Chap.

Many-electron interactions: the homogeneous

interacting electron gas and beyond
6 Introduction: Hartree-Fock Approximation
Up until now, we have treated the electronic Hamiltonian of solids as an effective Hamiltonian of independent electrons
moving in a mean (or average) field due to the ions and other electrons.

One-electron Schrödinger equation with the mean-field potential V(r)

How best to approximate the effects of electron–electron interactions in the exact Hamiltonian H within the Born–
Oppenheimer approximation ?

Very often, an effective one-electron Schrödinger-like equation of the form Eq. (6.1) is obtained by seeking the ground-state
energy and wavefunction of an interacting many-electron system using a variational principle approach.

In the Hartree–Fock (HF) treatment: the ground-state many electron wavefunction is of the form of a single Slater determinant
6 Introduction: Hartree-Fock Approximation

Here N is the number of electrons in the system, and ri and σi denote the spatial and spin coordinates of the ith electron,
respectively. φλ are single-particle orbitals to be determined in the calculation.

Applying the variational principle, i.e. demanding the variation of the ground-state energy with respect to any of the
single-particle orbitals φλ to be zero,
6 Introduction: Hartree-Fock Approximation

Here N is the number of electrons in the system, and ri and σi denote the spatial and spin coordinates of the ith electron,
respectively. φλ are single-particle orbitals to be determined in the calculation.

Applying the variational principle, i.e. demanding the variation of the ground-state energy with respect to any of the
single-particle orbitals φλ to be zero,
6 Introduction: Hartree-Fock Approximation

Here N is the number of electrons in the system, and ri and σi denote the spatial and spin coordinates of the ith electron,
respectively. φλ are single-particle orbitals to be determined in the calculation.

Applying the variational principle, i.e. demanding the variation of the ground-state energy with respect to any of the
single-particle orbitals φλ to be zero,

H. W. derivation

Hartree where

Fock (exchange)
6 Introduction: Hartree-Fock Approximation
Fock (exchange) term

The exchange term arises from the antisymmetric form of the many-fermion wavefunction, which introduces an effective
interaction between electrons owing to the Pauli exclusion principle.

It is a nonlocal operator of the integral form when acting on a single-particle orbital φi(r, σ), i.e.

which, in general, is rather difficult to treat quantitatively.

The eigenvalues εi, for either occupied or empty orbitals, naturally seem to have the interpretation of eigenenergies of the
original Hamiltonian, namely the total energy of the system is approximated as ∑𝑖𝑖 𝜖𝜖𝑖𝑖 𝑛𝑛𝑖𝑖 , where ni is occupation number of
orbital i.

The eigenvalue εi in the Hartree–Fock equation (Eq. (6.5)) is a Lagrange multiplier in the variational procedure, constraining
the normalization of the single-particle orbitals. Thus, rigorously speaking, they are not electron excitation energies of the
system. However, through the use of Koopmans’ theorem, these eigenvalues may be roughly interpreted as electron
excitation energies to lowest order in the Coulomb interaction.
6.1 The homogeneous interacting electron gas
The uniform interacting electron gas (the jellium model)
In this model, we consider a system of interacting electrons in a uniform positive-charged background so that the total
system is neutral. (an idealized metal)

In the zero-order approximation, we neglect electron–electron interactions; the jellium model reduces to that of a non-
interacting charged Fermi gas, i.e. the Sommerfeld model.

the Hamiltonian the ground-state wavefunction

The positive background potential

with spin

The value of k is within the Fermi sphere, i.e. k ≤ kF, where the factor 2 is due to spin.
6.1 The homogeneous interacting electron gas
The uniform interacting electron gas (the jellium model)

The orbital energy is independent of λ given by

The total energy of the system is

𝑁𝑁
Only one variable physical parameter, the density of the electrons 𝑛𝑛 = or 𝑘𝑘𝐹𝐹 = 3𝜋𝜋 2 𝑛𝑛 1/3
H. W. derivation
Ω

the dimensionless electron gas parameter the effective radius of the volume occupied by
the electron in units of the Bohr radius

the Fermi wavevector

the Fermi energy

the plasma frequency

6.1 The homogeneous interacting electron gas
The uniform interacting electron gas (the jellium model)

The average kinetic energy per electron

H. W. derivation

The average potential energy per electron

(rs between 1 and 5) (electronic lattice)

is a measure of the ratio of the average potential energy to the kinetic energy in the electron gas.
Coulomb energy

A system with rs < 1 corresponds to one with high density and dominant kinetic energy; a system with rs > 1 corresponds to
one with low density and dominant potential energy.
6.2 Hartree–Fock treatment of the interacting electron gas
The single-particle orbitals are planewaves times a spin function, i.e. the orbitals are specified by the quantum
numbers k and λ.

plane wave

Spin function

We note that planewave solutions result in a uniform electron charge density. This results in Vion(r) + VH = 0,
since the electron charge density cancels the positively charged background density.
6.2 Hartree–Fock treatment of the interacting electron gas

equal occupation of k ↑ and k ↓ orbitals, i.e. a paramagnetic electron gas

6.2 Hartree–Fock treatment of the interacting electron gas

The planewaves φkλ are solutions to the Hartree–Fock equation in the case of the uniform electron gas.

with
6.2 Hartree–Fock treatment of the interacting electron gas

The planewaves φkλ are solutions to the Hartree–Fock equation in the case of the uniform electron gas.

with
𝑘𝑘𝐹𝐹 1
2𝜋𝜋 4𝜋𝜋𝑒𝑒 2
−� 𝑑𝑑𝑑𝑑𝑑 � 𝑑𝑑cos𝜃𝜃 3 ′ 2
𝑘𝑘𝑘2
0 −1 2𝜋𝜋 𝑘𝑘 − 𝑘𝑘
𝑒𝑒 2 𝑘𝑘𝐹𝐹 1
𝑘𝑘𝑘2
=− � 𝑑𝑑𝑑𝑑𝑑 � 𝑑𝑑cos𝜃𝜃 2 Let 𝑎𝑎 = 𝑘𝑘 2 + 𝑘𝑘𝑘2 and 𝑏𝑏 = 2𝑘𝑘𝑘𝑘 ′
𝜋𝜋 0 −1 𝑘𝑘 + 𝑘𝑘𝑘2 − 2𝑘𝑘𝑘𝑘 ′ cos𝜃𝜃
𝑒𝑒 2 𝑘𝑘𝐹𝐹 2
1
1 1 1 1 1 1 1
=− � 𝑑𝑑𝑑𝑑𝑑𝑘𝑘𝑘 � 𝑑𝑑cos𝜃𝜃 ∫−1 𝑑𝑑cos𝜃𝜃 𝑎𝑎−𝑏𝑏cos𝜃𝜃= −𝑏𝑏 ln(𝑎𝑎 − 𝑏𝑏cos𝜃𝜃) = ln(𝑎𝑎 − 𝑏𝑏) + ln(𝑎𝑎 + 𝑏𝑏)
𝜋𝜋 0 −1 𝑎𝑎 + 𝑏𝑏cos𝜃𝜃 −1 −𝑏𝑏 𝑏𝑏

𝑒𝑒 2 𝑘𝑘𝐹𝐹 2
1 2 + 𝑘𝑘 ′ 2 + 2𝑘𝑘𝑘𝑘 ′ ) −ln(𝑘𝑘 2 + 𝑘𝑘𝑘2 − 2𝑘𝑘𝑘𝑘 ′ )
=− � 𝑑𝑑𝑑𝑑𝑑𝑘𝑘𝑘 ln(𝑘𝑘
𝜋𝜋 0 2𝑘𝑘𝑘𝑘 ′
𝑒𝑒 2 𝑘𝑘𝐹𝐹 1
=− � 𝑑𝑑𝑑𝑑𝑑𝑘𝑘𝑘2 ′ ln 𝑘𝑘 + 𝑘𝑘 ′ 2 −ln
𝑘𝑘 − 𝑘𝑘 ′ 2 )
𝜋𝜋 0 2𝑘𝑘𝑘𝑘
6.2 Hartree–Fock treatment of the interacting electron gas

The planewaves φkλ are solutions to the Hartree–Fock equation in the case of the uniform electron gas.

with
𝑘𝑘𝐹𝐹 1
2𝜋𝜋 4𝜋𝜋𝑒𝑒 2
−� 𝑑𝑑𝑑𝑑𝑑 � 𝑑𝑑cos𝜃𝜃 3 ′ 2
𝑘𝑘𝑘2
0 −1 2𝜋𝜋 𝑘𝑘 − 𝑘𝑘
𝑒𝑒 2 𝑘𝑘𝐹𝐹 2
1
=− � 𝑑𝑑𝑑𝑑𝑑𝑘𝑘𝑘 ′
ln 𝑘𝑘 + 𝑘𝑘 ′ 2 −ln 𝑘𝑘 − 𝑘𝑘 ′ 2 )
𝜋𝜋 0 2𝑘𝑘𝑘𝑘
𝑒𝑒 2 𝑘𝑘𝐹𝐹 𝑘𝑘𝑘 𝑥𝑥 2 −𝑎𝑎2 1
= � 𝑑𝑑𝑑𝑑𝑑 − ln k + k′ +ln k − k′ using ∫ 𝑥𝑥ln 𝑥𝑥 + 𝑎𝑎 𝑑𝑑𝑑𝑑 = ln 𝑥𝑥 + 𝑎𝑎 − (𝑥𝑥 − 𝑎𝑎)2
𝜋𝜋 0 𝑘𝑘 2 4
𝑘𝑘𝐹𝐹
𝑒𝑒 2 𝑘𝑘 ′2 − 𝑘𝑘 2 𝑘𝑘 − 𝑘𝑘 ′ 1 ′ 2+
1
= ln| | − 𝑘𝑘 + 𝑘𝑘 𝑘𝑘 − 𝑘𝑘 ′ 2
𝜋𝜋𝜋𝜋 2 𝑘𝑘 + 𝑘𝑘 ′ 4 4 0

𝑒𝑒 2 𝑘𝑘𝐹𝐹 𝑘𝑘𝐹𝐹2 − 𝑘𝑘 2 𝑘𝑘 + 𝑘𝑘𝐹𝐹

=− 1+ ln | |
𝜋𝜋 2𝑘𝑘𝑘𝑘𝐹𝐹 𝑘𝑘 − 𝑘𝑘𝐹𝐹
6.2 Hartree–Fock treatment of the interacting electron gas

The planewaves φkλ are solutions to the Hartree–Fock equation in the case of the uniform electron gas.

with

where

If one invokes Koopmans’ theorem, then the eigenvalues are a rough estimate of the electron
(or quasiparticle) excitation energy of the system.

The eigenvalues describe the energy for exciting a single-particle-like excitation within the
Hartree–Fock approximation, including only the exchange interaction.
the self energy of the excited particle

with
6.2 Hartree–Fock treatment of the interacting electron gas

The planewaves φkλ are solutions to the Hartree–Fock equation in the case of the uniform electron gas.

with

where

with

the self energy of the excited particle

The single-particle states are lowered in energy and the occupied

bandwidth WHF = εHF (kF)−εHF (0) increases.
6.2 Hartree–Fock treatment of the interacting electron gas
Limitation of the HF approximation

with

F′(1)= ∞

This divergence leads to behaviors for the electrons that are peculiar to the Hartree–Fock results.

- Infinite velocity and zero effective mass

using the relation

∗
the effective mass 𝑚𝑚HF → 0 as k → kF.
6.2 Hartree–Fock treatment of the interacting electron gas
Limitation of the HF approximation

with

F′(1)= ∞

This divergence leads to behaviors for the electrons that are peculiar to the Hartree–Fock results.

- Infinite velocity and zero effective mass

- A vanishing of the electron density of states D(ε) at ε = εF

- An electronic specific heat that behaves like Cel(T) ∼ T/ln|T| instead of a linear T
dependence at low temperature T
6.2 Hartree–Fock treatment of the interacting electron gas
Limitation of the HF approximation

with

F′(1)= ∞

This divergence leads to behaviors for the electrons that are peculiar to the Hartree–Fock results.

The above unphysical behavior can be traced back to the divergence of the Fourier
4𝜋𝜋𝑒𝑒 2 𝑒𝑒 2
transform of the Coulomb interaction , as k goes to zero.
Ω𝑘𝑘 2 𝑟𝑟

𝑒𝑒 2
Physically, in a metallic system, will be screened by the other electrons, especially at
𝑟𝑟
large values of r. A screened Coulomb potential would eliminate the divergence as k → 0.

Thomas–Fermi screening
(see Chapter 8)

where Ks is the Thomas–Fermi screening wavevector

6.3 Ground-state energy: Hartree–Fock and beyond
The Hartree–Fock approximation, being a variational approach for the ground-state wavefunction, is expected to
be more accurate for the ground-state energy as opposed to individual single-particle excitation energies.

The ground-state energy per particle for an N-particle system

where H is the full Hamiltonian and is Ψ the many-body ground-state wavefunction.

The ground-state energy per particle EKE: the kinetic energy per particle

The exchange energy per particle

Double counting

𝑘𝑘
using � sin𝜃𝜃𝜃𝜃𝜃𝜃𝜃𝜃𝜃𝜃 = 4𝜋𝜋, Let = 𝑥𝑥
𝑘𝑘𝐹𝐹

𝑛𝑛 = 𝑘𝑘𝐹𝐹3 /3𝜋𝜋 2

=1/4
6.3 Ground-state energy: Hartree–Fock and beyond
The Hartree–Fock approximation, being a variational approach for the ground-state wavefunction, is expected to
be more accurate for the ground-state energy as opposed to individual single-particle excitation energies.

The ground-state energy per particle for an N-particle system

where H is the full Hamiltonian and is Ψ the many-body ground-state wavefunction.

The ground-state energy per particle EKE: the kinetic energy per particle

1 1 2)
1 (1 − 𝑥𝑥 1 + 𝑥𝑥 𝐼𝐼
� 𝑥𝑥 2 𝐹𝐹 𝑥𝑥 𝑑𝑑𝑑𝑑 = � 𝑥𝑥 2 + 𝑥𝑥ln 𝑑𝑑𝑑𝑑
0 0 2 4 1 − 𝑥𝑥
1 1 1 2
1 + 𝑥𝑥
= + � (1 − 𝑥𝑥 )𝑥𝑥ln 𝑑𝑑𝑑𝑑
6 4 0 1 − 𝑥𝑥
1 1
= + 𝐼𝐼
6 4
using 1
1
� 𝑥𝑥 2 𝐹𝐹 𝑥𝑥 𝑑𝑑𝑑𝑑 =
0 4
6.3 Ground-state energy: Hartree–Fock and beyond
The Hartree–Fock approximation, being a variational approach for the ground-state wavefunction, is expected to
be more accurate for the ground-state energy as opposed to individual single-particle excitation energies.

The ground-state energy per particle for an N-particle system

where H is the full Hamiltonian and is Ψ the many-body ground-state wavefunction.

The ground-state energy per particle EKE: the kinetic energy per particle

The exchange energy per particle

Double counting

We expect that a more accurate theoretical treatment of the many-body wavefunction, going
beyond Hartree–Fock, would result in more terms.

In the limit of small rs (i.e. high density), the kinetic energy dominates, and we may expect the ground-state energy per particle
(and other quantities) to be expandable in a series in powers of rs.

The correlation energy

6.3 Ground-state energy: Hartree–Fock and beyond
The Hartree–Fock approximation, being a variational approach for the ground-state wavefunction, is expected to
be more accurate for the ground-state energy as opposed to individual single-particle excitation energies.

The ground-state energy per particle for an N-particle system

where H is the full Hamiltonian and is Ψ the many-body ground-state wavefunction.

The ground-state energy per particle EKE: the kinetic energy per particle

The exchange energy per particle

breakdown of perturbation theory
Double counting

The expression for EGS

For the interacting electron gas at small rs

Many-body perturbation theory is accurate in the limit of rs < 1. Correlation energy of the interacting electron gas as
a function of rs from different calculations.
6.3 Ground-state energy: Hartree–Fock and beyond
One of the simplest and yet highly accurate expressions for EC is the Wigner interpolation formula. (low density, i.e. rs >>1 .)

In this limit, the potential energy dominates over the kinetic energy and the electrons condense into an electron crystal in three
dimensions at rs ≥ 100.

To calculate the ground-state energy of the electron crystal, let us divide the crystal into Wigner–Seitz cells.

The energy per electron is the energy of the electron in one of the Wigner–Seitz spheres, which consists of a kinetic energy term
EKE and a potential energy term EPE.

- EKE comes from the zero-point motion of the electron and scales like rs−3/2, which may be neglected in the rs → ∞ limit.

- EPE comes from the Coulomb interaction of the charges involved.

Ee−p is the interaction energy between the electron at the center of the Wigner–Seitz sphere and the positive
neutralizing background density np = 3/(4π rs3),
6.3 Ground-state energy: Hartree–Fock and beyond
One of the simplest and yet highly accurate expressions for EC is the Wigner interpolation formula. (low density, i.e. rs >>1 .)

In this limit, the potential energy dominates over the kinetic energy and the electrons condense into an electron crystal in three
dimensions at rs ≥ 100.

To calculate the ground-state energy of the electron crystal, let us divide the crystal into Wigner–Seitz cells.

The energy per electron is the energy of the electron in one of the Wigner–Seitz spheres, which consists of a kinetic energy term
EKE and a potential energy term EPE.

- EKE comes from the zero-point motion of the electron and scales like rs−3/2, which may be neglected in the rs → ∞ limit.

- EPE comes from the Coulomb interaction of the charges involved.

Ep−p, is the self energy of the positive neutralizing background charge, given by
6.3 Ground-state energy: Hartree–Fock and beyond
One of the simplest and yet highly accurate expressions for EC is the Wigner interpolation formula. (low density, i.e. rs >>1 .)

In this limit, the potential energy dominates over the kinetic energy and the electrons condense into an electron crystal in three
dimensions at rs ≥ 100.

To calculate the ground-state energy of the electron crystal, let us divide the crystal into Wigner–Seitz cells.

The energy per electron is the energy of the electron in one of the Wigner–Seitz spheres, which consists of a kinetic energy term
EKE and a potential energy term EPE.

- EKE comes from the zero-point motion of the electron and scales like rs−3/2, which may be neglected in the rs → ∞ limit.

- EPE comes from the Coulomb interaction of the charges involved.

Using the fact that EPE = EX + EC and the Hartree–Fock result of EX = −0.916/rs,
6.3 Ground-state energy: Hartree–Fock and beyond
One of the simplest and yet highly accurate expressions for EC is the Wigner interpolation formula. (low density, i.e. rs >>1 .)

In this limit, the potential energy dominates over the kinetic energy and the electrons condense into an electron crystal in three
dimensions at rs ≥ 100.

To calculate the ground-state energy of the electron crystal, let us divide the crystal into Wigner–Seitz cells.

The energy per electron is the energy of the electron in one of the Wigner–Seitz spheres, which consists of a kinetic energy term
EKE and a potential energy term EPE.

- EKE comes from the zero-point motion of the electron and scales like rs−3/2, which may be neglected in the rs → ∞ limit.

- EPE comes from the Coulomb interaction of the charges involved.

Using the fact that EPE = EX + EC and the Hartree–Fock result of EX = −0.916/rs,

Using an estimate of EC (rs = 1)=−0.10 Ry from experiment, Wigner arrived at a simple interpolation formula of
6.3 Ground-state energy: Hartree–Fock and beyond
One of the simplest and yet highly accurate expressions for EC is the Wigner interpolation formula. (low density, i.e. rs >>1 .)

In this limit, the potential energy dominates over the kinetic energy and the electrons condense into an electron crystal in three
dimensions at rs ≥ 100.

To calculate the ground-state energy of the electron crystal, let us divide the crystal into Wigner–Seitz cells.

The energy per electron is the energy of the electron in one of the Wigner–Seitz spheres, which consists of a kinetic energy term
EKE and a potential energy term EPE.

Using an estimate of EC (rs = 1)=−0.10 Ry from experiment, Wigner arrived at a simple interpolation formula of

The Wigner interpolation formula has been proved to be accurate and useful in many applications, and although many other
correlation energy formulas have been derived over the years, the Wigner interpolation formula remains among the best.
6.4 Electron density and pair-correlation functions
An N-body fermion wavefunction

Normalization of the wavefunction

The electron density operators

The electron density

all the particles being identical.

6.4 Electron density and pair-correlation functions
The pair-correlation function

g(r, r')ρ(r') is the density of electrons at r', given that there is one electron at r.

The distribution of particle pairs

NC2 integration over the spin degree of freedom of particles 1 and 2

Two-particle density matrix
6.4 Electron density and pair-correlation functions
The pair-correlation function

General properties of the pair-correlation function

The sum rule: Integrating it over all volume gives the total number of
pairs in the system, which is N(N − 1)/2
the integrand is the distribution of particle pairs

The reciprocity relation comes from the symmetry.

The density of electrons at r' would be uncorrelated to the occupation

of an electron at r in a material if r' were very far away from r.
6.4 Electron density and pair-correlation functions
Parallel and antiparallel spin components

By separating the integration over spin coordinates of

6.4 Electron density and pair-correlation functions

The Pauli exclusion principle vanishes the wavefunction.

This may be proven by noting that

(N − 1) − N = −1
6.5 g (r, r') of the interacting electron gas
For a homogeneous system like an electron gas, the pair-correlation function g(r, r') depends only on the distance
between r and r' and not on the positions separately:

We now evaluate these correlation functions for the interacting electron gas within the Hartree–Fock approximation.
For simplicity, we work with a paramagnetic electron gas, i.e. the number of spin-up electrons is equal to the number
of spin-down electrons.

The pair-correlation function The electron density

(N-2)!
Use of one-electron orthonormal orbitals in the Hartree–Fock Slater determinant

The sum over (i, j) is carried out over

all occupied orbitals.
6.5 g (r, r') of the interacting electron gas
For the homogeneous electron gas,

where the spin function

Thus,

We make use of the result 2

H. W. derivation

Four combinations for (si, sj): (↑, ↑), (↓, ↓), (↑, ↓), and (↓, ↑)

Using and N↑ = N↓ = N/2

6.5 g (r, r') of the interacting electron gas
For the homogeneous electron gas,

where the spin function

Thus,

We make use of the result 2

H. W. derivation

H. W. derivation 𝑘𝑘𝐹𝐹 𝜋𝜋 2𝜋𝜋

Σ=� � � 𝑘𝑘 2 sin𝜃𝜃𝜃𝜃𝜃𝜃𝜃𝜃𝜃𝜃𝜃𝜃𝜙𝜙
where 0 0 0

𝐤𝐤 � 𝐫𝐫 = 𝑘𝑘𝑘𝑘cos𝜃𝜃
the first Bessel function
6.5 g (r, r') of the interacting electron gas

where

Since f(r) goes to 1 as r → 0

For r→∞, we have f → 0

6.5 g (r, r') of the interacting electron gas

𝐻𝐻𝐻𝐻
The extent over which 𝑔𝑔↑↑ deviates from 1/2 is dictated by 1/kF.

The Hartree– Fock Slater determinant ground-state wavefunction only has

exchange correlation between electrons of like spin; there is no correlation
between electrons of unlike spin.

In a more accurate description of the many-body wavefunction than that of

Hartree–Fock, there is correlation between unlike spins, due to the Coulomb
repulsion between the electrons.

→These correlations reduce g↑↓(r) to less than 1/2 as r → 0.

𝐻𝐻𝐻𝐻
g↑↑ (r) would change from that of 𝑔𝑔↑↑ since the sum rule is required.

The difference between the exact g and gHF leads to a change in the ground-state energy and gives rise to
the correlation energy.
6.6 The exchange-correlation hole
The electron density at position r' as seen by an electron at position r is given by

The difference between the electron density seen by an electron at r, ρr(r'), and that of the electron density ρ(r') measured,
for example, in an X-ray scattering experiment, is given by

Integrating over volume and making use of the sum rule yields

There is a depletion of electron density near the position of an electron,

and that the integrated missing density is exactly equal to one electron.

The Pauli exclusion principle, as well as Coulomb repulsion, reduces the

density of electrons near a specific site already occupied by an electron.

Schematic of ρ(r’) and ρr(r’). (a) charge density ρ(r) of a

This depletion of (or hole in) electron density given by is called system as measured by an external probe. (b) charge
the exchange-correlation hole. density ρr(r’) of the same system as seen by an electron at r.
6.6 The exchange-correlation hole
The exchange-correlation hole moves with the electron and, in general, changes size and shape as the electron travels
through a system of inhomogeneous electron density.

The electron density ρr(r') seen by a given electron at r is always different from that of ρ(r'). For example, in an
interacting homogeneous electron gas, although ρ(r') is a constant, ρr(r') is highly inhomogeneous.

The evaluation of the pair-correlation function and the exchange-correlation hole for a real material is non-trivial
since it requires knowing the many-electron ground-state wavefunction. reasonable description of g↑↓(r, r') means
incorporation of correlation effects going beyond Hartree–Fock. →variational quantum Monte Carlo simulations

r r

A suppression from the value of ½ for g↑↑ , as r → r' The suppression of g↑↓ is not nearly as pronounced because
showing a deep exchange hole diamond is only a moderately correlated electron system.
6.7 The exchange-correlation energy
The exchange-correlation energy of an electron is the amount of lowering of the energy of the electron, as the many-body
wavefunction describes more correlations of the motions of the electron as we improve on the Hartree approximation (that is,
the energy gain due to both exchange and correlation effects).

We may calculate this energy difference by evaluating the potential energy using a charge density around an electron at r,
with and without the exchange-correlation hole contribution.

Without the exchange-correlation hole, With the exchange-correlation hole,

The energy difference, which is the exchange-correlation energy for the electron at r,

the interaction energy of the electron with its exchange-correlation hole which can be considered positively
charged with an integrated charge of −e
6.7 The exchange-correlation energy
The quantity εxc(r) is also called the exchange-correlation energy density of the system.

We may apply this concept to the case of the interacting electron gas within the Hartree– Fock approximation. Since the
system is homogeneous, εxc(r) is the same everywhere and is the exchange-correlation energy per particle.

𝑘𝑘𝐹𝐹3 /3𝜋𝜋 2
6.7 The exchange-correlation energy
The quantity εxc(r) is also called the exchange-correlation energy density of the system.

We may apply this concept to the case of the interacting electron gas within the Hartree– Fock approximation. Since the
system is homogeneous, εxc(r) is the same everywhere and is the exchange-correlation energy per particle.

𝑘𝑘𝐹𝐹3 /3𝜋𝜋 2
6.7 The exchange-correlation energy
The quantity εxc(r) is also called the exchange-correlation energy density of the system.

We may apply this concept to the case of the interacting electron gas within the Hartree– Fock approximation. Since the
system is homogeneous, εxc(r) is the same everywhere and is the exchange-correlation energy per particle.

𝑘𝑘𝐹𝐹3 /3𝜋𝜋 2

A particularly interesting situation is the case in which the electron leaves the crystal and stays near the surface. The
exchange-correlation hole by necessity stays behind in the crystal

The Coulomb interaction between the positively charged exchange-correlation hole and the electron that leaves the crystal
results in an attraction that scales like the inverse of the distance from the surface, giving rise to the classical image force.
6.7 The exchange-correlation energy
The exchange-correlation energy plays an important role in determining conceptually and quantitatively many other
properties of solids.

the cohesive energy of the alkali metals

The cohesive energy per atom is defined by

Eatom: the binding energy of the outer active electrons in the atom for the isolated case
EGS: the energy per atom in the solid state

For the alkali metal elements

Eatom is just the negative of the ionization potential.

i)
the conduction band effective mass

ii) Ee-ion is the bottom of the conduction band.

iii) the homogeneous electron gas

6.7 The exchange-correlation energy
The exchange-correlation energy plays an important role in determining conceptually and quantitatively many other
properties of solids.

the cohesive energy of the alkali metals

The cohesive energy per atom is defined by

Eatom: the binding energy of the outer active electrons in the atom for the isolated case
EGS: the energy per atom in the solid state

For the alkali metal elements

Eatom is just the negative of the ionization potential.

If a crystal is formed, EGS is a minimum at the equilibrium lattice constant,

We see that the alkali metals are unbounded within the Hartree only approximation.
Both exchange and correlation effects are needed to give a reasonable description of
the cohesive properties of this simplest class of metals.