Quantum Mechanics A

PHY-319

Note Set No. 3

The Finite square well.

We have already solved the problem of the infinite square well. Let us now solve the
more realistic finite square well problem. Consider the potential shown in fig.1, the
particle has energy, E , less than V0 , and is bound to the well.

V = V0
E < V0
region 1

region 2

region 3

V=0
-L/2

L/2

Figure 1: A finite square well, depth, V0 , width L .

Region 1
L
x , V ( x) V0 , substituting into TISE:
2

h 2 d 2

V0 E
2m dx 2

d 2 2m
2 V0 E
dx 2
h

yielding
d 2
2
2
dx

with 2

2m
V0 E 0 ,so
h2

2m
V0 E ,which is real.
h2

The solutions to this differential equation are:

Bex De x
but since 0 as x , D 0
we get Bex for region 1
Region 3
L
x , V ( x ) V0 , .
2
similarly to region 1, the solutions are:

 Ae x D' ex
but since 0 as x , D ' 0
x
we get Ae for region 3
Region 2
L
L
x , V ( x) 0 , substituting into TISE:
2
2
2
2
2m
h d
d 2
2 E

2
2
2m dx
dx
h
meaning

2mE
d 2
k 2 with k 2 2 0 , so k
2
dx
h

2m
E , which is real.
h2

which we have already encountered in the infinite square well, since the potential is
symmetric, we have even and odd parity solutions, namely:

C coskx for even parity (n odd)

and D sin kx for odd parity (n even)

Parity +1
Even

Parity 1
Odd

Wavefunction

Region

Wavefunction

Bex

B' e x

C coskx

D sin kx

Ae x

A' e x

Table 1: summary of wavefunctions of the finite potential well.

Clearly, for parity +1

BA

and for parity -1

B' A'

Now we have to match both the wavefunction, (x) , and its derivative,
the well boundaries, namely at x

d ( x)
, at
dx

L
. Of course, we have to do this twice, since we
2

have even and odd parity solutions

L
So, at x , for even parity:
2

L
equating :
C cos k Ae 2
(1)
2
L

d
L
:
Ck sin k Ae 2
(2)
and equating
dx
2
Dividing eq(2) by eq(1) to eliminate C and A gives:

kL
(3)
tan
k
2
We can carry out the same analysis for the negative parity solutions and obtain:

kL
(4a)
cot
k
2

kL
tan
(4b)
k
2 2
L

2m
V0 E and k 2m2 E , in both equations (3) and (4),
2
h
h
there is only one unknown, the energy, E ; so we should be able to solve for the
energy. It transpires that both equation (3) and equation (4) are transcendental, that is,
they cannot be solved analytically, we can, however, solve them numerically. In order
to do so, we shall rewrite them in a more convenient form, using dimensionless
parameters, and 0 .
is simply the argument of the trigonometric function:

Note that since

kL L 2m

E
2
2 2
and is the variable we will be solving for since it contains the energy

Now, 0 is defined by:

L 2m
V0
2 2
and is called the potential-strength parameter, since it contains the depth of the
potential well, V0 , ( and its width, L ).

Recall that

2m
V0 E and k
h2

2m
E , so
h2

2m
V0 E 2m2 V0 k 2
2

2 1

L k

2 k 2 02

02
02

k
k 2 2
2

So our transcendental equations (3) and (4) become:

tan 0 1

(5)

for even parity, and,

(6)

for odd parity, note that cot tan

2

cot 0 1

kL L 2m

E , if we can solve these for , we will obtain the

2
2 2
energy eigenvalues E n for our finite well.

Reiterating, since

2
5
8

0
0

10

Figure 2: A plot of

0
1 versus for three values of 0 (2, 5 and 8).

EVEN
ODD

10

Figure 3: Plots of tan for even parity solutions and of cot tan for
2

odd parity solutions versus .

4

We can superimpose a plot of

0
1

onto plots of

tan

and

cot tan to graphically solve the two equations. For example taking an
2

electron in a well, width 4 and depth to be 14eV we can calculate 0 3.83 . A plot
2

of 0 1 for 0 3.83 is shown in figure 4, together with the trigonometric

function plots. The curves intersect at three values of (circled), corresponding to
the energy levels of the three bound states.

EVEN
ODD
Z = 3.83

/2

3/2 5

Figure 4: a graphical solution for the energy eigenvalues of the three bound states of
an electron in a 4, 14eV finite potential well.
The values of obtained are:
1.24 , corresponding to n 1 ,even parity
2.45 , corresponding to n 2 ,odd parity
3.54 , corresponding to n 3 ,even parity
and the corresponding energy eigenvalues are:
E1 1.47eV , E 2 5.74eV and E3 11.99eV
Note that these are much lower than the corresponding energy eigenvalues for an
infinite square well of the same width ( E 1 2.36eV , E 2 9.43eV and
E 3 21.24eV ). This is not surprising as the wavefunction in the finite potential
well extends into the classically forbidden region, so the corresponding wavelengths
are longer than those in the infinite well, resulting in lower energies (see figure 5).

V0 = 14eV
12eV

9.43eV

5.74eV
2.36eV
1.47eV

L = 0.4nm

L = 0.4nm

Figure 5: The three bound states in a 0.4nm, 14eV one-dimensional finite quantum
well. The wavefunctions are shown schematically. Note how the corresponding
energy levels of an infinite well are much higher.