MORE

PHYSICS WITH MATLAB

MORE PHYSICS WITH MATLAB
Dan Green
Fermi National Accelerator Laboratory, USA
Published by
World Scientific Publishing Co. Pte. Ltd.
5 Toh Tuck Link, Singapore 596224
USA office: 27 Warren Street, Suite 401-402, Hackensack, NJ 07601
UK office: 57 Shelton Street, Covent Garden, London WC2H 9HE
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library.
MORE PHYSICS WITH MATLAB
(with Companion Media Pack)
Copyright © 2015 by World Scientific Publishing Co. Pte. Ltd.
All rights reserved. This book, or parts thereof, may not be reproduced in any form or by any means,
electronic or mechanical, including photocopying, recording or any information storage and retrieval
system now known or to be invented, without written permission from the publisher.
For photocopying of material in this volume, please pay a copying fee through the Copyright Clearance
Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA. In this case permission to photocopy is
not required from the publisher.
ISBN 978-981-4623-93-3
ISBN 978-981-4623-94-0 (pbk)
Printed in Singapore
“It should be possible to explain the laws of physics to a barmaid.”
— Albert Einstein
“Not only is the Universe stranger than we think, it is stranger than
we can think.”
— Werner Heisenberg
 Preface

“Computers are like Old Testament gods; lots of rules and no mercy.”
— Joseph Campbell
“Part of the inhumanity of the computer is that, once it is competently programmed and
working smoothly, it is completely honest.”
— Isaac Asimov

There are only a very few analytically solvable problems in physics. They are
extremely useful because the equations for the solutions can be plotted and
the parameters defining the solutions can be varied in order to explore the
dependence of the solutions on the variables of the problem, although that
exercise may become tedious. In that way the student can build up an
intuition about the Kepler problem, for example.
However, this can only be done in a few cases and even then the effort
needed is large. For the others, numerical methods are needed and the
computation becomes somewhat cumbersome. As a result, it is more difficult
to vary the parameters of the problem numerically rather than symbolically
and develop an intuition about the dependence of the solution on those
parameters. In particular the time development of the system is often obscure
and “movies” can be a welcome tool in improving physical intuition.
Because of these issues, the advent of powerful personal computing has
considerably reduced the difficulties. Indeed, the aim of this book is to use the
ensemble of symbolic and numeric tools available in the MATLAB suite of
programs to illustrate representative numerical solutions to more than one
hundred problems spanning several physics topics. The student typically
works through the demonstration and alters the inputs through a menu driven
script. In that way the user driven menu allows for parametric variation. The
more advanced student can edit the script or, indeed, write a new script
covering a wholly new problem. It is crucial that the student read the script
which is supplied with the text. There are many “comment” lines in the script
which more fully explain what the procedures are in each specific case.
MATLAB is a good vehicle for the computational tasks. It has a compiler,
editor and debugger which are very useful and user friendly. The HELP utility
is very extensive. The MATLAB language is similar to a modern C++
language and it is a vector/matrix language which makes coding simpler than
older languages such as FORTRAN.
MATLAB contains many special functions. Matrices and linear algebra are
covered well. Curve fitting, polynomials and fast Fourier transforms are
supplied. Numerical integration packages are available. Differential
equations, symbolic, ordinary and partial, as well as numerical solutions are
available for both initial value and boundary value versions.
As an additional package, MATLAB has symbolic mathematics. Within
that package, calculus, linear algebra, algebraic equations and differential
equations are covered. It is easy to combine a symbolic treatment of a
problem with a numerical display of the solution when that is desirable. In
this way converting from symbols to numbers is easily achieved.
Finally, and very importantly, MATLAB has an extensive suite of display
packages. One can make bar, pie, histogram and simple data plots. There are
several contour and surface plots which are possible. There are two- and
three-dimensional plots of all types available. The time evolution of solutions
can be made into “movies” that illustrate the changes in speed during an
evolving process. These extensive visualization tools are crucial in that the
student can plot and then vary and re-plot.
There are many topics which are explored in the text. But, it is not what
topic is specifically explored that is important but how it is examined. A
useful methodology that one can apply to any specific problem is, in outline:
1. First look at the dimensions of an equation to see if they are correct.
For example, the Schrödinger equation has units of energy
[−h2 (∂2/∂x2 )/2m + V = ih∂/∂t]ψ.
Since the Planck constant has dimensions of position times momentum or
energy times time, it is clear that all the terms in the equation have the
dimensions of energy.
2. Then see what are the characteristic lengths and times implied by the
equation.
For the previous equation there is characteristic length times energy,
For electrons, that length is 8 eV * A. The second term
has dimensions eV. The third term has a time scale of 0.66 eV × 10−15
sec. Therefore for problems with scales of V of a few eV, it is expected
that distances for electrons will be a few Angstroms and times will be a
few 0.001 psec.
3. See if there are physically reasonable limits implicit in the equation.
 For example, if the section on “Rocket Drag” the equation of motion
is:

In the limit where k and g are zero, the simple rocket equation is
recovered as a known limit. This can be checked by choosing the
parameters of the problem appropriately.
4. Try to solve the equation symbolically, if not use numerical techniques.
5. For either type of solution look for limits which simplify the results and
are physically reasonable.
For example, the analytic solution of the Kepler problem for an orbit
with eccentricity e is:
r = rc(1 + e)/(1 + e cos θ)
Clearly, a circular orbit has a constant radius, so that such an orbit must
have e of zero. At the other limit, if the radius is to diverge, then the
cosine of the angle must be 1/e.
Therefore unbound orbits must have e > 1 and the cosine of the angle
of the orbit at large r, the asymptote, is 1/e.
6. Finally, vary parameters of the problem in order to establish an
“intuition” about the dependencies of the result on the parameters.
For example, the “Spring Pendulum” is governed by the equations:
d2r/dt2 = g cos θ + (k/m)(Lo − r) + r(dθ/dt)2
d2θ/dt2 = −g sin θ/r − 2(dr/dt)(dθ/dt)/r
These are solved numerically. Possible checks are to set g to zero when a
simple spring should result. Another limit would be to set k to zero, when
a simple pendulum should result. Both limits can be checked by choosing
the parameters, including the initial condition, of the problem.
The aim of using these tools is to create intuition, not to solve a specific
problem or to complete a specific number crunching exercise. Indeed, the aim
of the text is not to teach the basic physics but to give the user a sense of how
the solutions of a given physics problem depend on the parameters of that
problem. If the reader has questions, a dip into Wikipedia and/or Google
should be very helpful. Details are also available in the “comments” lines
within the specific script.
The script for these demonstrations is made available on storage media and
examples for classical mechanics appear in the Appendix. The student is
strongly advised to read the script before executing the script. In that way the
script is not a “black box” but is part of the learning process. Using that
material the student can write her own additions and explorations with the
scripts supplied used as jumping off points. In this way, a path is made
available to extend well beyond the specific demonstrations provided in the
book itself, making the search for further possible insights open ended.
 Contents

Preface
1. Mathematics
1.1 Chaos
1.2 Malthus
1.3 Missile Tracking
1.4 Monte Carlo Analytic
1.5 Monte Carlo Numeric
1.6 Monte Carlo Muon Beam
1.7 Fourier and Laplace Transforms
2. Classical Mechanics
2.1 Angular Momentum
2.2 Classical Scattering
2.3 Impact Parameter and DCA
2.4 Foucault Pendulum
2.5 Spring Pendulum
2.6 Spherical Pendulum
2.7 Driven Pendulum
2.8 Random Walk
2.9 Rotating Hoop
2.10 Tides
2.11 Action
2.12 Rocket Drag
2.13 Two Stage Rocket
 2.14 Table Top
2.15 Kepler Eccentric
2.16 Non-Central Force
2.17 Euler Angles
2.18 Forceless Top
2.19 Top Motion
3. Electromagnetism
3.1 Electron e/m
3.2 Potential and Fields
3.3 Electrostatics — Plates
3.4 Electrostatics — Angled Plates
3.5 Electrostatic Helmholtz
3.6 Electric Quadrupoles
3.7 Cylinder Boundaries
3.8 Laplace Fourier
3.9 Rotating Charged Sphere
3.10 Dielectric Image Charge
3.11 Dielectrics
3.12 Dielectric Sphere
3.13 Induction
3.14 Magnetic Bottle
3.15 A Fusion Reactor
3.16 Dipole Radiation
3.17 Two Antennas
3.18 Dispersion
3.19 Waveguide
3.20 Skin Depth
4. Gases and Fluids
4.1 Leaky Box
 4.2 Rectangular Flow
4.3 Debye Temperature
4.4 Times Arrow
4.5 Compressibility
4.6 Viscosity
4.7 Water Waves
4.8 Semiconductor
4.9 Semiconductor Junction
4.10 n-p Diode
4.11 Freezing Pipes
4.12 Cooling Earth − Interior
4.13 Cooling Earth − Exterior
4.14 Heat and pde
4.15 Heat Diffusion
4.16 Heat — Initial
5. Waves
5.1 Longitudinal Slinky
5.2 Transverse Slinky
5.3 SR — Doppler
5.4 Step Response
5.5 Pulsating Sphere
5.6 Rectangular Drum
5.7 Piano
5.8 Aperture Diffraction Off Axis
5.9 Antenna
5.10 Antenna Array
5.11 Lissajous
5.12 Plane Waves — 2d
5.13 Damped, Driven Oscillator
6. Quantum Mechanics
6.1 Box
6.2 Simple Harmonic Oscillator
6.3 Bound States — 3d Well
6.4 Identical Particles
6.5 Stark Effect
6.6 Square Well Scattering — 3d
6.7 Photoelectric Effect — Continuum
6.8 Line Width
6.9 Barriers
6.10 General Eigenvalues
6.11 Koenig–Penny
6.12 Decay Chain
6.13 Casimir Effect
7. Astrophysics
7.1 Transfer Orbit
7.2 Flyby
7.3 Lagrange Points
7.4 Binary Orbits
7.5 Three Body Orbits
7.6 Polytropic Star
7.7 Pulsating Stars
7.8 White Dwarf
7.9 Boltzmann and the Sun
8. General Relativity
8.1 Light Deflection
8.2 Circular Geodesic
8.3 General Geodesic
8.4 Kerr Photons
 8.5 Kerr General
8.6 Gravity Wave Radiation
8.7 Stellar Pressure
8.8 Light Propagation in the Universe
Appendix: Scripts for the Chapter on Classical Mechanics
Graphical User Interface (GUI)
References
Index
 Chapter 1

Mathematics

“The scientist does not study nature because it is useful; he studies it because he delights in it,
and he delights in it because it is beautiful. If nature were not beautiful, it would not be worth
knowing, and if nature were not worth knowing, life would not be worth living.”
— Jules-Henri Poincaré
“The chief forms of beauty are order and symmetry and definiteness, which the mathematical
sciences demonstrate in a special degree.”
— Aristotle

1.1. Chaos
The world is a complex place and the relationship for complex systems
between cause and effect is difficult to untangle. A famous statement about a
fluttering butterfly wing “causing” a weather storm is an example of this
perception. Complex systems have “chaotic” solutions. Predictions for
macroscopic systems with 1023 particles or so by tracking each particle are
clearly impossible. Even simple systems with few particles are not simply
predictable if the initial conditions are changed very slightly.
A very simple system, consisting of a particle starting from a position and
angle and bouncing around a totally rigid circular container is explored in the
script “Chaos_Butterfly”. Indeed, small changes in the initial conditions lead
to large changes in the tracked trajectory after only twenty bounces. Plots for
an initial position of y equal to zero and x equal to −0.5 (circular container of
radius one) and initial angles of 20 and 21 degrees are shown in Figures 1.1
and 1.2 respectively. The patterns are quite different with a small change in
the initial conditions. The user can “play” with different initial conditions and
watch the subsequent behavior of the system in order to see how little the
starting conditions need to be changed to make a large change in the
trajectory after only a few bounces off the container walls.
Figure 1.1: Trajectory for 20 bounces with an initial angle of 20 degrees. The starting point is the red
circle, while the center of the circle is the black star.

Figure 1.2: Trajectory for 20 bounces with an initial angle of 21 degrees. The starting point is the red
circle, while the center of the circle is the black star.

1.2. Malthus
Malthus long ago predicted that, with a fixed maximum food supply possible
and with a growing population, famine would be unavoidable. However, so
far, using fertilizers and crop improvements, this fate has been avoided by
increasing the food supply. In the long run that approach has limitations
which need to be considered and the predictions of Malthus will not be
evaded forever.
The simplified equations explored in the script “Malthus” are:

The function N is the population as a function of time. If only the first

equation in Equation (1.1) were active the number would grow exponentially
with a characteristic time of 1/a.
The “Malthus” equation assumes the number grows without limit, with
exponential growth of the population which is a consequence of the first of
Equation (1.1). A slightly more realistic model assumes the growth rate is
time dependent and that rate, has a concentration of nutrients, c, which
decreases with the population due to the spread of limited resources over an
expanding population. The modified dN/dt value has two terms, bN/c which is
growth with a time scale set by b/c and −bN2/c which decreases the growth
with a rate which depends on the population N at any given time.
The symbolic solution to the last differential equation shown in Equation
(1.1) is obtained using the utility “dsolve” in the script “Malthus”, where the
printout is shown in Figure 1.3.
Note that for late times the modified solution, Figure 1.3, approaches an
asymptotic limited value, N → c and the population is self-limited by the
availability of nutrients as the population grows in size. An example of the
plots that can be formed by “Malthus” is shown in Figure 1.4, where the input
parameters supplied by the user are a = 1, b = 1, c = 5. The reduction in the
rate of growth with time is clearly evident as is the asymptotic approach of
N(t) to c, five in this specific example.
Figure 1.3: Solutions to the Malthus equation and modified equation, where Nmo is the initial
population and Nlo is the initial population in the modified case which has limited resources.

1.3. Missile Tracking

A mathematical problem arises when a missile acquires a target and moves to
intercept it. The target is first seen at (x, y) of (L, 0). The missile tracks to
intercept by always aiming at the present location of the target at each
moment. Presumably that position is instantly available by radar tracking, for
example. The resulting differential equation to describe the missile trajectory,
y as a function of x, is non-linear but solvable. The parameter q is the speed
ratio of the missile to the target.
Figure 1.4: Plot of the time dependence of a population in the case of unconstrained resources,
Malthus, and for limited resources. Both populations start at t = 0 with N =1.

The script “Missile_Track” uses MATLAB “dsolve” to solve the equation

symbolically. The actual form of the solution is not very transparent however.
In the user dialogue, Figure 1.5, the input is the parameter n and the initial
distance L when the target is first acquired. A movie of the pursuit is provided
and the last frame for the specific case of L equal to ten and a speed ratio of
two appears in Figure 1.6. The expression for the y coordinate of the intercept
at x equal to L is (L * q)/(q2 − 1) which is plotted in Figure 1.7 where the
distances are in units of L. The interception point appears in the symbolic
solution, part of which appears in Figure 1.5.

1.4. Monte Carlo Analytic

The Monte Carlo technique is widely used in making models of complex
systems. In the simplest view it consists of choosing variables from specified
distributions, assembling a system from these choices and then studying the
resulting performance of the system. In a few cases, the choice of a variable
with a specific form of the distribution can be done analytically. Some of
these cases are shown in the script “MC_Analytic” which uses the MATLAB
random number generator “rand”.
Figure 1.5: Printout of the user dialogue for “Missile_Track”. The first term in the symbolic solution,
ys, is the intercept point.

Figure 1.6: Last frame of a movie for the target, red line on the y axis, and the missile, blue line,
intercept at y of 6.67.
Figure 1.7: Plot of the y intercept in L units as a function of the speed ratio. The intercept diverges as
the speed ratio approaches one, as expected.

One solvable problem is the power law distribution of a variable x, where

α is the desired power. The range of the variables is xmin to xmax and r is a
random variable giving uniformly distributed values from zero to one.

A second example is an exponential distribution with a lifetime τ.

The Brite-Wigner resonance line shape can also be solved analytically and is
shown in Equation (1.5). The central value is xo and the width, full width at
half maximum is Γ.
A Gaussian distribution can be achieved by using two random numbers and
the fact that the joint probability of two uncorrelated variables is the product
of the probabilities. In Equation (1.6) the standard deviation of the Gaussian
is σ. The joint probability can have a “radius” chosen from an exponential and
an azimuthal angle chosen randomly from zero to 2π. The means of the
Gaussian can simply be added to the resulting x and y values.

The script, “MC_Analytic” gives the user a choice of these distributions and
the parameters which specify the distributions, such as the desired power
where the menu presented by the script appears in Figure 1.8. An example
printout for a resonance shape appears in Figure 1.9 and the output is a
histogram of the distribution, displayed in Figure 1.10.

Figure 1.8: Menu choices presented by the script, “MC_Analytic”.

Figure 1.9: Example of the dialogue for the choice of a resonance distribution.

In this script the MATLAB utility “hist” is used to create histograms of the
numerical results. The options available can easily be seen in the Command
Window by typing in “help hist”.

1.5. Monte Carlo Numeric

In the vast majority of cases the integral which is needed in the analytic case
either cannot be done analytically or the result cannot be simply inverted. In
such cases a variety of numeric techniques are available. The simplest is to
choose two random numbers. The first, r1, uniformly populates the x range
desired, while the second, r2, weights the chosen x by the probability
distribution P(x). The user is encouraged to make a graph of the procedure
which will show why it works. It is here assumed that P(x) has a maximum in
the x range being used. Using the maximum makes the choices of the random
numbers more efficient. The procedure is shown in Equation (1.7).

Figure 1.10: Example of the output plot for the choice of a resonance distribution with 1000 generated
entries.
An example from the script “MC_Numeric” appears in Figure 1.11. In this
case the angular distribution for Compton scattering with user defined initial
photon energy is produced. The user can vary the energy and see how the
distribution varies, as desired. This distribution is complex and cannot be
found analytically.

Figure 1.11: Angular distribution in Compton scattering for a specific, user defined, photon energy.

1.6. Monte Carlo Muon Beam

As an example of a somewhat more complex modeling problem, an initial
beam of muons is defined by the user and the particles in the beam are tracked
through a block of iron where they lose energy by ionization and also multiple
scatter in the material. The script is called “Monte_Muon”. This package
attempts to give the user a picture of how a complex system might be
modelled and then tracked through a particular environment.
The script uses the functions “Gauss”, “PowerLaw”, “Mult_ Scatt” and
“Euler.” The first two functions were described above and provide Gaussian
distributed or power law distributed variables. The function “Euler”
transforms a momentum vector, p, from a frame with the z axis oriented along
spherical angles ϑ and φ to the “laboratory” frame, p′, where the z axis has
those angles. A check for the special case where the momentum is along the z
axis, px = py = 0, yields the momentum components in the primed system
which are as expected, seen as the third column of the transformation matrix.

The Euler angles define a slightly more general transformation and they will
be defined and used in the next chapter of this text. They relate the frame
attached to a solid body to a frame related, in general, by three angles
defining the orientation of an extended body.
The script “Mult_Scatt” uses “Euler” and “Gauss”.
A muon is a charged particle with a rest mass of 106 MeV. It traverses a
length of material and suffers a scattering angle ϑ by an amount chosen from a
Gaussian distribution with a defined mean of zero and a standard deviation
fixed by the length of material being travelled and by the muon momentum.
Because the angles of scattering are not small at low energies or for thick
material and the energies may change substantially, the block of material to be
traversed is divided into segments, the number of which is chosen by the user.
The segments should be such that a substantial fraction of the muon energy is
not lost in any single segment nor should the scattering angles be large. A
specific dialogue is shown in Figure 1.12.
The beam starts at normal incidence to the block of iron. Assuming a path
strictly normal the beam would lose 8 MeV in energy and scatter by 38
degrees after traversing the full 50 cm of iron. However, the path is more
complex than that, which is what necessitates the more detailed Monte Carlo
modelling.
Figure 1.12: Specific choice of the muon beam of 200 MeV mean momentum.

The 200 MeV mean momentum beam loses about 15 MeV on average and
about 14 MeV in energy on average. In addition, the multiple scattering may
lead to long distances in the iron block, so that “End of Range” or the loss of
all kinetic energy of a given muon must be protected against. The energy loss
is highly dependent on the muon velocity, going as the inverse square, so that
behavior should be modelled properly.
The starting energy distribution is shown in Figure 1.13, while the exiting
distribution appears in Figure 1.14. Note the long, low energy tail due to the
traversal of many cm of iron. As before, the utility “hist” is used to make the
plots.
There is a long, low energy tail which arises from multiple scattering in the
block, whose transverse size is assumed to be very large. The actual path
length for the muon with large scattering angles then becomes very long. The
initial energy was confined to the region (205,250) MeV as seen in Figure
1.13.
Figure 1.13: Distribution in initial energy for a 200 MeV momentum mean muon beam with a 3%
momentum spread.

The x and y locations of beam particles for the initial beam and final beam
are shown in Figure 1.15. Note the scale change in the two plots by a factor of
20. These two plots underscore the need to do correct modelling. In particular,
note that a lower energy muon has more multiple scattering for a given
thickness of material.

1.7. Fourier and Laplace Transforms

Fourier and Laplace transforms are useful quantities in several applications in
physics. The Fourier transform is defined in Equation (1.9). The inverse has
the same definition save that the argument of the exponential changes sign.
For the Laplace transform, the inverse, fILT(x), has a different integration
contour, one where the singularities of F(s) are assumed to be to the left of the
parameter a.
Figure 1.14: Distribution in final energy for a 200 MeV muon beam momentum with 3% momentum
spread having traversed a 50 cm block of iron. The entries at zero energy have, in fact, stopped
completely.

The transforms can be explored using the script “Fourier_Laplace_Trans.”

The dialogue first asks if another transform is desired and then asks that
Fourier of Laplace be specified. Then, the user specifies whether the
transform or the inverse is desired. The printout of a dialogue to find the
Laplace transform of a specific function appears in Figure 1.16. The integrals
often appear with limits because they are found symbolically using the
MATLAB utility “int”. When that happens the user is asked to supply the
correct result, as seen in Figure 1.16 for example, and then plots of f(x) and
F(y) are created. The specific plot appears in Figure 1.17.
Figure 1.15: Distributions of muon positions in the beam at the start of the traversal of the iron (left),
and after complete traversal of the iron block (right). The scales are different by a factor of 20.

Figure 1.16: Example of the Laplace transform for sin(x). The user must input the correct limit in
order to invoke plots of f(x) and F(y).
Figure 1.17: Laplace transform, F(y) of f(x) = sin(x).

The Fourier transform for an exponential and a Gaussian can be obtained.

The Laplace transform for a power of x, square root of x and 1/x, exponential
I x, power of x times an exponential in x, and sin(x) or cos(x) all give correct
results. The inverse Laplace transforms are less successful, but the user can
obtain the pairs for the tranforms by asking for the Laplace transform itself.
These are continuum results. In the discrete case the MATLAB utilities “fft”
and “ifft” yield the desired results.
 Chapter 2

Classical Mechanics

“The most exciting phrase to hear in science, the one that heralds new discoveries, is not
‘Eureka!’ but ‘That’s funny’ …”
— Isaac Asimov
“Measure what can be measured, and make measureable what cannot be measured.”
— Galileo Galilei

2.1. Angular Momentum

Angular momentum, L, is fundamental to many aspects of physics. For
example, since the vector L is the cross product of the position and the
momentum, it is a conserved quantity for central forces. These forces cause
orbits confined to a plane perpendicular to L, as for example in the Kepler
planetary orbit problem. This constant of the motion allows the motion in the
plane to be reduced to a one-dimensional problem for central forces.

Two-dimensional motion in the (x, y) plane is assumed in the script

“Angular_Momentum”. The solution is symbolic using the user supplied
symbolic trajectories x(t) and y(t) to compute L(t). A specific printout is
displayed in Figure 2.1.
A movie of the time development of the trajectory and the Lz component is
then made. In the specific example the plot is shown in Figure 2.2. Note that
Lz is not here a constant of the motion because the trajectory does not
correspond to a central force result. The user should try several symbolic trial
functions and see how Lz evolves.
Figure 2.1: Printout from the script “Angular_Momentum” where Lz is calculated symbolically given
the user input for x(t) and y(t).

Figure 2.2: Time development of the (x, y) trajectory and the angular momentum. Note that Lz is not a
constant on the trajectory.

2.2. Classical Scattering

Scattering off a force center is used classically to explore the forces between
the target and the scattered particle. However, the total scattering cross section
is not unique to a given force law. An example is the three cases which give
the same total cross section πa2; a hard sphere of radius a, a barrier of height
Vo and width a′, and a well of depth Vo and width a. For a hard sphere, the
differential cross section, dσ = 2πbdb, with no scattering for radii greater than
a, yields the cross section
The relationships between the impact parameter, b, and the scattering
angle, θ, for the three cases are given below. For the hard sphere the angle is
zero when b is a and π when b is zero.

A plot of the three scattering angles with energy E equal to twice the
magnitude of the potential, where appropriate, as a function of impact
parameter is made in the script “Classical_Scatt” and shown in Figure 2.3. It
appears that they are rather different. However, since the differential cross
section goes as bdb, the total cross section contributions are very heavily
weighted toward large b, and the small b scattering contributes little to the
total cross section. The conclusion of the plot is that one must measure the
differential cross section in order to make a more incisive determination of the
force law which is active in the scattering.

2.3. Impact Parameter and DCA

The impact parameter, b, is the perpendicular distance between an incident
particle direction and the force center location at large initial separations. The
distance of closest approach, a, is that point on the particle trajectory closest
to the force center, which is assumed to be fixed. They are related by applying
the conservation of energy and the conservation of angular momentum for a
general power law central force as shown in Equation (2.3).
Figure 2.3: Scattering angle vs. impact parameter for the three classical interactions. The specific
choice for the barrier has a′ larger than a by 5%.

Printout from the script “CM_b_DCA” is shown in Figure 2.4 for α = 2,

attractive and repulsive and 3 for an attractive force. The script uses the
MATLAB utility “solve” to find an explicit closed form solution if one exists.
Figure 2.4: Printout of “CM_b_DCA” for inverse power law of 2, repulsive and attractive and power
of 3, repulsive.

Other power laws are possible and the user can try them. Plots of the ratio
of a/b for the case of an inverse cubic power law force as a function of the
incident velocity are shown in Figures 2.5 and 2.6 for repulsive and attractive
forces respectively. In the case of a repulsive force the ratio a/b is always
greater than one and only approaches one at high incident velocity. The
opposite behavior is seen for attractive force. At low velocity the projectile is
pulled to the force center and small a/b values are seen. Only at large velocity
does the point of closest approach reach the limit of the impact parameter.
Figure 2.5: Ratio of a/b as a function of incident velocity (arb. units) for a repulsive force. At low
velocity the particle is repelled to large values of a/b, while at high velocity the power 3 repulsive force
is relatively weaker and a > b.

2.4. Foucault Pendulum

A Foucault pendulum is one installed on the Earth’s surface which thus exists
in an accelerated coordinate system due to the Earth’s rotation. One possible
way to think about it is that the pendulum oscillates in “absolute space” while
the Earth rotates under it. In the Earth frame the equations of motion have an
added term due to the Coriolis force, where Ω is the circular frequency of the
Earth and φ is the latitude. In the absence of rotation the equations of motion
describe simple harmonic motion in the (x, y) plane of the Earth’s surface.
With the rotation, the motion in x and y is coupled due to the terms
proportional to velocity. In Eq. (2.4) the symbol ω refers to the frequency of
the pendulum in the absence of accelerated motion.
Figure 2.6: Ratio of a/b as a function of incident velocity (arb. units). At low velocity the particle is
attracted to the force center and small values of a/b are observed, while at high velocity the power 3
attractive force is relatively weaker and a > b.

These equations are solved symbolically using the “dsolve” MATLAB utility
in the script “Foucault_exact”. The approximate solution for the case where
ω, the pendulum harmonic frequency, is much less than Ω is also shown for
comparison. The script is quite slow because at least an entire day is mapped
out; be patient. The exact x and y positions in the Earth frame are shown in
Figure 2.7 for the case where the pendulum period is 0.05 days and it sits at
thirty degrees north latitude. The approximate solution is also plotted by the
script as a movie and illustrates that the motion is not re-entrant on a daily
basis in this case because of the approximations made.
Figure 2.7: Plot of x and y locations of a Foucault pendulum as a function of time in days. These
trajectories are re-entrant in time with period of 1 day.

The approximate solution factorizes into simple harmonic motion times a

modulation of the amplitude due to the Coriolis force.

The plot in Figure 2.8 shows the exact solution in the pendulum frame in blue
and a view in the “absolute” frame where the pendulum exhibits simple
harmonic motion. The movie of the two motions is plotted and Figure 2.8 is
the last frame of that movie.
A plot of the (x, y) motion of the pendulum (exact) is shown in Figure 2.9
for the case where the pendulum period is 0.1 days at 30°N latitude. The
motion is clearly nicely re-entrant after one day of motion.
Figure 2.8: Plot of the pendulum position, in the exact case, on Earth (blue) and in absolute space (red
stars) as the last frame of a movie of the Foucault motion.

2.5. Spring Pendulum

The simple pendulum problem can be extended by making the mass support
have spring properties. The equations of motion are then:

where r is along the spring with coefficient k and un-stretched length Lo and
the angle θ is the angle with respect to the vertical axis. The results of a user
dialogue with the script “Spring_Pendulum” which uses the MATLAB script
“ode45” are shown in Figure 2.10. The spring elongates and compresses and
the mass point follows while simultaneously being pulled toward the vertical
by gravity. The resulting radius as a function of time with respect to the un-
stretched spring at Lo is displayed in Figure 2.10. A movie is provided by the
script and the last frame of the movie is shown in Figure 2.11.
Figure 2.9: Plot of the pendulum position, in the exact case, on Earth, as it evolves in time. The
pendulum period is 0.1 days. The trajectory is plotted for 2 days by which time the pendulum has
returned to the initial configuration.

Figure 2.10: Dialogue for “Spring_Pendulum” defining the spring and the pendulum aspects of the
problem.

The user can check the solutions by going to the limiting cases of g equal
to zero, the spring alone, and the case where k is equal to zero, the simple
pendulum. In general the motion is not simple harmonic. For example, the
plot in Figure 2.11 displays a fairly complex pattern where the maximum
amplitude in each oscillation period changes as does the frequency itself. The
trajectory is illustrated in the last frame of a movie for the (x, y) motion of the
mass point for a slightly different set of parameters in Figure 2.12. The initial
velocities are assumed to be zero in this script.
Figure 2.11: Plot of the radius as a function of time. The unstretched length is 10 in this example,
where the parameters chosen appear in Figure 2.10, where the initial r is 12. Changes in r indicate
stretching or compression of the spring which supports the mass point.

2.6. Spherical Pendulum

The spherical pendulum is examined in the script “Spherical_Pend” which
employs the MATLAB tool “ode45” to provide a numerical solution to the
problem. The pendulum has a mass point on a fixed length r0 with gravity, g,
acting on the mass point. The mass point moves on the surface of a sphere
and there is a conserved angular momentum so that the problem is reduced to
one dimension. The equations of motion for the spherical coordinates are:
Figure 2.12: Last frame of a movie of the motion of the spring pendulum for parameters defined to be
[kmLo] = [1 1 10] and initial r and angle = [12 50′].

As can be seen in Eq. (2.7) the polar angle θ has acceleration terms due to the
centrifugal force, the L2 term, and due to gravity, the g term, pulling the mass
term toward the vertical. The azimuthal angle term simply expresses the
conservation of the angular momentum L, and is included to easily evaluate
the results for that angle.
The script asks for L and the initial angle and initial angular velocity.
After defining the initial conditions a movie of the subsequent motion is
provided and plots for a specific case are shown in Figures 2.13 and 2.14.
Figure 2.13: Plot of the polar angle as a function of time for the specific case where g/ro = 1, L = 0.01,
θ = 50° and dθ(0)/dt = 4 degrees/sec.

The user is encouraged to try a full range of possible initial conditions and
a range of parameters defining the spherical pendulum. For example, another
set of parameters and initial conditions are the input to the plot of Figure 2.15.

2.7. Driven Pendulum

A system that displays chaotic behavior is the driven pendulum. The script
“Driven_Pendul” allows for an exploration of this behavior. The pendulum is
damped and harmonically driven, so that the equation of motion is:

The pendulum moves in one dimension only. The damping factor is Γ, the
amplitude of the driving force is Fo and the driving frequency is ωd while the
natural frequency ωo is taken to be one in the script.
Figure 2.14: Plot of the (x, y) pendulum trajectory for the specific case where g/r0 = 1,L = 0.01, θ =
50° and dθ(0)/dt = 4 degrees/sec.

There is a region of operation where the motion is largely harmonic. In the

case where the initial angle is fifty degrees, initial angular velocity is zero,
damping is 0.5, Fo is 0.2 and ωd is 0.6 the motion is shown in Figure 2.16. A
movie is also provided by the script. If the damping is changed to 0.2, the
amplitude to 1.5 and the frequency moved to 0.7, the behavior changes and
the pendulum loops over the horizontal as seen in Figure 2.17. Each problem
may have a region of validity governed by the parameters of the problem and
the user can explore over which region of parameters the characteristic
behavior occurs.
In the quasi-harmonic regime, Figure 2.16, the pendulum itself has
transient time structure which dies out so that at long times the behavior is
controlled by the driving term frequency. In the chaotic regime, the
connection between the response and the driving harmonic function is less
obvious.
Figure 2.15: Plot of the (x, y, z) pendulum trajectory for the specific case where g/ro = 1,L = 1, θ = 60°
and dθ(0)/dt = 5 degrees/sec.

2.8. Random Walk

The random walk refers to steps taken in a random direction. This is the
“drunkards walk” problem; if a drunkard starts out from a lamppost and takes
100 steps, how far does he get from the lamppost? The path taken is displayed
using the script “rand_walk”. The user chooses the number of steps and
subsequently a movie of the path of those steps in made. In this specific case
the walk is in two dimensions and the length of each step is one unit. The
walk starts from the origin.
A plot for one hundred steps is shown in Figure 2.18. This is clearly a
stochastic problem with a random path. A physical example of such a path is
the Brownian motion of a particle scattered randomly by the fluid in which it
is immersed. The distance traveled is not zero, but goes as the square root of
the number of steps. This behavior is characteristic of stochastic processes. A
plot of the radius of the walk for one hundred random steps is displayed in
Figure 2.19.
Figure 2.16: Angular velocity in the stable regime of parameters where the motion is quasi-harmonic.

Note that the radius is not zero, but does display the square root
dependence which is expected. For one hundred steps a mean radius of ten
might be expected and is approximately observed. Another example of
stochastic behavior is that exhibited in multiple Coulomb scattering where the
mean overall scattered angle goes as the square root of the distance travelled
by the charged particle. This was assumed in the muon beam exercise
previously considered as coded explicitly in the script “Mult_Scatt” where the
mean multiple scattering angle goes as the square root of the material distance
travelled in radiation length units.

2.9. Rotating Hoop

The normal pendulum problem is set up in an inertial system and displays
simple harmonic motion for small oscillations about the vertical equilibrium
position. An extension of this situation looks at a hoop, radius a, with a
sliding mass point, m, on it, where the hoop rotates with a circular frequency
ω. There are two equilibrium positions possible, one at the bottom vertical of
the hoop and one at a non-zero angle with respect to the vertical.
Figure 2.17: Angular position for parameters appropriate to chaotic motion. Note that the range of
angles is greatly expanded showing that the pendulum has crossed the horizontal several times during
the time period set for the display.

The equations of motion follow from the Lagrangian formed from the
potential, V, and two contributions to the kinetic energy, T. The first term in
the expression for T is due to the mass point velocity. The second term in T is
due to the rotation of the hoop due to the centrifugal effect.
Figure 2.18: The path of a random walk in x and y with 100 steps of length one taken.

Using the Lagrangian, L = T − V, and the Lagrange equations of motion, the

resulting angular acceleration is zero if The critical frequency
squared for the second equilibrium position to exist is g/a and harmonic
motion about the second equilibrium position occurs for small angles with
respect to that position with a frequency
The script “Rotating_Hoop2” is used to solve this problem numerically
using the MATLAB utility “ode45”. The radius and mass are taken to be
equal to one. A specific example, with frequency four, no initial velocity for
the mass, and an initial angle with respect to the second equilibrium position
of 0.3 radians is displayed below. Printout of the dialogue appears in Figure
2.20.
A movie of the motion of the hoop and mass point is displayed, showing
the approximately harmonic motion of the mass point for this choice of
parameters. The last frame of the movie appears in Figure 2.21. The user can
create several movies in order to get a feeling for the parameter dependence
of the problem.
Figure 2.19: The radius from the starting point of a random walk with 100 steps of length one taken.

Figure 2.20: Printout of the script “Rotating_Hoop2” for a user supplied input of the hoop rotation
frequency. In this case the stable angle is 0.91 radians.

2.10. Tides

Newton first explained the tides. As was understood much later, it is the tidal
forces that are important in General Relativity because any overall
acceleration can be removed by going to a free falling frame as stated by the
Equivalence Principle. The tidal forces are defined locally at a mass point.
They both stretch and compress an extended body.
Figure 2.21: Frame of a movie of the rotating hoop. The hoop is drawn in green, the mass point in red,
and the equilibrium angle is a blue o. The hoop rotates in the movie in order to show the relationship of
the hoop rotation to the mass point oscillation.

Due to the attraction of the moon there are tidal forces which are in
directions and of magnitude such as to elongate the oceans in a line with the
moon. For the moon-earth distance of r0, the forces Fz and Fx at a location
specified at a point (x, z) near radius r0 with z along the moon-earth direction
and x transverse to z are:

The forces can be integrated to yield the gravitational potential, V(x,z)

describing the tidal forces.
The additional forces cause the tides to rise to a height h, whose potential
gh balances the potential added by the moon where the tides are assumed to
be equipotentials of the system. The tidal height, low to high, is about, ∆h ~
0.6 m in this approximation and depends on the masses of the Earth and moon
and their radii, Mm, Me, Rm, Re. There are clearly two tides per day. The angle
θ defines the angle between a point fixed on the earth at radius r and the
location of the moon in its orbit. In polar coordinates the potential is:
A movie of the tides is provided by the script “Tides”. A frame of that movie
appears in Figure 2.22, where the spherical earth is red, while the ocean tides
are in dotted blue. The location of the moon is a blue circle and its location
appears to rotate about the fixed Earth.

2.11. Action
There have been two fundamental descriptions of dynamics in physics, the
equations of motion approach and the minimum action approach. They are
equivalent, although the equations of motion, for example those of Newton,
are rather more familiar. They yield second order differential equations which
are typically solved by specifying initial positions and velocities. Several
examples in classical mechanics appear in this chapter in fact.
The other approach is the “action principle”. Simply stated the action, S,
along the true path is an extremal, minimum, and the end points of the path
are fixed. Alternative paths have an action larger than the path “chosen” by
the system.

Figure 2.22: A frame from the movie provided by “Tides”. The earth is red, the oceans are dashed
blue and the Moon is a blue O.

Two examples are shown in the script “Lagrange_Act2”, for motion in a

uniform gravity field and for simple harmonic motion. The mass and
harmonic frequencies are taken to be one for simplicity. The action in the two
cases has the same kinetic energy, T, but differs in the potential energy, V:

In the script the starting values at the twenty sample points are randomly
chosen around the points of the exact solution. Then the MATLAB utility
“fminsearch”, is used to minimize the action. Therefore, each use of the script
will yield a slightly different solution and the minimization, which is
numerical, may not converge to strict limits. Typical results for both simple
harmonic motion and for motion in a uniform gravity field appear in Figures
2.23 and 2.24 respectively. The minimization search is somewhat unstable, so
that initial positions and velocities, number of sample points, and time span
are defined by the script itself. The user may vary the parameters, but only by
modifying the script. Aside from these caveats, the minimization of the action
is made plausible by using this script. Using an enhanced number of points
would presumably also smooth out the behavior, but at a cost of rather longer
execution time for the search.

Figure 2.23: Exact and least action solution for simple harmonic motion. Twenty points in time are
selected and the action is minimized with respect to them using “fminsearch”.
2.12. Rocket Drag

The simple rocket, as modeled, has no external forces acting on it. That model
can be improved by taking the gravity field at the launch site into account. For
a uniform field with acceleration g, an analytic solution exists. An additional
force arises from the drag due to the viscosity of the air through which the
rocket passes. The height above the earth is y and vo is the exhaust velocity.
The acceleration drag depends on the rocket velocity, v, squared and the
density of the local atmosphere, which falls off with altitude.

Figure 2.24: Exact and least action solution for motion in a uniform gravity field. Twenty points in
time are selected and the action is minimized with respect to them using “fminsearch”.

The rocket mass is M and the mass burn rate, dM/dt, is assumed to be a
constant. The gravitational acceleration is g and the air density is ρ, which
falls off with altitude on an approximate scale β = 1/9140m. The drag
constant, k, in MKS units is about 5000. This equation is solved numerically
in the script “cm_rocket_drag” using the MATLAB utility “ode45”. The
advanced user could replace g by the more correct gravitational force, g →
g/(1 + y/R)2, where R is the radius of the earth and study the resultant changes
in the rocket trajectory.
Figure 2.25: Velocity of a rocket which is free, in a constant gravity field, and subject to gravity and
drag. The black horizontal lines are the escape velocity for the earth (solid) and sun (dashed).

The velocity of the rocket as a function of time is shown in Figure 2.25 for
a particular case chosen by the user, near to the parameters of a Saturn V but
with only a 1000 kg payload. In this example the payload attains escape
velocity from the Earth. The free case goes fastest as expected. In the case
with an added gravity field, the initial acceleration must become positive,
leading to a delay in gaining altitude. Adding drag does nothing until the
speed builds up as expected from the dependence on velocity squared.
However, at later times when the velocity increases the air density falls off
and the drag force drops.
The position of the rocket in the three cases is displayed in Figure 2.26.
The free rocket is always highest at a given time. The effect of viscosity is not
negligible and needs to be taken into account in any real application.
Figure 2.26: Height of a rocket which is free, in a constant gravity field, and subject to gravity and
drag. The effect of viscosity drag is non-negligible.

2.13. Two Stage Rocket

The basic rocket equation for exhaust velocity v0 and initial and final masses
of the rocket of mi and mf is found from applying momentum conservation to
the system of the rocket itself and the ejected mass:

Gravity and drag are now ignored in the comparison of two and one stage
rockets.
For a single stage rocket with initial mass m0, payload mass mp and ratio of
plumbing mass to fuel mass rat the final mass is the payload plus the
plumbing, so that the final velocity, with zero initial velocity is:

A two stage rocket can improve on the magnitude the final velocity by
jettisoning the first stage dead weight of plumbing. Assuming the two stages
have the same rat value, the same launch weight m0 and payload mass mp, the
final velocity for a two stage rocket can be derived by repeated application of
the basic rocket equation. It is:
For the one stage the initial mass is m0 and the final mass is the payload mass
plus the plumbing for the total fuel load. The two stage case allows some of
the plumbing to be shed with the first stage, which is assumed to be all dead
weight plumbing when it is cast away. The plumbing masses are mr1 and mr2
for the two stages. The total second stage mass of payload plus plumbing plus
fuel is m2. The final two stage velocity is the velocity gain from the first stage
plus the gain from the second stage.
The comparison is made in the script “Two_Stage”. Dialogue from that
script appears in Figure 2.27. The velocity comparison for the parameters
given in Figure 2.28 and is plotted in Figure 2.29. A clear gain in final
velocity is seen, which is why most space missions use multi-stage launch
vehicles. The time unit T is the total fuel burning time assuming no plumbing,
so that the actual times are less since dead weight is now explicitly assumed.
The specific assumption is that the plumbing weight is 30% of the fuel
weight.
There is an optimization that can be done for each particular set of
parameters. For a 30% ratio of plumbing to fuel the results are shown in
Figure 2.29. There is a weak optimal second stage weight to total launch
weight for these specific parameters which occurs at a second stage fractional
weight of approximately 10%.

Figure 2.27: Typical input parameters to compare a two stage and a one stage rocket.
Figure 2.28: Comparison of the final velocities attained by a one and a two stage rocket.

Figure 2.29: Final velocity as a function of the fractional mass taken by the complete second stage.

2.14. Table Top

This exercise consists of exploring the case of a mass m1 on a horizontal

plane surface connected to a mass m2 by a string of length L hanging below
the plane. There is a competition between the gravitational pull on the second
mass and the centrifugal force due to the conserved angular momentum of the
first mass point. There are well behaved solutions for moderate values of the
angular momentum L. For too small values the lower mass pulls the upper
mass down below the table. For too high values of L the upper mass pulls the
lower mass up above the table.
The mass points obey the equations of motion, Eq. (2.18), which are
solved numerically in the script “Table_Top” using the MATLAB tool
“ode45”. The user is encouraged to find the L value where the lower mass is
just pulled up to z of zero and the other L value where the upper mass point
approaches r of zero. The upper mass point has coordinates r and φ, while the
lower mass point is at a distance z below the table.

A plot of the motion of the two mass points is supplied as a movie and also as
a three-dimensional plot using “plot3”. As seen in Figure 2.30, the blue lines
indicate the positions of the two mass points.

Figure 2.30: Plot of the positions of the two mass points. The upper mass point moves in the plane at z
= 0, while the lower mass point moves up and down along the vertical axis z.

2.15. Kepler Eccentric

The simplest gravitational system to study is the motion of a light “test

particle” in the field of a very heavy body. In that case, the motion of the
heavy body can be neglected and the problem is solved when the motion of
the light planetary body is specified. Since the force is central, motion of the
planet is in a plane and angular momentum is conserved, leaving only two
variables, the planetary radius and polar angle to be specified. It is fortunate
that the solar system well approximates this effective one body planetary
motion, to that the resulting simple Keplerian laws can be extracted from the
motion of several, almost independent, planets.
The orbits for an inverse square law are conic sections; circles, ellipses,
parabolas and hyperbolae. The character of the orbit depends on the
eccentricity, e, which depends on the angular momentum of the planet. The
circular orbits, with radius rc, and radial velocity equal to zero are simplest.

The orbits for several values of e are shown in Figure 2.31. The orbit at e = 0
is a circle, which evolves with increasing e into an ellipse and then a
hyperbola. The script used is called “eccentric_Kepler”. The printout of the
script for a case where e is greater than one appears in Figure 2.32 which
gives the asymptotic angle.
The boundary between bound and unbound orbits occurs at e = 1 when the
degenerate ellipse opens up into a parabola. Note that, looking at Eq. (2.19),
the radius r becomes large when cos θ = 1/e, which has a real angle as
solution only if e is greater than one. The angles in question are the
asymptotes. The velocities in the bound case are vy = vc(1 + e cos θ) and vx =
vce sin θ both scaled to the circular velocity

2.16. Non-Central Force

The solutions for a specific non-central force are explored in the script
“Non_Central_force”. There is an additional harmonic force with amplitude b
and harmonic frequency ω which is shown in Eq. (2.20).
Figure 2.31: Orbits with eccentricity 0, 0.3, 0.6, 0.9 and 1.01, taking rc = 1.

Figure 2.32: Printout from the script “eccentric_Kepler” for an unbound trajectory with e = 1.2.

The gravitational force appears with unit strength. In the script, a symbolic
solution is first tried but it is not successful. Then a numerical solution using
“ode45” is adopted. The user dialogue allows for a choice of the strength of
the added force and the angular frequency. The major axis of a Keplerian
orbit and the angular velocity are then chosen, which defines an elliptical
orbit in the absence of the additional force.
The perturbing force distorts the ellipse, elongates it and causes it to be
non-reentrant. The angular momentum is no longer conserved, Eq. (2.20), so
that the problem is intrinsically two-dimensional. For small amplitudes the
effect is small. For high frequencies with respect to the main orbital motion
the effect is also washed out since the averaged force approaches zero.
Typically, for a high enough frequency and a small enough amplitude, the
force only causes the ellipse to not close on itself, as illustrated in Figure 2.33.
At larger amplitudes the effect of the added force becomes more pathological,
as the user can easily explore.

Figure 2.33: Movie of the orbit for a strength of 0.3 and a frequency of 1. A major axis of 4 is
assumed for the Keplerian ellipse.

Figure 2.34: Movie of the orbit for a strength of 0.6 and a frequency of 1. A major axis of 4 is
assumed as the starting point of a force free ellipse. The actual orbit is nearly unbound due to the
additional force.
 A stronger force with a low frequency will distort the orbit even more. A
low frequency means that the sign of the force does not change rapidly during
the orbit and thus the force pushes the orbit to large radii. The strength is then
sufficient to make the bound orbit into a free orbit. A movie for this situation
is shown in Figure 2.34.

2.17. Euler Angles

The Euler angles were created to relate the position of a rigid body in its own
rest frame (body frame) to the position in a frame where the body is in motion
(space frame). Indeed, the transformation has already been invoked in the
multiple scattering script which was used in the exercise in the study of a
muon beam, where the multiple scattering with respect to the muon direction
was transformed to the angles in the laboratory (space) frame through which
the muon propagated. In that script a special case of the general
transformation, only two angles needed, was used.
The general transformation first rotates about the z axis, followed by a
rotation about the x axis and then a second rotation about the z axis. The
cosine and sin, c and s, of the three angles are indicated in Eq. (2.21) below.
In the figure, all angles are 35°.

A visualization of the effect of these rotations is provided by the script

“Euler_Angles” where the script output plot appears in Figure 2.35. The
locations on a circle and the x and z axes are indicated by blue, solid → red,
dash dot → green, dashed → black dotted for the initial axes and then the
results of the three successive rotations. The first rotation is in the (x, y) plane,
while the second is about the transformed x axis. The final third rotation is
about the transformed z axis. The specification of the position of a rigid body
requires three angles in Euclidian three-dimensional space, assuming the
C.M. velocity is zero.
Figure 2.35: Euler angle rotations from body → space frames showing the x and z axes and a circle in
the (x, y) plane for the three successive rotations, blue is initial, red is after the first rotation, followed by
green and then black. The orientation is indicated by the x and z axes and a circle in the (x, y) plane in
each case.

The Euler angles will be illustrated by the study of top motion. However,
they are perfectly general and can be applied to any situation describing rigid
body motion.

2.18. Forceless Top

A simple example of rigid body motion is that of a top spinning without
external forces acting on it. The shape of the top is defined by the moment of
inertia about the rotation axis, Iz, and the moments about the other principle
axes, I1, assuming that the top is rotationally symmetric. The top axis of
rotation precesses in this case with a frequency ωpre = [(Iz — I1)/I1]ωcos θ
where ω is the top rotational frequency and the top axis is inclined by the
angle θ with respect to the vertical. The precession frequency depends on the
top being non-spherical.
An illustration of the motion is given in the script “Top_Forceless”. A
specific output plot is shown in Figure 2.36. The user can change the shape of
the top and its inclination angle and gain some insight on the subsequent
motion. In this case the top rotational axis simply precesses with circular
frequency ωpre about the z axis at a constant angle of inclination.

2.19. Top Motion

An illuminating use of the Euler angles is the study of top motion. The top is
acted on by a torque due to gravity which then changes the angular
momentum of the top. The top is assumed to be rotationally symmetric which
leads to a constant of the motion, ωs. There remain five variables to solve, for
the three angles specifying the top and their derivatives, less the constant of
motion. The torque is defined by the mass of the top, m, and the height of the
C.M. of the top above the fixed point of rotation, h.

Figure 2.36: Precession of a top not acted on by any forces. The axis of rotation is indicated by the red
lines as a function of time in the movie provided. The top has a shape with Iz = 1.5 and I1 = 1 and the
inclination angle is 30 degrees.

The rotation axis is z with moment of inertia in the body frame defined to be
Iz, while the other axis is defined to be, as before, I1. In general the top
rotational axis precesses, as in the forceless case, although the precession rate
is now not a constant. It also nutates, with a complex path swept out in the
polar angle variable. This motion has the top initially dropping under gravity
and then rising in a periodic fashion. In fact, with extreme values of the user
supplied parameter, the top will “tumble” and then right itself later.
The equations to be solved in the space frame are:
Since the top is not translating there are three variables specifying the top; the
polar and azimuthal angles of the top axis in the space frame and the angle ψ
defined by the spin of the top itself about its axis of rotation.
This set of equations is solved numerically in the script “Top_Euler” using
the MATLAB utility “ode45”. The initial azimuthal angle φ is zero, while the
initial polar angle θ, which defines the inclination of the top, is defined by the
user. The initial value of the time rate of change of the azimuthal angle sets
the behavior of the nutation, and it is also a user supplied input. The spin is
defined by the angle ψ. In general there are five functions to determine. The
dialogue is shown in Figure 2.37.

Figure 2.37: User definition of the motion of a top. The top mass and shape are user defined as are the
top spin and angle of inclination from the vertical.
Figure 2.38: Precession of a top with gravitational torque acting on it. The structure in time is no
longer simple as it was in the forceless case.

A typical result for the precession in azimuth is shown in Figure 2.38. The
rate is not constant, as in the forceless case, but is modulated by the nutational
motion in polar angle.
In Figures 2.39 and 2.40 the nutational pattern for the polar angle is shown
for two values of the initial time rate of change of the azimuthal angle. In
Figure 2.41 the last frame of a movie of the motion of the top axis is shown,
illustrating the precession and nutation.
If the time rate change of initial azimuthal angle is set to zero, the top
merely drops initially in polar angle until if rises and executes simple
harmonic motions. There is another special case, where the top is initially
vertical, the “sleeping top”, where the top then simply spins about the z axis
as in the forceless case because the torque is then zero. The user has several
parameters to explore in this exercise. The motion is complex and the user is
highly encouraged to vary the top defining parameters and see what the
subsequent motion is. The top may “sleep”, or simply precess, or fall below z
= 0.
Figure 2.39: Nutation in the case of a small positive value for the time rate change of the azimuthal
angle.

Figure 2.40: Nutation in the case of a larger positive value for the time rate change of the azimuthal
angle.
Figure 2.41: Motion of the axis of the top for the case of a 30-degree initial angle and initial azimuthal
time rate of change as in Figure 2.39.
 Chapter 3

Electromagnetism

“One day sir, you may tax it.”

— Michael Faraday’s reply to William Gladstone, then British Chancellor of the
Exchequer, when asked of the practical value of electricity
“Magnetism, as you recall from physics class, is a powerful force that causes certain items to
be attracted to refrigerators.”
— Dave Barry, Pulitzer Prize-winning American author and humor columnist

3.1. Electron e/m

The electron was the first fundamental particle of matter to be discovered.
Since electrons are bound in atoms only by a few electron-volts, they can be
relatively easily liberated from atoms. The apparatus is shown schematically
in the output of the script “Electron_e_m”. Electrons are boiled off the
cathode of a vacuum tube. They are then accelerated in an electric field
indicated in Figure 3.1 below as the black boundary lines. The acceleration
due to the vertical electric field causes the path to curve, shown as a red line.
Applying a magnetic field using the Helmholtz coil indicated as four blue o,
the path can be restored to be a straight line as confirmed using a phosphor
screen at the rightmost end of the vacuum tube.
The user can derive the condition for no net force with a known electric, E,
field along y and a magnetic, B, field along z. The measured length is L and
the measured displacement in the case where the magnetic field is zero is y.
The initial x velocity is v set by applying a voltage to first accelerate the
electrons before they enter the region with the fields. Therefore the charge to
mass ratio of the electron can be measured using this technique. The distance
between source and screen is L and t is the transit time.
Figure 3.1: Schematic of the apparatus used to determine e/m for the electron. The e without the B
field is deflected by the vertical E field by y. When the B field is adjusted so that there is no deflection,
the e/m ratio is determined.

In Eq. (3.1) all quantities are measured on the right-hand side of the
expression for e/m. In this fashion J. J. Thomson was able to determine the
charge to mass ratio of the electron.

3.2. Potential and Fields

Conservative forces like gravity are derivable from potentials. The fields are
the gradient of the potentials, . The script “Potential-Field” asks for
a symbolic input potential in two dimensions, plots it and then finds the x and
y fields due to that potential. This is done symbolically using the MATLAB
utility “diff” to find the fields with a defined potential. The script is therefore
useful in visualizing the shapes for any potential that can be expressed
analytically. The potential, V = tanh(xy) is plotted in Figure 3.2, and the x
field for that potential appears in Figure 3.3. The user is encouraged to
explore other potential shapes which might interest him/her.
3.3. Electrostatics — Plates
The problem of a charge located at x = 0 and y = d between two grounded
plates at y = 0 and y = a is considered in the script “EM_Plates”. The potential
is assumed to factorize into a function of x times a function of y. The solution
is a series of terms which vanish at the conducting plates. The solution for x >
0 has a sign change in the argument of the exponential compared to the
solution for x < 0 which is shown in Eq. (3.2).

Figure 3.2: Potential and contours for V = tanh(xy).

Figure 3.3: Field in x for the potential V = tanh(xy).

The user chooses a value of d/a and the number of terms in the series. The
potential for the specific choice d/a of 0.8 and 20 terms appears in Figure 3.4
for the potential. That function peaks at the location of the point charge and
vanishes on the plate boundaries as expected. The electric field in the y
direction is evaluated using the MATLAB numerical utility “gradient” and is
shown in Figure 3.5.
The user should vary the number of terms to get a feeling of how quickly
the series converges to a smooth behavior as a function of x Potential and y.
The potential must vanish at the plate boundaries at y equal to zero and a and
must fall of rapidly as x differs from the charge location at x of zero.

Figure 3.4: Potential for the case where d/a = 0.8 for a series with 20 terms.

3.4. Electrostatics — Angled Plates

A different but similar geometry is explored in the script “EM_Angled”. In
this case the grounded plates intersect at the origin and are at an angle α with
respect to one another. The point charge is located between the plates at an
angle β at a radius ro. The observation point is defined by the radius r and the
angle θ. The potential is again solved as a series. However, the geometry is no
longer Cartesian but polar so that the series is not a Cartesian Fourier series
but a power series in radius.
Figure 3.5: Electric field in the y direction. There is a normal field which reflects the induced charge
on the upper plate.

The solution for r greater than ro inverts the power to be (ro/r). This
ensures that the potential vanishes at the origin and falls off at large r values,
as is physically reasonable. The potential is displayed as a function of x and y
after the user defined values of α and β and the number of terms of the series
are given. The result for ro = 1 with a thirty degree angle of the plates and a
ten degree angle for the charge are shown in Figure 3.6.
The contours indicate that the plates are equipotentials which therefore
satisfy the boundary conditions. There is a general falloff with distance from
the point charge. Another example is shown in Figure 3.7. In this case α is
sixty degrees and β is forty degrees. The change in the equal potential
contours is very evident with respect to the previous example. The vertical
field near the lower plate, Ey(x, 0), is shown in Figure 3.8. It was derived
using the potential and the MATLAB utility “gradient”. The x electric field at
the lower plate is small, zero within the numerical accuracy as is required by
the boundary conditions. The user is encouraged to try several values of the
two angles.
Figure 3.6: Potential contours for the case where α = 30 degrees and β = 10 degrees. The plates are
located at y = 0 and the second plate is indicated by the red dashed line.

3.5. Electrostatic Helmholtz

The Helmholtz coil consists of two magnetic coils such that the magnetic
field between them is rather uniform spatially. The electrostatic analogue is
two rings, radius a, of charge plus and minus Q with an axial distance L
between them. The dialogue for the script “Electric_Helmholtz” for a radius a
of one appears in Figure 3.9. The field is axial by symmetry.
For the specific relationship between the separation L and the radius a,
, the first and second derivatives both vanish in the center of the
two loops. The solution is found using “factor” on the second derivative
evaluated at z of zero in the script. The field and the derivatives for the special
relationship are shown in Figure 3.10. The field is quite constant over the full
L range. Both the first and second derivatives vanish in the center of the rings
which implies a rather constant field along the axis.
Figure 3.7: Potential contours for the case where α = 60 degrees and β = 30 degrees. The first plate is
located at y = 0 and the second plate is indicated by the red dashed line.

3.6. Electric Quadrupoles

The electric field has various multipole components which have characteristic
angular and radial dependencies. The field due to an isolated charge goes as
1/r2 and is isotropic. For a dipole the field goes as 1/r3 and has a cosθ
distribution for a dipole moment oriented along the z axis. For a quadrupole
field the radial dependence is 1/r4 with an angular dependence, for Qzz the
sole quadrupole moment component, of 3cos2θ – 1.
The script “EM_Quadrupole” treats the quadrupole field symbolically,
using the utility “diff” to find the fields. The basic charge distribution can be
thought of as a superposition of two dipoles oriented in opposite directions.
There are single charges at +z and −z and an opposite doubles charge at z of
zero. The printout made by the script appears in Figure 3.11. A plot of the
radial dependence of a quadrupole, both potential and field appears in Figure
3.12. The inverse fourth power of the field appears even rather close to the
quadrupole structure.
Figure 3.8: Field normal to lower plate for the potential shown in Figure 3.7. The field Ey for y = 0
vanishes at the origin and peaks around the x location of the charge between the plates.

Figure 3.9: Printout of the script “Electric_Helmholtz”. The first and second derivatives of the field
are prints by the script but are not shown here. The MATLAB symbolic utility “diff” is used to derive
the fields and derivatives so as to explore the uniformity of the field.
Figure 3.10: Plots of the field and its derivatives for the special case of a uniform field, where
where a is =1 in this plot.

A plot of the spatial shape of a quadrupole with only a z, z component of

the quadrupole moment is plotted in Figure 3.13. The behavior clearly has
two lobes, while a monopole shape has no poles and a dipole shape has only
one pole.

3.7. Cylinder Boundaries

Consideration of the electrostatic problem of a long cylinder, radius a, where
the positive y half cylinder is at potential Vo while the bottom is at –V0 is
given in the script “Laplace_Cyl”. The Laplace equation is appropriate as
there are no sources, only specified boundary values.
Figure 3.11: Symbolic calculation of the potential and the radial electric field for a simple quadrupole
arrangement of charges. An expansion of the field in small values of a/r reveals the inverse fourth
power of the radius as the leading term in the field.

Figure 3.12: Loglog plot of the potential and radial field for a simple quadrupole layout. The x axis is
(1, 3) in units of a. The line with 1/r3 dependence is to compare to the potential r dependence. The
power law behavior sets in quite rapidly, with r/a less than 2.

Figure 3.13: The solid shape derived for the spatial dependence of the radial field of a quadrupole
with a moment possessing only a z, z component.

Both the region interior and exterior to the cylinder is solved using a series
method. The small radius solution vanishes at the origin, while the large
radius solution falls off rapidly with radius. The potential, in Vo units, with
indices that are odd is:

The script assumes that the radius, a, is one as is the potential magnitude at
that radius. The user chooses the number of terms in the series, which allows
a visual inspection of the convergence of the series. The electric fields are
derived numerically using the MATLAB “gradient” tool. The potential for a
series with twenty odd terms is shown in in Figure 3.14, while the y
component of the electric field appears in Figure 3.15. With twenty terms the
representation on the radial boundaries appears to be quite good in that the
constancy of the potential is manifest.

Figure 3.14: Potential both interior and exterior to a cylinder of radius one and potential one. The rise
in V at interior r and the falloff at large r are evident.

3.8. Laplace Fourier

Boundary problems with the Laplace equation can be solved for Cartesian
boundary conditions by variants of the Fourier series technique. The related
problem for cylindrical boundary conditions appears in Section 3.7. For the
case of a conducting box of length 2a in x and 2b in y, with the left and right
sides at x = −a and x = a respectively held at voltages VL and VR and top and
bottom grounded the solution for the interior potential is a series with only
odd terms

Figure 3.15: The field derived from the series from the potential of Figure 3.14. The behavior is
complex at the boundary along y = 0.

The script “Laplace_Series” covers this situation. The user supplies the
parameters, a, b and the left and right voltages. In addition the number of
terms in the series is asked for so that the user can see how rapidly the
convergence of the series is. A mesh plot for the case of a equal to two, b
equal to one, left voltage of 10 and right voltage of 5 with twenty terms in the
series is shown in Figure 3.16. The left and right constant voltages on the
boundaries are clearly seen, as well as the voltage fall from left to right.
Figure 3.16: Voltage inside the box with sides at y = 0, 1 grounded and sides at x = 0, 2 held at 10
volts and 5 volts, respectively.

The electric field in the x direction for this specific case is shown in Figure
3.17. It is large in the four corners of the box where the potential changes
rapidly as expected.

3.9. Rotating Charged Sphere

A rotating charged sphere, radius a, will have an associated current which will
produce a magnetic field, both inside and outside the sphere. The solution for
this problem appears in Eq. (3.6). The axis of rotation is z. The solution can
be thought of as a superposition of many current loops, found by integration.
The units are indicated in Eq. (3.6) which factor out the rotational frequency
and charge as well as numerical factors
Figure 3.17: Ex inside the box with sides at y = 0, 1 grounded and sides at x = 0, 2 held at 10 volts and
5 volts, respectively.

The current depends on the charge, Q, and the rotational frequency ω. The
coordinates r and θ are spherical polar, and the problem is azimuthally
symmetric. The magnetic field is shown in Figure 3.18. The field has a dipole
like angular and radial dependence for radii greater than a, taken to be equal
to one in the mesh plots. The interior field at the origin is cosθ(5/2a), while it
falls to cosθ/a at the surface of the sphere, matching the exterior solution for
the field.
Figure 3.18: Mesh plot for the interior and exterior radial magnetic field of a rotating charged sphere.

The field in the polar angle direction is defined in Eq. (3.7) and is shown in
Figure 3.19.

The angular field peaks at ninety degrees to the axis of rotation which is
typical of a dipole field. At larger radii the field falls as 1/r3. The interior field
is largely negative and rises as r2, matching the exterior solution at radius of
a.

3.10. Dielectric Image Charge

Dielectrics have polarization charges which change the boundary conditions

which are appropriate to vacuum and to perfect conductors. There is a
displacement vector field D which is the dielectric constant times the electric
field. The tangential electric field is continuous across a boundary between
differing dielectric objects as is the potential, while the normal electric field is
not. The normal D field is continuous.
Figure 3.19: Mesh plot for the interior and exterior angular magnetic field of a rotating charged
sphere.

The simple case of a point charge, q, located at a distance a above a

dielectric medium filling the half space defined by z less than zero. The
dielectric constant for that region is K. At the boundary Ex is continuous as is
the potential, V, but Ez is not. The problem is explored in the script
“Dielectric_Image”. The method of image charges is used to satisfy the
boundary conditions. The symbolic MATLAB utility “diff” is used to derive
the fields from the potential and some of the results which are printed out
appear in Figure 3.20.
For z greater than zero the potential has a term due to the origional charge
q located at z of a, and an image charge –q(K – 1)/(K+1) located at z of –a.
The upper plane is assumed to be vacuum. In the lower half plane there is the
potential due to charge q at z of a and an image term as before, but now also
located at z of a. This potential is shown in Figure 3.21, in the specific case of
a of one and K of four.
Figure 3.20: Printout of “Dielectric_Image” which gives the potentials appropriate for z > 0 and for z
< 0 and Ex for the two regions. The field Ez is also printed out by the script but is not shown here.

The fields Ex and Ey are evaluated after being derived symbolically from
the potential. Both fields on the z axis at x equal to zero are shown in Figure
3.22. Clearly the tangential field is continuous, while the normal field is
discontinuous, though the potential itself is continuous. The net result is to
decrease the fields in the dielectric.
The user is encouraged to try some limiting cases and see if intuition is
verified. For K equal to 1, there should be no boundary. For finite K, Ez will
be discontinuous at the boundary while Ex is continuous. For a very large K, a
conductor is approached as a limit, so that Ex (0) approaches zero in the upper
half plane.
Figure 3.21: Potential as a function of x and z for a charge at z = 1 above a half plane with dielectric
constant of 4.

3.11. Dielectrics

The problem of a conductor surrounded by a dielectric cylinder immersed in a

uniform electric field, E, is explored using the script “Dielectric”. The user
chooses the two radii, a inner and b outer, and the dielectric constant, K, of
the cylinder. The solutions for the potentials outside and inside the dielectric
cylinder are:
Figure 3.22: Plot of the z dependence of the tangential (left) and normal (right) field at x = 0. The log
of the field is plotted to bring out the behavior at the boundary at z = 0.

The constants A, B and C are found by matching the boundary conditions

at the two radii, and the results can be found in the script, “Dielectric”. The
exterior potential approaches the applied field Ez since the term A/r falls off at
large r. Inside the dielectric terms that go as 1/r and r both exist. For r less
than a, the conductor, the fields are zero.
The potentials for a choice of one for radius, a, of the conductor (red), and
two for the dielectric radius, b, (green) and a dielectric constant of five are
shown in Figure 3.23. There is no field in the conductor as in true for all
electrostatic shielding and the surface at radius a is an equipotential. From the
potential plot it is clear that there are both normal and tangential fields at the
interface at a radius of b.
Note that the potential lines are not perpendicular to the green boundary as
they would be for a conductor and are at a radius of a for the surface of the
conductor. The user can vary the dielectric constant in order to see how the
boundary conditions change the potential.
Figure 3.23: Equipotentials for a specific example of a dielectric cylinder surrounding a conductor and
immersed in a field which is uniform in z at large distance.

The electric fields derived from the potentials are plotted in Figure 3.24,
which used the MATLAB plotting utility “quiver”. The fields are not
perpendicular to the outer, green boundary. They also have a large angular
dependence. Inside the conductor the fields are zero to the level of numerical
accuracy.

3.12. Dielectric Sphere

A dielectric sphere immersed in a uniform electric field, Ez does not result in a
zero field inside the sphere as is the case for a conductor. For the case of a
dielectric constant K the potentials inside the sphere and outside and the
induced surface charge, σ, normalized to the external electric field are:
Figure 3.24: Electric fields for a specific example of a dielectric cylinder surrounding a conductor and
immersed in a field which is uniform at large distance. The field inside the conductor is numerically
zero.

At a radius of a the potentials match to a normalized value of −3/(K + 2)*a

and the potential is continuous across the boundary. If the dielectric constant
is equal to one, vacuum, then the induced charge is zero as expected. The
script “Dielectric_Sphere” plots the solution for a user defined dielectric
constant. The equipotentials for a constant K equal to five appear in Figure
3.25 and the derived electric fields are shown in Figure 3.26. Clearly, the field
penetrates into the sphere. The user can view the depth of penetration while
varying the dielectric constant. The special case of K equal to one is also
instructive.
Figure 3.25: Equipotential contours for a dielectric sphere with constant K equal to five. There is no y
gradient inside the sphere. On the boundary there are both normal and tangential fields.

The electric fields of Figure 3.26 are again derived from the potential using
the numerical MATLAB utility “quiver”. The interior fields are horizontal,
and the vertical field is essentially zero. The induced surface charge goes as
the cosine of the angle as does the external electric field.

3.13. Induction
An electromotive force, EMF, is induced in a circuit when magnetic flux lines
are time dependent, an effect first shown by Faraday. Consider the simple
case of a moving charge, q, which creates both electric, E, and magnetic, B,
fields
Figure 3.26: Electric fields for a dielectric sphere with constant K equal to five immersed in a uniform
electric field along the z axis. The field inside the sphere is non-zero and largely along the z axis.

The EMF is proportional to the magnetic field (flux lines) and the particle
velocity (rate of change of flux lines), which makes it proportional to the
square of the particle velocity. A simplified model is provided by the script
“EM_Induction”. A charged particle crosses the centerline of a rectangular
loop of conductor at a user supplied height and velocity. Non relativistic
motion is assumed and low velocity forms for the fields are used. Plots of the
EMF induced by the moving charge appear in Figures 3.27 and 3.28 for a
velocity of two and heights of 0.2 and 0.5 above the loop, where the loop has
area one, respectively. Note the reduced value of the EMF and the wider time
structure for the case of a height of 0.5 compared to a height of 0.2. The EMF
appears as a movie tracking the motion of the charge.
Figure 3.27: Induced EMF for a charged particle crossing the center of a conducting loop with
velocity 2 and at a height 0.2 above the loop.

The user can establish the velocity dependence of the induced EMF
through the user supplied velocity in the dialogue and by viewing the
resulting plots which are movies with equal time difference between the
frames shown as blue o.

3.14. Magnetic Bottle

For many years physicists have designed “magnetic bottles” which use
shaped magnetic fields to contain a hot plasma in a search for fusion
containment. In this way, the high temperatures needed to initiate and control
a fusion reaction for a plasma fuel are to be achieved in a safe fashion. The
latest attempts favor magnetic confinement over other techniques such as
inertial confinement. A test reactor with magnetic confinement (a “tokamak”)
is being built in France with the research goal of producing net fusion power.
Figure 3.28: Induced EMF for a charged particle crossing the center of a conducting loop with
velocity 2 and at a height 0.5 above the loop.

The script “Mag_bottle” makes a very simplified threedimensional model

of such a bottle. The basic equation is the Lorentz force, . The
fields are chosen to “reflect” the charged particles from regions of high field
gradient. The radial field goes as the hyperbolic cosine of the radius, while the
z field approaches [z] far from the origin. The fields appear in Figure 3.29.
The equations of motion for a charged particle in an arbitrary magnetic
field are solved using “ode45” and the positions and velocities are displayed
using “plot3”. For the specific case of initial positions, [x y z] = [0.5 0 0] and
velocities, [vx/vyvz] = [0 −0.2 2] the last frame of the trajectory movie which is
provided is shown in Figure 3.30.
Note that the particle tends to be “reflected” from the high field regions. In
this case the charged particle is contained for this time interval. However, the
user is encouraged to vary the parameters in order to see how “leaky” the
bottle is. If it were easy the world would have a many clean fusion reactors
operating for some time by now. For example, the same initial positions and
with initial velocities [0, −0.4,3] the particle velocities appear in Figure 3.31
and they imply that the charge will escape.
Figure 3.29: Radial and axial fields of the bottle used as an example.

3.15. A Fusion Reactor

The sun is a fusion reactor, where the plasma is contained by gravity. The
required temperatures are very high for high fusion rates because the thermal
energies must overcome the Coulomb repulsion of the charged reaction
products. Therefore, confinement on earth must be accomplished by non-
material means, such as magnetic fields. We simply assume that a plasma is
confined in a magnetic bottle and the temperature is raised by external electric
fields. The script appears in “Fustion_ITER” where ITER refers to an
experimental fusion reactor being built in France.
Figure 3:30: Trajectory for a specific example of a particle in a “magnetic bottle”.

Figure 3.31: Velocities in the magnetic bottle for initial positions, [0.5, 0, 0] and velocities [0, −0.4, 3].

The normalized Maxwell–Boltzmann proabability, dN/dE, is given in Eq.

(3.11) where the plasma energy is E and the plasma temperature is T.

The reaction rate density, r is proportional to the square of the deuterium

plasma density, the deuterium velocity and cross section. Each plasma particle
can react with nD others, which explains the square dependence of the total
rate. The cross section sets the reaction probability scale and the velocity is
needed to initiate a reaction. Overall the energy dependent cross section is
weighted by the distribution of plasma energies, as set by the Maxwell–
Boltzmann factor.
The results of a dialogue initiated by this script appear in Figure 3.32.

Figure 3.32: The printout made by the script “Fusion_ITER” showing the exothermic reactions. The
user chooses the temperature and density of the deuterium plasma.

The reaction rate per unit volume is computed by using the MATLAB
facility “quad” to integrate over the spectral reaction rate which is formed
from the Maxwell–Boltzmann energy distribution convoluted with the d + d
(d denotes deuterium) cross section energy dpenendence. The d + d reaction
cross section is roughly constant at about 0.001 barns for energies above
about 20 keV. The reaction rate as a function of deuterium energy is shown in
Figure 3.33. Also shown, red *, is the mean energy at the user defined
temperature, 〈E〉 = 3kT/2. It is clear that the maximum of the reaction rate
occurs at a higher temperature due to the sharp rise in the d-d cross section
with energy.
The reaction rate for d-d fusion as a function of both temperature, which
sets the thermal energy distribution, and energy, which reflects the reaction
cross section, appears in Figure 3.34. Clearly the fusion rate increases very
rapidly with energy, which puts a premium on achieving the highest contained
plasma temperature possible. For example, a 100 million degree plasma with
a density of 1022 deuterons per cubic meter would generate thirteen
megawatts in a cubic meter of fuel. A factor of two more temperature raises
that to about one hundred watts per cubic meter. That is a truly impressive
power density.
Figure 3.33: Reaction rate density as a function of d-d C.M. energy. The peak is at about 50 keV,
substantially above the mean energy at the chosen temperature, shown as a red *.

Figure 3.34: Reaction rate density as a function of d-d C.M. energy. At high temperatures, the rate
increases very rapidly.

3.16. Dipole Radiation

Dipole electromagnetic radiation is explored in the regime where the radius

and inverse wave vector, r and 1/k, are much larger than the size of the dipole
in the script “Dipole_Power”. The velocity c is taken to be one. The
expression for the dipole power angular distribution is shown in Eq. (3.12).
The dipole angular distribution is the sin squared of the polar angle of k with
respect to the dipole direction. There is a wave outgoing at the speed of light
which falls as a radiated energy as inverse of radius, so that the power
crossing a sphere of radius r is independent of the size of r. This is the basic
characteristic of a radiation field.

A movie of the power is provided in Figure 3.35 where the outgoing

contour of the power in the (x, y) plane is shown. The dipole is assumed to be
oriented along the y axis. Contours of constant power appear in a movie as a
function of time for the dipole oscillations which drive the radiation. The
radiative nature of the power is quite evident.
A second movie for the radiated power appears in Figure 3.36. In this case
the power is plotted as the height above the (x, y) plane. There is a trough at y
of zero due to the dipole nature of the radiation. The outgoing wave crests
move at the speed of light and the frequency of the waves is defined by the
oscillations of the dipole moment, which are visible near the origin.

Figure 3.35: Radiated angular pattern for a dipole when r and 1/k are greater than the size of the
dipole itself.
Figure 3.36: Radiated angular pattern for a dipole when r and 1/k are greater than the size of the
dipole itself. The wave crests move out with the velocity of light and reflect the dipole oscillation
frequency, seen here in red.

3.17. Two Antennas

The relative phase of two antennas and the distance between them allow the
designer to set up the enhanced directionality of the overall power with
respect to the angular distribution of a single dipole antenna. This ability can
be useful when it is desired to “beam” a signal to a specific direction. The
electric field, expressed in Eq. (3.13), has a phase set by the wave vector, k,
the distance between the two quarter wave antennas, d, located at ±d/2 on the
x axis and the relative phase of the driving function for the two antennas, δ.
The resultant is a radiative solution, so that the electric field times radius, r,
has no r dependence

A contour of the field is shown in Figure 3.37 for the case when the script
“Two_Dipoles” is run with kd = 5 and zero relative phase. The shape is shown
in the surface plot in Figure 3.38.
A surface plot with a choice of kd = 5 and 3 radians of relative phase is
shown in Figure 3.39. The direction of the resultant radiation is indeed
changed very significantly, thus illustrating how one can beam a signal by
controlling the phase between the driving power of the two antennas.

Figure 3.37: Contour plot of the resultant electric field for two antennas with spacing, kd = 5, and no
relative phase. The pattern is distinctly not a simple dipole pattern.

Figure 3.38: Surface plot of the resultant electric field for two antennas with spacing, kd = 5, and no
relative phase.
Figure 3.39: Surface plot of the resultant electric field for two antennas with spacing, kd = 5, and a
relative phase of 3 radians.

3.18. Dispersion

Waves often encounter dissipative forces such as friction or viscosity. In those

cases the wave equation acquires an added term proportional to the first
partial derivative with respect to time, analogous to velocity dependent forces
in the study of ballistic motion or rocket drag.

The solutions, by separation of variable, are exponentials in space and

time. The imaginary part corresponds to a traveling wave with a velocity, v,
no longer c, but with a velocity which is frequency dependent. There is also a
real part corresponding to attenuation, which is also frequency dependent.
Note that, as δ approaches zero, the non-dispersive wave solutions, with no
attenuation and velocity of c, are recovered. The attenuation is μ which is zero
for δ of zero, as expected, and is approximately the imaginary part of the
propagation vector, k. The wave travels in x with velocity v and circular
frequency ω. Note that with larger δ values the velocity will decrease and the
attenuation will increase. Larger frequencies also mean increased attenuation.
 The solutions are explored in the script “Dispersion2”. The user chooses
both the constant δ and the circular frequency ω. A movie of the traveling
wave is produced and the first and last frames are superimposed in Figures
3.40 and 3.41. The velocity for the choice of parameters is only 58% of that in
the case without dispersion. Light would have travelled to x of five at the end
of the movie without dissipation.

Figure 3.40: Superimposed movie frames, first red, last blue of the travelling wave in the case of δ = 1
and ω = 0.1, showing minor attenuation.

The travelling waves are shown just to illustrate the attenuation and the
velocity changes that occur with a change of the parameter δ. The wave is
monochromatic. For a wave packet made up of waves of different
frequencies, the differing attenuation and velocity of the different frequency
components would distort the initial packet with time.

3.19. Waveguide
Waveguides are metallic structures in which radio frequency electromagnetic
radiation propagates. For rectangular hollow conductors of transverse size an
in x and b in y with waves traveling along z the electric and magnetic fields in
z are:

Figure 3.41: Superimposed movie frames, first red, last blue of the travelling wave in the case of δ = 1
and ω = 1. The velocity is the same as Figure 3.40, but the attenuation is increased.

Here it is assumed that the fields are on the inner surface and inside the
conductor. There is a cutoff velocity, cut below which no travelling wave can
exist. The modes are defined by the integers (n, m) one of which must be
greater than zero. The transverse electric mode has the lowest cutoff
frequency and it is the most common mode employed in waveguides. For m =
1 and n = 0 the lowest frequency possible is ω > πc/a.
In the lowest transverse electric mode the z and x electric fields are zero as
is the y magnetic field. The spatial dependence of the non-zero fields is:
 The transverse electric modes in a rectangular waveguide appear in the
script “Waveguide_TE”. The non-zero fields in this mode appear in Figure
3.42 for the users choice of aspect ratio, a/b, of one.
The lowest possible frequencies in the different modes appear in Figure
3.43. Basically, the waves must satisfy the boundary conditions and a
minimum frequency is needed if the wave is to oscillate from one wall to the
wall on the opposite side. Larger mode numbers mean that more oscillations
are needed which, in turn, means higher cutoff frequencies. In the figure a
specific choice of aspect ratio, a/b of one was made and the case was set by a
waveguide of one cm, which has a TE cutoff of 15 GHz.

Figure 3.42: Plots of the transverse, (x, y) dependence of the non-zero fields in the case of the lowest
(1,0) TE mode for a/b = 1.
Figure 3.43: Cutoff frequencies and contours in the TE mode for a specific, user chose, a/b aspect
ratio of the waveguide, in this case a/b of one. Modes up to (10,10) are plotted.

3.20. Skin Depth

Previously it had been assumed that the fields inside a “perfect” conductor do
not penetrate the conductor. That is true in the case of static fields for perfect
conductivity. However, an electromagnetic wave exists in a conductor
although it is damped strongly.
The conductivity σ relates the current density, J, and the electric field, E, in
the microscopic form of Ohm’s law. The wave in the conductor has a complex
wave vector, k, which means that there is an exponential penetration of the
wave into the conductor by a characteristic distance d which is proportional to
the inverse of the imaginary component of the wave vector k. The form for k
in Eq. (3.18) is closely related to the previous discussion of dispersion, with σ
playing the role of the parameter δ in Eq. (3.14)

Figure 3.44: Printout for a specific user input specifying the two important parameters, material and
frequency.
Figure 3.45: Skin depth for copper as a function of the frequency of the incident wave.

This effect is called the “skin depth” for the conductor. The script which
looks at this effect is called “Skin_Depth”. The user is invited to choose the
conductor and the frequency of the incident wave. The printout of the script
appears in Figure 3.44.
A plot for copper over a wide range of frequency is also provided as shown
in Figure 3.45. It is clear that high frequency waves populate only the “skin”
of the conductor with typical penetration distances in microns. This is why
waveguides for MHz waves and higher frequencies need not be made of thick
metal. On the other hand, 60 Hz home wiring has fields that penetrate
centimeter distances with resulting implications for the hardware.
 Chapter 4

Gases and Fluids

“Big whorls have little whorls, which feed on their velocity, and little whorls have lesser
whorls, and so on to viscosity.”
— Lewis Richardson
“Turbulence is the most important unsolved problem of classical physics.”
— Richard Feynman

4.1. Leaky Box

An interesting problem in statistical mechanics is the behavior of a container
of gas molecules with an aperture through which a gas molecule may escape.
As is intuitively obvious, faster molecules have a larger chance to escape, so
that the gas remaining in the container cools as time goes on. The energy flow
out of the box depends on temperature T, average velocity, 〈v〉 and the
aperture size A, so that the flow should be proportional to kT〈v〉A.
In three dimensions the number of molecules per magnitude of velocity
dn/dv goes as the square of v, so that the number of wall collisions goes as the
cube of v. The model in two dimensions is supplied in the script “Box_Leak”.
A two-dimensional Maxwell–Boltzmann energy distribution is a simple
exponential in the energy. Particles are picked out of this distribution, using
the Monte Carlo technique explored earlier, and then their momentum is
given a random direction. The initial positions are randomly assigned inside
the box.
A specific example appears in Figure 4.1. The user chooses the
temperature, the size of the box and the aperture. In that way one can confirm
that in this simple model, the total number of collisions with the walls
depends on the temperature and that the fraction escaping through the
aperture does also. The printout shows that the momentum of those molecules
exiting the aperture, is, on average, higher than those molecules that remain in
the box. A frame of the movie in a specific example appears in Figure 4.2. In
this case about one-third of the molecules escape during the one hundred time
steps over which the molecules are tracked. Other parameter choices should
be tried by the interested user.

Figure 4.1: Printout of a user choice for “Box_Leak”. The reduction in the mean momentum of the
molecules remaining in the box is clear.

4.2. Rectangular Flow

In looking at flow in a rectangular pipe the boundary conditions are that the
fluid has zero longitudinal velocity at the transverse boundaries. The flow is
driven by a pressure drop, ΔP of a fluid with viscosity η over a length L. The
Fourier coefficients for different modes defined by n and m require that m and
n are odd to satisfy the boundary conditions at the rectangular walls. The
Fourier coefficients are defined by Cm,n in Eq. (4.1) for a given choice of
modes and a given geometry.
The velocity is driven by the pressure drop, but falls with distance along
the pipe, set by L, due to the dissipative properties of the viscosity. The pipe is
square, size a on a side
Figure 4.2: Plot of the “gas” molecules at the end of 100 time steps with 100 initial molecules. The
aperture is in red and the escaped molecules are pinned to the center of the aperture. The mean
momentum for particles remaining in the box is reduced compared to the initial mean, while the mean
of the escaped particles is increased.

The behavior of specific modes is provided by the script “Flow_Rect2”. The

velocity contour for the lowest mode appears in Figure 4.3 while the velocity
profile for n = 1, m = 3 appears in Figure 4.4. The user can specify n and m
and explore the higher order modes in this fashion.
Figure 4.3: Velocity contour for the n = 1 and m = 1 mode. The velocity is monotonically decreasing
from the center of the pipe and vanishes on the boundaries.

4.3. Debye Temperature

Classically, the specific heat at constant volume, CV, of a solid is a constant,
in contradiction to the experimental data at low temperatures. The classical
prediction is about 25 Joules per degree Kelvin. Einstein first achieved a
partial solution by quantizing the “oscillators” that he assumed made up the
solid. However, the low temperature behavior of the resulting prediction did
not reproduce the data. At low temperatures the specific heat of solid
experimentally goes as the cube of the temperature.
Debye used Bose-Einstein statistics to treat the “phonons” or quantized
thermal oscillations of the atoms in the solid. In contrast to Einstein, Debye
allowed modes with several atoms participating rather that the Einstein
approach of independent atomic oscillators. The resulting theory has a
spectral energy density, du(E)/dE with a similar dependence as the black-body
Stefan-Boltzmann theory, du(E)/dE ∼ E3(eE/kT − 1) but has a parameter which
must be measured, the Debye temperature, TD. Having chosen it both the low
and high temperature behaviors are in agreement with experimental facts. The
expression for the specific heat in this case is found by differentiating u(E)
and is:
Figure 4.4: Velocity contour for n = 1 and m = 3. The plot peaks at the center of the pipe and vanishes
on the boundaries but has intermediate structure as opposed to the lowest (1, 1) mode.

Printout from the script “Debye” is shown in Figure 4.5. Both the Einstein
and Debye expressions for specific heat are computed and plotted. The
integral expression in the Debye case is evaluated numerically using the
MATLAB utility “integral” since the expression does not have a closed form
symbolic solution.

Figure 4.5: Printout from the script “Debye” with the user defined temperatures for the Einstein and
Debye specific heats.
Figure 4.6: Einstein and Debye specific heat as a function of the absolute temperature of the sample
expressed in thermal energy units.

The specific heat vs. temperature for the two models is shown in Figure
4.6. Note the rapid falloff of the specific heat in the Einstein case, while the
Debye theory displays the power law behavior of the experimental data in this
log-log plot.

4.4. Times Arrow

Entropy for a closed system always increases or stays the same. The earth is
not closed and we can violate this entropy increase because the sun supplies
us with energy. In the case considered here a closed box of gas molecules is
used to illustrate the principle of entropy increase. A “gas” is prepared to be
localized in the bottom half of a container in the script “Times_Arrow2”. The
user chooses the number of molecules, the number of time steps to track
them, the temperature of the molecules and the height of the container. Note
that the energies of the molecules are given a two-dimensional Maxwell-
Boltzmann distribution in energy using Monte Carlo methods. The initial box
for a specific example, 100 molecules, 100 time steps, and temperature of
0.01 and box height of one is shown in Figure 4.7.
The distribution of molecules after one hundred time steps is displayed in
Figure 4.8. The number in the top half is now 36, while 64 remain in the
bottom. The movie which is provided allows the user to follow the diffusion
of the molecules. The rate of diffusion is controlled by specifying the
temperature. It is quite clear that it is highly unlikely that the molecules will
later all once again appear in the bottom of the container. The entropy of the
final state of the system is clearly larger than that of the initial state, even
though the basic physics is time reversible. Indeed the laws of microscopic
physics obey time reversal. For example, a Keplerian orbit would be the same
if time ran “backwards”. It is in complex, many particle, systems that entropy
becomes evident.

Figure 4.7: Initial distribution of molecules confined to the lower half of a container.
Figure 4.8: Distribution of 100 molecules in a container after 100 time steps.

4.5. Compressibility
The compressibility of materials can be related to the fundamental structure of
the solid. A simple case is that of an ionic crystal where the positive ions have
a regular structure, with a onedimensional separation Ro. Overall ionic
neutrality is assumed and a simplified one-dimensional model is used. The
ionic cores repel strongly with an inverse power of separation with exponent
n. The electron-ion attraction falls off much more slowly, with an energy, E,
which goes as the inverse of the separation. The electromagnetic interaction
involves several ionic cores due to the long range of the interaction.

The Coulomb part has a sumable series, b = 2 ∗ ln(2). The minimum of the
energy, Eo, for a given ion occurs when E is minimized with respect to R:

The relationship of the crystal structure and the compressibility of the

material is explored in the script “Compress”. Printout from the script is
shown in Figure 4.9. Small radii result in a larger compressibility. A value of
exponent n = 10 is assumed in the script for the short range radial dependence
of the core repulsion.
Experimentally the compressibility, K, is a macroscopic quantity defined to
be the fractional change in volume per applied pressure at constant
temperature. It is also known as the bulk modulus. Typical values are
approximately 1011/Pa. At low temperature the change in internal energy dU
is −P dV, so that dP/dV is the second derivative of U with respect to V and V
goes as R3. The value of 1/K for the material is called the modulus and has
dimensions of pressure.

Figure 4.9: Printout from the script “Compress.” The user chooses an equilibrium radius in Å.
The value of Ro can be obtained by x-ray scattering patterns. The
compressibility can then be predicted for different n values. It is amusing that
a macroscopic measure of K allows for the measure on the microscopic core
repulsion power law n. The energy as a function of R is shown in Figure 4.10.
There is a clear minimum where the core repulsion is balanced by the
Coulomb attraction of the nearest neighbors, diluted by the next to nearest
neighbors.

Figure 4.10: Potential energy for an ionic crystal showing the core repulsion at small R and the
Coulomb attraction at large values of radius.
Figure 4.11: Contour of the modulus in Pa as a function of n and R. The modulus increases with n and
decreases with R.

The modulus, 1/K, is shown in Figure 4.11 as a function of the separation

of the crystal ions and the repulsive exponent. Experimentally the radius is
typically about 3 Å, while the exponent is about 10. The modulus, plotted
logarithmically, increases rapidly with decreasing radius and increases rapidly
with the value of the repulsive index. The magnitude of the modulus, 1/K, is
about 1011/Pa as expected.

4.6. Viscosity
Viscosity, η, is a measure of resistance to motion through a medium. In
ballistics, it leads to a terminal velocity when acceleration becomes zero. It
can be measured, for example, by measuring the terminal velocity and also be
several other means. It depends on temperature through the mean velocity of
the medium, 〈v〉, and it depends of the diameter, d, and mass, m, of the
objects which make up the medium. That means knowing the viscosity, the
size, diameter d, of the molecules in a gas can be indirectly measured

A typical value of viscosity in air is 18.1/μPa*sec, where one atmosphere at

STP is 0.1 Pa. The mean thermal velocity is assumed to be 3kT/2. Taking the
atmosphere to be all nitrogen N2, Å is 28. The mean velocity is then 0.51
km/sec. In this case, the molecules are inferred to have a diameter of 3.7 Å.
Again, a measured macroscopic quantity, the viscosity, allows for a
fundamental determination of the diameter of a molecule.
Printout from the script “Viscosity_Molecule” is given in Figure 4.12. The
user supplies the atomic weight of the air given the measured air viscosity.
The temperature is assumed to be STP. In Figure 4.13 the apparent size
dependence of the molecules is shown as a function of kT. The dependence is
weak, since d scales as the square root of T.

Figure 4.12: Printout from the script “Viscosity_Molecule” for two choices of Å made by the user.

Figure 4.13: Change in apparent molecular size with temperature. A slow square root dependence on T
is shown.
4.7. Water Waves
Water waves are explored for waves in deep water, where the relationship of
the frequency of the wave to the wave vector is or in shallow
water where and h is the water depth. The expression for the
incremental wave height z due to an initial disturbance, A(k), as a function of
r and t is:

The water waves are studied in the script “Water_Waves” which employs the
MATLAB functions “besselj” and “int” to perform the necessary integration.
The amplitude, A(k), of the initial wave is a sharply peaked Gaussian in k2,
which gives an approximate delta function in r at t equal to zero as an initial
condition. The dispersion of water waves is taken into account by the
relationship shown in Eq. (4.8), where the frequency approaches the deep
water result without dispersion for long wavelengths, kh much less than one.

In the script a sharply peaked Gaussian initial wave at small radius is

generated. The wave is then followed out in radius r and time t and a movie of
the motion is created. Results for shallow water height, no dispersion, as a
function of radius for the last movie frame are shown in Figure 4.14.
The script allows for the user defined choice for shallow water, with and
without dispersion and for deep water. The deep water result, the last movie
frame of wave height vs. radius, is shown in Figure 4.15. The wave velocity
in this case differs from the shallow water case and the wave is less peaked.
The mesh plot of the deep water wave, again last movie frame, is shown in
Figure 4.16. Dispersive results can also be generated by the user as desired.
Figure 4.14: Last frame of a movie for a shallow water wave, h = 0.1 m, without dispersion.

Figure 4.15: Wave height of water wave as a function of r for deep water; last frame of a movie. The
velocity is less than that of Figure 4.14 and the radial structure is more oscillatory.

4.8. Semiconductor
The use of semiconductors in the electronics industry is ubiquitous. A very
small introduction to the topic appears in the script “Semiconduct”. By
intrinsic, is meant a perfect crystal of silicon or germanium without crystal
defects and without any impurities. In this case there are, at a temperature of
zero, all valence electrons in a “valence band” of allowed energies and no
electrons in a “conduction band” of allowed energies the lowest energy of
which is at an energy Eg higher than the top valence band energy.
At elevated temperatures, some electrons receive sufficient thermal energy
to exist in the conduction band and to contribute to electrical conductivity. At
zero temperature such electrons are “frozen in” the valence band by the Fermi
exclusion principle since all available states in that band are filled. At low
temperatures the Fermi-Dirac statistical factor is approximately,
while the density of states factor for electrons arises from the statement that
all momentum components are equally probable, so that
is the density in momentum space, while in
energy space the density is proportional to

Figure 4.16: Mesh plot for a water wave in deep water. The frame corresponds to that of the previous
figure, Figure 4.15. There is no dispersion and the wave crests are quite sharp.

In the approximation that the Fermi level, EF, is at Eg/2 at low

temperature, the number of electrons per unit volume in the conduction band
is obtained by integrating the density of states over energy weighted by the
Fermi-Dirac statistical factor from Eg to infinity:

The number of holes remaining in the valence band, assuming the same
effective mass for holes in the crystal as for the electrons in the crystal is:
Using the values for the energy gap, Eg, in silicon, 1.1 eV, and germanium,
0.7 eV, at STP a conduction band density of charge carriers is expected to be
about 1.4 × 1016/m3 and 2.5 × 1019/m3 respectively. This density should be
contrasted with Avogadro’s number of about 1023. The product of the hole
and electron densities depends only on the temperature and the gap energy as
can be seen from Eqs. (4.9) and (4.10). This fact will be used later in
discussing a junction between two “doped” semiconductors.
It is clear that the addition of impurity states at the level of greater than
1018/m3, if these states are near the top of the valence band or the bottom of
the conduction band, but still in the gap would completely dominate the
intrinsic conductivity of the crystal. Indeed, this is how the n type and p type
semiconductors are created.
The script “Semiconduct” explores these concepts. Printout from the script
appears in Figure 4.17. The user chooses the gap energy, in this case that of
silicon. The addition of more electrons in the conduction band, due to added
impurity states changes the Fermi level, in this case increasing it, as can be
seen from Eq. (4.9). The equation shows that the Fermi energy depends
logarithmically on the density of charges. The change is small in most cases,
however, and a constant Fermi energy appropriate to T of zero can often be
assumed.

Figure 4.17: Printout from the script “Semiconduct” where the user has chosen a gap energy
appropriate to silicon. STP conditions are assumed.

The log of the number of intrinsic carriers is shown in Figure 4.18 as a

function of temperature. The strong dependence on the gap energy indicates
why germanium, with a lower gap energy than silicon, is not favored by
manufacturers because the intrinsic conductivity at high temperatures
competes with the doped impurity conductivity.
The dependence of the log of the intrinsic number density of electrons is
shown in Figure 4.19. The strong dependence on gap size is evident, as is the
strong temperature dependence. It is evident that it can be expected that
electronic devices will typically work better at reduced temperatures where
the doping effects will dominate.

Figure 4.18: Log of the electron number density as a function of temperature for silicon.

Figure 4.19: Log of the intrinsic number density of electrons for a semiconductor as a function of
temperature and gap energy. STP has a kT of 0.025.

4.9. Semiconductor Junction

The intrinsic semiconductors can be “doped” with impurity states with
energies near the bottom of the conduction band or the top of the valence
band but in the forbidden gap. These states then will define the mobile charge
carriers to be either electrons, n type, or holes, p type. A junction diode is a
device with both types in contact. There are ions which are not mobile and
there are mobile charge carriers. The Poisson equation for a region with only
positive ions and both positive and negative charged carriers is shown in Eq.
(4.11) which is appropriate to the n region of the junction

The electron charge is e, the charge density is ρ, ε is kεo, where k is twelve for
silicon, and εo is the vacuum permittivity. The voltage V exists can, in first
approximation be defined purely by the positive ions with density Po. Since
the Fermi energies on the n and p sides of the device differ, Eqs. (4.9) and
(4.10), a diffusion of charge carriers across the n/p boundary defines a dipole
layer of charge at the border which acts to equalize the Fermi energies. Both
the positive p and negative n charge carriers diffuse at the location x of zero,
where the n/p junction occurs and where they have equal densities no.
Starting with the approximation of ignoring the mobile charge carriers, the
field and potential in the case of full depletion, solving the Poisson equation,
when the field extends from the junction throughout the diode, of full width d,
is shown in Eq. (4.12). The coordinate x is zero at the junction. The field
extends throughout the size of the junction at full depletion. Equal doping
densities on both sides of the junction are assumed as are equal lengths of the
n and p sides.
The field is largest at the junction at x of zero and decreases linearly to zero at
the device boundary at +d/2 and −d/2 at full depletion. The potential is
quadratic in x and vanishes at the origin. The printout for the script
“n_p_Junction2” is shown in Figure 4.20 for a specific user set of choices.
The field at x of zero is printed out as is the voltage. The device voltage is
twice that.
The approximate voltage for the junction at depletion, where the mobile
carriers are swept out of the device, is shown in Figure 4.21. The voltage is
quadratic and vanishes at the junction. The fixed ionic charge density is also
shown schematically. The specific parameters chosen are a one hundred
micron device with an ion density of 5 × 1017 ions per cubic meter. The
density of donor and acceptor ions is assumed to be the same.
Taking the mobile charge carriers into account requies a numerical solution
to the full equation, Eq. (4.11). This is accomplished by using the MATLAB
ode45 utility. The user supplies the density of mobile carriers at the junction
boundary as well as the amount by which the maximum electric field is
reduced from the approximate full depletion value. That value defines the
equation and sets the initial conditions for ode45, along with V(0) = 0. The
actual density at the boundary is only estimated by the user.
Figure 4.20: Printout for a specific user chosen set of junction parameters.

Figure 4.21: The approximate voltage at depletion for the 100 micron junction. In this case the device
depletion voltage is about 2 × 1.9 = 3.8 volts. The left side is N type and the right side is P type —
referring to the fixed ions or p type and n type, referring to the mobile carriers.

Results for the electric field appear in Figure 4.22. Note that the field at
small x is linear as before, and an approximate proportionality of the field at x
of zero and the depletion region holds. In the example the depletion of mobile
carriers by the field is not complete but extends only over about half of the
full device. At large distances from the junction the density of mobile carriers
apporaches the donor density. Other sets of parameters can easily be explored.
The field and potential and n and p densities are calculated only for the
positive x or P type region. For negative x, symmetry can be invoked.
Figure 4.22: The electric field found numerically in the case where mobile carriers are taken into
account. The field near x = 0 is linear. The parameters for the example are given in Figure 4.20.

4.10. n-p Diode

The np junction has rectifying behavior when an external voltage is applied.

The diode is covered in the script “n_p_Diode” which looks at the response to
an applied voltage. The dialogue for that script in a specific example appears
in Figure 4.24.
The doping by donors and acceptors, P and N ions and n and p charge
carriers shifts the Fermi level, as mentioned previously. The Fermi levels for a
doping concentration given in Figure 4.24 are shown in Figure 4.25. Donors
put electrons in the conduction band and raise the Fermi energy. The opposite
is true for acceptors. The values follow from Eqs. (4.9) and (4.10) EFn = Eg +
kT log(Po/ne), EFn = −kT log(Po/ne). The diffusion of carriers at the junction
creates a dipole layer which then equalizes the Fermi levels, resulting in a
“built in” voltage Vbi which exists without an external applied voltage. The
value for Vbi is just the difference in Fermi levels in the n and p sides of the
diode.
Figure 4.23: The n and p charge densities found numerically in the case where mobile carriers in
thermal equilibrium are taken into account. The densities are equal at x = 0, but on the P side, x > 0, the
n are about 107 times larger than the p density. The p are minority carriers, and the n are majority
carriers in this region.

Figure 4.24: Output resulting from a dialogue with the script “n_p_Diode”. The user chooses a
junction thickness and doping density. An applied voltage is also chosen.
Figure 4.25: Fermi energies for the intrinsic case (dotted line) and for n and p type doping (dashed
line) for silicon and doping levels as given in Figure 4.24. The bottom of the valence band (VB) and the
top of the conduction and (CB) are indicated by solid lines.

The depletion region depends quadratically on the applied voltage. Away

from that region, it is assumed that all the donors are singly ionized so that the
n density is Po. On the other side the n density is reduced and is found by
assuming equal doping, where ni or ne is the intrinsic carrier
density. A simple interpolation in the depleted region results in Figure 4.26. In
this script, the contribution of the mobile carriers to the field and potential has
been ignored, while it was approximately taken account previously in the
script “n_p_Junction2”.
The diode acts as a rectifier. Almost independent of voltage there exists a
generated current, Ig, where holes thermally generated in the n region diffuse
into the p region. Although the dipole layer impedes diffusion there is also a
recombination current, Ir, with the opposite sign, where holes from the p
region flow to the n region and recombine. With no bias the two currents are
equal and there is no net current. For forward bias, with external voltage Vo,
the Fermi level is reduced for diffusing majority carriers to get through the
dipole layer at the junction boundary. They are then are swept up by device
and For reverse voltage applied, only the small reverse current,
−Ig, due to the other minority carriers exists since Ir is now reduced by the
exponential factor, not enhanced. The resulting idealized current versus
voltage relationship is:
Figure 4.26: Approximate density of n carriers for a 100 micron device with 2V externally applied.
The depletion depth for this example (Figure 4.24) is 77 microns.

The script asks for an applied voltage. It can treat forward values, V > 0, and
small values of reverse voltage if it is less than Vbi.

4.11. Freezing Pipes

The practical problem of how deep in the ground to lay in water pipes is
explored in the script “Pipe_Freeze”. There is an analytic solution to this
problem. The heat equation is shown in Eq. (4.13) below where T is
temperature, t is time and x is position. The basic parameter is a which has
dimensions length squared over time which is called the diffusion coefficient.
In this case a is about 2.4m2/year or 7.6 × 10−8 for a typical soil. A basic time
scale to transfer temperature by one meter is then 0.24 yr.
The initial condition is a temperature at all depths, by assumption, of To.
The boundary condition on the surface, x of zero, is a seasonal change of
temperature, ΔT from summer to winter, with t in years.

The printout of the initial dialogue for “Pipe_Freeze” is shown in Figure 4.27.
The temperature below ground is tracked for a number of years and the depth
for which the temperature just reaches zero is printed which is about 0.76
meters. The surface temperatures are assumed go from −15°C to 35°C
through the seasonal cycle.
The solution is shown in Eq. (4.15). In this case To is the mean surface
temperature of 10°C and the temperature swing is plus and minus 25°C. The
frequency ω is the seasonal driving frequency.

Figure 4.27: Printout of the dialogue in “Pipe_Freeze”.

Figure 4.28: Time dependence of T at a depth where the minimum temperature is just at freezing.
Note the time lag with respect to surface temperatures at this depth.

The result for a minimum temperature of just freezing is shown in Figure

4.28. That depth is evaluated using the MATLAB utility min for argument
(abs(T)). Note the time lag with respect to surface temperature as the heat
propagates into the ground.
Figure 4.29: Mesh plot of the temperature as a function of time and depth. The seasonal temperature
peak propagates in both depth and time, which causes a time lag with respect to the surface. At large
depths the mean surface temperature is approached.

The contour for temperature as a function of time and depth appears in

Figure 4.29. The temperature is followed over a two-year cycle. There is a
clear temperature delay with time as the changes in temperature propagate
underground.

4.12. Cooling Earth — Interior

The cooling of the Earth as heat diffuses into space is considered in the script
“Cooling_Earth_Int”. A series solution to the heat equation is evaluated with
thirty terms in the series shown in Eq. (4.16). The initial condition is that the
Earth’s temperature, TE, is uniform on the surface at r equal to R. The
temperature difference between the Earth and space, Ts, is taken to be three
hundred degrees absolute at time zero. There is a characteristic time, τ, which
depends on the radius of the Earth and the diffusion coefficient. For
reasonable values of the coefficient a the time is one hundred and thirty
billion years, even neglecting internal heating as is done here
The solution is a series in radius and time factors. The radial terms are
spherical Bessel functions, using the MATLAB utility “besselj”, of argument
(nπr/R) as is expected in the spherical geometry. The temporal terms are
exponentials with arguments (n2t/τ), so that the series converges rapidly in
time.
The results for thirty terms in the series are plotted below. In Figure 4.30
the last frame of a movie shows the temperature in the interior. The surface
has cooled to zero absolute but the central interior is still above one hundred
degrees absolute. A mesh plot of the temperature as a function of time and
depth appears in Figure 4.31, showing rapid surface cooling.

Figure 4.30: Temperature distribution, absolute scale, inside the Earth after 200 billion years.
Figure 4.31: Temperature distribution inside the Earth as a function of radius and time. Initially the
Earth is at a uniform temperature. As time goes on, the surface temperature falls rapidly, while the core
temperature is maintained for a long time.

4.13. Cooling Earth — Exterior

The exterior temperature as the Earth cools is formally very similar to the
previously quoted interior solution. The Earth is taken to be a uniform sphere
and the time scale is the same as before. The initial conditions are that the
temperature at the surface is TE at time zero and that at late times the
temperature at radius R approaches Ts. The solution is again a series of
spherical Bessel function terms spatially, and exponential decaying factors in
time, just as before.
The results for the exterior, with the same parameter values as taken for the
interior solution shown previously appear in Figure 4.32. The mesh plot is for
a series with thirty terms and with radial values greater than R. At radius of R
there is a smooth decrease in temperature as there was in the interior case. At
values of r greater than R, the temperature falls, with structure due to the
Bessel functions. Only the very first few terms are important because the
exponentially decreasing time factors damp out the higher order terms.
Figure 4.32: Surface of exterior temperatures for a cooling earth. The radii are greater than the earth’s
radius R. The characteristic time scale is 130 billion years as before.

4.14. Heat and pde

The script “Heat_Diffuse_pde” uses the MATLAB one-dimensional partial

differential equation solver, “pdepe” to address the problem of tracking the
temperature over an x region, (−L, L) in reaction to a pulse of heat initially in
the x region (−ab, ab). There is a menu driven choice of different functions
for the initial pulse as shown in Figure 4.33. A movie is shown of the time
development of the temperature. The script supplies functions defining the
heat equation (source less), the initial conditions and the boundary conditions.

Figure 4.33: Menu of initial temperature functions (left). Dialogue for the “Heat_Diffuse_pde” script
for a specific set of user choices (right).
Figure 4.34: Temperature as a function of x and t for an initially uniform distribution.

The temperature response as a function of x and t to a constant temperature

initial pulse is shown in Figure 4.34, while the response to an initial
temperature which is quadratic in x is displayed in Figure 4.35. In both cases
distinct shapes in position for the temperature distribution quite rapidly
smooth out and the overall temperature scale dies off in time. All the possible
shapes available to the user have the same general behavior.

Figure 4.35: Temperature as a function of x and t for an initial distribution quadratic in x.

4.15. Heat Diffusion

The script “Heat_Cond2” uses the MATLAB solver for a partial differential
equation much as in the “Heat_Diffuse_pde” script and describing a bar of
conductor of length L. However the initial conditions in this case are that the
bar temperature is zero, while the boundary conditions at x near L is that the
end heats very rapidly to TL, while the temperature at x of zero is held at zero.
The problem approximates the heating of a rod as heat flows into the rod
which has fixed temperatures at both ends of the rod.
The printout for “Heat_Cond2” appears in Figure 4.36, where the user
chooses the length of the copper rod and TL. The TL heating time is small with
respect to the time to relax to equilibrium. The characteristic relaxation time
is 1/a(π/L)2, in this case 250,000 seconds. A movie is provided of the time
development of the temperature. The printout defines the initial conditions,
IC, and boundary conditions, BC.
A mesh plot for the parameter choices made in the dialogue, Figure 4.36 as
a function of x and t is shown below in Figure 4.37. As can be seen the ends
of the rod are fixed at five hundred degrees and zero degrees. As time
increases, the temperature along the rod approaches the equilibrium solution
of a linear temperature distribution.

Figure 4.36: Printout of the dialogue for a set of specific parameter choices for the script
“Heat_Cond2.
Figure 4.37: Mesh plot of the temperature as a function of position along the rod and time since initial
heating of one end of the rod. The temperature along the rod near x = 0 increases until a linear
temperature dependence on x is obtained.

4.16. Heat — Initial

In this script, “Heat_IC”, the heat equation is solved for the case of a
symbolic, user supplied, initial temperature distribution in x from x of zero to
x of L. The solution, T(x, t) is by separation of variables and series solution.
The assumed boundary conditions are that the partial derivative of T as a
function of x for all t are zero at x of zero and L. No temperature flows out of
the defined region. The series has Cartesian Fourier spatial terms and
exponential temporal terms as seen in Eq. (4.17). The characteristic time for
the problem is (L2/απ2). The A coefficients are solved for symbolically using
MATLAB “int”.

A movie of the x dependence of the temperature as a function of t is provided

by the script. Printout for a specific choice of initial condition is shown in
Figure 4.38. The user chooses the length of the rod and the initial symbolic
temperature distribution. The number of terms of the series is also a user
choice so that the convergence of the series can be explored. The mesh plot of
temperature as a function of x and t is shown in Figure 4.39. The initial
distribution in x is smoothed and reduced as time increases. However, the
solution at large t is a constant because the initial distribution had a non-zero
average value, and heat does not flow out of the rod by construction.

Figure 4.38: Printout provided by “Heat_IC”. The initial temperature distribution is symbolic and
provided by the user as a function of x.

Figure 4.39: Mesh plot for temperature as a function of x and t for a user defined initial distribution,
T(x, 0) = |x − 1/2|2. Note that the equilibrium temperature is non-zero since the initial temperature has a
non-zero average value and no heat flows out of the rod.
 Chapter 5

Waves

“If you want to find the secrets of the universe, think in terms of energy, frequency and
vibration.”
— Nikola Tesla
“Sit in reverie and watch the changing color of the waves that break upon the idle seashore of
the mind.”
— Henry Wadsworth Longfellow

5.1. Longitudinal Slinky

It is useful to consider longitudinal waves in a continuous medium first as
oscillations in a finite discrete set of springs and then go to the continuum
limit. Such waves are well known in the “Slinky” toy. The problem is
approached in a simplified case of seven springs connecting eight mass points
in the script “Slinky” which solves for the motion numerically using the
MATLAB script “ode45”.
The motion of a mass point with coordinate δi due to nearest neighbor
displacements is governed by the equation:

where m is the point mass, k the spring constant and a is the equilibrium
separation of the mass points. It is amusing to notice that the grid method of
solving partial differential equations uses the spatial part of the second order
partial spatial derivative. In the present case the aim is to go to the continuum
limit, but actual numerical solutions go the other way on a discrete grid.
The longitudinal motion of the mass points is displayed as a “movie”. Only
the fourth mass point is displaced initially and all initial velocities are zero.
The equilibrium separation, a, is defined to be equal to one. The user supplies
the ratio k/m and that defines the characteristic time scale, The
motion of the springs is followed for five such times. The user also supplies
an initial displacement of spring #4. In the example shown it is 0.01. The last
frame of that specific movie is shown in Figure 5.1 where the displacements
are magnified by a factor fifty in order to make them more visible. The movie
makes the outgoing left and right waves very clear.

Figure 5.1: Last frame of a movie with initial displacement of the fourth mass point by 0.01a.

Going to a continuum limit, the coordinates are the fractional

displacement, η → (δi+1 − δi)/a. The Young’s modulus Y → ka and the density
ρ → m/a. The wave equation in this case follows plausibly from Eq. (5.1) for
the discrete case.

As an example, the speed of sound in the atmosphere can be estimated using

the wave equation. Take the ratio of specific heats, appropriate to adiabatic
motion where there is not sufficient time to carry away the heat generated, γ =
CP/CV = 5/3. The adiabatic approximation assumes dP/P = −γ(dV/V) which
relates the fractional change in pressure to the fractional change in volume.
The atmosphere at STP has a pressure of 0.101 MPa and a density of 1.29
kg/m3. The velocity, can be read off from the wave equation, Eq.
(5.2) since the adiabatic Young’s modulus is γP with units of pressure.
Numerically the result for the speed of sound is 360 m/sec.
Plots of the displacements of individual mass points with time are shown
in Figures 5.2 and 5.3. A longitudinal wave has clearly passed from one mass
point to the next. The wave arrives first at mass #2, elongates it and then
compresses it while the elongation appears then at #1, followed by
compression. The wave velocity can be taken for the two figures. In this
specific case the wave moves by a distance a of one in about one
characteristic time unit.

Figure 5.2: Plot of the position of mass point #2. Initially it does not move and then the longitudinal
wave reaches it and it is elongated, then compressed. Note that the displacement is greatly magnified.

Figure 5.3: Plot of the position of mass point #1. Initially it does not move and then the longitudinal
wave reaches the mass, after the wave has hit mass #2. There is first elongation followed by
compression. Note that the displacement is greatly magnified.
5.2. Transverse Slinky
The transverse oscillation modes for a short “slinky” are studied in the script
“Transverse_Slinky”. This script uses the MATLAB function “eig” to find the
eigenvalues and eigenfunctions for N = 2, 3 mass points connected by springs
with the end mass springs fixed to rigid walls and with k/m = 1 for the
springs. The eigenvalues for an N of two and three are shown in Figure 5.4,
while the eigenfunctions are graphically displayed in Figure 5.5 for N = 2 and
Figure 5.6 for N = 3.
In the case of N equal to two, the eigenfunctions are transverse motion in
the same or opposite directions. In the case of N equal to three, the central
spring is stationary for the lowest state and the two outer springs have
opposite displacements. For the next lowest eigenvalue, the two outer springs
have the same displacement, while the middle mass has the opposite
displacement. With the largest eigenvalue, all three mass points have the same
sense of displacement.

Figure 5.4: Eigenvalues, D, for N = 2 (left) and N = 3(right) mass points.

Figure 5.5: Eigenfunctions for N = 2 mass points. The end points are fixed by springs attached to rigid
walls. The lowest eigenvalue is shown as a dotted line.

5.3. SR — Doppler
The Doppler effect is well known in everyday life. However, there are
additional changes to that picture as the source velocity approaches that of
light. The ratio of the frequencies in two inertial frames moving with a
relative velocity v = β/c is:

Figure 5.6: Eigenfunctions for N = 3 mass points. The end points are fixed by springs attached to rigid
walls. The eigenfunction for the lowest eigenvalue is dashed, followed by dashed and solid for the
higher eigenvalues.

In the classical theory there is no shift in frequency at 90 degrees to the

direction of motion. In the relativistic case there is a shift, as illustrated in
Figure 5.7 which is output of the script “SR_Doppler”. This transverse
Doppler shift is due to the time dilation factor between the two frames. The
longitudinal shift is while the transverse shift is ω*/ω
= 1/γ. The emission angles of the two frames are related by, tan(θ*) = sinθ/
γ(cosθ − β). The result is easily derived using the fact that the wave vector k
and the frequency, ω/c are the components of a four-vector which transforms
using the Lorentz transformation.
A contour plot of the ratio of the frequencies as a function of β and the
emission angle of the light in the ∗ frame is displayed in Figure 5.8. If the
source is receding, the light is red shifted to lower frequencies while a source
moving toward the frame is blue shifted. The velocity is a signed quantity as
is the emission angle since emission can be toward the direction of relative
motion or away from it. The emission is symmetric if the sign of the velocity
and the direction of the light emitted are both changed.
Figure 5.7: Ratio of frequencies of the light observed in 2 inertial frames moving with a velocity
factor for the case where the factor is 2 as a function of the angle of the light
with respect to the axis of relative motion.

5.4. Step Response

A damped oscillator has a complicated response to a step driver which starts
at time zero. The differential equation has a damping term, a, a natural
frequency in the undamped case, b, and a driving term c.
Figure 5.8: Contour of the ratio of frequencies in 2 inertial frames for different relative velocities and
for different emission angles in the ∗ frame.

The problem is solved symbolically using the MATLAB utility “dsolve” in

the script “Step_Transient”. The solution is reproduced in Figure 5.9. At long
times after the transients die out the amplitude is a constant. Examining the
differential equation, in this case y → c/b2 is expected.
The amplitude for a specific, user defined, example is shown in Figure
5.10. There is initial transient behavior which is damped out on a time scale
set by the scale of the natural frequency τ ∼ 1/ω ∼ 1/b. The solution
asymptotically approaches a constant value, as expected. A constant driving
term, c, is assumed, but other functional forms could be tried, such as a
harmonic driver.

5.5. Pulsating Sphere

An ideal, incompressible fluid will propagate waves driven by a harmonically
varying object, for example water is a good transmitter of sound waves. A
specific case is considered in the script “Pulsating_Sphere”. The radial
velocity of the sphere is harmonic with a circular frequency ω. The resulting
velocity of the perfect fluid as the wave propagates and dissipates is:
Figure 5.9: Printout of the script “Step_Transient” showing the symbolic solution, y, when the initial
position and velocity of the oscillator are zero.

Figure 5.10: Amplitude for the damped, driven oscillator in the specific case of a = 0.2, b = c = 1.
Figure 5.11: Radial velocity of the wave in the fluid for a driving frequency of 3. The sphere is of
radius 1 and the velocity is zero for r < 1.

The amplitude of pulsation, vo, is taken to be one as is the radius a of the

sphere and the sound velocity in the fluid, c. A movie of the radial
dependence of the velocity as a function of time is plotted by the script. A
frame of that movie in the case of a driving frequency of ω equal to three, a
particular user choice, appears in Figure 5.11. Note the damping with radius
as the wave spreads as indicated by the inverse square of the radius behavior
shown in Eq. (5.5). The frequency of the wave is the driving frequency in the
case of a perfect fluid, without viscosity or other dissipation.

5.6. Rectangular Drum

The rectangular drum is constrained to have a vanishing displacement on the
boundaries at x of zero and a and in y at zero and b. The frequencies thus are
set for the modes n and m in the spatial solutions which are printed out in the
dialogue with the script. Temporally, the choice was made that the tension and
density of the drum were both defined to be one, so the drum wave velocity is
one from the wave equation… . The user defines the extent in b and the
modes m and n. The dialogue for the script “Rectangular_Drum” appears in
Figure 5.12.
The plotted output is a movie of the motion of the drum in time displayed
at the displacement z as a function of x and y as time progresses. A frame of
that movie is shown in Figure 5.13. The user is encouraged to vary b, m and n
in order to see the changes induced in the drum by watching the resulting
movie.

Figure 5.12: Printout of an example of the dialogue from the script “Rectangular Drum”.

Figure 5.13: Solutions to the 2-dimensional wave equation in rectangular coordinates and with
rectangular boundary conditions imposed. The choices made are: size is 1 in x and 2 in y. The modes are
2 in x and 3 in y. There are 2 extreme values in x and three in y.

5.7. Piano
The script “Piano” approximates the motion of a piano string which is struck
by an impulsive force, the piano hammer, at location axo along the string. The
string is held at x = 0 and x = L where L is taken to be one in the script. The
velocity of waves on the string is c, taken here to be equal to one. The
solution for the displacement of the string, y(x, t) depends on the location of
the hammer and the time t. The impulse is approximated by a function which
is large only for times < π/q where q is chosen to be very large. The time t is
zero at the end of the impulse and the string position, y, is plotted only for t
greater than zero. The series solution for the piano string displacement
appears in Eq. (5.6). The boundary conditions are satisfied by the sin terms in
the series, while the initial quasiimpulse is set by the cos term.
A movie is produced after the initial impulse position is chosen and the
number of terms in the series is defined. The last movie frame in the cases of
xo = 0.2 and ten and fifty terms respectively are plotted in Figures 5.14 and
5.15. Note that the waveform is much sharper in the latter case since the
Fourier series is more accurate because of the increased number of terms. No
dissipation has been assumed in this script.
The movie for the time development of the string exhibits the travelling
wave behavior both to increasing x and decreasing z and the reflections off the
ends of the string where it is held to zero displacement.

5.8. Aperture Diffraction Off Axis

Diffraction by an aperture has two distinct modalities. For wave vector k
times aperture size a, when ka is much greater than one the regime of
geometric optics is operative where light appears to go in straight lines. For
ka small with respect to one the diffraction of light by any aperture becomes
very evident.

Figure 5.14: Initial amplitude of the piano string using 10 terms in the series representation and initial
x = 0.2.

For a circular aperture of radius a and angle of incidence with respect to

the normal of plate containing the aperture of angle ψ, the angular distribution
of the power, dP/dΩ downstream of the aperture is:

The azimuthal angle, φ, is the angle of the electric field of the incident wave
with respect to the outgoing wave vector x axis and is taken to be zero. The
angle θ is the spherical polar angle of the wave vector downstream of the
aperture. The MATLAB utility “besselj” is used to evaluate the numerical
values of the radiated power.

Figure 5.15: Initial amplitude of the piano string using 50 terms in the series representation and initial
x = 0.2.

The user can check on expectations in regards to geometrical optics by

entering in the script “Diffract_Angle” a zero angle and a large value, say 20,
for ka. There is an azimuthally independent small angle peak within the
diffractive cone of angle approximately 1/ka. The situation for ka = 1 and
normal incidence is shown in Figure 5.16. There is, indeed, a forward
diffractive peak, but also azimuthal dependence. The azimuth of the incident
electric field is important and the power peaks at that azimuth.
For ka = 1 and an incident angle of forty-five degree, the pattern of
radiated power is shifted toward larger angles with respect to normal
incidence (Figure 5.17). There remains a forward peak, but the overall pattern
is shifted toward larger polar angles with a modified azimuthal dependence.
The user can select both ka and the incident angle and thus explore the
parameter space.

Figure 5.16: Radiated power for the case of ka = 1 and zero incident angle. There is a forward
diffractive peak in θ and a dependence on φ.

Figure 5.17: Radiated power for the case of ka = 1 and a 45 degree incident angle. The peak is still
forward but wider angles are now more probable.

5.9. Antenna
The power radiated by a center fed linear antenna is displayed using the script
“Antenna”. The antenna length is L; the antenna resistance is Ro; and the crest
current at the antenna midpoint is Io. The length L is in units of the driving
current wavelength. Printout of the script appears in Figure 5.18. The printout
displays the angular distribution of the radiated power for a user defined
choice of wavelength.
The angular distribution of the radiated power in the case where kL = 0.1 is
shown in Figure 5.19. The pattern goes as the square of the sin function as
expected for dipole radiation. The total radiated power is shown in Figure
5.20 as a function of kL. A closed form is not available in this case, so that the
MATLAB numerical utility “quad” is used to integrate the angular
distribution. The power increases rapidly with kL. For kL < 1 the approximate
results for the angular distribution and the total power are quoted in Eq. (5.8).
The total radiated power goes as the square of the antenna length, with a scale
set by the wavelength of the radiated waves.

Figure 5.18: Printout of the script “Antenna”.

Figure 5.19: Angular distribution of the radiated antenna power for kL = 0.1.

Figure 5.20: Radiated antenna power as a function of kL. The semilog plot in y indicates the strong L
dependence.

5.10. Antenna Array

Simple dipole antennas are quite isotropic in their radiation pattern as seen
above for small values of kL. The use of two antennas with controlled phase
between them was already explored in Section 3.17. In that case, substantial
directionality could be obtained. However, often one wishes to more tightly
“beam” a radio or other signal to a specific angular range. This can be
accomplished by having different waves from different parts of an antenna
array interfere with one another. In the specific case of an array of N quarter
wave, λ/4, antennas separated by spacing a and with relative phases φo, the
resulting intensity, I, is shown in Eq. (5.9).
The full array pattern can be thought of as N slit diffraction due to the
contributions of the N line sources. There is a maximum of I at the sin of the
angle equal to 2(1/N − φo/π). The pattern depends on N, the total number of
antennas, and the relative phase of the antennas, φo.

The script “Antenna_Array” allows the user to choose both the number of
members of the array, and the phase of the elements. Radiated power patterns
are shown in Figures 5.21 and 5.22. It is obvious that with a large number of
elements in the array, very tightly beamed power can be obtained. By varying
N, the user can watch the change in the power angular distribution. Clearly,
this beaming technique has important practical implications.

5.11. Lissajous

A particle may be confined in two dimensions with different confining

potentials in the two directions. In that case there can be simple harmonic
motion in each dimension, but with different amplitudes, phases and circular
frequencies. It can then happen that the trajectories in the (x, y) plane can
become rather complex. The x and y trajectories differ in frequency, ωx, ωy,
phase φ and the amplitude ratio which is defined to be a.
Figure 5.21: Intensity for N = 2 and phase equal 0. For N = 1 the power is emitted isotropically in this
approximation.

Figure 5.22: Intensity for N = 100 and phase equal 0.

The solutions are shown in the script “Lissajous” as movies where the user
chooses the parameters for the x and y harmonic oscillations. The patterns
may be re-entrant, but need not be. A plot of (x, y) for a specific case is the
last frame of the movie generated with an amplitude ratio of two, a phase
difference of π/2 and a frequency ratio of four is shown in Figure 5.23. Such
plots are familiar to devotees of old science fiction movies where
oscilloscopes are driven in (x, y) to make amusing “laboratory” displays.

Figure 5.23: Plot of (x, y) for simple harmonic motion in 2 dimensions with differing amplitudes (2×),
phases (π/2) and frequencies (4×).

5.12. Plane Waves — 2d

Two waves propagating on an infinite membrane display some interesting

interference effects. A simple case is that of two waves of equal amplitudes,
frequencies, ω, and wave vectors of the same magnitude, k, with the only
difference being that the angle between the wave vectors is finite. The x axis
is chosen as the bisector of the angle between the two wave vectors. The
superimposed amplitude is easily shown to be, with kx = k cos(θ), ky = k sinθ
and θ the angle between the two waves.

The superimposed waves travel in a new, x, direction with a modulated

amplitude in the y direction and with an increased wave velocity ω/kx. The
script “Plane_Wave_2d” is used to choose the angle between the two
individual waves and the resultant movie is shown, the last frame of which
appears in Figure 5.24. The nodal lines for the standing wave which exists in
the y direction are:
Figure 5.24: Superimposed waves in two dimensions for a 20 degree angle between the 2 constituent
waves. There are two nodal lines in y which are evident.

5.13. Damped, Driven Oscillator

A general second order inhomogeneous differential equation appears in Eq.
(5.13). It can be simplified by expressing time in units of the undriven and
undamped circular frequency ωo, or τ. When the natural frequency is defined
to be one, there remain three parameters defining the equation, a damping
factor b, a driving amplitude C and the ratio of the driving frequency to the
natural frequency k.

The equation is explored using the script “Damped_Forced_SHO”. The

solutions are created at four levels of complexity using the MSATLAB
symbolic solutions “dsolve” with initial conditions the y equals one and dy/dt
equals zero at t of zero. First the case of no driving force and no damping
results in purely harmonic motion. The specific waves associated with Figure
5.25 appear in Figure 5.26. The second condition is for damping but no
driving force, while the third has no damping but a driving force. Finally, the
general case is considered, with the parameters b, C and k supplied by the
user.
The simplest case is just harmonic. The case with damping can be
understood by assuming a harmonic solution to Eq. (5.13) and then solving
for the frequency ωb. The frequency is shifted from that for the free and
undamped case and acquires an imaginary part illustrating the damping of the
wave, Eq. (5.14).

Figure 5.25: Results of the dialogue for the script “Damped_Forced_SHO”. There are 4 cases; no
damping, no driver, damping, no driver, driver no damping and finally damping with a driving force.
Figure 5.26: Wave forms for the 4 specific cases given in Figure 5.25.

For the driven case, after the transients have died down, the amplitude y
follows the driving force amplitude and frequency. In the absence of damping
the amplitude can become very large if the driving frequency is near the
natural frequency (resonance), while damping limits the amplitude, as seen in
Figure 5.26.
 Chapter 6

Quantum Mechanics

“Quantum physics thus reveals a basic oneness of the universe.”

— Erwin Schrödinger
“If quantum mechanics hasn’t profoundly shocked you, you haven’t understood it yet.”
— Niels Bohr

6.1. Box
The script “QM_Box” uses the MATLAB symbolic mathematics utilities to
evaluate expectation values for a particle in a onedimensional box in the
ground state or n = 2, 3 or 4. The box boundaries are (−a/2, a/2). The
expectation values for x, momentum p, x2, and p2 are calculated symbolically.
The energy is evaluated and the uncertainty of the ground state, defined in Eq.
(6.1), is explicitly evaluated.

Printout from the script appears in Figure 6.1. The energy is printed along
with the wave function and normalization. Additional printout appears in
Figure 6.2, where the expectation values of position and momentum are
computed for the ground state, n = 1. Other states are available for the user to
choose from. In the ground state the basic Heisenberg uncertainty relation is
confirmed:

6.2. Simple Harmonic Oscillator

The simple harmonic oscillator is easily solvable classically. The potential
energy is V = kx2/2 or Mω2x2/2. At the turning points the velocity is zero, so
the turning points are at ±E/k, where E is the total energy. The classical
velocity as a function of x is then:
Figure 6.1: Printout from the script “QM_Box”. All quantities are calculated symbolically and then
evaluated numerically.

The probability to observe the classical particle goes as 1/v, which means it
is mostly near the turning points. Quantum mechanically, the solutions are
Hermite polynomials, which are available using the MATLAB Maple
functions, mfun(‘H’, n, x).
Figure 6.2: Additional printout from the script “QM_Box” for the case of the ground state, n = 1. The
uncertainty principle is explicitly confirmed.

The solutions are:

The correspondence between the classical and quantum mechanical

solutions is developed in the script “QM_SHO_Hermite”. The user chooses a
quantum number n and the two solutions, classical and quantum, are plotted.
Results for n = 0, the ground state, and n = 20, a highly energetic state, are
shown in Figures 6.3 and 6.4, respectively. The solutions are radically
different at low values of n but approach one another at high values of n. This
behavior is an example of the “correspondence principle” which argues that
quantum solutions approach classical ones in the limit of large quantum
numbers.
Figure 6.3: Plots of probability for a classical simple harmonic oscillator and for the n = 0 quantum
case. The wave function is outside the classical turning points at x of + and −1.

The quantum solution is not limited to be at x values less than the turning
points. Indeed looking at the Schrödinger equation and ignoring the energy E,
the solution is clearly the Gaussian appearing in Eq. (6.4), d2ψ/dx2 ~
2m(kx2)ψ/2. The Hermite polynomials are corrections to the basic Gaussian
behavior of the wave functions.

Figure 6.4: Plots of probability for a classical simple harmonic oscillator and for the n = 20 quantum
case.
6.3. Bound States — 3d Well
The bound state of a three-dimensional square well of radial size a can be
solved, but not in closed form in analogy to the onedimensional square well
of finite depth. The potential is equal to Vo for r less than a and zero
otherwise. The solutions are spherical Bessel functions inside and Hankle
functions outside the well which can be accessed using the MATLAB
functions “besselj” and “besselh”. For the ground state the radial wave
function R, is not used but rather the function u:

The solution for the ground state is found using the script QM_Square_3d.
The script applies the MATLAB utility “fminsearch” to evaluate the energy E,
less than zero for a bound state, that satisfies the condition that arises when
requiring the wave function and derivative to match at the boundary at r = a.
The starting value given to “fminsearch” is taken to be the solution for an
infinite square well

Printout from the script is shown in Figure 6.5. Several solutions for a
value of a of 1A are shown. For a well depth of 10 eV the state is barely
bound, while for a 24 eV well depth the binding energy is 6.46 eV.

Figure 6.5: Printout for some specific choices well depth and extent for the ground state.

The radial wave function rR for the barely bound case is shown in Figure
6.6. The wave function falloff outside the well is very slow since the barely
bound state can easily “leak” outside the well. More deeply bound states are
more localized to be in or near the well.
To make the comparison concrete, the deeply bound example of a one
Angstrom well 24 eV deep is shown in Figure 6.7. In this case the falloff of
the wavefunction outside the well is much faster than in the previous plot, as
expected.

Figure 6.6: Wavefunction for the ground state of a 3d square well when the state is barely bound.

Figure 6.7: Wavefunction for the ground state of a 3d square well when the state is bound by 6.47 eV.
6.4. Identical Particles
In quantum mechanics identical particles obey restrictions on the wave
functions describing their states. A simple example is to consider two
particles confined to a “box” in one dimension. The solutions are sin and cos
functions which vanish at the boundaries of the box. Independent particles
would be described by product wave functions since the joint probability in
such a case is simply the product of the individual probabilities.
However, in quantum mechanics identical particle wave functions must be
either symmetric or antisymmetric under the interchange of a pair of particles.

The Greek subscripts indicate the state and the labels 1 and 2 identify the
particle in that state. There are two possibilities; bosons have the plus sign for
symmetric wave functions under particle interchange while fermions are anti-
symmetric.
The wave function contours for several cases are explored using the script
“QM_Identical”. The user chooses the states to be explored for the box and a
plot is made of the resulting wave function contour. The contour for both
particles in the ground state is not shown since it vanishes for the
antisymmetric case which is the Fermi exclusion principle. For the n = 1 and
n = 2 states of the box, the product contour for classical considerations
appears in Figure 6.8. In this case particle 1 and particle 2 are uncorrelated. In
contrast the contour for the antisymmetric case is shown in Figure 6.9. In this
case there is a sort of “repulsion” when the coordinates are equal. Finally, for
the symmetric case, n = 1 and n = 3 states are plotted in Figure 6.10. Clearly
the symmetries imposed on the states of quantum identical particles have
profound implications for the physics of a system of such particles.
Figure 6.8: Contour of the system wave function for n = 1 and n = 2 for a product wave function. The
axes are uncorrelated.

6.5. Stark Effect

The Stark effect occurs when an external electric field is applied to an atom.
In the case of the hydrogen atom in the ground state the first order effect in
perturbation theory vanishes. That is because the ground state has no angular
momentum while the perturbation defines a vector direction in space, so that
the matrix element evaluated in the ground state vanishes. The atomic ground
state does however have an induced dipole moment in second order where
matrix elements between the ground state and other states contribute.
The effect is first order for degenerate states, the lowest energy being the n
= 2 hydrogenic state. The perturbation mixes the states with l = 0 and l = 1
because the operator has vector properties. Printout for this case appears in
Figure 6.11. The mixing matrix of the perturbation is A in the printout. The
MATLAB utility “eig” is used to find the eigenfunctions V and eigenvalues
with the diagonal matrix D of the perturbation.
Figure 6.9: Contour of the system wave function for n = 1 and n = 2 for an antisymmetric wave
function. The wave function is highly correlated and is the modulus is small for equal coordinate values.

More printout of the script appears in Figure 6.12. The MATLAB solutions
for the eigenvectors, V, and eigenvalues, D, are shown.
A plot of the eigenvectors is shown in Figure 6.13. There is an isotropic
part with l = 0 and a part with angular dependence. The two perturbed
solutions are not symmetric about ninety degrees as the unperturbed
eigenfunctions are. That asymmetry means that the electrons have a centroid
away from the origin and the displaced charge means that there is an induced
dipole moment which is first order in this case. The alignment of the electrons
can be either parallel or antiparallel to the external electric field.
Figure 6.10: Contour of the system wave function for n = 1 and n = 3 for a symmetric wave function.
The wave function is correlated, being large when the coordinates are similar.

6.6. Square Well Scattering — 3d

The Born approximation to a scattering amplitude consists in approximating
the incoming and outgoing wave functions by plane waves. In that case, the
scattering amplitude, f, is essentially the Fourier transform of the interaction
potential with respect to the magnitude of the momentum transfer between the
initial and final states, q. The explicit result for a plane wave vector k,
scattering by an angle θ appears in Eq. (6.8).
Figure 6.11: Printout from the script “QM_Stark”. There are two eigenvectors and two diagonal
eigenfunctions for the n = 2 mixing of the l = 0 and l = 1 states caused by the perturbation matrix A of
the interaction Hamiltonian HI. The Bohr radius is ao.

Figure 6.12: Solutions of the perturbation theory eigenvalue equations for the n = 2 hydrogenic atom.

The three-dimensional well is of depth Vo and is located at radii less than

a. In the regime where ka is much greater than one the scattering cross section
has a strong forward maximum at angles less than 1/ka which is typical for
the “diffraction” of classical waves.
The square well scattering in three dimensions is covered in the script
“QM_Scatt_3dwel”. Printout is shown in Figure 6.14. The scattering
amplitude is printed, the differential cross section, and the total cross section
all of which are solved for symbolically.
Figure 6.13: Plot of the unperturbed wavefunction |2, 1, 1 > and the eigenvectors of the perturbation
|2, 1, 0 > +|2, 0, 0 > where the numbers define the states n, l and m. The unperturbed state is symmetric
about 90°.

The differential cross section for ka = 1 is shown in Figure 6.15. The

distribution is forward peaked but also somewhat isotropic. The cross section
for ka = 10 is shown in Figure 6.16. In this case there is a sharp forward peak
with characteristic angles about 0.1 radians or 5.7 degrees. This behavior is
similar to that for classical diffraction of waves.

6.7. Photoelectric Effect — Continuum

The photoelectric effect refers to the process whereby an incident photon
interacts with a bound atomic electron, is absorbed, and the final state
electron is in a higher energy state. In the present case it is assumed that the
final state is unbound, in the energy continuum. In this case the final state |f >
can be approximated as a plane wave.
Figure 6.14: Printout from the script for Born scattering off a 3d square well.

In the approximation that the initial binding energy is small with respect to
the initial photon energy and the final electron energy the continuum electron
takes all the initial photon energy.

The cross section for this process, dσ/dΩ dω, is evaluated in the script
“Continuum_pe_effect”. Printout from that script appears in Figure 6.17. The
atom is assumed to be initially in the ground state, with a wavefunction which
is hydrogenic except that the full Z attractive protons are assumed to be
active. No screening of the nuclear charge by the other electrons is assumed,
so that the Bohr radius, ao, is reduced by a factor Z.
Figure 6.15: Angular distribution for scattering off a 3d square potential with ka = 1.

The user supplies values for the parameters Z and the photon energy in
mec2 units. Results for the choices Z = 90 and /mec2 = 0.01 and 10, 0.4 are
shown in Figures 6.18 and 6.19, respectively. In the first case the angular
distribution is a dipole like form reflecting the transverse nature of the electric
fields associated with the photon. In the second case the electrons are thrown
forward in a relativistic effect, the “searchlight” effect.
The dependence on the incident photon energy and the atomic number Z is
shown in the contour plot of Figure 6.20. The energy dependence is not very
strong, but the Z dependence shows a strong enhancement at low Z. This
effect arises because the “target” is smaller at high Z since the ground state
radius for an inner electron is reduced by Z because of the strength of the
attraction of the Z protons.
Figure 6.16: Angular distribution for scattering off a 3d square potential with ka = 10. In this case
there is a sharp, forward, diffractive peak.

Figure 6.17: Printout which shows the differential angular distribution as a function of the photon
energy, where alpha is the fine structure constant, and t is the polar angle.
Figure 6.18: Angular distribution for the case where Z is assumed to be 90 and is taken to be
0.01 and the process is in the non-relativistic regime.

6.8. Line Width

The wave function in quantum mechanics, ψ, can be taken to be a function of
momentum or position or energy or time because they are complementary
variables. One can freely convert from one description to the other, such as
conversion from time, t, to energy, E.

Suppose that the state being described decays in time with a lifetime τ. The
reciprocal of the lifetime is called the width of the quantum state, Γ because
any unstable state will have a spread in energies of a magnitude about the
width. Expressed in the energy variable, the state has a spread in energies due
to the uncertainty principle and since the time of the state’s existence is
limited.
Figure 6.19: Angular distribution for the case where Z is assumed to be 10 and is taken to be
0.4.

That spread is called the natural width, which has a Lorentzian shape when
described using energy as the variable. In the time domain the wave function
is an exponential function.

A script for atomic electric dipole decays called “line_width” is provided.

For example, the decay of the hydrogen state 2p → 1s (or n = 2, l = 1 to n = 1,
l = 0) has a lifetime of 1.6 nsec and emits a photon of 10.2 eV energy. The
natural width is a fraction 4 × 10−8 of the photon energy, very narrow indeed
scaling as approximately the cube of the fine structure constant, α = 1/137.
Printout from the script appears in Figure 6.21.
Figure 6.20: Contour plot of the continuum photoelectric angular distribution as a function of atomic
number Z and incident photon energy.

The experimentally measured width as a function of temperature is shown

in Figure 6.22. The natural width is temperature independent, but very
narrow. The total width is dominated by the Doppler broadening of the
spectrum due to thermal motion of the hydrogen gas until low temperatures
are achieved in the laboratory. In fact, only fairly recently have the natural
atomic widths been measured directly for a variety of gas molecules. As seen
in Figure 6.22, the temperature when the natural atomic width is comparable
with the thermal Doppler broadening is approximately one 100th of a degree
absolute.

6.9. Barriers
The problem of a series of barriers appears in solid state physics for ionic
crystals where they form a periodic potential. It can be approximately solved.
There is assumed to be a region of zero potential, where the solutions are
plane waves and a region where the potential Vo is greater than the energy E
so that there the solutions are exponentially damped in that region. Therefore,
the electrons are bound to the crystal. The coefficients at an interface between
the two regions can be found by requiring the continuity of the wave function
and its first derivative. There is oscillatory behavior with wave vector k and
exponential behavior with coefficient K. The incident waves with left going
and right going coefficient C and D connect to −x and +x exponentials with
coefficients A and B as;

Figure 6.21: Printout from the script “Line_width”. The transformation of the wave function from the
time domain to the energy domain is performed symbolically using MATLAB “int”.
Figure 6.22: Plot of the approximate energy width for hydrogen in the 2p → 1s transition. The
Doppler shift due to atomic thermal velocities dominates except at low temperatures.

The script “Barriers2” allows the user to set the geometry of the array of
potentials whose number is also to be chosen. The solution assumes the
barriers are quite “thin”, so that the exponent Ka is set to zero in the argument
of the exponential. The barrier height is fixed at 7 eV.
Results for two “thin” barriers are shown in Figure 6.23 while the
transmission coefficient for three barriers is displayed in Figure 6.24. In this
case the resonances are quite sharp at low energies where E is less than the
potential. The user is encouraged to explore a single barrier and also an array
of barriers where the exponential region is less “thin”. For example, a single
barrier with a region of size 4 Angstroms is shown in Figure 6.25 and looks
rather like the familiar transmission result for a single barrier. Looks
approximately like the solution of a barrier.
Figure 6.23: Transmission coefficient for 2 thin barriers, 1Å wide, separated by a 6Å free space. The
barrier height is 7 eV indicated by the red line.

Figure 6.24: Transmission coefficient for 3 thin barriers, 1Å wide, separated by a 6Å free space.
Figure 6.25: Transmission coefficient for a single barrier of width 4Å as a function of incoming wave
energy E for a barrier of height 7 eV.

6.10. General Eigenvalues

The time independent Schrödinger equation can be solved numerically for

any potential configuration. Examples are shown in the script “Gen_Eigen2”
for a single well, two wells, the simple harmonic oscillator and a hyperbolic
sin confining potential.
The Schrödinger equation on a numerical grid with x labeled by index j
becomes:

In this formula the grid spacing is Δx. The grid form of the second
derivative was already seen when the discrete “slinky” was taken to the
continuum limit when the wave equation was “derived” in Section 3.
Figure 6.26: Menu for the script “Gen_Eigen2”. The printout for a specific SHO gives the first 3
eigenvalues.

The equation could be solved using the MATLAB partial differential

equation utility for a time independent special case. Alternatively, it is
approached here as a grid or discrete solution. The span of x is set to be 50
Angstroms sampled 1000 times. The solutions are obtained using the
MATLAB utility “eigs” with the sparse matrix, the Hamiltonian matrix which
is almost diagonal. As can be seen from Eq. (6.12), the matrix has only 2 × 2
off diagonal terms. The time independent Schröldinger equation leads to
finding the energy eigenvalues of the Hamiltonian matrix.
A menu choice is available to the user as shown in Figure 6.26 along with
the printout in the specific case of a SHO. The script prints the first three
eigenvalues, largest first.
The results for a specific two well potential shape is shown in Figure 6.27,
while the eigenfunctions are plotted in Figure 6.28. The eigenvalues are 1.62,
0.51 and 0.46 eV in the two well example. The wells are two electron volts
“deep”.
Other choices are possible. In fact, it should be very simple for the
interested user to edit the script and set up any problem of interest. The
possibilities are open ended.
The eigenfunctions for a hyperbolic sin potential are displayed in Figure
6.29. The number of oscillations is proportional to the momentum and larger
numbers lie higher in energy. That behavior in similar to that displayed by the
simple harmonic oscillator.
Figure 6.27: Plot of the potential in the specific two well menu choice.

Figure 6.28: Plot of the eigenfunctions for the first three solutions in the example of the specific two
well problem. The green function is the most deeply bound, followed by red and then blue.

Figure 6.29: Plot of the eigenfunctions for the case of a hyperbolic sin potential. The green function is
the most deeply bound, followed by red and then blue. The solutions are like those of the SHO.

Note that the “eigs” script orders the results with the highest energy first,
which means that asking for the first three eigenvectors is ordered with blue
as the highest energy, followed by red and then green. The user can choose
parameters defining all the other menu choices.
6.11. Koenig–Penny

The Koenig–Penny model is a simplified, one-dimensional, constant and

periodic potential approximation to the electron wavefunction in a solid
where the ions are fixed and provide the binding forces. The wave function is
assumed to be the free particle plane wave, wave vector k, modulated by a
periodic function u(k) which accounts for the fixed ions.
The width of the potential Vo is b, while the spacing between the regions of
constant potential and of zero potential is a. Requiring continuity of the
wavefunction and its first derivative at both boundaries, x of zero x of a,
results in the constraint that the four homogeneous equations for u and du/dx
matching have determinant zero. The solution is k, while the wave vectors for
region a is ka, while that for region b is kb.

This constraint requires a numeric solution, and the MATLAB utility

“fminsearch” is used to find k in the script “Koenig_Penny2”. An
approximate solution for narrow potentials is given in many textbooks, where
there is a forbidden zone for the energy where no solution exits. The zone
occurs approximately when kaa = nπ. The dialogue for an example with a = 4
and b = 1 Angstroms with a potential of 5eV is shown in Figure 6.30. The
energy solutions are plotted in Figure 6.31 It is notable that there are
forbidden energies, near which the “fminsearch” results are a bit numerically
unstable because of the discontinuities while “fminsearch” assumes the
function to be minimized is well behaved. Nevertheless, the predictions of
bands of energy where the wavefunctions are allowed and move freely and
energy bands where they are excluded are predicted. This is basically the
simplest model for the energy band theory of solids.

Figure 6.30: Printout for a specific choice of the geometry and energy for the barrier height.
Figure 6.31: Relationship between energy, E, and wave vector, k, for a free particle (blue) and for a
particle bound into a one-dimensional periodic, constant potential (red). Energies in a band
approximately 1 eV wide near 2.5 eV are not possible.

The wavefunctions are required to be periodic, u(a) = u(-b). The assumed

form of the wavefunctions for the regions with zero potential and for binding
potential, Vo > E, are shown in Eq. (6.15)

Having found the wave vector k, the wave functions can be solved for at
any particular, user chosen, energy. For a specific energy in the allowed band,
1.5 eV, the wavefunctions are shown in Figure 6.32. The exponential region
for x = (−b, 0) and the oscillatory region, x = (0, a) match at x = 0. The
regions where the potential exists are indicated in red. The wave functions are
continuous and periodic. The solutions are derived from the matching
conditions, assuming A equals one, and using three of the equations to solve
for B, C and D using the MATLAB matrix inversion utility, “inv”, to solve the
three equations in three unknowns by matrix multiplication.
Figure 6.32: Wave functions for the specific case of parameters chosen for Figure 6.30. The function
is periodic and continuous across the 2 regions of potential.

6.12. Decay Chain

In the radioactive decay of an initial state, there may be several sequential

“daughters”, whose abundance is then time dependent. For example, U238 →
Th234 → Pa234. A simple case, assuming a single daughter species for each
parent, appears in the script “Decay_Chain” where the species a decays into
species b and hence b decays into species c. The decay rates are respectively
f, g and h for species a, b and c. The first order differential equations appear
in Eq. (6.6). The species a simply decays exponentially.

Figure 6.33: Symbolic solution of the populations for the 3 species, a, b and c.
 Species b and c are “fed” by species a and b respectively while decaying
with rates g and h. The equations are solved symbolically using the MATLAB
utility “dsolve”. The printout of that script appears in Figure 6.33. The user
can change the rates. For example, it is possible to depress the population of
species c by reducing the rate of species b which is the necessary “gateway”
for species c. A specific example appears in Figure 6.34.

6.13. Casimir Effect

The Casimir effect is due to quantum fluctuations in the vacuum. Classically,

two parallel plates with no external electric field applied experience no force.
However, there are quantum zero point fluctuations, for a harmonic
oscillator in the ground state, that induce a force on the plates. The fields can
be imagined as arising from the vacuum fluctuating into an electron–positron
pair which then goes back into the vacuum on a time scale consistent with the
uncertainty principle.

Figure 6.34: Populations of the 3 species in the specific case for decay rates of 1, 0.6 and 0.5. Species
2 builds up first, followed by species 3, while species 1 decreases exponentially.

This effect has been measured and has physical implications. It is, as will
be seen, important only at very small plate separation. In fact even today, as
microelectronics feature sizes continue to decrease, the Casimir effect must be
taken into account by chip designers. This is truly an impact of quantum
virtual pairs of particle-antiparticle in what is now everyday life.
One estimate of the size of the effect is to sum all zero point energies
subject to boundary conditions that the wave functions vanish at the surface
of the conducting plates. The quantized frequency for plates normal to the z
direction, with separation d is:

Figure 6.35: Numerical results for a user defined input of plate separation.

Figure 6.36: Energy per unit plate area as a function of plate spacing d, showing the inverse cubic
dependence on d. With a spacing of 5Å the energy per area is −0.22 in eV and A units.

The sum of all n and the integration over the transverse wave vectors,
however, diverges and convergence factors, in this first quantization
treatment, must be applied. The result is an attractive force between the plates
which creates a pressure, or energy density, of:

The pressure is, numerically, an energy density of 27.1 eV/A3. The Casimir
effect is explored in the script “QM_Casimir”. Printout from that script
appears in Figure 6.35.
The Casimir pressure as a function of separation is shown in Figure 6.36. It
is clear that only on the Angstrom scale is the effect going to be important.
However, it has been measured and it shows that there is quantum energy in
the “vacuum”.
 Chapter 7

Astrophysics

“Astronomy compels the soul to look upwards and leads us from this world to another.”
— Plato
“And yet it moves.”
— Galileo Galilei
“I ask you to look both ways. For the road to a knowledge of the stars leads through the atom;
and important knowledge of the atom has been reached through the stars.”
— Sir Arthur Eddington

7.1. Transfer Orbit

The solar system offers many planets to study. As a first example a movie is
provided by the script “Mars_More3” for the orbits of the inner planets
Mercury, Venus, Earth and Mars. In Figure 7.1 is shown the locations of the
four inner planets of our solar system at the end of a two-year time span, the
last frame of the movie, when all the planets are started with an azimuthal
angle of zero.
The visual experiment of watching the correlation of the orbital radii and
the periods is illuminating. For a circular orbit the period τ, the energy E and
the angular momentum L for an orbital radius arc, taking the planetary mass
equal to one;
Figure 7.1: Positions of the 4 inner planets, red stars, after a time span of 2 years. Orbits are in blue
and the sun is a green O.

Figure 7.2: Printout from the script “Mars_More3” for choices of Mars and Neptune as outer planets
to travel to.

Therefore, if we specify the radius we know everything about the orbit. The
dialogue printout is shown in Figure 7.2. The user chooses an outer planet.
In planning for planetary exploration it is useful to establish a feeling for
the size of the times and velocities involved. For example, Mars has a period,
velocity and orbital energy with respect to Earth of 1.87, 0.81 and 0.66,
respectively. A “transfer orbit” starts with a velocity increase to put the
vehicle into an elliptical orbit, semi-major axis a = r1/(1 − e), where e is the
eccentricity. The eccentricity of this orbit is (r1 − r2)/(r1 + r2) where r1 and r2
are the turning points, the perigee and apogee.
Figure 7.3: Approximate circular orbits for Earth (red) and Mars (green) and the elliptical transfer
orbit between them.

In Figure 7.3. the radii refer to the orbital size of the earth and Mars in
astronomical units, (AU). The eccentricity of the transfer orbit in this case is
0.21. The launch is at transfer orbit perigee and the rendezvous is at apogee.
The velocities during the trip are shown in Figure 7.4. As can be seen, boosts
in velocity are needed both at launch and at the end of the trip. The one way
trip time is 0.71 years. Faster orbits are possible but more costly in terms of
the fuel required and the reduced payload ratio. Clearly, there is also a
“launch window” since Mars needs to be there at the end of the transfer orbit.
For other planetary trips the size and time scale gets quite large. For
example, a trip to Neptune along a transfer orbit has an eccentricity of 0.90
and takes 16.0 years. The orbits are shown in Figure 7.5. To arrive at Uranus
requires a very elliptical orbit. The user is encouraged to try a few more of the
outer planets which are available on the menu given by the script “Mars
_More3”.
Figure 7.4: Velocities for the Mars transfer orbit. A boost is needed to begin. The red and green stars
show the orbital velocities for the Earth and Mars, respectively.

It is of interest to speed up the trips. Firstly, the passengers will thank you
for the speedy trip. In addition, there are fierce radiation fields to which the
passengers will be exposed during the trip. Shielding would be heavy, so
perhaps the orbit is not optimal for speed. Indeed, it is a minimal energy orbit,
but other options cost additional fuel. At present, this problem is not fully
solved.

7.2. Flyby
The velocity for an object to escape the gravity field of the earth is 11.2
km/sec, while for the sun it is 42.1 km/sec. A more economical way to escape
the solar system than constructing an enormous rocket has been used by
making a “gravity assist” or “slingshot” around the outer planets, such as
Jupiter. In this fashion the Voyager probes escaped the solar system. The basic
idea is that in an encounter with a very massive object, the spacecraft will
essentially be dragged along and pick up a maximum of approximately twice
the velocity of the object. In this case the momentum the object picks up is
extracted from the gravity field of Jupiter. No actual collision occurs. The
script uses a very simple numerical integration of the force on the rocket due
to a planet with fixed orbit, unaffected by the spacecraft.
Figure 7.5: Approximate circular orbits for Earth (red) and Neptune (green) and the highly elliptical
transfer orbit between them.

The printout of the script, “Jupiter_Flyby” appears in Figure 7.6.

The script supplies a movie of the encounter. The final frame of a specific
example appears in Figure 7.7. On the time scales in question, the orbit of
Jupiter is approximated as an unperturbed straight line with a velocity defined
to be one and to be constant. The trajectory of the spacecraft is shown as a
blue dashed line.

Figure 7.6: Printout of “Jupiter Flyby”. The user defines the closeness of the encounter of the
spacecraft with a point like Jupiter by choosing the initial velocity.
Figure 7.7: Trajectories of Jupiter, red, and a spacecraft, blue dashed, for a specific flyby which makes
a close encounter between the spacecraft and Jupiter.

The success of this flyby can be seen in Figure 7.8 where the velocity of
the spacecraft throughout the flyby is plotted. The y velocity, which starts at
0.2 initially, becomes −1 in the encounter which is approximately the assumed
velocity of Jupiter and illustrates the “slingshot” aspect of the flyby. A simple
numerical integration of the encounter with Jupiter’s gravity is used. The
actual flyby which was used trails Jupiter rather than leading it because then
the direction of the spacecraft is not reversed as shown in Figure 7.9.
Figure 7.8: Velocity in the y direction of the model spacecraft as a function of time. The spacecraft
gets a velocity increment comparable to the velocity of “Jupiter”, defined to be 1, in the flyby encounter.

7.3. Lagrange Points

The use of satellites as astronomical observatories has resulted in the
requirement of a stable observation platform. That is accomplished by using
centrifugal force to balance gravitational attraction for a three body system.
The script “Roche_Lagrange” explores the location of these stable Lagrange
points for binary systems, where the user chooses the masses of the large
bodies. The CM is at x = 0, while the masses are at positive and negative x,
heavier mass to the left. The printout of the script is shown in Figure 7.10.
The total potential is the sum of the potentials of the two large bodies plus
the centrifugal potential due to fact that the test body is assumed to rotate with
the binary orbital period about the two body center of mass. The mass ratio
m1/m2 is q. There exist five stable points for this potential.

Figure 7.9: Trajectories of Jupiter, red, and a spacecraft, blue dashed, for a specific flyby which makes
a close encounter between the spacecraft and Jupiter and is chosen to lag behind Jupiter, thus largely
preserving the initial direction.
Figure 7.10: Printout of the script “Roche_Lagrange” with a specific mass ratio chosen for the two
large masses.

There are three collinear points on the x axis joining the two masses. Two are
near the lighter mass; one toward the heavier mass and one away. The third
point is well beyond the heavier mass, located at negative x. There are two
additional points at sixty degrees to the heavier mass and at a radius which is
the same as the third Lagrange point. A contour plot of the equipotentials is
displayed in Figure 7.11.

Figure 7.11: Equipotentials for the binary system with mass ratio equal to 2. The location of 3
collinear Lagrange points is visible, as are the 2 points which are at 60 degrees to the heavier mass.
Figure 7.12: Force in the x coordinate for a binary mass ratio of 2. There are 3 collinear points. Two
are near the lighter body, while the third is at large negative x, nearer the heavy body. The locations of
the 2 bodies are indicated by the green and black stars.

The forces can easily be derived from the potential. The force along the x
axis is plotted below in Figure 7.12. The locations of the three points where
the force is zero are clear. The script prints out the locations in x for the ten
minimum values of |fx|. In this specific case they occur at x = −1.15, 1.23, and
0.22. The minima are a bit shallow, so care is needed in interpreting the use of
the MATLAB utility “sort”.

7.4. Binary Orbits

The simplest gravitationally bound system is a system of two objects. In the
case of one very heavy object, such as the sun in the Kepler problem, the
system can be taken to be a one body problem. However, in general the
motion can be complex and the script to deal with it uses the MATLAB utility
“ode45” to solve the problem numerically. In two dimensions the problem has
eight variables; position and velocity of body1 and the same for body2. The
force law is quite simple, even though the actual motion can be quite
complex.

The discovery and understanding of the force law benefitted from the
existence of several planets with different radii and periods and the fact that
the sun has almost all the mass of the solar system. The solar system is
effectively a series of one body, independent, planetary orbits as assumed
previously.
One application of the general problem comes from the present search for
extra solar planets. The Kepler problem solution assumes that the sun is much
more massive than any planet. In that case the problem only requires the
motion of the planet and the sun can be considered to be at rest. However, in
general there is motion of both bodies. An example using the script “Binary2”
for a separation of 1 astronomical unit (AU) with body1 having the mass of
the sun and body2 having one percent of that mass is performed. The initial
position and velocity of the sun are taken to be zero, while that for the
“planet” has an angular velocity of a circular orbit with no radial velocity.
There is a movie of the motion provided and the total paths are shown in
Figure 7.13. Note that there is a wobble to the sun which could be determined
using the Doppler effect to measure the velocity of the sun. This example is
clearly made to enhance the effect, but nearby planets of extra-solar stars
have, in fact, been observed using this technique.
At the other extreme, equal mass binary pairs can be studied. In our galaxy
it appears that about ten percent of the star systems are binary, so the exercise
is useful. An example with equal mass objects with a solar mass each at 1 AU
is shown in Figure 7.14. For body1 the initial velocity was 6 radians/AU and
the radial velocity was 2. For body2 the angular velocity was −5 and the
radial velocity was 2. The resulting evolution of the binary system is
complex. The reader is encouraged to play with other combinations of the six
input parameters to find interesting orbits.

Figure 7.13: Motion of a solar mass being orbited by a “planet” with a mass of 1% of the sun. The
wobble of the sun and the subsequent perturbation of the planetary orbits are very evident.
7.5. Three Body Orbits
An extension of the previous script, “Binary2”, is to consider a heavy sun
with solar mass that is fixed at the origin and is orbited by two “planets” in
unperturbed circular orbits. The script is “Three_Body” although a general
three body problem requires more variables. The forces are central, so that
motion is in a plane. The user chooses a radius and mass for both “planets”
with prompts for the earth and for Jupiter, as seen in Figure 7.15
The script uses the MATLAB “ode45” numerical integration od
differential equations to solve for the subsequent motion of the two planets.
The script is run twice, once for no perturbative force between the two planets
and then again with the gravitational interaction turned on. A movie is shown
in the unperturbed case. This procedure is adopted because the actual changes
due to Earth-Jupiter interactions is small and even integrating over five years
leads to small effects. The difference of the two orbits is shown in Figure
7.16, where the size of the effect is a few parts per thousand of the size of the
orbit after five years. The time dependence of the difference appears in Figure
7.17.

Figure 7.14: Orbital motion for equal mass binary stars. The initial conditions define roughly opposite
angular velocities and equal radial velocities.
Figure 7.15: Dialogue for the choice of the mass and radius of 2 planets in circular orbit about the sun.

Figure 7.16: Differential effect of the perturbation due to Earth-Jupiter attraction during tracking over
5 years. Initially Earth and Jupiter are given the same azimuthal value.

A study which is less sensitive to numerical roundoff errors is to put

Jupiter at a radius of two AU. In that case the perturbation of the orbit is about
2% per year relative to the unperturbed circular radius. Indeed, the interested
user can explore how the radii and masses of the two planets lead to rather
different orbits. For example, the heavier planet suffers smalled fractional
perturbation, as expected. In extreme cases the lighted planet can be captured
by the heavier planet, causing the script to terminate prematurely.

7.6. Polytropic Star

Two basic conditions for a star in equilibrium under gravitational attraction
are that mass, M(r) is conserved and that a pressure P(r) resists the attraction
and keeps the star in equilibrium
Figure 7.17: Difference between the unperturbed and perturbed orbit of Earth for 5 years scaled to the
radius of the orbit of the Earth.

The source of the pressure is not specifically specified. The Lane–Emden

equation describes a form of the Poisson equation for the gravitational
potential of a self-gravitating, spherically symmetric, polytropic fluid. There
is assumed to be an equation of state relating pressure, P, and density, ρ, and
which defines the polytrope index, n, which is a free parameter. The ideal gas
law need not be assumed in this model which relates pressure and density
through the index, n.

The temperature in the printout follows from the ideal gas law, P = kρT/m
where k is the Boltzmann constant and m is the mean mass of the molecules
making up the star.
The Lane–Emden equation is solved numerically using the MATLAB
utility “ode45”. The two constraints in Eq. (7.4) are imposed with a defined
relationship between pressure and density set by n, Eq. (7.5). The input
needed is the core, r = 0, temperature. The pressure and density are derived
quantities in this formulation. The solutions are integrated to larger r and
terminate at R where the temperature is defined to be zero as a boundary
condition. The solution is fully defined by the normalized temperature φ
treated as a function of normalized radius r/R.

Printout from the script “Star_Polytropic” is shown in Figure 7.18. The index
n of 3.5 gives a reasonable representation of the solar data which is plotted on
the same graph as the polytropic solution. The user chooses the index and can,
therefore, see how the choice changes the density, temperature and pressure
radial profiles.
The temperature profile is shown in Figure 7.19, while the pressure
distribution appears in Figure 7.20. The overall agreement of this very simple
model with the solar data is not too bad and is clearly a large improvement
over assuming a uniform density for the star.

Figure 7.18: A printout of the dialogue for a particular, approximately, solar index.

Figure 7.19: Radial distribution of temperature for a polytropic star with index 3.5. The data points,
‘o’, are solar data.

7.7. Pulsating Stars

It is not always the case that the hydrostatic equilibrium for a star is stable.
The pressure which resists gravity is due to the elevated temperature of fusion
reactions and the structure of the star may become unstable. Red giant stars
are observed to pulsate, growing larger and then smaller on time scales of
order months. The behavior can, in first approximation, be ascribed to a
competition between the interior stellar pressure, increasing r, and the force of
gravity, decreasing r. The expansion is assumed to be adiabatic, which for a
perfect gas keeps PV5/3 constant. After cooling of the shell which expands,
contraction takes place and then re-heating by the core and again expansion of
the shell occurs.

Figure 7.20: Radial distribution of the density for a polytropic star with index 3.5. The data points, ‘o’,
are solar data.

As seen from Eq. (7.7), depending on the core pressure and radius there may
be either expansion or contraction. The pulsations are considered in the script
“Stellar_Pulse”. The user is given the core mass, the outer shell mass and the
radius. The user chooses the pressure P0. The printout of a specific dialogue is
shown in Figure 7.21.
The heat from the interior is absorbed by the shell when the hydrogen in
the shell is ionized. The shell moves outward, cools, de-ionizes, and the shell
is pulled back in by gravity. The balance is quite critical and values of the
pressure from 0.4 to 0.6, in units of 105 Nt/m2, show stable oscillations, while
outside this range the behavior is not stable.
 Plots of radius as a function of time are shown in Figures 7.22 and 7.23 for
pressures of 0.4 and 0.6, respectively. The pulsation period in the former case
is fairly slow, on the scale of months, while in the latter case the scale is much
more rapid, on the scale of weeks. The amplitude is much smaller in the latter
case than in the former, as can be seen in the plots which are provided.
Indeed, this simple picture gives an approximate model for pulsating red giant
stars.

Figure 7.21: Dialogue for stellar pulsations. Note that the sun has a radius of 7 × 108 m and a mass of
2 × 1030 kg. In comparison the script defines a red giant star.

Figure 7.22: Pulsations as a function of time (sec) for a red giant with pressure 0.4 × 105 Nt/m2.
Figure 7.23: Pulsations as a function of time (sec) for a red giant with pressure 0.6 × 105 Nt/m2.

7.8. White Dwarf

Stars fuse light nuclei and the resulting gas pressure and radiation pressure
balance the attraction of gravity. This mechanism was implicit in the
polytropic star but not explicitly modelled. However, when the final nuclei
are iron, fusion is no longer exothermic and the star cools. In that case there is
a contraction due to gravity which is no longer opposed by gas and light
pressure but by the Fermi pressure of the electrons (fermions) in the star.
The nuclei have all the mass, but the electrons supply all the pressure. As
the density increases, the star shrinks. When the electrons begin to become
relativistic, the equation of state for the electrons changes. The Fermi
momentum, pf, is related to the electron number density, n. The fermions
cannot occupy the same state and the number of possible states in three
dimensions goes as the third power of the size of the star, V. Therefore, as the
density n increases, the momentum of the fermions must increase.
The electron equation of state for pressure P depends on density, ρ, with
the fermion energy going as the square of pf at low density and becoming
linear in pf at high density. The electron to nucleon fraction is Ye, is assumed
to be for iron in this script. Since pressure is an energy density, dP/dρ is
dimensionless, c equals one unit.
The typical values of density are no, the number density in the transition
region where the electron momentum becomes comparable to the mass. The
total mass density at that point, ρo, is determined by the nuclei and leads to a
typical radius Ro and mass Mo.

Numerical values of the scale factors for radius and mass appear in the
printout of the script, shown in Figure 7.24.

Figure 7.24: Printout from the dialogue of the script “White_Dwarf”. Numerical values for the scale
of electron number density, mass density, radius, stellar mass and stellar radius are given.
Figure 7.25: Scaled mass as a function for scaled radius for a white dwarf with a scaled central density
of one.

The white dwarf is explored in the script “White_Dwarf” where the

differential equations are solved numerically using the MATLAB “ode45”
utility. A central density is chosen by the user and then the density is
integrated in radius until it reaches zero at R, where the total mass is defined
to be contained within radii less than R. A plot of the mass, scaled to Mo as a
function of radius, scaled to Ro for a central density of ρo equal to one appears
in Figure 7.25.
The white dwarf is not like a typical star, for example, the polytropic star
displayed above, because the attraction of gravity is opposed only by the
Fermi pressure of the electrons. A plot of the mass and radius for different
central densities appears in Figure 7.26. At low densities the mass grows with
density and the star contracts. At higher densities the mass levels off because
the electrons are becoming relativistic and cannot support the gravitational
attraction. This level is roughly at the mass of the sun. However, the radius is
approximately 1/10000 the solar radius, indicating the compact nature of a
white dwarf. Above about one solar mass the Fermi pressure of the electrons
cannot resist the formation of a black hole and the heavier neutrons must take
over with their pressure.
Figure 7.26: Radius and mass of a white dwarf as a function of the central, core, density of the star.
The radius shrinks with density and at high density the mass levels off at values approximately one solar
mass.

7.9. Boltzmann and the Sun

It is perhaps amusing to connect the fact that the sun is yellow with the solar
constant on earth which is fundamental to solar power feasibility studies. The
energy per unit area radiated by a black-body is proportional to the fourth
power of the temperature. The constant of proportionality is the Stefan–
Boltzmann constant σ. The value of σ is 5.67 × 10−8 watts per square meter
per degree4. The energy density is u. It has a distribution of energy which is
where x = E/kT

The distribution of energies of the emitted photons follows the Bose–Einstein

distribution, but with a zero chemical potential. The chemical potential is zero
because all the photons can sit in the lowest quantum state since the number
density of photons is not constant because the system is in a “bath” at
constant temperature. The density of states, or phase space for photons goes
as so that the spectral energy density du(E)/dE
goes as the cube of the energy times the Bose–Einstein statistical factor.

Figure 7.27: Printout of the script “Boltz_Sun2” for two samples of solar surface temperature. The
quoted formula is for energy density per wavelength.

The emission spectrum of the sun peaks at a wavelength, yellow, that

corresponds to a temperature on the surface of the sun of about 5700 K. That
means a radiation of about 60 million watts per square meter. Scaling by the
inverse radius of the sun and the earth radius of one AU, the solar constant for
energy falling on the earth is 1.27kW/m2. The script “Boltz_Sun2” makes
these calculations and invites the user to try different temperature, perhaps
appropriate to red giant stars or blue stars. Printout for two examples is given
in Figure 7.27. It is clear that the radiated power is very strongly dependent
on the temperature.
Figure 7.28: Spectral energy density for a temperature of 4000 kelvins. Wavelength is in Å units.

Figure 7.29: Spectral energy density for a temperature of 5700 kelvins, appropriate to the sun.

The energy density per unit of wavelength follows from a change of

variables from energy to wavelength for photons and is:

A plot of the spectral distribution is shown in Figure 7.28 for a temperature of

4000 kelvins and in Figure 7.29 for the standard solar surface temperature of
5700 kelvins.
 Chapter 8

General Relativity

“God abhors a naked singularity.”

—Stephen Hawking
“It is impossible to imagine a four-dimensional space. I personally find it hard enough to
visualize a three-dimensional space!”
— Stephen Hawking

8.1. Light Deflection

Light is deflected by nearby masses because light has energy and energy has
mass and all mass gravitates. The geodesic equation for light with the variable
change u = 1/r is given below in Eq. (8.1). This equation does not have a
closed form solution. Nevertheless, an approximate solution for small light
deflections consists of assuming a Newtonian straight line path for the term
proportional to GM.

The approximate equation has solutions:

The approximate solutions are displayed by the script “Light_Deflect” which

uses the MATLAB utility “dsolve” as shown in Figure 8.1. There is a
symmetry in the solutions before and after the point of closest approach. In
the printed evaluations, the initial conditions are a radius at infinity, u of zero
at angle of zero, and an initial angular velocity of 1/R. The light deflection
angle is evaluated for a given value of the impact parameter, R with respect to
the Schwarzschild radius rs = 2GM/c2.

Figure 8.1: Printout from the script “Light_Deflect”. Showing the symbolic solution of the
approximate equation for the trajectory of light.

The asymptotes can be read off from the approximate equation above for
zero angle, and u of zero, ±2GM/Rc2. The small angle solution for the angle
between the asymptotes is 2rs/R. This appears in the printout as “4m” plus a
small correction for the approximate initial conditions used, since m = GM
and R = 1 in the script. For the sun the deflection is rather small, but light
passing near a black hole can have major angular deviations from a straight
line. The plot made by the script is shown in Figure 8.2.
The light orbits for Newtonian and general relativistic orbits are shown in
Figure 8.3. The blue line is the Newtonian straight line, while the red dashed
line is for rs/R = 0.02 and the black line is a distorted representation of the
surface of the sun, radius R.

8.2. Circular Geodesic

In classical mechanics, any radius is possible for a circular orbit. The classical
radius, a, depends quadratically on the angular momentum, L, of the orbit, as
mentioned in Eq. (7.1) above
Figure 8.2: Light deflection angle as a function of the ratio of impact parameter to rs. The result is
strictly only valid for small deflections.

The orbits in a Schwarzschild space are examined in the script

“circle_geodesic2”. The equation for a circular orbit has a classical term plus
an additional term due to general relativity for the metric, as quoted
previously in Eq. (8.1), 3GM/u2. The classical equation for u = 1/r has a
constant of the motion L. The general relativistic geodesic has a related
constant, h, which is approximately L/c. The radius of light, aγ, can be read
off Eq. (8.1), u = 3GMu2, or aγ = 3 GM/c2 = 3rs/2
Figure 8.3: Light trajectories for Newtonian light (undeflected, blue) and light with a small ratio of the
impact parameter and the Schwarzschild radius, 0.02. The black line is the solar sphere.

Light is “attracted” to the sun, and it has an infinite h which means the
circular orbit for light has a radius 3/2 the Schwarzschild radius, rs = 2GM/c2.
In Eq. (8.4) the first term is the classical result u = 1/a ~ GM/(hc)2.
For a material particle, h is finite and the solution to the geodesic equations
is a function of the parameter h which is a constant of the motion, has the
dimensions of length and is approximately L/c. In the classical limit, a rs
and h rs, h approaches L/c and a approaches L2/GM

In the script, the user picks an initial value of the constant of motion h. It is
clear from Eq. (8.5) that the value of h must be greater than and therefore
the radius must be greater than three times the Schwarzschild radius or twice
the photon orbital radius. An example of the printout for the script appears in
Figure 8.4.
The script supplies the radius as a function of the constant h. The radius,
both classical and relativistic, appears in Figure 8.5. The relativistic radius
approaches three times the Schwarzchild radius, while the classical radius is
smaller and has no minimum. Note the photon point which has a radius twice
as small as the smallest possible orbit for a material particle. However, that
orbit is unstable as are all orbits for material particles below three
Schwarzschild radii unless external power is applied.

8.3. General Geodesic

The Schwarzschild metric is the solution in General Relativity (GR) for the
space-time due to a point mass at the origin. Particles and light travel on
geodesics in such a metric. Radial geodesics have solutions which exist in
closed form;

Figure 8.4: Dialogue for the circular geodesics.

Figure 8.5: Radius of a circular orbit for a classical and a relativistic treatment. Note the GR minimum
radius is 3 in rs units. There is no classical minimum. The photon orbit at 3/2 units is also shown.
The coordinate radius is r and the coordinate time, measured at large r, is t,
while the proper time is s. The initial radius is ro, the initial velocity is zero
and rs is the Schwarzschild radius. The proper time is well behaved at all r,
while the coordinate time t diverges as the radius r approaches rs. Clearly as r
approaches rs, ct must approach infinity and an object falling into a black hole
takes forever to reach the Schwarzschild radius.
In the general case, analogous to the Kepler problem in classical
mechanics there is a constant of motion, h, like angular momentum. There is
another choice, basically that proper and coordinate clocks at large r run at the
same rate.

Using the constants of motion, the motion is in a plane as it is classically and

this fact is used to reduce the problem to one with only one dimension. The
equation of motion for u = 1/r is:

The first three terms arise in classical mechanics and define the Kepler
equation. The last term is added by considering the general relativistic
geodesic motion in the Schwarzchild space corresponding to a point particle
of mass M. For photons, the constant h approaches infinity.
The solutions are calculated in the script “gen_geodesic” which uses the
MATLAB utility “ode45”. The user chooses classical, relativistic or photon
solutions. The trajectory is defined by user inputs for the initial position in r,
the impact parameter (phi angle) and the L/c value in m units. The orbit is
plotted for an angular range also defined by the user.
The orbits for photons are almost straight lines for large r as is expected
classically because the metric becomes flat at large distances. However, for
smaller initial r, the deflection of light is very evident. A sample plot for light
is shown in Figure 8.6.
 The orbits of bound systems are not re-entrant in general relativity as are
Kepler orbits. This causes the famous perihelion advance of Mercury. Setting
the phi range of six (pi), the initial position of 5 and impact parameter of 5
with L/c of 6000 m results in the path shown in Figure 8.7. The path is a
bound state for a material particle, but it is far from the familiar Keplerian re-
entrant ellipse.

Figure 8.6: Trajectory for light with initial radius of 10 with impact parameter of 3. Note that b = 2 is
not a stable orbit and the light falls into the origin (red star). The Schwarzchild radius is plotted with
dashed red lines.

An unbound trajectory for initial r of ten, impact parameter of five and L/c
of twenty thousand is shown in Figure 8.8. The orbit is unbound and
corresponds classically to a hyperbolic orbit. The time samples are uniform so
that the slowing down of the particle at large distances is observable. The
relationship between coordinate time, t, and proper time on the orbit appears
in Figure 8.9. At large r values the two are approximately the same, while for
small radii, the proper time remains well behaved and the coordinate time is
larger. This effect has, indeed, been measured and confirms the relativistic
calculation.
Figure 8.7: Multiple orbits for a particle started on a circular orbit at radius 5. The non re-entrant
nature of the orbit is very evident in this extreme case.

Just as with Kepler orbits, the relativistic orbits have many parameters to
vary. The interested user is encouraged to try many combinations and explore
the subsequent motion.

8.4. Kerr Photons

The Einstein equations are non-linear because the field itself gravitates
because it has energy. For this reason very few solutions to the equations have
been found analytically. The first was the Schwarzschild metric and the
second is the Kerr metric.
The Kerr metric is the solution for a point mass M rotating with a signed
parameter a, which is proportional to the angular momentum of the mass with
the sign of a defining the sense of the rotation. The L/M value is
approximately a. There are two possible photon equatorial circular orbits, Eq.
(8.9), with radii depending on whether the orbit is in the sense of rotation or
against it. The differences are due to “frame dragging” where the Kerr
rotation drags the coordinates along. Spacetime is swept along by the rotation.
Figure 8.8: Unbound trajectory for relativistic particle motion, classically a hyperbola. At large r the
particle velocity is reduced.

Because of frame dragging there are no purely radial geodesics. In general

radial geodesics are impossible while circular geodesics are possible. There
are two event horizons. The mass is M, or GM/c2 = rs/2.

Figure 8.9: Relationship of coordinate time to proper time during the quasi hyperbolic orbit shown in
Figure 8.8. At small r, there is a major difference in t and s.

The geodesic equations for photons in equatorial orbits can be cast into a
simple first order differential equation. Note that a can be positive or
negative, depending on the sense of the rotation. The angular variable in the
equatorial plane is φ which depends on the Kerr rotation parameter a and M.

Printout from the script “Kerr_Photon2” is shown in Figure 8.10. Units with
M = 1 are used in this script. Values for the two event horizons and the two
possible circular orbits are part of the printout.
The equation for the orbit is solvable and is found using the MATLAB
utility “dsolve”. The symbolic solutions are also printed out, for both positive
and negative a. The constants C2 and C4 depend on the initial conditions.
Note that these particular photon solutions are for orbits which approach the
outer event horizon for the specific user input. Note that the trajectories are
incoming toward smaller r.

Figure 8.10: Printout from the script “Kerr_Photon2”. The equatorial photon orbits are solved for
symbolically.

An orbit for a = 0.98 and a minimum r of 1.25 is shown in Figure 8.11.

With these parameters the outer redshift boundary occurs at a radius of 1.20 in
M units. Note that the Schwarzschild radius, rs, in these M equals one units
has a value two, so that an observer can get closer to a Kerr mass than a non-
rotating mass using the frame dragging spin of the Kerr metric. The frame
dragging of the photon is very evident in the figure. Recall that a photon
minimum radius for a circular orbit is 1.5 in rs units, Eq. (8.4).
8.5. Kerr General
General geodesic motion in a Kerr metric is quite complicated. The script
“Kerr4” is restricted to the equatorial plane because only equatorial motion
remains planar due to frame dragging. Because the space is time symmetric
and axially symmetric in this case, there are two constants of the motion,
which can be identified with E, the energy at large radius and L, the angular
momentum as in the classical Kepler problem. Units with M = 1 are again
used, GM/c2 = 1, rs = 2. There is a distance where rotation classically has a
velocity of the speed of light which is L/c. As in the Kepler case, negative E
orbits are bound, while positive energy orbits are usually free but may be
captured by the dragging effect.

Figure 8.11: Geodesic for an equatorial photon traveling close to the outer event horizon. Units are for
M = 1, rs = 2.

There are then differential equations for the proper time, azimuth and
radius as a function of coordinate time t, with r limited to be larger than the
larger event horizon. There is, in addition, a sign for the radial solutions as
incoming or outgoing. The equations themselves are not very transparent, but
the user can read them off from the script “Kerr4” as desired.
In “Kerr4” the user chooses the orbital parameters analogous to the Kepler
parameters, E and L along with a, the initial radius, and the incoming or
outgoing sense of the trajectory. An example for a = 0.98, E = 1, L = 3, ro =
1.5 and outgoing appears in Figure 8.12. An orbit with the same parameters
except the sense of a appears in Figure 8.13.
Figure 8.12: Kerr trajectory for an outgoing path, E = 1, L = 3 and initial r = 1.5. The green circle is
the outer event horizon. The sense of the spin a and the trajectory reinforce.

Recall that Kepler trajectories are circles and ellipses for bound states and
parabolas and hyperbolae for unbound states and that the plots in Figures 8.12
and 8.13 are far from the simple conic sections. Thus it is clear that the Kerr
solution can have strong relativistic aspects, especially the “frame dragging”
due to the spin of the mass. The solutions in Kerr4 are for material particles
where L and thus h are finite. A movie is provided for the (x, y) motion along
the chosen trajectory as a function of equal steps in coordinate time t.
There are also interesting bound states with negative energy. An example
with E = −1, L = 10 initial r = 3 and incoming sense is shown in Figure 8.14,
where the frames of the movie are plotted as open circles. In this case the
frame dragging aspect of the orbit is very clear, since the sense of the
trajectory changes during the orbit. The user is encouraged to play with the
many available parameters which define a general equatorial orbit in the Kerr
geometry as only a few are illustrated here.
Figure 8.13: Trajectory for the same parameters as the previous figure except that the spin direction a
has been reversed. The trajectory is bent toward the sense of a, reversing the initial motion.

8.6. Gravity Wave Radiation

A rotating binary star system is a good candidate to be an observable source
of gravitational radiation. There are also many binary stars, so that their
radiation could serve as a “standard candle” for calibrating a successful
gravitational wave detector. For a binary with an individual stellar mass M
and radius R, the power radiated is;
Figure 8.14: Bound state trajectory with E = −1, L = 10, a = 0.98 and an incoming orbit. The particle
is captured by the rotation and cast into a quasire-entrant stable orbit.

The maximum radiated power is set by the quantity c5G which is 3.8 × 1052
watts. The emitted gravitational waves cause a metrical distortion
(dimensionless) at an observation distance r which is approximately:

That distortion is detected in a gravitational wave detector which can be

thought of as responding to the tidal forces, Eq. (2.10), carried by the wave
which elongate and compress a test body.
The distortion has a frequency of the orbital motion of the binary, which is
typically in the kHz range. The quantities needed to detect gravitational
waves are shown in the script “GR_Rad”. If these waves can be detected, then
observational cosmology can go to times shorter than the cosmic ray
background decoupling time which now blocks observation since the plasma
is opaque to visible light. Numerical values for a user chosen M, R and r
choice are shown in Figure 8.15.
Figure 8.15: Printout for a particular user choice of parameters for a rotating binary star system.

Clearly, the metrical distortion is very small and sets the scale for
fractional elongations in any gravitational wave detector. A contour for the
orbital frequency as a function of M and R appears in Figure 8.16. The
detectors must be sensitive to a range from 102 Hz to perhaps 105 Hz if they
are to use binaries to calibrate. The radiated power as a function of M and R is
shown in Figure 8.17 scaled to the maximum possible power. Detectors
should be sensitive to metrical changes several orders of magnitude below the
maximum value.

8.7. Stellar Pressure

The interior of a uniform density sphere can be explored in Newtonian and
general relativistic mechanics. The Newtonian interior potential is:

where M is the total mass and R is the radius. A plot of the interior and
exterior potentials is shown in Figure 8.18.
The Newtonian pressure, P, can be written in terms of the Schwarzschild
radius for a uniform density star. The equation of hydrostatic equilibrium was
already shown for polytropic stars and is reproduced here:
Figure 8.16: Orbital frequency as a function of M and R. Detectors need to be sensitive to disturbances
at those frequencies since the radiation occurs at the orbital frequency.

The pressure is an energy density, so that P/ρoc2 is dimensionless. It is easily

solved for
Figure 8.17: Radiated power in watts scaled by the maximum power possible as a function of M and
R. Only extreme binary systems with large M and small R approach the maximum.

The pressure is largest at the origin, P(0)/ρoc2 = rs/4R, and disappears at

the surface by definition from the boundary conditions.
In the general relativistic case, the tensor for dust without velocity is
assumed. In that case a closed form solution is possible which approaches the
classical solution when rs/R approaches zero. The equation for pressure in a
Schwarzschild interior space is:

which is the Oppenheimer–Volkoff–Tolman equation for a uniform density

sphere. There are two new pressure terms and a correction term in the
denominator, but in the limit of small Schwarzschild radius and low pressure
Eq. (8.17) yields the Newtonian equation.
Figure 8.18: Newtonian potential energy for a uniform sphere of radius R. For r > R the potential
scales as 1/r. Arbitrary units, GM = 1, are used.

The pressure values are compared in a script called “GR_P_Interior2”. The

Oppenheimer–Volkoff–Tolman equation is solved symbolically using
“dsolve” and the boundary condition P(R) equals zero is imposed.
Printout from that script is shown in Figure 8.19. The symbolic solution
for pressure is part of the printout, and the pressure at r equals to zero for
classical and GR pressure are compared.
The pressure as a function of radius is shown in Figure 8.20. After some
manipulation, the solution of the equation shown in Figure 8.18 can be put
into the following form.
Figure 8.19: Printout for the script for a user supplied ratio of rs/R. The symbolic solution of the
equation is also printed out.

Figure 8.20: Pressures as a function of r for the ratio rs/R = 0.5. The GR contribution is substantial
compared to a Newtonian estimate for pressure.
Figure 8.21: Pressures at r = 0 as a function of the ratio rs/R. At low values, GR does not contribute
much, but at high values it dominates over the Newtonian estimate.

The maximum pressure, at r equals zero, as a function of the radius R in rs

units is shown in Figure 8.21. As expected, the pressure contributed little at
small values. However, at large values the pressure term actually dominates.
Clearly, star models for very dense stars must take this into account. In fact,
as discussed above, the uniform density approximation badly underestimates
the central density and pressure, so that a more realistic model would be a
polytropic extension using the Oppenheimer–Volkov–Tolman equation.
The relativistic pressure can be much larger than the pressure exerted by
the matter density because pressure acts as a source of gravity, recalling that
pressure is an energy density and that all energy has mass and therefore
gravitates. Parenthetically, in the collapse of a star of radius R and mass M to
a singularity, the change in binding energy is approximately the energy
release and is a substantial fraction of the rest energy of the star GM2/R →
GM2/rs = Mc2/2. Hence a supernova collapse releases a truly amazing amount
of energy as opposed to fission or fusion which release only a small fraction
of the rest mass of their reaction products.

8.8. Light Propagation in the Universe

Since the Universe is expanding, the transmission of light signals is affected
because of evolution during the transit time. The expansion means that the
apparent velocity of recession, v is roughly proportional to the distance L,
with proportionality defined to be the present Hubble constant Ho.

The wavelength at emission relative to the wavelength at reception is the ratio

of the size parameters, R since distances scale with R. The difference in
wavelengths defines the z parameter which can be measured by looking at
redshifts. Emission is at the time t with size R and reception is indicated by
the o subscript. The positive z value means the light is redshifted. For
example, the cosmic microwave background has a z value of about 1000.
Beyond those z values, earlier in t, electromagnetic information is not
available because the universe is opaque. At present size scale at emission, z
is zero while at very small size scale at emission, z becomes very large.

The Hubble parameter is defined by the expansion rate, set by the rate of
change of the scale parameter R, so that it has the dimensions of inverse time.

The Hubble length is defined by the distance where the recession velocity is
c. Any sources beyond that distance are unobservable. The emission time
when that occurs is the inverse of the present Hubble parameter. A matter
dominated Universe with an R scaling as a power of t, in this case n = 2/3, is
assumed in the script “Cosmo_Light”. Printout from that script appears in
Figure 8.22.

Figure 8.22: Printout showing the Hubble parameters for a given choice of present Hubble parameter.

The photons always travel on a path where the local velocity is that of
light. Then the present reception and emission radial coordinates, r, are:

The reception time is to, while the emission time is t. Assuming a power law
behavior in t for R over the past history, the integrals can be performed and
the coordinates at emission and reception can be calculated as a function of
the z parameter (redshift).

Figure 8.23: The distance at emission and reception for light with redshift z. The distances are scaled
to the Hubble length where the recession velocity is c. The present distance at reception is Ror, while
the distance at emission is Rr.

The present distance to the source is Ror, while the distance to the source at
emission is Rr. There is a maximum present distance of 3cto which occurs at
large z. The emission distance does not increase monotonically with z, but has
a maximum. There is a maximum distance that can be observed because of
the expansion of the reception point. Note that light can also be received at
present distances greater than the Hubble length, also due to the expansion.
The Hubble length and velocity are both divided by n in the expanding
geometry. The Hubble length is ct/n. The results for a matter dominated
Universe, n = 2/3, are shown in Figure 8.23. Present distances larger than the
Hubble length are accessible.
 The look back times for reception and the emission times are shown in
Figure 8.24 as a function, again, of the redshift parameter, z. These plots are,
in principle, nothing new since the look back positions are already known as
shown in Figure 8.23. However, they give additional visual insights. The
emission time is the reception time at z = 0 by definition. The emission time
falls steadily as the reception time increases because light arrives from earlier
times. The look back time is then just the time of reception with respect to the
time of reception and increases monotonically.

Figure 8.24: Emission time and look back time as a function of the signal redshift. At high z the
emission time decreases steadily to zero while the look back time increases toward one.
Appendix
 Scripts for the Chapter on Classical
Mechanics

All the scripts are available to the user using the enclosed media. However, it
is useful to be able to quickly jump to a written version in order to see what
MATLAB commands are used. To that end, the script text for the chapter on
Classical Mechanics is enclosed below.

A.1. Angular Momentum

%
% Symbolic Angular Momentum
%
clear all;
help Angular_Momentum % Clear the memory
and print header
%
syms x y Lz t vx vy
% define variables as needed
%
fprintf('Find L for Symbolic x(t), y(t) Input \n')
%
iloop = 0;
irun = 1;
while irun > 0
%
krun = menu('Another x(t), y(t) Trajectory','Yes','No');
if krun == 2
irun = -1;
 break
end
%
if krun == 1
%
fprintf('Enter x(t) symbolic \n');
x = input(': ','s'); % Read input as a text string
fprintf('Enter y(t) symbolic \n');
y = input(': ','s');
%
m = 1;
vx = diff(x,sym('t'));
vy = diff(y,sym('t'));
Lz = x *vy - y *vx;
fprintf('Angular momentum (z-component) = \n');
Lz
%
tt = linspace(0,5);
for i = 1:length(tt)
t = tt(i);
xx(i) = eval(x);
yy(i) = eval(y);
LLz(i) = eval(Lz);
z1(i) = 0;
x1(i) = 0;
y1(i) = 0;
end
%
iloop = iloop + 1;
figure(iloop)
xmax = max(xx);
ymax = max(yy);
zmax = max(LLz);
xmin = min(xx);
ymin = min(yy);
 zmin = min(LLz);
for i = 1:length(tt)
%
plot3(xx(i), yy(i), z1(i),'o');
axis([xmin xmax ymin ymax zmin zmax])
%
title('x,y and Lz ')
xlabel('x')
ylabel('y')
zlabel('Lz')
pause(0.2)
hold on
plot3(x1(i),y1(i),LLz(i),'r*')
legend('x,y','Lz')
end
hold off
%
end
end
%

A.2. Classical Scattering

%
% Program to explore classical scattering
% where in 3 cases the cross section is the same
%
clear all; help Classical_Scatt; % Clear memory and print header
%
fprintf('All Three Scatterings have the Same Total Cross Section
\n')
%
iloop = 0;
%
b = linspace(0,1.0); % impact parameter
a = 1; % all results have sigma = pi*a^2
%
% hard sphere , radius a
% barrier, Vo with E = 2Vo
% n = sqrt(1-Vo ./E), bo = n/a
n = sqrt(1- 0.5);
bo = n .*a;
nw = sqrt(1 + 0.5);
%
for i = 1:length(b)
th_hs(i) = 0;
if b(i) < a
th_hs(i) = pi - 2.0 .*asin(b(i) ./a);
end
%
th_bar(i) = 0;
if b(i) < bo
th_bar(i) = 2.0 .*(asin(b(i) ./(n .*a))
- asin(b(i) ./a));
end
if b(i) > bo && b(i) < a
th_bar(i) = pi - 2.0 .*asin(b(i) ./a);
end
%
th_well(i) = 0;
if b(i) < a
th_well(i) = -2.0 .*(asin(b(i)
./(nw .*a)) - asin(b(i) ./a));
end
%
end
%
iloop = iloop + 1;
figure(iloop)
 plot(b,th_hs,'-b',b,th_bar,':r',b,th_well,'-.g')
title('Scattering Angle for Three Scattering Centers')
xlabel('b/a')
ylabel('\theta')
legend('Hard Sphere, 'Barrier','Well')
%

A.3. Impact Parameter and DCA

%
% Program to look at the relationship of impact parameter
% and distance of closest approach - force power laws
%
clear all; % Clear memory
help CM_b_DCA; % Print header
%
irun = 1;
iloop = 0;
%
fprintf('Particle with impact parameter b scatters off a force
center \n');
%
syms b a v vo m r alf aa
%
% central forces, L conserved, mo*vo*b = m * v * a
%
fprintf('a = vo*b/v, a is DCA, initial velocity vo, v at DCA \n')
%
fprintf('f(r) = m /r^a, V(r) = -m/(a-1)r^a-1 \n) % force and
potential
%
fprintf(' E conservation; vo = sqrt(v ^2 + 2 /((alf-1) *(r ^(alf-
1)))) \n')
%
while irun > 0
kk = menu(' Another value of the power law a ?
','Yes','No');
if kk == 2
irun = -1;
break
end
if kk == 1
%
ALF = input('Enter force law power: ');
syms b a v vo m r alf aa
ll = menu('Force Law ? 'attractive','repulsive');
if ll == 1
mm = -1;
else
mm = 1;
end
%
alf = ALF ;
aa = solve(a == sqrt(b^2 + mm * (2 *a ^2)/(vo ^2*(alf -
1) * (a ^(alf - 1)))),a)
%
b = 1;
vvo = linspace(0.1, 5);
for i = 1:length(vvo)
vo = vvo(i);
AAA(i) = eval(aa(1));
end
%
iloop = iloop + 1;
figure(iloop)
loglog(vvo,AAA)
title('DCA as a function of vo')
xlabel('vo')
ylabel('DCA/b')
%
 end
end
%

A.4. Foucault Pendulum

%
% Program to compute precession of a Focult pendulum
%
clear;
help foucault_exact;
%
% Clear memory and print header
%
syms
% k1 is natural
X1 frequency =
X2 w^2, k2 is
x1 2*omega*sin
x2 (phi)
k1
k2 ;
% phi is
latitude,
omega is
Earth
frequency
% pick time in
days, omega =
2*pi
% pendulum
precesses by
2*pi*sin(phi)
per day
% x is East and
y is North in
Earth system

%
fprintf('Focault pendulum, Gravity plus Coriolis Causes
Precession \n')
fprintf('D2x = -w^2x + (2*omega*sin(phi))*dydt \n')
fprintf('D2y = -w^2y - (2*omega*sin(phi))*dxdt \n')
%
[X1,X2]=dsolve('D2x1=-k1*x1+k2*D1x2','D2x2=-k1*x2-k2
*D1x1','x1(0)=0','x2(0)=1','Dx1(0)=0','Dx2(0)=1');
%
iloop = 0;
irun = 1;
while irun > 0
%
krun = menu('Another Set of Parameters to
Solve?','Yes','No');
if krun == 2
irun = -1;
break
end
%
if krun == 1
Tpen = input('Enter Pendulum Period (days): ');
w = (2.0 .*pi) ./Tpen; % convert to frequency
phi = input('Enter Latitude (degrees): ');
phir = (2 .*pi .*phi) ./360; % radians
omega = 2.0 .*pi; % Earth rotation
%
k1 = w .^2;
k2 = 2.0 .*omega .*sin(phir);
%
tt = linspace(0,2,200);
ntt = length(tt);
for i = 1:ntt
t = tt(i);
xxx1(i) = real(eval(X1));
xxx2(i) = real(eval(X2));
end
%
iloop = iloop + 1;
figure(iloop)
%
plot(tt,xxx1,'b-',tt,xxx2,'r-')
title('Exact x(t) and y(t) of Focault Pendulum - Earth Frame
')
xlabel('t')
ylabel('x, y')
%
iloop = iloop + 1;
figure(iloop)
plot(xxx1,xxx2,'-b')
title('Exact x vs y of Focault Pendulum - Earth Frame ')
axis([-1 1-1 1])
xlabel('x')
ylabel('y')
%
% approximate result
%
ta = linspace(0, tt(ntt),ntt);
%
% in absolute frame, pendulum has SHO,
%
yaf = sin(w .*ta);
%
% rotating - Earth - frame
%
wpre = omega .*sin(phir);
cp = cos(wpre .*ta);
sp = sin(wpre .*ta);
xear = sp .*yaf;
year = cp .*yaf;
%
iloop = iloop + 1;
figure(iloop)
for i = 2:length(ta)
%
x1(1) = xear(i-1);
y1(1) = year(i-1);
x1(2) = xear(i);
y1(2) = year(i);
 plot(x1,y1,'-')
axis([-1 1 -1 1])
title('Approximate Earth Frame - Precession, Movie')
xlabel('x')
ylabel('y')
pause(0.2)
hold on
end
hold off
%
iloop = iloop + 1;
figure(iloop)
for i = 2:length(ta)
%
x1(1) = xxx1(i-1);
y1(1) = xxx2(i-1);
z1(1) = 0;
x1(2) = xxx1(i);
y1(2) = xxx2(i);
z1(2) = 0;
plot3(x1,y1,z1,'-')
grid ON
axis([-1 1-11-14])
titled('Exact Earth Frame - Precession and Absolute
Frame, Movie')
xlabel('x')
ylabel('y')
pause(0.2)
hold on
plot3(0,yaf(i),3,'r*')
end
hold off
%
end
end
%

A.5. Spring Pendulum

function Spring_Pendulum
%
% Pendulum in plane with length variable due to spring connection
%
global g k m Lo
%
irun = 1;
iloop = 0;
%
while irun > 0
kk = menu('Pick Another Spring Pendulum?','Yes','No');
if kk == 2
irun = -1;
break
end
if kk == 1
%
BB = input('Enter k, m and Lo - [k,m,Lo]: ');
fprintf('Spring contributes as (Lo-r), >0 and <0 \n')
%
g = 9.8; % m/sec^2
k = BB(1);
m = BB(2);
Lo = BB(3);
%
wspr = sqrt(k ./m); % for SHM, period = 2*pi/w
wpen = sqrt(g ./Lo); % period for simple pendullum
fprintf('w for simple spring = %g and simple pendulum =
%g \n' , wspr, wpen)
%
AA = input('Enter Initial Positions , r and theta(deg) -
[r(0),theta(0)]: ');
fprintf('Initial Velocities Assumed to be Zero \n')
%
ro = AA(1);
tho = (AA(2) .*2 .*pi) ./360.0;
%
tspan = linspace(0,4 .*2.0 .*pi ./wpen);
% time frame - 4 pendulum periods
%
[t,y] = ode45(@pend_spring,tspan,[0 ro 0 tho]);
% vr, r vth, th - initial conditions
%
iloop = iloop + 1;
figure(iloop)
plot(t,y(:,2))
title('r vs t')
xlabel('t')
ylabel('r')
%
iloop = iloop + 1;
figure(iloop)
plot(t,y(:,4))
title('\theta vs t')
xlabel('t')
ylabel('\theta')
%
for i = 1:length(t)
yyy(i) = -y(i,2) .*cos(y(i,4));
xxx(i) = y(i,2) .*sin(y(i,4));
end
iloop = iloop + 1;
figure(iloop) xmax = max(xxx);
xmin = min(xxx);
ymax = max(yyy);
 ymax = 0;
ymin = min(yyy);
%
% movie
%
for i = 1:length(t)
plot(xxx(i),yyy(i),'o')
title('Spring - Pendulum Movie - x vs y')
xlabelC'x')
ylabel('y')
axis([xmin, xmax, ymin, ymax])
hold on
penx(1) = 0.0;
peny(1) = 0.0;
penx(2) = xxx(i);
peny(2) = yyy(i);
plot(penx,peny,'g')
plot(0.0,0.0,'r*')
pause(0.1)
hold off
end
%
iloop = iloop + 1;
figure(iloop)
plot(xxx,yyy,'-')
title('Spring - Pendulum - x vs y')
xlabel('x')
ylabel('y')
axis([xmin, xmax, ymin, ymax])
%
end
end
%---------------------------------------------------------
function dy = pend_spring(t,y)
%
global g k m Lo
dy = zeros(4,1);
% y1 = drdt, 2 = r, 3 = dthetadt, 4 = theta
%
ct = cos(y(4));
st = sin(y(4));
dy(1) = g .*ct + (k ./m) .*(Lo - y(2)) + y(2) .*y(3) .*y(3);
dy(3) = -(g .*st) ./y(2) - (2.0 .*y(1) .*y(3)) ./y(2);
dy(2) = y(1);
dy(4) = y(3);
%

A.6. Spherical Pendulum

%
% Solve for motion of a spherical pendulum
%
function Spherical_Pend
clear all;
help Spherical_Pend % Clear the memory and print
header
global Q w
%
fprintf(' Spherical Pendulum, r = 1, m = 1 \n');
fprintf(' Azimuthal Angular Momentum is Conserved - 1-d Problem
\n');
%
irun = 1;
iloop = 0;
while irun > 0
%
krun = menu('Another Pendulum IC','Yes','No');
if krun == 2
irun = -1;
 break
end
if krun == 1;
%
ww = input('Enter g/l for the pendulum: ');
w = sqrt(ww); % SHO
%
Q = input('Enter angular momentum, L/mr^2 : ');
tho = input('Enter initial angle (deg) : ');
tho = (2.0 .*pi .*tho) ./360.0 ;
vtho = input('Enter initial angular velocity (deg/sec) :
');
vtho = (2.0 .*pi .*vtho) ./360.0 ;
%
tspan = linspace(0, (2.0 .*2.0 .*pi) ./w);
% time frame - 4 pendulum periods
%
[t,y] = ode45(@sph_pend,tspan,[vtho tho 0]);
% y1 = dthdt, 2 = th, 3 =
phi, initial conditions
%
iloop = iloop + 1;
figure(iloop)
plot(t,y(:,2))
title('theta vs t')
xlabel('t')
ylabel('\theta')
%
iloop = iloop + 1;
figure(iloop)
plot(t,y(:,3))
title('phi vs t')
xlabel('t')
ylabel('\phi')
%
 iloop = iloop + 1;
figure(iloop)
for i = 1:length(y(:,2))
xx(i) = sin(y(i,2)) .*cos(y(i,3));
yy(i) = sin(y(i,2)) .*sin(y(i,3));
end
xmax= max(xx);
xmin = min(xx);
ymax = max(yy);
ymin = min(yy);
for i = 1:length(y(:,2))
plot(xx(i),yy(i),'o')
title('Movie for Spherical Pendulum')
xlabel('x')
ylabel('y')
axis([xmin xmax ymin ymax])
pause(0.2)
end
iloop = iloop + 1;
figure(iloop)
plot(xx,yy)
title('x vs. y')
xlabel('x')
ylabel('y')
end
end
end
%
% ---------------------------------------------------------
%
function dy = sph_pend(t,y)
%
global Q w
dy = zeros(3,1);
%
% y1 = dthdt, 2 = th , 3 = phi
%
dy(1) = -w .*sin(y(2)) + (Q .^2 .*cos(y(2))) ./(sin(y(2)) .^3);
dy(2) = y(1);
dy(3) = Q ./(sin(y(2)) .^2);
%
end
%

A.7. Driven Pendulum

%
% Program to compute a single pendulum, arbitrary angle, driven
% using MATLAB tools
%
function Driven_Pendul
%
clear all;
help Driven_Pendul % Clear the memory and print header
%
global vv Adr wdr
%
fprintf('Pendulum - Large Oscillations, Driven, 1-d Angle \n');
%
irun = 1;
iloop = 0;
%
while irun > 0
kk = menu('Pick Another Driven Pendulum?','Yes','No');
if kk == 2
irun = -1;
break
end
if kk == 1
%
% Set initial position and velocity of pendulum
%
tho = input('Enter Initial Angle (degrees): ');
tho = (tho .*pi) ./180.0; % Convert angle to radians
vo = input('Enter Initial Angular Velocity
(degrees/sec): ');
vo = (vo .*pi) ./180.0;
vv = input('Enter Damping in rad/sec: ');
fprintf('Take k/m = 1 - natural frequency units \n')
Adr = input('Enter Driving Amplitue: ');
wdr = input('Enter Driving Frequency: ');
%
tt = linspace(0, 12.0 .*pi);
%
% numerical solution using ODE tools
%
[t,y] = ode45(@pend_dr,tt,[vo tho]);
%
iloop = iloop + 1;
figure(iloop)
yy = y(:,1);
plot(t,yy,'-')
title('Angular Velocity')
xlabel('t(sec)')
ylabel('d\theta/dt')
%
iloop = iloop + 1;
figure(iloop)
yy = y(:,2);
plot(t,yy/'-')
title/'Angular Position')
xlabel('t(sec)')
ylabel('\theta')
%
 iloop= iloop + 1;
figure(iloop)
th = y(:,2);
for j = 1:length(t)
%
yy(j) = -cos(y(j,2)); xx(j) = sin(y(j,2));
end
xmax = max(xx); ymax = max(yy); xmin = min(xx);
ymin = min(yy);
for j = 1:length(t)
xL(1) = 0; yL(1) = 0; xL(2) = xx(j) ; yL(2) =
yy(j);
plot(xx(j),yy(j),'o',xL,yL,'-g')
hold on
plot(0,0,'r*')
axis([xmin xmax ymin ymax])
titled ('Movie of Angular Position')
xlabel('x')
ylabel('y')
pause(0.1)
hold off
end
%
end
end
%
function dy = pend_dr(t,y)
%
global vv Adr wdr
%
dy = zeros(2,1);
dy(1) = -sin(y(2)) - vv .*y(1) + Adr .*cos(wdr .*t);
dy(2) = y(1);
A.8. Random Walk

%
% Look at a random walk in 2-d
%
clear all;
help rand_walk % Clear the memory and print header
%
fprintf('Look at a Random Walk in 2-d n')
%
irun = 1;
iloop = 0;
%
while irun > 0
kk = menu('Pick Another Number of Random
Steps?','Yes','No');
if kk == 2
irun = -1;
break
end
if kk == 1
%
Ns = input('Enter Number of Steps of Size 1: ');
%
x(1) = 0;
y(1) = 0;
NN(1) = 1;
for i = 2:Ns
phi = 2 .*pi .*rand;
x(i) = x(i-1) + cos(phi);
y(i) = y(i-1) + sin(phi);
NN(i) = i;
end
iloop = iloop + 1;
figure(iloop)
%
xmax = max(x);
xmin = min(x);
ymax = max(y);
ymin = min(y);
%
% the movie first, to see how the stochastic
distance evolves
%
for i = 1:Ns
plot(x(i),y(i),'o')
hold on
title('Random Walk Steps')
xlabel('x')
ylabel('y')
axis([xmin xmax ymin ymax])
pause(0.1)
end
hold off
%
iloop = iloop + 1;
figure(iloop)
plot(x,y,'-')
title('Random Walk, Full Time Range')
xlabel('x')
ylabel('y')
axis([xmin xmax ymin ymax])
%
iloop = iloop + 1;
figure(iloop)
plot(sqrt(NN), sqrt(x .^2+y .^2),'-*')
title('Distance of Walk vs Number of Steps')
xlabel('sqrt(N)')
ylabel('radius')
%
 end
%
end
%

A.9. Rotating Hoop

%
% Hoop, radius = 1, Rotates with frequency w with mass point
located at
% theta, wrt vertical
%
function Rotating_Hoop2
%
clear all;
help Rotating_Hoop2 % Clear the memory and print header
%
% Initialize
%
global gg ww tho
fprintf('Hoop Rotates with frequency w, w^2 > g/a for equilibrium
not equal to 0 \n ')
%
gg = 9.8; % a = 1 m, pick w
aa = 1;
%
irun = 1;
iloop = 0;
%
while irun > 0
%
krun = menu('Another Frequency','Yes','No');
if krun == 2
irun = -1;
break
end
if krun == 1;
%
ww = input('Enter Frequency of Oscillation, > sqrt(g)
(~4): ');
tho = acos(gg /(ww*ww)); % equilibrium angle for this w
- rad
fprintf('Stable Angle for Small Oscillations = %g
(rad)\n' ,tho');
%
ro = input('Enter Initial Angle w.r.t. Stable Angle
(rad): ');
vo = input('Enter Initial Velocity (rad/sec):');
%
tspan = linspace(0, 5);
[t,y] = ode45(@Hoop,tspan,[vo ; ro]);
%
iloop = iloop + 1;
figure(iloop)
ymax = max(y(:,2));
ymin = min(y(:,2));
for i = 1:length(tspan)
plot(t(i),y(i,2),'*r')
title('Angle of Mass Point - Released Near
Equilibrium Point')
xlabel('t')
ylabel('\theta - \theta_o')
axis([0 max(tspan) ymin ymax])
pause(0.1)
end
%
iloop = iloop + 1;
figure(iloop)
plot(t,y(:,2),'-b')
title('Angle of Mass Point - Released at Equilibrium
Point')
xlabel('t')
 ylabel('\theta')
%
iloop = iloop + 1;
figure(iloop)
rr = linspace(-1, 1);
zz = sqrt(1- rr .^2);
%
% mass point
%
for i = 1:length(tspan)
zzm(i) = cos(tho+y(i,2));
rrm(i) = sin(tho +y(i,2));
xxx(i) = rrm(i) .*cos(ww .*t(i));
rrc = rr .*cos(ww .*t(i));
plot(rrc,zz,'g',rrc , -zz,'-g',xxx(i),-zzm(i)
,'*r')
hold on
plot(sin(tho) .*cos(ww .*t(i)), -cos(tho),'o')
title('Hoop Rotation, w, and Mass Point
Oscillatiion')
xlabel('x')
ylabel('y')
axis([-1.5 1.5 -1.5 1.5])
hold off
pause(0.1)
end
hold off
%
end
end
%
%---------------------------------------------------------
%
function dydt = Hoop(t,y)
% Equation of mass point on the hoop
%
global gg ww tho
dydt = zeros(2,1);
dydt = [-sin(tho+y(2)).*(gg-ww .*ww .*cos(tho+y(2))) ;
y(1)];
%

A.10. Tides
%
% Program to display tides
%
clear all; help Tides; % Clear memory and print header
%
fprintf('Lunar Tides a la Newton \n')
%
iloop = 0;
%
t = linspace(0,2 .*pi); % loction of the Moon
te = linspace(0,2 .*pi,20); % location on earth surface
cth = cos(t);
sth = sin(t);
%
rm = 2;
re = 1;
%
iloop = iloop + 1;
figure(iloop)
%
xm = rm .*cth; % Moon coordinates
ym = rm .*sth;
%
for i = 1:length(t)
% points for the circle - plots
 x1 = linspace(-re,re);
y1 = sqrt(1 - x1 .^2);
hold on
%
plot(x1,y1,'r-',x1,-y1,'r-')
axis([-2 2 -2 2])
plot(0,0,'g*')
%
plot(xm(i), ym(i),'o')
for j = 1:length(te)
ang = te(j) - t(i);
tide(j) =0.1 .*(3.0 .*cos(ang) .*cos(ang) - 1);
end
plot((re+tide) .*cos(te),(re+tide) .*sin(te) ,'b:')
pause(0.2)
hold off
clf
end
for i = 1:1
% points for the circle - plots
x1 = linspace(-re,re);
y1 = sqrt(1 - x1 .^2);
hold on
%
plot(x1,y1,'r-',x1,-y1,'r-')
axis([-2 2 -2 2])
plot(0,0,'g*')
%
plot(xm(i), ym(i),'o')
for j = 1:length(te)
ang = te(j) - t(i);
tide(j) =0.1 .*(3.0 .*cos(ang) .*cos(ang) - 1);
end
plot((re+tide) .*cos(te),(re+tide) .*sin(te) ,'b:')
title('Lunar Tides - Twice per Day')
 pause(0.2)
hold off
end
%

A.11. Least Action

%
% Least Action examples
%
function Lagrange_Act2
%
clear all;
help Lagrange_Act2 % Clear the memory and print header
%
% Initialize
%
global g delt itype yend1 yend2
%
iloop = 0;
irun = 1;
%
g = 9.8;
%
while irun > 0
kk = menu('Pick a Least Action,'Simple
Harmonic','Gravity','Quit');
if kk == 3
irun = -1;
break
end
if kk == 1
%
itype = kk;
tf = 2.0;
 vo = -0.1;
xo = 0;
fprintf(' SHO - time = 2.0 sec, vo = -0.1 m/sec, x(0) =
0 \n ')
%
t = linspace(0,tf,20);
%
nt = length(t);
delt = t(2) - t(1);
%
% exact trajectory
%
ye = xo .*cos(t) + vo .*sin(t); % k = m = 1, scale to
max amplitude
yend1 = ye(1); % fixed end points
yend2 = ye(nt-1);
%
% closer conversion
for i = 1:nt
ysmear(i) = ye(i) + (2.0 .*rand - 1) .* 0.02;
end
%
yysm = fminsearch(@Lla2,ysmear);
%
iloop = iloop + 1;
figure(iloop)
plot(t,ye,'-',t,yysm,'*r')
title('Trajectory for Simple Harmonic Motion')
xlabel('t(sec)')
ylabel('y(m)')
legend('Exact','Least Act')
%
end
if kk == 2
%
 itype = kk;
vo = 10;
tf = 1;
fprintf('Gravity - y(0) = 0, t(0) = 0, v(0) = 10
m/sec, tf = 1 sec \n')
%
t = linspace(0,tf,20);
%
nt = length(t);
delt = t(2) - t(1);
%
% exact trajectory
%
ye = vo .*t - (g .*t .*t) ./2.0;
yend1 = ye(1); % fixed end points
yend2 = ye(nt-1);
%
% closer conversion
for i = 1:nt
ysmear(i) = ye(i) + (2.0 .*rand - 1) .*
0.02;
end
yysm = fminsearch(@Lla2,ysmear);
%
iloop = iloop + 1;
figure(iloop)
plot(t,ye,'-',t,yysm,'*r')
title('Trajectory in Gravity Field')
xlabel('t(sec)')
ylabel('y(m)')
legend('Exact','Least Act')
%
end
end
%
%---------------------------------------------------------
%
function y = Lla2(x)
global g delt itype yend1 yend2
% evaluate action T, V along time
%
nt = length(x);
vla = diff(x); % velocity, kinetic and
potential, nt-1
vla = vla ./delt;
if itype == 1;
for i = 1:nt-1
Tla(i) = (vla(i) .^2) ./2;
Vla(i) = -x(i) .^2;
end
end
if itype == 2;
for i = 1:nt-1
Tla(i) = (vla(i) .^2) ./2;
Vla(i) = g .*x(i);
end
end
%
y = sum(Tla - Vla) .*delt;
%
% penalize end points not fixed
y = y + 1.0 .*(abs(x(1)-yend1) + abs(x(nt-1) - yend2));
%

A.12. Rocket Drag

%
function cm_rocket_drag
%
% Solve non-relativistic rocket, numerically
% allow for uniform gravity field and air drag
%
clear all;
help cm_rocket_drag % Clear the memory and print
header
%
% solve the rocket equation
%
global vo dmdt mo g dragg zo
%
g = 9.8; % accel at earth surface m/sec^2
re = 6.378 .*10 .^6; % earth radius - m
veq = (2.0 .*pi .*re) ./(24 .*3.6 .*10^3); % equatorial launch
velocity km/sec
rs = 1.5 .*10 .^11; % distance to sun - m
me = 6.0 .*10 .^24; % earth mass - kg
ms = 2.0 .*10 .^30; % sun mass, - kg
%
vorb = sqrt(g .*re); % orbital velocity - circular, low orbit
ve =sqrt(2.0 .*g .*re); % escape velocity for Earth ~ 11.2 km/sec
vs = ve .*sqrt(ms .*re ./(me .*rs)); % escape velocity to leave
solar system ~ 42 km/sec
%
% air density falloff with altitude - m
zo = 9140;
dragg = 0;
%
fprintf('Velocity, Satellite Low Circular Orbit (m/sec) = %g
\n',vorb);
fprintf('Escape Velocity - Earth (m/sec) = %g \n',ve);
fprintf('Escape Velocity - Solar System (m/sec) = %g \n', vs);
fprintf('Equatorial Launch Velocity (m/sec) = %g \n', veq);
%
irun = 1;
iloop = 0;
%
while irun > 0
kk = menu('Pick Another Drag Coefficient?','Yes','No');
if kk == 2
irun = -1;
break
end
if kk == 1
%
% total possible burn time T is mo/(dm/dt)
% payload ratio mp/mo = 1-tp/T, tp = burn time for this
payload
%
mo = input('Input the Rocket Mass (in 10^6 kg units) -
Saturn = 4x10^6 kg: ');
mo = mo .*10 .^6;
mp = input('Input the Payload Mass (in kg) - Saturn
Escape Module = 24610 kg: ');
vo = input('Input the Exhaust Velocity (in m/sec) -
Saturn = 2200 m/sec: ');
dmdt = input('Input Burn Rate (in kg/sec) - Saturn =
15000 kg/sec: ');
%
T = mo ./dmdt ; % max possible burn rate, with no
payload
tp = T .*(1 - mp ./mo); % burn time for this payload
tlift = -vo ./g + T; % t=0 is ignition, t = tl is lift
time, when acceleration > 0
%
fprintf('Maximum Burn Time (sec) = %g \n',T');
fprintf('Burn Time for This Payload (sec) = %g \n',tp);
%
tspan = linspace(tlift,tp,100);
g = 0;
[t,yfree] = ode45(@drag,tspan,[0 0]);
g = 9.8;
[t,yg] = ode45(@drag,tspan,[0 0]);
 %
dragg = input('Input Air Drag - Viscosity: ');
[t,ydrag] = ode45(@drag,tspan,[0 0]);
%
iloop = iloop + 1;
figure(iloop)
semilogy(t ./T,yfree(:,2),'-b',t ./T,yg(:,2),'-r',t
./T,ydrag(:,2),'-g')
legend('free','g','g+drag')
title('Rocket Motion With Gravity and Air Drag')
xlabel('Fractional Burn Time')
ylabel('y(m)')
%
iloop = iloop + 1;
figure(iloop)
semilogy(t ./T,yfree(:,1),'-b',t ./T,yg(:,1),'-r',t
./T,ydrag(:,1),'-g')
hold on
semilogy([-0.4, 0.4], [ve, ve],'-k')
semilogy([-.4 0.4 ], [vs, vs],':k')
legend('free','g','g+drag')
title('Rocket Motion With Gravity and Air Drag')
xlabel('Fractional Burn Time')
ylabel('v(m/sec)')
hold off
end
end
%
%---------------------------------------------------------
%
function dy = drag(t,y)
%
% rocket in uniform gravity field with air drag - vertical motion
%
global vo dmdt mo g dragg zo
%
dy = zeros(2,1);
%
m = mo - dmdt .*t;
dy(1) = (vo .* dmdt) ./m - g - ( dragg .*y(1) .*y(1) .* exp(-y(2)
./zo)) ./m;
dy(2) = y(1); % 2 position, 1 velocity
%

A.13. Two Stage Rocket

%
% Rockets - 2 stage vs. 1 stage
%
clear all;
help Two_Stage % Clear the memory and print header
%
% solve the rocket equation - no gravity
% compare 2 and 1 stage approaches for the same final payload
%
gg = 9.8; % accel at earth surface m/sec^2
re = 6.378 .*10 .^6; % earth radius - m
rs = 1.5 .*10 .^11; % distance to sun - m
me = 6.0 .*10 .^24; % earth mass - kg
ms = 2.0 .*10 .^30; % sun mass, - kg
%
vorb = sqrt(gg .*re); % orbital velocity - circular, low
orbit
ve =sqrt(2.0 .*gg .*re); % escape velocity for Earth ~ 11.2
km/sec
vs = ve .*sqrt(ms .*re ./(me .*rs)); % escape velocity to leave
solar system ~ 42 km/sec
%
fprintf('Velocity, Satellite Low Circular Orbit (m/sec) = %g
\n',vorb);
fprintf('Escape Velocity - Earth (m/sec) = %g \n',ve);
fprintf('Escape Velocity - Solar System (m/sec) = %g \n', vs);
%
irun = 1;
iloop = 0;
%
while irun > 0
kk = menu('Pick Another Two Rockets?','Yes','No');
if kk == 2
irun = -1;
break
end
if kk == 1
%
% vf-vi = vo*log(mf/mi) - valid generally
% m(t) = mi - dmdt*(t-ti), assume constant burn rate
%
mo = input('Input the Rocket Mass (in 10^6 kg units) -
Saturn = 4x10^6 kg: ');
mo = mo .*10 .^6;
mp = input('Input the Payload Mass (in kg) - Saturn
Escape Module = 24610 kg: ');
vo = input('Input the Exhaust Velocity (in m/sec) -
Saturn = 2200 m/sec: ');
dmdt = input('Input Burn Rate (in kg/sec) - Saturn =
15000 kg/sec: ');
m2 = input('Input Mass for Second Stage(in 10^6 kg
units) : ');
m2 = m2 .*10 .^6;
rat = input('Input the Plumbing to Fuel Mass Ratio for
Both Stages : ');
%
% defines the masses, same payload in both cases
% single stage
%
Mfu = (mo - mp) ./(1.0 + rat); % fuel is reduced by
plumbing deadweight
vf1 = -log((mp + rat .*Mfu) ./mo);
tf1 = Mfu ./dmdt;
%
% 2 stage assuming same plumbing ratio in both cases
%
mfu2 = (m2 - mp) ./(1.0 + rat); % second stage mass
mfu1 = (mo - m2) ./(1.0 + rat); % first stage, same
total mass = mo
mr2 = m2 - mp - mfu2;
mr1 = mo - m2 -mfu1;
tfu1 = mfu1 ./dmdt;
tfu2 = mfu2 ./dmdt;
vfu1 = -log((m2 + mr1) ./mo);
vfu2 = vfu1 - log((mp + mr2) ./m2);
%
T = mo ./dmdt ; % max possible burn time
%
tt = linspace(0,tf1); % one stage
%
% the free rocket
%
for i = 1:length(tt)
%
v1(i) = 0;
if tt(i) < tfl
v1(i) = - vo .*log(1.0 - dmdt .*tt(i)
./mo); % 1 stage
end
%
end
for i = 1:length(tt)
v2(i) = 0;
if tt(i) < tfu1 + tfu2
if tt(i) < tfu1; % first stage
v2(i) = -vo .*log(1.0 - (dmdt .*tt(i))
./mo); % first stage
else
 v2(i) = vfu1 .*vo - vo .*log(1.0 - (dmdt .*
(tt(i)-tfu1)) ./m2);
end
end
end
%
iloop = iloop + 1;
figure(iloop)
semilogy(tt,v1 ./vo,'-b',tt, v2./vo,'-g')
title('Rocket - Velocity in vo Units')
xlabel('Burn Time (sec)')
ylabel('velocity/vo')
legend('1 stage','2 stage')
%
hold on
%
semilogy(tt,vorb./vo,'r-',tt,ve ./vo,'r:',tt,vs ./vo,'r-
-')
legend('One Stage','Two Stages','Orbital Velocity,'Earth
Escape
Velocity, 'Sun Escape Velocity')
hold off
%
iloop = iloop + 1;
figure(iloop)
plot(tt ./T,v1 ./vo,'-b',tt ./T, v2./vo,'r:')
title('Rocket - Velocity in vo Units')
xlabel('Burn Time in T Units')
ylabel('velocity/vo')
legend('1 stage','2 stage')
%
end
end
%
A.14. Table Top
%
% Program to look at particles on a plane table, (r, theta) and
hanging
% below (z) with force g
%
function Table_Top
clear all; % Clear memory
help Table_Top; % Print header
%
global g m1 m2 L
irun = 1;
iloop = 0;
%
while irun > 0
kk = menu(' Another Set of masses, m1 and m2 ?
','Yes','No');
if kk == 2
irun = -1;
break
end
if kk == 1
%
m1 = input('Enter Mass on the Table (kg): ');
a = input('Enter m1 Initial Radius (m): ');
dadt = input('Enter m1 Initial Radial Velocity (m/sec):
');
m2 = input('Enter Mass Hanging Below the Table (kg) at r
= 0: ');
L = input('Enter m1 Angular Momentum (m^2/sec):');
g = 9.8; % m/sec^2
%
fprintf('Masses Connected by String, lenghth LL, z = LL-
r \n')
LL = input('Enter String Length (m): ');
%
tspan= linspace(0,10);
[t,y] = ode45(@top_table,tspan,[dadt a 0]);
% initial radial velocity and position
%
iloop = iloop + 1;
figure(iloop)
%
plot(t,y(:,2))
title('r vs t')
xlabel('t')
ylabel('r')
%
iloop = iloop + 1;
figure(iloop)
plot(t,y(:,l))
title('dr/dt vs t')
xlabel('t')
ylabel('dr/dt')
%
iloop = iloop + 1;
figure(iloop)
plot(t,LL - y(:,2))
title('z of m2 vs t')
xlabel('t')
ylabel('z')
iloop = iloop + 1;
figure(iloop)
plot(t,y(:,3))
title('polar angle vs t')
xlabel('t')
ylabel('\theta')
%
% movie
%
for i = l:length(t)
 yyy(i) = y(i,2) .*cos(y(i,3));
xxx(i) = y(i,2) .*sin(y(i,3));
zzz(i) = 0;
xx(i) = 0;
yy(i) = 0;
zz(i) = -(LL - y(i,2));
end
iloop = iloop + 1;
figure(iloop)
xmax = max(xxx);
xmin = min(xxx);
ymax = max(yyy);
ymin = min(yyy);
zmax = 0.1;
zmin = - LL;
%
% movie
%
for i = l:length(t)
plot3(xxx(i),yyy(i),zzz(i),'bs')
title('Plane and Hanging Weight')
xlabel('x')
ylabel('y')
zlabel('z')
axis([xmin, xmax, ymin, ymax, zmin, zmax])
grid
hold on
penx(1) = 0.0;
peny(1) = 0.0;
penz(1) = 0.0;
penx(2) = xxx(i);
peny(2) = yyy(i);
penz(2) = zzz(i);
plot3(penx,peny,penz,'g')
plot3(0,0,0,'*r')
 plot3(xx(i),yy(i),zz(i),'bs')
wtx(1) = 0.0;
wty(1) = 0.0;
wtz(1) = 0.0;
wtx(2) = xx(i);
wty(2) = yy(i);
wtz(2) = zz(i);
plot3(wtx,wty,wtz,'g')
grid
pause(0.2)
hold off
%
end
iloop = iloop + 1;
figure(iloop)
plot3(xxx,yyy,zzz,'b-')
title('Plane and Hanging Weight')
xlabel('x')
ylabel('y')
zlabel('z')
grid
axis([xmin, xmax, ymin, ymax, zmin, zmax])
hold on
plot3(xx,yy,zz,'b-')
plot3(0,0,0,'*r')
hold off
%
end
%
end
%
end
%---------------------------------------------------------
function dy = top_table(t,y)
%
global g m1 m2 L
dy = zeros(3,1);
%
% y1 = drdt, 2 = r , 3 = th
%
dy(1) = -g .*m2 + (L .^2 .*m1) ./(y(2) .^3);
dy(2) = y(1);
dy(3) = L ./(y(2) .^3);
%
end
%

A.15. Eccentric_Kepler
%
% Program to compute conic sections with different eccentricity
%
clear;
help eccentric_Kepler;
%
% Clear memory and print header
%
fprintf('Keplerian Orbits - defined by Eccentricity \n')
fprintf('r = ro*(1+e)/(1+e*cos(theta)) \n')
%
ro = 1;
th = linspace(0,2 .*pi,100);
st = sin(th);
ct = cos(th);
%
iloop = 1;
irun = 1;
while irun > 0
%
krun = menu(Άnother Eccentricity?','Yes','No');
if krun == 2
irun = -1;
break
end
%
if krun == 1
%
ecc = input('Enter Eccentricity, (0,2): ');
if ecc > 1
thasy = (360 .*acos( 1.0 ./ecc)) ./(2.0 .*pi);
fprintf('e > 1 - Aysmptotic Angle = %g \n', thasy);
end
r = ro .*(1.0 + ecc) ./(1+ecc .*ct);
x = r .*ct;
y = r .*st;
%
figure(l)
%
plot(x,y,'-' ,0,0, 'r*')
title('Orbit/Conic Section for All Eccentricities')
xlabel('x')
ylabel('y')
axis([-10 10 -10 10])
hold on
iloop = iloop + 1;
figure(iloop)
plot(x,y,'-' ,0,0,'r*')
title('Orbit/Conic Section for This Eccentricity')
xlabel('x')
ylabel('y')
%
end
hold off
end
 %

A.16. Non-Central Force

function Non_Central_Force
%
% orbits - Kelperian except that L is not conserved - driving
harmonic
% force
%
global pert w
%
irun = 1; iloop = 0;
%
while irun > 0
kk = menu('Pick Another Strength of Perturbation?',
'Yes','No');
if kk == 2
irun = -1;
break
end
if kk == 1
%
pert = input('Enter Strength of Perturbation for
Elliptical Orbit (~0.3): ');
w = input(Έnter Frequency, w, of Perturbation (~ 1): ');
a = input('Enter Initial radius (~ 4): ');
LL = input('Enter Initial Angular Velocity 1/a^3/2
Circular:');
fprintf('Angular velocity < circular is not stable and
hits r = 0 \n')
%
% circular orbits
%
Lc = sqrt(a);
vc = 1.0 ./(a .^1.5); % angular velocity 1/a^2/3
tauc = 2.0 .*pi .*(a .^1.5);
%
fprintf('Circular Orbits for Initial Radius a \n')
fprintf('L = %g, angular velocity = %g, period = %g
\n',Lc,vc,tauc)
%
tspan = linspace(0,4 .*tauc); % time frame - 4 circular
periods
%
[t,y] = ode45(@mech,tspan,[0 a LL 0]); % vr, r vth, th -
initial
conditions
%for unperturbed ellipse
%
iloop = iloop + 1;
figure(iloop)
plot(t,y(:,2))
title('r vs t')
xlabel('t')
ylabel('r')
%
iloop = iloop + 1;
figure(iloop)
plot(t,y(:,4))
title('\theta vs t')
xlabel('t')
ylabel('\theta')
%
for i = 1:length(t)
xxx(i) = y(i,2) .*cos(y(i,4));
yyy(i) = y(i,2) .*sin(y(i,4));
end
iloop = iloop + 1;
figure(iloop)
xmax = max(xxx);
xmin = min(xxx);
 ymax = max(yyy);
ymin = min(yyy);
%
% movie
%
for i = 1:length(t)
plot(xxx(i),yyy(i),'o')
title('Perturbed Orbit Movie - x vs y')
xlabel('x')
ylabel('y')
axis([xmin, xmax, ymin, ymax])
hold on
plot(0.0,0.0,'r*')
pause(0.1)
end
%
end
end
%---------------------------------------------------------
function dy = mech(t,y)
%
global pert w
dy = zeros(4,1);
% y1 = drdt, 2 = r, 3 = dthetadt, 4 = theta
%
dy(1) = (y(3) .^2) .*y(2) - 1.0 ./y(2) .^2;
dy(3) = (-2 .*y(1) .*y(3) .*y(2) + pert .*sin(w*t)) ./(y(2) .^2);
dy(2) = y(1);
dy(4) = y(3);
%

A.17. Euler_Angles
%
% Graphics for Euler Angles
%
clear all;
help Euler_Angles % Clear the memory and print header
%
% define variables as needed
%
% circle orientiation for coord system
phi = linspace(0,2 .*pi);
x = cos(phi);
y = sin(phi);
nph = length(phi);
for i = 1:nph
z(i) = 0;
end
%
% x axis
%
xax(1) = 0;
xax(2) = 1;
yax(1) = 0;
yax(2) = 0;
zax(1) = 0;
zax(2) = 0;
%
% z axis
%
xzx(1) = 0;
xzx(2) = 0;
yzx(1) = 0;
yzx(2) = 0;
zzx(1) = 0;
zzx(2) = 1;
%
% now successive rotations
%
ang =35 .*2.0 .*pi ./360.; % all 30 degrees;
sa = sin(ang);
ca = cos(ang);
D = [ca -sa 0; sa ca 0; 0 0 1];
C = [1 0 0; 0 ca sa; 0 -sa ca];
B = [ca sa 0; -sa ca 0; 0 0 1];
%
for i = 1:nph
xx(1) = x(i);
xx(2) = y(i);
xx(3) = z(i);
xD(:,i) = D*xx';
xCD(:,i) = C*D*xx';
xBCD(:,i) = B*C*D*xx';
end
%
% x axes %
for i = 1:2
xxx(1) = xax(i);
xxx(2) = yax(i);
xxx(3) = zax(i);
xaxD(:,i) = D*xxx';
xaxCD(:,i) = C*D*xxx';
xaxBCD(:,i) = B*C*D*xxx';
end
%
for i = 1:2
xxzx(1) = xzx(i);
xxzx(2) = yzx(i);
xxzx(3) = zzx(i);
xzxD(:,i) = D*xxzx';
xzxCD(:,i) = C*D*xxzx';
xzxBCD(:,i) = B*C*D*xxzx';
end
%
plot3(x,y,z,'b-'); % initial orientation
hold on
plot3(xax,yax,zax,'b-')
plot3(xD(1,:),xD(2,:),xD(3,:),'-.r') % rotate on (x,y) plane
plot3(xCD(1,:),xCD(2,:),xCD(3,:),'g--')
plot3(xBCD(1,:),xBCD(2,:),xBCD(3,:),':k')
plot3(0,0,0, 'r*')
plot3(xaxD(1,:),xaxD(2,:),xaxD(3,:),'r-.')
plot3(xaxCD(1,:),xaxCD(2,:),xaxCD(3,:),'g--')
plot3(xaxBCD(1,:),xaxBCD(2,:),xaxBCD(3,:),'k:')
plot3(xzx,yzx,zzx,'b-')
plot3(xzxD(1,:),xzxD(2,:),xzxD(3,:),'r-.')
plot3(xzxCD(l,:),xzxCD(2,:),xzxCD(3,:),'g--')
plot3(xzxBCD(l,:),xzxBCD(2,:),xzxBCD(3,:),'k:')
xlabel('x')
ylabel('y')
zlabel('z')
title('Euler Angles')
axis([-1 1-11 -0.4 0.4])
hold off
%

A.18. Top Forceless

%
% Top Motion with no forces
%
clear all;
help Top_Forceless % Clear the memory and print header
%
% define variables as needed
%
fprintf('Top Motion Depends on Top moments of I \n')
fprintf('and Angular Velocity w \n')
%
iloop = 0;
irun = 1;
while irun > 0
%
krun = menu('Another Set of Top I3, I','Yes','No');
if krun == 2
irun = -1;
break
end
%
if krun == 1
%
I1 = 1;
Iz = input('Enter Top Shape w = 1, Iz, I1 = 1 (~1.5) :
') ; % moments of 2 principal axes
%
tho = input('Enter Initial Top Polar Angle (deg) (~20) :
');
tho = (tho .*2 .*pi) ./360; % theta (0) - rad, must
be non-zero(sleeping)
%
w = 1;
w3 = w .*cos(tho);
wt = w .*sin(tho);
omeg = ((Iz - I1) .*w3) ./I1;
%
tt = linspace(0, 4 .*pi,50);
w1 = wt .*cos(omeg .*tt);
w2 = wt .*sin(omeg .*tt);
for i = 1:length(tt)
w3t(i) = w3;
end
%
iloop = iloop + 1;
 figure(iloop)
xmax = max(w1);
ymax = max(w2);
zmax = max(w3t);
xmin = min(w1);
ymin = min(w2);
zmin = min(w3t);
for i = 1:length(tt)
x1(1) = 0;
x1(2) = w1(i);
y1(1) = 0;
y1(2) = w2(i);
z1(1) = 0;
z1(2) = w3t(i);
%plot3(w1(i), w2(i), w3t(i),'o');
plot3(xl,yl,zl,'-r*')
axis([xmin xmax ymin ymax 0 zmax])
title('Top Precession')
xlabel('wx')
ylabel('wy')
zlabel('wz')
pause(0.2)
hold on
end
hold off
%
end
end

A. 19. Top_Euler
%
% Top Motion using explicit solution, Euler Equations
%
function Top_Euler
%
clear all;
help Top_Euler % Clear the memory and print header
global B Iz I1 ws
%
% define variables as needed
%
fprintf('Top motion depends on spin, Top moments of I, mass \n')
fprintf('and initial posiition and velocity (4 of) \n')
fprintf('Special Case of "drop", theta = phi = phidot = 0, only
thetao \n')
%
iloop = 0;
irun = 1;
while irun > 0
%
krun = menu('Another Set of Top Parameters - Euler Angle
Solution?','Yes','No');
if krun == 2
irun = -1;
break
end
%
if krun == 1
B = input('Enter m*g*h -Torque at \theta (~1.0) = 0 :
');
I1 = 1;
Iz = input(Έnter Top Shape Iz, I1 = 1 (~1.5): ');
% moments of 2 principal axes
%
tho = input('Enter Initial Top Polar Angle (deg) (~20) :
');
tho = (tho .*2 .*pi) ./360; % theta (0) - rad, must
be non-zero(sleeping)
phio =0; % phi(0)
phio = (phio .*2 .*pi) ./360;
psio = 0;
thdto =0; % thetadot(0)
%
phidto = input('Enter Initial Derivative of Top
Azimuthal Angle - Defines
Nutation (deg) (~0,0.2,0.8) : ');
%
ws = 1.5 + sqrt(4 .*I1 .*B .*cos(tho))/Iz;
%
% time scale for precession is ~ 1/sqrt(B), integrate
eqs for theta(t),
% phi(t)using Euler Angles in body axes, transform to
space axes
%
range = 1.0 ./sqrt(B);
%
tspan = linspace(O, 8 .*range,200);
% y(1) = phi, y(2)= phid, y(3) = th, y(4) = thd, y(5) =
psi, y(6) = s
% s is a constant of the motion = top spin, [tt,yt] =
ode45(@top_euler_eqs,tspan,[phio, phidto, tho, thdto,
psio] );
% constant of the motion
psidot = ws - yt(:,2) .*cos(yt(:,3));
%
iloop = iloop + 1;
figure(iloop)
plot(tt ./range,yt(:,3).*360 ./(2.0 .*pi))
title('\theta - nutation')
xlabel('time in 1/sqrt(B) units')
ylabel('theta degrees')
%
iloop = iloop + 1;
figure(iloop)
plot(tt ./range,yt(:,1).*360 ./(2.0 .*pi))
title('\phi - precesion')
xlabel('time in 1/sqrt(B) units')
ylabel('phi degrees')
%
iloop = iloop + 1;
figure(iloop)
plot(yt(:,1).*360 ./(2.0 .*pi) ,yt(:,3).*360 ./(2.0
.*pi));
title('\phi vs \theta')
xlabel('\phi (deg)')
ylabel('\theta(deg)')
%
iloop = iloop + 1;
xx = sin(yt(:,3)) .*cos(yt(:,1));
yy = sin(yt(:,3)) .*sin(yt(:,1));
zz = cos(yt(:,3));
xmax = max(xx);
xmin = min(xx);
ymax = max(yy);
ymin = min(yy);
zmax = max(zz);
zmin = min(zz);
for i = l:length(xx)
plot3(xx(i),yy(i),zz(i),'r*')
axis([xmin xmax ymin ymax zmin zmax])
title('CM Motion of the Top')
xlabel('x')
ylabel('y')
zlabel('z')
hold on
pause(0.1)
end
hold off
plot3(xx,yy,zz)
axis([xmin xmax ymin ymax zmin zmax])
title('CM Motion of the Top')
 xlabel('x')
ylabel('y')
zlabel('z')
end
end
%
%---------------------------------------------------------
%
function dy = top_euler_eqs(t,y)
%
global B Iz I1 ws
dy = zeros(5,1); % one constant of motion, top spin
%
% y(1) is phi of top, y(2) is phidot , y(3) is th, y(4) is thdot,
y(5) is psi
dy(1) = y(2);
dy(3) = y(4);
dy(2) = (Iz .*ws - 2.0 .*I1 .*y(2).*cos(y(3))).*y(4)
./(I1*sin(y(3))) ;
dy(4) = (B - (Iz .*ws - I1 .*y(2).*cos(y(3))).*y(2))
.*sin(y(3))/I1 ;
dy(5) = ws - y(2).*cos(y(3));
%
 Graphical User Interface (GUI)

All the scripts used in this volume are command line driven in that the user
makes input via the command window using text. In fact, MATLAB has a
suite of tools allowing the use of GUIs in order to wrap the scripts and have a
graphical input of user choices for a specific problem. There are several
“help” items in the “product help” giving GUI details. In particular there is a
video demo which illustrates how to build a GUI using the MATLAB
supplied tools.
When creating a GUI, there is a palette of possible GUI items. The choices
are shown below. Having chosen an object, the “property editor” allows the
GUI designer to set the properties, for example, the “Tag” to identify the
object or the String defining the text on the pushbutton.
The use of a GUI rather than a command line driven input dialogue is
illustrated with two examples. These examples are “wrappers” for the script
in Section 4.1 (Leaky Box), and 5.13 (Damped, Driven Oscillator), First the
problem of gas leaking from a container is wrapped with a GUI which
features editable text boxes which receive user input rather than command
line input. Default values are given as a text string. The number of molecules,
their temperature and the size of the aperture are supplied by the user, or the
defaults can be run using the start pushbutton.
The gas problem is one where number, temperature and “volume” for the
ideal gas are defined but the “pressure” must be approximately evaluated by
running the script. In another problem, the harmonic oscillator with damping
being driven harmonically by a wave with specified amplitude and frequency
can be done in real time. In that case, sliders are more appropriate. The slider
limits are provided in text boxes and the calculation starts and stops with
pushbuttons. During the calculation the user can choose the values using the
sliders. That implementation makes the response of the oscillator very
intuitively realizable. A snapshot of the GUI for this problem is shown below.
Figure GUI.1: Tools for GUI design in MATLAB. The pallette is on the left and the property editor
for a pushbutton is on the right.
Figure GUI.2: GUI for the leaking box problem showing the use of editable text, text boxes and a
pushbutton. The result for the mean momentum of the remaining and escaped gas appears as annotated
text on the figure of the box.

Figure GUI.3: Snapshot of the waveform for a harmonic oscillator with a particular user chosen set of
slider parameters for damping and driven amplitude and frequency.
 References

Numerical Methods for Physics, 2e, Alejandro L. Garcia, Prentice Hall, 2000,
ISBN: 0-13-906744-2. This text has companion script which is very useful as
a starting point for some exercises.
Numerical Recipes in Fortran 77. The Art of Scientific Computing, W. H.
Press, S. A. Teukolsky, W. T. Vetterling, and B. P. Flannery, Cambridge
University Press, 1986, ISBN: 0-521-43064-X. This book is a veritable
encyclopedia of numerical methods.
There are many excellent textbooks available for use as references. However,
as technology has evolved, the increasing use of online resources is at hand.
Therefore, a complete set of paper textbooks is not quoted here. Rather the
search engines and compiled online knowledge bases are invoked.
Google is an enormously useful search engine and many specific searches
will yield a great variety of information. Wikipedia has a large store of
interesting Physics topics, and a search through them will very often start the
user on a good path. Indeed, a particular article often has many links that can
be followed deeper into the topic. An example of the first page of a search for
“Foucault pendulum” is shown below. There are additional links and
references provided that give the user a very good reference experience.
Indeed, while looking at the MATLAB scripts for this text, the user can easily
dip into the online resources and gain further knowledge.
Foucault pendulum
From Wikipedia, the free encyclopedia
(Redirected from Focault pendulum)
This article is about the physics expeπment and implement. For the novel by Italian philosopher
Umberto Eco. see Foucaults Pendulum
The Foucault pendulum
(English pronunciation: /fu:’kou/foo-KOH; French pronunciation: [fu’ko]), or Foucault’s pendulum,
named after the French physicist Léon Foucault, is a simple device conceived as an experiment to
demonstrate the rotation of the Earth. While it had long been known that the Earth rotated, the
pendulum in 1851 was the first simple proof of the rotation in an easy-to-see experiment. Today,
Foucault pendulums are popular displays in science museums and universities.

A Foucault pendulum installed at the California Academy of Sciences. The Earth’s rotation causes the
trajectory of the pendulum to change over time, knocking down pins at different positions as time
elapses and the Earth rotates
 Index

absolute space, 23, 26

acceptor, 122
adiabatic, 140, 141, 212
Angstrom, ix
angular momentum, 18, 19, 21, 28, 29, 46, 48, 50, 53, 171, 196, 223, 227, 230, 233
antenna, 92
aperture, 102, 104, 150, 151
apogee, 198
array, 156
asymptote, x, 48
asymptotic, 3
attenuation, 95
beam, 11
Bessel, 131
“besselh”, 167
“besselj”, 114, 167
binary pair, 206
binding energy, 168
black hole, 218, 223, 227
body frame, 51, 54
Born approximation, 173
bosons, 170
boundary conditions, 64, 65, 71, 75, 76, 79, 80, 99, 103, 133, 135, 137, 149, 150, 193, 240
bound states, 235
Brite-Wigner, 8
burn time, 44
C++, viii
calibrate, 238
carriers, 122
Casimir effect, 192
central forces, 18, 21, 48
chaotic, 1, 30, 31, 34
charge to mass ratio, 59, 60
circular aperture, 151
collinear points, 204, 205
command line, viii, 303
compressibility, 109
Compton scattering, 10, 11
conduction, 116–120, 124, 126
band, 116
conductor, 78
confinement, 84
conservative forces, 61
continuum, 175
contour, 64
coordinate time, 232, 234, 235
Coriolis, 23, 25
correspondence principle, 166
cross section, 20, 174
current loop, 73
cutoff frequency, 97, 99
cylinder, 68
damping, 30, 31, 145, 148, 160–162
“daughters”, 191
Debye, 105
deep, 114–117, 128, 168, 186
degenerate states, 171
demo, 303
density of states, 117
depletion, 121
deuterium, 88, 89
dialogue, 12
dielectric constant, 75, 76, 78–82
“diff”, 61
diffraction, 150
diffusion, 109, 121, 124, 127, 128, 130, 135
coefficient, 128
diode, 120
dipole antenna, 92, 156
dipole radiated angular distribution, 90
dipole radiation, 90
dispersion, 95, 100, 114, 115, 117
dissipation, 96, 148, 150
distance of closest approach, 22
donor, 122
doping, 124
Doppler, 143, 144, 181, 183, 206
broadening, 181
drag, 40
driven, 30, 103, 133, 146, 147, 158, 160, 162
oscillator, 158
drum, 148–150
“dsolve”, 3, 5, 24, 146, 160, 192, 222, 232, 241
eccentricity, 48, 49, 198, 199
“eig”, 142
eigenfunctions, 142–144, 172, 174, 186–188
eigenvalues, 142
Einstein, 105
electric dipole decay, 180
electron, 59, 60, 110, 116–121, 124, 172, 175–177, 182, 186, 188, 193, 215–218
ellipse, 229
EMF, 82–85
energy band, 189
energy density, 218
energy gap, 118
entropy, 108, 109
equations of motion, 38
equatorial, 230, 232–234, 236
equilibrium solution, 137
equipotential, 38, 64, 80, 81, 204
escape velocity, 42
Euler angle, 51
Euler transformation, 11, 12
eV, ix
event horizon, 231
extrasolar planet, 206
“factor”, 65
feature size, 193
Fermi level, 117
Fermi pressure, 215, 217, 218
fermions, 170
fft, 17
flow, 102
“fminsearch”, 39–41, 168, 189
Foucault, 24, 308
Fourier, 71
transform, 14
four-vector, 144
frame dragging, 231, 233, 235
freeze, 128, 129
freezing, 128
fusion, 84–86, 88, 89, 212, 215, 244
Gaussian, 8
general relativity, 36
geodesic, 222–226, 228, 231–234
Google, 307
“gradient”, 61
gravity assist, 200
gravity wave, 236
GUI, 303
Hamiltonian matrix, 186
Hammer, 150
Hankle, 167
heat, 128–130, 133–138, 141, 213
equation, 133
Heisenberg uncertainty, 163
Helmholtz, 65
Hermite, 164
“hist”, 9
holes, 120
home wiring, 101
Hubble parameter, 244
ideal gas law, 210, 211
identical particles, 170
“ifft”, 17
image charge, 75
impact parameter, 20–22, 223–225, 228, 229
impurity states, 120
induced EMF, 83
inertial system, 33
initial conditions, 1, 29, 30, 123, 132, 133, 135, 160, 208, 223, 232
“integral”, 106
intercept, 4
intuition, x
“inv”, 191
inverse, 14
ionic core, 109
ionic crystal, 109, 111, 181
ITER, 88
junction, 118
Jupiter, 200
Kepler, 48
Kerr metric, 230, 233
Kerr rotation, 232
Koenig-Penny, 188
Lagrange points, 202
Lagrangian, 34, 35
Lane-Emden, 210, 211
Laplace, 14–17, 68, 71, 72
transform, 14
latitude, 25
launch window, 198
least action, 40, 41
lifetime, 179
light deflection, 223
limiting case, 27
line width, 179
Lissajous, 156
look back time, 246
Lorentz, 144
magnetic bottle, 84
magnetic field, 59, 65, 73–76, 83–86, 97, 98
Malthus, 3
Maple, 164
Maxwell-Boltzmann, 88, 89, 102, 108
membranes, 159
metric, 224, 226, 228, 230, 233
“min”, 129
missile, 4–6
molecule, 108
moment of inertia, 53
momentum, 12
Monte Carlo, 5, 9, 11, 12, 102, 108
movies, 36, 84, 158
multiple scattering, 13, 14, 33, 51
muon, 11
natural frequency, 145
natural width, 180, 181
neutron, 218
Newton, 36, 38
nodal lines, 159, 160
non-central, 48
force, 48
number density, 119, 120, 215, 216, 220
numerical methods, 307
nutation, 56, 58
Ode, 45, 26, 28, 35, 41, 47, 50, 57, 85, 122, 123, 139, 205, 207, 211, 217, 228
Ohms law, 99
Oppenheimer-Volkoff-Tolman, 240, 241, 243
optimization, 44
palette, 303
payload, 42–44, 198
“pdepe”, 133
pendulum, 23–34
perigee, 198
perihelion, 228
periodic potential, 181, 188
perturbed solutions, 172
phonon, 105
photoelectric effect, 175
photon orbit, 226, 227
piano, 150
Planck, ix
plasma, 84, 86, 88, 90
plot3, 47
plumbing, 43, 44
Poisson equation, 120
polytropic, 209
population growth, 3
precession, 53–55, 58
pressure, 110
proper time, 227
property editor, 303
psec, ix
“quad”, 89
quadrupole, 66, 68–70
quarter wave, 92, 156
“quiver”, 80
radial geodesic, 226
radiation length, 33
radiation pressure, 215
rand, 7
random walk, 32
range, 7, 9, 10, 13, 30, 34, 66, 101, 110, 156, 213, 228, 237, 238
reception time, 244
redshift, 233, 244–247
re-entrant, 25
orbit, 230, 237
relative phase, 92–94, 156
relaxation time, 135
resonances, 183
rest energy, 244
rigid body, 51, 53
Roche, 202, 203
rocket, 40
stages, 280
scattering, 20, 173
Schrödinger, ix
equation, 166, 185, 186
Schwarzchild radius, 223
script, x
semiconductor, 116
series, 62
shallow, 114, 115, 205
skin depth, 99
slider, 303
slinky, 139
solar constant, 218
solar data, 211
solar system, 206
“solve”, 21
“sort”, 202
sound velocity, 148
space frame, 51, 52, 57
species, 191
specific heat, 105–107, 140
standing wave, 160
stark effect, 171
stellar pressure, 238
stellar pulsation, 213
step driver, 145
stochastic, 32, 33
Sun, 84
supernova, 244
surface charge, 82
symbolic math, viii
symbolic solution, 3, 223
target, 4–6, 19, 177
temperature, 128
test body, 237
“thin” barriers, 183, 184
Thomson, J. J., 60
three body orbit, 207
tidal forces, 237
tide, 36
top, 53
trajectory, 1
transfer orbit, 196
transforms, 14
transient, 31
transverse electric, 97, 98
two stage, 43
valence band, 116
viscosity, 103
of air, 41, 113
water wave, 114–117
waveguide, 96
wavelength, 220
waves, 114
white dwarf, 215
Wikipedia, 307
wobble, 206
wrapper, 303
Young’s modulus, 140