1

CHAPTER 15

1. With the perturbing potential given, we get

C
(
1
s→2
p
)=
e
E
h
i
0
〈
21
0|
z|0∫
1
∞
〉d
0
t
eω
i
eγ
t−
t
φφ
where ω = (E21 – E10). The integral yields 1
/γ−
( ω
i )so that the absolute square of
C(1s2p) is

→ =
φ
|〈0|
2
1 zφ
|10〉
0
2
|
22
P
(
1s 2
p) e
E22 2
h(ω +γ) 0

1
5
2 2 2
φ
We may use |〈2
1
0 |z|φ1
0
0 〉
| =0a
1 0
to complete the calculation.
3

2. Here we need to calculate the absolute square of

1
h
i ∫
T
d
t
0
eω
i2
1t
si
n ×
t
2
a∫
a
d
x
0
s
i
2
n(ωλ π
x ax
−
x)s
i
n
a 2a
π
Let us first consider the time integral. We will assume that at t = 0 the system starts in the
ground state. The time integral then becomes

∞
eωs ω ωω ωω 1∞ ω
∫d =∫d + −
− =
i2t i
(2 )
t i
(2 )
t
t i
n t t{
e e }
ω ω
1 1 1

0 2
i − 0 2 2
2
1

We have used the fact that an finitely rapidly oscillating function is zero on the average.
In the special case that ω matches the transition frequency, one must deal with this
integral in a more delicate manner. We shall exclude this possibility.

The spatial integral involves

2 a 2πx πx a
a∫0
d
xsin
a
sin (x− )=
a 2
1 a πx 3πx a
∫ co
a 0 a
s −c
o s
a
(x− )
2
1 a  d a πx a 3πx a a πx a 3πx
= ∫dx  sin − sin (x− )− sin − sin 
a 0 dx π a 3π a 2  π a 3π a 
a
1a2 πx a2 3πx a8
=  2co
s − 2co
s  =−2 2
aπ a 9π a 0 π9

The probability is therefore

 2

λ

1
6  ω
a
2 2 2

2=
P
1 
h
9π(
2
ω
2
21−ω
22
)

(b) The transition from the n = 1 state to the n = 3 state is zero. The reason is that the
eigenfunctions for all the odd values of n are all symmetric about x = a/2, while the
potential (x – a/2) is antisymmetric about that axis, so that the integral vanishes. In fact,
quite generally all transition probabilities (even  even) and (odd  odd) vanish.

(c) The probability goes to zero as ω  0.

3. The only change occurs in the absolute square of the time integral. The relevant one is

ω τ
πω
τ
∞
∫d t−
22
−22
e =i2 t/ /4
t
e e1

−
∞

which has to be squared.

When τ  ∞this vanishes, showing that the transition rate vanishes for a very slowly
varying perturbation.

4. The transition amplitude is

λ h ∞
C =〈 m
| (
A ++
A 〉
)|n∫deiω
t (m−
nt−
)
eα
t
csω
o t
→
n m
h
i 2
M ω 0
1

α ω
=
−λ 1δ
i
2
Mhωm −
,n1nδ
+m +
,n(+
1n1
α
( −
i
−
ω
i
(
m
(
m
−n
−n
)+
)2
)
ω2
1
)
(a) Transitions are only allowed for m = n ± 1.

(b) The absolute square of the amplitude is, taking into account that (m – n)2 = 1,

λ(δ
2

δ αω
+ 2 2

hω α ωω αω
n +
(
n+1
)
2
M (+−
m
,
n−
)+
1
4 m
,
n+1
) 2 2 2
1
2 22

When ω 1  ω , nothing special happens, except that the probability appears to exceed
unity when α 2 gets to be small enough. This is not possible physically, and what this
suggests is that when the external frequency ω 1 matches the oscillator frequency, we
get a resonance condition as α approaches zero. Under those circumstances first order
perturbation theory is not applicable.
When α  0, then we get a frequency dependence similar to that in problem 2.

5. The two particles have equal and opposite momenta, so that

 3

i= +
2 24
E (
p
c ) m
ic

The integral becomes

( ) δ
1 ∞

π∫∫
Ω − −
2 2
d pd
pM
c E(p
)E(p
)
(
2h)
6 0 1 2

and it is only the second integral that is of interest to us. Let us change variables to

u = E1(p) + E2(p)

then
2 2
p
c pc p
d
p
d
u=d p+dp=
(E
1+E
)
2
E
1 E
2 EE
12

and the momentum integral is

( ) E
E du
δ δ
∞ ∞
∫pd c2− − =∫ c2−
2
p M E(
1p) E
2(
p) p1 2
2
(M u)
0 (
m +
1 m
2)
c2
u
c
E
E
=p14
2

M
c

To complete the expression we need to express p in terms of the masses.

We have

(
m2
2
c2
)+22
p
c =(
Mc2
−(
m1
2
c2
)+22
p
c 2
)
=
(c2
M)2
−c2
2
M E
1(
p)+(
m1
2
c2
)+22
p
c

This yields

+ −
22 22 22
(M
c) (
mc) (
mc)
E
(
1p
)= 1
2
2
2M
c

and in the same way

+ −
22 22 22
(M
c) (
mc) (
mc)
E
(
2p=
) 2
2
1
2M
c

By squaring both sides of either of these we may find an expression for p2.
The result of a short algebraic manipulation yields
 4

2
c
= −− −+ +− ++
2
p (
M
2 m
1m)
(
2Mm
1m)
(
2Mm
1m)
(
2Mm
1m)
2
4
M

6. The wave function of a system subject to the perturbing potential

λ V(t) = V f(t)

where f(0) = 0 and L

i
mf (t)=1
, with df(t)/dt << ω f(t), is given by
t→∞

ψ ∑ − h
φ
0
〉=C 〉
i
E t
/
| (
t) m(
t)
e m
|m
m

and to lowest order in V, we have

1
∫ ω
φ φ
t
= 〈 〉
it'
C
(t
) dt'
e f(
t
') |
V |
m
h 0
i m 0

where ω = − h and at time t = 0 the system is in the ground state. The time
0 0
(
EmE)
0/
integral is
ω
ω ω ω
it
'

ωω ω
de 1
t d 1
∫ ∫ ∫ ∫
t t t
= ) = −
it
' it
' it
'
d
t
'
e f
(
t'
) d
t '
f(
t' d
t
'(e f(
t
'
) d
t
'ed
f (
t
'
)/d
t
'
0 0 d
t
'i i0d
t' i0

The second term is much smaller than the term we are trying to evaluate, so that we are
/iω
iω
left with the first term. Using f(0) = 0 we are left with e t
,since for large times
f(t) = 1. When this is substituted into the expression for Cm(t) we get
ω
it

C
(
mt
)=
(
e
−0 0〈
E− E
φ|
V
)m 0
φ
| 〉m
≠0
m 0

Insertion of this into the expression for |ψ (t)> yields

ψ φ φ
〈 |
V φ
| 〉
φ
− h
∑
0
〉
= 〉
+ 〉
i
E t
/ m 0
| (
t
) | e 0
0 0|
0
≠
m
0 E
0− Em
m

On the other hand the ground state wave function, to first order in V is

φ φ
〈 |V|φ〉
φ
〉
|w
0=| 〉
0+∑ n
0
−
0
0| 〉
n
≠
n0E0 En

It follows that
〈
w|(ψ
〉
t
)=+
1−
i
E
e
0
0t
/h
∑
〈|
0V|m〉
〈m|
V 〉
|0 φφφφ
0
≠
m
0 (E−
0 0
0Em
2
)
 5

Thus to order V the right side is just one.

A fuller discussion may be found in D.J.Griffiths Introduction to Quantum Mechanics.i

7. The matrix element to be calculated is

2 −
ip/h
.r
e 3e
i=
M − ∫dr3
∫dr2..∫dr
3 3
Φ*
f(r
,r,.
.r ∫r V
)d
f
π
4ε0
1 A 1 2 A

Z
1
∑
|r−
ψ(r
r|
)Φ(r
,r,.
.r)
1
0
0 i 12 A
i=
1 i

The summation is over I = 1,2,3,..Z , that is, only over the proton coordinates. The
outgoing electron wave function is taken to be a plane wave, and the Φ are the nuclear
wave functions. Now we take advantage of the fact that the nuclear dimensions are tiny
compared to the electronic ones. Since |rI | << |r |, we may write

1 1r •r
=+ 3i +
..
−
|r r
i| r r

The 1/r term gives no contribution because 〈 Φ f|Φ i〉= 0. This is a short-hand way of
saying that the initial and final nuclear states are orthogonal to each other, because they
have different energies. Let us now define
Z
=
d∑∫
dr
1∫
3 3
dr
2.∫
dr
AΦ
3 *
f(r
,
r
1,
2.
)rΦ
j (
ir,
r
1,
2.
)
=
j1

The matrix element then becomes

− h
•
2 i
p.r
/

3ψ
e 3e dr
i=− ∫
M
f
4π
ε0
d
r
Vr 1 0
0(
r)

The remaining task is to evaluate this integral.

First of all note that the free electron energy is given by
2
p
=∆+
E |E
1
00|
2
m

where ∆ E is the change in the nuclear energy. Since nuclear energies are significantly
larger than atomic energy, we may take for p the value p =2 m ∆E .
To proceed with the integral we choose p to define the z axis, and write p/h= k . We
write the r coordinate in terms of the usual angles θ and φ . We thus have
 6

d
.
r
∫dr
e3 −
i
p.r/h
3ψ0(
1
0 r
)=
r
2
3
/2
∞ Z
∫d
Ω∫d
r
e −
i
k
rcsθ
o
(
d s
i
n θ
c
osφ
+d s
i
n θ
s
inφ+
d c
osθ
)  e−
Zr/a
0

0
x y z
π0
4a

2π
The solid angle integration involves ∫0
dφ, so that the first two terms above disappear.
We are thus left with

1
3
/2
Z ∞
π
 2 z∫ d sθ
)∫ d sθ
1
sθ −
d (
c
o rc
o −
ei
krc
o
eZr/a
0
=
πao −1 0

1 sθ
3
/2

Z 2 c
o
1
π(dˆ)∫ d
.p (
c
os θ
)
πao
−1
0+
(Z/a i
kcsθ
o )

The integral, with the change of variables cosθ = u becomes

u
∫
1
du =
−1 Z/ a0 +iku
u(Z/a −iku)
∫−1du (Z/a0)20 +k2u2 =
1

u2
−ik∫−1du
1

(Z/ a0)2 +k2u2

−i k w2 2i  a0 ak 
2 ∫−k dw 2 =− 2 k− arctan( 0 )
(Z/ a0) +w k  Z 
2
k Z

k
a kh ∆
2 E1 ∆ E
= =2 22 =
Note now that
αZ αZ . If Z is not too large, then the
0
Zm c
Z m
c (
1
3.6
e
V )
factor is quite large, because nuclear energies are in the thousands or millions of electron
volts. In that case the integral is simple: it is just

 p  π
3
/
2
 • hZ
1Z
 (π
2
d
−
) 2(2h)
i 1− 
π
a0 p 
2ap
0

We evaluate the rate using only the first factor in the square bracket. We need the
absolute square of the matrix element which is


3

2 2
e  Z (
d.
p)
(− 2
)16 π h 2 
4πε 0V
a p
0
4

The transition rate per nucleus is

 7

πd
2 3
p
V p 2

i= ∫ δ( −∆ 2
R E)|M
i|
f
h (
2πh) 2
3
m f

πd e2 2 Z 3
2 3
p
V p 2
1  
 ( •
dp)2
= ∫ δ
( −∆
E) 1
6πh2

h (
2πh)3 2m V4π
ε0 a
0 p 4

In carrying out the solid angle integration we get

π 22
4
∫Ω •)= |d 2
d (
d p |p
3

so that we are left with some numerical factors times ∫

d
p δ
(p/
22
m−
∆E=
)
∆
2E
m

Putting all this together we finally get

2
1
6 3d c2m
m c2
i= (
R Zα) 2
f
3 a ∆
0 2Eh

We write this in a form that makes the dimension of the rate manifest.