Physics 215 Final Exam Solutions Winter 2018

1. Consider a particle of mass m subject to a one-dimensional potential of the following form,

(
1
mω 2 x2 , for x > 0,
V (x) = 2
∞, for x ≤ 0.

(a) Determine the possible the bound state energy values of the particle.

The Schrödinger equation for this problem in the interval 0 < x < ∞ is,

~2 d 2
 
1 2 2
− + mω x ψ(x) = Eψ(x) . (1)
2m dx2 2
This is the Schrödinger equation for the one-dimensional harmonic oscillator, whose energy
eigenvalues and eigenfunctions are well known. However, in this problem, there is an infinite
barrier at x = 0, so we must impose an additional boundary condition, ψ(0) = 0.
The energy eigenvalues of the one-dimensional harmonic oscillator without the infinite bar-
rier at the origin are En = ~ω(n + 21 ), where n is any non-negative integer. The corresponding
energy eigenstates, ψn (x), are also eigenstates of parity and satisfy ψn (−x) = (−1)n ψn (x).
That is, the energy eigenstates are parity even for even n and parity odd for odd n. In
particular, if n is odd then ψ(0) = 0. In contrast ψn (x) 6= 0 for even values of n.
We can see this behavior explicitly from the expressions for the harmonic oscillator wave
functions. From eq. (B.4.3) of Sakurai and Napolitano,

mωx2
 
−1/2 mω
 1/4 p
n
ψn (x) = (2 n!) exp − Hn (x mω/~) , (2)
π~ 2~

where H0 (z) = 1, H1 (z) = 2z, . . . are the Hermite polynomials. The recursion relation satisfied
by the Hermite polynomials, given in eq. (2.5.25) of Sakurai and Napolitano,

Hn+1 (z) = 2zHn (z) − 2nHn−1 ,

can be used to compute Hn (z) for all n = 2, 3, . . ., given H0 (z) = 1, H1 (z) = 2z. It immediately
follows that Hn (0) = 0 for odd n and Hn+2 (0) = −2(n + 1)Hn (0) for even n, which yields
H2n (0) = (−2)n (2n − 1)!! 6= 0 for all non-negative n, after noting that H0 (0) = 1.
We conclude that only the odd parity harmonic oscillator wave functions vanish at the
origin. These are the allowed square integrable solutions to eq. (1) subject to the boundary
condition ψ(0) = 0. The corresponding energy eigenvalues are En = ~ω(n+ 21 ) for odd positive
integers n. Writing n = 2N + 1, we conclude that the possible bound state energies are

EN = ~ω(2N + 32 ) , for N = 0, 1, 2, . . .

1
 (b) What is the normalized ground state wave function in the coordinate representation?

The ground state wave function for this problem is proportional to the n = 1 parity odd
energy eigenstate of the one-dimensional harmonic oscillator. Using eq. (2) with H1 (z) = 2z,

mωx2
  
Ax exp −
 , for x > 0,
ψN =0 (x) = 2~

0, for x < 0.


where A is determined by the normalization condition,

Z ∞
|ψN =0 (x)|2 dx = 1 . (3)
−∞

For our problem, due to the fact that ψN =0 (x) = 0 for x < 0, the square of the normalization
constant A must be twice that of the corresponding n = 1 harmonic oscillator wave function
in order to satisfy eq. (3), i.e., |A|2 = 2|AHO |2 . Since eq. (2) implies that (for n = 1),
 1/4 
4 mω 3/4
AHO = ,
π ~

it follows that normalized ground state wave function for our problem is,
 1/4
mωx2
 
 16  mω 3/4

 x exp − , for x > 0,
ψN =0 (x) = π ~ 2~


 0, for x < 0.

after setting the overall phase (by convention) to unity. One can verify explicitly that this
result satisfies eq. (3).

2. Consider a one-dimensional problem with a Hamiltonian,

P2
H= + V (X) , (4)
2m
where X is the position operator and P is the momentum operator. Assume that the spectrum
of the Hamiltonian is discrete. Denote the normalized energy eigenstates by |ni, and the
corresponding energy eigenvalues by En .
(a) Show that the matrix elements of X and P , with respect to a basis consisting of
orthonormal energy eigenstates, satisfy the following relation,

hk| P |ni = c(k, n) hk| X |ni , (5)

where c(k, n) depends on Ek and En . Determine an explicit expression for c(k, n).

2
We begin by considering the commutator [H, X], where H is given by eq. (4). Using the
canonical commutation relation, [X, P ] = i~I, and noting that X commutes with any function
of X, it follows that
 2 
P P P i~
[H, X] = ,X = [P, X] + [P, X] =− P, (6)
2m 2m 2m m
after employing the commutator identity, [AB, C] = A[B, C] + [A, C]B. Hence, it follows that
im
P = [H, X] .
~
Inserting this result into hk| P |ni yields,
im im im
hk| P |ni = hk| [H, X] |ni = hk| (HX − XH) |ni = (Ek − En ) hk| X |ni ,
~ ~ ~
where we have used the fact that |ki and |ni are energy eigenstates, e.g., H |ni = En |ni.
Comparing with eq. (5), we conclude that
im
c(k, n) = (Ek − En ) .
~

(b) Evaluate the infinite sum,

mX  2
S≡ 2
(Ek − En )hk| X |ni , (7)
~ k

where the sum is taken over the complete set of energy eigenstates, |ki. Show that the sum
is equal to a dimensionless constant and determine its value.

Note that the equation H |ni = En |ni implies that the energy eigenstates |ni are dimension-
less. Moreover, X has the same dimensional units as ~/P (cf. the canonical commutation
relations). Hence, it follows that S is dimensionless since P 2 /m has the dimensions of energy.
It is convenient to introduce the following shorthand notation,
~ωkn ≡ Ek − En , Xkn ≡ hk| X |ni .
Then, the result of part (a) can be rewritten as,
hk| P |ni = imωkn Xkn . (8)
Since X is a self-adjoint operator, it follows that Xkn
∗
= Xnk . Hence, eq. (7) can be written
as,
mX
S= ωkn Xkn Xnk . (9)
~ k
Noting that ωkn = −ωnk , we can rewrite eq. (9) as,
mX
S= (ωkn − ωnk )Xkn Xnk . (10)
2~ k

3
 We now can use eqs. (8) and (10) to obtain,
!
i X X
S=− Xnk hk| P |ni − hn| P |ki Xkn . (11)
2~ k k

Inserting the definitions of Xnk and Xkn in eq. (11) and employing the completeness relation,
P
k |ki hk| = I (where I is the identity operator), it follows that
!
i X X
S=− hn| X |ki hk| P |ni − hn| P |ki hk| X |ni
2~
k k
i 1
= − hn| (XP − P X) |ni = ,
2~ 2
after using [X, P ] = i~I and hn|ni = 1 (since the energy eigenstates are normalized to unity).

REMARK:
The identity,
2m X hk| X |ni2 = 1 ,
 
(Ek − En ) (12)
~2 k
is called the Thomas-Reiche-Kuhn sum rule, and it plays an important role in quantum
radiation theory. It is mentioned on p. 368 of Sakurai and Napolitano, where another method
of derivation is proposed. Sakurai and Napolitano suggest considering,
  i~ ~2 ~2
hn| X, [X, H] |ni = hn| [X, P ] |ni = − hn|ni = − , (13)
m m m
after making use of eq. (6) and [X, P ] = i~I. Then after inserting a complete set of states in
appropriate places,

hn| X, [X, H] |ni = hn| (X 2 H − 2XHX + HX 2 |ni

 

X
= hn| X |ki hk| XH |ni − 2 hn| XH |ki hk| X |ni + hn| HX |ki hk| X |ni
k
X
=2 (En − Ek )| hk| X |ni |2 , (14)
k

after using H |ki = Ek |ki and H |ni = Ek |ni. Comparing eqs. (13) and (14), we end up with
eq. (12).
There are numerous related sum rules that are similar to the Thomas-Reiche-Kuhn sum
rule, which can be derived with similar techniques. For more details, see pp. 211–217 of Hans
A. Bethe and Roman Jackiw, Intermediate Quantum Mechanics, 3rd edition (Westview Press,
Boulder, CO, 1997).

(c) Verify your calculation of S in part (b) in the case of the one-dimensional harmonic
oscillator. That is, assuming V (X) = 21 mω 2 X 2 , compute explicitly hk| X |ni and then use the
energy eigenvalues of the one-dimensional harmonic oscillator to explicitly evaluate S.

4
Using eq. (2.3.25a) of Sakurai and Napolitano,
r
~ √ √
 
hk| X |ni = n δk,n−1 + n + 1 δk,n+1 .
2mω
Thus, for fixed n, the only two non-zero values of hk| X |ni are,
r r
~n ~(n + 1)
hn − 1| X |ni = , hn + 1| X |ni = . (15)
2mω 2mω
Note that if n = 0, only the second matrix element above is relevant.
The energy eigenvalues of the one-dimensional harmonic oscillator are En = ~ω(n + 12 ).
Hence, it follows that
En−1 − En = −~ω , En+1 − En = ~ω . (16)
Once again, if n = 0, only the second energy difference above is relevant.
Employing eqs. (15) and (16) in evaluating the sum S [eq. (7)], only two terms in the sum
survive if n 6= 0,  
m ~   1
S= 2 ~ω −n + (n + 1) = .
~ 2mω 2
If n = 0, we replace −n + (n + 1) with 1 since only one term contributes to the sum. The end
result is the same, and we again confirm that S = 21 .

3. One would perhaps conclude from the lack of angular momentum in the ground state of
the hydrogen atom that the electron is stationary.
(a) To show that this is not so, calculate the probability that the electron’s momentum, if
measured, would be found to lie in a momentum element d3 p centered at momentum p ~.

The wave function for the ground state of hydrogen is

1 −3/2
ψ(r, θ, φ) = √ a0 e−r/a0 ,
π
where a0 ≡ ~2 /(µe2 ) is the Bohr radius and µ is the reduced mass. To compute the probability
that the electron has momentum between p ~ and p ~ + d3 p, we need to compute the momentum
space wave function,
1
Z
φ(~p) = d3 x e−i~p ·~x/~ ψ(~
x) .
(2π~)3/2
To perform the above integral, we shall make use of spherical coordinates and choose the
z-axis to lie along p
~. Then, p x = pr cos θ, where r ≡ |~
~ ·~ x|. Hence,
1 1
Z
φ(~
p) = r 2 dr d cos θ dφ e−ipr cos θ/~ e−r/a0
(2π~) (πa30 )1/2
3/2

Z 1
1 1
Z ∞
2 −r/a0
= 2π r e dr d cos θ e−ipr cos θ/~
(2π~)3/2 (πa30 )1/2 0 −1

1 2~ pr
Z ∞  
= 3
r e −r/a0
sin dr . (17)
(2π 2 ~3 a0 )1/2 p 0 ~

5
To evaluate the remaining integral, consider the following function,
Z ∞
F (a) = e−ay sin(my)dy ,
0

where a > 0 (the latter is needed for the integral to be convergent). Then,
   
1 a + im m
Z ∞ Z ∞
−ay imy
F (a) = Im e e dy = Im e−(a−im)y
dy = Im = Im 2 2
= 2 .
0 0 a − im a +m a + m2
Taking the derivative of the above equation and multiplying by −1 yields,
dF 2am
Z ∞
− = y e−ay sin(my)dy = 2 , for a > 0.
da 0 (a + m2 )2
Applying this result to eq. (17), we end up with
√  a0 3/2 1
φ(p) = 8 π   a p 2 2 . (18)
2π~ 0
1+
~
Note that φ(~
p) depends only on the magnitude p ≡ |~ p|. That is, the momentum space ground
state wave function is spherically symmetric (which is also true for its coordinate space wave
function).
The probability that the electron lies between p
~ and p~ + d3 p is given by |φ(~
p)|2 d3 p.

REMARK 1: A useful check of our calculation is to verify that the momentum space wave
function is normalized to unity (this is a consequence of Parseval’s theorem of Fourier analysis).
 a 3 Z ∞ p2 dp
Z Z ∞
3 2 2 2 2 0
d p |φ(~
p)| = 4π p dp |φ(~ p)| = 256π   a p 2 4 .
0 2π~ 0 0
1+
~
The integral is recognized as a Beta function,1
Γ(p)Γ(q) ∞
up−1 du
Z
B(p, q) = = .
Γ(p + q) 0 (1 + u)p+q
After a change of variables, u = x2 , it then follows that
Γ 12 (p + 1) Γ r − 12 (p + 1)



xp dx
Z ∞
= . (19)
0 (1 + x2 )r 2Γ(r)
√
Recall that Γ(n) = (n − 1)! for positive integer n, Γ( 21 ) = π and Γ(x + 1) = xΓ(x). It follows
that
x2 dx π
Z ∞
2 4
= .
0 (1 + x ) 32
One then readily verifies that Z
d3 p |φ(~
p)|2 = 1 .
1
See, e.g., N.N. Lebedev, Special Functions and Their Applications (Dover Publications, Inc., New York,
1972).

6
REMARK 2: The following alternative derivation of eq. (17) is especially noteworthy, since
this method is quite useful in a number of applications. We begin with
1 1
Z ∞ Z
2 −r/a0 ~
φ(~p) = 3/2 3 1/2
r e dr dΩ e−ik·~x , (20)
(2π~) (πa0 ) 0

where ~ ~/~. Using the expansion of the plane wave in terms of spherical waves,2
k≡p
ℓ
∞ X
i~
X
e k ·~
x
= 4π iℓ Yℓm (r̂)Yℓm
∗
(k̂)jℓ (kr) ,
ℓ=0 m=−ℓ

where r̂ ≡ ~
x/r (with r ≡ |~
x). Integrating over the solid angle Ω = (θ, φ), where θ and φ are
the polar and azimuthal angles of r̂ with respect to a fixed z axis,
Z ∞ X ℓ Z
i~
X
k ·~
x ℓ ∗
dΩ e = 4π i Yℓm (k̂)jℓ (kr) dΩ Yℓm (r̂) , (21)
ℓ=0 m=−ℓ

Recall that the spherical harmonics satisfy the orthonormality relation,

Z
dΩ Yℓm(r̂)Yℓ∗′ m′ (r̂) = δℓℓ′ δmm′ .

Using the fact that Y00 (r̂) = (4π)−1/2 , it follows that

Z √
dΩ Yℓm (r̂) = 4π δℓ0 δm0 .

Using this result in eq. (21), it follows that

4π sin(kr)
Z
~
dΩ eik·~x = 4πj0 (kr) = . (22)
kr
Taking the complex conjugate of eq. (22) and inserting this result into eq. (20) (with k = p/~)
yields
1 2~ ∞ −r/a0  pr 
Z
φ(~
p) = r e sin dr ,
(2π 2 ~3 a30 )1/2 p 0 ~
which confirms the result of eq. (17).

(b) What are the electron’s mean kinetic and potential energy?

To compute the mean kinetic energy, it is convenient to perform the calculation in momentum
space. Thus,
128π 2  a0 3 ∞ p4 dp
  Z
p
~2 p
~2 2π ∞ 4
Z Z
2 3 2
= |φ(~
p)| d p = p dp |φ(~
p)| =   a p 2 4 .
2µ 2µ µ 0 µ 2π~ 0 0
1+
~
Employing the result of eq. (19), we end up with,
~2
 
p
~2
= . (23)
2µ 2µa20
2
See the class handout entitled, Expansion of plane waves in spherical harmonics.

7
 Next, we turn to the computation of the mean potential energy. The potential energy of
the hydrogen atom is
e2
V (r) = − .
r
Hence, the mean potential energy is
 

 2 1
V (r) = −e .
r
It is straightforward to compute
  Z
1 2 1 3 1 1
Z ∞
= |ψ(~ x)| d x = 4π 3 r e−2r/a0 dr = ,
r r πa0 0 a0
after making use of the well known integral
Γ(n)
Z ∞
xn−1 e−ax dx = n , (24)
0 a
under the assumption that a > 0. For positive integer n, Γ(n) = (n − 1)!.
Hence, it follows that

 e2 ~2
V (r) = − = − 2 ,
a0 µa0
where we have replaced e2 = ~2 /(µa0 ) using the definition of the Bohr radius, a0 . Comparing
with eq. (23), we obtain  
p
~2
= − 12 V (r) .


(25)
2µ

REMARK 3: An alternative calculation of the mean kinetic energy

One can also compute the mean kinetic energy by evaluating the expectation value in
position space,   Z  2 
p
~2 ~ ~2
= ψ (~ ∗
x) − ∇ ψ(~ x) d3 x .
2µ 2µ
Note that in an ℓ = 0 state, ψ(~
x) depends only on the radial coordinate r ≡ |~
x|. Hence,
!
2 ∂ 2
2 ∂ ~2
L
 2
∂ 2 ∂

~
∇ ψ(~ x) = + − 2 ψ(r) = + ψ(r)
∂r 2 r ∂r r ∂r 2 r ∂r

~ 2 is a purely angular differential operator. Hence,

since L
   2 Z ∞  2  
p
~2 ~ 1 d 2 d
= 4π − 3 2
+ e−r/a0
r 2 e−r/a0 dr
2µ 2µ πa 0 dr r dr
 2 0 Z ∞
~2

~ 1 1 2 2 −r/a0
= 4π − − r e dr = ,
2µ πa30 0 a20 ra0 2µa20
after employing eq. (24).

8
 (c) Show that the results of part (b) are consistent with the Virial Theorem.

Using the results of problem 4 of Problem Set 2, the quantum Virial Theorem is,
* +
P~2 D E
= X ~ · ∇V
~ . (26)
µ

For a spherically symmetric potential,

~ = r dV .
~ · ∇V
X
dr
In particular, for the Coulomb potential, V (r) = −e2 /r,

dV
r = −V .
dr
Hence, the eq. (26) yields * +
~2
P
= − 12 V (r) ,


(27)
2µ
which confirms the result of eq. (25) obtained above.

4. Consider the Hamiltonian for a free particle of mass µ that is confined to a two dimensional
cylindrical surface embedded in three dimensions. The cylinder is infinite in extent with a
radius R centered around a symmetry axis that coincides with the z-axis. The Hamiltonian
for this problem is the sum of a kinetic energy term and a rotational kinetic energy term,

Pz2 L2z
H= + , (28)
2µ 2µR2
where Pz and Lz are the z-components of the momentum and angular momentum operators,
respectively.

(a) The Hamiltonian given by eq. (28) is invariant under which (continuous) symmetries?
Identify the complete set of mutually commuting observables for this problem.

A particle confined to the surface of a infinite cylinder centered around the z-axis possesses two
symmetries: (i) translations in the z direction; and (ii) rotations about the z-axis. Translations
in the z direction are generated by Pz and rotations about the z axis are generated by Lz .
Thus, we expect that
[H, Pz ] = [H, Lz ] = 0 . (29)
Indeed, eq. (29) is a consequence of the commutation relations,

[Pi , Lj ] = i~ǫijk Pk . (30)

9
Eq. (30) is just the statement that P ~ is a vector operator. Moreover, eq. (30) implies that
[Pz , Lz ] = 0. It follows that
 2
L2z
  
Pz 1 2 1
[H, Pz ] = + , Pz = [L , Pz ] = Lz [Lz , Pz ] + [Lz , Pz ]Lz = 0 .
2µ 2µR2 2µR2 z 2µR2
Likewise,
Pz2 L2z
   
1 2 1
[H, Lz ] = + , Lz = [P , Lz ] = Pz [Pz , Lz ] + [Pz , Lz ]Pz = 0 .
2µ 2µR2 2µ z 2µ
On the other hand, a similar computation implies that [H, Px ] 6= 0, [H, Py ] 6= 0, [H, Lx ] 6= 0 and
[H, Ly ] 6= 0, where we have again made use of eq. (30). It therefore follows that the complete
set of mutually commuting observables for this problem consists of the three operators, H, Pz
and Lz .

(b) Employing cylindrical coordinates (ρ, φ, z), where x = ρ cos φ and y = ρ sin φ, the
surface of the cylinder corresponds to ρ = R. Hence, the Hamiltonian of eq. (28) describes an
effective two dimensional problem with coordinates z and φ. Write down the time-independent
Schrödinger equation in the coordinate representation for this problem, and determine the
allowed energy eigenvalues and the corresponding eigenfunctions of the Hamiltonian H.

In the coordinate representation,

h~ ~ |ψi = −i~∇ψ(~
x| P ~ x) , h~ ~ |ψi = −i~ ~
x| L ~ x) .
x × ∇ψ(~

Employing cylindrical coordinates,

∂ψ ∂ψ
hρ, φ, z| Pz |ψi = −i~ , hρ, φ, z| Lz |ψi = −i~ . (31)
∂z ∂φ
Note that the operator representation of Lz is the same as in two dimensions with polar
coordinates. Indeed, in class, I showed that for x = ρ cos φ and y = ρ sin φ,
∂ ∂x ∂ ∂y ∂ ∂x ∂ ~ z.
= + = −y +x = (~
x × ∇)
∂φ ∂φ ∂x ∂φ ∂y ∂φ ∂y
Moreover, the fact that ρ = R is fixed does not alter the results quoted in eq. (31).
Hence, starting from eq. (28), the the time-independent Schrödinger equation in the coor-
dinate representation is given by
~2 ∂ 2 1 ∂2
 
− + ψ(z, φ) = Eψ(z, φ) . (32)
2µ ∂z 2 R2 ∂φ2
One can solve this equation using the separation of variables technique. Equivalently, in light
of [H, Lz ] = 0, one can choose ψ(z, φ) to be a simultaneous eigenstate of H and Lz . This
means that one can replace L2z in eq. (28) with its eigenvalue, m2 ~2 , where m is any integer.
Hence,
ψ(z, φ) = Z(z)eimφ , (33)

10
where we are using the same eigenfunctions of Lz that were obtained in class when we examined
spherically symmetric two-dimensional problems using polar coordinates.
Inserting eq. (33) into eq. (32), we obtain an ordinary differential equation for Z(z),

~2 d 2 m2
 
− − 2 Z(z) = EZ(z) . (34)
2µ dz 2 R

Writing E = ~2 k 2 /(2µ), eq. (34) yields

 2
m2

d 2
+ k − 2 Z(z) = 0 . (35)
dz 2 R

We consider two cases. First, if k 2 < m2 /R2 , then the solution to eq. (35) is
(  1/2 ) (  1/2 )
m2 m2
Z(z) = A exp z − k2 + B exp −z − k2 ,
R2 R2

where A and B are constants. But, this solution is not acceptable since it is exponentially
growing either as z → ∞ if B = 0 or as z → −∞ if A = 0.
Second, if k 2 > m2 /R2 , then the solution to eq. (35) is
(   ) (  )
2 1/2 2 1/2

m m
Z(z) = A exp iz k 2 − 2 + B exp −iz k 2 − 2 ,
R R

where A and B are constants. This solution corresponds to plane waves propagating in the
positive and negative z directions, as one would expect for a free particle.
Thus, we have found that the possible energy eigenvalues are any non-negative value that
satisfies,
~2 k 2 ~2 m2
E= ≥ ,
2µ 2µR2
for any integer value of m, with the corresponding eigenfunction,
" (   ) (  )#
2 1/2 2 1/2

2µE m 2µE m
ψ(z, φ) = eimφ A exp iz − 2 + B exp −iz − 2 .
~2 R ~2 R

11