Infinite square well

We now turn to the most straightforward (and therefore educational) non-zero potentials.
We just discussed a free particle; we now turn to a bound particle, and will shortly
discuss potentials that can lead to both.

The infinite square well potential is given by:

( )

> <

=
a x x
a x
x V
, ,
,
0
0 0




A particle under the influence of such a potential is free (no forces) between x = 0 and x
= a, and is completely excluded (infinite potential) outside that region.

Now, the first lesson to take from this problem is that one does not have to solve the
TISE everywhere at once one can solve for the wave function piecemeal in the different
regions where the potential is given by one function. However, the wave function must
behave as a whole in certain ways as we discussed earlier.

We first look for the wavefunction in the region outside of 0 to a. Here, where the
particle is excluded, the wave function must be zero. Great, half of the problem solved.

Inside the region from zero to a, the wavefunction must be a solution of

( )
( ) x E
dx
x d
m

=
2
2 2
2
h


since V(x) = 0. The solution to this simple differential equation is simply the free particle
solutions:

( )
h
mE
k Be Ae x
ikx ikx
2
= + =

,

But now we have to apply whatever boundary conditions we have. We know that
continuity of the wave function insists that (0) = (a) = 0, so,
( )
( ) ( )
( ) ( )
( ) ... , , , sin
... , , ,
sin
sin
2 1 0
2 1 0
0
0
0 0
=

=
=
= =
= =
= =
=
+ =
+ =

n x
a
n
A x
a
n
k
n n ka
ka A a
kx A Ae Ae x
A B
B A
Be Ae
n n
ikx ikx
ik ik

But, n=0 is a trivial solution, so we have:

( ) ... , , , sin 3 2 1 =

= n x
a
n
A x
n n

These are our stationary state solutions to the infinite square well potential. We choose to
normalize these state solution for reasons that will become evident later:

( )

i
n
n
a
n n
e
a
A A
a
A
dx x
a
n
A dx x
2
1
2
1
2
0
2
2 2
= =
=
=

sin


where beta is an arbitrary phase, and since it has no consequence on measurement, we
take it to be zero:

a
A
2
=

So that the normalized stationary states are:

( ) ... , , , sin 3 2 1
2
=

= n x
a
n
a
x
n

Lets take a moment to study some of the properties of these solutions, as they are in
general applicable to solutions to most potentials that you will encounter:

1) These functions are alternatively even and odd about the center of the potential well.
This will be true for any symmetric potential.

2) With increasing n, each successive state has one more node in the wavefunction. This
is true regardless of the shape of the potential.

3) The states are orthogonal. This means that:
( ) ( ) n m dx x x
n m
=

, 0

We show a proof:

( ) ( )
( ) ( )
( ) [ ]
( )
( ) [ ]
( )
0
1
1 1
1
2
0
0
0
=

+
+

+
+

n m
n m
n m
n m
x
a
n m
n m
x
a
n m
n m
dx x
a
n m
x
a
n m
a
dx x
a
n
x
a
m
a
dx x x
a
a
a
n m

sin sin
sin sin
cos cos
sin sin


This, together with the fact that when m = n, the same integral gives one (this is why we
normalized the state functions) gives:

( ) ( )

=
=

n m
n m
dx x x
mn
mn n m
,
,
1
, 0

where
mn
is called the Kronecker delta. These statements are combined to say that the
states are orthonormal (both orthogonal and normalized). This is also a general result of
state functions regardless of the potential (assuming you normalize them).

4) The state functions form a complete set. This is perhaps the most interesting and
useful of the properties. Although not universal, it will usually be the case for potentials
you will encounter. This means that any other function f(x) can be written in terms of
these functions:

( ) ( )


=

= =
1 1
2
n
n
n
n n
x
a
n
c
a
x c x f

sin

Then, finding the values of the constants c
n
is straightforward using Fouriers trick (just
using the orthonormal nature of the state functions):

( ) ( ) ( ) ( )
( ) ( )

=

=
= = =
dx x f x c
c c dx x x c dx x f x
n n
m
n
mn n
n
n m n m

1 1

OK, we will get back to these properties of the state functions in a minute, but lets put
the whole picture together first. The general state solutions are given by:

( ) ... , , , sin , 3 2 1
2
=

=

n e x
a
n
a
t x
t iE
n
n
h

where we can determine the energies E
n
from above:

2
2 2 2
2
2
ma
n
E
a
n mE
k
n
h
h

=
= =


so that the boundary conditions impose specific, allowed values of the energy.



Now, if someone asks you: What is the wavefunction for a particle in an infinite square-
well potential? what would you answer? No, lets say you dont know its energy. The
answer is a superposition of all the
n
s:

( ) ( )
2
2 2 2
1
2
2
ma
n
E e x
a
n
c
a
t x c t x
n
n
t iE
n
n
n n
n
h
h

=

= =


=

, sin , ,

But how do you find the c
n
s? To determine these, you have to have an additional
boundary condition, namely at some time, call it t = 0, you have to know something else
about the particle.

Example: A particle in an infinite square well has as an initial wave function

( )
( )

> <

=
a x
a x x a Ax
x
0 0
0
0
,
, for some constant A. Find (x,t).

First, we must determine A using the normalization condition (since if (x,0) is
normalized, (x,t) will stay normalized, as we showed earlier):

( ) ( ) ( )
5
5
2
5 5 5
2
0
5 4 3
2 2
0
4 3 2 2 2
0
2 2 2
0
2
30
1
30 5 4
2
3 5 4
2
3
2 0
a
A
a
A
a a a
A
x x
a
x
a A
dx x ax x a A dx x a x A dx x
a
a a a
=
= =

+ =

+ =
+ = =

,


Now we turn to our method for determining the coefficients c
n
:

( ) ( ) ( ) ( )

=
= = dx x f x c x c x f
n n
n
n n
,
1
so,
( ) ( )
( ) ( )
( )
( )
( )
( )
( )
( )
( )
( )
( )
( )
( ) ( ) [ ]
( )

=
=

+ =

=
=
=

=
odd n if
15 8
even n if 0
0
15 4
0
2 2 15 2
2
2
15 2
15 2
30 2
0
0
3
3
3
3
3
2
3
3
3
0
3
2
2
0
2
3
0
2
0
3
0
5
1
,
,
cos cos
cos cos cos
cos sin
cos sin
sin sin
sin
,
,

n
n
n
n
a n
n
n
a n
n
a
a
x
a
n
a n
a x n
x
a
n
x
n
a
x
a
n
n
ax
x
a
n
n
a
a
a
dx x
a
n
x dx x
a
n
x a
a
dx x a x
a
x
a
n
a
dx x x c
x c x
a
a
a a
a
n n
n
n n


So that now we can form our time dependent wavefunction by putting this all together:

( )

=

=
... , ,
sin ,
5 3 1
2
3
3
2 2 2 1 2 30
n
ma t in
e x
a
n
n a
t x
h

Now, lets study this wavefunction, and its implications on measurable quantities. We
first ask what the expectation value of the energy is.

( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
( ) ( )
( ) ( )

= =
=
=

= =

n
n n
m n
mn n m n
m n
n m n m n
m n
n n m m n
n
n n n n
m
m m m
n
n n n
m
m m m
n
n n n
m
m m m
E c c c E
dx x x c c E
dx x c x c E
dx t x E c t x c
dx t x H c t x c
dx t x c H t x c dx t x H t x E
2

)
) )
, ,


So, we can think of the |c
n
|
2
as the probability that one will measure E
n
, remembering
again that you will only measure one of these values, and then the total probability to
measure one of the E
n
s must be one or:

1
2
=

m
n
c

We can show this since the wave function is normalized, each stationary state is
normalized and the initial condition is also normalized, then:

( ) ( ) ( ) ( )
( ) ( )
( ) ( )
1
0 0
2
= = =
=
=

m
n
m n
mn n m
m n
n m n m
m n
n n m m
n
n n
m
m m
c c c
dx x x c c
dx x c x c
dx x c x c dx x x

, ,


The c
n
s can also be thought to tell you the amount of
n
that is in the total wavefunction.

In the example above, you can see that the initial wavefunction closely resembles
1
, and
if we look at c
1
:

... .998555 0
15 8
2
3
2
1
=

c

very close to one indicating that the 1
st
state dominates.