FIRST ORDER NON-DEGENERATE PERTURBATION THEORY

Link to: physicspages home page.

To leave a comment or report an error, please use the auxiliary blog.
References: Griffiths, David J. (2005), Introduction to Quantum Mechan-
ics, 2nd Edition; Pearson Education - Problem 6.1.
All the quantum systems we’ve seen so far are special in the sense that
we could obtain exact solutions for the wave function and energy levels.
Needless to say, such systems are the exception rather than the rule, so we
need methods for handling more realistic cases.
Of course, we could jump straight to numerical solutions of the Schrödinger
equation, and this is, in fact, done in many cases. However, there are several
techniques by which we can find approximate solutions to systems where
the hamiltonian is a relatively minor modification of an exactly-solved sys-
tem. We’ll have a look at first order perturbation theory for non-degenerate
systems here. (That’s the simplest case, if you’re wondering.)
The idea is that we begin by assuming that we have a complete set of
solutions to the equation

H0 ψn0 = En0 ψn0 (1)

The notation gets a bit clumsy here, but the subscript 0 indicates that this
is exactly-known system, and the subscript n refers to the energy level, as
usual. We’re assuming that the system is non-degenerate, that is, that there
is exactly one state for each energy.
Now suppose we introduce a perturbation into the hamiltonian, so we
now have

H = H0 + λV (2)

where λ is a real value, assumed small, and V is a time-independent func-

tion. We would like to solve

Hψn = En ψn (3)

but in most cases, this won’t be possible, at least exactly.

We make the assumption at this point that the energies and wave func-
tions can be expanded in powers of λ :
1
 FIRST ORDER NON-DEGENERATE PERTURBATION THEORY 2

En = En0 + λ En1 + λ 2 En2 + . . . (4)

ψn = ψn0 + λ ψn1 + λ 2 ψn2 + . . . (5)
The subscripts n0, n1 and so on indicate the nth energy level and the 0th,
1st, etc order term in the expansion. We can plug these expansions back
into the original equation to get

(H0 + λV ) ψn0 + λ ψn1 + λ 2 ψn2 + . . . = En0 + λ En1 + λ 2 En2 + . . . ψn0 + λ ψn1 + λ 2 ψn2 + . . .




(6)
Equating individual powers of λ we get

H0 ψn0 = En0 ψn0 (7)

H0 ψn1 +V ψn0 = En0 ψn1 + En1 ψn0 (8)
H0 ψn2 +V ψn1 = En0 ψn2 + En1 ψn1 + En2 ψn0 (9)
The first of these is just the original unperturbed equation. The second
equation can be fiddled a bit as follows. Using the bra-ket notation ψni =
|nii, we can multiply both sides by hn0| and get

hn0| H0 |n1i + hn0|V |n0i = hn0| En0 |n1i + hn0| En1 |n0i (10)
Since H0 is hermitian, we can swap it to the bra end in the first term and
then use hH0 n0| = En0 , and since the Eni s are just constants, they come out
of the brackets on the right, and since |n0i is normalized, we get

En0 hn0 |n1i + hn0|V |n0i = En0 hn0 |n1i + En1 (11)
En1 = hn0|V |n0i (12)
That is, the first-order correction to the energy is given by the expecta-
tion value of the perturbation to the hamiltonian in the unperturbed wave
function ψn0 .
To get the first order correction to the wave function, we expand ψn1 in
terms of the complete set of eigenfunctions of H0 :

ψn1 = ∑ cn j ψ j0 (13)
j
We can plug this into the first order equation above and get

H0 ∑ cn j ψ j0 +V ψn0 = En0 ∑ cn j ψ j0 + En1 ψn0 (14)

j j
 FIRST ORDER NON-DEGENERATE PERTURBATION THEORY 3

Since the ψ j0 form an orthonormal set, we can use H0 ψ j0 = E j0 ψ j0 and

take the inner product with ψk0 for some specific index k. If we choose
k 6= n, then

cnk Ek0 + hk0|V |n0i = cnk En0 (15)

hk0|V |n0i
cnk = (16)
En0 − Ek0
If k = n, then

cnn En0 + hn0|V |n0i = cnn En0 + En1 (17)

En1 = En1 (18)
Thus the k = n term tells us nothing new. In fact, we can just leave out
this term from the expansion, since the contribution of the k = n term is
already present in the original unperturbed wave function, which is entirely
ψn0 . Thus the first-order perturbation to the wave function is

h j0|V |n0i
ψn1 = ∑ ψ j0 (19)
j6=n En0 − E j0
Note that the assumption of non-degeneracy is crucial here, since if En0 =
E j0 for some j 6= n, we would be dividing by zero.
As an example of these formulas, we can consider the case of a delta-
function perturbation in the centre of the infinite square well. That is, we
introduce
 a
V = αδ x − (20)
2
The exact wave functions in the unperturbed square well are
r
2 nπx
ψn0 = sin (21)
a a
The energy perturbation is

En1 = hn0|V |n0i (22)

2α 2 nπ
= sin (23)
a 2
For odd n this is En1 = 2α/a, while for even n En1 = 0. The latter result
is due to the fact that for even n ψn0 (a/2) = 0 so the delta function isn’t
seen.
 FIRST ORDER NON-DEGENERATE PERTURBATION THEORY 4

We can work out the perturbation in the wave function for the case n = 1.
We have

h j0|V |10i
ψ11 = ∑ ψ j0 (24)
j6=1 E10 − E j0
The unperturbed energy levels are

( jπ h̄)2
E j0 = (25)
2ma2
so

π 2 h̄2
2
E10 − E j0 = 1 − j (26)
2ma2
The matrix elements are
 a
h j0|V |10i = h j0| αδ x − |10i (27)
2
2α π jπ
= sin sin (28)
a 2 2
2α jπ
= sin (29)
a 2
Putting it all together, we get
r  
mα a 3πx 1 5πx 1 7πx
ψ11 = sin − sin + sin +... (30)
π 2 h̄2 2 a 3 a 6 a
So the complete wave function is
r r  
2 πx mα a 3πx 1 5πx 1 7πx
ψ= sin + sin − sin + sin + . . . (31)
a a π 2 h̄2 2 a 3 a 6 a
P INGBACKS
Pingback: Bosons in the infinite square well
Pingback: Second order non-degenerate perturbation theory
Pingback: Harmonic oscillator in an electric field
Pingback: Degenerate perturbation theory: two states
Pingback: Perturbing a particle on a circular wire
Pingback: Perturbation of 3-d square well
Pingback: Degenerate perturbation in 3 state system
Pingback: Perturbation theory for higher-level degenerate systems
Pingback: Fine structure constant
 FIRST ORDER NON-DEGENERATE PERTURBATION THEORY 5

Pingback: Fine structure of hydrogen: relativistic correction

Pingback: Harmonic oscillator: relativistic correction
Pingback: Perturbing the 3-d harmonic oscillator
Pingback: Van der Waals interaction
Pingback: Feynman-Hellmann theorem and the harmonic oscillator
Pingback: Stark effect in hydrogen for n = 1 and n = 2
Pingback: Hydrogen atom in a crystal lattice
Pingback: Perturbing the wave function (Stark effect and proton electric
dipole moment)
Pingback: Perturbation theory and the variational principle
Pingback: Variational principle with a two-state hamiltonian
Pingback: WKB approximation
Pingback: Time-dependent Schrödinger equation: switching a perturba-
tion on and off
Pingback: Creation and annihilation operators in the harmonic oscillator