University of Connecticut
DigitalCommons@UConn
Chemistry Education Materials
Department of Chemistry
6-27-2006
The Particle in a Box (and in a Circular Box)
Carl W. David
University of Connecticut, Carl.David@uconn.edu
Follow this and additional works at: http://digitalcommons.uconn.edu/chem_educ
Part of the Chemistry Commons
Recommended Citation
David, Carl W., "The Particle in a Box (and in a Circular Box)" (2006). Chemistry Education Materials. Paper 12.
http://digitalcommons.uconn.edu/chem_educ/12
This Article is brought to you for free and open access by the Department of Chemistry at DigitalCommons@UConn. It has been accepted for
inclusion in Chemistry Education Materials by an authorized administrator of DigitalCommons@UConn. For more information, please contact
digitalcommons@uconn.edu.
The Particle in a Box (and in a Circular Box)
C. W. David
Department of Chemistry
University of Connecticut
Storrs, Connecticut 06269-3060
(Dated: June 27, 2006)
I.
SYNOPSIS
The particle in a box is the simplest problem in elementary quantum mechanics, and as such is useful in more
places than one can easily enumerate. For instance, the
molecular orbital theory is explained easily in terms of
particle in a box wavefunctions, easier than using standard atomic orbitals. Hence, the attention to this particular problem.
II.
INTRODUCTION
We start with a choice of coordinate systems, which, by
the way, influences the form of the solutions we are going
to get, but not the substance. Here, we choose to use
0<x<L for the region on the x-axis where the particle is
said to exist. for regions x<0, i.e., the negative x-axis, we
say the particle is forbidden, and mathematically, we do
this by stating that V (x) = , (x) = 0 in this region.
We say exactly the same thing in the region x>L, i.e.,
the potential energy is infinite, and the wave function
vanishes.
In the domain 0<x<L, we expect the wave function
to exist and have a value different from zero, but at the
boundaries, we declare (0) = 0, (L) = 0.
Then, the Schrodinger Equation (inside the domain)
becomes
2 2
h
(x) = E(x)
2m
where m is the mass of the particle, h is, of course,
Plancks constant divided by 2, and E is the energy,
the eigenvalue, the allowed value of the energy that this
particle can have. We know the solution to this differential equation from elementary calculus, since there are
very few functions which resurrect themselves after being differentiated twice. One of these is the exponential,
and the other is the (sine/cosine) combination (which are
really forms of the exponential). Assume the sines and
cosines form, we have
and, taking the second derivative, we have
d2 (x)
= A 2 cos(x) B 2 sin(x)
dx2
which is, of course,
2 (x)
This means that
2
h
2 (x) = E(x)
2m
or, said in the most straightforward manner, E is related
to i.e.,
2 =
. Later, we will obtain values for and thereby obtain
2 2 2
h
values of E. They will turn out to be En = n2mL
2 , a very
famous result.
To obtain this result, we note the boundary conditions, that the wave function vanish on the left and
right boundaries. The left boundary condition, x=0 says
(0) = A cos(0) + B sin(0) = A cos(0) = A i.e., A must
be chosen to be zero (as cos 0 equals 1).
The right boundary condition now reads (L) =
B cos(L) and it is a famous argument that if B is not
to be zero, then the cosine must. This can only occur
if has values such that the argument L equates to
, 2, 3, . i.e. n. This is the infamous quantization
which takes place to force discrete values of E, the energy.
Every text book says this better than I do, so you are referred to standard texts for alternative presentations of
this material.
The various vave functions, now indexed with n, are
orthogonal to each other, i.e., perpendicular in function space. We see that this is just a Fourier discussion
in mufti. Thus, we have
L
1 (x)2 (x)dx = 0
(x) = A cos(x) + B sin(x)
where A, B, and are unknown (to be determined) constants. Taking the first derivative of this solution, we
have
d(x)
= A sin(x) + B cos(x)
dx
Typeset by REVTEX
2mE
h2
And we can change 1 to 17, and 2 to 43, and the same
holds. Formally,
Z
n (x)m (x)dx = 0
0
2
if n6=m. Of course, when n=m, we have the normalization
integral
L
n2 (x)dx 6= 0
i.e., something other than zero. This last condition is
used to choose the (so far) arbitrary constant, A, to force
the integral to have a value of 1, in accord with the probabilistic interpretation of the wave function. Since these
are, in the case of the particle in a box, integrals of products of sines, we have
Z
A
n
L x
e
2
n
L x
2
dx
q
which are trivial. A turns out to be L2 if we force the
integral to be 1. Again, this is explained in every text
book. Of course, the orthogonality integrals may also be
evaluated using DeMoivres theorem. All of this turns
out to be extraordinarily elementary.
III.
A RECTANGULAR BOX
When we talk about rectangular and square boxes in
two dimensional particle in a box problems, we are getting set up to discuss degeneracy. The wave equation
itself offers no hint of the complexity that is coming, but
the boundary conditions, which are inextricably bound
to the solutions of the Schr
odinger Equation, are the key
here.
Assuming we are working in the x-y space (2 dimensions), so (x, y) is what is being sought, we have
h2
2 2
+
x2
y 2
= E
(3.1)
where x and y are the standard Cartesian coordinates.
The solution to this variable separable differential equation is discussed in all texts, and has the form
X(x)Y (y) = N`x ,`y sin
nx x
ny y
sin
`x
`y
where nx and ny are integer quantum numbers, ranging
from 1 to . The box is rectangular if `x 6= `y . N`x ,`y is
the normalization factor, which has the form
r s
2
2
N`x ,`y =
`x `y
Once substituted into the original Schr
odinger Equation, the solution suggests that the energy is given by the
formula:
2 2 !
h2 2
nx
ny
Enx ,ny =
+
(3.2)
2
`x
`y
(a well known result).
The degeneracy appears when we allow the length of
the two sides of the box to become equal. At that
point, we can factor the common ` = `x = `y out of this
formula to obtain
Enx ,ny =
2 2 2
h
nx + n2y
2
2`
It is clear that the sum of two squares can add up to the
same value, e.g., 52 +22 = 22 +52 , i.e., n2x +n2y = reverse.
We are seeing a double degeneracy emerging. (In three
dimensions, we would have a three fold degeneracy developing, when the box became cubical.) Notice that when
nx = ny , we have a loss of degeneracy.
IV.
A CIRCULAR BOX
Squares are nice, but we learn more from circles. Consider a two dimensional particle in a circular box. The
particle is restricted to be within r=R, where R is a
constant. This is a polar coordinate problem, since the
boundary condition will be that (R, ) = 0, i.e., the
wave function will be required to vanish at the edge of
the disk region.
The first thing we have to do is transform the
Schrodinger Equation from Cartesian to polar coordinates. We are going from (x, y) (r, ). The transformation equations are
x = r cos
y = r sin
(4.1)
which says, given r and , we can compute x and y. The
reverse equations are
p
r = x2 + y 2
y
= tan1
(4.2)
x
which says the reverse, i.e., given x and y, we can compute
r and .
is terms of partial derivaFirst we need to express x
tives with respect to r and . From the chain rule we
have
=
+
x y
x y r
x y r
where we are carefully attempting to indicate what is
constant during each partial differentiation. We find that
!
p
r
x2 + y 2
1 2x
=
=
= cos
x y
x
2 r
y
and
=
y
tan1
x
y
x
y
3
This latter differential is a little more difficult than the
former one was, and we attack it in a special manner,
guaranteed to bring a smile of recognition to weary readers. Specifically, we use implicit differentiation. Thus
tan =
dy ydx
sin
sin2
2 =d
= d +
d
x
x
cos
cos2
d tan =
y
x
so
which is
dy ydx
2 =
x
x
so, holding y constant, one has
1
ydx
d
d = 2
=
2
cos
x
dx
sin2
1+
cos2
d =
y cos2
1
2
x2
cos2 x
y
cos2 + sin2
cos2
d
tan cos2
=
dx
K
1
cos2
d
dx
y
dy
d
=
x
cos2
cos
=
r
2
x2
= cos
d
ydx
d
=
2
x
dx
y
1
cos2
x2
y cos2
x2
= cos
y
= sin
x
sin
cos
+
r
The second partial derivatives then must be
so
cos2
=
r cos
1
cos2
so
Going the other way i.e., holding x constant, we have
d =
d =
tan cos2
sin
=
=
r cos
r
cos2
=
x
x
x
2
x2
2
y 2
=
cos
=
sin
sin
r
+ cosr
y
(4.3)
(4.4)
Equation 4.3 expands to become
x
cos
sin
r
!
sin
cos
sin
r
!
i.e.,
2
x2
= cos2
y
2
sin cos
sin cos 2
+
2
2
r
r
r
r
sin2
sin cos 2
sin cos
sin2 2
+
+
+
r r
r
r
r2
r2 2
(4.5)
4
For the y term we have
2
y 2
sin r
+
r
= sin
x
cos
r
!
sin r
+
cos
+
r
cos
r
!
(4.6)
i.e.,
2
y 2
= sin2
x
cos sin
sin cos 2
2
+
r2
r2
r
r
cos2
cos sin 2
cos sin
cos2 2
+
+
+
r r
r
r
r2
r2 2
(4.7)
so, adding the two relevant results we have
z
}|
{
2
sin cos
sin cos 2
+
= 2+
2
r
r
r
x
y
| r {z }
z
}|
{
sin2
sin cos 2
sin cos sin2 2
+
+
+ 2
r r
r
r | r2{z }
r 2
2
y 2
2
x2
z
}|
{
cos sin
sin cos 2
+
2
r
r
| r {z }
z
}|
{
cos2
cos sin 2
cos sin cos2 2
+
+
+ 2
r r
r
r | r2{z }
r 2
which becomes
2
1
1 2
+
+ 2 2
2
r
r r r
V.
where = 2mE/h2 .
This is one of the forms of the Bessel differential equation. To bring it to standard form, we change variables
from r to k = r, so
1
=
=
r
r
k
NOW, THE QUANTUM MECHANICS
The Schrodinger Equation now becomes
h2 2 1
1 2
+
+ 2 2 + zero = E
2m r2
r r
r
(why zero? because a particle in a box just feels its
boundaries, i.e., the model is of a particle free to roam
(linearly) until it hits a wall.) which is going to be related
to Bessels equation. We know this equation is variable
separable, between r and , so we will write the solution
as
= Rm` (r)eme ll
which, when we substitute this into the Schrodinger
Equation, gives us
h2 2 Rm` (r) 1 Rm` (r) m2`
+
2 Rm` (r) = ERm` (r)
2m
r2
r
r
r
which is
r2 R00 (r) + rR0 (r) + (r2
2m 2
m` )R(r) = 0
h2
(4.8)
so, choosing
k2 = 1
we have
r
k=
1
and we have
2 R00 () + R0 ()0 + (2 m2` )R() = 0
It is traditional to solve this equation separately for
different values or m, and in fact, it is rare to see solutions
for m` > 0 anywhere, since the problem, from the point
of view of quantum mechanics, is quote silly.
For m` = 0 we have
2 R00 () + R0 ()0 + 2 R() = 0
and we start the solution by assuming an Ansatz
X
R() =
ai i
i=0
5
leaving the question of the indicial equation to more advanced study. We then have
X
R0 () =
iai i1
so
i=1
and
R00 () =
(i)(i 1)ai i2
i=2
2 R00 ()
(i)(i 1)ai i 2a2 2 + (3)(2)a3 3 + (4)(3)a4 4 +
i=2
+R0 ()0
iai i a1 + 2a2 2 + 3a3 3 +
i=1
2
+( )R()
ai i+2 a0 2 + a1 3 + a2 4 +
i=0
=0
Gathering terms in the standard manner, we have
ematics, John Wiley and Sons. New York, 1962, page
548). So
a2 = a0 /2
((3)(2) + 3)a3 = a1
a0 /2
((4)(3) + 4)a4 = a2 =
((4)(3) + 4)
It is fairly obvious that, contrary to other Frobenius
Schemes in standard Quantum Chemistry, this one does
not lead to quantization through truncation of a power
series into a polynomial.
Instead, this Bessel Functions expansion never terminates, is never truncated, i.e., remains an infinite power
series.
The quantization occurs when the boundary condition
that the wave function vanish when we are at the rim of
the enclosure is invoked.
The requirement that the function be zero at the
boundary (r=R), results in requiring that the Bessel
function have a zero, call it 0 . This can be found in
tables of the Bessel Function.A For instance, the first
root, J0 (0 ) occurs at 0 = 2.4048 (and 5.5207, 8.6537,
11.7915, etc., E. Kreyszig, Advanced Engineering Math-
0 =
R
R
=q
k
1
i.e.,
2
0
==
2mE
h2
or,
2
2 0 2
h
h2 2.4048
E0 =
E0 =
2m R
2m
R
where you have specified the value of R when setting up
the problem.
The zeros of the Bessel function act to quantize the
energy via the boundary conditions, in a slightly different mode than what we are used to, but never the less,
we obtain a result which is quite comfortable inside the
context of standard quantum mechanical constructs.
