Problem Set 5 Solutions

8.04 Spring 2013 March 12, 2013

Problem 1. (10 points) The Probability Current

We wish to prove that

dPab
= J(a, t) J(b, t). (1)
dt
Since Pab (t) is the probability of nding the particle in the range a < x < b at time t it is
mathematically equal to
b b
Pab (t) = |(x, t)|2 dx = (x, t)(x, t)dx. (2)
a a

Its time derivative is therefore given by

b b b
dPab d
= (x, t)(x, t)dx = [ (x, t)(x, t)] dx = + dx, (3)
dt dt a a t a t t
where we were allowed to take the time derivative inside the integral because the integral is
in time whereas the integral is over space. Note, though, that the total derivative became a
b
partial derivative when we took it inside the integral, because whereas a (x, t)(x, t)dx
is a function of t only (since x has already been integrated over), is in general a function
of both x and t, and we only want to take its time derivative.

We now make use of the fact that regardless of what the wavefunction (x, t) happens
to be, it must obey the Schrodinger equation:
2
(x, t) 2 (x, t)
i = + V (x)(x, t) (4a)
dt 2m x2
(x, t) 2
2 (x, t)
i = 2
+ V (x) (x, t) (4b)
dt 2m x
We can thus eliminate the partial derivatives in Equation 3:
b
dPab 1 2 2 2
2
= 2
+ V 2
+ V dx (5a)
dt i a 2m x 2m x
b 2 2 2
1
= 2 2
+ ( V V ) dx (5b)
i a 2m x x
b
i 2 2
= 2 dx, (5c)
2m a x x2
where in the last equality we used the fact that V = V because in position space
the potential energy operator commutes with both and . To proceed, we integrate the
two remaining terms by parts. Recall that
b b
g b f
f dx = f g g dx. (6)
a x a a x

1
This means
Z b Z b
2
2
    
dPab i i
= dx = dx (7a)
dt 2m a x2 x2 2m a x x x x
Z b Z b
b
    
i b
=   dx + dx (7b)
2m x a x a a x x a x x
b
    
i b
=  (7c)
2m x a x

a


    
i
= , (7d)
 
2m x x x=a x x x=b
 

so if we let

 
i
J(x, t) , (8)
2m x x
then
dPab
= J(a, t) J(b, t), (9)
dt
which is what we were trying to prove. From this equation, we can see that the units of
J(x, t) must be (time)1 , because Pab is a probability and is therefore a pure number, and t
has units of time.

2
Problem Set 5 Solutions
8.04 Spring 2013 March 12, 2013
Problem 2. (20 points) Visual Observation of a Quantum Harmonic Oscillator

(a) (5 points) The energy of a classical harmonic oscillator is given by

1
E = m02 A2 , (10)
2
where 0 is the angular frequency, m is the mass, and A is the amplitude of the
oscillation. The quantum harmonic oscillator1 , on the other hand, has energy
 
1
En = ~0 n + . (11)
2
Equating these expressions and rearranging gives
1
2
m02 A2 1 m0 2 m0 2
n = A = A , (12)
0 2
where we have converted to using the frequency 0 = 0 /2 instead of the angular
frequency 0 , and have assumed that n is so large that n + 12 n (we can check
our result against this assumption later). Plugging in m = 1012 g, 0 = 103 Hz, and
A = 103 cm, we get
n 3 1012 . (13)
From this result we see that we were justied in neglecting the 1/2 term in n + 12 .
(b) (5 points) If this oscillator were in its ground state, its energy would be
1
E0 = ~0 = 0 = 2.1 1012 eV, (14)
2
where once again 0 = 103 Hz. This is smaller than the average thermal energy of air
molecules (25 meV) by a factor of
Eair 25 103 eV
= 1010 . (15)
E0 2.1 1012 eV

(c) (5 points) We can perform the same manipulations as we did in part (a), except this
time we cannot neglect the 1/2 term in the oscillators energy, since we are dealing
with the ground state:

1 1 2 2 ~ ~
E0 = 0 = m0 A A= = . (16)
2 2 m0 2m0
1
There is really no dierence between a classical harmonic oscillator and a quantum one the Universe is
governed by quantum mechanics, so in principle all oscillators are quantum mechanical. The only dierence
is whether or not the quantum eects are obvious.

3
 Note that this is width of the Gaussian ground state wavefunction. Numerically, this
comes out to be A = 0.00401 nm, which is much smaller than the wavelength of visible
light:
400 nm
= 105 . (17)
A 0.00401 nm
(d) (5 points) I would not recommend award of a grant to carry out this research. From
part (a), we can see that the experimenters proposed system corresponds to a high
quantum number state, which means the experiment is unlikely to really probe quan
tum mechanical eects. At best, it would be able to investigate the approach of the
quantum system to the classical limit, but even this is unlikely, given the results of
(b) and (c). In part (b), we saw that air molecules possess energies that are much
larger than the spacings between energy levels of the quantum harmonic oscillator.
Thus, unless the experimenter can perform his/her experiment in a perfect vacuum
(something that nobody can achieve), the air molecules will likely interact with the
system and cause the oscillator to make transitions to higher energy levels. Finally,
the calculation in part (c) shows that using visible light to probe the system is not
practical. Even if one neglects the collapse of the wavefunction caused by measuring
the system with light, the fact that the relevant spatial scales ( 0.004 nm) are so
much smaller the wavelength of the light means that the light will essentially ignore
features on those scales. This is why, for instance, mirrors for telescopes that operate
in the visible wavelength band have to be grounded to such high precision, whereas
those for microwave telescopes are about as smooth as a nice piece of cardboard.

4
Problem Set 5 Solutions
8.04 Spring 2013 March 12, 2013
Problem 3. (30 points) Harmonic Oscillators Oscillate Harmonically

(a) (4 points) Our wavefunction is initially

1
(x, 0) = [0 (x) + i1 (x)], (18)
2
where 0 and 1 are the normalized eigenstates for the ground and rst excited states of
the harmonic oscillator respectively. The eigenstates evolve in time in the usual fashion
(phase factor with angular frequency equal to En /~), so the principle of superposition
tells us that
1 1
(x, t) = eiE0 t/ 0 (x) + ieiE1 t/ 1 (x) = ei0 t/2 0 (x) + iei30 t/2 1 (x) ,
2 2
(19)
th
where in the last step we used the fact that the n excited state of a harmonic oscillator
has energy En = ~ 0 (n + 12 ). The probability distribution |(x, t)|2 is given by

1
|(x, t)|2 = ei0 t/2 0 (x) + iei30 t/2 1 (x) ei0 t/2 0 (x) + iei30 t/2 1 (x) (20a)
2
1
= ei0 t/2 0 (x) iei30 t/2 1 (x) ei0 t/2 0 (x) + iei30 t/2 1 (x) (20b)
2
1
= |0 (x)|2 + |1 (x)|2 iei0 t 1 0 + iei0 t 0 1 (20c)
2
1
= |0 (x)|2 + |1 (x)|2 + i0 1 ei0 t ei0 t (20d)
2
1
= |0 (x)|2 + |1 (x)|2 + 20 1 sin 0 t , (20e)
2
where in the second last step we took advantage of the fact that energy eigenstates
can be taken to be real (see Problem Set 3), so 0 = 0 and 1 = 1 .

(b) (10 points) The expectation value (x) is dened in the usual fashion:
Z Z Z Z Z 
2 1 2 2
(x) = xdx = x|(x, t)| dx = x|0 (x)| dx + x|1 (x)| dx + 2 sin 0 t x0 1 dx .
2
(21)
To proceed, we need to know something about the form of n , the nth excited energy
eigenstate of the harmonic oscillator. As discussed in lecture, the general form of n
is given by
x2
n (x) = Nn Hn (x/a)e 2a2 , (22)

5
 where Hn is the nth Hermite polynomial and a ~/m0 . Since the Hermite poly
nomials are either even or odd, the eigenstates are also either even and odd2 , which
means that |0 (x)|2 and |1 (x)|2 are both even. In turn, x|0 (x)|2 and x|1 (x)|2 are
odd, so Z Z
2
x|0 (x)| dx = 0 and x|1 (x)|2 dx = 0, (23)

since the integrals are over symmetric intervals. This leaves the last term, for which
we will need the precise forms of the energy eigenstates:

m0 1/4 x22 m0 1/4 x 2 x22
0 (x) = e 2a and 1 (x) = e 2a (24)
a
This gives
Z
r
Z
2m0 x 2 ( xa )2
(x) = sin 0 t x0 1 dx = a sin 0 t e dx (25a)
a

r Z r
r

2m 0 2 m 0 ~
= a2 sin 0 t u2 eu du = a2 sin 0 t = sin o t. (25b)
2 2m0
The reader is encouraged to t check that this expression has the right units. The ampli
~
tude of the oscillation is 2 2m 0
, and the angular frequency of the oscillation is 0 .
It is instructive to see how to calculate (x) also via the operator method. Taking the
result from Problem 4 (e), we can express x in terms of a, a :

r
~
x = (a + a ). (26)
2m0
Thus
r Z
~
(x) = dx (a + a ) =
2m0
r
Z Z
~ 1 i i0 t
= 0 (a + a )0 + e 0 (a + a )1 +
2m0 2 2
Z Z 
i i0 t 1
e 1 (a + a )0 + 1 (a + a )1 . (27)
2 2
Now, still from Problem 4 we know that, for any energy eigenstate n ,
Z Z Z Z

n a
n = n a n = n a n1 = n1 a n = 0, (28)

and thus we have

r
 Z Z  r
~ i i0 t i i0 t ~
(x) = e 1 e
0 a 1 a
0 = sin 0 t. (29)
2m0 2 2 2m0
2
Of course, we know from Problem Set 3 that we can always take the energy eigenstates to be even or
odd, but its nice to see it explicitly in Equation 22 as well.

6
 (c) (10 points) For (p) we have
Z Z
~
(p) = pdx
= dx (30a)
i x
~  i0 t/2
Z
0 (x) iei30 t/2 1 (x) ei0 t/2 0 (x) + iei30 t/2 1 (x) dx(30b)
 
= e
2i
~ 
Z
0 0 + 1 1 iei0 t 1 0 + iei0 t 0 1 dx,

= (30c)
2i

where we have once again used the fact that the energy eigenfunctions can be taken to
be real. Let us examine this term-by-term. Since the derivative of an even function is
odd (and vice versa3 ), terms like n n must be odd overall. We thus have
Z Z
0 0 dx = 0 and 1 1 dx = 0. (31)

Plugging in the forms of the energy eigenfunctions gives

~ 
Z
iei0 t 1 0 + iei0 t 0 1 dx

(p) = (32a)
2i

2

x2

r
1 m~0
Z 2 2
ei0 t x 2 e a2 ei0 t x 2 e a2 + ei0 t 2 e
x 2 x 2 a
= 3 3
dx (32b)
2 a a a

r
1 m~0 i0 t 2 u2 2
Z
2
= e 2u e ei0 t u2 2eu + ei0 t eu 2 du (32c)
2

r

 
1 m~0 i0 t i0 t i0 t
= e e +e 2 (32d)
2 2 2

r
1 m~0
= i sin 0 t 2 + (cos 0 t i sin 0 t)2 (32e)
2

r
m~0
= cos 0 t. (32f)
2
Again, lets obtain the expression for (p) also through the operator method. From
Problem 4 (e), we know that

r
m~0
p = i (a a ), (33)
2
and hence,
r Z
m 0
(p) = i dx (a a ). (34)
2
3
If youre not sure about this, try proving it from the denition of the derivative: f ' (x) =
limx0 f (x+x)f
x
(x)
. It may also help to compute the derivatives of the energy eigenstates of the har
monic oscillator explicitly. Finally, try drawing a sketch! A picture is worth 103 words, especially in physics.

7
 Now, expanding like in Equation (27) and keeping into account Equation (28), we
obtain

r Z  Z Z  r
m~0 i i0 t i i0 t m 0
(p) = i e 0 a
1 + e 1 a
0 = cos 0 t. (35)
2 2 2 2

(d) (6 points) Any wavefunction can be expanded in terms of the energy eigenstates of
the system:
(x, t) = cn n (x)eiEn t/ , (36)
n
th
where En is the energy of the n excited state. With the harmonic oscillator, we have
 
1
En = 0 n + , (37)
2
so the wavefunction takes the form

(x, t) = ei0 t/2 cn n (x)ein0 t . (38)

n

2
Each term in the sum oscillates with period Tn = n 0
, i.e. some integer fraction of
the classical period T = 2/0 . For example, aside from the overall the ei0 t/2 in
front, the rst term is constant, the second term oscillates with period T , the third
term oscillates with period T /2, and so on. Thus, after a time T all the phases will be
aligned once again, except for phase in the prefactor, so that

(x, t + T ) = ei0 T /2 (x, t). (39)

Taking the norm squared of both sides gives

|(x, t + T )|2 = |(x, t)|2 , (40)

which is a mathematical way of saying that the probability distribution returns to its
original shape after period T = 2/0 .

The probability distribution is periodic because the dierence between angular fre
quencies (En Em )/~ are rational multiples of a common value. With the harmonic
oscillator in particular, this arises because the energy levels are equally spaced apart
(with E = 0 ).
Consider instead a system whose rst three energy levels are E0 , 2E0 , 2E0 , with as
sociated eigenstates 0 , 1 , 2 . Then the state

1  i E0 t 2E
i 0 t
2E
i 0 t

(x, t) = e 0 + e 0 + e 2
3
is not periodic in time, or more precisely there is no value of T for which

(x, t) = ei (x, t + T ). (41)

8
To see this, lets rewrite as
E0
ei t  i
( 21)E0
t
E
i 0 t

(x, t) = 0 + e 0 + e 2 ,
3
so that (41) is equivalent to
E E0
21) 0 T
ei( = 1, together with ei
T
= 1,

which is equivalent to
~
( 2 1)T = 2n , together with T = 2m ~
E0 E0
for some integers n and m. Dividing one equation by the other we get
n
21= ,
m
which has no solution.

9
Problem Set 5 Solutions
8.04 Spring 2013 March 12, 2013
Problem 4. (40 points) Operators for the Harmonic Oscillator

(a) (10 points) First remember what it means when we take the Hermitian conjugate
(dagger) of an operator. The operator O is dened by the following equation:
where 1 and 2 are arbitrary wavefunctions that satisfy that usual requirements of
continuity, normalizability, and so on. With this, the norm squared of our state is
given by
Z Z Z
n n
(n , n ) = n n dx = [(a ) 0 ] (a ) 0 dx = [(a )n1 0 ] aa [(a )n1 0 ]dx

(42)
Now, aa can be rewritten in the following way:

a = a
a a a
a a = [
+a ] + a
a, a a = 1 + a
a,
(43)

where we have used the identity [a, a ] = 1 (prove this by expressing a and a in terms
of p and x). Inserting this into our expression, we have
Z
(n , n ) = (n1 , n1 ) + [(a )n1 0 ] a a[(a )n1 0 ]dx (44a)

Z
= (n1 , n1 ) + n1 a an1 dx. (44b)

Now, recall that the energy operator for a harmonic oscillator can be written as
 
1
E = ~ a a + , (45)
2

which means we can rewrite the last line in terms of the energy operator:
Z
1 1
(n , n ) = (n1 , n1 ) (n1 , n1 ) + En1 dx. (46)
2 n1

This is nice because n1 is an energy eigenstate (it may not be normalized, but
its still an energy eigenstate, because a constant times an eigenfunction is still an
eigenfunction). This means that E acts on it in a very simple way, because it satises
the energy eigenvalue equation:
 
1
H n1 = En1 n1 where En = n + ~. (47)
2

10
 In our case, then, we have
1 En1
(n , n ) = (n1 , n1 ) + (n1 , n1 ) (48a)
2  
1 1
= (n1 , n1 ) + (n 1) + (n1 , n1 ) (48b)
2 2
= n(n1 , n1 ). (48c)

We have thus related the norm square of n to the norm square of n1 . We can repeat
the process until we reach the ground state:

(n , n ) = n(n1 , n1 ) = n(n 1)(n2 , n2 ) = = n(n 1) . . . 2 1(0 , 0 ) = n!,

(49)
where in the last step we used the fact that 0 = 0 , so 0 is in fact the normalized
ground state. The norm of |n) is thus
t
(n , n ) = n! . (50)

(b) (6 points) The normalized energy eigenstates are given by

1
n (a )n 0 . (51)
n!
First we test for the normality of the states:
1 n n
Z
n!
(n |n ) = 0 a (a ) 0 dx = = 1, (52)
n! n!

where we made use of the result we proved in (a). For orthogonality, we have
Z
1
(m |n ) = 0 am (a )n 0 dx, (53)
m!n!

where m 6= n. Now, after acting on 0 with the creation operator a n times, we end
up with a state that is proportional to the nth excited state n . Our equation then
calls for the annihilation operator a to act on the state m times. If m > n, we have
more annihilation operators than we do creation operators, and at some point we get
to a0 , which gives us zero. If m < n, we can say
Z Z
1 m n 1
0 a (a ) 0 dx = (an (a )m 0 ) 0 dx, (54)
m!n! m!n!
and again we end up lowering a state more than we raise it, only this time its the left
copy of the wavefunction. We thus conclude that (m |n ) = 0 unless m = n, in which
case (m |n ) = 1. The states n are therefore orthonormal.

11
(c) (6 points) We wish to show that

n = nn1
a n =
and a n + 1n+1 . (55)
So far, we only know that an is proportional to n1 . In other words,
a
n = Dn1 , (56)
where D is some proportionality constant. Let us dot this equation with itself and
perform manipulations similar to what we did in part (a):
Z
[an ] [an ]dx = D2 (n1 |n1 ) (57a)

Z
n a a
n dx = D2 (n1 |n1 ) (57b)

1
Z
1
n En dx (n |n ) = D2 (n1 |n1 ) (57c)
~ 2
En 1
(n |n ) (n |n ) = D2 (n1 |n1 ) (57d)
~ 2
n(n |n ) = D2 (n1 |n1 ) (57e)

n = D, (57f)
where we have used the fact that En = n + 12 and that n and n1 are normalized

energy eigenstates of the harmonic oscillator. We thus have

a
n = nn1 . (58)
Similarly, let
a n = F n+1 , (59)
where F is some proportionality constant. Performing analogous manipulations, we
have:
Z
[a n ] [a n ]dx = F 2 (n1 |n1 ) (60a)

Z
n aa n dx = F 2 (n1 |n1 ) (60b)

Z
n [a, a ] + a a
n dx = F 2 (n1 |n1 ) (60c)

Z
(n |n ) + n a a
n dx = F 2 (n1 |n1 ) (60d)

(n |n ) + n(n |n ) = F 2 (n1 |n1 ) (60e)
2
(n + 1)(n |n ) = F (60f)

n + 1 = F, (60g)
R
where we have used the fact that [a, a ] = 1 and the result D2 n a a
n dx = n
that we obtained above. We thus have

a n = n + 1n+1 . (61)

12
 (d) (6 points) The number operator is dened as
= a a.
N (62)
We now examine the commutators of this operator with the annihilation and creation
operators:
, a]
[N a
= N a N =a
a a a
a a a a
= (a a)
a = [ ]a =
a, a a. (63)
Similarly,

, a
[N a
] = N a =a
N a
a a
a (a
a = a a a
a) [
=a ] = a .
a, a (64)
From lecture, we know that for the harmonic oscillator the energy operator E can be
written as4

1
 
1

E = ~0 a a

+ = ~0 N+ N = E 1. (65)
2 2 ~0 2
This means that the eigenfunctions of N are the eigenfunctions of E,
i.e. the energy
eigenstates of the harmonic oscillator. We can see this explicitly by having N act on
n :  
1 1 ~0 1 1
N n = En n = n+ n n = nn , (66)
~0 2 ~ 0 2 2
since we know that the energy eigenvalues of the harmonic oscillator are given by
En = 0 n + 12 . We can also see from this that the eigenvalues of N are simply the
various values of n (i.e. the quantum number corresponding to each energy eigenstate).
(e) (6 points) Our original denitions of a and a were
x p x p
a p + i and a p i . (67)
2~/m0 2m 0 2~/m0 2m~0
We can solve these to get expression for x and p:

r
r
~ m~0
x = (a + a ) and p = i (a a ). (68)
2m0 2
Now, for an eigenstate n , we have
Z r
Z
~
(x) = n x
n dx = n ( )n dx = 0,
a+a (69)
2m0

because when a acts on n , the result is proportional to n+1 , which is orthogonal to

n , and similarly for a. For the momentum, we have
Z r

m~0
Z

(p) = n dx = i
n p aa
n ( )n dx = 0, (70)
2

using exactly the same reasoning.

If youre not sure where this comes from, express a and a in terms of x and p, and make use of the
4
of the harmonic oscillator in terms of a and a .
commutation relation [x, p] = i~ to express E

13
(f) (6 points) First we express x2 and p2 in terms of a and a :


r
r
~ ~ ~
2
x = (a + a ) (a + a ) = (a + a )(a + a ) (71a)
2m0 2m0 2m0
~ ~
= (a2 + aa + a a + (a )2 ) = (a2 + (a )2 + 2a a + [a, a ]) (71b)
2m0 2m0
~
= (a2 + (a )2 + 2N + 1). (71c)
2m0
Similarly, for p2 , we have

r r

m~0 m~0 m~0
p2 = i (a a ) i (a a ) = (a a )(a a ) (72a)
2 2 2
m~0 2 m~0 2
= (a aa a a + (a )2 ) = (a + (a )2 2a a [a, a ]) (72b)
2 2
m~0 2
= (a + (a )2 2N 1). (72c)
2
Plugging these operators into (x2 ) and (p2 ) gives
Z  
2 ~ 2 2
(x ) = n (a + (a ) + 2N + 1) n dx (73a)
2m0
Z
~
= (2N + 1)n dx (73b)
2m0 n
~
= (2n + 1), (73c)
2m0
where in going from the rst line to the second line we used the same reasoning as
we did in part (e) i.e. orthogonality of n and (a )2 n n+2 or a2 n n2 . In
going from the second line to the third line we used our results from part (d). The
expectation value of the momentum looks similar:
Z  
2 m~0 2 2
(p ) = n (a + (a ) 2N 1) n dx (74a)
2
m~0
Z
= n (2N + 1)n dx (74b)
2
m~0
= (2n + 1). (74c)
2
p
Now,
p recall that the uncertainties are dened as x (x2 ) (x)2 and p
(p2 ) (p)2 . Since (x) and (p) are zero from part (e), we get
p ~
xp = (x2 )(p2 ) = (2n + 1), (75)
2
which is our desired result.

14
MIT OpenCourseWare
http://ocw.mit.edu

8.04 Quantum Physics I

Spring 2013

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.