Computer Simulations of Schroedinger Equation using Maple

Revised 4/14/09
The purpose of this lab project is to: 1) solve Schroedinger equation numerically. 2) learn how to solve differential equations using Maple. Introduction 1. In this lab we will solve Schroedinger equation h2 d 2 ( x) + U ( x) ( x) = E ( x) . 2m dx 2

It will help to change units so that we can apply the results to real world examples. We can rewrite Schroedinger equation units where length is measured in nm and energy is measured in eV by the following c d2 U E h2e2 ( x ) + u ( x ) ( x ) ( x ) = 0 , where , , u = = c = , and a = nm . dx 2 e e 2ma 2

First it makes Schrdingers equation simpler and more like a mathematical equation. But, it also makes the units appropriates for nano size objects. To start the solution we need the initial conditions or boundary conditions. There are two way to do this. The first used the boundary conditions at the origin. For a symmetric potential the solutions are always symmetric or anti-symmetric about the origin. We can start the solution on the axis of symmetry or for us x = 0. On the axis of symmetry the symmetric solutions all have a slope of zero or d(0)/dx = 0. We can choose any value for (0) because the wavefunctions are not normalized. So choose (0) = 1. For the anti-symmetric solution all the solutions are zero or (0) = 0 on the axis of symmetry. We can choose any value for the slope again because the wavefunctions are not normalized. Choose d(0)/dx = 1. The second way uses the boundary conditions at infinity. At infinite the wavefunction, d ( x ) ( ) = 0 and the derivative of the wavefunction, = 0 has to approach zero. Of dx x = course, if we start the integration of Schrdingers equation with these boundary conditions we will get a null wavefunction. Thus, we want to start with a small value for the wavefunction and its derivative (something like 0.00005). Well how far is infinity? All we need to do is be five or six times the distance to the turning point. That will work. The real test is that if you try a distance and move a little further away the wavefunction should not change.

Now we make a guess as to the energy and use the initial conditions on the wavefunction to find a mathematical solution. Thus, there is a solution for each energy but for a physical solution the wavefunction must be finite. We see that all solutions diverge unless the energy is correct. It is easy to find the correct energy because as you cross the correct energy the solution diverges in

the opposite direction (ie if it was diverging negatively then it will change to diverge positively). 2. In this lab we will use Maple to solve Schrdingers equation. You can use Maple to solve differential equations without knowing how it works but if you would like to know read chapter 8 section 1 on "Ordinary Differential Equations" in Numerical Mathematics and Computing by Ward Cheney and David Kincade. You can find the pdf on the course web page.
Infinite Square Well

3. I have provided a simple Maple program to get you started. The program called Finite_Square_Well amd will solve Schrdingers equation for the finite square well with different widths and potential energy walls. We can also use the program for the infinite square well. I have entered a large potential energy so that the solutions closely approximate an infinite square well. Now lets use it to find the ground state energy. Calculate the ground state energy for a well 20 nm wide. Do this in your notebook. Now run the program and see that it works. I have set the energy so that it is close to the eigenvalue. Also I have set the potential energy walls to 200 eV so that it approximates an infinite square well. This is not infinity but close enough to give us the same answer. Using numeric solution it is too slow if the potential energy gets too large. Next vary the energy a little above and below to see the behavior of the wavefunction. What do you observe? Scroll down through the program and make sure you understand all the commands. 4. First lets find the first six energy levels and eigenfunctions for the infinite well. We dont want to just guess so calculate the energy for the first six eigenvalues using the results found in class for the infinite square well. Do they agree? Are the wavefunctions the same? 5. Use your Maple program 6. Now we will change the well potential energy height to explore the finite square well. Keep the same well width of 20 nm and change the potential energy to U0 = 3eV. We can make finite wells in semiconductor wires at Harvard this size. If you are interested ask me. 7. Now find the first six energy levels. You should find the energy levels just below the values for the infinite square well. Start the search at the value for the infinite square well and lower the energy until you find the eigenvalue for the energy. 8. Compare the infinite to the finite square well and comment on and differences in your notebook. Compare the wavefunctions you found to the expressions from class. Check to see that the wavefunctions are normalized by comparing the normalization constant found in class to the one you found using Maple. Do this in your notebook. 9. Change this section. Next we should change the program so that it will find the energies for us. You should have been using the bisection algorithm. This is the algorithm we will use. Open and resave the Schroedinger_General program and rename it Finite_Square_Well_with_Search. Enter the potential energy function for the finite square well where the program has U:= potential energy function. Find the eigenvalues (energy) and eigenfunctions (wavefunction) for the first six states. Draw an energy level diagram for the finite and infinite square well next to each other. Comment on any differences in your notebook. -2-

10. As the depth of the well is increased you would expect the energy levels to approach the value for the infinite square well. Look at the quantized energy for the infinite square well and make sure you understand why the energy changes. We will come back to this question later. Find the energy for the first (n = 1) eigenvalue for different well depths (up to 500 eV). Plot the energy for the first eigenvalue as a function of well depth using about 10 different values. Try to see what value the first eigenvalue extrapolates too for an infinite well depth. Compare this to the value we found for the infinite square well in class. 11. In this part of the program we can find the expectation value for different operators. For example we can find the momentum expectation value <p> or the potential energy expectation value. This is the average value you would get if you measured the momentum or potential energy of the electron a number of times. First we will find the expectation value for the electrons position <x> in the ground state. This is
x
* = n ( x ) x n ( x )dx

Can you think of a simple classical argument to find the value for the position expectation? Write your answer in your lab book. 12. Change this section if 9 changes. Reopen your program Finite_Square_Well. At the end of the program after you find the energy and wavefunction and add a command to calculate the position expectation. Use the program to find the position expectation for the ground state. Does it agree with your argument? Find the expectation value for the position for two or three other eigenvalues. Is this what you expected? 13. Lets find the momentum expectation for the finite square well. First can you use classical arguments to find the average momentum? Write your argument in your lab book. Add a command to calculate the momentum expectation after the position expectation. will recall that the momentum operator is You

[ p] =

ih 2 d 2me dx

h2 d * = n ( x ) n ( x )dx 2 m dx e

Again we will set h = 1 to keep the momentum value from being extremely small. Thus to find the actual momentum you need to multiple you value by h . Find the momentum expectation for the ground state. Do you get the classical results? What about higher energy states? Does the classical argument work? Check it out for a couple of other states. 14. Lets find the expectation for the energy and see if it is the same as the eigenvalue energy. What do you think? Use the momentum expectation and position expectation to find the kinetic energy expectation, <K>, potential energy expectation, <P>, and total energy expectation, <E>.

-3-

Does the expectation for the total energy agree with the energy value you found earlier? 15. Using the wave functions for the infinite square well calculate the kinetic energy expectation. How does it compare to the kinetic energy expectation for the finite square well? Is this difference about the same as the total energy difference between the finite and infinite square well? 16. Look at the probability of finding the electron outside the well for all eigenstates. Which state has the largest probability of being found outside the well? Does this agree with the energy values you found in section 5? You should be able to explain the difference in energy for the finite and infinite square well using the probability plots. Recall <E> = <K> + <V> The kinetic energy is related to the curvature of the wave function. Does it change? Can you estimate the change in curvature? From the probability calculate the change in potential energy. This difference should be the difference between the energy for the infinite and finite square well. The region outside the well is called the classically forbidden region. Why is this region referred to in this way? 17. you find the potential energy expectation for the region outside the well? 18. Finally, lets calculate x and p and see if x p is equal to h/2. We will use the root-meansquared for x and p. The root-mean-squared is defined for a continuous function as

x =

* n ( x x ) n dx 2

19. where (x) is a normalized probability density and <x> is the average value. From this show that

x =

x2 x

20. Use this expression for the root-mean-squared to find the uncertainty in x for the ground state. Use a similar expression to find the uncertainty in the momentum in the ground state. What is x p? This is the explanation for the zero point energy or why the ground state energy is not and cannot be zero.

-4-

Quantum Ball for a Linear Restoring Potential Energy

21. Lets think about a new example. Consider a constant restoring force toward the origin. Can you use the definition for the potential energy in terms of the work? For a constant restoring force you would find a linear potential energy function as shown.
V(x)

k|x|

22.

Quantum Ball in a Gravitational Field

23. Lets think about a new example. Consider a constant force toward the origin. The force would be F = kx/|x|. The term x/|x| causes the force to point toward the origin. The potential energy is V(x) = k|x|. This is shown below. From your understanding of Schroedinger equation draw the first four wavefunctions. Try to guess how the energy levels will compare with the harmonic oscillator. Higher, lower, why?.

7. Use Maple to solve Schrdingers equation for a constant force or u(x) = k|x| for x > 0. Find the first four wavefunctions and energy levels. Compare them with the harmonic oscillator. Explain why the energies are higher or lower than the harmonic oscillator. For the wavefunctions think about what ahalf quantum harmonic might look like. -5-

Finite Square Well with an Electric Field

12. Lets think about a new example. Consider an electric field applied to the electron in the box. The force would be F = -eE and the potential energy is U(x) = eEx. This is shown below. From your understanding of Schroedinger equation draw the first four wave functions. Try to guess what the energy will be for the first four levels compared to the finite square well.. Use your Maple program to solve Schrdingers equation for a constant force F = -eE or V(x) = eEx. Find the first four wave functions and energy levels. Compare them with the finite square well.
V(x)

Two Finite Square Well: A Diatomic Molecule

9. The potential energy for a double well is shown in the figure on the next page. Try to guess the first six wavefunctions for the double well. Draw them in your notebook. Now try to guess the energy for the first six levels. Think about the kinetic and potential energy changes. 10. Use your program with graphics to plot the wave function and find the energy levels for a square well with a bump in it. Again we will look at this problem in class. We will use perturbation theory to solve this in class. How do the energy levels and wave functions compare to the same well with out the bump?

Quantum Harmonic Oscillator

3. First we will use Maple on a problem we can solve analytically (we know the solutions and -6-

energies from class), the harmonic oscillator. The potential energy for the harmonic oscillator is U(x) = (1/2)kx2 = (1/2)m2 x2. In our units u(x) = (1/2)m2 x2/C. Dont forget You can use the Maple commands diff and dsolve or the command that does it all DEplot. Dont forget to add the plotting functions (with (plot): and with (DEtools) : Find the energy levels (just the first six) for a harmonic oscillator. How do the energy levels compare with the analytical values, En = _(n+1/2) Show both on an energy level diagram. 4. How do the wave functions compare? 5. Lets compare the analytic solution to the approximation (computer solution) found with Maple. Plot the wavefunction for n = 3 and overlay the analytical solution (y) = Aexp(y2/2)H3(y), where y=x??. How do they compare.
Radial Hydrogen Equation

radial dimension for Hydrogen

-7-