DFT Presentation

The many-body problem From Hartree to Thomas Fermi DFT HK theorems KS scheme and bands Summary
Key concepts in Density Functional Theory

From the many body problem to the Kohn-Sham scheme
Silvana Botti
ILM (LPMCN) CNRS, Universite Lyon 1 - France

European Theoretical Spectroscopy Facility (ETSF)
December 12, 2012 Lyon
Key concepts in Density Functional Theory Silvana Botti

Outline
1 The many-body problem
2 Earlier solutions: Hartree, Hartree-Fock, Thomas-Fermi
3 The solution: density functional theory
4 Hohenberg-Kohn theorems
5 Practical implementations: the Kohn-Sham scheme
6 Summary

Outline
6 Summary

The many-body problem
Schrodinger equation for a quantum system of N interacting particles:
How to deal with N 1023 particles?

H ({R} , {r}) = E ({R} , {r})
Ne electrons H = Tn ({R}) + Vnn ({R}) +

Nn nuclei Te ({r}) + Vee ({r}) + Uen ({R} , {r})

The many-body Hamiltonian
H = Tn ({R}) + Vnn ({R}) + Te ({r}) + Vee ({r}) + Uen ({R} , {r})

Nn N
X 2I Xe
2
Tn = , Te = i,
2MI 2m
I =1 i=1
Nn Ne
1 X ZI ZJ 1 X 1
Vnn = , Vee = ,
2 |RI RJ | 2 |ri rj |
I ,J,I 6=J i,j,i6=j
NX
e ,Nn
ZJ
Uen =
|RJ rj |
j,J

Starting approximations
Born-Oppenheimer separation
In the adiabatic approximation the nuclei are frozen in their equilibrium
positions.
Example of equilibrium geometries

Starting approximations
Pseudopotential and
pseudowavefunction
Concept of pseudopotentials
The chemically intert core electrons are
frozen in their atomic configuration and
their effect on chemically active valence
electrons is incorporated in an effective
potential.

Pseudopotentials: generation criteria
A pseudopotential is not unique, several methods of generation also exist.

1 The pseudo-electron eigenvalues must be the same as the valence
eigenvalues obtained from the atomic wavefunctions.
2 Pseudo-wavefunctions must match the all-electron wavefunctions
outside the core (plus continuity conditions).
3 The core charge produced by the pseudo-wavefunctions must be the
same as that produced by the atomic wavefunctions (for
norm-conserving pseudopotentials).
4 The logaritmic derivatives and their first derivatives with respect to
the energy must match outside the core radius (scattering properties).
5 Additional criteria for different recipes.

Pseudopotentials: quality assessment
It is important to find a compromise between

1 Transferability: ability to describe the valence electrons in different
environments.
2 Efficiency: softness few plane waves basis functions.
Moreover:
Which states should be included in the valence and which states in the
core? Problem of semicore states.

The many-body Hamiltonian
Applying the Born-Oppenheimer separation. . .

Nn 2 Ne 2
X
0 = Tn =
X

2M
I
, T =
I
e
2m
, i
I =1 i=1
Nn Ne
1 X ZI ZJ 1 X 1
constantVnn = , Vee = ,
2 |RI RJ | 2 |ri rj |
I ,J,I 6=J i,j,i6=j
NX
e ,Nn eN
ZJ X
Uen = = v (rj )
|RJ rj |
j,J j

The many-body problem
Schrodinger Equation for a quantum-system of Ne interacting electrons:
Still, how to deal with Ne 1023 particles?

H ({r}) = Ee ({r})
Ne Ne
2i

X 1 X 1
H = + v (ri ) +
Ne electrons 2m 2 |ri rj |
i=1 i,j,i6=j

Why we dont like the electronic wavefunction
How many DVDs are necessary to store a wavefunction?
Classical example: Oxygen atom (8 electrons)

(r1 , . . . , r8 ) depends on 24 coordinates
Rough table of the wavefunction:
10 entries per coordinate: = 1024 entries

1 byte per entry : = 1024 bytes
5 109 bytes per DVD: = 2 1014 DVDs
10 g per DVD: = 2 1015 g DVDs
= 109 t DVDs

Outline
6 Summary

The Hartree equations: Hartree potential
Hartree introduced in 1927 a procedure, which he called the

self-consistent field method, to approximate the Schrodinger equation
variational principle to an ansatz (trial wave function) as a product of
single-particle functions: H = 1 (r1 ) 2 (r2 ) . . . N (rN )

Z
| (r 0 )|2
1 2i + vext (r) + j
X
d 3r 0 i (r) = i i (r)
2 |r r0 |
j
If we do not restrict the sum to j 6= i: self-interaction problem

and the antisymmetry of the wavefunction?

The Hartree equations: Hartree potential
Hartree introduced in 1927 a procedure, which he called the

self-consistent field method, to approximate the Schrodinger equation
variational principle to an ansatz (trial wave function) as a product of
single-particle functions: H = 1 (r1 ) 2 (r2 ) . . . N (rN )

Z
| (r 0 )|2
1 2i + vext (r) + j
X
d 3r 0 i (r) = i i (r)
2 |r r0 |
j
If we do not restrict the sum to j 6= i: self-interaction problem

and the antisymmetry of the wavefunction?

The approach of the best wavefunction
In 1930 Slater and Fock independently pointed out that the Hartree
method did not respect the principle of antisymmetry.
A Slater determinant trivially satisfies the antisymmetry of the exact
solution and hence is a suitable ansatz for applying the variational
principle
The original Hartree method can then be viewed as an approximation
to the Hartree-Fock method by neglecting exchange.
The Hartree-Fock method was little used until the advent of
electronic computers in the 1950s (need of self-consistency!).

The Hartree-Fock equations: exchange potential

The Hartree-Fock method determines the set of (spin) orbitals which
minimizes the energy and give us the best single-determinant:

1 (x1 )
2 (x1 ) N (x1 )
1 1 (x2 )
2 (x2 ) N (x2 )
HF (x1 , x2 , . . . , xN ) = . .. ..
N! .. . .
1 (xN ) 2 (xN ) N (xN )

Z
| (r 0 )|2
1 2i + vext (r) + j
X
d 3r 0 i (r)
2 |r r0 |
j
j (r) i (r0 )
X Z
i j d 3 r 0 = i i (r)
|r r0 |
j

Some remarks on Hartree-Fock
The HF potential is self-interaction free.

Ignoring relaxation (i.e. the change in the self-consistent potential)
the first ionization energy is equal to the negative of the energy of the
HOMO: Koopmans theorem.
The approximate solution of the HF equations cannot be exact
because the true many-body wavefunction is not a single-determinant.
In quantum chemistry the correlation energy is defined as the
difference between the HF energy and the exact ground-state energy.
To go beyond HF a trial wavefunction can be built as a linear
combination of HF orbitals, including excited orbitals:
configuration interaction (CI).

Thomas-Fermi theory
Formulated in 1927 in terms of the electronic density alone, the TF

theory is viewed as a precursor to density functional theory.
Z
TF 3 2
T [] = 3 2 3 5/3 (r)d 3 r
10
The power of can be deduced by dimensional analysis, while the

coefficient is chosen to agree with the uniform electron gas.
Adding the classical expressions for the nuclear-electron and
electron-electron interactions we obtain the original TF functional:
F TF [] = T TF [] + VHartree [] + V []

Problems of Thomas-Fermi theory

The kinetic energy is a sizable part of the total energy and it is here
described by a too poor approximation.
The original formulation did not include the exchange energy (Pauli
principle): an exchange energy functional was added by Dirac in 1928.
However, the Thomas-Fermi-Dirac theory remained rather inaccurate
for most applications. The largest source of error was in the
representation of the kinetic energy, followed by the errors in the
exchange energy, and to the neglect of electron correlation.
In 1962, Edward Teller showed that TF theory cannot describe
molecular bonding the energy of any molecule calculated with TF
theory is higher than the sum of the energies of the constituent
atoms. More generally, the total energy of a molecule decreases when
the bond lengths are uniformly increased.

Outline
6 Summary

Ground state densities vs potentials
Question at the heart of DFT

Is there a 1-to-1 mapping between different external potentials v (r) and
their corresponding ground state densities (r)?

Density functional theory (DFT): the essence
If we can give a positive answer, then it can be proved that

(i) all observable quantities of a quantum system are completely
determined by the density.
(ii) which means that the basic variable is no more the many-body
wavefunction ({r)} but the electron density (r).
P. Hohenberg and W. Kohn, Phys. Rev. 136, B864 (1964).
You can find all details in R. M. Dreizler and E.K.U. Gross, Density
Functional Theory, Springer (Berlin, 1990).

Density functional theory (DFT): the essence
If we can give a positive answer, then it can be proved that

(i) all observable quantities of a quantum system are completely
determined by the density.
(ii) which means that the basic variable is no more the many-body
wavefunction ({r)} but the electron density (r).
You can find all details in R. M. Dreizler and E.K.U. Gross, Density
Functional Theory, Springer (Berlin, 1990).

Outline
6 Summary

Density functional theory (DFT)
Hohenberg-Kohn (HK) theorem I

The expectation value of any physical observable of a many-electron
system is a unique functional of the electron density .
Hohenberg-Kohn (HK) theorem II

The total energy functional has a minimum, the ground state energy E0 ,
at the ground state density 0 .

Density functional theory (DFT)
Restrictions:
In practice, only ground state properties.
The original proof is valid for local, spin-independent external
potential, non-degenerate ground state.
There exist extensions to degenerate ground states, spin-dependent,
magnetic systems, etc.

Hohenberg-Kohn theorem I
~
A A
v(r) ({r}) (r)
singleparticle groundstate
groundstate
potentials having densities
wavefunctions
a nondegenerate
ground state
G : v (r) (r) is obvious.

HK theorem states that G is invertible.

Proof:
1 A is invertible: the Schrodinger equation can be always solved for the
external potential, yielding the potential as a unique function of .

X E T Vee T
V = v (ri ) = = Vee + const.

i
2 A is invertible (proof for non-degenerate ground state):

T + Vee + V = E

T + Vee + V 0 0 = E 0 0
Now what is left to show is that 6= 0 6= 0

Applying the variational principle (Rayleigh-Ritz):

Z
E = h|H|i < h0 |H|0 i = E 0 + d3 r 0 (r) v (r) v 0 (r)

Z
E 0 = h0 |H 0 |0 i < h|H 0 |i = E + d3 r (r) v 0 (r) v (r)

Proof by contradiction:
If = 0 it has to be E + E 0 < E + E 0 , which is absurd.
Therefore, we deduce that 6= 0 6= 0

Direct consequence of the 1st HK theorem

The expectation value of any physical observable of a many-electron
system is a unique functional of the electron density .
G 1 solving S.E.
Proof: v [] 0 []
Then an observable O [] = h0 [] |O|0 []i is a functional of .

Reminder: what is a functional?
A functional maps a function to a number
E[]
(r) R
functional
set of functions set of real numbers

vr [] = v [] (r) is a functional that depends parametrically on r
r1 ...rN [] = [] (r1 . . . rN ) is a functional that depends
parametrically on r1 . . . rN

2nd HK theorem: Variational principle

h0 |H|0 i = min {EHK []} = E0 [0 ]
Euler-Lagrange equation:
Z
3
EHK [] d r (r) =0
(r)
E
where is the chemical potential since = N .

2nd HK theorem: Variational principle

h0 |H|0 i = min {EHK []} = E0 [0 ]
EHK []
=
(r)
It yields the exact ground-state energy E0 and density 0 (r).


Formal construction of EHK []:
A1
For an arbitrary ground state density it is always true (r) []
we can define the functional of the density:
EHK [] = h [] |T + Vee + V | []i
EHK [] >E0 for 6=0

EHK [] =E0 for =0
EHK [] = FHK [] + d3 r v (r)(r), where FHK [] is universal, as it
R
does not depend on the external potential.

FHK []
(r) + vext (r) =

In principle: the Euler-Lagrange equation allows to calculate 0 (r) without

introducing a Schrodinger equation.
The HK theorem proves the existence of the universal functional
Z
EKS [] = FHK [] + d3 r v (r)(r)
FHK [] = h|T + Vee |i

but it does not tell us how to determine it.
In practice: FHK [] needs to be approximated and approximations of T []
lead to large errors in the total energy.
(See again the same problem as in Thomas-Fermi theory).

Outline
6 Summary

Reformulation: Kohn-Sham scheme
HK 11 mapping for HK 11 mapping

interacting particles noninteracting particles
vext[](r) (r) vKS[](r)
Essence of the mapping

The density of a system of interacting particles can be calculated exactly
as the density of an auxiliary system of non-interacting particles
= Reformulation in terms of single-particle orbitals!
W. Kohn and L. J. Sham, Phys. Rev. 140, A1133 (1965).

Back to the Hohenberg-Kohn variational principle
For the non-interacting system:

Z
EKS [] = h[]|Ts + VKS |[]i = Ts [] + d3 r (r)vKS (r)
Euler-Lagrange equation for the non-interacting system

Z

EKS [] d3 r (r) = 0
(r)
Ts []
+ vKS (r) =
(r)

Using a one-particle Schrodinger equation
Kohn-Sham equations
2

+ vKS (r) i (r) = i i (r)
2
X
0 (r) = |i (r) |2
i, lowest i
i = KS eigenvalues, i (r) = KS single-particle orbitals
Can we always build vKS for the non-interacting electron system?

Uniqueness of vKS follows from HK 1-1 mapping.
Existence of vKS is guaranteed by V-representability theorem.

Problem of V-representability
Definition
(r) is V-representable if it is the ground-state density of some potential V .
Question
Are all reasonable functions (r) V-representable?
Answer: V-representability theorem

On a lattice (finite or infinite) any normalizable positive function (r), that is
compatible with the Pauli principle, is both interacting and non-interacting
V-representable. For degenerate ground states such a (r) is ensemble
V-representable, i.e. representable as a linear combination of the degenerate
ground-states densities.
Chayes, Chayes, Ruskai, J Stat. Phys. 38, 497 (1985).

Reformulation: Kohn-Sham scheme
Kohn-Sham one-particle equations

2

+ vKS (r) i (r) = i i (r)
2
to be solved self-consistenlty with 0 (r) = i |i (r) |2

P
i = KS eigenvalues, i (r) = KS single-particle orbitals

Which is the form of vKS for the non-interacting electrons?
Hartree potential
.
vKS (r) = v (r) + vH (r) + vxc (r)
%
unknown exchange-correlation (xc) potential

Kohn-Sham scheme: Hartree and xc potentials
Hartree potential
(r0 )
Z
vH [] (r) = d3 r 0
|r r0 |
vH describes classic electrostatic interaction
Exchange-correlation (xc) potential

Exc []
vxc [] (r) =
(r)
vxc encompasses many-body effects

Kohn-Sham scheme: xc potential

Proof:
Knowing that we want the total energy of the real system,
we rewrite FHK [] = Ts [] + EH [] + Exc [], after defining
Exc [] = FHK [] EH [] Ts [].
EH []
We use the variational principle (and (r) = vH (r))
FHK []
+ v (r) =
(r)
Ts []
+ vKS (r) =
(r)
Exc []
To obtain vKS (r) = v (r) + vH (r) + vxc (r), (r) = vxc (r)

Approximations for the xc potential

LDA: Z
LDA
Exc [] = d3 r (r) HEG
xc ( (r))
LSDA: Z
LSDA
Exc [ , ] = d3 r (r) HEG
xc ( , )
GGA:
Z
GGA
Exc [ , ] = d3 r (r) GGA
xc ( , , , )
meta-GGA:
Z
MGGA
d3 r (r) MGGA , , , , 2 , 2 , ,

Exc [ , ] = xc
EXX, SIC-LDA, hybrid Hartree-Fock/DFT functionals, . . .

Approximations for the xc potential

LDA: Z
LDA
Exc [] = d3 r (r) HEG
xc ( (r))
LSDA: Z
LSDA
Exc [ , ] = d3 r (r) HEG
xc ( , )
GGA:
Z
GGA
Exc [ , ] = d3 r (r) GGA
xc ( , , , )
meta-GGA:
Z
MGGA
d3 r (r) MGGA , , , , 2 , 2 , ,

Exc [ , ] = xc
EXX, SIC-LDA, hybrid Hartree-Fock/DFT functionals, . . .

Can we calculate excited states within static DFT?
Density functional theory in the Kohn-Sham scheme

gives an efficient and accurate description of GROUND STATE
properties (total energy, lattice constants, atomic structure, elastic
constants, phonon spectra . . .)
is not designed to access EXCITED STATES
however . . .


however . . .


however . . .

Kohn-Sham band structure: some facts
One-electron band structure

is the dispersion of the energy levels n as a function of k in the Brillouin
zone.
The Kohn-Sham eigenvalues and eigenstates are not one-electron
energy states for the electron in the solid.
However, it is common to interpret the solutions of Kohn-Sham
equations as one-electron states: the result is often a good
representation, especially concerning band dispersion.
Gap problem: the KS band structure underestimates systematically
the band gap (often by more than 50%).

Discontinuity in Vxc
Band gap error not due to LDA, but to the discontinuity in the exact Vxc .
.6

.6
Eg = (E(N+1) E(N) ) (E(N) E(N1) )
.6
11 Eg = KS KS
N+1 (N + 1) N
;&
.6
1 (JDS
(JDS
.6

.6
1
EgKS = KS KS
N+1 N
1HOHFWURQV N 1HOHFWURQV N

xc = Eg EgKS = Vxc
+
(r) Vxc (r)
L. J. Sham and M. Schlter, Phys. Rev. Lett. 51, 1888 (1983); L. J. Sham and M. Schlter, Phys. Rev. B 32, 3883 (1985).
J. P. Perdew and M. Levy, Phys. Rev. Lett. 51, 1884 (1983).
R. W. Godby, M. Schluter and L. J. Sham, Phys. Rev. Lett. 56, 2415 (1986).

GaAs band structure

8
L3,c
4
15,c
X3,c
2
X1,c
Experimental gap: 1.53 eV
L1,c 1,c
0 15,v DFT-LDA gap: 0.57 eV
Energy (eV)
L3,v
-2 X5,v
-4
-6
Applying a scissor operator
L2,v X3,v (0.8 eV) we can correct the
-8
GaAs
band structure.
-10 X1,v
L1,v
-12
1,v
L X K

DFT in practice
1 Pseudopotential or all-electron?
2 Represent Kohn-Sham orbitals on a basis (plane waves, atomic
orbitals, gaussians, LAPW, real space grid,..)
3 Calculate the total energy for trial orbitals. For plane waves:
1 kinetic energy, Hartree potential in reciprocal space,
2 xc potential, external potential in real space
3 FFTs!
4 Sum over states = BZ integration for solids: special k-points
5 Iterate or minimize to self-consistency

Software supporting DFT

Abinit EXCITING OpenMX
ADF Fireball ORCA
AIMPRO FSatom - list of codes ParaGauss
Atomistix Toolkit GAMESS (UK) PLATO
CADPAC GAMESS (US) PWscf
CASTEP GAUSSIAN (Quantum-ESPRESSO)
CPMD JAGUAR Q-Chem
CRYSTAL06 MOLCAS SIESTA
DACAPO MOLPRO Spartan
DALTON MPQC S/PHI/nX
deMon2K NRLMOL TURBOMOLE
DFT++ NWChem VASP
DMol3 OCTOPUS WIEN2k
http://en.wikipedia.org/wiki/Density_functional_theory

The code ABINIT
http://www.abinit.org
First-principles computation of material properties : the ABINIT software

project.
X. Gonze et al, Computational Materials Science 25, 478-492 (2002).
A brief introduction to the ABINIT software package.
X. Gonze et al, Zeit. Kristallogr. 220, 558-562 (2005).

Tutorial III
Ground state geometry and band structure of bulk silicon

1 Determination of the total energy
2 Determination of the lattice parameter a
3 Computation of the Kohn-Sham band structure

Equilibrium geometry of silicon
Our DFT-LDA lattice parameter:

a = 10.217 Bohr = 5.407 A
Exp. value: a = 5.431 A at 25 .

Kohn-Sham band structure of silicon
10
4 Indirect gap
2 Good dispersion of
Energy [eV]
0
~ 0.5 eV bands close to the
-2 gap
-4
Exp. gap = 1.17 eV
-6
-8
Scissor operator =
-10
0.65 0.7 eV
-12
U W X W L

Kohn-Sham band structures
Application of standard DFT to solids:

band structure calculations (Kohn-Sham bands)
We know that the Kohn-Sham bands are not quasiparticle states.
However they turn out to give a qualitative picture in many cases.
When they are completely wrong we need to go beyond standard
DFT.

Outline
6 Summary

Summary
The electron density is the key-variable to study ground-state

properties of an interacting electron system.
The ground state expectation value of any physical observable of a
many-electron system is a unique functional of the electron density .
The total energy functional EHK [] has a minimum, the ground state
energy E0 , in correspondence to the ground state density 0 .
The universal functional FHK [] is hard to approximate.
The Kohn-Sham scheme allows a reformulation in terms of
one-particle orbitals.

Suggestion of essential bibliography
W. Kohn and L. J. Sham, Phys. Rev. 140, A1133 (1965).
W. Kohn, Rev. Mod. Phys. 71, 1253 - 1266 (1999).
R. M. Dreizler and E.K.U. Gross, Density Functional Theory, Springer (Berlin,

1990).
R.G. Parr and W. Yang, Density Functional Theory of Atoms and Molecules,
(Oxford, New York, 1989).
Eds. C. Fiolhais, F. Nogueira, M. A. L. Marques, A primer in Density Functional

Theory, Springer-Verlag (Berlin, 2003).
K. Burke, Lecture Notes in Density Functional Theory,

http://dft.rutgers.edu/kieron/beta/

Suggestion of essential bibliography
Some additional items:

R. M. Martin, Electronic structure: Basic Theory and Practical Methods,
Cambridge University Press (2004).
http://www.abinit.org and references there.

DFT Presentation

Încărcat de

Informații document

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

DFT Presentation

Încărcat de

Drepturi de autor:

Formate disponibile

The many-body problem From Hartree to Thomas Fermi DFT HK theorems KS scheme and bands Summary

Key concepts in Density Functional Theory

ILM (LPMCN) CNRS, Universite Lyon 1 - France

December 12, 2012 Lyon

Key concepts in Density Functional Theory Silvana Botti

1 The many-body problem

2 Earlier solutions: Hartree, Hartree-Fock, Thomas-Fermi

3 The solution: density functional theory

5 Practical implementations: the Kohn-Sham scheme

Key concepts in Density Functional Theory Silvana Botti

1 The many-body problem

2 Earlier solutions: Hartree, Hartree-Fock, Thomas-Fermi

3 The solution: density functional theory

5 Practical implementations: the Kohn-Sham scheme

Key concepts in Density Functional Theory Silvana Botti

The many-body problem

Schrodinger equation for a quantum system of N interacting particles:

How to deal with N 1023 particles?

Ne electrons H = Tn ({R}) + Vnn ({R}) +

Key concepts in Density Functional Theory Silvana Botti

The many-body Hamiltonian

H = Tn ({R}) + Vnn ({R}) + Te ({r}) + Vee ({r}) + Uen ({R} , {r})

Key concepts in Density Functional Theory Silvana Botti

Example of equilibrium geometries

Key concepts in Density Functional Theory Silvana Botti

Key concepts in Density Functional Theory Silvana Botti

Pseudopotentials: generation criteria

A pseudopotential is not unique, several methods of generation also exist.

Key concepts in Density Functional Theory Silvana Botti

Pseudopotentials: quality assessment

It is important to find a compromise between

Key concepts in Density Functional Theory Silvana Botti

The many-body Hamiltonian

Applying the Born-Oppenheimer separation. . .

Key concepts in Density Functional Theory Silvana Botti

The many-body problem

Schrodinger Equation for a quantum-system of Ne interacting electrons:

Still, how to deal with Ne 1023 particles?

Key concepts in Density Functional Theory Silvana Botti

Why we dont like the electronic wavefunction

How many DVDs are necessary to store a wavefunction?

Classical example: Oxygen atom (8 electrons)

Rough table of the wavefunction:

10 entries per coordinate: = 1024 entries

Key concepts in Density Functional Theory Silvana Botti

1 The many-body problem

2 Earlier solutions: Hartree, Hartree-Fock, Thomas-Fermi

3 The solution: density functional theory

5 Practical implementations: the Kohn-Sham scheme

Key concepts in Density Functional Theory Silvana Botti

The Hartree equations: Hartree potential

Hartree introduced in 1927 a procedure, which he called the

If we do not restrict the sum to j 6= i: self-interaction problem

Key concepts in Density Functional Theory Silvana Botti

The Hartree equations: Hartree potential

Hartree introduced in 1927 a procedure, which he called the

If we do not restrict the sum to j 6= i: self-interaction problem

Key concepts in Density Functional Theory Silvana Botti

The approach of the best wavefunction

Key concepts in Density Functional Theory Silvana Botti

The Hartree-Fock equations: exchange potential

Key concepts in Density Functional Theory Silvana Botti

Some remarks on Hartree-Fock

The HF potential is self-interaction free.