CHEM 373

Physical Chemistry: Quantum Chemistry

Table of Contents
Part I. Light – wave or particle?.......................................................................................................................3
1. Newton: light is corpuscular...........................................................................................................................3
2. Young: light is a wave.....................................................................................................................................4
3. Maxwell: light is an electromagnetic wave.....................................................................................................7
Part II. Thermal radiation and Planck’s postulate............................................................................................7
1. Blackbody radiation........................................................................................................................................7
2. Planck’s postulate.........................................................................................................................................10
3. Boltzmann distribution (Laboratory 1, Part 1)..............................................................................................12
4. Photoelectric effect......................................................................................................................................17
Part III. Schrödinger theory of quantum mechanics.......................................................................................20
1. De Broglie’s hypothesis................................................................................................................................20
2. The Schrödinger equation............................................................................................................................21
3. Operators and commutators (Tutorial 1)......................................................................................................22
4. Solving the Schrödinger equation by separating the variables.....................................................................24
5. Statistical interpretation of wavefunction....................................................................................................25
6. Time independent Schrödinger equation.....................................................................................................26
7. Solution of the time independent Schrödinger equation for a free particle.................................................26
8. Solving a second order differential equation with constant coefficients (Tutorial 2)...................................27
9. Creating a wavepacket (Laboratory 1, Part 2)...............................................................................................28
10. The uncertainty principle............................................................................................................................48
11. Expectation values (Tutorial 3)...................................................................................................................49
12. Particle in the infinite potential well...........................................................................................................52
13. Modeling cyanine dye using the particle-in-the-box approach; superposition principle (Laboratory 2)....58
14. Quantum mechanical harmonic oscillator..................................................................................................65
15. Solving the Hermite differential equation (Tutorial 4)................................................................................67
16. Modeling quantum harmonic oscillator (Laboratory 3)..............................................................................68
17. Quantum harmonic oscillator exercises (Tutorial 5)...................................................................................75
18. Barrier potentials........................................................................................................................................76
19. Scanning tunneling microscopy..................................................................................................................89
Appendices...................................................................................................................................................89
Appendix A. Differentiation and integration rules............................................................................................89

1
Appendix B. Complex numbers........................................................................................................................90

2
Part I. Light – wave or particle?
1. Newton: light is corpuscular
In an attempt to disprove the wave theory of light, Newton conducted a two prism
experiment. First, white light was sent through a prism and separated into what
we know today as the visible light spectrum. Next, a second prism was used to
combine light of different colours back into white light.

Newton’s experiments showed that

a) white light is made of different colours;

b) if one colour is chosen, it cannot be split any further;

c) even if an object appears of a certain colour in sunlight, one can make it appear
any colour by illuminating with a “pure” spectral colour in a dark room;

d) if two objects appear of two different colours in sunlight, and then we pick light
of one of these two colours and illuminate the objects in a dark room, they would
not appear the same.

3
All of the above lead to an important conclusion that Newton arrived at: colour is
a property of light reflected off of an object, not the object itself. It is important to
remember that Newton’s two prism experiment did not prove the existence of
photons. The corpuscular nature of light was simply postulated by Newton.
When travelling between two points in different media, light does not always
select the shortest path but the path that takes the least time. In 1662 Fermat
was the one who showed this. This principle of the least time was accepted as a
useful mathematical concept but also a concept that has no physical meaning
because light does not have free will and, therefore, may not choose one way or
another.
Fermat's Principle can be used to understand the law of reflection and refraction.

2. Young: light is a wave

Experiment: two narrow parallel slits separated by a very small distance; to the
left from the slits a source of light; to the right from the slits a light sensitive
screen. Every time a photon would hit the screen, it would leave a mark on it.
Both the light source and the screen need to be at a comparatively large distance
from the slits.
Thomas Young’s argument: if light is composed of particles, they should pass
through separate slits, and create two bright spots on the screen. The positions of
the bright spots should be aligned with the position of the slits.

However, if light is a wave, the waves will reach the slits first. According to
Huygens' principle, each slit would become a source of new spherical waves:

4
Let us track the waves originating from the slits at the same time and moving
along the black lines in the figure above. Before reaching the screen, each wave
travels distance d . When they meet at the screen, the amplitude of one wave will
be the same as the amplitude of the other wave. Waves are known to interfere
with each other. Upon interference, their amplitudes add, and the resulting wave
is the sum of the two. When two waves are in phase (depicted in the schematic
below), upon interference their sum amplitude will be twice the amplitude of the
original waves. This is known as constructive interference.

In the context of the double slit experiment, the constructive interference should
be observed in the centre of the screen to which both waves travelled the same
distance d .

5
Are there any other positions on the screen where the constructive interference is
expected to occur?
Let us figure out a general condition for constructive interference. We want the
crests and the troughs of two waves to coincide at the screen. This will happen
when the waves originated at the slits either travel the same distance d , or travel
distances that differ by an integer number of wavelengths λ :

Thus, the condition for constructive interference may be expressed as

At the locations of constructive interference, the amplitude of the resulting wave

is expected to be the sum of the amplitudes of the individual waves.
The higher the wave amplitude is (the taller its crests and the deeper its troughs),
the higher its energy is (at least classically). Therefore, the spots on the screen
where two waves interfere constructively should appear as bright spots.
Similarly, we can figure out the condition for destructive interference, when the
amplitudes of two waves cancel out:

6
The locations on the screen that satisfy this condition should appear dark because
the wave sum amplitude becomes zero.
Coming back to Thomas Young’s argument, if light is a wave, we should observe
an interference pattern on the screen that will consist of bright spots, where the
waves interfere constructively, and dark spots, where the waves interfere
destructively.
Thomas Young performed the experiment, observed the interference pattern, and
concluded that light is indeed a wave.
3. Maxwell: light is an electromagnetic wave
James Clerk Maxwell put together the laws of electricity and magnetism and
showed that electromagnetic waves propagate at the same speed as the speed of
light.
It was one of Maxwell’s greatest achievements to connect electricity, magnetism,
and light.
Thus, by the end of the nineteenth century, light was accepted not just
as a wave but as an electromagnetic wave, and Maxwell’s theory of
electromagnetism became a new theory of light.

7
Part II. Thermal radiation and Planck’s postulate
1. Blackbody radiation
Thermal radiation is radiation emitted by a body due to its temperature. At
moderate temperatures most bodies are visible to us not because they emit light
but because they reflect light. That is why we do not see objects in a dark room.
As the temperature increases, the bodies become self-luminous and start glowing
in the dark.
Imagine that we are heating an iron poker in a flame. When the poker is at a
relatively low temperature, it radiates heat but does not glow. Therefore, it is easy
to get burnt by such a poker – there is no visible sign that it may be dangerous to
touch it. What does it mean in terms of radiation, in terms of its emission
spectrum (wavelengths)?
As the temperature increases, the poker becomes dull red, then it becomes bright
red, and at very high temperatures it becomes blue-white colour. What happens
to the emission spectrum of the poker as its temperature goes up?

In general, the wavelengths present in the spectrum depend on the object.

However, blackbodies, i.e. bodies that absorb all radiation, produce thermal
spectra of a universal character. An ideal blackbody does not reflect light and
appears black at low temperatures, thus, the origin of the name. An object coated
with black paint is not a perfect blackbody but a good approximation of it.
We, on the other hand, will consider a different object that is also a blackbody.
The object is simply a small hole that connects the inside of some cavity with the
outside.
Why does this hole have the property of a blackbody? Radiation penetrating
through the hole is reflected back and forth by the walls of the cavity and,
eventually, gets absorbed by these walls. By making the hole small enough, we
can ensure that essentially all radiation is absorbed. Thus, the hole must have the
properties of the surface of a blackbody.

8
If we reverse the process and have the walls of the cavity heated, they will start
emitting thermal radiation that will fill the cavity. Only small fraction of this
radiation will be able to escape through the hole. Thus, the hole, which, as we just
established, has the properties of a blackbody, must emit thermal radiation. The
emission spectrum corresponding to this radiation is called the blackbody
spectrum.

The shape of the blackbody spectrum, i.e. the spectrum of the energy density ρT
as a function of frequency ν , is predicted in the classical theory by the Rayleigh-
Jeans formula:
2 2
8πv 8πv
ρ (v )= 3
< E≥ 3 kT
c c
The important point to note is that the average energy is kT , where k is the
Boltzmann constant and T is the absolute temperature. This comes from the law
of equipartition of energy, which states that for a system of gas molecules in
thermal equilibrium at temperature T , the average kinetic energy of a molecule
kT
per degree of freedom is . In our case, the standing waves inside the cavity
2
also have one degree of freedom, which is their electric field amplitudes.
kT
Therefore, their average kinetic energy is . A physical system that has one
2
degree of freedom and executes simple harmonic oscillations in time is known to
have a total energy that is twice its average kinetic energy. Therefore, each
standing wave in the cavity has the average total energy equal to kT .
9
In the figure below the Rayleigh-Jeans formula is plotted in red. The experimental
spectrum of the blackbody radiation is plotted on the same graph in blue.
7.E-17

6.E-17

5.E-17

4.E-17
r(u)

3.E-17

2.E-17

1.E-17

0.E+00
0 1E+14 2E+14 3E+14 4E+14 5E+14
u, Hz

The discrepancy between the theory and experiment is apparent. In the limit of
low frequencies the classical theoretical spectrum (red line), i.e. the one built
using the Rayleigh-Jeans formula, approaches the experimental results. As the
frequency increases, the theoretical prediction goes to infinity due to the
quadratic dependence of the energy density on frequency. The experiments, on
the other hand, show (blue line) that the energy density remains finite. Moreover,
the energy density goes to zero at very high frequencies. Thus, the Rayleigh-Jeans
formula summarizes the failure of classical physics. It suggests that regardless of
the temperature, there should be infinitely high energy density at infinitely high
frequencies (or infinitely short wavelengths). This unrealistic behaviour predicted
by the classical theory at high frequencies was termed the ultraviolet
catastrophe by Ehrenfest.
2. Planck’s postulate
To resolve the discrepancy between theory and experiment, Planck considered
the possibility of violating the law of equipartition of energy.
From the figure above, it is clear that the law gives satisfactory results for small
frequencies. Thus, we can write that the average total energy approaches kT
when the frequency approaches 0:
lim ¿ E>¿=kT ¿
υ→ 0

The discrepancy at high frequencies could be eliminated if we assume that for

whatever reason the average energy is not the same for all frequencies; and it
approaches 0 as the frequency approaches infinity:
10
lim ¿ E >¿=0 ¿
υ →∞

In other words, Planck proposed that the average energy of the standing waves
inside the cavity is a function of frequency. This, obviously, contradicts the Law
of Equipartition of Energy which states that the average energy is independent of
frequency.
The Law of Equipartition of Energy arises from the Boltzmann distribution:
−E
e kT
P(E)=
kT
where E is the energy, k is Boltzmann’s constant, and T is the absolute
temperature.
One can plot the above distribution for a given temperature. The one for T =293 K
is shown below.

As you can see, this is a continuous distribution, i.e. any values of energy are
allowed.
A probability distribution must be normalized in a way that
∞

∫ P( E)dE=1
0

This requirement stems from the fact that all probabilities must add to 1. You can
easily verify that the Boltzmann distribution is indeed properly normalized.

11
The next question we may ask is what the average energy is. In other words, what
is the average of the energy distribution?

Thus, to find the average energy we need to calculate the integral

∞
¿ E≥∫ ❑
0

Planck proposed that not all energy levels are allowed. If this is the case, the
continuous Boltzmann distribution turns into a discrete one:

The above distribution does abide by the Boltzmann distribution formula. As a

matter of fact, the figure above was built using the formula. However, Planck’s
hypothesis is that we can only calculate the probability values for some allowed
energies. Which ones are allowed? We do not know yet. The important message
though is that some energy values are forbidden. i.e. the system can only change
its energy in some chunks. This idea clearly violates the classical theory in
general, and the continuous Boltzmann distribution that allows all energies in
particular.
In the figure above the allowed energy values are depicted by red diamonds. They
are spaced equally, i.e. the energy gaps between the adjacent allowed energy
values are the same. In other words, the chunks of energy by which the system
can change its energy are of equal magnitude. The interval ∆ E between the

12
allowed energy values was arbitrarily chosen to be 2.02 ×10−21 J , which is
approximately one half of kT .
3. Boltzmann distribution (Laboratory 1, Part 1)
Note that in the laboratory manuals (typed in green in this set of notes)
all the formulas that you need to memorize before the laboratory
session are highlighted in bold font. You need these formulas for closed
book quizzes that will be administered during the laboratory exercises.
Finding the average of a distribution
Imagine that there are 100 students in a class. On a quiz that was marked out of
10 points, 15 students got 10 points, 30 students got 9, 25 students got 8, 20
students got 7, and 10 students got 6.
We can represent the above data (distribution of grades) in a form of a table:
Grade Number of students
0 0
1 0
2 0
3 0
4 0
5 0
6 10
7 20
8 25
9 30
10 15
Or we can represent the same data as a graph:

13
 30

25
Number of students

20

15

10

0
0 1 2 3 4 5 6 7 8 9 10
Grade

What is the average grade for this class distribution?

If all 100 students got 10 points, the average would just be 10. If 50 students got
10 and the other 50 got 9, we would intuitively say that the average should be
exactly between 10 and 9, i.e. 9.5; but how is this number obtained numerically?
We need to multiply the grade points by their corresponding probabilities and
add. Since we have the equal probability of getting 9 points (50%) and 10 points
(50%), the average grade is
¿ grade> ¿ grade value 1 × probability 1+ grade value 2 × probability 2
¿ grade> ¿ 9 ×0.5+10 × 0.5=9.5
Although sometimes it may be convenient to think of probabilities in percentages
(30% chance of rain tonight), in statistics the probability does not have units.
Furthermore, the total probability must always be 1. 50% chance corresponds to
the probability of 0.5.
Note that the angle brackets are used to denote average. We will maintain this
notation for average values throughout the course.
For the distribution in the table and the graph above, the probability of getting
grade i is calculated as
number of students with grade i
P(grade=i)=
total number of students ∈class
Grade Probability
0 0/100
1 0/100
2 0/100
3 0/100

14
 4 0/100
5 0/100
6 10/100
7 20/100
8 25/100
9 30/100
10 15/100
You can easily verify that all probabilities in the second column add to 1.
We are ready to write the formula for the average of a distribution
N
¿ x≥∑ x i × P(x i)
i=1

where ¿ x >¿ is the average value of a discrete distribution containing N different

values of x ( x i) and P(x i ) is the probability of obtaining the value x i. Note that
individual probabilities P(x i ) must add to 1:
N
P=∑ P (x i)=1
i=1

For the distribution in question:

10 20 25 30 15
¿ grade> ¿ 6 × +7 × +8 × +9 × +10 × =8.2
100 100 100 100 100
For a continuous distribution the sum needs to be replaced by the integral:
∞
¿ x≥∫ x P ( x ) dx
0

In the Excel file in tab “Part 1”, in column H calculate the probability values P ( En )
that correspond to the energy values in column G. The discrete distribution will
appear on the graph with the continuous Boltzmann distribution. Sum the
calculated probabilities in cell H22.
Next, in column I calculate the product of the allowed energy values (column G)
and their corresponding probabilities (column H). Add the calculated values in cell
I22.
You can now calculate the average energy for this discrete distribution shown
above using the following formula:
∞

∑ En P ( En)
¿ E≥ n=0∞
∑ P ( En)
n=0

15
The numerator of this formula is identical to the formula that we used to calculate
the class average. However, we also need to divide the result by the sum of all
probabilities. Why?
Calculate the value ¿ E>¿ in cell I25 and compare it to the kT value.
Next, we are going to choose ∆ E to be 5 times kT . In this case, the distribution will
look like the one in the figure:

In this case, we only see two allowed energy values (at this scale, obviously. If we
wanted to, we could calculate an infinite number of probability values P ( E ) for
larger and larger values of energy E , i.e. keep extending the horizontal axis
further and further) but this would not change our result significantly.
In the Excel file, you will need to repeat the calculations for this case – populate
columns L and M, calculate the required values in cells L15 and M15, and finally
calculate the average value of energy in cell M18. Compare this value to kT as
well.
Note, that in the Excel file we use 9 values compared to the 2 values shown in the
figure above. This is done to illustrate that the horizontal axis representing energy
has no upper boundary. We are free to calculate an infinite number of
probabilities corresponding to higher and higher values of energy. However, these
values will not change the value of the average energy drastically as they are
getting smaller and smaller. You can verify this by deleting the last five values
from your calculations and observing a change in the value of ¿ E>¿ in cell M18.
Finally, you can formulate the trend.

16
In order to fix the discrepancy between the theoretical and experimental
observations of the blackbody radiation, we want to have ¿ E>→ kT for small
frequencies, and ¿ E>→ 0 for large frequencies.
We just showed that by selecting small ∆ E (∆ E ≪ kT ¿ , we obtain ¿ E>¿ close to kT ;
and by selecting large ∆ E (∆ E ≫ kT ), we can obtain ¿ E>¿ close to 0 . This implies
that ∆ E should increase with frequency.
Planck chose the simplest, linear dependence and suggested that the energy
difference is proportional to frequency:
∆ E∝ν
If we write it as an equation, we will obtain
∆ E=hν ,
where h is the proportionality constant.
Planck determined the value of the proportionality constant by looking for a
number that would produce the best fit of the theoretical and experimental data.
The value that he obtained was very close to the currently accepted one of
−34
6.626 ×10 J s. This famous constant is now called Planck’s constant.

Formula analysis:
What happens when hν ≪ kT ?

What happens when hν ≫ kT ?

In both cases, we recover the desired outcome.

Substituting this formula into the formula for the energy density of the blackbody
spectrum, we obtain
2
8πν hν
ρ ( ν ) dν = 3 hν
dν
c kT
e −1

17
This is Planck's blackbody spectrum. The experimental results are in complete
agreement with Planck's formula for all temperatures.
You should remember that Planck did not change the Boltzmann distribution. He
changed the character of that distribution from continuous to discrete.
4. Photoelectric effect
Photoelectric effect, ejection of electrons from metal surface by light, was
discovered by Heinrich Hertz at the end of the 19 th century.
Experiment: two electrodes are located inside a tube in vacuum and one of the
electrodes is exposed to light. When the energy from light is absorbed by
electrons in the cathode and the absorbed energy is sufficient for the electrons to
escape, they can be detected as photocurrent using the ampermeter.

Let us conduct the experiment.

First, we irradiate the cathode with light of some arbitrary frequency ν . The
amperemeter shows no current. We increase the light intensity, no current.
Increase the intensity further – no current. Only when we increase the frequency
of light above a certain threshold ( ν> ν threshold ), we start detecting photocurrent.
Above the threshold, the photocurrent does depend on light intensity: the higher
the intensity – the higher the photocurrent.

18
Next, we start applying a positive bias to the anode. The photocurrent increases
and plateaus at some point. Intuitively, this makes sense – positive bias attracts
electrons, and, therefore, more of them reach the electrode.
When we reverse the bias, the negatively charged electrode starts repulsing the
photoelectrons and the photocurrent is reduced. At some point, it drops to zero
which implies that no electrons reach the electrode. In other words, the potential
energy associated with the electric potential V stop exceeds the kinetic energy of
electrons:
e V stop > K e
where V stop is the negative potential applied to the anode, e is the charge of an
electron, and K e is the kinetic energy of electrons.
We can irradiate the photocathode with light of higher intensity. In this case, we
will see a higher maximum photocurrent (larger blue circles in the figure below).
However, the stop potential (V stop ) will not change. This implies that the
maximum kinetic energy of photoelectrons does not depend on light
intensity. This contradicts the classical wave theory of light, according to which
the electric vector ℇ of the light wave increases in amplitude with the light
intensity. Since the force applied to the electron is eℇ , the kinetic energy of the
photoelectrons should also increase with the light intensity. Next, the classical
theory predicts that photoelectric effect should be possible for any frequency of
light provided that the light is intense enough to give enough energy needed to
eject the photoelectrons. However, a characteristic cut-off frequency is observed
experimentally. Below this frequency, the photoelectric effect is not observed, no
matter what the light intensity is.

19
We can resolve these issues if we assume that light exists in bundles of energy,
later named photons by G.N. Lewis, each having the energy content equal to
E=hν
Here E is the energy [ J ], ν is the frequency [ s−1 ], and h is Planck’s constant [ J s ].
It was Einstein who argued that for an electron to be ejected it needs to absorb
energy that is higher than the amount of work required to remove it from the
metal.
hν=W + K
Thus, energy can only take on discrete values that are multiples of h. When an
electron in metal absorbs a photon that has energy that is higher than the work
function of metal W (i.e. energy that is needed to escape), that energy is used to
leave the metal (W ) and the rest becomes the kinetic energy K of the electron.
In this case, it is clear why light of certain wavelengths cannot eject electrons no
matter how high its intensity is – if the frequency of light times Planck’s constant
is smaller than the work function (W ), the electron simply does not gain enough
energy to escape.
According to Einstein’s formula, the electromagnetic radiation is quantized and
occurs in finite "bundles" of energy which we call photons. The quantization
implies that a photon of any given frequency (wavelength) always has the same
amount of energy. For example, a photon of blue light with a wavelength of 450
nm will be of energy:
8
c −34 3× 10
E=h =6.626 ×10 −9
λ 450 ×10
20
Photons occur in quantized chunks and you cannot have half a photon of blue
light.
The available frequency is, of course, continuous, and has no upper or lower
limits.
This is an important concept. “Energy is quantized” does not mean that you can
only have photons with energy values in some increments. Energy can take upon
any values because you can have light of any frequency (wavelength). λ can be
500 nm or 500.23 nm or 500.78561 nm, etc. “Energy is quantized” means that
when you have photons of a given wavelength (500.78561 nm, for the sake of
argument), you cannot break individual photons into smaller chunks of energy.
We can discuss this in terms of light intensity as well. If you have light with a
wavelength of 500.78561 nm, you have a bunch of photons, each with an energy
of
8
c −34 3 ×10
E=h =6.626 ×10 −9
λ 500.78561× 10
When you reduce the intensity of such light, all you are doing is reducing the
number of photons. If you reduce light intensity to one photon only, you cannot
reduce it any further. It is either one photon, or no light at all. That is the meaning
of “energy is quantized”.
Thus, physicists had no choice but to settle with the idea that in some
experiments light behaves as a wave and in some experiments it exhibits a
particle-like behaviour. Wave-particle duality was accepted as a physical
phenomenon.
Part III. Schrödinger theory of quantum mechanics
1. De Broglie’s hypothesis
We have shown that in certain experiments light displays a particle like behavior.
Lois de Broglie proposed that if light (a wave in classical physics) can display a
particle like behavior, then classical particles should likewise display wave-like
behavior. This proposal brought an interesting symmetry to nature. De Broglie
connected the wavelength, λ and momentum, p:
h
λ=
p
where h is Planck’s constant.
h
λ= =
Js
[
p kg m s−1
=
][ ]
Nm
N
=[m]

Just a reminder, p=m v

where m is the mass of a particle and v is the velocity.
Thus, de Broglie introduced the concept of matter waves. Intuitively, it is difficult
to combine a particle and a wave into a single concept. When we think of a
21
particle, we always imagine something that is localized. A wave, on the other
hand, is always spread out. This is not the only philosophically unpleasant issue
with matter waves. More are coming but we will deal with them in due time.
2. The Schrödinger equation
Before we introduce the equation itself, let us recap what we have discussed up to
this point.
1. In order to explain the photoelectric effect, Einstein hypothesized that light
is made up of particles that we call photons nowadays. Furthermore, he
assigned energy to a photon
E=hν
2. Since light manifests itself as waves in certain experiments and as particles
in other experiments, wave-particle duality was accepted as a phenomenon
of nature.
3. In an attempt to bring certain symmetry to nature, De Broglie postulated
that if light may manifest itself both as waves and particles, particles should
also be able to manifest themselves as waves. He assigned wavelength to
particles that have momentum p:
h
λ=
p
Thus, De Broglie postulated the existence of matter waves – a completely
novel concept that did not exist in classical theory.
This brief summary brings us to the next point. If matter waves do exist, we
should have an equation whose solutions would be wavefunctions representing
these matter waves.
Let us consider a few potential candidates for such wavefunctions.
We will start with the wavefunctions that are solutions of wave equations in
classical physics. These wavefunctions most certainly have familiar forms:
Ψ ( x ,t )=¿
Ψ ( x ,t )=¿
Are they good candidates to represent matter waves? These functions are not
localized, instead, they are spread out in space. This may be fine for a particle
with a defined momentum. If the exact momentum is known, the uncertainty in
position is infinite (this is a direct consequence of the Heisenberg uncertainty
principle that we will discuss later).
Another potential cause for concern is that both functions are zeros for certain
values of x and t . To find these values is easy – all you need to do is to make the
argument of the functions (kx −ωt ) equal to nπ for the sine and (2 n+1)π /2 for the
cosine. The wavefunction disappearing for certain values of x is not too
troublesome – we can always explain it by saying that the particle is not at that
22
position. However, there is nothing we can do to reconcile the fact that both
functions are zeros for certain values of t . This would imply that there exist times
when matter disappears because the wavefunction disappears.
We are left with no choice but to abandon the option of using real trigonometric
functions to describe matter waves. Instead, we are going to try a complex
exponential function of the kind
Ψ ( x ,t )=e i(kx−ωt )
At first sight it is not obvious why an exponential function would be descriptive of
a wave. However, we can show that it is not a bad choice using Taylor’s series.

To summarize, the wavefunction of the form Ψ ( x ,t )=e i(kx−ωt ) does satisfy the formal
requirements to represent matter waves. For more information on complex
numbers, see Appendix C at the end of this document.
3. Operators and commutators (Tutorial 1)
To continue with the story, we need to familiarize ourselves with two concepts –
operators and commutators.
Operators are neither functions nor constants. The notation O ^ Ψ ( x , y , z , t ) should not
be treated or read as widehat {O} multiplied by Ψ left (x,y,z, t right ) . This would be
^ Ψ ( x , y , z , t ) means that O
incorrect. O ^ operates on Ψ ( x , y , z , t ), i.e. an operator is a set of
instructions embedded in the equation on what we need to do with the function
that it operates on. From now on, we will always use a hat ( O ^ ) to indicate
operators in equations.
Write down the result of operation in the last column:
^)
Operator (O Function ( f ) ^f
O

√… x +3
2
sin ⁡(…) x
∂… t×e
x

∂x
∂… f ( x , t ) =sin ( t ) e
x

∂t
A special case of an operator is a position operator ^x . When the position operator
operates on functions of x , it gives another function of x as follows:
23
^x f ( x ) ≡ xf (x)
Note that it follows from this equation and successive applications of the operator
that
^x k f ( x ) ≡ x k f (x)
If the potential V (x , t) can be written as series expansion in terms of x , then the
potential energy operator is
^ (x , t)f (x , t)≡ V (x ,t) f (x , t) .
V
A major difference between ordinary algebra and operator algebra is that
numbers obey the commutative law of multiplication, i.e. ab=ba , where a and b are
any numbers, but operators do not necessarily do so and, in general, order of
operation matters.
We can define the commutator of operators ^A and ^B as
[ ^A , B^ ] ≡ ^A B^ −B^ ^A
^
A ^B− ^B ^
A is just another operator. If ^ A , then [ ^
A ^B= ^B ^ ^ ] =0 and we say that the
A,B
operators ^ A and ^B commute.
d
A=3 x and ^B= . Do these
1. Calculate the commutator for the operators ^
dx
operators commute?

^
d
2. Calculate the commutator for the operators and ^x . Do these operators
dx
commute?

d
A=x and ^B= . Do these operators
3. Calculate the commutator for the operators ^ 3
dx
commute?

d
4. Calculate the commutator for the operators ^
A= and ^B=5 x 2+ 3 x + 4. Do these
dx
operators commute?

24
5. Calculate the commutator for the operators ^x and ^p: [ ^x , ^p ] . Do these operators
commute?

We now come back to the quest of finding the wave equation for matter waves.
We suspect that the solution of the equation may be the wavefunction
i(kx−ωt )
Ψ ( x ,t )=e .

The time dependent Schrödinger equation:

^ ∂
H Ψ ( x , y , z , t ) =iℏ Ψ ( x , y , z ,t )
∂t
Ψ ( x , y , z , t ) is a wave function, a function that Schrödinger found to describe matter
waves postulated by Louis de Broglie. Since we live in a 3 dimensional world, in
general, the wave function depends on three coordinates and time.
^ is an operator. H
H ^ stands for Hamiltonian and the operator is called the full
energy operator.
The Schrödinger equation is the equation that tells us the form of the
wavefunction Ψ ( x , y , z , t ), if we know what the Hamiltonian looks like, i.e. if we
know the form of the potential V ^ (x , t).
4. Solving the Schrödinger equation by separating the variables
What is a solution of a differential equation? It is always a function.
If we plug the function (the solution) into the equation and differentiate
accordingly, the right hand side should be equal to the left hand side.
In the case of the time dependent Schrödinger equation
2
−ℏ2 ∂ Ψ ( x , t ) ^ ∂
2
+ V ( x ) Ψ ( x , t )=iℏ Ψ ( x , t )
2m ∂x ∂t
we need to find the wavefunction Ψ ( x ,t ) that depends on both the spatial
coordinate and time.
25
5. Statistical interpretation of wavefunction
The wavefunction is complex. This behavior is forced upon us by the Schrödinger
equation that relates the first order time derivative to the second order space
derivative.
The fact that the wavefunction is complex makes it apparent that we should not
give it a physical existence. We should not attempt to find any correspondence
between the wavefunction and any physical waves or anything measurable at all
because a complex quantity cannot be measured by any instrument. The real (in
a non-mathematical sense) world is the world of real (in a mathematical sense)
quantities. Therefore, we should not ask what is waving. The wavefunctions are
mathematical constructs that serve as computational devices and are
meaningful only in the context of the Schrödinger theory.
If the wavefunction is not measurable, what is its physical meaning? The answer
was given by Max Born, who proposed the statistical interpretation of the
wavefunction. This is another postulate of quantum mechanics.
If Ψ ( x ,t ) is the wavefunction describing a particle, then the probability to find this
particle (measure it) at position x at time t is given by
¿
P(x , t)=Ψ ( x , t ) Ψ ( x , t )
¿
where Ψ ( x , t ) is the complex conjugate of Ψ ( x ,t ).
This is not too bizarre of an assumption. It seems logical to say that the particle
and the corresponding wavefunction must be associated in space. In other words,
the particle must be at some location where the wavefunction has an appreciable
amplitude. We cannot equate the probability and the wavefunction simply
because the probability must be real (in a mathematical sense) and positive; and
the wavefunction is complex. On the other hand, the wavefunction times its
complex conjugate is always real and positive.
¿
We can show that Ψ (x , t)Ψ (x , t) is necessarily real and either positive or zero.

26
Considering that the space is continuous and the wavefunction is also continuous,
it is more appropriate to talk about the probability density rather than the
probability. The probability density P(x , t)dx is the probability of finding the
particle in some infinitesimal space between x and d x . If you integrate the
probability density over the entire space (in one dimensional case x ∈(−∞ ,+ ∞)¿,
you will obtain the total probability. Since the total probability should always be 1,
for the properly normalized wavefunction we have
+∞

∫ Ψ ¿ Ψ dx=1
−∞

Requirements for eigenfunctions:

6. Time independent Schrödinger equation

Solving the time dependent Schrödinger equation resulted in two equations – one
that depends only on time and the other that depends only on the spatial
coordinate. The latter is called the time independent Schrödinger equation:
^
H ψ ( x )=Eψ (x )
The time independent Schrödinger equation is an eigenequation, i.e. an equation
in which the operation on a function gives the same function times a constant:
when you operate with the Hamiltonian on the function ψ ( x ), you obtain the same
function ψ ( x )multiplied by the constant E . This constant, energy, is called an
eigenvalue and the function itself is called an eigenfunction.
The solution of the time independent Schrödinger equation is the eigenfunction
ψ ( x ) that describes a state that the system may be found in. In addition to the
eigenfunction, the solution of the equation contains the eigenvalue E that is the
energy that the system in question has in that state.
7. Solution of the time independent Schrödinger equation for a free
particle
At this point, we will move to considering systems for which we can find the exact
solutions of the Schrödinger equation.
The simplest time-independent Schrödinger equation is the one for the case
V ( x )=const . A particle moving under the influence of a constant potential is free
because the force acting on it is
−dV (x )
F= =0.
dx
This is true regardless of the value of the constant. We can always choose an
arbitrary additive constant to make V ( x )=0. This is a legitimate step considering

27
that in order to get the potential energy we need to integrate the force and an
arbitrary constant always appears during integration.
We know that classically a free particle is either at rest or moving with the
constant momentum p. In both cases its total energy E is constant.
According to de Broglie, the momentum may be found from
h hk
p= = = ℏk
λ 2π
The only energy that the particle has is the kinetic energy (remember that the
potential is zero):
2 2 2
p ℏ k
E= =
2m 2m

There are several points to take into consideration. This is a one-dimensional

problem, i.e. it only makes sense to discuss the position of the electron along the
horizontal axis. The vertical axis does not represent a spatial coordinate. The
vertical axis is energy.
Since the potential is zero everywhere, the potential term in the time independent
Schrödinger equation is also zero, and the equation is simplified to
2
−ℏ2 ∂ ψ ( x )
=Eψ (x)
2 m ∂ x2
This is a second order differential equation with constant coefficients, i.e. an
equation of the kind
2
d ψ (x ) d ψ (x)
a +b +c ψ (x)=0
d x2 dx
8. Solving a second order differential equation with constant coefficients
(Tutorial 2)
Solve a second order linear differential equation
2
d ψ (x ) d ψ (x)
a +b +c ψ (x)=0
d x2 dx

28
The general solution of an ordinary (non-partial) second order differential equation
contains two arbitrary constants. The reason is that obtaining the solution
amounts to performing two successive integrations to remove the second
derivative. Each step yields a constant of integration.
You may recall encountering a similar situation in solving Newton’s equation of
motion, which is also a second order differential equation. In that case, two
arbitrary constants are the initial position and momentum.

Thus, the general solution is a linear combination of two linearly independent

solutions ψ 1 ( x ) =eikx and ψ 2 ( x ) =e−ikx, which contains two arbitrary constants:
ikx −ikx
ψ= A ψ 1 ( x )+ B ψ 2 ( x )= A e +B e
Linearly independent simply means that neither ψ 1 ( x ) nor ψ 2 ( x ) can be obtained by
multiplying the other function by a constant. For instance, the functions e ikx and
5 e are linearly dependent because one can be obtained by multiplying the other
ikx

by 5; but the functions e ikx and e−ikx are linearly independent.

9. Creating a wavepacket (Laboratory 1, Part 2)

Part 2.1. Creating wavepackets by adding two waves.
Imagine that a particle is moving along the x axis. According to the de Broglie
postulate, there exists a matter wave associated with the particle. Intuitively we
understand that the matter wave, which is designated by the symbol Ψ ( x ,t ),
should look similar to the one shown in the figure below, i.e. the amplitude of the
matter wave must be modulated in such a way that it is non-zero only over some
finite region of space in the vicinity of the particle. This condition appears to be
necessary because somehow matter waves must be associated in space with
29
particles that they describe. It is also clear that the matter wave must move along
the x axis with the same velocity as its corresponding particle. The figure below is
a snapshot of the matter wave at some time t .
1.0
Y(x,t)

0.8

0.6

0.4

0.2

0.0
0 1 2 3 4 5 6 7 8 9 10
-0.2 x

-0.4

-0.6

-0.8

-1.0

However, you saw in lecture, that the velocity of the de Broglie wave is half the
velocity of the particle.
This apparent paradox may be resolved if instead of individual waves we will
consider wave packets.
Localizing a wave by creating a wave packet
Let us start with a well-known traveling wave equation:
Ψ ( x ,t )=Ψ 0 sin(kx−ωt )
where Ψ 0 is the amplitude, ω is the angular frequency, and k is the wave vector.
The angular frequency is related to the ordinary frequency by
ω=2 π ν
The wave vector is related to the wavelength by
2π
k=
λ
A wave may be localized in space by invoking the superposition principle.
Mathematically, the superposition principle may be written as
Ψ ( x ,t )=∑ Ψ 0 ( k n ) sin(k n x−ω(k n) t)
n

The above equation simply represents the sum of n waves with different
amplitudes Ψ 0 ( k n ) corresponding to different values of the wave vector k n. The

30
amplitudes Ψ 0 ( k n ) define the weights that different waves contribute to the overall
amplitude of the combined wave.
We are going to start this exercise by adding two waves. For simplicity, let us
assume that their amplitudes Ψ 0 ( k n ) are equal to 1. Furthermore, they have wave
vectors and angular frequencies that are different by infinitesimal amounts dk and
dω , respectively.
Ψ 1 ( x , t )=sin(kx−ωt )
Ψ 2 ( x , t )=sin( [ k +dk ] x−[ ω+dω ] t )
In Excel template for Lab 1 downloaded from D2L, plot both Ψ 1 ( x , t ) and Ψ 2 ( x , t )
waves:
First, you need to create a column of x values. In order to do so, enter the value
“0” into cell A14 followed by entering the formula “=$A14+0.1” into cell A15.
After pressing the Enter (in Windows) or Return (in Mac) key, the value “0.1”
should appear in cell A15. Every time you input a formula in Excel, you need to
start with the equal sign “=” followed by the mathematical expression. The dollar
sign in front of A indicates that the column will not change if you copy this formula
into a different cell. A lack of the dollar sign in front of 14 indicates that the row
will change if you copy and paste this formula into a different cell. Copy cell A15
(“Ctrl” plus “C” in Windows, “Command” plus “C” in Mac), select cells A16
through A583, and paste (“Ctrl” plus “V” in Windows, “Command” plus “V” in
Mac)) cell A15 into the selected cells. Now, if you make, let us say, cell A16 active,
you will see “=$A15+0.1” in the formula line. If you keep moving down column A,
you will notice that every time you change the row, the number in the formula
line also changes – in row 35 it is going to be “=A34+0.1”.
Enter the formula “=sin($B$7*$A14-$B$8*$B$9)” into cell B14. The formula uses
the values of k , ω , and t located in cells B7, B8, and B9, respectively. Also, note
the use of the dollar sign in front of the letters indicating columns and the
numbers indicating rows. After pressing Enter or Return key, the value of Ψ 1 ( x , t )
for x=0 and t=0 will be calculated in cell B14. Copy cell B14, select cells B15
through B583, and paste cell B14 into the selected cells. This operation will result
in populating these cells with values of Ψ 1 ( x , t ) for different values of x located in
column A and t=0 . As soon as the cells are populated, the graph of the function
Ψ 1 ( x , t ) will be plotted automatically on the top plot.
In a similar way, calculate the values of Ψ 2 ( x , t ) in column C. The second wave will
appear on the top plot.
Next, we are going to plot the sum of two waves:
Ψ ( x ,t )=Ψ 1 ( x , t ) +Ψ 2 ( x , t )
Ψ ( x ,t )=sin ( kx −ωt ) +sin( [ k+ dk ] x −[ ω+dω ] t )

31
Recall from trigonometry that

sinα +sinβ=2 sin ( α +2 β ) cos( α −β

2 )
Therefore,

Ψ ( x ,t )=2 sin ( 2 ) (
kx −ωt + [ k +dk ] x−[ ω+ dω ] t
cos
kx−ωt −[ k +dk ] x+ [ ω+dω ] t
2 )
Ψ ( x ,t )=2 sin ( ) cos( −dkx+2 dωt )
2 kx−2 ωt +dkx −dωt
2
Considering that dk ≪ 2 k and dω ≪ 2 ω, and also that cosα=cos ⁡( −α ), the above
equation may be simplified to

Ψ ( x ,t )=2 sin ( 2 kx−22 ωt ) cos ( dkx−dωt

2 )
Ψ ( x ,t )=2 sin ( kx−ωt ) cos ( x−
2 )
dk dω
t
2
Use the above formula to calculate the values of Ψ ( x ,t ) in cells E14 through E583
for different values of x (column A) and t=0 . Use the same values of k , ω , and t as
for plotting Ψ 1 ( x , t ) and Ψ 2 ( x , t ). These values and the values of dk and dω are in
cells B7 through B11.

Finally, calculate values of the function Ψ ' ( x ,t )=2 cos ( dk2 x− dω2 t ) in column F. As you
can see, the function Ψ ' ( x ,t ) is the second term in the function Ψ ( x ,t ). Populate
column G with the values of the function of Ψ ' ( x ,t ) but with the opposite sign. You
can do it by using the following syntax in cell G14 “=-$F14”, and then copying
and pasting it into the rest of the cells in column G.
This is what you should have on your screen after completing all the steps:

32
The first term of the combined wave Ψ ( x ,t ) is a wave of the same form as the
single wave Ψ 1 ( x , t ). However, the wave Ψ ( x ,t ) is further modulated by the second
term so that the oscillations of Ψ ( x ,t ) fall within an envelope of periodically
varying amplitude. This happens because two waves of slightly different
wave number k and angular frequency ω interfere constructively and
destructively to produce a succession of groups.
To illustrate the above statement, consider two waves:

33
 1.5

1.0

0.5

0.0
0 10 20 30 40 50 60
-0.5

-1.0

-1.5

They were plotted using the equations

Ψ 1 ( x , t )=sin(kx−ωt )
Ψ 2 ( x , t )=sin( [ k +dk ] x−[ ω+dω ] t )
with k =0.5 and dk =0.05 for t=0 . Since t=0 , the values of ω and dω are irrelevant for
the moment.
You can see that both waves coincide perfectly at the beginning (they are in
phase). As x increases, they shift more and more relative to each other. This
happens because their wave vectors are slightly different, which means that their
wavelengths are different. Now imagine that you add the amplitudes of these two
waves for all values of x . In the regions where the waves are in phase (red arrow),
their amplitude will double; in the regions where they are out of phase (blue
arrows), their amplitudes will cancel out; and in the regions in between the
amplitude will range from double the individual wave amplitude to zero. The
pattern will repeat itself with a certain period as you can see above in the
snapshot of the Excel sheet.
What we have constructed so far are the wavepackets at time t=0 . As the time
goes on, these wavepackets, and the individual waves which they contain, are
both moving in the direction of increasing x .
In order to see that, we are going to animate the graphs.
First, make sure that macros are enabled in the Excel file that you are using as a
template. By default, they should be as the template was created as an Excel
Macro-Enabled Workbook with the extension .xlsm.
Under the “Developer” tab, click “Visual Basic”. A Microsoft Visual Basic for
Applications window will open. Click “View”, then “Code”. Another window into
which to input the code will open.

34
Enter the following code into the window.
Sub wave()
Range("b9").Value = 0
Do While Range("b9").Value < 50
Range("b9").Value = Range("b9").Value + 0.25
Worksheets(1).Calculate
DoEvents
ActiveSheet.ChartObjects(1).Activate
ActiveSheet.ChartObjects(1).Chart.Refresh
ActiveSheet.ChartObjects(2).Activate
ActiveSheet.ChartObjects(2).Chart.Refresh
Application.ScreenUpdating = True
DoEvents
Loop
End Sub
The program tells Excel to start with a value of 0 in cell B9 and increase its value
in increments of 0.25. After each step Excel recalculates all formulas and updates
the graphs. Excel will keep on doing this as long as the entry in cell B11 is less
than 50.
Close the Visual Basic window.
On the menu bar click View → Macros. Select your macro and click run. You can
see the wave travelling in time. Once the animation is finished, reset the value in
cell B9 to 0.
Next, we are going to determine the velocity of the wave and velocity of the
group. To help you visualize the difference between the wave and the group,
follow the animation in Excel. You can track the velocity of the wave by looking at
the blue wave. To track the velocity of the group or the wave packet, you need to
follow the red wave.
Let us start with the wave equation for the sum of two waves.

Ψ ( x ,t )=2 sin ( kx−ωt ) cos ( dk2 x− dω2 t )

The first term is identical to the wave equation of the single wave Ψ 1 ( x , t ). Thus,
the velocity v wave can be evaluated by considering the first term of Ψ ( x ,t ), i.e.
sin(kx−ωt ). This is the simplest type of wave motion, a sinusoidal wave of the
frequency ν=ω /2 π and the wavelength λ=2 π /k , which is of constant unit amplitude
from −∞ to + ∞, but which is moving with uniform velocity in the direction of

35
increasing x . You can see this wave (shown in green) in the upper graph of your
Excel spreadsheet.
The zeros of the function correspond to the nodes of the wave. The sine function
is zero when its argument is equal to nπ :
k x n−ωt =nπ
n=0 , ±1 , ±2 , …
ωt nπ
x n= +
k k
The values x n represent the positions of the nodes, points on the x axis where the
function sin ( kx −ωt )=0 . In fact, these points are moving in the direction of
increasing x . Recall from classical mechanics that in order to find velocity we need
to take the time derivative of the position, i.e.

( )
dx n ωt nπ ' ω
v wave= = + =
dt k k k
Note that this is identical to the expression that we saw in lecture for the velocity
of the classical wave:
2π ω ω
v wave=λϑ = =
k 2π k
Therefore, we have not resolved the observed paradox yet.
Now, we are going to evaluate the velocity of the group (or wave packet). We will
proceed as before for a single wave but in this case we are going to use the

second term of the wave function Ψ ( x ,t ), which is cos (

dk
2
x−
dω
2 )
t . If you go back to

your Excel plots, you will see that this function, shown in red, represents the wave
packet.
Similarly to the approach we implemented above, we need to find the nodes first,
i.e. the values of x n when the amplitude is zero. The cosine function is zero when
(2 n+1)π
its argument is equal to , i.e.
2
dk dω (2 n+1)π
x n− t=
2 2 2
2 dω 2 ( 2 n+1 ) π
x n= t+
2 dk 2 dk
Then the velocity of the wave packet v wavepacket is the first derivative of the
coordinate with respect to time:

v wavepacket=
dt (dk
t+
dk )
dx n dω (2 n+1 ) π ' dω
= =
dk

36
Finally, we are in position to connect the group velocity (or the velocity of the
wave packet) v wavepacket to de Broglie’s matter wave associated with the moving
particle. From the Einstein and de Broglie relations, we have
dE
dω ℏ dE h dE
v wavepacket= = = =
dk 2 π ℏ 2π d p d p
dλ
Differentiating the kinetic energy of the particle
2
m v particle
E=
2
with respect to its velocity gives
dE
=m v particle
d v particle
At the same time, the particle momentum is given by
p=m v particle
and its differential with respect to velocity is
dp
=m
d v particle
Substitutive the above formulas into the formula for the wavepacket velocity we
obtain
d E m v particle d v particle
v wavepacket= = =v particle
dp md v particle
The velocity of the group of matter waves is just equal to the velocity of the
particle whose motion they govern, and de Broglie's postulate is internally
consistent.
To make the concept of the wave velocity and the group velocity more intuitive,
change the value of ω in cell B8 from 1 to 5, and animate the graphs (you will
need to set the value of t in B9 back to 0). Higher ω corresponds to a higher wave
velocity, which you should be able to notice as the motion of the blue wave in the
bottom graph. The group velocity does not depend on ω , and therefore the
velocity of the red wave is not affected. Next, change ω back to 1, and increase
the value of d ω from 0.2 to 0.7. Animate the graphs. Now you should see how the
velocity of the wavepacket changed – it moves faster.
It can be shown that, for an infinitely large number of waves that combine to form
one moving group, the dependence of the wave velocity v wave, and the group
velocity v wavepacket is exactly the same as for the simple case of two waves that we
have just considered.
Part 3.1. Creating a wavepacket using an infinite number of waves.

37
Up to this point, we have dealt with waves which are represented by
mathematical equations containing only real numbers. However, it is very
common to represent a plane wave with the following equation:
i(kx−ωk t )
Ψ ( x ,t )=Ψ 0 e
Why this equation is appropriate to represent waves was discussed in class
previously. Moreover, we already saw that the above equation is a valid solution
of the Schrödinger equation.
In the first part of this laboratory exercise you constructed the wave packets by
summing two waves with slightly different wave vectors k and angular frequency
ω . Two waves were not enough to localize a particle in space, i.e. create a
wavepacket shown below:
1.0
Y(x,t)

0.8

0.6

0.4

0.2

0.0
0 1 2 3 4 5 6 7 8 9 10
-0.2 x

-0.4

-0.6

-0.8

-1.0

It turns out that no matter how many waves you add, you will never be able to
create a single wavepacket. When adding a finite number of waves, there will
always be another, and then another region of space where these waves interfere
constructively.
The only way to create a single wavepacket is to add an infinite number of waves
with different values of the wave vector k . The interval for the k values can be
both finite and infinite. Below you will see two examples of this: in the former case
k values range from −1 to +1, in the latter case (Gaussian distribution) they
change from minus to plus infinity. Even when the interval for k is finite, the
number of waves is infinite because there is an infinite number of k that one can
fit between −1 and +1.
When moving from the finite to the infinite number of waves, we need to change
the sum sign in the formula that you saw in part 2 of this lab
38
Ψ ( x ,t )=∑ Ψ 0 ( k n ) sin(k n x−ω(k n) t)
n

to the integral
+∞
1
Ψ ( x ,t )=
√ −∞
2 π
∫ Φ (k )e
i (kx −ω(k)t )
dk

Note that in the above integral the integration limits are given from −∞ to + ∞. As
already mentioned, it does not have to be the case. The integration limits are
determined by the range for k .
This is how you should interpret the above integral:
1
First, we will ignore the factor , which is a constant, for now. The sine function
√2 π
that you see in the sum was replaced with the complex exponential. As discussed
above, complex exponentials are used to represent waves. Furthermore, we
already found out that the complex exponentials in the above integral indeed
represent quantum mechanical wavefunctions. Φ (k ) is a function that represents
the distribution of k vectors in the wave packet. In other words, the wave packet
consists of an infinite number of waves, each with the wave vector k whose value
is governed/given by Φ (k ), i.e. Φ (k ) shows how these k s are distributed. Φ (k ) may
be of different shapes. For instance, it may have the form of a square or the
Gaussian:

39
In the former case, the wave packet contains waves with the wave vectors in the
range between −1 and +1. All these waves contribute equally to the wave packet,
1
i.e. with equal amplitudes of ≈ 0.707.
√2
In the latter case, the wave packet contains waves with the wave vectors centred
around some value k 0. In the example above, this value is 20. The distribution of k
follows a well-known Gaussian distribution. Note that in this case not all waves
contribute equally. For instance, the wave with the wave vector k =10 enters the
integral with less weight compared to the wave with k =20.
Note that in both distributions we have an infinite number of wavevectors. It may
not be so intuitive with the square distribution because the k values are restricted
between −1 and 1. These boundaries should not mislead you. You should ask
yourself the following question: how many values of k I can select between −1 and
1. The answer is obviously infinite.
The Gaussian distribution is defined in the range from minus infinity to plus
infinity. Therefore, it contains an infinite number of wavevectors that is not limited
by any boundaries.
Before we continue with wave packets, we need to explore the Gaussian
distribution in greater detail. The simplest form of the Gaussian distribution is
given by
2
−α x
G ( x )=e
This distribution has a bell shape centred around x=0 . The width of the
distribution is determined by the parameter α .
Let us determine the area under the Gaussian bell curve, i.e. let us integrate the
function G ( x ):
40
 +∞ +∞
I = ∫ G(x)dx= ∫ e
2
−α x
dx
−∞ −∞

Instead of calculating I , we will calculate its square using two variables instead of
one:
+∞ +∞ +∞
I =∫ e dx ∫ e dy =∬ dxdy e
2 2 2 2
2 −α x −α y −α ( x + y )

−∞ −∞ −∞

This is a perfectly legitimate decision because the value of any integral should not
depend on the name of a variable that we use.
Next, we transform this integral into polar coordinates:
+∞ 2 π +∞
I =∬ e rdrdθ=2 π ∫ e
2 2
2 −α r −α r
rdr
00 0

Here we used the following well-known formulas connecting Cartesian and polar
coordinates:
x=r cos θ
y=r sin θ
Therefore, x 2+ y 2=r 2
Furthermore, dxdy representing an infinitesimal area on a plane is given by rdrdθ in
polar coordianates.
Introducing a new variable u, we obtain
2
u=r
du=2 rdr
du
dr =
2r
+∞
I =2 π ∫ e
2

0
−αu 1
2
du=π
−1 −αu + ∞ π
α
e
0 α (
= (−e −(−e ) )=
−∞ 0 π
α | )
√
+∞ +∞
π
I = ∫ G(x)dx= ∫ e
2
−α x
dx=
−∞ −∞ α
If you need to shift this distribution along the x axis, you modify it as
2
−α ( x−x 0)
G ( x )=e
In this case the distribution will maintain its shape and width but will be shifted
along the x axis by x 0. Intuitively, you should see that the integral
+∞

∫ e−α ¿ ¿ ¿
−∞

also equals to

area under it.

√ π
α
– by shifting the curve along the x axis you do not change the

41
In this lab you are going to work with the Gaussian distribution of the following
form:

( )
1 /4
2α −α (k−k 0 )2
Φ (k )= e
π
The coefficient ensures that the distribution is normalized in a way that
+∞

∫ |Φ (k )| dk =1
2

−∞
2
|Φ (k )| is a norm squared of the function Φ (k ). For real functions norm squared is
just the square of the function. The Gaussian function is real and we can verify the
normalization condition by using the value of the Gaussian integral that we just
calculated

[( ) ] √ √ √
+∞ 1/ 4 2 + ∞ +∞
2α 2 2α 2α π
∫ |Φ (k )| dk = π ∫ [ e−α (k−k ) ] ∫
2 2
−2 α (k−k ) 2
0
dk= e 0
dk = =1
−∞ −∞ π −∞ π 2α
We are going to plot the Gaussian distribution Φ (k ) on sheet 2 of the lab template
in Excel.
To do this, populate cells A12 through A412 with the values of the wave vector k .
Start with 0 and use the increment of 0.1, i.e. in cell A412 the value of k should be
40.
Calculate the values of Φ (k ) in column B using the above Gaussian formula. By
now you should be able to turn the formula into a correct Excel syntax by yourself
but in case you need to doublecheck, below is what the syntax in cell B12 should
look like:
=((2*$B$6/PI())^0.25)*EXP(-$B$6*(($A12-$B$7)^2))
Copy and paste the value in cell B12 into cells B13 through B412. The distribution
should automatically appear on the first plot. The value k 0 that you see in cell B7
is the centre of the distribution. The value α that you see in cell B6 defines the
width of the distribution. Change the values in B6 and B7 to see how they affect
the Gaussian function.
Let us go back to our initial problem. We want to integrate (sum) the infinite
number of waves with the wave vectors k distributed according to the Gaussian
distribution. If we look at the Gaussian distributions that you plotted in Excel, it
may appear that it decays to zero fast. Although it is true that only waves with the
wave vectors around the values of k corresponding to the highest amplitude (k =20
in our case) are expected to have substantial contribution to the overall value of
the integral (sum), it is important to remember that the exponential function
never decays to zero and waves with the wave vectors varying from minus to plus
infinity are present within the distribution with amplitudes that are maybe very
small but, nevertheless, different than zero. Once again, the wave amplitudes are

42
not constant (like we chose to have in part 1 of this lab for the sum of two waves
and for the square distribution shown above) but they are distributed according to
the Gaussian function, i.e. enter the integral with different weights.
We are now ready to plug the Gaussian distribution into the equation for the wave
packet:
+∞ +∞

( )
1/ 4
1 1 2α
∫ ∫
2
i (kx −ω(k)t ) −α (k−k 0) i(kx−ω(k )t )
Ψ ( x ,t )= Φ (k )e dk = e e dk
√ 2 π −∞ √2 π −∞ π
First, we are going to solve this integral for t=0 :
1/ 4 + ∞
1 2α
( ) ∫ e−α(k−k ) e ikx dk
2

Ψ ( x ,0 )= 0

√2 π π −∞

In order to do it, we will make a few transformations to arrive at the standard

Gaussian integral for which we already know the answer.
First, we are going to multiply and divide the entire integral by a constant e i k x . 0

This is indeed a constant because it does not contain the variable k . Therefore, we
can move it in and out of the integral:
+∞ +∞

( ) ( )
1/ 4 ikx 1/ 4
1 2α e α
∫e ∫ e−α(k−k ) e i(k−k ) x dk
2 2
i k0 x −α (k− k0 ) i k0 x
Ψ ( x ,0 )= e dk = e 0 0

√2 π π −∞ e
ik x
0
2π
3
−∞

Next, let us introduce a new variable k ' :

'
k =k−k 0
Then,
dk ' =dk
and Ψ ( x ,0 ) becomes
+∞ +∞

( ) ( )
1/ 4 1/ 4
α α
∫e ∫ e−α k +ik ' x dk '
2 '2
i k0 x −α k ' ik ' x i k0 x
Ψ ( x ,0 )= 3
e e dk ' = 3
e
2π −∞ 2π −∞

A common way of solving such integrals is by completing the square in the

exponent. It is a simple algebraic technique that you may be familiar with from
school when you were shown how to solve quadratic equations. We want to turn
the exponent in the above integral into one of the form:
+∞ +∞

∫ e−α k dk '=∫ e−α k + β

2
( '2
+2 k β+ β )+const
' 2
( '
) +const
dk '
−∞ −∞

Let us see what we need to do with the exponent to achieve this:

First, we factor out α
'2
−α k +i k x=−α k +k
'
( '2 ' ix
α )
Second, we multiply and divide the second term by 2

43
 (
−α k '2 +k '
ix
α ) (
=−α k '2+2 k '
ix
2α )
( )
2
ix
Third, to the expression in parenthesis we add and subtract the term
2α

( ) ( ( ) ( ))
2 2
'2 ' ix '2 ' ix ix ix
−α k +2 k =−α k +2 k + −
2α 2α 2α 2α
Regrouping and taking into account that i 2=−1, we obtain

( ( )) ( ) ( )
2 2 2 2
'2 ' ix ix ix ' ix x
−α k +2 k + +α =−α k + −
2α 2α 2α 2α 4α
Substituting the above expression into the exponent in the integral, we obtain
( )
2 2
+∞ +∞ ix x

( ) ( )
1/ 4 1/ 4 −α k ' + −
α α
∫ e−α k ∫e
'2
i k0 x +ik ' x i k0 x 2α 4α
Ψ ( x ,0 )= 3
e dk '= 3
e dk '
2π −∞ 2π −∞

( )
2 2
−x +∞ ix

( 2απ )
1/ 4 −α k ' +
Ψ ( x ,0 )= 3
e
i k0 x
e 4α
∫e 2α
dk '
−∞
+∞
The above integral is exactly the standard Gaussian integral of the kind ∫ e−a¿ ¿ ¿.
−∞

We know that it is equal to

√ π
α
Thus, the final expression for the wave packet at t=0 is

( ) [√ ](
2 2
−x −x

)
1/ 4 1 /4
α π ik x 4α 1 i k0 x 4α
Ψ ( x ,0 )= e e = 0
e e
2π3 α 2 πα
We are going to see if the wave packet (wavefunction) is properly normalized. The
wavefunction is complex. We defined how to calculate a norm squared for real
functions. Here we are going to define the norm for complex numbers and
functions.
The norm of a complex number z=a+bi is defined as |z|=√ z ¿ z , where z ¿ denotes the
complex conjugate of z :
√ z ¿ z=√(a−bi)(a+bi)=√ a2 −b2 i2= √ a2+ b2
Thus, the norm squared is given by
2
|z| =z ¿ z=a2 +b2
Geometrically, a 2+ b2 is a distance from the origin to the point on the complex
plane representing the complex number (see the schematic of a complex plane
above).
The norm squared of a complex function is defined as the product of the function
times its complex conjugate:

44
 2
|Ψ ( x , 0 )| =Ψ ¿ ( x , 0 ) Ψ ( x , 0 )
2
1 /4 −x

( ) e
If Ψ ( x ,0 )= 1
2 πα
i k0 x
e 4α
, then its complex conjugate is
2
−x
Ψ ( x , 0 )=(
2 πα )
1/ 4
¿ 1 −i k 0 x 4α
e e

All we needed to do is to change the sign in front of i.

2
+∞ 1 /2 + ∞ −x 1 /2

∫ |Ψ ( x , 0 )| dx= 2 1πα
−∞
2
( ) ∫e
−∞
2α
( )
dx =
1
2 πα
√2 πα=1
As you can see, the function is complex. As we will see later, the wavefunction
must be complex. This is a requirement imposed on it by the Schrödinger
equation. However, we can still visualize the real part of the wave using the Euler
formula:
2
−x

( )
1/ 4
1
ℜ { Ψ ( x , 0 ) }= e 4α
cos(k 0 x )
2 πα
We are going to plot this function on spreadsheet 2 of the laboratory template.
Populate cells D12 through D512 with the values of x . Start from x=−5 and keep
increasing this value by 0.02 until you reach x=5 in cell D512. Calculate the
2
1 /4 −x
functions ℜ { Ψ ( x , 0 ) }, Ψ ' ( x ,0 )= 1
2 πα ( )
e 4 α , and −Ψ ' ( x ,0 ) in columns E, F, and G. The

graphs will be plotted automatically.

Important points:
1. The Gaussian distribution is a function of k . Therefore, the distribution exists in
a so-called k -space that describes the wavepacket in terms of its wavenumbers
(or momenta). The wavefunction Ψ ( x ,t ) exists in a so-called real space and
describes the position of the particle in that real space. The space is real in a
sense that the coordinate x represents a real spatial coordinate that we can use to
indicate a position of the particle.
The position space and the momentum space are intrinsically connected. You
already saw how we moved from the Gaussian distribution (momentum space)
into the position space by calculating the integral of the kind
+∞ +∞
1 1
Ψ ( x )= ∫
√2 π −∞
ikx
Φ( k )e dk = ∫
√2 πℏ −∞
ipx /ℏ
Φ(k )e dk

The inverse transform may be performed to obtain

+∞ +∞
1 1
Φ (k )=
√ −∞
2 π
∫ −ikx
Ψ ( x ) e dx=
√ 2 πℏ
∫ Ψ ( x )e
−ipx /ℏ
dx
−∞

This integration is known as Fourier transform.

45
2. Change the value of α in cell B6 from 1 to 0.5 to 0.08. This value governs the
width of the Gaussian distribution. As you observe, selecting smaller and smaller
values of α increases the spread of the Gaussian distribution Φ (k ). The broader
the Gaussian distribution is, the narrower the wavefunction Ψ ( x ) is. The narrower
the Gaussian distribution is, the broader the wavefunction is. It is instructive to
think of a broader Gaussian distribution as of a distribution representing a higher
uncertainty in momentum. Indeed, the wavenumber (momentum) may be viewed
as poorly defined for such a distribution because many waves with different
wavevectors contribute to the wavefunction. This higher uncertainty in
momentum translates into in a lower uncertainty in position – the wavefunction
becomes narrower and narrower. We can envision the extreme cases for which
either the momentum or the position is defined exactly. For the momentum
specified precisely, we will have an infinite uncertainty in position, i.e. the
wavefunction will have a sizable amplitude in the interval from minus to plus
infinity. For the position specified precisely, we will have an infinite uncertainty in
the momentum. Quantum mechanics tells us that it is impossible to specify both
the position and momentum precisely. This is the essence of Heisenberg’s
uncertainty principle that states
ℏ
∆ x ∆ px ≥
2
Bonus material (extra 2 points will be added to the lab grade if completed
correctly)
Instead of using the Gaussian distribution, use a square function to build a
wavepacket:

Φ (k )=
1
√a
a
, k ∈ k0− , k0 +
2
a
2( )
(a
2
a
¿ 0 , k ∈ −∞ , k 0 − ∪ k 0 + ,+ ∞
2 ) ( )
1. Select the value of a randomly (choosing a=2 may simplify the problem slightly)
and plot the function Φ (k ) on a separate spreadsheet in Excel.
2. In a separate Word document show that the function Φ (k ) is properly
normalized, i.e.
+∞

∫ |Φ (k )| dk =1
2

−∞

Use the Word built-in equation editor to enter formulas.

3. In the same Word document calculate the Fourier integral
a
k 0+
2
1
Ψ ( x )=
√2 π
∫ Φ (k ) e ikx dk
a
k 0−
2

46
and find the expression for the wavefunction.
Use the Word built-in equation editor to enter formulas.
The integration should be straightforward as Φ (k ) is just a constant. The resulting
wavefunction will have a complex component.
4. Plot the real part of the wavefunction ℜ { Ψ ( x ) } in Excel. Vary the width of Φ (k ) by
changing the a value. How does it affect the width of the wavefunction ℜ { Ψ ( x ) } and
why?
Part 2.3. Time evolution of the wavepacket.
What happens to the wave packet at times later than the time t=0 ?
Imagine that we start with the same Gaussian distribution of the wave vector k :
2
−α (k−k 0)
Φ (k )=e
Then the wavefunction becomes
+∞
Ψ ( x ,t )= ∫ Φ(k )e
i(kx−ω (k)t )
dk
−∞

( )
1/ 4
2α
Note that for simplicity of calculations the normalization constant and the
π
constant (1/ √ 2 π ) in front of the integral were dropped. The complexity of the
problem arises from the fact that ω is not a constant. It is a function of the wave
vector k , i.e. each wave has its own angular frequency. From the physical
perspective, this simply implies that for a free particle, the case we are
considering, the energy depends on the momentum. Recall that for a photon
E p h oton = pc
This formula comes from the famous energy-momentum relation (sometimes
called relativistic dispersion relation) introduced by Dirac:
2 2
E =( pc) + ( m0 c )
2 2

Since the mass of a photon is 0, the equation is reduced to E photon= pc .

If we consider a non-relativistic free particle, for example, a non-relativistic
electron moving in the absence of any forces, its energy is simply a kinetic energy
given by
2
p
E e=
2m
Substituting the expressions for the energy ( E=ℏω) and momentum ( p=ℏk ¿ into
the above formulas we obtain
ℏω=E photon= pc=ℏkc
ω=kc
and
47
 2 2 2
p ℏ k
ℏω=E e= =
2 m 2m
2
ℏk
ω=
2m
Next, we are going to expand ω (k ) using Taylor’s series
2
ω ( k )=ω 0 +
dω ❑
|
dk k 0
( k−k 0 ) +
1d ω ❑
2 d k k0
2
(k−k 0 )
2
|
For a photon, ω=kc , thus

|
dω ❑
=c
dk k 0
and
2
d ω ❑
d k k0
2
=0
|
dω ω
Recall from part 2 of this laboratory exercise that is a group velocity and is
dk k
a wave velocity. Our result shows that for light the wave velocity and the group
velocity are the same and are just equal to the speed of light:
ω dω
= =c
k dk
2
ℏk
For an electron ω= and the first and second derivatives with respect to k are
2m
given by
dω ❑ ℏ k 0
|
=
dk k 0 m
and
2
d ω ❑ ℏ
=
d k k0 m
2 |
Thus
ℏ k0 ℏ 2
ω ( k )=ω 0 +
m
( k−k 0 ) +
2m
(k−k 0 )

We are now ready to do the integral

+∞
Ψ ( x ,t )= ∫ Φ(k )e
i(kx−ω (k)t )
dk
−∞

Ψ ( x ,t )= ∫ e
+∞
−α(k−k 0)
2

e
[
i(kx− ω0 +
ℏk 0
m
ℏ
]
2
(k−k 0 )+ 2 m (k−k 0) t )
dk
−∞

First, we are going to make the substitution k '=k−k 0

48
 +∞ ℏk 0 ℏ 2
−i k ' t −i k' t
Ψ ( x ,t )= ∫ e
2
−α k ' ik ' x i k0 x −i ω0 t m 2m
e e e e e dk '
−∞
+∞ ℏk 0 ℏ 2
−i k ' t −i k' t
∫e
2
i k0 x −i ω0 t −α k ' ik ' x
Ψ ( x ,t )=e e e e m
e2m
dk '
−∞

Ψ ( x ,t )=e i(k 0 x−ω0 t )

+∞

∫e
[ − α +i
ℏ
2m ]
2
t k ' i x−
ℏ k0
m
e
t k' [ ] dk '
−∞

Let us compare this integral to the one that we already solved

√
2
+∞ −x
π ik x 4 α
∫e
2
i k0 x −α k ' ik ' x
Ψ ( x ,0 )=e e dk ' = e e 0

−∞ α
The integral with time dependence is identical to the one we solved with the
following substitutions:
k 0 x ↔k 0 x−ω0 t
ℏ k0
x ↔ x− t
m
ℏ
α ↔ α +i t
2m
Therefore, it is easy to write down the answer
( )
2
ℏk 0
− x− t
m

√
π i(k 0 x−ω0 t ) (
4 α +i
ℏ
2m
t )
Ψ ( x ,t )= e e
ℏ
α +i t
2m
For the sake of simplicity during visualization, we are going to introduce a new
ℏ
constant β= . Then the wavefunction becomes
2m
2

√
− ( x−2 β k 0 t )
π
Ψ ( x ,t )= ei (k x−ω t) e 0 0 4 ( α+iβt )
α+iβt
We cannot plot this wavefunction because it is not real but we can certainly find
its absolute value squared and plot that
2 2
− ( x−2 β k 0 t ) − ( x−2 β k 0 t )
2 π
|Ψ ( x ,t )| = e
i(k x−ω t ) −i (k
e 0 0 0 x−ω 0 t)
e 4 ( α +iβt )
e 4 ( α −iβt )

√(α +iβt )(α −iβt )

2
−α ( x−2 β k 0 t )
2 π
|Ψ ( x ,t )| = 2 ( α +β t )
2 2 2

e
√α +β 2 2 2
t
This function is plotted on Sheet 3 of your lab template for time t=0 .
Change the values of time in cell B11 on Sheet3 between 0 and 10 with an
increment of 1 and observe changes. Explain why the wave packet spreads out in
time.
49
Let us calculate the expectation value of the momentum p . According to the
general expectation value formula
+∞ +∞
¿ p≥ ∫ Ψ ^p Ψ dx= ∫ Ψ −iℏ
−∞
¿ ¿

−∞
( dΨ
dx
dx )
( )
+∞ i ( kx−ωt ) +∞
d(Ae )
¿ p≥ ∫ Ψ −iℏ¿
dx=∫ Ψ ¿ (−iℏ ( ik ) A ei ( kx−ωt ) ) dx
−∞ dx −∞
+∞ +∞ +∞
¿ p≥ ∫ Ψ (−iℏ ( ik ) ) Ψ dx =∫ Ψ ( ℏk ) Ψ dx=ℏk ∫ Ψ Ψ dx
¿ ¿ ¿

−∞ −∞ −∞

The integral on the right is the probability density integrated over the entire range
of the x axis. This is just the probability that the particle will be found somewhere,
which must equal one. Therefore, we obtain
¿ p≥ℏk = √2 mE
The eigenfunctions and wave functions just considered represent the idealized
situations of a particle moving, in one direction or the other, in a beam of infinite
length. Its x coordinate is completely unknown because the amplitudes of the
waves are the same in all regions of the x axis. That is, the probability densities,
for instance
¿ ¿ −i (kx−ωt) i (kx−ωt) ¿
Ψ Ψ=A e Ae =A A
are constants independent of x . Thus the particle is equally likely to be found
anywhere, and the uncertainty in its position is ∆ x=∞. The uncertainty principle
states that in these situations we may know the value of the momentum p of the
particle with complete precision, since
ℏ
∆ p∆ x ≥
2
can be satisfied for an uncertainty in its momentum of ∆ p=0, if ∆ x=∞. Perfectly
precise values of p are also indicated by the de Broglie relation, p=ℏk , because
these wave functions contain only a single value of the wave number k . Since
there is an infinite amount of time available to measure the energy of a particle
traveling through a beam of infinite length, the energy-time uncertainty principle
ℏ
∆ E ∆t≥
2
allows its energy to be known with complete precision. This agrees with the
presence of a single value of the angular frequency ω in these wave functions,
because the de Broglie-Einstein relation E=ℏω shows this means a single value of
the energy E .
A physical example approximating the idealized situation represented by these
wave functions would be a proton moving in a highly monoenergetic beam

50
emerging from a cyclotron. Such beams are used to study the scattering of
protons by targets of nuclei inserted in the beam.
It is possible to obtain a much more realistic sense of motion than is seen in
figures showing infinite sinusoidal waves moving to the left or to the right by
using a large number of wave functions of the form of Ψ ( x ,t )=e i(kx−ωt ) to generate a
group of traveling waves as you did in lab 1. We can show its motion in the
direction of increasing x , and the ever increasing width of the group. At any
instant the location of the group can be well characterized by the expectation
value ¿ x >¿ calculated from the probability density. The constant velocity of the
¿
group, d < x > dt ¿, equals the constant velocity of the free particle,
v= p/m=√ 2 mE /m=√ 2 E /m. The spreading of the group is a characteristic property of
waves that is intimately related to the uncertainty principle. Of course, the
behavior of the group wave function is easier to interpret than the behavior of a
purely sinusoidal wave function, because the corresponding probability density is
closer to the description of particle motion we are familiar with from classical
mechanics. However, the mathematics required to describe the group, and treat
its behavior analytically, is much more complicated. The reason is that a group
must necessarily involve a distribution of wave numbers k , and therefore a
distribution of energies E=ℏ2 k 2 /2 m. In order to compose even as simple a group, a
very large number of sinusoidal waves, with very small differences in wave
numbers or energies, must be summed. These mathematical complications far
outweigh any advantages involved in the ease of interpretation. Consequently,
groups are rarely used in practical quantum mechanical calculations, and most
such calculations are performed with wave functions involving a single wave
number and energy.
10. The uncertainty principle
In classical theories, probabilities and statistical analysis are simply practical
devices for treating complicated systems. In quantum physics the probabilistic
view is fundamental.
In classical mechanics, if you know the initial position and momentum of a
particle, and also the forces exerted on it, solving the equations of motion will
give you the position and momentum of the particle at any time. In other words,
the future of motion is determined.
In quantum mechanics, the position and momentum may not be determined more
accurately than allowed by the Heisenberg uncertainty principle:
∆ p x ∆ x ≥ ℏ/2
where ∆ p x and ∆ x are the uncertainties in the momentum and position along x , ℏ
is the reduced Planck’s constant.

51
It is important to understand that this principle has nothing to do with our ability
to perform measurements accurately. No improvement in instrumentation will
lead to a better determination of p x and x simultaneously. In other words, we
cannot do better than ∆ p x ∆ x ≥ ℏ/2 in principle.
Due to the fact that the product of uncertainties is involved, the more accurately
we measure, let us say, p x, the more we lose the ability to determine x . We will
encounter situations in which p x will be known exactly. For these situations we will
know nothing at all about x (i.e., if ∆ p x =0, ∆ x=∞). Therefore, the Heisenberg
uncertainty principle imposes restrictions not on how accurately x or p x can be
measured, but on the product ∆ p x ∆ x in a simultaneous measurement of both x
and p x.
11. Expectation values (Tutorial 3)
Imagine conducting a number of measurements on identical systems described
by the same wavefunction Ψ ( x ,t ), always at the same time t ; and recording the
observed values of position x where the particle is found/observed/measured. We
want to find the average value of x , which is called the expectation value written
as x or ¿ x >¿.
In statistics, the average of a continuous distribution is given by
+∞
x= ∫ x P ( x ,t ) dx
−∞

Recall Born’s postulate that connects the probability and the wavefunction as
2
P ( x , t )=Ψ ( x , t ) Ψ ( x , t )=|Ψ ( x ,t )|
¿

Therefore, we can write

+∞ +∞
x= ∫ x P ( x ,t ) dx=∫ x |Ψ ( x ,t )| dx
2

−∞ −∞

In a similar way we can evaluate the expectation value of any function of x :

+∞
x =∫ Ψ ( x ,t ) x Ψ ( x , t ) dx
2 ¿ 2

−∞
+∞
f (x)=∫ Ψ ( x , t ) f (x)Ψ ( x , t ) dx
¿

−∞

where f (x) is any function of x .

Now consider the average value of some physical property B(x , p) that, in general,
is a function of the particle’s coordinate and momentum. How do we find the
average value of B quantum mechanically? We postulate that
+∞
B= ∫ Ψ ( x ,t ) B
^ Ψ ( x ,t ) dx
¿

−∞

52
where ^B is the quantum mechanical operator for the property B. In other words,
the average of a measurable quantity, an observable, is obtained by sandwiching
the operator between the wavefunction and its complex conjugate and integrating
over the entire space.
For a stationary state
¿ ^ Ψ ( x ,t )=eiEt /ℏ ψ ¿ ^B e−iEt/ ℏ ψ =ψ ¿ B
Ψ ( x ,t) B ^ψ
+∞
B= ∫ ψ ( x ) B
^ ψ ( x ) dx
¿

−∞

Consider the special case when ψ is an eigenfunction of ^B, i.e. ^Bψ ( x )=bψ ( x )
In this case
+∞ +∞ +∞
B= ∫ ψ ( x ) B
^ ψ ( x ) dx= ∫ ψ ¿ ( x ) bψ ( x ) dx=b ∫ ψ ¿ ( x ) ψ ( x ) dx =b
¿

−∞ −∞ −∞

This result is reasonable, since b is the only possible value that we can find for B
when we make a measurement.
Next, we are going to evaluate the expectation value of momentum and energy.
The expectation value of momentum is given by
+∞
p= ∫ Ψ ( x , t ) ^p Ψ ( x , t ) dx
¿

−∞

( ) ( )
+∞ +∞
∂Ψ ( x , t ) ∂Ψ ( x , t )
p= ∫ Ψ ( x , t ) −iℏ dx=−iℏ ∫ Ψ ( x , t )
¿ ¿
dx
−∞ ∂x −∞ ∂x
Using the energy operator, we can evaluate the expectation value of the total
energy E of a particle in a state described by the wave function Ψ ( x ,t ):

( )
+∞ +∞ 2 2
−ℏ ∂
E=∫ Ψ ( x , t ) ^
H Ψ ( x , t ) dx =∫ Ψ ( x , t ) + V^ ( x ) Ψ ( x ,t ) dx
¿ ¿
2
−∞ −∞ 2 m ∂ x
In fact, the expectation value of any dynamical quantity f (x , p , t) is given by
+∞

−∞
¿
(
f (x , p , t)= ∫ Ψ ( x ,t ) f^ x ,−iℏ
∂
∂x )
,t Ψ ( x ,t ) dx

where ^f x ,−iℏ(∂
∂x ) ∂
is obtained from f (x , p , t) by replacing p with −iℏ .
∂x
Thus, in addition to the probability, the wavefunction Ψ ( x ,t ) contains information
on the expectation value of x , the momentum p, the total energy E , and, in
general, on the expectation value of any dynamical quantity f (x , p , t). In fact, the
wavefunction contains all the information that the uncertainty principle allows us
to learn about the system.
∞
d ⟨x⟩ d
=m ∫ x|Ψ | dx
2
⟨ p ⟩=m
dt dt −∞
53
 [ ]
∞ ∞ ¿
d ¿ dΨ ¿ dΨ
⟨ p ⟩=m ∫ x Ψ Ψ dx=m ∫ x Ψ +Ψ dx
−∞ dt −∞ dt dt
2
d Ψ iℏ ∂ Ψ iV
= − Ψ
dt 2 m ∂ x 2 ℏ
¿ 2 ¿
d Ψ −iℏ ∂ Ψ iV ¿
= + Ψ
dt 2 m ∂ x2 ℏ

[( ) ( )]
∞ 2 ¿ 2
⟨ p ⟩=m ∫ x −iℏ ∂ Ψ2 + iV Ψ ¿ Ψ + Ψ ¿ i ℏ ∂ Ψ2 − iV Ψ dx
−∞ 2m ∂x ℏ 2m ∂x ℏ

[ ]
∞ 2 ¿ 2
⟨ p ⟩=m ∫ x −iℏ ∂ Ψ2 Ψ + iV Ψ ¿ Ψ + i ℏ ∂ Ψ2 Ψ ¿− iV Ψ Ψ ¿ dx
−∞ 2m ∂ x ℏ 2m ∂x ℏ

[ ]
∞ 2 2 ¿
⟨ p ⟩= iℏ ∫ x ∂ Ψ2 Ψ ¿ − ∂ Ψ2 Ψ dx
2 −∞ ∂ x ∂x

( )
¿ 2 ¿ 2 ¿ ¿ 2 2 ¿
∂ ∂Ψ ¿ ∂Ψ ∂ Ψ ¿ ∂Ψ ∂Ψ ∂ Ψ ∂Ψ ∂Ψ ∂ Ψ ¿ ∂ Ψ
Ψ − Ψ = 2Ψ + − 2
Ψ− = Ψ − Ψ
∂x ∂x ∂x ∂x ∂x ∂x ∂x ∂ x ∂ x ∂ x2 ∂x
2

[ ( )]
∞ ¿
iℏ
⟨ p ⟩= ∫ x ∂ ∂ Ψ Ψ ¿− ∂∂Ψx Ψ dx
2 −∞ ∂ x ∂ x
u=x
du=dx

( )
¿
∂ ∂Ψ ¿ ∂Ψ
dv = Ψ − Ψ
∂ x ∂x ∂x
¿
∂Ψ ¿ ∂Ψ
v= Ψ − Ψ
∂x ∂x

[( )| ]
∞ ∞

( )
¿ ¿
⟨ p ⟩= iℏ x ∂ Ψ Ψ ¿ − ∂Ψ Ψ −∫
∂Ψ ¿ ∂Ψ
Ψ − Ψ dx
2 ∂x ∂x −∞ −∞ ∂x ∂x
∞

( )
¿
−i ℏ ∂Ψ ¿ ∂Ψ
⟨ p ⟩= ∫
2 −∞ ∂ x
Ψ −
∂x
Ψ dx

∞ ¿
∂Ψ
∫ ∂x
Ψ dx
−∞

u=Ψ
dΨ
du= dx
dx
¿
dΨ
dv = dx
dx
¿
v=Ψ

54
 ∞ ¿ ∞
∂Ψ ∂Ψ ¿
∫ ∂x
Ψ dx=−∫ Ψ dx
−∞ −∞ ∂ x
∞

(
⟨ p ⟩= −i ℏ ∫ ∂ Ψ Ψ ¿ + ∂Ψ Ψ ¿ dx
2 −∞ ∂ x ∂x )
∞
⟨ p ⟩=−iℏ ∫ ∂ Ψ Ψ ¿ dx
−∞ ∂x

( )
∞ ∞
d ⟨ p⟩
( )
2 ¿
∂ ∂Ψ ¿ ∂ Ψ ¿ ∂Ψ ∂Ψ
=−i ℏ ∫ Ψ dx=−iℏ ∫ Ψ + dx
dt −∞ ∂ t ∂x −∞ ∂ t ∂ x ∂ x ∂t

( ) ( )
∞
d ⟨ p⟩ 2
∂ Ψ iℏ ∂ Ψ iV
2 ¿
¿ ∂ Ψ −iℏ ∂ Ψ iV ¿
=−i ℏ ∫ − Ψ Ψ + + Ψ dx
dt −∞ ∂ x 2 m ∂ x
2
ℏ ∂x 2m ∂ x 2
ℏ

12. Particle in the infinite potential well

We are going to move to a slightly more complicated case and consider an
electron trapped inside a potential well of an infinite height. We will assume that
the potential inside the well is zero, i.e. no force is exerted on the electron inside
the well, and the potential outside the well is infinity, i.e. the electron may never
penetrate the barrier, and be found anywhere else, except inside the well.

55
Since the potential inside the well is zero, the Schrödinger equation will have the
same form as for a free particle (see the section above):
2
−ℏ2 ∂ ψ ( x )
=Eψ (x)
2 m ∂ x2
We already found the general solution for this equation:
ikx −ikx
ψ ( x )=A e + B e
As a useful exercise, we are going to verify that the general solution indeed
works. We can do this by simply substituting the solution into the original
equation:

We can always write the wavefunction as

−iE −iE
t t
Ψ ( x ,t )= A e ikx φ ( t ) + B e−ikx φ ( t )=A e ikx e ℏ
+ B e−ikx e ℏ
= A e ikx e−iωt + B e−ikx e−iωt
i(kx−ωt ) −i (kx+ωt )
Ψ ( x ,t )= A e +B e
The first term may be viewed as a wave traveling in the direction of increasing x ,
and the second term describes a wave traveling in the direction of decreasing x .
If we picture the electron classically, we can say that it is a particle bouncing back
and forth within the well. This, in turn, suggests that the waves travelling left and
right should have equal amplitudes. Recall that two waves of equal amplitudes
traveling in the opposite directions combine to form a standing wave. Thus, we
expect to have a standing wave inside the well with nodes at the boundaries.

56
Mathematically we are bound to say that A=B or − A=B , i.e. the amplitudes of the
waves are equal in magnitude (but not necessarily in sign).
For the former case:
ikx −ikx
ψ 1 ( x ) =B ( e + e
ikx −ikx
) =2 B e +e
2
We can rewrite the above equation using the Euler’s formula:

For the latter case ( B=− A )

Both ψ 1 ( x ) =α cos ⁡(kx ) and ψ 2 ( x ) =β sin ⁡(kx ) represent two linearly independent
solutions to the same time-independent Schrödinger equation, and since the
differential equation is linear in ψ (x ), their sum represents the general solution
(similarly to what we saw for the solution with complex exponentials):
ψ ( x )=α cos ( kx ) + β sin ( kx )
To reiterate, the above equation is the general solution of the time independent
Schrödinger equation for an electron trapped inside the potential well of an
infinite height. It is equivalent to the general solution written with complex
exponentials. It also looks “wavy” because it contains oscillating functions.
Next, we need to find the constants α and β . In order to do this, we need to
consider the boundary conditions.
It is given in our problem that the barriers are infinite. Therefore, the electron
cannot escape. If the electron may only be inside the well, it can never be found
outside, which means that the probability of finding it outside the well must be
zero.
Recall, that according to a quantum mechanical postulate, the probability of
finding a particle somewhere in space is given by the product of the eigenfunction
times its complex conjugate. If the probability of finding the electron outside the
well is zero, then the eigenfunction (and consequently the wavefunction) must be
zero outside the well. But if the eigenfunction is 0 outside the well, then it must be

57
0 at the boundaries as well because one of the conditions the eigenfunction must
fulfil is that it must be continuous.
Thus, the eigenfunction must satisfy the following boundary conditions:

Note that the solution is no longer a single eigenfunction. Instead, we obtained n

different solutions for the same electron. We will see later that these
eigenfunctions correspond to different states of different energy.
The coefficient β is unknown but we can find it from the normalization condition of
the wavefunction.
Since the amplitude of the wavefunction is used to calculate the probability
distribution AND we know that the electron is inside the well (it has to be
somewhere between 0 and L; we just do not know where exactly), the total
probability of finding the electron inside the well should add up to 1 (the total
probability is always 1). In mathematical form:
L

∫ β 2 sin2 ⁡( nπL x ) dx=1

0

2 1−cos ⁡(2 x)
To solve the integral we need to recall that sin ( x )= .
2

The solution is
2
β
L=1
2

β=
√ 2
L
Now we can re-write the eigenfunctions as

ψ n ( x )=
√ 2
L ( )
sin ⁡
nπ
L
x

Finally, let us recall the substitution that we made:

58
 2 2 mE
k = 2
ℏ

Remember that the potential is zero inside the box, so the only energy that the
electron has is the kinetic energy.
Key points:
1. De Broglie hypothesized that matter (electron in this case) has wave-like
properties. We have accepted De Broglie’s idea as a hypothesis first. Now, we
have solved the Schrödinger equation and have shown that the functions
obtained as the solution of the equation are indeed oscillating waves.
2. Inside the potential well, the electron may only have certain values of energy.
Not all energies are allowed. The allowable energy levels depend on the size of
the box. The smaller the box – the higher the corresponding energy level is.
Let us consider energy level 1 for two boxes. If box 1 is twice the size of box 2
then

( )
2 2
box 1 h 1
E1 =
8 m 2× L
and

()
2 2
box 2 h 1
E1 =
8m L
box 2
E1
The ratio of box 1 is 4.
E1
3. The lowest energy of the electron corresponds to energy level 1, i.e. n=1. This
energy is NOT zero and the electron may not have energy lower than that. Thus,
we can introduce the concept of zero point energy, i.e. the lowest possible energy
that a quantum mechanical system may have. The zero point energy is purely
quantum mechanical concept. It has no counterpart in classical physics.
4. In order for the trapped electron to move from energy level n to energy level
n+1, or n+2 , etc., it needs to gain/absorb energy. We know that light is a form of
energy and light can be absorbed. The trapped electron may absorb a photon only
if the energy of the photon matches the energy gap between the corresponding
levels. When absorption happens, the photon gets destroyed (a more proper term
is annihilated) and the electron transitions to a higher energy level.

59
Now let us have a closer look at eigenfunctions corresponding to different energy
levels.
First, consider a sine function
f ( x )=sin ( kx )
Its wavelength is
2π
λ=
k
You can easily verify this by plotting sine functions for different k . For example, for
k =1 , λ=2 π
Thus, the amplitude of the wavevector k is
2π
k=
λ
Comparing this amplitude to the equation for k that we obtained from considering
the second boundary condition, we can write
2 π nπ
k= =
λ L
Rearranging for the wavelength we obtain
2L
λ n=
n
The equation above allows one to calculate the eigenfunctions for an electron
trapped inside the potential well of infinite height when it resides on different
energy levels.
For example,
for n=1, λ 1=2 L ;
2L
for n=2, λ 2= =L
2
2L
for n=3, λ 3=
3
2L L
for n=4 , λ 4= =
4 2
The higher the energy level – the shorter the wavelength and the higher the
frequency, which agrees well with what we saw before – higher frequency means
higher energy.
We can plot those eigenfunctions for different values of n inside a well of a given
width.

60
Note that the waves and eigenfunctions that we are discussing are associated
with the electron in the well. The wave describes the electron residing on different
energy levels.
An electron may be promoted from a lower energy level to a higher energy level
by absorbing a photon when the energy of the photon matches the energy gap
between the two energy levels, i.e.

()
' 2

()
2 2 2
c h n h n
E p h oton =h ν =h =∆ E=En −En =
' −
λ 8m L 8m L
where n' and n are integers corresponding to different energy levels. Naturally, n'
should be larger than n (you are promoting an electron from a lower energy level
to a higher one).
13. Modeling cyanine dye using the particle-in-the-box approach;
superposition principle (Laboratory 2)
Now we are going to move away from an electron trapped in an abstract potential
well and have a look at a real life system that we can model using the particle-in-
the-box model.
Part 1

61
Consider a simple system of alternate double and single bonds in a cyanine dye
molecule:

Note that k in the chemical structure is not the wavevector k discussed above. It is
just an integer indicating the number of repeating units in the polymethine chain
between two nitrogen atoms that may be longer or shorter.
The bonds between two nitrogen atoms are shown as alternating double and
single. Each single bond is a σ bond, each doble bond consists of one σ and one π
bond. Although this representation is convenient, it is not necessarily accurate. In
reality, the π electrons involved in the formation of the π bonds are delocalized
along the length of the chain. Furthermore, the delocalization of the π electrons is
so uniform that there is almost no bond length alteration in the polymethine
chain. Although these electrons are free to move between the nitrogen atoms,
these atoms also limit the freedom of the π electrons. Of course, one could always
provide enough energy for a π electron to escape the molecule altogether, but
this would simply be the case of breaking the chemical bond and ionization, which
is of no interest to us here. To use proper quantum mechanical terminology, the π
electrons are confined in space. This kind of confinement, often called quantum
confinement, is at the origin of very interesting and peculiar optical phenomena,
which are going to be the focus of exploration in this laboratory.
If we allow ourselves to leave out of consideration three-dimensional molecular
geometry of a cyanine dye and, instead, reduce it to one dimension, we can use
the particle-in-the-box model to make certain predictions about the allowable
energy values and optical properties of the compound. For example, we can adopt
the model in which the confined π electrons may only move in one dimension,
along an imaginary axis that connects two nitrogen atoms in the molecule. This is
our box. Let us estimate its size first. We are going to assume that in cyanine
dyes under consideration all C–C bonds equal 1.44 Å , which is the average bond
length in ethane and ethene. Next, we need to consider the length of two C–N
bonds. Finally, although the nitrogen atoms do act as potential barriers to the π
electrons, they are not fully impenetrable, and the π electrons may penetrate past
these barriers. The distance of penetration depends on the end groups – stronger
electron withdrawing groups allow for farther penetration. To account for the last
two factors we are going to introduce an empirical parameter α that is simply the
length of two C–N bonds plus whatever distance the π electrons may move past
the nitrogen atoms.
You are going to determine this empirical parameter α first, using the
experimental absorption data for 3,3'-diethylthiacarbocyanine iodide.
62
Next, you are going to use this empirical parameter α to predict the maximum
absorption of four more cyanine dyes: 3,3'-diethylthiadicarbocyanine iodide, 3,3'-
diethylthiatricarbocyanine iodide, 3,3'-diethyloxacarbocyanine iodide, and 1,1'-
diethyl-3,3,3',3'-tetramethylindocarbocyanine Iodide.
The structures and absorption spectra of all five dyes were downloaded from
https://www.photochemcad.com/databases/common-compounds/cyanine-dyes
Go to tab “dye1” in the Excel lab template. 3,3'-diethylthiacarbocyanine iodide
has four C–C bonds, we assume that each bond is 1.44 Å and also at an angle of
120° . Calculate this length in cell K10.
Next, you need to determine between which levels the electronic transition
occurs. In other words, we need to determine the highest occupied molecular
orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO). The
electronic transition is going to happen between these two levels. Determine how
many π electrons 3,3'-diethylthiacarbocyanine iodide has. This is going to be the
total number of π electrons that you should distribute among the energy levels
inside the box. Remember that according to the Pauli exclusion principle you may
only place two electrons on each level. Record the quantum numbers n for the
LUMO and HOMO energy levels in cells K11 and K12.
Finally, using the experimental spectra record the wavelength at which maximum
absorption occurs and record this value in K15. Convert nm into m in cell K16.
You are now equipped to calculate the empirical parameter α :

( ) ( )
2 2 2 2
c h nLUMO h nHOMO
E p h oton =h ν =h =∆ E=En −E n = −
λ LUMO HOMO
8 m LC – C +α 8 m LC – C + α
Rearrange the above formula for α and calculate its value in cell K18.
Use the calculated empirical parameter α to predict where the other four dyes
absorb light.
Go to tab 2. Similarly to tab 1, determine the values in cells K9 through K12. For
the empirical parameter use the value determined in the previous step (tab 1).
Calculate the wavelength of maximum absorbance in cell K14. Convert m into nm
in cell K15. Using the experimental absorption spectrum, find out the wavelength
of maximum absorbance, record it in cell K16. Compare the values predicted by
the particle-in-the-box model with the experimentally measured value.
Repeat the calculations for the remaining three dyes in tabs “dye3”, “dye4”, and
“dye5”.
Checklist:
Each tab should contain: 1) two columns with the experimental values of
absorption downloaded from www.photochemcad.com; 2) graph of the absorption
spectrum with labelled axes; 3) structure of the dye; 4) tab “dye1” should have a

63
formula for the empirical parameter α programmed in cell K18; tabs “dye2”,
“dye3”, “dye4”, and “dye5” should have a formula for λ programmed in cell K14.
Part 2
We are returning to the particle trapped inside the potential well of an infinite
height.
One of the most fundamental concepts of quantum mechanics is that any
wavefunction ψ ( x ) may be expressed as a linear combination of the
wavefunctions:
∞
ψ ( x )=∑ c n ψ n ( x )
n=1

In the case of a particle trapped inside the potential well of an infinite height the
above equation becomes

√ ( )
∞ ∞
2
ψ ( x )=∑ c n ψ n ( x )= ∑ c sin nπ
L n=1 n L
x
n=1

The proper term for this is completeness, i.e. the eigenfunctions form a complete
set, or a basis, from which any other wavefunction can be built as a liner
combination.
The question is how to find the coefficients c n.
To answer this question we need to remember that the eigenfunctions are
orthogonal, i.e.
+∞

∫ ψ ¿m ( x ) ψ n ( x ) dx=0
−∞

for m ≠n .
Properly normalized eigenfunctions are called orthonormal and the above
condition may be written as
+∞

∫ ψ ¿m ( x ) ψ n ( x ) dx=δ mn
−∞

where δ mn is the Kronecker delta function that is defined as

δ mn= {1if0 if m=n

m≠n
To find the coefficients we are going to multiply the equation
∞
ψ ( x )=∑ c n ψ n ( x )
n=1

by a complex conjugate of one of the eigenfunctions, an arbitrary one:

∞
ψ ¿m ( x ) ψ ( x )=ψ ¿m ( x ) ∑ c n ψ n ( x )
n=1

Next, we are going to integrate the RHS and the LHS of the equation:
64
+∞ +∞ ∞

∫ψ ¿
m ( x ) ψ ( x ) dx=∫ ψ ¿m ( x ) ∑ c n ψ n ( x ) dx
−∞ −∞ n=1

We are allowed to move the sum sign in and out of the integral. Additionally, c n
are constants, so they can be moved around as well:
+∞ ∞ +∞ ∞

∫ψ ¿
m ( x ) ψ ( x ) dx=∑ c n ∫ ψ ( x ) ψ n ( x ) dx=∑ c n δ mn =c m
¿
m
−∞ n=1 −∞ n=1

This is a significant result! To find the m th coefficient c m all we need to do is to

multiply the wavefunction by the complex conjugate of the m th eigenfunction and
integrate.
Now, let us work through an example that uses the ideas just discussed.
Suppose that a particle is trapped inside a potential well of infinite height and its
wavefunction at time t=0 is given by
( x−0.25
0.05 )
2
−
ψ ( x )=A e
You are familiar with this function. It is the Gaussian function that you saw in Lab
1. It is shifted from the origin by 0.25 and its half width is given by 0.05. You can
think of this value as of a parameter that squeezes the bell shaped curve.
Calculate the values of this function in column B in the tab “Gauss”. Use the x
values in column A. The function will be automatically plotted on the top graph.
¿
Column C will contain values of ψ ( x ) ψ ( x ). Since ψ ( x ) is a real function, its complex
¿
conjugate is the function itself. Therefore, in order to get ψ ( x ) ψ ( x ), we simply need
to square ψ ( x ), which will happen automatically after you calculate column B. The
¿
function ψ ( x ) ψ ( x ) will be plotted automatically on the bottom graph.
Next, we want to find the value of the integral
+∞

∫ ψ ¿ ( x ) ψ ( x ) dx
−∞

i.e. the area under the curve on the bottom graph.

We can implement the trapezoid rule to evaluate this integral in Excel. The area
under any curve can be broken into small trapezoids with one side being
∆ x n=x n −x n−1 .

65
The area of each trapezoid is given by
f ( xn)+ f ¿ ¿
Use the above formula to calculate the areas of each trapezoid in column D. The
total area, i.e. the sum of all trapezoids, is approximately equal to the area under
¿
the function ψ ( x ) ψ ( x ), or the integral
+∞

∫ ψ ¿ ( x ) ψ ( x ) dx
−∞

The total area (the sum of the areas of all trapezoids (sum of values D8 through
D107)) will automatically appear in cell E8.
Recall that this integral is the total probability of finding the particle somewhere in
space. We know that the total probability should be equal to 1 for properly
normalized wavefunction.
As you can see, it is not equal to 1 yet as we expect for a normalized
wavefunction. The normalization is achieved by a properly chosen value of the
pre-exponential factor A (cell A2). Change that value to achieve normalization – it
should be slightly less than 4.
We have the wavefunction ψ ( x ); we also know what the eigenfunctions are

ψ m ( x )=
√ 2
L ( )
sin ⁡
mπ
L
x

In the example that we are considering the size of the well is 1. Therefore,
ψ 1 ( x ) =√ 2 sin ⁡( πx )
ψ 2 ( x ) =√ 2 sin ⁡( 2 πx )
ψ 3 ( x ) =√ 2 sin ⁡( 3 πx )
etc.
The first 30 eigenfunctions are calculated in tab “eigen”. The first four are plotted
in the same tab.
In the tab “coeff” we use the formula

66
+∞

∫ ψ ¿m ( x ) ψ ( x ) dx=c m
−∞

to calculate the first 30 coefficients.

¿
To calculate c 1, we need to multiply ψ m ( x ) and ψ ( x ), and integrate.
¿
ψ m ( x ) is real. Therefore, ψ m ( x )=ψ m ( x ) . For our case, the limits of integration is
between 0 and 1 because the particle may only be found inside the potential well.
In column B we multiply ψ 1 ( x ) (its values are in column B in tab “eigen”) by the
wavefunction ψ ( x ) (its values are in column B in tab “Gauss”).
Next, we need to calculate the integral
1

∫ ψ 1 ( x ) ψ ( x ) dx
0

which is equivalent to finding the area under the function ψ 1 ( x ) ψ ( x ) . We utilize the
same approach as before, namely, we use the trapezoid rule. The areas of
trapezoids are calculated in column C, and their sum is calculated in cell C4. This
sum represents the area under the curve ψ 1 ( x ) ψ ( x ) , and consequently is equal to
1

∫ ψ 1 ( x ) ψ ( x ) dx =c 1
0

By analogy, the first 30 coefficients are calculated. They are given in cells C4, E4,
G4, I4, K4, etc.
It is interesting to follow the changes in the coefficients as m increases. This
dependence is plotted in the tab “coeff cont”. As you can see, the coefficients
oscillate but they also become smaller and smaller as m increases. This suggests
that although the sum in the superposition equation
∞
ψ ( x )=∑ c n ψ n ( x )
n=1

is infinite, we may be able to get a desirable outcome with just a few terms.
The first 30 functions c 1 ψ 1 ( x ), c 2 ψ 2 ( x ), c 3 ψ 3 ( x ), etc. are calculated in the tab “cnpsin”.
The sum of the first ten
10

∑ cn ψn ( x )
n =1

is given in column AG.

The sum of the first twenty
20

∑ cn ψn ( x )
n =1

is given in column AH.

The sum of the first thirty
67
 30

∑ cn ψn ( x )
n =1

is given in column AI.

The functions given in columns AG, AH, and AI are plotted next to these columns.
As you can see, summing the first ten eigenfunctions multiplied by the
corresponding coefficients does not quite recover the original wavefunction ψ ( x ).
Summing the first twenty improves the results. Summing the first thirty recovers
the original wavefunction ψ ( x ) with an excellent accuracy.
In Tab “orthonorm” show graphically that the functions
ψ 1 ( x ) =√ 2 sin ⁡( πx )
ψ 2 ( x ) =√ 2 sin ⁡( 2 πx )
are orthonormal in a box of a unit size.
Start by plotting these two functions and ask yourself what the conditions for
normalization and orthogonality are. Right it down and proceed with plotting
these conditions.
14. Quantum mechanical harmonic oscillator
Quantum mechanical harmonic oscillator is the first example of a more realistic
potential which is a continuous function of position. It turns out that there are only
a limited number of such potentials for which it is possible to obtain solutions to
the Schrödinger equation by analytical techniques. Coulomb potential is another
example.
−1
V Coulomb ( r ) ∝r
2
V h armonic oscillator ( r ) ∝ r
At a position of stable equilibrium, the potential function V (x ) must have a
minimum. Since any realistic potential function is continuous, the function in the
region near its minimum can almost always be approximated by a parabola
2
kx
V ( x )=
2
where k is a constant.

In such a potential a particle is moving under the influence of a linear restoring

force (Hooke’s law):

68
 −dV ( x )
F ( x )= =−kx
dt
According to Newton’s second law
2
d x
−kx =F ( x )=m a x =m 2
dt
2
d x −k
2
= x
dt m
This is a second order differential equation with constant coefficients. We
encountered the equation of this form when dealing with the Schrödinger
equation for a free particle. You can go back in your notes and see how the
equation is solved. Alternatively, we can simply guess the solution. The equation
tells us that the second derivative of the function is equal to a constant times the
function itself. Both the sine and cosine functions fit as a solution. We can write
the general solution as a sum of individual solutions, similar to what we did
before:

x (t )= A sin ⁡ (√ mk t )+ B cos ⁡( √ mk t )=A sin ⁡( ωt ) + B cos ⁡( ωt )

You can verify that this is indeed the right solution by plugging it into the equation
itself and differentiating twice.

We see that the oscillator oscillates with the angular frequency ω=

√ k
m
.

The energy of the moving oscillator constantly changing between the kinetic
energy and potential energy. However, the maximum energy can be found at the
maximum displacement, where the kinetic energy is zero and the total energy is
equal to the potential energy
2
k x max
Etotal =
2
Quantum mechanics predicts that the total energy E can assume only a discrete
set of values because the particle is bound by the potential to a region of
finite extent. What are the allowed energy values predicted by quantum
mechanics for this potential? To find out, the time-independent Schrödinger
equation for the simple harmonic oscillator potential must be solved.
The first step is to write down the Hamiltonian, which is not very straightforward
considering we need to decide on the form of the potential. We can take the
classical expression for the total energy of the harmonic oscillator which is
2 2 2 2
p kx p mω 2
H= + = + x
2m 2 2m 2

69
and simply declare that ^p and ^x will be operators that abide by [ ^x , ^p ]=iℏ and the
Hamiltonian is given by

^ ^p2 m ω2 2
H= + ^x
2m 2
Then the time-independent Schrödinger equation for this Hamiltonian becomes
2 2 2
−ℏ d ψ mω ^ 2
+ x ψ=Eψ
2 m d x2 2
Operating with the position operator ^x 2 on the function is equivalent to operating
on the function with the position operator ^x twice, i.e. this operation is simply
multiplication of the function times x 2:
2 2 2
−ℏ d ψ mω 2
+ x ψ=Eψ
2 m d x2 2

2
d H ( u) dH ( u )
2
−2 u + ( ℇ −1 ) H ( u ) =0
du du
Solving this differential equation allows one to determine the function H (u).
Let us recap. We started with the time-independent Schrödinger equation
2
d ψ ( 2
2
= u −ℇ ) ψ
du
This equation cannot be solved directly. However, by writing the solutions as the
2
product of the function A e−u /2 (which as we established works as the solution for
|u|→ ∞ ) times some function H (u), we transform the problem to solving
2
d H ( u) dH ( u )
2
−2 u + ( ℇ −1 ) H ( u ) =0
du du
The equation of this kind is called the Hermite differential equation.
15. Solving the Hermite differential equation (Tutorial 4)
Solve the Hermite differential equation
2
d H ( u) dH ( u )
2
−2 u + ( ℇ −1 ) H ( u ) =0
du du

70
The solutions for the differential equation in terms of the dimensionless parameter
u are given by
2

ψ n ( u )=H n ( u ) A e−u / 2
where H ( u ) is the polynomial of the kind
∞
H ( u )=∑ an un
n=0

The polynomial coefficients are found using the recursion relationship

an +2 ( 2 n+1−ℇ )
=
an ( n+1 )( n+2 )
The relation allows us to calculate, successively, the coefficients a 2, a 4, a 6, ... in
terms of a 0, and the coefficients a 3, a 5, a 7, ... in terms of a 1. The coefficients a 0 and
a 1 are not specified by the recursion relation, but this is expected since the
differential equation for H (u) contains a second derivative and, therefore, its
general solution should contain two arbitrary constants.

16. Modeling quantum harmonic oscillator (Laboratory 3)

For any diatomic molecule, the movement of the two atoms can be divided into
three distinct types of motion: translational, vibrational, and rotational. The figure
below depicts vibrational (a) and rotational (b) motions, which are internal
movements that leave the center of mass of the molecules unchanged.

In quantum mechanics, they are modeled by the harmonic oscillator and the rigid
rotor models, respectively. The quantum harmonic oscillator is the focus of this
71
laboratory. It is the first example of a potential which is a continuous function of
position. As you will see, there is a limited number of such potentials for which the
Schrödinger equation may be solved by analytical techniques. Another example is
Coulomb potential.
−1
V Coulomb ( r ) ∝r
2
V h armonic oscillator ( r ) ∝ r
At a position of stable equilibrium, the potential function V (x ) must have a
minimum. Since any realistic potential is continuous, almost always the function
in the region near its minimum can be approximated by a parabola
2
kx
V ( x )=
2
where k is a constant.
In such a potential a particle is moving under the influence of a restoring force
(Hooke’s law):
−dV ( x )
F ( x )= =−kx
dx
According to Newton’s second law
2
d x
−kx =F ( x )=m a x =m 2
dt
2
d x −k
2
= x
dt m
As discussed in lecture the general solution is given by

x (t )= A sin ⁡ (√ mk t )+ B cos ⁡( √ mk t )
Thus, the oscillator oscillates with the angular frequency ω=

between −X and + X .
√ k
m
in some region

In this laboratory, you will use the HCl molecule as a model to get different
eigenfunctions that are the solutions of the Schrödinger equation.
Since we model the HCl molecule, we will have to use a so-called reduced mass of
the molecule defined as
m1 m2
μ=
m1 +m2
where m1 and m2 represent masses of each atom.
The reduced mass of the HCl molecule is given in cell B2 in tab “eigenfunctions”.
Rearrange the formula for ω to calculate the spring constant k .

72
Copy and paste this formula into cell A4 of the Excel file (note that all formulas
must be typed using the Word Equation editor). Using the formula, calculate
the value of k in cell D2.
Type up the formula for the total energy of the system as the sum of the kinetic
and potential energy.
Copy and paste this formula into cell C4 of the Excel file.
Note how the relative contribution is constantly changing for a moving oscillator:
the maximum of the potential energy is achieved at the maximum displacement
−X or + X , where the kinetic energy is zero and the total energy is equal to the
potential energy.
Write down the formula for the total energy for the maximum displacement X .
Copy and paste this formula into cell E4.
When the oscillator is in the lowest point, the potential energy is zero and the
kinetic is at its maximum
2
pmax
Etotal =
2μ
However, the classical oscillator can never be found beyond the turning points −X
and + X and the total energy of the system will depend only on how far the
oscillator has been displaced.
Importantly, since the classical oscillator slows down and reverses its direction of
motion at the turning points, it will spend more time near these points compared
to any other points on its trajectory.
Quantum mechanics predicts that the total energy E can assume only a discrete
set of values because the particle is bound by the potential to a region of finite
length. To find out the allowed energy values, we need to solve the time-
independent Schrödinger equation. The equation was solved in lecture. As a
recap, the classical expression for the potential was used in the equation:
2 2
−ℏ d ψ k 2
+ x ψ =Eψ
2 μ d x2 2
Next, a dimensionless variable u was introduced as
x=αu
2 ℏ
where α = .
μω
The solutions for the differential equation in terms of the dimensionless parameter
u are given by
2

ψ n ( u )=H n ( u ) A e−u / 2
where H ( u ) is the polynomial of the kind

73
 ∞
H ( u )=∑ an un
n=0

The polynomial coefficients are found using the recursion relationship

an +2 ( 2 n+1−ℇ )
=
an ( n+1 )( n+2 )
As discussed in class, in order for the series to be finite, the recursion relationship
needs to be terminated at ℇ =2 n+1.
Write down the allowed energy of the quantum harmonic oscillator obtained from
solving the time independent Schrödinger equation.
Copy and paste this equation into cell H4. Combine the equations in cells E4 and
H4 and calculate the maximum displacement X n for different n .
Copy and paste this formula into cell J7. Calculate the value X n in cells E2, F2, G2,
and H2 for different values of n.
Next, we are going to use the recursion relationship to find the coefficients a n for
different n.
The algorithm below describes the process for n=0:
1. Find the value of ℇ for n=0:
ℇ =2 n+1=2 ×0+ 1=1
2. Plug the value of ℇ into the recursion equation:
( 2 n+1−ℇ ) ( 2× 0+1−1 )
a 2= a 0= a =0
( n+1 )( n+2 ) ( 0+1 ) ( 0+2 ) 0
Thus, for n=0, we only have one coefficient a 0, which is the initial constant.
We are going to set this coefficient equal to 1.
3. Plug the coefficient into the expression for H ( u ):
∞
H 0 ( u )=∑ an un=a0 u 0=a 0=1
n=0

4. Plug H ( u ) into the expression for the eigenfunction:

2 2 2

ψ 0 ( u )=H 0 ( u ) A 0 e−u / 2=a 0 A 0 e−u /2= A 0 e−u / 2

Repeatn=0
the above algorithm for n=1 ,2 , 3.
ℇ =2× 0+1=1
( 2 n+1−ℇ ) ( 2× 0+1−1 )
a 2= a 0= a =0
( n+1 )( n+2 ) ( 0+1 ) ( 0+2 ) 0
0
H 0 ( u )=a0 u =a 0=1
n=1
ℇ =2 n+1=¿
( 2 n+1−ℇ )
a 3= a =¿
( n+1 ) ( n+2 ) 1
74
 H 1 (u )=¿

ℇ =2 n+1=¿
( 2 n+1−ℇ )
a 2= a =¿
( n+1 )( n+2 ) 0
( 2 n+1−ℇ )
a 4= a =¿
n=3 ( n+1 ) ( n+2 ) 2
H 2 ( u )=¿

ℇ =2 n+1=¿
( 2 n+1−ℇ )
a 3= a =¿
( n+1 ) ( n+2 ) 1
( 2 n+1−ℇ )
a 5= a =¿
( n+1 ) ( n+2 ) 3
H 3 ( u )=¿

The polynomials calculated in the table above are called the Hermite polynomials.
They are named after the French mathematician Charles Hermite. The general
expression for these polynomials is given by
n
n u
2
d −u 2

H n (u)=(−1) e n
e
du
where n=0 , 1, 2 , 3 , …
Using this formula, calculate H 0 (u), H 1 (u), H 2 (u), and H 3 (u).
H 0 (u)=¿
H 1 (u)=¿
H 2 (u)=¿
H 3 (u)=¿
The polynomials calculated in the table using the recursion relation and calculated
using the general formula for the Hermite polynomials are equivalent. The ones
calculated using the general formula can be obtained from the ones calculated in
the table by multiplying the latter by constants. If your math is correct, these
constants for H 0 (u) through H 3 (u) are 1, 2, −2, −12. Multiplying these polynomials
by a constant is a legitimate operation since these constants can be absorbed by
the constants A0 through A3 in the expressions of the eigenfunctions.
Re-write the expressions for the eigenfunctions using the Hermite polynomials
calculated using the general formula for H (u):
2

ψ 0 ( u )= A 0 e−u /2
75
ψ 1 (u )=¿
ψ 2 (u )=¿
ψ 3 ( u )=¿
Copy and paste these functions into cells G7, G 10, G13, G16.
Finally, we need to determine the constants A0 , A1, A2 , and A3 . This is done
through the normalization of the functions. Recall that an eigenfunction times its
complex conjugate represent probability, and the total probability must be 1. In
the integral form
+∞

∫ ψ ¿n ( x ) ψ n ( x ) dx =1
−∞

The solution of the equation is

2

ψ n ( u )=H n ( u ) A e−u / 2
or

ψn ( αx )=H ( αx ) A e
n
−x 2 /2 α 2

Then, the solution times its complex conjugate is

() () ()
x x x x x x −x / α
() () ()
2 2 2 2 2 2

ψ ¿n ψn =H n A n e− x / 2α H n A n e− x / 2α = A 2n H n Hn e
α α α α α α
and the normalization integral becomes
+∞
An ∫ H n ( αx ) H ( αx ) e
2 2
2 −x /α
n dx
−∞

Since x=αu , dx=αdu. Therefore, we can write the above integral using the variable
u instead of x :
+∞ +∞

∫ H n( u ) H n ( u) e α du= A α ∫ H n ( u ) H n ( u ) e
2 2
2 −u 2 −u
A n n du
−∞ −∞

We will use the general equation to calculate the Hermite polynomials

n
n u d 2
−u
2

H n ( u )=(−1 ) e n
e in the above integral
du

[ ] [ ]
+∞ n +∞ n
d −u −u d −u
A α ∫ H n ( u ) (−1) e e e du= An α (−1) ∫ H n ( u )
2 2 2 2
2 n u 2 n
n n n
e du
−∞ du −∞ du
Applying integration by parts n times

[ ]
+∞ n
d
∫
2
2 n n −u
An α (−1) (−1) n
H n ( u ) e du
−∞ du

76
The second (−1)n appears in the above expression because every time we do
integration by parts, we change sign. If we do it n times, we change the sign (−1)n.
Thus, the integral becomes

[ ]
+∞ n
d
A α∫
2
2 −u
n n
H n ( u ) e du
−∞ du
The next question we want to answer is what the highest order term (HOT) for
2
H n ( u ) is. Every time we differentiate the expression e−u we obtain −2 u term. If we
differentiate n times, we obtain (−2 u)n term. Thus,
2 2
n n
H nHOT (u )=(−1 ) e u (−2 u ) e−u =2n un
If we come back to the integral

[ ]
+∞ n
d
A α∫
2
2 −u
n n
H n ( u ) e du
−∞ du
we see that we need to differentiate H n ( u ) n times. You can easily verify that if we
do, we’ll only preserve the highest order term because all other terms will become
zero. For illustration, go to the expression for H 3 (u) above and differentiate it three
times. You will see that only the HOT remains non-zero.
Let us differentiate the HOT H nHOT (u )=2n un
n times:
n− ( n−1 ) −1
2 n ( n−1 )( n−2 ) … ( n− ( n−1 ) ) u
n n
=2 n!
Thus,

[ ]
+∞ n +∞
d
A α∫ H n ( u ) e du= A n α 2 n ! ∫ e du= A n α 2 n ! √ π
2 2
2 −u 2 n −u 2 n
n n
−∞ du −∞

Since the above integral must be 1,

2 1
An =
α 2 n! √ π
n

Considering
2 ℏ
α = .
mω

A2n = √mω
2 n !√π ℏ
n

( )( )
1 /2 1 /4
1 mω
An = n
2 n! πℏ
Finally, you can replace An in the eigenfunction equations with the proper
expression, and replace u with x :
ψ n ( x )=¿

77
ψ 0 ( x )=¿
ψ 1 ( x ) =¿
ψ 2 ( x ) =¿
ψ 3 ( x ) =¿
Copy and paste these equations into the Excel spreadsheet next to the
eigenfunction equations that depend on u.
Using the formulas for ψ n ( x ), calculate the values of the functions in columns B, C,
D, and E. These functions will automatically be plotted in tab “plots of
eigenfunctions”.
Go to tab “n=0”. You will see the values of the function ψ 0 ( x ) in column B.
A properly normalized eigenfunction should meet the requirement
+∞

∫ ψ ¿n ( x ) ψ n ( x ) dx =1
−∞
¿
Calculate the values ψ 0 ( x ) ψ 0 ( x ) in column C. This probability distribution function
will be automatically plotted in tab “Chart n=0” (red line). Using the trapezoid rule
calculate the integral
+∞

∫ ψ ¿0 ( x ) ψ 0 ( x ) dx
−∞
¿
To do so, calculate the values ψ 0 (x )ψ 0 ( x ) dx in column D. Each value will represent
the area of a small trapezoid with the base dx . Use dx=2.67 ×10−13 , which is the
difference between two adjacent points on the horizontal axis (column A in all
tabs). Sum these areas in cell E8. What should be the value If you did everything
correctly?
Using the formula for X n, the maximum displacements of the oscillator if it was
treated classically were calculated in cells E, F, G, H in tab “eigenfunctions”. They
are also duplicated in corresponding cells in tabs “n=0”, “n=1”, “n=2”, and
“n=3”. For n=0, the maximum displacement is ± 1.07 ×10−11 m. The values of the
2
kx
potential that we approximate by the parabola V = between −1.07 ×10−11 m and
2
−11
+1.07 ×10 m are given in column G in tab n=0, and are plotted in tab “Chart n=0”
(blue line).
¿
The function ψ 0 (x )ψ 0 ( x ) shows the probability where the oscillator may be found.
You see that quantum mechanically the allowed region extends beyond the
classical region [−X ¿ ¿ n , + X n ]¿. In cell E11 calculate the probability of locating the
oscillator inside the classical region, and in cell E13 calculate the probability of
locating the oscillator outside the classical region.
Repeat these steps for tabs “n=1”, “n=2”, and “n=3”.
78
17. Quantum harmonic oscillator exercises (Tutorial 5)
If we consider a system of two particles (a diatomic molecule, for example), the
internal motion energy of such system is the sum of the potential energy of
interaction between two particles and the kinetic energy of a hypothetical particle
whose mass is
m1 m2
μ=
m1 +m2
and whose coordinates are the coordinates of one particle relative to the other.
This quantity is called the reduced mass.
The internal motion of a diatomic molecule consists of vibration (change in the
distance between the two nuclei), and rotation (change in the spatial orientation
of the imaginary line connecting the nuclei). To a good approximation, the
vibrational and rotational motions may be treated separately. Quantum harmonic
oscillator is to model the vibrational motion and predict the vibrational energy
levels.
The relative population of two molecular energy levels in a system in thermal
equilibrium is given by the Boltzmann distribution law
N i −(E − E )/ kT
=e i j

Nj
where energy levels i and j have energies Ei and E j and are populated by N i and
N j molecules, and where k is Boltzmann’s constant and T is the absolute
temperature.

The magnitude of ν=
√
1 k
2π μ
is such that for light diatomics (for example, H 2, HCl,
CO) only the n=0 vibrational level is significantly populated at room temperature.
For heavy diatomics (for example, I2) there is significant room-temperature
population of one or more excited vibrational levels.
1. The n=0 → n=1 band of 1H35Cl in the IR spectrum occurs at 2885.98 cm−1. Calculate
N n=1
ratio.
N n=0

2. The strongest infrared band of C16O occurs at ~ν=2143 cm−1. Find the force
12

constant of 12C16O.

79
3. The force constant for the HCl molecule is 516 N m−1 and the bond length is
127.5 pm. Calculate the vibrational zero point energy of this molecule. If this
amount of energy were converted to translational energy, how fast would the
molecule be moving? Calculate the frequency of the light corresponding to the
lowest energy vibrational transition. Which region of the electromagnetic
spectrum does this frequency lie in?

4. The vibrational frequency of N 2 molecule expressed in wavenumbers is 2358 cm−1

. What is the force constant associated with the N 2 bond? How much would a
classical string with this force constant be elongated if a mass of 1.75 kg were
attached to it? Use the gravitational acceleration on Earth at sea level for
calculations.

Infra-Red (IR) Spectroscopy

Consider the following trends:

1. For a stronger bond (larger k value), the frequency ν=

1 k
2π μ √
increases.

As examples of this, in order of increasing bond strength compare:

CC bonds: C−C ¿ ), C=C (1600 cm−1) and C ≡ C(2200 cm−1),
CH bonds: C−C−H (2900 cm−1 ), C=C−H (3100 cm−1 ) and C ≡ C−H (3300 cm−1)

2. For heavier atoms attached (larger μ value), ν=

1 k
2π μ √
decreases.

As examples of this, in order of increasing reduced mass:

−1
C−H (3000 cm )
−1
C−C (1000 cm )
−1
C−Cl (800 cm )
−1
C−Br (550 cm )
−1
C−I (about 500 cm )

80
5. A coin with a mass of 6.21 g suspended on a rubber band has a vibrational
frequency of 3 s−1. Calculate the force constant of the rubber band, the zero point
energy, and the total vibrational energy if the maximum displacement is 0.45 cm ,
and the vibrational quantum number corresponding to this energy.

6. Two 2.75 g masses are attached by a spring with a force constant of k =325 N m−1.
Calculate the zero point energy of the system and compare it with the thermal
energy kT at 298 K . If the zero point energy were converted to translational
energy, what would be the speed of the masses?

18. Barrier potentials

Step potential

In the previous section we saw that a particle in the potential well of an infinite
height is a good model predicting optical properties of cyanine dyes. The step
potential is also a useful model. For instance, it can be used to model electrons at
the metal surface where the potential energy increases sharply and goes from
essentially some constant inside the metal to a higher constant outside.

81
Imagine that a particle of mass m and total energy E is moving in the region x <0
towards the point x=0 at which the potential V goes from zero to V 0.
In classical mechanics, the behaviour of the particle upon reaching x=0 depends
on its own energy and the value of the potential V 0. This is also true in quantum
mechanics but the consequences are different.
Case 1. E<V 0, i.e. the total energy of the particle is less than the height of the
potential barrier.

Let us consider the region x ≤ 0 . As we discussed previously, the first term in the
eigenfunction should be viewed as a traveling wave (and associated with it
particle) in the direction of increasing x . The second term represents the wave
and the associated particle travelling in the direction of decreasing x . Thus, we
can say that the first term represents the particle incident on the potential step
and the second term represents the particle reflected from the potential step.
Then, the probability of reflection may be expressed as
¿
B B
R= ¿ =¿
A A
where B¿ B is the probability of reflection and A¿ A is the total probability of
incidence.
The fact that this ratio is equal to 1 means that the particle with the total energy
that is less than the energy of the potential step is reflected from the potential
step with the probability 1, i.e. always reflected. This is in agreement with our
intuition and predictions of classical theory.
Using the Euler’s formula, it is easy to show that the eigenfunction may be written
as

{
k II
C cos k I x−C sin k I x x ≤ 0
ψ ( x )= kI
¿ x≥0
¿ C e−k xII

Below is the plot of this eigenfunction generated for arbitrarily selected values of
k I =k II =2 and C=0.1. Note that the grey area represents space inside the barrier.

82
 0.15 (x)
0.1

0.05

0
-10 -5 0 5
-0.05

-0.1

-0.15

We can further generate the probability density function as

¿
Ψ ( x ,t) Ψ ( x,t )

{( )
k II −iEt/ ℏ
C cos k I x−C sin k I x e x≤0
Ψ ( x ,t )= kI
¿ x ≥0
¿ (C e )e
−k x −iEt /ℏ
II

{( )
¿ k II
¿ iEt/ ℏ
¿ C cos k I x −C sin k I x e x≤0
Ψ ( x , t )= kI
¿ x ≥0
¿ (C e ) e iEt/ ℏ
¿ −k x II

Thus,

{ ( )
2
k II
C¿ C cos k I x− sin k I x x≤0
Ψ ¿ ( x , t ) Ψ ( x , t )= kI
−k II x 2
¿x ≥0
¿ C C (e )
¿

¿
The plot of the function Ψ ( x , t ) Ψ ( x , t ) for the same values of k I =k II =2 and C=0.1 is
given below

83
 0.02

0.015

0.01

0.005

0
-10 -5 0 5
-0.005

Here we observe a phenomenon that has no classical counterpart and is in sharp

contrast to classical predictions. Although the probability density does decrease
rapidly in the region x >0, there is a finite probability of finding a particle there. In
classical mechanics it would be absolutely impossible because the total energy is
less than the potential energy, which, in turn, makes the kinetic energy in that
region negative. This phenomenon is called penetration into the classically
excluded region and it is one of the most striking predictions of the quantum
theory.
A few important considerations:
1. You should not think of the excluded region as of space where the particle may
be stored. As a matter of fact, you already saw that the particle is always
reflected from the barrier with the probability of 1.
2. There is no contradiction with our observations on a macroscale. The
probability in the excluded region goes down as e−2 k x .
II

3. How can we reconcile with the fact that the total energy is less than the
potential barrier?

84
Next, we consider Case 2, in which E>V 0, i.e. the total energy that the particle has
is more than the height of the potential barrier. Classically, the particle is allowed
to move past the point x=0 . As a matter of fact, it will move past this point
without changing the direction of motion but its kinetic energy will be reduced by
exactly the value of the potential barrier.
If you consider the case E>V 0 and compare the Schrödinger equations for region I
and II , you will see that they are identical and the only difference is that the total
energy E is reduced by some constant V 0 but the solutions for both regions will
contain complex exponentials.

Next we are going to plot the eigenfunctions for the case k I =2 k II .

First, let us have a look at the eigenfunction in region I :
i kI x k I −k II −i k x ik x A −i k x
ψI = A e + Ae I
=Ae + e
I I

k I + k II 3
Using Euler’s formula
i kI x A −i k x A A
ψI = A e + e I
=A cos k I x+ A i sin k I x + cos k I x− i sin k I x
3 3 3
4A 2A
ψI = cos k I x+ i sin k I x
3 3
Next, we are going to consider region II :
i k II x 2 kI ik x 4A kI 4A kI
ψ II =C e = Ae =
II
cos x + isin x
k I +k II 3 2 3 2
The functions are complex but we can plot the real part. The graph below depicts
the function for arbitrarily selected values of A=1 and k I =2 . The grey area
represents the space inside the barrier.

85
 1.5

0.5

0
-10 -5 0 5 10
-0.5

-1

-1.5

Note that the amplitude of the eigenfunction does not change but its frequency
does. As the particle transitions into the barrier, it loses energy, which is reflected
in the frequency of the eigenfunction. Lower frequency implies lower energy.
Next we can calculate the probability

(
{( )
A −i k x −iωt
i kI x
Ae e +e I

3 x≤0
Ψ ( x ,t )=
)
k
4 A i 2 x −iωt ¿ x ≥ 0
II

¿ e e
3

(
{( )
¿
A i k x iωt
A¿ e−i k x + e I
e I

¿ 3 x≤0
Ψ ( x , t )=
¿x ≥0
)
¿ −i k x
4A
II

¿ e 2 eiωt
3

(
{ )( )
¿
A ik x A
A ¿ e−i k x + e I
A ei k x + e−i k x I I I

¿ 3 3 x ≤0
Ψ ( x , t ) Ψ ( x , t )=
( ¿ x≥0
)( )
¿ −i k x k
4A 4A i2 x
II II

¿ e 2 e
3 3
Considering

(e−i k I x 1
3
I

)( 1
3
I
1
3
1
+ e i k x ei k x + e−i k x =1+ e 2 i k x + e−2i k x +
3
1
9
I

) I I

and using Euler’s formula we write

86
 { ( 13 e ) ( )
¿ 2 i kI x 1 −2 i k x 10 ¿ 2 10
A A + e +
I
=A A cos 2 k I x +
¿ 3 9 3 9 x≤0
Ψ ( x , t ) Ψ ( x , t )= ¿
16 A A ¿x ≥0
¿
9
¿
Below is a graph of Ψ ( x , t ) Ψ ( x , t ) for A=1

1.9

1.4

0.9

0.4

-0.1
-10 -5 0 5

We can calculate the reflection coefficient in the same fashion we did for the case
E<V 0:

( ) (
k I −k II ¿ ¿ k I −k II
A A
)
( )
B B
¿ k I +k II k I +k II k −k
2

R= ¿ = = I II
A A A¿ A k I +k II
As you can see from the result, R<1 for the case E>V 0. Of course, the surprising
fact is not that it is less than 1 but that it is larger then zero. This finding implies
that even when the energy of the particle is larger than the height of the potential
barrier, there is a finite probability that the particle will be reflected and will not
be able to overcome the barrier.
Tunneling
Next, we are going to consider the case of a rectangular barrier for a particle with
E<V 0.

87
If the energy of the particle is less than the height of the step, the particle
incident from the left can appear on the right side of a barrier. This is a non-
classical phenomenon known as barrier penetration or tunneling.
The potential can be written as follows
V =V 0 for −a< x< a
V =0 for x ←a∧x> a
Our strategy is going to be similar to the one that we implemented for the barrier
potential. First, we divide the space into three regions and solve the time
independent Schrödinger equation for each. Next, we find the coefficients using
the requirements that the wavefunction and its first derivative must be single
valued and continuous.
ikx −ikx
ψI = A e + B e
' ikx −ikx
ψ I ( x )= Aik e −Bik e
κx −κx
ψ II =C e + D e
' κx −κx
ψ II ( x )=Cκ e −Dκ e
ikx
ψ III =F e
' ikx
ψ III ( x ) =Fik e
88
where

k=√
2 mE
ℏ

κ=
√ 2 m(V 0−E)
ℏ
1. Appling the boundary conditions yields four equations.
ψ I (−a )=ψ II (−a)
−ika ika −κa κa
Ae +B e =C e +D e
ψ II ( a )=ψ III ( a)
κa −κa ika
C e +D e =F e
' '
ψ I (−a )=ψ II (−a )
−ika ika −κa κa
Aik e −Bik e =Cκ e −Dκ e
' '
ψ II ( a )=ψ III ( a )
κa −κa ika
Cκ e −Dκ e =Fik e
2. First, we eliminate B from the two equations for the left boundary
ika −κa κa −ika
B e =C e +D e − A e
ik [ A e−ika−C e−κa −D eκa + A e−ika ]=Cκ e−κa −Dκ e κa
−ika −κa κa κ −κa κ κa
2Ae −C e −D e =C e −D e
ik ik

2 A e−ika=C e−κa 1+ ( ikκ )+ D e (1− ikκ ) κa

Using two equations at the right boundary

κa ika −κa
C e =F e −D e
κ [ C e κa −D e−κa ] =Fik e ika
κ [ F e ika−D e−κa− D e−κa ]=Fik e ika
ika −κa ika
κF e −2 κD e =Fik e
−κa κ ika ik ika
De = F e −F e
2κ 2κ

D e−κa =
1
2
1− (
ik
κ
F e ika )
D e κa =
1
2
1− (
ik
κ
F eika e 2 κa )
−κa ika κa
De =F e −C e
κ [ C e κa −D e−κa ] =Fik e ika
κ [ C e κa −F e ika+C e κa ] =Fik eika
89
 κa ika ika
2 κC e −κF e =Fik e
κa ik ika κ ika
C e =F e + Fe
2κ 2κ

C e−κa =
1
2( )
1+
ik
κ
F e ika e−2 κa

= (1+ )(1+ ) F e ( )( )
1 ik κ 1 ik κ
2 A e−ika ika −2κa
e + 1− 1− F e ika e2 κa
2 κ ik 2 κ ik

[ ] [ ]
ika ika
−ika Fe ik κ ik κ −2 κa F e ik κ ik κ 2 κa
2Ae = 1+ + + e + 1− − + e
2 κ ik κ ik 2 κ ik κ ik

[ ]
ika
−ika Fe −2κa ik −2 κa κ −2 κa 2 κa ik 2 κa κ 2 κa
2Ae = 2e + e + e +2 e − e − e
2 κ ik κ ik

[ ]
ika 2 2 2 2
Fe i k −2 κa iκ −2 κa i k 2 κa i κ 2 κa
2 ( e +e ) +
−ika −2 κa 2 κa
2Ae = e − e − e + e
2 kκ kκ kκ kκ

[( ]
ika 2 2
−ika Fe −2 κa 2 κa
) − ik ( e2 κa −e−2 κa ) + i κ ( e 2 κa −e−2 κa )
2Ae = 2e +e
2 kκ kκ

2Ae
−ika
=F e
ika
[ 2
( e 2 κa + e−2 κa ) i ( κ 2−k 2 ) ( e 2κa −e−2 κa )
2
−
kκ 2 ]
2Ae
−ika
=F e
ika
[ 2cosh ⁡(2 κa)−
i ( κ 2−k 2 )
kκ
sinh ⁡(2 κa) ]
Ae
−ika
=F e
ika
[ cosh ⁡(2 κa)−
i ( κ 2−k 2 )
2 kκ
sinh ⁡(2 κa) ]
The transmission coefficient is the ratio of intensity transmitted, represented by
2 2
|F| , to the intensity incident, represented by | A| . It is conventional to calculate
the reciprocal of this ratio to arrive at a compact expression. Both sides are in the
polar form of complex numbers, i.e., magnitude and phase. The last result can be
arranged as the ratio

| || i ( κ2−k 2 )
|
−ika
Ae
ika
= cosh (
⁡ 2κa)− sinh ⁡(2 κa)
Fe 2kκ

| || | | |
ika 1/ 2 ¿ 1 /2 1/ 2
Ae
−ika
Ae A e
−ika ¿
AA | A A¿| | A|
= = ¿ = =
|F F ¿| |F|
ika ika ¿ −ika 1 /2
Fe Fe F e FF

| | | | ( )
2 2 2
| | i ( κ −k ) ( κ 2−k 2 )
2 2
A 2 2
2
= cosh ⁡(2 κa)− sinh ⁡(2 κa) =cosh ( 2 κa ) + sinh ⁡(2 κa)
F 2 kκ 2kκ
In the last expression we used the fact the product of complex conjugates is sum
of the squares of the real part and the coefficient of the imaginary part.
Recalling that cosh 2 ( x ) =1+sinh2 ⁡(x )

90
 2
| A| 2
4 2 2
κ −2 κ k +k
4
2
2
=1+ sinh ( 2 κa ) + 2 2
sinh ⁡(2κa)
|F| 4k κ

( )
2
| A| 4 2 2
κ +2 κ k + k
4
2
2
=1+ 2 2
sinh ⁡(2 κa)
|F| 4k κ

( )
2
| A| 2
κ +k
2 2
2
2
=1+ sinh ⁡(2 κa)
|F| 2 kκ
We can substitute k =√ 2mE /ℏ and κ= √ 2 m(V 0−E)/ℏ into the above equation

( )
2
2m 2m
(V 0−E)+ 2 E
√ 2 m( V 0−E) a
( )
2 2
| A| ℏ 2 ℏ
=1+ sinh ⁡ 2
( ℏ )( ℏ )
|F|
2
2m 2m ℏ
4 2 E 2
(V 0−E)

2
4m 2
V0
(√ )
2
| A| ℏ
4
2 2 m(V 0−E)
=1+ sinh ⁡ 2 a

( )
|F|
2
4m
2 ℏ
4 E(V 0 −E)
ℏ4

√ 2m(V 0−E) a
( )
2 2
| A| −1 V0 2
=T =1+ sinh ⁡ 2
|F|
2
4 E(V 0−E) ℏ
To determine the transmission coefficient for a particle with E=V 0 incident from
the left on the rectangular barrier
ikx −ikx
ψI = A e + B e
' ikx −ikx
ψ I ( x )= Aik e −Bik e
ψ II =C + Dx
'
ψ II ( x )=D
ikx
ψ III =F e
' ikx
ψ III ( x ) =Fik e
where

k=
√2 mE
ℏ
−ika ika
ψ I (−a )= A e + B e =C−Da=ψ II (−a )
ika
ψ II ( a )=C + Da=F e =ψ III ( a )
' −ika ika '
ψ I (−a )= Aik e −Bik e =D=ψ II (−a )
' ika '
ψ II ( a )=D=Fik e =ψ III ( a )
Multiplying the first boundary condition by ik and solving for the term with B
ika −ika
Bik e =Cik−Daik −A ike
91
Substituting it into the third equation
−ika
2 Aik e −Cik + Daik=D
Eliminating C using the second equation
2 Aik e−ika −( F e ika−Da ) ik+ Daik =D
2 Aik e−ika =( F e ika−Da ) ik−Daik + D
−ika ika
2 Aik e =ikF e −2 Daik + D
−ika ika
2 Aik e =ikF e −D (2 aik−1)
Substituting the fourth equation into the previous one eliminates D
−ika ika ika
2 Aik e =ikF e −Fik e (2 aik−1)
2 Aik e−ika =F e ika ( ik−ik 2 aik +ik )
2 Aik e−ika =F e ika ( 2 ik+ 2a k 2 )
Aik e−ika=F e ika ( ik+a k 2 )
−ika ika
Ae =F e ( 1−ika )
−ika
Ae
ika
=1−ika
Fe
2
| A| −1 2
=T =|1−ika| =1+ k a
2 2
2
|F|

Substituting k =
√ 2 mE
ℏ
2
| A| −1 2 mE 2
2
=T =1+ 2 a
|F| ℏ
For E>V 0
2 2
| A|
|F|
2
−1
=T =1+
V0
4 E(E−V 0 )
sin
2a
ℏ (
√2 m(E−V 0) )
According to classical mechanics, a particle of total energy E in the region x ←a,
which is incident upon the barrier in the direction of increasing x , will have a
probability of one to be reflected if E<V 0, and a probability of one to be
transmitted into the region x >a if E>V 0.
As you can see from the above considerations, neither of these statements
describes accurately the quantum mechanical results.
We already discovered in the case of a step potential that even if E is larger than
V 0, the theory predicts that reflection is possible. If E is smaller than V 0 , the theory
predicts that the particle may penetrate into the classically forbidden region. Now
in the case of the barrier potential we see that there exists certain probability that
the particle will be transmitted through the barrier into the region x >a.
92
In "tunneling" through a barrier whose height exceeds its total energy, a particle
is behaving purely like a wave. But in the region beyond the barrier it can be
detected as a localized particle.
The graph of the probability density corresponding to the eigenfunction obtained
is shown below. In the region x >a the wave function is a pure traveling wave and
so the probability density is constant. In the region x ←a the wave function is
primarily a standing wave with a small traveling wave component because the
reflected traveling wave has an amplitude less than that of the incident wave.
Therefore, the probability density in this region oscillates but has minimum values
always above zero. In the region between – a and a the wavefunction is a
decreasing exponential.

2
| A|
Using the formulas for 2 we can plot the transmission coefficient as a function of
|F|
E
.
V0

93
 1
0.9
0.8
0.7
0.6
R&T

0.5
0.4
0.3
0.2
0.1
0
0 1 2 3 4 5 6 7 8 9 10
E/V0
E
For >1, the transmission coefficient T is in general somewhat less than one,
V0
owing to reflection at the discontinuities in the potential. However, it can be seen
2a
ℏ √
that T =1 whenever 2 m ( E−V 0 )=π , 2 π , 3 π ,… This is simply the condition that the

length of the barrier region, 2 a, is equal to an integral or half-integral number of

2π
de Broglie wavelengths λ II = in that region. The effect is a result of destructive
k II
interference between reflections at x=0 and x=a .
19. Scanning tunneling microscopy

Appendices
Appendix A. Differentiation and integration rules
constant rule d
const=0
dx

94
 power rule d n n−1
x =n x
dx
trigonometric rule d
sin(x )=cos (x)
dx
d
cos (x)=−sin( x )
dx
logarithmic rule d 1
ln( x)=
dx x
product rule d df (x) dg(x )
( f (x) g(x ) )= g (x)+ f (x )
dx dx dx
chain rule df df du
=
dx du dx
Integration by parts
Some integrals can be evaluated through integration by parts.
If we integrate the differentiation product rule
( uv )' =u' v +uv '
We will obtain
uv=∫ vu ' +∫ uv '
or
∫ vu '=uv−∫ uv '
Let us consider an example:
π

∫ x cos ( x ) dx
0

If we let x=v , and cos (x)=u' ; then

v =1 and u=sin ⁡(x )
'

π π
π π
∫ x cos ( x ) dx= [ x sin ( x ) ]|0 −∫ 1 sin ( x ) dx=[ π sin ( π )−0 sin ( 0 ) ]− [−cos ( x ) ]|0 =[ 0−0 ] −[ 1+ 1 ] =−2
0 0

Remember, every time you employ integration by parts, you need to be very
careful with the integration limits.
Appendix B. Complex numbers
By definition, i is the imaginary number. i=√ −1, i.e. i 2=−1. The imaginary numbers
are part of complex numbers.
Any complex number is of the form:
z=a+bi
where a and b are real numbers.
95
If you think of an a as of the real part of a complex number, and b as of the
imaginary part of a complex number, then each complex number may be
visualized on a plane using orthogonal axes along which you plot real and
imaginary parts:

The above plot shows the imaginary number z :

z=5+3 i
An important concept is of a complex conjugate. Each complex number has a
complex conjugate that is its mirror image relative to the Re axis.
To obtain the complex conjugate (usually denoted with a star: z ¿ ¿ of a complex
number, you simply need to change the sign in front of the imaginary part, i.e.:
¿
z =5−3 i
Here are a few examples:
¿
z=5+3 i z =5−3 i
¿
z=−7+ i z =−7−i
¿
z=−2 i z =2i
The complex conjugate of any real number is the number itself:
¿
z=3 z =3
You can complex conjugate complex functions as well:
−ix ¿ ix
ψ ( x )=e ψ ( x )=e
The complex conjugate of a real function is always the function itself:
¿
ψ ( x )=cos ⁡(x) ψ ( x )=cos ⁡( x )

96