1

Physics 316 Solution for homework 6 Spring 2005

Cornell University
I. EXERCISE 1
Compute the position uncertainty x in the ground state
0(x) =
1
_

a
e

2
2
, =
x
a
, a =
_
h
m0
. (1.1)
for the harmonic oscillator potential. Use x =
_
(x x)
2
=
_
x
2
x
2
and the expectation values
x =
_

x|0(x)|
2
dx ,

x
2
_
=
_

x
2
|0(x)|
2
dx . (1.2)
(prepared by Steve Drasco)
First we will compute the average position x
x =
_

dx x|0|
2
= 0, (1.3)
since x|0|
2
is an odd function, and the integration boundaries are symmetric about the origin. Here we have used a
general rule which applies to any odd function f(x) = f(x)
_
A
A
dx f(x) =
_
0
A
dx f(x) +
_
A
0
dx f(x),
=
_
0
A
dx f(x) +
_
A
0
dx f(x),
=
_
A
0
dx f(x) +
_
A
0
dx f(x),
= 0. (1.4)
We now compute

x
2
_

x
2
_
=
_

dx x
2
|0|
2
,
=
1
a

dx x
2
e
Ax
2
, (1.5)
where we have dened A = a
2
. Now note that
x
2
e
Ax
2
=

A
e
Ax
2
, (1.6)
so that we have

x
2
_
=
1
a

A
_

dx e
Ax
2
,
=
1
a

A
A
1/2
,
=
a
2
2
, (1.7)
2
where we have used the known Gaussian integral
_

dx e
Ax
2
=
_

A
. (1.8)
To prove this relation let I =
_

e
Ax
2
. Then we have
I
2
=
_

dx
_

dy e
A(x
2
+y
2
)
,
=
_

0
dr
_
2
0
rd e
Ar
2
,
= 2
_

0
dr re
Ar
2
,
= /A, (1.9)
where in the second line we have converted from cartesian coordinates with area element dxdy, to polar coordinates
with area element rddr. These results, Eqs. (1.3) and (1.7), give
x =
_
x
2
x
2
=
a

2
0.7a. (1.10)
So on average, measurements should nd the particle at the origin, and these measurements should uctuate on a
scale of about 70% of the ground states classical maximum extension a.
II. EXERCISE 2
What is the expectation value V and the uncertainty V =
_
V
2
V
2
of the harmonic oscillator potential
V =
1
2
m
2
0
x
2
for the ground state wave function? This corresponds to the fact that the energy of a classical harmonic
oscillator is on average split in equal parts between kinetic an potential energy.
(prepared by Steve Drasco)
Using Eq. (1.7), we have V = m
2
0

x
2
_
/2 = a
2
m
2
0
/4. Then since a
2
= h/(m0), we have
V =
1
4
h0 =
E0
2
, (2.1)
where E0 = h0/2, is the ground state energy. We now compute

V
2
_
= m
2

4
0

x
4
_
/4. The expected value of x
4
is
given by

x
4
_
=
_

dx x
4
|0|
2
,
=
1
a

dx x
4
e
Ax
2
, (2.2)
where we have again dened A = a
2
. Now note that
x
4
e
Ax
2
=

2
A
2
e
Ax
2
, (2.3)
so that we have

x
4
_
=
1
a

2
A
2
_

dx e
Ax
2
,
=
1
a

2
A
2
A
1/2
,
=
3
4
a
4
, (2.4)
3
where in the second line we have again used the Gaussian integral (1.8). Then since a
4
= h
2
/(m0)
2
, this means that

V
2
_
=
3
16
h
2

2
0
=
3
4
E
2
0
, (2.5)
where again E0 = h0/2, is the ground state energy. This means that
V =
_
V
2
V
2
=
E0

2
0.7E0. (2.6)
So on average, measurements should nd the particles potential energy to be half of the ground state energy E0, and
these measurements should uctuate by about 70% of the ground state energy.
III. EXERCISE 3
A.
An electron is in a rectangular box with innite potential walls with width a, length b, and depth c. The potential is
chosen so that in the inside of the box there is V = 0. Determine the possible energy eigenvalues and the corresponding
eigenfunctions.
We want to nd the energy levels and the wavefunctions for a particle in a box of dimensions (a, b, c). In other
words, we want to solve the Schrodinger equation for this problem:

h
2
2m

2
E +V E = EE. (3.1)
The box is modelized by a 3D innite well:
V (x, y, z) =
_
0 if (x, y, z) [0, a1] [0, a2] [0, a3]
otherwise.
(3.2)
This problem can be solved using separation of variables. Quantized wavevectors (kx, ky, kz) = (
nx
a1
,
ny
a2
,
nz
a3
) are
introduced to satisfy the constraint E = 0 everywhere on the boundary of the box. Since you have seen the solution
in class, we just cite here the nal result. The eigenstates nx,ny,nz
= Enx,ny,nz
are of the form
nx,ny,nz
=
_
8
a1a2a3
sin (kxx) sin (kyy) sin (kzz) , (3.3)
and the corresponding eigenenergies are
Enx,ny,nz
=
h
2
k
2
nx,ny,nz
2m
=
h
2

2
2m
_
n
2
x
a
2
1
+
n
2
y
a
2
2
+
n
2
z
a
2
3
_
. (3.4)
B. (bonus)
Give an expression N(E) for the number of energy eigenstates per energy interval for energies much larger than
the ground state energy, i.e. there are N(E)dE energy eigenstates in the energy interval dE.
We now want to nd N(E), with N(E)dE dened as the number of eigenstates having an energy between E and
E +dE. This is done recalling the treatment we used to count the modes of an electomagnetic eld in a cavity.
Lets call N0 the number of particles with energy equal or less than E0:
N0 =
_
E0
0
N(E)dE. (3.5)
N0 can also be written as a sum of all possible eigenstates:
N0 =

nx=0

ny=0

nz=0
H(E0 Enx,ny,nz
), (3.6)
4
H being the Heaviside step function. The volume element in N-space is related to the volume element in k-space by
nxnynz =
a1a2a3

3 kxkykz. Hence, we get

N0 =

nx=0

ny=0

nz=0
H(E0 Enx,ny,nz
)nxnynz
=
a1a2a3

nx=0

ny=0

nz=0
H(E0 E(kx, ky, kz))kxkykz

a1a2a3

3
_

0
_

0
_

0
H(E0 E(kx, ky, kz))d
3
k. (3.7)
This integral describes 1/8th of the volume in a sphere with radius k0 =

2mE0/h,
N0 =
a1a2a3

3
4
3
k
3
0
=
a1a2a3
6
2

2mE0
3
h
3
. (3.8)
Therefore,
N(E) =
dN0
dE0
=
a1a2a3m

2
h
3

2mE . (3.9)
IV. EXERCISE 4
Show that the spherical nite well potential, V (r) = V0H(r R) with H(x) = if(x> 0, 1, 0), does not have a
spherically symmetric bound state if V0 <
h
2

2
8mR
2 .
We consider the spherical nite potential well V (r) = V0H(r R). In general, spherically symmetric states are
found with the radial part of the Schrodinger equation,

h
2
2m
u

(r) +V (r)u(r) = Eu(r), (4.1)

the actual wavefunction being given by (r) =
u(r)
r
. We can solve (4.1) in the two domains r < R and r > R
independantly and then make sure that u(r) and its rst derivative are continuous at r = R:

h
2
2m
u

1
(r) = Eu1(r) (4.2)

h
2
2m
u

2
(r) = (V0 E)u2(r). (4.3)
Both solutions are already known:
u1(r) = A1 sin(k1r) +B1 cos(k1r) (4.4)
u2(r) = A2e
k2r
+B2e
k2r
, (4.5)
with k1 =

2mE/h and k2 =
_
2m(V0 E)/h. In order to have
u1(r)
r
well dened at the origin, we must have
u1(0) = 0, that is B1 = 0. Since we want to be normalizable, we also need to put A2 = 0.
We now apply the continuity conditions:
u1(R) = u2(R) A1 sin(k1R) = B2e
k2R
(4.6)
u

1
(R) = u

2
(R) k1A1 cos(k1R) = k2B2e
k2R
. (4.7)
Therefore, dividing (4.7) by (4.6), one gets
k1 cot(k1R) = k2. (4.8)
k2 can be dened as
_
k
2
V
k
2
1
, where kV =

2mV0/h. The condition then becomes

k1
_
k
2
V
k
2
1
= tan(k1R). (4.9)
5
0 2 3
1
A B
C
k R
1
FIG. 1: Graphical solution of equation (4.10) for three dierent values of kV . A: kV R <

2
(too small for any possible solution
for k1R). B: kV R =

2
; this is the limiting value in order to get a solution. C: kV >

2
; At least one solution is possible (in
this example there are exactly three).
Since the left-hand side of this equation is always positive, we know that k1R can be a solution only if tan k1R is
negative, that is, in the ranges [

2
, ], [
3
2
, 2], ... Moreover, assuming that k1 and
_
k
2
V
k
2
1
are the sides of a right
triangle with hypotenuse kV , we can transform (4.9) into:
k1R
kV R
= | sin(k1R)|, k1R
_
(2n 1)
2
, n
_
, n = 1, 2, 3, . (4.10)
Figure 1 summarizes our results. From the gure, we conclude that there is no solution if
1
kV R
>
2

, that is,
V0 <
h
2

2
8mR
2
. (4.11)
It is worth pointing out that you could have obtained this result from previous calculations for the symmetric one-
dimensional square well. The mathematics is almost the same, except that here we ask that u(0) = 0. By working out
the details, you can show that this means our solutions here are the antisymmetric solutions from that problem. The
branches of gure 1 are the branches of the antisymmetric solutions of the square well. While the one dimensional
square well always has a bound ground state, which is symmetric, it does not always have the rst antisymmetric
state. Therefore the spherical potential well does not always have a bound state.