October 17

Lectures #12 and #13

October 21

Second quantization
Evaluation of one and twotwo-particle matrix elements
Coulomb matrix elements: example of twotwo-particle
operator (second quantization form)
Helium atom: LS coupled states

Chapter 4, pages 9595-99, 102102-106, Lectures on Atomic Physics

Chapter 11, pages 241241-246, Atomic manymany-body theory
I. Lindgren and J. Morrison

First & second quantization

First quantization

Classical particles are

assigned wave amplitudes

Second quantization

Wave fields are quantized to

describe the problem in
terms of quanta or particles.

1
2

Second quantization:
atomic electrons (fermions)
One-electron state |k>
Described by the wave function k(ri)

k = ak 0
Creation operator

0 0 =1

Vacuum state
(no electrons)

Second quantization:
atomic electrons (fermions)
One-electron state |k>
Described by the wave function k(qi)

k = ak 0
k = 0 ak
ak 0 = 0 :

Vacuum state
(no electrons)

Creation operator

0 0 =1

Annihilation operator
there are no electrons to annihilate in a vacuum

0 ak = 0

Anticommutation relations

{a , a } = a a
{a , a } = a a
{a , a } = a a

i j

+ a j ai = 0

+ a j ai = 0

+ a j ai = ij

Orthonormality of one-electron states:

ak 0 = 0
0 ak = 0

i j = 0 ai a j 0 = 0 ij a j ai 0 = ij

Two-electron states
First quantizaiton:
quantizaiton:
Slater determinant
jk (q1 , q2 ) =

Second quantizaiton

1 j (q1 ) k (q2 )
2 j (q2 ) k (q2 )

jk = a j ak 0

Orthonormality of two-electron states:

ij ( q1 , q2 ) kl ( q1 , q2 ) = ik jl il jk

ij kl = 0 a j ai ak al 0
NOTE: reversed order

Need to evaluate

Orthonormality of two-electron states

ij kl = 0 a j ai ak al 0 = ?
Lets transform
Move:

a j ai ak al

Creation operators
left

to normal form
form:
Annihilation operators
right

Use anticommutation relations to move ai to the right

a j ai ak al = a j al ik a j ak ai al = ik jl al a j ik jk ai al + ak a j ai al
= ik jl al a j ik jk il + jk al ai + ak a j il ak a j al ai =
= ik jl al a j ik jk il + jk al ai + ak a j il jl ak ai + ak al a j ai

Orthonormality of two-electron states

ij kl = 0 a j ai ak al 0 = ?

Lets transform a j ai ak al to normal order
order:
Normal
order with
respect to
vacuum

Creation operators
left

Annihilaiton operators
right

Use anticommutation relations to move ai and aj to the right

a j ai ak al = a j al ik a j ak ai al = ik jl al a j ik jk ai al + ak a j ai al
= ik jl al a j ik jk il + jk al ai + ak a j il ak a j al ai =
= ik jl al a j ik jk il + jk al ai + ak a j il jl ak ai + ak al a j ai

ij kl = 0 a j ai ak al 0

Normal order of operators

Why do we want to transform to normal order?
To calculate expectation values: expectation value of
normal ordered operators is zero.

a j ai ak al = ik jl al a j ik jk il + jk al ai + ak a j il jl ak ai + ak al a j ai
0

0 a a a a 0 = ik jl 0 a a j 0 ik jk il + jk 0 a a 0

j i k l

l i

+ 0 a a j 0 il jl 0 a a 0 + 0 a a a j ai 0

k i

k l

= ik jl jk il
Exactly the result we had in first quantization

Many-particle operators in
second quantization
ai 0

one-particle state

two-particle state

States ai a j 0

ai a j

an 0

N-particle state

described
by Slater determinants
in first quantization

Operators
First quantizaiton
One-particle
operator
Two-particle
operator

F=

N
i =1

G=

f (ri )

1
g (rij )
2 i j

Second quantizaiton

F=

fij ai a j
i, j

G=

1
2

gijkl ai a j al ak
i, j

gijkl = g jilk

Many-particle operators in
second quantization
First and second quantization description must produce identical
results; however, is it more convenient to use second quantization
quantization
for calculation of properties of manyelectro
systems
as the
many
antisymmetrization properties are carried by the operators rather
than the wave functions as in the case of Slater determinants.

F=

G=

fij ai a j
i, j

1
2

gijkl ai a j al ak
i, j

f ab = a f b = d 3 r a (r ) f (r ) b (r )
g abcd = ab g cd = d 3 r1 d 3 r2 a (r1 ) b (r2 ) g (r12 ) c (r1 ) d (r2 )
The rules for calculating matrix elements are equivalent, for example:

ab G ab =

1
2

gijkl 0 ab aa ai a j al ak aa ab 0 = g abab g abba

ijkl

Examples of one and two-particle

operators
First quantizaiton
1

H0 =

Second quantizaiton

N
i =1

H0 =

hi
1
Z
hi = i2
2
ri

N
i =1

i ai ai

eigenvalue of hi

Coulomb (1/r12) interaction: two-particle operator

H'=

1
r12

H'=

1
2

gijkl ai a j al ak
ijkl

two-electron matrix element

of the Coulomb potential 1/r12

g abcd = d 3 r1 d 3 r2 a (r1 ) b (r2 )

1
c (r1 ) d (r2 )
r12

He-like systems revisited

H0 =

N
i =1

i ai ai

Two-electron state ab = aa ab 0 is an eigenfunction of

(0)
H0 with an eigenvalue Eab
= a + b:

H 0 ab = ( a + b ) ab
However, it is not necessarily an angular momentum eigenstate.
Lets construct eigenstates of {H, L2, Lz, S2,Sz } from |ab> states.

L = l a + lb
S = sa + sb
S=0: singlet state
S=1: triplet states

First, we couple la and lb to construct eigenstates

of L2, Lz with eigenvalues L(L+1) and ML
Second, we couple sa and sb to construct eigenstates
of S2, Sz with eigenvalues S(S+1) and MS
sa=sb=1/2

S=0,1

He-like systems:
construction of coupled states
ab = aa ab 0
Therefore, the coupled states are given by the sum over magnetic
moments which contains two Clebsch-Gordon coefficients:
a

lama

ab, LM L , SM S =

ma mb a b

LML

SMS

aa ab 0

lbmb

Normalization factor

ab, LM L , SM S ab, LM L , SM S = 2 1 + (1) L + S na nb lalb

1
1
2

na nb , la lb
Identical orbitals na = nb , la = lb

He-like systems:
First-order energy
a

lama

ab, LM L , SM S =

ma mb a b

LML

SMS

aa ab 0

lbmb

The first-order energy is given by

E (1) = ab, LM L , SM S H ' ab, LM L , SM S

H'=

1
2

gijkl ai a j al ak
ijkl

g abcd = d 3 r1 d 3 r2 a (r1 ) b (r2 )

1
c (r1 ) d (r2 )
r12

Coulomb matrix element

First, lets consider Coulomb matrix element

g abcd = d 3 r1 d 3 r2 a (r1 ) b (r2 )

1
c (r1 ) d (r2 )
r12

The functions are given by nlm (r ) = Rnl (r )Ylm ( , )

The 1/r12 can be expanded as

1
4 r<k k *
Ykq (1 , 1 )Ykq ( 2 , 2 )
=
r12 k =0 2k + 1 r>k +1 q = k

This expression may be re-written using C-tensors defined by

Cqk (r) =

4
Ykq ( , )
(2k + 1)

1
1
r<k k
=
=
(1) q Cqk (r1 ) Ckq (r2 )
r12 r1 r2 k =0 r>k +1 q = k

Coulomb matrix element

We now substitute the expressions for and 1/r12 back into our matrix
element and separate dr and d integrals

r<k k
1
=
(1) q Cqk (r1 ) Ckq (r2 )
r12 k =0 r>k +1 q = k

g abcd = d 3 r1 d 3 r2 a (r1 ) b (r2 )

k =0

1
c (r1 ) d (r2 )
r12

dr1 r12 dr2 r22 Rnala (r1 ) Rnblb (r2 )

0

nlm (r ) = Rnl (r )Ylm ( , )

Rk (abcd )

r<k
Rn l (r1 ) Rnd ld (r2 )
r>k +1 c c

radial
integral

d 1Yla ma (1 , 1 ) C kq (1 , 1 )Ylc mc (1 , 1 )

la ma C kq lc mc

d 2Ylb mb ( 2 , 2 ) C k q ( 2 , 2 )Yld md ( 2 , 2 )

lb mb C k q ld md

(1) q

q = k

Coulomb matrix element

g abcd =

k =0

Rk (abcd )

q = k

(1)q la ma C kq lc mc lb mb C k q ld md

Next, we use Wigner-Eckart theorem for both of the matrix elements:

g abcd =

k =0

Rk (abcd )

q = k

lama

(1)q

kq
lcmc

lbmb

We use

k -q
ldmd

lbmb

k -q

la C k lc lb C k ld

ldmd

lbmb

= (1)k q

kq

Note: k and q are integers

ldmd

Coulomb matrix element

g abcd =

lbmb

lama

k =0

Rk (abcd )

(1) k

q = k

la C k lc lb C k ld

kq

kq

ldmd

lcmc

lama

k =0

(1) k Rk (abcd ) la C k lc lb C k ld

lbmb

k
lcmc

ldmd

X k (abcd )
lama

g abcd =

lbmb

k =0

lcmc

X k (abcd )

ldmd

Summary: Coulomb matrix element

(non-relativistic case)
lama

g abcd =

lbmb

k =0

lcmc

X k (abcd )

ldmd

X k (abcd ) = (1) k Rk (abcd ) la C k lc lb C k ld

Rk (abcd ) = dr1 r12 dr2 r22 Rnala (r1 ) Rnblb (r2 )

l1 C k l2 = (1)l1 (2l1 + 1)(2l2 + 1)

r<k
Rn l (r1 ) Rnd ld (r2 )
r>k +1 c c

l1 k l2
0 0

10

Back to He-like systems

a

lama

ab, LM L , SM S =

LML

ma mb a b

SMS

aa ab 0
Assume
la=la
lb=lb

lbmb

E (1) = ab, LM L , SM S H ' ab, LM L , SM S

a

lama

=2

LML

ma mb a b ma 'mb ' a ' b '

g a 'b ' ab a a ' b b ' g a 'b 'ba a b ' b a '

H'=

1
2

SMS

lbmb

lama

lbmb

gijkl ai a j al ak ;

LML

SMS
b

ab G ab = g abab g abba

ijkl

Sum over s: Term 1

Term 1:
a b a ' b '

a a ' b b '

= (2S + 1)

SMS

SMS
b

SMS

SMS
b

=1

We used orthogonality relation to evaluate this diagram

j1
j3 m3

j3m3

1
j ' j m 'm
2 j3 +1 3 3 3 3

j2

11

Sum over s: Term 2

Term 2 :
a b a ' b '

a b ' b a '

SMS

SMS

= (2S + 1)(1)

SMS

S +1

SMS

= (1)

Extra phase factor

First-order energy: He-like systems

lama

E (1) = 2
ma 'mb ' ma mb

lama
LML

lbmb
lama

Assume
la=la
lb=lb

LML
lbmb
lama

lbmb

k
lama

X k (abab) + (1)S

+
k
lbmb

lbmb

lbmb

X k (abba ) )

lama

where we substituted Coulomb matrix elements g into our formula

lama

g abcd =

lbmb

k =0

lcmc

X k (abcd )

ldmd

12

Sum over ms: Term 1

lama

lama
LML

ma 'mb ' ma mb

LML

lbmb

lama

lama

1) Remove triangle (get 6-j symbol)

LML

2) Use orthogonality relation

lbmb

lb
la

la
lb

lbmb

rotate
to the right

LML

= (1) L + k +la +lb

lama

lbmb

+
k

lama

lbmb

lbmb

lbmb

lama

=(2L+1)

lama

(get factor 1/(2L+1)

lbmb

L
k

Sum over ms: Term 2

lama

lama
LML

ma 'mb ' ma mb

lbmb

= (1) k +la +lb

la
la

lb
lb

lama
LML

lbmb

lbmb

k
lbmb

lama

L
k

The calculation is very similar to Term1, only need to switch

a and b lines in the first angular diagram.

13

ab, LM L , SM S

Final expression:
First order energy for He-like system
E (1) = 2

(1) L + k +la +lb

k

+(1) S + k +la +lb

la

lb

lb

la

X k (abab)

la

lb

la

lb

X k (abba)

X k (abcd ) = (1) k Rk (abcd ) la C k lc lb C k ld

Rk (abcd ) = dr1 r12 dr2 r22 Rnala (r1 ) Rnblb (r2 )

l1 C k l2 = (1)l1 (2l1 + 1)(2l2 + 1)

r<k
Rn l (r1 ) Rnd ld (r2 )
r>k +1 c c

l1 k l2
0 0

Special case : if la = lb = 0 ( s states ) s C s = k 0

k

0 0 0
0 0 0

=1

14