Perturbation theory Diverting from the normal procedure we normalize

Stationary perturbation methods hn(0) |ni = cn (λ) = 1.

Let us suppose that
We write
• we have solved completely the problem
|ni = |n(0) i + λ|n(1) i + λ2 |n(2) i + · · ·
(0)
H0 |n i= En(0) |n(0) i. ∆n = λ∆(1) 2 (2)
n + λ ∆n + · · · .

The basis {|n(0) i} is now complete. Because

hn(0) |λV − ∆n |ni = 0,
• the states |n(0) i are non degenerate.
we have, on the other hand
• we want to solve the problem
∆n = λhn(0) |V |ni.
(H0 + λV )|niλ = En(λ) |niλ .
Thus we get
Usualy the index λ is dropped off.
λ∆(1) 2 (2)
n + λ ∆n + · · ·
When we denote
= λhn(0) |V |n(0) i + λ2 hn(0) |V |n(1) i + · · · .
∆n ≡ En − En(0) ,
Equalizing the coefficients of the powers of the parameter
the eigenvalue equation to be solved takes the form λ we get

(En(0) − H0 )|ni = (λV − ∆n )|ni. O(λ1 ) : ∆n

(1)
= hn(0) |V |n(0) i
(2)
(0) O(λ2 ) : ∆n = hn(0) |V |n(1) i
Note Because the expression (En − H0 )−1 |n(0) i is .. ..
undefined the operator
(0)
− H0 )−1 is not well defined.
(En . . .
(N )
So, in the equation above we cannot invert the operator O(λN ) : ∆n = hn(0) |V |n(N −1) i
(0)
(En − H0 ). .. ..
. .
Now
hn(0) |λV − ∆n |ni = hn(0) |En(0) − H0 |ni, We substitute into the expression
so in the state (λV − ∆n )|ni there is no component along φn
the state |n(0) i. |ni = |n(0) i + (λV − ∆n )|ni
En(0) − H0
We define a projection operator as
X for the state vector the power series of the state vector
φn = 1 − |n(0) ihn(0) | = |k (0) ihk (0) |. and the energy correction and we get
k6=n
|n(0) i + λ|n(1) i + λ2 |n(2) i + · · ·
Now
φn
1 X 1 = |n(0) i + (0) (λV − λ∆(1) 2 (2)
n − λ ∆n − · · ·)
φn = |k (0) ihk (0) | En − H0
En(0) − H0 (0)
k6=n En − Ek
(0)
×(|n(0) i + λ|n(1) i + · · ·).
and Equalizing the coefficients of the linear λ-terms we get in
(λV − ∆n )|ni = φn (λV − ∆n )|ni. the first order
Since in the limit λ → 0 we must have
O(λ) : |n(1) i
(0)
|ni → |n i, φn ∆(1)
n
= V |n(0) i − φn |n(0) i
the formal solution is of the form En(0) − H0 En(0) − H0
φn
1 = V |n(0) i,
|ni = cn (λ)|n(0) i + φn (λV − ∆n )|ni, En(0) − H0
En(0) − H0
because
where
φn ∆(1)
n |n
(0)
i = 0.
lim cn (λ) = 1
λ→0
We substitute |n(1) i into the expression
and
cn (λ) = hn(0) |ni. ∆(2)
n = hn
(0)
|V |n(1) i,
so of one of these states to the energy corrections of the
φn other are
∆(2)
n = hn (0)
|V V |n (0)
i.
En(0) − H0
(2) |Vij |2
We substitute this further into the power series of the ∆i = (0) (0)
<0
Ei − Ej
state vectors and we get for the coeffients of λ2 the
condition (2) |Vij |2
∆j = (0) (0)
>0
φn φn Ej − Ei
O(λ2 ) : |n(2) i = V V |n(0) i
En(0) − H0 En(0) − H0 and the energy levels move apart from each other.
φn (0) (0) φn (0)
− hn |V |n i V |n i. Perturbation expansions converge if |Vij /(Ei − Ej )| is
(0) (0)
En(0) − H0 En(0) − H0
”small”. In general, no exact convergence criterion is
Likewise we could continue to higher powers of the known.
parameter λ. This method is known as the The state |ni is not normalized. We define the normalized
Rayleigh-Schrödinger perturbation theory. state
The explicit expression for the second order energy |niN = Zn1/2 |ni,
correction will be
so that
φn hn(0) |niN = Zn1/2 hn(0) |ni = Zn1/2 .
∆(2)
n = hn
(0)
|V V |n(0) i
En(0) − H0 Thus the normalization factor Zn is the probablility for
X φn the perturbed state to be in the unperturbed state.
(0) (0) (0) (0) (0) (0)
= hn |V |k ihk | |l ihl |V |n i
k,l
En(0) − H0 The normalization condition
X hk (0) |l(0) i N hn|niN = Zn hn|ni = 1
= Vnk (0)
Vln
k,l6=n En(0) − El
gives us
X |Vnk | 2
= .
En(0) − Ek
(0) Zn−1 = hn|ni = (hn(0) | + λhn(1) | + λ2 hn(2) | + · · ·)
k6=n
×(|n(0) i + λ|n(1) i + λ2 |n(2) i + · · ·)
Thus, up to the second order we have = 1 + λ2 hn(1) |n(1) i + O(λ3 )
X |Vkn |2
∆n ≡ En − En(0) = 1 + λ2 + O(λ3 ).
(0) (0) 2
X |Vnk |2 k6=n (En − Ek )
= λVnn + λ2 (0)
+ ···.
k6=n En(0) − Ek Up to the order λ2 the probability for the perturbed state
to lie in the unperturbed state is thus
Correspondingly, up to the second order the state vector
is X |Vkn |2
Z n ≈ 1 − λ2 (0)
.
X Vkn k6=n (En(0) − Ek )2
(0) (0)
|ni = |n i+λ |k i (0)
k6=n En(0) − Ek
 The latter term can be interpreted as the probability for
X X Vkl Vln the ”leakage” of the system from the state |n(0) i to other
+λ2 |k (0) i  (0) (0) states.
(0)
k6=n l6=n (En − Ek )(En(0) − El ) Example The quadratic Stark effect.
! We consider hydrogen like atoms, i.e. atoms with one
Vnn Vkn
− (0)
electron outside a closed shell, under external uniform
(En(0) − Ek )2 electric field parallel to the positive z-axis. Now
+ ···.
p2
We see that the perturbation mixes in also other states H0 = + V0 (r) and V = −e|E|z.
2m
(than |n(0) i).
We see that We suppose that the eigenstates of H0 are non degenerate
(not valid for hydrogen). The energy shift due to the
• in the 1st order we need only the matrix element Vnn . external field is
X |zkj |2
• in the 2nd order the energy levels i and j repel each ∆k = −e|E|zkk + e2 |E|2 + ···,
(0) (0) (0) (0)
other. Namely, if Ei < Ej , then the contributions j6=k Ek − Ej
where Now
zkj = hk (0) |z|j (0) i. · ¸
(0) (0) (0) (0) e2 1
Since we assumed the states |k (0) i to be non degenerate −E0 + Ek ≥ −E0 + E1 = 1− ,
2a0 4
they are eigenstates of the parity. So, according to the
parity selection rule the matrix element of zkk vanishes. so
Indeces k and j are collective indeces standing for the 8a0 16a30
α < 2e2 a20
= ≈ 5.3a30 .
quantum number triplet (n, l, m). According to the 3e2 3
Wigner-Eckart theorem we have The exact summation gives

zkj = hn0 , l0 m0 |z|n, lmi 9a30

α= = 4.5a30 .
hn0 l0 kT (1) knli 2
= hl1; m0|l1; l0 m0 i √ ,
2l + 1
Degeneracy
where we have written the operator z as the spherical (0)
tensor Let’s suppose that the energy state ED is g-foldly
(1)
T0 = z. degenerated:

In order to satisfy zkj 6= 0 we must have (0)

H0 |m(0) i = ED |m(0) i, ∀|m(0) i ∈ D, dimD = g.
• m0 = m and We want to solve the problem
0 0
• l = l − 1, l, l + 1. From these l = l is not suitable
due to the parity selection rule. (H0 + λV )|li = El |li

So we get with the boundary condition

½ X
l0 = l ± 1 lim |li → hm(0) |l(0) i|m(0) i,
hn0 , l0 m0 |z|n, lmi = 0 unless . λ→0
m0 = m m∈D

We define the polarizability α so that i.e. we are looking for corrections to the degenerated
states. With the help of the energy correction we have to
1 solve
∆ = − α|E|2 . (0)
2 (ED − H0 )|li = (λV − ∆l )|li.
As a special case we consider the ground state We write again
|0(0) i = |1, 00i of hydrogen atom which is non degenerate
when we ignore the spin. The perturbation expansion |li = |l(0) i + λ|l(1) i + λ2 |l(2) i + · · ·
gives (1) (2)
X∞ ∆l = λ∆l + λ2 ∆l + ···,
|hk (0) |z|1, 00i|2
α = −2e2 (0) (0)
,
k6=0 E0 − Ek so we get
where the summing must be extended also over the (0)
(ED − H0 )(|l(0) i + λ|l(1) i + λ2 |l(2) i + · · ·)
continuum states.
(1) (2)
Let us suppose that = (λV − λ∆l − λ2 ∆l − · · ·)
(0) (1) 2 (2)
(0) (0) ×(|l i + λ|l i + λ |l i + · · ·).
E0 − Ek ≈ constant,
Equalizing the coefficients of the powers of λ we get in
so that the first order
X X
|hk (0) |z|1, 00i|2 = |hk (0) |z|1, 00i|2 (0)
(ED − H0 )|l(1) i
k6=0 all k’s
(1)
= (V − ∆l )|l(0) i
= h1, 00|z 2 |1, 00i. " #
(1)
X
(0) (0) (0)
In the spherically symmetric ground state we obviously = (V − ∆l ) |m ihm |l i .
have m∈D
1
hz 2 i = hx2 i = hy 2 i = hr2 i. (0)
3 Taking the scalar product with the vector hm0 | and
Using the explicit wave functions we get recalling that
(0) (0)
hz 2 i = a20 . hm0 |(ED − H0 ) = 0,
we end up with the simultaneous eigenvalue equations 1◦ Identify the degenerated eigenstates. We suppose
X (1) (0)
that their count is g. Construct the
Vm0 m hm(0) |l(0) i = ∆l hm0 |l(0) i. g × g-perturbation matrix V .
m
2◦ Diagonalize the perturbation matrix.
(1)
The energy corrections ∆l are obtained as eigenvalues.
From the equation 3◦ The resulting eigenvalues are first order corrections
for the energy shifts. The corresponding eigenvectors
(0) (1)
(ED − H0 )|l(1) i = (V − ∆l )|l(0) i are those zeroth order eigenvectors to which the
corrected eigenvectors approach when λ → 0.
we also see that
(1)
4◦ Evaluate higher order corrections using non
hm(0) |V − ∆l |l(0) i = 0 ∀|m(0) i ∈ D. degenerate perturbation methods but omit in the
(1) summations all contributions coming from the
Thus the vector (V − ∆l )|l(0) i has no components in the degenerated state vectors of the space D.
subspace D. Defining a projection operator as
Example The Stark efect in the hydrogen atom.
g
X X
(0) (0) (0) (0) The hydrogen 2s (n = 2, l = 0, m = 0) and 2p
φD = 1 − |m ihm |= |k ihk |
(n = 2, l = 1, m = −1, 0, 1) states are degenerate. Their
m∈D k6∈D
energy is
(0)
we can write ED = −e2 /8a0 .
(1) (1)
(V − ∆l )|l(0) i = φD (V − ∆l )|l(0) i = φD V |l(0) i. We put the atom in external electric field parallel to the
z-axis:
We get the equation V = −ez|E|.
(0) (1) Now z is the q = 0 component of a spherical tensor:
(ED − H0 )|l(1) i = φD (V − ∆l )|l(0) i,
(1)
(0)
where now the operator (ED − H0 ) can be inverted: z = T0 .

φD According to the parity selection rule the operator V now

|l(1) i = (0)
V |l(0) i has nonzero matrix elements only between states with
ED − H0
l = 0 and l = 1, and due to the m-selection rule all states
X |k (0) iVkl
= . must have the same m:
(0) (0)
k6∈D ED − Ek 2s 2p, 0 2p, 1 2p, −1
 
When we again normalize 2s 0 × 0 0
2p, 0 
 × 0 0 0  .
hl(0) |li = 1, V =
2p, 1  0 0 0 0 
2p, −1 0 0 0 0
we get from the equation
(0)
The nonzero matrix elements are
(ED − H0 )|li = (λV − ∆l )|li
h2s|V |2p, m = 0i = h2p, m = 0|V |2si = 3ea0 |E|.
for the energy shift
The eigenvalues of the perturbation matrix are
∆l = λhl(0) |V |li.
(1)
∆± = ±3ea0 |E|
We substitute the power series and get
and the eigenvectors
λhl(0) |V (|l(0) i + λ|l(1) i + λ2 |l(2) i + · · ·)
= λ∆l
(1)
+ λ2 ∆l
(2)
+ ···. 1
|±i = √ (|2s, m = 0i ± |2p, m = 0i).
2
The second order energy correction is now
Note The energy shift is a linear function of the electric
(2) (0) (1) (0) φD (0) field. The states |±i are not parity eigenstates so it is
∆l = hl |V |l i = hl |V (0)
V |l i
E D − H0 perfectly possible that they have permanent electric
X dipole moment hzi 6= 0.
|Vkl |2
= (0) (0)
.
k6∈D ED − Ek Nearly degenerated states
Let the states m ∈ D to be almost degenerate. We write
Thus the perturbation calculation in a degenerate system
proceeds as follows: V = V1 + V2 ,
where So the potential couples states
X X
V1 = |m(0) ihm(0) |V |m0(0) ihm0(0) | |ki, |k + Ki, . . . , |k + nKi, . . . .
m∈D m0 ∈D
V2 = V − V1 . The corresponding energy denominators are
(0) (0)
We proceed so that Ek − Ek+nK .

1. we diagonalize the Hamiltonian H0 + V1 exactly in In general

(0) (0)
the basis {|m(0) i} and Ek 6= Ek+nK
2. handle the term V2 like in an ordinary non except when
degenerate perturbation theory. This is possible since nK
. k≈−
2
hm0(0) |V2 |m(0) i = 0 ∀m, m0 ∈ D. We suppose that the condition
nK
Example Weak periodic potential. k 6= −
Now the dominant term in the hamiltonian is 2
holds safely. The first order state vectors are then
p2
H0 = X
2m Vn
|k (1) i = |ki + |k + nKi (0) (0)
,
and the perturbation satisfies n6=0 Ek − Ek+nK

V (x) = V (x + a). and the wave functions

1 0 X 1 0 V
We denote the unperturbed eigenstates by the wave (1)
ψk (x0 ) = √ eikx + √ ei(k+nK)x (0) n (0) .
vector: L n6=0 L Ek − Ek+nK
h̄2 k 2
H0 |ki = |ki, Correspondingly the energy up to the second order is
2m
so that (2) (0)
X |Vn |2
(0) h̄2 k 2 Ek = Ek + V0 + (0) (0)
.
Ek = . Ek − Ek+nK
2m n6=0

We impose the periodic boundary conditions Let us suppose now that

hx0 |ki = ψk (x0 ) = hx0 + L|ki = ψk (x0 + L), nK
k≈− .
2
for the wave function and get
We diagonalize the Hamiltonian in the basis
1 0 2π
ψk (x0 ) = √ eikx , k = n, n ∈ I.
L L {|ki, |k + nKi},

Because the potential V is periodic it can be represented i.e. we diagonalize the matrix
as the Fourier series
|ki |k + nKi
∞
X Ã !
(0)
V (x) = e inKx
Vn , |ki Ek + V0 Vn∗
(0) .
n=−∞ |k + nKi Vn Ek + V0
where Its eigenvalues are
K = 2π/a
(0) (0)
is the reciprocal lattice vector. The only nonzero matrix E + Ek+nK
Ek± = V0 + k
elements are now v 2
uà (0) !2
u E − E (0)
hk + nK|V |ki = Vn , ± t k k+nK
+ |Vn |2 .
2
because
Z (0) (0)
1X 0 0 0 0 When |Ek − Ek+nK | À |Vn |, it reduces to two solutions
hk 0 |V |ki = Vn dx0 e−ik x einKx eikx
L n (0)
X Ek+ = Ek + V0
= Vn δk+nK,k0 . (0)
n
Ek− = Ek+nK + V0 ,
which are first order corrected energies. In the limiting
case we get
(0)
lim Ek± = EnK/2 + V0 ± |Vn |.
k→−nK/2

Brillouin-Wigner perturbation theory

We start with the Schrödinger equation
(En − H0 )|ni = λV |ni,
and take on both sides the scalar product with the state
|m(0) i, and get
(0)
(En − Em )hm(0) |ni = λhm(0) |V |ni.
We correct the state |n(0) i. We write the corrected state
|ni in the form
X
|ni = |m(0) ihm(0) |ni = |n(0) ihn(0) |ni + φn |ni
m
X
= |n(0) i + |m(0) ihm(0) |ni,
m6=n

which has been normalized, like before,

hn(0) |ni = 1.
We substitute into this the scalar products
λhm(0) |V |ni
hm(0) |ni = (0)
,
En − Em
and end up with the fundamental equation of the
Brillouin-Wigner method
X λ
|ni = |n(0) i + |m(0) i (0)
hm(0) |V |ni.
m6=n
En − Em

Iteration gives us the series

X 1
|ni = |n(0) i + λ |m(0) i (0)
hm(0) |V |n(0) i
m6=n
En − Em
XX 1
+ λ2 |l(0) i (0)
hl(0) |V |m(0) i
m6=n l6=n En − El
1
× (0)
hm(0) |V |n(0) i
En − Em
X XX 1
+ λ3 |k (0) i (0)
hk (0) |V |l(0) i
m6=n l6=n k6=n En − Ek
1 1
× (0)
hl(0) |V |m(0) i (0)
hm(0) |V |n(0) i
En − El En − Em
+ ···.
Note This is not a power series of the parameter λ
because the energy denominators
(0)
En − Em = En(0) − Em
(0)
+ ∆n
depend also on the parameter λ according to the equation
∆n = λ∆(1) 2 (2)
n + λ ∆n + · · ·