Lecture Notes

Download as pdf or txt
Download as pdf or txt
You are on page 1of 99

Quantum Mechanics

of
Atoms and Molecules
Lecture Notes & Exercises
3rd edition, 2021

Dr.ir. M.A. van der Hoef


Department of Science & Technology
University of Twente.
0
Preface

Quantum mechanics of atoms and molecules is a deepening course. While most students will be acquainted
with the general rules and concepts of the periodic table and molecular bonds, this course will give an
insight into the deeper physical laws that are behind it. In the beginning of the 20th century it became
increasingly clear that physics and chemistry have the same fundamental basis, and that electrons are the
key to understanding chemical bonds. For this reasons, chemist should have some understanding of the
physics that dictates the behavior of electrons: quantum mechanics. But this is not the only reason that this
course is essential for chemists. Quantum mechanics is at the basis of much of our modern day equipment
(electronic devices, lasers) and is of pivotal importance in the area of nanotechnology. For this reason, some
basic knowledge of the concepts and rules of quantum mechanics is essential for any modern engineer. In
that sense, the course can also be viewed as broadening.

These lecture notes are a translation of the (dutch) lectures notes “Quantum Mechanica van Atomen en
Moleculen” that have been used in Chemical Engineering in the past decade, which were again based on the
older lectures notes “Chemische Fysica” by Dick Feil and Dick van den Ham, and “Fysische Eigenschappen
van Moleculen” by Wim Briels. While most of the material is the same as in the Dutch lecture notes, with
the translation to English I have also taken the opportunity to update and rewrite some of the contents,
in particular with respect to notation. For this reason I would advice students not to use the older, Dutch
edition to avoid confusion.

Quantum mechanics is by nature a rather abstract theory, and the formulation is very “mathematical”.
Since this lecture is an introduction, and to avoid getting lost in mathematical details and subtleties (which
may hinder an understanding of what it is actually about), I have tried to not make it unnecessary formal.
This means that students should at times just accept a result without derivation (for instance for “hermitian
operators”), and at times I may also be “cutting corners” in the eyes of a theoretical quantum chemist
(for instance only real (and not complex) state functions are used). For additional reading (which I would
really advice the students to do) any modern textbook on Physical Chemistry will be useful, for instance by
Thomas Engel and Philip Reid (they go somewhat deeper than these lecture notes, but still connect to it).

Enschede, January 2021

Martin van der Hoef

1
A comment on the notation

• Expectation value of the energy. The expectation value of the energy, calculated with an (approximate)
state function, is formally written as ⟨E⟩. In the MO-LCAO approach we use this quantity very often,
and it also becomes adorned with super- and sub-scripts (like ⟨E⟩+ or ⟨E⟩1 ). For this reason we leave
out the brackets: so it becomes simply E + of E1 . But this is then a different E1 that is used for
instance in chapter 2 for the particle in a box, where it is an eigenvalue of Ĥ.
• Use of ψ, ϕ, etc. In principle the symbol ψ is used for the state function of the system. When the
system is stationary, then the state function is any of the eigenfunctions ψ1 , ψ2 , ... of the Hamiltonian
Ĥ. For systems with more than one particle, these functions cannot be determined exactly, so in that
case ψ will always be an approximate state function, often described as the product of the single-particle
state functions (in that context called orbitals). For a molecule, we indicate these orbitals with the
symbol ϕ. In the MO-LCAO approach, this ϕ is made out of the atomic orbitals, which are denoted
as χ. This makes it difficult to maintain a consistent notation throughout these lecture notes, as in
chapter 3 the atomic orbitals are the eigenfunctions of Ĥ, so are denoted as ψ.
• Complex vs Real statefunctions. In principle, quantum mechanics theory is formulated in terms of
complex operators and functions. And for a profound understanding of the theory, it is unavoidable to
use complex functions. However, it is perfectly well possible to get a good grasp of the basis, and to
calculate properties like the charge on an atom, with using real functions only, which is done in these
lecture notes. I practice this means that we drop the ∗ whenever an expectation value is calculated.

2
Contents

Preface 1

1 Introduction to Quantum Mechanics; Postulates 5


1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 The photoelectric effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 The Compton effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.4 Particle diffraction & de Broglie’s hypothesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.5 The postulates of quantum mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.6 The uncertainty relation of Heisenberg . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2 Particle in a Box and Harmonic Oscillator 13


2.1 Particle on a line of length L . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Harmonic oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.3 Systems in 3-D; particle in a box . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3 Hydrogen Atom and Helium Atom; Periodic System 23


3.1 The hydrogen atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.2 Quantum mechanics for systems with two particles . . . . . . . . . . . . . . . . . . . . . . . . 27
3.3 The helium atom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.4 Higher atomic numbers; Periodic system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

4 The Molecular Hydrogen Ion 33


4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.2 Exact calculation of H+ 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.3 Approximate solution method: MO-LCAO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.4 Excited states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

5 Simple Molecules 47
5.1 Introduction: the H2 -molecule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.2 LiH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
5.3 General formulation of the MO-LCAO method . . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.4 N2 ,O2 and F2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
5.5 Water, Ammonia, Methane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.6 Ethane, Ethylene, Ethyne . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

6 Hückel Theory and Delocalized Bonds 63


6.1 1,3-Butadiene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
6.2 Benzene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
6.3 Pyridine and Pyrrole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68

3
Appendix A: Classical Mechanics 75

Appendix B: Functions and Operators 79

Appendix C: Atomic Units 83

Appendix D: Orbitals of the Hydrogen atom 85

Appendix E: Anti-symmetry (not required) 87

Appendix F: Quantitative Results for H2+ (not required) 91

Appendix G: Derivation of the Compton Equation (not required) 93

Answers to the Exercises 95

4
Chapter 1

Introduction to Quantum Mechanics;


Postulates

In the first part of this chapter we will sketch some of the historical background of the development of
quantum mechanics. In the beginning of the 20th century scientist became concerned about a number of
phenomena that could not be explained by classical theory, most notably the spectrum of the hydrogen atom
and the photoelectric effect. It turned out that ad-hoc adaptions of the classical theory did not work, and
that a fundamentally new theory had to be developed. This theory has been formulated in terms of a set of
postulates. which will be introduced in the second part of the chapter. The logic behind these postulates will
not be immediately clear, and their usefulness for explaining the behavior of for instance electrons even less.
This will slowly become more apparent in the next chapters of these lecture notes.

1.1 Introduction
In the beginning of the 20th century it became clear that an increasing number of experimental observations
could not be explained by the known laws of physics (such as Newton’s law), in particular when it came to
the behavior of electrons. As a classic example: in the framework of Newton’s law, the moon-earth system is
completely equivalent to an electron-proton system. While the nature of the force between the two objects
is different (moon-earth: gravitational force, electron-proton: electrostatic force), the functional form of
the interaction potential is the same: in both case it decays as 1/r, where r is the distance between the
objects. “The same equations have the same solutions” is a well-known quote from the famous physicist
Richard Feynman, which means that Newton’s law would give that the electron would reside in a stable orbit
around the proton, just as the moon does around the earth. However, the experimental observations are
contradicting this. It turns out that the electron energy can only have certain specific values (the energy is
quantized), which would mean that the electron can only reside in specific orbits. The solution from Newton’s
law does not give such a quantization: any energy, and hence any orbit is possible. A second problem is
that the laws of electrodynamics give that an accelerated charge (such as an electron moving in an orbit)
will emit radiation - a radio antenna is based on this principle. This would imply that the orbiting electron
would constantly loose energy, move in smaller orbits, and eventually collapse onto the proton.
A second example of the quantization of energy is the explanation of the photoelectric effect by Einstein, for
which he received the Nobel prize. This experiment also showed that light can behave as a stream of particles
(photons), which was corroborated by Compton who showed that the scattering of light with electrons could
be well described by a “classical” collision between a photon and an electron. Both experiments are explained
in more detail in the next two sections. While these experiments show that light, which is traditionally
described as a wave, can take on the characteristics of particles, other experiments (e.g. the double slit
experiment) show that electrons can take on the characteristics of waves. The general conclusion can be that
at the level of electrons, the boundary between waves and particles gets blurred, and that a new mechanics is

5
required to accommodate for this. This mechanics is know as quantum mechanics (and the “old” mechanics
is since then referred to as classical mechanics). Quantum mechanics cannot be “derived” or “proven”: it is
more that the collective new insights have lead to a set of postulates, which could explain all experimental
observations. While some may find it surprising that the laws of quantum mechanics cannot be proven, the
same is true for Newton’s celebrated 2nd law: also that law is a postulate which was found to perfectly fit
the experimental observations at the time.
In the following two sections we will first briefly discuss the aforementioned experiments, and then introduce
the postulates. How the scientists have exactly come to these postulates is not easy to explain, and may
lead to more confusion than to enlightening. It is also important to realize that this has been a process of
many years with contributions from many scientists, although the key equation (the Schrödinger equation)
has been proposed in a relatively short span from december 1925 to july 1926. In this context, it is good
to realize that it is difficult to understand quantum mechanics at all. What is hindering our understanding
is that we have no direct (visual) experience with the phenomena that happen in the world of quantum
mechanics, like we have with classical mechanics (such as colliding snooker balls, etc). So phenomena like
the two-slit experiment that would be completely logic to an inhabitant of quantum world, are baffling to us.
This has given rise to many (philosophical) discussions about the interpretation of the theory. While these
discussions can be interesting, it remains questionable if they are really helpful for understanding. Another
famous quote of Richard Feynman is: “I think I can safely say that nobody understands quantum mechanics”
(note that Feynman has received the Nobel price for his work on quantum mechanics), and maybe we should
just accept that. In these lecture notes, we just give the postulates as the absolute truth (note that the
postulates have never been contradicted by any experiment), and simply use them. The postulates itself
are very clear (albeit a bit abstract), and using the postulates is straightforward - to difficulty is only in the
mathematics that may be involved. But first we will discuss some of the key experiments that have led to
the development of quantum mechanics.

1.2 The photoelectric effect


When a beam of light is hitting a metal, its energy can be used to “break electrons free”, and the emitted
electrons can be detected by some device. The interesting observation is that when the frequency of the
light is below a certain threshold value νo , no electrons are emitted at all, independent of how strong the
intensity of the beam is. If the frequency is above νo , electrons start emitting. It turns out that the number
of electrons increases with the intensity of the beam, but is independent of the frequency, while the kinetic
energy of electrons increases with the frequency of the beam, and is independent of the intensity. From a
classical point of view, this is remarkable. Namely, the energy of the beam increases with the frequency and
with the intensity. A higher energy means that more electrons can be “freed”, and/or the electrons that
are freed obtain a higher kinetic energy. Why would increasing the intensity only affect the number, and
increasing the frequency only the kinetic energy?
A solution was proposed by Einstein. He suggested that light could be modeled by a beam of
particles (photons), each carrying an amount of energy hν, were h is Planck’s constant. An electron can use
the energy of only one photon to get freed, and the photon will cease to exist. Suppose that the potential
energy of the electron in the metal is −Φ, that is, an energy of Φ is needed to make the electron free. This
means that the threshold frequency is set by hνo = Φ. For lower frequencies, the photon’s energy is not
high enough to overcome the potential energy (remember that the electron can only use the energy of one
photon). If the frequency is above the threshold, then the excess energy is converted into the electron’s
kinetic energy. That is, the electron gets an kinetic energy

1
me v 2 = hν − Φ (1.1)
2
where me is the electron mass. So the kinetic energy of the electrons scales linearly with the frequency ν.
Since the electrons can only use the energy of one photon, the number of electrons that are emitted will
scale with the number of photons, and thus with the intensity. Of course, this explanation fits a very specific

6
problem, and cannot yet be the basis of a universal theory. However, it shows in which direction one should
think for such a theory. For now the conclusion is that light can be modeled as a beam of particles with
energy hν each.

1.3 The Compton effect


When high intensity light is shone upon free electrons, a scattering is observed: the light beam is scattered
by an angle θ, and the electrons are scattered into an opposite direction with angle φ (see fig. 1.1). Free
electrons are of course not present in nature, however the electrons in the outermost shell of a heavy atom
are so weakly bound to the atom, that they can almost be considered as free.


Figure 1.1: The Compton effect: when
ϕ θ a beam of light with frequency νo hits
an electron, it is scattered with angle θ,
and the frequency changes to ν. In the
Z
experiments it is found that the vari-
ables θ, νo = c/λo and ν = c/λ are
related according to (1.2).

hν0

It turns out that in this scattering process, also the wavelength of the light is changed, from λo (the original
wavelength) to λ, and the change is found to depend on θ as

h
λ(θ) − λo = (1 − cos θ) (1.2)
mc
in which c is the speed of light. Without further explanation we state that these experiments cannot be
explained by classical mechanics and electrodynamics. However, it can be explained by using the particle
model for light. For this, it is assumed that the photons not only have energy hν, as in the photoelectric
effect, but also a momentum hν/c, pointing into the direction of the beam. The scattering happens because
a photon with energy hνo and momentum hνo /c collides with an electron that is at rest, making the electron
shoot away. So the process is that of a single photon - electron collision. As is shown in Appendix G,
conversion of energy and momentum then leads to the Compton Equation (1.2).

The conclusion of both the Compton effect and Photoelectric effect is that a model where photons
have energy hν and momentum hν/c give a perfect agreement with the experimental findings.

1.4 Particle diffraction & de Broglie’s hypothesis


If a crystal is hit by a beam of particles such as electrons, neutrons, or helium atoms, a diffraction is observed,
similar to ordinary X-ray diffraction for crystals. This means that the diffraction pattern can be explained

7
by modeling the beam of particles as a wave, for which we can apply Huygen’s principle. Experimentally it
follows that the wavelength of the “particle waves” is equal to
h
λ= (1.3)
p
with p the momentum of the particles. So we can conclude that not only waves can represent themselves
as particles, but also particles can represent themselves as waves. In both cases the relation between the
momentum of the particle p and the wavelength λ is given by p = h/λ, and the momentum is in the direction
of the wave. We can also write this as (using h̄ = h/2π):

p⃗ = h̄⃗k (1.4)

Here ⃗k the so-called wave vector, which is a vector in the direction of the wave, and with a length 2π/λ. De
Broglie proposed that this relation holds for any particle.

We have now seen that particles can also be considered as waves. Typical for a wave is that it is not
localized in space; this means that if the particle “becomes a wave”, it can be considered as “being smeared
out” over space. However, if you do a measurement, you will find the particle in a particular location. But in
another measurement (for a new system, but prepared identically) you will find a different location. “Being
smeared out” should rather be thought off as that the position of the particle is not definite, but that there
are many possibilities; when measured, the particle will “project” itself into one of these possibilities. The
same is true for the momentum and energy of a particle. So probability (and uncertainty) are key concepts
for the new theory. With this in mind, we have a good starting point for presenting the postulates. We will
do this for stationary systems, that is, systems that do not change in time. So any dependency on time is
not considered.

1.5 The postulates of quantum mechanics


Position, velocity, momentum, and energy are dynamic variables of particles: they change under the
influence of forces acting on the particles. An important aspect of quantum mechanics - and consistent with
the wave nature of particles - is that there is a chance aspect in the value that these variables will have.
Hence, unlike Newton’s law, the laws of Quantum Mechanics do not (and cannot) predict the precise value
for the dynamics variables; instead, they predict the probability of finding a value. As any physical theory,
quantum mechanics is based on a number of postulates. On the basis of these, one can in principle predict
the (average) outcome of experiments. Up to now, no experimental observation has been in disagreement
with these predictions. In order to keep it as simple as possible, we will formulate the postulates for one
particle in one dimension. The extension to many particles in three dimensions is relatively straightforward
once you are familiar with the concepts, and will be done in the later chapters in relation to specific examples
such as the hydrogen atom.
Postulate 1:
All information about a system is provided by the state function ψ(x). The physical meaning of ψ
is that ψ ∗ (x)ψ(x)dx is equal to the probability to find a particle in the interval (x, x + dx); that is,
ψ ∗ (x)ψ(x) is the probability density. We make a few remarks to clarify this statement:
• In this, the ∗ refers to “complex conjugated”. In general the state function is a complex function
(“complex” meaning that it contains an imaginary part). However, the probability density should
be a real value, which is guaranteed by multiplying ψ with ψ ∗ . In these lecture notes, we will
mainly use real state functions (hence ψ = ψ ∗ ) in which case the probability density is just ψ 2 .
Nevertheless, the notation with ∗ is kept to make the equations more general.
• This postulate says that maximum information we have of the system is the probability to find a
particle somewhere. With this information we should predict all other properties.

8
∫b
• The probability to find a particle between x = a and x = b is given by a ψ ∗ (x)ψ(x, t)dx. Since
∫∞
the probability is 1 that particle is in the universe, it must hold that −∞ ψ ∗ (x)ψ(x)dx = 1.
Mathematically this means that the function is normalized.
• In most literature the state function is called the wave function, because it is has the characteristics
of a wave (like having maxima, minima, nodes). In these lecture notes the term state function is
used.
• Since there is no time dependence, the state ψ(x) is called a stationary state. As we will later
see, there is a special condition for a state to be stationary. The time-dependent state function is
usually referred to as Ψ(x, t).

Postulate 2(a):
Each dynamic variable (or “observable”) A is associated with a linear operator Â. An operator in
this context is defined as some mathematical manipulation (like differentiation) of a function, which
changes it to a new function. When an operator Λ̂, applied to a function f (x), changes it to g(x), we
write this as g(x) = Λ̂f (x). More information on operators can be found in appendix B. The following
operators are associated with the dynamic variables position (x), momentum (p) and total energy (E):

x → x̂ = x (1.5)
h̄ d
p → p̂ = (1.6)
i dx
2
h̄ d2
E → Ê = Ĥ = − + V (x) (1.7)
2m dx2

In this, V (x) is the potential energy of the particle. Instead of Ê, it is custom to use the notation
Ĥ. This operator is called the Hamilton operator, or Hamiltonian, and is of central importance in
quantum mechanics. Note that the expression for Ĥ(x) is consistent with the expression for the total
energy in classical mechanics, E = p2 /2m + V (x), with p replaced by p̂ and x by x̂.

Postulate 2(b):
The only possible outcome of a measurement of a dynamic variable A is one of the eigenvalues an of the
operator Â. (see appendix B for more information on eigenvalues and eigenfunctions). For instance,
the eigenvalues of the Hamiltonian are the only possible outcomes when measuring the energy.

Postulate 2(c):
In general it is not possible to predict the outcome of a single measurement, so it is not possible to
predict exactly which eigenvalue you will find. The only information that you have about the system
is ψ, and with that knowledge you can only predict the probability of finding an individual outcome,
and thus also the average outcome of many measurements. The average outcome is given by1

notation
⟨A⟩ = dx ψ ∗ (x)Â(x)ψ(x) = ⟨ψ|Â|ψ⟩ (1.8)

while the probability to find a particular outcome an is given by c2i , where the c’s are the coefficients
of the expansion of ψ(x) in the eigenfunctions of Â:

ψ(x) = cn ϕn (x)

This last part is quite abstract. In Appendix B it is explained what “expansion in eigenfunctions”
means. Note that only in the special case that ψ is equal to one of the eigenfunctions of Â, (say
ψ = ϕ3 ), then you known for certain that the outcome of a measurement on A will be a3 .
1 This is the formally correct expression. In these lectures we will only consider real state functions, so then ψ ∗ = ψ

9
Postulate 2(d):
Postulate 2(b) states that a measurement on variable A will yield one of the eigenvalues. Suppose
you have found eigenvalue ai . From that moment on, the state function of the system is equal to
ψ(x) = ϕi (x). We say that the measurement has the result that the state function is projected into
one the eigenfunctions. As a result, each next measurement on the system2 will give the eigenvalue ai .
(For a popular parable on projection and measurement, search for “ Schrödinger’s Cat”).

Postulate 3: The only possible state functions of a stationary system are the eigenfunctions ψn of
the Hamiltonian:

Ĥψn (x) = En ψn (x) (1.9)

which means that the only options are ψ(x) = ψ1 (x) or ψ(x) = ψ2 (x), etc. The eigenvalue En is
the energy of this state. The state with the lowest energy is called the ground state, the others are
excited states. Note that the well-kown 1s, 2s, 2px , 2py , 2pz functions of the hydrogen atom are in
fact eigenfunctions of the Hamiltonian of that system. The whole quantum mechanics of atoms and
molecules revolves around finding the eigenfunctions and eigenvalues of the Hamiltonian of the electron.
Equation (1.9) is known as the time-independent Schrödinger Equation; for completeness, we will also
show the full Schrödinger Equation for a time-dependent state function:

dΨ(x, t)
ĤΨ(x, t) = ih̄
dt
which could be regarded as the “Newton’s law” of quantum mechanics. To solve it, again the eigen-
functions of Ĥ are crucial. Since we will only consider stationary states, we limit ourselves to the
time-independent Schrödinger equation (1.9) to which we will refer to as “Schrödinger equation”.

What is described above is quite abstract, and for now it seems difficult to see how these rules could be at
the basis of the behavior of atoms and molecules as we known it. In order to gain some understanding of
how these postulates should be used, we will treat two very simple system in the next chapter, for which
the mathematics is quite doable. From these system, it is a small step to the hydrogen atom, which only
requires more advanced math skills. By contrast, the step from hydrogen to helium, so from a single particle
(electron) to two particles (two electrons), does not require more math, but rather an additional postulate
which might be even more abstract than those given above. Yet it is essential for the build up of the periodic
table. In the last three chapters (4,5,6) we discuss molecules, so chemical bonds. While it is well possible
to formulate the equations which exactly describe these systems, it is impossible to solve them, as is done
in chapters 2 and 3. So we have to resort to approximate methods, which can still serve to give a good,
qualitative insight into the nature of the bonds. But before doing so, we discuss a very special aspect of the
quantum theory.

1.6 The uncertainty relation of Heisenberg


Given the state function ψ, we can calculate the expectation value ⟨p⟩, that is, the average outcome of a
very large number of measurements of the momentum. But then it is also possible to calculate the standard
deviation: the average deviation of the individual measurements with the mean. We call this ∆p, and it is
defined as √
∆p = ⟨(p − ⟨p⟩)2 ⟩
∆p is a measure of the uncertainty in our prediction of the momentum. If it is large, it means that there
is a large spreading in the individual results, so there is a large uncertainty in the prediction of a single
measurement. If ∆p would be zero, this would mean that the individual results would not differ from the
2 This is assuming that ϕ is a stationary state. This would normally only be the case if it is an eigenfunction of Ĥ, see
i
postulate 3

10
mean - in other words, we could predict the momentum exactly, there is no uncertainty.
Similarly, we can calculate the standard deviation of measurements on the position:

∆x = ⟨(x − ⟨x⟩)2 ⟩
It can now be derived that

∆p∆x ≥ (1.10)
2
This means that position and moment cannot be measured with infinite accuracy in the same system (∆x
and ∆p cannot both be zero). Equation (1.10) is known as Heisenberg’s uncertainty relation, which expresses
most clearly the fundamental uncertainty of quantum mechanics. This principle and its consequences have
been the topic of much debate; for instance Einstein could not believe that it was true. The uncertainly
relation is not limited to position and momentum, there are other pairs of variables this satisfy this relation.

1.7 Exercises
General properties of functions and operators
Exercise 1.1
Given is that ϕ1 (x) are ϕ2 (x) both eigenfunctions of the operator  with eigenvalues a1 and a2 , respectively.
Are the following functions also eigenfunctions of Â, and if so, what is the eigenvalue?
a) ψ = 3ϕ1 b) ψ = ϕ1 + ϕ2

Exercise 1.2
d2
Show that f (x) = exp(−ax2 ) is an eigenfunction of the operator 1 d
x dx , and also of the operator dx2 − 4a2 x2 .

Exercise 1.3
1 d
In exercise 1.2 it was shown that f (x) = exp(−ax2 ) is an eigenfunction of the operator x dx .

a) Show that g(x) = C exp(−ax2 ), is also an eigenfunction of this operator, with the same eigenvalue. C
is an arbitrary constant.
b) Find the value for C, so that the function g(x) is normalized.
The conclusion from a) is that you when multiply an eigenfunction by an arbitrary constant, it remains
an eigenfunction with the same eigenvalue (see also exercise 1.1a). This gives the option to normalize the
eigenfunction, what you did under b).

Exercise 1.4
∂ ∂
We have a 2-D system (coordinates x and y) and the operators  = and B̂ = .
∂x ∂y
a) Show that f (x) = exp(ax) is an eigenfunction of  with eigenvalue a, and g(y) = exp(by) is an
eigenfunction of B̂ with eigenvalue b.
b) Show that the 2-D function h(x, y) = f (x)g(y) is an eigenfunction of the operator Ĉ = Â + B̂, and
give the eigenvalue.
The conclusion is that when an operator can be written as the sum of two independent operators (by
independent we mean that  only depends on x, and B̂ only on y), the eigenfunction is a product of the
eigenfunctions of the two independent operators. We will use this quite often, for example also in exercise 1.8.

11
Exercises with the eigenfunctions of the Hamiltonian
In the following exercises it is given that {ψ1 , ψ2 , ψ3 , ...} are the normalized eigenfunctions of the Hamiltonian
Ĥ, with corresponding eigenvalues E1 , E2 , E3 , ...

Exercise 1.5
a) Show that φ = √1 [ψ1 + ψ2 ] is also normalized.
2

b) Find the normalizing constant for ψ1 + ψ2 + ψ3 .


c) Show that φ is only an eigenfunction of Ĥ when E1 = E2 (we then say that ψ1 and ψ2 are degenerate
states).

Exercise 1.6
Show that when the state function ψ is equal to ψ = ψ3 that the expectation value for the energy ⟨E⟩ is
equal to E3 .

Exercise 1.7
We have prepared a system such that the state function ψ is given by ψ = 12 ψ1 + 12 ψ2 + √1 ψ3 .
2

a) Show that ψ is normalized.


b) Determine the value of ⟨E⟩
c) What is the probability that you will find the value E2 when you measure the energy of the system?

Exercise 1.8
2 2
We have given that ψn (x) is an eigenfunction of Ĥx = − 2mh̄ d
dx2 + V (x) with eigenvalue En , and ψl (y) is
2 2
an eigenfunction of Ĥy = − 2m dy2 + V (y) with eigenvalue El . Show that ψn (x)ψl (y) is an eigenfunction of
h̄ d

Ĥx + Ĥy , with eigenvalue En + El .

12
Chapter 2

Particle in a Box and Harmonic


Oscillator

In this chapter, we will apply the postulates to a number of simple cases. In this, we come to insights that
are not just limited to these cases, such as the relation between the energy and the number nodes in the state
function. Another important new concept will be the zero-point energy. We will start with two simple systems
in 1-D: a particle that is allowed to move freely on a line of finite length L, and a particle that is bounded
to the origin by a spring (a so-called harmonic oscillator). This is followed by a general description of 3-D
systems, which is applied to the specific case of a particle that can freely move in a box of size L × L × L.
At first sight the “particle-in-box” system seems still quite abstract and not yet linked to reality; however, it
turns out that you can actually derive the ideal gas law (P V = nRT ) from it, showing that the postulates of
quantum mechanics are at the basis of one of the most widely-used empirical laws in chemistry. We will not
show the derivation in this course. However, we will use the system of a particle in a box as the stepping
stone for the description of the hydrogen atom in the next chapter.

2.1 Particle on a line of length L


We consider a system in which the particle is free to move in the x direction, yet it should remain within
the boundaries x = 0 and x = L. To ensure that the particle cannot cross these boundaries, we define a
potential energy for the particle that is infinite for x ≤ 0 and x ≥ L. Since the particle is free to move in
between these boundaries, the potential energy is zero there. So

V (x) = ∞ for x ≤ 0 , x ≥ L , V (x) = 0 for 0 < x < L .

So for 0 ≤ x ≤ L, the Hamiltonian equals


h̄2 d2
Ĥ = −
2m dx2
and the Schrödinger equation becomes:
h̄2 d2 ψ
− = Eψ (2.1)
2m dx2
We write this equation as
d2 ψ 2mE
= −k 2 ψ with k2 = (2.2)
dx2 h̄2
So we need to find a function which, when differentiated twice, becomes −k 2 times this function. It can be
readily verified that both sin(kx) and cos(kx) fulfill this condition, so the most general solution is

ψ(x) = C sin(kx) + D cos(kx) (2.3)

13
where C and D can take any value. Yet, there is a second condition on ψ: not only should it satisfy the
Schrödinger equation, it should also be continuous for any x. Since the particle cannot be present outside
the domain, the function has to go to zero when x approaches 0, and when x approaches L. In other words,
we have the following boundary conditions for ψ:
ψ(0) = 0 and ψ(L) = 0
Since cos(0) = 1 the first boundary condition dictates that D = 0, so the wave equation becomes ψ(x) =
C sin(kx). The second boundary condition gives that C sin(kx) = 0. This is always true for C = 0, but
then we are left with a function ψ(x) = 0 for any x. The second boundary condition can also be fulfilled for
C ̸= 0, namely when
sin(kL) = 0 → kL = nπ with n = 1, 2, 3...
So C sin(kx) can fulfill the second boundary condition when k = nπ/L, which gives that the only solutions
that are possible are ( nπx )
ψ(x) = C sin
L
Since k was linked to E via k 2 = 2mE
h̄2
, the corresponding eigenvalue is

E = h̄2 k 2 /(2m) = h̄2 n2 π 2 /(2mL2 )


Now remember that n can only take the values 1, 2, 3, .... This means that we have naturally found the set
of solutions to the Schrödinger equation Ĥψ = Eψ, and we label each solution with n:
( nπ ) ( )
h̄2 n2 π 2 h2 n2
Eigenfunction: ψn (x) = Cn sin x Eigenvalue: En = = (2.4)
L 2mL2 8mL2

where each function satisfies Ĥψn = En ψn . The number n is called the quantum number. The energy can
h̄2 π 2 4h̄2 π 2 9h̄2 π 2
thus only take the values 2mL 2 , 2mL2 , 2mL2 , ....: the energy is quantized. The constant Cn we can finally

obtain from the condition that ψn should be a normalized function:


∫ √
∗ 2
dx ψ (x)ψ(x) = 1 → Cn = (see exercises)
L
Two remarks to the solutions (2.4):
(i) Another option for a solution we could consider is ekx , however differentiating this twice yields k 2 ekx ,
so no minus sign (check yourself that using e−kx will also not yield the minus sign). However, the
functions eikx and e−ikx (with i the imaginary number) would work, and this is also found in many
textbooks. Yet, these solutions are exactly the same as we derived above. The most general solution
in the case of exponentials would be ψ = A eikx + B e−ikx . Since e±ikx = cos(kx) ± i sin(kx), we can
write this solution as ψ = (A+B) cos(kx) + (A−B)i sin(kx). If we now define the constants D = A+B
en C = (A−B)i, we have arrived at the same general solution as in (2.3).
(ii) The integers n = −1, −2, −3.. will also satisfy the boundary conditions. However, since it holds that
sin(a) = − sin(a), the solutions for n = 1 en n = −1 are identical, except for a minus sign in front
of the function. All values that are calculated using ψ (e.g. expectation values) involve using ψ 2 , so
functions that only differ a minus sign give identical results, and are not counted separately.
A number of the solutions (2.4) are sketched in 2.1. Clearly, the solutions ψn correspond with waves (hence
often the term “wave function” is used for ψ) with specific values for the wavelength λn , such that the wave
“fits” in between 0 and L:
ψ1 → λ1 = 2L
ψ2 → λ2 = L
2L
ψn → λn =
n

14
n=1

n=2

n=3

Figure 2.1: Solutions of Ĥψn = En ψn for a particle on a line of length L.

We see that the number of nodes (points where the wave function is zero) increases with n. The energy also
increases with n, and hence also with the number of nodes. Later we will see that this relation generally
holds. The lowest possible value for the energy is

h̄2 π 2
E1 =
2mL2
When the system is in the state corresponding to this energy, it is said to be in the ground state. When
the system is put in thermal contact with its surrounding, its energy may increases or decrease, yet its value
will always have to be one of the set {E1 , E2 , E3 , ....}. So when energy is added (system is heated), the
system will move “up” the energy ladder En ; when energy is removed (system is cooled) it will descent
on this ladder. But it is not possible to get lower than E1 , even not when cooled to 0 K. So even at the
absolute zero, the particle will still have some kinetic energy, the zero-point energy. The random motion
that is associated with the kinetic energy is called the zero-point motion. In agreement with the uncertainty
relation, the particle cannot be completely at rest.

Now that we know the state function of the system, we can use it to predict the expectation values of
dynamic variables. The expectation value of the position (= the average position of the particle) we can
calculate by
∫ ( nπx ) ( nπx )
2 L
⟨x⟩ = ⟨ψn | xψn ⟩ = sin x sin dx
L 0 L L
It is convenient to evaluate this integral by switching to a new variabele u, defined as
nπx L
u= → dx = du
L nπ
It follows that
x=0 → u=0, x = L → u = nπ
so that the upper and lower boundary for u become 0 and nπ. In terms of the variable u, the integral is
given by
( )2 ∫ nπ
2 L
⟨x⟩ = u sin2 u du
L nπ 0
( )2 [ 2 ]nπ [ ]
2 L u u sin 2u cos 2u 2L n2 π 2 L
= − − = 2 2 = (2.5)
L nπ 4 4 8 0 n π 4 2
This answer for the average position seems logical, since the particle would not have a preference for a
particular side of the box. Note that the average position is not necessarily the most likely position to find
( )1/2
a particle. For instance, for the function ψ2 (x) = L2 sin 2πx
L the probability to find the particle at L/2
(see figure 2.1) is zero! However, the probability function is always symmetric around x = L/2. So while

15
the particle is not equally likely to be at any position in the box (which would mean a “flat” probability
function), it is equally likely to be in the left or the right half of the box. Hence the average of many
measurements on its position would yield the middle of the box, in agreement with (2.5).
The expectation value of the moment px follows from

h̄ d
⟨px ⟩ = ⟨ψ ψ⟩
i dx
∫ ( nπx ) d ( nπx )
2 h̄ L
= · sin sin dx
L i 0 L dx L
∫ ( nπx ) ( nπx )
2 h̄ nπ L
= · · sin cos dx
L i L 0 L L
∫ ( )
2 h̄ nπ 1 L 2nπx
= · · · sin dx
L i L 2 0 L

1 h̄ nπ 1 h̄ nπ
= · sin(2u)du = − · [cos(2u)]0 = 0
L i 0 2L i

So the particle moves to the left as often as it moves to the right, which is again in agreement with the
particle being equally likely in the left and right half of the box. We can show that the particle is not
standing still by calculating the expectation value for p2x , which yields (see exercises):

⟨p2x ⟩ = 2mEn

We summarize what we have learned so far for the particle on a line of length L:

• In order to get the stationary states, we should solve Schrödinger’s equation (2.1).

• The first boundary condition (state should be zero at x = 0) reduces the general solution (2.3) to a
sine function.

• The second boundary condition (state should be zero at x = L) gives that the function should “fit” on
the line: it dictates that the wavelength of the sine can only be 2L, 2L/2, 2L/3, 2L/4, ...

• The condition that ψ should be normalized gives the amplitude of the wave.

• Each wave has a different energy, so the energy is quantized.

• The energy increases with the number of nodes in the wave.

2.2 Harmonic oscillator


There are many systems in nature which can be described by particles that oscillate around an equilibrium
position. We can consider such a particle as attached to a spring, which applies a force F that scales linearly
with the distance x of the particle from the equilibrium position, and directed towards this position. In
equation form:
F = −kx
where k is the spring constant. To describe such a system quantum-mechanically, we need the potential
energy V (x) associated with this force. The relation between force and potential energy is by definition
F = −dV (x)/dx (see appendix A), so that for the harmonic oscillator
1 2
V (x) = kx + C
2

16
The potential energy is in principle a relative energy, so you always have to define where its value is zero.
We choose V (x) = 0 for x = 0 (so at the equilibrium position), which corresponds to setting C = 0. The
Hamiltonian for this system then reads:

h̄2 d2 1
Ĥ = − + kx2
2m dx2 2
so that Schrödinger’s equation becomes

h̄2 d2 1
− ψ + kx2 ψ = E ψ (2.6)
2m dx2 2
Finding a solution to this equation is less straightforward than for a particle on a line. Clearly a sine, cosine
or e−ax do not satisfy equation (2.6). In exercise 1.2 we have seen that the function e−ax is an eigenfunction
2

2
2 − 4a x , which is of a form that is similar to Ĥ, so that could be a good starting point.
d 2 2
of the operator dx
Let us work this out properly. We pose as solution:
2
ψ(x) = e−ax ,

where a is some arbitrary constant, apply Ĥ to it, and see if we can arrive at some constant times ψ(x).
[ ]
h̄2 d2 1 2 −ax2
Ĥ ψ(x) = − + kx e
2m dx2 2

h̄2 d2 −ax2 1 2
= − 2
e + kx2 e−ax
2m dx 2
2 ( )
h̄ 2 2 1 2
= − −2a e−ax + 4a2 x2 e−ax + kx2 e−ax
2m 2
2 ( 2 2 )
h̄ a −ax2 2h̄ a 1 ax2
= e| {z } − − k x2 |e−{z }
m m 2
ψ(x) ψ(x)
So Ĥ operating on ψ(x) = exp(−ax2 ) yields
( )
h̄2 a 2h̄2 a2 1
Ĥ ψ(x) = ψ(x) − − k x2 ψ(x)
m m 2

Our choice for ψ(x) would be an eigenfunction (with eigenvalue h̄2 a/m) if the last term would be zero. This
can be arranged, however, since we are still free to choose a. That parameter thus should satisfy

2h̄2 a2 1 km
= k → a=
m 2 2h̄
The eigenvalue then becomes √
h̄2 a 1 k 1
E = = h̄ = h̄ω
m 2 m 2

where we introduced ω = k/m, which also has a physical meaning: it is the natural frequency of the mass
+
√ spring system (if you have a mass on a spring, you pull it, and let it go, it will oscillate with a frequency
k/m).
So we managed to find one eigenfunction, yet there are many more. It can be shown that these function are
given as e−ax times a polynomial fn (x) of degree n, where n = 0, 1, 2, ...:
2

( )
−ax
2 1
Eigenfunction: ψn (x) = Cn fn (x) e Eigenvalue: En = n + h̄ω ,
2

17
The set of polynomials fn (x) are the so-called Hermite polynomials, and are given in table 2.1. So again we
naturally get to a discrete set of functions, where the index n is now set by the degree of the polynomial.
Note that the discretization does not follow from the boundary condition as for the particle on the line,
but just follows from the math when solving the eigenvalue equation (2.6). Why this is exactly so is quite
technical, and goes beyond these lectures. We find that the lowest possible energy (the zero-point energy)
of the harmonic oscillator is h̄ω/2. So the solution that we proposed as a starting point turns out to be the
ground state. De normalizing constant Cn is found to be equal to


Table 2.1: Hermite polynomials fn (x), with p = (8a)1/2 x, and a = km/(2h̄).

n fn
0 1
1 p
2 p2 − 2
3 p3 − 6p
4 p4 − 12p2 + 12
5 p5 − 20p3 + 60p
6 p6 − 30p4 + 90p2 − 120

( )1/2
(2a)1/2
Cn =
π 1/2 2n n!
which we will prove for the ground state in the exercises. The functions ψo , ψ1 and ψ2 are shown in figure 2.2.
On the y-axis we show two quantities: i) the energy, where the solid line is the potential energy V (x) = 12 kx2 ,
while the dashed lines give the level of eigenvalues: 12 h̄ω, 32 h̄ω, and 52 h̄ω), and ii) the value of ψ en ψ ψ ∗ ,
where the zero level is at the dashed line. It is useful to consider how the particle would move in a classical
picture. Suppose you stretch the spring to a point that the potential energy of the particle is equal to h̄ω/2.
The particle’s position is then xe . Since you hold the particle static, the kinetic energy is zero, so all the
energy of the particle is potential energy. You then release the particle. It starts oscillating between +xe and
−xe , and the potential energy follows the solid curve in 2.2. In the equilibrium point the potential energy is
zero, and the kinetic energy has a maximum. In the end points it is vice versa. In this picture, the particle
will never get out of the range [−xe , xe ], since it needs more energy to do so. Now the ψ 2 function shows
that a quantum mechanical particle with energy h̄ω/2 can move outside the range set by the parabola, since
ψ 2 is non-zero outside this range. The particle is somehow able to reach a position, for which in a classical
picture it would have not have the energy. This phenomena is called tunneling: the QM particle has found
a kind of tunnel through the potential energy barrier - where a classical particle would need extra kinetic
energy to get over the barrier. The figure shows that ψ 2 decays for x going to ±∞: the further you are in
the “classical forbidden area”, the less likely it is to find the particle there.
An important measuring technique in the nano-technology, Scanning Tunneling Microscopy (STM), is based
on the principle of tunneling. When the tip of a needle is held very close to a solid surface, electrons can
jump from the surface to the tip, while in a classical picture they would not have enough energy to do so. So
this is again a tunneling effect, and the probability that the electrons make the jump (and so the magnitude
of the electric current through the tip) is a very sensitive function of the distance between the tip and the
surface. With this principle, small variations in the surface height can be measured via the current - even
to a level that you can see where atoms are.

2.3 Systems in 3-D; particle in a box


The extension of the postulates to three dimensions is straightforward. The coordinates of the particle are
then given by x, y, z, and the state function will depend on all three coordinates: ψ(x, y, z). The physical

18
ψ2 ψ2ψ2∗

ψ1 ψ1ψ1∗

ψ0 ψ0ψ0∗

Figure 2.2: Eigenfunctions of the harmonic oscillator

meaning of ψ is that ψ ∗ (x, y, z)ψ(x, y, z)dxdydz is equal to the probability to find the particle in the volume
dxdydz around x, y, z. Since the probability is 1 that the particle is somewhere in space, we have the
normalizing condition forψ:
∫ ∞∫ ∞∫ ∞
⟨ψ|ψ⟩ = ψ ∗ (x, y, z)ψ(x, y, z)dxdydz = 1
−∞ −∞ −∞

The extension of eqs. (1.5) - (1.7) to 3-D are:

x, y, z → x̂ = x, ŷ = y, ẑ = z (2.7)
h̄ d h̄ d h̄ d
px , p y pz → p̂x = , p̂y = , p̂z = (2.8)
i dx i dy i dz
2 2 2 2 2
h̄ d h̄ d h̄ d2
E → Ĥ = − − − + V (x, y, z) (2.9)
2m dx2 2m dy 2 2m dz 2

The expression for Ĥ follows again by replacing the variables by operators in the classical expression for
the Hamiltonian in 3-D: H = p2x /2m + p2y /2m + p2z /2m + V (x, y, z). Note the difference between the kinetic
energy and potential energy in 3-D: the kinetic energy can be written as three independent terms, for the
three directions, while the potential energy V (x, y, z) is one function, which in general cannot be written as
three independent terms (e.g Vx (x) + Vy (y) + Vz (z)). This has important consequences, as we will see later.
The Hamiltonian in (2.9) can be written in a more compact form:

h̄2 2
Ĥ = − ∇ + V (x, y, z) (2.10)
2m
with ∇2 = ∇
⃗ ·∇
⃗ = d2 /dx2 + d2 /dy 2 + d2 /dz 2 .

We will now apply these postulates to the most simple 3-D system: a particle that can freely move within
a box of size L × L × L. As for the particle on a line, the particle is kept within the box by making the
potential energy outside it infinitely large, while within the box V = 0:

V (x, y, z) = 0 if 0 < x < L and 0 < y < L and 0 < z < L


V (x, y, z) = ∞ all other values of x, y, z

19
The Hamiltonian for 0 < x < L , 0 < y < L , 0 < z < L is then

h̄2 d2 h̄2 d2 h̄2 d2


Ĥ = − − −
2m dx2 2m dy 2 2m dz 2

We want to find the eigenfunctions of Ĥ, that is, we should solve Ĥψ = Eψ for the given boundary conditions.
These conditions will again give that we arrive at a discrete set of functions, which will be labeled by three
quantum numbers n, l, m. So we want to find the set ψnml (x, y, z) which satisfy
( )
h̄2 d2 d2 d2
− 2
+ 2+ 2 ψnlm (x, y, z) = Enlm ψnlm (x, y, z) (2.11)
2m dx dy dz

Since the operator is written as the sum of three terms which work on x, y and z separately, we can use
the method of separation of variables (see also exercise 1.4 and 1.8). This implies that the solution can be
written as
ψnlm (x, y, z) = ψn (x)ψl (y)ψm (z) (2.12)
where ψn , ψl en ψm satisfy the 1-D eigenvalue equations:

h̄2 d2 ψn (x)
− = En ψn (x) (2.13)
2m dx2
h̄2 d2 ψl (y)
− = El ψl (y) (2.14)
2m dy 2
h̄2 d2 ψm (z)
− = Em ψm (z) (2.15)
2m dz 2
(2.16)

with Enlm = En + El + Em . The solutions of the 1-D equations have already been determined in section
2.1: the solution of (2.13), with boundary conditions ψn (0) = 0, ψn (L) = 0 is given by (2.4); the solution of
(2.14) is similar, only with x replaced by y and n by l, and likewise (but then with z and m) for (2.15). So
the total solution is
( )1/2 ( nπx ) ( ) ( mπz )
8 lπy
ψn,l,m (x, y, z) = 3
sin sin sin
L L L L

and
h̄2 π 2 [ 2 2 2
]
Enlm = n + l + m (2.17)
2mL2
Note that the functions ψ211 , ψ121 , ψ112 all have the same energy: the degeneracy of this level is 3. To
recapitulate: since the Hamiltonian does not contain “mixed terms” (terms in which for instance both x and
y appear, like xy or x1 dy
d
), but only independent terms (containing only x, y, or z), the full 3-D problems can
be written as three 1-D problems, which can be solved independent of each other, meaning that the solutions
for for instance the x direction do not depend on those for the other directions. This also means that the
quantum numbers n, l, m can be chosen freely, and in any combination. Unfortunately, there are very few
systems that allow for such a separation of variables. In the next chapter we will see that even for the most
simple atom, the hydrogen atom, this is not possible, which means that the quantum numbers n, l, m are
connected (for instance n > l).

20
2.4 Exercises
Exercise 2.1
a) Show that Schrödinger’s equation for a single particle confined to a line of the length L can be written
d2 2mE
as ψ(x) + k 2 ψ(x) = 0 with k 2 = .
dx2 h̄2
b) The general solution of this differential equation is ψ(x) = C sin kx + D cos kx. Check that this is true.
c) For which values of C, D and k does ψ(x) satisfy the boundary conditions? Which values can the
energy have? Give a general expression for the normalized function ψ.
d) Sketch the probability density ψ 2 (x) for the ground state and the first excited state.
e) Suppose the particle is in the ground state. Determine ⟨x⟩ and ⟨p⟩, that is, the expectation value of a
measurement of the x position, and the x-momentum. Consider if the result is what you expect.

Exercise 2.2
(a) We have the same system as in exercise 2.1. Calculate ⟨x2 ⟩ and ⟨p2 ⟩ for the ground state and the first
excited state.
(b) The variation in measurements (or ”uncertainty”) of the position ∆x and of the momentum ∆p are
given by
( )1/2 ( )1/2
∆x = ⟨x2 ⟩ − ⟨x⟩2 , ∆p = ⟨p2 ⟩ − ⟨p⟩2
Show that for the ground state ∆x, ∆p satisfy Heisenberg’s uncertainty relation.

Exercise 2.3
We
[ 1 again ] have the same system as in exercise 2.1. What is the probability to find the particle in the interval
3
4 L, 4 L when it is in the ground state?

Exercise 2.4
This exercise concerns that harmonic oscillator: a single particle that can only move in the x-direction, and
the force which drives the particle back the origin is equal to F = −kx. The Hamiltonian of this system is

h̄2 d2 1
Ĥ = − + kx2
2m dx2 2

a) Give the normalized eigenfunction ψo (x) of Ĥ with the lowest eigenvalue, in terms of a = km/2h̄.
b) Show that ψo (x) is normalized.
c) Determine ⟨x⟩ and ⟨x2 ⟩ in terms of a for when the particle is in the ground state ψo .

We have the operator p̂ corresponding with the momentum, and the operator p̂2 corresponding with the
momentum squared. These operators acting on ψo give the following results:

p̂ ψo (x) = − 2ax ψo (x)
i
p̂2 ψo (x) = 2h̄2 a(1 − 2ax2 ) ψo (x)
d) Is ψo an eigenfunction of p̂? And of p̂2 ?

21
e) Determine ⟨p⟩ and ⟨p2 ⟩ in terms of a and h̄, when the particle is in the ground state.

f) Show that ∆x∆p = 21 h̄ (see also exercise 2.2).

Exercise 2.5
Show that the ground state function and the first excited state function for the harmonic oscillator are
orthogonal.

Exercise 2.6
Given a particle confined to a rectangular box with dimensions a × b × c.
a) Give the Hamiltonian Ĥ for the particle.
nx πx ny πy nz πz
b) Verify that ψ(x, y, z) = N sin · sin · sin is an eigenfunction of Ĥ, with eigenvalue
( ) a b c
h̄2 π 2 n2x n2y n2z
E= + 2 + 2
2m a2 b c

c) Normalize ψ(x, y, z).


In the following we take a = b = c = L.
d) Give the energy of the ground state.

e) Give the first three (different) energy levels above the ground state. Consider which combinations of
quantum numbers give these excited state energies, and from this determine the degeneracy of each of
these levels.
f) Give the number of nodal planes for each of these energies.

22
Chapter 3

Hydrogen Atom and Helium Atom;


Periodic System

In order to get to an understanding of concepts, it is useful to consider them first for very simple (yet
perhaps not so realistic) systems. We have done this in the previous chapter, where we have learned how
Schrödinger’s equation and the boundary conditions lead to a set of specific states that the particle can have.
In this chapter we will deal with a system that is much more realistic, while the mathematics of the problem is
still not excessively difficult: the hydrogen atom. Like for the particle in the box and the harmonic oscillator,
we aim to find the eigenfunctions of the Hamiltonian of the hydrogen atom - which are the stationary states
that an electron can have. These states are at the basis of the structure of the periodic table, and so our
understanding of the hydrogen atom is of immense importance to chemistry in general. We will discuss this
in the second part of this chapter, starting with the helium atom. There we will see that by going from a
one-electron system (the H-atom) to a two-electron system (the He-atom) a new and unexpected condition
for quantum mechanical systems is required: namely that particles are indistinguishable. To this end, a 4th
postulate is introduced.

3.1 The hydrogen atom


The study of the emission/absorption spectrum of atomic hydrogen has been pivotal for the development of
quantum mechanics. In the mid 19th century these studies showed that only light with specific wavelengths
could be absorbed. Long before the development of the atomic model by Bohr, it was observed that there
was a certain pattern in the possible wavelengths; for instance, Balmer found in 1888 that λ ≡ n2f /(n2f − 4),
where nf = 3, 4, ..., would exactly fit a certain subset of the wavelengths. This equation was extended by
Rydberg, who suggested the general empirical equation λ−1 ≡ n−2 i − n−2
f , with ni = 1, 2, 3, .. and nf an
integer > ni . The set of wavelengths for ni = 1 is called the Lyman series, for ni = 2 the Balmer series.
In 1913 Bohr gave a physical interpretation to these relations: since the energy of a photon is equal to
hc/λ, Rydberg’s equation gives that the possible energies of an electron state En should be ∼ n−2 , with
n = 1, 2, 3.... The integers ni and nf in Rydberg’s equation are then the values of n of the initial state
and the final state of the electron when the energy of a photon is absorbed by it. Bohr combined this with
the classical picture of the hydrogen atom: a nucleus, and an electron moving in a orbit around it (see also
appendix A). The attractive Coulomb force provides precisely the force that is needed to constantly change
the direction of velocity of the particle when moving in such an orbital. It can be shown that the energy
of the electron is then equal to E is ∼ −L−2 , where L is the angular momentum. So the spectrum could
be explained by assuming that L ∼ n. Filling in all the constants, it turns out that L = nh̄ would also
quantitatively predict the spectrum. However, this model cannot be a true representation of reality, since
measurements show that for the H-atom in the ground state the angular momentum is zero, while the Bohr
model predicts that it is equal to h̄. Also, according to electromagnetic theory, the orbiting electron should

23
emit radiation. This has also not been observed.
In 1926 Schrödinger developed the quantum mechanics on the basis of the hydrogen atom, and showed that
the solutions of his equation Ĥψ = Eψ could explain all the experimental observations of the hydrogen
atom. A detailed derivation of this solution is outside the scope of this introductory course. Below we will
just give the essential steps and results, without going through all the explicit math. However, we will be
able to verify some of the result (see exercise 3.1).

The hydrogen atom consist of two elementary particles: a proton and an electron, which are both free
to move. However, this motion can be described as the motion of the proton, and the motion of the electron
relative to that of the proton1 . In the following, we just consider the latter motion, and put the origin of
the coordinate system at the center of the proton. The only force that the electron feels is the attractive
Coulomb force, the potential energy of which is given by (see appendix A):

−e2
V (r) = (3.1)
4πϵ0 r

where we have defined the zero of potential energy at r = ∞. With this potential, the Hamiltonian of the
H-atom becomes (see equation (2.10)):

h̄2 2 e2
Ĥ = − ∇ − (3.2)
2me 4πϵo r

with r = x2 + y 2 + z 2 the distance of the electron to the proton. Before we proceed any further, it is
convenient to switch to a different system of units. Equation (3.2) has an abundance of physical constants
h̄, ϵo , e, me , etc. It will also be clear that SI units (meter, kg) are not really convenient for the atomic
scale. Rather than having a platinum mass of 1 kg as a reference, it is far more logical to use the mass of
an electron as reference, and express all other masses in units of the electron mass. This brings us to the
introduction of atomic units. In this system, the unit of

Mass = me = 9.10953 10−31 kg


2
Length = ao = 4πε o h̄
me e2 = 5.29177 10−11 m
Time = me a20 h̄−1 = 2.41887 10−17 s
−2
Energy = h̄2 m−1
e a0 = 4.35984 10−18 J
Charge = e = 1.60219 10−19 C
These units have been chosen such that the numerical values of the various physical constants are ex-
actly equal to one: e = me = h̄ = 4πεo = 1 (see also appendix C). In these units, the Hamiltonian of the
H-atom becomes
1 1
Ĥ = − ∇2 − (3.3)
2 r
or written out in Cartesian coordinates x, y, z:
( )
1 d2 d2 d2 1
Ĥ = − + 2+ 2 − √
2 dx2 dy dz x + y2 + z2
2

The last term makes it impossible to write the Hamiltonian as three independent terms Ĥ = Hˆx + Ĥy + Ĥz
where e.g. the Hˆx term would not contain any y, z. Given the spherical symmetry of the potential, the
x, y, z does also not seem the most logical coordinate system. We therefore switch to spherical coordinates
r, θ, ϕ, defined by:
x = r sin θ cos ϕ , y = r sin θ sin ϕ , z = r cos θ
1 formally it is split into the motion of the center of mass, and that of both particles relative to the center of mass. Since the

proton is so much heavier than the electron, the motion of the proton is essentially equal to the motion of the center of mass.

24
With some straightforward but tedious algebra you can show that the ∇2 operator in spherical coordinates
becomes [ ( )]
d2 2 d 1 1 d2 1 d d
∇ =
2
+ + + sin θ
dr2 r dr r2 sin2 θ dϕ2 sin θ dθ dθ
so that the Hamiltonian takes the following form:
[ ( )]
1 d2 1 d 1 1 −1 d2 −1 d d
Ĥ = − − − + + sin θ (3.4)
| 2 dr 2 r
{z dr r} 2r2 sin2 θ dϕ2 sin θ dθ dθ
| {z }
depends only on r depends only on θ,ϕ

This is almost the sum of two independent terms: one depending only on r, the other only on θ, ϕ. Almost,
since unfortunately there is still a r2 present in the last term. This means that a full separation of variables
is not possible; still, a solution of the form ψ(r, θ, ϕ) = R(r)Y (θ, ϕ) to Ĥψ = Eψ can be found, but the
functions R and Y are not independent (owing to the r2 in the last term). The precise derivation is rather
long and technical, and not given here. The result is that, like for the particle in the box, there is a set
of eigenfunction ψnlm of the Hamiltonian (so again characterized by three quantum number n, l, m), the
eigenvalue of which only depends on n:
1
Ĥψnlm = En ψnlm , En = − where n = 1, 2, 3, ... (3.5)
2n2
So the quantum number n can take the values 1, 2, 3, ... as for the particle in the box. However the other
two number l, m cannot take any value: they are bound by the conditions that

l<n ; | m |≤ l (3.6)

So contrary to the particle-in-box system, the three quantum numbers cannot be chosen independent from
each other, which is due to the fact that no full separation of variables was possible. Condition (3.6), and
the fact that the energy depends only on n, is the key to the periodic table. Note that this condition follows
purely from the math when solving Ĥψ = Eψ for Ĥ given by equation (3.3).

The eigenfunctions ψnlm are the well-known s, p en d orbitals of the H-atom, see table 3.1(for the n = 1 and
n = 2 orbitals. The orbitals for n = 3 are shown in Appendix D). These are thus the stationary states of
the atom. Note that for instance 2px is made up out of two different eigenfunctions: 2px = ψ211 − ψ21−1 .
Namely, for m ̸= 0 the function ψnlm is complex. However, the difference of ψ211 and ψ21−1 will be a real
function (to see this you need to know the specific form of ψnlm ), which is very much the same as the 2pz
function, only with x instead of z. In the same way we can make the 2py function by adding. Note that
both 2px and 2py are eigenfunctions of Ĥ, with eigenvalue E2 (why?), so these are also stationary states of
the H-atom. These states - also referred to as the orbitals of the hydrogen atom - are characterized by three
quantum numbers: the principle quantum number n, the azimuthal quantum number l, and the magnetic
quantum number m (in literature often denoted as ml to distinguish it from the secondary spin quantum
number ms ). As stated earlier, the energy only depends on the principle quantum number n. This means
that the states 2s, 2px , 2py , 2pz all have the same energy E2 = −1/8, so these are degenerate states. Since

l = 0, 1, · · · , n − 1 and m = −l, −l + 1, · · · , +l

the degree of degeneracy (= number of degenerate states) of the energy level En equal to


n−1
(2l + 1) = n2 (3.7)
l=0

The azimuthal quantum number l is an indicator of the shape of the orbital. For l = 0 the orbital has a
spherical shape, for l = 1 it has the shape of a “dumbbell”, directed along the x, y or z axis. For higher values

25
Table 3.1: State functions (or orbitals) and of the H-atom in atomic units for n = 1 and n = 2 (for n = 3
see appendix D). Cartesian and spherical coordinates are mixed in these equations. The constants C1 , C2
are such that the functions are normalized.

n, l, m En

1, 0, 0 1s = ψ100 = C1 e−r -1/2

2, 0, 0 2s = ψ200 = C2 (2 − r) e−r/2 -1/8

2, 1, 0 2pz = ψ210 = C2 z e−r/2 -1/8

2px = ψ211 − ψ21−1 = C2 x e−r/2 -1/8


2, 1, ±1
2py = ψ211 + ψ21−1 = C2 y e−r/2 -1/8

Normalizing constants: C1 = (π)−1/2 , C2 = 14 (2π)−1/2

of l, the shapes become more complex.


√ The azimuthal quantum number also sets the size √ of the angular
momentum of the electron: |L| = l(l + 1) (if we would not use atomic units, it would be l(l + 1)h̄). The
magnetic quantum number is more abstract. We just mention that is sets the z-component of the angular
momentum: Lz = m (or mh̄ if we are not using a.u.). So only the length and the z − component of the
angular momentum vector L ⃗ are known for a stationary state. That we do not know all three components
turns out to be in agreement with the uncertainty principle.

A final remark about the constants in front of the states in table 3.1. These are chosen such that the
functions are normalized, that is, ⟨ψnlm |ψnlm ⟩ = 1 for each state n, l, m. For instance, C2 is such that
∫∫∫

⟨ψ210 |ψ210 ⟩ = ψ210 ψ210 dV = 1
∫∫∫
where dV stands for the integration
∫∫∫ ∫over the
∫ ∞total∫ ∞
space, up to infinity. For Cartesian coordinates

this integral can be written as dV = −∞ dx −∞ dy −∞ dz (which was used for the particle-in-a-box).
∫∫∫ ∫∞ ∫π ∫ 2π
For spherical coordinates, however, it is equal to dV = 0 r2 dr 0 sin θdθ 0 dϕ, so the normalization
condition gives (using ψ210 = 2pz = C2 r cos θ e−r/2 ):
∫ ∞ ( )2 ∫ π ∫ 2π
−r/2
C2 r cos θ e 2
r dr sin θdθ dϕ = 1 →
0 0 0
∫ ∞ ∫ π ∫ 2π
−r 1
C22 4
r e dr 2
(cos θ) sin θdθ dϕ = 1 → C22 · 32π = 1 → C2 = (2π)−1/2
0 0 0 4

26
Thus far, the results are for an electron in the presence of a single proton, so for an atom with a nucleus
that has a charge +1 (in a.u.). For systems where the nucleus as a charge +Z, the Hamiltonian becomes
Ĥ = − 12 ∇2 − Z/r. The eigenfunctions of this Hamiltonian are almost identical to those listed in table 3.1,
only with r replaced by Zr, and the constant should be multiplied by Z 3/2 to ensure normalization. So for
instance:
1s(Z) = ψ100 = C1 Z 3/2 e−Zr , 2s(Z) = ψ200 = C2 Z 3/2 (2 − Zr) e−Zr/2
(Z) (Z)
(3.8)
(Z)
The eigenvalue is equal to En = −Z 2 /(2n2 ).

The eigenfunctions of the H-atom will also provide the basis for our description of atoms with more than one
electron, starting with Helium, so two electrons. Before doing so, we must first extend the theory of quantum
mechanics - which was thus far for only one particle - to two particles. This is partly straightforward, but it
also turns out that an extra postulate is needed to make the theory work.

3.2 Quantum mechanics for systems with two particles


We consider two particles in one dimension, so on a line. The coordinates of particle 1 and 2 are given
by x1 en x2 , respectively, and the state function is now dependent on both coordinates: ψ(x1 , x2 ). The
extension of the postulates to two particles is similar to the extension to 3D as discussed in section 2.3, only
that instead of going from x → x, y, z we now go from x → x1 , x2 . The physical meaning of ψ is now that
ψ ∗ (x1 , x2 )ψ(x1 , x2 )dx1 dx2 is equal to the probability to find particle 1 in the section dx1 around x1 , and
simultaneously particle 2 in the section dx2 around x2 . Since the probability is 1 to find the particles at
some position in space, the normalizing condition for ψ is now:
∫ ∞∫ ∞
⟨ψ|ψ⟩ = ψ ∗ (x1 , x2 )ψ(x1 , x2 )dx1 dx2 = 1
−∞ −∞

The extension of equations (1.5) - (1.7) to two particles is:

x1 , x 2 → x̂1 = x1 , x̂2 = x2 (3.9)


h̄ d h̄ d
p x1 , p x2 → p̂x1 = , p̂x2 = (3.10)
i dx1 i dx2
h̄2 d2 h̄2 d2
E → Ĥ = − − + V (x1 , x2 ) (3.11)
2m dx21 2m dx22

But this is not all. In case the particles are electrons, it turns out that the state function ψ(x1 , x2 ) should
satisfy an extra condition, which we formulate as a postulate:

Postulate 4:
The state function of a two-electron system must be anti-symmetric, that is, when changing the labels 1 and
2, the state function should switch sign: ψ(x1 , x2 ) = −ψ(x2 , x1 )

This is a very abstract condition, of with both the logic and the consequences are not directly clear. Yet, it is
a crucial postulate for atomic and molecular chemistry: one of the consequences is namely that two electrons
cannot have exactly the same state (the Pauli exclusion principle), which is is that basis of the periodic table.
Already in 1924, so before the quantum mechanic theory was developed by Schrödinger and Heisenberg,
Wolfgang Pauli had concluded on the basis of spectroscopic measurements that no two electrons can be
in the same state of motion. With the development of the quantum theory, the Pauli exclusion principle
could be explained in the terms of the condition of anti-symmetry of the state function for electrons. The
derivation is not given here, but interested students can find it in appendix E; it is not part of the learning
goals of this course.

27
3.3 The helium atom
The He-atom consists of a nucleus with charge +2e and two electrons. As for the hydrogen atom, we consider
the nucleus fixed in the origin, and provides the force field in which the electrons moves. We are thus dealing
with a 3-D system with two particles, for which the Hamiltonian is generally given by

h̄2 2 h̄2 2
Ĥ = − ∇1 − ∇ + V (⃗r1 , ⃗r2 ) (3.12)
2m 2m 2
where the first two terms represent the kinetic energy, and the last term the potential energy of the electrons,
which reads
2e2 2e2 e2
V (⃗r1 , ⃗r2 ) = − − +
4πϵ0 r1 4πϵ0 r2 4πϵ0 |⃗r1 − ⃗r2 |
In this expression, the first and the second term represent the attraction of electron 1 and 2 by the nucleus,
while the last term represent the repulsion between the two electrons. Switching to atomic units, the
expression becomes
1 1 2 2 1
Ĥ(⃗r1 , ⃗r2 ) = − ∇21 − ∇22 − − + where r12 = |⃗r1 − ⃗r2 | (3.13)
2 2 r1 r2 r12
If the Hamiltonian could be written as the sum of two terms - one which only contains label 1, the other
only label 2 - then we could apply separation of variables, and treat the “two-electron problem” as two
“one-electron problems” which could be solved independent of each other. Unfortunately2 the Hamiltonian
contains the mixed term 1/r12 , which means that we cannot use separation of variables. Up to now, no
exact solution has been found for this problem, and we should resort to approximations3 . We start by
simply ignoring the last term in 3.13: so we consider a He-atom in which the electrons are attracted to the
nucleus, but do not feel each others presence. It will be clear that each electron will then be in hydrogen-like
state (it only feels a nucleus), but with a charge +2 (so the states like those of (3.8) with Z = 2). We will
do it formally: if the variables r1 and r2 can be fully separated, then the solution of Schrödinger’s equation
is the product function:
ψ(⃗r1 , ⃗r2 ) = ψa (⃗r1 ) · ψb (⃗r2 )
where the functions ψa and ψb are the solutions of the one-particle problems:
[ ]
1 2
− ∇21 − ψa (⃗r1 ) = ϵa ψa (⃗r1 ) (3.14)
2 r1

and [ ]
1 2
− ∇22 − ψb (⃗r2 ) = ϵb ψb (⃗r2 ) (3.15)
2 r2
where E = ϵa + ϵb , the sum of the one-particle energies. The one-particle equations (3.14) and (3.15) are
Schrödinger’s equation for the He-ion (single electron the field of a nucleus with charge 2). The label a
in these equation stands for the three quantum numbers na , la , ma , and likewise b for the three quantum
numbers nb , lb , mb ; ϵa and ϵb are the energies of the He-ion, and only depend on na and nb respectively.
Since the total energy E = ϵa + ϵb , we get the lowest energy when both states ψa and ψb are in the ground
(2) (2) (2)
state of the the He-ion, so ψa = ψ100 = 1s(2) , ψb = ψ100 = 1s(2) , for which ϵa = ϵb = E1s . So the lowest
energy state of the two-(independent)-electron system is

ψo (⃗r1 , ⃗r2 ) = 1s(2) (⃗r1 ) · 1s(2) (⃗r2 ) (3.16)

in other words, both electrons have the same state: the 1s state with Z = 2. This would be in violation of
the Pauli principle. However, we are “saved” by the spin state, as we discuss next.
2 At least for solving it. Fortunately the term is there, otherwise the world would not exist as we know it.
3 Note that also classically a three-body problem like sun-earth-moon cannot be solved exactly.

28
In the previous sections we have assumed that the state function of a single particle is only a function
of the spatial coordinate(s) (x, or r). In the one-dimensional case this resulted in a state function charac-
terized by one quantum number, in the three-dimensional case this resulted in a state function characterized
by three quantum numbers. However, there is an additional state of the particle, we which have not con-
sidered up to now: the spin state. An electron can have two possible spin states: positive (or “up”) and
negative (or “down”), so the state can be characterized by an additional quantum number s which has the
value +1/2 or -1/2 (there is a reason why it is not just +1 or -1, which we will not discuss here). So the
total single-particle state function of for instance electron 1, is given by a state function ψa,sa (1), where a
represents the “spatial” state (so na , la , ma ) and sa the spin state. The single-particle state including the
spin state is called a spin-orbital. What is now essential is that the spin state should be included with regard
to Pauli’s principle that no two electrons can have the same state. That is, electrons are allowed to have
the same spatial state (so na = nb , la = lb , ma = mb ) as long as the spin state is different (sa ̸= sb , so they
should have opposite spins). Since there are only two different spin states, there can be at most two electrons
with the same spatial state. At first sight it seems that the spin state is rather abstract, “invented” to allow
two electrons in the same orbital under Pauli’s principle. However, the spin is really a property inherent to
the electron, and can be measured: the Stern-Gerlach experiment shows that electrons posses a magnetic
moment which can have two possible values, and which cannot be caused by their spatial motion. Since
classically a magnetic moment is associated with a rotation of charge, the image of a “spinning” electron
emerges. However, there is no physical basis for such a model. We just observe that the electron has a
magnetic moment, in the positive or negative direction.

The conclusion is that the ground state ψo of the He-atom as given by (3.16) does not violate Pauli’s
exclusion principle, as long as the electrons have different spins. However, the function (3.16) is still not the
true eigenfunction of (3.13), but of the Hamiltonian without the last term (the electron-electron interaction).
(2)
We had seen that the approximate function has a total energy 2E1s ; we can get a better estimate of the
energy of the He-atom by calculating the expectation value of the energy using the true Hamiltonian (3.13)
with the approximate function state (3.16), which can be shown to be equal to (exercise 3.5):
(2) 1
⟨E⟩ = ⟨ψo |Ĥ|ψo ⟩ = 2E1s + ⟨ψo | |ψo ⟩ (3.17)
r12
Calculating the term ⟨ψo | r112 |ψo ⟩ requires some math skills, and gives 1.25 a.u. You can check for yourself
(2)
(see section 3.1) that E1s = −2 a.u. So the total result is ⟨E⟩ = −4 + 1.25 = −2.75 a.u., which is only 5%
larger than the experimental value of -2.90 a.u.! The origin for this difference is that we have not properly
included the electron-electron repulsion, since it was completely ignored in the approximate function ψo that
we used to calculate ⟨E⟩. We can get to a better approximate function by considering what the effect of the
second electron would be on the state of the first electron. If the first electron would be far away from the
nucleus, it would feel the second electron as shielding the +2 charge of the proton, and it would effectively
feel a charge +1. So in that case, the 1s(1) function (a hydrogen 1s function) would be more appropriate as
state function. When the first electron get closer to the nucleus, it feels the shielding of the second electron
less and less, where very close to nucleus the shielding would be almost absent. In that case, the electron feels
the full +2 charge, and the 1s(2) would be more appropriate. For an average distance somewhere in between,
you would expect that a value somewhere in between 1 and 2 would be the best choice. So to effectively
take into account the effect of the other electron, we let the electron feel an effective (screened) charge Z ′ ,
′ (Z ′ )3/2 −Z ′ r
which would mean that the orbital would be 1s(Z ) = e . We now determine the optimal value
π 1/2 ′ ′

for Z by calculating the expectation value as in (3.17), but now with ψo (⃗r1 , ⃗r2 ) = 1s(Z ) (⃗r1 ) · 1s(Z ) (⃗r2 ), and

find the value of Z which gives the lowest result for ⟨E⟩. This yields
Z ′ = 1.6875 for which E = −2.85 a.u.
which is very close (within 2 %) to the experimental value of -2.90 a.u. In general, the method where
which you have a “trial” state function, which depends on some parameter (in this case Z ′ ) and where you

29
minimalize the energy with respect to this parameter, is called the variational method. It is based on the
principle that any approximate state will always yield an expectation value for the energy that is higher
than that of the “true” state. So the lower you can get the expectation value with your approximate state,
the closer you are to the true energy.

3.4 Higher atomic numbers; Periodic system


We start with the lithium atom. We again use the independent particle model, where we have an approximate
solution which is the product4 of single-electron functions, and the electron-electron interaction is included
as an effective charge. We will not give the full details of such a description; the independent particle model
now just means that each electron “occupies” a single electron state (oribtal); taking into account Pauli’s
principle, this gives that the occupied orbitals for the ground state of lithium are:

Li : 1s, 1s, 2s shortly (1s)2 , 2s

where we omitted the superscript which we used before to indicate the (effective) nuclear charge. We now just
assume that the orbitals are those that have the most effective Z ′ for this particular atom. The configurations
of other atoms we can arrive at by adding more atomic orbitals:

Be : (1s)2 (2s)2
B: (1s)2 (2s)2 , 2px or
(1s)2 (2s)2 , 2py or
(1s)2 (2s)2 , 2pz

which is also written as [Be],2px or [Be],2py or [Be],2pz . For carbon we arrive at a problem, since the energies
of the px and py orbital are the same. Should we then put two electrons in the px orbital, with opposite
spin ((1s)2 (2s)2 (2px )2 ) or should we put one electron in the px orbital, and the other in the py orbital
((1s)2 (2s)2 , 2px , 2py )? It turns out that when we have electrons in different spatial orbitals, they are on
average further away from each other as compared to when they are in the same spatial orbital. So electrons
that are in different orbitals have on average less Coulomb repulsion, so this is energetically more favourable.
So the rule is that if the energy levels of the orbitals are the same, the electrons will distribute themselves
as much as possible over different orbitals. This is the essence of Hund’s rule. A more solid explanation
can be found in appendix E.

For the hydrogen atom all states with the same primary quantum number have the same energy, so E2s = E2p .
For atoms with more electrons this is not true anymore. Electrons in different orbitals are shielded differently
from the nucleus by the other electrons. Because of this the 2s orbital will have a lower energy than the 2p
orbital. Similarly, the 3p orbital will have a lower energy than the 3d orbital; for the same primary quantum
number n, the energy of the orbitals increases with l. How much it increases depends also on the number of
electrons present in the atom, so on the atomic number (Z). In figure 3.1 we show the energy of the different
orbitals as a function of the atomic number. Note that for K (Z = 19) we get that E3d > E4s , so the orbital
with the lower primary quantum number has a higher energy! The configuration for K is thus K: [Ar], 4s.
Two atomic numbers “further” we have scandium. For this atom the 3d-orbital turns out to have a lower
energy than the 4p orbitals, so the configuration is

Sc : [Ar] , (4s)2 , 3d

When increasing the atomic number further, the d orbitals will be filled before the 4p orbitals, so that

Zn : [Ar] , (4s)2 (3d)10 , Ga : [Ar] , (4s)2 (3d)10 , 4p

4 formally this product should be made anti-symmetric

30
0,1
1/2
(E / EH)

10
1 10 100
Atomic number Z

Figure 3.1: Energy levels E of the various atomic orbitals as function of the atomic number Z. The energies
(1)
are normalized by EH = E1s = −1/2 a.u., the energy of the ground state of hydrogen.

31
Exercises
Exercise 3.1
Show that the 1s function (see table 3.1) is an eigenfunction of the Hamiltonian of the hydrogen atom, with
eigenvalue -1/2 (in atomic units). Also show that the function is normalized.

Exercise 3.2
When an electron falls back from the first excited state to the ground state, the energy “gain” is used to
create a photon. Determine the wavelength λ of the photon, on the basis of the quantum mechanical model
of the hydrogen atom (see chapter 1 for the energy of a photon). The experimental value is 121.6 nm.

Exercise 3.3
Give the Hamiltonian and the 1s function for the He+ ion. Show that this function is an eigenfunction of
this Hamiltonian, and give the eigenvalue. Also show that the function is normalized.

Exercise 3.4
We have given a single particle in 1-D (so on a line) confined to an interval of length L, that is, the potential
energy V (x) = 0 for x ∈ [0, L] and V (x) = ∞ for x ̸∈ [0, L].
√ ( )
a) Show that the function ψn (x) = L2 sin nπx L is an eigenfunction of the Hamiltonian of this system
and show that this function is normalized.
We now consider two non-interacting particles confined to the same interval on the line. We indicate the
positions of particle 1 and 2 by x1 and x2 , respectively.
b) Give Hamilton operator Ĥ for this system.

c) Show that ψna ,nb (x1 , x2 ) = ψna (x1 )ψnb (x2 ) is an eigenfunction of Ĥ, and give the eigenvalue.

Exercise 3.5
The Hamiltonian for the helium atom is given by (in atomic units):
1 1 2
Ĥ = Ĥ1 + Ĥ2 + , Ĥi = − ∇2i − (3.18)
|r1 − r2 | 2 ri

a) Show that if we ignore the last term in Ĥ (so the term with |r1 − r2 |), that the function ψ(r1 , r2 ) =
(2)
1s(2) (r1 )1s(2) (r2 ) is an eigenfunction of Ĥ, with eigenvalue 2E1s . In this, 1s(2) is the ground state of
(2)
He+ , and E1s the corresponding energy (see exercise 3.3.). Note that the superscript “(2)” refers to
the value Z = 2 for helium.

b) We now use ψ(r1 , r2 ) to calculate the expectation value of the energy, for which we use the “true”
Hamiltonian (3.18), so without ignoring the last term. Show that the expectation value is equal to
∫∫
1
⟨E⟩ = ⟨ψ|Ĥψ⟩ = 2E1s + (1s(r1 ))2 (1s(r2 ))2 dr1 dr2
|r1 − r2 |

32
Chapter 4

The Molecular Hydrogen Ion

In this chapter we make the step from atoms to molecules. We learn the quantum mechanical basis of the
covalent bond for the most simple molecule: the molecular-hydrogen ion. Since this ion contains only one
electron, the states can still be calculated exactly. Nevertheless, we will use an approximate solution method,
the MO-LCAO method, since it allows for a good qualitative insight (which is the main goal of these lectures),
and can also be used for more complex molecules (like O2 ), which cannot be solved exactly. In this chapter,
we will learn some important concepts of the MO-LCAO method such as the secular equations and the secular
determinant, and Coulomb and resonance integrals, which we will use often in the two later chapters.

4.1 Introduction
Already in the 6th century BC the Greek philosopher Democritos argued that matter was build up of atoms.
For various reasons, but foremost because there was not any solid empirical basis for this hypothesis, the
idea went of out fashion for more than 2000 years. In the 17th century there was a renewed interest in the
atomic hypothesis (Descartes, Boyle), and gradually the number of phenomena that could be explained by
such a model increased. Boyle was the first to clearly suggest the concept of a molecule: namely that matter
is composed of clusters of atoms, which can rearrange. Note that the terms “molecule” and “atom” were
used interchangeably. For instance, in 1811 Avogadro wrote1 : “Suppose that the constituent molecules of
any simple gas (....) are not formed of a solitary elementary molecule, but are made up of a certain number
of these molecules united by attraction to form a single one.”. In this, “constituent molecule” is what we
now call a molecule, while “solitary elementary molecule” is what we now call an atom. The question that
of course arises is why and how these atoms attract each other to form a molecule, in other words: what
is the origin of the chemical bond? In 1812 Berzelius suggested that electrical forces (which had just been
discovered) were at the basis of chemical bonds, which would be confirmed nearly a century later. Yet,
for homo-nuclear molecules like N2 and O2 it seemed unlikely that the two atoms would have an opposite
charge, so the model could not be universally true. The discovery of the electron provided a gateway for new
theories. The bonding between two atoms with a large difference in electronegativity could be explained by
a model where an electron jumps from one atom to the other. The resulting ions would be held together by
the attractive Coulomb force: the ionic bond. Immediately after the development of quantum mechanics,
Heitler and London provided in 1927 an explanation for the bonding in the hydrogen molecule, the so-called
Valence Bond (VB) model. In these lecture notes, we will take a different approach, and use the Molecular
Orbital (MO) model to explain covalent bonds. For this we start with the most simple molecule for which
the molecular orbitals can still be calculated exactly, the H+
2 molecule.

1 Amedeo Avogadro (1811). ”Masses of the Elementary Molecules of Bodies”. Journal de Physique. 73: 58–76.

33
Figure 4.1: Definition of the coordinates that are used in eqs. (4.1) and (4.2).

4.2 Exact calculation of H+


2
As stated in the introduction, the H+ 2 -ion provides the simplest example of a covalent bond. The energy
of the ion proves to be a function of the distance R between the nuclei, with a minimum at R = Re . The
energy that is required to fully separate the two nuclei (R → ∞) is called the dissociation energy. In the
quantum mechanical description of the ion we again treat the nuclei as classical objects, like we did for the
hydrogen and helium atom. We put the nuclei at a distance R from each other, and keep them fixed in the
frame of reference in which the electron moves. We now taking the same elementary steps as before:

1. Construct the Hamiltonian of the electron in the H+


2 -ion;

2. Solve Schrödinger’s equation → gives state functions of the electron;

3. Calculate expectation values of variables using the state function


⃗ A en R
If the positions of nucleus A and B are given by R ⃗ B , respectively, then the Hamiltonian of the electron
is equal to (in atomic units):
1 1 1
Ĥ = − ∇2 − − (4.1)
2 | ⃗r − R⃗ A | | ⃗r − R ⃗B |

Using the short-hand notation rA =| ⃗r − R


⃗ A |, rB =| ⃗r − R
⃗ B | gives

1 1 1
Ĥ = − ∇2 − − (4.2)
2 rA rB

It should be stressed that while it seems that we have two coordinates (rA , rB ), there is only 1 coordinate:
⃗r, the position of the electron. This in contrast to the Hamiltonian of the Helium atom (3.13), which has
a similar form, but with r1 and r2 instead of rA and rB . There r1 and r2 is the distance of electron 1 and
electron 2, respectively, to the single nucleus. In equation (4.2) rA and rB are the distance of the single
electron to nucleus A and to nucleus B, respectively.

We have to solve Schrödinger’s equation


Ĥϕ = Eϕ
with Ĥ given by 4.2. Note that we have replaced ψ by ϕ, which is the more customary symbol for the single-
particle state function in a molecule. Solving the eigenvalue problem (4.2) can be done analytically, but it
requires quite some advanced math skills: you should rewrite the equations in terms of elliptical coordinates,
where the focal points of the ellipse are the position of nucleus A and B. This will again give a discrete set
of states, ϕexact
1 , ϕexact
2 , ϕexact
3 .... (which we call molecular orbitals) with eigenvalues E1exact , E2exact , E3exact , ....
Note that we add the superscript “exact”, to distinguish from the approximate solutions that we introduce

34
Figure 4.2: The state function of the electron in the H+ exact
2 -ion for the ground state ϕ1 and the first excited
exact
state ϕ2 , for an inter-atomic distance of 2 a.u.. On the top the contours are shown in the plane of the
molecule, below the value of ϕ is plotted along the inter-nuclear axis. Note that ϕexact
2 is zero in the plane
halfway (and perpendicular to) the inter-nuclear axis: the function has a node (or actually a nodal plane),
as we had also seen for the first excited state in the examples of chapter 2.

later. The lowest two molecular orbitals ϕexact1 and ϕexact


2 (so the ground state and first excited state) are
shown in figure 4.2. The eigenvalue (so the energy) for the ground state E1exact can be seen in figure 4.3.
Not surprisingly, the electron energy depends on R (the distance between the nuclei), which was a parameter
in the calculation. It is instructive to considering the two limits R → ∞ and R → 0. In the first limit the
electron will stay bounded to one of the two nuclei, so that the other nucleus will become an isolated proton.
So in the limit R → ∞ the electron energy of H+ 2 in the ground state should approach that of a hydrogen
atom in the ground state, which is -1/2 (in a.u.). The scale in figure in 4.3 does not go far enough to clearly
exact
observe this, but the actual data for Eelectron for larger R shows that it indeed approaches the value -0.5.
For the other limit, R → 0, the two protons are on top of each other, so the electron feels a single nucleus
with charge +2. So the energy should become equal to (see chapter 3) −1/2 · 22 = −2. It can be seen in
figure 4.3 that indeed this limit is reached. Note also that the approximate LCAO result has the wrong limit
(-1.5). We will come back to this when discussing the method.

35
1

Erepulsion
0.5

0 LCAO
EH +
Energy (atomic units)

-0.5
exact
EH +
2

+
-1 E1LCAO

-1.5
exact
E1

-2
0 1 2 3 4
R (atomic units)

Figure 4.3: Energy of the H+ 2 -ion for the ground state. Shown is both the exact result for the electron energy
E1exact and the result of the approximate LCAO calculation E1LCAO , as given by eq. (4.32). The total energy
of the ion in the ground state is EH+ = E1 + Erepulsion
2
.

36
To get the total energy of the system, we should add the repulsion energy of the two nuclei:
exact
EH+ = E1exact + Erepulsion (4.3)
2

Since both nuclei have a charge +1 (in a.u.), the repulsion energy is

1
Erepulsion = (4.4)
R
exact
The value of EH + as function of R is also shown in figure 4.3. The curve has a minimum - 0.60263 for
2
R = Re = 1.9972, which is indeed equal to the experimental value, and yet another validation of the quantum
mechanical theory. The dissociation energy is then De = −0.5 − 0.60263 = 0.10263 a.e., or 2.793 eV.

4.3 Approximate solution method: MO-LCAO


If we look at the shape of ϕexact
1 in figure 4.2, it looks very much like the sum of two 1s functions separated
by a distance R. In figure 4.4 we take a closer look at this, and the conclusion is that ϕexact
1 can indeed be
reasonably well described by the sum of a 1s hydrogen function centered at nucleus A (which we call 1sA )
and a 1s hydrogen function centered at nucleus B (which we call 1sB ). This makes sense: if the electron
would only feel nucleus A, it would be in the 1sA state, and likewise it would be in the 1sB state if it would
only feel nucleus B. Now if the electron feels both nuclei, a sum of these states seems a reasonable first
approximation. So we pose that we can approximate ϕexact 1 by

ϕexact
1 ≈ ϕ1 ≡ N1 (1sA + 1sB ) (4.5)

Figure 4.2 also suggests that the first excited state ϕexact
2 could be approximated by the difference:

ϕexact
2 ≈ ϕ2 ≡ N2 (1sA − 1sB ) (4.6)

In this 1sA en 1sB are hydrogen orbitals with their center on nucleus A and B, respectively:

1sA = π −1/2 e−rA , 1sB = π −1/2 e−rB (4.7)

and N1 and N2 are normalizing constants. In mathematical terms, equations (4.5) and (4.6) represent
molecular orbitals (ϕ1 , ϕ2 ) that are approximated by a linear combination of the atomic orbitals (1sA , 1sB ),
hence the term MO-LCAO approximation. With the approximate LCAO orbitals ϕ1 and ϕ2 we can calculate
the energy by
E1LCAO = ⟨ϕ1 Ĥ ϕ1 ⟩ , E2LCAO = ⟨ϕ2 Ĥ ϕ2 ⟩ (4.8)

with Ĥ given by 4.2. These integrals can be evaluated exactly (the results are in appendix F for the interested
student), and the resulting curve for E1LCAO is shown in figure 4.3. We observe that the energy of H+ 2 as
calculated with the exact and approximate state function are in qualitative agreement. However, the position
of the minimum for the total energy is somewhat shifted (LCAO: Re = 2.49, Exact: Re = 2.00) and is less
deep (LCAO: −0.565, Exact: −0.603).
In equations (4.5) and (4.6) we have given 1sA and 1sB equal weight in the sum: this is clear from
the exact solution, and in agreement with the symmetry in the molecule: there is no reason why the electron
would prefer to vicinity of nucleus A to that of B, or vice versa. That the AO’s should be subtracted to
form ϕLCAO
2 is less obvious, although you could argue that this way you keep the electron density (∼ ϕ22 )
symmetric, and at the same time create a nodal plane in ϕ2 . We will now show a more systematic approach,
in which we write the MO as a general linear combination of AO’s, where the coefficients are yet to be
determined, and show that the variational method will indeed yield (4.5) and (4.6). So we write

ϕ = cA 1sA + cB 1sB (4.9)

37
exact
φ1
0,4
φ1 = 1sA + 1sB
1sA , 1sB

φ1 0,2

0
-3 -2 -1 0 1 2 3
Position along inter-nuclear axis (a.u.)

Figure 4.4: The exact molecular orbital ϕexact


1 (solid line) for H+
2 in the ground state, compared with 1sA +1sB
(dashed line), where 1sA , 1sB are the hydrogen 1s functions centered at atom A and B, respectively.

Note that from now on we will stick to the LCAO approximation, so we drop the superscript “LCAO”
for the energy in the remainder of the lecture notes, with the understanding that all results are calculated
within the framework of the LCAO method. The coefficients cA and cB can be viewed as the parameters in
an approximate state function (see also section 3.3), that can be used to minimize the energy (variational
method):
∂⟨E⟩ ∂⟨E⟩
= =0 (4.10)
∂cA ∂cB
Note that we have not yet given an index to ϕ: as long as cA and cB in (4.9) can take any value, there
are an infinite amount of ϕ′ s possible. The calculation that follows will show that there are only two sets
of coefficients that satisfy (4.10), so we naturally arrive at a ϕ1 and ϕ2 . So we introduce the labels 1 and
2 a posteriori. A second remark is that the function ϕ as given by (4.9) is in principle not normalized: cA
and cB are not yet known (so the normalization can also only be done a posteriori). So to calculate ⟨E⟩ we
should use the general expression for a non-normalized state function:

⟨ϕ Ĥ ϕ⟩
⟨E⟩ = (4.11)
⟨ϕ|ϕ⟩

Since we have ⟨ϕ|ϕ⟩ in the denominator (where normally it is equal to 1), the normalizing condition is now
an implicit element in the variational approach. Inserting (4.9) for ϕ gives

⟨cA 1sA + cB 1sB Ĥ cA 1sA + cB 1sB ⟩


⟨E⟩ = (4.12)
⟨cA 1sa + cB 1sB |cA 1sA + cB 1sB ⟩

Before we proceed, it is convenient to use the following symbols for the integrals:

HAA ≡ ⟨1sA Ĥ 1sA ⟩ (4.13)

HAB ≡ ⟨1sA Ĥ 1sB ⟩ = ⟨1sB Ĥ 1sA ⟩ (4.14)

HBB ≡ ⟨1sB Ĥ 1sB ⟩ (4.15)


SAB ≡ ⟨1sA |1sB ⟩ = ⟨1sB |1sA ⟩ (4.16)

38
We omit the proof that HBA = HAB . HAA and HBB are called the Coulomb integrals, HAB the
resonance integral, and SAB the overlap integral. With this new notation, equation (4.12) becomes

c2A HAA + c2B HBB + 2cA cB HAB P


E= = (4.17)
c2A + c2B + 2cA cB SAB Q

where we have written ⟨E⟩ simply as E, and in the second step defined the nominator as P , and the
denominator as Q, which is handy for evaluating the derivative of E with respect to cA :
( ) [ ] [ ]
∂E ∂ P 1 ∂P P ∂Q 1 ∂P P ∂Q 1 ∂P ∂Q
= = − 2 = − = − E (4.18)
∂cA ∂cA Q Q ∂cA Q ∂cA Q ∂cA Q ∂cA Q ∂cA ∂cA

Differentiating P and Q with respect to cA gives

∂P ∂Q
= 2cA HAA + 2cB HAB and = 2cA + 2cB SAB
∂cA ∂cA
so that
∂E 2
= [cA HAA + cB HAB − E(cA + cB SAB )]
∂cA Q
∂E
The condition was that ∂cA = 0, so

cA (HAA − E) + cB (HAB − ESAB ) = 0 (4.19)

The derivation of ∂E/∂cB = 0 is very similar, and we skip the steps. The result is:

cA (HAB − ESAB ) + cB (HBB − E) = 0 (4.20)

The set of equations (4.19) and (4.20) are called the secular equations; in matrix form:
[ ][ ] [ ]
HAA − E HAB − ESAB cA 0
=
HAB − ESAB HBB − E cB 0

There is only a non-trivial solution for cA and cB if the determinant of the matrix (the secular determinant)
is equal to zero:
HAA − E HAB − ESAB
=0 (4.21)
HAB − ESAB HBB − E
Equation (4.21) has two solutions for E, which we call E1 and E2 :

HAA + HAB
1) E1 = (4.22)
1 + SAB
HAA − HAB
2) E2 = (4.23)
1 − SAB
where we used that HAA = HBB , since A and B are the same atoms. Both integrals HAA and HAB have
a negative value (we come back to this later), so E1 has the lowest value. We have now found two values
for E, for which the secular equations (4.19) and (4.20) have a non-trivial solution, and thus can be solved.
Inserting E = E1 into the equations gives:
( ) ( )
HAA + HAB HAA + HAB
cA HAA − + cB HAB − SAB = 0 →
1 + SAB 1 + SAB

HAA SAB − HAB HAB − HAA SAB


cA + cB =0 → cA = cB
1 + SAB 1 + SAB

39
Since we inserted E1 , these are the coefficients of lowest molecular orbital ϕ1 in the LCAO, in agreement
with what we expect. You may show yourself that inserting E = E2 into the secular equations leads to

cA = −cB (4.24)

which yields ϕ2 . It is important to realize that we only get a relation between cA and cB , and not an actual
value. The latter we can get from the normalization condition. Using cA = cB for ϕ1 gives:

ϕ1 = cA (1sA + 1sB )

So cA can be viewed as a normalizing constant. We can get the value by requiring

1 = ⟨ϕ1 |ϕ1 ⟩ = ⟨cA 1sA + cA 1sB |cA 1sA + cA 1sB ⟩


= c2A ⟨1sA + 1sB |1sA + 1sB ⟩
= c2A [⟨1sA |1sA ⟩ + ⟨1sA |1sB ⟩ + ⟨1sB |1sA ⟩ + ⟨1sB |1sB ⟩]
| {z } | {z } | {z } | {z }
1 SAB SAB 1
−1/2
= c2A [2 + 2SAB ] → cA = (2 + 2SAB ) , cB = cA = (2 + 2SAB )−1/2

Likewise, starting from


ϕ2 = cA (1sA − 1sB )
we get the value for cA by requiring

1 = ⟨ϕ1 |ϕ1 ⟩ = ⟨cA 1sA − cA 1sB |cA 1sA − cA 1sB ⟩


= c2A ⟨1sA − 1sB |1sA − 1sB ⟩
= c2A [⟨1sA |1sA ⟩ − ⟨1sA |1sB ⟩ − ⟨1sB |1sA ⟩ + ⟨1sB |1sB ⟩]
| {z } | {z } | {z } | {z }
1 SAB SAB 1
−1/2
= c2A [2 − 2SAB ] → cA = (2 − 2SAB ) , cB = −cA = −(2 − 2SAB )−1/2

So the normalized MO’s are:


1sA + 1sB
ϕ1 = √ (4.25)
2 + 2SAB
and
1sA − 1sB
ϕ2 = √ (4.26)
2 − 2SAB
The Coulomb, resonance and overlap integrals are of central importance for understanding molecular bonds,
and we consider them in some more detail. These integral can be written as (see exercise 4.1):
1
HAA = ⟨sA |Ĥ|sA ⟩ = E1s − ⟨1sA | |1sA ⟩ (4.27)
rB
1
HAB = ⟨sA |Ĥ|sB ⟩ = E1s SAB − ⟨1sA | |1sB ⟩ (4.28)
rA
where E1s is the energy of an electron in the hydrogen 1s state, so -1/2 a.u.. In principle, the remaining
integrals ⟨1sA | r1B |1sA ⟩ and ⟨1sA | r1A |1sB ⟩ can be calculated exactly (the results are found in appendix F);
below we will consider them qualitatively, which is more useful for getting an understanding of chemical
bonds. HAA represents an integral over space of the product 1s2A · 1/rB . When the atoms are at their
equilibrium distance, this product is always small: at a point in space where 1s2A is large (close to the
nucleus A), the term 1/rB is small, and vice versa. For a purely qualitative understanding, we can neglect
this term as compared to E1s . The second term of HAB is an type of integral that is reminiscent of the

40
ϕ2 6+ 1
E2 R
6
K·SAB
+ 1−SAB

1s
E1s
− 1+S
K·SAB
AB
ϕ1 6+ 1
E1 ? R

H H+
2

Figure 4.5: Schematic representation of the energy levels of the H2+ -molecule.

overlap integral, only with an extra term 1/rB . We now assume that this second term behaves similar to
SAB : so where SAB is small, this term is small. So we approximate

HAA ≈ E1s , HAB ≈ E1s SAB − K · SAB (4.29)

with K some positive constant. Inserting this into (4.22) and (4.23) gives:

E1s + E1s SAB − K · SAB K · SAB


E1 = = E1s − (4.30)
1 + SAB 1 + SAB
E1s − E1s SAB + K · SAB K · SAB
E2 = = E1s + (4.31)
1 − SAB 1 − SAB

These energies are shown schematically in figure 4.5. We find that the energy level E1 is lower than the E1s
level, and the energy level E2 is higher than the E1s level. The distance between the anti-bonding level E2
and E1s is larger than the distance between E1 and E1s , because 1−SAB < 1+SAB . This becomes even more
pronounced when we consider the total energy of the system, so with the repulsive energy 1/R of the nuclei
added, in which case both levels rise with the same amount as compared to the E1s (see levels on the far right
in figure 4.5). We now consider how these levels change when changing the inter-nuclear distance R. When
R → ∞, the overlap integral SAB goes to zero, and so E1 = E2 = E1s : there are no bonding or anti-bonding
levels. As the nuclei approach each other, the overlap integral SAB increases, and thereby the depth of the
bonding level E1 relative to E1s , attaining a maximum when the nuclei are on top of each other, for which
SAB = 1. However, the 1/R repulsive energy of the nuclei increases when the atoms get closer, becoming
infinitely large for R = 0. At some distance, there is a minimum in the total energy: the equilibrium distance.

We stress again that the equations (4.30) are only useful for getting a qualitative picture, and should not
be used for quantitative results. As stated earlier, it is possible to evaluate the Coulomb-, resonance- and
overlap-integral analytically (see appendix F). Such a calculation gives that
(1 )
1
− R +1 e−2R + (1 + R) e−R
E1 = E1s − R
(4.32)
1 + (1 + R + 31 R2 )e−R

which is the curve E1LCAO shown in figure 4.3. It can be easily checked that with the repulsion 1/R added
to E1 , the energy curve has a minimum -0.5648 at R = 2.49, so a dissociation energy of 0.0648 (all in a.u.).
Note that the exact solution gave a minimum - 0.60263 for R = 1.9972, so a dissociation energy of 0.10263.
In appendix F it is shown that with an effective nuclear charge, the prediction from the LCAO method
can be much improved, giving a minimum - 0.5867 for R = 2.00, very close to the exact (and so also the
experimental) result.

41
4.4 Excited states
Of course, ϕexact
1 and ϕexact
2 are not the only molecular orbitals of the H+ 2 -ion: in principle there is an
infinite set of eigenfunctions of Ĥ, which are the excited states. Again all these orbitals can be reasonably
well approximated as linear combinations of the H-atom orbitals, where we should then also include the
excited states of the atoms (2s, 2px , 2py , 2pz , ...). In general we can write:

ϕi = N (χA + χB ) (4.33)

where χA and χB are orbitals of the H-atom, centered on the nucleus of atom A and B, respectively. The
first question that arises is: should χA and χB always be equal orbitals (like 1sA and 1sB )? Let us consider
a 1s-orbital on nucleus A, and a 2px orbital on nucleus B (see figure 4.6) where we have chosen the inter-
nuclear axis along the z-direction. We again use the variational method. Starting from (4.9), but now with

Figure 4.6: Schematic representation of a 1s orbital on A and a 2px orbital on B.

ϕ = cA 1sA + cB 2px,B , and going through the same steps leads again to the secular determinant

HAA − E HAB − ESAB


=0 (4.34)
HBA − ESAB HBB − E

where now

HAA = ⟨1sA |Ĥ|1sA ⟩ HBB = ⟨2px,B |Ĥ|2px,B ⟩


HAB = ⟨1sA |Ĥ|2px,B ⟩ HBA = ⟨2px,B |Ĥ|1sA ⟩ (4.35)
SAB = ⟨1sA |2px,B ⟩

We consider the overlap integral



SAB = ⟨1sA |2px,B ⟩ = 1sA 2px,B dV (V = whole (infinite) space) (4.36)
V

more closely, for which we again refer to figure 4.6. Now consider an arbitrary point P above the yz plane,
and the mirror image of it (point Q) below the plane, that is, if P has coordinates x, y, z, then Q has
coordinates −x, y, z. It will be clear from symmetry that the value of 1sA in point P and Q are the same,
and both positive. However, the value of 2px,B in point Q will be equal to the value of 2px,B point P , only
with an opposite sign. This means that the value of the product 1sA 2px,B in point P and Q are equal, but
with opposite sign. Generally, for any point in the top half, there is a mirror point in the bottom half where

42
the value of 1sA 2px,B is the same, but with opposite sign. This means that integrating the product over the
whole space (which is the overlap integral, see (4.36)) gives exactly zero: the integral over the “top half”
and ‘bottom half” will be equal, but with opposite sign, so the sum is zero. So

SAB = ⟨1sA |2px,B ⟩ = 0 (4.37)

It was possible to come to this conclusion because we could find a plane (the yz plane) for which the relevant
functions (1sA and 2px,B ) have a different symmetry: 1sA is symmetric, 2px,B is anti-symmetric. Note that
the overlap between 1sA and 1sB is not zero: they are both symmetric. We will often use this symmetry
argument to argue if the overlap is zero or not: it is a key tool for understanding chemical bonds. In exercise
4.3 you will get some practice with this. We now consider

HAB ≡ ⟨1sA |Ĥ|2px,B ⟩ = 1sA Ĥ2px,B dV (V = whole (infinite) space)
V

When Ĥ operates on px,B , the function changes it shape, but it keeps its original symmetry with respect to
the xy plane: so Ĥpx,B will still be anti-symmetric, so that the product of sA and Ĥpx,B , integrated out
over space will also give zero:
HAB ≡ ⟨sA |Ĥ|px,B ⟩ = 0 (4.38)
With this, the secular determinant becomes

HAA − E 0
=0
0 HBB − E

which has two solutions for E (which we again call E1 and E2 ):

E1 = HAA , E2 = HBB

As before, HAA is approximately equal to the energy of an electron in atomic orbital 1sA , so HAA ≈ E1s .
Likewise, HBB is approximately equal to the energy of an electron in atomic orbital 2px,B , so HBB ≈ E2p .
So we have that
E1 = E1s , E2 = E2p , (4.39)
So the energy levels in the molecule are equal to the original energy levels of the atomic orbitals: no bonding
or anti-bonding states are being formed (see also figure 4.7). Note that this is consistent with our earlier
statement that the “depth” of the bonding MO scales linearly with the overlap integral: if it is zero, the
“depth” is zero.

Figure 4.7: Schedule of energy levels in H+


2 in case of (a) non-zero overlap, (b) zero overlap.

43
Using the symmetry argument again, it follows that the overlap between 2px orbitals on both atoms is
non-zero: so this gives rise to bonding and anti-bonding molecular orbitals (see figure 4.8). As can be seen
in the figure, these MO’s have nodal planes. In general, when there is no nodal plane coinciding with the
inter-nuclear axis, as is the case for ϕ1 and ϕ2 , we call them σ-orbitals: the ϕ1 orbital is often referred to as
σg , and the ϕ2 orbital as σu∗ . In this, the “g” and “u” stand for “gerade” (even) and “ungerade” (uneven),
respectively, referring to the presence or absence of a nodal plane perpendicular to the axis (see figure 4.2),
while the ∗ stands for anti-bonding. In case there is a nodal plane coinciding with the inter-nuclear axis, as
in the MO’s shown in figure 4.8, we call them π-orbitals. Note that the anti-bonding π-orbital also has a
nodal plane perpendicular to the axis, so we could indicate it as πu∗ . In general we again see that when the
number of nodal planes increases, the energy increases. Up to now we have assumed that we can use the
pure, original hydrogen orbitals to build our MO’s. However, the presence of the nucleus B does influence
the atomic orbitals on A. This can be partly remedied by again introducing an effective charge in the atomic
orbitals, and find the optimum charge via the variational approach (see appendix F). So the 1sA and 1sB
are then “hydrogen-like” atomic orbitals.

Figure 4.8:

44
Exercises
Exercise 4.1
Derive equations (4.27) and (4.28), that is, show that the Coulomb integral HAA for H+
2 is equal to

1
HAA = E1s − ⟨1sA | | 1sA ⟩
rB
and the resonance integral is equal to
1
HAB = ⟨sA |Ĥ|sB ⟩ = E1s SAB − ⟨1sA | |1sB ⟩
rA
where E1s is the ground state energy of the hydrogen atom.

Exercise 4.2

Consider the overlap integral SAB = ⟨1sA | 1sB ⟩ = 1sA 1sB dV .

a) What is the dimension of SAB ?


b) What is the value of SAB for R = 0, and for R → ∞, where R is the distance between the nuclei?

Exercise 4.3
Given the H+ 2 ion, with the inter-nuclear axis is along the z-direction. On the basis of symmetry, argue
whether the following overlap integrals SAB are zero or non-zero (use the same argument as was done for
SAB = ⟨1sA |2px,B ⟩ with the help of figure 4.6):
a) ⟨1sA |2sB ⟩, b) ⟨1sA |2px,B ⟩, c) ⟨1sA |2py,B ⟩, d) ⟨1sA |2pz,B ⟩, e) ⟨2px,A |2sB ⟩,
f) ⟨2px,A |2px,B ⟩, g) ⟨2px,A |2py,B ⟩, h) ⟨2px,A |2pz,B ⟩, i)⟨2pz,A |2pz,B ⟩

Exercise 4.4
A molecular orbital for a two-atomic molecule AB is given by ψ = N (χA + λχB ), where χA en χB are
normalized AO’s centered on nuclues A and B, respectively. Determine the normalizing constant N of this
MO in terms of λ and the overlap integral SAB .

45
46
Chapter 5

Simple Molecules

In this chapter we use the MO-LCAO method to describe and understand the chemical bonds in a number
of common molecules such as N2 , O2 , H2 O, NH3 , CH4 , etc., where the “aufbau” principle that was used for
atoms, is now also used to “fill” the molecular orbitals. Students will recognize much from what has been
learned in earlier courses, such as that N2 has one σ bonds and two π bonds. However, we can put this
understanding now in a more solid, quantum-mechanical framework, based on the symmetry properties of
the atomic orbitals.

5.1 Introduction: the H2 -molecule


In this previous chapter, the molecular orbitals of the H+
2 molecule were derived. Since this molecule contains
only one electron, the use of molecular orbitals (= 1 electron state) is in principle correct, even though the
orbitals itself may follow from approximate methods. When going from H+ 2 to H2 , the molecule becomes a
two-electron system, and in principle an exact solution cannot be found due to the electron-electron interac-
tion that makes separation of variables impossible. We now make the same approximations that we did for
helium, which we will show in more detail below.

The model that we use for the H2 -molecule is very similar to the H+ ⃗ ⃗
2 -molecule (nuclei fixed at a RA , RB ) ,
only now we have two electrons, the position of which are given by ⃗r1 en ⃗r2 . The Hamiltonian is given by
(in a.u.):

1 1 1 1 1 1 1
Ĥ = − ∇21 − ∇22 − − − − + (5.1)
2 2 |⃗r1 − RA | | ⃗r1 − RB |
⃗ ⃗ |⃗r2 − RA | |⃗r2 − RB |
⃗ ⃗ |⃗r2 −⃗r1 |
| {z } | {z } | {z }
interaction electron 1 interaction electron 2 electron-electron
with nucleus A and B with nucleus A and B interaction

The first two terms are the kinetic energy of the electrons; the other terms are the Coulomb attraction and
repulsion between the various charges in the molecule. Using the notation rAi =| ⃗ri − R
⃗ A |, rBi =| ⃗ri − R
⃗ B |,
and r12 =| ⃗r2 − ⃗r1 | (see figure on the right) this Hamiltonian becomes:

1 1 1 1 1 1 1
Ĥ = − ∇21 − ∇22 − − − − + (5.2)
2 2 rA1 rB1 rA2 rB2 r12

Because of the last term, the electron-electron interaction, the Hamiltonian cannot be separated with respect
to coordinates ⃗r1 en ⃗r2 . If the last term would be absent, then the eigenfunction of Ĥ would be product
function: ϕI (⃗r1 )ϕJ (r⃗2 ), where ϕI (⃗r1 ) is the eigenfunction of the part of the Hamiltonian that only contains

47
⃗r1 , in other words, an eigenfunction of the H+
2 Hamiltonian, as discussed in the previous chapter, and the
same for ϕJ . Postulate 4 requires that the state function is anti-symmetric. This means that electron 1
and electron 2 can be in in the same spatial state, but with different spins. So the ground state can be
approximated by a state in which both electrons occupy the lowest molecular orbital, so by the product

ψ(1, 2) = ϕ1 (r1 )ϕ1 (r2 ) (5.3)

where ϕ1 is the ground state of H+ 2 . Like we did for the He atom, we can include the effect of the other
electron in the approximate state function ψ by taking the charge of the nuclei as a parameter, and minimize
the energy with respect to this parameter. In the remainder of this chapter, we assume that the MO’s that
we use are optimized for the effective electron-electron interaction in this way. Using the approximate state
function (5.3) we can calculate the energy of the electrons ⟨E⟩ as ⟨ψ|Ĥψ⟩, where we use the “true” Ĥ as
given by (5.3) . If we add the proton-proton repulsion energy, we get the total energy in the system. Doing
this for various inter-nuclear distances R we get graph like shown in figure 5.1. Using the most optimal MO’s

Re
0
Energy

De

Inter-nuclear distance

Figure 5.1: Total energy of the H2 molecule as function of the inter-nuclear distance.

ϕ1 in the product 5.3 gives an equilibrium distance of Re = 0.74Å, and the following values for the various
contribution to the total energy in the H2 molecule: :

Total energy of both electrons = - 50.0 eV


Repulsion energy of nuclei = + 19.3 eV
+
Total energy of H2 molecule = - 30.7 eV
Energy of two separate H-atoms = - 27.2 eV
-
Dissociation energy De in the MO-approx. = 3.5 eV
The experimental values are Re =0.74 Å and De = 4.75 eV. When we use MO’s that are not optimized,
we arrive at Re =0.85 Åen De = 2.68 eV. Advanced calculations, in which the electron-electron interaction
is treated more explicitly (instead of just via shielding the nuclear charge) give a value for the dissociation
energy that is within 3 decimal places equal to the experimental value.

Note that the approximations that are needed for the step from H+ 2 to H2 are very similar to the ap-
proximations in the step from He+ to He, the only difference is that instead atomic orbitals we now use
molecular orbitals. This will also hold for the molecular with more electrons: we will again use the same

48
σu∗
@
@
@
@
6 @
1s Q  1s
Q  ?
Q 6 
Q
Q 
 σg
?

H H2 H

Figure 5.2: Energy diagram of the H2 -molecule. It is common to call the ϕ1 and ϕ2 state σg and σu∗ , after
their symmetry properties (see also chapter 4).

6
σu∗
? @
@
@
@
6 @ 6
1s Q  1s
? Q  ?
Q 6 
Q
Q 
 σg
?

He He2 He

Figure 5.3: Energy diagram of the (hypothetical) He2 -molecule.

aufbau principle to fill the molecular orbitals with electrons, taking the Pauli principle and Hund’s rule into
account. The most simple example is the (hypothetical) He2 molecule, which contains 4 electrons, for which
we need 2 molecular orbitals For this we choose the two H+ 2 -orbitals with the lowest energy, ϕ1 (or σg ) and
ϕ2 (or σu∗ ). The energy diagram is shown in figure 5.3. We see that occupying the anti-bonding orbital costs
more energy than what is gained by occupying the bonding orbitals, so it is energetically unfavorable to
make a molecule from two separate He atoms: He2 does not exist.

5.2 LiH
Our aim is to find the MO’s build up out of the AO’s in LiH, and then fill them with the available electrons.
We start by considering an MO, build up out of a 1s-orbital on H, and a 1s-orbital on Li:

ϕ = c1 1sH + c2 1sLi (5.4)

Note that we now use labels 1 and 2, which is more convenient than A and B when using the general
MO-LCAO method. The constant are determined via the variational method. This leads to the secular
equations: [ ][ ] [ ]
H11 − E H12 − ES12 c1 0
= (5.5)
H21 − ES12 H22 − E c2 0

49
which only has a non-trivial solution if the determinant is zero:

(H11 − E)(H22 − E) − (H12 − ES12 )(H21 − ES12 ) = 0 (5.6)

In these equations E = ⟨E⟩, and

H11 = ⟨1sH |Ĥ|1sH ⟩ H12 = ⟨1sH |Ĥ|1sLi ⟩


H22 = ⟨1sLi |Ĥ|1sLi ⟩ H21 = ⟨1sLi |Ĥ|1sH ⟩ (5.7)
S = ⟨1sLi |1sH ⟩

Because of the larger charge of the Li-nucleus, the 1sLi is much more “compact” than the 1sH orbital, and for
this reason the two orbitals have little overlap: S12 is small. When the overlap integral is small, the resonance
integrals are in general also small (see e.g. equation (4.29)). So we use that H12 − ES12 = H21 − ES12 = δ,
a small number. Using this in equation (5.6) gives that

(H11 − E)(H22 − E) = δ 2 δ = small (5.8)

so that
E = E2 ≈ H11 or E = E1 ≈ H22 (5.9)
Now H11 is (close to) the energy of an electron in the 1sLi state, while H22 is (close to) the energy of an
electron in the 1sH state, so that we find that the energies E1 and E2 of the MO’s are thus almost equal to
the energies of the original AO’s. We indicated the solution E ≈ E1s(Li) as E1 , since this energy is lower
than E1s(H) .
The conclusion is as in chapter 4: when two atomic
orbitals have little to no overlap (for instance because
one is much more compact than the other), they do
E2 hhh
not give rise to bonding and anti-bonding MO’s. We H11
show this schematically in the figure to the right.
A slight bonding/anti-bonding can be seen (as com- H22 hhh
E1
pared to figure 4.7, right) because the overlap is not
absolutely zero.
By contrast, the MO ϕ = c1 1sH + c2 2sLi will cer- Li LiH H
tainly lead to bonding and anti-bonding states: the
two AO’s are similar in size, and there will be a sig-
nificant overlap. Very roughly we can say that AO’s
that are similar in size, have similar energy: so in
general we find bonding and non-bonding MO’s are
made out of AO’s that have similar energy levels. Li H
However, the 1sH orbital not only has overlap with
the 2sLi , but also with the 2pz orbital of lithium (see
figure on the right), where the z-axis is the inter-
nuclear axis. So also 2pLi z and 1sH lead to binding
and anti-binding MO’s. The 2pLi Li
x and 2py have no
overlap with 1sH because of anti-symmetry.

50
So more generally, for LiH we should consider MO’s which are formed out of three AO’s:

ϕ = c1 1sH + c2 2sLi + c3 2pLi


z (5.10)

∂⟨E⟩
A similar derivation as shown in chapter 4 (expressing ⟨E⟩ in terms of c1 , c2 , c3 and setting ∂c1 = 0,
∂⟨E⟩ ∂⟨E⟩
∂c2 = 0, ∂c3 = 0) gives the secular equations in the MO-LCAO approximation with three AO’s:
    
H11 − E H12 − ES12 H13 − ES13 c1 0
 H21 − ES21 H22 − E H23 − ES23   c2  =  0  (5.11)
H31 − ES31 H32 − ES32 H33 − E c3 0

The secular matrix is now 3x3, meaning that Det[] = 0 will yield three solutions for E, which we denote as
E1 , E2 and E3 . Inserting each of these into the secular equations, and solving them, will give values for
c1 , c2 , c3 , which describes the MO belonging to that particular energy level. Since we now have three energy
levels, each of which gives three coefficients, it is useful to have an extra label for the coefficients indicating
to which energy level they belong. So the coefficients c1 , c2 , c3 that follow from inserting the value E = E1
into the secular equations we will denote as c11 , c12 , c13 . The first MO is then given as

ϕ1 = c11 1sH + c12 2sLi + c13 2pLi


z

3σ A
The coefficients that follow from inserting the value 
 2σ\ A
 \A
E = E2 into the secular equations we will denote as 2pz \A
\
c21 , c22 , c23 , and the corresponding MO is then \A
2s Z\ \A
Z\ 
1s
ϕ2 = c21 1sH + c22 2sLi + c23 2pLi
z Z\\
Z 

Likewise, the value E = E3 gives
Li LiH H
ϕ3 = c31 1sH + c32 2sLi + c33 2pLi
z

The energy schedule is shown in the figure on the


right, where ϕ1 = 1σ, ϕ2 = 2σ, ϕ3 = 3σ. Each of the
lines represents one of the coefficients.
Since 1sLi , 2pLi Li
x and 2py do not combine with the 1sH , their energy level remains unchanged when a
molecule is formed. So in the molecule these AO’s become non-bonding MO’s, and are denoted by 1s, πx ,
and πy . The total energy schedule for LiH, including all AO’s, is shown in figure 5.4. The molecule has 4
electrons. According to the “aufbau principle” we fill the molecular orbitals with a maximum of two electrons
per level, which gives that the electron configuration for Li in the ground state is equal to (1sLi )2 (1σ)2 . The
electron density is shown in figure 5.5. When compared to the distribution of electrons in the separate atoms
Li and H, we observe that in the molecule there is a shift of the electron density from Li to H: the molecule
has a dipole moment: Li+δ − H−δ .

We close this section with two remarks:

• From symmetry it follows that some terms in (5.11) are zero:

S23 = ⟨2sLi |2pLi


z ⟩ = 0 = S32 , H23 = ⟨2sLi |Ĥ|2pLi
z ⟩ = 0 = H32

which simplifies the calculations.

51
3σ A
A
 2σ \ A
  \ A

πx πy \ A
2px 2py 2pz
\ \A
6 \ \A
2s Z \
Z \ \A 1s
Z\   ?
ZZ
\ 6 
?1σ

1s 6 6
? ?1s
Li LiH H

Figure 5.4: Total energy schedule for LiH.

Figure 5.5: Distribution of electrons in LiH.

• For the construction of MO’s we are free to use AO’s that are not “used” (occupied) in the isolated
atom, like the 2pz orbital of Li atom. In fact, if we would not have used it, that is, we would have
limit ourselves to just linear combinations of 1sH en 2sLi , then the E1 level would be less deep. The
variational approach finds that lowest energy: so any extra variational parameter (in this case c3 )
can only lower the energy, never raise the energy. This does not mean that it is useful to include all
possible AO’s. We have already discussed that the 2px,Li will not contribute to the binding because of
symmetry. Including for the 2s of the H-atom will give only a marginal difference in energy, given the
large energy difference with the 2sLi (for very accurate calculations it can be included, however).

5.3 General formulation of the MO-LCAO method


In the previous section we have seen that we can also construct MO’s out of 3 AO’s, in which we can also
include “empty” AO’s like the 2pz orbital of Li atom. We will now formulate the MO-LCAO equations, as
introduced in section 4.3 for only two AO’s, for the general case of N MO’s. In the MO-LCAO method, we

52
write an MO ϕ as a linear combination of all (relevant) AO’s χ of the atoms present in the molecule:


N
ϕ= cj χj (5.12)
j=1

In the case of LiH, if we are only interested in the bonding MO’s, we would have (see 5.10): χ1 = 1sH ,
χ2 = 2sLi , χ3 = 2pLi Li H Li Li
z . In principle we could include all AO’s (so also 1s , 2s , 2px , 2py ), which would also
yield the non-bonding MO’s.

Note that as before, at this point the MO as defined in (5.11) does not have an index yet. We now follow
the same procedure as in section 4.3: we calculate the expectation value of the energy

⟨ϕ Ĥ ϕ⟩
E = ⟨E⟩ =
⟨ϕ|ϕ⟩

which again can be written in terms of

Hii ≡ ⟨χi Ĥ χi ⟩

Hij ≡ ⟨χi Ĥ χj ⟩
Sij ≡ ⟨χi |χj ⟩ = ⟨χj |χi ⟩

The requirement that


∂E
=0 for all j
∂cj
gives N equations, which can be written in matrix form:
     
H11 − S11 E H12 − S12 E H13 − S13 E ... H1N − ES1N c1 0
 H21 − S21 E H − S E H23 − S23 E ... H2N − ES2N   c2   0 
 22 22     
 H31 − S31 E H − S E H 33 − S33 E ... H3N − ES3N · c3  =  0  (5.13)
 32 32     
 ... ... ... ... ...   .   . 
HN 1 − SN 1 E HN 2 − SN 2 E HN 3 − SN 3 E ... HN N − ESN N cN 0

This set of equations does only have a non-trivial solution when the determinant of the matrix is zero. This
requirement gives N solutions for E, which we list as E1 , E2 , E3 , ..... Each Ei , when inserted into (5.13),
gives a different set of equations, and thus also a different set of solutions for the coefficients, and thus a
different MO ϕi . So we now also label the coefficients with the index i of the energy that was inserted into
the secular equations, since the coefficients are unique to this energy level: cij . The i-the MO is then given
as
∑N
ϕi = cij χj
j=1

For example: the coefficients of ϕ3 = c31 χ1 + c32 χ2 + c33 χ3 + ... + c3N χN follow from solving:
     
H11 − S11 E3 H12 − S12 E3 H13 − S13 E3 ... H1N − E3 S1N c31 0
 H21 − S21 E3 H22 − S22 E3 H23 − S23 E3 ... H2N − E3 S2N     
   c32   0 
 H31 − S31 E3 H32 − S32 E3 H33 − S33 E3 ... H3N − E3 S3N  ·  c33 
   0 
  =  
 ... ... ... ... ...   .   . 
HN 1 − SN 1 E3 HN 2 − SN 2 E3 HN 3 − SN 3 E3 ... HN N − E3 SN N c3N 0

In general these set of equations can only be solved “by hand” for very simple molecules like LiH, and even
then only when some further approximations are made. Also for describing π type bonding in hydrocarbons

53
we can often solve the secular equations with pen and paper. In general, however, one has to resort to nu-
merical methods, and the result that we will show in the next sections for N2 , O2 , H2 O have been calculated
by a computer.

The coefficient cij dictates “how much” the AO χj contributes to the MO ϕi ; So for an electron that
occupies ϕi , the coefficient cij then also gives information about “how much” of this electron we can at-
tribute to the atom to which χj belongs. In other words, we can use the coefficients to give an estimate
of the distribution of the electrons over the different atoms. More specifically, if an electron occupies ϕi ,
then the fraction of the electron that we can attribute to atom A can be estimated as the fraction of the
coefficients squared belonging to A, from the total sum of the coefficients squared:
∑ 2
j∈A cij
∑N 2
j=1 cij

where the sum in the nominator is over all AO’s that belong
∑N to atom A, the sum in the denominator is over
all AO’s. In practice, the coefficient are normalized, so j=1 c2ij = 1. In order to get the total number of
electrons on atom A, we should add the contributions of all the MO’s, taking into account their electron
occupation (0, 1 or 2). In conclusion, the total number of electrons present on atom A in the MO-LCAO
approximation, is equal to ∑ ∑
nA = Ni c2ij
i j∈A

with i the sum over all MO’s, Ni the occupation number of the i-th MO (Ni = 0,1 or 2), while the sum
over j is over all AO’s the belong to atom A. The net charge on atom A (in a.u.) within the molecule is
then QA = ZA − nA , with ZA the nuclear charge. How this will work in practice will become more clear in
the exercises.

54
5.4 N2 ,O2 and F2
In this section we discuss the bonding in the common homo-nuclear molecules N2 ,O2 and F2 , starting with
N2 . For calculations that do not require a very high accuracy it is sufficient to limit ourselves to the core
and valence AO’s of N in the MO-LCAO description. This means that each N atom “brings” the following
5 AO’s into the calculation:
1s, 2s, 2px , 2py , 2pz
So in total there are 10 AO’s in the calculations, from which we can construct 10 MO’s. The result of the
variational approach (as outlined in the previous section) is shown in table 5.1. As an example: 1σg =
.47 · 2sA + .23 · 2pA
z + .47 · 2s + .23 · 2pz ;
B B
E = −42 eV (to verify that the MO’s are normalized, information
on the overlap integrals is needed). The AO’s and MO’s in table 5.1 are organized such, that a block
structure becomes visible: the 2s and 2pz on both atoms (4 AO’s in total) are forming binding combinations
leading to 4 σ MO’s; the two px AO’s lead to two πx MO’s; the two py AO’s lead to πy MO’s; because of
symmetry there is no “mixing” between all these sets. Also both “1s” core orbitals form a separate set. The
set of secular equations (5.13) for N2 , which are at the basis of these results, has a similar block structure.
From linear algebra we know that each of these blocks can be solved independent of each other, leading to
the blocks of table 5.1. The energy schedule of N2 is shown in figure 5.6. Note that each gray line between
AO levels and MO levels can be associated with a (non-zero) coefficient of table 5.1. Next, we fill the MO’s
according to the aufbau principle. The N2 molecule has 14 electrons, hence for the ground state we need to
fill the 7 MO’s that have the lowest energy. This leads to the following configuration for the ground state:

(1sB )2 (1sA )2 (1σg )2 (1σu∗ )2 (1πx )2 , (1πy )2 , (2σg )2

MO
E (eV) ↓ MO/AO → 1sA 1sB 2sA 2pA
z 2sB 2pB
z 2pA
x 2pB
x 2pA
y 2pB
y
-427 1SA 1.00
-427 1SB 1.00
-42 1σg .47 .23 .47 .23
-21 1σu∗ .72 -.27 -.72 .27
-15 2σg .40 -.60 .40 -.60
+30 2σu∗ 1.17 1.22 -1.17 -1.22
-16 πx .62 .62
+7 πx∗ .83 -.83
-16 πy .62 .62
+7 πy∗ .83 -.83

Table 5.1: Coefficients that determine the contribution of the AO’s to the MO’s in N2 . The coefficients are
scaled such that the MO’s are normalized (note: to varify this you need to know the values of the various
overlap integrals

The 1sA and 1sB AO’s are so “deep” inside the atom, that for all practical purposes they do not have any
overlap with the other orbitals, so they are not changed in the molecule: the molecular orbital 1SA = 1sA and
the molecular orbital 1SB = 1sB , and they do not play any role in the chemical bond. Usually these “core
orbitals” are ignored in the MO-LCAO approximation. If you look at the energy levels of the fully occupied
1σg and 1σu orbitals with respect to the 2s level, you can argue (in a simple, qualitative picture) that the
bonding and anti-bonding effects of these MO’s cancel, and so they do not contribute to the bonding. This
leaves three bonding MO’s (as compared to the 2p level) that are occupied by 6 electrons: 2σg , πx , and πy ;
so in a simple picture we can say that N2 has a three covalent bonds: one of the σ type, and two of the π type.

55
2σu +30 2σu

πx∗ πy∗
+7 πx∗ , πy∗
 

 
2pz
H
2px  2py 
Z
JH \ \
JZHZHH \ \
H 2σg π s s
J Z  1σ \\ πx \ 2σg
-15
\ y ss ss
 Z
Z
-16
s s πx , πy
J  

u
-21 1σu
2s 
 
 J
H
HH J
HJ 1σg
H
H
J -42 s s 1σg

1s 1SA 1SB
-427 ss s s 1SA ,1SB

N N2 N N2 N2

Figure 5.6: Energy schedule of N2 , in eV. Since the molecule is fully symmetric, the energy levels of only
one of the N atoms are shown. For clarity, the σ and π levels are shown separately. On the far right the
“filling up” of the energy levels with electrons is shown, where each electron is represented by a dot.

For O2 the MO’s are constructed in the same way as for N2 . Since
the nuclear charge is larger, and the molecule contains two more
electrons, the energy levels of the MO’s are somewhat shifted as 2σu
compared to N2 , but qualitatively the picture is the same (see
scheme on the right). One difference is that the 2σg level for O2
is just below the πx , πy level, instead of just above it as for N2 . s s πx∗ , πy∗
We now have to fill the MO’s with 16 electrons, meaning that two
electrons have to be placed in the anti-bonding π ∗ -orbitals. Also
here Hund’s rule is applied: in order to minimize their repulsion, s s s s πx , πy
s s 2σg
the electrons are put in two different MO’s (πx∗ and πy∗ ). By s s
1σu
filling the anti-bonding π ∗ -orbitals, the bonding effect of one of
the π orbitals is nullified, leaving only two binding MO’s: one s s 1σg
of a σ type, and one of a π type. For this reason, the molecular
bond in O2 is less strong then in N2 :
N2 O2
Dissociation energy, De : 711 494 kJ · mol−1 s s s s 1SA ,1SB
Equilibrium distance, Re : 1,1 1,2 Å
Since there are two electrons with parallel spins, the O2
molecule possesses a magnetic moment, that is, it can be aligned
in a magnetic field: O2 is paramagnetic.
The MO scheme for F2 is again qualitatively similar to N2 and O2 . As compared to O2 , the MO’s need to
accommodate two more electrons, meaning that both the πx∗ and πy∗ orbitals are completely filled, so that
all bonding effect of the π orbitals are canceled. What is left is a single bonding MO of the σ type: the
chemical bond of F2 will again be less strong than that of O2 . Finally, for the (hypothetical) Ne2 molecule,
also the 2σu∗ will be filled, so that all binding effects are canceled: this molecule does not exist.

56
5.5 Water, Ammonia, Methane
In the previous sections the MO theory has provided us with a simple, but very effective picture of chemical
bonds: when two single-occupied atomic orbitals are combined to bonding and anti-bonding MO’s, the two
electrons can occupy the bonding MO, which is the energetically more favorable state. If the atomic orbitals
contain more than one electron, also the anti-bonding MO will be filled, which reduces (or even completely
cancels) the gain in energy of the bonding MO’s. We will use this principle to come to qualitative insight in
bonding in non-linear molecules.

We start with water, H2 O (see also exercise 5.4).

O-configuration: (1s)2 (2s)2 2px , 2py , (2pz )2

In the simple MO picture, only the 2px and 2py AO’s O H


can provide bonding (see also the figure on the right,
where we only considered the single-occupied AO’s).
Each of these AO’s will combine with a 1s AO of
hydrogen, to come to a bonding MO. We have seen
that the bond strength increases with the overlap, x
which has a maximum when both H-atoms approach
the oxygen with 90o angle. Hence we expect a H2 O- y
molecule with an ̸ HOH = 90o . Experiments show H
that the angle is 104o .
More extensive calculations show that there is a shift of the electron density from the H-atoms to the
oxygen, so that there is a negative overall charge −2δ on O, and a positive overall charge +δ on each of the
H atoms. This can directly explain why the angle is larger than 90o : there is a slight repulsion between the
two H atoms because of the positive charge on both. Note that when the angle is not exactly 90o , the x-y
coordinate system as shown in the figure is not the most convenient. In exercise 5.4 we use a different coor-
dinate system, which can allow for any angle. Also, it turns out that the 2s of oxygen also plays a role in the
chemical bond (which shows that the picture of combining only single occupied AO’s is an oversimplification).

For the ammonia molecule, NH3 , we limit ourselves to predicting the geometry. To this end, we consider
the possible options for doubly occupied MO’s. The configuration of the nitrogen atom is

(1s)2 (2s)2 2px , 2py , 2pz

so there are three, half-filled AO’s, each perpendicular to the other. To achieve the largest overlap, the
H-atoms should approach the N atom along the x, y, and z direction, where the 1s-orbital of each H forms a
two-atom type binding and anti-binding MO with one of the 2p-orbitals of N. So we expect a pyramid-type
structure, with ̸ HNH = 90o . Experiments show a larger angle: 107.3o . The NH3 molecule is very similar to
the PH3 molecule. The configuration of P is: [Ne], (3s)2 , 3px , 3py , 3pz . Also here we expect the pyramid-like
structure. The repulsion between the H atoms is less, however, since the PH distance is somewhat larger
than the NH distance. Experiments show that ̸ HPH = 93.3o .

Last, we consider the methane molecule, CH4 . The configuration of the C atoms reads

(1s)2 (2s)2 2px , 2py

on the basis of which one would expect a molecule CH2 with ̸ HCH = 90o . In reality, a CH4 molecule with
a tetrahedral symmetry is formed, that is, a molecule with C in the center of mass, and all HCH angles
equal to 109o . The reason is that it is energetically more favorable for the C atom to first go to an excited
state, where one of the 2s electrons is promoted to the empty 2pz , so (1s)2 (2s)2 2px 2py → (1s)2 2s2px 2py 2pz .

57
Figure 5.7: Left: a single sp3 -hybrid orbital. The larger part on the right is positive, the smaller part on the
left is negative. The line represents lη = 0. Right: the four sp3 -hybrid orbitals in methane, where only the
large, positive parts are shown.

While this “costs” 4 eV, the atom now has 4 single occupied AO’s, and can participate in 4 bonds instead
of just 2, which will more than make up for the loss of 4 eV. These 4 bonds are similar in nature, yet the 2s
orbital of carbon clearly has different symmetry properties as the three 2p orbitals. Here it shows that our
picture of each MO being formed out of two AO’s is too simple. In a more complete picture, the four AO’s
on Carbon (2s, 2px , 2py , 2pz ) are combined with the four 1s AO’s of the hydrogen atoms, giving 8 MO’s,
of which 4 have the same (bonding) level, and 4 have the same (anti-bonding) level, and thus providing 4,
similar type of chemical bonds. While this would be the formally correct approach, for insight it is more
useful to try to stick to a picture with “two AO’s leading to a bonding and anti-bonding MO”. This can be
done by constructing “new” AO’s on Carbon as linear combinations of the 2s, 2px , 2py , 2pz , the so-called
hybrid orbitals:
ηi = ci1 2s + ci2 2px + ci3 2py + ci4 2pz
It should be stressed that these are not MO’s, but rather a new set of AO’s that are constructed such to
support the tetrahedral symmetry of molecule, which the original AO’s did not. In case of methane, the
following 4 hybrid AO’s are formed:
η1 = 1/2(2s + 2px + 2py + 2pz )
η2 = 1/2(2s + 2px − 2py − 2pz ) (5.14)
η3 = 1/2(2s − 2px + 2py − 2pz )
η4 = 1/2(2s − 2px − 2py + 2pz )
Since these orbitals are formed out of one s orbital and three p orbitals, there are called sp3 -hybrid orbitals.
These orbitals are aligned along the 4 directions of the tetrahedron, and are not symmetric or anti-symmetric
with respect to the origin. The hybrid orbitals have a shape where a large, positive part is directed towards
each of the 4 corners of a cube in a tetrahedral structure (see figure 5.7, right). Each of the single-occupied
hybrid orbitals can make a bonding and anti-bonding MO with the 1s orbitals of a hydrogen atom, sustaining
the picture of two AO’s leading to binding/anti-binding MO’s. When CH4 is formed out of pure carbon
(graphite C(s)) and hydrogen gas, an energy 0.77 eV per molecule is released. This process can be analyzed
in terms of the following steps:
• Dissociation of two molecules H2 : 9.02 eV
• Vaporization of Carbon: 7.43 eV
• (1s)2 (2s)2 2px 2py → (1s)2 2s2px 2py 2pz : 3.98 eV
• 4 x CH-bonding energy: -21.20 eV

58
Figure 5.8: Ethane. The CH3 groups are 60o rotated with respect to each other, when viewed along the C-C
axis.

5.6 Ethane, Ethylene, Ethyne


Ethane, C2 H6
The bond between atoms in ethane is very similar to those in methane. Again the configuration of each
of the C-atoms changes from (1s)2 (2s)2 2px 2py → (1s)2 2s2px 2py 2pz , and the 4 single-occupied AO’s are
combined to 4 hybrid sp3 orbitals. Three of those are used for bonds with the H-atoms. The remaining sp3
orbital on each of the C atoms combine, leading to the C-C bond. All H-C-H and H-C-C angles are around
109o . For the overlap between the two sp3 orbitals it makes no difference if one of the CH3 groups is rotated
with respect to the other. There is a slight interaction, however, between the two groups, and the orientation
where the H-atoms have the largest separation will have the lowest energy. The molecule is shown in figure
5.8.

Ethylene, C2 H4
We arrive at a new situation when only four H-atoms are available to form a molecule with two C-atoms.
The bonds are constructed according to the following steps1 . First the single occupied AO’s are formed on
each of the C-atoms, as before: (1s)2 (2s)2 2px 2py → (1s)2 2s2px 2py 2pz . Then only three of these (2s, 2px ,
2py ) are combined to (sp2 ) hybrid orbitals:

1 √
η1 = √ (2s + 2 2px )
3
1 √ √
η2 = √ (2s − 1/2 2 2px + 1/2 62py )
3
1 √ √
η3 = √ (2s − 1/2 2 2px − 1/2 62py )
3
The sp2 -hybrid orbitals η1 , η2 , η3 are directed along three lines that lie in the xy plane, and make an angle of
1 Of course, these are imaginary steps, which the atoms do not actually go through

59
120o with respect to each other. The configuration of each C-atom is then (1s)2 , η1 , η2 , η3 , 2pz . The orbitals
η2 and η3 will form bonds with the H-atoms. The remaining η1 -orbitals on each of the C-atoms are directed
towards each other, and form a σ-bond.

The two single occupied 2pz -orbitals on each of the


C-atoms are not “used” yet. They do also overlap,
to form a bonding π-orbital. Note that since both
2pz ’s do not overlap with any of the other orbitals,
this π bond can be calculated independently of any
of the other bonds (as was also the case in N2 etc.).

When one of the CH2 -groups rotates with respect to the other group, the overlap between the 2pz -orbitals
will decrease, and the energy of the system will increase. This is why the double bond between the two
C-atoms makes the molecule somewhat rigid.

Ethyne (Acetylene), C2 H2
This molecule has the same number of electrons as N2 , and has many similarities with N2 . We make the
same steps as before, where now only the 2px and 2s on each C-atom are combined to make hybrid orbitals:

(1s)2 (2s)2 , px , py → (1s)2 (2s), px , py , pz → (1s)2 , η1 , η2 , py , pz

where η1 and η2 are the sp-hybrid orbitals:


1 1
η1 = √ (2s − 2px ) , η2 = √ (2s + 2px )
2 2
These hybrid orbitals combine with the 1s-orbitals of the H-atoms to make (localised) σ-MO’s:

C
ϕ1 = c1 sH1 + c2 η1 1
C C
ϕ2 = c2 η2 1 + c3 η2 2
C
ϕ3 = c1 sH2 + c2 η1 2

The py and pz AO’s that do not participate in the sp hybridization are used to form πy and πz orbitals:
( ) ( )
C C C C
πy = N py 1 + py 2 , π z = N pz 1 + pz 2

So in total there are three bonding MO’s that are localized between C1 and C2 : ϕ2 , πy , πz . Like in N2 , we
can speak of a triple bond. The table below gives an overview of the different C−C bonds:

Type of bond strength length


C−C 368 kJ mol− 1 1.54 Å
C−C 652 ” 1.34 Å
C−−C 965 ” 1.20 Å

60
Exercises
Exercise 5.1
The simplest picture of bonding in the LiH molecule is provided by constructing MO’s out of linear combi-
nations of only the 1s AO of H, and the 2s AO of Li. Furthermore, to simplify the calculations, the overlap
integral S will be set to zero.
Given: αLi = −10 eV, αH = −13 eV and β = -2 eV, where αLi and αH are the Coulomb-integral of Li and
H, respectively, and β is the resonance-integral.

a) Give the secular equations in matrix form, and determine for which energies the determinant of the
matrix is zero.

b) Calculate the bonding energy of LiH

c) Calculate the (normalized) coefficients of the bonding MO.

d) Calculate the charge on the H-atom when the LiH molecule is in the ground state.

e) Same as d), but now for the first excited state.

Exercise 5.2
We have a two-atom molecule AB, the MO’s of which are constructed as a linear combination of an AO
centered on A, and an AO centered on B.
Given: the ionization energy for atom A is +15 eV, and for atom B is +7 eV. The resonance integral for the
two AO’s has a value of -3 eV. The value of the overlap integral can be neglected as compared to the other
values in the secular equation.

a) Calculate the energy level of the MO’s.

b) Calculate the (normalized) coefficients of both MO’s.

Exercise 5.3
This exercise concerns the CO2 molecule, which is a linear molecule with a structure O = C = O. TWe
choose the z direction along the inter-nuclear axis. The molecular orbitals ϕ are approximated as the linear
combination of the atomic orbitals χ. The 10 atomic orbitals that we include in the calculation are the
2sC , 2pC C C 1 1 1
x , 2py en 2pz on carbon (indicated by the superscript “C”), the 2s , 2px , 2py of the first oxygen
2 2 2
atom (indicated by the superscript “1”), and the 2s , 2px , 2py of the second oxygen atom (indicated by the
superscript “2”). We do not consider the deep 1s orbitals of the atoms.

a) Indicate for each of the atomic orbitals on C if they have overlap with the atomic orbitals on oxygen.

b) Indicate for each of the atomic orbitals on C if they have overlap with the symmetric and anti-symmetric
combination of the 2s orbitals of oxygen: 2s1 + 2s2 , 2s1 − 2s2 .

The table below shows the coefficients that determine the contribution of the atomic orbitals to the lowest
8 molecular orbitals. The order of the atomic orbitals is random, and also the order of the MO’s is random
(so not ordered by increasing energy).

61
2sC 2p2x χ3 χ4 χ5 χ6 χ7 χ8 χ9 χ10
ϕ1 .40 .00 .65 .00 .00 .00 .65 .00 .00 .00
ϕ2 .00 .00 .51 -.70 .00 .00 -.51 .00 .00 .00
ϕ3 .48 .00 -.62 .00 .00 .00 -.62 .00 .00 .00
ϕ4 .00 .56 .00 .00 .56 .00 .00 .61 .00 .00
ϕ5 .00 .00 .00 .00 .00 .61 .00 .00 .56 .56
ϕ6 .00 .00 .57 -.59 .00 .00 -.57 .00 .00 .00
ϕ7 .00 -.71 .00 .00 .71 .00 .00 .00 .00 .00
ϕ8 .00 .00 .00 .00 .00 .00 .00 .00 .71 -.71

c) Give, on the basis of the values for the coefficients, and with the help of your answers for a) and b),
which atomic orbitals correspond with χ3 to χ10 .
d) Which MO’s are of the σ type, and which are of the π type?
e) If the molecule is in the ground state, how many MO’s are filled: 6, 7 or 8? Explain your answer.

Exercise 5.4
We consider the water molecule. We choose the (x, y) coor-
O x
dinate system as indicated in the figure. The MO’s ϕ are ap-
proximated as a linear combination of AO’s χ. The 7 AO’s
that we bring into the calculation are 1s(H1 ) of the first hy-
drogen atom, the 1s(H2 ) of the second hydrogen atom, and H1 H2
(O) (O) (O)
the 1s(O) , 2s(O) , 2px , 2py and 2pz of oxygen.
y

a) Why is it not meaningful to also include the d orbitals of oxygen in the calculation?
Because of symmetry, we expect that the coefficients of 1s(H1 ) and 1s(H2 ) will be the same, or the same
but with opposite sign (see also H2 ). Anticipating this, it makes sense to consider (in our mind) the hybrid
orbitals η + = 1sH1 + 1sH2 and η − = 1sH1 − 1sH2
b) Sketch η + and η −
c) Argue (on the basis of symmetry) which AO’s of oxygen have no overlap with η + . Do the same for η − .
With the help of a computer program the coefficients have been calculated. In this we have neglected the
overlap integral, as we did in the previous exercises. See table below.
energy (eV) χ1 χ2 χ3 χ4 χ5 χ6 χ7
ϕ1 -557 0.00 0.00 0.00 0.00 0.00 0.00 1.00
ϕ2 -36 0.15 0.17 0.00 0.96 0.00 0.17 0.00
ϕ3 -19 0.00 0.53 0.00 0.00 -0.66 -0.53 0.00
ϕ4 -13 -0.81 0.27 0.00 -0.45 0.00 0.27 0.00
ϕ5 -12 0.00 0.00 1.00 0.00 0.00 0.00 0.00

d) The atomic orbitals χ are in random order (note: they are the original AO’s, so also include 1sH1 and
1sH2 ). Indicate which χ corresponds with which AO.
e) Calculate the charge on the oxygen atom, when it is in the ground state.

62
Chapter 6

Hückel Theory and Delocalized Bonds

In the previous chapter we have seen that due to symmetry, the π bonds can be treated independently from
the σ bonds. Furthermore, the π bonds are less strong than the σ bonds, so the electrons that occupy the
π MO’s are often central to understanding the reactivity of a molecule. In this chapter, we focus on the π
MO’s in hydrocarbons (although at the end we will also discuss examples with a C replaced by N). We will
find that a typical aspect of the π MO’s in the molecules that we discuss is that they are not localized between
two atoms, as for the σ MO’s, but are spread out over the whole molecule: the bonds are delocalized. We
start with the 1,3-butadiene molecule.

6.1 1,3-Butadiene
The atoms of the 1,3-butadiene molecule C4 H6 all lie in the same plane (which we define as the xy-plane).

We start by assuming that sp2 -hybrid orbital are


C2 C3
formed from the 2px , 2py and 2s on each carbon
atom, which are “used” for forming the σ skeleton
C1 C4
of the carbon atoms, the geometry of which can be
cis
seen on the right.

The hydrogen atoms are located such that there is a maximum overlap between their 1s orbital, and the
remaining sp2 -hybrid orbitals on the C atoms. The (single-occupied) pz -orbitals of the carbon atoms are
perpendicular to the xy plane (for convenience we denote the 2pz by just pz ), and do not mix with the σ
(2) (3)
bonds of the skeleton: they will form the π bonds in the molecule. The overlap between pz and pz , that is,
(2) (1)
the pz -orbitals of C2 and C3 , respectively, and the overlap between pz and pz , is expected to be of similar
magnitude. So there is no reason to assume that the configuration of 1,3-butadiene is C1 −C2 −C3 −C4 .

Figure 6.1: Overlap of pz orbitals in 1,3-butadiene.

63
To get to a better understanding of the π MO’s in this molecule, we resort to the MO-LCAO approach,
where the π orbitals are constructed as linear combinations of all the available pz orbitals:

π = c1 p(1) (2) (3) (4)


z + c2 pz + c3 pz + c4 pz

Since these AO’s do not mix with any of the orbitals that constitute the σ skeleton, the secular equations of
the π-system can be solved independently from the σ-system. The secular equations of the π system can be
written as:
     
H11 − S11 E H12 − S12 E H13 − S13 E H14 − ES14 c1 0
 H21 − S21 E H22 − S22 E H23 − S23 E H24 − ES24   c2   0 
     
 H31 − S31 E H32 − S32 E H33 − S33 E H34 − ES34  ·  c3  =  0 
H41 − S41 E H42 − S42 E H43 − S43 E H44 − ESN 4 c4 0

To makes some headway in solving these equations, we make a number of assumptions, as proposed by
Hückel:

H11 ≡ H22 ≡ H33 ≡ H44 ≡ α, α ≈ E2p


H12 ≡ H23 ≡ H34 ≡ H21 ≡ H32 ≡ H43 ≡ β, the resonance integral of two neighboring 2pz orbitals,
Sij = 0 when i ̸= j
Hij = 0 when the carbon atoms i and j are not nearest neighbors.

With these approximations, the secular equations for 1,3-butadiene take the following form:
     
α−E β 0 0 c1 0
 β α − E β 0   c2   0 
 ·   
 0 β α−E β   c3  =  0 
0 0 β α−E c4 0

This set of equations can be written in a more simple form by dividing left and right by β, giving
     
x 1 0 0 c1 0
 1 x 1 0   c2   
 ·  =  0 
 0 1 x 1   c3   0 
0 0 1 x c4 0

where we defined x = (α − E)/β. There is only a non-trivial solution to the secular equations when the
determinant of the matrix is zero:

x[x(x2 − 1) − 1(x)] − 1(x2 − 1) = 0


x2 (x2 − 1) − x2 − (x2 − 1) = 0
(x2 − 1)(x2 − 1) − x2 = 0
(x2 − 1)2 = x2

which gives two sets of solutions:


1 1√ 1 1√
x2 + x − 1 = 0 → x1 = − − 5 = −1.618 , x3 = − + 5 = 0.618
2 2 2 2
1 1√ 1 1√
x2 − x − 1 = 0 → x2 = − 5 = −0.618 , x4 = + 5 = 1.618
2 2 2 2

64
Inserting the value x = x1 = −1.618 into the secular equations gives the set
    
−1.618 1 0 0 c1 0
 1 −1.618 1 0   c2   
 ·  =  0 
 0 1 −1.618 1   c3   0 
0 0 1 −1.618 c4 0

which can be solved, giving

c1 = 0.37 , c2 = 0.60 , c3 = 0.60 , c4 = 0.37 (6.1)

Note that when solving the set of equations you basically aim to express three of the coefficients (say c2 , c3 ,
c4 ) in terms of one the coefficients (c1 ). The only unknown coefficient c1 then follows from the condition
that the coefficients are normalized: c21 + c22 + c23 + c24 = 1 (you can check this is the case for the set of (6.1).
Since x = (α − E)/β, or E = α − βx, the coefficients (6.1) belong to the energy level E1 = α + 1.618β.
Inserting the other values of x gives the coefficients of the other energy levels. The final result is:

c1 c2 c3 c4
x4 = 1.618 E4 = α − 1.618 β 0.37 -0.60 +0.60 -0.37
x3 = 0.618 E3 = α − 0.618 β 0.60 -0.37 -0.37 0.60
x2 = -0.618 E2 = α + 0.618 β 0.60 0.37 -0.37 -0.60
x1 = -1.618 E1 = α + 1.618 β 0.37 0.60 0.60 0.37
Since β represent the exchange integral, and thus has a negative value, we can conclude that the MO
corresponding to the solution x1 = −1.618 has the lowest energy, and x2 is the next lowest level, etc. The
energy scheme is shown in figure 6.2, under “delocalized”. In this we have used that α represents the coulomb
integral of the 2pz orbital, so its value is about equal to the energy of a 2p AO in carbon. Each of the four
contributing pz -orbitals “contains” one electron, that is, the MO’s should accommodate four electrons. This
means that in the ground state, the MO’s π1 and π2 are fully filled. Note that all 2pz ’s contribute to π1 and
π2 : we cannot localize the MO in the molecule. All the coefficients have the same sign in π1 , hence there
are no nodal planes: ϕ1 is a bonding MO with respect to all C atoms. For the π2 -orbital the coefficients
c2 and c3 have a different sign: this MO has a nodal plane between C2 and C3 , meaning that the π-orbital
is anti-bonding with respect to the atoms C2 and C3 . This partly cancels the bonding effect of π1 . With
respect to C1 and C2 , both π1 and π2 are bonding MO’s . We thus may expect a different bond length;
experimentally it is found that C1 −C2 : 1.34 Å, C2 −C3 : 1.47 Å.

While we have neglected this in our calculations, there is still a slight interaction between non-nearest neigh-
bors. For this reason, there is some difference in energy between the cis and the trans isomer of 1,3-butadiene
(fig. 6.3); the trans isomer has the lowest energy.

The highest occupied σ orbital has an energy of about -11 eV, roughly equal to that of the lowest π-orbital.
The π2 orbital has an energy of -9 eV. Since chemical reactions involve electrons that are placed in the
highest occupied MO’s, the electrons that are in the π orbitals play a central role in the reactivity of the
molecule, and the Hückel method as described above can give a quick and simple insight into this.
The structural formula of 1,3-butadiene is often given as H2 C−CH−CH−CH2 , which indicates that
the π bonds are localized between C1 −C2 and C3 −C4 , and that there is no overlap between the pz ’s on C2
en C3 . The secular equations corresponding to this picture are
    
α−E β 0 0 c1 0
 β α−E 0 0   c2   
 ·  =  0 
 0 0 α−E β   c3   0 
0 0 β α−E c4 0

65
separate C4 H 6 C4 H 6 energy
C atoms delocalized localized

π4 α − 1.62β
π3 π4 α−β
π3 α − 0.62β
2pz 6 6 6 6 α ≈ E2p
π2 6 α + 0.62β
? π1 6 π2 6 α+β
? ?
π1 6 α + 1.62β
?

Figure 6.2: Energy levels for butadiene of the delocalized and the localized π-MO’s, relative to the energy
levels of the 2pz AO’s. Note that β has a negative value.

C4

C2 C3 C2 C3

C1 C4 C1
cis trans

Figure 6.3: Geometry of the van het cis and trans isomer of 1,3-butadiene.

or      
x 1 0 0 c1 0
 1 x 0 0   c2   
 ·  =  0 
 0 0 x 1   c3   0 
0 0 1 x c4 0
Setting the determinant equal to zero gives [x2 − 1]2 = 0, so the solutions are x1 = x2 = −1, x3 = x4 = 1.
The corresponding energies are

E1 = α + β , E2 = α + β , E3 = α − β , E4 = α − β

In figure 6.2 we compare these energy levels, so of the localized MO’s, with the energy levels of the delocalized
MO’s that we derived at the beginning of this section. It shows that the π electrons have a lower energy in
(2) (3)
the delocalized π system, where we allow for overlap between pz and pz . The difference in energy of the
electrons in the two systems is

4(α + β) − [2(α + 1.62β) + 2(α + 0.62β)] = −0.48β

which is known as the delocalization-energy.

66
In the localized state there is only a σ bond between C2 −C3 ,
so the two C2 H3 groups can rotate freely with respect to H H
each other. For the delocalized state, there is a rotation
barrier of 24 kJ.mol−1 . The structure formula for 1,3- H2C C C CH2
butadiene that reflects the delocalized nature of the π bonds
is given on the right.

6.2 Benzene
While the chemical formula for benzine (C6 H6 ) was long known, the
structure has puzzled chemist for quite some time: how to combine six
carbon atoms (with a valence of 4) with 6 hydrogen atoms? In 1861
Kekulé managed to find a solution with the famous model of a cyclic
molecule (see left), with only one hydrogen atom per carbon atom, and
alternating single and double bonds (not shown in the graph). Since
the molecule turned out to be much less reactive than you would expect
from the presence of the double bonds, it was assumed that the double
bonds would switch rapidly within the molecule, so from {C1 −C2 ,
C3 −C4 , C5 −C6 } ↔ {C2 −C3 , C4 −C5 , C6 −C1 }. This has not been
confirmed by any experiment.

We are now in a position to make a model that can explain the properties of benzene well. We again
assume that all C and H atoms are located in one plane (which is confirmed by experiments). On each C
atom sp2 hybrid orbitals are formed, which provides the two σ bonds with its nearest neighbors, and one σ
bond with an H atom: the σ skeleton. Perpendicular to this plane are the pz -orbitals, which are overlapping
and thereby combining to π-orbitals. We again use the variational approach with the Hückel approximations,
to get to the secular equations for the π system:
     
α−E β 0 0 0 β c1 0
 β α−E β 0 0 0   c2   0 
     
 0 β α−E β 0 0     
  ·  c3  =  0 
 0 0 β α−E β 0     
   c4   0 
 0 0 0 β α−E β   c5   0 
β 0 0 0 β α−E c6 0

We again introduce x = (α − E)/β, by use of which the equations can be written as:
     
x 1 0 0 0 1 c1 0
 1 x 1 0 0 0   c2   0  x = ±2
     
 0 1 x 1 0 0   c3   0 
      → ±1
 0 0 1 x 1 0  ·  c4  =  0  Det[] = 0: 6 solutions: x =
     
 0 0 0 1 x 1   c5   0 
x = ±1
1 0 0 0 1 x c6 0

This gives the following energy scheme (in which we have already placed the 6 π-electrons):

67
α − 2β

α−β

α ≈ E2p

6 6 α+β
? ?
6 α + 2β
?

Figure 6.4: Energy levels of the π MO’s for benzene.

As before, a simple tool to understand the nature of


the π-orbitals is by looking at the nodal planes (see
figure on the right). For the π1 -orbital all coefficients
have the same sign and and size. There is no nodal
plane (except for the xy plane, which is by definition
a nodal plane since it concerns a π orbital). For the
π2 - and π3 -orbitals there is one nodal plane, they
have the same energy. For the π4 and π5 orbitals
there are two nodal planes. For the π6 orbital, all
C-C connection are “cut” by nodal planes: this MO
has a strong anti-bonding character.

It is useful to introduce a terminology for the MO’s that play a role in chemical reactions. The highest
occupied molecular orbital is refered to as the HOMO, while lowest unccupied molecular orbital is refered
to as the LUMO. In case of benzene, π2 en π3 are HOMO’s, while π4 en π5 are LUMO’s.

6.3 Pyridine and Pyrrole

The Hückel approach can also be used to study bonding in hetero-cyclic


compounds, that is, a cyclic compound where also atoms other than
carbon are present in the ring. As an example we consider the pyridine
molecule, the geometry of which is shown on the right. Like in benzene,
the C atoms form sp2 -hybrid orbitals which - together with the 1s orbitals
of the H atoms - combine to form σ skeleton.

The ground state configuration of the N atom is (1s)2 (2s)2 px , py , pz . In order to be able to fully participate
in the bonding, also this atom forms sp2 -hybrid orbitals, so that the configuration becomes:

(1s)2 , spI2 , spII III 2


2 , (sp2 ) , pz

68
The two single-occupied sp2 -orbitals combine with the sp2 orbitals on the
neighboring carbon atoms to form σ-bonds. The double-occupied sp2 -
orbital is directed outward, and does not participate in the bonding. The
two electrons in this sp2 orbital are called a “lone pair”. Their energy
is higher than that of the electrons in the σ bonds, and comparable to
that of the π-electrons,

The remaining, single-occupied pz -orbital of the N-atom will participate in the π-MO’s:

πi = c1 p(N)
z + c2 p(2) (3) (4) (5) (6)
z + c3 pz + c4 pz + c5 pz + c6 pz

The secular equations are very similar to those of benzene:


     
αN − E β 0 0 0 β c1 0
 β α − E β 0 0 0   c2   0 
     
 0 β α−E β 0 0   c3   0 
    =  
 0 0 β α − E β 0 · c4   0  (6.2)
     
 0 0 0 β α−E β   c5   0 
β 0 0 0 β α−E c6 0
(N) (N) (C ) (C )
Note that the integral αN = ⟨pz |Ĥ|pz ⟩ is not equal to α = ⟨pz i |Ĥ|pz i ⟩. Since αN ≈ E2pN (the
energy level of a 2p orbital of N), and α ≈ E2pC (the energy level of a 2p orbital on C), we expect that
(C ) (N) (C ) (C )
αN < α. In principle the resonance integral ⟨pz i |Ĥ|pz ⟩ will also be different for ⟨pz i |Ĥ|pz j ⟩ (with i, j
next neighbors, and i the next neighbor of N). Nevertheless, in the secular equations (6.2) we have set both
integrals equal to the same value β. Dividing left and right of (6.2) by β gives:

     
x + hN 1 0 0 0 1 c1 0
 1 x 1 0 0 0   c2   0 
     
 0 1 x 1 0 0   c3   0 
 ·  =  
 0 0 1 x 1 0   c4   0 
     
 0 0 0 1 x 1   c5   0 
1 0 0 0 1 x c6 0
where we defined x and hN by
α−E αN − α
x= and hN =
β β
The solutions to the secular equations are:

xi C1 C2 C3 C4 C5 C6
π1 -2.107 .52 .42 .36 .34 .36 .42
π2 -1.167 .57 .19 -.35 -.60 -.35 .19
π3 -1.000 0 .5 .5 0 -.5 -.5
π4 0.841 .55 -.37 -.24 .57 -.24 -.37
π5 1.000 0 -.50 +.50 0 -.50 +.50
π6 1.934 .32 -.39 .44 -.45 .44 -.39

The molecule has a mirror plane perpendicular to the xy plane (the plane of the molecule itself). We
note that the orbitals π1 , π2 , π4 and π6 are symmetrical with respect to this plane. The orbitals π3 and π5
switch sign with respect to this plane: so it represents nodal plane. In these two MO’s there is no partici-
(N)
pation from pz ; for this reason π3 and π5 are indistinguishable from the corresponding MO’s in benzene.

69
(N) (C)
The coefficient of pz in orbital π1 is larger than those of the pz -orbitals: there is a higher probability to
find the electron at the N atom than at a C atom. The overall distribution of the six π electrons over the
atoms follows by summing the coefficients squared of∑ all the occupied MO’s (see section 5.3). That is, the
(π)
amount of π electrons on atom j is given by qj = 2 3(cij )2 , where we used that π1 , π2 , and π3 are each
i=1
occupied by 2 electrons. This results in the following distribution of the 6 electrons:

atom: N C2 C3 C4 C5 C6 total
charge: 1.21 0.92 1.00 .95 1.00 .92 6.00
We observed that there is a shift of electrons towards the N atom.
We have not yet discussed the presence of the lone pair of electrons on the
N atom. This has the effect that there is a cumulation of negative charge on
the “outside” of the N atom, which can attract positive ions. In a solution
where there are also H+ -ionen available, a cation is formed: pyridinium, the
conjugate acid of pyridine. So the presence of the lone pair causes pyridine to
act as base. The N-H bond in pyridinium is indistinguishable from a “regular”
covalent bond where each of the atoms contributed one electron. This type of
bond, where one of the partners brings in two electrons (instead of each one)
is called a dative bond.

Pyrrole - At first sight pyrrole (see figure 6.5) is very similar to pyridine, however, there are some es-
sential differences. The carbon atoms, as well as the nitrogen atoms, have sp2 hybrid orbitals. Overlap
between these orbitals leads to a flat five-membered ring. The angles of such a ring (108o ) are different from
the 120o angles that the hybrid orbitals have on each atom. For this reason, the ring has strain, which raises
the energy. To reach maximum overlap, the H-atoms are located in the plane of the ring. The pz orbitals,
perpendicular to this plane, combine to form π-orbitals, where each C atom contributes one electron. The N
atom now has two “choices”: i) put one electron in the π orbitals, and like for pyridine, two electrons in an
sp2 hybrid orbital (which then becomes a lone pair), or ii) two electron in the π orbitals, and one electron
in the sp2 hybrid orbital, which then, being single occupied, can form a bond with an hydrogen atom. The
latter option is energetically favorable. In that case there are 6 electron in the π system. When they are
more or less equally distributed over the 5 atoms, this means that the N atom “gets” 6/5 = 1.2 electron
out of the π system, while it “donated” 2 electrons to the system: a loss of 0.8 electrons. This is partly
compensated by the fact that there is a slight shift of the σ-electrons towards the N atom.

π5
π4
π3 2p(C)
π2
2p(N )

π1

M.O.’s A.O.’s

Figure 6.5: Pyrrole: geometry of the σ-skeleton (left) and the energy of the orbitals (left).

70
Exercises
Exercise 6.1
Calculate the delocalization energy of benzene in terms of β.

Exercise 6.2
In this exercise we consider 3-methylene-1,4-pentadiene, which is a “flat” molecule:

For the π-system the following coefficients and energies have been found, where we have made use of the
Hückel approximations:
(1) (2) (3) (4) (5) (6)
xi pz pz pz pz pz pz
π1 -1.932 0.230 0.444 0.628 0.325 0.444 0.230
π2 -1.000 ? ? ? ? ? ?
π3 -0.518 0.444 0.230 -0.325 -0.628 0.230 0.444
π4 0.518 -0.444 0.230 0.325 -0.628 0.230 -0.444
π5 1.000 0.500 -0.500 0.000 0.000 0.500 -0.500
π6 1.932 0.230 -0.444 0.628 -0.325 -0.444 0.230

a) Give the secular equations of the π-system in matrix form, making use of the Hückel approximations.
Use the numbering as indicated in the graph.

b) Draw the nodal planes in the molecule for π3 , π4 , π5 and π6 . Where do you expect the nodal plane for
π2 ?

c) Calculate the coefficients for π2 , and check if the nodal plane agrees with your answer at b).

Exercise 6.3
We consider the allyl radical H2 C−CH−CH2 . The molecule is “flat”, and the angles of the sigma bonds are
120o .

a) Determine the energy levels and the normalized coefficients of the MO’s of the π-system.

b) Give the distribution of the π electrons over the atoms, both for the ground state and the first excited
state.

Exercise 6.4
By using the Hückel method, the following coefficients and energies have been found for the MO’s that
describe the π system of the Methylenecyclopropane molecule (structure: see above). In this, piz is the 2pz
orbital on the carbon atom with number i).

71
H
@
@
C4 H
@
@
C2 C1
@
@
C3 H

E1 = α + 2.17β ϕ1 = 0.28 p1z + 0.61 p2z + 0.52 p3z + 0.52 p4z


E2 = α + 0.31β ϕ2 = −0.81 p1z − 0.25 p2z + 0.36 p3z + 0.36 p4z
E3 = α − 1.00β ϕ3 = ...
E4 = α − 1.48β ϕ4 = 0.50 p1z − 0.74 p2z + 0.30 p3z + 0.30 p4z

a) Give the secular equations of the π-system in matrix form, making use of the Hückel approximations.
Use the numbering as indicated in the graph.
b) Calculate the normalized coefficients of ϕ3 .
c) Show that ϕ1 and ϕ4 are orthogonal.
d) Calculate the delocalization energy.
e) How much energy is needed to bring the atom into the first excited state?
f) Calculate the charge on atom 1 for the first excited state. For this, you may assume that the “σ
electrons” are equally distributed over the atoms, so that only the π electrons may give rise to a shift
of charge.

Exercise 6.5
This question concerns the butalene molecule C6 H4 , which is composed of two fused cyclobutadiene rings.
The structure of the σ-skeleton is given below, which also indicates the numbering of the C atoms that
should be used (keep to this numbering!):

We want determine the energy levels and the coefficients of the π-electron system, by making use of the
Hückel approximations.
a) Give the secular equations of the π-system in the form of a matrix, in terms of x = (α − E)/β, where
α, β are the Coulomb- and resonance integral, respectively, while E is the energy of a π molecular
orbital. The rings are under strain. Why?

72
Setting the determinant of the√secular matrix equal to √
zero gives the following 6 values for x (you do not
need to show this): x = ±(1 + 2), x = ±1, x = ±(1 − 2). The corresponding normalized coefficients are
given in the table below:
x c1 c2 c3 c4 c5 c6
ϕ1 -2.4142 0.3536 0.3536 0.5000 0.5000 0.3536 0.3536
ϕ2 - 1.0000 -0.5000 -0.5000 0.0000 0.0000 0.5000 0.5000
ϕ3 - 0.4142 -0.3536 0.3536 -0.5000 0.5000 -0.3536 0.3536
ϕ4 + 0.4142 0.3536 0.3536 -0.5000 -0.5000 0.3536 0.3536
ϕ5 + 1.0000
ϕ6 + 2.4142 0.3536 -0.3536 -0.5000 0.5000 0.3536 -0.3536

b) Calculate the normalized coefficients of the ϕ5 molecular orbital, so the coefficients for the solution
x = 1. To facilitate the calculation, it is given that c3 = 0.
c) Draw the nodal planes of ϕ1 up to ϕ6 in the molecule.

d) Calculate the delocalization energy of the molecule.


e) The molecule is in the ground state. Calculate the charge on the first carbon atom, on the basis of the
distribution of the π electrons.

73
74
Appendix A: Classical Mechanics

In this appendix we present some basic principles and concepts of classical mechanics that are also used in
quantum mechanics, such as kinetic & potential energy, momentum, force and Hamiltonian. We will do this
for a single particle, and start in 1-D. The position, velocity and acceleration are given by

dx(t) dvx (t) d2 x(t)


x(t) , vx (t) = , ax (t) = =
dt dt dt2
The momentum of a particle is equal to px = mvx (t), with m the inertial mass. If we apply a force Fx on the
particle, the momentum will change according to Fx = dpx /dt, which is the second law of Newton. Inserting
px = mvx (t), with m independent of time, we get the famous “F = ma” equation:

dvx d2 x
Fx = m or Fx = m = max (6.3)
dt dt2
Two well-known examples of a force are:

a) Gravity force: Fx = −m′ g , with g = 9.81m/s2 the gravitational constant, and m′ the heavy mass.
Experimentally it follows that m′ = m, so we will use only one symbol m. Newton’s second law then
gives

d2 x d2 x dx 1
−mg = m → = −g → = −gt + c1 → x = − gt2 + c1 t + c2
dt2 dt2 dt 2
where c1 and c2 are integration constants, which are not known. Hence we need additional information,
which makes sense: it is clear that we need to know the initial position xo and initial velocity vo to
predict the position some time later. The third equation (vx (t) = −gt + c1 ) gives that vx (0) = c1 , so
c1 = vo . In the same way it follows that x(0) = c2 so that c2 = xo . The final equation for the position
as function of time is then
1
x(t) = − gt2 + vo t + xo
2
b) Force from a spring: Fx = −kx , where k is the spring constant, which indicates the spring’s stiffness.
In this, we have chosen that at the origin (x = 0) the spring is in its equilibrium position, that is, it
applies no force on the particle. When the particle moves away from the origin, the spring applies a
force that scales linearly with the distance from the origin, and directed towards the origin. For this
force, Newton’s second law gives that

d2 x d2 x
−kx = m → = −ω 2 x → x(t) = c1 cos(ωt + c2 )
dt2 dt2

where we introduced the parameter ω = k/m. The last step can be checked by differentiating twice.
The constants c1 en c2 can be determined again from the initial values for position xo and velocity vo ,
which we do not work out any √ further. The solution x(t) shows that the particle will be oscillating
with an angular frequency ω = k/m.

75
In case a force is constant, or only depends on position, as in the examples given above, we can define a
function V such that
dV
Fx = − (6.4)
dx
The function V (x) is known as the potential energy. It can be easily checked that for the gravity force
V (x) = mgx + C, and for the spring force V (x) = 21 kx2 + C, with C some constant. So from the force
we cannot derive an absolute value for V (x), only how it changes with x. This is fine, since it are only
differences in the potential energy which are relevant. This means that you can choose an arbitrary point
where you define the value to be zero, which sets the constant. For instance, when you use V (x) = mgx for
gravity, and V (x) = 12 kx2 for the spring, you have implicitly defined that in both cases V = 0 for x = 0.
We now insert (6.4) into Newton’s equation (6.3), multiply both sides by vx and use that d(vx2 )/dt =
2vx dvx /dt:
[ ]
dV dvx dV dvx dx dV d 1
− =m → −vx · = mvx · → − · = mv 2
dx dt dx dt dt dx dt 2 x
[ ]
d 1
→ 2
mv + V (x) = 0 (6.5)
dt 2 x
The term in brackets (called the Hamiltonian) is thus independent of time. We refer to 21 mvx2 as the kinetic
energy. We have thus proven that for forces which are constant or only depend on x, the sum of the kinetic
and potential energy is constant in time. For this reason these type of forces are called conservative forces.
Note that for instance a friction force (like the friction with the air for a falling object) depends on the
velocity of the particle, and so that is not a conservative force. Indeed, this force will give that the particle’s
velocity when reaching the bottom is lower than without this force: kinetic energy is lost (of course, it is not
really lost: it is converted into heat).

We now make the extension to 3-D. The position is now not only described by x(t), but also by two
other coordinates: y(t), z(t); likewise we have now also velocities and forces in the other directions: vy (t),
vz (t), and Fy , Fz . For each direction Newton’s law holds:

dvx dvx dvz d⃗v


Fx = m , Fx = m , Fz = m in vector form: F⃗ = m
dt dt dt dt
For conservative forces we can define an expression similar to (6.4):

dV dV dV
Fx = − , Fy = − , Fz = − in vector form: F⃗ = −∇V
⃗ (6.6)
dx dy dz

where V (x, y, z) is now a function of all three coordinates, In the same way as for one dimension, you can
show that the 3-D Hamiltonian is constant in time:
[ ]
d 1 2 2 2
m(vx + vy + vz ) + V (x, y, z) = 0
dt 2

At first sight, there are very many different forces in nature: gravity, magnetic force, spring force, fric-
tion force. However, many of these forces are effective forces, by which we mean that they are build up out
of more fundamental forces. For instance, the friction force that a particle feels when falling through air,
follows from the collisions with the air molecules - so it is an effective collision force. However, the collision
between an object and an air molecule can be analyzed in terms of interactions (forces) between electrons
and protons. Eventually, all forces in nature can be traced back to two elementary forces1 :
1 except for the forces that are present within a nucleus

76
1. The gravity force. Any two bodies attract each other. If one body has mass m and the other mass M ,
then the potential energy of this force is equal to
1 √
V = −G M m with r= x2 + y 2 + z 2 (6.7)
r
where we defined the origin on one of the particles, and G a fundamental constant of nature (the
gravity constant). A few remarks:
– We have defined the potential such, that it is zero for r → ∞.
– By applying F⃗ = −∇V ⃗ you can show that close to the earth’s surface, this force is F = −mg,
pointed to the earth’s center, with g = M G/R2 where M and R are the earth’s mass and radius
respectively. The first accurate estimate of G was obtained by Cavendish in 1798 by measuring
the force between two large lead balls in close proximity; he called this experiment “the weighing
of the earth” (can you think why?).
– The potential (6.7) can give rise to a stable motion where one mass orbits around the other mass,
like for instance the moon around the earth. The force needed to keep the particle in the orbit
(the centrifugal force) is then exactly equal to the attractive force.
2. The electrostatic force. Two bodies with opposite charge attract each other. If one body has charge
−q and the other charge Q (where q and Q are positive numbers), then the potential energy of this
force is equal to
1 1 √
V =− Qq with r = x2 + y 2 + z 2 (6.8)
4πϵo r
where we defined the origin on one of the particles, and ϵo is a fundamental constant of nature (the
dielectric constant).
It is directly obvious that both forces are very similar, which of course rises the question if there is a common,
deeper origin. Up to now this has not been proved.

If equation (6.7) would give a stable orbit, then this should also also hold for (6.8), in case of an elec-
tron and a proton. The image of the hydrogen atom where the electron is orbiting around the nucleus
seems therefore logical, and indeed this is the picture that you will mainly find when searching for “hydro-
gen atom”. However, as also discussed in chapter 1, equation (6.8) would in principle allow for any orbit.
The emission spectrum2 of hydrogen is discrete, which indicates that there are only a discrete set of orbits
possible. So clearly something is missing when applying the classical equations to the system of a proton
and an electron. Initially some “ad-hoc” rules were added by Bohr, such as that the angular moment could
only be multiples of h̄ so, h̄, 2h̄, 3h̄, .... While this rule would explain the spectrum, the rationale behind it
was lacking; moreover, it could only predict the spectrum of hydrogen, and not of for instance helium. Also,
it predicts that the angular momentum of the ground state (n = 1) is h̄, while experimentally it is found to
be zero. The conclusion must be that classical mechanics is simply not valid at the level of atoms.

2 The emission spectrum gives the wavelength (and thus the energy) of the radiation that an electron emits when “falling”

to a lower energy orbit.

77
78
Appendix B: Functions and Operators

Function and operators play a central role in quantum mechanics. In this appendix we introduce some
elementary concepts, which are essential for understanding the theory. We will do this for functions that
depend only one one variable (x) (which in QM means one dimensional systems).

Functions
You will all be familiar with the concept of a function: the function f (x) is process or relation which couples
a value f to a value x (for a formal definition, see math textbooks). New will be the concept of a “function
space”, formally called the Hilbert space. In this, an analogy is made between vectors and functions, which
is very useful since it allows for using the notation and language of vector algebra. Before doing so, it
is useful to first briefly discuss the classic 3-D vector space.3 To describe the 3-D space we make use of
three orthonormal vectors: ⃗ex , ⃗ey , ⃗ez , where ⃗eα · ⃗eβ = 0 if α ̸= β (“ortho”), and ⃗eα · ⃗eβ = 1 if α = β
(“normal”). These three vectors form a complete set, meaning that any arbitrary vector ⃗a be written as a
linear combination of these three vectors:

⃗a = cx⃗ex + cy ⃗ex + cz ⃗ez

We now make the analogy with functions: we speak of a complete set of functions f1 , f2 , f3 , .. if any arbitrary
function h(x) can be written as a linear combination of the functions f :

h(x) = c1 f1 (x) + c2 f2 (x) + c3 f3 (x) + c4 f4 (x) + ...

A typical example of a complete set of functions are the polynomials f1 = 1, f2 = x, f3 = x2 , ...: any function
can be written as a polynomial series (known as the Taylor series), for instance cos x = 1 − x2 /2 + x4 /24 + ...,
1 2 1 3
or ex = 1 + x + 2! x + 3! x + ..... We can thus think of the functions f1 , f2 , ... as the “unit vectors” of the
function “space”. Of course it is not a real space, it is just an analogy (for instance, there are an infinite
amount of functions f1 , f2 , .. needed, not just 3 as in the real space).
Why is this useful? In quantum mechanics, functions play a key role, and it is very convenient to describe the
manipulation of these functions in terms of vector algebra, for instance projecting a function, normalizing a
function, taking the inner product of two functions, showing that two functions are perpendicular, etc. We
start with the example of the inner product and being perpendicular. If two vectors ⃗a and ⃗b are perpendicular
(= orthogonal), it means that the inner product ⃗a · ⃗b = 0. We now switch to function space. For the inner
product of two (complex) function f (x) and g(x) we use the notation ⟨f |g⟩, and define it as
∫ ∞
⟨f |g⟩ = f ∗ (x)g(x)dx
−∞


where stands for the complex conjugated. Note that the inner product is different from just the product
of two functions: the latter would yield a new function. The inner product is the product, followed by
integrating out over space, so that the result is a number (just like for vectors: the result of the inner
3 Note that the “x” in the argument of the function has nothing to do with the x direction of the vector analogy

79
product is not a new vector, but is also a number). We say that two functions f and g are perpendicular
(orthogonal) when the inner product is zero, just like for vectors: √
Another useful concept from the vector space is the norm: the norm of a vector ⃗a defined as |a| = ⃗a · ⃗a,
which is the length of the vector. In function space, the norm of a function f is defined in the same way:
√∫
√ ∞
|f | = ⟨f |f ⟩ = f ∗ (x)f (x)dx
−∞

∫In∞quantum

mechanics we mainly work with functions that are normalized, meaning that the norm is 1:
−∞
f (x)f (x)dx = 1.
Note that the unit vectors for the vector space are orthonormal. We can now imply the same condition to
the set of functions f1 , f2 , f3 , ...:

⟨fn |fm ⟩ = 0 if n ̸= m , ⟨fn |fm ⟩ = 1 if n = m

In this case we speak of a complete, orthonormal set of functions. Check for yourself that while the polyno-
mials are a complete set, they are certainly not an orthonormal set.

Operators
An operator Λ̂ is a mathematical manipulation of a function, say f (x), which changes it to a function g(x):
Λ̂f = g. In quantum mechanics we only have linear operators. By definition, an operator Λ̂ is linear when

Λ̂(cf ) = cΛ̂f (6.9)

(where c is a constant) and


Λ̂(f1 + f2 ) = Λ̂f1 + Λ̂f2 (6.10)

For the sum and product of two linear operators Λ̂1 and Λ̂2 the following rules hold:

(Λ̂1 + Λ̂2 )f = Λ̂1 f + Λ̂2 f (6.11)


(Λ̂1 Λ̂2 )f = Λ̂1 (Λ̂2 f ) (6.12)
(cΛ̂)f = c(Λ̂f ) (6.13)


Examples of linear operators are ∂x and the operator h(x) which multiplies a function f by h(x). In quantum
mechanics, the eigenfunctions of an operator are of special importance. When the following relations holds:

Λ̂f = λf (6.14)

then f is said to be an eigenfunction of Λ̂, and λ is called the eigenvalue. The eigenfunction is thus that
function which is unchanged under the action of the operator. In general there is an infinite set of eigen-
functions for each operator Λ̂. In case two or more different eigenfunctions have same eigenvalue, these
functions a are called degenerate. Note that functions which only differ by a constant are not considered
as “different”.

80
Hermitian Operators
The operators that we encounter in quantum mechanics are always hermitian. We skip the exact definition of
when a operator is hermitian, but implications of it are very relevant. These mainly concern the eigenvalues
and eigenfunctions. Suppose we have an operator  which has a set of eigenfunction ϕn , and corresponding
eigenvalues an , so Âϕn = an ϕn . If the operator  is hermitian, then
i the eigenvalues are real.
ii the eigenfunctions are an orthogonal set, meaning that


⟨ϕn |ϕm ⟩ = dx ϕn (x) ϕm (x) = 0 when n ̸= m


⟨ϕn |ϕm ⟩ = dx ϕn (x) ϕm (x) ̸= 0 when n=m

iii all eigenfunctions together make a complete set. This means that any arbitrary function χ(x) can be
written as a linear combination of the eigenfunctions ϕn :

χ(x) = cn ϕn (x) (6.15)

The eigenfunctions of a hermitian operator thus form a complete, orthogonal basis for the function space.

81
82
Appendix C: Atomic Units

Atomic units are chosen such that the various constants of nature which play a role in quantum mechanics
take the value 1:

h̄ = 1
me = 1
e = 1
4πϵ0 = 1
c = 137.035
From this follows that:
mp = 1836.154
mn = 1838.679

Since in SI units the constants have the values:

h = 6.62618 10−34 J · s
me = 9.10953 10−31 kg
e = 1.60219 10−19 C
c = 2.99793 108 m/s
Na = 6.02205 1023

we get the following conversion from a.u. to SI units for the various quantities:

1 atomic units of: is in SI units other name


Mass = me = 9.10953 10−31 kg
Length = a0 = 5.29177 10−11 m Bohr
Time = me a20 h̄−1 = 2.41887 10−17 s
−2
Energy = h̄2 m−1
e a0 = 4.35984 10−18 J Hartree
Charge = e = 1.60219 10−19 C

For energy there is a variety of units in literature. Below is a “conversion” table:

83
1 a.u. of energy = 4.35984 10−18 J
= 27.21498 eV
= 2625.72 kJ/mol
= 627.503 kcal/mol
= 4.35984 10−11 erg
= 3.15803 105 K (×kB )
= 2.19502 105 cm−1 (×2πh̄)

84
Appendix D: Orbitals of the
Hydrogen atom

State functions (or orbitals) of the Hydrogen atom in atomic units. Cartesian coordinates (x, y, z) and spher-
ical coordinates (r, θ, ϕ) are mixed in these equations, where x = r sin θ cos ϕ , y = r sin θ sin ϕ , z = r cos θ.
The constant C3 is such that the functions are normalized.

n, l, m En

3, 0, 0 3s = ψ300 = √1 C3
6
(27 − 18r + 2r2 ) e−r/3 -1/18

3, 1, 0 3pz = ψ310 = C3 (6 − r) z e−r/3 -1/18

3px = ψ311 − ψ31−1 = C3 (6 − r) x e−r/3 -1/18


3, 1, ±1
3py = ψ311 + ψ31−1 = C3 (6 − r) y e−r/3 -1/18

3,2,0 3dz2 = ψ320 = √1 C3


12
(3z 2 − r2 ) e−r/3 -1/18

3dxz = ψ321 − ψ32−1 = C3 xz e−r/3 -1/18


3, 2, ±1
3dyz = ψ321 + ψ32−1 = C3 yz e−r/3 -1/18

3dxy = ψ322 − ψ32−2 = C3 xy e−r/3 -1/18


3, 2, ±2
3dx2 −y2 = ψ322 + ψ32−2 = 12 C3 (x2 − y 2 ) e−r/3 -1/18


6 −1/2
Normalization constant: C3 = 81 (3π)

85
86
Appendix E: Anti-symmetry (not
required)

In this appendix we show the relationship between anti-symmetry, the Pauli exclusion principle, and Hund’s
rule. We will do this for the most simple case of two particles in one dimension. We start with the Schrödinger
equation for such a system:

h̄2 d2 h̄2 d2
Ĥψ(x1 , x2 ) = Eψ(x1 , x2 ) met Ĥ = − − + V (x1 , x2 ) (6.16)
2m dx21 2m dx22
We now assume that we can write V (x1 , x2 ) as the sum of two independent terms:

V (x1 , x2 ) = V (x1 ) + V (x2 )

so both particles feel the same potential V . Then the total Hamiltonian can also be written as the sum of
two terms that only depend on x1 en x2 :

−h̄2 d2 −h̄2 d2
Ĥ = 2 + V (x1 ) + + V (x2 ) = Ĥ1 + Ĥ2
2m dx1 2m dx22
| {z } | {z }
Ĥ1 Ĥ2

The operators Ĥ1 and Ĥ2 are the single-particle Hamiltonians of particle 1 and 2 respectively. It will be clear
that we can now use the method of separation of variables: since there are no mixed terms, the eigenfunction
of Ĥ is equal to the product of the eigenfunctions of Ĥ1 and Ĥ2 :

Ĥ1 ψna (x1 ) = Ena ψna (x1 )


ψnI a ,nb (x1 , x2 ) = ψna (x1 )ψnb (x2 ) with (6.17)
Ĥ2 ψnb (x2 ) = Enb ψnb (x2 )

and with eigenvalue Ena + Enb . It is important to realize that Ĥ1 and Ĥ2 are identical Hamiltonians, and
their eigenfunctions ψna and ψnb are the same set. We use different labels na en nb since particle 1 and 2
do not have to be in the same state; it is only that their possible states are of the same set. So particle one
could be in the ground state (na = 1), particle two in an excited state (nb = 3), which would give the state
I
ψ1,3 (x1 , x2 ). The reason that we also included a label “I” is that this function is not the only eigenfunction
of Ĥ with eigenvalue Ena + Enb . The function in which particle 1 is in a state with quantum number nb ,
and particle 2 in a state with quantum number na :

ψnIIa ,nb (x1 , x2 ) = ψnb (x1 )ψna (x2 ) (6.18)

is also an eigenfunction with eigenvalue Ena + Enb . Yet, when ψ I and ψ II are eigenfunctions of Ĥ with the
same eigenvalue Ena + Enb , then any linear combination of ψ I and ψ II is also an eigenfunction with that
eigenvalue (see exercise 1.1). Below we suggest two of them (why these will become clear later):
1
ψnIII
a ,nb
(x1 , x2 ) = √ [ψna (x1 )ψnb (x2 ) + ψnb (x1 )ψna (x2 )] (6.19)
2

87
1
ψnIVa ,nb (x1 , x2 ) = √ [ψna (x1 )ψnb (x2 ) − ψnb (x1 )ψna (x2 )] (6.20)
2

The factor 1/ 2 is required to keep the function normalized. So we now have found 4 eigenfunctions. Is each
of these possible as a state of the two particles, and what are the differences? The answer to this question
is given by the 4th postulate of quantum mechanics:

Postulate 4a:
Particles are indistinguishable: the probability density may not change when the labels of two particles are
changed.

This means that in general only ψ III and ψ IV are possible options as state functions4 : it is easy to show
that
ψnIII
a ,nb
(x1 , x2 ) = ψnIII
a ,nb
(x2 , x1 ) and ψnIVa ,nb (x1 , x2 ) = −ψnIVa ,nb (x2 , x1 )

so for both state holds that the probability density (= ψ 2 ) does not change when changing labels 1 and
2. The function ψ III is called the symmetric state function, the function ψ IV the anti-symmetric state
function. Note that while any linear combination of ψ I and ψ II would be an eigenfunction of Ĥ with
eigenvalue Ea + Eb , only (6.19) and (6.20) satisfy postulate 4a (check for yourself that for instance a
ψnVa ,nb (x1 , x2 ) = ψna (x1 )ψnb (x2 ) + 2ψnb (x1 )ψna (x2 ) is not symmetric or anti-symmetric).

At first sight, indistinguishability seems a rather esoteric condition, yet it is an essential part of quan-
tum mechanics. We are used to the fact that we can pin an identity to particles. Even while to the naked
eye two red snooker balls do not show any difference, we can still follow their individual paths on the snooker
table. This is not the case for elementary particles like electrons: there is not any distinction between them,
and their individual paths cannot be followed; in other words, they do not have any identity. So we already
“went wrong” at the start of this appendix by putting the labels 1 and 2 to the particles. However, it
becomes impossible to do any mathematical manipulations without such labels, but we then have to “force”
indistinguishability via functions like (6.19) and (6.20).
Particles which have a symmetric state function are called bosons, particles that have an anti-
symmetric state function are called fermions.

Postulate 4(b):
Electrons are fermions, that is, their state function is anti-symmetric.

There is no further explanation for why electrons are fermions. They just are. But there is an impor-
tant consequence: for an anti-symmetric state function (see 6.20) na cannot be equal to nb , otherwise the
function would become zero. So two electrons cannot be in the same state. This is known as Pauli’s ex-
clusion principle. Another property of the anti-symmetric state function is that the two particles tend to
stay away from each other: it looks like they repel each other, even while there is not repulsion term in the
Hamiltonian. This is known as the “fermi hole”,

In chapter 3 we have seen that we should not only consider the spatial state of the electrons, but also
the spin state. Before discussing this, it is convenient to first adopt a more compact notation. Instead of
the labels na , nb we just use a, b, and the positions x1 , x2 we write as 1, 2. Equation 6.17 then becomes

I Ĥ1 ψa (1) = Ea ψa (1)


ψab (1, 2) = ψa (1)ψb (2) where (6.21)
Ĥ2 ψb (2) = Eb ψb (2)

In this notation, particle 1 is in the state a, and particle 2 in state b. Note that this can then also hold for
3-D systems, where particle 1 has coordinates x1 , y1 , z1 and state a is characterized by na , la , ma . In this
4 in the special case that na = nb , ψ I and ψ II are also possible, but they are then identical to ψ III

88
notation, the anti-symmetric state function is
1
IV
ψab (1, 2) = √ [ψa (1)ψb (2) − ψb (1)ψa (2)] (6.22)
2
As explained in chapter 3, an electron can also have two spin states, “up” and “down”. We use an extra
quantum number s which takes the value -1/2 or +1/2 to indicate this state, so the total single-particle
state (= spin orbital) of particle 1 is then given by ψa,sa (1), and of particle 2 by ψb,sb (1), in which a and b
represent the quantum numbers of the spatial state.

Since the spatial coordinates and the spin coordinates never appear as a mixed term in the Hamiltonian, we
can write the spin orbital as the product of the spatial orbital and a “spin function” which we call α for a
positive spin, and β for a negative spin:
ψa,+ 21 (1) = ψa (1)α(1) ψa,− 21 (1) = ψa (1)β(1)

so ψa (1)α(1) represents a state in which particle 1 is in the spatial state “a” (for instance the 1s function
if the system is a hydrogen atom), and with a spin of +1/2. The spin gives the possibility to make an
anti-symmetric state function in which the electrons have the same spatial state (from now on we omit
the superscript IV, since it is clear that we are only discussing anti-symmetric functions; we also omit the
subscript a, b, since a and b only refer to the spatial state, not the total state):
1 1
ψ(1, 2) = √ [(ψa (1)α(1))(ψa (2)β(2)) − (ψa (1)β(1))(ψa (2)α(2))] = ψa (1)ψa (2) · √ [α(1)β(2) − β(1)α(2)]
2 2
In the second step we have separated the spatial and spin part. It is clear that the spatial part is symmetric,
while the spin state is anti-symmetric. The total function is thus anti-symmetric.

In case we have different spatial states, there are 4 possibilities for an anti-symmetric state function
1
ψ(1, 2) = √ [ψa (1)ψb (2) − ψb (1)ψa (2)] · α(1)α(2)
2
1
ψ(1, 2) = √ [ψa (1)ψb (2) − ψb (1)ψa (2)] · β(1)β(2)
2
1 1
ψ(1, 2) = √ [ψa (1)ψb (2) − ψb (1)ψa (2)] · √ (α(1)β(2) + α(2)β(1))
2 2
1 1
ψ(1, 2) = √ [ψa (1)ψb (2) + ψb (1)ψa (2)] · √ (α(1)β(2) − α(2)β(1))
2 2
For the first three functions, the spatial part is anti-symmetric, the spin part symmetric. For the 4th function
it is vice versa. The first three states are called the triplet state, the 4th function the singlet state. The
system will “prefer” the triplet state, since for an anti-symmetric spatial function the electrons will stay
away from each other, which is energetically more favorable. Now suppose that ψa en ψb have the same
energy, so they are degenerate states (for instance the px and py state for the H-atom). At first sight, it
does not seem the matter if you have two electrons in state ψa , or one electron in ψa and one electron in
ψb . However, only in the latter case you can have an anti-symmetric spatial state (one of the triplet states),
which is energetically more favorable as we just stated. This is the essence of Hund’s rule: for degenerate
spatial states, the electrons will distribute themselves as much as possible over the different states. Note
that of the three triplet states, there are two in which the electrons have the same spin (both α or β), and
one in which they have opposite spin (one α, the other β and vice versa, since it should be a symmetric
function). Two additional remarks
• A function in which both the spatial part and the spin part is anti-symmetric would be a symmetric
function. So ψ(1, 2) = √12 [ψa (1)ψb (2) − ψb (1)ψa (2)] · √12 (α(1)β(2) − α(2)β(1)) would not be a correct
state.

89
• The argument is given that the system prefers an anti-symmetric spatial state because then the particles
stay away from each other, which is energetically more favorable when they repel each other. Now note
that formally, when they repel each other, this term should be included into the Hamiltonian. This
would introduce a mixed term, and so separation of variables (and thus the whole picture of singe-
particle states or orbitals) would not be valid. So there is an inconsistency.
• As the rationale behind Hund’s rule it is often given that the electrons prefer to have parallel spins, so
they should occupy different orbitals. However, this is a consequence (and not the origin) of Hund’s
rule: if particles are in different spatial states, it is possible for them to have parallel spins.

90
Appendix F: Quantitative Results for
H2+ (not required)

The analysis in Chapter 4.3 can be regarded as only qualitative, owing to the approximation that was made
in the step from eqs. (4.27) and (4.28) to (4.29). However, it is possible to calculate the integrals ⟨sA | r1B |sA ⟩,
⟨sA | r1A |sB ⟩ and SAB exactly as function of the inter-nucleus distance R. The results are:
1 1 1
⟨sA | |sA ⟩ = (1 − e−2R ) − e−2R , ⟨sA | |sB ⟩ = (1 + R)e−R , (6.23)
rB R rA
and
SAB = ⟨sA |sB ⟩ = (1 + R + 13 R2 )e−R (6.24)
It is convenient to define the functions
F (R) = 1
R (1 − e−2R ) − e−2R , S(R) = (1 + R + 31 R2 )e−R , G(R) = 31 R2 e−R
in terms of which the Coulomb-, resonance- and overlap-integrals can be written as
HAA = E1s − F (R) , HAB = E1s S(R) − S(R) + G(R) , SAB = S(R) (6.25)
So the energy of the electron for the ground state in the LCAO approximation is
HAA + HAB E1s −F (R)+E1s S(R)−S(R)+G(R) F (R) S(R)−G(R)
E1 = = = E1s − − (6.26)
1 + SAB 1+S(R) 1+S(R) 1+S(R)
while the energy of the electron in the first excited state is
HAA − HAB E1s −F (R)−E1s S(R)+S(R)−G(R) F (R) S(R)−G(R)
E2 = = = E1s − + (6.27)
1 − SAB 1−S(R) 1−S(R) 1−S(R)
Note that if we could neglect F (R) you can most clearly see the bonding and anti bonding levels: E1 =
E1s − S(R)−G(R)
1+S(R) , E2 = E1s +
S(R)−G(R) LCAO
1−S(R) . Equation (6.26) is shown in figure 5.3 as E1 . It is interesting
to check the limiting behavior of this function. For R → ∞ the functions F (R), G(R) and S(R) all go to
zero, giving E1 = E1s = −1/2, which is indeed the correct value,
( namely
) the energy of the ground state of an
isolated H-atom. For the limit R → 0 you can show that R1 1 − e−2R → 2, so that F (0) = 1. Furthermore
and S(0) = 1, G(0) = 0, so that so that E1 = −1/2 − 1 = 0 = −3/2, while it should be -2, the energy of the
ground state of He+ . That you do not get the value -2 can indeed be expected: the atomic orbitals in the
LCAO approximation are given by (4.7), which are the state functions of hydrogen. These could never be
able to describe the state He+ , which is a much faster decaying function of r:
3/2 −1/2 −2rA 3/2 −1/2 −2rB
1she
A =2 π e , 1she
B =2 π e (6.28)
A logical step for improving would be to include a parameter λ in the atomic orbital, which would allow to
switch the AO from H-like to He-like.
1sλA = λ3/2 π −1/2 e−λrA , 1sλB = λ3/2 π −1/2 e−λrB

91
and determine the optimal value for λ for each R, knowing that for R → ∞: λ = 1 (hydrogen AO), and
R → 0: λ = 2 (helium AO). With these λ-dependent AO’s, the analytic solutions of the integral will also
become function of λ. The result of such a calculation is

HAA = λ2 E1s + λ2 − λ − λF (λR) , HAB = λ2 E1s S(λR) − λ(2 − λ)(S(λR) − G(λR) , SAB = S(λR)

Inserting this into E1 = (HAA + HAB )/(1 + SAB ) gives

HAA + HAB λ2 E1s +λ2 −λ−λF (λR)+λ2 E1s S(λR) − λ(2 − λ)(S(λR) − G(λR)
E1λ = =
1 + SAB 1+S(λR)

λ2 −λ λF (λR) λ(2 − λ)(S(λR) − G(λR)


= λ2 E1s + − −
1+S(λR) 1+S(λR) 1+S(λR)
You can easily check that for λ = 1 this expression reduces to (6.26). The λ in the argument of the functions
F , G, and S does not affect the limiting behavior, so for R → ∞: F = 0, G = 0, S = 0, and R → 0,
F = 1, G = 0, S = 1 so that we get

λ2 −λ λ λ(2 − λ) λ2 −λ − 2
lim E1λ = λ2 E1s , lim E1λ = λ2 E1s + − − = λ2 E1s +
R→∞ R→0 2 2 2 2

so for λ = 2, limR→0 E1λ = 4E1s = −2, which is the correct result. For arbitrary R we can get the optimal
value for λ from the variational principle, by requiring that dE1 /dλ = 0. If we re-calculate E1 by this
LCAO
procedure, we find that the total energy for EH + has a minimum at R = 2.00, almost equal to the value
2
that followed from the exact calculation, and a big improvement over the value 2.49 that was found with
using λ = 1 for every R. The dissociation energy is now De = 0.087 a.u., which is still too small as compared
to the exact value (0.103 a.u.), but again big improvement over the value 0.065 obtained before.

92
Appendix G: Derivation of the
Compton Equation (not required)

The process that is underlying the Compton effect is that of the collision between a single photon and a
single electron (see figure 1.1). We now assume that we can use the laws of classical mechanics to describe
this collision process, where the photon has energy hν and momentum hν/c. Conservation of energy gives
that:
1
hν0 = hν + mv 2 (6.29)
2
Conservation of momentum in the direction of the incident beam (see figure 1.1) gives that:
hν0 hν
= cos θ + mv cos ϕ (6.30)
c c
Conservation of momentum in the perpendicular direction (in the plane of the figure) gives that

0= sin θ − mv sin ϕ (6.31)
c
where v is the electron’s velocity after the collision. Eliminating ϕ and v from these 3 equations gives
( )2 ( )2
hνo hν hν
− cos θ + sin θ = m2 v 2 = 2m(hνo − hν) (6.32)
c c c
The first “= ” follows from (6.30) and (6.31), the second “=” from (6.29). Since a wave travels a distance
λ in a time T = 1/ν, we can write the speed of light as c = λν or c = λo νo , so that we can write (6.32) as
( )2 ( )2
1 1 1 2mc 1 1
− cos θ + sin θ = ( − )
λo λ λ h λo λ

Since (a − b cos θ)2 + (b sin θ)2 = (a − b)2 + 2ab(1 − cos θ) we can write this as
( )2
1 1 1 1 2mc 1 1
− +2 (1 − cos θ) = ( − )
λo λ λo λ h λo λ
( )
If we now assume that λ1o − λ1 << 2mc h (which is true for the typical values for λ) we can ignore the first
term on the LHS, giving
1 1 mc 1 1
(1 − cos θ) = ( − )
λo λ h λo λ
Multiplying left and right by λo λ gives
mc
(1 − cos θ) = (λ − λo )
h
which is equal to the experimental expression by Compton (1.2).

93
94
Answers to the Exercises

1.1
a) Yes, eigenvalue a1 b) No

1.3
b) C = (2a/π)1/4

1.4
Eigenvalue: a + b.

1.5
b) √1
3

1.7
b) ⟨E⟩ = 14 E1 + 41 E2 + 12 E3 , c) 1
4

2.1
e) ⟨x⟩ = L/2, ⟨p⟩ = 0.

2.2
( ) 2 2 2
a) ⟨x2 ⟩ = L2 31 − 2n12 π2 ; ⟨p2 ⟩ = h̄ L
n π
2 .
n = 1 gives the results for the ground state, n = 2 for the first excited state.
Tip: because V (x) = 0 you can write p̂2 = 2mĤ. It then follows directly that ⟨p2 ⟩ = 2mEn .
b) Using the results from 2.1e) gives: ∆x = 0.180756 L en ∆p = h̄π 1
L , so ∆p∆x = 0.568h̄ > 2 h̄.

2.3
1 1
2 + π

2.4
( )1/4 −ax2
a) ψo (x) = 2aπ e
c) ⟨x⟩ = 0, ⟨x2 ⟩ = 4a1

d) No, No
e) ⟨p⟩ = 0, ⟨p2 ⟩ = h̄2 a

95
2.6

8 3h2
c) N = abc d) Ground state: E = 8mL 2

e) and f)
6h2
1st excited state: E = 8mL 2 , degeneracy = 3, nodel planes: 1
9h2
2nd excited state: E = 8mL 2 , degeneracy = 3, nodal planes: 2
rd 11h2
3 excited state: E = 8mL2 , degeneracy = 3, nodal planes: 2
12h2
4th excited state: E = 8mL 2 , degeneracy = 1, nodal planes:3

3.2
The QM model gives the same result as the experiment.

3.3
23/2 −2r
1s = π 1/2
e , E1s = −2 (in a.e.)

4.2
a) Dimensionless
b) For R = 0: SAB = 1. For R → ∞: SAB = 0.

4.3
Only for b), c), e), g) and h) are the overlap integrals zero.

4.4
N = (1 + 2λS + λ2 )−1/2 .

5.1
a) E = -9 eV and E = -14 eV.
b) 5 eV √ √
c) cH = 0.8, cLi = 0.2.
d) Charge H = −0.6e, Charge Li = +0.6e. √ √
e) Coefficients of the anti-binding MO: cH = 0.2, cLi = − 0.8. There is no net charge on the atoms.

5.2
a) E = −16 eV and E = −6
√ eV. √ √ √
b) For E = −16 eV: c1 = 0.9, c2 = 0.1. For E = −6 eV: c1 = 0.1, c2 = − 0.9.

5.3
a,b) In the table is indicated when the overlap integral is zero.
2s1 2s2 2p1x 2p2x 2p1y 2p2y 2s1 +2s2 2s1 −2s2
C
2s 0 0 0 0 0
2pCz 0 0 0 0 0
2pCy 0 0 0 0 0 0
2pCx 0 0 0 0 0 0
1/2 1/2
c) χ3 = 2s1/2 , χ4 = 2pC 1 C
z , χ5 = 2px , χ6 = 2py , χ7 = 2s
1/2
, χ8 = 2pC
x , χ9 = 2py , χ10 = 2py .

96
d) σ orbitals: ϕ1 , ϕ2 , ϕ3 , ϕ6 . π orbitals: ϕ4 , ϕ5 , ϕ7 , ϕ8 .
e) All MO’s are filled.

5.4
c) χ1 = 2pOy , χ2 = 1s
H1
or 1sH2 , χ3 = 2pO O O
z , χ4 = 2s , χ5 = 2px , χ6 = 1s
H1
or 1sH2 , χ7 = 1sO
d) -0.4766 in a.u.

6.1
−2β

6.2
c) c1 = 0.5, c2 = 0.5, c3 = 0, c4 = 0, c5 = −0.5, c6 = −0.5

6.3
√ c1 c2 √ c3
E1 = α + 2β 1/2√ 1/ 2 1/2 √
a)
E2 = α √ 1/ 2 0 √ −1/ 2
E3 = α − 2β 1/2 -1/ 2 1/2
b) Ground state 1 : 1 : 1, first excited state 5/4 : 1/2 : 5/4.

6.4
√ √
b) c1 = c2 = 0, c3 = 1/ 2, c4 = −1/ 2
d) −0.96β, e) −1.31β, f) 1-0.8129 = 0.1871 in a.u.

6.5
b) c1 = 0.5 , c2 = −0.5 , c3 = 0, , c4 = 0 , c5 = −0.5 , c6 = 0.5
c)

d) −1.6568β
e) There is no net charge on C1 .

97

You might also like