many electrons atoms_2012.12.04 (pdf with links
DESCRIPTION
From the course PHYS261, University of Bergen. This is a PDF with links. The links do not work here on slideshare preview.TRANSCRIPT
Many electron atoms
dV =
x − r
( )xρ
x
r
d x3
Autumn 2012Version: 04.12.2012
List of topics
(1)
Pauli-principle Hund’s rule 1 slide + 3 slidesAntisymmetric functions for n-particles - Slater Determinants 3 slidesFilling Shells Figure 2 slidesIonization energies - Figure 2 slidesHartree - Selfconsistent field - Iterations 8 slides
only Iterations part 4 slidesSelfconsistent field Result - Figures 3 slidesConfigurations - Ionization potentials of atoms - Tables 3 tablesScreened potential and Centrifugal barrier - Explains Tables 2 slidesEnergy for N-particles - Using Slater Determinants 5 slides
Helium example 1 slideLithium example Counting nonzero terms 2 slides
N particles - Energy Summary 1 slideSchrodinger equation from variational method 1 slideVariational method - deriving Hartree-Fock Equations 2 slidesHartree-Fock Equations 1 slideTotal energy and the selfconsistent orbital energies 3 slidesEvaluation of electron repulsion Configuration mixing 2 slides + 2slides
2
1 Filling up the shells with electrons
In order to know and to, more or less, understand in which state an electronis bound we can use some basic rules. The generel idea is that the lowestenergy-state is the most stable one. The exited states can in most cases fall inthe lower state or ground state by emission of light.The principles one need to know how to fill up the shells are the Pauli-principleand the Hund’s rule.
List of topics
3
2 Pauli-principle
The filling of the shells observed was explained by Pauli by formulating thePauli exclusion principle:
Two electrons can not be in the same state - defined here by possible quantumnumbers n, l, m and ms
n, the principal quantum number of the shell,
l, the angular momentum quantum number of the subshell,
m, the magnetic quantum number
ms denotes spin ”up” and spin ”down”, ms = 1/2 or ms = −1/2
The Pauli principle mathematical: require the multiparticle wavefunc-tion antisymmetric with respect to exchange of two particles.
If two particles coordinates are exchanged, the function must change sign.
If two particles are in the same state (function),
this requirement leads to zero wavefunction, thus impossible.
List of topics
4
Figure 1: Atomic levels (shells)List of topics
5
Figure 2: Atomic levels (shells) filled up to the element ....List of topics
6
3 Antisymmetric product function for n-particles
Ψ(x1, x2, ...xn) =∑
perm(α,β,...ν)
(−1)P (perm(α,β,...ν))perm (φαφβ....φν) (x1)(x2)....(xn)
where each term in the sum looks as φβ(x1)...φν(x2)...φα..., summing over allpermutations, and P (perm(α, β, ...ν)) is the number of swaps of the givenpermutation perm(α, β, ...ν)
This is very close to the definition of the determinant
det(A) =∑
σ∈Sn
(sgn(σ)
n∏
i=1
Ai,σ(i)
)
The above in this notation
Ψ(x1, x2, ...xn) =∑
σ∈Sn
sgn(σ)n∏
i=1
φασ(i)(xi)
List of topics
7
Slater determinant
The antisymmetric combination for n-particles can be written as a determinantin this way: ∣∣∣∣∣∣∣∣
φα(x1) φα(x2) ... φα(xn)φβ(x1) φβ(x2) ... φβ(xn)... ... ... ...
φν(x1) φν(x2) ... φν(xn)
∣∣∣∣∣∣∣∣
For 3 particles:
3 particle Slater determinant
∣∣∣∣∣∣
φα(x1) φα(x2) φα(x3)φβ(x1) φβ(x2) φβ(x3)φγ(x1) φγ(x2) φγ(x3)
∣∣∣∣∣∣List of topics
8
3 × 3 determinant
∣∣∣∣∣∣∣
φα(1) φα(2) φα(3)φβ(1) φβ(2) φβ(3)φγ(1) φγ(2) φγ(3)
∣∣∣∣∣∣∣
φα(1) φα(2) φα(3)
φβ(1) φβ(2) φβ(3)
φγ(1) φγ(2) φγ(3)
φα(1) φα(2) φα(3)
φβ(1) φβ(2) φβ(3)
φα(1) φβ(2) φγ(3) + φβ(1) φγ(2) φα(3) + φγ(1) φα(2) φβ(3)
− φγ(1) φβ(2) φα(3) − φα(1) φγ(2) φβ(3) − φβ(1) φα(2) φγ(3)
List of topics
9
4 Hund’s rule
Hund’s rule is the manifestation of the same effect as we have seen in theparahelium - orthohelium effect ( the ”second” rule )
Hund’s first rule
Full shells and subshells do have a total circular momentum of zero.
This can be calculated and is allways valid.
Hund’s second rule
The total spin S should allways have the highest possible value. So as manyof the single electron spins as possible should be parallel.
The second rule appears more empirical and applies to a different magneticquantum number of the electrons with parallel spins. If is of course not allowedto break the Pauli-principle.
List of topics
10
5 Example for the Hund’s rule
Figure 3: Electron spins of the first elements.
The elements carbon C, nitrogen N and oxygen O are those where the Hund’srule have the biggest influence. We see electrons with parallel spins in thestates m = −1, 0,+1, depending on the element. The magnetic quantumnumber m gives more or less the “direction” of the circular moment l. Thes-states are the states where l = 0 and the p-states those where l = 1. Thenames for l = (2, 3, 4, ...) are (d, e, f, ...).The shell-name of the shell n = (1, 2, 3, 4, ...) is (K,L,M,N, ...).
List of topics
11
In cases with more electrons do we get some exceptions. For example is the4s subshell earlier filled than the 3d. This is caused by the smaller distanceof the 4s to the core and thus by the lower energy, which is more stable. Thementioned rules work well for general considerations and for atoms with notto many electrons.
6 Number of states
Usefull to know is the largest possible number of states for a given n whichmeans until a special shell is filled.
Nmax = 2 ·n−1∑
l=0
(2l + 1) = 2n2. (2)
We get this formula by adding all the possible quantum number configurations:We get for each n every l in the range (0, 1, ..., n− 1),for each l every m in the range (−l, ...,−1, 0, 1, ..., l)and for each of these states 2 spin possibilities.
List of topics
12
7 Ionization energies
Figure 4: Ionization energies. The shell properties would explain the structurein general - Periodic table and the Selfconsistent field List of topics
13
Especially at the first elements, we see the minimums at the elements withjust one electron in the last shell and a maximum at the noble gas elements.The weak maxima are caused by filled subshells or by the maximum of parallelstanding spins. After argon, it is more difficult to observe general tendences.
The shell properties would explain the structure in general
But there should be no closed shell at argon
Details - Periodic table and the Selfconsistent field
List of topics
14
8 Hartree - Selfconsistent field
Interaction energy of two charges depends on their distance|~r − ~x|:
W (|~r − ~x|) =q1q2|~r − ~x|
The two charges are an electron and a little volume dV at ~x containing charge cloudof density ρ
q1 → (−e) q2 → ρ(~x)dV → ρ(~x)d3x
The interaction energy of these twocharges is
dW (|~r1 − ~r2|) =(−e)ρ(~r2)
|~r − ~x|d3r2
dV =
x − r
( )xρ
x
r
d x3
List of topics
15
Interaction with a cloud; summing over all the small volume elements - it meansintegrating over the whole volume of the cloud gives the potential energy
W (~r) =
∫(−e)ρ(~x)
|~r − ~x|d3x
dV =
x − r
( )xρ
x
r
d x3
List of topics
16
If the charge cloud represents one elec-tron in state ψi(~x)
ρ(~x) = (−e)|ψi(~x)|2
If we have N electrons, each in its state,the total density becomes
ρ(~x) = (−e)N∑
i=1
|ψi(~x)|2
and again integrating gives the poten-tial energy due to the interaction witha (probability based density) cloud ofelectrons
W (~r) =
∫ (−e)2N∑i=1|ψi(~x)|2
|~r − ~x|d3x
dV =
x − r
( )xρ
x
r
d x3
List of topics
17
Now solving the Schrodinger equation with W (~r),
(T + V +W )ψi(~x) = Eiψi(~x)
We first need to know the W (~r), but that depends on all the other N solutions
W (~r) =
∫ (−e)2N∑i=1|ψi(~x)|2
|~r − ~x|d3x
Approximation chain: First we choose some simple approximation, e.g. the hydrogen-like states, or we might know the states foranother atom. We call it
ψ(0)i (~x)
From the set of all N ψ(0)i we construct
W (1)(~r) =
∫ e2N∑i=1|ψ(0)i (~x)|2
|~r − ~x|d3x
In atomic units the whole Schrodinger equation is(−1
2∇2 − Z
r+W (1)(~r)
)ψ(1)i (~x) = E
(1)i ψ
(1)i (~x)
List of topics
18
1. step - choose arbitrary set of ψ(0)i
ψ(0)i (~x) → W (1)(~r) =
∫N∑i=1|ψ(0)i (~x)|2
|~r − ~x|d3x (3)
(−1
2∇2 − Z
r+W (1)(~r)
)ψ(1)i (~x) = E
(1)i ψ
(1)i (~x)
2. step: take the set of ψ(1)i from the 1. step
ψ(1)i (~x) → W (2)(~r) =
∫N∑i=1|ψ(1)i (~x)|2
|~r − ~x|d3x
(−1
2∇2 − Z
r+W (2)(~r)
)ψ(2)i (~x) = E
(2)i ψ
(2)i (~x)
List of topics
19
3. step: take the set of ψ(2)i from the 2. step
ψ(2)i (~x) → W (3)(~r) =
∫N∑i=1|ψ(2)i (~x)|2
|~r − ~x|d3x
(−1
2∇2 − Z
r+W (3)(~r)
)ψ(3)i (~x) = E
(3)i ψ
(3)i (~x)
4. step: take the set of ψ(3)i from the 3. step
ψ(3)i (~x) → W (4)(~r) =
∫N∑i=1|ψ(3)i (~x)|2
|~r − ~x|d3x
(−1
2∇2 − Z
r+W (4)(~r)
)ψ(4)i (~x) = E
(4)i ψ
(4)i (~x)
List of topics
20
5. step: take the set of ψ(4)i from the 4. step
ψ(4)i (~x) → W (5)(~r) =
∫N∑i=1|ψ(4)i (~x)|2
|~r − ~x|d3x
(−1
2∇2 − Z
r+W (5)(~r)
)ψ(5)i (~x) = E
(5)i ψ
(5)i (~x)
6. step: take the set of ψ(5)i from the 5. step
ψ(5)i (~x) → W (6)(~r) =
∫N∑i=1|ψ(5)i (~x)|2
|~r − ~x|d3x
(−1
2∇2 − Z
r+W (6)(~r)
)ψ(6)i (~x) = E
(6)i ψ
(6)i (~x)
List of topics
21
Iteration chain
ψ(n)i (~x) → W (n+1)(~r) =
∫N∑i=1|ψ(n)i (~x)|2
|~r − ~x|d3x → ψ
(n+1)i (~x)
This chain can continue, until the set of ψ(n)i produces a potential W (n+1) which
is the same as W (n), which was the one to determine ψ(n)i . The potentials and
functions become consistent, hence the name Selfconsistent field.
Criterium for self-consistency: the (n+1)-th solution does not differ from th n-thsolution ∫ N∑
i=1
∣∣∣ |ψ(n+1)i (~x)|2 − |ψ(n)
i (~x)|2∣∣∣ d3x < ε (4)
where ε ∝ 10−8 as a typical value.
List of topics
22
9 Ionization potentials of atoms
!"##$%
&'(#)%
!"# $%#& '()& *(+,-./)%0-(+
1 2 2"3)(.&+ 1456 7189
: 2& 2&;-/# :<5=>
4 ?- ?-0@-/# =54> A2&B 7:89
< C& C&)";;-/# >54: A2&B
= C C()(+ D54 A2&B 7:E9
6 * *%)F(+ 115:6 A2&B
G $ $-0)(.&+ 1<5=4 A2&B
D H HI".&+ 1456: A2&B
> J J;/()-+& 1G5<: A2&B
1K $& $&(+ :15=6 A2&B
11 $% !(3-/# =51< A$&B 7489
1: L. L%.+&8-/# G56= A$&B
14 M; M;/#-+/# =5>> A$&B 74E9
1< !- !-;-'(+ D51= A$&B
1= N N@(8E@()/8 1K5<> A$&B
16 ! !/;,/) 1K546 A$&B
1G *; *@;()-+& 1:5>G A$&B
1D M) M).(+ 1=5G6 A$&B
OOP O O OOOQA&RB O O )
O O O ) O ) )
O O O ) *%+,: O ) )
O O O O ) )
O O O *-+,: O ) )
O O O *-+,: O )
O O O *-+,: *-.,: O )
O O O *-+,: *-.,4 O )
O O O *-+,: *-.,< O )
O O O *-+,: *-.,= O )
O O O *-+,: *-.,6 O )
O O O O ) )
O O O */+,: O ) )
O O O */+,: O )
O O O */+,: */.,: O )
O O O */+,: */.,4 O )
O O O */+,: */.,< O )
O O O */+,: */.,= O )
O O O */+,: */.,6 O )
List of topics
23
!"##$%
&'(#)%
!" #$ #$%&' !()*+ ,-./
!0 1 2&3455678 9):9 ,#$/ ;95<
=> ?4 ?4@A678 +)!! ,#$/
=! BA BA4'C678 +)(9 ,#$/ ;:C<
== D6 D634'678 +)"= ,#$/
=: E E4'4C678 +)*9 ,#$/
=9 ?$ ?F$&8678 +)** ,#$/ ;95<
=( G' G4'%4'.5. *)99 ,#$/
=+ H. I$&' *)"* ,#$/
=* ?& ?&J4@3 *)"+ ,#$/
=" -6 -6AK.@ *)+9 ,#$/
=0 ?7 ?&LL.$ *)*: ,#$/ ;95<
:> M' M6'A 0):0 ,#$/
:! N4 N4@@678 + ,#$/ ;9L<
:= N. N.$84'678 *)0 ,#$/
:: #5 #$5.'6A 0)"! ,#$/
:9 B. B.@.'678 0)*( ,#$/
:( O$ O$&86'. !!)"! ,#$/
:+ 1$ 1$PL3&' !9 ,#$/
Q Q Q *+,-= *+.-+ Q )
Q Q Q Q ) )
Q Q Q */,-= Q ) )
Q Q Q */,-= Q )
Q Q Q *+0-= */,-= Q )
Q Q Q *+0-: */,-= Q )
Q Q Q *+0-( Q )
Q Q Q *+0-( */,-= Q )
Q Q Q *+0-+ */,-= Q )
Q Q Q *+0-* */,-= Q )
Q Q Q *+0-" */,-= Q )
Q Q Q *+0-!> Q )
Q Q Q *+0-!> */,-= Q )
Q Q Q *+0-!> */,-= Q
Q Q Q *+0-!> */,-= */.-= Q
Q Q Q *+0-!> */,-= */.-: Q
Q Q Q *+0-!> */,-= */.-9 Q
Q Q Q *+0-!> */,-= */.-( Q
Q Q Q *+0-!> */,-= */.-+ Q
List of topics
24
!"##$%
&'(#)%
!" #$ #$%&'() *+ ,-$.
!/ 01 02134325 +6*7 ,#$. 89:;
!7 <$ <'$()'325 96/ ,#$.
!= > >''$325 "6!7 ,#$. 8+4;
+? @$ @3$A()325 "67+ ,#$.
+* B1 B3(1325 "677 ,#$. 89:;
+C D( D(E%14F)25 /6* ,#$. 89:;
+! GA GFAH)F'325 /6C7 ,#$.
++ 02 02'HF)325 /6!/ ,#$. 89:;
+9 0H 0H(4325 /6+" ,#$. 89:;
+" I4 IJEEJ4325 76!+ ,#$.
+/ -K <3ELF$ /697 ,#$. 89:;
+7 M4 MJ45325 76== ,#$.
+= N) N)4325 96/= ,#$. 89&;
9? <) G3) /6!+ ,#$.
9* <1 -)'35()% 76"+ ,#$.
9C GF GFEE2$325 =6?* ,#$.
9! N N(43)F *?6+9 ,#$.
9+ OF OF)() *C6*! ,#$.
99 M: MF:325 !67= ,OF. 8":;
9" PJ PJ$325 96C* ,OF.
9/ QJ QJ)'HJ)25 9697 ,OF. 894;
97 MF MF$325 969+ ,OF. 8+R; 894;
9= I$ I$J:F(4%5325 96+" ,OF.
"? B4 BF(4%5325 969! ,OF.
"* I5 I$(5F'H325 9699 ,OF.
"C <5 <J5J$325 96"+ ,OF.
"! S2 S2$(&325 96"/ ,OF.
"+ T4 TJ4(E3)325 "6*9 ,OF. 894;
"9 G1 GF$1325 967" ,OF.
"" U% U%:&$(:325 96=+ ,OF.
"/ V( V(E5325 "6?C ,OF.
"7 S$ S$1325 "6* ,OF.
"= G5 GH2E325 "6*7 ,OF.
/? >1 >''F$1325 "6C9 ,OF.
/* Q2 Q2'F'325 96+! ,OF. 894;
/C VR VJR)325 "6"9 ,OF.
/! GJ GJ)'JE25 /67= ,OF.
W W W *+,-*? *./-C *.0-" W
W W W W ) )
W W W *1/-C W ) )
W W W *1/-C W )
W W W *.,-C *1/-C W )
W W W *.,-+ W )
W W W *.,-9 W )
W W W *.,-9 *1/-C W )
W W W *.,-/ W )
W W W *.,-7 W )
W W W *.,-*? W ) )
W W W *.,-*? W )
W W W *.,-*? *1/-C W )
W W W *.,-*? *1/-C W
W W W *.,-*? *1/-C *10-C W
W W W *.,-*? *1/-C *10-! W
W W W *.,-*? *1/-C *10-+ W
W W W *.,-*? *1/-C *10-9 W
W W W *.,-*? *1/-C *10-" W
W W W W ) )
W W W *2/-C W ) )
W W W *2/-C W )
W W W *2/-C W
W W W *.3-! *2/-C W )
W W W *.3-+ *2/-C W )
W W W *.3-9 *2/-C W )
W W W *.3-" *2/-C W )
W W W *.3-/ *2/-C W )
W W W *.3-/ *2/-C W
W W W *.3-= *2/-C W )
W W W *.3-*? *2/-C W )
W W W *.3-** *2/-C W )
W W W *.3-*C *2/-C W )
W W W *.3-*! *2/-C W )
W W W *.3-*+ *2/-C W )
W W W *.3-*+ *2/-C W
W W W *.3-*+ *1,-C *2/-C W
W W W *.3-*+ *1,-! *2/-C W
List of topics
25
Screened potential and Centrifugal barrier
0 1 2 3 4 5 6 7 8−0.5
0
0.5
L=0
L=1
L=2Coulomb potential
0 1 2 3 4 5 6 7 8−0.5
0
0.5
L=0
L=1L=2
Hartree potential
List of topics
(−Zr
)+L(L+ 1)
r2vs.
(−Z − 1
re−αr − 1
r
)+L(L+ 1)
r2(5)
26
0 1 2 3 4 5 6 7 8−0.5
0
0.5
L=0
L=1
L=2Coulomb potential
0 1 2 3 4 5 6 7 8−0.5
0
0.5
L=0
L=1L=2
Hartree potential
27
Figure 5: Selfconsistent field calculations compared to coulomb - real results
Here we note the position of the turning point - same for both Coulomb andSCF case. This is true only for the 1s states. List of topics
28
Figure 6: Hartree Atomic potential
This figure shows results for oxygen with the potential and energies obtainedfrom a calculation with SCF program
List of topics
29
results for oxygen with the potential and energies obtained from a calculationwith SCF program compared with hydrogen-like levels List of topics
30
10 Expectation value of energy for N-particles
Hamiltonian for N-particles - the repulsion - we sum over pairs of coordinates
H (r1, r2, ...rN) =
N∑
i=1
(−1
2∇2ri− Ze2
ri
)+
∑
(i,j)pairs
e2
|~ri − ~rj|
Without antisymmetrization - independent particles - product function -only 1 term -each particle has its orbital ψa,ψb,ψc.....ψn
Φ (r1, r2, ...rN)→ ψa(r1)ψb(r2)ψc(r3)......ψn(rN)
With antisymmetrization - the Slater determinant
ΦHFa,b,...n (r1, r2, ...rN)→ 1√
N !
∣∣∣∣∣∣∣∣
ψa(r1)ψb(r1)......ψn(r1)ψa(r2)ψb(r2)......ψn(r2).......ψa(rN)ψb(rN)......ψn(rN)
∣∣∣∣∣∣∣∣
List of topics
31
For N-particles - the repulsion - we sum over pairs of coordinates
N∑
i=1
(−1
2∇2ri− Ze2
ri
)+
∑
(i,j)pairs
e2
|~ri − ~rj|
Φ (r1, r2, ...rN) = EΦ (r1, r2, ...rN)
Expectation value with the independent particles - product function - only1 term
Φ (r1, r2, ...rN)→ ψa(r1)ψb(r2)ψc(r3)......ψn(rN)
becomes
E = 〈Φ|H |Φ〉 →n∑
α=a
〈ψα|T −Ze2
r|ψα〉+
∑
(α,β)pairs
〈ψβψα|e2
|~r − ~r ′ ||ψβ ψα 〉
The sum over coordinate pairs becomes sum over pairs of orbitalsThis remains true for Slater determinants - but it must be shown.There are N ! terms in the Slater determinant, the left and right give(N !)2 terms, and there are N(N − 1)/2 coordinate pairs.Thus (N !)2N(N − 1)/2 terms.Due to the normalization, this becomes only N(N − 1) terms,N(N − 1)/2 direct terms and N(N − 1)/2 exchange terms List of topics
32
It should be shown that when the Slater determinant is used
ΦHFa,b,...n (r1, r2, ...rN)→ 1√
N !
∣∣∣∣∣∣∣∣
ψa(r1)ψb(r1)......ψn(r1)ψa(r2)ψb(r2)......ψn(r2).......ψa(rN)ψb(rN)......ψn(rN)
∣∣∣∣∣∣∣∣the expectation value will give (final result shown here first)
⟨ΦHF
∣∣H∣∣ΦHF
⟩=
n∑
j=α
〈ψα|T −Ze2
r|ψα〉
+∑
(α,β)pairs
〈ψβψα|e2
|~r − ~r ′||ψβ ψα 〉
−∑
(α,β)pairs
〈ψβψα|e2
|~r − ~r ′||ψα ψβ 〉 (6)
The last term - exchange energy. On the previous page product function (1term) is shown - no exchange energy
Next 3 slides: 1. Helium, 2. Lithium examples, 3. Counting nonzero terms
List of topics
33
For Helium
2
r1r
2
r12
+Ze
r
−e−e
1
[− h2
2me
∇ 2r1− Z e2
r1− h2
2me
∇ 2r2− Z e2
r2+
e2
|r1 − r2|
]Ψ (r1, r2) = E Ψ (r1, r2)
ΦHFa,b (r1, r2) =
1√2
∣∣∣∣ψa(r1) ψb(r1)ψa(r2) ψb(r2)
∣∣∣∣ =1√2
[ψa(r1)ψb(r2)− ψb(r1)ψa(r2)]
and the energy becomes
⟨ΦHF
∣∣H∣∣ΦHF
⟩= 〈ψa|T −
Ze2
r|ψa〉+ 〈ψb|T −
Ze2
r|ψb〉
+ 〈ψaψb|e2
|~r − ~r′ ||ψa ψb 〉 − 〈ψaψb|
e2
|~r − ~r′ ||ψb ψa 〉 (7)
The last term - exchange energy - and there is only one pair.List of topics Lithium example
34
r1r
2
r12
rr3
23r
r13
−e3
−e1
−e2
+Ze
r
Lithium 6× 6× 3-pair terms
[− h2
2me∇ 2r1 −
Z e2
r1− h2
2me∇ 2r2 −
Z e2
r2− h2
2me∇ 2r3 −
Z e2
r3
+e2
|r1 − r2|+
e2
|r1 − r3|+
e2
|r2 − r3|
]Ψ (r1, r2, r3) = E Ψ (r1, r2, r3)
(8)
1√3! ψα(1) ψβ(2) ψγ(3) + ψβ(1) ψγ(2) ψα(3) + ψγ(1) ψα(2) ψβ(3)
− ψγ(1) ψβ(2) ψα(3) − ψα(1) ψγ(2) ψβ(3) − ψβ(1) ψα(2) ψγ(3)
List of topics Helium example Next slide: Counting nonzero terms
35
+c a b
+b c a
−−b a c
−−c b a
−a c b
+a b c
+a b c
+c a b
+a b c
r2 3
e2
r2 3
e2
r2 3
e2
−b a c
+a b c r2 3
e2−
+a b c r2 3
e2−
+a b c r2 3
e2−
−−c b a
−a c b
+a b c
r2 3
e2
r2 3
e2
r2 3
e2
r2 3
e2
r2 3
e2
r2 3
e2
b c
b c
b c
b c
b c
b c
+b c a
+a b c
+ + −
+
+
1
3!
b c
a b
c a
a c
b a
c b
a
a
a
a
a
c
b
b
c
a
a a
Evaluation of the first six terms of the total 36 terms. For each further groupof 6 a very similar procedure would follow. From each six, two remain, totaling12 nonzero terms List of topics
36
11 N particles in Slater determinant - Total
Energy Summary
〈Φ| H |Φ〉 =N∑
j=1
〈ψj|T −Ze2
r|ψj〉+
∑
(i,j)pairs
〈ψjψi|e2
|~r − ~r′||ψj ψi 〉
−∑
(i,j)pairs
〈ψjψi|e2
|~r − ~r′ ||ψi ψj 〉 (9)
We will use also a more compact notation, instead of orbitals ψi we use orbitals|α〉, i.e. the index becomes the name of the function
〈Φ| H |Φ〉 =∑
α
〈α|(T + V )|α〉+∑
pairs αβ
[〈αβ|Vee|αβ〉 − 〈αβ|Vee|βα〉] (10)
List of topics
37
Schrodinger equation from variational principle: minimize
〈Φ|H |Φ〉〈Φ | Φ〉
(≥ Eg.s.)
which can be written as minimalization with constraint 〈Φ | Φ〉 = 1 ; minimize
〈Φ|H |Φ〉 − λ 〈Φ | Φ〉 〈Φ | Φ〉 − 1 = 0 λ Lagrange multiplier
Derivative - differential
f(x)→ df(x)
dx→ df =
df(x)
dxdx
Functional derivative - functional F(f); variation of the functions f(x), δf
F(f) → δF(f)
δf→ δF =
δF(f)
δfδf (11)
Variation of 〈Φ|H |Φ〉 is taken (considering the complex nature of |Φ〉 ) as
δ (〈Φ|H |Φ〉) = 〈δΦ|H |Φ〉
With arbitrary variations |δΦ〉
〈δΦ|H |Φ〉 − λ 〈δΦ | Φ〉 −→ H |Φ〉 − λ |Φ〉 = 0 −→ H |Φ〉 = λ |Φ〉
Thus the Lagrange multiplier plays the role of energy eigenvalue List of topics
38
Deriving Hartree-Fock: For N electrons this energy must be minimized
〈Φ| H |Φ〉 =∑
α
〈α|(T + V )|α〉+∑
pairs αβ
[〈αβ|Vee|αβ〉 − 〈αβ|Vee|βα〉] (12)
with respect to the N orbitals |α〉, |β〉, etc. ,
with N conditions 〈α|α〉 = 1, 〈β|β〉 = 1, i.e.
with N Lagrange multipliers εα, εβ .......
〈δΦ(α→ δα)| H |Φ〉 − εα 〈δα | α〉 = 0
Which results in N equations (one for each of the N orbitals)
〈δα|(T + V )|α〉+∑
β
[〈(δα)β|Vee|αβ〉 − 〈(δα)β|Vee|βα〉]− εα 〈δα | α〉 = 0 (13)
Leaving for a while out the exchange-related term, we can as in prev. slide reducethis by removing 〈δα|
(T + V )|α〉+∑
β
[〈β|Vee|β〉] |α〉 − εα | α〉 = 0
which is exactly the Hartree method. The exchange term leads to complications →List of topics
39
Reducing the full equation
〈δα|(T + V )|α〉+∑
β
[〈(δα)β|Vee|αβ〉 − 〈(δα)β|Vee|βα〉]− εα 〈δα | α〉 = 0 (14)
we get the Hartree Fock Equations (N orbitals)
(T + V )|α〉+∑
β
[〈β|Vee|β〉] |α〉 −∑
β
[〈β|Vee|α〉] |β〉 − εα | α〉 = 0
The complexity of the exchange term becomes clear when we write it explicitelywith coordinates and integrations.Hartree Fock Equations: The direct term and the exchange term:
WHF = Wd − Wex
WHFψa(r) =
[occ∑
b
∫ψ∗b (x)
e2
|r− x|ψb(x)d3x
]ψa(r)−
[occ∑
b
ψ∗b (r)
∫ψb(x)
e2
|r− x|ψa(x)d3x
]
where the exchange potential is nonlocal
Wdψa →Wd(r)ψa(r) Wexψa(r)→∫Wex(r,x)ψa(x)d3x
List of topics
40
Hartree-Fock total energy and sum of orbital energies
with a shorthand notation
ϕα(r) → |ϕα〉 → |α〉
Slater determinantΦα,β,....,ν → |α, β, ...., ν〉
We have started our work with the Hartree-Fock equations by evaluating
〈Φα,β,....,ν |
[T + V +
∑
pairs ij
e2
|ri − rj|
]|Φα,β,....,ν〉 (15)
This we have evaluated as
〈Φ| H |Φ〉 =∑
α
〈α|(T + V )|α〉+∑
pairs αβ
[〈αβ|Vee|αβ〉 − 〈αβ|Vee|βα〉] (16)
From this expression we have obtained Hartree-Fock equations by the varia-tional procedure.
List of topics
41
The Hartree-Fock equation can be then written as[T + V +
∑
β
〈β| e2
|r− r′||β〉
]|α〉−
∑
β
[〈β| e2
|r− r′||α〉]|β〉 = εα |α〉 (17)
We form the matrix element with the given 〈α|
〈α|
[T + V +
∑
β
〈β| e2
|r− r′||β〉
]|α〉 − 〈α|
[∑
β
〈β| e2
|r− r′||α〉
]|β〉 = εα
(18)since 〈α|α〉 = 1. This can be rewritten as
εα = 〈α|(T + V )|α〉+∑
β 6=α
[〈βα|Vee|βα〉 − 〈βα|Vee|αβ〉] (19)
And now we can explore what is the sum of all εα∑
α
εα =∑
α
〈α|(T + V )|α〉+∑
α
∑
β 6=α
[〈βα|Vee|βα〉 − 〈βα|Vee|αβ〉] (20)
which we can compare with the above 〈Φ| H |Φ〉
〈Φ| H |Φ〉 =∑
α
〈α|(T + V )|α〉+∑
pairs αβ
[〈αβ|Vee|αβ〉 − 〈αβ|Vee|βα〉] (21)
42
The two expressions are very similar, but they differ in fact by all the interac-tion term, since it is counted twice in the sum:
∑
α
∑
β 6=α
Fαβ = 2∑
pairs αβ
Fαβ
for any set of objects that are symmetric Fαβ = Fβα
Our objects are symmetric, because they are in fact of the type Hαβ,αβ. Thewavefunctions are antisymmetric, |αβ〉 = −|βα〉.Thus, surprisingly perhaps
〈Φ| H |Φ〉 6=∑
α
εα
but as we derived here List of topics
〈Φ| H |Φ〉 =∑
α
εα −∑
pairs αβ
[〈αβ|Vee|αβ〉 − 〈αβ|Vee|βα〉] (22)
43
Evaluation of the repulsion term using the multipole expansion
1
|r1 − r2|=∑
LM
4π
2L+ 1
rL<rL+1>
Y ?LM (r1)YLM (r2) (23)
wherer< = r1, r> = r2 for |r1| < |r2|
r< = r2, r> = r1 for |r1| > |r2|
Evaluation of the matrix element in general case∫d3r1
∫d3r2ψ
?n1l1m1
(r1)ψ?n2l2m2
(r2)1
|r1 − r2|ψn1l1m1 (r1)ψn2l2m2 (r2) (24)
is performed separately over the radial and angular parts∫r21dr1
∫dr1
∫r22dr2
∫dr2 R?
n1l1(r1)Y
?l1m1
(r1)R?n2l2
(r2)Y?l2m2
(r2)
1
|r1 − r2|Rn1l1(r1)Yl1m1 (r1)Rn2l2(r2)Yl2m2 (r2)(25)
where dri means the integration over dΩi = sin θidθidϕi. List of topics
44
The evaluation of general case - angular integrals of three Ylm’s
CL =
∫Y ?limi
(θ, ϕ)YLM(θ, ϕ)Ylimi(θ, ϕ)dΩ (26)
For the case of both s-states, li = 0 mi = 0 only L = 0 M = 0 are non-zero; The sum reduces to one term. The angular factors give value one, sincethe (YL=0M=0)
2 = (4π)−1 cancels the corresponding factor in the multipoleexpansion and due to the normalization.
List of topics
45
12 Configuration mixing
Consider the usual:Hx(x)ϕα(x) = Eαϕα(x)
Hy(y)χβ(y) = Eβχβ(y)
For any Φ(x)
Φ(x) =∑
cαϕα(x)
For any Ξ(x)
Ξ(y) =∑
dβχβ(y)
Take now a general Ψ(x, y). First look at y as a parameter, Ψ(x, y0)
Ψ(x, y0)→ Φ(x) =∑
cα(y0)ϕα(x)
for every y0 ; Thus we get a new function of y;
cα(y) =∑
dβ(α)χβ(y)
Inserting back:
Ψ(x, y) =∑
dβ(α)χβ(y)ϕα(x)
46
Or, with a simpler notation
Ψ(x, y) =∑
dβαχβ(y)ϕα(x)
In the case of Helium, for example, the H(x) and H(y) are identical
and so are the χβ(y) and ϕα(x). This becomes configuration mixing.
Ψ(x, y) =∑
dβαϕβ(y)ϕα(x)
The coefficients are found by diagonalization.
For three coordinate sets - e.g. for Lithium :
Ψ(x, y, z) =∑
Dγβαϕ(z)ϕβ(y)ϕα(x)
List of topics
47
END OF THE PRESENTATION
List of topics
48