vF , which is the frequency-wavevector relation one
expects for particle-hole excitations in a degenerate one-dimensional Fermi
gas. This observation makes us wonder whether a better description than
a description in terms of particles is possible. Can one make a theory that
uses the collective modes as the building blocks, rather than the individual
fermions?
Another reason to look for a theory that does not start from a system of
non-interacting fermions is that, in one dimension, the quasiparticle lifetime
scales inversely proportional to the excitation enery of the quasiparticle. This
is different from three dimensions, where the life time is inversely proportional
to the square of the excitation energy. In three dimensions, this dependence
was the basis for “Fermi Liquid theory”, the statement that a picture based
on non-interacting electrons is a good starting point. In one dimension,
quasiparticle decay is much faster, and Fermi Liquid theory is not valid.
What description one used instead follows from the above considerations, as
we’ll see in the next sections.
46
The difference between a “solid” and a “liquid” has to do with the presence of trans-
verse rigidity (shear). In one dimension, there is no transverse direction, so that this formal
difference between “solid” and “liquid” disappears. We usually denote the one-dimensional
electron system as a “liquid”, because of the lack of long-range order. However, as we see
from the present discussion, one-dimensional electron systems have characteristics of both
solids and liquids.
229
12.2 Effective Hamiltonian for a one-dimensional elec-
tron liquid
Based on the considerations of the previous section, we now write down an
effective Hamiltonian that has the same (collective) excitations as the one-
dimensional electron liquid. For simplicity, we consider the case of spinless
fermions, so that the only collective excitations are the plasmon modes. Also,
we assume that the long-range part of the Coulomb interaction in the wire
is screened by a nearby piece of metal, and, hence, take the plasmon velocity
vc to be independent of q.
With these observations, we are tempted to write down the following
Hamiltonian as an effective Hamiltonian for the one-dimensional electron
liquid, X
H= ωq (a†q aq + 1/2), (12.12)
q
where the summation over the wavenumber q extends over both positive and
negative q, with ωq = vc |q|, and where a†q and aq are boson creation and
annihilation operators that obey commutation relations,
h i h i
aq , aq′ = a†q , a†q′ = 0
h i− −
†
aq , aq′ = δq,q′ .
−
As we discussed above, the Hamiltonian (12.12) is valid for long wavelengths

and low frequencies only. At high wavenumbers |q| ∼ kF , particle-hole exci-
tations become important, and our description in terms of collective modes
only ceases to be valid.
Although there is nothing wrong with the Hamiltonian (12.12) as an
effective Hamiltonian for the spinless one-dimensional electron liquid, we can
gain considerably more insight if we replace the creation and annihilation
operators a† and a by “momentum” and “displacement” operators Π and φ,
where
s
~
φq = (aq + a†−q ),
2Kωq
r
K~ωq †
Πq = iaq − ia−q . (12.13)
2
230
We will use the freedom of the arbitrary constant K later in order to give
a physical interpretation of the operators Π and φ. You verify that these
“momentum” and “displacement” operators satisfy the usual commutation
rules for canonically conjugate variables,
[Πq , Πq′ ]− = [φq , φq′ ]− = 0

[Πq , φq′ ]− = −i~δqq′ . (12.14)
In terms of the new variables Π and φ we write the Hamiltonian as

X 1 1

2
H= Πq Π−q + Kωq φq φ−q . (12.15)
q
2K 2
Next, we perform a Fourier transform to real space,

r
1X
Π(x) = Πq e−iqx ,
L q
r
1X
φ(x) = φq eiqx . (12.16)
L q
Since the effective Hamiltonian (12.12) is valid for low q only, the summation
over q should be restricted to q ≪ kF only. This means that delta functions
and divergencies appearing in the real-space formulation should be cut off at
distances x ∼ 1/kF . In the real-space formulation, the effective Hamiltonian
becomes
1 1
Z
2 2 2
H = dx Π(x) + Kvc (∂x φ(x)) . (12.17)
2K 2
The commutation relations of the fields φ(x) and Π(x) are
[φ(x), φ(x′ )]− = [Π(x), Π(x′ )]− = 0

[Π(x), φ(x′ )]− = −i~δ(x − x′ ). (12.18)
Let us now calculate time-derivative of φ(x),

∂ i 1
φ(x) = [H, φ(x)]− = Π(x). (12.19)
∂t ~ K
We want the operators Π and φ to be related to the electron current density
and particle density, respectively. If such an identification is to hold, Eq.
231
(12.19) should represent the continuity equation ∂t n = −∂x j. Taking a
derivative to x on both sides of the equation, we see that such an identification
holds, if we identify the (excess) particle density n(x) with the x-derivative
of φ(x) and the current with Π(x). In fact, the following identifications are
made,
1
n(x) = ∂x φ(x), (12.20)
π
j(x) = −vF Π(x), (12.21)
which implies K = 1/πvF . However, different identifications are possible,

depending on the microscopic details of the one-dimensional electron liquid.
Combining everything, we see that we have arrived at the effective Hamil-
tonian " 2 #
π2

∂φ(x)
Z
~vF 2 2
H= dx 2 Π(x) + g , (12.22)
2π ~ ∂x
where we introduced the dimensionless parameter
vc
g= . (12.23)
vF
You may recognize Eq. (12.22) as the Hamiltonian of an elastic string, where
φ is the displacement and Π is the momentum density. The Hamiltonian
(12.22) is quite different from the Hamiltonian you would write down for non-
interacting fermions in one dimension. It reflects the fact that all excitations
are collective and, hence, bosonic, rather that quasi-particle like. In order
to stress the difference of the collective dynamics of the one-dimensional
electron liquid and the quasi-particle dynamics of the Fermi liquid in higher
dimensions, the former case is referred to as “Luttinger Liquid”.
The Hamiltonian (12.22) is a great starting point if one wants to study
the excitation spectrum and thermodynamic quantities of a one-dimensional
spinless electron liquid. One example is the calculation of the compressibility
κ, which is the derivative of the particle density to the chemical potential
µ = ∂E/∂n, where E is the energy of the electron liquid,
2 −1
∂n ∂ E
κ= = . (12.24)
∂µ ∂n2
According to the effective Hamiltonian (12.22), one has κ = vF /πvc2 . Without
interactions, when vc = vF , ths simplifies to κ = 1/πvF = ν, ν being the
232
density of states. With interactions, κ is decreased by a factor g 2 = (vc /vF )2 ,
reflecting the increased energy cost for addition of charge. Similarly, you
verify that the specific heat is decreased by a factor 1/g 2 with respect to the
non-interacting electron case.
For some applications it is important to be able to make reference to
individual electrons. An example is a tunneling experiment, where electrons
tunnel into a one-dimensional electron liquid from a weakly coupled elec-
trode. Since we employ a long-wavelength description, we will not attempt
to describe a process in which an electron that is created or annihilated is lo-
calized within a wire segment of length comparable to the Fermi wavelength.
Instead, we’ll want to write down an operator that creates an electron that
is delocalized over a piece of wire of length λ ∼ λF . This spatial “smear-
ing” amounts to a momentum cut off factor exp(−|q|λ/2) in Eq. (12.16) or,
equivalently, replacing the fields φ and Π by
2λ
Z
φ(x) → dx′ φ(x′ ),
π[λ + 4(x − x′ )2 ]
2
2λ
Z
Π(x) → dx′ Π(x′ ). (12.25)
π[λ + 4(x − x′ )2 ]
2
On the other hand, since we give up spatial resolution on length scales below
λF , we can specify the momentum of the electron to within kF , i.e., we
can specify whether the electron moves left (L) or right (R). The creation
operator of a right moving electron at position x then becomes47
1
ψR† (x) = √ UR† e−ikF x+iφ(x)−iθ(x) . (12.26)
2πλ
Similarly, for a left-moving particle one has
1
ψL† (x) = √ UL† eikF x−iφ(x)−iθ(x) . (12.27)
2πλ
Here the auxiliary field θ(x) is defined as
π
Z
θ(x) = dx′ Π(x′ )sign (x − x′ ). (12.28)
2~
47
In the next section, where we give a more formal derivation of the Hamiltonian (12.22),
you can find more details of the formal relationship between the fields φ(x) and Π(x) and
the underlying electron creation and annihilation operators ψ † and ψ.
233
φ
x x’
Figure 74: Upon creation of an electron at position x, the field φ acquires a kink,
corresponding to a peak in the excess density ρ.
You verify that θ commutes with itself, whereas the commutator with φ is
given by
π
[θ(x), φ(x′ )]− = [φ(x), θ(x′ )]− = −i sign (x − x′ ). (12.29)
2
In these equations, the fast exponential factor exp(±ikF x) corresponds to
the orbital phase of left and righ moving electrons. The operator U † is
known as the “Klein factor”. It is a formal operator that increases the total
number of left or right moving particles by one. It is necessary, because the
operators φ(x) and Π(x) conserve the total number of particles. The operator
exp[−iθ(x)] produces a shift of the field φ(x′ ) by −π/2 at x′ < x and by π/2
at x′ > x. Hence, the derivative ∂x φ is peaked near x, corresponding to
an excess particle density at that point, see Fig. 74.48 The operator eiφ(x)
changes sign each time φ is increased by π, i.e., each time a particle passes
through the point x. This property enforces the fermion anticommutation
rules for the ψ-operators. Furthermore, e±iφ(x) displaces the field Π(x) by one,
corresponding to a current density ∓vF δ(x). Finally, the prefactor (2πλ)−1/2
is chosen in conjunction with the regularization (“smearing”) procecure for
the fields φ and Π. It is chosen such that a calculation of the fermion Green
function for non-interacting electrons gives the same result in boson and in
fermion language.
48
This statement does not conflict with the fact that the operator φ conserves particle
number. At the upper and lower ends of the wire there is an “antikink” in φ, corresponding
to a decrease of the particle number by one (in total). It is the operator U † that keeps
track of the total particle number.
234
The information of the boson fields Π and φ is contained in their Green
functions. The calculation of these Green functions is quite similar to that
of the phonon Green functions of Ch. 3. For technical reasons, we calculate
Green functions for the fields φ and θ. In terms of the fields θ and φ, the
Hamiltonian reads
" 2 2 #
∂θ(x) ∂φ(x)
Z
~vF
H= dx + g2 . (12.30)
2π ∂x ∂x
In order to calculate the Green functions of the fields θ and φ, we use the
equation of motion approach. Hereto we need the imaginary time evolution,
∂ ∂φ(x, τ )
θ(x, τ ) = −ivF g 2 ,
∂τ ∂x
∂ ∂θ(x, τ )
φ(x, τ ) = −ivF . (12.31)
∂τ ∂x
We define the temperature Green functions
Dθθ (x, τ ) = −hTτ θ(x, τ )θ(0, 0)i,

Dθφ (x, τ ) = −hTτ θ(x, τ )φ(0, 0)i,
Dφφ (x, τ ) = −hTτ φ(x, τ )φ(0, 0)i,
Dφθ (x, τ ) = −hTτ φ(x, τ )θ(0, 0)i. (12.32)
Using Eq. (12.31) we find that these Green functions satisfy the equations of
motion
∂ ∂Dφθ (x, τ )
Dθθ (x, τ ) = −ivF g 2 ,
∂τ ∂x
∂ π ∂Dφφ (x, τ )
Dθφ (x, τ ) = i δ(τ ) sign (x) − ivF g 2 ,
∂τ 2 ∂x
∂ π ∂Dθθ (x, τ )
Dφθ (x, τ ) = i δ(τ ) sign (x) − ivF ,
∂τ 2 ∂x
∂ ∂Dθφ (x, τ )
Dφφ (x, τ ) = −ivF . (12.33)
∂τ ∂x
The solution to these equations is easily found by inspection,
Dθφ (x, τ ) = Dφθ (x, τ )

1
= ln sin[πT (τ + ix/vc )sign (τ )]
4
235
1
− ln sin[πT (τ − ix/vc )sign (τ )],
4
vF vc
Dθθ (x, τ ) = Dφφ (x, τ )
vc vF
1
= ln sin[πT (τ + ix/vc )sign (τ )]
4
1
+ ln sin[πT (τ − ix/vc )sign (τ )]. (12.34)
4
Regularization amounts to smearing of the boson fields by a Lorentzian factor
2λ/[π(λ2 + 4(x − x′ )2 ], see Eq. (12.25) above. In the Green functions such a
smearing amounts to the substitution x → x ± iλsign (τ ). This corresponds
to the addition of λπT /vF to the argument of the sine function, so that the
divergence of the Green functions at x → 0 and τ → 0 is cut off at distances
of order λ and times of order λ/vF .
As an example, let us now calculate a fermion Green function using the
boson language. We consider the temperature Green function for right-
moving electrons,
GRR (x, τ ) = −hTτ ψR (x, τ )ψR† (0, 0)i. (12.35)
Using Eq. (12.26), we write this as
eikF x
hTτ UR (τ )UR† (0)i Tτ e−iφ(x,τ )+iθ(x,τ )+iφR (0,0)−iθ(0,0) .

GRR (x, τ ) = −
2πλ
(12.36)
Since the total Green function needs to be antiperiodic in τ , whereas the
boson part of the Green function is manifestly periodic in τ , we require that
the time-ordering for the Klein factors UR and UR† is that of fermions. Since
the operators UR and UR† are simply ladder operators, we have
hTτ UR (τ )UR† (0)i = sign (τ ). (12.37)
Calculating the boson part of the Green function is easier than it seems
at first sight. Since the Hamiltonian of the boson fields is quadratic, we can
use the cumulant expansion, which, for hφ(x, τ )i = hθ(x, τ )i = 0, reads

−iφ(x,τ )+iθ(x,τ )+iφ(0,0)−iθ(0,0) 1 2
Tτ e = e 2 h(iφ(0,0)−iθ(0,0)−iφ(x,τ )+iθ(x,τ )) i
= eDφφ (0,0)−Dφθ (0,0)−Dθφ (0,0)+Dθθ (0,0)
× e−Dφφ (x,τ )+Dφθ (x,τ )+Dθφ (x,τ )−Dθθ (x,τ ) .
(12.38)
236
Without regularization, the equal-time and equal-position Green functions
Dφφ (0, 0) + Dθθ (0, 0) are divergent (and negative). With regularization, this
divergence is cut off at distances ∼ λ (see above). Substituting Eq. (12.34)
and putting everything together, we find
ν/2−1/2
1 T /2vc
GRR (x, τ ) = − sign (τ )
(2πλ)1−ν sin[πT (τ + ix/vc )sign (τ )]
ν/2+1/2
T /2vc
× , (12.39)
sin[πT (τ − ix/vc )sign (τ )]

1 1 1 vc vF
ν= g+ = + . (12.40)
2 g 2 vF vc
It is instructive to compare this result to the electron Green function
calculated in the fermion language,
T /2vF
GRR (x, τ ) = − . (12.41)
sin[πT (τ − ix/vF )]
Taking the result obtained in the boson language, and setting g = 1, we
recover the fermion Green function. This argument, a posteriori, fixes the
cut-off dependent prefactor (2πλ)−1/2 in the relations (12.26) and (12.27)
between the fermion creation/annihilation operators and the boson fields φ
and Π.
There is an important difference between the single-electron Green func-
tion for g = 1 and for g 6= 1. In the absence of interactions, the single-
electron Green function has a simple pole at −iτ = x/vF . With interactions,
the location of the singularity shifts, and the singularity acquires a different
analytical structure. This may be brought to light by a calculation of the re-
tarded Green function, i.e., by performing a Fourier transform to Matsubara
frequencies followed by analytical continuation iωn → ω + iη. For simplicity
we look at the equal-position Green function,
Z 1/T
GRR (0, iωn ) = GRR (0, τ )eiωn τ , (12.42)
0
which is related to the spectral density ARR for right-moving electrons. We

deform the τ -integration as shown in Fig. 75, avoiding the branch cuts for
237
0 1/T τ
Figure 75: Integration contour for the calculation of the density of states in a
Luttinger liquid.
Re τ = 0 and Re τ = 1/T . After the deformation of the contours, the integral

reads
Z ∞ ν
2i −ωn t T /2vF
GRR (0, iωn ) = − dte Re
(2πλ)1−ν 0 sin[πT (it + λ/vF )]
Z ∞
2i
= − dte−ωn t
(2πλ)1−ν 0
ν
−iπν/2 T /2vF
× Re e . (12.43)
sinh[πT (t − iλ/vF )]
Now we can take the analytical continuation iωn → ω + iη and calculate the
retarded Green function,
Z ∞ ν
R 2i iωt −iπν/2 T /2vF
GRR (0, ω) = − dte Re e .
(2πλ)1−ν 0 sinh[πT (t − iλ/vF )]
(12.44)
Evaluation of the remaining integral depends on whether the energy ω is

small or large in comparison to T . If ω ≪ T , we find
ν−1
R 2i πT λ
GRR (0, ω) ≈ − . (12.45)
2πvF vF
238
(The numerical prefactor is valid for ν close to unity.) You verify that one
obtains GRRR (0, ω) = −i/2vF for the non-interacting case ν = 1, which implies
ARR = 1/vF , the result that we expected for the spectral density of right-
moving electrons.. If, on the other hand, T ≪ ω, one finds a result similar
to Eq. (12.45) with πT replaced by ω.
We conclude that for an interacting one-dimensional electron liquid the
density of states at the Fermi level has a power-law singularity, and vanishes
proportional to max(ω, T )ν−1. This behavior is quite different from the case
of a Fermi liquid, where the density of states is non-singular at the Fermi
level. Also note that the cut-off length λ enters into the expression for the
density of states if ν 6= 1. For a quantitative estimate, one should replace λ
by λF .
12.3 Luttinger’s model

12.3.1 Formulation of the model
Although the discussion of the previous section provided the physical moti-
vation of the effective Hamiltonian (12.22), it might leave some uneasyness
with those of you who want a “constructive” description of a field theory of
interacting electrons in one dimension. Therefore, we now return to a de-
scription of a one-dimensional electron liquid in fermion language, and derive
the effective Hamiltonian (12.22) using formal manipulations. The discus-
sion of this section follows that of F. D. M. Haldane, J. Phys. C 14, 2585
(1981), although it is less rigorous at points. Those readers who want a fully
rigorous treatment are referred to Haldane’s article, or to the tutorial review
by J. von Delft and H. Schoeller, Annalen Phys. 7, 225 (1998).
As a starting point, we use a description of the one-dimensional elec-
tron system in which we have linearized the electron spectrum, see Fig.
76. We shift momenta by an amount −kF (kF ) for right (left) moving
electrons, so that the Fermi points are shifted to k = 0, and denote the
creation/annihilation operators for right moving (left moving) electrons with
shifted momentum k by c†kR and ckR (c†kL and ckL), respectively. We then
write the kinetic energy as
X
Hkin = vF (k : nkR : −k : nkL :), (12.46)
k
where nkR = c†kR ckR is the density of right-moving electrons and nkL = c†kL ckL
239
ε ε
left right
movers movers
−kF kF k k
ε=ε F− vF (k−kF ) ε=ε F+ vF (k−kF ) k=0 for k=0 for

left movers right movers
Figure 76: Linearized spectrum for electrons in one dimension. In our description,
we separate right and left moving electrons and shift momenta such that the Fermi
points correspond to k = 0.
is the density of left-moving electrons. The symbol : . . . : refers to normal

ordering, i.e., to subtraction of the densities in the (non-interacting) ground
state. The summations over k should be truncated at |k| . kF because the
linear dispersion is valid for a window of wavevectors of size . kF around
k = 0 only and because the distinction between left movers and right movers
implies that the (translated) wavevector of a right mover cannot be smaller
than −kF , whereas the (translated) wavevector of a left mover cannot be
larger than kF , see Fig. 76. In terms of the fields
1 X ikx 1 X ikx
ψR (x) = √ e ckR , ψL (x) = √ e ckL, (12.47)
L k L k
the Hamiltonian (12.46) reads49

Z h i
Hkin = vF dx : ψL† (x)(i∂x )ψL (x) : − : ψR† (x)(i∂x )ψR (x) : . (12.48)
49
The normal ordering in Eq. (12.48) means that ψ † ψ has to be replaced by −ψψ † for
states with momentum smaller than kF . This prescription is not transparent in a real
space formulation.
240
12.3.2 Bosonization
We’ll now show that Hkin can be written in terms of the densities ρqR and
ρqL of right and left moving electrons,
X † X †
ρqR = ck,R ck+q,R , ρqL = ck,L ck+q,L. (12.49)
k k
We consider the case q = 0 separately; note that we do not need to use normal
ordering for q 6= 0. The commutation relations of the density operators are
qL
[ρq,R , ρ−q′ ,R ]− = δqq′ ,
2π
qL
[ρq,L , ρ−q′ ,L ]− = − δqq′ ,
2π
[ρq,R , ρ−q′ ,L ]− = 0. (12.50)
To see how this result is obtained, let us look at the first line of Eq. (12.50)
in detail,
Xh † †
i
[ρq,R , ρ−q ,R ]− =
′ ck,R ck+q,R , ck′,R ck′ −q′ ,R
−
k,k ′
X
= c†k,R ck+q−q′,R − c†k+q′ ,R ck+q,R . (12.51)
k
′
If q 6= q , this is zero, as one can see from relabeling the summation index k.
Otherwise, if q = q ′ we find
X
[ρq,R , ρ−q′ ,R ]− = (nk,R − nk+q,R ) . (12.52)
k
The summation on the right hand side of Eq. (12.52) might appear ambigu-
ous. It can be computed unambiguously by normal ordering of the summand.
Normal ordering of the operators between brackets gives
qL X
[ρq,R , ρ−q,R ]− = − : nk,R − nk+q,R :
2π k
qL
.= (12.53)
2π
The summation of normal ordered terms gives zero after relabeling of the
summation index k. We had to normal order first, since relabelling the
summation index is allowed for normal-ordered summands only.
241
With these commutation relations, the density operators can serve as
boson creation and annihilation operators. Operators ρqR with q < 0 play the
role of creation operators, whereas operators ρqR with q > 0 are annihilation
operators. Similarly, for left-moving particles, the creation operators are the
ρqL with q > 0 and annihilation operators are ρqL with q < 0.50 When
regarded as creation and annihilation operators, the density operators span
the entire Hilbert space with fixed numbers NR and NL of left moving and
right moving electrons. You can verify this statement by explicit construction
of the states, or by comparing the partition functions for the electron liquid in
fermion and boson representations, see footnote 51 below. Since the density
operators span the entire Hilbert space, the kinetic energy can be represented
in terms of the density operators, rather than the fermion operators. In order
to achieve this, we look at the time derivative of ρq,R and ρq,L , for which we
find
∂ i X h † †
i
ρq,R = vF k ck,R ck,R , ck′,R ck′ +q,R
∂t ~ −
k,k ′
i X
† †

= vF k ck,R ck+q,R − ck−q,R ck,R
~ k
−iqvF
= ρq,R . (12.54)
~
Similarly, for the left-moving particles we get
∂ iqvF
ρq,L = ρq,L . (12.55)
∂t ~
Utilizing the commutation relations of the density operators, we find that we
can write the kinetic energy in terms of the boson operators ρq,R and ρq,L ,
2πvF X πvF
Hkin = (ρ−q,R ρq,R + ρq,L ρ−q,L ) + (NR2 + NL2 ), (12.56)
L q>0 L
where the additive constant reflects the additional cost of adding electrons
to the system.51 Another way to obtain Eq. (12.56) is to note that, by
Eq. (12.49), the operator ρq creates a multitude of electron-hole pairs, all
50
Formally, this identification requires that we multiply the density operators by
(L|q|/2π)−1/2 .
51
Now we are in a position to prove that the set of states that is spanned by the density
242
with the same energy vF q. Hence, the energy of the “boson” created by ρq
is vF q. Taking into account the normalization factor (2π/qL)1/2 from the
commutation relations (12.50), one arrives at Eq. (12.56).
The interaction Hamiltonian is, in fact, easier to deal with. Since it
contains four fermion creation or annihilation operators, it is quadratic in
the boson operators ρq,R and ρq,L . Using the observation that the interaction
couples to the total density only, one arrives at the Hamiltonian
1 X
Hint = (V0 − V2kF ) (ρq,R + ρq,L )(ρ−q,R + ρ−q,L ). (12.57)
2L q6=0
The prefactor (V0 − V2kF ) follows from the difference of Hartree and Fock
type interactions, assuming that the interaction Vq depends only weakly on
q on scales ≪ kF . For a point-like interaction, one has V2kF = V0 , and Hint
vanishes, as is required by the Pauli principle.
The above manipulations, in which the fermion operators in the Hamil-
tonian are replaced by boson operators, are known as “bosonization”.
The total Hamiltonian H = Hkin +Hint can be rewritten in terms of boson
fields φ(x) and Π(x) that depend on the coordinate x only. These fields are
related to the particle density and current density as in Eqs. (12.20) and
(12.21). In terms of the density operators, the field φ(x) reads
Z x
φ(x) = π dx′ : ρ(x) :, (12.58)
operators is complete. Calculating the grand canonical partition function for the fermions,
we find
"∞ #2
Y
2n−1 2
Z= (1 + w ) ,
n=1
where we abbreviated w = exp(−πvF /T L) and put the chemical potential precisely be-
tween two successive energy levels. The square within the square brackets corresponds
to states with different signs of k, i.e., to particle and hole-like excitations, whereas the
overall square reflects the identical contributions from left movers and right movers. For
the boson basis, we find
∞ ∞
!2
m2
Y X
2n −2
Z= (1 − w ) w .
n=1 m=−∞
Here, the first factor is the partition function corresponding to the boson degrees of free-
dom, whereas the second factor gives the contribution of the total number of particles to
the energy. You verify that these two partition functions are, indeed, equal.
243
where ρ(x) is the total electron density,
1 X iqx
ρ(x) = e (ρq,R + ρq,L )e−|q|λ/2 , (12.59)
L q6=0
whereas the field Π(x) is given by

X
Π(x) = − eiqx (ρq,R − ρq,L) e−|q|λ/2 . (12.60)
q6=0
(The exponential high-momentum cutoff is the same as in the previous sec-

tion.) Using the commutation relations between the density operators ρq,R
and ρq,L , you verify that the fields φ(x) and Π(x) obey canonical boson
commutation relations, cf. Eq. (12.18). With this change of variables, the
Hamiltonian H = Hkin + Hint acquires the form (12.22), with
r
V0 − V2kF
vc = vF 1 + . (12.61)
πvF
Instead of a formulation in terms of the fields φ(x) and Π(x), which refer
to total electron density and current density (after normal ordering), one
often uses fields φR and φL that are related to the densities of right and left
moving electrons,
1 X iqx 1 X iqx
ρR (x) = e ρq,R e−|q|λ/2 , ρL (x) = e ρq,L e−|q|λ/2 , (12.62)
L q6=0 L q6=0
as
Z x Z x
′ ′
φR (x) = 2π dx : ρR (x ) :, φL(x) = 2π dx′ : ρL (x′ ) : . (12.63)
These fields have commutation rules

[φR (x), φR (x′ )]− = iπsign (x − x′ ),
[φL (x), φL (x′ )]− = −iπsign (x − x′ ), (12.64)
[φR (x), φL (x′ )]− = 0.
They are related to the original fields φ(x) and Π(x) as
π
Z
φR (x) = φ(x) − dx′ Π(x′ )sign (x − x′ ), (12.65)
2~
π
Z
φL (x) = φ(x) + dx′ Π(x′ )sign (x − x′ ). (12.66)
2~
244
In terms of the fields φR (x) and φL , the Hamitonian reads
" 2 2 #
vF ∂φR ∂φL
Z
H = dx +
4π ∂x ∂x
2
V0 − V2kF ∂φR ∂φL
Z
+ dx + . (12.67)
2π ∂x ∂x
12.3.3 Fermion creation and annihilation operators

Having formally rewritten the electron Hamiltonian in terms of boson fields
φ(x) and Π(x), we still want to be able to express the creation and anni-
hilation operators of a single electron in terms of the boson fields. Before
we do this, let us recall the form of the Hilbert space in the boson repre-
sentation. In the boson representation, any electronic state can be written
as density operators ρq,R and ρq,L acting on the ground state with NR right
moving electrons and NL left moving electrons. Hence, every electronic state
is represented by occupation numbers nq,L with q > 0 and nq,R with q < 0
for the boson modes, and the numbers NR and NL ,
|statei = |{nq }, NR , NL i. (12.68)
What is the action of a single-fermion creation operator ψL† (x) or ψR† (x) in
this Hilbert space? First of all, ψR† (x) (ψL† (x)) increases NR (NL ) by one.
Inside the boson Hilbert space, this is achieved by “ladder operators” UR†
and UL† ,
UR† |{nq }, NR , NL i = |{nq }, NR + 1, NL i,

UR |{nq }, NR , NL i = |{nq }, NR − 1, NL i,
UL† |{nq }, NR , NL i = |{nq }, NR , NL + 1i,
UL |{nq }, NR , NL i = |{nq }, NR , NL − 1i. (12.69)
In addition to this, the operator ψR† (x) creates a multitude of electron-

hole pairs, so that, in the end, the change in particle and current densities
is sharply peaked around x. In order to capture these particle hole pairs in
terms of the boson fields, the notation that uses the separate fields φR (x) and
φL (x) for right and left moving electrons is particularly useful. Returning
to the qualitative arguments of the previous section — the addition of an
245
electron corresponds to a “kink” in the fields φR (x) and φL (x) —, we now
make an educated guess for ψR† (x) and ψL† (x),
1 1
ψR† (x) = √ UR† e−iφR (x) , ψL† (x) = √ UL† eiφL (x) , (12.70)
2πλ 2πλ
where the cut-off dependent prefactor (2πλ)−1/2 has already been inserted.
We verify that this gives, indeed, the correct commutator with the density
fields ρR (x) and ρL (x),

h
† ′
i 1 † −iφR (x) ∂ ′
ψR (x), ρR (x ) = √ UR e , ′ φR (x )
− 2π 2πλ ∂x −
1 ∂
= √ UR† e−iφR (x) ′ sign (x − x′ )
2 2πλ ∂x
†
= −ψR (x)δ(x − x′ ),
h i
ψL† (x), ρL (x′ ) = −ψL† (x)δ(x − x′ ). (12.71)
−
In deriving this result, we used the fact that the commutator [eX , Y ]− =
eX [X, Y ]− if [X, Y ]− is proportional to the identity operator. Similarly, you
verify that the ψR† (x) and ψR† (x′ ) anticommute, as well as ψL† (x) and ψL† (x′ ).
There is a problem, however, with the calculation of the anticommutator of
ψR† (x) and ψR (x′ ), which becomes ambiguous when x → x′ . The ambiguity
is lifted once the regularizing momentum cut off in Eqs. (12.59) and (12.60)
or Eq. (12.62) is taken into account. In coordinate representation, this mo-
mentum cut-off amounts to the replacement of the boson fields φR (x) and
φL (x) in the exponents in Eq. (12.70) by the “smeared” fields
2λφR (x′ )
Z
φR (x) → dx′ ,
π[λ2 + 4(x − x′ )2 ]
2λφL (x′ )
Z
′
φL (x) → dx , (12.72)
π[λ2 + 4(x − x′ )2 ]
where λ ∼ λF is the length scale over which the electrons are delocalized.
This way, the “kink” in the fields φR (x) and φL (x) that is created by the
operators ψR† (x) and ψL† (x) is smeared out over a length λ, and, hence, the
added electron is delocalized over a segment of length ∼ λ.52 For the regu-
52
We have chosen a regularization integral in the exponent of Eq. (12.70), instead of a
regularization integral before the exponent. The latter choice would amount a superpo-
sition of sharply localized electrons, which does not need to be the same as a delocalized
electron.
246
larized fields, the sign-function in the commutator is smeared over a distance
∼ λ,
[φR (x), φR (x′ )]− = 2i arctan[(x − x′ )/λ],

[φL (x), φL (x′ )]− = −2i arctan[(x − x′ )/λ]. (12.73)
With this smearing of the boson fields, it can be shown that the ansatz (12.70)
provides the correct anticommutation relation for the fermion creation and
annihilation operator. For details, see the original paper F. D. M. Haldane, J.
Phys. C 14, 2585 (1981) or the tutorial review J. von Delft and H. Schoeller,
Annalen Phys. 7, 225 (1998). Writing the fields φL (x) nd φR (x) in terms of
the original fields φ(x) and Π(x) and undoing the momentum shift q → q±kF
we performed at the beginning of this section, you recover Eqs. (12.26) and
(12.27) of the previous section.
Finally, we have to ensure that operators for left moving and right moving
fermions anticommute. There are various ways to achieve this. The simplest
method is to postulate that the ladder operators UR† and UL† anticommute.
An alternative way is to postulate that the commutator between the left-
moving and right-moving boson fields φL and φR is not zero, but iπ instead.
You are referred to the specialized literature for details.
12.4 Electrons with spin

For electrons with spin, one can take the same route as we took in Sec.
12.2, and write down an effective Hamiltonian based on the spin and charge
velocities in the one-dimensional electron liquid. Introducing fields Πc and
φc that are related to the particle current density and particle density as
1 √
nc (x) = n↑ (x) + n↓ (x) = ∂x φc (x) 2 (12.74)
π √
jc (x) = j↑ (x) + j↓ (x) = −vF Πc (x) 2, (12.75)
together with fields Πs and φs that are related to the spin current density
and spin density,
1 √
ns (x) = n↑ (x) − n↓ (x) = ∂x φs (x) 2 (12.76)
π √
js (x) = j↑ (x) − j↓ (x) = −vF Πs (x) 2, (12.77)
247
we can write the Hamiltonian as
2
π
Z
~vF
dx 2 Πc (x)2 + Πs (x)2

H =
2π ~
+ gc (∂x φc (x))2 + gs2 (∂x φs (x))2 ,
2

(12.78)
where gc = vc /vF and gs = vs /vF . The boson fields Πc,s and φc,s are normal-
ized such that they satisfy canonical commutation relations,
[φc,s (x), φc,s (x′ )]− = [Πc,s (x), Πc,s (x′ )]− = 0
[Πc (x), φc (x′ )]− = [Πs (x), φs (x′ )]− = −i~δ(x − x′ ), (12.79)
[Πc (x), φs (x′ )]− = [Πs (x), φc (x′ )]− = 0.
The form of the fermion operators is found in the same way as for the spinless
case,
† 1 †
ψRσ (x) = √ URσ e−ikF x−iφRσ (x) ,
2πλ
† 1 †
ψLσ (x) = √ ULσ eikF x+iφLσ (x) , (12.80)
2πλ
where σ =↑, ↓ and where the fields φLσ and φRσ are defined as

1
φRσ (x) = √ φc (x) + σφs (x)
2

π
Z
′ ′ ′ ′
− dx [Πc (x ) + σΠs (x )]sign (x − x ) ,
2~

1
φLσ (x) = √ φc (x) + σφs (x)
2

π
Z
′ ′ ′ ′
+ dx [Πc (x ) + σΠs (x )]sign (x − x ) .
2~
(12.81)
Alternatively, one can find a description starting from a microscopic pic-

ture, as in the previous section. In this case, bosonization of the kinetic
energy is done as before. For this, we refer to the references cited above for
more details.
One important aspect of the one-dimensional electron liquid with spin, is
that spin excitations and charge excitations travel at different speeds, see our
248
discussion in Sec. 12.1. We again emphasize that this phenomenon, which is
known as “spin-charge separation”, also exists for collective modes in higher
dimensions. The difference between the higher-dimensional case and the
one-dimensional case is that in higher dimensions there are quasiparticle ex-
citations that carry both spin and charge, in addition to the spin-charge sep-
arated collective excitations. These quasiparticles dominate the low-energy
physics of higher dimensional electron liquids, so that, in the end, spin and
charge always travel together. In one dimension, there are no quasiparticle
excitations, and the spin-charge separated collective modes dominate the low
energy physics.
To illustrate the phenomenon of spin-charge separation, let us create a
right-moving electron with spin up at position x = 0 at time t = 0. As we
discussed above, creating an electron corresponds to making a kink of size π
in the field φR↑ around x = 0. The time-dependence of the charge density
and the spin density then follows from decomposing φR↑ in terms of the spin
and charge fields, and solving for their time-evolution using the Hamiltonian
(12.78). A kink in φR↑ corresponds to kinks in both φc and φs , but these two
kinks travel at different velocities. Hence, although initially charge and spin
were located at the same point, they are separated after a finite time. Only
for a non-interacting electron gas, for which vc = vs = vF , spin and charge
are not separated.
12.5 Does a one-dimensional electron liquid exist?

Now that we have seen that electrons in one dimension behave quite differ-
ently from electrons in three dimensions, it is natural to ask whether all of
this was an academic exercise, or whether one-dimensional electron systems
exist in nature. Fortunately, the answer is positive.
One example of a one-dimensional electron liquid is realized in extremely
thin and clean wires formed in the two-dimensional electron gas that exists
at the interface in a GaAs/GaAlAs heterostructure. This is the example
that comes closest to the physics we have discussed in this chapter. Spin-
charge separation in these thin wires has been observed directly, in the form
of interference between charge and spin excitations in a wire of finite length.
You can find more about the experiment and the theory in Y. Tserkovnyak,
B. I. Halperin, O. M. Auslaender, and A. Yacoby, cond-mat/0302274 and
references therein.
Another example is that of a carbon nanotube. Nanotubes are very good
249
candidates to observe Luttinger Liquid behavior, since they are thin and
clean by virtue of their chemistry. (The wires, on the other hand, have to
be carefully engineered.) However, carbon nanotubes have two propagating
modes at the Fermi level (per spin direction), which slightly modifies some of
the properties. The anomalous density of states has been observed in carbon
nanotubes, see, e.g., M. Bockrath et al., Nature 397, 598 (1999).
The third example is that of edge states in the quantum Hall effect. The
main difference between edge states and the one-dimensional electron liquid
we considered here is that edge states are “chiral”: all electrons move in one
direction. For more details, you are referred to the original papers by Wen
[Phys. Rev. B 41, 12838 (1990); 43, 11025 (1991)].
In addition to the question of real existence, one-dimensional electron
systems have been used as a model for quite a variety of problems in con-
densed matter physics. One example is the Kondo problem with a point-like
magnetic impurity, where it can be argued that the impurity interacts only
with s-wave states in the metal, which, in turn, can be modeled as a one-
dimensional electron gas. Another example is that of a Coulomb blockaded
quantum dot, where the electron gas in the point contacts between the dot
and the bulk electrodes is modeled as a one-dimensional electron gas. In
these applications, most of the electron liquid is non-interacting, the role of
the impurity or of the quantum dot being to introduce a “localized interac-
tion” in the Luttinger Liquid.
250

Notes 2003

Transféré par

Informations du document

Description originale:

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Notes 2003

Transféré par

Droits d'auteur :

Formats disponibles

Many-particle physics

Lecture notes for P654, Cornell University, spring 2003.

hAi = tr Aρ, (1.1)

|ψ(t)i = U(t, t0 )|ψ(t0 )i. (1.7)

U(t, t′ ) = U † (t′ , t). (1.8)

The time-dependence of the evolution operator U(t, t0 ) is given by the Schrö-

U(t, t′ ) = exp[−iH(t − t′ )/~]. (1.10)

If the Hamiltonian H is time dependent, there is no such simple result as Eq.

A(t) = U(t0 , t)AU(t, t0 ), (1.11)

In the interaction picture, the operators evolve according to the unperturbed

A(t) = eiH0 (t−t0 )/~A(t0 )e−iH0 (t−t0 )/~, (1.16)

whereas the quantum state |ψi evolves according to a modified evolution

1.3 imaginary time

Figure 1: Relation between real time t and imaginary time τ .

Show that the time dependence of the expectation value of an observable A

if the operator A has no explicit time dependence.

Exercise 1.2: Free particle, Schrödinger picture.

Exercise 1.3: Free particle, Heisenberg picture

Exercise 1.4: Harmonic oscillator

a† = (ω0 m/2~)1/2 x − i(2~ω0 m)−1/2 p

Exercise 1.5: Imaginary time

Derive and solve the imaginary-time Heisenberg picture evolution equations

2.1 Temperature Green functions

Tτ [A(τ1 )B(τ2 )] = θ(τ1 − τ2 )A(τ1 )B(τ2 ) − (±1)θ(τ2 − τ1 )B(τ2 )A(τ1 ). (2.2)

As discussed previously, the + sign applies if the operators A and B satisfy

GA;B (τ ) = −(±1)GA;B (−τ ) if A, B, H symmetric. (2.3)

Using the periodicity of G, one may write G as a Fourier series,4

with Matsubara frequencies ωn = 2πn/T , n integer, for bosons and ωn =

2.2 Green functions with real time arguments

where the + sign applies to fermions and the − sign to bosons.

Inversely, the greater and lesser Green functions can be expressed as

2.3 Relations between Green functions

ℜGR (ω) = ℜGA (ω), (2.31)

hA(t)B(t′ )i∗ = hB † (t′ )A† (t)i, (2.33)

ℜGR (ω) = ℜG(ω), (2.34)

whereas for the time-ordered Green function one has

AA;B (ω) = −2ℑGR (ω) = 2ℑGA (ω). (2.44)

J1 (ω) = iG> <

see Eq. (2.46) below.

Figure 2: Integration contours for derivation of Eq. (2.48).

we write Eq. (2.51) as

Performing the Fourier transform to time, one finds

With the Lehmann representation, the general relations we derived in Sec.

(b) Derive Lehmann representations of the retarded, advanced, lesser, time-

Exercise 2.2: Spectral density

(d) Now consider a Hamiltonian that is quadratic in creation and annihi-

For a one-dimensional harmonic oscillator with mass m and frequency ω0 ,

(a) Both A and B are equal to the position operator x.

(b) The operator A is equal to the position operator x, while B is equal to

(c) The operator A is equal to the annihilation operator a, while B is equal

In that case, one has the expressions

The momenta and displacements corresponding to opposite wavevectors ~k

3.2 Phonon Green functions

Diα;jβ (τ1 ; τ2 ) = −hTτ [uiα (τ1 )ujβ (τ2 )]i, (3.23)

Since the displacement operators u obey commutation relations (phonons are

where the ωn = 2πn/T , n integer, are the boson Matsubara frequencies.

3.3 Calculation of phonon Green functions

× (hb~λk (t)b~kλ† (0) + bλ†

× hn~kλ ieiω~kλ |t| + hn~kλ + 1ie−iω~kλ |t| .

Performing a Fourier transform to time, we have

Diα;jβ (τ ) = −huiα (τ )ujβ (0)i

3.3.2 Calculation using the equation of motion

Then, using the definition of the temperature Green function