Solutions to Lectures on

Quantum Mechanics

Steven Weinberg
The University of Texas at Austin
Prepared by Dr Joel Meyers.
 Chapter 1 Problem Set Solutions 1

Chapter 1 Problem Set Solutions

1. Consider a non-relativistic particle of mass M in one dimension, confined in

a potential that vanishes for −a ≤ x ≤ a, and becomes infinite at x = ± a,
so that the wave function must vanish at x = ± a.
• Find the energy values of states with definite energy, and the corresponding
normalized wave functions.
• Suppose that the particle is placed in a state with a wave function pro-
portional to a 2 − x 2 . If the energy of the particle is measured, what is the
probability that the particle will be found in the state of lowest energy?
The states of definite energy are those which are solutions of the time-
independent Schrödinger equation
2 d 2
− ψ(x) + V (x)ψ(x) = Eψ(x).
2M d x 2
The potential for the infinite square well is given by

0 for |x| < a
V (x) =
∞ for |x| ≥ a.
This potential requires that the wave function vanishes outside the well
ψ(x) = 0 for |x| ≥ a,
while inside the well the Schrödinger equation becomes
d2 2M E
ψ(x) = − 2 ψ(x) for |x| < a.
dx 2 
The solution of this differential equation takes the form
ψ(x) = A sin(kx) + B cos(kx),
where we have defined
√
2M E
k≡ .

Continuity of the wave function requires that we impose the boundary
conditions
ψ(x = ± a) = 0,
and so
0 = A sin(ka) + B cos(ka),
0 = −A sin(ka) + B cos(ka).
2 Solutions to Lectures on Quantum Mechanics

Adding these two equations gives

0 = 2B cos(ka),
which requires that
nπ
B = 0 or k = where n = 2, 4, 6, . . . ,
2a
and subtracting the equations gives
0 = 2A sin(ka),
which requires that
nπ
A = 0 or k = where n = 1, 3, 5, . . . .
2a
In either case, the energy levels are given by
2 k 2 n 2 2 π 2
En = = .
2M 8Ma
Next, we need to normalize the wave functions in the sense of Eq. (1.5.4),
such that
 ∞
|ψ(x)|2 d x = 1.
−∞

For n odd, we have

 a  nπ x 2  a  
  2 nπ x
 B cos  d x = |B| 2
cos d x = |B|2 a = 1 ,
−a 2a −a 2a
which requires that
1
B=√ ,
a
where we have fixed the arbitrary complex phase for convenience. For n even,
the normalization condition takes the form
 a  nπ x 2  a  nπ x 
 
 A sin  d x = |A| 2
sin2 d x = |A|2 a = 1 ,
−a 2a −a 2a
which requires that
1
A= √ ,
a
where we have again fixed the arbitrary complex phase for convenience.
Summarizing our results, the normalized wave functions of states with
definite energy are
 Chapter 1 Problem Set Solutions 3
⎧ 1  nπ x 
⎪
⎨ √a cos 2a  for |x| < a and n = 1, 3, 5, . . .
ψn (x) = √1a sin nπ x
for |x| < a and n = 2, 4, 6, . . .
⎪
⎩
2a
0 for |x| ≥ a

and the energy levels are

n 2 2 π 2
En = .
8Ma
Since these states are normalized and the energy level is determined by n,
the argument below Eq. (1.4.26) proves that the states of definite energy are
orthonormal
 ∞
ψm∗ (x)ψn (x) d x = δmn .
−∞

Let us define a state which vanishes for |x| ≥ a, while for |x| < a it is
given by
 
(x, t = 0) = C a 2 − x 2 .

We must normalize this state in order to determine C.

 a 5 a
  2  2 3
16a 5 2
C a − x 2 2 d x = |C|2 x − 2a x + a 4 x = |C| = 1,
−a 5 3 −a 15
and so we find (after fixing an arbitrary phase)

15
C= ,
16a 5
which then gives

15  2 
(x, t = 0) = a − x 2
.
16a 5
Since the states of definite energy that we found above form a complete
orthonormal set on the interval −a < x < a, we can express  in terms
of those states

(x, t = 0) = cn ψn (x),
n

and the probability of measuring the state  to have energy E m , following

Eq. (1.5.18), is given by
 a 2
 
P ((x, t = 0) → ψm (x)) = |cm | = 
2  ψm (x)(x, t = 0) d x  .
∗
−a
4 Solutions to Lectures on Quantum Mechanics

For the state of lowest energy this gives

 a
1  π x   15  
c1 = √ cos 5
a2 − x 2 d x
−a a 2a 16a
  
15 1 2a 3  π x  8xa 2 πx  π2x2
− 2 πx  a
4a 2
= sin − cos − sin
16 a 3 π 2a π2 2a π3 2a
8a 3 −a
 
15 1 4a 3 4a 3 32a 3
= − + 3
16 a 3 π π π
√
8 15
= ,
π3
and so the probability of finding  in the lowest energy state is
960
P ((x, t = 0) → ψ1 (x)) = |c1 |2 = ≈ 0.9986.
π6
2. Consider a non-relativistic particle of mass M in three dimensions, described
by a Hamiltonian

P2 Mω02 2
H= + X.
2M 2
• Find the energy values of states with definite energy, and the number of
states for each energy.
• Find the rate at which a state of next-to-lowest energy decays by photon
emission into the state of lowest energy.
Hint: You can express the Hamiltonian as a sum of three Hamiltonians for
one-dimensional oscillators, and use the results given in Section 1.4 for the
energy levels and x-matrix elements for one-dimensional oscillators.
First, we will rewrite the Hamiltonian in components

p2 Mω02 2
H= + x
2M 2
 2   2   2 
p1 Mω02 2 p2 Mω02 2 p3 Mω02 2
= + x + + x + + x
2M 2 1 2M 2 2 2M 2 3
= H1 + H2 + H3 ,

where H1 , H2 , and H3 are the one-dimensional harmonic oscillator Hamil-

tonians for x1 , x2 , and x3 , respectively. Now, we will assume a separable
solution of the form

ψ(x) = ψ1 (x1 )ψ2 (x2 )ψ3 (x3 ),

 Chapter 1 Problem Set Solutions 5

where
H1 ψ1 (x1 ) = E n 1 ψ1 (x1 )
H2 ψ2 (x2 ) = E n 2 ψ2 (x2 )
H3 ψ3 (x3 ) = E n 3 ψ3 (x3 )
H1 ψ2 (x2 ) = 0
H1 ψ3 (x3 ) = 0
..
.
and the E n are given by the energies of the one-dimensional harmonic
oscillator, Eq. (1.4.15)
 
1
En = n + ω0 ,
2
where n = 0, 1, 2, . . .. We therefore find for the three-dimensional harmonic
oscillator
H ψ(x) = E N ψ(x)
= H1 ψ1 (x1 )ψ2 (x2 )ψ3 (x3 ) + ψ1 (x1 )H2 ψ2 (x2 )ψ3 (x3 )
+ ψ1 (x1 )ψ2 (x2 )H3 ψ3 (x3 )
= E n 1 ψ1 (x1 )ψ2 (x2 )ψ3 (x3 ) + ψ1 (x1 )E n 2 ψ2 (x2 )ψ3 (x3 )
+ ψ1 (x1 )ψ2 (x2 )E n 3 ψ3 (x3 ),
and so
 
3
E N = En1 + En2 + En3 = n1 + n2 + n3 + ω0 ,
2
where n 1 , n 2 , and n 3 are each non-negative integers. If we define
N = n1 + n2 + n3,
then the energy levels for states of definite energy are
 
3
EN = N + ω0 ,
2
where N = 0, 1, 2, . . ..
Now we must count the number of states with each energy. For a definite
value of N , the integer n 1 can take values 0, 1, . . . , N , then n 2 will take
values 0, 1, . . . , N − n 1 (which represents N − n 1 + 1 possibilities), and n 3
is fixed to be N − n 1 − n 2 . Then for each energy E N there is a degeneracy

N
1 1
gN = (N −n 1 +1) = N (N +1)− N (N +1)+N +1 = (N +1)(N +2).
n1 =0
2 2
6 Solutions to Lectures on Quantum Mechanics

The rate of spontaneous emission of photons carrying energy ωnm = E m −

E n is given by Eq. (1.4.5) to be

4e2 ωnm
3
Anm = |[x]nm |2 .
3c3 
In three dimensions, we have E 1 = 52 ω0 and E 0 = 32 ω0 , so ω01 = ω0 .
The relevant matrix elements for the one-dimensional harmonic oscillator
are given by Eq. (1.4.15) to be

(n + 1)
[x]∗n+1,n = [x]n,n+1 = e−iω0 t .
2m e ω0

Since the wave function for the three-dimensional harmonic oscillator is just
a product of three one-dimensional harmonic oscillator wave functions, we
can also take these matrix elements to represent the matrix elements of a
single spatial component x1 , x2 , or x3

[xi ]mn = [x]mn .

The first excited state has n 1 , n 2 , or n 3 equal to 1 with the others equal to
zero, and so the rate of spontaneous emission from the first excited state to
the ground state for the three-dimensional harmonic oscillator is given by
3  
4e2 ω01  2e2 ω02
A1 =
0
= .
3c3  2m e ω0 3c3 m e
3. Suppose the photon had three polarization states rather than two. What
difference would that make in the relations between Einstein’s A and B
coefficients?
For black-body radiation in a cubical box with side L, the frequency of a
normal mode is given by Eq. (1.1.2) as ν = |n|c/L. The number of normal
modes N (ν)dν in a range of frequencies between ν and ν + dν is three times
the volume of a spherical shell in frequency space (the factor of three here
comes from the assumed three polarization states of the photon)
 3
L
N (ν) dν = 3 × 4π|n| d|n| = 12π
2
ν 2 dν.
c
Assuming that the energies of the light quanta are integer multiples of hν,
the mean energy is
  
−nhν
n exp k B T nhν hν
Ē =    =   .
−nhν
n exp kB T
exp hν
kB T
− 1
 Chapter 1 Problem Set Solutions 7

Then the energy density in radiation between ν and ν + dν is given by

Ē N (ν) dν 12π h ν 3 dν
ρ(ν) dν = =   .
L3 c3 exp hν − 1
kB T

Now assume that we have black-body radiation in equilibrium with atoms at

a temperature T . The transition rate for atoms to go from state m to state n
must equal the rate for the transition from state n to state m, so
 
Nm Anm + Bmn ρ(νnm , T ) = Nn Bnm ρ(νnm , T ).
Using the Boltzmann distribution for the atoms gives
   
Nm (E m − E n ) −hνnm
= exp − = exp .
Nn kB T kB T
We can then rearrange the condition of equilibrium to give
  
12π h ν 3 dν −hνnm
Anm =   exp Bn
m
− Bm .
n
c3 exp hν − 1 kB T
kB T

Requiring that the Einstein coefficients A and B are temperature independent

then gives
Bmn = Bnm ,
and
 3 
12π hνnm
Anm = Bmn .
c3
This conclusion gives a value for A which is larger than the usual expression
Eq. (1.2.16) by a factor of 3/2.
4. Show that the solution ψ(x, t) of the time-dependent Schrödinger equa-
tion for a particle in a real potential has the property that ∂|ψ|2 /∂t is the
divergence of a three-vector.
We wish to calculate
∂
|ψ(x, t)|2 ,
∂t
for some ψ(x, t) which is a solution of the time-dependent Schrödinger
equation, which reads
∂ −2 2
i ψ(x, t) = ∇ ψ(x, t) + V (x, t)ψ(x, t).
∂t 2M
Taking the complex conjugate of the Schrödinger equation gives
∂ ∗ −2 2 ∗
−i ψ (x, t) = ∇ ψ (x, t) + V (x, t)ψ ∗ (x, t).
∂t 2M
8 Solutions to Lectures on Quantum Mechanics

Recall that |ψ(x, t)|2 can be rewritten as ψ(x, t)ψ ∗ (x, t), and so we have
∂ ∂ ∂
|ψ(x, t)|2 = ψ ∗ (x, t) ψ(x, t) + ψ(x, t) ψ ∗ (x, t).
∂t ∂t ∂t
Using the Schrödinger equation to replace the time derivatives (now dropping
space and time arguments), this becomes
   
∂ ∗ i 2 i −i 2 ∗ i ∗
|ψ| = ψ
2
∇ ψ − Vψ + ψ ∇ ψ + Vψ
∂t 2M  2M 
i ∗ 2 i i i
= ψ ∇ ψ− ψ∇ 2 ψ ∗ − V |ψ|2 + |ψ|2
2M 2M  
i  ∗ 2 
= ψ ∇ ψ − ψ∇ 2 ψ ∗
2M
i  
= ∇ · ψ ∗ ∇ψ − ψ∇ψ ∗ .
2M
We see that we can identify a three-vector
−i  ∗ 
j≡ ψ ∇ψ − ψ∇ψ ∗ ,
2M
such that
∂
|ψ|2 = −∇ · j.
∂t
Notice that this is a continuity equation which implies that probability is con-
served in quantum mechanics. The time rate of change of probability density
|ψ|2 in some infinitesimal volume is equal to the rate at which the probability
current j flows into the same infinitesimal volume.
 Chapter 2 Problem Set Solutions 9

Chapter 2 Problem Set Solutions

1. Use the method described in Section 2.2 to calculate the spherical harmonics
(aside from constant factors) for
= 3.
As discussed in Section 2.2, Y
m is given by a sum of terms, each of which
contains ν± factors of x̂± such that m = ν+ − ν− , and the total number of
factors of x̂+ , x̂− , and x̂3 in each term is
. The unit vectors are defined as in
Eq. (2.2.4)
x̂± ≡ x̂1 ± i x̂2 = r sin θ e±iφ ,
x̂3 = r cos θ.
Also, the spherical harmonics satisfy
 
∇ 2 r
Y
m = 0,
and the condition


d 2 Y
m∗ (x̂)Y
m (x̂) = δ

 δmm  .

We will now fix

= 3 and compute the spherical harmonics for each value
of m. Beginning with m = 3, we must have ν+ = 3 and ν− = 0, so
 3
Y33 ∝ x̂1 + i x̂2 = sin3 θ e3iφ .
For m = 2, we have ν+ = 2 and ν− = 0, which gives
 2
Y32 ∝ x̂1 + i x̂2 x̂3 = sin2 θ cos θ e2iφ .
For m = 1, we must have either ν+ = 2 and ν− = 1, or ν+ = 1 and
ν− = 0, so
 2    
Y31 ∝ A x̂1 + i x̂2 x̂1 − i x̂2 + B x̂1 + i x̂2 x̂32 .
 
We can find A and B by requiring that ∇ 2 r 3 Y3m = 0, which in this case
gives
      
∇ 2 r 3 Y31 ∝ ∇ 2 A x13 + i x12 x2 + x1 x22 + i x22 + B x1 x32 + i x2 x32
= A (6x1 + 2i x2 + 2x1 + 6i x2 ) + B (2x1 + 2i x2 )
= 8A (x1 + i x2 ) + 2B (x1 + i x2 )
= 0,
and so we find that B = −4A. Therefore, we can write
 2    
Y31 ∝ x̂1 + i x̂2 x̂1 − i x̂2 − 4 x̂1 + i x̂2 x̂32
 
= sin3 θ − 4 sin θ cos2 θ eiφ
 
= 1 − 5 cos2 θ sin θ eiφ .
10 Solutions to Lectures on Quantum Mechanics

For m = 0, we must have either ν+ = 1 and ν− = 1, or ν+ = 0 and ν− = 0,

so we find
  
Y30 ∝ C x̂1 + i x̂2 x̂1 − i x̂2 x̂3 + D x̂33 .

We can find C and D by the same method as above

     
∇ 2 r 3 Y30 ∝ ∇ 2 C x12 + x22 x3 + Dx33
= 4C x3 + 6Dx3
= 0,

and so we find D = − 23 C. Therefore, we can write

   2
Y30 ∝ x̂1 + i x̂2 x̂1 − i x̂2 x̂3 − x̂33
3
2
= sin θ cos θ − cos θ
2 3

 3

5
= 1 − cos2 θ cos θ.
3
For m = −1, we must have either ν+ = 1 and ν− = 2, or ν+ = 0 and
ν− = 1, so
  2  
Y3−1 ∝ E x̂1 + i x̂2 x̂1 − i x̂2 + F x̂1 − i x̂2 x̂32 .

We again find the coefficients by imposing the Laplace equation

      
∇ 2 r 3 Y3−1 ∝ ∇ 2 E x13 − i x12 x2 + x1 x22 − i x22 + F x1 x32 − i x2 x32
= E (6x1 − 2i x2 + 2x1 − 6i x2 ) + F (2x1 − 2i x2 )
= 8E (x1 − i x2 ) + 2F (x1 − i x2 )
= 0,

and so we find that F = −4E. Therefore, we can write

  2  
Y3−1 ∝ x̂1 + i x̂2 x̂1 − i x̂2 − 4 x̂1 − i x̂2 x̂32
 
= sin3 θ − 4 sin θ cos2 θ e−iφ
 
= 1 − 5 cos2 θ sin θ e−iφ .

For m = −2, we have ν+ = 0 and ν− = 2, which gives

 2
Y3−2 ∝ x̂1 − i x̂2 x̂3 = sin2 θ cos θ e−2iφ .

Finally, for m = −3, we must have ν+ = 0 and ν− = 3, so

 3
Y3−3 ∝ x̂1 − i x̂2 = sin3 θ e−3iφ .
 Chapter 2 Problem Set Solutions 11

Now, we must impose the normalization condition which reads

  π  2π
 2  2
d 2 Ylm (x̂) = sin θ dθ dφ Ylm (x̂) = 1.
0 0

For
= 3, we will need to make use of the following integrals
 π
32
(sin θ)7 dθ =
35
0 π
16
(sin θ)5 (cos θ)2 dθ =
105
0 π
 2 32
(sin θ)3 1 − 5 cos2 θ dθ =
21
0 π
  2 8
sin θ (cos θ)2 3 − 5 cos2 θ dθ = .
0 7
Finally, we arrive at the spherical harmonics for
= 3
 
35  3 35
Y3 = −
3
x̂1 + i x̂2 = − sin3 θ e3iφ
64π 64π
 
105  2 105
Y3 =
2
x̂1 + i x̂2 x̂3 = sin2 θ cos θ e2iφ
32π 32π

21     2  
Y31 = − 4 x̂1 + i x̂2 x̂32 − x̂1 + i x̂2 x̂1 − i x̂2
64π

21  
=− 5 cos2 θ − 1 sin θ eiφ
64π

7  3    
Y30 = 2x̂3 − 3 x̂1 + i x̂2 x̂1 − i x̂2 x̂3
16π

7  
= 5 cos2 θ − 3 cos θ
16π

21      2 
Y3−1 = 4 x̂1 − i x̂2 x̂32 − x̂1 + i x̂2 x̂1 − i x̂2
64π

21  
= 5 cos2 θ − 1 sin θ e−iφ
64π
 
105   105
Y3−2 = sin2 θ cos θ e−2iφ
2
x̂1 − i x̂2 x̂3 =
32π 32π
 
−3 35  3 35
Y3 = x̂1 − i x̂2 = sin3 θ e−3iφ .
64π 64π
We have fixed the arbitrary phases here to match the convention used in
Section 2.2, a convention which will be explained in Chapter 4.
12 Solutions to Lectures on Quantum Mechanics

2. Derive a formula for the rate of single photon emission from the 2 p to the 1s
state of hydrogen.
The rate of single photon emission is given in the dipole approximation by
Eq. (1.4.5) as
4e2 ωmn
3
Anm = |[x]nm |2 ,
3c3 
where the matrix element [x]nm is defined as

[x]nm = d 3 x ψn∗ (x)xψm (x).

In order to calculate the relevant matrix element for the hydrogen atom, we
need to work out the hydrogen wave functions using the method described in
Sections 2.1–2.3. We begin as in Section 2.3 with the wave function in the
form
u(r ) m
ψ(x) ∝ Y (θ, φ),
r

where
u = ρ
+1 e−ρ F(ρ),
and we have defined
ρ ≡ κr.
According to Eq. (2.3.17) F = 1 for the 2 p state, since
= 1 and n = 2. We
also have F = 1 for the 1s state, since
= 0 and n = 1. Using Eqs. (2.3.18)
and (2.3.19), we have
μZ e2 1
κn = 2
= ,
n na
where the Bohr radius a appearing in the second equality is given by
2
a= ,
μZ e2
and we have included the effect of the finite mass of the nucleus by using the
reduced mass in place of the electron mass as discussed in Section 2.4
m Nm e
μ= .
mN + me
For the case of hydrogen, we have Z = 1 and m N = m p . Putting this together,
the wave functions take the form
1
ψ1s ∝ e−r/a Y00 (θ, φ)
a
r
ψ2 p ∝ 2 e−r/2a Y1m (θ, φ).
4a
 Chapter 2 Problem Set Solutions 13

We need to normalize these wave functions so that


d 3 x |ψ(x)|2 = 1.

For the 1s state this gives


1 = d 3 x |ψ1s (x)|
   
2
 1 −r/a 2  2
= dr r  N1 e  d 2 Y00 (θ, φ)
a 
2  ∞
|N1 |
= dr r 2 e−2r/a
a2 0
 
|N1 |2 a 3
=
a2 4
a|N1 | 2
= ,
4
where in the third line, we used the fact that the spherical harmonics are
normalized. We find that the normalization constant N1 is given, up to an
arbitrary phase, by
2
N1 = .
a
Recalling from Section 2.2 that

1
Y00 = ,
4π
we find for the normalized 1s wave function

1 1 −r/a
ψ1s = e .
π a 3/2
Following the same procedure for the 2 p state gives

1 = d 3 x |ψ2 p (x)|
  
2 r −r/2a 2  2
= dr r  N2 2 e  d 2 Y1m (θ, φ)
4a

|N2 |2 ∞
= dr r 4 e−r/a
16a 4 0
|N2 |2  
= 4
24a 5
16a
3a|N2 |2
= ,
2
14 Solutions to Lectures on Quantum Mechanics

which gives up to an arbitrary phase factor


2
N2 = ,
3a
and so
1 r −r/2a m
ψ2 p = √ e Y1 (θ, φ).
24 a 5/2
Using the definition of the spherical harmonics from Section 2.2

3
Y1 =
0
cos θ
4π

±1 3
Y1 = ∓ sin θe±iφ ,
8π
we find for the normalized 2 p wave functions
1 r −r/2a
ψ2m=0
p =√ 5/2
e cos θ
32π a
1 r
ψ2m=±1 = ∓ √ 5/2 e−r/2a sin θ e±iφ .
p
8 πa
We are now in a position to calculate the matrix elements appearing in
the transition rate. Let us begin by calculating the matrix element for the
transition beginning in the 2 p state of hydrogen with m = 0
  
1 1
[x](1s)(2 p) = r 2 sin θ dθ dφ dr e−r/a
π a 3/2
 
× x̂1r sin θ cos φ + x̂2r sin θ sin φ + x̂3 cos θ
 
1 r −r/2a
× √ e cos θ
32π a 5/2
1
=√ dθ dφ dr r 4 e−3r/2a sin θ cos θ
32πa 4
 
× x̂1r sin θ cos φ + x̂2r sin θ sin φ + x̂3 cos θ

x̂3
=√ dθ dr r 4 e−3r/2a sin θ cos2 θ
8a 4
 ∞
x̂3
=√ dr r 4 e−3r/2a
18a 4 0
256
= √ a x̂3 .
243 2
 Chapter 2 Problem Set Solutions 15

We will now repeat the calculation beginning from the 2 p state of hydrogen
with m = ±1
  
1 1 −r/a
[x](1s)(2 p) = r sin θ dθ dφ dr
2
e
π a 3/2
 
× x̂1r sin θ cos φ + x̂2r sin θ sin φ + x̂3 cos θ
 
1 r
× ∓ √ 5/2 e−r/2a sin θ e±iφ
8 πa

1
= ∓ dθ dφ dr r 4 e−3r/2a sin2 θ
8πa 4
 
× x̂1r sin θ cos φ + x̂2r sin θ sin φ + x̂3 cos θ

1  
= ∓ 4 dθ dr r 4 e−3r/2a sin3 θ x̂1 ± i x̂2
8a
 ∞
1  
= ∓ 4 dr r 4 e−3r/2a x̂1 ± i x̂2
6a 0
128  
= ∓ a x̂1 ± i x̂2 .
243
Squaring these matrix elements, we see that for any m, we have
  15
[x](1s)(2 p) 2 = 2 a 2 .
310
To calculate the transition rate, we also need the frequency of the emitted
photon, which is given by
E (2 p) − E (1s)
ω(1s)(2 p) = .

According to Eq. (2.3.20), the energy levels of the hydrogen atom are
given by

2
En = − ,
2μa 2 n 2
and so we see that
3
ω(1s)(2 p) = .
8μa 2
Putting everything together, we arrive at the rate of single photon emission
from the 2 p to the 1s state of hydrogen
16 Solutions to Lectures on Quantum Mechanics

22 e2 33 3 215 2
A(1s)
(2 p) = × × a
3c3  29 μ3 a 6 310
28 e2 2
= 8 3 3 4
3c μa
28 e10 μ
= 8 6 3.
3c
3. Calculate the expectation values of the kinetic and potential energies in the
1s state of hydrogen.
For a given state ψ, the expectation value of the kinetic energy in that state
is given by
  
∗ 2 2
K ψ = d x ψ (x) −
3
∇ ψ(x),
2M
and the expectation value of the potential energy in that state is given by

V ψ = d 3 x ψ ∗ (x) (V (x)) ψ(x).

In the previous problem, we derived the normalized wave function for the 1s
state of hydrogen

1 1 −r/a
ψ1s = e .
π a 3/2
The expectation value of the kinetic energy in the 1s state of hydrogen is
given by
  
∗ 2 2
K (1s) = d x ψ(1s) (x) − ∇ ψ(1s) (x)
3
2μ
  
 2
−r/a 1 ∂ 2 ∂ −r/a
=− dθ dφ dr r sin θ e
2
r e
2πa 3 μ r 2 ∂r ∂r
  
−22 ∞ 1  r  −2r/a
=− dr − 2a − e
μa 3 0 a a2
 
22 a 2 a 2
= −
μa 4 2 4
 2
=
2μa 2
μe4
= 2.
2
In the second line we have used the Laplacian operator in the form of
Eq. (2.1.16), and we used the fact that
= 0 in the 1s state of hydrogen.
The expectation value of the potential energy in the 1s state of hydrogen is
given by
 Chapter 2 Problem Set Solutions 17
  2
∗ e
V (1s) = d x 3
ψ(1s) (x) − ψ(1s) (x)
r
  
e2 1 −2r/a
= − 3 dθ dφ dr r 2 sin θ e
πa r

−4e2 ∞
=− 3 dr r e−2r/a
a 0
e2
=−
a
μe4
=− 2

= −2K (1s) .
Notice that this is a special case of the virial theorem, which states that for
any potential of the form V (r ) = Cr n , we have the relation
2
V  = K .
n
4. Calculate the expectation values of the kinetic and potential energies in the
lowest energy state of the three-dimensional harmonic oscillator, using the
algebraic methods that were used in Section 2.5 to find the energy levels in
this system.
In order to calculate the expectation values of the kinetic and potential energy
for the harmonic oscillator using the algebraic methods of Section 2.5, we
will first need to write the position and momentum operators in terms of the
raising and lowering operators defined in Eq. (2.5.7)
 
1 ∂
ai ≡ √ −i − i Mωxi ,
2Mω ∂ xi
 
† 1 ∂
ai ≡ √ −i + i Mωxi .
2Mω ∂ xi
Solving these expressions for xi and ∂/∂ xi , we find
√
i   
xi = √ ai − ai† ,
2Mω
√
∂ i Mω  
= √ ai + ai† .
∂ xi 2
Recall the commutation relations for the raising and lowering operators given
in Eqs. (2.5.8) and (2.5.9)
 
ai , ai† = δi j ,
 
[ai , ai ] = ai† , ai† = 0.
18 Solutions to Lectures on Quantum Mechanics

We will work with the normalized wave function for the lowest energy state
ψ0 , so that

d 3 x ψ0∗ ψ0 = 1.

Recall also that the lowest energy state is annihilated by the lowering
operators
ai ψ0 = 0.
The expectation value of the kinetic energy in the lowest energy state is
given by
  
∗ 2 2
K 0 = d x ψ0 −
3
∇ ψ0
2M
  
ω   
= d 3 x ψ0∗ ai + ai† ai + ai† ψ0
4 i
  
ω 
= d 3 x ψ0∗ ai2 ψ0 + ai ai† ψ0 + ai† ai ψ0 + ai†2 ψ0 .
4 i
The first and third terms vanish because the lowering operator annihilates the
lowest energy state. The fourth term can be rewritten by using the definition
of the adjoint as
 
ω  ω   ∗
d 3 x ψ0∗ ai†2 ψ0 = d 3 x ai2 ψ0 ψ0 = 0,
4 i 4 i
which vanishes because the lowering operator annihilates the lowest energy
state. The only remaining term in the expression for the expectation value of
the kinetic energy can be rewritten in terms of a commutator

ω 
K 0 = d 3 x ψ0∗ ai ai† ψ0
4 i
   
ω 
= d 3 x ψ0∗ ai , ai† ψ0 + ai† ai ψ0
4 i

ω 
= d 3 x ψ0∗ ψ0
4 i
3
= ω,
4
where the second term in the second line vanishes because the lowering oper-
ator annihilates the lowest energy state, and we have used the commutation
relation of the raising and lowering operators in the first term. Proceding in
the same way for the expectation value of the potential energy gives
 Chapter 2 Problem Set Solutions 19
  
 Mω2
∗
V 0 = d x ψ0
3
xi ψ0
i
2
   
ω 
=− d 3 x ψ0∗ ai − ai† ai − ai† ψ0
4 i
  
ω 
=− d 3 x ψ0∗ ai2 ψ0 − ai ai† ψ0 − ai† ai ψ0 + ai†2 ψ0
4 i
    
ω 
=− d 3 x ψ0∗ − ai , ai† ψ0 − ai† ai ψ0
4 i

ω 
= d 3 x ψ0∗ ψ0
4 i
3
= ω
4
= K 0 .
Notice that like the previous problem this gives a specific case of the virial
theorem.
5. Derive the formula for the energy levels of the three-dimensional harmonic
oscillator by using the power-series method (with suitable modifications) that
was used in Section 2.3 for the hydrogen atom.
For the harmonic oscillator, the potential takes the form
Mω2r 2
V (r ) = .
2
We begin with a wave function of the form given by Eqs. (2.1.21) and
(2.1.28),
u(r ) m
ψ(x) = Y (θ, φ),
r

where u(r ) is a solution of Eq. (2.1.29)


2 d 2 u(r ) Mω2r 2
(
+ 1)2
− + + u(r ) = Eu(r ).
2M dr 2 2 2Mr 2
Following closely the steps of Section 2.3, this can be rewritten in the form
2 2 2 
d 2 u(r ) M ωr
(
+ 1)
− + + u(r ) = −κ 2 u(r ),
dr 2 2 r2
where κ is defined by
2 κ 2
E =− .
2M
20 Solutions to Lectures on Quantum Mechanics

After defining
ρ ≡ κr,
and dividing by κ 2 , the equation for u becomes

d 2u
(
+ 1)
− 2 + ξρ + 2
u = −u,
dρ ρ2
where we have also defined
M 2 ω2
ξ≡ .
2 κ 4
In the limit ρ → 0, the differential equation for u becomes
d 2u
≈
(
+ 1)ρ −2 u,
dρ 2
which implies that in this limit, we must have u ∼ ρ
+1 . In the limit ρ → ∞,
the equation for u becomes
d 2u
≈ ξρ 2 u,
dρ 2
√
which implies that in this limit, we must have u ∼ e± ξ ρ /2 . The requirement
2

that the wave function be normalizable eliminates the exponentially growing

solution, and so we keep only the exponentially decaying solution. We can
then write u in the form
 

+1 1√
u = ρ exp − 2
xiρ F(ρ).
2
In order to plug this into the Schrödinger equation, we must find the second
derivative of u. The first derviative is
    
du
+1 1 2
+1  dF
= ρ exp − ξρ − ξρ F + ,
dρ 2 ρ dρ
and the second derivative is
 
d u 2

+1 1 2
= ρ exp − ξρ
dρ 2 2
    

(
+ 1) 
+1  dF d2 F
× − ξ (2
+ 3) + ξρ F + 2
2
− ξρ + .
ρ2 ρ dρ dρ 2
Plugging this into the Schrödinger equation and dividing out common factors
gives a differential equation for F
 
d2 F 
+ 1 d F  
− 2 ξ ρ − − ξ (2
+ 3) − 1 F = 0.
dρ 2 ρ dρ
 Chapter 2 Problem Set Solutions 21

We can simplify this equation further if we make the identification

√
ρ = σ,

so that we have
d dσ d d √ d
= = 2ρ =2 σ ,
dρ dρ dσ dσ dσ
d2 d d
= 4σ 2 + 2 .
dρ 2 dσ dσ
The differential equation for F now becomes

d2 F dF   d F  
4σ + 2 − 4 ξ σ − (
+ 1) − ξ (2
+ 3) − 1 F = 0,
dσ 2 dσ dσ
which can be simplified slightly to read
  √
d2 F 
+ 32 d F ξ (2
+ 3) − 1
− ξ− − F = 0.
dσ 2 σ dσ 4σ

We now look for a series solution of the form

∞

F= as σ s ,
s=0

where we require that a0 = 0 in order to maintain the behavior of the wave

function in the limit ρ → 0 that we found above. Using this ansatz in the
equation for F gives
∞  
 s−1 3
as s(s − 1)σ s−2
− s ξσ +s
+ σ s−2
s=0
2

1  
− ξ (2
+ 3) − 1 σ s−1
= 0.
4
In the first and third terms, we will make the replacement s → s + 1. This
does not affect the limits of the sum, because initially, the s = 0 contribu-
tion of the first and third term vanishes. After this replacement, the equation
becomes
∞  
 3
as+1 s(s + 1) − as s ξ + as+1 (s + 1)
+
s=0
2

1  
− ξ (2
+ 3) − 1 σ s−1 = 0.
4
22 Solutions to Lectures on Quantum Mechanics

This expression must vanish for values of σ , and therefore the sum must
vanish term by term. This results in a recurrence relation for the coefficients
of the series solution
   √ 
3  ξ (2
+ 3) − 1
as+1 (s + 1)
+ + s = as s ξ + .
2 4
√
For large values of s, this expression gives as+1 /as → ξ /s and so
√ s
a ξ
as → ,
(s + b)!
which gives for F at large s
√ s  −b √  −b √ 2
 ξσ
F →a =a ξσ e ξσ = a ξ ρ2 e ξρ ,
s
(s + b)!

which is the exponentially growing solution we threw out above in order to

require normalizability of the wave function. In order to avoid this exponen-
tially growing solution, we require that the power series must terminate at
some finite s = smax . We therefore require that
√
 ξ (2
+ 3) − 1
smax ξ + = 0,
4
which we can rearrange to read
1
√ = 2
+ 3 + 4smax .
ξ
√
Since
and smax are both non-negative integers, we find that 1/ ξ must be
an odd integer greater than or equal to 3
1
√ = 2n + 3 for n = 0, 1, 2, . . . .
ξ
Returning to the definition of ξ we find
M 2 ω2 1
ξ= = ,
κ
2 4 (2n + 3)2
and so
M 2 ω2 4M 2 2
κn4 = (2n + 3)2 = E .
2 4 n
Finally, we arrive at the expression for the energy levels
 
3
En = n + ω.
2
 Chapter 2 Problem Set Solutions 23

This is exactly what we expect based on the fact that a three-dimensional

harmonic oscillator behaves just like three one-dimensional harmonic oscil-
lators. As a side note, we have also arrived at an expression for the wave
functions of the three-dimensional harmonic oscillator
   1  
Mω 2
+ 2 Mω 2

ψn ∝ r exp − r L 1 (n−
) r Y
m (θ, φ),
2 2 
where the L (α)
k (y) are called generalized Laguerre polynomials, and are
generated by the recurrence relation derived above.
6. Find the difference in the energies of the Lyman α transitions in hydrogen
and deuterium.
The Lyman α transition refers to the transition from n = 2 to n = 1. The only
relevant difference between hydrogen and deuterium for the Lyman α transi-
tion is the reduced mass of the system. For hydrogen, the reduced mass is
m pme
μH = ,
m p + me
while for deuterium the reduced mass is slightly higher
2m p m e
μD = .
2m p + m e
According to Eq. (2.3.20), the energy levels of hydrogenic atoms are given by
2 μZ 2 e4
En = − = − .
2μa 2 n 2 22 n 2
The energy of a Lyman α photon is given by
3μZ 2 e4
Eα = E2 − E1 = .
82
The isotopic shift is given by
 
3Z 2 e4 3Z 2 e4 m p m 2e
E α = (μ D − μ H ) =   
82 82 2m p + m e m p + m e
≈ 2.77644085 × 10−3 eV.
This is to be compared with the energy of a Lyman α photon emitted from a
hydrogen atom
3μ H Z 2 e4
E α(H ) = ≈ 5.21153244 eV,
82
and so the isotopic shift represents about a 0.05% shift in the energy of the
emitted photon. This shift in energy is large enough to be detected with mod-
ern spectroscopic equipment, which allows hydrogen and deuterium to be
distinguished by their emission spectra.
24 Solutions to Lectures on Quantum Mechanics

7. Calculate the wave function (aside from normalization) of the 3s state of the
hydrogen atom.
The 3s state of hydrogen has n = 3,
= 0, and m = 0. In order to find the
wave function for the 3s state of hydrogen, we begin with the form given by
Eq. (2.1.21)
ψ(x) = R(r )Y
m (θ, φ),
where R(r ) can be put in the form of Eq. (2.1.28)
u(r )
R(r ) = .
r
As discussed in Section 2.2, Y00 is just a constant, and since we are not
concerned with the normalization here, we can simply include it in the unde-
termined normalization constant for the purposes of this problem. For the
hydrogen atom, u takes a form given by Eq. (2.3.7)
u = ρ
+1 exp(−ρ)F(ρ),
where we have introduced the notation
ρ ≡ κr,
where κ is defined by Eqs. (2.3.3) and (2.3.18) to be
√
−2μE 1
κ= = ,
 na
and a = 2 /μZ e2 is the Bohr radius. According to Eq. (2.3.9), F(ρ) takes
the form of a power series
∞

F(ρ) = as ρ s ,
s

where the coefficients as satisfy a recurrence relation, which is given by

Eq. (2.1.12)
−2n + 2s + 2
+ 2
as+1 = as .
(s + 2
+ 2)(s + 1)
For the 3s state of hydrogen, this gives
a1 = −2a0
1 2
a2 = − a1 = a0
3 3
ak = 0 for k ≥ 3.
Therefore, F(ρ) takes the form
 
2 2
F = a0 1 − 2ρ + ρ ,
3
 Chapter 2 Problem Set Solutions 25

and so u takes the form

     r 
2 2 r 2r 2 2r 3
u = a0 ρ exp(−ρ) 1 − 2ρ + ρ = a0 − 2+ exp − .
3 3a 9a 81a 3 3a
Putting this together, we find the wave function for the 3s state of hydrogen
   r 
1 2r 2r 2
ψ(x) = N − 2+ exp − ,
3a 9a 81a 3 3a
where N is the normalization constant.
26 Solutions to Lectures on Quantum Mechanics

Chapter 3 Problem Set Solutions

1. Consider a system with a pair of observable quantities A and B, whose

commutation relations with the Hamiltonian take the form [H, A] = iw B,
[H, B] = −iw A, where w is some real constant. Suppose that the expec-
tation values of A and B are known at time t = 0. Give formulas for the
expectation values of A and B as a function of time.
According to Eq. (3.6.7), for any Heisenberg picture operator O H (t), the time
derivative is given by
i
Ȯ H (t) = [H, O H (t)].

In this case, this gives
i w
Ȧ H (t) = [H, A H (t)] = − B H (t),
 
i w
Ḃ H (t) = [H, B H (t)] = A H (t).
 
Taking another time derivative, we find

w w2
Ä H (t) = − Ḃ H (t) = − 2 A H (t),
 
w w 2
B̈ H (t) = Ȧ H (t) = − 2 B H (t).
 
These differential equations have a well-known solution of the form
w  w 
A H (t) = C1 sin t + C2 cos t ,
 
w  
w
B H (t) = C3 sin t + C4 cos t .
 
Now we must impose the condition that Ȧ H (t) = − w B H (t), which for this
solution takes the form
w w  w w  w w  w w 
C1 cos t − C2 sin t = − C3 sin t − C4 cos t .
       
In order for this to be satisfied for all times, we require

C4 = −C1 , C3 = C2 .

Evaluating these solutions at t = 0, we find

C2 = A H (0), C1 = −B H (0),
 Chapter 3 Problem Set Solutions 27

and so we can write the operators as

w  w 
A H (t) = −B H (0) sin t + A H (0) cos t ,
 w   w 
B H (t) = A H (0) sin t + B H (0) cos t .
 
Now if we define

A H (0) ≡ a, B H (0) ≡ b,

then we find for the expectaion values

w  w 
A H (t) = −b sin t + a cos t ,
 w   w 
B H (t) = a sin t + b cos t .
 
2. Consider a normalized initial state  at t = 0 with a spread E in energy,
defined by

 2 
E ≡ H − H  .


Calculate the probability |((δt), )|2 that after a very short time δt the
system is still in the state . Express the result in terms of E,  and δt, to
second order in δt.
We can use Eq. (3.6.3) to write

(δt) = exp (−i H δt/) ,

where we have defined  ≡ (0), and we will assume that  is normalized

so that (, ) = 1. Now let us expand (δt) to second order in δt
i 1  
(δt) =  − H δt − 2 H 2 δt 2  + O δt 3 .
 2
We can then calculate the probability that the system remains in the initial
state
i 1  
| ( (δt) , ) |2 = | (, ) + δt (H , ) − 2 δt 2 H 2 ,  |2
 2
i i
= |(, )| + δt (H , ) (, ) − δt (H , ) (, )
2
 
1 2 1    
+ 2 δt | (H , ) |2 − 2 δt 2 H 2 ,  (, ) + O δt 3
 
δt 2 δt 2  
= 1 + 2 H 2 − 2 H 2  + O δt 3 .
 
28 Solutions to Lectures on Quantum Mechanics

Now we note that

 !
 E ≡ (H − H  )2

 ! !
= H 2  − 2 H H  + H 2
 
"
= H 2  − H 2 ,

and so we can rewrite the inner product above as

( E)2 δt 2  
| ((δt), ) |2 = 1 − + O δt 3 .
 2

3. Suppose that the Hamiltonian is a linear operator with

H  = g, H  = g ∗ , H ϒn = 0,

where g is an arbitrary constant,  and  are a pair of normalized indepen-

dent (but not necessarily orthogonal) state vectors, and ϒn runs over all state
vectors orthogonal to both  and . What are the conditions that  and 
must satisfy in order for this Hamiltonian to be Hermitian? With these con-
ditions satisfied, find the states with definite energy, and the corresponding
energy values.
In order for H to be Hermitian, we must have

(, H ) = (H , ) .

Using the relation H  = g, we can rewrite the left-hand side as

(, H ) = (, g) = g (, ) ,

and the right-hand side as

 ∗
(H , ) = (g, ) = g ∗ (, ) = g ∗ (, )∗ = g (, ) .

We find that the condition for H to be Hermitian is that the product g (, )
must be real. 
We will now look for states χi = ai  + bi  + n cin ϒn such that H χi =
E i χi . Applying the Hamiltonian to such a state χi , we find
 

H χi = H ai  + bi  + cin ϒn = gai  + g ∗ bi  + 0
n
 

= E i ai  + bi  + cin ϒn = E i χi .
n
 Chapter 3 Problem Set Solutions 29

First, we notice that either cin = 0 or E i = ai = bi = 0. We will set cin = 0

and look for states with non-zero energy. We find that for E i = 0, we must
have
E i ai = g ∗ bi , E i bi = gai ,
which we can combine to find that
E i2 = |g|2 .
Let us distinguish two states which we will call χ+ and χ− such that
E + = |g|, E − = −|g|.
We then find that the states are given by
 
g
χ± = N ±  ±  ,
|g|
where N± is a normalization constant, which we can determine by the
condition that

g
1 = (χ± , χ± ) = |N± | (, ) + (, ) ± 2
2
(, ) .
|g|
So we find, up to an arbitrary overall phase the normalized states of definite
energy are given by
−1/2  
g g
χ± = 2 ± 2 (, ) ±  ,
|g| |g|
and these states satisfy the condition
H χ± = ±|g|χ± ,
and of course there are the states of zero energy
H ϒn = 0.
4. Suppose that a linear operator A, though not Hermitian, satisfies the condi-
tion that it commutes with its adjoint. What can be said about the relation
between the eigenvalues of A and of A† ? What can be said about the scalar
product of two eigenstates of A with unequal eigenvalues?  
We are told that the operator A commutes with its adjoint A, A† = 0, and
so there exists a state , which we will take to be normalized (, ) = 1,
which is an eigenstate of both A and A†
A = a, A†  = b.
Now let us calculate the expectation value of the product A A† in the state 
     
, A A†  = ba = A† , A†  = |b|2 = , A† A
= ab = (A, A) = |a|2 .
30 Solutions to Lectures on Quantum Mechanics

We can see from this that we must have a = b∗ .

If we now define a second state   such that
A  = a    , A†   = b   ,
we can use the definition of the adjoint to write
       
0 =   , A − A†   ,  = a   ,  − b∗   ,  .
 
We see that this equation implies that either a = b∗ or   ,  = 0. Since
b∗ = a  because 
  A commutes with its adjoint, we find that for a = a , we
must have  ,  = 0.
5. Suppose the state vectors  and   are eigenvectors of a unitary operator
with eigenvalues λ and λ , respectively. What relation must λ and λ satisfy if
 is not orthogonal to   ?
We have two states  and   such that
U  = λ, U   = λ   ,
with U a unitary operator U †U = 1. Let us first consider the inner product
of  with itself
 
(, ) = , U †U  = (U , U ) = (λ, λ) = λ∗ λ (, ) ,
which implies that λ∗ λ = 1. The same procdedure applied to   requires that
λ∗ λ = 1. Now let us consider the inner product of  with  
         
,   = , U †U   = U , U   = λ, λ   = λ∗ λ∗ ,   ,
 
which implies that λ∗ λ = 1 or ,   = 0. Putting this together we find
that for  and   not orthogonal, we must have
λ = λ∗ = eiφ
for some real constant φ.
6. Show that the product of the uncertainties in position and momentum takes
its minimum value /2 for a Gaussian wave packet of free particle wave
functions.
We wish to show that

 X  P =
2
for a Gaussian wave packet, where the spread of any observable correspond-
ing to an operator A in a state  is defined as in Eq. (3.3.23)
 !
 A ≡ (A − A )2 .


A Gaussian wave packet can be written in a coordinate representation as

 Chapter 3 Problem Set Solutions 31
 
exp (−i p0 x/) (x − x0 )2
ψ(x) =  1/4 exp − .
2πa 2 4a 2

Let us begin by calculating X ψ for the Gaussian wave packet

 ∞
X ψ = ψ ∗ (x)xψ(x) d x
−∞
 ∞  
1 (x − x0 )2
= 1/2 x exp − dx
2πa 2 −∞ 2a 2
= x0 .
We follow a similar procedure to calculate X 2 ψ
 ∞
X 2 ψ = ψ ∗ (x)x 2 ψ(x) d x
−∞
 ∞  
1 (x − x 0 )2
= 1/2 x 2 exp − dx
2πa 2 −∞ 2a 2
= x02 + a 2 .
This gives for the spread in position
 "
 2 !
ψ X = X − X ψ = X 2 ψ − X 2ψ
ψ
"
= (x02 + a 2 ) − (x0 )2 = a.
Next, we will calculate Pψ for the Gaussian wave packet
 ∞
d
Pψ = ψ ∗ (x)i ψ(x) d x
−∞ d x
 ∞   
1 x − x0 (x − x0 )2
= 1/2 p0 − i exp − dx
2πa 2 −∞ 2a 2 2a 2
= p0 .
For P 2 ψ we find
 ∞  2 
∗ 2 d
P  =
2
ψ (x) − ψ(x) d x
−∞ dx2
 ∞     
1 x − x0 2 2 (x − x0 )2
= 1/2 p0 − i + 2 exp − dx
2πa 2 −∞ 2a 2 2a 2a 2
2
= p02 + .
4a 2
32 Solutions to Lectures on Quantum Mechanics

This gives for the spread in momentum

 "
 2 !
ψ P = P − Pψ = P 2 ψ − P2ψ
ψ
 
2 
= p02 + 2 − ( p0 )2 = .
4a 2a
Now, calculating the product of the spread in position and momentum, we
find

ψ X ψ P = ,
2
which saturates the bound from the Heisenberg uncertainty principle.
 Chapter 4 Problem Set Solutions 33

Chapter 4 Problem Set Solutions

1. Suppose that an electron is in a state of orbital angular momentum

= 2.
Show how to construct the state vectors with total angular momentum
j = 5/2 and corresponding 3-components m = 5/2 and m = 3/2 as lin-
ear combinations of state vectors with definite values of S3 and L 3 . Then
find the state vector with j = 3/2 and m = 3/2. (All state vectors here
should be properly normalized.) Summarize your results by giving values
for the Clebsch–Gordan coefficients C 1 2 ( jm; m s m
) in the cases ( j, m) =
2
(5/2, 5/2), (5/2, 3/2) and (3/2, 3/2).
We would like to construct eigenstates of the operators J2 , J3 , L2 , and S2
which are linear combinations of eigenstates of the operators L2 , L 3 , S2 , and
S3 , just as in Eq. (4.3.18)


ms j = C
s ( j m ; m
m s )ψ
ms
m s .
m
,m s

The state with the maximum values of j and m is given simply by the state
with the maximum values of m
and m s , as shown in Eq. (4.3.20), and so in
the case
= 2, s = 1/2, this reads
5 1
2
22 1 5 = ψ2 12 .
2 2 2

To find the state with m = 3/2, we can apply the lowering operator with
proper normalization as given in Eq. (4.2.16)
    2 −1/2
3 1 5 5 5 5 5
2 1 5 =
2
+1 − + (J1 − i J2 ) 22 1 5
2 2  2 2 2 2 2 2

1 21
= √ (L 1 − i L 2 + S1 − i S2 ) ψ2 12
 5 2

1  1/2 1 12
= √  2 (2 + 1) − (2)2 + 2 ψ2 1
 5 2

   ⎞
  2 1/2
1 1 1 1 2 − 12
+ +1 − + ψ2 1 ⎠
2 2 2 2 2

 
4 1 12 1 2 − 12
= ψ2 1 + ψ 1 .
5 2 5 22
2 − 12 2 − 12
The properly normalized state vector constructed from ψ2 1 and ψ2 1
2 2
3 3
which is orthogonal to 2 12
5 must be 2 1 3 , as in Eq. (4.3.23), and so we
2
2 2 2 2
find
34 Solutions to Lectures on Quantum Mechanics
 
3 1 1 12 4 2 − 12
2 1
2
3 =− ψ 1 + ψ 1 ,
2 2 5 22 5 22
up to an arbitrary overall phase factor. We can read off the Clebsch–Gordan
coefficients by comparing with the first equation above, and we find for the
non-vanishing coefficients
 
55 1
C2 1 ;2 =1
2 22 2
  
53 1 1
C2 1 ;2 − =
2 22 2 5
  
55 1 4
C2 1 ;1 =
2 22 2 5
  
53 1 4
C2 1 ;2 − =
2 22 2 5
  
55 1 1
C2 1 ;1 =−
2 22 2 5
2. Suppose that A and B are vector operators, in the sense that
 
[Ji , A j ] = i i jk Ak , [Ji , B j ] = i i jk Bk .
k k

Show that the cross-product A × B is a vector in the same sense.

Given that
     
Ji , A j = i i jk Ak , Ji , B j = i i jk Bk ,
k k

we wish to show that

  
Ji , (A × B) j = i i jk (A × B)k .
k

First, we note that the cross-product can be written as


(A × B)i = i
m A
Bm ,

m

and so we can write the right-hand side of the desired result as

 
i i jk (A × B)k = i jk k
m A
Bm
k k
m
 
= i δi
δ jm − δim δ j
A
Bm

m
 
= i Ai B j − A j Bi ,
 Chapter 4 Problem Set Solutions 35

where we have used the well-known relation for the Levi–Civita symbols

ki j k
m = δi
δ jm − δim δ j
.
k

Now we can calculate the commutator of Ji with the cross-product (A × B) j

  
Ji , (A × B) j =  j
m ([Ji , A
] Bm + A
[Ji , Bm ])

m
 
  
=  j
m i i
k Ak Bm + A
i imk Bk

m k k
     
   
= i  j
m i
k Ak Bm +  j
m imk A
Bk
km
k
m
 
   
= i δmk δ ji − δmi δ jk Ak Bm + δ jk δi
− δi j δk
A
Bk
km k

 
 
= i δi j Ak Bk − A j Bi + Ai B j − δi j Ak Bk
k k
 
= i Ai B j − A j Bi

= i i jk (A × B)k .
k

We find that if A and B are vector operators, then (A × B) is also a vector

operator.
3. What is the minimum value of the total angular momentum J2 that a state
must have in order to have a non-zero expectation value for an operator Omj
of spin j.
Let us define a state
 
mj jm ≡ Omj  mj  ,

which as shown in Eq. (4.4.10) acts as a state vector describing two parti-
cles with spins j and j  and spin 3-components m and m  . We can express
this state vector as a sum over states with total angular momentum J and
3-component M as in Eq. (4.4.11)

 
mj jm = C j j  ( j  m  ; m m  )ωmj j  j  .
j  ,m 

As discussed in Section 4.3, the range of j  is

j  = | j − j  |, | j − j  | + 1, . . . , j + j  .
36 Solutions to Lectures on Quantum Mechanics
  

We are interested in the inner product  mj  , ωmj j  j  . Note that because J2 is
an Hermitian, this inner product vanishes unless j  = j 
  
   
   

 mj  , J2 ωmj j  j  = 2 j  ( j  +1)  mj  , ωmj j  j  = 2 j  ( j  +1)  mj  , ωmj j  j  .

A non-vanishing inner product therefore requires that j  is in the range

spanned by j  , and so the minimum value that j  can take is given by
  
jmin = | j − jmin | = j − jmin ,

where we have taken j − jmin to be positive because for j < j  there is always
a smaller value of j  for which the inner product does not vanish. Solving this

equation for jmin we find

 1
jmin = j.
2
We therefore conclude that
  
 1
 mj  , Omj  mj  = 0 for j ≥ j,
2

where j is restricted to be an integer. For half-integer j, one of  mj  and

Omj  mj  will be bosonic while the other is fermionic, and so will always have
a vanishing inner product.
4. The Hamiltonian for a free particle of mass M and spin S placed in a
magnetic field B in the 3-direction is
p2
H= − g|B|S3 ,
2M
where g is a constant (proportional to the particle’s magnetic moment). Give
the equations that govern the time dependence of the expectation values of
all three components of S.
We can find the time derivative of S(t) in the Heisenberg picture by taking
the commutator with the Hamiltonian

i i p2 iμB
Ṡi = [H, Si ] = , Si − [S3 , Si ]
  2M 
 
iμB 
=− i 3ik Sk
 k
= μB (δi1 S2 − δi2 S1 ) .
Therefore, the individual components of the spin have time derivatives
Ṡ1 = μB S2 , Ṡ2 = −μB S1 , Ṡ3 = 0.
 Chapter 4 Problem Set Solutions 37

We can take second time derivatives of S1 and S2 to find

S̈1 = μB Ṡ2 = −μ2 B 2 S1 ,
S̈2 = μB Ṡ1 = −μ2 B 2 S2 .
These differential equations have well-known solutions of the form
S1 (t) = A1 cos(μBt) + A2 sin(μBt),
S2 (t) = A3 cos(μBt) + A4 sin(μBt).
Imposing the condition that Ṡ1 (t) = μB S2 (t), we find that
−A1 μB sin(μBt) + A2 μB cos(μBt) = A3 μB cos(μBt) + A4 μB sin(μBt),
which must be true for all times, and so this requires
−A4 = A1 , A2 = A3 .
Evaluating these solutions at t = 0, we find
A1 = S1 (0), A3 = S2 (0),
and so we find
S1 (t) = S1 (0) cos(μBt) + S2 (0) sin(μBt),
S2 (t) = S2 (0) cos(μBt) − S1 (0) sin(μBt),
S3 (t) = S3 (0).
5. A particle of spin 3/2 decays into a nucleon and pion. Show how the ang-
ular distribution in the final state (with spins not measured) can be used to
determine the parity of the decaying particle.
First, we note that nucleons have spin 1/2 and pions have spin 0. In order for
the decay of a spin−3/2 particle (which we will call X ) into a nucleon and
pion to conserve angular momentum, the final state must have orbital angular
momentum
= 1 or
= 2. Assuming that the decay conserves parity, we
can write
η X = (−1)
η N ηπ .
According to Section 4.7, nucleons have intrinsic parity η N = +1 while
pions have intrinsic parity ηπ = −1 and so we find
η X = (−1)
+1 .
Now, if the angular distribution of the final state can be used to determine the
orbital angular momentum
, we can use this expression to find the intrinsic
parity of the spin−3/2 particle. To be specific, we find

+1 if
= 1
ηX = .
−1 if
= 2
38 Solutions to Lectures on Quantum Mechanics

6. A particle X of isospin 1 and charge zero decays into a K and a K . What is

− 0
the ratio of the rates of the processes X 0 → K + + K and X 0 → K 0 + K ?
First, we note that the isospin of all of the kaons is T = 1/2. The T3 quantum
number for each of the kaons is given below

K+ T3 = +1/2
−
K̄ T3 = −1/2
K0 T3 = −1/2
K̄ 0
T3 = +1/2.

Because the particle X decays into a K and a K̄ , the conservation of T3 in

strong interactions requires that the particle X has T3 = 0. If we label the T3
quantum numbers of the final state particles as m and m  , the amplitudes of
these decays will be proportional to Clebsch–Gordan coefficients of the form
 
M(0, m, m  ) = M0 C 1 1 1 0; m m  .
2 2

In both cases the relevant Clebsch–Gordan coefficient is

  
1 1 1
C 1 1 1 0; + − = .
2 2 2 2 2
We therefore find for the ratio of the decay rates
    
(X → K + + K̄ − )  M 0 + 1 − 1 2 |M0 |2 12
=  2 2
 =   = 1.
(X → K 0 + K̄ 0 )  M 0 + 1 − 1 2 |M0 |2 12
2 2

7. Imagine that the electron has spin 3/2 instead of 1/2, but assume that the
one-particle states with definite values of n and
in atoms are filled, as the
atomic number increases, in the same order as in the real world. What ele-
ments with atomic numbers in the range from 1 to 21 would have chemical
properties similar to those of noble gases, alkali metals, halogens, and alkali
earths in the real world?
The values taken by the 3-component of spin range from −s to s in integer
steps giving 2s + 1 values. There are four values taken by m s for a spin−3/2
particle
3 1 1 3
ms = − , − , , .
2 2 2 2
As a result, for multi-electron atoms constructed from spin−3/2 electrons
there will be 4(2l + 1) distinct states for each value of the prinicipal quantum
number n and the orbital angular momentum l. Then if atoms are filled in the
order described by Eq. (4.5.11), we can count states as
 Chapter 4 Problem Set Solutions 39

1s 4 states
2s, 2 p 4 + 12 = 16 states
3s, 3 p 4 + 12 = 16 states
.. ..
. .
Noble gases are those which have a full outer shell of electrons, and in this
case are atoms with atomic number Z = 4, 20. Alkali metals are those
which have only one electron in the outermost shell, and here are atoms with
Z = 1, 5, 21. Halogens are elements having an outermost shell which is
one electron short of being full, and in this case are atoms with Z = 3, 19.
Finally alkali earths have two electrons in their outer shell, and here are the
atoms with Z = 2, 6.
8. What is the commutator of the angular momentum operator J with the
generator K of Galilean transformations?
Since the generator of Galilean transformations K is a vector operator, we
expect that
  
Ji , K j = i i jk K k ,
k
according to Eq. (4.1.13). To be more explicit, we may write the generator of
Galilean transformations as  
1 
K= Pt − m n Xn ,
 n
which satisfies the commutation relations given by Eqs. (3.6.18) and (3.6.20).
Now, using the commutation relations of the angular momentum operator
with Xn and P given in Eq. (4.1.22), we may compute the commutator
directly
    
  1   
Ji , K j = Ji , Pj t − mn Xnj =i i jk Pk t − m n X nk
 n k n

= i i jk K k .
k

9. Consider an electron in a state of zero orbital angular momentum in an atom

whose nucleus has spin (that is, internal angular momentum) 3/2. Express
the states of the atom with total angular momentum z-component m = 1
(of electron plus nucleus) and each possible definite value of the total angu-
lar momentum as linear combinations of states with definite values of the
z-components of the nuclear and electron spins.
We would like to construct eigenstates of the operators J2 , J3 , I2 , and S2
which are linear combinations of eigenstates of the operators I2 , I3 , S2 , and
S3 , just as in Eq. (4.3.18)
40 Solutions to Lectures on Quantum Mechanics

 Ims j = C I s ( j m ; m I m s )ψ ImsI m s ,
m I ,m s

where we are labeling the nuclear spin quantum number as I and its three-
component as m I . The state with j = 2 and m = 1 is given by
   
3 1 3
−1 11 1 1
 13 1 2 = C 3 1 2 1 ; − ψ 32 1 2 + C 3 1 2 1 ; ψ 32 12 ,
2 2 2 2 2 2 2 2 2 2 22 2 2

and the state with j = 1 and m = 1 is given by

   
11 1 1 11 −1 1
3 1 1 = C3 1 1 1 ;
1
ψ3 1 + C3 1 1 1 ; −
2 2
ψ 3 12 2 .
2 2 2 2 22 2 2 2 2 22 2 2

Now all that remains is to find the relevant Clebsch–Gordan coefficients. The
first two coefficients can be read off from Eqs. (4.3.24) and (4.3.25), and the
others can be obtained by requiring the orthogonality of the two states. We
therefore find

1 32 − 12 3 12 12
3 1 2 = ψ3 1 +
1
ψ3 1 ,
2 22 4 22

2 2

3 32 − 12 1 12 12
 13 1 1 = ψ3 1 − ψ3 1 .
2 2 4 22 2 22
 Chapter 5 Problem Set Solutions 41

Chapter 5 Problem Set Solutions

1. Suppose that the interaction of the electron with the proton in the hydrogen
atom produces a change in the potential energy of the electron of the form
V (r ) = V0 exp(−r/R),
where R is much smaller than the Bohr radius a. Calculate the shift in the
energies of the 2s and 2 p states of hydrogen, to first order in V0 .
The first-order shift in energy due to a perturbation is given by Eq. (5.1.6) as
δ1 E a = (a , δ H a ) .
In this case the perturbation is given by
V (r ) = V0 exp(−r/R),
where R is much smaller than the Bohr radius a. In order to calculate this
energy shift for the 2s and 2 p states of hydrogen, we need to work out the
form of the unperturbed wave functions. Recall from Section 2.3 that these
wave functions are given by
ψ2s = R2 0 (r )Y00 ,
ψ2 p = R2 1 (r )Y0m ,
where the radial part of the wave function is
Rn
∝ r
exp(−r/na)Fn
(r/na).
According to Eq. (2.3.17) in this case we have
r
F2 0 = 1 − ,
2a
F2 1 = 1,
and so the radial parts of the wave functions take the form
 r 
R2 0 = A exp(−r/2a) 1 − ,
2a
R2 1 = Br exp(−r/2a).
We can normalize the wave functions to determine A and B. For A we find
  ∞ 
∗
d x ψ2s ψ2s =
3
dr r (R2 0 )
2 2
d 2 Y00∗ Y00
0
 ∞  r 2
= |A|2 dr r 2 1 − exp(−r/a)
0 2a
= 2a 3 |A|2
 −1/2
A = 2a 3 ,
42 Solutions to Lectures on Quantum Mechanics

and for B we find

  ∞ 
∗
d x ψ2 p ψ2 p =
3
dr r (R2 1 )
2 2
d 2 Y1m∗ Y1m
0
 ∞
= |B|2 dr r 4 exp(−r/a)
0
= 24a 5 |B|2
 −1/2
B = 24a 5 .
We find for the 2s and 2 p hydrogen wave functions
 r
ψ2s = (2a)−3/2 2 − exp(−r/2a)Y00 ,
a
r
ψ2 p = (2a)−3/2 √ exp(−r/2a)Y0m .
3a
We can now calculate the first-order shift in energy of the 2s state as

∗
δ1 E 2s = d x ψ2s V0 exp(−r/R)ψ2s
 ∞  
−3 r 2  r r
= V0 dr r (2a)
2
2− exp − − d 2 Y00∗ Y00
0 a a R
R 3 (a 2 − a R + R 2 )
= V0 ,
(a + R)5
and for the 2 p state we find

δ1 E 2 p = d x ψ2∗p V0 exp(−r/R)ψ2 p
 ∞  r 
−3 r
2
r
= V0 dr r (2a)
2
exp − − d 2 Y1m∗ Y1m
0 3a 2 a R
R5
= V0 5 .
a
In the limit R  a, we find for the energy shifts
R3
δ1 E 2s → V0,
a3
R5
δ1 E 2 p → V0 5 .
a
2. It is sometimes assumed that the electrostatic potential felt by an electron in
a multi-electron atom can be approximated by a shielded Coulomb potential,
of the form
Z e2
V (r ) = − exp(−r/R),
r
 Chapter 5 Problem Set Solutions 43

where R is the estimated radius of the atom. Use the variational method to
give an approximate formula for the energy of an electron in the state of
lowest energy in this potential, taking as the trial wave function
 
ψ(x) ∝ exp − r/ρ ,
with ρ a free parameter.
Given that the potential takes the form
Z e2
V (r ) = − exp(−r/R),
r
we will try to minimize the ground state energy by taking an ansatz for the
wave function of the form
ψ(x) ∝ exp(−r/ρ)Y
m (x̂),
by varying the parameter ρ. The Hamiltonian takes the form
p2 Z e2
H= − exp(−r/R),
2M r
which can be written in the form
2 ∂ 2 ∂ 1 Z e2
H =− r + L 2
− exp(−r/R),
2Mr 2 ∂r ∂r 2Mr 2 r
as was done in Chapter 2. Since we are examining the ground state, we will
take
= 0, because non-zero values of
will have higher energy. First, we
will normalize our wave function
  ∞ 
∗
1 = d xψ (x)ψ(x) =
3
dr r |A| exp(−2r/ρ) d 2 Y00∗ Y00
2 2
0
ρ3
= |A|2
4
2
A= .
ρ 3/2
We find for our normalized trial wave function
2 1
ψ(x) = 3/2 exp(−r/ρ)Y00 = √ 3/2 exp(−r/ρ).
ρ πρ
The expectation value of the energy in this state is given by

H ψ = d 3 x ψ ∗ (x)H ψ(x)
 ∞  
4 2 ∂ 2 ∂ Z e2
= dr r 2 3 exp(−r/ρ) − r − exp(−r/R)
ρ 2Mr 2 ∂r ∂r r
0

× exp(−r/ρ) d 2 Y00∗ Y00
44 Solutions to Lectures on Quantum Mechanics
 ∞  
4 2 2r r2
= 3 dr − − + 2 exp(−2r/ρ)
ρ 0 2M ρ ρ
 
2r r
−Z e2r exp − −
ρ R
⎡ ⎤
2  
4 ⎢  ρ ρ Ze 2
⎥
= 3 ⎣− − + − 2 ⎦
ρ 2M 2 4 2
ρ
+R 1

2 4Z e2 R 2
= − .
2Mρ 2 (2R + ρ)2 ρ
We can find the value of ρ that will minimize the energy by setting the
derivative with respect to ρ equal to zero
∂ 2 8Z e2 R 2 4Z e2 R 2
0= H ψ = − + + .
∂ρ Mρ 3 (2R + ρ)3 ρ (2R + ρ)2 ρ 2
We can simplify this expression slightly by multiplying both sides by
− 2 (2R+ρ)
M
3 ρ 3 and using the definition of the Bohr radius a =  /Me
2 2

   
Z R2 Z R3
0 = ρ + 6R − 12
3
ρ + 12R − 8
2 2
ρ + 8R 3 .
a a
Solving this equation for ρ, we find that the only real root is given by

2R 2R 3 Z (−5a + 6R Z )
ρ0 ≡ − (a − 2R Z ) +
a a
−1/3
27 4 3 2 8 2
× R Z (a − 4R Z )(a − R Z ) + 3a R Z (27a − 20a R Z − 9R Z )
2 2 2
4 4
1/3
2 2 2 8 2
+ 2R Z (a − 4R Z )(a − R Z ) +
4
3a R Z (27a − 20a R Z − 9R Z )
2 2 2 .
a 9

This gives the energy in the ground state as

2 4Z e2 R 2
H ψ = − .
2Mρ02 (2R + ρ0 )2 ρ0
3. Calculate the shift in energy of the 2 p3/2 state of hydrogen in a very weak
static electric field E, to second order in E, assuming that E is small enough
so that this shift is much less than the fine-structure splitting between the
2 p1/2 and 2 p3/2 states. In using second-order perturbation theory here, you
can consider only the intermediate state for which the energy-denominator
is smallest.
The perturbation to the Hamiltonian in this case takes the form
δ H = −eE · X,
 Chapter 5 Problem Set Solutions 45

and for convenience, we will define our coordinate system so that E points in
the 3-direction E = |E|x̂3 , and so the perturbation becomes
δ H = −e|E|X 3 .
Notice first that the first-order perturbation to the energy vanishes for n = 2
and j = 3/2 because
    
nm
 j , δ H nm
j = −e|E|δm  m nm
 j , X 3 nm
j ,
and due to the fact that X 3 changes sign under space inversion this must

vanish unless (−1)
(−1)
= −1. For the n = 2, j = 3/2 state, we must
have
= 1, and so this inner product vanishes. The second-order shift in the
energy is given by Eq. (5.4.5) as
 |(b , δ H a )|2
δ2 E a = .
b=a
Ea − Eb

The largest contribution to this sum comes from states b for which |E a − E b |
is the smallest. For the 2 p3/2 state of hydrogen, the states nearest in energy
are the 2s1/2 and 2 p1/2 states which would be degenerate with the 2 p3/2 state
in the absence of fine-structure splitting. The relevant matrix element for the
2 p1/2 state is given by
  
−e|E|δm  m 2m1 1 , X 3 2m1 3 = 0,
2 2

which vanishes due to the behavior of X 3 under space inversion because

(−1)1 (−1)1 = 1. This leaves only the 2s1/2 intermediate state. The oper-
ator X 3 acts on orbital angular momentum, but not on spin, so we need to
express the state vectors in terms of state vectors with definite values for the
m
and m s quantum numbers

nm
j = C
1 ( j m; m
m s ) ψnm

m s .
2
m
ms

After making this change of basis, we can write the matrix elements as
  
m 
m s
m
ms
ψn
, X 3 ψn 
 = δm s m s d 3 x Rn
(r )Y
m
∗ (θ, φ)r cos θ
m
× Rn;
 (r )Y

(θ, φ),
as in Eq. (5.3.5). The state vector 2m0 1 must have m
= 0 so the matrix
2
elements of X 3 between 2m0 1 and 2m1 1 receive contributions only from the
2 2
m
= 0 components of both wave functions. The relevant matrix element
then takes the form
     
± 21 ± 21 1 1 1 3 1 1
2 0 1 , X 3 2 1 1 = C 0 1 ± ;0± C1 1 ± ;0± I,
2 2 2 2 2 2 2 2 2 2
46 Solutions to Lectures on Quantum Mechanics

where we have defined


I ≡ d 3 x r cos θ R2 0 (r )Y00 (θ, φ)R2 1 (r )Y10 (θ, φ),

as in Eq. (5.3.15). This integral was evaluated in Eq. (5.3.15), which gives
I = −3a.
The relevant Clebsch–Gordan coefficients are
 
1 1 1
C0 1 ± ;0± =1
2 2 2 2
  
3 1 1 2
C1 1 ± ;0± = ,
2 2 2 2 3
and so we find
 
± 12 ± 12 √
2 0 1 , X 3 2 1 1 = − 6a.
2 2

Putting this together, we find for the second-order shift in the energy of the
2 p3/2 state of hydrogen
  2
 1 1 
−e|E|  ± 2 1 , X 3  ± 2 1 
 20 2 21 2  6e2 |E|2 a 2
δ2 E (2 p3/2 ) = = .
E (2 p3/2 ) − E (2s1/2 ) E (2 p3/2 ) − E (2s1/2 )
The fine-structure splitting gives an energy difference of
 
m Z 4 e8 1 m Z 4 e8 Z 4 e6
E (2 p3/2 ) − E (2s1/2 ) = − 1 = − = − ,
164 c2 2 324 c2 32a2 c2
and so we find for the second-order energy shift
1922 c2 a 3 |E|2
δ2 E (2 p3/2 ) = − .
Z 4 e4
4. The spin-orbit coupling of the electron in hydrogen produces a term in the
Hamiltonian of the form
H = ξ(r )L · S,
where ξ(r ) is some small function of r . Give a formula for the contribution
of V to the fine-structure splitting between the 2 p1/2 and 2 p3/2 states in
hydrogen, to first order in ξ(r ).
The first-order shift in energy is given by
 
δ1 E = nm
j , H nm
j ,
 Chapter 5 Problem Set Solutions 47

where H = ξ(r )L · S. First, we recall that because [L, S] = 0, we may

write

J2 = L2 + S2 + 2L · S,

and so
1 2 
L·S= J − L2 − S2 .
2
We can therefore write the first-order energy shift as
 
1  2  m
δ1 E = n
j , ξ(r ) J − L − S n
j
m 2 2
2
2  
 3  m 
= j ( j + 1) −
(
+ 1) − n
j , ξ(r )nm
j
2 4
2  
 3
= j ( j + 1) −
(
+ 1) − dr r 2 ξ(r ) (Rn
(r ))2 .
2 4
For both the 2 p1/2 and 2 p3/2 states of hydrogen, the radial integral takes the
form
    3/2   2
1 1 r
dr r ξ(r ) (R2 1 (r )) = dr r ξ(r ) √
2 2 2
exp(−r/2a)
3 2a a

r4
= dr ξ(r ) exp(−r/a).
24a 5
The energy shift for the 2 p3/2 state is given by
   
2 3 3 3 r4
δ1 E (2 p3/2 ) = + 1 − 1 (1 + 1) − dr ξ(r ) exp(−r/a)
2 2 2 4 24a 5

2 r4
= dr ξ(r ) exp(−r/a),
2 24a 5
and the energy shift for the 2 p1/2 state is given by
   
2 1 1 3 r4
δ1 E (2 p1/2 ) = + 1 − 1 (1 + 1) − dr ξ(r ) exp(−r/a)
2 2 2 4 24a 5

r4
= − 2
dr ξ(r ) exp(−r/a).
24a 5
This gives a splitting between the states of

2
δ1 E (2 p3/2 ) − δ1 E (2 p1/2 ) = dr ξ(r )r 4 exp(−r/a).
16a 5
48 Solutions to Lectures on Quantum Mechanics

5. Using the WKB approximation, derive a formula for the energies of the bound
s states of a particle of mass m in a potential V (r ) = −V0 e−r/R , with V0 and
R both positive.
Eq. (5.7.33) gives the condition for a bound state for a central potential with

= 0 as
 bE  
  3
k(r ) dr = n + π,
0 4

where the function k(r ) is defined as in Eq. (5.7.2)


2m
k(r ) = (E − U (r )),
2
where U (r ) is given by

2
(
+ 1)
U (r ) = V (r ) + ,
2m r2
and b E is the turning point, defined by the condition that U (b E ) = E. In this
case, we are considering only states with
= 0 and a potential given by

V (r ) = −V0 exp(−r/R).

We find that b E is given by


E
b E = −R ln − ,
V0

and k(r ) is

2m
k(r ) = (E + V0 exp(−r/R)).
2
The integral appearing in the condition for a bound state becomes
  
bE bE
2m
k(r ) dr = (E + V0 exp(−r/R))
dr
0 0 2
    b E
2mV0 E E −1 E
= −2R + exp(−r/R) + − sin − exp(−r/R)
2 V0 V0 V0
0
      
2m  √ E |E|
= 2R E + V0 + −E sin−1 − − sin−1 .
 2 V0 V0
 Chapter 5 Problem Set Solutions 49

The condition for a bound state is then a transcendental equation for E which
reads
      
2m  √ −1 E −1 |E|
2R E + V0 + −E sin − − sin
2 V0 V0
 
3
= n+ π,
4
whose solutions give the values of the energy levels.
50 Solutions to Lectures on Quantum Mechanics

Chapter 6 Problem Set Solutions

1. Consider a time-dependent Hamiltonian H = H0 + H  (t), with

H  (t) = U exp(−t/T ),
where H0 and U are time-independent operators, and T is a constant. What
is the probability to lowest order in U that the perturbation will produce a
transition from one eigenstate n of H0 to a different eigenstate m of H0 during
a time interval from t = 0 to a time t  T ?
If we expand the solution of the full Hamiltonian H in terms of eigenstates
of the unperturbed Hamiltonian H0 as in Eq. (6.1.4)

(t) = cn (t) exp(−i E n (t)/)n ,
n

then the coefficients in this expansion can be found to first order in the
perturbation H  (t) as in Eq. (6.1.6)
 t
i 
c j (t) ≈ c j (0) − ck (0) dt  H jk

(t  ) exp(i(E j − E k )t  /),
 k 0


 
where H jk (t  ) ≡  j , H  (t  )k . To find the probability of a transition from
a state n to a state m, we set cn (0) = 1 and ck (0) for all k = n, so the state
is described at time t = 0 by n . If we take H  (t) = U exp(−t/T ), we can
then calculate the coefficient of the the state m after a time t as

i t   
cm (t) ≈ − dt Hmn (t ) exp(i(E m − E n )t  /)
 0

i t 
=− dt Umn exp(−t  /T ) exp(i(E m − E n )t  /),
 0
where we have defined Umn = (m , U n ). In the limit where t  T , this
becomes
 ∞   
i  Em − En 1 
cm (t) = − Umn dt exp i − t
  T
0 
i −1
= − Umn Em −En
 i  − T1
iUmn
= 
.
i(E m − E n ) − T
The probability of finding the system in the state m is then given in the limit
of large t by
|Umn |2
|(m , (t))|2 = |cm (t)|2 → 2
.
(E m − E n )2 + T2
 Chapter 6 Problem Set Solutions 51

2. Calculate the rate of ionization of a hydrogen atom in the 2 p state in

a monochromatic external electric field, averaged over the component of
angular momentum in the direction of the field. (Ignore spin.)
Without loss of generality, we can choose our coordinate system such that
the electric field lies along the 3-direction so that the perturbation in the
Hamiltonian takes the form of Eq. (6.3.2)
H  (t) = −eE X 3 exp(−iωt) − eE ∗ X 3 exp(iωt),
where X is the operator for the electron position. This means that the operator
U appearing in Eq. (6.2.1) is
U = −eE X 3 .
We need to calculate the matrix element of this perturbation between the
normalized wave function describing the 2 p state of hydrogen and the free
electron wave function, given by Eq. (3.5.12)
ψk (x) = (2π)−3/2 exp(ik · x),
where the momentum of the electron is k which we take to be much larger
than the hydrogen binding energy such that we may ignore the effect of the
Coulomb field of the proton in the final state. Recall the form of the wave
functions for the 2 p state of hydrogen
1 r
ψ2m=0
p =√ e−r/2a cos θ
32π a 5/2
1 r
ψ2m=±1 = ∓ √ 5/2 e−r/2a sin θ e±iφ ,
p
8 πa
which were worked out for the Chapter 2 problem set. The relevant matrix
elements are then

eE
Uk 2 pm=0 = − √ d 3 x e−ik·x x3r e−r/2a cos θ
(2π)3/2 32πa 5

eE
Uk 2 pm=±1 = ± √ d 3 x e−ik·x x3r e−r/2a sin θ e±iφ .
(2π)3/2 64πa 5

We can rewrite these integrals as


eE
Uk 2 pm=0 =− √ d 3 x e−ik·x x32 e−r/2a
(2π)3/2 32πa 5

eE
Uk 2 pm=±1 =± √ d 3 x e−ik·x x3 (x1 ± i x2 )e−r/2a ,
(2π)3/2 64πa 5
and then we can perform the angular integral by recalling that
 
−ik·x 1 ∞
3
d xe f (r ) = dr 4πr f (r ) sin kr.
k 0
52 Solutions to Lectures on Quantum Mechanics

Differentiating both sides twice with respect to k3 , we find

 
−ik·x 1 ∞
− d xe 3
f (r )x3 = 3
2
dr 4πr f (r )
k 0

k32 k32r
× − sin kr + kr cos kr + 3 2 sin kr − 3 cos kr − k3 r sin kr ,
2 2
k k
 
or applying ∂∂x3 ∂∂x1 ± i ∂∂x2 to both sides, we find
 
−ik·x 1 ∞
− d xe 3
f (r )x3 (x1 ± i x2 ) = 3 dr 4πr f (r )
k 0

k3 (k1 ± ik2 ) k3 (k1 ± ik2 )r
× 3 sin kr − 3 cos kr − k (k
3 1 ± ik 2 )r 2
sin kr .
k2 k
Using these expressions in the matrix elements gives
 ∞
4πeE −r/2a
Uk 2 pm=0 = √ dr r e − sin kr + kr cos kr
(2π)3/2 k 3 32πa 5 0

k32 k32r
+ 3 2 sin kr − 3 cos kr − k3 r sin kr ,
2 2
k k
 ∞
4πeE −r/2a k3 (k1 ± ik2 )
Uk 2 pm=±1 = ∓ √ dr r e 3 sin kr
(2π) k 64πa 0
3/2 3 5 k2

k3 (k1 ± ik2 )r
−3 cos kr − k3 (k1 ± ik2 )r sin kr .
2
k
The radial integrals can be evaluated to give
 ∞
−r/2a k2
dr r e − sin kr + kr cos kr + 3 32 sin kr
0 k

2
k r 256a 5 k 3 (1 + 4a 2 (k 2 − 6k32 ))
− 3 3 cos kr − k32r 2 sin kr = − ,
k (1 + 4a 2 k 2 )4
 ∞
k3 (k1 ± ik2 ) k3 (k1 ± ik2 )r
dr r e−r/2a 3 2
sin kr − 3 cos kr
0 k k

6144a 7 k 3 k3 (k1 ∓ ik2 )
− k3 (k1 ± ik2 )r 2 sin kr = ±i .
(1 + 4a 2 k 2 )4
Plugging these into the matrix elements, we find
64a 5/2 eE(1 + 4a 2 (k 2 − 6k32 ))
Uk 2 pm=0 = ,
π(1 + 4a 2 k 2 )4
√ 9/2
768 2a eEk3 (k1 ∓ k2 )
Uk 2 pm=±1 = −i .
π(1 + 4a 2 k 2 )4
 Chapter 6 Problem Set Solutions 53

Using Eq. (6.2.6), the differential ionization rate is

2π
d(2 p → k) = |Uk 2 p |2 δ(−B2 p + ω − E(k))k 2 dk d ,

where B2 p ≈ 3.4 eV is the binding energy of the 2 p state of hydrogen,
d is the differential element of solid angle of the final electron direction,
and E(k) = 2 k2 /2m e . Now, dk = m e d E/2 k, and the effect of the factor
d E δ(−B2 p + ω − E(k)) in the integral over k is just to set k equal to the
value fixed by the conservation of energy

k = 2m e (ω − B2 p )/,
and so the differential ionization rate is
d(2 p → k) 2πm e k
= |Uk 2 p |2 ,
d 3
with k given by the expression above. In order to calculate the rate averaged
over the component of angular momentum in the direction of the field, we
need to average the squares of the matrix elements we calculated above. The
norm square matrix elements are given by
212 a 5 e2 E 2 (1 + 4a 2 (k 2 − 6k 2 cos2 θ))2
|Uk 2 pm=0 |2 = ,
π 2 (1 + 4a 2 k 2 )8
217 32 a 9 e2 E 2 k 4 cos2 θ sin2 θ
|Uk 2 pm=±1 |2 = .
π 2 (1 + 4a 2 k 2 )8
These formulas only apply in the limit of large k, and so we may drop the
terms which are suppressed by k 2 to find
e2 E 2 (1 − 6 cos2 θ)2
|Uk 2 pm=0 |2 = ,
π 2 a 7 k 12
18e2 E 2 cos2 θ sin2 θ
|Uk 2 pm=±1 |2 = ,
π 2 a 7 k 12
and so the average for large k is
1 
|Uk 2 p |2 = |Uk 2 pm=0 |2 + |Uk 2 pm=1 |2 + |Uk 2 pm=−1 |2
3
e2 E 2 (1 + 24 cos2 θ)
= .
3π 2 a 7 k 12
Finally, we find for the differential ionization rate
d(2 p → k) 2m e e2 E 2 (1 + 24 cos2 θ)
= .
d 3π3 a 7 k 11
3. Consider a Hamiltonian H [s] that depends on a number of slowly varying
parameters collectively called s(t). What is the effect on the Berry phase
54 Solutions to Lectures on Quantum Mechanics

γn [C] for a given closed curve C, if H [s] is replaced with f [s]H [s], where
f [s] is an arbitrary real numerical function of the s?
We begin with the formula for the Berry phase given in Eq. (6.7.8)
  
γn [C] = i d Ai j
A[C] i j m=n
  ∗   
∂ H [s] ∂ H [s]
× n [s], m [s] n [s], m [s] (E m [s] − E n [s])−2 .
∂si ∂s j
If we replace H [s] by f [s]H [s], the Schrödinger equation appearing in
Eq. (6.6.1) is modified to

f [s]H [s]n [s] = f [s]E n [s]n [s],

and the derivatives appearing in the Berry phase become

∂ f [s]H [s] ∂ f [s] ∂ H [s]

= H [s] + f [s].
∂si ∂si ∂si

The relevant matrix elements are then

     
∂ f [s]H [s] ∂ H [s]
n [s], m [s] = n [s], m [s] f [s]
∂si ∂s j
∂ f [s]
+ (n [s], H [s]m [s])
∂si
  
∂ H [s]
= n [s], m [s] f [s]
∂s j
∂ f [s]
+ (n [s], m [s]) E m
∂si
  
∂ H [s]
= n [s], m [s] f [s]
∂s j
∂ f [s]
+ δnm E m .
∂si
Using this to calculate the Berry phase, we find
 
γn [C] = i d Ai j
A[C] i j


∂ H [s]
 ∗ 
∂ H [s]
 
× n [s], m [s] n [s], m [s]
m=n
∂si ∂s j
× f [s]2 ( f [s]E m [s] − f [s]E n [s])−2
 Chapter 6 Problem Set Solutions 55
  
=i d Ai j
A[C] i j m=n
  ∗   
∂ H [s] ∂ H [s]
× n [s], m [s] n [s], m [s]
∂si ∂s j
× (E m [s] − E n [s])−2 ,

where the terms involving ∂ f [s]/∂si have gone away because the sum
excludes the term n = m, and the factors of f [s] cancel due to the fact
that both the Hamiltonian and the energy eigenvalues are rescaled. We there-
fore find that the Berry phase is unchanged by the transformation H [s] →
f [s]H [s] for a real function f [s].
56 Solutions to Lectures on Quantum Mechanics

Chapter 7 Problem Set Solutions

1. Use the Born approximation to give a formula for the s-wave scattering
length as for scattering of a particle of mass μ and wave number k by an
arbitrary central potential V (r ) of finite range R, in the limit k R  1. Use
this result and the optical theorem to calculate the imaginary part of the
forward scattering amplitude to second order in the potential.
In the Born approximation, Eq. (7.4.2) gives the scattering amplitude for a
central potential

2μ ∞ sin qr
f k (x̂) = − 2 dr r 2 V (r ) ,
 0 qr
where q is defined as in Eq. (7.4.3)
q ≡ |k − k x̂| = 2k sin(θ/2).
For k R  1 where R is the range of the potential, we can expand
 
sin qr ≈ 2kr sin(θ/2) + O (kr )3 ,
and in this limit the scattering amplitude becomes
  
2μ ∞ 2kr sin(θ/2) + O (kr )3
f k (x̂) ≈ − 2 dr r 2 V (r )
 0 2kr sin(θ/2)

2μ ∞  
=− 2 dr r 2 V (r ) + O (k R)2 .
 0
We can also expand the scattering amplitude in Legendre polynomials as in
Eq. (7.5.10)
∞
1 
f k (x̂) (2
+ 1)P
(cos θ)(e2iδ
− 1).
2ik
=0

For k → 0, the quantity tan δ

vanishes as k 2
+1 , so δ0 dominates, and so in
this limit,
1
k cot δ0 → − + O(k 2 ),
as
tan δ0 ≈ δ0 → −as k.
We can therefore write the scattering amplitude in terms of the scattering
length as
1 δ0
f k (x̂) ≈ (2 × 0 + 1)P0 (cos θ)(1 + 2iδ0 − 1) = ≈ −as .
2ik k
Comparing this with the expression above, we find that in the Born approxi-
mation the s-wave scattering length for a central potential is given by
 Chapter 7 Problem Set Solutions 57

2μ ∞  
as ≈ 2 dr r 2 V (r ) + O (k R)2 .
 0
The cross-section is given by

σ = d | f k (x̂)|2 ,

which in this limit goes to

 ∞ 2
16μ2
σ → 4πas2 = dr r V (r ) .
2
4 0

Then, according to Eq. (7.3.8) the optical theorem gives

 ∞ 2
k 4kμ2
Im f k (0) = σ = dr r V (r ) .
2
4π 4 0

2. Suppose that in the scattering of a spinless non-relativistic particle of mass

μ by an unknown potential, a resonance is observed at energy E R for which
the elastic cross section at the peak of the resonance is σmax . Show how to use
this data to give a value for the orbital angular momentum of the resonant
state.
We begin with the Breit–Wigner formula given in Eq. (7.6.12)
π(2
0 + 1) 2
σscat = .
k2 (E − E R )2 +  2 /4
Setting E = E R , we find
4π(2
0 + 1)
σmax = ,
k 2R
where we have defined
2μE R
k 2R = .
2
We can rearrange this equation to solve for
0 , which gives
 
1 μE R σmax

0 = − 1 .
2 2π2
3. Give a formula for the tangent of the
= 0 phase shift for scattering by a
potential
+
−V0 r < R
V (r ) = ,
0 r≥R
for all E > 0, and to all orders in V0 > 0.
58 Solutions to Lectures on Quantum Mechanics

We are dealing here with a central potential of the form


−V0 r < R
V (r ) =
0 R≥R
We will use Eq. (7.5.17) which reads
k j
 (k R) − 
(k) j
(k R)
tan δ
(k) = ,
kn 
(k R) − 
(k)n
(k R)
where 
(k) is defined in Eq. (7.5.16)
R
 (R)

(k) = ,
R
(R)
and R
(r ) is the radial wave function. For u(r ) = r R(r ), the radial part of
the Schrödinger equation for r < R takes the form
d 2 u
(
+ 1) 2μ
− u + 2 V0 u + k 2 u = 0.
dr 2 r 2 
Setting
= 0 this gives
 
d 2u 2μ 2μ
= − k + 2 V0 = − 2 (E + V0 ) ≡ −k̃ 2 u.
2
dr 2  
We can then solve for u(r ) which gives
u(r ) = A sin k̃r + B cos k̃r,
and so R0 (r ) is
A B
R0 (r ) = sin k̃r + cos k̃r.
r r
The radial wave function must be finite at r = 0, and so we must set B = 0.
We therefore find for the radial wave function for
= 0 in the range r < R
A
R0 (r ) = sin k̃r,
r
and for its derivative we get
A A
R0 (r ) = k̃ cos k̃r − 2 sin k̃r.
r r
We can then calculate 0 (k)
R0 (R) 1
0 (k) = = k̃ cot k̃ R − .
R0 (R) R
The relevent Bessel functions and their derivatives are
sin z
j0 (z) =
z
 Chapter 7 Problem Set Solutions 59

cos z sin z
j0 (z) = − 2
z z
cos z
n 0 (z) = −
z
sin z cos z
n 0 (z) = + 2 .
z z
Putting this together, we find for the tangent of the
= 0 phase shift
 cos k R sin k R   
R
− kR 2 − k̃ cot k̃ R − 1 sin k R
R kR
tan δ0 (k) =    
sin k R
R
+ cos kR
k R2
+ k̃ cot k̃ R − R1 cosk Rk R
k cos k R − k̃ cot k̃ R sin k R
= ,
k sin k R + k̃ cot k̃ R cos k R
"
2μ(E+V0 )
where k̃ ≡ 2
.
4. Suppose that the eigenstates of an unperturbed Hamiltonian include not only
continuum states of a free particle with momentum p and unperturbed energy
E = p2 /2μ, but also a discrete state of angular momentum
with a negative
unperturbed energy. Suppose that when we turn on the interaction, the con-
tinuum states feel a local potential, but remain in the continuum, while also
the discrete state moves to positive energy, thereby becoming unstable. What
is the change in the phase shift δ
(k) as the wave number k increases from
k = 0 to k = ∞?
If we place the system in a sphere of radius R, the number of states with
angular momentum
and energies between 0 and E is given by Eq. (7.8.2)
1
N
(E) = (k R + δ
(E) − δ
(0)).
π
The change in the number of scattering states with energy between 0 and E
due to the interaction is given by Eq. (7.8.3)
1
N
(E) =
(δ
(E) − δ
(0)).
π
Because one bound state becomes a positive energy scattering state when the
interaction is turned on, we must have
1
N
(∞) = (δ
(∞) − δ
(0)) = 1,
π
and as a result we must have
δ
(∞) − δ
(0) = π.
5. Find an upper bound on the elastic scattering cross section in the case where
the scattering amplitude f is independent of angles θ and φ.
60 Solutions to Lectures on Quantum Mechanics

The scattering cross-section is given by

 π  2π
σscattering ≡ sin θ dθ dφ | f k (θ, φ)|2 ,
0 0

which by the optical theorem Eq. (7.3.8) is equal to

4π
σscattering =Im f k (0).
k
In the case where the scattering amplitude is independent of angle, we have
f k (θ, φ) = f k (0),
and so the scattering cross-section is
 π  2π
σscattering = sin θ dθ dφ | f k (0)|2
0 0
= 4π| f k (0)|2 ,
and since the absolute value square of any function is always greater than the
absolute value square of its imaginary part, we may write
 2 
k
4π| f k (0)| ≥ 4π|Im f k (0)| = 4π
2 2
σ 2
,
16π 2 scattering
where we have again used the optical theorem. We therefore find that
k2 2
σscattering ≥
σ ,
4π scattering
and so we find an upper bound on the scattering cross-section
4π
σscattering ≤
.
k2
6. Use the eikonal approximation to calculate the phase of the scattering ampli-
tude for the scattering of a non-relativistic charged particle by a Coulomb
potential.
We are interested in calculating the phase of the scattering amplitude for the
scattering of a non-relativistic particle by a Coulomb potential. We begin
with Eq. (7.10.3) which in this case reads
(∇ S)2 Z 1 Z 2 e2
+ = E.
2m r
We imagine starting with a particle very far from the scattering center with
energy 2mk moving in the z-direction toward the scattering center with an
2 2

impact parameter b. Since the eikonal approximation is only valid where

the wave function ψ(x) varies much more rapidly with x than the potential,
we require that E  V . In this limit, the deflection of the particle will be
 Chapter 7 Problem Set Solutions 61

small, and we can approximate the classical trajectory of the particle as a

straight line. We can then solve the above expression for the phase of the
wave function as
 z 1/2
S 2m Z 1 Z 2 e2
= k − 2 √
2
dz  + S0 .
 −∞  b +z
2 2

In the case when the particle is uncharged, this expression should give S =
kz, and so we can choose the constant S0 such that
 z  2 1/2

S 2m Z Z e
− k dz  + kz.
1 2
= k2 − 2 √
 −∞  b2 + z 2
√
Since we are working in the regime where 2 k 2 /2m  Z 1 Z 2 e2 / b2 + z 2 ,
we can expand the square root to get
 z
S m Z 1 Z 2 e2
≈ kz − 2 √ dz  .
  k −∞ b2 + z 2
Now, recalling Eq. (7.10.2) we find that the wave function takes the form
  
1 im z Z 1 Z 2 e2 
ψ(x) = e ikz
exp − √ dz ,
(2π)3/2 2 k −∞ b2 + z 2
where we have neglected the spatial variation of N (x) at this order of approx-
imation. Now recall that the scattering amplitude is given in Eq. (7.2.6) as

m
f k (x̂) = − (2π) 3/2
d 3 y e−ik x̂·y V (y)ψk (y).
2π2
In this case, the expression for the scattering amplitude takes the form
 2
m 3  −ik x̂·x Z 1 Z 2 e 
f k (x̂) = − d x e √ eik·x
2π 2
b +z
2 2
  z 
im Z 1 Z 2 e2 
× exp − 2 √ dz .
 k −∞ b2 + z 2

Now we will use cylindrical coordinates to do the integral over x for which
d 3 x  = b db dφ dz  . Let us examine the factor (k − k x̂) · x appearing in the
exponential, which here becomes
(k − k x̂) · x = (k − k x̂) · (b + z  ẑ) ≈ (k − k x̂) · b,
because (k − k x̂) · ẑ is negligible for small deflection angle. We can then
express this in cylindrical coordinates as
 
 θ
(k − k x̂) · x ≈ −2kb sin cos φ,
2
62 Solutions to Lectures on Quantum Mechanics
 
where we have used the fact that k − k x̂ = q where q = 2k sin θ2 . Our
expression for the scattering amplitude now becomes
 ∞  2π
m
f k (x̂) = − b db e−iqb cos φ dφ
2π2 0
 ∞
0
   
 Z 1 Z 2e
2
im z Z 1 Z 2 e2 
× dz √ exp − 2 √ dz .
−∞ b2 + z 2  k −∞ b2 + z 2
Next, we notice that we can write
 ∞   z 
2 2
Z Z e im Z Z e
dz  √ dz 
1 2 1 2
exp − 2 √
−∞ b2 + z 2  k −∞ b2 + z 2
   z  =∞
i2 k im z Z 1 Z 2 e2 

= exp − 2 √ dz   ,
m  k −∞ b + z
2 2 
z =−∞

and the integral over φ can be done to give

 2π
dφ e−iqb cos φ = 2π J0 (qb).
0

The scattering amplitude then takes the form

 ∞
 
f k (θ) = −ik db b J0 (qb) eiχ (b) − 1 ,
0

where we have defined the Eikonal phase

 ∞
m Z 1 Z 2 e2
χ (b) = − 2 √ dz  .
 k −∞ b2 + z 2
This integral is logarithmically divergent, but it can be computed by intro-
ducing a screening radius and then taking the limit in which the screening
radius tends to infinity. The result is
2m Z 1 Z 2 e2
χ (b) = − ln(kb).
2 k
The scattering amplitude is then
 ∞
 
f k (θ) = −ik db b J0 (qb) eiχ (b) − 1
0
 ∞   
ik kx 2iξ
=− 2 J0 (x) − 1 x dx
q 0 q
    ∞ 
ik k 2iξ ∞
=− 2 J0 (x)x 2iξ +1 d x − J0 (x)x d x ,
q q 0 0
 Chapter 7 Problem Set Solutions 63

where we have defined ξ as in Eq. (7.9.5)

m Z 1 Z 2 e2
ξ≡ .
2 k
We can then evaluate these integrals by using the identity

x J0 (x) d x = x J1 (x),

and integrating by parts in the first term, and so we find

  ∞  2iξ ∞
ik k 2iξ 2iξ +1 
 k
f k (θ) = − 2 x J1 (x) + 2iξ x 2iξ J1 (x) d x + x J1 (x)|∞
q q  q 0
0

 0
 2iξ 
ik  2iξ ln(kx/q) ∞ k (1 + iξ )
= − 2 x J1 (x) e − 1 0 − 2iξ 2 2iξ
q q (1 − iξ )
 2iξ
2kξ 2k (1 + iξ )
=− 2
q q (1 − iξ )
  
2m Z 1 Z 2 (1 + iξ ) 2 θ
=− 2 2 exp −iξ ln sin .
 q (1 − iξ ) 2
Now, comparing with Eqs. (7.9.13) and (7.9.14), in the limit of large r
where we can neglect the logarithmic term in the exponential, we find that
the eikonal approximation has reproduced the exact expression for the scat-
tering amplitude valid at all angles. This is a bit surprising considering
that we approximated the classical path as a straight line, but it should be
noted that this feature is specific to the Coulomb potential, and the eikonal
approximation does not give the exact scattering amplitude for general
potentials.
64 Solutions to Lectures on Quantum Mechanics

Chapter 8 Problem Set Solutions

1. Consider a general Hamiltonian H0 +V , where H0 is the free-particle energy.

Define a state α0 by the modified Lippmann–Schwinger equation
(E α − H0 )
α0 = α + V α0 ,
(E α − H0 )2 +  2
where α is an eigenstate of H0 with eigenvalue E α , and  is a positive
infinitesimal quantity. Define
 
Aβα ≡ β , V α0 .

(a) Show that Aβα = A∗αβ for E β = E α .

(b) For the simple case of a non-relativistic particle with energy k2 2 /2μ in
a local potential V (x),
 calculate
 the asymptotic behavior of the coordinate-
space wave function x , k of the state k0 for x → ∞. Express the result
0

in terms of matrix elements of A.

(a) We begin with the definition
(E α − H0 )
α0 = α + V α0 ,
(E α − H0 )2 +  2
where α is an eigenstate of the free-particle Hamiltonian H0 with eigen-
value E α and  is a positive infinitesimal quantity. We will find it useful
to rearrange this equation to get an expression for α in terms of α0
(E α − H0 )
α = α0 − V α0 .
(E α − H0 )2 +  2
The quantity Aβα is defined as
 
Aβα ≡ β , V α0 .
We can then write
 ∗  
A∗αβ = α , V β0 = V β0 , α
 
(E α − H0 )
= V β , α −
0 0
V α0
(E α − H0 )2 +  2
 
  (E α − H0 )
= V β , α − V β ,
0 0 0
V α0
(E α − H0 )2 +  2
 
  (E α − H0 )
= β0 , V α0 − V  0
β , V  α ,
0
(E α − H0 )2 +  2
where in the last line we have used the fact that V and H0 are Hermitian,
and that E α and  are real. We can now combine these inner products and
use the fact that E α = E β to find
 Chapter 8 Problem Set Solutions 65
  
(E α − H0 )
A∗αβ = β0 − V  0
β , V  0
α
(E α − H0 )2 +  2
 
= β , V α0
= Aβα .

(b) We
 are0 interested in calculating the coordinate-space wave function
x , k for a non-relativistic particle with energy k2 2 /2μ in a local
potential V (x). We will use the modified Lippmann–Schwinger equation
to write
⎛  2 2  ⎞
k
  − H
⎜ 2μ 0
⎟
x , k0 = (x , k ) + ⎝x ,  2
V k0 ⎠ .
2 k2
2μ
− H0 +  2

The first inner product is just the free-particle wave function given in
Eq. (3.5.12), and we will insert a complete set of states p in the second
inner product which gives
⎛  2 2  ⎞
 k
− H
  e ik·x  ⎜ 2μ 0
⎟
x , k0 = + d 3 p x , p ⎝p ,   V k0 ⎠
(2π) 3/2
2 k2
2
2μ
− H0 +  2
 22 
 k
− 2 p2
e ik·x   2μ 2μ  
= + d 3 p x , p   p , V k0
(2π) 3/2 2
2 k 2
− 2μp
2 2
2μ
+ 2
  
eik·x 2μ   k 2
− p 2  
= + 2 d 3 p x , p  2 p , V k0 ,
(2π) 3/2  k 2 − p 2 +  2

where in the second expression, we have defined   = 2μ 2

 which is also
a real positive infinitesimal quantity. Now we insert another complete set
of states y in the second term, and we find
  2 
  eik·x 2μ   k − p2
x , k0= + 2 d p d y x , p 
3 3
2
(2π)3/2  k 2 − p 2 +  2
  
× p , y y , V k0
  2 
eik·x 2μ d 3 p 3 ip·(x−y) k − p2  
= + 2 d ye  2 y , V k0 .
(2π) 3/2  (2π) 3
k 2 − p 2 +  2

Now we can perform the integral over p by the same method used in
Eq. (7.4.2), that is
66 Solutions to Lectures on Quantum Mechanics
  2   ∞  2 
d 3 p ip·(x−y) k − p2 4π p 2 dp sin( p|x − y|) k − p2
e  2 =  2
(2π)3 k 2 − p 2 +  2 0 (2π)3 p|x − y| k 2 − p 2 +  2
 ∞    2 
−i −i p|x−y| k − p2
= dp p e i p|x−y|
− e  2
4π 2 |x − y| 0 k 2 − p 2 +  2
 ∞  2 
−i k − p2
= d p pe i p|x−y|
 2 ,
4π 2 |x − y| −∞ k 2 − p 2 +  2

where in going to the third line we have taken p → − p in the second

term in parentheses in order to combine the integrals. Now we can write
   2 
k 2 − p2 k − p2
 2 = 2  ,
k 2 − p 2 +  2 k − p 2 − i  k 2 − p 2 + i 

and then we can do the integral by closing the contour in the upper
half complex plane and calculating the contribution of the residues at
p = k + i  and p = −k + i  . This gives
  2  ik|x−y| 
d 3 p ip·(x − y) k − p2 −i ke ke−ik|x−y|
e  2 = 2 (−2πi) −
(2π)3 k 2 − p2 +  2 4π |x − y| 2k −2k
−1  
= eik|x−y| + e−ik|x−y| ,
4π|x − y|

which we can plug into the coordinate-space wave function to get


  eik·x   2μ 1  ik|x−y| 
x , k0 = − d 3 y y , V k0 2 e + e−ik|x−y| .
(2π)3/2  4π|x − y|

Now we can insert another complete set of states q to get


  eik·x    2μ 1
x , k0 = − d 3 y d 3 q y , q q , V k0 2
(2π)3/2  4π|x − y|
 ik|x−y| −ik|x−y|

× e +e

eik·x eiq·y 2μ 1
= − d 3 y d 3q Aqk 2
(2π)3/2 (2π) 3/2  4π|x − y|
 ik|x−y| −ik|x−y|

× e +e .

In the limit of large x, we can expand |x − y| to get


y 2y · x̂
|x − y| = r |x̂ − | → r 1 − ≈ r − y · x̂,
r r
 Chapter 8 Problem Set Solutions 67

where r ≡ |x|. In this limit, the wave function takes the form

  eik·x eiq·y 2μ 1
x , k =
0
− d 3
y d 3
q Aqk 2
(2π) 3/2 (2π) 3/2  4π|x − y|
 
ikr −ik x̂·y −ikr ik x̂·y
× e e +e e
 √
eik·x 2π μ
= − d q 3
(2π)3/2 2 r
 
× δ 3 (q − k x̂)eikr Aqk + δ 3 (q + k x̂)e−ikr Aqk
√
eik·x 2π μ  ikr −ikr

= − e A +k x̂ k + e A −k x̂ k .
(2π)3/2 2 r
2. Consider a separable interaction, whose matrix elements between free-
particle states have the form
 
β , V α = f (α) f ∗ (β),
where f (α) is some general function of the momenta and other quantum
numbers characterizing the free-particle state α .
(a) Find an exact solution of the Lippmann–Schwinger equation for the “in”
state in this theory.
(b) Use the result of (a) to calculate the S-matrix.
(c) Verify the unitarity of the S-matrix.
(a) The Lippmann–Schwinger equation is given in Eq. (8.1.6) which reads
α+ = α + (E α − H0 + i)−1 V α+ ,
where α is an eigenstate of H0 with eigenvalue E α . We can insert a
complete set of states α1 in the second term and then expand using the
Lippmann–Schwinger equation again to find

 
α = α + dα1 α1 α1 , (E α − H0 + i)−1 V α+
+


= α + dα1 α1 (E α − E α1 + i)−1
   
× α1 , V α + α1 , V (E α − H0 + i)−1 V α+ .
 
We can then use the fact that β , V α = f (α) f ∗ (β) in the first term
in brackets, and insert another complete set of states α2 in the second
term in brackets to get

f (α) f ∗ (α1 )
α+ = α + dα1 α
(E α − E α1 + i) 1
  
α1 , V (E α − H0 + i)−1 α2  
+ dα1 dα2 α2 , V α+ α1
E α − E α1 + i
68 Solutions to Lectures on Quantum Mechanics

f (α) f ∗ (α1 )
= α + dα1 α
(E α − E α1 + i) 1

f (α2 ) f ∗ (α1 ) f (α) f ∗ (α2 )
+ dα1 dα2 α
(E α − E α1 + i)(E α − E α2 + i) 1
  
f (α) f ∗ (α1 ) α2 , V (E α − H0 + i)−1 V α+
+ dα1 dα2 α1 .
(E α − E α1 + i)(E α − E α2 + i)
We can then iterate this process of inserting complete sets of states and
using the Lippmann–Schwinger equation to get

f (α) f ∗ (α1 )
α+ = α + dα1 α
(E α − E α1 + i) 1

f (α) f ∗ (α1 )| f (α2 )|2
+ dα1 dα2 α
(E α − E α1 + i)(E α − E α2 + i) 1

f (α) f ∗ (α1 )| f (α2 )|2 | f (α3 )|2
+ dα1 dα2 dα3 α + · · ·
(E α − E α1 + i)(E α − E α2 + i)(E α − E α3 + i) 1

f (α) f ∗ (α1 )
= α + dα1 α
(E α − E α1 + i) 1
   2 
| f (β)|2 | f (β)|2
× 1 + dβ + dβ + ··· .
(E α − E β + i) (E α − E β + i)

The term in square brackets is just an infinite power series which we can
sum to get an explicit expression for α+
0 f ∗ (α1 )
dα1 (Efα(α)
−E α1 +i) α1

+
α = α + 0 .
1 − dβ (Eα| −E f (β)|2
β +i)

(b) The S-matrix is given by Eq. (8.1.10) as

Sβα = δ(β − α) − 2πiδ(E α − E β )Tβα ,
where Tβα is defined in Eq. (8.1.11) as
 
Tβα ≡ β , V α+ .
Using our expression for α+ from above, we can calculate Tβα in this
case to be
0 f ∗ (α1 )  
  dα1 (Efα(α)
−E α1 +i)
 β , V α1
Tβα = β , V α + 0
1 − dγ (Eα| −E f (γ )|2
γ +i)
0 f (α) f ∗ (α1 ) f (α1 ) f ∗ (β)
dα1 (E α −E α1 +i)
= f (α) f ∗ (β) + 0
1 − dγ (Eα| −E f (γ )|2
γ +i)
 Chapter 8 Problem Set Solutions 69
⎡ 0 ⎤
f (α1 )|2
dα1 (Eα| −E +i)
= f (α) f ∗ (β) ⎣1 + ⎦
α1
0 | f (γ )|2
1 − dγ (Eα −Eγ +i)
f (α) f ∗ (β)
= 0 .
1 − dγ (Eα| −E f (γ )|2
γ +i)

We can then plug this into the expression for the S-matrix to find
f (α) f ∗ (β)
Sβα = δ(β − α) − 2πiδ(E α − E β ) 0 .
1 − dγ (Eα| −E f (γ )|2
γ +i)

(c) In order to verify unitarity of the S-matrix, we must show that


∗
dγ Sγβ Sγ α = δ(α − β).

Using our expression for the S-matrix from part (b), we can write this
out as

∗ S
dγ Sγβ γα
⎡

⎢ f ∗ (β) f (γ )
= dγ ⎣δ(γ − β)δ(γ − α) + 2πiδ(E γ − E β )δ(γ − α) 0
dα1 (E | −E
f (α1 )| 2
1−
β α1 −i)

f (α) f ∗ (γ )
− 2πiδ(E γ − E α )δ(γ − β) 0 f (α1 )|
dα1 (E | −E
2
1−
α α1 +i)
⎤
f (α) f ∗ (β)| f (γ )|2
⎥
+4π 2 δ(E γ − E α )δ(E γ − E β )  0  0 ⎦.
| f (α1 )|2 | f (α2 )|2
1 − dα1 (E −E −i) 1 − dα2 (E −E +i)
β α1 α α2

We can use the δ-functions in the first three terms to do the integral over
γ , and so this expression becomes

∗ S
dγ Sγβ γα

= δ(β − α)
⎡ ⎤
⎢ 1 1 ⎥
+ 2πiδ(E α − E β ) f (α) f ∗ (β) ⎣ 0 − 0 ⎦
1− f (α1 )|2
dα1 (E | −E 1− f (α1 )|2
dα1 (E | −E
β α1 −i) α α1 +i)

δ(E γ − E α )δ(E γ − E β )| f (γ )|2
+ 4π 2 f (α) f ∗ (β) dγ  0  0 .
1− f (α1 )|2
dα1 (E | −E −i) 1 − dα 2
| f (α2 )|2
(E −E +i)
β α1 α α2

We can set the energies equal in the terms in square brackets on the sec-
ond line because they are multiplied by δ(E α − E β ), and then these terms
can be combined to give
70 Solutions to Lectures on Quantum Mechanics
⎡ ⎤
⎣ 1 1 ⎦
0 − 0
1− f (α1 )|2
dα1 (Eα| −E α1 −i)
1− f (α1 )|2
dα1 (Eα| −E α1 +i)
0  
−1
dγ | f (γ )|2 E α −E γ +i
+ 1
E α −E γ −i
= 0  0 
1− f (α1 )|2
dα1 (Eα| −E α1 −i)
1 − dα | f (α2 )|2
2 (E α −E α +i)
0   2

dγ | f (γ )| (Eα −Eγ )2 + 2
2 2i

= 0  0 .
f (α1 )|2
1 − dα1 (Eα| −E α −i)
1 − dα2
| f (α2 )|2
(E α −E α +i)
1 2

As discussed in the text above Eq. (8.1.15), for infinitesimal , the func-
tion /(x 2 +  2 ) is negligible away from x = 0, while its integral over
all x is π, and so it can be replaced in any integral by πδ(x). The above
expression then becomes
⎡ ⎤
⎣ 1 1 ⎦
0 | f (α1 )|2
− 0
1 − dα1 (Eα −Eα −i) 1 − dα1 (Eα| −E f (α1 )|2
α1 +i)

1

δ(E α − E γ )| f (γ )|2
= 2πi dγ  0  0 .
f (α1 )|2
1 − dα1 (Eα| −E α −i)
1 − dα | f (α2 )|2
2 (E α −E α +i)
1 2

We therefore see that


∗
dγ Sγβ Sγ α

= δ(β − α)

δ(E γ − E α )δ(E γ − E β )| f (γ )|2
− 4π 2 f (α) f ∗ (β) dγ  0  0 
1 − dα1 (E β| −E
f (α1 )|2
α −i)
1 − dα2
| f (α2 )|2
(E α −E α +i)
1 2

δ(E γ − E α )δ(E γ − E β )| f (γ )|2
+ 4π 2 f (α) f ∗ (β) dγ  0  0 
1 − dα1 (E β| −E
f (α1 )|2
α −i)
1 − dα2 (E α| −E
f (α2 )|2
α +i)
1 2

= δ(β − α),
and so we see that the S-matrix is indeed unitary.
3. The scattering of π + on protons at energies less than a few hundred MeV
is purely elastic, and receives appreciable contributions only from orbital
angular momenta
= 0 and
= 1.
(a) List all the phase shifts that enter in the amplitude for π + -proton scat-
tering at these low energies. (Recall that the spins of the pion and proton are
zero and 1/2, respectively.)
(b) Give a formula for the differential scattering cross-section in terms of
these phase shifts.
 Chapter 8 Problem Set Solutions 71

(a) As discussed in Chapter 4, when adding two angular momenta

and s,
the values taken by the resulting angular momentum J ranges from |
−s|
to
+ s in integer steps. For
= 0 and s = 1/2, the only possibility is
that J = 1/2. For
= 1 and s = 1/2, we may have J = 1/2 or J = 3/2.
If we adopt the notation δ J
s for the phase shifts, then the phase shifts
that contribute to the amplitude for low energy π + -proton scattering are
δ 1 0 1 , δ 1 1 1 , and δ 3 1 1 .
2 2 2 2 2 2

(b) Using Eq. (8.2.14), we can write a formula for the differential cross-
section as
 
pβ
dσ (α → β) = (2π) 2
μα μβ |Mβα |2 d β ,
pα
where μ = E 1 E 2 /c2 (E 1 + E 2 ). If we work in the center of momentum
frame in a coordinate system in which the initial state π + momentum
is in the three-direction, we have an a matrix element like that given in
Eq. (8.4.12)
1 1
Mpπ ,0,−pπ ,σ p ;pπ ,0,−pπ ,σ p =  √
μ |pπ |
  μ|p π|
  
× Y
m ( p̂π )C0 1 (s  σ  ; 0 σ p )Cs 
 (J M; σ  m  )
2
J M
 m  s  σ 

 2
+ 1  
× C0 1 (s σ ; 0 σ p )Cs
(J M; σ 0) S J (E) − 1
 s  ;
s ,

sσ
4π 2

where
 
S J (E)
 s  ;
s = exp(2iδ J
s (E))δ

δs  s .
The relevant Clebsch–Gordan coefficients take the form
C0 1 (s  σ  ; 0 σ p ) = δ 1 s  δσ  σ p
2 2

C0 1 (s σ ; 0 σ p ) = δ 1 s δσ σ p ,
2 2

and so the matrix element simplifies to

1 1
Mpπ ,0,−pπ ,σ p ;pπ ,0,−pπ ,σ p =  √
μ |pπ | μ|pπ |
 
 
× Y
m ( p̂π )C 1
 (J M; σ p m  )
2
J M
 m 
  
2
+ 1
×
C 1
(J M; σ p 0) exp(2iδ J
1 (E)) − 1 .
4π 2 2

When the initial and final proton spin 3-components are both +1/2 this
becomes
72 Solutions to Lectures on Quantum Mechanics
1 1
M p  1 1 = √
π ,0,−pπ ,+ 2 ;pπ ,0,−pπ ,+ 2
μ |pπ | μ|pπ |

0  1 1 1 1 1 1 1  
× Y0 ( p̂π )C 1 0 ( + ; + 0) C 1 0 ( + ; + 0) exp(2iδ 1 0 1 (E)) − 1
2 2 2 2 4π 2 2 2 2 2 2

1 1 1 3 1 1 1  
+ Y01 ( p̂π )C 1 1 ( + ; + 0) C 1 1 ( + ; + 0) exp(2iδ 1 1 1 (E)) − 1
2 2 2 2 4π 2 2 2 2 2 2
  
3 1 1 3 3 1 1
+ Y01 ( p̂π )C 1 1 ( + ; + 0) C 1 1 ( + ; + 0) exp(2iδ 3 1 1 (E)) − 1 .
2 2 2 2 4π 2 2 2 2 2 2

Plugging in the relevant Clebsch–Gordan coefficients and spherical

harmonics, this gives
1 1
Mpπ ,0,−pπ ,+ 1 ;pπ ,0,−pπ ,+ 1 =  √
2 μ |pπ | μ|pπ |
2 
   
1 1
× exp(2iδ 1 0 1 (E)) − 1 + cos θ  exp(2iδ 1 1 1 (E)) − 1
4π 2 2 4π 2 2

  
1
+ cos θ  exp(2iδ 3 1 1 (E)) − 1 ,
2π 2 2

which is the same expression for the case when the initial and final proton
spin are both −1/2
Mpπ ,0,−pπ ,− 1 ;pπ ,0,−pπ ,− 1 = Mpπ ,0,−pπ ,+ 1 ;pπ ,0,−pπ ,+ 1 .
2 2 2 2

For the case when the initial and final proton spins are −1/2 and +1/2,
respectively, we find

1 1
Mp  1 1 √= 
π ,0,−pπ ,+ 2 ;pπ ,0,−pπ ,− 2 μ|pπ |
μ |pπ |

1 1 1 e 1 1 1  
× Y1−1 ( p̂π )C 1 1 ( − ; + − 1) C 1 1 ( − ; − 0) exp(2iδ 1 1 1 (E)) − 1
2 2 2 2 4π 2 2 2 2 2 2
 
3 1 1 3 3 1 1 
+ Y1−1 ( p̂π )C 1 1 ( − ; + − 1) C 1 1 ( − ; − 0) exp(2iδ 3 1 1 (E)) − 1 ,
2 2 2 2 4π 2 2 2 2 2 2

and plugging in the relevant Clebsch–Gordan coefficients and spherical

harmonics gives

1 1 −1 
Mpπ ,0,−pπ ,+ 1 ;pπ ,0,−pπ ,− 1 =  √ sin θ  e−iφ
2 2 μ |pπ | μ|pπ | 4π
  1  
 −iφ 
× exp(2iδ 1 1 1 (E)) − 1 + sin θ e exp(2iδ 3 1 1 (E)) − 1 .
2 2 4π 2 2

The matrix element is identical up to a minus sign for the case when the
initial and final proton spins are +1/2 and −1/2, respectively
 Chapter 8 Problem Set Solutions 73

Mpπ ,0,−pπ ,+ 1 ;pπ ,0,−pπ ,− 1 = −Mpπ ,0,−pπ ,− 1 ;pπ ,0,−pπ ,+ 1 .

2 2 2 2

Finally, we can use these matrix elements to find the differential cross-
sections by using Eq. (8.2.14)
  
1  1 2  1  
dσ σ p = + → σ p = + = exp(2iδ 1 1 (E)) − 1
2 2 |pπ |2  2 2 0 2

1     2
+ cos θ exp(2iδ 1 1 1 (E)) − 1 + cos θ exp(2iδ 3 1 1 (E)) − 1  d 
 
2 2 2 2 2
 
1 1
= dσ σ p = − → σ p = − ,
2 2
and
    
1  1 2  1 
dσ σ p = − → σ p = + = − sin θ exp(2iδ 1 1 (E)) − 1
2 2 |pπ |2  2 2 1 2

1   2
+ sin θ exp(2iδ 3 1 1 (E)) − 1  d 

2 2 2
 
1 1
= dσ σ p = + → σ p = − .
2 2
(1)
4. Let Tβα be the term in Tβα including contributions of all orders in Vs but
only of first order in Vw , where the total interaction is Vs + Vw . Show that
(1)
the complex conjugate of Tβα satisfies a relation of the form

Tβα = dβ  dα  cβαβ  α Tβ(1)
(1)∗
 α .

Give a formula for the coefficients cβαβ  α in terms of S-matrix elements,

including contributions of all orders in Vs but zeroth order in Vw .
We begin by writing the Lippmann–Schwinger equation for the in and out
states in the absence of Vw as in Eq. (8.6.16)
±
sα = α + (E α − H0 ± i)−1 Vs sα
±
,

where α is an eigenstate of H0 with eigenvalue E α . We can then write the

T-matrix as in Eq. (8.6.17)
 −   − 
Tβα = sβ , Vw α+ + sβ , Vs α ,

and we will assume that the process α → β cannot occur in the absence of
Vw , so that the second term vanishes. In this case we can write
 − 
Tβα = sβ , Vw α+ ,
74 Solutions to Lectures on Quantum Mechanics

and to first order in Vw , we can neglect the difference between α+ and sα
+

and so the T-matrix at first order in Vw is given by

(1)  − +

Tβα = sβ , Vw sα .
We can write the complex conjugate as
(1)∗  + −
  + −

Tβα = Vw sβ , sα = sβ , Vw sα ,
+ −
where we have used the fact that Vw is Hermitian. Now, because sα and sα
each form a complete set of states, we may expand each in terms of the other
   
− − − (0)
sα = dβ sβ  sβ  , sα = dβ  sβ
+  +
 Sβ  α

 
− +
 + −
 − (0)∗
sβ = dα  sα  sα  , sβ = dα  sα  Sβα  ,

(0)
where Sαβ is the S-matrix evaluated to all orders in Vs , but to zeroth order in
Vw . Using these expressions in our formula above we find
  
(1)∗ − (0) (0)∗ −
Tβα = dβ  dα  sβ S
 βα , V S
w βα   sα 
  
= dβ  dα  Sβ(0)∗ (0)∗
 α Sβα 
−
sβ −
 , Vw sα 


= dβ  dα  Sβ(0)∗ (0)∗ (1)∗
 α Sβα  Tβ  α 


= dβ  dα  cβαβ  α Tβ(1)∗
 α ,

where cβαβ  α = Sβ(0)∗ (0)∗

 α Sβα  .

5. By direct calculation, show that the terms of first and second order in
the interaction in time-dependent perturbation theory give the same results
for the S-matrix as the first- and second-order terms in old-fashioned
perturbation theory.
In old-fashioned perturbation theory, the S-matrix is given up to second order
in V by plugging Eq. (8.6.1) into Eq. (8.1.10)
Sβα = δ(β − α) − 2πiδ(E α − E β )Tβα
 
= δ(β − α) − 2πiδ(E α − E β ) β , V α+
= δ(β − α) − 2πiδ(E α − E β )
    
× β , V + −1
α + β , V (E α − H0 + i) V α + · · · ,

where the dots represent terms of higher order in V . In time-dependent

perturbation theory, the S-matrix is given by Eq. (8.7.4), which we can write
up to second order as
 Chapter 8 Problem Set Solutions 75
    2  +∞
i +∞ −i
Sβα = β , 1 − dτ VI (τ ) + dτ1
 −∞  −∞
 τ1  
× dτ2 VI (τ1 )VI (τ2 ) + · · · α ,
−∞

where VI (τ ) is defined in Eq. (8.7.8)

VI (t) ≡ ei H0 t/ V e−i H0 t/.

The term of zeroth order in V is simply given by

 
β , α = δ(β − α).

The term of first order in V is

     
i +∞ i +∞
β , − dτ VI (τ )α = − dτ β , ei H0 τ/ V e−i H0 τ/α
 −∞  −∞

i +∞  
=− dτ ei Eβ τ/e−i Eα τ/ β , V α
 −∞
 
= −2πiδ(E β − E α ) β , V α ,

where in the second line we have used the fact that α is an eigenstate of H0
with eigenvalue E α . The term which is second order in V is given by
     τ1 
−i 2 +∞
β , dτ1 dτ2 VI (τ1 )VI (τ2 )α
 −∞ −∞
  τ1  
1 +∞
=− 2 dτ1 dτ2 β , ei H0 τ1 / V e−i H0 τ1 /ei H0 τ2 / V e−i H0 τ2 /α
 −∞ −∞
 +∞  τ1  
1
=− 2 dτ1 dτ2 ei Eβ τ1 /e−i Eα τ2 / β , V e−i H0 τ1 /ei H0 τ2 / V α
 −∞ −∞
   τ1 
1 +∞ i E β τ1 / −i H0 τ1 / −i(E α −H0 )τ2 /
=− 2 dτ1 e β , V e dτ2 e V α .
 −∞ −∞

Now, we can do the integral over τ2 by noticing that

 τ1  τ1
−i(E α −H0 )τ2 /
dτ2 e = lim+ dτ2 e−i(Eα −H0 +i)τ2 /
−∞ →0 −∞

= (E α − H0 + i)−1 e−i(Eα −H0 )τ1 /,
−i
76 Solutions to Lectures on Quantum Mechanics

where  is taken to be a positive infinitesimal in the second line. Using this

result in the expression above, we find
  2  +∞  τ1 
−i
β , dτ1 dτ2 VI (τ1 )VI (τ2 )α
 −∞ −∞

1 +∞
=− 2 dτ1 ei Eβ τ1 /
 −∞
 

× β , V e−i H0 τ1 / (E α − H0 + i)−1 e−i(Eα −H0 )τ1 / V α
−i
 +∞
i  
=− dτ1 ei(Eβ −Eα )τ1 / β , (E α − H0 + i)−1 V α
 −∞
 
= −2πiδ(E β − E α ) β , V (E α − H0 + i)−1 α .
Putting these results together, we find for S-matrix calculated to second order
in V in time-dependent perturbation theory
 
Sβα =δ(β − α) − 2πiδ(E β − E α ) β , V α
 
− 2πiδ(E β − E α ) β , V (E α − H0 + i)−1 V α + · · · ,
which matches the result we found for old-fashioned perturbation theory.
6. Assume isospin conservation, and suppose that the only appreciable phase
shift in the scattering of pions on nucleons is the one with quantum numbers
J = 3/2,
= 1, and T = 3/2. Calculate the differential cross-sections for
the reactions π + + p → π + + p, π + + n → π + + n, π + + n → π 0 + p,
and π − + n → π − + n in terms of this phase shift.
The differential cross-section of each process will be very similar to that
worked out in Problem 3 above, but we will need to work with phase shifts
for states with definite J ,
, s, and T , whcih we will deonte δ J
s T . For
pion–nucleon scattering, assuming the conservation of isospin, Eq. (8.4.12)
is modified to
1 1
Mpπ ,0,T3 π ,−pπ ,σ N ,T3 N ;pπ ,0,T3 π ,−pπ ,σ N ,T3 N =  √
μ |pπ | μ|pπ |
  
× Y
m ( p̂π )C0 1 (s  σ  ; 0 σ p )Cs 
 (J M; σ  m  )
2
J M
 m  s  σ 

 2
+ 1
× C0 1 (s σ ; 0 σ p )Cs
(J M; σ 0)

sσ
4π 2

  
× C1 1 (T T3 ; T3 π T3 N )C1 1 (T T3 ; T3 π T3 N ) S J T (E) − 1
 s  ;
s ,
2 2
T

where
 
S J T (E)
 s  ;
s = exp(2iδ J
s T (E))δ

δs  s .
 Chapter 8 Problem Set Solutions 77

In the case when only δ 3 1 1 3 is appreciable, this becomes

2 2 2

1 1
Mpπ ,0,T   1  1 = √
3 π ,−pπ ,+ 2 ,T3 N ;pπ ,0,T3 π ,−pπ ,+ 2 ,T3 N μ |pπ | μ|pπ |
 
   
1 3
× cos θ  exp(2iδ 3 1 1 3 (E)) − 1 C1 1 T3 ; T3 π T3 N
2π 2 2 2 2 2
 
3
× C1 1 T3 ; T3 π T3 N
2 2
= Mpπ ,0,T  ,−pπ ,− 1 ,T  ;pπ ,0,T3 π ,−pπ ,− 1 ,T3 N ,
3π 2 3N 2

when the initial and final nucleon spin 3-components are equal and

1 1
Mpπ ,0,T   1  1 = √
3 π ,−pπ ,+ 2 ,T3 N ;pπ ,0,T3 π ,−pπ ,− 2 ,T3 N μ |pπ | μ|pπ |
   
1  −iφ  3  
× sin θ e exp(2iδ 3 1 1 3 (E)) − 1 C1 1 T3 ; T3 π T3 N
4π 2 2 2 2 2
 
3
× C1 1 T3 ; T3 π T3 N
2 2
= Mpπ ,0,T  ,−pπ ,− 1 ,T  ;pπ ,0,T3 π ,−pπ ,+ 1 ,T3 N ,
3π 2 3N 2

when the initial and final nucleon spin 3-components are opposite. We can
then write the differential cross-sections in this case by using Eq. (8.2.14)
which gives
   2
  2 
 
dσ π N → π N ; σ N = σ N =
cos θ exp(2iδ 3 1 1 3 (E)) − 1 

|pπ |2  2 2 2

   2
3 3
× C1 1 T3 ; T3 π T3 N C1 1 T3 ; T3 π T3 N d  ,
2 2 2 2

when the initial and final nucleon spins are equal, and
   2
 2  
dσ π N → π N ; σ N =  
= sin θ exp(2iδ 3 1 1 3 (E)) − 1 
−σ N

4|pπ |2  2 2 2

   2
3 3
× C1 1 T3 ; T3 π T3 N C1 1 T3 ; T3 π T3 N d  ,
2 2 2 2

when the initial and final nucleon spins are opposite. The isospin quantum
numbers for each of the relevant particles are given by the following table.
78 Solutions to Lectures on Quantum Mechanics

T T3
π+ 1 1
π0 1 0
π− 1 −1
p 1/2 1/2
n 1/2 −1/2
The differential cross-section for the process π + + p → π + + p is then
given by
  2 cos2 θ  2
dσ π + + p → π + + p; σ p = σ p = sin δ 3 1 1 3 (E) d 
|pπ |2 2 2 2
 

≡ dσ σ N = σ N ,
when the initial and final proton spins are equal, and
  2 sin2 θ  2
dσ π + + p → π + + p; σ p = −σ p = sin δ 3 1 1 3 (E) d 
4|pπ |2 2 2 2
 

≡ dσ σ N = −σ N ,
when the initial and final proton spins are opposite. The differential cross-
sections for each of the pion–nucleon processes are then just multiples of
these expressions, given by the appropriate Clebsch–Gordan coefficients
associated with isospin
  1  
dσ π + + n → π + + n; σn = σn = dσ σ N = σ N ,
9
 +  2  
dσ π + n → π + p; σn = σ p = dσ σ N = σ N ,
0 

 −  9  
dσ π + n → π + n; σn = σn = dσ σ N = σ N ,
− 

and similarly for the processes where the initial and final nucleon spin are
opposite
  1  
dσ π + + n → π + + n; σn = −σn = dσ σ N = −σ N ,
9
 +  2  
dσ π + n → π 0 + p; σn = −σ p = dσ σ N = −σ N ,
  9  
dσ π − + n → π − + n; σn = −σn = dσ σ N = −σ N .
 Chapter 9 Problem Set Solutions 79

Chapter 9 Problem Set Solutions

1. Consider the theory of a single particle with Lagrangian

m 2
L= ẋ + ẋ · f(x) − V (x),
2
where f(x) and V (x) are arbitrary vector and scalar functions of position.
• Find the equation of motion satisfied by x.
• Find the Hamiltonian, as a function of x and its canonical conjugate p.
• What is the Schrödinger equation satisfied by the coordinate-space wave
function ψ(x, t)?
The equations of motion may be derived from Eq. (9.1.3), which in this case
reads
∂ L (x(t), ẋ(t)) d ∂ L (x(t), ẋ(t))
= .
∂ xi dt ∂ ẋi
Taking the appropriate derivatives of L, the equations of motion are

ẋ × (∇ × f(x)) + (ẋ · ∇) f(x) − ∇V (x) = m ẍ + (ẋ · ∇) f(x),

which can be simplified to read

ẋ × (∇ × f(x)) − ∇V (x) = m ẍ.

Notice that this is very similar to the equation of motion of a charged particle
in an external magnetic field, given by Eq. (10.1.1) with vanishing electric
field.
The canonical conjugate of x is defined in Eq. (9.3.3) as
∂L
pi = = m ẋi + f i (x),
∂ ẋi
and the Hamiltonian is given in Eq. (9.3.4) as

H = ẋ(x, p) · p − L (x, ẋ(x, p)) ,

which in this case gives

1 1 1
H= (p − f(x)) · p − (p − f(x))2 − (p − f(x)) · f(x) + V (x)
m 2m m
1
= (p − f(x)) + V (x).
2
2m
Notice that this is very similar to the Hamiltonian for a charged particle in an
external magnetic field, given by Eq. (10.1.9) with vanishing scalar potential.
80 Solutions to Lectures on Quantum Mechanics

The Schrödinger equation satisfied by the coordinate-space wave function

ψ(x, t) is given by
∂
H ψ(x, t) = iψ(x, t),
∂t
where the operator p acts on the coordinate-space wave function as −i∇.
The Schrödinger equation therefore takes the form
∂ 2 2 i
i ψ(x, t) = − ∇ ψ(x, t) + f(x) · ∇ψ(x, t)
∂t 2m m
i f2 (x)
+ (∇ · f(x)) ψ(x, t) + ψ(x, t) + V (x)ψ(x, t).
2m 2m
2. Show that Poisson brackets and Dirac brackets both satisfy the Jacobi
identity.
The Poisson bracket is defined in Eq. (9.4.19)
   ∂ f ( p, q) ∂g( p, q) ∂g( p, q) ∂ f ( p, q) 
f (q, p), g(q, p) P ≡ − ,
N
∂q N ∂ pN ∂q N ∂ pN
which we will write in a slightly more compact notation as
   
f, g P = ∂q N f ∂ p N g − ∂q N g∂ p N f .
N

We wish to show by direct calculation that the Poisson bracket satisfies the
Jacobi identity, which is given in Eq. (9.4.22)
           
f, g, h P P + g, h, f P P + h, f, g P P = 0.
Let us first work out just one of the terms appearing in the Jacobi identity
 
     
f, g, h P P = f, ∂q N g∂ p N h − ∂q N h∂ p N g
N P
  
= ∂q M f ∂ p M ∂q N g∂ p N h − ∂q N h∂ p N g
MN

 
− ∂q M ∂q N g∂ p N h − ∂q N h∂ p N g ∂ p M f

= ∂q M f ∂ p M ∂q N g∂ p N h + ∂q M f ∂q N g∂ p M ∂ p N h
MN
− ∂q M f ∂ p M ∂q N h∂ p N g − ∂q M f ∂q N h∂ p M ∂ p N g
− ∂q M ∂q N g∂ p N h∂ p M f − ∂q N g∂q M ∂ p N h∂ p M f

− ∂q M ∂q N h∂ p N g∂ p M f − ∂q N h∂q M ∂ p N g∂ p M f .
 Chapter 9 Problem Set Solutions 81

Now, putting this together to calculate the combination appearing in the

Jacobi identity, we find
           
f, g, h P P + g, h, f P P + h, f, g P P
 
= ∂q M f ∂ p M ∂q N g∂ p N h + ∂q M f ∂q N g∂ p M ∂ p N h
MN
− ∂q M f ∂ p M ∂q N h∂ p N g − ∂q M f ∂q N h∂ p M ∂ p N g
− ∂q M ∂q N g∂ p N h∂ p M f − ∂q N g∂q M ∂ p N h∂ p M f

− ∂q M ∂q N h∂ p N g∂ p M f − ∂q N h∂q M ∂ p N g∂ p M f

+ ∂q M g∂ p M ∂q N h∂ p N f + ∂q M g∂q N h∂ p M ∂ p N f

− ∂q M g∂ p M ∂q N f ∂ p N h − ∂q M g∂q N f ∂ p M ∂ p N h
− ∂q M ∂q N h∂ p N f ∂ p M g − ∂q N h∂q M ∂ p N f ∂ p M g

− ∂q M ∂q N f ∂ p N h∂ p M g − ∂q N f ∂q M ∂ p N h∂ p M g

+ ∂q M h∂ p M ∂q N f ∂ p N g + ∂q M h∂q N f ∂ p M ∂ p N g

− ∂q M h∂ p M ∂q N g∂ p N f − ∂q M h∂q N g∂ p M ∂ p N f
− ∂q M ∂q N f ∂ p N g∂ p M h − ∂q N f ∂q M ∂ p N g∂ p M h

− ∂q M ∂q N g∂ p N f ∂ p M h − ∂q N g∂q M ∂ p N f ∂ p M h .

Now, due to the fact that partial derivatives commute and also that we can
relabel the summation variables, we can begin canceling terms. The pairs of
terms which cancel are the 1st and 22nd, 2nd and 12th, 3rd and 16th, 4th and
18th, 5th and 23rd, 6th and 9th, 7th and 13th, 8th and 19th, 10th and 20th,
11th and 24th, 14th and 17th, and finally the 15th and 21st. We see that all
terms in the sum cancel, and we have
           
f, g, h P P + g, h, f P P + h, f, g P P = 0,
and so the Poisson bracket satisfies the Jacobi identity.
The Dirac bracket is defined in Eq. (9.5.16) as

[ f, g] D ≡ [ f, g] P − [ f, χr ] P Cr−1
s [χs , g] P ,
rs

where the matrix Cr s is defined in Eq. (9.5.12)

Cr s ≡ [χr , χs ] P .
We will now work out [ f, [g, h] D ] D in terms of Poisson brackets
82 Solutions to Lectures on Quantum Mechanics
 

[ f, [g, h] D ] D = [ f, [g, h] P ] D − f, [g, χr ] P Cr−1
s [χs , h] P
rs D
 
 
= [ f, [g, h] P ] P − [ f, χr ] P Cr−1
s [χs , [g, h] P ] P − f, [g, χr ] P Cr−1
s [χs , h] P
rs rs P
 
 
−1
+ [ f, χt ] P Ctu χu , [g, χr ] P Cr−1
s [χs , h] P
tu rs P
 
= [ f, [g, h] P ] P − [ f, χr ] P Cr−1
s [χs , [g, h] P ] P − [ f, [g, χr ] P ] P Cr−1
s [χs , h] P
rs rs
     
− [g, χr ] P f, Cr−1
s [χs , h] P − [g, χr ] P Cr−1
s f, [χs , h] P P
P
rs rs

−1
+ [ f, χt ] P Ctu [χu , [g, χr ] P ] P Cr−1
s [χs , h] P
tu rs
  
−1
+ [ f, χt ] P Ctu [g, χr ] P χu , Cr−1
s [χs , h] P
P
tu rs

−1
+ [ f, χt ] P Ctu [g, χr ] P Cr−1
s [χu , [χs , h] P ] P .
tu rs

Using the Jacobi identity for Poisson brackets, which we proved above, we
can rewrite the 6th term in this last expression as

−1
[ f, χt ] P Ctu [χu , [g, χr ] P ] P Cr−1
s [χs , h] P
tu rs

−1
=− [ f, χt ] P Ctu [g, [χr , χu ] P ] P Cr−1
s [χs , h] P
tu rs

−1
− [ f, χt ] P Ctu [χr , [χu , g] P ] P Cr−1
s [χs , h] P .
tu rs

Now we can use the definition of Cr u and the fact that Cr u = −Cur to
simplify the 1st term
      
−1 −1 −1
− Ctu g, Cr u P = − g, Ctu Cr u P + g, Ctu C
P ru
u u u
   
−1
= − g, (−δtr ) P + g, Ctu C
P ru
u
 
−1
= g, Ctu P
Cr u .
u

Using this in the expression above gives

   
−1
[ f, χt ] P Ctu [χu , [g, χr ] P ] P Cr−1
s [χ s , h] P = − [ f, χ ]
r P g, C −1
rs [χs , h] P
P
tu rs rs

−1
− [ f, χt ] P Ctu [χr , [χu , g] P ] P Cr−1
s [χs , h] P
tu rs
 Chapter 9 Problem Set Solutions 83
  
= [h, χr ] P g, Cr−1
s [χs , f ] P
P
rs

−1
− [ f, χt ] P Ctu [χr , [χu , g] P ] P Cr−1
s [χs , h] P ,
tu rs

where in the last equality, we have used the antisymmetry of Cr−1 s and also
the antisymmetry of the Poisson bracket to rearrange the first term on the
right-hand side. Returning now to [ f, [g, h] D ] D , we have

[ f, [g, h] D ] D = [ f, [g, h] P ] P − [ f, χr ] P Cr−1
s [χs , [g, h] P ] P
rs

− [ f, [g, χr ] P ] P Cr−1
s [χs , h] P
rs
     
− [g, χr ] P f, Cr−1
s P [χs , h] P − [g, χr ] P Cr−1
s f, [χs , h] P P
rs rs
  
+ [h, χr ] P g, Cr−1
s P [χs , f ] P
rs

−1
− [ f, χt ] P Ctu [χr , [χu , g] P ] P Cr−1
s [χs , h] P
tu rs
  
−1
+ [ f, χt ] P Ctu [g, χr ] P χu , Cr−1
s P [χs , h] P
tu rs

−1
+ [ f, χt ] P Ctu [g, χr ] P Cr−1
s [χu , [χs , h] P ] P .
tu rs

Now, the combination appearing in the Jacobi identity reads

[ f, [g, h] D ] D + [g, [h, f ] D ] D + [h, [ f, g] D ] D (1)


= [ f, [g, h] P ] P − [ f, χr ] P Cr−1
s [χs , [g, h] P ] P
rs

− [ f, [g, χr ] P ] P Cr−1
s [χs , h] P
rs
     
− [g, χr ] P f, Cr−1
s P [χs , h] P − [g, χr ] P Cr−1
s f, [χs , h] P P
rs rs
  
+ [h, χr ] P g, Cr−1
s P [χs , f ] P
rs

−1
− [ f, χt ] P Ctu [χr , [χu , g] P ] P Cr−1
s [χs , h] P
tu rs
  
−1
+ [ f, χt ] P Ctu [g, χr ] P χu , Cr−1
s P [χs , h] P
tu rs

−1
+ [ f, χt ] P Ctu [g, χr ] P Cr−1
s [χu , [χs , h] P ] P
tu rs
84 Solutions to Lectures on Quantum Mechanics

+ [g, [h, f ] P ] P − [g, χr ] P Cr−1
s [χs , [h, f ] P ] P
rs

− [g, [h, χr ] P ] P Cr−1
s [χs , f ] P
rs
     
− [h, χr ] P g, Cr−1
s P [χ s , f ] P − [h, χr ] P Cr−1
s g, [χs , f ] P P
rs rs
  
+ [ f, χr ] P h, Cr−1
s P [χs , g] P
rs

−1
− [g, χt ] P Ctu [χr , [χu , h] P ] P Cr−1
s [χs , f ] P
tu rs
  
−1
+ [g, χt ] P Ctu [h, χr ] P χu , Cr−1
s P [χs , f ] P
tu rs
  
−1
+ [g, χt ] P Ctu [h, χr ] P Cr−1
s χu , [χs , f ] P P
tu rs

+ [h, [ f, g] P ] P − [h, χr ] P Cr−1
s [χs , [ f, g] P ] P
rs

− [h, [ f, χr ] P ] P Cr−1
s [χs , g] P
rs
     
− [ f, χr ] P h, Cr−1
s P [χs , g] P − [ f, χr ] P Cr−1
s h, [χs , g] P P
rs rs
  
+ [g, χr ] P f, Cr−1
s P [χs , h] P
rs

−1
− [h, χt ] P Ctu [χr , [χu , f ] P ] P Cr−1
s [χs , g] P
tu rs
  
−1
+ [h, χt ] P Ctu [ f, χr ] P χu , Cr−1
s P [χs , g] P
tu rs
  
−1
+ [h, χt ] P Ctu [ f, χr ] P Cr−1
s χu , [χs , g] P P .
tu rs

We can see immediately that the 1st, 10th, and 19th terms sum to zero
because the Poisson brackets obey the Jacobi identity. The 12th term in
Eq. (1) can be rewritten as
 
− [g, [h, χr ] P ] P Cr−1
s [χ s , f ] P = − −1
[χr , f ] P Csr [g, [h, χs ] P ] P
rs rs

−1
=+ [ f, χr ] P Csr [g, [h, χs ] P ] P
rs

=− [ f, χr ] P Cr−1
s [g, [h, χs ] P ] P ,
rs
 Chapter 9 Problem Set Solutions 85

and after this rearrangement, we can see that the 2nd, 12th, and 23rd terms
of Eq. (1) sum to zero because the Poisson brackets obey the Jacobi identity.
Likewise, we can do a similar rearrangement of the 3rd and 21st terms in Eq.
(1), and then we see that the 3rd, 13th, and 20th terms of Eq. (1) sum to zero,
and also the 5th, 11th, and 21st terms of Eq. (1) sum to zero. We can also see
that the 4th and 24th terms of Eq. (1) cancel, as well as the 6th and 13th, and
the 15th and 22nd. At this point, we are left with

[ f, [g, h] D ] D + [g, [h, f ] D ] D + [h, [ f, g] D ] D (2)


−1
=− [ f, χt ] P Ctu [χr , [χu , g] P ] P Cr−1
s [χs , h] P
tu rs
  
−1
+ [ f, χt ] P Ctu [g, χr ] P χu , Cr−1
s P [χs , h] P
tu rs

−1
+ [ f, χt ] P Ctu [g, χr ] P Cr−1
s [χu , [χs , h] P ] P
tu rs

−1
− [g, χt ] P Ctu [χr , [χu , h] P ] P Cr−1
s [χs , f ] P
tu rs
  
−1
+ [g, χt ] P Ctu [h, χr ] P χu , Cr−1
s P [χs , f ] P
tu rs
  
−1
+ [g, χt ] P Ctu [h, χr ] P Cr−1
s χu , [χs , f ] P P
tu rs

−1
− [h, χt ] P Ctu [χr , [χu , f ] P ] P Cr−1
s [χs , g] P
tu rs
  
−1
+ [h, χt ] P Ctu [ f, χr ] P χu , Cr−1
s P [χs , g] P
tu rs
  
−1
+ [h, χt ] P Ctu [ f, χr ] P Cr−1
s χu , [χs , g] P P .
tu rs

We can rewrite the 1st term of Eq. (2) as


−1
− [ f, χt ] P Ctu [χr , [χu , g] P ] P Cr−1
s [χs , h] P
tu rs

=− [ f, χr ] P Cr−1 −1
s [χu , [χs , g] P ] P C ut [χt , h] P
tu rs

−1
=− [h, χt ] P Ctu [ f, χr ] P Cr−1
s [χu , [χs , g] P ] P ,
tu rs

and so we can see that the 1st and 9th terms of Eq. (2) cancel. After a sim-
ilar rewriting of the 4th and 7th terms, we find that the 3rd and 4th terms
of Eq. (2) cancel, as well as the 6th and 7th. This leaves only three terms
remaining
86 Solutions to Lectures on Quantum Mechanics

[ f, [g, h] D ] D + [g, [h, f ] D ] D + [h, [ f, g] D ] D (3)

  
−1
= [ f, χt ] P Ctu [g, χr ] P χu , Cr−1
s P [χs , h] P
tu rs
  
−1
+ [g, χt ] P Ctu [h, χr ] P χu , Cr−1
s P [χs , f ] P
tu rs
  
−1
+ [h, χt ] P Ctu [ f, χr ] P χu , Cr−1
s P [χs , g] P .
tu rs

Next, notice that

     
χu , Cr−1
s = χ u , δr v C −1
vs = χ u , C −1
rw C wv C −1
vs
P P P
v vw
  
= χu , Cr−1
w C wv C −1
vs + Cr−1 −1
w [χu , C wv ] P C vs
P
vw vw
  
+ Cr−1
w C wv
−1
χu , Cvs
P
vw
     
= χu , Cr−1
w δ ws + C −1
rw [χ u , C wv ] P C −1
vs + δr v χ u , C −1
vs
P P
w vw v
  
= 2 χu , Cr−1
s + Cr−1 −1
w [χu , C wv ] P C vs
P
vw
     
χu , Cr−1
s = − C −1
rw [χ u , C wv ] P C −1
vs = − Cr−1 −1
w χu , [χw , χv ] P P C vs .
P
vw vw

Using this identity in each of the terms appearing in Eq. (3), we can rewrite
this expression as
[ f, [g, h] D ] D + [g, [h, f ] D ] D + [h, [ f, g] D ] D
  
−1 −1 −1
=− [ f, χt ] P [g, χr ] P [χs , h] P Ctu Cr w Cvs χu , [χw , χv ] P P
tu rs vw
  
−1 −1 −1
− [g, χt ] P [h, χr ] P [χs , f ] P Ctu Cr w Cvs χu , [χw , χv ] P P
tu rs vw
  
−1 −1 −1
− [h, χt ] P [ f, χr ] P [χs , g] P Ctu Cr w Cvs χu , [χw , χv ] P P ,
tu rs vw

using the antisymmetry of the Poisson bracket this becomes

[ f, [g, h] D ] D + [g, [h, f ] D ] D + [h, [ f, g] D ] D
  
−1 −1 −1
= [ f, χt ] P [g, χr ] P [h, χs ] P Ctu Cr w Cvs χu , [χw , χv ] P P
tu rs vw
  
−1 −1 −1
+ [g, χt ] P [h, χr ] P [ f, χs ] P Ctu Cr w Cvs χu , [χw , χv ] P P
tu rs vw
  
−1 −1 −1
+ [h, χt ] P [ f, χr ] P [g, χs ] P Ctu Cr w Cvs χu , [χw , χv ] P P ,
tu rs vw
 Chapter 9 Problem Set Solutions 87

after relabeling summation indices this becomes

[ f, [g, h] D ] D + [g, [h, f ] D ] D + [h, [ f, g] D ] D
  
−1 −1 −1
= [ f, χt ] P [g, χr ] P [h, χs ] P Ctu Cr w Cvs χu , [χw , χv ] P P
tu rs vw
  
+ [g, χr ] P [h, χs ] P [ f, χt ] P Cr−1 −1 −1
u C sw C vt χu , [χw , χv ] P P
tu rs vw
  
−1 −1 −1
+ [h, χs ] P [ f, χt ] P [g, χr ] P Csu Ctw Cvr χu , [χw , χv ] P P ,
tu rs vw

and after some more relabeling this gives

[ f, [g, h] D ] D + [g, [h, f ] D ] D + [h, [ f, g] D ] D
  
−1 −1 −1
= [ f, χt ] P [g, χr ] P [h, χs ] P Ctu Cr w Cvs χu , [χw , χv ] P P
tu rs vw
  
+ [g, χr ] P [h, χs ] P [ f, χt ] P Cr−1 −1 −1
w C sv C ut χw , [χv , χu ] P P
tu rs vw
  
−1 −1 −1
+ [h, χs ] P [ f, χt ] P [g, χr ] P Csv Ctu Cwr χv , [χu , χw ] P P ,
tu rs vw

and we can use the antisymmetry of Cr−1

s to express this as

[ f, [g, h] D ] D + [g, [h, f ] D ] D + [h, [ f, g] D ] D


−1 −1 −1
= [ f, χt ] P [g, χr ] P [h, χs ] P Ctu Cr w Cvs
tu rs vw
      
× χu , [χw , χv ] P P + χw , [χv , χu ] P P + χv , [χu , χw ] P P
= 0,
which vanishes because the Poisson brackets obey the Jacobi identity. We
have thus shown that the Dirac brackets obey the Jacobi identity.
3. Consider a one-dimensional harmonic oscillator, with Hamiltonian
p2 mω2 x 2
H= + .
2m 2
Use the path-integral formalism to calculate the probability amplitude for a
transition from a position x at time t to a position x  at time t  > t.
The Hamiltonian of the harmonic oscillator is
p2 mω2 x 2
H= + ,
2m 2
and since this Hamiltonian is quadratic in the momentum, we can write the
path integral in the form given in Eq. (9.6.20)
88 Solutions to Lectures on Quantum Mechanics
 
x  ,t  , x,t
     2 
1 i t m ẋ (τ ) mω2 x 2 (τ )
=C d x(τ ) exp dτ − ,
x(t)=x; x(t  )=x  τ
 t 2 2
where we have used the fact that the Lagrangian of the harmonic oscillator is
m ẋ 2 (τ ) mω2 x 2 (τ )
L (x(τ ), ẋ(τ )) =− .
2 2
The Lagrangian is quadratic in x(τ ) and ẋ(τ ), and so we can do the integral
by setting x(τ ) equal to the classical solution, which is given by
d ∂ L(τ ) ∂ L(τ )
= ,
dτ ∂ ẋ(τ ) ∂ x(τ )
with the constraint that x(t) = x and x(t  ) = x  . The equation of motion
takes the form
m ẍ(τ ) = −mω2 x(τ ),
which has a well-known solution of the form
x(τ ) = Aeiωτ + Be−iωτ .
Imposing the conditions x(t) = x and x(t  ) = x  we find
x = x(t) = Aeiωt + Be−iωt
 
x  = x(t  ) = Aeiωt + Be−iωt ,
and so A and B are given by

xeiωt − x  eiωt
A=
e2iωt − e2iωt 

xe−iωt − x  e−iωt
B = −2iωt .
e − e−2iωt 
We can then calculate ẋ(τ ) to be
ẋ(τ ) = iω Aeiωτ − iωBe−iωτ .
Plugging in the classical solution, we can evaluate the integral appearing in
the exponential of the path integral
 t  t  2 
m ẋ (τ ) mω2 x 2 (τ )
dτ L(τ ) = dτ −
t t 2 2
 t
m  2 2 2iωτ 
= dτ −ω A e + 2ω2 AB − ω2 B 2 e−iωτ
t 2
2  
mω 
− 2 2iωτ
A e + 2AB + B e 2 2iωτ
2
 Chapter 9 Problem Set Solutions 89
 t  
= dτ −mω2 A2 e2iωτ − mω2 B 2 e−2iωτ
t
imω A2  2iωt   imωB 2  

= e −e 2iωt
− e−2iωt − e−2iωt .
2 2
Plugging in the expressions for A and B, this becomes
 t  
imω x 2 e2iωt − 2x x  eiω(t+t ) + x 2 e2iωt
dτ L(τ ) =
t 2 e2iωt  − e2iωt
 
x 2 e−2iωt − 2x x  e−iω(t+t ) + x 2 e−2iωt
−
e−2iωt  − e−2iωt
 
imω 1  
2iω(t−t  ) −2iω(t−t  )
=     x 2
e − e
2 e2iωt  − e2iωt e−2iωt  − e−2iωt
  
   

+ x 2 e2iω(t−t ) − e−2iω(t−t ) − 2x x  2eiω(t−t ) − 2e−iω(t−t )

imω 2i(x 2 + x 2 ) sin(2ω(t − t  )) − 8i x x  sin(ω(t − t  ))
=
2 2 − 2 cos(2ω(t − t  ))

mω 2(x + x ) sin(ω(t − t  )) cos(ω(t − t  )) − 4x x  sin(ω(t − t  ))
2 2
=−
2 2 sin2 (ω(t − t  ))
2 
mω (x + x 2 ) cos(ω(t − t  )) − 2x x 
=− .
2 sin(ω(t − t  ))
The expression for the path integral then becomes

  imω (x 2 + x 2 ) cos(ω(t  − t)) − 2x x 
x  ,t  , x,t = C exp ,
2 sin(ω(t  − t))
where C is a proportionality constant which depends on t  − t. We can find
C by requiring that
  
lim

 x  ,t  ,  x,t → δ(x − x),
t →t

or in other words that

 +∞ 
imω (x 2 + x 2 ) cos(ω(t  − t)) − 2x x 
lim d x C exp f (x) → f (x  ).
t  →t −∞ 2 sin(ω(t  − t))
In the limit t  → t, the exponential function oscillates very rapidly, and so
the only appreciable contribution to the integral come from the value x = x  .
It only remains to show that
 +∞ 
imω (x 2 + x 2 ) cos(ω(t  − t)) − 2x x 
lim d x C exp → 1.
t  →t −∞ 2 sin(ω(t  − t))
90 Solutions to Lectures on Quantum Mechanics

We can do this integral by using Eqs. (9.6.17) and (9.6.18) which give
 +∞ 
imω (x 2 + x 2 ) cos(ω(t  − t)) − 2x x 
lim d x C exp
t  →t −∞ 2 sin(ω(t  − t))
   
2iπ sin(ω(t  − t)) imω sin(ω(t  − t))
= lim C exp − x 2 .
t  →t mω cos(ω(t  − t)) 2 cos(ω(t  − t))
In order for this expression to approach 1 in the limit t  → t, its absolute
value square must also approach 1 in this limit, and so we see that up to an
arbitrary overall phase

mω cos(ω(t  − t))
C= .
2iπ sin(ω(t  − t))
We therefore find the probability amplitude for the transition from a position
x at t to a position x  at t  to be
 
x  ,t  , x,t
 
mω cos(ω(t  − t)) imω (x 2 + x 2 ) cos(ω(t  − t)) − 2x x 
= exp .
2iπ sin(ω(t  − t)) 2 sin(ω(t  − t))
 Chapter 10 Problem Set Solutions 91

Chapter 10 Problem Set Solutions

1. Consider a system in an external electromagnetic field. Suppose that the part

of the Lagrangian that depends on the scalar potential φ and vector potential
A takes the form

L int (t) = d 3 x [−ρ(x, t)φ(x, t) + J(x, t) · A(x, t)] ,

where ρ and J depend on the matter variables but not on φ or A. What

condition must be satisfied by ρ and J for the action to be gauge-invariant?
We will work with an interaction Lagrangian density of the form

L int = d 3 x [−ρ(x, t)φ(x, t) + J(x, t) · A(x, t)],

and we wish to find the the conditions such that the action is invariant under
a gauge transformation, defined in Eqs. (10.2.1) and (10.2.2) as
A(x, t) → A (x, t) = A(x, t) + ∇α(x, t)
1 ∂
φ(x, t) → φ  (x, t) = φ(x, t) − α(x, t).
c ∂t
Under this transformation, the interaction part of the action becomes
 ∞
Sint = dt L int
−∞
∞   
1 ∂
→ dt d 3 x − ρ(x, t) φ(x, t) − α(x, t) + J(x, t) · (A(x, t) + ∇α(x, t))
−∞ c ∂t
 ∞    ∞ 
∂ 1 1 ∂ρ(x, t)
= Sint + dt d 3 x ρ(x, t)α(x, t) − dt d3x α(x, t)
−∞ ∂t c −∞ c ∂t
 ∞   ∞ 
+ dt d 3 x ∇ · [J(x, t)α(x, t)] − dt d 3 x ∇ · Jα(x, t).
−∞ −∞

The second and fourth terms are both total derivatives which integrate to zero
as long as α(x, t) → 0 as t → ±∞ and J(x, t) → 0 as x → ∞, respectively.
The transformation then takes the form
 ∞  
1 ∂ρ(x, t)
Sint → Sint − dt 3
d x + ∇ · J(x, t) α(x, t),
−∞ c ∂t
and so in order for the action to be invariant for arbitrary α(x, t), we require
1 ∂ρ(x, t)
+ ∇ · J(x, t) = 0,
c ∂t
which is the continuity equation for electric charge.
2. Consider a homogeneous rectangular slab of metal, with edges L x , L y , and
L z . Assume that the electric potential φ vanishes within the slab, and that the
wave functions of conduction electrons in the slab satisfy periodic boundary
92 Solutions to Lectures on Quantum Mechanics

conditions at the slab faces. Suppose that the slab is in a constant magnetic
field in the z-direction that is strong enough so that the cyclotron frequency ω
is very much larger than /m e L 2z . Suppose that there are n e conduction elec-
trons per unit volume in the slab. Calculate the maximum energy of individual
conduction electrons, in the limit ωm e L 2z / → ∞.
The Hamiltonian for an electron in a general electromagnetic field is given in
Eq. (10.3.1)
1  e 2 2μ
e
H= p + A(x, t) − s · H(x, t) .
2m e c 
For a constant magnetic field in the z-direction, the vector potential can be
taken to have the form

A y = x Hz , A x = A z = 0,

and so the Hamiltonian becomes

   
1 ex Hz 2 2μe
H= px + p y +
2
+ pz −
2
sz Hz .
2m e c 

This Hamiltonian commutes with p y , pz , and sz , so there exist states  such

that

p y  = k y , pz  = k z , sz  = ± , H  = E.
2
The Schrödinger equation can then be written as
  2   
1 eH z x 2 k z2
px + k y +
2
= E− ± μe Hz ,
2m e c 2m e

which we can rewrite as

  
1 2 m e ω2 2 k z2
px + (x − x0 )2  = E − ± μe Hz ,
2m e 2 2m e
where we have defined ω ≡ eHz /m e c and x0 ≡ −k y c/eHz . In this form, it
is clear that this is just the Schrödinger equation for the harmonic oscillator
which was discussed in Section 2.5, and so the energy levels are given by
 
2 k z2 1
E= ∓ μe Hz + ω n + for n = 0, 1, 2, . . . .
2m e 2
For electrons in a metal slab with dimensions L x , L y , and L z , we must have
Lx Lx Ly Ly Lz Lz
− ≤x≤ , − ≤y≤ , − ≤z≤ ,
2 2 2 2 2 2
 Chapter 10 Problem Set Solutions 93

and as a result we must have

Lx
|x0 | <
  2
 k y c  Lx
 
 eH  < 2
z
  eHz L x
k y  < .
2c
Due to the periodic boundary conditions at the slab faces, we have
2πn y 2πn z
ky = , kz = ,
Ly Lz

where n x and n y are integers. The number of states with a given n, 2 k z2 /2m e ,
and sz having |k y | less than eHz L x /2c is the number
 of positive
 and negative
integers with magnitude less than (eHz L x /2c) L y /2π which is
eHz A
N = ,
2πc
where A = L x L y is the area of the slab face perpendicular to the mag-
netic field. In the limit ωm e L 2z / → ∞ all of the conduction electrons
will be in the state n = 0, and with spin-z component along the direction
of the magnetic field, so the energy of individual conduction electrons is
E = 2 k z2 /2m e , where we are neglecting corrections which come from quan-
tum electrodynamics. The Pauli exclusion principle allows one electron in
each state, which will be filled in order of increasing energy until N /V times
the number of filled energy states equals the number of conduction electrons
per unit volume n e . The integer n z will take values 0, ±1, ±2, . . . , n z,max
with one electron for each n z , or two electrons for each value of |n z |, so that
2|n z,max |N
ne = .
Lx L y Lz
Solving this expression for |n z,max | we find
 
ne L x L y L z 2πc n e L z πc
|n z,max | = = .
2 eHz L x L y eHz
Therefore the maximum energy of individual conduction electrons is
 
2 2πn z,max 2 2π 4 4 c2 n 2e
E max = = .
2m e Lz m e e2 Hz2
3. Consider an non-relativistic electron in an external electromagnetic field.
Calculate the commutators of different components of its velocity.
94 Solutions to Lectures on Quantum Mechanics

Recall that for a non-relativistic electron in an external electromagnetic field,

the velocity is given in terms of the canonical momentum as in Eq. (10.1.8)
1  e 
ẋ = p + A(x, t) ,
m c
where we have inserted the charge of the electron en = −e. The posi-
tion and its canonical conjugate obey the commutation relations given in
Eqs. (10.1.10) and (10.1.11)
 
xi , p j = iδi j
   
xi , x j = pi , p j = 0.
The commutation relations of two components of the velocity are then
given by
  1  
  1 e e
ẋi , ẋ j = pi + Ai (x, t) , p j + A j (x, t)
m c m c
1    e  e 
= 2 pi , p j + pi , A j (x, t) + Ai (x, t), p j
m c  c
e e
+ Ai (x, t), A j (x, t) .
c c
The first and last terms vanish, and we can calculate the remaining commu-
tators by using the canonical commutation relations
e  e ∂ Ai (x, t)
Ai (x, t), p j = i ,
c c ∂x j
and so we find

  e ∂ A j (x, t) ∂ Ai (x, t)
ẋi , ẋ j = −i 2 −
m c ∂ xi ∂x j
e 
= −i 2 i jk (∇ × A(x, t))k .
m c k
 Chapter 11 Problem Set Solutions 95

Chapter 11 Problem Set Solutions

1. Calculate the rates for emission of photons in the transitions 3d → 2 p and

2 p → 1s in hydrogen. Give formulas and numerical values. You can use the
facts that the proton is much heavier than the electron, and the wavelength
of the photon emitted in these processes is much larger than the atomic size,
and neglect electron spin.
For the transitions 3d → 2 p and 2 p → 1s, assuming that kr  1, and
ignoring electron spin, we can use the dipole approximation and calculate
the rate by using Eq. (11.7.28)
 2
4ω2  

= 3  en (b |x̄n | a) ,
3c   n 
where the matrix element is given by Eq. (11.7.27)
 1  
 m m x̄m

(b |x̄n | a) = d x̄m δ
3 3
ψb∗ (x̄)x̄n ψa (x̄).
m m
M
For hydrogen, the delta function becomes
   
 m m x̄m m p x̄ p + m e x̄e
δ 3
= δ3 ≈ δ 3 (x̄ p ),
m
M m e + m p

since the mass of the proton is much larger than the mass of the electron
m p  m e . Therefore, the integral over x̄ p just sets x̄ p = 0. Now, we need to
calculate the matrix elements

(b |x̄n | a) = d 3 x̄ ψb∗ (x̄)x̄ψa (x̄),

where x̄ is the position of the electron relative to the proton. Clearly, the tran-
sition rates will be independent of the m quantum number of the initial state,
because the rate cannot depend upon how we choose to orient our coordinate
system, but we will show this explicitly below. The relevant wave functions
for hydrogen are given by
 3/2  
1 1 r 2 −r/3a 2 ±2iφ
ψ3d ±2 = √ e sin θe ,
162 π a a
 3/2  
1 1 r 2 −r/3a
ψ3d ±1 = √ e sin θ cos θe±iφ ,
81 π a a
 3/2  
1 1 r 2 −r/3a  
ψ3d 0 = √ e 3 cos2 θ − 1 ,
81 6π a a
 3/2  
1 1 r 2 −r/2a
ψ2 p±1 = √ e sin θe±iφ ,
8 π a a
96 Solutions to Lectures on Quantum Mechanics
 3/2  
1 1 r 2 −r/2a
ψ2 p0 = √ e cos θ,
4 2π a a
 3/2
1 1
ψ1s 0 =√ e−r/a .
π a

The only non-vanishing matrix elements for the transition 3d → 2 p are

given by
  r 3
 ±1 ±2
 1 ∞
2 p |x̄|3d = 4 4 3 e−5r/6a
dr r 3
2 3 πa 0 a
 π  2π
× dθ sin5 θ dφ e±iφ cos φ
0 0
211 34
= a
57  ∞
 ±1  1  r 3
±2
2 p | ȳ|3d = 4 4 3 dr r 3 e−5r/6a
2 3 πa 0 a
 π  2π
× dθ sin θ 5
dφ e±iφ sin φ
0 0
211 34
= ±i 7 a
5 
 ±1  1 ∞  r 3
2 p |z̄|3d ±1 = 3 4 3 dr r 3 e−5r/6a
2 3 πa 0 a
 π  2π
× dθ sin θ cos θ
3 2
dφ
0 0
211 34
=a
57  ∞
 0  1  r 3
2 p |x̄|3d ±1 = 5/2 4 3 dr r 3 e−5r/6a
2 3 πa 0 a
 π  2π
× dθ sin θ cos θ
3 2
dφ e±iφ cos φ
0 0
1 211 34
=√ a
2 57
 ∞  r 3
 0 ±1
 1
2 p | ȳ|3d = 5/2 4 3 dr r 3 e−5r/6a
2 3 πa 0 a
 π  2π
× dθ sin θ cos θ
3 2
dφ e±iφ sin φ
0 0
i 211 34
= ±√ a
2 57
 Chapter 11 Problem Set Solutions 97
 ∞  r 3
 0  1
2 p |z̄|3d 0 = 3 9/2 3 dr r 3 e−5r/6a
2 3 πa 0 a
 π  2π
× dθ sin θ cos θ(3 cos θ − 1)
2 2
dφ
0 0
2 211 34
=√ a
3 57
 ∞  r 3
 ±1  1
2 p |x̄|3d = 7/2 9/2 3
0
dr r 3 e−5r/6a
2 3 πa 0 a
 π  2π
× dθ sin θ(3 cos θ − 1)
3 2
dφ e∓iφ cos φ
0 0
1 211 34
= −√ a
6 57
 ∞  r 3
 ±1  1
2 p | ȳ|3d = 7/2 9/2 3
0
dr r 3 e−5r/6a
2 3 πa 0 a
 π  2π
× dθ sin θ(3 cos θ − 1)
3 2
dφ e∓iφ sin φ
0 0
i 211 34
= ±√ a,
6 57
and the only non-vanishing matrix elements for the transition 2 p → 1s are
 ∞    π  2π
 0  1 3 r
1s |x̄|2 p ±1 = dr r e −3r/2a
dθ sin θ
3
dφ e±iφ cos φ
23 πa 3 0 a 0 0
27
= 5a
3
 ∞ r   π  2π
 0  1
1s | ȳ|2 p ±1 = 3 3 dr r 3 e−3r/2a dθ sin3 θ dφ e±iφ sin φ
2 πa 0 a 0 0
27
= ±i 5 a
3
 ∞    π  2π
 0  1 3 r −3r/2a
1s |x̄|2 p = 5/2 3
0
dr r e dθ sin θ cos θ
2
dφ
2 πa 0 a 0 0
√ 27
= 2 5 a.
3
Squaring these matrix elements, we find for the 3d → 2 p transition
     
 2 p|x̄|3d ±2 2 =  2 p ±1 |x̄|3d ±2 2 +  2 p ±1 | ȳ|3d ±2 2
223 38 2
= a
514
98 Solutions to Lectures on Quantum Mechanics
       
 2 p|x̄|3d ±1 2 =  2 p ±1 |z̄|3d ±1 2 +  2 p 0 |x̄|3d ±1 2 +  2 p 0 | ȳ|3d ±1 2
223 38 2
= a
514
        
 2 p|x̄|3d 0 2 =  2 p 0 |z̄|3d 0 2 + 2  2 p ±1 |x̄|3d 0 2 + 2  2 p ±1 | ȳ|3d 0 2
223 38 2
= a ,
514
and for the 2 p → 1s transition
     
 1s|x̄|2 p ±1 2 =  1s 0 |x̄|2 p ±1 2 +  1s 0 | ȳ|2 p ±1 2
213 2
= a
310
    
 1s|x̄|2 p 0 2 =  1s 0 |z̄|2 p 0 2
215 2
= a .
310
As stated above, we see that the transition rate does not depend upon the
initial value of the m quantum number. Next, we need the frequencies for
the emitted photons in the 3d → 2 p and 2 p → 1s transitions of hydrogen.
The energy of the nth level is
m e e4
En = − ,
22 n 2
and so the frequency is given by
 
Ea − Eb m e e4 1 1
ω= = − 2 .
 23 2
nb na
For the 3d → 2 p transition, this gives
ω3d→2 p = f rac5m e e4 622 ,
and for the 2 p → 1s transition, this gives
3m e e4
ω2 p→1s = .
83
Putting this together and using the definition of the Bohr radius a = 2 /m e e2 ,
we find for the rate of the 3d → 2 p transition
 3 2 23 5
4 5m e e4 e 2 3 2
3d→2 p = a
3 72 2 c3  514
216 3 m 2 e14
= 11 3e 10 a 2
5 c 
216 3 m e e10
= 11 3 6
5 c 
= 6.4677 × 107 s−1 ,
 Chapter 11 Problem Set Solutions 99

and for the rate of the 2 p → 1s transition

 3 2 15
4 3m e e4 e 2 2
2 p→1s = a
3 8 3 c3  310
28 m 3 e14
= 8 3e 10 a 2
3 c 
28 m 310
= 8 3e 6
3 c 
= 6.2674 × 108 s−1 .

2. What power of the photon wave number appears in the rate for single photon
emission in the decay of the 4 f state of hydrogen into the 3s, 3 p, and 3d
states?
We will begin by examining Eq. (11.7.18) which gives the differential rate
for single photon emission
 2
k  ∗  en 

d(k̂, σ ) = e (k̂, σ ) · Dn ba (k̂) d ,
2π  n
m n c 

where Dn ba is defined in Eq. (11.7.13)

 1  
 m m x̄m

Dn ba (k̂) = d x̄m δ
3 3
ψb∗ (x̄)e−ik·x̄n (−i∇n ) ψa (x̄).
m m
M

In the case when the wavelength of the emitted photon is much larger than
the atomic size kr  1, we can expand the exponential appearing in Dn ba
and keep only the leading non-vanishing term
1
e−ik·x̄n ≈ 1 − ik · x̄n − (k · x̄n )2 + · · · .
2
The operators Pn and X̄n are spatial vectors, and so they behave as operators
with j = 1 under rotations. The matrix elements of such operators between
states ψa and ψb vanish unless the angular momentum quantum numbers ja
and jb satisfy | ja − jb | ≤ 1 with ja and jb not both zero. Furthermore, these
operators change sign under a spatial reflection, so these matrix elements
vanish unless the states ψa and ψb have opposite parity.
The states 4 f and 3d have opposite parity, and |3 − 2| = 1 ≤ 1, so for the
transition 4 f → 3d, we can use the leading term in the expansion of e−ik·x̄n .
In this case, the transition rate is given by Eq. (11.7.28)
 2  2
4ω3  
 4k 3  

4 f →3d = 3  en (3d|x̄n |4 f ) =  en (3d|x̄n |4 f ) ,
3c   n  3  n 
100 Solutions to Lectures on Quantum Mechanics

and so we find
4 f →3d ∝ k 3 .
The states 4 f and 3 p have the same parity, and |3−1| = 2  1, so we need
to use the second term in the expansion of e−ik·x̄n . In this case, the transition
rate is given by Eq. (11.7.34)

2k 3  3  2 5  2
4 f →3 p = 3 p|Q i j |4 f  +  3 p|Mi j |4 f  .
15 im 4 4
The second term, which represents the magnetic dipole contribution vanishes
because the operator Mi j acts under rotation as an operator with j = 1 and
therefore vanishes between states ψa and ψb unless | ja − jb | ≤ 1 ≤ ja + jb .
The first term, whcih represents the electric quadrupole moment, does not
vanish for this transition because the operator Q i j acts as an operator
 with
j = 2 under rotation, and therefore the matrix elements b|Q i j |a vanish
unless | ja − jb | ≤ 2 ≤ ja + jb . The electric quadrupole matrix element can
be written as in Eq. (10.7.38)
     1 2 

b|Q i j |a = ick en b|x̄ni x̄n j |a − b|x̄n |a .
n
3

This means that the rate 4 f →3 p will have an additional factor k 2 relative to
the rate 4 f →3d
⎡  ⎤
2
2k 3  ⎣ 3 2 2     1  
4 f →3 p = c k i en 3 p|x̄ni x̄n j |4 f − 3 p|x̄2n |4 f  ⎦ ,
15 i j 4  n 3 

and so we find
4 f →3 p ∝ k 5 .
The states 4 f and 3s have opposite parity and |3 − 0| = 3, so we must
use the third term in the expansion of the exponential. The matrix element
appearing in the rate for the transition 4 f → 3s is then
1  
− ki k j 3s|x̄ni x̄n j pn |4 f .
2 ij

The product of three spatial vectors contains a term which behaves as an

operator with j = 3 (and also terms with lower values of j), and so this
matrix element will be non-vanishing if | ja − jb | ≤ 3. Following the same
logic as above, the rate for the transition 4 f → 3s then has an additional
factor k 4 as compared to the rate 4 f →3d , and so we find
4 f →3s ∝ k 7 .
 Chapter 11 Problem Set Solutions 101

3. Consider the theory of a real scalar field ϕ(x, t), interacting with a set of
particles with coordinates xn (t). Take the Lagrangian as
   
1 ∂ϕ(x, t) 2  2
L(t) = 3
d x − c ∇ϕ(x, t) − μ ϕ (x, t)
2 2 2
2 ∂t
  mn  2  
− gn ϕ(xn (t), t) + ẋn (t) − V x(t) ,
n n
2

where μ, m n and gn are real parameters, and V is a real local function of the
differences of the particle coordinates.
a. Find the field equations and commutation rules for ϕ.
b. Find the Hamiltonian for the whole system.
c. Express ϕ in the interaction picture in terms of operators that create and
destroy the quanta of the scalar field.
d. Calculate the energy and momentum of these quanta.
e. Give a general formula for the rate of emission per solid angle of a sin-
gle ϕ quantum in a transition between eigenstates of the matter part of
the Hamiltonian (that is, the part of the Hamiltonian involving only the
coordinates xn and their canonical conjugates).
f. Integrate this formula over solid angles in the case where the wave-
length of the emitted quanta is much larger than the size of the ini-
tial and final particle system. What are the selection rules for these
transitions?
We begin with a Lagrangian of the form
  2 
1 ∂ϕ(x, t)
L(t) = d3x − c2 (∇ϕ(x, t))2 − μ2 ϕ 2 (x, t)
2 ∂t
  mn
− gn ϕ(xn (t), t) + (ẋ(t))2 − V (x(t)).
n n
2

We can rewrite this Lagrangian as

  mn
L(t) = d 3 x L(x, t) + (ẋ(t))2 − V (x(t)),
n
2

where we have defined the Lagrangian density for the field ϕ

 
1 ∂ϕ(x, t) 2 c2 μ2 2
L(x, t) = − (∇ϕ(x, t))2 − ϕ (x, t)
2 ∂t 2 2

− gn ϕ(x, t)δ 3 (x − xn (t)).
n
102 Solutions to Lectures on Quantum Mechanics

(a) The Euler–Lagrange field equations are given by Eq. (11.1.9)

∂L ∂ ∂L ∂ ∂L
− = ,
∂ϕ ∂ xi ∂(∂i ϕ) ∂t ∂ ϕ̇
which in this case read

−μ2 ϕ(x, t) − gn δ 3 (x − xn (t)) + c2 ∇ 2 ϕ(x, t) = ϕ̈(x, t) .
n

We can also find the field equation for xn (t) which is given by
 
∂ δL δL
= ,
∂t δ ẋni δxni
which gives in this case
∂ V (xn (t)) ∂ϕ(xn (t), t)
ẍni (t) = − − gn .
∂ xni ∂ xni
The canonical conjugate to ϕ is
∂L
π(x, t) = = ϕ̇(x, t).
∂ ϕ̇
Since there are no constraints, the field ϕ and its conjugate π obey
canonical commutation relations
 
ϕ(x, t), π(y, t) = iδ 3 (x − y)
   
ϕ(x, t), ϕ(y, t) = π(x, t), π(y, t) = 0.
(b) The Hamiltonian for the system is

H = d 3 xπ(x, t)ϕ̇(x, t) + pn (t) · xn (t) − L(t),

where we have defined the canonical conjugate of xn

δL
pni (t) = = m n ẋni (t).
δ ẋni
The Hamiltonian is then given by

1  
H= d 3 x π 2 (x, t) + c2 (∇ϕ(x, t))2 + μ2 ϕ 2 (x, t)
2
  (pn (t))2
+ gn ϕ(xn (t), t) + + V (x(t)) .
n n
2m n

We can rewrite this in terms of a free Hamiltonian and an interaction

H = H0 ϕ + H0 mat + Hint ,
 Chapter 11 Problem Set Solutions 103

where we have defined


1  
H0 ϕ = d 3 x π 2 (x, t) + c2 (∇ϕ(x, t))2 + μ2 ϕ 2 (x, t)
2
 (pn (t))2
H0 mat = + V (x(t))
n
2m n

Hint = gn ϕ(xn (t), t).
n

(c) Now we will introduce the interaction picture field ϕ I whose time-
dependence is governed by H0 ϕ
i  
ϕ̇ I = H0 ϕ , ϕ I = π I

i  
π̇ I = H0 ϕ , π I = c2 ∇ 2 ϕ I − μ2 ϕ I = ϕ̈ I .

The interaction picture field and its canonical conjugate obey the same
commutation relations as the Heisenberg picture fields
 
ϕ I (x, t), π I (y, t) = iδ 3 (x − y)
   
ϕ I (x, t), ϕ I (y, t) = π I (x, t), π I (y, t) = 0.
The wave equation for ϕ I reads
ϕ̈ I = c2 ∇ 2 ϕ I − μ2 ϕ I .
This can be solved by writing ϕ I as a Fourier integral

 
ϕ I (x, t) = d 3 k eik·x e−iωt α(k) + e−ik·x eiωt α † (k) .

Plugging this expression into the wave equation gives

−ω2 ϕ I = −c2 k 2 ϕ I − μ2 ϕ I ,
and so we can solve for ω

ω= c 2 k 2 + μ2 .
The canonical conjugate for ϕ can then be expressed as

 
π I (x, t) = ϕ̇ I = −i d 3 k ω eik·x e−iωt α(k) − e−ik·x eiωt α † (k) .

We can work out the commutators of α(k) and α † (k) by requiring that
we recover the correct commutation relations for ϕ I and π I
  
     
ϕ I (x, t), π I (y, t) = −i d k d 3 k  ω ei(k·x+k ·y) e−i(ωt+ω t) α(k), α(k )
3

   
+ ei(−k·x+k ·y) ei(ωt−ω t) α † (k), α(k )
104 Solutions to Lectures on Quantum Mechanics
   
− ei(k·x−k ·y) ei(−ωt+ω t) α(k), α † (k )
   
− ei(−k·x−k ·y) ei(ωt+ω t) α † (k), α † (k )
= iδ 3 (x − y).
This expression is time-independent, so the first and the fourth terms
must vanish, which requires
   
α(k), α(k ) = α † (k), α † (k ) = 0,
and this condition immediately guarantees that
   
ϕ I (x, t), ϕ I (y, t) = π I (x, t), π I (y, t) = 0.
For the remaining terms, we can express δ 3 (x − y) as a Fourier integral

1
δ (x − y) =
3
d 3 k eik·(x−y) .
(2π)3
We can now read off the remaining commutator
  
α(k), α † (k ) = δ 3 (k − k ).
(2π)3 2ω
Now, we can rewrite α(k) in terms of a(k) which has simple communta-
tion relations


α(k) = a(k).
(2π)3 2ω

The commutation relations for a(k) and a † (k) are then

 
a(k), a † (k ) = δ 3 (k − k )
   
a(k), a(k ) = a † (k), a † (k ) = 0,
which we recognize as the commutation relations of creation and annihi-
lation operators. We can then express the interaction picture field and its
conjugate in terms of these creation and annihilation operators
 
  ik·x −iωt 
ϕ I (x, t) = d 3 k e e a(k) + e−ik·x eiωt a † (k)
(2π) 2ω
3

 
  ik·x −iωt −ik·x iωt †

π I (x, t) = −i d k ω
3
e e a(k) − e e a (k) .
(2π)3 2ω
(d) Let us first write out the free part of the Hamiltonian in terms of cre-
ation and annihilation operators. The terms appearing in H0 ϕ are, after
performing the integrals over x and k
 Chapter 11 Problem Set Solutions 105
   

d 3 x π I2 (x, t) = − d 3 k ω2 e−2iωt a(k)a(−k) + e2iωt a † (k)a † (−k)
2ω

− a(k)a † (k) − a † (k)a(k)
    

d x c (∇ϕ I (x, t)) = c
3 2 2 2 3
d k k 2 e−2iωt a(k)a(−k)
2ω

+ e2iωt a † (k)a † (−k) + a(k)a † (k) + a † (k)a(k)
   

d xμ ϕ I (x, t) = μ
3 2 2 2 3
d k e−2iωt a(k)a(−k) + e2iωt a † (k)a † (−k)
2ω

+ a(k)a † (k) + a † (k)a(k) .

Using the dispersion relation ω2 = c2 k 2 + μ2 , we can see that all of the

time-dependent terms in H0 ϕ cancel, and we have

1  
H0 ϕ = d 3 k ω a(k)a † (k) + a † (k)a(k) .
2
Now, we can calculate the commutators of the Hamiltonian H0 ϕ with the
operators a † (k) and a(k)
  
H0 ϕ , a † (k) = ωa † (k) =  c2 k 2 + μ2 a † (k)
  
H0 ϕ , a(k) = −ωa(k) = − c2 k 2 + μ2 a(k).
This confirms that a † (k) and a(k) are raising and lowering operators
for the energy, so if a state  is an eigenstate of H0 ϕ with eigen-
 then the state a (k) is an eigenstate of H0 ϕ with eigenvalue
†
value E,
E + c2 k 2 + μ2 and the state a(k) is an eigenstate with eigenvalue
E −  c2 k 2 + μ2 . If we define a state 0 such that a(k)0 = 0, then
we can construct states of the form
k1 ,k2 ,··· ,kn ∝ a † (k1 )a † (k2 ) · · · a † (kn )0 ,
which have energy (ignoring the infinite constant E 0 like the one
appearing in Eq. (11.6.5))
" " "
 c2 k12 + μ2 +  c2 k22 + μ2 + · · · +  c2 kn2 + μ2 .
We interpret this as a state with n scalar field quanta, with energies
" " "
 c2 k12 + μ2 ,  c2 k22 + μ2 , . . . ,  c2 kn2 + μ2 .
By using the general formula given in Eq. (9.4.4) we can construct a
momentum operator Pϕ which is given by

Pϕ = − d 3 x π I (x, t)∇ϕ I (x, t).
106 Solutions to Lectures on Quantum Mechanics

The commutators of this operator with the scalar field and its conju-
gate are
 
Pϕ , ϕ I (x, t) = i∇ϕ I (x, t)
 
Pϕ , π I (x, t) = iπ I (x, t).
From these relations, we can calculate the commutators with the creation
and annihilation operators
 
Pϕ , a † (k) = ka † (k)
 
Pϕ , a(k) = −ka(k).
So we can see that the state k1 ,k2 ,··· ,kn has momentum
k1 + k2 + · · · + kn ,
and thus the quanta of the scalar field have momenta k1 , k2 , . . . , kn .
Tosummarize the quantum of scalar field created by a † (k) has energy
 c2 k 2 + μ2 and momentum k.
(e) Let us define states a and b which are eigenstates of the matter
Hamiltonian H0 mat
H0 mat a = E a a , H0 mat b = E b b ,
and are annihilated by a(k)
a(k)a = a(k)b = 0,
for all k. If we wish to calculate the rate of single ϕ quantum emission
from b → a, then we need a state bϕ = −3/2 a † (k)b which contains
a single ϕ quantum. To first order in the interaction, the S-matrix takes
the form
   
Sbϕ,a = −2πiδ E a − E b −  c2 k 2 + μ2 bϕ , V (0)a
  
= −2πi−3/2 δ E a − E b −  c2 k 2 + μ2 (b , a(k)V (0)a ) ,
where the potential is given by

V (0) = gn ϕ I (xn ).
n

Plugging this in and using the commutation relation of a(k) with ϕ, we

find
 
   
−3/2
Sbϕ,a = −2πi δ E a − E b −  c2 k 2 + μ2 b , a(k) gn ϕ I (xn )a
n
√
2πi    
= −" δ(E a − E b − ω) gn b , e−ik·xn a .
(2π )¯ 2ω
3
n
 Chapter 11 Problem Set Solutions 107

If we shift to center-of-mass coordinates

 m n xn 
X= , M≡ mn, x̄n ≡ x − X,
n
M n

we can rewrite the matrix element

   
b , e−ik·xn a = b̄ , e−ik·xn a ,
where we have defined
b̄ = eik·X b ,
which has momentum
Pb̄ = (pb + k) b .
We can then write
 
b̄ , e−ik·x̄n a = δ 3 (pb + k − pa ) Dn ba (k̂),
and so the S-matrix element becomes
Sbϕ,a = δ(E a − E b − ω)δ 3 (pb + k − pa ) Mbϕ,a ,
where we have defined
√
−2πi  
Mbϕ,a = gn Dn ba (k̂).
(2π)3 2ω n
The differential rate of single ϕ emission is then given by
1  2
d = Mbϕ,a  μb k d ,
2π
where the quantity μb is defined as
E b ω ω
μb ≡ ≈ 2,
c2 (Eb + ω) c
and the approximate equality holds when E b ≈ μc2  ω. Putting this
all together we arrive at an expression for the emission rate per solid
angle
 2
d(k̂) 1 4π 2  ω  

= k  gn D n ba (k̂)
d 2π (2π) 2ω c
3  n 
 2
 
k  
=  gn D n ba (k̂) .
8π c
2 2  n

108 Solutions to Lectures on Quantum Mechanics

(f) Assuming that wavelength of the emitted quantum is much larger than
the size of the initial and final system, we can expand the exponential
that appears in Dn ba (k̂) to get
 
Dn ba (k̂) = b|e−ik·x̄n |a = (b| (1 − ik · x̄n + · · · ) |a) ,

where for any operator O we can write the matrix element of O as

 
b̄ , Oa = δ 3 (pb − pa − k) (b|O|a) .

The states a and b̄ are orthogonal, so the first term in this expansion
vanishes, and the leading term is given by

Dn ba (k̂) = (b| − ik · x̄n |a) = −ik k̂i (b|x̄ni |a) .
i

The matrix element (b|x̄ni |a) is independent of the direction k̂, so we can
integrate the differential rate to find the total decay rate

d(k̂) k3   ∗
= gn gm k̂ i k̂ j (b| x̄ ni |a) b| x̄ m j |a .
d 8π 2 c2 nmi j
0
Using the formula d k̂i k̂ j = 4π
δ ,
3 ij
we find for the total decay rate
 2
k3  
 
=  gn (b|x̄n |a) .
6πc2  n 

The operator x̄n is a spatial vector and so acts as an operator with j = 1

under rotations. As a result, matrix elements (b|x̄n |a) will vanish unless
| ja − jb | ≤ 1 ≤ ja + jb . Also, x̄n changes sign under spatial reflection,
so (b|x̄n |a) will vanish unless the states a and b have opposite parity.
4. Express the coherent state A as a superposition of states (11.6.7) with
definite numbers of photons.
The coherent state is defined in Eq. (11.6.17) as

a(k, σ )A = A(k, σ )A .

As discussed at the end of Section 11.6, the state A must contain an infinite
number of photons, and so we expect the coherent state to take the form
∞
  n
A = Cn a † (k, σ ) 0 ,
n=0
 Chapter 11 Problem Set Solutions 109

where the Cn are normalization constants which we will determine below

and 0 is the state defined in Eq. (11.6.3) as the state which obeys
a(k, σ )0 = 0.
Recall the commutation relations for a(k, σ ) given in Eqs. (11.5.24) and
(11.5.25)
 
a(k, σ ), a † (k , σ  ) = δσ σ  δ 3 (k − k )
   
a(k, σ ), a(k , σ  ) = a † (k, σ ), a † (k , σ  ) = 0.
We will make the assumption that we can put the system in a large but finite
box so that k takes discrete values and the commutator contains a Kronecker
delta function for momenta rather than a Dirac delta function. This means
that the commutator for equal momenta is 1 rather than being infinite. The
continuum limit can be recovered by taking the volume of the box to infin-
ity. With this subtlety aside, and temporarily dropping the momentum and
helicity arguments, the defining equation for the coherent state reads
∞
 ∞

a Cn (a † )n 0 = A Cn (a † )n 0 ,
n=0 n=0

and using the commutation relations of a and a † we find a recursion relation

for the Cn
nCn = ACn−1 .
We can then write all of the Cn in terms of C0 as
An
Cn = C0 .
n!
To find C0 , we require that the coherent state is normalized
 ∞ ∞

 A ∗m  A n
1 = (A , A ) = |C0 |2 0 , am (a † )n 0 .
m=0
m! n=0
n!
Each term for which m = n vanishes in this double sum because a0 = 0,
and so the expression becomes a single sum
 ∞

 |A|2n n † n
1 = |C0 | 0 ,
2
a (a ) 0 .
n=0
(n!)2

Now we can move each of the factors of a past all of the factors of a † picking
up commutators and use the fact that a annihilates the state 0 to replace all
of the factors of a and a † in this expression with n!
 ∞
 ∞
 |A|2n  |A|2n
1 = |C0 | 0 ,
2
0 = |C0 | 2
= |C0 |2 exp(|A|2 ).
n=0
n! n=0
n!
110 Solutions to Lectures on Quantum Mechanics

Up to an arbitrary phase, we can then write C0 as

C0 = exp(−|A|2 /2).
Now we can express the coherent state (restoring momentum and helicity
arguments) as
∞
 (A(k, σ )a † (k, σ ))n
A = exp(−|A(k, σ )|2 /2) 0
n=0
n!
= exp(−|A(k, σ )|2 /2) exp(A(k, σ )a † (k, σ ))0 .