Lecture Notes on Quantum Mechanics

A.N. Njah,
Department of Physics,
University of Agriculture,
Abeokuta, NIGERIA

PHS 411: Symmetry in quantum mechanics. Fundamentals of quantum mechanics - okperations in Hilbert space. Matrix formulation of quantum mechanics. Angular momentum in quantum mechanics; approximation methods
in collision theory. Many electron systems. Scattering theory.
Suggested reading.
1. Quantum Mechanics by L. I. Schi
2. Quantum Mechanics by E. Merzbacher
3. Schaums Outline of Theory and Problems of Quantum Mechanics by
Y. Peleg, R. Pnini and E. Zaatur

Contents

Chapter 1
APPROXIMATION METHODS
1.1

VARIATIONAL METHOD

This method is applicable to conservative systems. Consider a physical system with time-independent Hamiltonian H. We assume for simplicity that
the entire spectrum of H is discrete and non degenerate
(H|n
= En |n
),

n = 1, 2, 3...

(1.1)

Let Eo be the smallest eigenvalue of H (i.e, the smallest energy of the system).
An arbitrary state |
can be written in the form
|
=
Then
|H|
=

n

n

cn |n

|cn |2En Eo

On the other hand,

|
=

n

(1.2)

|cn |2

|cn |2

(1.3)
(1.4)

Thus, we can conclude that for every ket,

H
=

|H|

Eo
|

(1.5)

Eq.(6.5) is the basis of the variational method. A family of kets |()
is

chosen, the so called trial kets. The mean value of H in the states |()

is calculated, and the expression H
() is minimized with respect to the
parameter . The minimum value obtained is an approximation of the ground
state energy Eo .
3

Equation (6.5) is actually a part of a more general result called the Ritz
theorem. The mean value of the Hamiltonian H is stationary in the neighborhood of its discrete eigenvalues.
The variational problem can therefore be generalized to provide an estimation for other energy levels based on the ground state. If the function
H
() obtained from the trial kets |()
has several extrema, they give
approximate values of some of its energies En .
Examples 1. Consider a 1D harmonic oscillator
1
h2 d2
mw2 x2
+
H=
2
2m dx
2
2

(a) For the one-parameter family of wave functions (x) = ex ( > 0),
nd a wave function that minimizes H
. What is the value of H
min?
2
(b) For another one parameter family of wave functions (x) = xex ( >
0), nd a wave function that minimizes H
and compute the value of H

min.
(c) Repeat the same procedure for
(x) =

1
x2 +

( > 0)

Solution
(a) We begin by considering H
.
H
=




1
h2 d2
2 2
2m dx2 + 2 mw x (x)dx


(x) (x)dx

h
2
1 mw2
=
+
2m
8
We dierentiate H
with respect to
h
2
1 mw2
dH

=

d
2m 8 2
From the condition

dH

d =o

= 0, we have

h
2
mw2
mw

;
=
0

=
o
2m
8o2
2h
4

thus o gives the minimum value of H
. The wave function that minimizes
mwx2
H
is o (x) = e 2h and
H
min

h
2
mw2
1
o +
w
=
= h
2m
8o
2

Thus H
min coincides with the energy of the n=0 level of a 1D harmonic
oscillator. Note that the family of functions we are studying coincides with
the ground state wave function of the harmonic oscillator.
(b) We proceed with the same method as in part (a).
H
=




1
h2 d2
2 2
2m dx2 + 2 mw x (x)dx


(x) (x)dx

3 mw2
h
2
=3 +
m
8
and

dH
3h2 3m 2 1
=

=0
d
2m
8 2
2

mx /2h
We obtain o = 12 m
; so H
min equals the energy
h and o (x) = xe

of the n = 1 level of the one-dimensional harmonic oscillator.

(c) Applying the procedures of parts (a) and (b), we obtain

H
=




1
h2 d2
2 2
2m dx2 + 2 mw x (x)dx


(x) (x)dx

h
2 1 1
+ mw2
=
2m 2
and
o (x) =

x2 + h
2/ 2m

1 h
o =
2 m

2. Consider a particle in a one-dimensional potential V(x) - x4. Using the

variational method, nd an approximate value for the energy of the ground
h2 13
state. Compare it to the exact value Eo = 1.06 2m
k where k= 2m/h2 .
1
2
Choose as a trial function = (2/) 4 ex .
5

Solution: First note that the trial function = (2/) 4 ex is normal

h2 d2
4
ized to unity; that is
||2 dx=1. The Hamiltonian is H =
2m dx2 + x ;
thus



h2 d2
4

+
x
(x)dx

2m dx2
H
=


(x)(x)dx
The denominator equals one since (x) is normalized; thus


1
  2  1
  2  1
h
2  d2  2  14 x2
4 x2
4 x2
4 2 4 x2
e

e
dx+
e
x
e
dx
H
=


2m dx2




2


2 
h

2
2
2x2
2
e
2[2x 1]dx +

x4e2x dx
=
2m






2
2
2




2 

h
(2)
2
h
(4 )
2
2
2 2x2
2x2
xe
dx+
e
dx+

x4 e2x dx
=
2m

2m

The rst integral is



h
2
h
2 (42)
h
2 (42)
2x2 2

2 1
=
e
x dx =
I1
2m

2m
4 2
2m

The second integral is:



h
2
h
2(2)
h
2 (2)
2x2

2
=
e
dx =
I2
2m

2m
2
m

and the third integral is:

I3



2 
2

x4e2x dx



2

3
3
=
4(2)2 2 162

Substituting these integrals we obtain

2
3
3
h
2
h
2 h
+
+

+
=
H
=
2m
m
162
2m
16 2
h2

3
=
Hence, dH

d
2m 8 3o .

1
3m 3
o = 4h2 . In terms of

Since


dH

d =

= 0, we obtain

2
h
2m
1

3
8 3o

= 0

k =2m/h2, we have o = (3/8) 3 k 3 . Substituting

this value in , we obtain

H
min

3 1 h
h
2 1
2 1
k 3 = 1.082 k 3
= 33
4 2m
2m
6

Comparing the last result to the exact value of Eo we see that we have quite
a good approximation. The error is approximately 2 percent.
3.(a) Taking a trial wave function proportional to exp(r) , where is a
variable parameter, use the variational method to obtain an upper limit for
the energy of the ground state of the hydrogen atom in terms of atomic constants.
(a) The upper limit of the ground state is the minimum value of the expectation value of the energy as varies. Denote the trial wave function
by
= A exp(r)
where A is a constant. Its value is obtained from the normalizing condition.
Since depends only on r, we put d3 r = 4r2 dr. Therefore, the normalization
condition yields,
4A


r2 exp(2r)dr
0

8A2
=1
=
(2)3)

Thus

3
A =
(1)

The Hamiltonian for the hydrogen atom is

2

e2 1
h
2 2

(2)
H=
2m
4 o r
The expectation value of the energy for the function is
h
2 2
e2 1 
E
= 4A

exp(r)
exp(r)r2dr
(3)
0
2m
4 o r
Since exp(r) depends only on r, we need only the r- dependent part of the
operator
 2
2 

2
exp(r) =
+
exp(r
(4)
r2 r r
Inserting (1) and (4) in (3) gives
2

E
=

=
=

h2 2  2

2m 0 r exp(2r)dr
 2




h
e2
+ m 4o 0 r exp(2r)dr

2 2
h
e2

m
4o
2
h
e2
2
2m 4o

h
2m
2 +

Then

2
e2
dE
h
=

d
m
4 o

which is

me2
0, when =
4 o h
2
Inserting this in the expression for E
gives
E
min

me  e2 2
= 2
2h 4 o

e2 1
8 o ao

Where ao is the Bohr radius.

(b) The expression obtained above is the correct value of the ground-state
energy of the hydrogen atom. The expression is correct because the rtrial
function is the correct form of the ground-state wave function.

1.2

THE WKB APPROXIMATION (or Semiclassical

Approximation)

WKB stands for the names of the proponents: Wentzel, Kramer and Brillouin. The method is suitable for obtaining solutions of the one-dimensional
..
Schrodinder equation as well as the 3-dimensional problems, if the potential
is spherically symmetric and a radial deerential equation can be separated.
That is, it is applicable to situations in which the wave can be separated
into one or more total dierential equations each of which involves a single
independent variable.
..
Consider the Schrodinger equation in one-dimention:
d2 2m
+ 2 [E V (x)](x) = 0
dx2
h

Let

k(x) =
(x) =

1
h

(1)

2m[E V (x)] for E
V (x)

i
h

(2)

2m[E V (x)] for EV (x)

If V (x) is constant eq.(1) has solution eikx

However if V (x) varies slowly with x we assume a solution of the form
= Aeiu(x)/h

(3)

Subsituting this solution in eq.(1) we nd that u(x) satises the equation



du 2
d2 u
+ [hk(x)]2 = 0
(4)
ih 2
dx
dx
In the WKB approximation we expand u(x) in power series of h
as follows:
u(x) = u0 + h
u1 + h
2 u2 + . . .

(5)

where un , n = 0, 1, 2, . . . are approximate solutions of eq.(4). Neglecting

expansion terms of order h
2 and putting eq.(5) in eq.(4) we have

d2 u 0
du0
h
i 2 + 2
dx
dx



du1
dx

du0

dx

2

+ [hk(x)]2 = 0

(6)

where again the expansion terms of order h

2 are neglected and the terms in
eq.(6) are grouped in powers of h
. Now equating equal powers of h
we have
u2
hk(x)]2 = 0
0 + [
9

(7)

iu0 2u0u1 = 0

(8)

Integrating eqs.(7) and (8) we have

u0(x) = h

 x

k(x)dx + c1

(9)

1
u1(x) = i ln k(x) + c2
(10)
2
where c1 and c2 are arbitrary constants of integration. Substituting eqs.(9)
and (10) into eq.(5) and sustituting eq.(5) in eq.(3) one gets
(x) = Ak

21

(x) exp(i

 x

k(x)dx)

for E > V

(11)

where c1 and c2 are absorbed into A. In a similar fashion, the case EV gives
12

(x) = B

(x) exp(

 x

(x )dx)

for EV

(12)

The region in which E > V (x) is called a classically allowed region of motion,
while a region in which EV (x) is called classically inaccessible. The points
in the boundary between these 2 kinds of regions are called turning points
[where E = V (x)].
Applicability condition: The WKB approximation is based on the condition


1  dk 
2

 |k (x)|
(13)


2 dx
(i.e. V(x)varies very slowly with x). k(x) is the wave number so k(x) = 2
,
where is the wavelength. We can then write eq.(13) as


1  dk 
2


k(x)
2  dx 



 dk 

 k(x)
(14)
or
4  dx 
Solution near a turning point: Adjacent to the turning points for which
dk
k(x0), we have k dx
|x0 (x x0). Thus the WKB approximation is applicable
for a distance from the turning point satisfying the condition
|x x0| 

(15)

The connection formulas

The connection formulas at the turning point depends on whether the classical region is to the left (p1) or to the right (p2) of it.
10

In the rst case we have x > a:

 x
A1

1(x) =
cos( k(x)dx B1)
a
k

(16)

while in the second case, for xb

 b
A2

2(x) =
cos( k(x)dx B2)
x
k

(17)

(The results in eqs.(16 and 17) are quoted)

Application to the bound state: The WKB approximation can be applied
to derive an equation for the energies of a bound state. Since the wave function is zero at the boundaries the argument of the cosine in the connection
formulas eqs.(16 and 17) is given by
 b
a

1
k(x)dx B = (n + ) n = 0, 1, 2 . . .
2
 b

1
k(x)dx = (n + ) n = 0, 1, 2 . . .
a
2
if the initial phase is chosen to be zero, i.e. B = 0.
or

1
p(x)dx = 2h(n + ) n = 0, 1, 2 . . .
2

(18)
(19)

(20)

where p(x) = h
k(x) Eq.(20) is called the Bohr-Sommerfeld quantization rule.
Barrier potential: For a potential barrier V(x) between x = a and x =
11

b and a particle of energy E, the transmission coecient T in the WKB

approximation is (quoted)


2  b 
T = exp
2m[V (x) E]dx
(21)
h
a
Solved examples:
1. Using the WKB approximation nd the bounded states for a one-dimensional
innite potential well. Compare your result with the exact solution.
Suppose that the boundaries of the potential well are at x = a. At the
boundaries the wave function has value zero and k(a) = 0. From eqs.(16)
and (17) we have

0 = cos(B1)

0 = cos(B2)

and therefore B1 = B2 = 12 . Thus we get, according to eq.(18)

 a

1
1
Kn dx = (n + )
a
2
2

i.e.

(n + 1)
2a
2 2
2 2
(n + 1)2
h
kn
1h
En =
=
2 m
8ma2

2akn = (n + 1)

From eq.(2) we get

kn =

2 2 2

h n
Recall that the exact solution is 8ma
2
2.Use the WKB approximation to obtain the energy levels of a harmonic oscillator.
Consider the Bohr-Sommerfeld quantization rule:
 b

1
(i)
p(x)dx = h
(n + ) (n = 0, 1, 2, . . .)
a
2

where p(x) = 2m[E V (x)] is the momentum of the oscillator, V (x) =

b
1
2 2
m
x
.
Since
pdx
=
2
a pdx holds for a linear harmonic oscillator we can
2
use eq.(i). The points a and b are the turning points that are determined by
the condition p(a) = p(b) = 0 or E V = 0; thus E 12 m 2 x2 = 0 so we
have 


2
2E
2E
a = m2 , b = m2 . We introduce the new variable z = x m
2E , and
obtain
 b
2E  1
E
(ii)
p(x)dx =
1 z 2 dz =
a
1

12

Comparing eqs.(ii) and (i) we obtain En = h

(n + 12 ) which is the exact
solution.

13

1.3

TIME-INDEPENDENT PERTURBATION THEORY

(FOR BOUND STATES) OR STATIONARY PERTURBATION THEORY

(SPT)
Time-independent perturbation theory is widely used in Physics. Perturbation theory is appropriate when the Hamiltonian H of the system can be put
in the form
H = H0 + h
(1)
where is a small parameter such that the perturbation (or disturbance)
h H0 , i.e., 1; and the matrix elements of the operator h are comparable in magnitude to those of H0. The eigenstates un and eigenvalues En
of the unperturbed Hamiltonian, H0 must be known. The SPT is concerned
with nding the changes in discrete energy levels and eigenfunctions of a system when a small disturbance (perturbation) is applied.
In the unperturbed state
H0 u n = E n u n

(2)

or H0 |n
= En |n

(2)

where un or |n
form an orthonormal basis of the state space.


(i.e., m|n
= mn and n |n
n| = 1).
In the perturbed state H = H() as in eq.(1) and its nth eigenstate n ()
and the corresponding eigenvalue En () are also functions of . Therefore
H()n () = En()n ()

(3)

We assume that En () and n () can be expanded as a power series of in

the form
En() = 0 + 1 + 2 2 + . . .
(4)
n () = 0 + 1 + 2 2 + . . .

(5)

Substituting eqs.(4) and (5) into eq.(3) we obtain

(H0 + h)(0 + 1 + 2 2 + . . .) = ( 0 + 1 + 2 2 + . . .)(0 + 1 + 2 2 + . . .)
(6)
Expanding eq.(6) and equating equal powers of we have for
0 :

(H0 0 )0 = 0
14

(7)

1 :

(H0 0 )1 = ( 1 h)0

(8)

2 :

(H0 0 )2 = ( 1 h)1 + 2 0

(9)

.................................
Note the following:
1. As 0, we have 0 = En (the unperturbed eigenvalue of the nth level)
and eq.(7) means that 0 is the corresponding unperturbed eigenfunction.
We, therefore put
0 = un, 0 = En
(10)
2. any arbitrary multiple of 0 can be added to any of the functions s
without aecting the RHS of eqs.(7) to (9) and the determination of s in
terms of lower order functions. Hence we choose the arbitrary multiple such
that
(0, s ) = 0 s > 0
(11)
3. Using the result


(r) (r)d3r

= [ (r)] (r)d3r = ||

(12)

(where is the matrix element of in the states , ) the inner product

of 0 and the LHS of eqs.(7) to (9) is zero in each case, and it follows, with
the help of eqs.(10) and (11) that
s =

(0, hs1 )
= (un , s1)
(0 , 0)

(13)

First-order perturbation theory: Eq.(13) with s = 1 shows that

1 = (un , hun) or n|h|n

(14)

which is the expectation value of h for the unperturbed state un or |n
. Any
function can be written as a linear combination of the orthonormal basis
functions um so


(15)
1 = a(1)
m um
m

Substituting eq.(15) into eq.(8) we have

(1)
am (H0
m

En )um = ( 1 h)un

15

where a(1)
n = 0 because of eq.(11) (free choice). Now replace H0 um by Em um ,
multiply by uk and sum over all m making use of the orthonormality of the
u s (only terms k = m survive) to have
(1)

ak =

(uk , hun )
k|h|n

or
En Ek
En Ek

k = n

(16)

m|h|n

m = n
En Em

m|h|n

|m

m = n
Hence 1 =
m En Em
The rst-order solutions are
i.e. a(1)
m =

(17)
(18)

En() = 0 + 1 = En + n|h|n

n () = 0 + 1 = un +

(19)

m|h|n

|m

En Em

m

m = n

(20)

Usually we set = 1
Second-order perturbation
When s = 2 eq.(13) gives
2 = (un, h1 ) = (un, h

(1)
am um )
m

(1)
am (un, hum)
m

(1)
am n|h|n

m

|n|h|m
|2
m|h|n

n|h|m
=
m = n
(21)
m En Em
m En Em
since h is assumed to be hermitian. Eq.(21) is the second-order approximation of the eigenvalue.
The second-order approximation of the eigenfunction 2 is calculated by expanding it in terms of um :

2 =

(2)
am um,
m

a(2)
n = 0

(22)

Substituting eq.(22) into eq.(9) we obtain

(2)
am (H0
m

En )um =

(1)
am ( 1
m

h)um + 2 un

Again, replace H0um by Em um , multiply by uk and sum over all m making

use of the orthonormality of the us (only terms k = m survive) to have
(2)

(1)

ak (Ek En ) = ak 1

(1)
am (uk , hum )
m

16

(1)

or ak 1

(1)
am k|h|m

m

k = n

This gives, with the help of eqs.(14) and (17),

(2)

ak =

k|h|n
n|h|n

k|h|m
m|h|n

+
(En Ek )2
m (En Ek )(En Em )

m=n
n=k

(23)

Inserting eq.(23) into eq.(22) we have

2 =

k

k|h|n
n|h|n

k|h|m
m|h|n

+
|m

(En Ek )2
m (En Ek )(En Em )

m=n
n=k

(24)

The second-order solutions are:

2

En() = 0 + 1 + 2 = En + n|h|n
+

|n|h|m
|2
En Em

m|h|n

|m

m En Em

k|h|m
m|h|n

k|h|n
n|h|n

|m

+
+2
2
(E

E
)
(E

E
)(E

E
)
m
n
k
n
k
n
m
k

m=n

(25)

n () = 0 + 1 + 22 = un +

m=n
n=k

(26)

Example 1: The eect of the nite size of the nucleus is to raise the energies
of the electronic states from the theoretical values based on a point nucleus.
Show from rst-order perturbation theory that, if the proton is regarded (for
simplicity) as a thin uniform spherical shell of charge of radius b, the frac4b2
tional change in the energy of the ground state is 3a
2 where the ground state
0

1
( a1 3 ) 2
0

eigenfunction for Hydrogen is us =

exp(r/a0) and b/a0 = 105.
Solution: For a point nucleus the potential is
e2
V0 =
4 0 r

(i)

For this uniform spherical shell the potential is

e2

40 r

V =

r>b

e2
4
0b

(ii)
rb

The perturbation potential is therefore

e2 1

40 r

V1 = V V0 =

17

1b )
r>b

rb
(iii)

The rst order correction to the energies of the ground state is



1 =

u0V1 u0d

(iv)
1

where d = 4r2 dr is the volume element and u0 = ( a1 3 ) 2 exp(r/a0)

0

Thus

4 e2  b 2 1 1
r ( ) exp(2r/a0)dr
a30 4 0 0
r b

(v)

Since b/a0 105, the exponential term may be replaced by unity over the
range of integration, and the integral becomes
1
r2
(r )dr = b2
0
b
6

 b

(vi)

The ground-state energy for a point nucleus is

e2
E0 =
8 0 a0

(vii)

From eqs.(v), (vi) and (vii)

E1
4b2
= 2
E0
3a0
Example 2: Consider a 3 dimensional problem. In a given orthonormal
basis, the Hamiltonian is represented by the matrix

H=

1 0 0
0 3 0
0 0 2

0 c 0
c 0 0
0 0 c

Here, H = H o + H  and c is a constant, c 1.

1. Find the exact eigenvalues of H.
2. Use second order perturbation to determine the eigenvalues.
3. Compare the results of 1 and 2.
Solution 1. The eigenvalue of H are roots of the equation det(H-I) =0.














1
c
0



1
c 

2
2

 = (c2) 
c
3
0
 = (c2)[ 4+3c ]


c
3 
0
0
c 2 

18


Therefore, = c 2, 2 1 + c2 .
2. The second order correction to the energy may be written as: En =
En(0) + En(1) + En(2) or
(E)i = (H )ii + (H )ii +



Hik
Hki

(0)

(0)

Ei Ek

k=i

It can be seen that (H o )ii = 1, 3 and -2. The rst order energy correction is



given by H11
= 0, H22
= 0 and H33
= c. For the second correction, we have
(2)
E1

(2)
E2



H12
H21
(0)

(0)

E1 E2


H12
H21
(0)

(0)

E2 E1

and
(2)
E3



H13
H31
(0)

(0)

(0)

E1 E3


H32
H23
(0)

(0)

E2 E3

(0)

E3 E1



H13
H31

0
c2
c2
+ =
=
2 3
2

0.0 c2
c2
+
=
=
31
3
2



H23
H32
(0)

(0)

E3 E2

= 0+0=0

c
c
Thus E1 = 1
2 , E2 = 3 + 2 ; E3 = 2 + c.
3. Expand 2 1 + c2 in a binomial series

c2
1
c2
1 + c2 = 2 (1 + c2 + ...) = 3 + , 1 (c2 1).
2
2
2

This gives the same result as the second order corrections.

Example 3: Consider a perturbation of the form 12 bx2 to the linear harmonic oscillator problem. Thus,
h2 d2
1 2
kx
+
Ho =
2m dx2 2
and H  = 12 bx2. The eigenvalues and eigenfunctions of Ho are well known
Ho u n = E n u n
where

1
En = n + )h; n = 0, 1, 2, 3...
2
1
un = Nn Hn() exp( 2 )
2
19


k
 12
=
, Nn = n
m
2 n!

m  mk 1/4
= x, =
=
h

h
2
Notice that the exact solution can easily be obtained; all that one has to do
is replace k by k+b. Thus if we write

Hn = Enn
H = Ho + H 

where


1
b 1/2

k
En = n + )h
1+
2
m
k

then

or



1
1 b 1 b2

En = n + )h
1+
+ ... ...
2
m
2 k 8 k2

the binomial expansion been valid when kb 1. If we compare (**) with equation (4), we obtain


o = n +


1 =



1

k
...............()
m



1

k b
1
...............()
n+ h

2
2
mk




1

k b2
1
n+ h

...............( )
8
2
m k2
where o represents the unperturbed eigenvalue, and 1 and 2 represent the
rst and second order perturbations respectively. We can now apply perturbation theory to compute 1 and 2 etc. Now,
2 =

1 
1
1 = b
un(x)x2un (x)dx = bn|x2|n

2
2


1
k b
1
= n+ h
...............(i)

2
2
mk
Equation (i*) agrees with eqn. (**). In order to calculate 2 , we note that
1

Hmn
= m|H  |n
= bm|x2 |n

2
20

and this is non zero , only when m=n-2, m=n or m=n+2. Thus
2 =




m



|Hmn
|2
1 2 |n 2|x2|n
|2 |n + 2|x2|n
|2 
= b
+
En Em
2
En En2
En En+2

1 2 h
1

= b 2 2 n+
8 m
2
which agrees with eqn (***).

21

1.4

TIME-DEPENDENT PERTURBATION THEORY

The time-dependent perturbation theory, which is sometimes called the method

of variation of constants, assumes as in the last section that
H(t) = H0 + h(t)

(1)

H0 u k = E k u k

(2)

where H0 is the unperturbed Hamiltonian with known eigenvalues Ek and

eigenstates uk , and h (the perturbation) is small, (i.e., 1) and depends
on time t. h has the eect of causing transitions between eigenstates of H0
that would be stationary in the absence of h. We use the adiabatic approximation which assumes that H(t) contains parameters which change very
slowly with time. For time-dependent perturbation we work with the time..
..
dependent Schrodinger equation. The time-dependent Schrodinger equation
for the unperturbed state is
ih

uk (t)
= H0uk (t) = Ek uk (t)
t

i.e.

uk (t)
i
= Ek uk (t)
t
h

uk (t) = uk (0)e

iEk t
h

(3)
..

uk (0) is the solution of the unperturbed time-independent Schrodinger eq.(2).

..
The time-dependent Schrodinger equation for the perturbed state is
ih

= H(t)
t

(4)

We express as an expansion in the eigenfunctions uk e

ak (t) that depend on time.
=

n

an (t)une

iEk t
h

with coecients

iEnt
h

(5)

ak (t) = uk

(6)

ak (t) is the probability amplitude for state uk , |ak (t)|2 is the probability that
the system is in the state uk . Substituting eq.(5) into eq.(4) we have


n

ih an un e

iEn t
h

n

an En une

iEnt
h

22

n

an (H0 + h(t))une

iEn t
h

where an = dadtn . We replace H0 un by En un on the RHS, multiply through on

the left by uk , and sum over all k, making use of the orthonormality of the
us to have


iEn t
iEkt
.
ih ak e h = an (t)e h k|h(t)|n

(7)
where k|h(t)|n
=

all
space

uk h(t)un d . We

kn =

dene the Bohr (angular) frequency

Ek En
h

(8)

and eq.(7) becomes

.

ak =

k|h(t)|n
an (t)eikn t
ih n

(9)

First-order Perturbation
We express the an (t) in eq.(9) as a power series in :
(1)
2 (2)
an (t) = a(0)
n (t) + an (t) + an (t) + . . .

(10)

Substitute eq.(10) into eq.(9), equate the coecients of corresponding powers

of , and set = 1 in nal result to have the set of equations
a(0)
k (t) = 0
.

(11a)

1

ikn t
k|h(t)|n
a(s)
(11b)
n (t)e
ih n
Eq.(11) can be integrated successively to obtain approximate solutions to any
desired order in the perturbation. For rst-order perturbation s = 0 and
a(s+1)
(t) =
k
.

a(1)
k (t) =
.

1

ikn t
k|h(t)|n
a(0)
n (t)e
ih n

(12)
(0)

where a(0)
n (t) are the solution of eq.(11a), the zero-order coecients ak , which
(0)
are constant with time. The ak are the initial conditions of the system before
the perturbation is applied.
(0)
Assuming, for simplicity, that all except one of the ak are zero, so that the
system is initially in a denite unperturbed energy state |m
.
We thus put

(0)

ak = k|m
= km

23

(13)

Put eq.(13) into eq.(12) to have

a(1)
k (t) =
.

(1)

ak (t) =

1
k|h(t)|m
eikm t
ih

1 t

k|h(t )|m
eikm t dt

ih or 0

(1)

ak (t) is the rst-order probability amplitude at any time t > 0 for state |k
,
(1)
|ak (t)|2 is the rst-order probability of nding the system in the state |k

i.e. the transition probability from state |m
to |k
after time t > 0.
Examples
1. A system of hydrogen atoms in the ground state is contained between the
plates of a parallel-plate capacitor. A voltage pulse is applied to the capacitor
so as to produce a homogeneous electric eld

E=

0 t0
e 0 exp(t/ )

t>0

Show that, after a long time, the fraction of atoms in the 2p (m = 0) state
is, to rst order
215 a20 e2 20
310 h
2( 2 + 12 )
where a0 is the Bohr radius, and h
is the energy dierence between the 2p
and the ground state.
Solution: First-order time-dependent perturbation theory gives the result
that if a Hamiltonian h(t) is applied at t = 0 to a system in an initial state
|m
with energy Em , the probability that a transition has occurred to a state
|k
, with energy Ek at time t is |ak |2 , where
ak =

1 t
k|h(t )|m
exp(it)dt ,
0
ih

h
= Ek Em

(i)

In this problem h(t ) is the potential of the electron in the applied eld, whose
direction we take as the z-axis, i.e.
h(t ) = e 0 z exp(t / ),

t > 0

(ii)

For t = , eq.(i) becomes


e 0
k|z|m

exp{(i 1/ )t}dt
ak =
0
ih

24

(iii)

The time integral is


0

exp{(i 1/ )t}dt =

1
1

[exp{(i 1/ )t}]0 =
(iv)
i 1/
i 1/

e 0
k|z|m

(v)
h
( + i/ )
The initial state is the ground state of the hydrogen atom, i.e., the quantum
numbers are n = 1, l = 0, m = 0 and the nal state has n = 2, l = 1 m =
0. We evaluate the matrix element in spherical polar coordinates r, , using
the relation z = r cos . Then
Therefore,

ak =

k|z|m
=


all
space

u210r cos u100dv

dv = r2 dr sin dd
2 1
u100 = R10 Y00 =
exp(r/a0)
4 a3/2
0
where

1
r
1
u210 = R21Y10 =
exp(r/2a0) cos
4 (2a0)3/2 a0

1 1  4
3r
Thus k|z|m
= 3/2 4
r exp(
)dr cos2 sin d
0
2 a0 0
2a0

(vi)
(vii)
(viii)
(ix)
(x)

 6

4! 2
= 3/2
2
3


a0

(We have used the result 0 r4 exp(r)dr =

probability of a transition is

(xi)
4!
5 ).

From eqs.(v) and (xi) the

215 a20 e2 20
|ak | = 10 2 2 1
3 h
( + 2 )
2

2. A time-varying Hamiltonian h(t ) brings about transition of a system from

a state |k
at t = 0 to |j
at t = t with probability pkj (t). Use rst-order
time-dependent perturbation theory to show that, if pjk (t) is the probability
that the same Hamiltonian brings about the transition j k in the same
time interval, then pjk (t) = pkj (t).
Solution: The probability of a transition from |k
(at time zero) to |j
(at
time t) is
pkj (t) = |akj (t)|2
(i)
25

where the rst-order expression for akj (t) is

1 t
j|h(t )|k
exp(ijk t )dt
akj (t) =
ih 0
1 t
k|h(t )|j
exp(ikj t )dt
Similarly,
ajk (t) =
0
ih

Now h(t ) is a Hermitian operator. Therefore
k|h(t )|j
= j|h(t )|k

Also

h
kj = Ek Ej = hjk

(ii)
(iii)

(iv)
(v)

Therefore the integral in eq.(iii) is the complex conjugate of the one in eq.(ii).
Thus
ajk (t) = {akj (t)}
giving

pjk (t) = |akj (t)|2 = pkj (t)

Comment: It is in general true that the probability of transition between

two states, due to an external stimulus represented by h(t), is the same for
transition in either direction. The result is known as the principle of detailed
balancing
3. Consider a one-dimensional harmonic oscillator with angular frequecy 0
and electric charge q. At time t = 0 the oscillator is in ground state. An
electric eld is applied for time , so the perturbation is

h(t) =

q x
0

0t
otherwise

where is a eld strength and x is a position operator. Using rst-order

perturbation theory, calculate the probability of transition to the state n = 1.

26

Chapter 2
SCATTERING THEORY
Introduction: Much of what we know about the forces and interactions in
atoms and nuclei has been learnt from scattering experiments, in which atoms
in a target are bombarded with beams of particles whose nature are known.
Theoretically, the most signicant aspect of scattering processes is that of
the continuous part of energy spectrum. Here intensities are the objects of
measurement and prediction. These intensities, being measures of the likelihood of nding a particle at certain places, are related to the eigenfunctions.
Establishing this relation is the rst problem of scattering theory.
In an idealized scattering experiment a single xed scattering center is bombarded by particles (projectiles) incident along the z-axis. In the classical
limit each particle can be assigned an impact parameter, and an azimuth
angle, , which together with z dene its position in cylindrical coordinates.

Let d/d be the area which, when placed at right angles to the incident
beam, would be traversed by as many particles as are scattered into the
unit solid angle around a direction characterized by the angles and (g.
27

above). If the scattering potential is spherically symmetric, V = V (r) the

scattering becomes independent of . Therefore, classically it can be shown
that the dierential cross-section is


d
 d 


=
(1)
d sin  d 
The impact parameter is a function of . To determine this function from
Newtons second law is the problem of classical scattering theory. Eq.(1) is
valid if the de Broglie wavelength of the projectile is much smaller than the
dimension of the scattering region. As the wavelength increases quantum
features appear. Quantum Mechanics represents the particles in the beam by
wave packets. That is, the particles do not have sharp boundaries. Therefore,
in Quantum Mechanics particles need not actually touch each other as they
can do in Classical Mechanics. When two particles are near each other they
feel the inuence of each other by the degree of freedom which is common to
both. E.g., an electron has the degrees of freedom (mass, charge, spin) while
a neutron has (mass, spin). So an electron and a neutron can feel or relate
with each other by (mass, spin) which is common to both.
Elastic Scattering
It is scattering without energy loss or gain by the projectile, i.e. kinetic energy of projectile is conserved. In elastic scattering the scattering center on
the particles can be represented by a potential energy V (r) of nite range d
such that V (r) = 0, if r > d. The disc of radius d is called the scattering
cross-section. Scattering cross-section is a measure of the probability of scattering of the projectile by the target.
Denition of dierential cross-section: A beam of particles of mass m travelling along the z-direction with velocity v is scattered by a short-range potential V (r) centred on the origin O (g. below). We use spherical polar
..
coordinates r, , taking z as the polar axis. The Schrodinger equation is
used to calculate the probability of a particle being scattered into a small
solid angle d in the direction , . The probability is expressed in terms of
a dierential cross-section, d/d, which is dened as the ratio of the number of scattered particles dn(, ) per unit time within the solid angle d to
the incident ux F ;
d/d = dn(, )/F d
(2)
where d/d has dimension of surface. We assume:
1. Any interaction between the scattered particles themselves is negligible
28

2. Possible multiple scattering processes are negligible. (A multiple scattering process is a process in which a scattered particle can be scattered multiple
times in the same target range)
3. The incident beam width is much larger than a typical range of the scattering potential, so that the particle will have a well-dened momentum.
The total scattering cross-section sc =(total number of particles scattered
 d
per second)/F = d
d. Absorption cross-section ab = (total number of
particles absorbed per second)/F .

..

The Schrodinger equation for scattering

1. Ignore the center of mass motion
2. Consider only the motion in the center of mass frame i.e. relative motion
which is important for collision.

h
2 2

+ V (r) (r) = E(r)

2

(3)

where = reduced mass. This is the equation for the projectile; it is not an
eigenvalue equation. E is positive and is the energy given to the projectile
by the experimenter. Therefore, it is arbitrary. Divide (3) through by h
2 /2
to obtain
(
'
2V
2E
2
+ 2 2
h

2
2
(4)
or [ + k ] = U
2 2

where E = h2k and U = 2V

h2

Greens function method: We seek solution (r) = (r) exp(it) that

satises the boundary conditions that
29

(r) is nite at r = 0 and (r) 0 as r . In that case

f (, )
exp(ikr)
(5)
r
This result is obtained by Greens function method applied to (4) (details
deferred). Therefore,
(r) = exp(ikz) +

f (, )
exp i(kr t)
(6)
r
The rst term represents the incident particles moving along the z-axis, and
the second term, which is a spherical wave travelling outwards from the origin,
represents the scattered particles. We are conning the dicussion to elastic
scattering hence the value of k is the same in the incident and scattered
terms in (6). The function f (, ) is known as the scattering amplitude. The
dierential cross-section is related to f (, ) by
(r) = exp i(kz t) +

d/d = |f (, )|2

(7)

Proof: The area dA subtends a solid angle d at the origin, where

d = dA/r2

(i)

The term ui = exp(ikz) in the wave function (6) represents the incident
particles with density
Gi = |ui|2 = 1
(ii)
While the term us = 1r f (, ) exp(ikr) represents the scattered particles with
density
Gs = |us |2 = |f (, )|2/r2
(iii)
The number of scattered particles passing through dA in unit time = (density
2
of particles at r)v dA = |f (,)|
= |f (, )|2vd
(iv)
2
r
where v is the velocity of the incident and scattered particles. The incident
ux is F = Gi v = v
(v).
Inserting (iv) and (v) into the denition of d/d, eq.(2) we have (7)
Born Approximation
This is a rst-order perturbation calculation, valid for fast particles and a
weak potential, which means that the incident wave function is only slightly
perturbed by the potential. The approximation gives
m 
f (, ) =
V (r) exp(i.r)dv
(8)
2h2 all
space
30

where the quantity , known as the scattering vector, is dened by  = k k ,

where k is the wave vector of the incident particles, and k  that of the scattered
particles. The angle between k and k  is the scattering angle . For elastic
scattering

(9)
|k| = |k  |, and  = 2k sin
2
Example 1: Show that for a spherically symmetric potential


4 

V (r) exp(ik.r) =
V (r) sin krdr

k 0

Solution: Since the potential is spherically symmetric, it depends only on

the magnitude of r, thus we have


V (r) exp( k.r)d3r = 2


0

V (r)r dr


0

exp(ikrcos) sin d

(i)

where is the angle between r and the polar axis. Put u = cos . Then
du = sin
and


0

 1

exp(ikr cos ) sin d =

exp(ikru)du =

2
sin kr
kr

(ii)

Inserting (ii) in (i), we have the expected result.

Example 2: a) Particles are incident on a spherically symmetric potential V (r) = r exp(r), where and are constants. Show that, in the Born
approximation, the dierential
scattering
cross-section for the scattering vec

2

tor  is given by d/d = h2 (2m

2 + 2 )
b) Use this result to derive the Rutherford formula for the scattering of particles, namely, that for -particles of energy E incident on nuclei of atomic
number Z, the dierential scattering cross-section
for scattering at an angle

Ze2
to the incident direction is d/d = 8 E sin2 ( )
0
2
Solution: a) The Born approximation gives the amplitude
f () =


where

I=

all
space

mI
2h2

V (r) exp(i.r)dv

31

(i)
(ii)

Using the result that



I=

all
space

V (r) exp(i.r)dv =

4 
V (r)r sin rdr
k 0

V (r) = (/r) exp(r)dr

4 
I=
sin r exp(r)dr
0
The integral is easily evaluated expressing the sine function as

(example1)

with

sin r =

1
{exp(ir) exp(ir)}
2i

(iii)

(iv)

and noting that the integrated expression vanishes at r = due to the factor
exp(r). The result is
4
I= 2
(v)
+ 2
Thus from (i) and (v)

2m
d/d = |f (0)| = 2 2
h
( + 2)
2

(vi)

b) -particles are scattered by the electrostatic interaction between the charge

2e on the -particle and the charge Ze on the nucleus. The potential is
2Ze2 1
V (r) =
4 0 r

(vii)

This is represented by the potential in part (a), with

2Ze2
=
,
4 0

=0

(viii)

Inserting these values in (vi) gives

mZe2
d/d =
0h
2 (2

(ix)

The energy of the -particle is

E=

h
2k2
p2
=
2m
2m
32

(x)

where p is the momentum. Since = 2k sin( 2 )

2 2

 

h
= 8mE sin

(xi)

Inserting (xi) in (ix) gives the required result

Partial waves and phase shifts
..
We look for solutions of the Schrodinger equation
h
2k2
h
2 2

+ V (r) =
2
2

(10)

(where 2 is the Laplacian in spherical polar coordinates and V (r) is a spherically symmetric potential elastic scattering) which have the asymptotic
form
eikr
(+)
(11)
k = eikrcos + fk ()
r
We establish the connection between these solutions and
(r, , ) = Rl,k (r)Ylm (, ) =

ul,k (r) m
Yl (, )
r

(12)

which are simultaneous (common) eigenfunctions of H, L2 and Lz . The radial

functions Rl,k and ul,k satisfy the dierential equations

l(l + 1) 2
1 d 2d

(r
)
+
+ 2 V (r) k 2 Rl,k (r) = 0
2
2
r dr dr
r
h

(13)

d2
l(l + 1) 2

+
+ 2 V (r) k 2 ul,k (r) = 0
(14)
2
2
dr
r
h

respectively as well as the boundary condition that Rl,k be nite at the origin
ul,k = 0
(15)
For the region r > d, V (r) = 0

l(l + 1)
1 d 2d

(r
)
+
k 2 Rl,k (r) = 0
2
2
r dr dr
r

(16)

The asymptotic solution of (16) for large kr is (quoted)

Rl,k (r) =

sin(kr l/2)
cos(kr l/2)
ul,k (r)
Al
Bl
r
kr
kr
33

(17)

The rst term of (17) is called the regular solution while the second term is
the irregular solution. The magnitude of Bl /Al is a measure of the intensity
scattering.
Bl /Al = tan l
(18)
where l is called scattering phase shift. Hence (17) can be written as
sin(kr l/2 + l )
ul,k (r)
Cl
r
kr
or

ul,k (r)
sin(kr l/2 + l )

r
kr

(19)
(+)

(where Cl = 1 in order to x the normalization). If we expand k in terms

of the separable solution of (12) (i.e. Legendre polynomials) we get
(+)

k (r, ) =

al (k)pl (cos )

l=0

ul,k (r)
r

(20)

(+)
where k depends only on the angle between k and r but not on their
directions. The asymptotic expansion of the plane wave eikz gives

eikz = eikrcos

(2l + 1)il

l=0

sin(kr l/2)
pl (cos )
kr

(21)

Substituting (21) into (11), and (19) into (20) and comparing coecients of
we have
the terms of the form pl (cos ) exp(ikr+il/2)
kr
(2l + 1)il il
e
al =
(2)3/2

(22)

Hence asymptotically, (i.e. using (22) and (19) in (20))

(+)

k (r, )

(2l + 1)il eil

l=0

sin(kr l/2 + l )
pl (cos)
(2)3/2kr

(23)

Eq.(23) diers from a plane wave by the presence of and is called distorted
ikr
plane waves. Put (21) into (11) and compare the coecients of e r in (11)
and (23) to have
fk () =

(2l + 1)

l=0

34

e2il 1
pl (cos )
2ik

(24)


1

(2l + 1)eil(k) sin l (k)pl (cos )
(25)
k l=0
Each l term in (24) or (25) is known as a partial wave. The l = 0 wave is
termed an s-wave, the l = 1 a p-wave, the l = 2 a d-wave and so on (same
nomenclature as for atomic states). The lth wave, being an eigenfunction of
the operator L2 , is a state in which
the particle has orbital angular momentum

about the origin of magnitude {l(l + 1)}h lh. Since the linear momentum
of the particle is h
k, a classical picture of the lth wave is that it corresponds
to the particle passing the potential at a distance l from its centre, where
h
kl = lh, i.e.
(26)
l = l/k

or

fk () =

Eq.(25) shows that, if l = 0, that partial wave gives no contribution to f ().

If l , the distance of closest approach, is much larger than d, the range of the
potential, the particle passes outside the potential and is not scattered, so
l = 0 for that partial wave. It follows from (26) that only those partial waves
with l dk have non-zero values of l , and contribute to the scattering.
If dk 1, as in the case for nuclear scattering at low energy all the l = 0,
except 0 . Since k = 2/ the condition dk 1 2d/ 1 
d i.e. the wavelength of the particle is large compared to the range of the
potential. In this situation, when only the s-wave is disturbed, the angular
dependence of f () is given by the Legendre polynomial p0(cos ) = 1 so the
scattering is spherically symmetric. As k 0, tan 0 is proportional to k,
i.e. 0 0 or n and the limiting value of tan 0 /k, is termed scattering
length. the concept is most commonly applied to the scattering of thermal
neutrons.
Example: 1. a) Verify that, outside the range of a short-range potential,
the wave function


1
i
(r, ) =
1+
exp(ikr) cos
r
kr
represents an outgoing p-wave.
b) A beam of particles represented by the plane wave e(ikz) is scattered by an
impenetrable sphere of radius d, where kd 1. By considering only s and p
components in the scattered wave, show that, to order (kd)2, the dierential
scattering cross-section for scattering at an angle is
1
d/d = d2{1 (kd)2 + 2(kd)2 cos }
3
35

[The value of cos2 averaged over all directions is 1/3]

Solution: a) The dependence on for a p-wave (l = 1) is given by the
Legendre polynomial p1 () = cos . So the angular part of the given function
(r, ) has the required form for a p-wave. It remains to show that the radial
part R(r) has the correct form. Put R(r) = u(r)/r. From (14), outside the
range of the potential, u(r) satises

d2 u 2 l(l + 1)
+ k
u = 0,
dr2
r2

(i)

with l = 1. For the given (r, ),



i
exp(ikr).
u(r) = 1 +
kr

(ii)

Dierentiating this twice with respect to r and substituting in (i) shows that
the equation is satised with l = 1.
b) Outside the range of the potential, the wave function for s and p waves
has the form


B
i
A
1+
exp(ikr) cos ,
(iii)
= exp(ikz) + exp(ikr) +
r
r
kr
where A and B are constants. The term in A is the spherically symmetric
s-wave, and the term in B is the p-wave. As r
= exp(ikz) +

f ()
exp(ikr)
r

(iv)

Equating terms in (1/r) exp(ikr) in (iii) and (iv) gives

f () = A + B cos

(v)

Since the scattering object is an impenetrable sphere, the wave function vanishes on the surface r = d. Thus


A
B
i
exp(ikd cos ) + exp(ikd) +
1+
exp(ikd) cos = 0.
(vi)
d
d
kd
The terms on the RHS that are independent of must sum to zero, and
similarly the terms proportional to cos . Expanding exp(ikd cos ) in powers
of kd we have
1
exp(ikd cos ) = 1 + ikd cos (kd)2 cos2 + 0(k 3d3).
(vii)
2
36

We cannot make the terms in cos2 sum to zero in (vi), because that would
require the contribution from d-waves, which are excluded. However, we need
to take into account the average value of cos2 , which is 1/3. Summing the
terms in (6) which are independent of gives
A
1
1 (kd)2 + exp(ikd) = 0
6
d
1
whence A = d{1 (kd)2} exp(ikd).
6
Summing the terms in cos gives


(viii)
(ix)

B
i
ikd +
1+
exp(ikd) = 0.
d
kd

(x)

Since kd 1, i/kd  1. Therefore to terms in (kd)2

B = d(kd)2.

(xi)

From (v),(ix), and (xi), the dierential scattering cross-section is

1
d
= |f ()|2 = |A + B cos |2 = d2{1 (kd)2 + 2(kd)2 cos },
d
3

(xii)

to terms in (kd)2
Example 2 A particle of mass and momentum P = h
k is scattered by
r/a
the potential V (r) = e r Vo a, where Vo and a>0 are real constants (Yukawa
potential).
(a) Using the Born approximation, calculate the dierential cross section (b)
obtain the total cross section.
Solution
(a) The range of the Yukawa potential is characterized by the distance a. We
2 /, so that the Born approximation is valid for all
assume that Vo a2 h
values of ka. The scattering amplitude is then given by:
1 2Vo a  iq.r er/ro 3
f (, ) =
dr
e
4 h
r
2
Since the potential has spherical symmetry V(r) = V(r), we can carry out
the integration using the relation


r2eiq.r V (r)drd =

4 
sin(qr)V (r)rdr
q 0

37

where r =|r| and d = sin dd. Therefore,

oa
f () = 2V
h2 k


r/a
dr
0 r sin(kr)e

2Vh2oa

Finally, the dierential cross section

d
d

= 2Vh2o a

1
1+q 2 a2

1
1+[2ka sin 2 ]2

= |f ()|2 is:

d() 42 Vo2 a6
1
=
4
d
h

[1 + 4k 2a2 sin2 2 ]2

Note that due to spherical symmetry, the cross section does not depend on
the azimuthal angle.
(b) The total cross section is obtained by integration


d()
4
42 Vo2 a6
d =
4
d
1 + 4k 2a2
h

Note that the innite range limit ( a , Vo 0 and Vo a = Z1 Z2 e2 =constant

of the Yukawa potential corresponds to the Coulomb interaction between two
ions of charges Z1 e and Z2 e. At this point (*) reduces to the well -known
Rutherford formula
d
42 Z12Z22 e4
Z12Z22 e4
= 4
=
d
h
16K 4 sin4 2
16E 2 sin4 2
2 2

where E = h2k is the energy of the particle in the center of mass frame and
is their reduced mass.
Example 3: A point particle is scattered by a second particle with rigid
core; that is, the scattering potential is V(r) = 0 for r> a and V(r) =
for r < a. The energy of the scattered particle satises ka = 1. (a) Find the
expression for l . Complete the table below (express l in radians).
tan l l sin l
l=0
l=1
l=2
(b) Calculate the dierential cross section d/d for angles 0 and , taking
into account only the waves l =0 and l =1. (c) Compute the total cross
38

section T taking into account only the waves l =0 and l =1. (d) What is the
accuracy of part (c)?
Solution
(a) The phase shifts for a rigid sphere is given by the equation:
tan l =

jl (ka)
nl (ka)

Using the known expressions for the Spherical Bessel functions:

sin x
no (x)= cosx x
x
sin x cos x
x
sin x
n1 (x) = cos
j1 (x) = 2
2
x
x
x
x




3
1
cos x
x
j2 (x) = 2
sin x 3 2 n2(x) = x32 x1 cos x 3 sin
2
x
x
x
x
jo (x) =

and substituting x=ka =1, we nd tan o = -1.56, tan 1 = -0.22 and tan 2 =
-0.02. Therefore
tan l
l = 0 -1.56
l = 1 -0.22
l = 2 -0.02

l
-1.00
-0.22
-0.02

sin l
-0.84
-0.22
-0.02

(b) The dierential cross section is given by:

1

d
= 2 | (2l + 1)eil sin l Pl (cos )|2
d k l=0

For l =0, 1 and k =a1

d
=
a2 | sin o eio + 3 sin 1 ei1 cos |2
d
= a2 [sin2 o + 6 sin o sin 1 cos(o 1) cos
+9 sin2 1 cos2 ]
Substituting = 0, , we obtain:
d
|0, = a2 | sin2 o 6 sin o sin 1 cos(0 1 ) + 9 sin2 1 |
d
with 1 and 0 given in the table above.
39

(c) The total cross section is given by:

T =

4

(2l + 1) sin2 l
2
k l=0

for l =0, 1 and k =1/a

T = 4a2 [sin2 o + 3 sin2 1 ] 0.854a2.
(d) A rough estimate on the accuracy of the calculation in part (c) is given
by calculating the additional term l =2.
T (0.85 + 0.002)4a2
Example 4: (a) Consider scattering from a spherical symmetric potential.
The solution of the Schrodinger equation is given by the expansion (r, ) =

l=0Rl (r)Pl (cos )), where R(r) is the solution of the radial wave equation
and Pl (cos ) is the Legendre polynomial of order l. In the limit r the
asymptotic form of the wave function is:
1
(r, )r eikz + f ()eikr
r

(1)

where f() is the scattering amplitude. Similarly, the asymptotic form of R(r)
is



Al sin kr 2 l + l
(2)
R(r)r
kr
where l are phase shifts. (i) use expressions (1) and (2) to obtain the Legendre expansion of f()
(b) Show that the total cross section is given by
T =

4

(2l + 1) sin2 l
2
k l=0

Solution (a) The asymptotic form of the wave function is as given above:


(r, )r

Al

sin kr 2 l + l

l=0

kr

1
Pl (cos ) = eikr + f ()eikr
r

Using the Legendre expansion of eikz ,

eikz = eikr cos =

(2l + 1)il jl (kr)Pl (cos )

l=0

40

we nd that


Al

sin kr

2l

+ l

kr

l=0



=
where f ()

1
Pl (cos ) = eikr + f ()eikr
r

(2l + 1)i

sin kr

2l

kr
l=0

=
l=0 fl Pl (cos ). Now,


i(

I Al e
II



1
ikr
+ fl ()e Pl (cos )
r

+ l

we write sin x =


kr 2 l+l

eix eix
,
2i

and obtain

kr 2l

(2l + 1)il e
= 2ikfl eikr
l
l
Al eikr 2 +l 2l + il eikr 2 = 0

Therefore from (II), we obtain Al = (2l + 1)il eil and then by subst back in
(II), we have
f () = (2ik)

1

(2l + 1)(e2il 1)Pl (cos )

l=0

(b) The total cross section is



T =

|f ()|2d = 2

 1

d(cos )

|
(2l + 1)(e2i1 1)Pl (cos )|2
2
1
4k
l=0

1
= 2
d(cos )
(2l + 1)(2l + 1)(e2il 1)(e2il 1)Pl (cos )Pl (cos )
2k 1
l,l =0

Now,

 1
1

d(cos )Pl (cos )Pl (cos ) =

2
ll .
(2l + 1)

Therefore
T =

4

2il
2il
(2l
+
1)(2

e
)
=
(2l + 1) sin2 l
2
2
k l=0
k l=0

41