Problems for Quantum Physics I (Phys.

591)
Version 1.4
Kirill Tuchin1
1

Department of Physics and Astronomy, Iowa State University, Ames, IA 50011

(Dated: September 22, 2016)

Contents

I. Basic principles of Quantum Mechanics

II. Schr
odinger equation

2
7

III. Motion in one-dimension

10

IV. Representation theory

18

V. Motion in central potential

VI. Angular momentum
VII. Motion in magnetic field
VIII. Stationary perturbation theory

26
31
38
41

2
I.

BASIC PRINCIPLES OF QUANTUM MECHANICS

1. Prove the completeness relation

X

k (r 0 )k (r) = (r 0 r) .

(1.1)

where k (r) =

1
V

eikr .

Solution
Denote the left-hand-side of (1.1) by f :
X

f (r r 0 ) =

k (r 0 )k (r) =

1 X ik(rr0 )
e
.
V
k

The right-had-side coincides with expansion of f into Fourier series with Fourier coefficients fk = 1/V , where
V = L3 is boxs volume. On the other hand, the Fourier coefficients can be computed as
1
fk = 3
L

L/2

f (r r 0 )eik(rr ) d3 r =

L/2

1
,
V

Therefore, f (r r 0 ) = (r r 0 ).

2. (a) In a state described by the wave function


(x) = A exp

ip0 x (x x0 )2

~
2a2


,

(1.2)

where
p0 ,
x0 , a are

 real parameters, find the average coordinate hxi, momentum hpi and their fluctuations
(x)2 , (p)2 .
(b) Suppose now that (1.2) with x0 =
at t = 0. Determine the wave function at

0 describes

a free particle

t > 0 and calculate hx(t)i, hp(t)i, (x(t))2 , (p(t))2 . Interpret the results.
Solution
(a) Normalization condition gives |A| = (a2 )1/4 . Using hF i =
hxi = x0 ,

hpi = p0 ,

2  a2
x =
+ x20 ,
2

2
~2
p = 2 + p20 ,
2a

F d3 r we derive

 a2
(x)2 =
,
2


~2
(p)2 = 2 ,
2a

Notice that
p

h(x)2 i h(p)2 i =

(b) Expand in a complete set of momentum eigenstates

Z
0 (x) = ap p (x)dp ,

~
.
2

p (x) =

1
eipx/~
2~

3
Z
ap =



aA
(p p0 )2 a2
0 (x)p (x) dx = exp
2~2
~

Time dependence:
Z

ip2 t

ap e 2m~ p (x)dp


A
ma2 ~2 (x v0 t)2 + i~3 x3 t + ia4 m2 v0 ~(2x v0 t)
=q
exp
2m(a4 ~2 + t2 ~4 /m2 )
i~t
1 + ma
2

(x, t) =

where v0 = p0 /m.
2
0 t)
a|A|2 (xv
a(t)
,
|(x, t)| =
e
a(t)

a (t) = a

~2 t 2
1+ 2 4
m a

Normalization of the wave function gives |A|2 = (a2 )1/2 . We have



Z

 a2
~2 t2
hx(t)i = x|(x, t)|2 dx = a(t) ,
(x(t))2 =
1+ 2 4
2
m a

Z
hp(t)i =

p|ap |2 dp = p0 ,


~2
(p(t))2 = 2
2a


Note, that hp(t)i and (p(t))2 are time-independent as it should be for a free particle.

3. Find the relation between the average values of coordinate and momentum in two states with wave functions
1 and 2 satisfying
(a) 2 (x) = 1 (x + a),
(b) 2 (x) = eip0 x/~ 1 (x).
Solution
(a)
Z
Z
hxi2 =
2 (x) x 2 (x) dx =
1 (x + a) x 1 (x + a) dx

Z
1 () 1 () d a = hxi1 a ,
=

where = x + a.
Z
hpi2 =

2 (x)(i~)

2 (x) dx =
x

1 ()(i~)

1 () d = hpi1 .

(b)
hxi2 = hxi1 .
Z
hpi2 =

1 (x)eip0 x/~ (i~)

i
h
1 (x)eip0 x/~ dx = hpi1 + p0 .
x

4
B
C]
and [AB,
C]
through [A,
B],
[A,
C],
[B,
C].

4. (a) Express commutators [A,

are Hermitian, L
is an arbitrary linear operator. Prove that the following operators
(b) Operators A and B

AL
, AB
+B
A,
i(AB
B
A).

are Hermitian: L L, L + L , i(L L ), L

(c) Determine eigenvalues F of a Hermitian operator F satisfying
i. F 2 = c2 ,
ii. F 2 = cF ,
iii. F 3 = c2 F ,
where c is a real non-zero number.
Solution
(a)
B
C]
= AB
C B
C A = AB
C B
AC + B
AC B
C A = [A,
B]
C + B[
A,
C]
.
[A,
=B
A and (A ) = A,
for arbitrary linear operators.
(b) You should use (AB)
(c) (a) Eigenvalue equation F = F . Applying F to both sides:
F 2 = F 2

(1.3)

On the other hand, F 2 = c2 , hence F = c.

(b) F 2 = cF = cF . Using (1): cF = F 2 . Solutions: F = 0 and F = c.
(c) F 3 = F 3 . Also F 3 = c2 F = c2 F . Thus, F 3 = c2 F , which has these solutions: F = 0, F = c.

i = i~ijk rj k , where ijk is the Levi-Civita

5. (a) Orbital angular momentum operator can be represented as L
i, L
j ] = i~ijk L
k,
symbol and summation over the repeated index is implied (see lecture notes). Prove that [L
2
2

and calculate the following commutators: [Li , r ], [Li , L ], [Li , rj ].

(b) Prove that components of the angular momentum operator in spherical coordinates are given by





z = i~
+ cot cos
cot sin
, Ly = i~ cos
, L
Lx = i~ sin

Solution
(a) Let f be a function of x, y, z.
x, L
y ]f = ~2 [yz zy , zx xz ]f = ~2 {[yz , zx ] + [zy , xz ]}f
[L
= ~2 {yz (zx f ) zx (yz f ) + zy (xz f ) xz (zy f )}
zf .
= ~2 {xy yx }f = i~L
2 ] = 0.
i , r2 ] = [L
i, L
[L
P
i , rj ] = ~
[L
l,m ilm [xl m , xj ] = i~ijl rl .

5
(b) Spherical coordinates:
x = r sin cos ,

y = r sin sin ,

z = r cos

or
r 2 = x2 + y 2 + z 2

cos =

z
,
r

tan =

y
x

We have
r
r
r
= cos ,
= sin sin
= sin cos
z
y
x

sin

cos sin

cos cos
=
,
=
,
=
z
r
y
r
x
r

cos

sin
= 0,
=
,
=
z
y
r sin
x
r sin
Therefore,



y
=
Lz = i~ x
y
x





r


r


i~ r sin cos
+
+
+
+
r sin sin
y r y
y
x r x
x

= i~ .

can be calculated similarly.

Other components of L

6. Consider the following operators defined on the entire real axis < x < :
(a) Inversion P : P (x) (x);
(b) Translation: Ta : Ta (x) (x + a);

c: M
c (x) c (cx), c > 0 ;
(c) Scale transformation: M
K(x)

(d) Complex conjugation K:

(x);
(e) Permutation of coordinates of two particles: P12 : P12 (x1 , x2 ) (x2 , x1 ).

Which of these operators are linear? Find the corresponding Hemitian conjugated and inverse operators.
Solution
are linear.
All operators, except K
To calculate the operator Hermitian conjugated of Ta use its definition
Z
Z
(x)Ta (x)dx (x)(x + a)dx
Changing variables x x a and employing the definition of Hermitian conjugation we get
Z
Z
Z
(x)Ta (x)dx = (x a)(x)dx = (Ta (x)) (x)dx
Therefore, Ta (x) = (x a) = Ta (x), i.e. Ta = Ta .
c = M
1/c , P = P12 . K
does not exist, since K
is not linear.
By the same token, P = P , M
12
c1 = M
1/c , K
1 = K,
P 1 = P12 .
Inverse operators: P 1 = P , Ta1 = Ta , M
12

6
P
7. Let F (z) be a function of z that can be expanded in series F (z) = n cn z n . Define operator F = F (f), such
P
P
n
c
that F = n cn fcn . (For example, ebx = n xn! ). Using this definition find the explicit form of the following
operators
n
o
(a) exp iaP ,
 d
(b) Ta = exp a dx
,
 d
a = exp ax
a xk = (ea x)k ),
(c) L
(Hint: prove first that L
dx ,
where a is a real number, P is the inversion operator.
Solution
(a) By definition

eiaP =

X
(ia)n cn
P
n!
n=0

Using Pb2 = 1 we get Pbn = P for odd n and 1 for even n.

eiaP =

X
(ia)2n X (ia)2n1
1+
P = cos a + iP sin a
(2n)!
(2n 1)!
n=1
n=0

Thus,

eiaP (x) = cos a (x) + i sin a (x)

(b)
d
Ta = ea dx =

X
an dn
(x) = (x + a)
n! dxn
n=0

(c) Consider first

a xk =
L


n

X
1
d
ax
xk
n!
dx
n=0

d k
x = akxk
dx
2

d
= a2 k 2 xk
ax
dx
...

n
d
ax
= (ak)n xk
dx
ax

Thus,
a xk =
L

X
1
(ak)n xk = eak xk
n!
n=0

Now, using the Taylor expansion near x = 0:

a (x) = L
a
L

X
dk (0) xk
k=0

dxk

k!

X
dk (0) 1 a k
(e x) = (ea x) .
dxk k!

k=0

7
II.

SCHRODINGER
EQUATION

= pF

1. For a particle moving under the action of a constant force F , prove that G
t is an operator of a conserved
quantity. Explain.
Solution

dG

i
= G
+ [H,
G]
dt
t
~
Using
2
= p F r
H
2m

we get

i
dG
=0
= F [(F r), p]
dt
~
Therefore hGi = const.

2. Given the following one-dimensional systems: (i) particle in constant field U = F x, (ii) harmonic oscillator.
(a) Find Heisenberg representations of the coordinate x
and momentum p operators using two methods: (A)

by unitary transformation S(t),

(B) by direct solution of equations of motion for Heisenberg operators.
(b) Suppose that systems (i),(ii) are in a state described by the wave function


ip0 x (x x0 )2
(x) = A exp

.
~
2a2


Using the Heisenberg operators calculate hx(t)i, hp(t)i, (x(t))2 , (p(t))2 .
(c) Calculate [
p(t), x
(t0 )] for each system. Interpret the value of this commutator in case (ii).
Solution
(a) (i) Particle in a constant field.
(A) First, calculate the commutators
x
[H,
] =

1 2
i~
[
p ,x
] =
p ,
2m
m

[H,
x
[H,
]] =

i~
~2
(F )[
x, p] = F
m
m

p] = F [
[H,
x, p] = F i~ .
Now

x
H = eiHt/~ x
S eiHt/~ = x
S +

it
(it)2
pS
F 2
[H, x
] +
[H, [H, x
]] . . . = x
S +
t+
t .
~
2~2
m
2m

pH = eiHt/~ pS eiHt/~ = pS +

Ft
~

8
(B) Equations of motion
d
i
pH
x
H = [H,
x
H ] =
,
dt
~
m

d
i
pH = [H,
pH ] = F t
dt
~

Integrating the second equation and substituting to the first one yields the same result.
= p2 /(2m) + m 2 x2 /2. Repeating the steps of (A,B) we get
(ii) Harmonic oscillator. H
x
H = x
S cos t +

pS
sin t ,
m

(b) Average (expectation) values are given by

Z
hx(t)i = (x)
xH (x)dx ,

pH = pS cos t m
xS sin t

(x)
pH (x)dx ,

hp(t)i =

etc.

Substituting expressions for x

H , pH from (a) and integrating we get the result. Actually, you did all these
integrals in a previous assignment. Namely,

h
xS i = x0 ,


a2
,
x
2S = x20 +
2

h
pS i = p0 ,


~2
p2S = p20 + 2 ,
2a

Another useful relation

h
xS pS + pS x
S i = 2x0 p0
We have for (i)
hx(t)i = x0 +

 a2
(x(t))2 =
2

p0 t F t2
+
,
m
2m

hp(t)i = p0 +



~2 t 2
1+ 2 4 ,
m a

Ft
,
m


~2
(p(t))2 = 2
2a

For (ii)
hx(t)i = x0 cos t +

p0
sin t ,
m

hp(t)i = p0 cos t mx0 p0 sin t



a2
~2
2
2
(x(t)) =
cos t + 2 2 4 sin t ,
2
m a


2


~
m2 2 a4
2
2
2
(p(t)) = 2 cos t +
sin t
2a
~2

(2.1)
(2.2)

(c) Using the Heisenberg representation we find

(i)

[
p(t), x
(t0 )] = i~ ,

(ii)

[
p(t), x
(t0 )] = i~ cos (t t0 )

0
Notice
Suppose that at t = 0

that for (ii) the commutator vanishes at t t = (n + 1/2)/ (n is an integer).

2
(x) is small so that the oscillator is in a state with almost definite x. Then, at t0 = (n + 1/2)/ the value
of momentum is almost definite (see e.g. (1),(8.2) with a 0).

3. Consider a Hermitian operator F () that depends on a real parameter . Let n () be its discrete eigenstates
and Fn () the corresponding eigenvalues. Denote hF in the average value of F in state n. Prove the following
theorem
*
+
Fn ()
F
=
.

9
Solution
Using the fact that F is Hermitian: F = F and hence Fn = Fn we have
+
*
Z
Z
Z
Z
F

n
n
F
= n
n dq =
(n F n ) dq
F n dq n F
dq

n
Z
Z
Fn
n
n
=

F n dq
F n dq

Z
Z
Fn
n
n
=
Fn
n dq Fn n
dq

Fn
Fn
Fn
||2 dq =
.
=

10
III.

MOTION IN ONE-DIMENSION

1. Calculate the transmission and reflection coefficients for the barrier U (x) = 0 if x 0, and U (x) = U0 if x > 0.
Investigate special cases E U0 and E .
Solution
Suppose that a particle is incident from the left with E > U0 . Denote
p
p
k = 2mE/~2 > 0 , k 0 = 2m(E U0 )/~2 > 0
The corresponding wave functions are
1 = eikx + Beikx ,

x<0

for the incident wave and

0

3 = Ceik x ,

x>0

for the outgoing (transmitted) wave. Continuity of the wave function and its first derivative at x = 0 implies
the following matching conditions:
1+B =C,

k(1 B) = k 0 C

Solution:
B=

k k0
,
k + k0

C=

2k
k + k0

Thus,
2

R(E) = |B| =

!2

E E U0

E + E U0

p
4 E(E U0 )
k0
2
T (E) = |C| =

k
( E + E U0 )2

Evidently, R + T = 1.
At E : R(E) U02 /16E 2 0 .
p
At E U0 : T (E) 4 (E U0 )/U0 0.
2. Consider a particle in an infinitely deep potential well:
(
0, 0 x a,
U (x) =
, x < 0, x > a ,
(a) Find energy spectrum.
(b) What is the symmetry of the Hamiltonian? Specify the corresponding operator P and its spectrum. Classify
energy levels with respect to the eigenvalues P.


(c) Calculate hxin , (x)2 n , hpin , (p)2 n in a stationary state n (n = 1, 2, 3, . . .).

(d) Assume that at t = 0 the particle is in a state described by the wave function
(
Cx(x a) , 0 x a ,
(x, 0) =
0,
x < 0, x > a ,
where C is a constant.
i. What is the wave function (x, t) at t > 0?
ii. What is the probability to find the particle in the nth stationary state at t = 0?

11
iii. What is the probability to find the particle in the nth stationary state at t > 0?
iv. Calculate hx(t)i, hx(t)i,

hp(t)i.
Solution
(a) Schr
odinger equation:
r
00

0 x a,

+ k = 0,

k=

2mE
.
~2

General solution:
(x) = A cos kx + B sin kx .
Boundary conditions
(0) = (a) = 0
are satisfied by A = 0 and kn a = (n + 1), n = 0, 1, 2, 3, . . .. The corresponding spectrum is


(n + 1)x
~2 (n + 1)
n (x) = B sin
, En =
.
a
2ma
p
Normalization: B = 2/a.
(b) The Hamiltonian commutes with the operator of inversion with respect to the center of the potential well:

P(x)
(a x). Since P 2 =
1, its eigenvalues are P = 1. Now




(n + 1)x
(n + 1)x
n (x a) = B sin (n + 1)
= B cos((n + 1)) sin
= (1)n n (x) .
a
a
Thus, states with even n are even, i.e. P = 1, while states with odd n are odd, i.e. P = 1.
(c)
a
hxin = ,
2

(x)

hpin = 0 ,
(d) Normalization of (x, 0) gives C =

=a

(p)2

1
1

12 2 2 (n + 1)2

= 2 (n + 1)2


.

~2
.
a2

30/a5 . Write

(x, 0) =


an B sin

n=1

(n + 1)x
a


.

The coefficients read

r
an =

60
a6


x(x a) sin

(n + 1)x
a


dx =

240 1 + (1)n
.
3 (n + 1)3

Probabilities are given by wn = |an |2 .

At t > 0:
(x, t) =

X
n=1

Clearly wn s are t-independent.


an B sin

(n + 1)x
a

eiEn t/~ .

12

Z
hx(t)i =
=

(x, t)x(x, t)dx

0
X

an an0 B 2 ei(En En0 )t/~

n=1 n0 =1

X
a2
=
|an |2 B 2
4
n=1

a
a
= + 2
2


sin

(n + 1)x
a


sin

(n0 + 1)x
a

"


xdx

1 (1)nn
1 (1)n+n

+
B e
(n + n0 + 2)2
(n n0 )2
n,n0 =1
"
#
0

n+n0
X
1 (1)nn
i(En En0 )t/~ 1 (1)

an an0 e
.
(n + n0 + 2)2
(n n0 )2
0
an an0

2 i(En En0 )t/~

a2
2 2

n,n =1

Coefficients an and energy levels En have been calculated above.

hp(t)i is calculated in a similar way. hx(t)i

= hp(t)i /m.
3. A particle of mass m is confined to a one-dimensional square potential:

0, 0 x a;
U (x) =
, otherwise .
At t = 0 its normalized wave function is
r
(x, t = 0) =
(a)
(b)
(c)
(d)
(e)

8
[1 + cos(x/a)] sin(x/a) .
5a

Calculate the spectrum of the Hamiltonian at t = 0 and at t > 0.

Is the spectrum discrete, continuous or a combination of both? What is its degeneracy?
What is the wave function (x, t) of the particle at t > 0?
What is the probability that the particle is in the first excited state?
What is the probability that the particle is found in the left half of the box, i.e. in the region 0 x a/2
at t > 0?

Solution
(a) Schr
odinger equation

~2 00
= E ,
2m

0 < x < a.

In terms of
k2 =

2mE
~2

the solution is
(x) = A sin(kx) + B cos(kx)
Boundary conditions (0) = (a) = 0 imply that B = 0 and ka = n. The corresponding wave functions
r
nx
2
n (x) =
sin
a
a
and energy levels
En =

n2 2 ~2
,
2ma2

n = 1, 2, . . .

Energy levels are energy independent, while the wave functions

n (x, t) = n (x)eiEn t/~

13
(b) The spectrum is discrete and non-degenerate (its degeneracy is one).
(c) Expand
r
r
X
8
2
x
2x
(x, 0) =
An n =
sin
+
sin
.
5a
a
5a
a
n
The only two non-vanishing coefficients are
1
A2 = ,
5

2
A1 = ,
5
Thus,
r
(x, t) =

r




8
x
2
2x
i 2 ~t
2i 2 ~t
sin
sin
exp
+
exp

5a
2ma2
a
5a
ma2
a

(d)
P1 = |A1 |2 =

1
5

(e)
Z
P =

a/2

|(x, t)|2 dx =

8
5a

a/2

sin2


 2 
x
3 ~t
x
x
1 + cos2
+ 2 cos
cos
dx
a
a
a
2ma2

1
16
= +
cos
2 15

3 2 ~t
2ma2

4. One-dimensional harmonic oscillator is initially (at t = 0) in a state with wave function

(x, 0) = N

cn n (x) ,

n=0

where n (x) are the stationary states of the harmonic oscillator and c is a complex parameter.
(a)
(b)
(c)
(d)

Calculate the normalization constant N .

Find the wave function of the system at a later time t > 0.
Calculate the probability of finding the system again in the initial state at a later time t > 0.
Compute the expectation value of energy.

Solution
(a)
Z

|(x, 0)|2 dx = N 2

|c|2n =

n=0

N2
1 |c|2

so that
N=

p
1 |c|2

(b) The time-evolved wave function is

(x, t) = N eit/2

X
n=0

cn eint n (x)

14
(c) The probability for finding the system again in the initial state is
2
Z




P =  (x, 0)(x, t)dx .
Z

(x, 0)(x, t)dx = |N |2 eit/2

|c|2n eint = eit/2

n=0

1 |c|2
1 |c|2 eit

Thus,

P = 1+

4|c|2
t
sin2
2
2
(1 |c| )
2

1

(d) The expectation value of energy is

"
#

X
~
2
2n
1 + (1 |c| )
hHi = |N |
En |c| =
2n|c|
2
n=0
n=0
"
!#

X
~

~ 1 + |c|2
2
2n
1 + |c|(1 |c| )
|c|
=
=
2
|c| n=0
2 1 |c|2
2

2n

5. Using the ladder operators calculate the following expectation values for one-dimensional harmonic oscillator in
the nth stationary state:
(a) hxin , hpx in ,



(b) x2 n , p2x n ,


(c) x4 n .


(d) Prove that (
x)2 n (
px )2 n = (n + 1/2)2 ~/.
Solution

r
x
=

~
(
a + a
) ,
2m

p = i

m~
(
aa
)
2

(a)
hxin = hpx in = 0
(b)

2
En
x n=
,
m 2

2
px n = mEn

(c)

(d)

=3

~
2m

2
[1 + 2n(n + 1)]

15
6. Particle of charge q performs one-dimensional harmonic oscillations in constant electric field of strength E
directed along the oscillation axis. Find the stationary states.
Solution
Potential energy
U=

kx2
qEx ,
2

where q is electric charge. S.E.

~ 00
kx2
+(
qEx) = .
2m
2

Changing variables z = x qE/k we obtain

~ 00 kz 2
q2 E 2
+
= (
) .
2m
2
2k

This is mathematically equivalent to the harmonic oscillator problem without electric field. Therefore,
En = ~(n + 1/2)

q2 E 2
,
2k

p
k/m ,

n = 0, 1, 2, . . .

n (x) = nosc (z) = nosc (x qE/k)

7. Particle moves in one-dimensional harmonic oscillator U = kx2 /2.

(a) Prove that in any state (x, t), the expectation values of kinetic and potential energies averaged over the
period 2/, satisfy the virial theorem hT it = hU it .
(b) Suppose that a thin impenetrable barrier is inserted at x = 0. What is the energy spectrum?
Solution
(a) Starting from equation
r
n =

n
n1 +
2

n+1
n+1
2

it is not difficult to derive that



1p
1
1p
2
n =
n(n 1)n2 + n +
n +
(n + 1)(n + 2)n+2 .
2
2
2
Using this formula we have

2
x =

x2 dx =

an an0 ei(n0 n )t/~

n x2 n0 dx

n,n0 =0

~
=
m
~
=
m
+

X
n=0

an an0 ei(n0 n )t/~

n 2 n0 d

n,n0 =0

"

1
an0 +2 an0 ei2t
2

n0 =0

1
an an+2 ei2t
2

(n0

2)(n0

n=0

#
p

(n + 2)(n + 1)

+ 1) +

|an |

1
n+
2

16
Oscillating terms vanish when averaged over the period yielding
hU it =




m 2
2 
~ X
1
x t=
|an |2 n +
.
2
2 n=0
2

The subscript t means time-average. On the other hand,

Z

X
X

i(En0 En )t/~
n0 dx =
0
hEi =
an an e
n H
|an |2 En = 2 hU it .
n,n0 =0

Finally, hT it = hEi hU it = hU it .
(b) Inserting the barrier imposed a boundary condition (x = 0) = 0 on the wave functions n of the harmonic oscillator. This condition is satisfied by odd-parity eigenfunctions 2n+1 (x), n = 0, 1, 2, . . .. The
corresponding energy levels are En = (2n + 3/2)~.

8. Particle moves in three-dimensional harmonic oscillator U = k(x2 + y 2 + z 2 )/2. Using separation of variables
find
(a) energy spectrum;
(b) degeneracy of nth level;
(c) parity of the nth level.
Solution
Particle in three-dimensional harmonic oscillator U = k(x2 + y 2 + z 2 )/2, k = m 2 .
(a) Solution can be found in the form (x, y, z) = X(x)Y (y)Z(z). Substituting into the S.E. and doing the
same manipulations as we deed for the square well in three dimensions we obtain for each coordinate the
one-dimensional harmonic oscillator problem:

~2 00 kx2
X +
X = x X , . . .
2m
2

such that x + y + z = . Solution to the one-dimensional problem was derived in class. Eigenfunctions:
2

nx ny nz (x, y, z) = Nnx Nny Nnz Hnx (ax)Hny (ay)Hnz (az)ea

(x2 +y 2 +z 2 )/2

where a2 = m/~. Energy levels:

1
1
3
1
E = ~(nx + ) + ~(ny + ) + ~(nz + ) = ~(n + )
2
2
2
2
with n = nx + ny + nz .
(b) Consider first degeneracy of 2d oscillator. We need to count all states that have different nx and ny
(0 nx n, 0 ny n) but the same sum nx + ny = n. For a given nx there is only one ny that satisfies this
condition (e.g. nx = 0, ny = n; nx = 1, ny = n 1, etc). Since there are n + 1 possible nx , the degeneracy is
n + 1.
In the 3d case we need to count all states that have different nx , ny , nz but the same sum nx + ny + nz = n.
As we explained, at fixed nz there are n nz + 1 different states. Summing over all nz s we get
n
X
nz

1
1
(n nz + 1) = (n + 1)2 n(n + 1) = (n + 1)(n + 2) .
2
2
=0

(c) If n is even, then either nx ,ny , nz are all even or one of then is even and the other two are odd. In either
case, the parity of the nth state is even. If n is odd, then either nx ,ny , nz are all odd, or one of them is odd
and the other two are even, in which case the state n is odd.

17
9. Consider an arbitrary potential U (x) 0 satisfying U 0 when x . Using the integral representation of
the Schr
odinger equation show that
|En |

m
2~2

Z

2
U (x) dx

for any energy level En . (Hint: first apply S.E. to the ground state and use the oscillation theorem).
Solution
By definition, |En | |E0 |, so we need to consider only the ground state. Let x = x0 be the maximum of the ground
state wave function 0 (x). S.E in the integral form reads
Z
0
m
0 (x0 ) =
e|x0 x | |U (x0 )|0 (x0 )dx0
2
0 ~
The wave function of the ground state has no zeros, so we can choose it to be positive 0 (x) > 0 (by fixing the phase).
0
Notice, that the expression under the integral is positive. So, if we replace e|x0 x | 0 (x0 ) 0 (x0 ), the integral
will not decrease. Consequently,
Z
m
0 (x0 )
|U (x0 )|dx0 0 (x0 ) .
0 ~2
Canceling 0 (x0 ) and using 20 = 2m|E0 |/~2 we derive the required result.

18
IV.

REPRESENTATION THEORY

1. Wave function of a particle in the coordinate representation is given by



ip0 x (x x0 )2
,
p (x) = C exp

~
2a2
where p0 , x0 , a are real parameters. Calculate the wave function x (p) in the momentum representation.
Solution
General rule:
Z
hp|ai =

dxhp|xihx|ai

Substituting hp|ai = x (p), hx|ai = p (x) and

hp|xi =

1
eipx/~
2~

we obtain
Z
x (p) =

eipx/~
dx
C exp
2~

ip0 x (x x0 )2

~
2a2



i(p p0 )x0
a2 (p p0 )2
aC

= exp
.
~
2~2
~

2. Particle is in a state described by the wave function (x, y, z). What is the probability to find the particle in
the intervals z1 z z2 and p1 py p2 ?
Solution
First, calculate the wave function in momentum representation with respect to y-direction
Z
1
(x, py , z) =
(x, y, z)eipy y/~ dy
2~
The the probability to find the particle in the intervals z1 z z2 and p1 py p2 is
Z z2
Z p2
Z
w=
dz
dpy
dx |(x, py , z)|2
z1

p1

3. Find explicit form of the following operators in momentum representation

(a) Inversion P : P (x) (x);
(b) Translation: Ta : Ta (x) (x + a);

c: M
c (x) c (cx), c > 0 ;
(c) Scale transformation: M
K(x)

(d) Complex conjugation K:

(x);
(e) Permutation of coordinates of two particles: P12 : P12 (x1 , x2 ) (x2 , x1 ).

Solution
(a) Consider inversion P : P (x) (x). Multiplying by hp|xi and integrating over x we have
Z
Z
1
1
P dx
eipx/~ (x) = dx
eipx/~ (x)
2~
2~

19
Changing the integration variable in the right-hand-side x x we obtain
P (p) = (p) .
By the same token:
(b) Ta (p) = eipa/~ (p) ,
c (p) =
(c) M

1 (p/c) ,
c

(d) K(p)
= (p) ,
(e) P12 (p1 , p2 ) = (p2 , p1 ).

4. Determine the form of the following operators:

d in x-representation,
(a) 1/p
d in p-representation.
(b) 1/x
Hint: in both cases use the general formulas connecting the x and p-representations. Given F in x-representation,
its p-representation can be calculated using
Z
Z
hp0 |bi = dphp0 |F |pihp|ai ,
hp0 |F |pi = dxhp0 |xiF hx|pi .
Conversely if you know F in the p-representation, then in x-representation
Z
Z
hx0 |bi = dxhx0 |F |xihx|ai ,
hx0 |F |xi = dphx0 |piF hp|xi .

d in p-representation. Hint: start with

(c) 1/r
Z
d (p) =
1/r

K1/r (p, p0 ) (p0 ) d3 p0

and compute the kernel.

Solution
Suppose we know an operator F in x-representation. Then in p-representation
Z
Z
0
0
0
hp |bi = dphp |F |pihp|ai ,
hp |F |pi = dxhp0 |xiF hx|pi .
Conversely if you know F in the p-representation, then in x-representation
Z
Z
hx0 |bi = dxhx0 |F |xihx|ai ,
hx0 |F |xi = dphx0 |piF hp|xi .
(a)
Z
Z
Z
0
1
1 1
1 x
dp ip(xx0 )/~
d
eipx /~
eipx/~ =
d
x
e
hx0 |1/p|xi
= dp
p 2~
i~
2~
2~
Z
1 x
1
=
d
x (
x x0 ) = (x x0 ) ,
i~
i~
where (x) is the Heaviside (step) function defined as follows: = 1 for x > 0 and = 0 for x < 0.
Z
Z
Z
1
1
d
hx0 |bi = dxhx0 |1/p|xihx|ai
=
dx (x x0 )hx|ai =
dx hx|ai
i~
i~ x0

20
This is equivalent to:
d a (x0 ) =
b (x0 ) = 1/p

1
i~

dx a (x) .
x0

Note, that

D
E Z
d =
1/p

1
|(p)|2 dp < ,
p

only if (0) = 0. This implies that

Z

(x)dx = 0 .

In other words,
Z

x0

(x)dx =

(x)dx .
x0

So we can write
d a (x0 ) = 1
b (x ) = 1/p
i~
0

x0

1
dx a (x) =
i~

x0

dx a (x) .

d = 1 or (d/dx)(1/p(x))
d
This problem can be solved differently, by noting that p 1/p
= (i/~)(x). Integrating over x we get the same result as before.
(b) Similarly,
d a (p0 ) = 1
1/x
i~

dp a (p) =
p0

1
i~

p0

dp a (p) .

(c) In coordinate representation

K1/r (r, r 0 ) =

1
(r r 0 ) .
r

Therefore, in momentum representation

Z
Z Z 1
0
1
1 i(pp0 )r/~ 3
1
1
0
d r=
dr
d ei|pp |r/~ 2r2
K1/r (p, p ) =
e
3
3
(2~)
r
(2~) 0
r
1


Z
4~
|p p0 |r
=
dr sin
(2~)3 |p p0 | 0
~


Z
|p p0 |r
4~
lim Im
dr exp i
+ i r
=
(2~)3 |p p0 | 0
~
0


4~
1
4~2
1
=
lim
Im

=
=
3
0
0
3
0
2
2
(2~) |p p | 0
i|p p |/~ + i
(2~) |p p |
2 ~|p p0 |2
where = cos , > 0.

5. Find all eigenvalues and eigenfunctions of the following operator

0 i2
0
i
y = ~
0 i2
L
2
,
i
0
0
2

21
acting in the Hilbert space of three-component vectors.
Solution
Equation

 i
0

 i 2 i
 2
2

i
 0





=0



y are Ly = 0, ~.
has three solutions = 0, 1, so that eigenvalues of L
To find the eigenvector 1 = (a, b, c), corresponding to = 1, substitute = 1 in to the system of equations
y = Ly . Since only two are linearly independent we have
L
i
a b = 0 ,
2

i
bc=0
2

Implying that the eigenvector is

i2
1 = b 1 .

i
2

Normalization: 1 1 = 1 implies b = 1/ 2, so finally

1 =

2i
1
2
i
2

By the same token,

1
2

0 = 0 ,
1
2

1 =

i
2
1
2
2i

6. Unitary operators.
c , K,
P12 (see problem sets #3 and #5)?
(a) Which of the following operators are unitary: P , Ta , M
2 = U
. What is U
?
(b) A unitary operator satisfies the equation U
, specify the values of parameter c for which operator U
0 = cU
is also unitary.
(c) Given unitary operator U
(d) Can a unitary operator (or matrix) be hermitian? Give two examples.
c in the form eiF where F is a corresponding Hermitian operator.
(e) Write operators P , Ta , M
Solution
c , P12 are unitary. For example, P = P , so that P P = 1 etc. K
is not unitary as it
(a) Operators P , Ta , M
is not linear.
(b)
U
= U
2U
= U
.
1 = U

22
(c)
0 U0 = |c|2 U
U
= |c|2 1 = 1 .
U
Therefore c = ei , where is any real number.
(d) Yes, it can if
U
= U
2 = 1.
U
2 = U 2 = , its eigenvalues are U = 1. So if the only eigenvalues of an operator are 1, it is
Since U
both hermitian and unitary. Examples: P , P12 and Pauli matrices.
(e) Inversion operator P satisfies

eiaP = cos a + i sin aP .

Choose a = /2. Using i = ei/2 we get

P = ei 2 (P 1) .

Translation operator
d

Ta = ea dx = eiap/~
.

Scale transformation: Mc (x) =

c (cx). Using formula

d

eax dx (x) = (ea x)

(see previous assignments) with a = ln c we get
c =
M
I replaced x by x
and

d
dx

d
ln c x dx

ce

i
1
= exp ln c x
p +
~
2



by i
p/~. Substituting now 1/2 = [
x, p]/(2i~) we finally derive


c = exp ln c i (
xp + px
)
M
2~

7. Discuss bound states in the potential U = [(x a) + (x + a)] using the Schrodinger equation in momentum
representation. In particular,
(a) What is the condition (threshold) for existence of the parity-odd bound state?
(b) Calculate the bound state energy () near the threshold and for large separation a between the delta-wells.
Solution
(a) We discussed even state in class. Odd state exists if


=
1 e2a .
Denoting = 2a we can write this equation as

= 2
a .
1 e
Since the left-hand-side is not smaller than one, solution exists only if

a > 1/2

23
(b) At the threshold
a = 1/2, = 0 and hence = 0. To find a finite contribution we need to expand at small
x and small = 2
a 1 > 0:
(1 + )(

2
),
2

which has a non-trivial solution = 2, or a = 2

a 1. Substituting 2 = 2m/~2 we find
() = (2
a 1)2

~2
.
2ma2

The case
a  1 (see my or yours lecture notes):
() =

m2

1 e2a
2
2~

8. Coherent states in the occupation number representation.

(a) Consider operator
= ea a ,
U
is unitary.
where a
|i = |i. Prove that U
= U
(
(b) Prove that a
U
a + ).
|0i.
(c) Show that |i = U
(d) Let us expand a coherent state in the occupation number representation:
X
|i =
Cn |ni .
n

Using the above results, compute Cn s and show that |Cn |2 satisfy the Poisson distribution. You will need
to use the following Weyl identity:

eA+B = e[A,B]/2 eA eB ,

[A,
B]]
= [B,
[A,
B]]
= 0.
provided that[A,

Solution
U
=
(a) U
1.
(b) Using [
a, a
] = 1 you can prove the following identity for an integer k:
a
(
a )k = (
a )k a
+ k(
a )k1
With its help we have

X
n  
X
X
1
n
(
a a
)n = a

(
a )k ( a
)nk
n!
k
n=0
n=0 k=0
X
n  
X
n  
X
X
n
n k
=
(
a )k ( a
)nk a
+
k(
a )k1 ( a
)nk
k
k
n=0
n=0

= a
a
U
ea

k=0

=a

k=0

X
X

1
1
=
(
a a
)n a
+
n(
a a
)n1 = ea a a
+ ea a
n!
n!
n=0
n=0

(
=U
a + ) .

24
(c)
|0i] = U
(
|0i = [U
|0i] .
a
[U
a + )|0i = U
|0i is an eigenstate of a
|0i.
Thus, U
: |i = U
(d) Let us expand a coherent state in the occupation number representation:
X
|0i .
|i =
Cn |ni Cn = hn|i = hn|U
n

Using the Weyl identity

[A,
B]]
= [B,
[A,
B]]
= 0.
provided that[A,

eA+B = e[A,B]/2 eA eB ,
we get

= e||2 /2 ea e a
U

Now, e a |0i = |0i and

hn|ea |0i = hn|

X
X
n
m m
m
|mi =
(
a ) |0i = hn|
m!
m!
n!
m=0
m=0

Finally,
2
n
Cn = e|| /2 .
n!

9. A particle of mass m is confined to a one-dimensional square potential:


0, 0 x a;
U (x) =
, otherwise .
At t = 0 its normalized wave function is
r
(x, t = 0) =

8
[1 + cos(x/a)] sin(x/a) .
5a

(a) What is the probability/probability density to find the particle with momentum p = 0 and with momentum
in the interval < p < +?
(b) Calculate the wave function of the particle in the energy representation hEm |i.
(c) What is the expectation value of energy hEi at t = 0 and t > 0?
Solution
(a) Wave function in the momentum representation is
Z a
Z
(p) = hpn |i =
dxhpn |xihx|i =
0

1
eipn x/~ (x)dx ,
a

where pn = ~kn = 2n~/a. The probability of p = 0:

Z a
2


1
32
(x)dx =
|(p = 0)|2 = 
2
5
a
0
Notice that eia~/a is not one of the momentum eigenfunctions.

X
n=

|(pn )|2 = 1

25
(b)
Z
hEn |i =

r Z a
 xn 
2
1
2
dxhEn |xihx|i =
dx sin
(x, t) = n1 + n2
a 0
a
5
5

(c)
hEi =

A2n En =

It does not depend on t (energy conservation).

4
1
4 2 ~2
E1 + E2 =
5
5
5ma2

26
V.

MOTION IN CENTRAL POTENTIAL

1. Potential energy of interaction of two particles of masses m1 and m2 is U (|r1 r2 |), where r1 and r2 are their
positions.
(a) Write down the Hamiltonian of the system.
(b) Introduce new vectors r = r1 r2 and R = (m1 r1 + m2 r2 )/(m1 + m2 ). What is the physical meaning of
r and R?
=L
1 + L
2 . Write L
in terms of r, R and the corresponding
(c) Operator of orbital angular momentum is L
derivatives.
=H
1 + H
2 in terms of r and R.
(d) Write the Hamiltonian H
(e) Using the method of separation of variables seek solution of S.E. in the form (r)(R) and determine
equations satisfied by (r) and (R).
Solution
(c)
=L
1 + L
2 = ir1 1 ir2 2
L
In terms of r and R we have
1 =

m1
R r ,
m1 + m2

2 =

m2
R r
m1 + m2

Therefore,
= ir r iR R ,
L
where the first term is the operator of the orbital angular momentum in the c.o.m. frame, while the second term
is the operator of the orbital angular momentum related to the motion of the center-of-mass.

2. For the following potentials derive a condition for existence of the nth bound s-state.
(a) U (r) = r4 for r > a and U = for r < a, > 0;

0a
2
2
(b) U (r) = (r2U+a
2 )2 , U0 ,a > 0. (Hint: make a substitution w = / r + a and r = a tan .)

Solution
(a) As in the one-dimensional case, new bound states emerge at = 0. For example, this happens as the
potential becomes deeper. The corresponding equation for the radial function is


~2

r
R 4R = 0.
2
2mr r
r
r
Changing variables = 1/r we obtain

~2 2
R() R() = 0 .
2m 2

Solution:
R() = A cos() + B sin() ,

2 =

2m
~2

27
In order that the wave function be finite at r ( 0) we have to require A = 0. Boundary condition at
r = a:
m
2 n2
=
~2 a 2
2

sin(/a) = 0 ,

(b) Making a substitution: w = / r2 + a2 and r = a tan we arrive at the following equation

d2 w
+ 2w = 0 ,
d 2

2 = 1 +

2mU0 a2
.
~2

Solution:
w() = A cos() + B sin() .
Now we apply two boundary conditions at r = 0 and r . At r 0, 0 thus
w(0) =

(0)
= 0,
a

A = 0.

At r , /2. Introduce = /2 and expand in : r = a tan(/2 ) a/,



a
w = B sin((/2 )) B sin(/2) cos(/2)
r
At large r w coincides with R, therefore in order that the wave function be finite we have to require that

= n ,
2

1+

2mU0 a2
= 4n2 .
~2

3. Calculate the degree of degeneracy g(n ) of the three-dimensional harmonic oscillator using the solution in the
spherical coordinates.
Solution
Energy eigenvalues are n = ~(2nr + ` + 3/2), nr = 0, 1, . . ., ` = 0, 1, . . .. Energy does not depend on magnetic
number m, therefore for each n and ` there are 2` + 1 independent wave functions. For a given even n = 2nr + `,
` can take only even values (because 2nr is always even). In this case
g(n) =

n
X

(2` + 1) =

`=0,2,4,...

n/2
X

(4k + 1) =

k=0

1
(1 + n)(2 + n) .
2

where I used ` = 2k. For an odd n, ` is always odd:

g(n) =

n
X

(n+1)/2

(2` + 1) =

`=1,2,3,...

(4k 1) =

k=1

1
(1 + n)(2 + n) ,
2

where ` = 2k 1. In both cases g(n) is the same.

4. For a particle moving in potential U = (r a).

(a) Derive an algebraic transcendental equation that determines the discrete energy levels for ` = 0.
(b) What is the condition for existence of at least one bound state?
(c) Repeat (a) for ` > 0.
(d) Derive an explicit expression for E(`) when ma/~2  1. Make sure to find the `-dependent term.

28
Solution
(a) Discrete spectrum corresponds to E < 0. Define
r

2m
~2

Consider first ` = 0. Solving S.E. with the appropriate boundary conditions yields at r 6= a:
= A sinh r ,

r < a,

= Ber ,

r > a.

Matching at r = a has been done before many times with the result
~2
= 1 e2a .
m

(5.1)

This is the desired transcendental equation for .

(b) Notice that the first solution appears only if ma/~2 1/2.

(c) Now, for ` 6= 0 we have for the function u = rR

u = AI`+1/2 (r) ,

r < a,

u = BK`+1/2 (r) ,

r > a,

where we took into account the boundary conditions. I and K are the modified Bessel functions. Matching
condition at r = a (check that these matching conditions can be applied either to R or to u):
0
0
BK`+1/2
(a) AI`+1/2
(a) =

2m
AI`+1/2 (a)
~2

BK`+1/2 (a) = AI`+1/2 (a) ,

where prime denotes derivative with respect to z = a. Excluding coefficients A and B gives
0
K`+1/2
(a)

K`+1/2 (a)

0
2m I`+1/2 (a)
+
~2
I`+1/2 (a)

(5.2)

If you reached this point you solved the problem.

One can further simplify (5.2) by means of the formula (the so-called Wronskian)
0
0
K`+1/2
(a)I`+1/2 (a) K`+1/2 (a)I`+1/2
(a) =

1
.
a

The result is
I`+1/2 (a)K`+1/2 (a) =

~2
2ma

(5.3)

(d) At ma/~2  1, for ` = 0 we can neglect the exponent in (5.1) with the result:
E0 =

m2
.
2~2

(5.4)

For ` > 0 we can use either (5.2) or (5.3). I will use (5.3). On its right-hand-side is a small number. Therefore
we need to choose the right asymptotic of the Bessel functions. At a ,
I`+1/2 (a)K`+1/2 (a)

1
 1,
2a

whereas at a 0 this product approach a constant. Thus, the right asymptotic is a . In order to find
the `-dependent term we need to expand the Bessel functions beyond the leading term keeping the next two
terms. The result is
I`+1/2 (a)K`+1/2 (a)

1
`(` + 1)

2a
4(a)3

29
So
~2
1
`(` + 1)
=

2a
4(a)3
2ma
If we neglect the `-dependent term we get 0 a = ma/~2 , which leads to (5.4). To find a correction write
= 0 + 1 with 1  0 . Expanding in 1 we derive:
1 = `(` + 1)

1
220 a2

Energy levels:
E(`) =

2 ~2
~2 `(` + 1)
2 ~2
0 1 ~2
m2
= 0
= 2 +
2m
2m
m
2~
2ma2

5. Prove that operator Tik = pi pk /m + k

xi x
k is an integral of motion (i.e. a conserved quantity) for a particle in
2.
the spherical harmonic oscillator. Show that Tik does not commute with L
Solution

6. For a hydrogen atom in the ground state calculate

(a) hrn i, where n is integer.
(b) Average kinetic and potential energy of electron.
(c) Probability density w(p) to find the electron with momentum p.
Solution
Hydrogen atom in the ground state.
(a)
0 = (a3 )1/2 er/a
hrn i =

rn |0 |2 dV =

(n + 2)!an
2n+1

(b)
Z
hU i =
0

e2
e2
|0 |2 dV =
r
a

Now, since hT i + hU i = 0 = e2 /2a, it follows that hT i = e2 /2a = hU i /2.

(c) Probability density w(p) to find the electron with momentum p:

Z
1
2 2~5/2
1
ipr/~
0 (p) =
e

(r)dV
=
0
(2~)3/2
a5/2 (p2 + ~2 /a2 )2
Therefore,
dw(p) = |0 (p)|2 d3 p

30
7. A particle is in two-dimensional potential U () = /, > 0, in a state with magnetic quantum number m
(projection of the orbital angular momentum on some axis).
(a) Derive an equation satisfied by the radial function R() using polar coordinates. Do the energy levels
depend on the sign of m?
(b) Calculate the discreet energy spectrum using the formulas on the back of this sheet.
(c) What is the degree of degeneracy of each level?
Solution
(a) Schr
odinger equation

~2 2
+ U =
2M

in polar coordinates takes form

1


+

1 2 2M
2 [U () ] = 0
2 2
~

Since U depends only on the wave function has form (, ) = R()(). By separation of variables we
obtain for the angular part
00 + m2 = 0 ,

() = ( + 2)

The periodic boundary condition requires that m be an integer. The normalized wave functions are
1
m = eim
2
For the radial part we have
m2
2M
1
R + R0 2 R + 2

~
00

+ R=0

(b) This equation is the same as the equation for u with U = /r in the list of formulas. We only need to
replace l + 1/2 |m|. Note that depends only on |m|, i.e. it is the same for m. Using the known
formula for the spectrum in the Coulomb potential we obtain:
En |m| =

M 2
2~2 (n + |m| + 1/2)2

As in the three-dimensional case we can introduce the principle quantum number N = n + |m| + 1:
EN =
(c) Degeneracy of each level is 2N 1.

M 2
,
1/2)2

2~2 (N

N = 1, 2, . . .

31
VI.

ANGULAR MOMENTUM

2 and L
z , find the explicit matrix form of operators L
y, L
z,
1. (a) In a representation with definite values of L
+, L
for l = 1.
L
in which its z-projection had a
(b) Earlier this semester we used another representation of the operator L
(b)

(a) . In the (b)

non-diagonal form. Denote that representation by L , while the representation of (a) by L
representation we showed that

0 0 0
0 0 1
0 1 0
(b)
(b)
z(b) = i~ 1 0 0 .
0 0 1 ,
0 0 0,
L
L
L
x = i~
y = i~
0 1 0
1 0 0
0 0 0
that transforms (b) to (a).
Determine the unitary transformation U
Solution
(a) Non-vanishing matrix elements:
h`m|L+ |`, m 1i = ~

h`, m 1|L |`mi = ~

(` + m)(` m + 1) ,

(6.1)

(` + m)(` m + 1) ,
1p
h`m|Lx |`, m 1i = h`, m 1|Lx |`mi = ~
(` + m)(` m + 1) ,
2
1p
(` + m)(` m + 1) ,
h`m|Ly |`, m 1i = h`, m 1|Ly |`mi = i~
2
h`m|Lz |`mi = ~m .

(6.2)
(6.3)
(6.4)
(6.5)

In matrix form

(a) = ~
L

0
1
2

1
2

0
1
2

0
1
2

(a)
L
+

0 i2
1

0 i2
= ~ i2
,
i
0
0
2

0
= ~ 0
0

(a)
L
y

2 0
0
2,
0 0

(a)
L
z

1 0 0
= ~ 0 0 0 ,
0 0 1

(a)
L

0 0 0
= ~ 2 0 0
2 0
0

(b) Earlier this semester we used the following representation

0 0 0
0 0 1
(b) = i~ 0 0 1 ,
(b) = i~ 0 0 0 ,
L
L
x
y
0 1 0
1 0 0

(b)
L
z

0 1 0
= i~ 1 0 0 .
0 0 0

(b)

Clearly, matrix L
z is not diagonal. A unitary matrix U that transforms (b) to (a) satisfies
(a) = U
(b) U
L

(6.6)

U
= 1 .
U

(6.7)

and

as follows:
Let us denote the matrix elements of U

a b c
=d e f ,
U
g h k

32
where a, b, . . . are complex numbers that we have to find.
(a) U
(b) . Substituting matrices L
=U
L
(a)
(b)
We can write eq. (6.6) in a more convenient form as L
and L
z
z
and multiplying them we obtain

a b
c
ib ia 0
0 0 0 = ie id 0
g h k
ih ig 0
Thus, e = d = c = k = 0, a = ib, g = ih and f is not fixed. To find the remaining coefficients a, h and
f , repeat this calculation for the y-components and apply eq. (6.7). The result is

=
U

1
2

0
12

i2 0
0 1
i2 0

2. In a state with a definite value of orbital angular momentum ` and its projection m on z-axis
(a) Prove that hLx i = hLy i = 0.



(b) Compute L2x , L2y .
(c) Let z 0 be an axis oriented at polar angle and azimuthal angle with respect to z. Prove that
hLz0 i = ~m cos ,

2
1
Lz0 = ~2 [l(l + 1) 3m2 ] sin2 + ~2 m2 .
2

Solution
:
(a) In view of properties of L
|`mi = 0 ,
h`m|L
It follows then that hLx i i hLy i = 0, so hLx i = hLy i = 0.
(b) Similarly, from
2 |`mi = 0
h`m|L
follow that
D
E

2

xL
y + L
yL
x = 0
Lx L2y i L
D
E



xL
y + L
yL
x = 0
Thus, L2x = L2y and L
On the other hand,
D
E D
E
2 L
2x + L
2y = L
2z = ~2 `(` + 1) ~2 m2
L
Therefore,

2 
2  ~2
[`(` + 1) m2 ]
Lx = Ly =
2
(c) We can write
=L
z ez + L
x ex + L
y ey ,
L

33
where ex ,ey ,ez are unit vectors. Now, projection on z 0 -axis Lz0 reads
ez 0 = L
z ez ez0 + L
x ex ez0 + L
y ey ez0
Lz0 = L
z cos + L
x sin cos + L
y sin sin
=L

(6.8)
(6.9)

Averaging over a state with fixed ` and m and using the results of 2b we derive
hLz0 i = hLz i cos
Taking square of (6.9), averaging and again using the results from 2b as well as form ??:




~2
L2z0 = L2z cos2 + L2x sin2 = ~2 m2 cos2 + [`(` + 1) m2 ] sin2
2
2
2
2 2
21
= ~ [l(l + 1) 3m ] sin + ~ m .
2

(6.10)
(6.11)

3. Suppose that a particle is in a state with the orbital angular momentum l = 1 and its projection m on z-axis.
Using solution to the previous home assignment, calculate the probabilities wm0 of orbital angular momentum
projections on z 0 -axis, which is oriented at polar angle and azimuthal angle with respect to z.
Solution
From the previous home assignment:
hLz0 i = ~m cos ,


1
L2z0 = ~2 [2 3m2 ] sin2 + ~2 m2 .
2

(6.12)

On the other hand,

hLz0 i = w2 (1~) + w0 (0~) + w2 (1~) ,

2
Lz0 = w2 (1~)2 + w0 (0~)2 + w2 (1~)2

This gives two equations for two unknowns w2 and w2 . Solving them we get


1 1
2
2
2
(2 3m ) sin + m + m cos ,
w2 =
2 2


1 1
2
2
2
w2 =
(2 3m ) sin + m m cos ,
2 2


1 1
2
2
2
w0 = 1 w2 w2 = 1
(2 3m ) sin + m
2 2

(6.13)
(6.14)
(6.15)

4. Spin-1/2 particle is in a state with definite l, m, sz . Calculate the probabilities of possible values of the total
angular momentum J = L + S.
Solution
S.
Thus
2 + S2 + 2L
There are two possible values of j: j = l 1/2. Now, J2 = L





2
1 1
J = ~2 l(l + 1) +
+ 1 + 2m .
2 2
= 21 . On the other hand

2
J = ~2 [wl+1/2 (l + 3/2)(l + 1/2) + wl1/2 (l + 1/2)(l 1/2)] ,
Solving these equations we derive
wl+1/2 =

l + 2m + 1
,
2l + 1

wl1/2 =

l 2m
.
2l + 1

wl+1/2 + wl1/2 = 1

34
= H0 + S n, where H0 and are constants. The operator S n
5. An electron is subject to the Hamiltonian H
represents the spin projection along the direction n. We can express this direction in spherical coordinate angles
as n = sin (cos ex + sin ey ) + cos ez , where ex , ey , ez are unit vectors.
(a) Find the possible values of the electron energy and prove that the corresponding spin wave functions +
and .
(b) At t = 0 the electrons spin wave function is an eigenvalue of Sz with eigenvalue +~/2. What are the
probabilities of finding the electron in the states you found in 1.
(c) Calculate the spin state of the system at later times.
(d) What is the probability of finding the electron again in the initial state?
(e) What is the probability of finding it with the inverted spin?
Solution
(a) The possible values of the electron energy are the eigenvalues of the Hamiltonian that can be written as


~
~
cos sin ei

H = H0 + [sin cos x + sin sin y + cos z ] = H0 +

2
2 sin ei cos
The eigenvalues of matrix


cos sin ei
sin ei cos

are found from the equation


 cos sin ei

 sin ei cos



=0


= 1

Therefore, the possible

values are = H0 ~/2. The corresponding normalized wave functions
 energy

a
are of the form =
with a and b satisfying
b
a(cos 1) + b sin ei = 0
2

(6.16)

|a| + |b| = 1

(6.17)

Using the identities

1 cos = 2 sin2 (/2)
1 + cos = 2 cos2 (/2)
sin = 2 sin(/2) cos(/2)
we find from (6.16) and (6.17)

+ =

cos 2 ei
sin 2


,

sin 2 ei
cos 2

(b) At t = 0 the electrons wave function is


(0) =

1
0

The probabilities w to find electron in the states are

w+ = |+
(0)|2 = cos2 (/2) ,

w = |
(0)|2 = sin2 (/2)

35
(c) {+ , } is a complete set of states. This means that spin wave function can be represented as a linear
combination of these states: (0) = c+ + + c , where c = h |(0)i. Therefore,
(0) = cos(/2) ei + sin(/2) ei
Time-evolution of the electron state:
(t) = ei+ t/~ cos(/2) ei + ei t/~ sin(/2) ei
(d) The probability of finding the electron again in the initial state is
w = | (0)(t)|2 = |ei+ t/~ cos(/2) ei cos(/2)ei ei t/~ sin(/2) ei ( sin(/2))ei |2
= |ei+ t/~ cos2 (/2) + ei t/~ sin2 (/2)|2 = |eit/2 cos2 (/2) + eit/2 sin2 (/2)|2
= | cos(t/2) i sin(t/2) cos |2 = cos2 (t/2) + sin2 (t/2) cos2
= 1 sin2 (t/2) sin2
(e)
w = 1 w = sin2 (t/2) sin2

6. Any 2 2 matrix A can be expanded as

.
A = a0 1 + ax
x + ay
y + az
z a0 + a
are Pauli matrices.
where
2a = Tr(
where Tr stands for trace, i.e. sum of diagonal elements.
A),
(a) Show that 2a0 = Tr(A),
n
n
if n is even, and (a )
n = an1 (a ),
if n ia odd.
= a 1,
(b) Prove that (a )
(c) Prove the following identities
a)(
b) = a b + i
(a b);
i. (
(
) = + i( );

ii.
Solution
(a)
.
A = a0 1 + ax
x + ay
y + az
z a0 + a
= 0 and Tr
Since Tr
1 = 2, and using the fact that trace is a linear operator
TrA = a0 Tr1 = 2a0 .
Next,
= a0 Tr()
A)
+ Tr[a
]
= aj ek Tr[
Tr(
k
j ]

= aj ek Tr[ijkl
l + jk 1] = aj ek 2jk = 2ak ek = 2a

(6.18)

2 = a2 1 ,
(
a )

(6.20)

(6.19)

(b) We want to prove that

where is a positive integer. Well do it by induction. For = 1

2 = aj ak
(a )
j
k = aj ak (ijkl
l + jk 1) = a2 1 ,

(6.21)

36
where in the last step we used the fact that jkl is anti-symmetric in jk, while aj ak is symmetric. Assume
now that (6.20) holds for some 1
2(1) = a2(1) 1.
(a )

(6.22)

2 = (a )
2(1) (a )
2 = a2(1) a2 1 = a2 1 ,
(a )

(6.23)

Lets prove that that it holds for .

where we used (6.21),(6.22).

The proof for odd nth is similar.
(c) The identities follow immediately from
i
j = iijk
k + ij 1.

= 2~
7. Spin-1 particle is subject to the Hamiltonian H
sz + 3
s2x . Disregard all other degrees of freedom.
(a) Calculate energy spectrum.
(b) If at time t = 0 the spin is in an eigenstate with sz = ~, calculate the expectation value of spin at time t.
Solution
= 2~
(a) H
sz + 3
s2x . Using the representation given in the list of Useful formulas we get

0 0 0
s2x = ~2 0 1 0
0 0 1
Thus,

0 2i~2 0
= 2i~2 3~2
0
H
0
0
3~2
are found from the equation
Eigenvalues of H

 2i~2
0

0
det  2i~2 3~2
 0
0
3~2




 = 0,



which can be written as

(3~2 )[( 3~2 ) 4~2 ] = 0
There are three solutions: 0 = 3~2 , + = 4~2 , = ~2 .
Seek for the eigenfunctions is the form

a
=b
c
For 0 = 3~2 :
a = b = 0,

c=1

0
0 = 0
1

37
For + = 4~2
c = 0,

a=

ib
2

Normalization: |a|2 + |b|2 = 1, so

2
b=
5

1 i
2
+ =
5
0
For = ~2
c = 0,

a = 2ib

Normalization: |a|2 + |b|2 = 1, so

1
b=
5

2i
1
= 1
5
0

(b) At time t = 0 the spin is in an eigenstate with sz = ~, which is given by

1 1
i
(0) =
2 0
This can be written as
(0) = 0 + + +
It is straightforward to determine the coefficients = 0, = i 310 , = i 110 . Now, the wave function at
any time t is
i t/~

i t/~
+
3e
+
2e

x
i
1
1
(t) = (3ei+ t/~ + ei t/~ ) = 6iei+ t/~ iei t/~ y .
10
5 2
5
2
0
0
Expectation value of spin at time t:
hsi = (t) s(t)
Substituting the corresponding spin matrices we get hsx i = hsy i = 0 and


i~
i~
~
18
2 5~t

hsz i =
(x y + x y ) =
2i Im (x y ) =
[16 + 9 cos(5~t)] = ~ 1
sin
50
50
25
25
2

38
VII.

MOTION IN MAGNETIC FIELD

1. Electromagnetic potentials A, are not observable quantities, unlike the electromagnetic fields E, B. Therefore,
Pauli equation must not change under the gauge transformation
A A0 = A + f ,

0 =

1 f
,
c t

(7.1)

with arbitrary f . On the other hand, Pauli Hamiltonian explicitly depends on electromagnetic potentials A,
. Therefore, in order that the (time-dependent) Pauli equation stays invariant under (1), the wave function,
which is also not an observable quantity, must transform concurrently with (1) as
0 = eig .

(7.2)

Determine function g for which the Pauli equation stays invariant under the transformations (1),(8.2).
Solution
Pauli equation:
i~

= H,
t

2 + e .
= 1 (p e A)
H
2m
c

(7.3)

Consider the following gauge transformation:

A A0 = A + f ,

0 =

1 f
,
c t

0 = eig .

(7.4)

The left-hand-side of the Pauli equation (1) transforms as

i~

i~
= i~eig
~ge
ig .
t
t
t

(7.5)

On the right-hand-side, a term involving the time derivative is ec feig . We expect that g is proportional to
f . Thus, we require that
e
~ge
ig = feig
c

g=

e
f.
~c

(7.6)

Now we need to check the terms involving the gradient, i.e. nabla operator :
e 0 0
e e
e
(p A
) = (p A
f )eig
(p A)
c
c
c
c
e e
ig
ig
= e (p A f ) + e (i~)ig
c
c

(7.7)
(7.8)

The two extra terms cancel out if

e
f + ~g = 0
c
which yields again g =

(7.9)

e
~c f .

2. For particle of mass m and charge e moving in magnetic field B,

(a) Compute the operator of velocity v and compare it with the classical expression.
(b) Calculate the commutators [
vi , vj ], where i, j = x, y, z and express them as a function of B. Can the three
components of v have definite values simultaneously?

39
Solution
Operator of velocity is given by
v =

d
r
1

= [
r , H]
dt
i~

(7.10)

v =

e
1
(p A)
,
m
c

(7.11)

Using (1) we find

which is the same as the corresponding classical formula.

Commutation relations:
[
vi , vj ] =

ie~
ijk Bk
m2 c

(7.12)

Because [
vi , vj ] 6= 0, components of velocity cannot have definite values simultaneously.

3. Spin-zero particle of mass M and charge e moves in constant magnetic field B = Bez .
(a) Choosing the vector potential in the symmetric gauge A = 12 (B r), calculate the energy spectrum.
2 = (
vx2 + vy2 )/ 2 . Show that
(b) Consider operators x
0 = x
vy /, y0 = y + vx /, r02 = x
20 + y02 and R
L
2
2
x
0 , y0 ] = [H,
r02 ] = [H,
R
L
L
[H,
0 ] = [H
] = [
r02 , R
] = 0,

[
x0 , y0 ] =

i~c
eB

(7.13)

2 .
(c) Find spectra of operators r02 and R
L
Solution
(a) In cylindrical coordinates the Pauli Hamiltonian reads
2
2 2
2
= p e~ B L
z + e B 2 H
+ pz .
H
2M
2M c
8M c2
2M

(7.14)

z , pz and H
all commute with each other, the eigenfunctions can be chosen in the following
Since operators L
form
Empz (, z, ) =

1
f ()ei(m+pz z/~) .
2 ~

(7.15)

Schr
odinger equation for f :


2
2M E
m2 1/4 emB
e2 B 2 2
f 00 + f 0 +

f = 0.

~2
2
~c
4~2 c2

(7.16)

This is the same equation as the one for the spherical harmonic oscillator. Denote
0 =

|e|B
,
Mc

a2 =

~
,
M 0

x=

2
.
2a2

(7.17)

Then, solution to (7.16) reads

f = Cex/2 x|m|/2 F (n , |m| + 1, x) ,

(7.18)

40
where F is the hypergeometric function with
n =

1
2

2E
em
+ |m| + 1
~0
|e|


.

(7.19)

Wave function is finite only if n = 0, 1, 2, . . .. Thus,

E,n = ~0 (n + 1/2),

n = 0, 1, 2, . . . ,

n = n +

|m| em/|e|
2

(7.20)

Notice, that the spectrum is a physical observable quantity independent of gauge.

(b)
(c) We can write

2 = 2H
R
L
M 02

(7.21)

= H
p2z /(2M ). Since En = (n + 1/2)~0 + p2z /(2M ), we find
where H
2
(RL
)n =

2
(n + 1/2)~0 = a2 (2n + 1) .
M 02

(7.22)

In the gauge A = (B r)/2 we can write

+ 2 e 0 L
z .
M 02 r02 = 2H
|e|

(7.23)

Hence eigenfunctions (7.15) are also eigenfunctions of r02 . Thus,

(
r02 )k = (2k + 1)a2 ,

k =n+

em
= 0, 1, 2, . . .
|e|

(7.24)

41
VIII.

STATIONARY PERTURBATION THEORY

1. One-dimensional electrically charged harmonic oscillator is placed in electric field of strength E. Calculate the
first two corrections to energy levels.
Solution
(0)

Unperturbed levels are En

element

= ~(n + 1/2). Potential energy of particle is V = qEx. The relevant matrix

r
hm|x|ni =

~n
m,n1 +
2m

~(n + 1)
m,n+1 .
2m

(8.1)

The first correction is

En(1) = Vnn = 0 .
The second correction:
En(2) =

X0 |hm|V |ni|2
m

(0)
En

(0)
Em

q2 E 2
2m 2

2. Particle moves in a two-dimensional potential U = k(x2 + y 2 )/2 + xy, || < k.

(a) Calculate exact energy spectrum and the degree of degeneration of each level.
(b) Treating V = xy as a perturbation find the first non-vanishing correction to the ground state energy.
Indicate the conditions for applicability of your result. Compare with (a).
(c) In the framework of the perturbation theory, calculate the splitting of the first excited level: determine
energy levels and the zeroth order eigenfunctions.
(d) Repeat (b) and (c) for V = xy(x2 + y 2 )/2 keeping only the first order in perturbation theory.
Solution
(a) We can write the potential as follows
U=

k1
k2
(x + y)2 + (x y)2
4
4

where k1,2 = k > 0. Now, introduce new variables

x+y
x1 = ,
2

y1 =

x + y

,
2

which is a rotation by /4 in the xy-plane. The Hamiltonian becomes

2
2
2
2
= ~ + k1 x21 ~ + k2 y12
H
2m x21
2
2m y12
2

This is an unperturbed two-dimensional harmonic oscillator with the following spectrum:

r
r
k+
k
En1 n2 = ~
(n1 + 1/2) + ~
(n2 + 1/2) ,
n1,2 = 0, 1, . . .
m
m

(8.2)

42
(b) The unperturbed spectrum is given by
(0)

En(0)
EN = ~(N + 1) ,
1 ,n2

n(0)
= nosc
(x) nosc
(y)
1 ,n2
1
2

N = 0, 1, 2, . . .. Under the perturbation V = xy the first order correction is

(1)

EN = 0
as in the previous problem. The second correction to the ground state:
(2)

E0

X 0 |hn1 , n2 |V |0i|2
(0)

n1 ,n2

(0)

E0 En1 ,n2

Using (8.1) we find that the only non-vanishing matrix element is

h11|V |0i =

~
2m

Thus,
(2)

E0

2 ~
8m2 3

To compare with the result of problem 1, note that the perturbation theory is valid when ||  m 2 = k.
Expanding (8.2) in powers of /k we obtain


p
~ p
~
2
E0 =
[ 1 + /k + 1 /k]
2 2 ,
2
2
4k
which agrees with the perturbation theory.
(c) Consider the first excited level, i.e. N = 1. It is two-fold degenerated. The corresponding unperturbed
(0)
(0)
wave functions are 10 and 01 . Now, we need to calculate
(0)

(0)

(0)

(0)

h10 |V |10 i = h01 |V |01 i = 0 ,

~
(0)
(0)
(0)
(0)
h10 |V |01 i = h01 |V |10 i =
2m

(8.3)
(8.4)

The corresponding secular equation is





(1)
~/2m 
 E1
 = 0.

 ~/2m E1(1) 
Its solution is
(1)

E1

~
.
2m

It indicates that the degeneracy is lifted. The corresponding wave functions read
1
(0)
(0)
(10
01 )
2
(d) V = xy(x2 + y 2 )/2. The oscillator wave functions are
2
2
1
00 = e( + )/2 ,

r
2 (2 +2 )/2
10 =
xe
,

r
2 (2 +2 )/2
01 =
ye
,

(0)

E0

= ~

(8.5)

(0)

= 2~

(8.6)

(0)

= 2~ ,

(8.7)

E1
E1

43
p
p
where = x m/~, = y m/~. The first excited level is two-fold degenerated.
First order corrections:
Z Z

(1)
|00 (, )|2 xy(x2 + y 2 ) dd = 0
E0 =
2

h10|V |10i = h01|V |01i = 0 ,
Z Z
h10|V |01i = h01|V |10i =


=

~
m

3/2

10 (, )01 (, )

xy(x2 + y 2 ) dd
2

3
4

The secular equation yields

E (1) = h10|V |01i

3. Particle moves in potential U = U0 aer/a /r, such that U0  ~2 /ma2 . Using the perturbation theory, calculate
the difference between the energy levels in this potential and in the Coulomb potential U = U0 a/r. Indicate,
the region of applicability of your result.
Solution
Similar problem was discussed in class. When U0  ~2 /ma2 the perturbation is


r2
r
V = U0 1
+
,
2a 6a2
which is obtained by subtracting the Coulomb potential U = U0 a/r from U = U0 aer/a /r and expanding
at r  a. Correction to the energy levels is


n2
3n2 l(l + 1)
(1)
2
+
[5n
+
1

3l(l
+
1)]
Enl = U0 1
4
12 2
where
= ma2 U0 /~2 .

4. Particle moves in one-dimensional potential U = F x for x 0 and U = for x < 0. Using the variational
method calculate the ground state energy:
(a) Argue that the following trial functions are a reasonable choice for the ground state
i. = Axex ,
2
ii. = Bxex /2 .
(b) For each trial function compute the ground state energy.
(c) Compare with the exact solution

En =

~2 F 2
2m

1/3
n+1 ,

n = 0, 1, . . .

where n are zeros of the Airy function. In particular, 1 2.338. Which trial function gives better
approximation?
Solution

44
(a) Particle moving in one-dimensional potential U = F x for x 0 and U = for x < 0 has E > 0. At x 0
Schr
odinger equation reads

~2 00
= E
2m

sin(kx) kx.

Therefore the trial function must also be proportional to x. At x the wave function must exponentially
decay. Both functions i and ii smoothly interpolate between these limits. Moreover, they do not have any
zeros between x = 0 and x = as it should be for the ground state.
(b) (i) = Axex .
Normalization:
Z

|(x)|2 dx = 1

A2 = 43

Compute the functional J():


~2 d2
~2 2
3F
(x)
+
F
x
(x)dx
=
J() =
+
.
2
2m dx
2m
2

(8.8)

Minimizing with respect to :

J 0 ()|=0 = 0

0 =

3mF
2~2

1/3

Therefore,

E0 = J(0 ) =

243
32

1/3 

~2 F 2
m

1/3


1.966

~2 F 2
m

1/3

(ii) = Bxex .
Repeating the same steps for (i) we have
r

1/3
3
3~2
2F
16m2 F 2
2
B =4
, J() =
+
, 0 =

4m
9~4

 2 2 1/3
 1/3  2 2 1/3
~ F
~ F
81
1.861
E0 =
4
m
m

(8.9)
(8.10)

(c) Exact solution


E0 = 2.338

~2 F 2
2m

1/3


= 1.856

~2 F 2
m

1/3

The variational approach gives an upper limit to the value of E0 . Therefore (ii) gives a better approximation.

5. Use the variational method to calculate the energy levels and the corresponding wave functions of the 1s and 2s
levels of the hydrogen atom.
Solution
Hamiltonian of the hydrogen atom:
2
2
= ~ 2 e .
H
2M
r

45
Since we are interested in the s-states, the corresponding wave functions are spherically symmetric, i.e. depend
only on r. For the ground state 1s we try the following wave function
1s = Aer .
This choice is motivated by: (i) large r behavior of any function satisfying the Schrodinger equation with the
above Hamiltonian (see lecture notes), (ii) the oscillation theorem that dictates that the ground state wave
function must have no zeros at 0 < r < , (iii) boundary condition at r = 0. Normalization of 1s yields
3
.

A2 =
Thus,
J1s () =

2 3 ~2
m

er (2 er )r2 dr 4 3 e2

e2r rdr =

~2 2
e2 .
2M

Minimum of J1s () is at 0 = 1/a, where a = ~2 /M e2 . Finally,

E1s = J1s (0 ) =

e2
,
2a

1s =

1
a3

er/a .

The wave function of the first excited state must have one zero. A reasonable choice is

r  r/a
2s = B 1 +
e
.
a
Orthogonality condition:

1s (r)2s (r)r2 dr = 0

1
(1 + ) .
3

Normalization condition yields:

B2 =

a3 (1

35
.
+ 2 )

The only remaining parameter is fixed by minimizing J2s given by



Z
e2
2
7 2
3

+
.
J2s () = 2s H2s d r =
a
2
6
2(2 + 1)
The result is
0 = 1/2 .
Thus,
E2s =

e2
,
8a

2s =

1
8a3

r  r/2a
e
.
2a