Anda di halaman 1dari 6

PHYSICAL

REVIEW

VOLUM E

140,

NUM B ER

4A

15

NOVEM B ER

1965

Self-Consistent

Equations

Including

Exchange

and Correlation

Unieersity

W. KOHN

of Ca/Bfornia,

AND

L.

J. SHAM

San Diego, la Jolta, California

(Received

21 June

1965l

Effects*

From a theory of Hohenberg and Kohn, approximation methods for treating an inhomogeneous system

of interacting

electrons are developed.

These methods

are exact for systems

of slowly varying

or high density.

For the ground

state,

they

lead to self-consistent

equations

analogous

to the Hartree

and Hartree-Fock

equations,

respectively.

In these equations

electron

gas appear

the exchange

effective

and correlation

portions of the chemical potential

of a uniform

as additional

potentials.

(The exchange portion of our effective

potential

differs

from that

due

to Slater

by

a factor

of -';. ) Electronic

systems

at finite

temperatures

and in

magnetic

lelds

are

also

treated

by

similar

methods.

An

appendix

deals

with

a further

correction for

systems with short-wavelength density oscillations.

I. INTRODUCTION

 

'N

recent

years

a great

deal

of attention

given

to the problem

of a homogeneous

has been

- -

gas of inter-

acting electrons

and its properties

have been established

 

with a considerable

degree

of confidence

over

a

wide

range of densities. Of course, such a homogeneous

gas

 

represents

only

a mathematical

model,

since

in

all

real

systeins

(atoms,

inolecules,

solids,

etc. )

the

electronic

density is nonuniform.

It is

then

a matter

of interest

of the homogeneous

gas

to

see

how

in theoretical

properties

can be utilized

studies of inhomogeneous systems. The well-known methods of Thomas-Fermi' and the Slater' exchange

 

hole

are

in

this

spirit.

In the present

paper

we use

the

formalism

approach

of

Hohenberg

and

further

and

we obtain

Kohn'

carry a set of self-consistent

to

this

equations

which

include,

in

an

approximate

way,

ex-

change

and

correlation

effects.

They'

require

only

a

knowledge

of

the

true

chemical

potential,

tie(e),

of

a

homogeneous

interacting

electron

gas

as a function

of

 

the density

n.

We

derive

two

alternative

sets

of

equations

[Eqs.

(2.8) and

(2.22)) which

are

analogous,

respec-

tively,

to

the

conventional

Hartree

and

Hartree-Fock.

equations,

and,

although

they

also include

correlation

effects, they The local effective

are

no more

difficult

potentials

to solve.

in these

equations

are

 

unique

in

a sense

which

is described

in

Sec. II.

In par-

ticular,

we And

that

the Slater exchange-hole

potential,

besides

its

by a factor

omission

'.

of —,

of correlation

effects,

is

too

large

Apart

from

work.

on

the

correlation

energy

of

the

homogeneous

electron

gas, most theoretical

many-body

 

studies

have

been

concerned

with

elementary

excita-

tions

and

as a result

there has been little recent progress

 

in

the

theory

of cohesive

energies,

elastic

constants,

etc. ,

of

real

(i.e. , inhomogeneous)

metals

and

alloys.

The

methods

proposed

here

offer

the

hope

of new

progress

in

this

latter

area.

~

Supported

in part

by the

U. S. Ofhce

of Naval

Research.

'L.

H.

Thomas,

Proc.

Cambridge

Phil.

Soc. 23, 542 (1927);

E.

'

Fermi,

Z. Physik

48, 73 (1928).

'

J.

C. Slater,

Phys.

Rev. 81, 385 (1951).

 

P.

Hohenberg

and

W.

Kohn,

Phys.

Rev. 136, 3864 (1964l;

referred

to hereafter

as HK.

In Secs. III and IV, we describe

the necessary

Inodid-

cations

to

deal

with

the

finite-temperature

properties

and with

the spin paramagnetism

of an inhomogeneous

electron

gas.

Of course,

the

simple

posed in general

involve

methods

which

are here

pro-

errors. These are of two general

origins4:

a too rapid

variation

of density

and,

for 6nite

systems, boundary effects. Refinements aimed at re-

ducing the

6rst type

of error

are

brieQy

discussed

in

Appendix

II.

II. THE

GROUND

STATE

A. Local Effective Potential

It has been shown'

that the ground-state

energy

of an

interacting

inhomogeneous

electron

gas

in

a static

po-

tential

n(r)

can be written

in

the

form

1

e(r)e(r')

Z=

tt(r)e(r) dr+

r

r'[—

dr

dr'+G[e),

i:, ,

(2.1)

where

e(r) is the

density

and G[e) is a universal

func-

tional

of the density.

This expression,

furthermore,

is a

minimum

for the correct density

function

section

we

propose

first

an

approximation

e(r). In this

for G[e),

which

leads

to a scheme

analogous

to Hartree's

method

but

contains

the major

part

of the effects of exchange

and correlation.

We first write

G[e) = T.[e)yZ,[e),

(2.2)

where

T,[e) is the kinetic

interacting

electrons

with

energy

density

of a system

e(r) and

of

F,[e) is,

non-

by our definition,

of an interacting

the exchange

and. correlation

system

with

density

energy

e(r). For an arbi-

trary

e(r),

of

course,

one can

pression for E,[e). However,

give

if

no

simple

exact ex-

e(r)

is

sufliciently

 

slowly varying, one can show' that

F,[e)=

e(r)e,(e(r)) dr,

(2.3)

4 W. Kohn

and L. J. Sham,

Phys.

Rev. 137, A1697 (1965).

~ For such

a system

it follows

from

HK that the kinetic

energy

is in

fact

a unique

functional

of the density.

1138

A &&34

 

AN

where

&x. (

electron

o

istheexcha

e

)

.

f a uniform

n

g

iectron

e e

e and

gas

prrelation

corre

ene

f densltyn.

O

gy

P

el

ur sole

l

apprpxima

matipn

con

ts

pf

assum~~g

that

2 3) consti-

tutes ana

e

q

uaterep

.

resentatipn

pf exchang

o

.

nd corre-

lation

e

cts

in

the

s

sy

stems

un

e

consl

era

tion.

e

 

I,npwn

n

from

theories

 

the

homo-

gene

eiectron

'

g

'

From

t

h

statio

ary

prop

rty

of Eq'

 

2.1) we now

obtain,

subject to th,e ,o„dition

 
 

be(r) dr=0,

 

(2.4)

the equation

 

here

be(r)

q(r)+

t(„. n

())

r

 

d

=,

0

(

2.

)

 

~(r) = ~(r)+

e(r')

dr',

 

(2 6)

and

 

'

is

the

exc

correlation

c

 

to

th

o

e obtain

 

om

t

e

po

Therefore

or

g

p~

h&ch

sat&sfies

these

e

uat

 

si

a

nd setting

 
 

e(r)=Z

I

'(r) I',

'

 

i=1

 

iv

is the

num

er

o

E. (2.8)

as

an

a

ii

dd't'

l

ff

g

familiar

Equatio

ns

from

(2.

.

6—

(2.9)

)

.

h

to

o

 

ne

 

begins

wi

ed

mr,

 
 

2. 6) an

(i

pxc

i'

o

(2.7,

and fin

s ane

 

f

f

(2.8) and

.

Z=P c,

——

(2.9 .

2.9

. The energy

is g'

is

is

iven

 

y

 

1

ir

r'

n

r

e,(n(r))

e

n

r

t(.(e(r))g

.

r

dr.

r.

(2.10)

The results

of ou r

p

roce du re are exactct inin twow

limiting

 

cases.

 

'

'

ensi

.

ime

 

is

c

aracter-

ized

by

n

see

D

..

Inc.

, ,

P',

Pines,

r,

ro~(1, where

r

elementary

w

ew

r, is

'

the

'

i

6F r

a review

(W. A. Benjamin,

in

x

'n

Solzds

SH~M d. ro» a

is

it.z, ra

ius

n

an

appreciab

l

e

c

yp

hange

in

ical lengt

th over which

'

dens

sit

ty

. In

t»s

se

as

 

can expan

nd

the

true

exc

hange

and

 

ene

gy

a

F,fej=

a .(e)e. dr

 

c

('&(e))Vn('dr+

~

~

~,

(2.11)

e

and

correlation

p ortion

o

the

n

rg

xpansi ansion

in powers

o

expand

T, e

in

T,Ln) =

,',

(3x'e)'"e dr

 
 

t('&(n)iVei'dr+

 

.

(2

..12)

2

 

From

HK expecially

Sec.III 2,

Sec.

,

we hhave the following

h

expression

for

o

the energy.

 

)

+

   
 

+

go(e) dr+

2&'&(e)

g2

[Ve)'dr+

 

(2.13)

 

3m'e)'"+e .e

e)}e,

(2.14)

6, ('& (n)+t(" (n) ) e.

 
 

'g

'y

 

of the

order

 

l

ou

re

determines

 

t

e

e

 

of

o

d

ivi4.

 

pp

 

d

and that

 

g

 

'

'

ter's

exc ange

 

begin

by

writing

w

f

eHartree-

the

-F

equivalent

po

oc

k

en

h

o erator

in

the form

function

 
 

N

.*(r)A '(r')4' (r)A(r'

 

w.,(r)=- Z

[r r'[

 

A*(r)4'(r)

 

( 2

..16)

ter

there

 

have

been

 

m

ts

to

add

corre a

'

R

Schrieffer

P

y

Rev. 136,

 

s

9

 

on

Phys.

 

S

and

C

W. UGord,

Phys. R

139

 

SELF CONSISTENT

EQUATIONS

 

A 1135

where

the

symbols

r

and

r' are understood

to include

an atomic

nucleus,

and

(2) it does

not

lead

to quantum

electron spin coordinates and integration is understood

density

oscillations, 4 such

as

the

density

fluctuations

to include

summation

over spin

coordinates.

One next

due

to atomic

shell

structures.

By not making

the

re-

assumes

that

the wave

functions

can be approximated

 

placement

(2.20), we avoid both of these shortcomings.

by plane

waves

which

results

in

 

Let us now

qualitatively

discuss

the appropriateness

kr(r)

w„i, (r) =

1+

krs(r)

k—

s

ln

k+kr

()

r

(2.17)

of

our

systems.

procedure

for

various

classes

of

electronic

2kkr

(r)

 

k kr(r)

 

In

atoms

and

molecules

one

can

distinguish

three

   

regions:

(1) A

region

near

the

atomic

nucleus,

where

where

kr(r)—

=

{3s rt(r)}Us.

Finally,

one

averages

 

v

&

the electronic

density

is high

and

therefore,

in view

of

over the occupied

state k, which

results

in

 

case

(h)

above,

we

expect

our

procedure

to

be

satis-

 

n„(r) = (3/2') {3s'I (r) }'"

 

(2.18)

factory.

(2) The main

"body" of the charge distribution

where

the

electronic

density

n(r)

is relatively

sjowly

In our procedure

(neglecting

correlation)

we obtain,

 

in

varying,

so that

our

approximation

(2.3) for

e, is

ex-

place of Slater's

v

 

pected to be satisfactory

as discussed

in case

(a) above.

     

(3) The "surface" of atoms

and

the overlap

regions

in

smaller

t *()=

(

/~){ ~'I()}'"

(2

)

by a factor

of

',

.

From the discussion in Appen-

mole cules. Here our approximation

validity

and

therefore

we expect

this

(2.3)

region

to

has

be

no

the

dix I, it follows

that

while

gives the exchange

correc-

main

source

of error.

We do not expect an accurate

de-

tion

of the density

p

correct to order

V~', inclusive, s

scription

of chemical

binding.

In large atoms, of course,

 

~

t as indeed

any other function

order

~ V ~'. The same comment

of e(r)] leads

to errors of

applies

to any extension

this "surface" region

becomes

of less importance.

(The

surface is more satisfactorily handled in the nonlocal

of Slater's

exchange

to include

 

correlation

in

the

self-

method described under

8 below. )

 

consistent

potential.

 

For metals, alloys, and small-gap

insulators

we have,

We may

note that

our result

 

is equivalent

to taking,

of course,

no

surface

problem

and

we

expect

our

ap-

not

the

average

of

(2.17), but

rather

its

value

 

at

proximation

(2.3) to

give

a good representation

of ex-

k=kr(r);

i.e. ,

the

effective

exchange

potential

for

a

change

and

correlation

effects. In large-gap

insulators,

state

at

the

top

of

the

Fermi

distributions.

This

 

is

however,

the

actual

correlation

energy

will

be

con-

physically understandable

since

density

adjustments

siderably

reduced

compared

to that

of a homogeneous

come about

by redistribution

 

of the electrons

near

the

electron

gas

of the

same

density.

Fermi level.

 

(b) High

dertsity.

This regime

is characterized

by

the

 

B. Nonlocal

Effective

Potential

condition

r,/a,((1, where

as

is

the

Bohr radius.

In this

   

case,

the

entire

exchange

and

correlation

 

energy

 

is

Instead

of the

Sec. IIA it is also

Hartree-type

procedure

discussed

to obtain a scheme

which

in

in-

smaller

(r,/as)

than

the

and

hence

kinetic

energy

our inaccuracy

by

a factor

of order

in representing

these

cludes

exchange

Eq. (2.3)

possible

effects

exactly.

We

write

in

place

of

portions

becomes

negligible.

   

The reader

will have

noticed

 

that

while

in Eq.

(2.3)

 

F,PN)=E Lrtj+

rt(r)e, (e(r)) dr

(2.21)

we approximate

the exchange

and. correlation

energy

by

 

the expression valid for a slowly varying

density,

we

made

no approximation

for

the

kinetic-energy

func-

where

Z

LNj

is

the

exchange

energy

of

a

Hartree-Fock

tional

T, Lnj of Eq. (2.2). This procedure

is responsible

system

per particle

of

density

I (r) and e,(m) is the correlation

energy

for the exactness

of the high-density

limit,

even

when

of a homogeneous

electron

gas. Applying

the

density

is rapidly

varying,

 

such

as

in

the

vicinity

this

ansatz

in conjunction

with

Eq.

(2.2) and

the

sta-

of an atomic nucleus.

 

tionary

property

of (2.1) leads

to the following

system

We now Inake

a few further

 

remarks

about

our

ap-

of equations:

 

proxirnation.

If

in

Eq.

(2.2), we

had

approximated

 

T, LN)

by

its form

appropriate

 

to

a system

of slowly

varying

density,

 

Ni(r, r')

 

T,Lm| ~

, ', (3x'I)'t'rt

dr,

 

(2.20)

where

lr r'I

,

4'(r')

d~= e'4'(r),

(2 22)

we

would

have

been

led

to

the

generalization

 

of the

p. = d (rte.)/dN,

(2.23)

Thomas-Fermi

method

shares

method

with

the

suggested

by

Thomas-Fermi

Lewis.

method

This

two

I

I (r r )=Z It't(r)A*(r

)

(2.24)

shortcomings:

(1) It

leads

to

an

in6nite

density

near

j~1

'

  • H. W. Lewis, Phys. Rev. 111, 1554 (1958).

 

and

p(r), rt (r) are defined

as before, Eqs. (2.6) and (2.9).

 

&&36

 

K

AND

The energy

is now

 
 

1V'

e(r)e(r')

 

E=P p,

——

 

1

 

1

ei(r, r')ei (r',r

 
 

e

p

1'

qg p

&o

(e(r)) pp(e( ))

dr.

(2' )

This

procedure

a

may

beregar

e

as

a Hartree-Foc

k

eth()d

( orrected

o

Beets.

It

is

no

mo

 

plicated

 

than

the nn

t d Hartree-Foc&»e

 
 

1

tor appearing»

 

l

described

in

 
 

n.ow

 

1

we

must

expect,

 

more

accurate resu ts

a

from

rticular,

near

the

e surface

the m

o

f

an

atom

the

ctive

effec iv

i

p

proaches

zero

muc hf

str . E

 

n

here,

howeve,

 

correlation

e

ec s

 

tl

described

near

t

e

surface.

 
 

III. FREE EENERGY; SPECIFIC HEAT

 
 

tion

 

o

gr

re ensembles

by using

temperature

'

generalization

He has

shown

that

t

a

o

E

q

.

the

gran

(2.1)

can

t

h

i

e

6

ni

m

b

tential

can

be wri

en

 

+G[e7

)

 

e(r) dr,r

3.1)

 

ni

ue

functiona 1 of

o

thee density

at

a

 

'

 

gi

"

p

erature

g and

p

is t

h

e c

h

e

1

rrect

n

this

quan

i

y

In analogy

witith

(2 2)

e now

 

write

 

here

G[e7=G,[e7+F,[e7;

 

(3.2)

 

G,[e7=

T,[e7 rS, [e7,

(3.3)

where,

energy

a

Ln

n

(1

t,

~

ar

are

respective

of noninteracting

el

y, 1

e

the

to

kinetic

th

R

 

py

 

tion,

ge

an

corre

1

a

tion

o

 

tib

to

free

energy.

 

FFor

tte

approximation

 

~-[ 7=

()f*.( ())d,

 

(3 4)

where,

n

an

e

and

correlation

conntribu-

 

ti

on to the

free energy

perr

electron

e

o auniUo

 

1

9 N. D. Mermin,

s. Rev. Rev.

h

Phys.

137, A1441 (19651.

 
L' I gHAM as of de»ity n (3.5) 0 ~here f and are are the free
L'
I
gHAM
as of de»ity
n
(3.5)
0
~here
f and
are
are the
free energie
er elect.ron
of an
fp
P
~
nd noninteracting
t
gas
respective
y.
.
g
1
( )+ (bG [e7/be(r))+I)~p(e
is given
iven,
as before,
b
Fq. (2.6)
and
"-"
)=q( f ())/d".
(3.7)
'
Equation
3
6) is identlca
]. to
the
corresPon
ndln
eq
~
t'
'
ion
for
a, system
o
f noninteracting
1"""
electr()ns
in
the
potent
+~-'
effective
1
tion
is therefo
W
.
ter»ined
by
t
tern of equat
o
{
i~2+ &(r)+pxo(e(r)))~'
~,
(3.8)
N
n&=
i 2/{
(e( —
p)/pe+ 1)
(3.9)
j, =1
is determined
d as usua
b
th
t talnumbero
r of
a,rticles
p
y
P
roxi-
'
teracting
allo
s
is
the
low-
p
n
ex
ansion,
in
g
po
A
ut
more
con-
1
ad
namics
8G
5 e7=
(0+—1)c—
'(T)—v=-
~()+
BT
be(r)
()e(r)
(BG[e7
(3.10)
The integral
i
vanishishes
ish
because
of thee stationary
property
of 0, so that
5[e7= — (()G[e7/
~~
3.11
)
n(r), V ~
)
ment,
applied
to
a system
of noni
oninter-
'
'
Coinbining
o
Eqs. s.
(3.11, 3.11,
3.12),
3.
(3.2),
and
(3.4),
we
obtai n
5
e7= 5 [e7+
e(r) (Bf
(n
n
87'
87
~(,),
vdr.
vdr.
(3.13)
(
)
'
For small
r it is we
ell known
that
5,
is
g'
iven by
5,[e7= Ã-', x'k'~g, (p),
(3.14)
le-
article
densityit
o
of
states
in
the
(bf
..()/~
).„
.. =-'ir'k'r[g . (d(o, (e)) —
=-.
-
e
gp(((p
(3.15)

SELF

Co N

S1STENT

EQ UATIONS

where

„( )

phi

and

Po(n)

'

respective

'

y r

the

chemical

tentials

of an

teracting

and

a n

teract»g

 

h

s of density

~

d

and

go

are

t- he respective

 

It follows immediately

t

a

low

 

P

erature

 

heat

capacity

is given

by

C,

, = yv-,

(3.16)

where

 
 

=-'ee Xg Q)+

«)(g6 ~(n)) ap(t p(n))) «

n(.

 

We

shall

no