Schrodinger

Doing Physics with Matlab Quantum Mechanics Bound States
1
DOING PHYSICS WITH MATLAB
QUANTUM MECHANICS
SCHRODINGER EQUATION
TIME INDEPENDENT
BOUND STATES

Ian Cooper
School of Physics, University of Sydney
[email protected]

MATLAB SCRIPTS
Goto the directory containing the m-scripts for Quantum Mechanics

The Matlab scripts are used to solve the Schrodinger Equation for a variety of potential
energy functions.

se_fdm.m
m-script se to solve [1D] time independent Schrodinger Equation using a finite difference
approach where E is entered manually to find acceptable solutions.

simpson1d.m
Function to evaluate the area under a curve using Simpsons 1/3 rule.

Colorcode.m
Function to return the appropriate colour for a wavelength in the visible range from 380
nm to 780 nm.

se_wells.m
First m-script to be run when solving the Schrodinger equation using the Matrix Method.
Most of the constants and all the well parameters are declared in this file. You can select
the type of potential well from the Command Window when the m-script is run. You
alter the m-script code to change the parameters that characterize the wells and you can
2
add to the m-script to define your own potential well. When this m-script is run it clears
all variables and closes all open Figure Windows.

se_solve.m
This m-script solves the Schrodinger equation using the Matrix Method after you have
run the m-script se_wells.m. The eigenvalues and corresponding eigenvectors are found
for the bound states of the selected potential well.

se_psi.m
To be run after se_wells.m and se_solve.m. A graphical output displays the total energy,
the potential energy function, kinetic energy function, eigenvector and probability
distribution for a given quantum state.

se_measurements.m
To be run after se_wells.m and se_solve.m. Evaluates: the expectation values for a set of
dynamic quantities; the inherent quantum-mechanical uncertainties in measurements and
gives a test of the Heisenberg uncertainty principle.

se_orthonormal.m
To be run after se_wells.m and se_solve.m. You can investigate the orthonormal
characteristic of the eigenvectors (stationary state wavefunctions).

se_infwell.m
Used to test the accuracy of the Matrix Method. Compares the analytical and numerical
results for an infinite square well potential of width 0.1 nm.

gausssiani.m
Produces a graphical display of a Gaussian shaped potential well and the corresponding
force graph.

se_stationary.m
To be run after se_wells.m and se_solve.m. You can investigate and view the time
evolution of an energy eigenstate and save the plot as an animated gif.

se_super.m
To be run after se_wells.m and se_solve.m. You can investigate and view the time
evolution of a compound state and save the plot as an animated gif.

3
SCHRODINGER EQUATION

On an atomic scale, all particles exhibit a wavelike behavior. Particles can be represented
by wavefunctions which obey a differential equation, the Schrodinger Wave Equation
which relates spatial coordinates and time. You can gain valuable insight into quantum
mechanics by studying the solutions to the one-dimensional time independent
Schrodinger Equation.

A wave equation that describes the behavior of an electron was developed by
Schrodinger in 1925. He introduced a wavefunction ( , , , ) x y z t + . This is a purely
mathematical function and does not represent any physical entity. An interpretation of the
wave function was given by Born in 1926 who suggested that the quantity
2
( , , , ) x y z t +
represents the probability density of finding an electron. For the one dimensional case,
the probability of finding the electron at time t somewhere between x
1
and x
2
is given by
(1)
2
1
*
Prob( ) ( , ) ( , )
x
x
t x t x t dx = + +
}

where
*
+ is the complex conjugate of the wavefunction + . The value of Prob(t) must lie
between 0 and 1 and so when we integrate over all space, the probability of finding the
electron must be 1.

*
( , ) ( , ) 1 x t x t dx
+
+ + =
}

In this instance the wavefunction + is said to be normalized.

We can see how the time-independent Schrodinger Equation in one dimension is
plausible for a particle of mass m, whose motion is governed by a potential energy
function U(x) by starting with the classical one dimensional wave equation and using the
de Broglie relationship
Classical wave equation
2 2
2 2 2
( , ) 1 ( , )
0
x t x t
x v t
c+ c+
=
c c

4
Momentum (de Broglie)
h
p mv k
= = =
Kinetic energy
2
1
2
K mv =
Total energy ( ) ( ) E K x U x =
Wavefunction ( , ) ( )
i t
x t x e
e

+ = periodic in time for t coordinate

Combining the above relationships, the time-independent Schrodinger Equation in one
dimension can be expressed as
(2)
2 2
2
( )
( ) ( ) ( )
2
d x
U x x E x
m dx
+ =

Our goal is to find solutions of this form of the Schrodinger Equation for a potential
energy function which traps the particle within a region. The negative slope of the
potential energy function gives the force on the particle. For the particle to be bound the
force acting on the particle is attractive. The solutions must also satisfy the boundary
conditions for the wavefunction. The probability of finding the particle must be 1,
therefore, the wavefunction must approach zero as the position from the trapped region
increases. The imposition of the boundary conditions on the wavefunction results in a
discrete set of values for the total energy E of the particle and a corresponding
wavefunction for that energy, just like a vibrating guitar string which has a set of normal
modes of vibration in which there is a harmonic sequence for the vibration frequencies.

The Schrodinger Equation can be solved analytically for only a few forms of the potential
energy function. So, we will consider two numerical approaches to solving the
Schrodinger Equation: (1) finite difference method in which the second derivative is
approximated as a difference formula and (2) a matrix method where the eigenvalues of a
matrix gives the total energies of the particle and the eigenfunctions the corresponding
wavefunctions.

5
FINITE DIFFERENCE METHOD
One can use the finite difference method to solve the Schrodinger Equation to find
physically acceptable solutions. One can also use the Matlab ode functions to solve the
Schrodinger Equation but this is more complex to write the m-script and not as versatile
as using the finite difference method. The finite difference method allows you to easily
investigate the wavefunction dependence upon the total energy. The heart of the finite
difference method is the approximation of the second derivative by the difference
formula
(3)
2
2 2
( ) ( ) 2 ( ) ( ) d x x x x x x
dx x
+A + A
=
A

and the Schrodinger Equation is expressed as
(4)
( )
2
2
2 2
( ) 2
( ) ( ) ( ) ( )
d x m
k x x k x E U x
dx

| |
= =
|
\ .

We will consider an electron trapped in a potential well of width L and depth U
0
as
shown in Figure 1. In regions where E > U(x), k(x) is real and ( ) x has a sinusoidal shape
and this corresponds to the classical allowed region (kinetic energy K > 0). In regions
where E < U(x), k(x) is imaginary and ( ) x has an exponential increasing or decreasing
nature and this corresponds to the classical forbidden (kinetic energy K < 0). The function
2
2
d
dx
is the curvature of the wavefunction and its negative is a measure of the kinetic
energy of the particle (Figure 1).

6

Fig. 1. Potential well defined by the potential energy function U(x). The
bound particle has total energy E and its wavefunction is ( ) x .

You can use a shooting method to find E that satisfies both the Schrodinger Equation and
the boundary conditions. We start with
min
( ) 0 x = and a given value for E and solve the
Schrodinger Equation. The value of E is increased or decreased until the other boundary
condition,
max
( ) 0 x ~ is satisfied. If the end boundary condition at x = x
max
is not
satisfied then the wavefunction will exponentially diverge (
max
( ) x x x ).
When a solution is found, the wavefunction decreases exponentially towards the zero as
shown in Figure 2.

e
n
e
r
g
y
position x
U(x)
U = 0
E
U = -U
0
E= K + U
E U
K > 0
classical forbidden
region
classical forbidden
region
classical
allowed
region
2
2
. .
d
curvature K E
dx

( ) x
0 =
positive
curvature
positive
curvature
negative
curvature
U(x)
7

Fig. 2. Physically acceptable solutions are found only when the
wavefunction ( ) x converges to zero at the extreme values of x. The
depth of the well is -400 eV and the width 0.1 nm. [se_fdm.m].

The process of finding physically acceptable values for E and the corresponding
wavefunctions ( ) x can be automated by counting the number of zero crossing of the
wavefunction and adjusting the value of E until the condition
max
( ) 0 x ~ is satisfied. For
the lowest energy state (ground state) E
1
, the wavefunction is zero only at the extreme
values for x and therefore, the number of crossing is zero. The 1
st
excited state E
2
will
have only 1 crossing and the n
th
excited state E
n+1
will have n crossings as shown in
Figure 3.

In the Finite Difference Method, we start with
( ) ( ) ( ) 1 , 2 , , x x x x N = (

( ) ( ) ( ) 1 , 2 , , k k k k N = (

where N is the maximum number of x coordinates, x(1) = x
min
and x(N) = x
max

( )
min
( ) (1) 0 x x = = and ( ) ( ) (2) (1) 1 x x x = + A =
then as x is incremented, the other values of ( ) x are calculated from the equation

( ) ( )
( )
( ) ( )
2
1 1
2 (
c c c c
x k x x x
+
= A for c = 2 to N-1

When a physically accepted solution is found for the n
th
state the wavefunction is
normalized by numerically integrating the wavefunction using Simpsons 1/3 rule
8

2
( )
(1)
( )
( ) ( )
x N
n n
x
n
x
x dx A x
A
= =
}

The m-script se_fdm.m was used to find the total energies and its corresponding
wavefunctions for a potential well of depth -400 eV and width 0.1 nm. The range for the
x-coordinates was from -0.1 nm to + 0.1 nm. The value of E was manually adjusted to
find the physically acceptable solutions as shown in Figure 3. For this potential well,
there are four bound states. The total energy for the ground state, n = 1 is E
1
= -373.84
eV. Thus, the binding energy of the electron or its ionization energy (energy need to free
the electron from its bound state) is E
B
= +373.84 eV.

Fig. 3. The four states for an electron confined by a potential well of
depth -400 eV and width 0.10 nm with x
min
= -0.10 nm, x
max
= +0.10 nm.
[se_fdm.m]

The smaller binding energy, the less accurate are the results because the range of x values
is not large enough to accurately define the exponential convergence to zero of the
wavefunctions as x approaches x
max
. For the case n = 4 as shown in Figure 3, E
4
= -21 eV
ground state n = 1, E
1
= -373.84 eV 1
st
excited state n = 2, E
2
= -296.63 eV
2
nd
3
= -173.5 eV 3
nd
4
= -21 eV
9
when x
max
= 0.1 nm and one can see that the exponential tail is not very well defined.
When x
max
is increased to 0.2 nm, the exponential tail is better defined and the total
energy is E
4
= -24.85 eV. Solutions of the Schrodinger Equation depend upon the width
of the well and the range of x values. One has to make a judgment based upon the
variation of (x) as x approaches x
max
in determining the most suitable range for the x
values. Figure 4 shows the ground state, for the potential well of depth -400 eV and width
0.10 nm when x
max
is increased from 0.10 nm to 0.20 nm. It is now more difficult to find
the total energy for this state since slight variations in E result in exponential diverging
tails.
Fig. 4. The potential well has a depth -400 eV and width 0.01 nm. The x
values range from -0.2 nm to + 0.2 nm. The wavefunction (x) is more
sensitive to the total energy E as the range of x values is increased.
[se_fd.m].

E= - 373.797300 eV E= - 373.797386 eV E= - 373.797396 eV
10
MATRIX METHOD
The one-dimensional time independent Schrodinger equation can be expressed as
(5)
2
2
2 2
2 2
( )
( ) ( ) ( )
2
( ) ( ) ( )
2
( ) ( )
d x
U x x E x
m dx
d
U x x E x
m dx
H x E x
+ =
(
+ =
(

=

where the H is the Hamiltonian operator, which is the operator that corresponds to the
total energy of the system. This is an eigenvalue equation. The action of the operator H
on the function returns the original function multiplied by a constant which could be
complex. This eigenvalue equation is generally satisfied by a particular set of functions
1 2
( ), ( ), , x x and a corresponding set of constants
1 2
, , E E . These are the
eigenfunctions and the corresponding eigenvalues of the operator H. The time
independent Schrodinger equation of a system is the energy eigenvalue equation of the
system. An eigenfunction ( )
n
x describes a state of define energy E
n
. When the energy
of the system in this state is measured, the result will always be E
n
. For the eigenfunction
to represent physical sensible solutions, we require
( ) 0 as
n
x x
so that the wavefunction can be normalized.

For atomic systems it is more convenient to measure lengths in nm (nanometers) and
energies in eV (electron volts). We can use the scaling factors
length: L
se
= 110
-9
to convert m into nm
energy: E
se
= 1.610
-19
to convert J into eV
and so we can write Equation (5) as
(6)
2
2
2 2
2
2 2
2
1
( ) ( ) ( )
2
1
( ) ( ) ( ) where
2
( ) ( )
se se
se se
se se
d
U x x E x
m L E dx
d
C U x x E x C
dx m L E
H x E x

( | | | |
+ =
( | |
\ .\ .
| | ( | |
+ = =
| | (
\ .\ .
=

11
Consider an electron in a potential well (see Figure 1). For energy values below the top of
the well, the physically acceptable solution of the time independent Schrodinger equation
give a discrete set of energies which are the energy eigenvalues and corresponding to
each eigenvalue there is the energy eigenfunction. Quantization of the energy levels of
bound particles arises naturally from the time independent Schrodinger equation and the
boundary conditions imposed for physically acceptable solutions. The spectral lines
observed in atomic systems are the result of transitions between such energy levels.

These eigenstates ( )
n
x represents stationary states and the total wavefunction can be
expressed as
(7)
/
( , ) ( )
n
i E t
n n
x t x e

+ =
This is a state of definite energy, if the energy is measured then the value obtained will be
E
n
. It is called a stationary state, because the probability of locating a particle in an
interval dx is time independent.

Many problems in physics reduce to solving an eigenvalue equation, for example, the
vibrations of a violin string. The eigenvalues and eigenfunctions can be easily found
using the Matlab command eig. The m-script se_solve.m is used to solve the Schrodinger
equation using the Matrix Methos. To solve the Equation (6), we first represent the
continuous functions of x by sets of N discrete quantities expressed as vectors and
matrices. The discrete set of x values is represented by the vector
n
x the corresponding
wavefunctions by the vector
n
. The potential energy is given by a (N-2)(N-2)
diagonal matrix [U] with diagonal element U
n
. A sample code [se_solve.m] for assigning
the diagonal elements for [U] is

U_matrix = zeros(N-2,N-2);

for c = 1 : N-2
U_matrix (c,c) = U(c);
end

12
Next, we have to represent the operator

2
2 2
2
d
m dx

as a matrix of size (N-2)(N-2). From the definitions of the first and second derivatives of
the function y(x), we can approximate them by the equations

1/ 2
1
n
n n
x
dy y y
dx x
~
A

1/ 2 1/ 2
2
1 1
2 2
1 2
n n
n
n n n
x x
x
d y dy dy y y y
dx x dx dx x
+
+
| |
+
~ =
|
|
A A
\ .

Hence, the second derivative matrix for N = 6 can be written as a 44

2
2 1 0 0
1 2 1 0
1
[ ]
0 1 2 1
0 0 1 2
x
(
(
(
=
( A
(

SD
The SD matrix size is (N-2)(N-2) and not NN because the second derivative of the
function cant be evaluated at the end points, n = 1 and n = N. The kinetic energy matrix
[K] is then defined as
| | [ ]
se
C = K SD
We can now define the Hamiltonian matrix as
| | | | [ ] = + H K U
A sample code [se_solve.m] for the generating the Hamiltonian matrix is
% Make Second Derivative Matrix ------------------------------------------
off = ones(num-3,1);
SD_matrix = (-2*eye(num-2) + diag(off,1) + diag(off,-1))/dx2;

% Make KE Matrix
K_matrix = Cse * SD_matrix;

% Make Hamiltonian Matrix
H_matrix = K_matrix + U_matrix;

13
Therefore, the Schrodinger Equation in matrix form is
[ ]
n n n
E = H
This is an eigenvalue equation in matrix form where the action of the Hamilton matrix
results in each value to be the vector
n
being multiplied by a multiplied by the set of
numbers E
n
. The set of numbers E
n
are called the eigenvalues and set of vectors
n
are
the eigenvectors.

This is a single Matlab function that finds both the eigenvalues and eigenvectors. The
syntax of the command is
[e_funct, e_values] = eig(H)
where e_funct is an (N-2)(N-2) matrix with the n
th
column corresponding to the n
th

eigenfunction and e_values is a column vector for the N eigenvalues in increasing order.
Only the negative values of e_values are significant. To obtain the complete eigenvector
we need to include the end points where
min max
( ) ( ) 0
n n
x x = = . A sample Malab code
[se_solve.m] to obtain the discrete set of eigenvalues and normalized eigenfunction is
% All Eigenvalues 1, 2 , ... n where E_N < 0
flag = 0;
n = 1;
while flag == 0
E(n) = e_values(n,n);
if E(n) > 0, flag = 1; end; % if
n = n + 1;
end % while
E(n-1) = [];
n = n-2;

% Corresponding Eigenfunctions 1, 2, ... ,n: Normalizing the wavefunction
for cn = 1 : n
psi(:,cn) = [0; e_funct(:,cn); 0];
area = simpson1d((psi(:,cn) .* psi(:,cn))',xMin,xMax);
psi(:,cn) = psi(:,cn)/sqrt(area);
prob(:,cn) = psi(:,cn) .* psi(:,cn);
end % for

A potential well as shown in Figure (5) has a minimum at x = 0 and tends to zero away
from the origin. A classical particle would be trapped by this potential well and oscillate
to and fro about x = 0 because the force on the particle is always directed towards origin
for position since F = -dU/dx.
14

Fig. 5. A potential well which traps a particle because the force acting on
the particle is always directed towards the origin. [guassian_p.m
annotation of figure done in MS Powerpoint].

We can solve the Schrodinger equation ( Equation 5) for a variety of different potential
energy functions. The m-script SE_wells.m defines most of the constants and potential
well parameters. Within the m-script you can change the code to modify the potential
wells and add new potential wells. The first step in solving the Schrodinger Equation
using the Matrix Method for a bound electron is to run the file SE_wells.m and select the
type of potential well. The default potentials include:
1 Square well
2 Stepped well
3 Double well
4 Sloping well
5 Truncated Parabolic well
6 Morse Potential well
7 Parabolic fit to Morse Potential
8 Lattice
x
x
U
U = 0
F = 0
x = 0
F
force on bound electron
dU
F
dx
=
15

Fig. 6. Some of the default potential wells that can be generated using the
m-script se_wells.m.

-1.5 -1 -0.5 0 0.5 1 1.5
x 10
-10
-400
-350
-300
-250
-200
-150
-100
-50
0
Potential Well: SQUARE
x (nm)
e
n
e
r g
y
( e
V
)
-1.5 -1 -0.5 0 0.5 1 1.5
x 10
-10
-400
-350
-300
-250
-200
-150
-100
-50
0
Potential Well: STEPPED
x (nm)
e
n
e
r g
y
( e
V
)
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-400
-350
-300
-250
-200
-150
-100
-50
0
50
100
Potential Well: DOUBLE
x (nm)
e
n
e
r g
y
( e
V
)
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15
-400
-350
-300
-250
-200
-150
-100
-50
0
Potential Well: SLOPING
x (nm)
e
n
e
r g
y
( e
V
)
-3 -2 -1 0 1 2 3
x 10
-10
-35
-30
-25
-20
-15
-10
-5
0
Potential Well: Truncated PARABOLIC
x (nm)
e
n
e
r g
y
( e
V
)
-1 0 1 2
x 10
-10
-5
0
5
10
15
20
25
Potential Well: MORSE potential
x (nm)
e
n
e
r g
y
( e
V
)
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
-350
-300
-250
-200
-150
-100
-50
0
Potential Well: Lattice
x (nm)
e
n
e
r
g
y

(
e
V
)
16
If you are not satisfied with the potential well that is displayed, you can change any of the
parameters in the m-script se_wells.m and run it again. Then to solve the Schrodinger
equation run the m-script se_solve.m.

Consider a sloping potential well with input parameters set in se_wells.m
% Input parameters
num = 801;
xMin = -0.1; % default value = -0.1 nm
xMax = +0.1; % default value = + 0.1 nm
U1 = -1200; % Depth of LHS well: default = -1200 eV;
U2 = -200; % Depth of RHS well: default = -200 eV;
x1 = 0.05; % 1/2 width of well: default = 0.05 nm;

The solution of the Schrodinger equation for this sloping potential well using se_solve.m
provides the following information:

(1) The energy eigenvalues are displayed in the Command window
No. bound states found = 5
Quantum State / Eigenvalues En (eV)
1 -894.67
2 -624.3
3 -403.46
4 -203.93
5 -13.546

The energies are stored in the variable E. E(1) is the first energy level (ground state),
E(2) is the second energy level, and so. The values can be displayed at any time in the
Command Window by simply typing E
>> E
E =
-894.6664 -624.3014 -403.4580 -203.9267 -13.5456

17
(2) A graph of the energy spectrum as shown in Figure (7) for a sloping potential well

Fig. 7. The energy spectrum for a potential well produced by the m-script
se_solve.m.

(3) The energy eigenvectors are given by the array psi with dimensions
(numN) where num is the number of data points and N is the number of
eigenvalues. For example, to display the eigenvector for quantum state n =
2, type psi(:,2) into the Command Window. A graphical display of the first 5
eigenvectors and corresponding probability density distributions are
displayed in a Figure Window as shown in Figure (8).

-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
-1200
-1000
-800
-600
-400
-200
0
position x (nm)
e
n
e
r
g
y

U
,

E
n

(
e
V
)
Potential Well: SLOPING
18

Fig. 8. The energy eigenvectors and probability distributions for sloping
potential well. [se_solve.m].

To view a graphical display of an eigenvector and the probability density graphs for a
given state, you can run the m-script se_psi.m from the Command Window. The
graphical output for a sloping potential well for quantum state n = 3 is shown in Figure
(9). Variations with position of the eigenvector ( )
n
x ; probability density
2
( )
n
x ;
energies U(x), K(x), E are shown along with a probability cloud where each point
displayed shows the position of the particle after a measurement is made on the system
resulting in the collapse of the wavefunction.

n = 1
2
n = 1
n = 2
2
n = 2
n = 3
2
n = 3
n = 4
2
n = 4
n = 5
2
n = 5
19

Fig. 9. The eigenvector, probability density and energy graphs for the
quantum state n = 3 of a particle trapped in a sloping potential well.
[se_psi.m] .

We will again consider the sloping well. The sloping potential well represents an external
electric field acting on the electron. Classically, the electron is subject to a very large
force to the right at the left edge of the well and a force to the left along the sloping
section with the magnitude of the force equal to the gradient of the slope ( / F dU dx = ).
We could also view the potential as a quantum mechanical version of a particle sliding
along a frictionless inclined plane with a high wall at the bottom. The classical motion of
such a confined particle is that it would oscillate up and down the incline, having its
largest kinetic energy at the bottom and decreasing kinetic energy the further up the
incline. The particle would be moving most rapidly at the bottom of the incline and most
slowly at the top. Therefore, it would spend less time at the bottom than at the top and
this implies that it is more likely to find the particle at the top than at the bottom. Also the
particle could be at rest at the bottom.

Figure (10) shows the quantum mechanical solution for the sloping well potential for the
state n = 1 (ground state). The results are very different from the classical predictions.
-0.1 -0.05 0 0.05 0.1
-6
-4
-2
0
2
4
6
Potential Well: SLOPING n = 3 E
n
= -403.458
position x (nm)
w
a
v
e

f
u
n
c
t
i
o
n

p
s
i
-0.1 -0.05 0 0.05 0.1
-1200
-1000
-800
-600
-400
-200
0
U

(
e
V
)
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
0
10
20
30
position x (nm)
p
r
o
b

d
e
n
s
i
t
y

p
s
i
?
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
-1000
-500
0
U

(
e
V
)
-0.1 -0.05 0 0.05 0.1
-1000
-500
0
500
1000
E
n
(black) U(red) K(green)
position x (nm)
e
n
e
r
g
i
e
s

(
e
V
)
20
The minimum energy of the particle is greater than zero. There is a non-zero probability
of the electron being in a classically forbidden region (E
K
< 0), i.e., the electron can
penetrate the walls of the potential well and the most likely location to find the electron is
towards the bottom and not at the top.

quantum state n = 1 (ground state) of a particle trapped in a sloping
potential well. [se_psi.m].

There are 5 bound states for the potential well shown in Figures (10, 11). As the quantum
number n increases by 1, an additional hump of width in the order of half a wavelength is
added to the shape of the wavefunction. For the state n = 1 there is only 1 hump and the
state n = 5, there are 5 humps. The characteristics of the highest bound state (n = 5) is
shown in Figure (11). This highly excited state has features more in-line with the
classical picture. The particle is more likely to be found towards the top of the well where
its kinetic energy is less. Since the particle moves at high speed at the bottom and less at
the top, the wavelength of the wavefunction must be smallest near the bottom and
-0.1 -0.05 0 0.05 0.1
0
5
10
n
= -894.6664
position x (nm)
w
a
v
e

f
u
n
c
t
i
o
n

p
s
i
-0.1 -0.05 0 0.05 0.1
-1000
0
U

(
e
V
)
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
0
20
40
position x (nm)
p
r
o
b

d
e
n
s
i
t
y

p
s
i
?
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
-1000
0
U

(
e
V
)
-0.1 -0.05 0 0.05 0.1
-1000
-500
0
500
1000
E
n
position x (nm)
e
n
e
r
g
i
e
s

(
e
V
)
21
increases towards the top. The ground state defies our classical expectations but the
highly excited states are often in broad agreement with the classical picture. This is not so
surprising since classical physics objects have very high quantum numbers. This idea is
called the correspondence principle.

Measurements of the half-wavelengths can be made using the ginput command
to measure the positions of the peaks and then the successive distances between
the peaks of the probability density curve, for example,
z = ginput
z(5) z(4)

The measured half-wavelengths (nm) from Figure (11) as the kinetic energy
decreases from left to right are 0.0184 0.0208 0.0241 0.0315 confirming that
as the kinetic energy decreases the wavelength must decrease since

2 2
2
2 2
K
h p h
p E
m m
= = =

excited state, n = 5 of a particle trapped in a sloping potential well.
[se_psi.m].

-0.1 -0.05 0 0.05 0.1
-6
-4
-2
0
2
4
6
n
= -13.5456
position x (nm)
w
a
v
e

f
u
n
c
t
i
o
n

p
s
i
-0.1 -0.05 0 0.05 0.1
-1200
-1000
-800
-600
-400
-200
0
U

(
e
V
)
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
0
10
20
position x (nm)
p
r
o
b

d
e
n
s
i
t
y

p
s
i
?
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
-1000
0
U

(
e
V
)
-0.1 -0.05 0 0.05 0.1
-1000
-500
0
500
1000
E
n
position x (nm)
e
n
e
r
g
i
e
s

(
e
V
)
22
EXPECTATION VALUES

The quantum behavior of bound particles is becoming more important in electronic
devices made using semiconductor materials such as silicon and gallium. The heart of
such devices is a tiny structure called a quantum dot that consists of a speck of one
semiconductor material (width ~ 1 nm, ~100s atoms) completely enclosed by a much
larger semiconductor material. Some of the electrons in the tiny speck become detached
from their parent atoms and behave as particles trapped in potential wells of 1 (quantum
wafer), 2 (quantum wire), or 3 (quantum dot) dimensions. These particles are often
referred to as particles in a box. The bound electrons have a discrete energy spectrum just
like electrons bound in atoms. Quantum dots are often referred to as artificial atoms. The
light emitting properties of quantum dots are used in solid state lasers such as DVD
players, lighting systems, solar cells, fluorescent markers in biomedical applications. Our
focus of finding the physically acceptable solutions of the Schrodinger equation for
bound particles is important to understanding real systems of increasing technological
importance.

The wavefunction that describes a quantum state provides the most complete description
that is possible of that state it contains all the information we can possibly know about
this quantum state. But the wavefunction is not something that can be measured and
therefore has no physical meaning. The information we can get from the wavefunction is
related only to the probabilities of the possible outcomes of measurements made on
identical systems. Hence, there is an intrinsic and unavoidable indeterminism in the
quantum world.

When a measurement is made on a quantum system, the wavefunction collapses, the
state of the system after the measurement is not the same as the state before the
measurement. The Schrodinger equation does not describe this collapse of the
wavefunction. So, for experiments designed to test the predictions of quantum mechanics,
you cant do a sequence of measurements on a single system, rather, you start with a
large number of identical systems all in the same state and then perform the same
measurement of each system. Although each system is initially identical, the outcomes of
23
the measurements typically give a spread of results, for example, the measurement of the
position of the electron in passing through an equivalent of a double slit.

We will consider systems of a single electrons confined by a one dimensional potential
wells. We already know that it is a simple matter to use the wavefunction to find the
probability distribution for the position of the electron. However, we can extend this
method to find the probability distributions of other measureable quantities such as
position, momentum and energy. This information is contained in the shape of the
wavefunction and we will show how this information can be extracted.

For the electron trapped in the sloping potential wells (Figures 7, 9, 10, 11) the solution
of the Schrodinger equation gave a set of energy eigenvalues and their corresponding
eigenvectors (stationary states). The energy eigenvalues are the only allowed energies of
the system. No matter what state the system is in, if we measure its energy, we always get
one of the energy eigenvalues. These stationary states are special, because they are states
of definite energy. So in this case, there is zero uncertainty in the measurements. If the
system is described by a stationary state wavefunction ( , )
n
x t + and we measure its
energy, we will certainty get the result E
n
for the energy eigenvalue.

The eigenvectors are said to be orthonormal: (1) the eigenvector is normalized so that the
probability is 1 of finding the electron, and (2) any two energy eigenvectors with
different energy eigenvalues are orthogonal to each other. Mathematically this is stated
as:

(8)
*
1
( ) ( ) Kronecker delta
0
j i ji ji
j i
x x dx
j i
o o
+
=
= =

=
}

One meaning of the wavefunction is that its magnitude measures the probability
amplitude of the particle being found at some point. If the wavefunction corresponds to
an eigenfunction ( , )
n
x t + , then the probability distribution prob
n
(x) is independent of
time ()and is given by

24

* *
( ) ( ) ( ) ( ) ( )
n n
iE t iE t
n n n n n
prob x x e x e x x
= =

and the probability P
n
of finding the particle between x
1
and x
2
is

2
1
*
( ) ( )
x
n n n
x
P x x dx =
}

We can also calculate probability distributions for other observable quantities in a similar
manner to above. The probability distribution is characterized by two measures its
expectation value (estimate of the mean value of the distribution) and its uncertainty
(spread of values around the mean). Suppose that an observable quantity A has a discrete
set of possible values a
1
, a
2
, . Then, we can prepare N identical systems, all in the
same state and measure the value of A in each system. In general, we will get a spread of
results. The best estimate of the quantity A is its mean value <A> and the standard
deviation A A can be used as a measure of the spread or uncertainty in the measurements.
The expectation value of an observable A is the quantum-mechanical prediction for the
mean value. In a stationary state described by the wavefunction ( )
n
x , the expectation
value of an observable A is given by

(9)
*
( ) ( )
n n
A x A x dx
+
=
}

where
A is the quantum-mechanical operator corresponding to A. The operator is

sandwiched between the two wavefunction
*
( )
n
x and ( )
n
x , and for this reason we
call the right hand side of Equation (9) a sandwich integral. The first part of the integrand
is called the bra (
*
( )
n
x ) and the second part the ket (
( )
n
A x ). Some of the dynamical
properties of a system that can be calculated are shown in Table 1.

25

Total probability
1 A=
*
1 ( ) ( )
n n
x x dx
+
=
}

Position
A x =
*
( ) ( )
n n
x x x x dx
+
=
}

Momentum
*

d
A i
dx
=
*
( ) ( )
n n
d
p x i x dx
dx

+
| |
=
|
\ .
}

Kinetic Energy
2 2
2
2
d
A
m dx
=
2 2
*
2
( ) ( )
2
K n n
d
E x x dx
m dx

+
| |
=
|
\ .
}

Potential Energy
( ) A U x =
*
( ) ( ) ( )
P n n
E x U x x dx
+
=
}

Total energy
2 2
2
( )
2
d
A H U x
m dx
= = +
where H is the Hamiltonian
( )
*
2 2
*
2
( ) ( )
( ) ( ) ( )
2
n n
n n
E x H x dx
d
E x U x x dx
m dx

+
=
| |
= +
|
\ .
}
}

Table 1. Operators and expectation value integrals for a range of dynamical
quantities.
*
Note: For the momentum, the integrand is imaginary, hence = 0.

The standard deviation A A is used to describe the spread of the measurements about the
mean value. It is measured by summing the square deviations from the mean and then
taking the square root. For a system in a given state ( )
n
x , the uncertainty of the
observable A is defined by
(10)
( )
( )
1/ 2 1/ 2
2 2
2
A A A A A A = =

The fundamentals of quantum mechanics were worked out in a remarkably short period
of time from 1925 to 1927. One of the fundamental consequences of quantum theory is
the Heisenberg uncertainty principle. One form of the uncertainty principle is:

In any system, the uncertainties of position x A and momentum component
x
p A obey the
inequality
(11)
2
x
x p A A >

26
Hence, it is impossible to find a state in which a particle has both definite values for
position and momentum. The classical picture of a particle following a well-defined
trajectory is no longer valid. A system described by a wavefunction implies a spread of
position and momentum values and that a narrow spread in one of these quantities is
offset by a wide spread in the other. The uncertainty principle is sometimes interpreted in
terms of the disturbance produced by a measurement, but, this is not correct. The
uncertainty principle tells us about the indeterminacy that is inherent in a given state
before the measurement is made.

When a systems wavefunction corresponds to one of its eigenfunctions
n
, the
uncertainty in the total energy E
n
is zero, 0
n
E A = . This is show by the following
calculation, in which the operator for the total energy E is the Hamiltonian
H :

( )
( )
* *
2
2
2 * * 2
1/ 2
1/ 2
2
2 2 2

0
n n n n
n n n n
H E
E H dx E dx E
E E
E H H dx E H dx E
E E E E E

+ +

+ +

=
= = =
=
= = =
A = = =
} }
} }

In our models, the range of x values is restricted to the range, x
Min
to x
Max
and the
wavefunction is forced to zero at these extreme values of x. So an expectation value is
calculated as
(12) ( ) ( )
Max
Min
x
x
A bra ket dx =
}

The integral is evaluated numerically using Simpsons 1/3 rule by using the m-script
simpson1d.m. The m-script se_measurements.m can be used to evaluate the integrals in
Table (1) and to evaluate the uncertainties. Sample codes from the m-script
se_measurements.m to evaluate the expectation value for total energy and the
uncertainty in position are:
27

% total energy need to convert units in calculations
psi1 = gradient(Psi,x);
ket = - (hbar^2/(2*me*e))*gradient(psi1,x)(Lse^2*Ese) + (U.*Psi')';
braket = bra .* ket;
Eavg = simpson1d(braket,xMin,xMax);

% position x & uncertainty in position
ket = x .* Psi;
xavg = simpson1d(braket,xMin,xMax);

ket = (x.^2) .* Psi;
x2avg = simpson1d(braket,xMin,xMax);

deltax = sqrt(x2avg - xavg^2) * Lse;
% length units nm to m

The m-script se_measurements.m outputs numerical values of the integrals to the
Command Window and graphical representations in a Figure Window (Figure (12)). For
example, the sloping well potential:

Quantum number, n = 3
Energy, E = -403.458 eV
Total Probability = 1
<x> = -0.000569644 nm
<x^2> = 0.000708057 nm^2
<ip> = -3.53767e-030 N.s
<ip^2> = 7.36845e-047 N^2.s^2
 = -655.895 eV
<K> = 252.787 eV
<E> = -403.107 eV
<K> + = -403.107 eV

deltax = 2.66032e-011 m
deltap = 8.58397e-024 N.s
(dx dp)/hbar = 2.16456

28
From the above result, we can conclude the following:
1 E = < E > ignoring the slight difference in numerical values due to numerical
inaccuracies in our modeling.
2 < E > = < K > + 
3 The average position of the electron is slightly to the left of the origin.
4 The uncertainty principle is satisfied 2.2 0.5
x p A A
= > .
5 The expectation for momentum is complicated by the factor i and so we have to
calculate which is then a real quantity and from the results above, we can
conclude that = 0. No measurable physical quantity can be imaginary, hence
the integral must be zero.

It is remarkable that a bound particle has definite set of discrete values for its total energy
and this follows simply and directly from applying boundary conditions to a
wavefunction that satisfies the Schrodinger equation. At first, it seems surprising that the
total energy has a set of unique values ( 0 E A = ) but the momentum does not, since the
uncertainty principle must be satisfied. But a classical vibrating string in a normal mode
has a unique frequency of vibration while not possessing a unique wavelength as the
transverse standing wave is sinusoidal along the string but is zero at and outside the fixed
end points. However, the shape of the string inside and outside the fixed ends can be
thought of a superposition of waves of different wavelengths with amplitudes and phases
chosen so that the disturbance cancels everywhere except in the region between the end
points. Therefore, the wavefunction can be thought of a superposition of different
wavelengths and through the deBroglie relationship / p h = ,the confined particle has a
range of momentum.

29

Fig. 12. The graphical output from the m-script se_measurements.m for the
quantum state n = 3 of a particle trapped in a sloping potential well. The lower
graphs shows the bra (wavefunction), the ket and the area under the braket curve
representing the integral for the evaluation of the expectation value for position
<x> The blue bar indicates the expectation value for position. You can modify the
m-script se_measurements.m to produce graphical outputs for operators other
than the position operator x .

-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
-1500
-1000
-500
0
P
E
,

U

(
e
V
)
Quantum State, n = 3
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
-5
0
5
w
a
v
e

f
u
n
c
t
i
o
n
,

-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
0
10
20
30
position x (nm)
p
r
o
b
.

d
e
n
s
i
t
y
,

2

(
n
m
-
1
)

-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
-5
0
5
w
a
v
e

f
u
n
c
t
i
o
n
<x> = -0.00056964 nm
-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
-0.2
-0.1
0
0.1
0.2
k
e
t
,

-0.1 -0.08 -0.06 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1
-1
-0.5
0
0.5
b
r
a

k
e
t

position x (nm)
30
ACCUARY OF THE NUMERICAL MODELLING

Numerical results may not always be accurate. Where possible you need to check that the
results are OK. Firstly, you should always vary the number of points used in the
modeling. In solving the Schrodinger equation by the Matrix Method, the variable num
which is set in the m-script se_wells.m determines the size of the vectors and arrays.
Generally, if this number is too low the results are not accurate. You should start with a
small number for num and increase it until there is a convergence in the results. If num is
too large, then the execution time in running the m-scripts could be excessive and/or too
much memory is used. Secondly, if possible you should compare the predictions of a
numerical method with the known results predicted by analytical methods. The energy
eigenvalues for an infinite square well can be easily found analytically, so we can
compare those results with our numerical model.

Consider a system with an electron of mass m
e
confined to a one dimensional region of
length L. The potential well called an infinite square well has infinitely high walls which
trap the particle, but the force on the electron is zero between the walls. The potential
energy function and boundary conditions on the wavefunctions are

( ) 0 / 2 / 2
( ) / 2 and / 2
( / 2) 0 ( / 2) 0
U x L x L
U x x L x L
L L
= s s
= < >
= + =

The classical predictions for such a system, is that the total energy of the trapped particle
can have any positive value and since the potential energy inside the well is zero, the
kinetic energy of the particle equals the total energy. The particle can have zero total
energy (zero kinetic energy) and if its kinetic energy is greater than zero the particle will
bounce back and forth between the impenetrable walls with constant speed. Increasing
the energy simply increases the speed but no matter how great the speed, the particle
cant be outside the walls, the region outside the walls is said to be classically forbidden.

The predictions of quantum mechanics are very different. In solving of the Schrodinger
equation it is found that the shape of the wavefunction inside the walls is sinusoidal and
31
the wavefunction must be zero at the walls since the potential energy is infinite at each
wall. Therefore, the only acceptable physical solutions are those wavefunctions where
integer multiple of half-wavelengths of a sine curve fit into the distance L between the
walls:

2
1, 2, 3,
2
L
L n n
n

| |
= = =
|
\ .

The total energy E is related to the momentum p and the momentum related to a
wavelength (de Broglie relationship) hence, we can determine an expression for the
total energy as a function of n (principle quantum number) and the length L

2
2
2
L h p
p E
n m
= = =
(13)
2
2
2
1, 2, 3,
8
n
h
E n n
mL
| |
= =
|
\ .

This is a momentous result and very different from the predictions of classical physical.
The principle quantum number n is restricted to integer values, therefore, the bound
particle has only a discrete set of values for its total energy and these are the only
possible values for the total energy of the system. The lowest value is n = 1, and the
lowest allowed energy is
( )
2 2
1
/ 8 0 E h mL = > . No particle in an infinite square well can
have an energy lower than this, there is no state of zero total energy. This energy is called
the ground state energy. If L , then the spacing between energy 0 E A which
implies that the particle is behaving more like a free classical particle which can have any
positive energy.

We can compare the predictions from the analytical model and the Matrix Method for an
electron confined to an infinite square well potential of width 0.1 nm. To use the Matrix
Method, modify m-script se_wells.m to produce an infinite square well potential then run
it:
Infinite square well parameters set in se_wells.m
num = 801
32
Option 1 (case 1) square well
xMin = -0.05
xMax = +0.05
U1 = 0

Then run the m-scipt se_solve.m. The screen output states that no eigenvalues are found.
This is because the code is expecting negative values for the total energy. However, we
can extract the required values for the total energy manually using the command
e_values(mm,mm) in the Command Window. Do this and record the first 10 eigenvalues
for mm = 1, 2, , 10. These values are stored in the m-script se_infwells.m. Run the m-
script se_infWells.m. Note, when this file is run, all variables are cleared from the
workspace. The file se_infWells.m calculates the first 10 energy levels using Equation
(13), outputs the results to the Command Window and plots the numerical and analytical
values in a Figure Window (Figure (13)):
n E (eV) Emm (eV)
1 37.74 37.69
2 150.95 150.74
3 339.64 339.17
4 603.80 602.97
5 943.44 942.13
6 1358.56 1356.60
7 1849.15 1846.50
8 2415.21 2411.70
9 3056.76 3052.30
10 3773.77 3768.10
33

Fig. 13. The energy eigenvalues for an infinite square well of width 0.10
nm calculated from Equation (13) and the Matrix Method using the m-
scripts se_wells.m and se_solve.m. [se_infWell.m]

The discrepancy between the two sets of results is better than 0.2%. Therefore, we can
have some confidence in the results from the Matrix Method for other potential well
functions.

The wavefunctions for the infinite square well can be plotted after running se_solve.m by
typing the following commands in the Command Window, for example, to display the
eigenvector for the state n = 8
plot(e_funct(:,8),'lineWidth',3)
axis off

Figure (14) show the eigenvectors for the first 10 quantum states of an infinite square
well. The eigenfunction is zero outside the well and is of a sinusoidal shape inside the
well. For the n
th
state, n half-wavelengths of a sinusoidal curve fit into the width of the
well. The shape of the wavefunction is the same as the shape of a vibrating guitar string
for normal modes of vibration.
1 2 3 4 5 6 7 8 9 10
0
500
1000
1500
2000
2500
3000
3500
4000
Quantum No. n
T
o
t
a
l

E
n
e
r
g
y

E

(
e
V
)

analytical
Matrix Method
34

Fig. 14. The energy eigenvectors for an infinite square well for the first 10
quantum states. [se_wells.m se_solve.m]

When a potential well is of finite depth, the wavefunction extends into the classically
forbidden region and so there is some probability of finding the particle outside the well
which in classical terms in impossible. For a particular state of an infinite square well, the
wavelength is constant (independent of x). This is not the case when the potential inside
the well changes with x, for example, the wavelength increases with decreasing potential
in the sloping well (Figure (9,10)) and so the shape of the wavefunction is not simply
sinusoidal.

1 2
4
5 6
7 8 9 10
3
35
STATIONARY STATES AND SUPERPOISITION PRINCIPLE

In wave mechanics, the state of a system is described by a wavefunction that satisfies the
Schrodinger equation and the boundary conditions imposed on the system. For a one
dimensional system, the Schrodinger equation for a potential energy function U(x) is
(14)
2 2
2
( , ) ( , )
( ) ( , )
2
x t x t
i U x x t
t m x
c+ c +
= + +
c c

We will consider only those solutions that are described by a state of definite energy by
the method of separation of variables where the wavefunction is expressed as
(15) ( , ) ( ) ( ) x t x T t + =
The substitution of this wavefunction into the Schrodinger equation, gives two ordinary
differential equations and for quantum state n:
(16)
2
2 2
( )
( ) ( ) ( )
2
n
n n n
d x
U x x E x
m dx
+ =

time independent Schrodinger equation
(17)
( )
( )
n
n n
T t
i E T t
t
c
=
c
n = 1, 2, 3 ,

The solution of the ordinary differential equation (17) is
(18) ( ) exp
n
n
i E t
T t
| |
=
|
\ .

For stationary states which have a definite total energy, the wavefunction is
(19) ( , ) ( ) exp
n
n n
i E t
x t x
| |
+ =
|
\ .

If the energy of the system is measured, the value E
n
will certainly be obtained. This is a
special type of solution of the Schrodinger equation.

Applying Borns rule, which states that the probability of finding the particles in a small
interval dx centered on x for the state n is

2 2
*
( , ) ( , ) ( , ) ( )
n n n n
probability x t dx x t x t dx x dx = + = + + =

36
Hence, the probability is independent of time t and this is why it is called a stationary
state. The real and imaginary parts of the wavefunction change with time but not the
modulus. All parts oscillate in phase (like a stationary normal mode of a vibrating guitar
string). The concept of stationary states is not one that is familiar to us in the classical
world as nothing is moving at all, the probability of detecting an electron in any given
region never changes. All the expectation values listed in Table (1) are also independent
of time.

Each stationary state wavefunction as described by Equation (19) describes a complex
standing wave. A standing wave is one that oscillates without propagation through
space, with all points of the standing wave vibrating in phase. The frequency and period
of the vibration are

n n
n n n
n
E E h
f T
h E
e = = =

The standing wave has nodes fixed in space. As n increases, the total energy E
n
and
number of nodes increases, while the period of vibration T
n
decreases.

There is zero probability of locating the particle bound in the well at each node. We can
no longer think of the particle moving from place to place in the well. In quantum
physics, a particle has no position until it is measured, so it makes no sense to think of the
particle moving about. There is no problem associated with the particle having to travel
through a nodal point, since the particle has no position and velocity. When a
measurement is made, it only tells us the position and velocity at the instant after the
measurement. Before the measurement, the state is simply described by the wavefunction
which gives us the most complete description that is possible of the state of the system,
but it does not give us information about the position and velocity of the particle.

37
Figure (15) show the time evolution for a number of time steps of the real and imaginary
parts of the wavefunction and the probability density for the stationary state n = 3 of the
truncated harmonic oscillator. The phase of the real and imaginary parts change with time
but the probability density is independent of time. An animation of the changes with time
of the wavefunction can be observed by running the m-script se_stationary.m.

Fig. 15. The time evolution of the wavefunctions real and imaginary
parts for the stationary state n = 3 for a truncated parabolic well. The
probability density does not change with time.
[se_wells.m se_solve.m se_stationary.m]

An animated gif [se_stationary.m] of the time evolution of the stationary
state shown in Figure (15) can be viewed at

www.physics.usyd.edu.au/teach_res/mp/quantum/gif/ag_stationary.gif

real +
imaginary +
prob. density
38
Wavefunctions have a very important property in that they obey the superposition
principle:

If a set of wavefunctions are solutions of the Schrodinger equation, then any linear
combination of the wavefunction is also a solution

n n
n
a + = +
where a
c
are arbitrary complex numbers
If the wavefunctions
n
+ are normalized wavefunction and + is also normalized then

2
1
n
n
a =

The probability of finding a particle in a given length element dx is

*
( , ) ( , ) x t x t dx + +

When the particle is in an eigenstates with a definite energy E
n
then

*
( , ) ( , ) *( ) ( ) constant
n n
x t x t dx x x dx + + = =

and so the probability distribution does not change with time as shown in Figure (15).
Also, the expectation value for the position of the particle x is not a function of time.
However, when the particle is in a compound state (more than one a
n
is nonzero), the
probability density
2
( , ) x t + , the expectation value for position x and probability of
locating the particle at each position all vary with time.

The expectation value for the total energy is

*
( , ) ( , ) E x t i x t dx
t
+
c | |
= + +
|
c
\ .
}

Using the fact that the eigenfunctions are orthonormal

*
1
( ) ( ) Kronecker delta
0
j i ji ji
j i
x x dx
j i
o o
+
=
= =

=
}

and doing some algebra it is not difficult to show that the expectation value for the total
energy is
39
(20)
2
n n
n
E a E =
where E
c
are the eigenvalues
Any measurement of the total energy of the system will always give one of the
eigenvalues E
n
and not E . If many identical system with the compound wavefunction
+ were prepared and the total energy measured for each system than the average value
those measurements would be very close to the expectation value E .

Run se_wells.m and se_solve.m for the default truncated parabolic well:
% parabolic *****************************************************
case 5
xMin = -0.2; % default = -0. nm
xMax = +0.2; % default = +0.2 nm
x1 = 0.2; % width default = 0.2 nm;
U1 = -400; % well depth default = -400 eV;

The energy eigenvalues for this potential well from se_solve.m are:
No. bound states found = 5
1 -360.92
2 -282.77
3 -204.68
4 -127.09
5 -52.272

The m-script se_super.m adds any two of the eigenfunctions to form a compound state.
To investigate the addition of the eigenstates 1 and 2, in the code for se_super.m enter
qn(1) = 1; qn(2) = 2; % states to be summed
ac(1) = 0.5; ac(2) = sqrt(1-ac(1)^2); % coefficients
to form the compound state
(21)
1 2
1 2
( , ) 0.5000 ( ) exp 0.8660 ( ) exp
i E t i E t
x t x x
| | | |
+ = +
| |
\ . \ .

The expectation value (eV) for this compound state is displayed in the Command
Window
Eavg_s = -302.3036 see equation (20)
Figure (16) shows the time evolution of the compound state at times t = 0, T/4, T/2, 3T/4,
T where T is the period of the oscillation of the wavefunction. The red curves are the real
parts and the blue curves the imaginary part of the compound wavefunction. The black
40
curves shows the variation with time of the probability density. For the compound state
where
2 1
2 1 1 n n n A = = = + the probability curve oscillates back and forth in the
potential well with a period T where
2 1
h
T
E E
=

parts for a compound state given by Equation (21) for a truncated
parabolic well. The probability density changes with time as the charge
distribution oscillates back and forward in the potential well. Red curves
real part + , blue curves imaginary part + and black curves probability
density.
[se_wells.m se_solve.m se_super.m]

An animated gif [se_super.m] of the time evolution of the stationary state
shown in Figure (16) can be viewed at

www.physics.usyd.edu.au/teach_res/mp/quantum/gif/ag_super12.gif

Therefore, the expectation of the position of the electron x oscillates with the period T
as shown in Figure (17).

real +
imaginary +
prob. density

real +
imaginary +
prob. density

real +
imaginary +
prob. density

real +
imaginary +
prob. density

real +
imaginary +
prob. density
t = 0 t = T/4
t = T/2 t = 3T/4
t = T
41

Fig. 17. The time variation of the expectation value for the position of the
electron x in the truncated harmonic oscillator potential well. The
oscillation period is T = 5 .29410
-17
s.

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
x 10
-16
-0.02
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
0.02
time t (s)
<
x
>

(
n
m
)
42
Radiation emitted by the system
In classical electromagnetism, an accelerated charge radiates electromagnetic radiation at
a rate that increases with its acceleration. In quantum mechanics, the acceleration of a
particle is not defined and an electron in a given state does not radiate energy. But, we
can explain the radiation due to transitions between states using a semi-classical approach
by considering the expectation value x instead of the classical well defined position
coordinate x. Consider an electron in a stationary eigenstate. Its expectation value x
does not depend upon time

*
( , ) ( , ) *( ) ( ) constant
n n
x t x x t dx x x x dx
+ +

+ + = =
} }

therefore,
2
2
0
d x
dt
= and no energy is radiated.
So, a stationary state does not radiate. However, the bound particle is not alone in the
universe, sooner or later the particle will interact with its surroundings (photon, or some
other particle, etc), perturbing the energy so that the particle is no longer in a stationary
state. The particle will start in its initial stationary state, absorb a photon for example, and
then exist in a compound state where the coefficients a
n
are now time dependent. In the
compound state, the system does not have a definite energy and the expectation value
x is now time dependent. Finally a transition will occur to a lower energy eigenstates
and a photon will be emitted. The product q x where q is the charge on the particle is
the x component of an oscillating electric dipole moment for which the frequency of the
radiation emitted is
i f
E E
f
h
=

For the our compound state in the harmonic potential well, the variation the expectation
values with time depends on the exponential term
( )
1 2
exp
i E E t | |
|
\ .

and the expectation value x varies with time as shown in Figure (17) because of the
variation with time of the probability distribution. The frequency of oscillation of x is y
43
given by
i f
E E
f
h
= . Hence, the frequency of the emitted photon is the same as the

frequency at which the charge distribution changes. Assume that the initial state of the
system is the eigenstates n = 2 and E
2
= 282.7 eV. Then system is disturbed to produce
the compound state given by Equation (21). In this complex state, the charge distribution
oscillates with a frequency f
21
and finally a photon is emitted with frequency f
21
as the
system loses energy and returns to the ground state n = 1, E
1
= - 360.92 eV. The values of
the frequency and wavelength of the emitted photon for the transition are:
n = 2 n = 1 An = +1 f
21
= 1.8910
16
Hz
21
= 1.5910
-8
m (UV)

We will now consider the compound state which is a summation of the two eigenstates
n = 1 and n = 3
(22)
3 1
1 3
( , ) 0.5000 ( ) exp 0.8660 ( ) exp
i E t i E t
x t x x
| | | |
+ = +
| |
\ . \ .

The time evolution of the real and imaginary parts of the wavefunction and the
probability density are shown in Figure (18). However, even though the probability
density changes with time, there is no sloshing back and forward and the expectation
value for position is 0 x = . Therefore, no radiation can be emitted in the transition n = 3
n = 1 for our harmonic oscillator potential well. So, in the case of the one-dimensional
harmonic oscillator, electric dipole radiation is emitted when n decreases by 1 and
absorbed only when n increases by 1. Thus, the harmonic oscillator obeys the selection
rule 1 n A = , and the only frequency emitted or absorbed is the classical oscillator
frequency 2
s
k
f
m
t = since the energy levels are equally spacedwhere
m mass of particle
k
s
spring constant for system
44

parts for a compound state given by Equation (22) for a truncated
parabolic well. The probability density changes with time as the charge
distribution oscillates back and forward in the potential well. Red curves
real part + , blue curves imaginary part + and black curves probability
density.

An animated gif [se_super.m] of the time evolution of the stationary state
shown in Figure () can be viewed at

www.physics.usyd.edu.au/teach_res/mp/quantum/gif/ag_super13gif

real +
imaginary +
prob. density
t = 0
t = T/4
t = T/2 t = 3T/4
t = T
45

INVESTIGATIONS (Matrix Method)

There are many aspects of quantum behavior for bound states you can investigate by
studying various potential wells with different parameters. A proven learning strategy
that you can apply to gain a deeper understanding of quantum physics is called POE
Predict Observe Explain
For a given potential well, some of the predictions you could make are: number of bound
states, approximate values of energy eigenvalues, spacing of energy levels, shape of
wavefunction, number of half-wavelengths inside well, changes in half-wavelengths and
size of exponential tails, the number of zero crossings of the wavefunction, and height of
peaks in wavefunction. For example, in a square well the energy levels are more crowded
towards the bottom of the well, in a parabolic well (shape bends away from x-axis) the
energy levels are equally spaced and in the Morse potential well or a Coulomb shaped
well (shape bends towards the x-axis) the energy levels are more crowed toward the top
of the potential well.

After you have made your predictions, run the m-scripts and observe the results and
compare them with your predictions.

Then, you explain the results you observed and rectify any discrepancies in the results
and your predictions.

To solve the Schrodinger equation using the Matrix Method and observe the solutions run
the following m-scripts:
se_wells.m
Need to change the m-script to set all the well parameters and select type of well
se_solve.m
Solves the Schrodinger equation using the matrix Method.
se_orthonormal.m
Investigate the orthonormal characteristic of wavefunctions.
se_psi.m
Displays energy functions and wavefunctions.
se_measurements.m
Calculates expectation values.
46
se_stationary.m
animation of stationary states
se_super.m
animation of compound states

1 Infinite Square Well and Finite Square Well

1.1 Find the energy eigenvalues for various widths of an infinite square well and
compare your results with theoretical values predicted by Equation (13). For each
state (n = 1, 2, . ) measure the wavelength , then using the de Broglie
relationship ( / p h = ), calculate the momentum of the electron and its kinetic
energy (
2
/ 2
K
E p m = ). Compare the energy eigenvalues with the kinetic energy
values. Repeat the calculations for different values of num (size of vectors and
arrays). Can use the m-script se_infwell.m.
1.2 Find the energy eigenvalues for various square wells of different depths and
widths. Observe the wavefunction and expectation values for different states.
Check the orthonormal character of the wavefunctions [se_othonormal]. Why
does the wavelength of the eigenfunctions decrease as the quantum number n
increases? Why is the kinetic energy of the electron in each eigenstate of a finite
square well less than the value of its kinetic energy in an infinite square well of
the same width? How does the number of zero crossing of the wavefunction vary
with the quantum number n? Comment on the curvature of the wavefunction
inside and outside of the potential well and explain. Comment on the expectation
values and is the Heisenberg Uncertainty Principle satisfied?
1.3 The simplest nuclear system is a deuteron - a single proton bound together with a
single neutron by the attractive strong nuclear force. This system has only one
bound state with a binding energy of 2.22 MeV near the top of the well. This is
the ground state, there are no excited states. This is a three dimensional system,
but we can model the strong nuclear force as a one-dimensional finite square well
potential where the solution of the Schrodinger equation gives the radial
component of the wavefunction. Experiments on the scattering of high energy
electrons from deuterons provides evidence of the existence of a strong repulsive
47
core in the nuclear potential, this means that the particles avoid the centre of the
deuteron, the proton and neutron cant get too close to each other, hence, the
radial wavefunction must approach zero towards the centre. The mass of the
bound particle is equal to the reduced mass of the system, hence
p n
p n
m m
m
m m
=
+
m = (m
p
+ m
n
) / 2
We can alter the code for the square well potential (case 1) to model the potential
for the deuteron. The modifications are:
m = mp*mn/(mp+mn); reduced mass - deuteron
xMin = 0 forces the wavefunction at the centre of the deuteron to be zero
x1 = 1.5e-6 width of square well in nm
xMax = 20*x1
U1 = -100e6 starting depth of square well in eV

Find the binding energy using se_solve.m. Then, adjust the value of U1 until the
binding energy is about 2.2 MeV to find an estimate of the well depth. View the
characteristics of the wavefunction using se_psi.m. In the ground state, the
eigenfunction describing the separation of the proton and neutron extends far
beyond the confines of the well. From the Command Window determine the
probability that the separation distance between the proton and neutron is greater
than the width of the well using the function simpson1d.m:
Check that the function is normalized, use
simpson1d(prob(:,1)',xMin,xMax) result should be 1.
The probability that the electron is found within the well, use
simpson1d(prob(1:min(find(x>=x1)),1)',0,x1)
to evaluate the integral
1 1
0 0
( ) ( ) ( )
x x
x x dx prob x dx =
} }
. The statement
min(find(x>=x1)) finds the index of x for the edge of the well at x
1
.

The probability density
2
( ) x specifies the probability of finding the two
nucleons in the deuteron with a separation in the vicinity of x.

Use se_measurements.m to find the average separation distance.
48

One book quoted that the mean separation of the proton and neutron as
410
-15
m (much larger than the width of the well) and that there was a 50%
chance of finding the particles in the classical forbidden region. How do your
results compare? Your results about probability imply that the attractive region
for the only bound state is just barely strong enough to overcome the effect of the
repulsive core and lead to binding. As a consequence, there is a high probability
that the two nucleons in the deuteron have a separation larger than the range of
nucleon forces. The values for xMax, x1 and U1 are not unique in giving the correct
value for the binding energy. If you check different references, you will get
different numbers quoted in each reference. Remember, that our model is a crude
one, and the numbers are not necessary accurate but the model does give an
insight to the deuteron nucleus.

2 Stepped Well (Asymmetrical Well)
2.1 Investigate the solutions of the Stepped Well for various widths and depths of the
left and right sides of the well by modifying the code in se_wells.m
case 2
xMin = -0.15; % default = -0.15 nm
xMax = +0.15; % default = +0.15 nm
x1 = 0.1; % Total width of well: default = 0.1 nm
x2 = 0.06; % Width of LHS well: default = 0.06 nm;
U1 = -400; % Depth of LHS well; default = -400 eV;
U2 = -200; % Depth of RHS well (eV); default = -250 eV;

Set U1 = -400, U2 = -200, x1 = 0.10 and x2 = 0, 0.25x1, 0.50x1, 0.75x1, 1.00x1
Set x1 = -0.10, x2 = -0.05, U1 = -400 and U2 = 0, 0.25U1, 0.50U1, 0.75U1, 1.00U1
2.2 Investigate an asymmetrical well by commenting the line as shown below in the
m-script for se_wells.m (case 2).
for cn = 1 : num
if x(cn) >= -x1/2, U(cn) = U1; end; %if
if x(cn) >= -x1/2 + x2, U(cn) = U2; end; %if
%if x(cn) > x1/2, U(cn) = 0; end; %if
% comment above line to give an asymmetrical well
end %for

49

Fig. 19. The energy spectrum for a stepped well.
[se_wells.m se_solve.m]

Fig. 20. The quantum state n = 3 for a stepped well.
[se_wells.m se_solve.m se_psi.m]

-0.15 -0.1 -0.05 0 0.05 0.1 0.15
-450
-400
-350
-300
-250
-200
-150
-100
-50
0
50
position x (nm)
e
n
e
r
g
y

U
,

E
n

(
e
V
)
Potential Well: STEPPED
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
-10
0
10
Potential Well: STEPPED n = 3 E
n
= -115.2244
position x (nm)
w
a
v
e

f
u
n
c
t
i
o
n

p
s
i
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
-400
-200
0
U

(
e
V
)
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
0
20
40
position x (nm)
p
r
o
b

d
e
n
s
i
t
y

p
s
i
?
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
-400
-200
0
U

(
e
V
)
-0.1 -0.05 0 0.05 0.1 0.15
-400
-200
0
200
E
n
position x (nm)
e
n
e
r
g
i
e
s

(
e
V
)
50

Fig. 21. The quantum state n = 4 for an asymmetrical well. The total
energy is greater than the potential on the left hand side of the well,
therefore, the kinetic energy of the electron is positive with the electron no
longer bound to the well and acts as a free particle.

3 Double Well
3.1 Investigate the properties of double potential wells. Remember to employ the
POE learning strategy. Start with the well parameters:
U1 = -400; % Depth of LHS well: default = -440 eV
U2 = -400; % Depth of RHS well: default = -400 eV
U3 = -300; % Depth of separation section: default = 100 eV
x1 = 0.10; % width of LHS well: default = 0.10 nm
x2 = 0.10; % Width of RHS well: default = 0.10 nm
x3 = 0.10; % Width of separtion section: default = 0.10 nm
xEnd = 0.05; % parameters to define range for x-axis: default = 0.05 nm

Then, increase the height of the centre section of the well U3 in steps of 50 to 200.

-0.15 -0.1 -0.05 0 0.05 0.1 0.15
-4
-2
0
2
4
Potential Well: STEPPED n = 4 E
n
= -161.9123
position x (nm)
w
a
v
e

f
u
n
c
t
i
o
n

p
s
i
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
-400
-300
-200
-100
0
U

(
e
V
)
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
0
10
20
position x (nm)
p
r
o
b

d
e
n
s
i
t
y

p
s
i
?
-0.15 -0.1 -0.05 0 0.05 0.1 0.15
-400
-200
0
U

(
e
V
)
-0.1 -0.05 0 0.05 0.1 0.15
-400
-200
0
200
400
E
n
position x (nm)
e
n
e
r
g
i
e
s

(
e
V
)
51
3.2 The results for the double well with U3 = 200 eV are:
1 -373.13
2 -373.13
3 -293.56
4 -293.56
5 -165.41
6 -165.4
7 -2.8576
8 -2.7866

Fig. 22. The energy spectrum for the double well. Why do you only see 4
energy levels and not 8? [se_wells.m se_solve.m]

-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
-400
-300
-200
-100
0
100
200
position x (nm)
e
n
e
r
g
y

U
,

E
n

(
e
V
)
Potential Well: DOUBLE
52

Fig. 23. The energy eigenvalues for the double well. [se_wells.m
se_solve.m]

Explain the spacing of energy levels.

3.3 Investigate a well with a deep centre section, try U3 = -600.

4 Sloping Well
4.1 Investigate the properties of sloping potential wells. Remember to employ the
POE learning strategy.

5 Truncated Parabolic Well
5.1 Investigate the properties of truncated parabolic potential wells. Remember to
employ the POE learning strategy. What can you say about the spacing between
energy levels?
5.2 The potential energy for a classical particle oscillating in a parabolic well (e.g.
particle vibrating at the end of a spring executing simple harmonic motion) is

2
1
2
( )
s
U x k x = where k
s
is the spring constant.
The eigenstates for this system are the solutions of the Schrodinger equation:
n = 1
2
n = 1
n = 2
2
n = 2
n = 3
2
n = 3
n = 4
2
n = 4
n = 5
2
n = 5
53

2 2
2
1
2 2
( ) ( )
2
s n sn n
d
k x x E x
m dx

| |
+ =
|
\ .

The energy eigenvalues for this system known as the quantum harmonic
oscillator are given by
( )
1
2 sn
E n e =
where n = 1, 2, 3, (N.B. most books have n = 0, 1, 2, 3, )
The classical frequency of oscillation e for a particle of mass m is
2
s s
k k
f
m m
e t = =
The Schrodinger equation for our truncated parabolic well is

2 2 2
1
1 2 2
1
1 4 ( ) ( )
2
n sn n
U d
U x E x
m dx x

| | | | | |
+ = | | |
| |
\ . \ . \ .

where U
1
is the depth of the well and x
1
is the full width of the well when U = 0.
Show that the spring constant k
s
is related to the well parameters U
1
and x
1
by the
relationship

1
2
1
8
s
U
k
x
=
Run the m-script se_wells.m for the parabolic well (case 5) with the parameters:
xMin = -0.2; % default = -0. nm
xMax = +0.2; % default = +0.2 nm
x1 = 0.2; % width default = 0.2 nm;
U1 = -400; % well depth default = -400 eV;

and verify the following results.
n Matrix
Method E
n
(eV)
Theory
E
sn
+ U
1

E
n+1
- E
n

1 -360.92 -360.94 ---
2 -282.77 -282.83 78.15
3 -204.68 -204.71 78.08
4 -127.09 -126.60 77.60
5 -52.272 -48.485 74.81
Explain why the discrepancies are larger for higher the quantum numbers.
54

Fig. 24. The eigenfunctions and probability distributions for the harmonic
oscillator.

Explain why the state n = 5 in Figure (24) is more like the classical prediction for
the probability of locating the particle.

5.2 Repeat the above, for different well parameters. Run se_measurements.m and
find the expectation values for a number of states. What is the significance for the
state n =1 by considering the value for x p A A ?

The quantum harmonic oscillator system is very important one in explaining the
vibrational states of molecules. You can have a more in-depth analysis of the quantum
harmonic oscillator and the vibrations of diatomic molecules from the link
www.physics.usyd.edu.au/teach_res/mp/quantum/doc/vibrations.pdf

n = 1
2
n = 1
n = 2
2
n = 2
n = 3
2
n = 3
n = 4
2
n = 4
n = 5
2
n = 5
55
6 Morse Potential Well
6.1 Investigate the properties of Morse parabolic potential wells. Remember to
employ the POE learning strategy. What can you say about the spacing between
energy levels?
6.2 The Morse potential is often used as the potential energy function for diatomic
molecules. One form of the Morse potential is

2
1
1
( ) 1 exp 1
x x
U x U
S
| |
| | | |
= |
| |
|
\ . \ .
\ .

where S is a positive constant which determines the width of the well and the
maximum depth of the well is at x
1
, U(x
1
) = U
1
. Consider the Morse potential with
parameters:
U1 = -1200;
x0 = 0.08;
S = 0.075;
xMin = 0;
xMax = 0.3
xMax = 0.3;

Fig. 25. The energy eigenvalues for the Morse potential.

0 0.05 0.1 0.15 0.2 0.25 0.3
-1000
-500
0
500
1000
1500
2000
2500
3000
position x (nm)
e
n
e
r
g
y

U
,

E
n

(
e
V
)
Potential Well: MORSE potential
56
Run the m-script se_measurements.m and record the expectation value
for position. Comment on the spacing of the energy levels and explain why
most materials expand upon heating.

The Morse potential system is very important one in explaining the vibrational
states of diatomic molecules. You can have a more in-depth analysis of the
quantum harmonic oscillator and the vibrations of diatomic molecules from the link
www.physics.usyd.edu.au/teach_res/mp/quantum/doc/vibrations.pdf

7 Stationary States
7.1 Run se_wells.m, se_solve.m and then se_stationary.m and explore the
characteristics of the time evolution of stationary states. Determine the frequency
and period of the oscillations.

8 Stationary States
8.1 Run se_wells.m, se_solve.m and then se_super.m and explore the characteristics
of compound wavefunctions.

8.2 Select the truncated harmonic potential well and verify that radiation is emitted
only for transitions where 1 n A = .

More solutions of the time independent Schrodinger equation for bound systems are
available from the following links:

Hydrogen like atoms
www.physics.usyd.edu.au/teach_res/mp/quantum/doc/hydrogen.pdf
Solid state systems: One dimensional lattices
www.physics.usyd.edu.au/teach_res/mp/quantum/doc/lattices.pdf
Alpha decay
www.physics.usyd.edu.au/teach_res/mp/quantum/doc/alpha_decay.pdf
Links to solutions of the time dependent Schrodinger equation:
www.physics.usyd.edu.au/teach_res/mp/quantum/doc/se_time.pdf

Schrodinger

Uploaded by

Copyright:

Available Formats

Schrodinger

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Schrodinger

Uploaded by

Copyright:

Available Formats

Doing Physics with Matlab Quantum Mechanics Bound States

A is the quantum-mechanical operator corresponding to A. The operator is

= . Hence, the frequency of the emitted photon is the same as the

You might also like