ATOMIC AND MOLECULAR PHYSICS

PHAS0023

2018 – 2019

Prof. Stephen Hogan and Prof. David Cassidy

Department of Physics and Astronomy
University College London

12.01.2019
 General Information

Prof. Stephen Hogan

Office: Room A10 Physics Building
Email: [email protected]
Phone: 34324

Prof. David Cassidy

Office: Room A10 Physics Building
Email: [email protected]
Phone: 31360

Course outline:

1. Structure and spectra of atoms

2. Lasers and laser cooling

3. Molecular structure - diatomic molecules

4. Atoms in magnetic and electric fields

5. Particle scattering

Recommended text books:

- Spectra of Atoms and Molecules, P. F. Bernath, Oxford University Press,

2nd Edition (2005)

- Atoms and Molecules Interacting with Light, P. van der Straten and
H. Metcalf, Cambridge University Press, 1st Edition (2016)

- The Fundamentals of Atomic and Molecular Physics, R. L. Brooks, Springer,

1st Edition (2013)

2
Contents

I Atomic Structure 7

1 Introduction 8

2 One-electron atoms 10
2.1 Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Energy levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.1 Emission and absorption spectra . . . . . . . . . . . . 12
2.2.2 Accounting for the finite mass of the ion core . . . . . 15
2.2.3 Spectroscopic units . . . . . . . . . . . . . . . . . . . . 18
2.3 Wavefunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.3.1 Angular momentum in quantum mechanics . . . . . 19
2.3.2 Spherical harmonic functions . . . . . . . . . . . . . . 22
2.3.3 Parity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.3.4 Spectroscopic notation . . . . . . . . . . . . . . . . . . 24
2.3.5 Radial wavefunctions . . . . . . . . . . . . . . . . . . 25
2.3.6 Wavefunction normalisation . . . . . . . . . . . . . . 28
2.3.7 Energy level degeneracy . . . . . . . . . . . . . . . . . 30
2.3.8 Electron spin . . . . . . . . . . . . . . . . . . . . . . . 30
2.3.9 Transitions between energy levels . . . . . . . . . . . 31

3 Many-electron atoms 33
3.1 Hamiltonian for many-electron atoms . . . . . . . . . . . . . 33
3.2 Approximate solutions to the Schrödinger equation for N-
electron atoms . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.2.1 Independent-particle model . . . . . . . . . . . . . . . 34

3
 3.2.2 Central-field approximation . . . . . . . . . . . . . . . 37
3.2.3 Quantum defects . . . . . . . . . . . . . . . . . . . . . 42
3.3 Indistinguishable particles . . . . . . . . . . . . . . . . . . . . 43
3.3.1 The Pauli principle . . . . . . . . . . . . . . . . . . . . 44
3.4 Effect of the Pauli principle on atomic structure . . . . . . . . 45
3.4.1 Spin wavefunctions . . . . . . . . . . . . . . . . . . . 46
3.4.2 He atom wavefunctions . . . . . . . . . . . . . . . . . 49
Ground state of He . . . . . . . . . . . . . . . . . . . . 49
Excited states of the He atom . . . . . . . . . . . . . . 50
3.4.3 Spin multiplicity . . . . . . . . . . . . . . . . . . . . . 50
3.5 Exchange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3.5.1 Formal analysis of the Exchange Interaction . . . . . 52
3.6 Electronic configurations . . . . . . . . . . . . . . . . . . . . . 57
3.6.1 Nomenclature . . . . . . . . . . . . . . . . . . . . . . . 57
3.6.2 Periodic table of the elements . . . . . . . . . . . . . . 58
3.6.3 Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.6.4 Terms for two-electron atoms . . . . . . . . . . . . . . 61
Non-equivalent electrons . . . . . . . . . . . . . . . . 61
Equivalent electrons . . . . . . . . . . . . . . . . . . . 63
3.6.5 Hund’s rules . . . . . . . . . . . . . . . . . . . . . . . 65
3.7 Spin-orbit interaction . . . . . . . . . . . . . . . . . . . . . . . 66
3.7.1 Orbital magnetic moment . . . . . . . . . . . . . . . . 67
3.7.2 Spin magnetic moment . . . . . . . . . . . . . . . . . 69
3.7.3 The spin-orbit interaction . . . . . . . . . . . . . . . . 70
3.7.4 Spin-orbit operator . . . . . . . . . . . . . . . . . . . . 73
3.7.5 LS coupling (Russell-Saunders coupling) . . . . . . . 77
3.7.6 j j-coupling . . . . . . . . . . . . . . . . . . . . . . . . 78
3.7.7 Parity of terms in many-electron atoms . . . . . . . . 79
3.8 Hyperfine structure . . . . . . . . . . . . . . . . . . . . . . . . 80

4 Atomic spectra 83
4.1 Selection rules for electric dipole transitions . . . . . . . . . . 83
4.2 Rates of photon emission and absorption . . . . . . . . . . . 88
4.2.1 Two-level rate equations . . . . . . . . . . . . . . . . . 89

4
 4.2.2 Excited state lifetimes . . . . . . . . . . . . . . . . . . 92
Two-level system . . . . . . . . . . . . . . . . . . . . . 92
Multi-level system . . . . . . . . . . . . . . . . . . . . 94
4.2.3 Effects of finite excited state lifetimes . . . . . . . . . 95
4.2.4 Metastable states . . . . . . . . . . . . . . . . . . . . . 96
4.3 Lasers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.3.1 Three-level lasers . . . . . . . . . . . . . . . . . . . . . 98
4.3.2 Four-level lasers . . . . . . . . . . . . . . . . . . . . . 100
4.3.3 Typical laser characteristics . . . . . . . . . . . . . . . 102
4.4 Laser cooling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
4.4.1 Particle momentum in a gas of hot atoms . . . . . . . 103
4.4.2 Laser cooling Cs . . . . . . . . . . . . . . . . . . . . . 104
4.4.3 Laser deceleration . . . . . . . . . . . . . . . . . . . . 105
4.4.4 Magnetic trapping and further cooling . . . . . . . . 107

II Molecules 110

5 Diatomic molecules 111

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.1.1 Molecular Schrödinger equation . . . . . . . . . . . . 112
5.2 Born-Oppenheimer approximation . . . . . . . . . . . . . . . 113
5.3 The molecular hydrogen cation H+2 . . . . . . . . . . . . . . . 117
5.3.1 H+2 wavefunctions . . . . . . . . . . . . . . . . . . . . 118
5.3.2 H+2 electronic energies . . . . . . . . . . . . . . . . . . 121
5.4 The neutral hydrogen molecule H2 . . . . . . . . . . . . . . . 126
5.4.1 H2 electronic wavefunctions . . . . . . . . . . . . . . 127
5.4.2 Molecular orbitals . . . . . . . . . . . . . . . . . . . . 136
5.5 Nuclear motion . . . . . . . . . . . . . . . . . . . . . . . . . . 138
5.5.1 Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . 141
5.5.2 Vibration . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5.6 Total energy of a diatomic molecule . . . . . . . . . . . . . . 145
5.7 Spectra of diatomic molecules . . . . . . . . . . . . . . . . . . 148
5.7.1 Franck-Condon principle . . . . . . . . . . . . . . . . 153
5.8 Ionically bound molecules . . . . . . . . . . . . . . . . . . . . 154

5
 5.8.1 Dissociation energy of alkali metal halides . . . . . . 155
5.8.2 Ionic character of a bond . . . . . . . . . . . . . . . . . 157

III Atoms in electric and magnetic fields 160

6 Introduction 161
6.1 Hamiltonian operator . . . . . . . . . . . . . . . . . . . . . . . 161
6.1.1 First order perturbation theory . . . . . . . . . . . . . 162

7 Atoms in magnetic fields 163

7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
7.2 Zeeman effect when S = 0 . . . . . . . . . . . . . . . . . . . . 164
7.3 Zeeman effect when S , 0 (weak-field regime) . . . . . . . . 165
7.4 Zeeman effect when S , 0 (strong-field regime) . . . . . . . . 171
7.5 The Stern-Gerlach experiment . . . . . . . . . . . . . . . . . . 172
7.6 Multistage Zeeman deceleration of atomic beams . . . . . . 174

8 Atoms in electric fields 177

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
8.2 Stark effect in weak fields . . . . . . . . . . . . . . . . . . . . 178
8.2.1 Quadratic Stark effect . . . . . . . . . . . . . . . . . . 178
8.2.2 Linear Stark effect . . . . . . . . . . . . . . . . . . . . . 181
Energy shifts . . . . . . . . . . . . . . . . . . . . . . . 182
Electron charge distributions . . . . . . . . . . . . . . 184
8.3 Ionisation in strong electric fields . . . . . . . . . . . . . . . . 184

IV Scattering 188

9 Low-energy particle scattering 189

9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
9.2 Scattering processes . . . . . . . . . . . . . . . . . . . . . . . . 190
9.3 Scattering cross-sections . . . . . . . . . . . . . . . . . . . . . 191
9.4 Quantum treatment of low energy particle scattering . . . . 193
9.5 Angular momentum partial waves . . . . . . . . . . . . . . . 197

6
 Part I

Atomic Structure

7
Chapter 1

Introduction

The objective of this course is to provide a detailed overview of the struc-

ture of atoms and diatomic molecules, and their interactions with electric,
magnetic and electromagnetic fields.

This information is of importance in many areas of research including, for

example1 :

1. Tests of fundamental physics: Comparison of high precision experi-

mental measurements of the energy-level structure of H, He, H2 and other
simple atoms and molecules, with the results of calculations provide a basis
for fundamental tests of atomic and molecular quantum mechanics. For
example, at present very precise measurements of energy intervals in the
hydrogen atom suggest that the radius of the proton may be different from
that which has been determined up to now2 . This is motivating new preci-
sion spectroscopy experiments involving hydrogen-like atoms which may
help understand the origins of this discrepancy.

2. Investigations of matter-antimatter asymmetries: Comparison studies

of the energy level structure of the hydrogen (H) and anti-hydrogen (H)
atoms are currently being developed with the aim of investigating why our
1
Note: the following list is only a very small subset of the research areas in which the
content of this course if of importance.
2
The Proton Radius Puzzle, J. C. Bernauer and R. Pohl, Scientific American, February 2014.

8
observable Universe appears to composed predominantly of matter, even
though at the time of Big Bang it is thought that equal amounts of matter
and antimatter should have been created. In this area, studies involving
the hydrogen-like positronium atom (Ps), composed of an electron bound
to its antiparticle, the positron, may also provide insight.

3. Studies of extrasolar planets: Accurate measured or calculated spec-

tra of small molecules are of importance in many areas of research ranging
from environmental sensing, to searches for signs of life on planetary objects
outside the solar system (extrasolar planets, or exoplanets). Comparison
of these spectra with those observed in emission or absorption from such
planets allow the molecular composition of their atmospheres to be deter-
mined. By precisely identifying the atmospheric composition it may be
possible to identify signs of life.

4. Atomic clocks: At present the second is defined with respect to the

inverse of the frequency of the electromagnetic radiation required to drive
a microwave transition between to energy levels in the caesium (Cs) atom.
However, more precise clocks have now been developed in which optical
transitions are driven in, e.g., strontium (Sr) atoms or aluminium (Al+ ) ions.
A detailed knowledge of the energy level structure of these atoms and their
susceptibility to electric and magnetic fields is essential for the identifica-
tion of appropriate transitions for the further development of more precise
atomic clocks such as these.

5. Quantum computing: Efforts in many laboratories around the world

to develop new quantum technologies, including those related to quantum
computation, center around the use of pairs of energy levels in neutral
atoms [e.g., helium (He) or rubidium (Rb)], or in atomic ions [e.g., the
calcium ion (Ca+ )], as quantum bits (qubits). A precise knowledge of the
lifetimes of these energy levels, and their sensitivity to electric, magnetic
and electromagnetic fields is crucial to take advantage of the opportunities
that these systems offer in this area.

9
Chapter 2

One-electron atoms

To begin we consider atoms and atomic ions with one electron bound to a
nucleus, or ion core, with a charge q = +Zcore e, where e is the charge of the
electron. Many of these atoms exist and include those listed in the table
below.

System Symbol Composition Zcore

Hydrogen atom H p+ – e− +1e
Deuterium atom D np+ – e− +1e
Positronium atom Ps e+ – e− +1e
He+ ion He+ He2+ – e− +2e
Li2+ ion Li2+ Li3+ – e− +3e

2.1 Hamiltonian

The two particles in these systems, the electron with a charge −e and the
ion core with charge +Zcore e, interact via the Coulomb interaction for which
the potential energy is given by

−Zcore e2
V(r) = , (2.1)
4πǫ0 r

where ǫ0 is the permittivity of the vacuum, and r = |~r | is the inter-particle

separation.
In this situation, if the ion core A+ (see Figure 2.1) is considered to be

10
 e-

A+
Ion core (charge +eZcore)

Figure 2.1: A one electron atom in which a single electron is bound to a

nucleus, or ion core, of charge +eZcore .

infinitely heavy, the Hamiltonian governing the motion of the electron is

~2 2 Zcore e2
Ĥ = −∇ − , (2.2)
2me 4πǫ0 r
where ~ = h/(2π) with h the Planck constant, and me is the mass of the
electron.

2.2 Energy levels

The energy eigenvalues, E, associated with this quantum system can be

obtained by solving the Schrödinger equation

Ĥ Ψ = E Ψ, (2.3)

where Ψ are the wavefunctions of the electronic states. This leads to a set
of energy eigenvalues, En , of the form
me Z2core e4
En = − , (2.4)
32π2 ǫ20 ~2 n2
where n = 1, 2, 3, . . . ∞ is the principal quantum number (see Figure 2.2).
From Equation 2.4 the energy of the n = 1 ground state of the H atom is
calc
En=1 = −2.179 87 × 10−18 J
≡ −109 737.316 cm−1 . (2.5)

However, the experimentally measured value for this ground state energy
is
expt
En=1 = −2.178 68 × 10−18 J
≡ −109 677.583 cm−1 . (2.6)

11
 0

n=3
-2

n=2
-4
Energy / e (eV)

-6

-8

-10

-12

-14
n=1
-40 -30 -20 -10 0 +10 +20 +30 +40
Radial position (10-10 m)

Figure 2.2: Electronic energy levels (red horizontal lines - see Equation 2.4)
and Coulomb potential (black curves) associated with the H atom.

The discrepancy between these calculated and measured values is a result

of the assumption that, in this simplified description of the H atom, the ion
core (proton) is treated as being infinitely heavy.

2.2.1 Emission and absorption spectra

Early emission spectra of H atoms were recorded by generating electric

discharges in cells containing hydrogen gas. The energies of the emitted
photons observed corresponded to the energy differences between the dis-
crete electronic energy levels of the atoms (see Figure 2.3).
In emission, if Eni is the energy of the initial state with principal quantum
number ni , and Enf is the energy of the lower-lying, final state, nf , where

me Z2core e4 1
Eni = − (2.7)
32π2 ǫ20 ~2 n2i

12
 (a) Photoemission

n=3 Ei = En = 3

n=2 Ef = En = 2
Energy

Emission of photon
h = Ei - Ef

n=1

(b) Photoabsorption

n=3 Ef = En = 3

n=2 Ei = En = 2
Energy

Absorption of photon
h = Ef - Ei

n=1
Figure 2.3: The processes of (a) emission, and (b) absorption of a photon
by a H atom.

13
and
me Z2core e4 1
Enf = − , (2.8)
32π2 ǫ20 ~2 n2f

the energy of the emitted photon is equal to the energy difference, ∆Ei f ,
between the two states
 
me Z2core e4  1 1 
∆Ei f = Eni − Enf =  − 2  . (2.9)
32π2 ǫ20 ~2  n2f ni

If ni > nf a photon is emitted [Figure 2.3(a)], while if ni < nf a photon can

be absorbed by the atom [Figure 2.3(b)].
Regular series of sharp spectral lines were first identified in H atom
emission spectra by Johann Balmer in 1885. Balmer realised that these
sharp lines indicated that photoemission must occur following transitions
between discrete energy levels within the atoms.
Following this, in 1888 Johannes Rydberg found an empirical formula
linking the wavelengths of the observed spectral lines, λi f , with the quan-
tum numbers of the states involved in the transitions
 
1  1 1 
= R  2 − 2  , (2.10)
λi f nf ni

where R is a constant.
Since

∆Ei,f = hνi f
hc
= , (2.11)
λi f
where νi f is the transition frequency, and c is the speed of light in vacuum,
by comparing Equation 2.9 and Equation 2.10

me Z2core e4
R = , (2.12)
64π3 ǫ20 ~3 c

and is known as the Rydberg constant.

Because this expression (Equation 2.12) for the Rydberg constant results
from the comparison of Rydberg’s formula (Equation 2.10) with an equation
obtained for a hydrogen atom in which the electron is considered to be

14
bound to an infinitely heavy ion core, when Zcore = +1 this value of the
Rydberg constant is denoted R∞ , such that

me e4
R∞ =
64π3 ǫ20 ~3 c
= 10 973 731.6 m−1
= 109 737.316 cm−1 (2.13)

2.2.2 Accounting for the finite mass of the ion core

In the above treatment of one-electron atoms the single electron of mass

me = 9.109 × 10−31 kg is considered to be bound to an infinitely heavy ion
core. The result of this is that the motion of the electron does not affect the
center-of-mass motion.
This is generally not a bad approximation since, e.g., for the H atom the
mass of the ion core Mcore = mp , where mp = 1.673 × 10−27 kg, so

mp ≫ me . (2.14)

and
mp
= 1836. (2.15)
me
However, corrections to the energy level structure arising from the finite
mass of the ion core are important. These effects give rise to energy level
shifts for different isotopes of the same atomic species, e.g., H, D, and T,
and as a result different transitions frequencies.
To treat the finite mass of the ion core in a one-electron atom we must
consider the reduced mass, µM , of the two-body system. The reduced mass
is the effective mass of the system in the frame of reference associated with
the center-of-mass.
In the case of the H atom, the forces acting on the electron and the proton
in this frame of reference are indicated in Figure 2.4, where

~p
~r = ~re − R (2.16)

and

F ~e, p .
~p, e = −F (2.17)

15
 CM
Fp, e Fe, p

p+ Rp re e-

mass = mp mass = me

Figure 2.4: The forces acting on the electron and proton in the H atom
in the frame of reference associated with the center-of-mass (CM) of the
system.

Given that
~p
d2 R 2
~p, e = mp
F and ~e, p = me d ~re ,
F
dt2 dt2

~p, e
F ~p
d2 R ~e, p
F d2~re
= and = . (2.18)
mp dt2 me dt2

With the result that

~e, p
F ~p, e
F 2~
d2~re d R p
− = − (2.19)
me mp dt2 dt2

d2~r
= . (2.20)
dt2
Using Equation 2.17 this leads to
!
~e, p 1 1 d2~r
F + = . (2.21)
me mp dt2

If
1 1 1
= + , (2.22)
µM me mp

an expression for the equation of motion of the two-body system is obtained:

2
~e, p = µM d ~r ,
F (2.23)
dt2

16
in terms of the reduced mass
me mp
µM = . (2.24)
me + mp

In the previous treatment of a one-electron atom with an infinitely heavy

ion core, the reduced mass is
me mp
µ∞ = lim (2.25)
mp →∞ me + mp

= me . (2.26)

Therefore Equation 2.13 can also be expressed in the form

µ∞ e4
R∞ = . (2.27)
64π3 ǫ20 ~3 c

To account for the effect of the finite mass of the proton in the H atom
the Rydberg constant, RH , must be determined, such that

µH e4
RH = , (2.28)
64π3 ǫ20 ~3 c

with
me mp
µH =
me + mp
= 0.999456 me . (2.29)

Therefore
µH
RH = R∞
me
= 0.999456 R∞
= 109 677.583 cm−1 . (2.30)

Or more generally for a one-electron atom of mass M, with an ion core of

charge +e Zcore ,
µM 2
RM = R∞ Z . (2.31)
me core

17
2.2.3 Spectroscopic units

The Rydberg constant1 , R∞ , is defined as a wave number in units of m−1 ,

or cm−1 , as

R∞ = 10 973 731.568 508 ± 0.000 065 m−1

= 109 737.315 685 08 ± 0.000 000 65 cm−1 . (2.32)

It is most common in atomic and molecular spectroscopy to quote the wave

number, ν̃, of the radiation emitted or absorbed, such that
1
ν̃ = , (2.33)
λ
where λ is the wavelength of the radiation. The wave number is there-
fore is the number of wavelengths per unit length (i.e., per meter, or per
centimeter).
Although SI units will generally be used throughout this course, there
are some occasions where it is convenient to express quantities in atomic
units. In this unit system

The atomic unit of mass is the electron mass, me

The atomic unit of charge is the charge of the electron, e

The atomic unit of length is the Bohr radius, a0

The atomic unit of energy is 2hcRM

and therefore

~ = 4πǫ0 = a0 = e = me = 1. (2.34)

For example, in this system of units the Coulomb potential associated with
the interaction of the electron and the proton in the hydrogen atom
e2
V(r) = − (2.35)
4πǫ0 r
is expressed as
1
V(r) = − . (2.36)
r
1
CODATA Internationally recommended 2014 values of the fundamental physical con-
stants (http://physics.nist.gov/cuu/Constants)

18
2.3 Wavefunctions

The Shrödinger equation associated with the Hamiltonian in Equation 2.2,

when expressed in spherical polar coordinates (r, θ, φ), where

x = r sin θ cos φ (2.37)

y = r sin θ sin φ (2.38)
z = r cos θ, (2.39)

results in wavefunctions Ψ(r, θ, φ), that are separable into radial, R(r), and
angular, Y(θ, φ), components such that

Ψ(r, θ, φ) = R(r) Y(θ, φ). (2.40)

2.3.1 Angular momentum in quantum mechanics

The classical orbital angular momentum, ~L = (Lx , L y , Lz ), of a particle can be

expressed in terms of the radial position vector, ~r = (x, y, z), and the linear
momentum, ~
p = (px , p y , pz ) as

~L = ~r × ~
p. (2.41)

therefore

Lx = ypz − zp y (2.42)
L y = zpx − xpz (2.43)
Lz = xp y − ypx . (2.44)

From the Correspondence Principle

~
p ⇒ ~ ~ˆ
pˆ = −i~∇ (2.45)

and

~r ⇒ ~rˆ. (2.46)

19
Therefore the quantum mechanical operators associated with angular mo-
mentum are

~ˆ
~Lˆ = −i~(~rˆ × ∇) (2.47)

!
∂ ∂
L̂x = −i~ ŷ − ẑ (2.48)
∂z ∂y
!
∂ ∂
L̂ y = −i~ ẑ − x̂ (2.49)
∂x ∂z
!
∂ ∂
L̂z = −i~ x̂ − ŷ (2.50)
∂y ∂x

and

~Lˆ 2 = L̂2x + L̂2y + L̂2z . (2.51)

This set of angular momentum operators obey the commutation relation

h i
L̂i , L̂ j = L̂i L̂ j − L̂ j L̂i
= i~L̂k , (2.52)

ˆ
where i, j, k ≡ x, y, z. Consequently, the individual x, y or z components of ~L
cannot be assigned definite values simultaneously unless ~L = (0, 0, 0).
ˆ
However, the operator ~L2 does commute with any one of the individual
components of ~L, i.e.,
 
~Lˆ 2 , L̂ j = 0. (2.53)

For example,
 
~Lˆ 2 , L̂z = 0. (2.54)

ˆ
Therefore it is possible to have simultaneous eigenfunctions of ~L2 and one
ˆ
component of ~L. By convention the component L̂z is chosen.
Under these conditions the angular momentum vector ~L can be consid-
ered to precess about the z-axis (see Figure 2.5). This precessional motion
ˆ
preserves ~L2 and L̂z while the time-averaged values of L̂x and L̂ y are zero.

20
 z

L
Lz
x

Figure 2.5: Precession of the angular momentum vector ~L about the z-axis.

Converting the expressions in Equation 2.48 – Equation 2.51 into spher-

ical polar coordinates leads to
!
∂ ∂ ∂
L̂z = −i~ x̂ − ŷ = −i~ (2.55)
∂y ∂x ∂φ
" ! #
~Lˆ 2 = −~2 1 ∂ ∂ 1 ∂2
sin θ + . (2.56)
sin θ ∂θ ∂θ sin2 θ ∂φ2
The simultaneous eigenfunctions of the eigenvalue equations involving
ˆ
the operators L̂z and ~L2 are the spherical harmonic functions, Yℓ,m (θ, φ).

~Lˆ 2 Yℓ,m (θ, φ) = ℓ(ℓ + 1)~2 Yℓ,m (θ, φ) (2.57)

L̂z Yℓ,m (θ, φ) = m~ Yℓ,m (θ, φ). (2.58)

ℓ = 0, 1, 2, 3, . . . is the orbital angular momentum quantum number

m = 0, ±1, ±2, · · · ± ℓ is the azimuthal quantum number

The discrete values of these quantum numbers indicate that the electron
orbital angular momentum is quantised. The value of m corresponds to
~ onto the
the projection of the electron orbital angular momentum vector, ℓ,
z-axis, and is also often referred to as the magnetic quantum number.

21
2.3.2 Spherical harmonic functions

The spherical harmonic functions, Yℓ,m (θ, φ), are separable functions, i.e.,

Yℓ,m (θ, φ) = Φm (φ) Θℓ,m (θ). (2.59)

The two components of these functions are

1
Φm (φ) = √ eimφ , (2.60)
2π
and

Θℓ,m (θ) ∝ Pm
ℓ (cos θ), (2.61)

where the functions Pm

ℓ
(x) are associated Legendre polynomials.
The complete set of normalised spherical harmonic functions can be
expressed explicitly as
" #1/2
1 imφ (2ℓ + 1)(ℓ − |m|)!
Yℓ,m (θ, φ) = √ e P|m|
ℓ
(cos θ), (2.62)
2π 2(ℓ + |m|)!

with examples for low values of ℓ and m in Table 2.1 below, and in Figure 2.6.

ℓ m Yℓ,m (θ, φ)

0 0 Y0,0 = √1
4π
q
3
1 0 Y1,0 = 4π cos θ
q
3
±1 Y1,±1 = ∓ sin θ e±iφ
8π
q  
5 2
2 0 Y1,0 = 16π 3 cos θ − 1
q
15
±1 Y2,±1 = ∓ 8π sin θ cos θ e±iφ
q
15 2 ±2iφ
±2 Y2,±2 = 32π sin θ e

Table 2.1: Spherical harmonic functions with ℓ = 0, 1 and 2, and m = 0, ±1

and ±2 as indicated.

22
 Figure 2.6: Spherical harmonic functions, Yℓ,m (θ, φ), for which ℓ = 0, 1
and 2. In each panel, the sign of the function is indicated by the colour –
positive values are indicated in red and negative values are indicated in
blue.

2.3.3 Parity

Parity is an important symmetry in atomic and molecular systems. When

applied to a function of the position vector ~r, the parity transformation
corresponds to an inversion of the function through the origin such that
~r → −~r. When the parity operator, P̂, associated with this transformation is
applied to a function f (~r ), the result can therefore be expressed as

P̂ f (~r ) = f (−~r ). (2.63)

i. If P̂ f (~r ) = f (~r )

i.e., if f (−~r ) = f (~r )

⇒ the function f (~r ) is said to have even parity.

23
 ii. If P̂ f (~r ) = − f (~r )

i.e., if f (−~r ) = − f (~r )

⇒ the function f (~r ) is said to have odd parity.

When applied to the spherical harmonic functions, Yℓ,m (θ, φ), in Equa-
tion 2.62:

The parity of eimφ is given by (−1)m

The parity of Pm
ℓ
(cos θ) is given by (−1)ℓ−m

Therefore

P̂ Yℓ,m (θ, φ) = (−1)ℓ Yℓ,m (θ, φ), (2.64)

and the parity of Yℓ,m (θ, φ) is given by (−1)ℓ so that

If (−1)ℓ = +1, the corresponding spherical harmonic has even parity

If (−1)ℓ = −1, the corresponding spherical harmonic has odd parity

2.3.4 Spectroscopic notation

In the spectroscopic notation used to label one-electron orbitals in atoms,

the orbital angular momentum quantum number, ℓ, is represented by a
lower case letter.

ℓ= 0 1 2 3 4 5 ...
s p d f g h ...

This labelling system arose from early observations of atomic spectra in

which emission from excited states with ℓ = 0 was seen to give rise to sharp
spectral features leading to these states been labelled s-states. The strongest,
principal, features in such spectra arose from emission from excited states
with ℓ = 1, hence these states were labelled p-states, while the emission
from states with ℓ = 2 was seen to be more diffuse, leading to the labelling
of these states as d-states.

24
For example:

- The ground state of the H atom has n = 1, ℓ = 0, m = 0 and is therefore

denoted the 1s state

- The excited state of the H atom for which n = 2, ℓ = 1, m = 0 is

denoted the 2p state (see Figure 2.7)

0
... I. P.

3s 3p 3d
Energy / hc (cm-1)

2s 2p

-50000

-100000

1s

Figure 2.7: Energy levels of the H atom labelled using the spectroscopic
notation for one-electron atomic orbitals.

2.3.5 Radial wavefunctions

For a one electron atom such as the H atom, the radial component, R(r),
of the total wavefunction in Equation 2.40 can be determined analytically2 ,
and takes the form

2 ℓ
     
2 2ℓ+1 2
Rn,ℓ (r) = Nn,ℓ r exp − r Ln−ℓ−1 r , (2.65)
na0 na0 na0
2
Quantum Mechanics of One- and Two-Electron Atoms H. A. Bethe and E. E. Salpeter,
Springer, Berlin, 1957

25
where Nn,ℓ is a normalisation constant,
 3/2 r
2 (n − ℓ − 1)!
Nn,ℓ = , (2.66)
na0 2n(n + ℓ)!

Lba (x) are the associated Laguerre polynomials which take the form
a
X (a + b)!
Lba (x) = (−1)k xk , (2.67)
(a − b)! (b + k)! k!
k=0

and a0 is the Bohr radius (the classical radius of the electron orbit in the
state with n = 1) for an electron bound to an infinitely heavy proton, such
that
4πǫ0 ~2
a0 = . (2.68)
e2 me
However, as in the case of the Rydberg constant, the Bohr radius can be
corrected for the reduced mass, µM , and charge of the ion core, Zcore , such
that in general

4πǫ0 ~2
aM = (2.69)
Zcore e2 µM

me 1
= a0 . (2.70)
µM Zcore
The radial wavefunctions in a one-electron atom therefore only depend
on the principal quantum number n, and the electron orbital angular mo-
mentum quantum number ℓ. Some examples of radial wavefunctions for
low values of n and ℓ are given in Table 2.2 below.
For each of these radial wavefunctions, as r → ∞ the exponential term
dominates, as a result for all n and ℓ,

Rn,ℓ (r) → 0 as r → ∞ (2.71)

From the exponential term in Equation 2.65 and the dependence of aM on

Zcore , it can be seen that:

· As the value of Zcore is increased for a particular value of n, aM de-

creases and the amplitude of the wavefunctions reduce more rapidly
with increasing r

26
 n ℓ Rn,ℓ (r)

1 3/2 − aM
  r
1 0 R1,0 = 2 aM e

1 3/2
    − r
2 0 R2,0 = 2 2aM 1− 2arM e 2aM
 3/2   − r
√1 1 r
1 R2,1 = 2aM aM e
2aM
3

Table 2.2: Radial wavefunctions for states with n = 1 and 2, and ℓ = 0

and 1 in one-electron atoms.

⇒ The electron is therefore localised more closely to the ion core

when the core is more highly charged

· As the value of n is increased for a selected value of Zcore , the ampli-

tude of the wavefunctions reduces more slowly with increasing r

⇒ The charge of the electron is therefore extended over a longer

range in more highly excited states

R3s(r)

R2s(r)

nℓm (r, , ) = R nℓ(r) · Yℓm ( , )

R1s(r)

Figure 2.8: Radial wavefunctions of the 1s, 2s and 3s states of the H atom
displayed at the location of each energy level (dashed horizontal lines) in
the Coulomb potential (black curves).

27
2.3.6 Wavefunction normalisation

The total wavefunctions for the one-electron atom must be normalized so

that
Z  2

Ψn,ℓ,m (r, θ, φ) dτ = 1, (2.72)

therefore
Z ∞ 2 Z 2πZ π  2
Rn,ℓ (r) r2 dr
 
Yℓ,m (θ, φ) sin θdθdφ = 1. (2.73)
0 0 0

Since the spherical harmonic functions are a normalised set of functions,

i.e.,
Z π 2

Yℓ,m (θ, φ) sin θdθdφ = 1, (2.74)
0
 2
the function Pn,ℓ (r) = Rn,ℓ (r) r2 , gives the radial probability distribution of
the electron. This is the probability of finding the electron at a distance r
from the ion core. Some examples of these radial probability distributions
for selected values of n and ℓ in the H atom are displayed in Figure 2.9.

0.6 0.6
Normalised probability density r 2 |Rnl(r)| 2

Normalised probability density r 2 |Rnl(r)| 2

0.4
1s 0.4
3d
0.2 0.2

0 0
0.6 0.6

0.4
2s 0.4
3p
0.2 0.2

0 0
0.6 0.6

0.4
3s 0.4
3s
0.2 0.2

0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
Radial position / a0 Radial position / a0

Figure 2.9: Electron radial probability distributions associated with the 1s,
2s, 3s, 3p, and 3d states in the H atom as indicated.

28
 As can be seen from this figure, for a selected value of ℓ (left column) the
spatial extent of the electron charge distribution increases with increasing
values of n. For a selected value of n (right column), the probability of
finding the electron near to the ion core decreases with increasing values of
ℓ. This occurs because the centrifugal barrier experienced by the orbiting
electron pushes it outwards to larger radial positions. As a result, low-ℓ
orbitals (e.g., the s- or p-orbitals) are said to be more core-penetrating than
orbitals with higher values of ℓ (e.g., the d-orbital).
The radial probability density distribution associated with the 1s ground
state of a hydrogen-like atom can be expressed explicitly in the form
"  3/2  #2
 2
2 1 r
Pn,ℓ (r) = Rn,ℓ (r) r = 2 exp − r2 . (2.75)
aM aM
From Figure 2.9 it can be seen that this function has a single intensity
maximum. The position of this intensity maximum corresponds to the
most probably radial position, rprob , of the electron. By calculating the
point in this function where the derivative with respect to r is zero, it is
found that

rprob = aM . (2.76)

Therefore, in the ground state of the H atom the orbital radius of the electron
determined in the semiclassical Bohr model corresponds to the radius at
which it is most likely to find the electron in the quantum treatment of the
atom.
The average radial position of the electron is given by the expectation
value, hri, where
Z
hri = Ψ∗n,ℓ,m (r, θ, φ) r̂ Ψn,ℓ,m (r, θ, φ) dτ (2.77)
Z ∞
= r Pn,ℓ (r) dr, (2.78)
0

and, after integration, for any nℓ state in a hydrogen-like atom is

aM h 2 i
hri = 3n − ℓ(ℓ + 1) . (2.79)
2
For the 1s state in the hydrogen atom,
3 aH
hr1,0 i = . (2.80)
2

29
This average radial position of the electron in the ground state of the H
atoms is larger than the rprob (see Equation 2.76) because of the long tail in
the radial probability distribution toward large values of r.

2.3.7 Energy level degeneracy

In a one-electron atom the energy eigenvalues, En , given by the Rydberg

formula
En RM
= − , (2.81)
hc n2
depend only on the value of the principal quantum number n. For each
value of n, there exist n corresponding values of ℓ [i.e., ℓ = 0, 1, 2, . . . (n − 1)],
and 2ℓ + 1 values of m (i.e., m = 0, ±1, ±2, · · · ± ℓ).
As a result the total number of degenerate eigenstates, i.e., states with
the same energy, En , in a one-electron atom is
n−1
X
(2ℓ + 1) = n2 . (2.82)
ℓ=0

and there is complete energy degeneracy with respect to ℓ and m.

m-degeneracy: The m-degeneracy of the eigenstates in a one-electron

atom is a feature of a central potential. Such a potential, V, depends
on the magnitude of ~r but not its direction, i.e., V = V(r).

ℓ-degeneracy: The ℓ-degeneracy of the eigenstates in a one-electron

atom is a characteristic of a potential, such as the Coulomb potential,
that depends on 1/r. The ℓ-degeneracy in the atom is removed if
the dependence of the potential experienced by the electron on r is
modified.

2.3.8 Electron spin

In addition to its orbital angular momentum when bound within an atom, a

relativistic treatment of a one-electron atom leads to an additional quantised
quantity. This is its intrinsic spin angular momentum, ~s.

30
 ˆ
As in the case of the orbital angular momentum operators ~L2 and L̂z , the
operators associated with the square of the magnitude of the spin vector,
~ˆ 2 , and the projection of the spin vector in the z-dimension, Ŝz , commute,
S
i.e.,
 
~ˆ 2 , Ŝz
S = 0. (2.83)

Considering the one-electron spin eigenfunctions, χs,ms , where s is the elec-

tron spin quantum number, and ms is the projection of the electron spin
vector in the z-dimension,

~ˆ 2 χs,m 1
S s = s(s + 1) ~2 χs,ms with s= , (2.84)
2
and
1
Ŝz χs,ms = ms ~ χs,ms with ms = ±s = ± . (2.85)
2
The total electronic wavefunction of a one-electron atom including elec-
tron spin can therefore be expressed in the form

Radial Angular
z}|{ z }| {
Ψn,ℓ,mℓ ,s,ms (r, θ, φ) = Rn,ℓ (r) Yℓ,mℓ (θ, φ) χs,ms , (2.86)
| {z } |{z}
Space Spin

where the projection of the orbital angular momentum, m, from Section 2.3.1
has been replaced by mℓ to distinguish it from the projection of the spin
angular momentum, ms .

2.3.9 Transitions between energy levels

In all quantum systems transitions between different eigenstates resulting

from emission or absorption of electromagnetic radiation are only allowed
to occur if they fulfill an appropriate set of selection rules. These selection
rules arise following the treatment of the interaction of the atom with an
electromagnetic field which will be discussed later in the course.
In a one-electron atom the selection rules for transitions between differ-
ent eigenstates require that:

31
 0
... I. P.
3s 3p 3d
Energy / hc (cm-1)

2s 2p

-50000

-100000

1s

Figure 2.10: Examples of allowed transitions between nℓ states of the H

atom.

• ∆n = any value: The value of n can change by any value when a

photon is emitted or absorbed.

• ∆ℓ = ±1: The value of ℓ must change by +1, or by −1 upon the emission

or absorption of a photon to conserve angular momentum.

• ∆mℓ = 0, ±1: The value of mℓ can change by 0 or by ±1 upon the

emission or absorption of a photon.

Examples of transitions in the H atom that fulfill these selection rules are
indicated by the arrows in Figure 2.10.

32
Chapter 3

Many-electron atoms

For one-electron atoms such as the H atom the electronic Schrödinger equa-
tion can be solved analytically. For more complex atoms with many elec-
trons, only numerical solutions to the Schrödinger equation are possible.
However, using simple models it is possible to estimate the energies of, and
classify, the eigenstates of atoms with more than one electron.

3.1 Hamiltonian for many-electron atoms

The electronic Hamiltonian for an atom with N electrons takes the form
Single e− K.E. e− − nucleus attraction
z }| { z }| {
N " #
X ~2 2 Znucl e2
Ĥ(~r1 , ~r2 , ~r3 , . . . ~rN ) = − ∇ − + ...
2me i 4πǫ0 ri
i=1
N
N X
X e2
+ , (3.1)
4πǫ0 ri j
i=1 j>i
| {z }
e− − e− repulsion

where ri j = |~ri − ~r j | is the separation between electron i and electron j.

For example, in the two-electron He atom, the electron-electron repul-
sion term, Vrep , is

e2
Vrep = , (3.2)
4πǫ0 r12

33
while in the three-electron Li atom,

e2 e2 e2
Vrep = + + . (3.3)
4πǫ0 r12 4πǫ0 r13 4πǫ0 r23
In general for an N-electron atom, Vrep , has N(N − 1)/2 contributing terms.
The Hamiltonian for an N-electron atom, Equation 3.1, can therefore be
expressed in the form of a sum over a set of single-electron Hamiltonians,
ĥ(~ri ), plus a sum over the full set of electron-electron interaction terms, i.e.,
X X e2
Ĥ(~r1 , ~r2 , ~r3 , . . . ~rN ) = ĥ(~ri ) + . (3.4)
4πǫ0 ri j
i i, j>i

Because the electron-electron repulsion depends on the instantaneous

positions of every pair of electrons, the Schrödinger equation associated
with Ĥ(~r1 , ~r2 , ~r3 , . . . ~rN ) cannot be completely separated into single electron
components and therefore cannot be solved analytically. However, it is
possible to separate the single electron contributions within certain sets of
approximations.

3.2 Approximate solutions to the Schrödinger equa-

tion for N-electron atoms

In the following, two approaches to obtaining approximate solutions to the

Schrödinger equation for N-electron atoms are considered. These are the
independent-particle model and the central-field approximation.

3.2.1 Independent-particle model

In the independent-particle model of a many electron atom, the electron-

electron repulsion is first ignored to obtain a very crude approximation for
the energies of the eigenstates. The neglected interactions are then added
to refine the result.
Without the electron-electron interaction term, the Hamiltonian for a
N-electron atom (Equation 3.4) takes the form
X
Ĥ(~r1 , ~r2 , ~r3 , . . . ~rN ) = ĥ(~ri ). (3.5)
i

34
This operator is separable because the term associated with each electron,
labelled with the index i, is dependent only on the position, ~ri , of that elec-
tron. The complete Hamiltonian is therefore the sum over N single-electron
Hamiltonians, ĥ(~ri ). Following from this each single-electron Hamiltonian
will have corresponding eigenfunctions ψ(~ri ), and eigenvalues Ei such that

ĥ(~ri ) ψ(~ri ) = Ei ψ(~ri ), (3.6)

where, as in the one-electron atom, ψ(~ri ) = Rni ,ℓi (ri ) Yℓi ,mℓi (θi , φi ) χsi ,msi is
separable into radial, angular and spin components characterised by the
single-electron principal quantum number ni , orbital angular momentum
quantum number ℓi , azimuthal quantum number mℓi , spin quantum num-
ber si and the projection of the spin msi .
The total electronic wavefunction, Ψ(~r1 , ~r2 , ~r3 , . . . ~rN ), for the N-electron
system is then the product of the N single-electron wavefunctions,

Ψ(~r1 , ~r2 , ~r3 , . . . ~rN ) = ψ(~r1 ) ψ(~r2 ) ψ(~r3 ) × · · · × ψ(~rN ), (3.7)

and the total energy, E, is the sum over the energies of the N single electrons,

E = E1 + E2 + E3 + · · · + EN , (3.8)

where each Ei is given by the Rydberg formula (Equation 2.81) for a one-
electron atom.
To test the accuracy of this independent-particle model we use it to
calculate the ground-state energy of the two-electron He atom, in which the
two electrons are in the 1s orbital. The total energy of the system is therefore

Egnd (He) = En=1 + En=1 (3.9)

Egnd (He) µHe+ µHe+

= −R∞ Z2nucl − R∞ Z2nucl , (3.10)
hc me me
where Znucl = +2 and
me mHe2+
µHe+ = . (3.11)
me + mHe2+
Therefore,
Egnd (He)
= −877 778 cm−1 (3.12)
hc
≃ −8 R∞ . (3.13)

35
However, the true ground state energy of the He atom determined from
experimental measurements is
Egnd (He) Eion (He) Eion (He+ )
= − − , (3.14)
hc hc hc

= −198 311 cm−1 − 438 909 cm−1

= −637220 cm−1 (3.15)
≃ −5.8 R∞ , (3.16)

where Eion (He) is the ground state ionisation energy of the neutral He atom,
and Eion (He+ ) is the ground state ionisation energy of the singly-charged
He+ ion. Therefore, as might be expected, neglecting the effects of electron-
electron repulsion when calculating the energy of the ground state of the
He atom results in the electrons being too tightly bound.
To crudely correct for the electron-electron repulsion we can consider
the typical separation between the two electrons in the 1s orbital of the He
atom and approximate the repulsion energy. Since the most probable radial
position of each individual electron in the 1s orbital is

rprob (He+ 1s) = aM (3.17)

a0
≃ (3.18)
Znucl
a0
= , (3.19)
2
where a0 is the Bohr radius, and because of the Coulomb repulsion between
them we may expect to find them typically located on opposite sides of the
nucleus, the most probably spatial separation between the two electrons is
r12 ≃ a0 . Therefore the typical electron-electron repulsion energy is

Erep (He) = Vrep (He) (3.20)

e2
≃ (3.21)
4πǫ0 a0
and
Erep (He)
= 219 475 cm−1 (3.22)
hc
≃ 2 R∞ . (3.23)

36
Therefore adding this to the value obtained when the electron-electron
repulsion is neglected (Equation 3.13), leads to the result that
Egnd (He) e2 1
= −877 778 cm−1 + (3.24)
hc 4πǫ0 r12 hc

= −877 778 cm−1 + 219 475 cm−1 (3.25)

≃ −6 R∞ (3.26)

This is closer to the measured value of ∼ −5.8 R∞ .

3.2.2 Central-field approximation

A more refined approach to the calculation of the electronic structure

of atoms with many-electrons involves the introduction of a position-
dependent correction to the electron-nucleus interaction for each individual
electron. This approach is known as the central-field approximation, and
allows the electron-electron interactions to be accounted for in a way that
the motion of the individual electrons can still be separated.
To achieve this, we consider a single electron Hamiltonian, ĥ(~ri ), as
above, to which we add a central potential Vc (ri ) that represents the average
interaction of the electron labelled with the index i, with all of the other
electrons. Therefore the complete N-electron Hamiltonian becomes
X
Ĥ(~r1 , ~r2 , ~r3 , . . . ~rN ) = ĥ′ (~ri ), (3.27)
i

where

ĥ′ (~ri ) = ĥ(~ri ) + Vc (ri ) (3.28)

~2 2 Znucl e2
= − ∇ − + Vc (ri ), (3.29)
2me 4πǫ0 ri
and
* +
e2
Vc (ri ) = . (3.30)
4πǫ0 ri j j,i

Therefore the central potential, Vc (ri ), depends only on ri j = |~ri j |, the magni-
tude of ~ri j , and not on its direction. As a result this potential has no angular
dependence.

37
 To determine the electronic structure within this central-field approxi-
mation it is therefore necessary to solve N one-electron Schrödinger equa-
tions of the form

ĥ′ (~ri ) ψ(~ri ) = Ei ψ(~ri ), (3.31)

where ψ(~ri ) are one-electron orbitals. The total electronic energy, Etot , is
then
X
Etot = Ei , (3.32)
i

and, as in the independent-particle model, the total electronic wavefunction,

Ψ(~r1 , ~r2 , ~r3 , . . . ~rN ), has a product form,

Ψ(~r1 , ~r2 , ~r3 , . . . ~rN ) = ψ(~r1 ) ψ(~r2 ) ψ(~r3 ) × · · · × ψ(~rN ). (3.33)

The one-electron orbitals obtained using this central-field approximation

are separable into radial, Fni ,ℓi (ri ), and angular, Yℓi ,mi (θi , φi ), parts, i.e.,

ψ(~ri ) = Fni ,ℓi (ri ) Yℓi ,mi (θi , φi ). (3.34)

As in a one-electron atom, the angular functions are the set of spherical

harmonic functions, however, because of the deviation of the potential from
a pure Coulomb potential following the introduction of the central-field, the
radial functions differ from those obtained in the case of the one-electron
atom.
Because the central potential, Vc (r), modifies the Coulomb potential at
short range, i.e., for small values of r, the ℓ-degeneracy associated with
eigenstates with equal values of n that was seen in the case of a one-electron
atom, is lifted. However, because the central potential depends only on the
magnitude of ~ri j and not its direction, the m-degeneracy of all states with
the equal values of n and ℓ remains. Therefore

Ei ≡ Eni ,ℓi . (3.35)

The above description of the electronic wavefunctions of many-electron

atoms in terms of products of one-electron orbitals leads to the notation in
which the occupation of each orbital in the many-electron atom is listed,
e.g., in the case of the C atom, in the form

38
 Figure 3.1: Periodic table of the elements, NIST1 .

C atom ground state: 1s2 2s2 2p2 .

In this notation the number of electrons in each orbital is indicated by the

right superscript. When listed in this way, the set of occupied orbitals is
called the electronic configuration.
Applicability of the central-field approximation

• The central-field approximation is most applicable to atoms or atomic

ions with a single electron occupying an outer orbital. These include:

The alkali metal atoms in Group I of the periodic table, e.g., Li, Na, K,
Rb,. . . (see Figure 3.1). These atoms have one single electron in
an outer s-orbital:
Li: 1s2 2s1
1
R. A. Dragoset, A. Musgrove, C. W. Clark, and W. C. Martin, The National In-
stitute of Standards and Technology (NIST), Physical Measurement Laboratory, 2010
(http://www.nist.gov/pml/data/periodic.cfm).

39
 Na: 1s2 2s2 2p6 3s1
K: 1s2 2s2 2p6 3s2 3p6 4s1
Singly-ionised alkaline earth metals in Group II of the periodic table,
e.g., Be+ , Mg+ , Ca+ ,. . . (see Figure 3.1). These atomic ions have
one single electron in an outer s-orbital:
Be+ : 1s2 2s1
Mg+ : 1s2 2s2 2p6 3s1

• Other atoms with one electron in a highly excited state, i.e., a state
with a high value of n. Atoms, such as these, with one electron in a
high-n state are often called Rydberg atoms.

In each of the these cases, N − 1 electrons in the many-electron atom are

tightly bound to the nucleus while one single electron is loosely bound. The
central-field approximation is most appropriate for calculating the eigen-
functions and eigenvalues of the this one loosely bound electron.
In these cases the inner electrons as said to screen the charge of the nu-
cleus such that

As ri → 0:

Znucl e2 Znucl e2
− + Vc (ri ) −→ − (3.36)
4πǫ0 ri 4πǫ0 ri

As ri → ∞:

Znucl e2 e2
− + Vc (ri ) −→ − (3.37)
4πǫ0 ri 4πǫ0 ri
Therefore when the one loosely-bound electron approaches ri = 0, the
Coulomb attraction of the positively charge nucleus dominates, while at
larger distances the nucleus and core electrons appear, from the perspective
of this outer electron, as a singly charged object with an effective charge of
Znucl = +1.

40
 H+ X+

continuum

continuum
Ionisation

Ionisation
Energy / hc (cm-1)

0 Eion Eion

n n

-500 15 15
14 14
13 13
12 12
11 11
-1000
10 10

l = 0.. n-1 l: 0 1 2 3 4 ≥5

-1500 H atom Other atoms

Figure 3.2: Series of eigenstates of the H atom in which the energies are
only dependent on the value of n and independent of the value of ℓ, and in
other non-hydrogenic many-electron atoms for which the energies depend
on n and ℓ.

As can be seen from the radial probability density distributions associ-

ated with the one-electron orbitals in the H atom displayed in Figure 2.9,
the proximity of the electron to the position of the screened charge of the
nucleus depends on the value of n and the value of ℓ. Because an electron in
a low-ℓ orbital is more core-penetrating than one in a high-ℓ orbital with the
same value of n, the effective charge of the nucleus and core electrons appear
greater. On the other hand, electrons in high-ℓ orbitals appear bound to an
ion core with an effective charge of Zcore = +1 and their binding energies
are similar to those of a one electron atom (see Figure 3.2).

41
Therefore:

H atom: E1s < E2s = E2p < E3s = E3p = E3d . . .

He atom: E1s < E2s < E2p < E3s < E3p < E3d . . .

3.2.3 Quantum defects

As discussed above, the central-field approximation is well suited to the

calculation of the eigenfunctions and eigenvalues of many-electron atoms
with a single (excited) outer electron. Such an electron is often called
an optically active electron since transitions involving it tend to lie in the
optical/visible regions of the electromagnetic spectrum.
Within the central-field approximation, the eigenenergies of a single
active electron can be expressed as
En,ℓ RM
= − , (3.38)
hc (n − ∆nℓ )2
where

∆nℓ : is known as the quantum defect

(n − ∆nℓ ): is often described as the effective principal quantum number

and as in Section 2 the Rydberg constant

µM
RM = R∞ Z2core , (3.39)
me
and in this case

Zcore : is the effective combined charge of the nucleus and core elec-
trons

This expression (Equation 3.38) for the energy of the single active electron
can be derived rigorously by solving the Schrödinger equation for a single
electron in a central potential that deviates from its dependence on 1/r at
small r.
The value of ∆nℓ , the quantum defect in Equation 3.38, depends on how
strongly the single active electron interacts with the non-hydrogenic ion
core.

42
 For high values of ℓ (typically ℓ ≥ 5): ∆nℓ → 0

This is because the centrifugal barrier associated with the high angular
momentum forces the electron away from the core and it therefore
does not penetrate the cloud of inner electrons.

For low values of ℓ (typically ℓ < 5): ∆nℓ > 0

This is because in lower-ℓ states the active electron penetrates the

cloud of inner electrons more significantly.

As can be seen in Figure 3.2 the values of ∆nℓ are strongly dependent
on ℓ and weakly dependent on the value of n. For example, in the case of a
single active electron in low-n states of the helium atom the corresponding
quantum defects are listed in the table below.

∆nℓ ℓ=0 ℓ=1 ℓ=2 ℓ=3

n=2 0.310 0.066 - -
n=3 0.302 0.066 0.002 -
n=4 0.299 0.066 0.002 0.0003

Table 3.1: Quantum defects for the low-n triplet states of He.

3.3 Indistinguishable particles

In a quantum mechanical treatment of many-electron atoms, it is necessary

to consider the effects that the indistinguishability of electrons has on the
total electronic wavefunction.
Consider two particles that are indistinguishable by any measurement,
e.g., two electrons, two photons, . . . . If the wavefunction of the two particle
system is ψ(1, 2), then

|ψ(1, 2)|2 = |ψ(2, 1)|2 . (3.40)

This condition can be satisfied in two ways. If

ψ(1, 2) = +ψ(2, 1), (3.41)

43
the wavefunction is said to be symmetric with respect to exchange of the two
particles, while if

ψ(1, 2) = −ψ(2, 1), (3.42)

the wavefunction is said to be antisymmetric with respect to exchange.

More generally any physically allowed wavefunction of a two-particle
system must be either symmetric or antisymmetric with respect to exchange
of the two particles.

3.3.1 The Pauli principle

To explain the way in which the one-electron orbitals are filled in a many-
electron atom, Wolfgang Pauli postulated that

• Wavefunctions must be antisymmetric with respect to exchange of

identical Fermions (particles with 1/2 integer spins, e.g., e− , p+ , n,
...)

• Wavefunctions must be symmetric with respect to exchange of identical

Bosons (particles with integer spins, e.g., photon, α-particle, 4 He atom,
...)

For example, for two electrons (two Fermions),

ψ(1, 2) = −ψ(2, 1). (3.43)

Within the orbital approximation, if φA and φB are two one-electron orbitals,

e.g., the 1s and 2p orbitals, then

ψ(1, 2) = φA (1) φB (2), (3.44)

and

ψ(2, 1) = φA (2) φB (1), (3.45)

where the numbers in brackets are arbitrary labels assigned to the two
electrons, e.g., electron number 1 and electron number 2. Therefore, from
Pauli’s postulate for two Fermions

φA (1) φB (2) = −φA (2) φB (1), (3.46)

44
which can only be true of

φ A , φB . (3.47)

Because the wavefunction φA cannot be the same as the wavefunction φB it

must be concluded that no to electrons can occupy the same (spin-) orbital
in a many-electron atom. For example, two electrons cannot occupy the 1s
orbital unless they have opposite spins.
When normalised, a generalised two-electron wavefunction that fulfills
the principle postulated by Pauli takes the form

1 h i
ψ(1, 2) = √ φA (1) φB (2) − φA (2) φB (1) (3.48)
2

= −ψ(2, 1). (3.49)

3.4 Effect of the Pauli principle on atomic structure

Consider the case of the He atom with two electrons (indistinguishable

Fermions). The total wavefunction has the form

Ψ(1, 2) = Φ(~r1 , ~r2 ) χ(σ1 , σ2 ) = −Φ(~r2 , ~r1 ) χ(σ2 , σ1 ), (3.50)

| {z } | {z }
Space Spin

where σi represent the spin coordinates.

~ˆ 2 and Ŝz , where
The function χ is an eigenfunction of the the operators S

~ˆ = ~sˆ1 + ~sˆ2 ,
S (3.51)

and

Ŝz = ŝz1 + ŝz2 , (3.52)

with ~sˆ1 , ~sˆ2 , ~sˆz1 and ~sˆz2 the spin operators associated with each the individual
electrons, 1 and 2.

45
3.4.1 Spin wavefunctions

For an individual electron the spin quantum number, s, is

1
s = . (3.53)
2
This single electron spin can be aligned along the positive z-axis, in which
case it is said to be spin-up and is denoted by the one-electron spin wave-
function

α or ↑ (3.54)

and the azimuthal quantum number

1
ms = + , (3.55)
2
or can be aligned along the negative z-axis, in which case it is said to be
spin-down and is denoted by the one-electron spin wavefunction

β or ↓ (3.56)

and the azimuthal quantum number

1
ms = − , (3.57)
2
i.e.,
1
Ŝz α = + ~ α (3.58)
2

1
Ŝz β = − ~ β. (3.59)
2
~ is the sum of the two
For a two-electron system the total spin vector, S,
individual electron spin vectors, ~s1 and ~s2 ,

~ = ~s1 + ~s2 .
S (3.60)

~ can then take values in the range

The total spin quantum number S = |S|

S = |s1 − s2 |, . . . , |s1 + s2 | (3.61)

46
in steps of 1, where si = |~si |. The total electron spin of the atom is therefore
quantised.
Since for the two electrons in the He atom s1 = s2 = 1/2,
   
 1 1   1 1 
S =  − ,..., +  (3.62)
2 2 2 2

= 0, . . . , 1 (3.63)
= 0 and 1. (3.64)

The spin quantum number S = 0 corresponds to the situation in which

the two electron spins are aligned antiparallel to each other, while S = 1
corresponds to the case in which the two electrons spins are aligned parallel
to each other.
In general the total electron spin in a many-electron atom is denoted by
~ while the individual electron spins are denoted by
the uppercase letter S,
the lowercase letter ~s, i.e.,
X
~ =
S ~si . (3.65)
i

The total spin quantum number and the projection of the total spin onto the
z-axis are denoted by the upper case letters S and MS , respectively.

Considering separately the cases in which S = 0 and S = 1.

S = 1: When S = 1, MS = −1, 0, +1.

Because there are three values of MS , the S = 1 state is known as a

triplet spin state.

MS = +1: The triplet state in which the spins of the two electrons are
both aligned parallel to the z-axis is described by the wavefunction

χT = α(1)α(2). (3.66)

MS = −1: The triplet state in which the spins of the two electrons are
both aligned antiparallel to the z-axis is described by the wavefunction

χT = β(1)β(2). (3.67)

47
 MS = 0: The two wavefunctions above are both symmetric with re-
spect to exchange of the two electrons. Therefore the wavefunction
associated with the final triplet state, for which MS = 0, must also be
symmetric with respect to exchange. This wavefunction is therefore
the symmetric combination of electron 1 with spin-up and electron 2
with spin-down, and electron 2 with spin-up and electron 1 with
spin-down, i.e.,

1  
χT = √ α(1)β(2) + α(2)β(1) , (3.68)
2
where the prefactor ensures that the two-particle wavefunction is
normalised.

S = 0: When S = 0, MS = 0.

Because there is only one single value of MS , the S = 0 state is known

as a singlet spin state.

MS = 0: As above the singlet spin state with MS = 0, is a combina-

tion of electron 1 with spin-up and electron 2 with spin-down, and
electron 2 with spin-up and electron 1 with spin-down. Since the
symmetric form of this function must correspond to the triplet state,
it is the antisymmetric form that corresponds to the singlet state, i.e.,

1  
χS = √ α(1)β(2) − α(2)β(1) . (3.69)
2

48
3.4.2 He atom wavefunctions

Since the spatial part of the electronic wavefunction of the He atom can be
expressed in a general form as
1 h i
Φ± (~r1 , ~r2 ) = √ φnℓmℓ (~r1 )φn′ ℓ′ m′ℓ (~r2 ) ± φnℓmℓ (~r2 )φn′ ℓ′ m′ℓ (~r1 ) , (3.70)
2
where under exchange of the two electrons

Φ+ (~r1 , ~r2 ) ⇒ symmetric

Φ− (~r1 , ~r2 ) ⇒ antisymmetric

to satisfy the Pauli principle, that the total wavefunction of the two-electron
system must be antisymmetric with respect to exchange of the two electrons,
the symmetric spatial wavefunction can only occur when the spins form a
singlet state, while the antisymmetric spatial wavefunction must only occur
when the spins form a triplet state.

S Spin Space Wavefunction

S = 0 ⇒ Singlet antisymmetric symmetric Φ+ (~r1 , ~r2 ) χS
S = 1 ⇒ Triplet symmetric antisymmetric Φ− (~r1 , ~r2 ) χT

Ground state of He

1s2 : Two electrons in the 1s orbital.

The spatial part of the wavefunction for the ground state of the He atom
can have the form
1 h i
Φ+ (~r1 , ~r2 ) = √ φ1s (~r1 )φ1s (~r2 ) + φ1s (~r2 )φ1s (~r1 ) , (3.71)
2
or
1 h i
Φ− (~r1 , ~r2 ) = √ φ1s (~r1 )φ1s (~r2 ) − φ1s (~r2 )φ1s (~r1 )
2

= 0. (3.72)

Because Φ− (~r1 , ~r2 ) = 0, the ground state of the He atom can only be de-
scribed by the Φ+ (~r1 , ~r2 ) spatial wavefunction which is symmetric with
respect to exchange of the two electrons.

49
 With a symmetric spatial component of the ground-state wavefunction,
to satisfy the Pauli principle the spin part must be antisymmetric with
respect to exchange of the two electrons, i.e., χS .
Therefore the wavefunction for the ground state of the He atom must
be:

He 1s2 : Ψ(1, 2) = Φ+ (~r1 , ~r2 ) χS . (3.73)

Excited states of the He atom

In excited states of the He atom, e.g., in the 1s2s configuration, the spatial
wavefunctions
1 h i
Φ+ (~r1 , ~r2 ) = √ φ1s (~r1 )φ2s (~r2 ) + φ1s (~r2 )φ2s (~r1 ) , (3.74)
2
and
1 h i
Φ− (~r1 , ~r2 ) = √ φ1s (~r1 )φ2s (~r2 ) − φ1s (~r2 )φ2s (~r1 ) (3.75)
2
are both non-zero. Therefore excited states in which the two electrons are
in distinct orbitals can be either singlet or triplet states.

3.4.3 Spin multiplicity

As seen above, for spin states with S = 1, there are three possible values
of MS : −1, 0 and +1. From this it can be concluded that for any value of S
there are 2S + 1 associated values of MS .
This value of the quantity 2S + 1 is known as the spin multiplicity of the
state. For S = 1, 2S + 1 = 3 leading to such states being denoted triplet states.
This nomenclature is also extended to all other values of 2S + 1, i.e.,

For S = 0: MS = 0
2S + 1 = 1 ⇒ singlet state

For S = 1/2: MS = −1/2, +1/2

2S + 1 = 2 ⇒ doublet state

For S = 3/2: MS = −3/2, −1/2, +1/2, +3/2

2S + 1 = 4 ⇒ quartet state

50
3.5 Exchange

Consider a triplet state of the He atom. In this state S = 1 and the spin part
of the wavefunction is χT , where

α(1)α(2)
 
χT ⇒ √1 α(1)β(2) + α(2)β(1)
2

β(1)β(2)

The spatial parts of the triplet wavefunctions are antisymmetric with

respect to exchange of the two electrons
1 h i
Φ(~r1 , ~r2 ) = Φ− (~r1 , ~r2 ) = √ φa (~r1 )φb (~r2 ) − φa (~r2 )φb (~r1 ) . (3.76)
2
However, if the two electrons are at the same position, i.e., ~r1 = ~r2 , the
expression in Equation 3.76 is equal to zero. Therefore in the triplet states of
He the two electrons tend to avoid each other. As a result they do not screen
the Znucl = +2 charge of the nucleus for each other as much as they would if
they overlapped spatially. They are therefore more strongly bound (lower
in energy) than would be expected if the symmetry restrictions imposed by
the Pauli principle were not considered.
On the other hand, in the singlet states of He, S = 0 and the spin part of
the wavefunction is χS , where
 
χS ⇒ √1 α(1)β(2) − α(2)β(1)
2

In this case the spatial parts of the wavefunctions are symmetric with
respect to exchange of the two electrons
1 h i
Φ(~r1 , ~r2 ) = Φ+ (~r1 , ~r2 ) = √ φa (~r1 )φb (~r2 ) + φa (~r2 )φb (~r1 ) (3.77)
2

, 0 if ~r1 = ~r2 . (3.78)

As a result there is a non-zero probability of finding the two electrons close

to each other where they can each screen the other from the Zcore = +2
charge of the nucleus. Consequently, the singlet states are more weakly
bound (higher in energy) than the triplet states as can be seen in Figure 3.3

51
 1s2s (S = 0)

-50000 1s2s (S = 1)
∆E -1
hc = 6421 cm
∆E
e = 0.8 eV
Energy / hc (cm-1)

-100000

-150000

-200000 1s2 (S = 0)

Figure 3.3: Energy splitting between the 1s2s singlet and triplet states of
He that results from the Exchange Interaction.

The energy shifts that arise as a result of considering the effects of

the Pauli principle on the electronic structure of an atom suggest that an
interaction occurs between the electrons within the atom, the sign of which
depends on the relative orientations of the their spins. This interaction is
known as the Exchange Interaction and is a direct result of the requirement of
the Pauli principle that the total wavefunction is antisymmteric with respect
to exchange of identical Fermions.

3.5.1 Formal analysis of the Exchange Interaction

To treat the Exchange Interaction in a more formal way we must consider

the effects of the symmetry of the total wavefunctions on the repulsion
between the two electrons. The repulsive interaction, Vrep , has the form

e2 1
Vrep = =C , (3.79)
4πǫ0 r12 r12

52
where C is a constant. The electron-electron repulsion therefore depends
directly on the expectation value of r−1
12
,
  ZZ
1 1
= ψ∗ (1, 2) ψ(1, 2) dV dσ, (3.80)
r12 r12

where ~r1 and ~r2 are denoted by the the numbers 1 and 2, and dV and dσ
represent the sets of spatial coordinates and spin coordinates, respectively.

Considering first the singlet states, the total wavefunctions take the form

Ψ(1, 2) = Φ+ χS , (3.81)

and therefore

  ZZ
1 1
= Φ∗+ χS ∗ Φ+ χS dV dσ, (3.82)
r12 r12
ZZ
1 h i∗
= √ φa (1)φb (2) + φa (2)φb (1) × . . .
2

1 
× √ α(1)β(2) − α(2)β(1) ∗ × . . .

2

1
× × ...
r12

1 h i
× √ φa (1)φb (2) + φa (2)φb (1) × . . .
2

1  
× √ α(1)β(2) − α(2)β(1) dV dσ. (3.83)
2

Because the operator r−1

12
only acts on the spatial parts of the wavefunctions
and not the spin parts, the space and spin parts of Equation 3.83 can be

53
separated, such that,
  Z h
1 1 i∗
= φa (1)φb (2) + φa (2)φb (1) × . . .
r12 2

1 h i
× φa (1)φb (2) + φa (2)φb (1) dV × . . .
r12
Z
1
α(1)β(2) − α(2)β(1) ∗ × . . .
 
×
2

 
× α(1)β(2) − α(2)β(1) dσ. (3.84)

Since the spin wavefunction is normalised, i.e.,

Z
χS ∗ χS dσ = 1, (3.85)

expanding the spatial parts of Equation 3.91 leads to the result that

 h
Z φ (1)φ (2) ∗ φ (1)φ (2)
i h i
  

1 1
 a b a b
=  dV + . . .
r12 2

 r12

Z φ (1)φ (2) ∗ φ (2)φ (1)

h i h i
a b a b
+ dV + . . .
r12
h i∗ h i
Z φa (2)φb (1) φa (1)φb (2)
+ dV + . . .
r12
h i∗ h i 
Z φa (2)φb (1) φa (2)φb (1) 



+ dV  (3.86)
r12 



54
and therefore
(Z

1

1 |φa (1)|2 |φb (2)|2
= dV + . . .
r12 2 r12

φ∗a (1)φ∗b (2)φa (2)φb (1)

Z
+ dV + . . .
r12

φ∗a (2)φ∗b (1)φa (1)φb (2)

Z
+ dV + . . .
r12
Z )
|φa (2)|2 |φb (1)|2
+ dV (3.87)
r12

φ∗a (1)φ∗b (2)φa (2)φb (1)

Z Z
|φa (1)|2 |φb (2)|2
= dV + dV (3.88)
r12 r12

= J + K (3.89)

where J is the Coulomb Integral, and K is the Exchange Integral. From this
expression it can be seen that the Exchange Interaction raises the expectation
value associated with the electron-electron repulsion in the singlet states. It
therefore increases the energy of the singlet states.
The Coulomb Integral, J, represents the repulsion between between the
two electron clouds. The Exchange Integral, K, corresponds to the energy
shift arising from the restrictions imposed on the wavefunctions by the Pauli
principle. Because of the positive signs of the two functions in Equation 3.88,
J and K are both repulsive electron-electron interactions for the singlet states
and increase their energies.
A similar treatment can be carried out for each of the triplet states. In
this case, considering for example the spin wavefunction

χT = α(1)α(2), (3.90)

55
then
  Z h
1 1 i∗
= φa (1)φb (2) − φa (2)φb (1) × . . .
r12 2

1 h i
× φa (1)φb (2) − φa (2)φb (1) dV × . . .
r12
Z
1
× [α(1)α(2)]∗ [α(1)α(2)] dσ. (3.91)
2
Proceeding as above, with the normalised spin wavefunction leads to the
result that
φ∗a (~r1 )φ∗b (~r2 )φa (~r2 )φb (~r1 )
Z Z

1
 |φa (~r1 )|2 |φb (~r2 )|2
= dV − dV
r12 r12 r12

= J − K

where J is the Coulomb Integral, and K is the Exchange Integral. From this
expression it can be seen that for the triplet states the Exchange Interaction
reduces the expectation value associated with the electron-electron repul-
sion, and therefore lowers the energy of these states.

56
3.6 Electronic configurations

The electronic configuration of a many-electron atom specifies the occupancy

of the atomic orbitals. For example:

Atomic number Symbol Electronic configuration

1 H 1s1
2 He 1s2
3 Li 1s2 2s1
4 Be 1s2 2s2
5 B 1s2 2s2 2p1
6 C 1s2 2s2 2p2
... ... ...

In each configuration,

• n; the principal quantum number is give by an integer number

• ℓ; the orbital angular momentum quantum number is give by a letter;

s, p, d, f,. . .

• The number of electrons occupying each orbital is indicated by the

superscripts

Since there are 2ℓ + 1 values of mℓ for each value of ℓ, and two values of ms ,
following from the Pauli principle each nℓ orbital can hold up to 2(2ℓ + 1)
electrons.

3.6.1 Nomenclature

The following nomenclature is used to describe an electronic configuration.

• Electrons with the equal values of n are said to be in the same shell

• Electrons with the equal values of n and ℓ are said to be in the same
sub-shell (or orbital)

• Electrons with equal values of n and ℓ are said to be equivalent

Pn−1
• Each shell can hold up to 2n2 electrons, i.e., 2 ℓ=0 (2ℓ + 1) electrons

57
 • Shells containing 2n2 electrons are said to be closed (or full, or complete)

• Shells containing < 2n2 electrons are said to be open (or incomplete, or
incompletely filled)

• Electrons in open shells are said to be optically active

For example, a sub-shell with ℓ = 0 can contain up to two electrons with

the quantum numbers:

ℓ = 0, mℓ = 0, and ms = ±1/2 .
|{z}
×2

On the other hand a sub-shell with ℓ = 2 can contain up to 10 electrons with

the quantum numbers:

ℓ = 2, mℓ = −2, −1, 0, +1, +2, and ms = ±1/2 .

| {z } |{z}
×5 ×2

3.6.2 Periodic table of the elements

In the periodic table of the elements, elements with the same valence struc-
ture are listed in the columns known as groups, while across each row, called
a period, the number of electrons increases by one at a time.
The chemical properties of the elements are governed by their outer
valence electrons therefore the elements in each group have similar charac-
teristics.

Group I: The group I elements are known as the alkali metals and in their
ground states all have one electron in an outer ns sub-shell (see table
below). This optically active outer electron is easily removed to form
a positively charged ion (cation), or in the formation of a molecule
where it is donated to the molecular bond.

Element Electronic configuration

H 1s1
Li 1s2 2s1
Na 1s2 2s2 2p6 3s1
K 1s2 2s2 2p6 3s2 3p6 4s1
... ...

58
Group VII: The group VII elements are known as the halogens and in their
ground states are missing one electron from their outer np sub-shell
(see table below). This incomplete sub-shell makes them extremely
chemically reactive and they easily gain an electron to form a nega-
tively charged ion (anion), or in the formation of a molecule where
they readily accept an electron to form the molecular bond.

Element Electronic configuration

F 1s2 2s2 2p5
Cl 1s2 2s2 2p6 3s2 3p5
Br 1s2 2s2 2p6 3s2 3p6 4s2 3d10 4p5
... ...

Group VIII: The group VIII elements are known as the nobel gases and in
their ground states contain fully closed shells (see table below). These
closed shells make them chemically inert.

Element Electronic configuration

He 1s2
Ne 1s2 2s2 2p6
Ar 1s2 2s2 2p6 3s2 3p6
Kr 1s2 2s2 2p6 3s2 3p6 4s2 3d10 4p6
... ...

3.6.3 Terms

Beyond the central-field approximation, which is valid for atoms with one
isolated outer electron, the complete many-electron Hamiltonian, Ĥ, does
not commute with the angular momentum operators associated with the
individual electrons, i.e.,
 
ˆ
Ĥ, ℓ~i , 0. (3.92)

This indicates that ℓi is not a good quantum number in the description of the
many-electron structure. Instead of considering the angular momentum of
the individual electrons separately, it is instead necessary to consider the

59
combined total orbital angular momentum, ~L, of all of the electrons together.
ˆ
In a many-electron system Ĥ does commute with ~L2 , and the square of the
~ˆ 2 , i.e.,
total electron spin operator, S
 
ˆ
Ĥ, ~L2 = 0 (3.93)

 
~ˆ 2
Ĥ, S = 0. (3.94)

To account for these effects in the description of the atomic structure,

electronic configurations are separated into terms that are labelled with the
total orbital angular momentum quantum number, L, and the total electron
spin quantum number, S.
For each configuration the total orbital angular momentum
X
~Lˆ = ˆ
ℓ~i , (3.95)
i

and the total spin

X
~ˆ =
S ~sˆi . (3.96)
i

The values of L and S are therefore determined by adding the angular

momenta of the individual electrons, rejecting combinations of L and S
forbidden by the Pauli principle.
The symbol describing the value of L and S for a given configuration
in known as a term symbol. In this nomenclature the total orbital angular
momentum quantum number is represented by a capital letter, i.e,

L = 0 1 2 3 4 ...
S P D F G ...

The term symbol is then

2S+1
L (3.97)

60
For example,

The ground state of Fe+ has the term symbol 6 D:

Therefore L = 2 and 2S + 1 = 6 ⇒ S = 5/2

The ground state of the He atom has the term symbol 1 S:

Therefore L = 0 and 2S + 1 = 1 ⇒ S = 0

An excited state of the He atom has the term symbol 3 S:

Therefore L = 0 and 2S + 1 = 3 ⇒ S = 1

The ground of the H atom has the term symbol 2 S:

Therefore so L = 0 and 2S + 1 = 2 ⇒ S = 1/2

In determining the allowed terms for a given electronic configuration

the sum over the orbital angular momenta and spins of the electrons in
closed sub-shells is zero, i.e., for a closed sub-shell
X
~Lˆ = ˆ
ℓ~i = 0 (3.98)
i

and
X
~ˆ =
S ~sˆi = 0. (3.99)
i

Therefore for a given configuration the possible values of L and S are deter-
mined by considering the electrons in open sub-shells.

3.6.4 Terms for two-electron atoms

Non-equivalent electrons

Consider two non-equivalent electrons, e.g., one in the orbital nℓ and the
other in the orbital n′ ℓ where n , n′ . Because these electrons are in different
spin-orbitals the Pauli principle will be satisfied for all possible terms.
ˆ
For these two electrons their total orbital angular momentum, ~L, is

~Lˆ = ℓ~ˆ1 + ℓ~ˆ2 (3.100)

61
therefore the possible values of the total orbital angular momentum quan-
tum number, L, are

L = |ℓ1 − ℓ2 |, . . . , |ℓ1 + ℓ2 | in steps of 1 (3.101)

~ˆ is
In a similar way, the total spin of the two electrons, S,

~ˆ = ~sˆ1 + ~sˆ2
S (3.102)

and as a result the possible values of the total spin quantum number, S, are

S = |s1 − s2 |, . . . , |s1 + s2 | in steps of 1 (3.103)

As an example, consider the configuration 2p3p. Here n , n′ so the

Pauli principle will be satisfied. For this configuration

ℓ1 = 1
s1 = 1/2
ℓ2 = 1
s2 = 1/2

and therefore

L = |ℓ1 − ℓ2 |, . . . , |ℓ1 + ℓ2 |
= 0, 1, and 2
≡ S, P, and D

and

S = |s1 − s2 |, . . . , |s1 + s2 |
= 0 and 1

so (2S + 1) = 1 or 3 and the possible allowed terms are

1S 1P 1D

3S 3P 3D

The number of states associated with each term, i.e., combinations of

ML and MS , is (2S + 1)(2L + 1) and so here the total number of electronic

62
states associated with this set of six terms is 36. For the configuration 2p3p
each p-orbital has six states, i.e., three possible values of mℓ combined with
two possible values of ms . Therefore with the two electrons in p-orbitals the
total number of states is 6 × 6 = 36, in agreement with that given by the set
of terms.

Equivalent electrons

Consider two equivalent electrons, i.e., electrons with the same values of n
and ℓ. From the Pauli principle if n and ℓ are the same for both electrons,
they must have different values of mℓ or ms .

Example 1: In an ns2 configuration in He,

Both electrons have the same value of n

Both electrons have the same value of ℓ, i.e., ℓ1 = ℓ2 = 0

Both electrons have the same value of mℓ , i.e., mℓ1 = mℓ2 = 0

Therefore the two electrons must have opposite spin, i.e., ms1 = +1/2
and ms2 = −1/2. Because s1 = s2 = 1/2, for this configuration S = 0,
and therefore the corresponding term is 1 S.

Example 2: In an np2 configuration in He (e.g., 2p2 ),

ℓ1 = 1
s1 = 1/2
ℓ2 = 1
s2 = 1/2

In this case the possible values of L are

L = |ℓ1 − ℓ2 |, . . . , |ℓ1 + ℓ2 |
= 0, 1 and 2
= S, P and D

63
and the possible values of S are

S = |s1 − s2 |, . . . , |s1 + s2 |
= 0 and 1

To determine which combinations of L and S are allowed by the Pauli

principle we must consider the form of the total electronic wavefunc-
tion of the atom

Ψ = Φ(~r1 , ~r2 ) χ(σ1 , σ2 ) . (3.104)

| {z } | {z }
Space Spin

Since this total wavefunction must be antisymmetric with respect to

exchange of the two electrons, when the space part is symmetric the
spin part must be antisymmetric and vice-versa.

The symmetry of the space part of the total wavefunction is deter-

mined by the symmetry (parity) of the spherical harmonic functions,
YL,ML (θ, φ) and given by (−1)L . If (−1)L = +1, the function is symmetric
while if (−1)L = −1 the function is antisymmetric.

For the three possible values of L above we can then conclude that

L = 0 ⇒ symmetric
L = 1 ⇒ antisymmetric
L = 2 ⇒ symmetric.

Therefore when:

L = 0; χ(σ1 , σ2 ) must be antisymmetric, i.e., S = 0;

L = 1; χ(σ1 , σ2 ) must be symmetric, i.e., S = 1;
L = 2; χ(σ1 , σ2 ) must be antisymmetric, i.e., S = 0;

and the terms allowed by the Pauli principle are

1
S,3 P and 1 D.

64
3.6.5 Hund’s rules

For a given electronic configuration the set of allowed terms can be ordered
in increasing energy by following Hund’s rules. These are a set of three
rules, which, although they are empirical, apply rigorously to all ground-
state electronic configurations.

1. For a given electronic configuration the term with the highest spin
multiplicity, i.e., largest value of S, is lowest in energy.
This rule is a consequence of the effect of spin-spin interactions on
the energy of a term. Terms with high spin multiplicity are those
for which the electron spins are aligned parallel to each other. As
discussed above in the case of the exchange interaction (Section 3.5),
electrons with parallel spins tend to avoid each other and cannot be
located in the same place. Because of this they are spatially separated
from each other and the Coulomb repulsion between them is reduced.
As a result they are more tightly bound (lower in energy).

2. For a particular value of S in a given electronic configuration, the

term with the highest total orbital angular momentum, i.e., largest
value of L, is lowest in energy.
When the total orbital angular momentum of a term is largest, the
orbital angular momentum vectors of the individual electrons in the
configuration must be aligned parallel to each other, and in a classical
description the electrons must be orbiting in the same direction about
the nucleus. When orbiting in the same direction as each other, two
electrons encounter each other less often than when orbiting in oppo-
site directions. They are therefore on average further apart from each
other reducing the Coulomb repulsion between them and lowering
the term energy.

Example: In the 1s2 2s2 2p1 3p1 excited configuration of carbon:

S = 0 and 1 [i.e., (2S + 1) = 1 and 3 ] (3.105)

L = 0, 1 and 2 (3.106)

65
 1
S
1
P
1
D

Energy 3
S
3
P
3
D

Figure 3.4: Example of the ordering of terms in increasing energy according

to Hund’s rules.

Because the two outer electrons are in different orbitals they are non-
equivalent and the Pauli principle will be satisfied for all possible
terms, which are

1S 1P 1D

3S 3P 3D

Following Hund’s rules the set of triplet terms will be lower in energy than
the set of singlet terms, and for each set the D-term will be lowest in energy,
followed by the P- and S-terms as indicated in Figure 3.4

3.7 Spin-orbit interaction

We have seen above that electronic configurations can be separated into

terms which arise as a result of electron-electron interactions. However,
this does not give a complete description of the energy-level structure of a
many-electron atom.
Further energy splittings of individual terms can arise as a result of
relativistic spin-orbit interactions, leading to a total Hamiltonian, Ĥ, which

66
takes the general form

Ĥ = Ĥ0 + ĤSO , (3.107)

where Ĥ0 is the Hamiltonian obtained before including the effects of spin-
ˆ ~ˆ
orbit interactions, and ĤSO ∝ ~L · S is the Hamiltonian operator associated
with the spin-orbit interaction which is proportional to the scalar product
ˆ
of the total electronic orbital angular momentum operator, ~L, and the total
~ˆ
electron spin operator, S.
Since the electron is charged, and a dipolar magnetic field distribution
(magnetic dipole moment) can be associated with a rotating charge, two
types of magnetic dipole moment can be associated with an electron bound
within an atom. These are (1) the magnetic moment associated with the
orbital motion of the electron, and (2) the magnetic moment associated with
the spin of the electron. The interaction of these two magnetic moments is
known as the spin-orbit interaction and the resulting shifts or splittings of
the atomic energy levels are known as fine structure.

3.7.1 Orbital magnetic moment

To investigate the origin of the spin-orbit interaction in atoms we first con-

sider a simple classical picture of an electron with charge −e undergoing
circular motion in an orbit with a constant radius, |~r | = r, and a constant
v | = v (see Figure 3.5).
tangential velocity, |~
The current, I, associated with the orbit of this electron is
e
I = − (3.108)
T
ev
= − , (3.109)
2πr
where T is the period of the orbit.
For an orbit of area, A = π r2 , this current gives rise to a dipolar magnetic
field distribution with a component, µz , in the z-dimension

µz = IA (3.110)
evπr2
= − (3.111)
2πr

evr
= − . (3.112)
2

67
 y

v
e-
r

Figure 3.5: Electron undergoing circular motion in an orbit with a constant

radius, |~r | = r, and a constant tangential velocity, |~
v | = v.

From Figure 3.5 the orbital angular momentum vector associated with the
electron, ℓ~ = (0, 0, ℓz ), acts in the z-dimension (i.e., into the page), and

ℓz = me vr, (3.113)

therefore writing the z-component of the magnetic dipole moment arising

from the orbital motion of the electron as, µℓz ,
e
µℓz = − ℓz . (3.114)
2me
Since the same approach can be applied to the calculation of magnetic
moments associated with orbital motion in other planes, more generally
e ~
~ℓ = −
µ ℓ. (3.115)
2me
Note: This orbital magnetic moment is often expressed in terms of the
gyromagnetic ratio of the electron, γe , where
e
γe = − (3.116)
2me
such that

µ ~
~ ℓ = γe ℓ. (3.117)

68
 In the case depicted in Figure 3.5 where the orbital angular momentum
vector acts in the z-dimension, since (in SI units)

ℓz = mℓ ~ = 0, ±1~, ±2~, . . . , ±ℓ~, (3.118)

e
µℓz = − mℓ ~ (3.119)
2me
= −µB mℓ , (3.120)

where
e~
µB = (3.121)
2me
= 9.274 009 994 × 10−24 J T−1 (3.122)

is the Bohr magneton2 .

3.7.2 Spin magnetic moment

A similar procedure to that outlined above (Section 3.7.1) can be followed

~ s , associated
to obtain a classical approximation for the magnetic moment, µ
with the electron spin, i.e.,
e
~ sClass = −
µ ~s. (3.123)
2me
However, because the electron spin is a consequence of relativistic quantum
mechanics, the above simple classical picture does not give exactly the
~ s . The connection between the intrinsic spin of
correct magnitude for µ
an electron and its magnetic moment derived from the relativistic Dirac
equation is
e
~ sDirac = −2
µ ~s, (3.124)
2me
exactly twice the classical result in Equation 3.123. This additional factor of 2
is known as the electron spin g-factor and denoted ge . However, experimental
measurements3 of the electron magnetic moment lead to a value for ge which
is slightly different from 2,

ge = 2.002 319 304 362 (3.125)

2
CODATA Internationally recommended 2014 values of the fundamental physical con-
stants (http://physics.nist.gov/cuu/Constants)
3
D. Hanneke, S. Fogwell, and G. Gabrielse, New Measurement of the Electron Magnetic
Moment and the Fine Structure Constant, Phys. Rev. Lett. 100, 120801 (2008)

69
This small difference of 0.002 319 304 362 arises from quantum electrody-
namic contributions. Formally,

µsz = −ge µB ms , (3.126)

where ms = ±1/2.

3.7.3 The spin-orbit interaction

The above treatment provides a classical description of how magnetic mo-

ments can be associated with an electron undergoing circular motion. This
motion can correspond to the electrons orbital motion when bound in an
atom, or to its spin.
To treat the spin-orbit interaction within an atom the most appropriate
approach is to first determine the magnetic field arising from the orbital
motion of the electron, and to then consider the interaction of the magnetic
moment associated with the electron spin with this magnetic field.
Again we begin by considering a classical electron of mass me , moving
with a velocity ~ ~ Such a charged particle moving in an
v in an electric field E.
~ where
electric field experiences an effective magnetic field B,

~×~
E v
~ =
B . (3.127)
c2

If the electric field arises from the gradient of an isotropic electric potential,
φ(r) then

~ = − ~r
dφ
E , (3.128)
r dr
where ~r/r represents the radial unit vector, and hence

dφ
~ = − 1
B ~r × ~
v. (3.129)
2
r c dr
However, since

ℓ~ = ~r × ~
p = me ~r × ~
v, (3.130)

~ℓ = − 1 dφ ~
B ℓ (3.131)
me r c2 dr

70
is the magnetic field associated with the electron with orbital angular mo-
mentum ℓ.~
~ mag , with a magnetic
The interaction energy, Vmag , of a magnetic dipole, µ
~
field, B, can be expressed in a general form as

Vmag = −~ ~
µmag · B. (3.132)

~ s , associated with the spin

Therefore in the case of the magnetic moment, µ
of the electron the spin-orbit interaction becomes (in SI units)

HSO = −~ ~
µs · B (3.133)

1 dφ
= ~ s · ℓ~
µ (3.134)
me r c2 dr

e dφ
= − ~
~s · ℓ, (3.135)
m2e c2 r dr

e dφ ~
= − ℓ · ~s, (3.136)
m2e c2 r dr
using Equation 3.126 and the properties of the scalar product.
However, this is exactly twice the result obtained by solving the rela-
tivistic Dirac equation. This deviation by a factor of two arises as result of
the assumption that the nucleus in the atom is stationary and the electron
is orbiting classically around it. To account for this interaction correctly it is
necessary to carry out a relativistic treatment of the electron motion. When
this is done, the expression for HSO given in Equation 3.136 is modified by a
factor of 1/2 because of an effect known as Thomas Precession which occurs
for an electron moving at speeds close to the speed of light.
In the relativistic treatment, the coordinate system as seen from the
nucleus appears to rotate in the plane of the electron motion with a rotation
by 180◦ when the electron has made one complete orbit. Therefore the
electron appears to be spinning in its own frame of reference at only half of
the rate it would if this reference-frame was stationary, thus reducing the
apparent motion by a factor of 1/2.
The complete spin-orbit Hamiltonian is therefore
e dφ ~
HSO = − ℓ · ~s. (3.137)
2m2e c2 r dr

71
 To investigate the dependence of the spin-orbit interaction potential on
the nuclear charge +Ze and the principal quantum number, n, we must con-
sider that when bound within an atom the electron experiences a Coulomb
potential, i.e., in the case of a one-electron atom,

Ze
φ = − (3.138)
4πǫ0 r
and so
dφ Ze
= , (3.139)
dr 4πǫ0 r2
with the result that the Hamiltonian
e Ze ~
HSO = − ℓ · ~s, (3.140)
2me c2 r 4πǫ0 r2
2

Z
∝ . (3.141)
r3
The dependence on n is determined by calculating the expectation value of
r−3 , i.e.,
  Z
1 1
= Ψ∗nℓmℓ Ψnℓmℓ dτ (3.142)
r3 r3
Z3
= , (3.143)
n3 a3M ℓ(ℓ + 21 )(ℓ + 1)

where aM is the Bohr radius. Therefore

Z4
HSO ∝ (3.144)
n3
and for a given value of n rapidly increases with increasing nuclear charge,
while as the charge distribution of the electron becomes more extended at
high n, it reduces.

72
3.7.4 Spin-orbit operator

The Hamiltonian associated with the spin-orbit interaction can be expressed

in terms of the total orbital angular momentum and total spin operators as

e 1 dφ ~ˆ ~ˆ
HSO = − L · S. (3.145)
2m2e c2 r dr
ˆ
The form of this operator indicates that the spin-orbit interaction couples ~L
~ˆ so that they no longer have fixed z components, i.e., ML and MS are
and S
not good quantum numbers in the presence of spin-orbit coupling.
ˆ ˆ ~ˆ
However, ~J = ~L+ S does have a fixed z component M J . Consequently, the
spin-orbit interaction splits terms into levels the energies of which depend
on the total angular momentum quantum number J and are eigenstates of
ˆ ˆ ˆ ~ˆ 2 operators. The resulting good quantum numbers are
the ~J 2 , ~Jz , ~L 2 and S
therefore J, M J , L, and S.
When expressed in terms of the spin-orbit coupling constant A(L, S),
which is generally written in spectroscopic units (m−1 or cm−1 ),

~
HSO = hc A(L, S) ~L · S, (3.146)

such that
1 e dφ
A(L, S) = − 2
(3.147)
hc 2me c ~ r dr
2 2

~ in Equation 3.146 are therefore in atomic units.

and the quantities ~L and S
However, since

~ˆ
~Jˆ = ~Lˆ + S (3.148)
 2
~Jˆ 2 = ~Lˆ + S ~ˆ (3.149)
ˆ ~ˆ 2 + 2 ~Lˆ · S,
= ~L 2 + S ~ˆ (3.150)

the scalar product of the total orbital angular momentum and total spin is
 
~ˆ = 1 ~Jˆ 2 − ~Lˆ 2 − S
~Lˆ · S ~ˆ 2 . (3.151)
2
Because the spin-orbit interaction is weak compared to the gross energy
level structure of the atom, i.e., the energy separation between levels with

73
different values of n, first order perturbation theory can be applied to eval-
uate its effect on the atomic structure. This requires the expectation value of
the operator ĤSO to be determined to give the first order energy shift, ∆E(1)
arising form the interaction,
(1)
D E
∆ESO = ĤSO (3.152)

Z
= Ψ∗J,L,S,MJ ĤSO Ψ J,L,S,MJ dτ, (3.153)

where Ψ J,L,S,MJ are the wavefunctions of the unperturbed system.

In general the first order energy shift, ∆E(1) , associated with a weak
perturbing potential, Vperturb , is given by first order perturbation theory as
D E Z
(1)
∆E = Vperturb = Ψ∗0 Vperturb Ψ0 dτ, (3.154)

where Ψ0 are the wavefunctions of the unperturbed system.

So since
 
hc A(L, S) ~ˆ 2 ~ˆ 2 ~ˆ 2
ĤSO = J −L −S , (3.155)
2
the first order correction to the energy is given by
(1) (Z Z Z )
∆ESO A(L, S) ˆ2
∗~ ˆ2
∗~ ˆ2
∗~
= Ψ J Ψ dτ − Ψ L Ψ dτ − Ψ S Ψ dτ
hc 2
(3.156)

But because, e.g.,

~Jˆ 2 Ψ = J(J + 1) Ψ (3.157)

which is equal to
ˆ
Ψ∗~J 2 Ψ = Ψ∗ J(J + 1) Ψ (3.158)
= J(J + 1)Ψ∗ Ψ (3.159)

because J(J + 1) is a constant,

Z Z
ˆ2
∗~
Ψ J Ψ dτ = J(J + 1) Ψ∗ Ψ dτ (3.160)
| {z }
=1
= J(J + 1). (3.161)

74
 ˆ
A similar procedure can be followed to obtain the expectation values of ~L 2
~ˆ 2 with the result that
and S
∆ESO A(L, S)
= [J(J + 1) − L(L + 1) − S(S + 1)] . (3.162)
hc 2
For a given term, i.e., for given values of L and S, the total angular momen-
tum quantum number, J, takes the values

J = |L − S|, . . . , |L + S| in steps of 1, (3.163)

and the projection of the total angular momentum vector on the z axis, M J
is

M J = −J, −(J − 1), . . . , +(J − 1), +J. (3.164)

Therefore the spin-orbit interaction results in sets of levels associated

with each term, the energies of which depend on L, S and J. These levels
are labelled in spectroscopic notation as

2S+1
LJ (3.165)

For Example: In the 2p1 configuration in the H atom:

L = ℓ=1 (3.166)
1
S = s= (3.167)
2
Therefore
   
 1  1 
J = 1 −  , . . . , 1 +  (3.168)
2 2
1 3
= and . (3.169)
2 2
leading to two levels

2
P1/2 and 2 P3/2 . (3.170)

If the spin-orbit coupling constant A2p (L, S) = 0.243 cm−1 ,

∆ESO (J = 1/2)
= −A2p (L, S) = 0.243 cm−1 (3.171)
hc

75
 and
∆ESO (J = 3/2) 1
= + A2p (L, S) = 0.121 cm−1 (3.172)
hc 2
and the spin-orbit splitting between the two levels is

3
A2p (L, S) = 0.364 cm−1 (3.173)
2

These energy splittings are on the order of 3 × 10−6 RH , where RH , the

Rydberg constant for hydrogen, sets the scale for the gross energy level
spacings between sates with different values of n in the atom. The smaller
energy splittings associated with the spin-orbit interaction are known as
fine structure.
From the above it can be seen that in general the energy difference
between level J and level J − 1 is

∆ESO (J) − ∆ESO (J − 1) = hc A(L, S) J. (3.174)

This is known as the Landé interval rule.

76
3.7.5 LS coupling (Russell-Saunders coupling)

In the above discussion of the spin-orbit interaction in atoms the most

appropriate description is provided by considering that the total orbital
ˆ ~ˆ
angular momentum, ~L, of all of the electrons is coupled to the total spin, S,
ˆ
of all the electrons to give the total orbital angular momentum, ~J, i.e,
X
~ˆ =
S ~sˆi (3.175)
i

and
X
~Lˆ = ˆ
ℓ~i (3.176)
i

combine to give

~ˆ
~Jˆ = ~Lˆ + S. (3.177)

This regime, in which the total orbital angular momentum and total spin
couple, is known as LS coupling or Russell-Saunders coupling and is most ap-
propriate for atoms with low nuclear charge, Z, (i.e., low-Z atoms) in which
the spin-orbit interaction is weaker than the electron-electron interactions.
In this case, Hund’s third rule can be applied to arrange the resulting
levels in order of increasing energy. This rule has three cases.

3a. Normal case: For a particular term of a given electronic configura-

tion, if the outer subshell of the configuration is less than half full
the level energies increase with increasing values of J. Therefore
the level with the smallest value of J is lowest in energy.

3b. Inverted case: For a particular term of a given electronic configura-

tion, if the outer subshell of the configuration is more than half full
the level energies decrease with increasing values of J and the level
with the largest value of J is lowest in energy.

3c. For a particular term of a given electronic configuration, if the outer

subshell is half full there is no multiplet energy splitting.

Therefore the energy degeneracy of terms with different values of J is lifted

by the spin-orbit interaction. However, the energy degeneracy of sublevels
with different values M J remains.

77
3.7.6 j j-coupling

In contrast to the above description of angular momentum coupling in low-

Z atoms in terms of LS coupling, in high-Z atoms, because the spin-orbit
coupling constant A(L, S) depends strongly on Z, i.e,

A(L, S) ∝ Z4 , (3.178)

ˆ
the spin-orbit interaction between ℓ~i and ~sˆi for each individual electron can
be very strong and combining them gives each individual electron its own
ˆ
value of ~ji . In such cases, it the most appropriate description of the angular
momentum coupling is given by combining the total angular momenta of
ˆ
the individual electrons to obtain ~J, i.e.,

~jˆi = ℓ~ˆi + ~sˆi (3.179)

and
X
~Jˆ = ~jˆi (3.180)
i

such that in this case

J = |j1 − j2 |, . . . , | j1 + j2 | in steps of 1. (3.181)

These coupling schemes represent a way in which to classify the en-

ergy level structure in many electron atoms, but neither the LS- nor the j j-
coupling scheme give a complete description of the structure. The changes
in the groupings of levels from the LS coupling regime for low-Z atoms to
the j j coupling regime for high-Z atoms in the ground state configurations
of the group IV elements of the periodic table can be seen in Figure 3.6. In
all cases the total number of levels is the same. However, when Z is low,
i.e., for C and Si, the grouping of 3+1 energy levels follows more closely
the LS-coupling description, while when Z is high, i.e., for Sn and Pb, the
grouping of 2+2 energy levels is better described by the j j-coupling scheme.

78
 Carbon Silicon Germanium Tin Lead
( j1, j2 )J
(Z = 6) (Z = 14) (Z = 32) (Z = 50) (Z = 82)
+6000
(3/2, 1/2)1
+4000
(3/2, 1/2)2
Energy / hc (cm-1)

+2000 1
P1

0
3
P0,1,2
-2000

(1/2, 1/2)1
-4000
(1/2, 1/2)0
1 1 1 1 1 1 1 1 1 1
2p 3s 3p 4s 4p 5s 5p 6s 6p 7s
-6000

Figure 3.6: Energy level structure in the ground sate configurations of the
group IV elements.

3.7.7 Parity of terms in many-electron atoms

As discussed in Section 2.3.3 the parity of a one-electron wavefunction is

given by (−1)ℓ , i.e., the parity of the spherical harmonic Yℓmℓ (θ, φ).
In the case of a many-electron atom, the parity of a particular term is
given by
P
(−1)ℓ1 (−1)ℓ2 (−1)ℓ3 × · · · × (−1)ℓi = (−1) i ℓi , (3.182)

where ℓi are the orbital angular momentum quantum numbers of the in-
dividual electrons. It is important to note that this is not equivalent to
(−1)L .
If a term has odd parity this is indicated by a superscript ’o‘after the
term symbol, i.e.,

2
Po (3.183)

If a term has even parity this is indicated by omitting this superscript, i.e.,

2
P (3.184)

79
 For example the ground state electronic configuration of the C atom is
1s2 2s2 2p2 so
X
ℓi = 2 ⇒ even parity (3.185)
i

and the ground level is then simply

3
P0 (3.186)

However, since the ground state configuration of the B atom is 1s2 2s2 2p1 ,
X
ℓi = 1 ⇒ odd parity (3.187)
i

and the ground level is

2 o
P1/2 (3.188)

3.8 Hyperfine structure

Atomic nuclei are composed of protons and neutrons. These are Fermions
with spin angular momenta of 1/2. When they bind to form the nucleus
of an atom the spin angular momenta of these individual particles can be
combined, giving rise to a total nuclear spin vector ~I.
In the same way that there is a magnetic dipole moment associated with
~ I , can be associated with
the spin of the electron, a magnet dipole moment, µ
a nuclear spin such that

~I
~ I = gN µN ,
µ (3.189)
~
where
me
µN = µB (3.190)
mp

is known as the nuclear magneton, and gN is the nuclear spin g-factor which
depends on the structure of the nucleus. For example, the proton spin g-
factor is gp = 5.6, while the neutron spin g-factor is gn = −3.8. The total
nuclear spin quantum number depends strongly on the nuclear structure,
with some examples given in Table 3.2.

80
 Isotope Nuclear spin
H I = 1/2
12 C I=0
16 O I=0
13 C I = 1/2
14 O I = 5/2

Table 3.2: Nuclear spins of selected atomic isotopes.

The nuclear magnetic moment in an atom can give rise to energy level
shifts that are analogous to those associated with the spin-orbit interaction.
However, in this case the energy shifts, ∆EHFS , depend on the interaction of
the nuclear magnetic moment with the total electron angular momentum
and total electron spin, i.e.,

∆EHFS ∝ ~I · ~J (3.191)

and

~
∆EHFS ∝ ~I · S. (3.192)

The energy shifts arising from these interactions are known as the atomic
hyperfine structure (HFS) of the atom. Hyperfine interactions lift the energy
degeneracy of levels with different values of the total angular momentum
quantum number including nuclear spin, F. The vector associated with this
this total angular momentum quantum number is F,~ where

~ = ~J + ~I,
F (3.193)

and the quantum numbers then take the values

F = |J − I |, . . . , |J + I | in steps of 1. (3.194)

As an example of the hyperfine structure of an atomic energy level,

consider the 1 2 S1/2 ground state of the H atom. This has

S = 1/2 (3.195)
J = 1/2, (3.196)

81
and because the proton has a spin of 1/2,

I = 1/2. (3.197)

Therefore the possible values of F are

F = 0, 1. (3.198)

As a result of the hyperfine interaction, the levels with these two differ-
ent total orbital angular momenta including nuclear spin are separated in
energy by the hyperfine splitting, ∆EHFS . For this particular case

∆EHFS
= 0.047 cm−1 . (3.199)
hc
The transition between these two hyperfine sublevels occurs at a wave-
length of

λ ≃ 21 cm, (3.200)

in the radio frequency region of the electromagnetic spectrum. The emission

of radiation in the decay of H atoms via this transition is of importance in
astronomical spectroscopy of interstellar gas clouds. Because it has a very
small Einstein A-coefficient of

AHFS ≃ 2.9 × 10−15 s−1 , (3.201)

the decay of a sample of ground state H atoms via this transition occurs
very slowly with a timescale on the order of

1
τfl = = 11 × 106 years. (3.202)
AHFS

82
Chapter 4

Atomic spectra

The transitions that can occur between atomic energy levels are governed
by sets of selection rules. These selection rules impose restrictions on the
transitions that are allowed.

4.1 Selection rules for electric dipole transitions

Consider an atom in the presence of an electromagnetic field. The oscillating

electric component of this field interacts most strongly with the atom so we
therefore neglect the much weaker interaction of the oscillating magnetic
component of the field. The oscillating electric field takes the form

~ = E
E ~ 0 eiωt , (4.1)

~ 0 = (E0 , E0 , E0 ) is the amplitude of the field, and ω = 2π ν is the

where E x y z

angular oscillation frequency.

The interaction of the atom with the oscillating electric field occurs via
~ = e ~r such that
its electric dipole moment, µ

Vdip = −~ ~
µ · E, (4.2)

and the Hamiltonian in the presence of the field is

Ĥ = Ĥ0 − µ ~ˆ
~ˆ · E, (4.3)

where Ĥ0 is the field-free Hamiltonian.

83
 When the time-dependent perturbation associated with the oscillating
electric field is small compared to Ĥ0 , first-order time-dependent perturba-
tion theory can be applied. This leads to Fermi’s golden rule by which the
transition probability, Tif , between an initial state, i, and final state, f, can be
determined
Z 2
 ∗ 
Tif =  Ψf Vdip Ψi dτ  . (4.4)

For each pair of initial and final states if Tif = 0 the transition is forbidden,
while if Tif , 0 the transition is allowed.
To obtain further insight into the probability of a transition occurring,
consider the interaction of the atom with the x, y and z components of the
oscillating electric field separately. In spherical polar coordinates,

~ˆ = e ~rˆ = e (r sin θ cos φ, r sin θ sin φ, r cos θ).

µ (4.5)

The z component of the interaction potential is therefore

Vdipz = e r cos θ E0z eiωt . (4.6)

For a one-electron atom, the total electronic wavefunction excluding

spin takes the form

Ψnℓmℓ = Rnℓ (r) Yℓmℓ (θ, φ). (4.7)

Therefore following Fermi’s golden rule, the probability for a transition

between an initial state Ψnℓmℓ and a final state Ψn′ ℓ′ m′ℓ to occur is
Z ∞
2
Tn′ ℓ′ m′ℓ , nℓmℓ = E0z  R∗n′ ℓ′ (r) r Rnℓ (r) r2 dr ×

0
Z 2πZ π
2

× Yℓ∗ ′ m′ (θ, φ) cos θ Yℓmℓ (θ, φ) sin θ dθ dφ (4.8)
ℓ 
0 0

The radial part of this integral imposes no restrictions on the values of nℓ

and n′ ℓ′ . However, for any give values of n and ℓ, the angular part of the
integral does limit the values of n′ and ℓ′ for which the transition probability
is non-zero.
The angular part of Equation 4.8 can be expressed completely in terms
of spherical harmonic functions by considering that
r
3
Y1,0 (θ, φ) = cos θ (4.9)
4π

84
and therefore
r
4π
cos θ = Y1,0 (θ, φ), (4.10)
3
and

∗
Yℓ,mℓ
(θ, φ) = (−1)mℓ Yℓ,−mℓ (θ, φ), (4.11)

leading to the integral

r
4π 2π π
Z Z
mℓ
(−1) Yℓ,−mℓ (θ, φ) Y1,0 (θ, φ) Yℓ,mℓ (θ, φ) sin θ dθ dφ (4.12)
3 0 0
which, from the properties of the spherical harmonic functions, is equal to
zero unless

ℓ′ = ℓ ± 1 ⇒ ∆ℓ = ±1 (4.13)
m′ℓ = mℓ ⇒ ∆mℓ = 0. (4.14)

For the x and y components of the oscillating electric field ∆ℓ = ±1 and

∆mℓ = ±1.
Because the electric dipole interaction potential, Vdip , does not depend
on ~s the electron spin cannot change in an electric dipole transition in a
one-electron atom, i.e.,

∆s = 0. (4.15)

In addition, since the corresponding operator

~
Vdip = −e ~r · E (4.16)

has odd parity (because ~r has odd parity), and the integrand

Ψ∗f Vdip Ψi (4.17)

must be even for Tif , 0, therefore

Ψ∗f Ψi (4.18)

must be odd. For this to be the case the initial and final states must have
opposite parity, i.e., one must be odd and one must be even. This is the

85
origin of the selection rule that parity must change in an electric dipole
transition.
The complete set of selection rules for electric dipole transitions in one-
electron atoms is therefore:

Electric dipole selection rules for one-electron atoms:

∆ℓ = ±1
∆mℓ = 0, ±1
∆s = 0
Parity must change

If there is a strong spin-orbit interaction:

∆j = 0, ±1 but j = 0 9 j′ = 0

Along with these electric dipole allowed transitions other transitions can occur
(e.g., magnetic dipole transitions, electric quadrupole transitions), but with
probabilities that are considerable smaller.
Examples of allowed and forbidden electric dipole transitions in the H
atom are:

Allowed transitions: 1s → 2p
1s → 3p
2p → 5d

Not allowed: 1s 9 2s
1s 9 3d

86
Electric dipole selection rules for many-electron atoms:
In many-electron atoms a similar set of selection rules hold for electric
dipole transitions. These include a subset of rigorous selection rules which
must hold for a transition to occur at all, and an additional subset of selection
rules which if they are also fulfilled give rise to strong transitions.

Rigorous selection rules:

∆J = 0, ±1 but J = 0 9 J′ = 0
∆M J = 0, ±1
Parity must change

In addition transitions are weak unless:

∆S = 0
∆L = 0, ±1 but L = 0 9 L′ = 0

For example, the C atom has an excited electronic configuration 1s2 2s2 2p1 4d1
with an associated level

3 o
F2 (4.19)

for which

(2S + 1) = 3 ⇒ S = 1, (4.20)
L = 3 (4.21)
J = 2 (4.22)
’o‘ ⇒ odd parity (4.23)

The following transitions from this level can be classified as:

3F o → 3 P1 Weak because ∆L , ±1
2
3F o → 3 P1o Forbidden because parity does not change
2
3F o → 1D Weak because ∆S , 0, (intercombination line)
2 2

generally stronger than first transition above

3F o → 3 D2 Strong allowed transition
2

87
 (a) (b) 0 I. P.
...
3s 3p 3d
Level 2

Energy / hc (cm-1)
E2
2s 2p
Energy

-50000
E21 = E2 - E1
Approximate
2-level system

E1
Level 1
-100000
1s

Figure 4.1: Two-level systems. (a) An ideal two-level system with the
energy difference, ∆E21 , between the two levels, 1 and 2, as indicated. (b)
Example of an approximate two-level system in the H atom. In this case
the two relevant levels are the 1s and 2p levels. This can be considered a
two-level system because after excitation from level 1 (1s) to level 2 (2p)
the system can only decay back to level 1.

4.2 Rates of photon emission and absorption

To treat the absorption and emission of photons by atoms we first consider

a simple two-level system as depicted in Figure 4.1(a). An example of
such as system is that associated with the 1s and 2p states in the H atom
[see Figure 4.1(b)].
If a two-level system is exposed to an electromagnetic field of frequency,
ν21 = ∆E21 /h, corresponding to the energy interval between the two levels,
three processes must be considered: Absorption, spontaneous emission, and
stimulated emission.

Absorption: This involves the system, initially in level 1, absorbing a pho-

ton from the electromagnetic field and being excited to level 2.

The rate of absorption, Γabs , is proportional to (i) the number of atoms

(population) in level 1, N1 , and (ii) the energy density, U(ν12 ), of the
electromagnetic field, i.e.,

Γabs ∝ N1 U(ν12 ). (4.24)

88
Spontaneous emission: This involves the system, initially in level 2, de-
caying spontaneously to level 1 with the emission of a photon of
frequency, ν21 .

The rate of spontaneous emission, Γspon , from a sample of atoms is

proportional to the number of atoms (population) in level 2, N2 , i.e.,

Γspon ∝ N2 . (4.25)

Stimulated emission: This involves a photon from the electromagnetic

field ‘stimulating’, the system, initially in level 2, to decay to level 1
with the emission of another photon.

The rate of stimulated emission, Γstim , is proportional to (i) the number

of atoms (population) of level 2, N2 , and (ii) the energy density, U(ν21 ),
of the electromagnetic field, i.e.,

Γstim ∝ N2 U(ν21 ). (4.26)

The photon emitted by the system following stimulated emission has

the same frequency, is coherent with, i.e., its electromagnetic field is
oscillating in phase with, and propagates in the same direction as
the photon that stimulated the emission. This process of simulated
emission, in which one photon incident on an excited system causes
the emission of a second photon with identical properties, is the basis
on which the LASER (Light Amplification by Stimulated Emission of
Radiation) operates.

4.2.1 Two-level rate equations

Consider an ensemble of two-level atoms with N1 , the number of atoms in

energy level 1, with energy E1 , and N2 the number in energy level 2 with
energy E2 . If this ensemble is exposed to a thermal electromagnetic field,
i.e., a field in which the photons are propagating in random directions with
random polarisations, with a spectral energy density U(ν12 ) ≡ U(ν21 ). The
time-dependence of the population of levels 1 and 2 in this system can be

89
expressed as
Absorption Spon. decay Stim. emiss.
z }| { z}|{ z }| {
dN1
= − C U(ν12 ) N1 + A N2 + B U(ν12 ) N2 (4.27)
dt

Absorption Spon. decay Stim. emiss.

z }| { z}|{ z }| {
dN2
= + C U(ν12 ) N1 − A N2 − B U(ν12 ) N2 , (4.28)
dt
where A, B, and C are constants.
When the ensemble of two-level atoms is in equilibrium with the thermal
electromagnetic field

dN1 dN2
= = 0, (4.29)
dt dt
i.e., in the steady state there is no change in the average population of either
level 1 or level 2. Therefore

C U(ν12 ) N1 = A N2 + B U(ν12 ) N2 , (4.30)

or
N1 A + B U(ν12 )
= . (4.31)
N2 C U(ν12 )
When the system is in thermal equilibrium, the ratio of the population
in level 1, to the population in level 2 is described by a Maxwell-Boltzmann
distribution, i.e,

N1 e(−E1 /kB T)
= (4.32)
N2 e(−E2 /kB T)

= e(hν12 /kB T) , (4.33)

where hν12 = E2 − E1 , and kB is the Boltzmann constant.

Substituting Equation 4.33 into Equation 4.31 and rearranging the re-
sulting equation leads to

C U(ν12 ) e(hν12 /kB T) = A + B U(ν12 ) (4.34)

90
and so
A
U(ν12 ) = (4.35)
C e(hν12 /kB T) −B

(A/B)
= (4.36)
(C/B) e(hν12 /kB T) − 1

However, the spectral energy density of the electromagnetic field asso-

ciated with a radiating blackbody can be expressed, following the Plank
formula, as

8π ν212 hν12
U(ν12 ) = (4.37)
c3 e 12 /kB T)
(hν −1

8π h ν312 1
= . (4.38)
c3 e(hν12 /kB T) −1
Therefore comparison of Equation 4.38 with Equation 4.36 leads to the
conclusion that

C = B (4.39)

and
8π h ν312
A = B. (4.40)
c3
The variables A and B are known as the Einstein coefficients for spontaneous
and stimulated emission, respectively.

A ≡ Einstein A-coefficient for spontaneous emission

B ≡ Einstein B-coefficient for stimulated emission

The Einstein B-coefficient can be expressed in terms of the transition

probability given by Fermi’s Golden Rule as
  Z 2
2π 1  ˆ~ Ψ dτ  ,
B =  Ψ∗2 µ 1 (4.41)
3 ~2 4πǫ0 

where, for electromagnetic radiation linearly polarised in the z dimension,

~ˆ = e ẑ.
µ

91
4.2.2 Excited state lifetimes

Two-level system

In the absence of an electromagnetic field to stimulate emission an isolated

two-level atom in its excited state (level 2) will decay spontaneously. The
rate, Γspon , at which this spontaneous decay occurs is equal to the Einstein
A-coefficient for the transition from the excited state (level 2) to the ground
state (level 1), i.e.,

Γspon = A12 . (4.42)

The lifetime of the excited state, τspon , is equal to the inverse of this decay
rate
1
τspon = . (4.43)
Γspon

For example:
Consider a H atom in the 2p state with mℓ = 0, i.e.,

Ψ n ℓ mℓ = Ψ2 1 0 . (4.44)

From this state the atom can only decay to the 1s ground state with mℓ = 0,
i.e.,

Ψn′ ℓ′ m′ℓ = Ψ1 0 0 . (4.45)

Because of this single decay channel this can be considered a two-level

system so to determine the lifetime of the excited state using Equation 4.40
and Equation 4.41

8π2 ν312 
Z 2
A12 = ˆ Ψ2 1 0 dτ .
 Ψ∗1 0 0 µ
~ (4.46)
3ǫ0 ~ c3 

Note: this expression for the Einstein A-coefficient is sometimes also written
in an equivalent form in terms of the angular frequency, ω12 = 2π ν12 , and
h = 2π ~ as
2 ω312 
Z 2
A12 = ∗ ˆ
~ Ψ2 1 0 dτ .
 Ψ1 0 0 µ (4.47)
3ǫ0 h c3 

92
 For the ∆mℓ = 0 transition between the states Ψ2 1 0 and Ψ1 0 0 the electric
~ˆ , is
dipole transition operator, µ

~ˆ = e ẑ
µ = er cos θ. (4.48)

and since

Ψ n ℓ mℓ = Rn ℓ (r) Yℓ mℓ (θ, φ) (4.49)

The A-coefficient for the transition is

8π2 ν312  ∞ ∗
Z
A12 =  R1 0 (r) er R2 1 (r) r2 dr × . . .
3ǫ0 ~ c3  0
Z 2πZ π  2

× Y0∗ 0 (θ, φ) cos θ Y1 0 (θ, φ) sin θ dθdφ  . (4.50)
0 0 

Using the radial wavefunctions for the H atom in Equation 2.65 and in Ta-
ble 2.2, together with the spherical harmonic functions in Equation 2.62 and
in Table 2.1, Equation 4.50 becomes
 r  2
8π2 ν312  1 
A12 =  1.29 ea0  . (4.51)
3ǫ0 ~ c3  3 

Since in the H atom the energy difference between the 1s and 2p states is
given by the Rydberg formula as
 
∆E12 1 1
= RH 2
− 2 (4.52)
hc 1 2
−1
= 82258 cm (4.53)
≡ 121.568 nm, (4.54)
ν12 = 2.466 × 1015 Hz, (4.55)

and

A12 = 6.25 × 108 s−1 (or 6.25 × 108 Hz). (4.56)

Therefore the lifetime, τ2p , of the 2p state is

1
τ2p = (4.57)
A12
= 1.60 × 10−9 s (4.58)
= 1.60 ns. (4.59)

93
 0
4s 4p 4d 4f
3s 3p 3d
Energy / hc (cm-1)

2s 2p

-50000

-100000

1s

Figure 4.2: Spontaneous decay pathways from the 4s state in the H atom.

Multi-level system

In a multi-level system the rate of spontaneous decay, Γi , of an excited state,

i, is obtained by summing the Einstein A-coefficients for all possible decay
pathways, Ai f to all allowed final states f , i.e., the total decay rate is the
sum of the rates of decay via each decay pathway,
X
Γi = Ai f . (4.60)
f

Consequently, the lifetime, τi , of an excited state in a multi-level system is

1
τi = (4.61)
Γi
 −1
X 
 
=  Ai f  (4.62)
 
f

For example:
Under the selection rules for electric dipole transition, a H atom in the 4s
state can decay to the 3p or 2p states (see Figure 4.2)

94
 Following a similar procedure to that outlined above the Einstein A-
coefficients for the 4s→3p and 4s→2p transitions can be determined to be

A 4s 3p = 1.83 × 106 s−1 (4.63)

A 4s 2p = 2.58 × 106 s−1 . (4.64)

From Equation 4.60, the total rate of spontaneous decay of the 4s state, Γ4s ,
is therefore

Γ4s = A 4s 3p + A 4s 2p (4.65)
= 4.41 × 106 s−1 . (4.66)

The lifetime of the 4s state, τ4s , can then be determined to be

1
τ4s = (4.67)
Γ4s

= 227 ns. (4.68)

4.2.3 Effects of finite excited state lifetimes

The finite lifetimes of excited states of atoms impose limits on the precision
with which the frequency (energy difference) associated with transitions to,
or from, these states can be determined. This follows from Heisenberg’s
time-energy uncertainty relation

~
∆E ∆t ≥ , (4.69)
2
where ∆E is the uncertainty in the energy, and ∆t is the uncertainty in the
time.
As a consequence of this, transitions involving excited states with short
lifetimes have large spectral widths, i.e., after excitation the short excited
state lifetime means that the time when the atom will spontaneously decay
can be determined precisely, but the frequency of the photons emitted when
the decay occurs will not be very precisely defined.
On the other hand, transitions to states with long lifetimes have narrow
spectral widths, i.e., after excitation the long excited state lifetime means
that the time when the atom will spontaneously decay cannot be precisely

95
determined, but the frequency of the emitted photons will be very precisely
defined.

4.2.4 Metastable states

If the spontaneous decay of an excited state in an atom is forbidden by the

selection rules for electric dipole transitions (see Section 4.1), i.e., if A f i = 0
for all final states, then the excited state is said to be metastable.
For example:
In the H atom the 2s state is a metastable state as it is forbidden to decay to
the 1s ground state by the selection rule that ∆ℓ = ±1 (see Figure 4.3).

In the 1s2s configuration of the He atom the 1 S0 and 3 S1 levels are both
metastable (see Figure 4.3).

• The 1s2s 1 S0 level cannot decay to the 1s2 1 S0 ground state because:

i. the parity does not change

ii. the ∆J = 0, ±1 but J = 0 9 J′ = 0 selection rule is not fulfilled
iii. the ∆L = 0, ±1 but L = 0 9 L′ = 0 selection rule is not fulfilled

• The 1s2s 3 S1 level cannot decay to the 1s2 1 S0 ground state because:

i. the parity does not change

ii. the ∆S = 0 selection rule is not fulfilled
iii. the ∆L = 0, ±1 but L = 0 9 L′ = 0 selection rule is not fulfilled

Each of these metastable states can decay by higher order processes,

i.e, by magnetic dipole transitions, electric quadrupole transitions, or the
emission of two photons, but these are very slow. Consequently the 1s2s 3 S1
level in He, lives for 7870 s and is the longest lived excited state of an atom1 .

1
S. S. Hodgman, R. G. Dall, L. J. Byron, K. G. H. Baldwin, S. J. Buckman and A. G.
Truscott, Metastable Helium: A New Determination of the Longest Atomic Excited-State
Lifetime, Phys. Rev. Lett. 103, 053002 (2009)

96
 Helium
Energy / hc (cm-1) Hydrogen

Energy / hc (cm-1)

Figure 4.3: The metastable 2s state in the H atom, and the metastable
1s2s 1 S0 and 1s2s 3 S1 levels in the He atom (circled). Electric dipole allowed
transitions to these levels from more highly excited states, and allowed
decay pathways to the ground states are indicated.

4.3 Lasers

When a sample of two-level atomic systems are in thermal equilibrium

with their environment the ratio of the population of level 1, N1 , to the
population of level 2, N2 , follows a Maxwell-Boltzmann distribution, i.e.,
N1
= e(hν12 /kB T) . (4.70)
N2
As a result, N1 > N2 , and there are always many more atoms in level 1 than
in level 2. Therefore if a photon of frequency ν12 passes through the sample
it is more likely to be absorbed that to stimulate emission.
To ensure that stimulated emission is more probable than absorption it
is necessary to have more atoms in level 2 than in level 1, i.e., N2 > N1 . This
situation is known as a population inversion.
If a population inversion is achieved, and the rate of stimulated emission
(the Einstein B-coefficient) is greater that the rate of spontaneous emission

97
(the Einstein A-coefficient), incoming photons can cause stimulated emis-
sion, and the resulting photons can stimulate further emission leading to
an intense, coherent beam of monochromatic photons all travelling in the
same direction – a laser beam.
However, it is in general not possible to generate a population inversion
in a pure two-level system. This is because:

- If the population is initially in level 1 (i.e., N1 ≫ N2 ) absorption is

most likely

- This leads to an increase in the population of level 2 until N1 = N2

at which time further absorption and stimulated emission are both
equally likely to occur

Because a population inversion cannot be achieved, stimulated emission

cannot become the dominant excited state decay process.
To achieve a population inversion a system with at least three levels is
required.

4.3.1 Three-level lasers

A schematic diagram of a three-level system that can be exploited in the

operation of a laser is presented in Figure 4.4.
The process of generating a population inversion in this system begins
with the sample of atoms all in level 1.

A: The first step of the process of generating the population inversion in-
volves exciting, ’pumping‘, atoms from level 1 to level 2 using photons
at frequency, ν12 .

B: This is then followed by fast spontaneous decay (internal conversion) to

level 3 which is metastable leading to a population inversion between
level 3 and level 1.

C: Lasing can then occur on the transition from level 3 to level 1 at

frequency ν31 . Because level 3 is long-lived (metastable) it is possible
to ensure that stimulated emission from level 3 to level 1 is faster than
spontaneous emission.

98
 Level 2
B
Level 3
Energy (metastable)

A 12
C
31

Level 1

Figure 4.4: Schematic diagram of a three-level system in which a popula-

tion inversion can be achieved between level 3 and level 1 and exploited
as a laser.

In this system, because the frequency of the electromagnetic radiation used

to prepare the population inversion, ν12 , is different to the frequency at
which the stimulated emission occurs, ν31 , a population inversion can be
readily achieved.
Having prepared a system with a population inversion, to ensure the
successful operation of a laser the rate of stimulated emission must be grater
than the rate of spontaneous emission (Einstein A-coefficient). Because the
Einstein A-coefficient is proportional to the cube of the transition frequency,
i.e.,

Ai f ∝ ν3i f , (4.71)

this is more easily achieved at lower transition frequencies. For this reason
the first laser operated at microwave frequencies using a transition in the
ammonia molecule (NH3 ) and was known as a MASER – Microwave Am-
plification by Stimulated Emission of Radiation2 (see also 3 ). However, as
2
J. P. Gordon, H. J. Zeiger and C. H. Townes, The Maser – New Type of Microwave
Amplifier, Frequency Standard, and Spectrometer, Phys. Rev. 99, 1264 (1955)
3
C. H. Townes, How the Laser Happened: Adventures of a Scientist, Oxford University Press
(2002)

99
in Figure 4.4 this problem can also be overcome by selecting a system in
which level 3 is long-lived (metastable).
Following the development of the MASER the first LASER to operate
in the visible region of the electromagnetic spectrum employed ruby as the
lasing medium4 . The ruby laser is an example of a three-level system with
an energy level structure of the kind depicted in Figure 4.4.
Ruby is an aluminium oxide (Al2 O3 ) crystal which contains chromium,
Cr3+ impurities. It has strong absorption bands at wavelengths of ∼ 400 nm
and ∼ 550 nm (corresponding to ν12 in Figure 4.4). Excitation at these
wavelengths is followed by rapid non-radiative decay to a metastable level 3
which has a lifetimes of ∼ 3 ms. The lasing transition from level 3 to level 1
lies at a wavelength of 694 nm in the red region of the electromagnetic
spectrum.
However, three-level lasers of the kind outlined in Figure 4.4 such as the
ruby laser are not so efficient because after being stimulated to decay from
level 3 to level 1, the photons generated in the stimulated emission process
can be reabsorbed if generated in sufficient quantities.

4.3.2 Four-level lasers

To overcome the problem of absorption of amplified photons by the lasing

medium, and therefore to obtain improved operation efficiency, lasers based
upon four-level systems can also be implemented. An example of such a
laser is the helium-neon (He-Ne) laser, the energy level structure of which
is displayed schematically in Figure 4.5.
In the case of the He-Ne laser, level 1 is the overall ground state of the
He and Ne system, i.e., the He atom in the 1s2 1 S0 level and the Ne atom
in the 1s2 2s2 2p6 1 S0 level. Level 2 is associated with the 1 S0 and 3 S1 levels
of the excited 1s 2s configuration of the He atom, and level 3 corresponds
to the excited 1s2 2s2 2p5 4s1 and 1s2 2s2 2p5 5s1 configurations of the Ne
atom, while level 4 is associated with the 1s2 2s2 2p5 3p1 and 1s2 2s2 2p5 4p1
configurations of Ne.
As in the case of the three-level laser the process of generating a popu-
4
T. Maiman, Stimulated Optical Radiation in Ruby, Nature 187, 493 (1960)

100
 He Ne
Level 2 B
Level 3
Energy
34 C
A 12

Level 4
D
Level 1
Figure 4.5: Schematic diagram of a four-level system in which a population
inversion is achieved between level 3 and level 4 and exploited in a laser.

lation inversion in this four-level system begins with the sample of atoms
all in level 1.

A: The first step of the process of generating a population inversion

involves exciting, ’pumping‘, atoms from level 1 to level 2. This
is sometimes achieved optically, but in the case of the He-Ne laser
involves the generation of an electric discharge in a mixture of He and
Ne gas.

B: Excitation of level 2 is then followed by rapid energy transfer to level 3.

In the case of the He-Ne laser this is achieved via collisions between
the excited He atoms and ground state Ne atoms in the gas mixture.

C: Lasing occurs on the transition from level 3 to level 4 at frequency ν34 .

D: Level 4 is selected to rapidly decay by spontaneous emission to the

ground state, level 1, such that following stimulated emission from
level 3 the atoms in the system do not remain for a long time in level 4.
Consequently, reabsorption of the amplified radiation cannot occur.

The He-Ne laser is a gas laser operated with a ∼ 10 : 1 ratio of He to

Ne gas, at pressures on the order of 10 mbar. Such a laser can be operated

101
 Pumping

Laser output
Lasing medium

Totally reflecting Partially reflecting

mirror mirror

Figure 4.6: Schematic diagram a laser resonator with one totally reflect-
ing mirror and one partially reflecting mirror surrounding the lasing
medium. In such a configuration the lasing medium is typically trans-
versely pumped to generate a population inversion.

at several wavelengths in the red or infrared regions of the electromagnetic

spectrum, the most common being 633 nm.

4.3.3 Typical laser characteristics

In a typical laser system the lasing medium is enclosed between two mirrors
which form an optical cavity, or resonator as depicted in Figure 4.6. One of
these mirrors is selected to be very highly (totally) reflecting at the wave-
length at which the laser is operated while the other is chosen to be partially
reflecting. In such a configuration the laser medium is transversely pumped
to generate a population inversion, or in the case of the He-Ne laser, in the
region between the two mirrors an electric discharge is generated. In such
a resonator, photons that are generated by stimulated emission are fed back
into the lasing medium when reflected at the cavity mirrors, to cause fur-
ther stimulated emission, while a small fraction of photons are extracted
through the partially reflecting mirror as the laser output.
Lasers can be operated in pulsed or continuous (CW) modes each with
different characteristics and potential applications. Since the frequency

102
 Continuous lasers Pulsed lasers
Wavelength range ir – vis ir – uv
Power (peak) . 100 W > 1010 W
Bandwidth (∆ν) < 1 MHz 100 MHz – 20 THz
Pulse duration (∆t) – 10 ns – 50 fs
Examples Diode lasers Nd:YAG lasers (λ = 1064 nm)
He-Ne lasers Ti:Sa lasers (λ = 800 nm)

Table 4.1: Typical parameters of continuous and pulsed lasers.

width (bandwidth), ∆ν, is limited by the temporal duration (pulse duration),

∆t, pulsed lasers have bandwidths that are approximately equal to the
inverse of the pulse duration, i.e., ∆ν ≃ (∆t)−1 .

4.4 Laser cooling

Precision measurements of transition frequencies in atoms, molecules and

ions requires narrow bandwidth (CW) lasers and cold samples so that the
interaction times between the quantum system and the laser radiation does
not affect the frequency resolution that can be achieved. Frequency mea-
surements at the highest precision are of importance in tests of fundamental
physics including, e.g., in searches for time-variations of the fundamental
constants, the implementation of atomic clocks, quantum computing, and
the use of cold atoms and molecules to model, or ‘simulate’, complex con-
densed matter quantum systems. The development of laser cooling as a
way to prepare very cold samples of alkali metal atoms, in the 1980s and
1990s, opened up many new possibilities in these areas of reseaerch.

4.4.1 Particle momentum in a gas of hot atoms

Consider a gas of 133 Cs atoms at room temperature. Since M 133 Cs = 2.21 ×

10−25 kg, at T = 300 K in one-dimension
1 1
M 133 Cs v2rms = kB T, (4.72)
2 2

103
and so

vrms ≃ 137 m s−1 . (4.73)

Moving at this speed the momentum, p Cs , of a single 133 Cs atom is therefore

p Cs ≃ M 133 Cs vrms (4.74)

= 3 × 10−23 kg m s−1 . (4.75)

4.4.2 Laser cooling Cs

Because alkali metal atoms, e.g., Li, Na, K, Rb, Cs, have one outer optically
active electron in a s-orbital they all possess 2 S1/2 → 2 P3/2 transitions that
approximate two-level systems. In the case of laser cooling this is beneficial
because when excited to the upper 2 P3/2 level the atoms will always decay
directly back to the starting 2 S1/2 level.
In the case of Cs this transition involves the electron initially in the
6s orbital and its excitation to the 6p orbital, i.e., the 6 2 S1/2 → 6 2 P3/2
transition. This occurs at a wavelength of λCs = 852 nm.
Since the momentum, phν , associated with a single photon of wave-
length, λ, is
h
pλ = , (4.76)
λ
In the case of a photon at 852 nm,

pλ = 7.8 × 10−28 kg m s−1 . (4.77)

When the atom absorbs the photon it experiences a small change in mo-
mentum. However, since

pλ ≪ p Cs (4.78)

the absorption of one single photon has little effect on the atoms momentum.
However, if many photons are scattered (absorbed and spontaneously
re-emitted) by an atom the effect of the small individual changes in momen-
tum add up. But because
p Cs
≃ 38 461 (4.79)
pλ
it is necessary that a single thermal Cs atom scatters 38461 photons in order
to be brought to rest.

104
 Laser beam
Atomic beam

pCs Spontaneous
emission p

Figure 4.7: Laser deceleration of an atomic beam. The atoms absorb

photons from a laser beam counter propagating along the axis of the atomic
beam. As each photon is absorbed the momentum of the atoms in their
direction of propagation is slightly reduced. When the atoms subsequently
spontaneously decay, the emission can occur in any direction. The effects
of the changes in the momentum of the atoms associated with these emitted
photons therefore averages to zero in each direction after many absorption-
emission cycles.

4.4.3 Laser deceleration

The momentum transfer associated with an atom absorbing a photon can

be exploited to decelerate the atom, or a beam of atoms, moving in one par-
ticular direction by arranging that a laser beam counter propagates along
the axis of the atomic beam as indicated in Figure 4.7. Every time one of the
atoms in the atomic beam absorbs a photon it experiences a change in its
momentum of −pλ in its direction of propagation. When the excited atom
subsequently decays, the spontaneously emission can occur in any direc-
tion. As a result the changes in momentum associated with the spontaneous
emission average to zero over many absorption-emission cycles.
Because the atoms are travelling in the direction opposite to the direction
of propagation of the laser beam the Doppler shift of the laser radiation
must be considered when setting the frequency of the laser radiation to be
resonant with the atomic transition. Since the effective frequency of the

105
laser radiation in the frame-of-reference of the atoms, ν′ , is
1
ν′ = ν L  , (4.80)
1 + vc
where νL is the frequency at which the laser is operated, v is the speed
at which the atoms travel with respect to the direction of propagation of
the laser beam, and c is the speed of light in vacuum. In order, for the
laser radiation to be resonant with the atomic transition frequency, νi f , it is
therefore necessary to adjust the frequency at which the laser is operated to
 
v
νL = ν i f 1 + . (4.81)
c
It is important to note that in this approach to decelerating a beam of atoms
it is necessary to continually adjust the laser frequency to account for the
changes in the Doppler shift as the atoms slow down.
The final momentum of an atom decelerated via this laser cooling
scheme is limited by the recoil momentum associated with the emission
of the last photon. Therefore the minimum speed to which a beam of 133 Cs
atoms can be decelerated, vmin
Cs
, is
pλ
vmin
Cs = (4.82)
M 133 Cs
h 1
= (4.83)
λ M 133 Cs
≃ 0.003 m s−1 . (4.84)

In this approach to laser deceleration, ‘cooling’, atoms the maximum

deceleration is achieved when the time between the atom absorbing two
consecutive photons in minimal. Because the excited state, the 6 2 P3/2 state
in the case of Cs, has a finite lifetime τ Cs = 30 ns, it is necessary to ensure that
it has decayed spontaneously before the atom can absorb another photon.
If the intensity of the laser beam is not selected to meet this criterion it
will cause stimulated emission, canceling the effect of the reduction in the
momentum of the atom following the initial absorption of the photon. To
achieve the maximum deceleration it is most optimal to set the intensity
of the laser beam so that the average time between the absorption of two
consecutive photons, ∆tabs , is twice the excited state lifetime, i.e.,

∆tabs ≃ 2 τ Cs . (4.85)

106
Hence, for maximum deceleration the absorption rate must be set to be
approximately have of the spontaneous emission rate.
To further ensure that stimulated emission does not limit the decel-
eration (cooling) efficiency, the laser radiation is chosen to be circularly
polarised, with one chosen helicity, in which case it will only drive, e.g.,
∆M J = +1 transitions. If the transition

6 2 S1/2 −→ 6 2 P3/2 (4.86)

(M J = +1/2) (M J = +3/2)

is driven using circularly polarised light ∆M J = +1 as indicated, and the

excited state with M J = +3/2 cannot be stimulated to emit by a further
∆M J = +1 transition because the maximum value of M J of ground level is
+1/2.
In the case of the Cs atom, since τCs = 30 ns, for maximal deceleration
the average time between the absorption of two consecutive photons is
therefore 60 ns. So for the atom to scatter 38 361 photons and be decelerated
from an initial speed u = 137 m/s to 0 m/s requires at time, t, of
p Cs
t = 2τ Cs (4.87)
pλ
= 2.3 ms. (4.88)

For a final speed of 0 m/s the corresponding acceleration, a Cs , is therefore

u
a Cs = − (4.89)
t
= −5.9 × 104 m s−2 (4.90)
≃ 6 000 g. (4.91)

4.4.4 Magnetic trapping and further cooling

When decelerated using the approach described above, the resulting cold
atoms can be trapped magnetically. This is achieved by generating a min-
imum of magnetic field strength in three-dimensions by operating two
solenoids with electrical current flowing in the opposite direction in each as
indicated in Figure 4.8. The resulting magnetic field distribution has a zero-
field point mid-way between the two solenoids around which cold atoms

107
 (a) Solenoid 1 Solenoid 2
I I

Raidal position

(b)
Magnetic field B

(c)
Magnetic field
strength |B|

Trap minimum

Figure 4.8: (a) Two co-axial solenoids in an anti-Helmholtz configuration

with current, I, flowing in the opposite direction in each. (b) Schematic
diagram of the the magnetic field on the axis of the two solenoids in (a).
(c) The magnitude of the magnetic field in (b) with the three-dimensional
trap minimum indicated.

can be trapped. Such a configuration of current carrying solenoids is known

as a anti-Helmholtz configuration and gives rise to a quadrupole magnetic field
distribution.
When trapped in such a magnetic quadrupole trap after deceleration the
atoms can continue to be laser cooled using six counterpropagating laser
beams. In this way temperatures of the trapped atoms on the order of
T ∼ 1 µK can be achieved, with approximately 108 atoms confined in a
volume of ∼ 1 mm3 .
From this point it is difficult to cool the atoms further using laser cooling

108
techniques because of the momentum transfer that occurs following the
photon recoil upon spontaneous emission. However, they can be further
cooled by a technique known as evaporative cooling. This involves gradually
lowering the depth of the magnetic trap to allow the fastest moving atoms
to escape. The remaining atoms can then rethermalise by collisions to a
lower temperature. In this way final temperatures in the range from 10−7 –
10−9 K (100 nK – 1 nK) can be achieved.
At these low temperatures the ultracold atoms can form a Bose-Einstein
Condensate (BEC) in which, because the de Broglie wavelength of each of
the atoms is large, the spatial wavefunctions of the atoms overlap and all
atoms collect in the same quantum state of the magnetic trap.

109
 Part II

Molecules

110
Chapter 5

Diatomic molecules

5.1 Introduction

Atoms rarely remain isolated. At sufficient density they are more likely to
bind with each other to form molecules or solids. For example, a gas of H
atoms will combine to form molecular hydrogen, H2 , and a gas of atomic
nitrogen will combine to form N2 . Because of this, in the laboratory we
cannot fill a gas cylinder with H or N atoms. The cylinders of gas that are
used only contain H2 or N2 molecules. The exceptions to this are the rare
gas atoms, He, Ne, Ar, Kr and Xe which are stable and unreactive because
of their closed-shell electronic structure. These atoms do not readily form
molecules at room temperature.
In molecules, as in atoms, the energy level structure is dominated by
Coulomb interactions. However, in molecules the motion (vibrational and
rotational) of the nuclei must also be considered. In the following we will
treat the covalently bound diatomic molecular hydrogen cation, H+2 , and
the neutral hydrogen molecule, H2 , and its isotopomers, e.g., HD, D2 , HT,
etc. . . . We will also discuss ionically bound diatomic molecules such as
NaCl, LiF, HF, etc. . . .
Simple diatomic molecules such as these are important in several areas
of research including

i Precision spectroscopic tests of molecular quantum mechanics and

fundamental physics

111
 A B
Vibrational motion

Rotational motion

Figure 5.1: Schematic examples of the vibrational and rotational motion

of a diatomic molecule AB.

ii Studies of simple chemical processes and dynamics

iii Atmospheric chemistry and sensing

iii Chemistry in the interstellar medium

Molecular spectra are in general more complicated than atomic spec-

tra. This additional complexity arises from the quantised vibrational and
rotational motion of the nuclei (see Figure 5.1) which must be considered
in addition to the electronic structure.

5.1.1 Molecular Schrödinger equation

In a molecule with N nuclei of charge ZN and mass MN located at positions

~ N , and i electrons located at positions ~ri , the time-independent Schrödinger
R
equation takes the form

~ 1, R
Ĥ Ψ(R ~ 2, . . . , R
~ N ; ~r1 , ~r2 , . . . , ~ri ) = E Ψ(R
~ 1, R
~ 2, . . . , R
~ N ; ~r1 , ~r2 , . . . , ~ri ),

(5.1)

~ N ; ~ri ) is the total wavefunction of the system, and the Hamiltonian

where Ψ(R
operator, Ĥ, accounts for the kinetic energy of the nuclei, the kinetic energy

112
of the electrons and the total potential energy of the system.
~ 1, R
Ĥ = Ĥ(R ~ 2, . . . , R
~ N ; ~r1 , ~r2 , . . . , ~ri ) (5.2)

~2 2 ~2 2 ~2 2
= − ∇1 − ∇2 − · · · − ∇ + ...
2M1 2M2 2MN N

~2 2 ~2 2 ~2 2
− ∇1 − ∇2 − · · · − ∇ + ...
2me 2me 2me i

+ ~ N ; ~ri )
V(R (5.3)
| {z }
Coulomb interactions

X " ~2 # X " ~2 #
= − ∇2
+ − ∇2 ~ N ; ~ri ).
+ V(R (5.4)
2MN N 2me i
N i
The potential energy term is then the sum over the electron-electron Coulomb
repulsion, the nucleus-nucleus Coulomb repulsion, and the electron-nucleus
Coulomb attraction, i.e,
 
e2
X X Z Z X Z 
1
~ N ; ~ri ) =  N M  − N
 
V(R    +    .
4πǫ0 ~ ~  N>M R ~N − R
~ M  N,i R ~ N − ~ri  
i>j ri − r j


(5.5)

Because of the last term in this expression that represents the interaction
between the electrons and the nuclei, it is in general not possible to simplify
the molecular Schrödinger equation by separating the nuclear, ν(R ~ N ), and
electronic, ψ(~ri ), parts of the total wavefunction, i.e.,
~ N ; ~ri ) , ν(R
Ψ(R ~ N ) ψ(~ri ). (5.6)

However, under certain conditions an approximation known as the Born-

Oppenheimer approximation can be applied and this does allow a partial
separation of nuclear and electronic motion.

5.2 Born-Oppenheimer approximation

The Born-Oppenheimer approximation is based on the notion that in gen-

eral because the nuclei in a molecule are much heavier than the electrons,

113
 A B
CM

R
Figure 5.2: A diatomic molecule with the relative positions of nuclei A and
~ CM is the center of mass of the molecule.
B represented by the vector R.

i.e.,
me
≪ 1 (5.7)
MN
they move much more slowly than the electrons. Therefore for every change
in the position of the nuclei the electrons can be considered to adjust their
positions instantaneously. Taking this into account it is possible to express
the total wavefunction of the system as a product of nuclear and electronic
parts such that the electronic wavefunction depends on the instantaneous
position of the nuclei but not on their motion.
Consider the diatomic molecule AB in Figure 5.2. For this system with
a reduced nuclear mass, µAB , of
MA MB
µAB = (5.8)
MA + MB
the Hamiltonian takes the form

~ ~ri ) = − ~2 2 X ~2
~ ~ri ).
Ĥ(R; ∇R + − ∇2 + V(R; (5.9)
2µAB 2me i
i

To this system the Born-Oppenheimer approximation can by applied in

6 steps.

1. First the nuclei must be fixed in space at a selected internuclear sepa-

ration, i.e.,

~ = const.
R (5.10)

114
 In this situation the nuclear kinetic energy term in Equation 5.9,

~2 2
− ∇ (5.11)
2µAB R

can be neglected.

2. The Schrödinger equation can then be solved for the motion of the
electrons in the field of the fixed nuclei.

~ ~ri ) = Eelec Ψ(R;

Ĥelec Ψ(R; ~ ~ri ), (5.12)

where
X ~2 2 ~ ~ri ),
Ĥelec = − ∇ + V(R; (5.13)
2me i
i

is the Hamiltonian, and Eelec are the electronic eigenenergies of the

system. The solution to Equation 5.12 is then obtained for a range of
~ giving a complete description of the electronic
different values of R
energies and wavefunctions in terms of the the internuclear separa-
tion.

3. The solution of the complete Hamiltonian of the molecule including

the nuclear motion is then assumed to take the form

~ ~ri ) = ν(R)
Ψ(R; ~ ψ(R;
~ ~ri ), (5.14)

i.e., the nuclear part of the wavefunction depends only on the internu-
clear separation, and the electronic part depends on the internuclear
separation and the positions of the electrons1 .

~ ~ri ) is then inserted into the

4. In this form the total wavefunction Ψ(R;
full Schrödinger equation including the nuclear kinetic energy term
in Equation 5.11, i.e.,
 
 ~2 X ~2 
− ∇2
+ − ∇2
+ V( ~ ~ri ) ν(R)
R; ~ ψ(R;
~ ~ri ) = E ν(R)
~ ψ(R;
~ ~ri )
 2µAB R 2me i 
i
(5.15)
1
Note: This dependence of the nuclear and electronic parts of the total wavefunction on
~
R and ~ri is different to that in Equation 5.6.

115
 5. Applying the nuclear kinetic energy operator to the total wavefunc-
tion in Equation 5.15 leads to

~2 2 2
~ ~ri ) = − ~ ∇2 ν(R)
~ ψ(R;~ ~ri )
− ∇R Ψ(R; (5.16)
2µAB 2µAB R
~2 h 2 i
= − ψ∇R ν + 2∇R ν∇R ψ + ν∇2R ψ ,
2µAB
(5.17)

~ and ψ = ψ(R;
where in this case ν = ν(R) ~ ~ri ), and for the two functions
ν and ψ

∇2 ν ψ = ∇(∇ν ψ)
= ∇(ψ ∇ν + ν ∇ψ)
= ∇ψ ∇ν + ψ ∇2 ν + ∇ν ∇ψ + ν ∇2 ψ
= ν ∇2 ψ + 2 ∇ν ∇ψ + ψ ∇2 ν (5.18)

The insight of Born and Oppenheimer2 was that the electronic wave-
functions are insensitive to changes in the positions of the nuclei, i.e.,
the terms involving the gradients of ψ with respect to the nuclear co-
ordinates, ∇R ψ and ∇2R ψ in Equation 5.17, can be neglected in solving
the problem. Therefore

~2 2 2
~ ~ri ) = − ~ ψ(R;
~ ~ri ) ∇2 ν(R)
~
− ∇R Ψ(R; R (5.19)
2µAB 2µAB

5. Using this result (Equation 5.19) the total wavefunction given in Equa-
tion 5.15 can be simplified leading to

~ ~ri ) ψ(R;
⇒ Ĥelec (R; ~ ~ri ) = Eelec ψ(R;
~ ~ri )
z }| {
" #  
~2 X ~2 
~ ~ri ) −
ψ(R; 2 ~ +
∇ ν(R)  − 2
∇i + V(R; ~ ψ(R;
~ ~ri ) ν(R) ~ ~ri )
2µAB R

2me 
i
~ ψ(R;
= E ν(R) ~ ~ri ).

(5.20)
2
M. Born and R. Oppenheimer, Zur Quantentheorie der Molekeln (On the quantum
theory of molecules), Annalen der Physik, 389, 457, (1927)

116
 e-
rA rB

p+ R p+
A B
Figure 5.3: Coordinate system describing the molecular hydrogen cation
H+2 .

Since the second term in brackets, operating on the electronic part

of the total wavefunction, gives the electronic energy, Eelec , of the
system Equation 5.20 reduces to
" #
~2 2 ~ ν(R)
~ = E ν(R),
~
− ∇ + Eelec (R) (5.21)
2µAB R
the Schrödinger equation for the nuclear motion. Hence, the electronic
~ act as the effective potential in which the nuclei
energies, Eelec (R),
move.

5.3 The molecular hydrogen cation H+2

To apply the Born-Oppenheimer approximation to a real system, consider

the case of the molecular hydrogen cation H+2 . This simple diatomic
molecular ion is composed of two protons and one electron. In using
the Born-Oppenheimer approximation it is necessary to solve two separate
Schrödinger equations, (1) an electronic Schrödinger equation, and (2) a
nuclear Schrödinger equation.
~ is
First the positions of the nuclei must be fixed, i.e., a constant value of R
selected and the electronic motion is treated. This leads to the Hamiltonian
~2 2 e2 e2 e2
Ĥelec = − ∇r − − + , (5.22)
2me 4πǫ0 rA 4πǫ0 rB 4πǫ0 R
in the coordinate system indicated in Figure 5.3. In atomic units, where

e = me = ~ = 4πǫ0 = 1, (5.23)

117
this can be simplified to the form

∇2r 1 1 1
Ĥelec = − − − + . (5.24)
2 rA rB R
The resulting electronic Schrödinger equation is therefore
!
∇2r 1 1 1 ~ ~rA , ~rB ) = Eelec ψ(R;
~ ~rA , ~rB )
− − − + ψ(R; (5.25)
2 rA rB R

5.3.1 H+2 wavefunctions

When rA is small: the electron is close to proton A. The proton and electron
at A appear as a neutral H atom while proton B is an isolated proton. In
this case the electronic wavefunction for the ground state of the molecule
reduces to that of a hydrogenic 1s orbital, φ1s , centred upon proton A, i.e.,

~ ~rA , ~rB ) −→ φ1s (~rA ) = q 1

ψ(R; e−rA /aH , (5.26)
π a3H

where aH is the Bohr radius or the H atom (see Table 2.1 and Table 2.2).

When rB is small: the electron is close to proton B and the electronic wave-
function for the ground state of the molecule reduces to that of a 1s orbital,
φ1s , centred upon proton B, i.e.,

~ ~rA , ~rB ) −→ φ1s (~rB ) = q 1

ψ(R; e−rB /aH . (5.27)
3
π aH

Between these two cases the total wavefunction is a superposition of the

situation in which the electron is at proton A, and that in which the electron
is at proton B, i.e.,

~ ~rA , ~rB ) ≃ C1 φ1s (~rA ) + C2 φ1s (~rB ),

ψ(R; (5.28)

where C1 and C2 are constants.

Since H+2 is a homonuclear diatomic molecule, i.e., it is composed of two
identical nuclei, the probability of finding the electron at proton A has to be
always equal to that of finding the electron at proton B. Therefore

|C1 |2 = |C2 |2 = |C|2 (5.29)

118
and the electronic wavefunction can therefore be symmetric, ψ+ , or anti-
symmetric, ψ− , with respect to exchange of A and B,
h i
ψ+ = C φ1s (~rA ) + φ1s (~rB ) ‘symmetric’ (5.30)

h i
ψ− = C φ1s (~rA ) − φ1s (~rB ) ‘antisymmetric’ (5.31)

The symmetry of molecular electronic wavefunctions of these kinds is la-

belled using the german words ‘gerade’ and ‘ungerade’,

symmetric ≡ ‘gerade’
antisymmetric ≡ ‘ungerade’

and this approach to constructing molecular orbitals from superpositions

of atomic orbitals is known as the LCAO method (LCAO ≡ Linear Combi-
nation of Atomic Orbitals).
The value of the constant C in Equation 5.30 and Equation 5.31 can be
obtained by normalisation of the functions. This requires that
Z  2
ψ± (~rA , ~rB ) drA drB = 1

(5.32)

and therefore
Z h i∗ h i
2
C φ1s (~rA ) ± φ1s (~rB ) φ1s (~rA ) ± φ1s (~rB ) drA drB = 1 (5.33)

and so
Z h
2
C φ∗1s (~rA )φ1s (~rA ) + φ∗1s (~rB )φ1s (~rB ) + . . .
i
± φ∗1s (~rA )φ1s (~rB ) ± φ∗1s (~rB )φ1s (~rA ) drA drB = 1 (5.34)

However, because the wavefunctions, φ1s (~r ), associated with the atomic
orbitals are normalised
Z Z
∗
φ1s (~rA )φ1s (~rA ) drA = φ∗1s (~rB )φ1s (~rB ) drB = 1. (5.35)

If the overlap integral, I(R), is defined as

Z
I(R) = φ∗1s (~rA )φ1s (~rB ) drA drB (5.36)
Z
= φ∗1s (~rB )φ1s (~rA ) drA drB (5.37)

119
the normalisation condition becomes

C2 [1 + 1 ± 2I(R)] = 1, (5.38)

so
1
C = p . (5.39)
2 [1 ± I(R)]

Therefore

(i) If

R −→ ∞ ⇒ I(R) −→ 0 (5.40)

because at large internuclear separations there is no overlap between

the wavefunctions associated with the electron being at proton A and
that at proton B, and

1
C = √ (5.41)
2

(ii) If

R −→ 0 ⇒ I(R) −→ ±1 (5.42)

because for small internuclear separations the wavefunctions associ-

ated with the electron being at proton A and that at proton B com-
pletely overlap, and

1
C = (5.43)
2

The resulting electronic wavefunctions for the H+2 molecular ion are
displayed in Figure 5.4. From the electron probability distribution |ψ+ |2
in this figure, it can be seen that the symmetric (gerade) state has electron
density between the two protons, A and B. This neutralises the proton-
proton Coulomb repulsion leading to a stable molecular state in which the
electron is shared between the two nuclei in a covalent bond. Because this
state corresponds to a stable bound state of the molecule the corresponding
molecular orbital is often known as a bonding orbital.

120
 +0.6 +0.6
Symmetric Antisymmetric
‘gerade’ ‘ungerade’
+0.4 +0.4
a03/2 and | |2 a03

a03/2 and | |2 a03

+0.2 +0.2

+0.0 0.0
p+ p+ p+ p+
A B A B
-0.2 -0.2

-0.4 + -0.4 -
| +
|2 | -|2
-0.6 -0.6
-6 -4 -2 0 +2 +4 +6 -6 -4 -2 0 +2 +4 +6
R / a0 R / a0

Figure 5.4: Gerade, ψ+ , and ungerade, ψ− , electronic wavefunctions and

electron probability distributions for the ground states of the molecular
hydrogen cation H+2 .

On the other hand, the antisymmetric ψ− (ungerade) state exhibits a

lower probability of finding the electron located between the two protons.
As a result the proton-proton Coulomb repulsion is not compensated and
pushes the two nuclei apart . Consequently in this state the molecule is
unstable and breaks up, dissociates. The corresponding molecular orbital is
therefore known as an antibonding orbital.

5.3.2 H+2 electronic energies

The electronic energies, E± (R), of the lowest-lying gerade and ungerade

eigenstates of H+2 are given by
R
ψ∗± Ĥelec ψ± dτ
E± (R) = R (5.44)
ψ∗± ψ± dτ

A± expectation value of Ĥelec

= ≡ . (5.45)
N± normalisation constant
By requiring that the correct normalisation of the wavefunctions as R −→ ∞

121
gives

1 h i
ψ± = √ φ1s (~rA ) ± φ1s (~rB ) . (5.46)
2
Therefore,
Z h
1 i∗ h i
N± = φ1s (~rA ) ± φ1s (~rB ) φ1s (~rA ) ± φ1s (~rB ) dτ (5.47)
2

1
= [1 + 1 ± 2I(R)] (5.48)
2

= 1 ± I(R) (5.49)

To calculate A± ,
Z
1 h i∗ 1 h i
A± = √ φ1s (~rA ) ± φ1s (~rB ) Ĥelec √ φ1s (~rA ) ± φ1s (~rB ) dτ,
2 2
(5.50)

so therefore
Z " #
1 ∇2r 1 1 1
A± = ψ∗1s (~rA ) − − − + ψ1s (~rA ) dτ + . . .
2 2 r A rB R
Z " 2 #
1 ∇r 1 1 1
± ψ∗1s (~rA ) − − − + ψ1s (~rB ) dτ + . . .
2 2 rA rB R
Z " 2 #
1 ∇r 1 1 1
+ ψ∗1s (~rB ) − − − + ψ1s (~rB ) dτ + . . .
2 2 rA rB R
Z " 2 #
1 ∇r 1 1 1
± ψ∗1s (~rB ) − − − + ψ1s (~rA ) dτ. (5.51)
2 2 rA rB R

However, in the H atom

" 2 #
∇r 1
− − ψ1s (~rA ) = E1s ψ1s (~rA ) (5.52)
2 rA

and
" 2 #
∇r 1
− − ψ1s (~rB ) = E1s ψ1s (~rB ), (5.53)
2 rB

122
where E1s is the energy of the 1s state in the H atom, with similar operators
indicated in red and blue in Equation 5.51, therefore
Z  
1 1 1
A± = ψ∗1s (~rA ) E1s − + ψ1s (~rA ) dτ + . . .
2 rB R
Z  
1 1 1
± ψ∗1s (~rA ) E1s − + ψ1s (~rB ) dτ + . . .
2 rA R
Z  
1 1 1
+ ψ∗1s (~rB ) E1s − + ψ1s (~rB ) dτ + . . .
2 rA R
Z  
1 1 1
± ψ∗1s (~rB ) E1s − + ψ1s (~rA ) dτ. (5.54)
2 rB R
which can be simplified to
" Z Z  
1 1 1
A± = E1s ψ∗1s (~rA ) ψ1s (~rA ) dτ + ψ∗1s (~rA ) − + ψ1s (~rA ) dτ + . . .
2 rB R
Z Z  
1 1
± E1s ψ∗1s (~rA ) ψ1s (~rB ) dτ ± ψ∗1s (~rA ) − + ψ1s (~rB ) dτ + . . .
rA R
Z Z  
1 1
+ E1s ψ∗1s (~rB ) ψ1s (~rB ) dτ + ψ∗1s (~rB ) − + ψ1s (~rB ) dτ + . . .
rA R
Z Z   #
1 1
± E1s ψ∗1s (~rB ) ψ1s (~rA ) dτ ± ψ∗1s (~rB ) − + ψ1s (~rA ) dτ .
rB R
(5.55)

If the integrals of the form

Z
1
J(R) = ψ∗1s (~rA ) ψ1s (~rA ) dτ, (5.56)
rB
which represent the interaction of the electron charge close to proton A with
proton B, are then defined as the Coulomb integrals, where3
1h i
J(R) = 1 − (1 + R) e−2R . (5.57)
R
And integrals of the form
Z
1
K(R) = ψ∗1s (~rA ) ψ1s (~rB ) dτ, (5.58)
rA
3
See, for example; Physics of Atoms and Molecules, B. H. Bransden and C. J. Joachain,
Prentice Hall, 2nd Edition (2003)

123
which represent the exchange of the electron between the two nuclei, are
defined as the exchange integrals, where

K(R) = (1 + R) e−R . (5.59)

Then from Equation 5.55


1 1
A± = E1s − J(R) + + ...
2 R

I(R)
± E1s I(R) ∓ K(R) ± + ...
R

1
+ E1s − J(R) + + ...
R
#
I(R)
± E1s I(R) ∓ K(R) ± (5.60)
R

[1 ± I(R)]
= E1s [1 ± I(R)] + − J(R) ∓ K(R). (5.61)
R
Therefore
A±
E± = (5.62)
N±

1 −J(R) ∓ K(R)
= E1s + + . (5.63)
R 1 ± I(R)
The general dependence of energies of the two lowest lying electronic states
of H+2 on the internuclear distance, R, are

(i) As R −→ ∞ the overlap, Coulomb and exchange integrals all tend

towards zero because of their dependences on e−R .

I(R), J(R), and K(R) −→ 0 (5.64)

⇒ E− = E+ = E1s (5.65)

(ii) When R < ∞ the correction to the energy that depends on the overlap,
Coulomb and exchange integrals

−J(R) ∓ K(R)
(5.66)
1 ± I(R)

124
 ⇒ Lowers the energy of the gerade state, E+ .

⇒ Raises the energy of the ungerade state, E− .

(iii) As R −→ 0 the energies of the gerade and ungerade states both tend
towards infinity

E± −→ ∞ (5.67)

because of the Coulomb repulsion between the positively charged

nuclei.

(iv) The function, E+ (R), associated with the gerade state, exhibits a mini-
mum. This potential energy minimum corresponds to an internuclear
separation for which a stable molecule can exist.

On the other hand, the function, E− (R), associated with the ungerade
state, lies higher than the energy of the separated atoms for all inter-
nuclear separations and does not exhibit a minimum. This potential
energy curve therefore indicates that a stable molecule cannot form in
the antisymmetric state. If excited to this state, the molecule will fall
apart, dissociate.

These features of the E± (R) functions can be seen in Figure 5.5 in which
the results obtained using the orbital approximation described above (con-
tinuous curve) are displayed together with the results of an exact calculation
of the E+ (R) function (dashed curve). Also indicated in this figure are the
equilibrium internuclear distance, Re , for which the molecule is stable when
in the gerade state, and the dissociation energy, De , of this state. This equi-
librium dissociation energy is the energy difference between the potential
energy minimum at Re and the energy of the separated atoms.

4
W. Kolos, Some Accurate Results for Three Particle Systems, Act. Phys. Acad. Sci. Hung.,
27, 241 (1969).

125
 +0.1
Orbital approximation
0 Exact calculation

-0.1
Energy (a.u.)

-0.2
:
-
Antisymmetric
‘ungerade’
-0.3

-0.4

-0.5
:
+
Symmetric De
-0.6 ‘gerade’
Re

0 2 4 6 8 10
R / a0

Figure 5.5: Gerade, ψ+ , and ungerade, ψ− , electronic potential energy

functions for the ground states of the molecular hydrogen cation H2+ . The
continuous lines represent the energies obtained using the procedure de-
scribed, while the dashed curve is the result of an exact calculation4 . Also
indicated are the equilibrium internuclear distance, Re , and the dissocia-
tion energy, De , corresponding to the dashed curve.

5.4 The neutral hydrogen molecule H2

The neutral hydrogen molecule, H2 , is the simplest neutral molecule. As in

the case of H+2 ;

- The energy level structure of H2 can be calculated by first determining

the electronic wavefunctions and energy eigenvalues within the Born-
Oppenheimer approximation

- The H2 molecule exhibits a covalent bond in which the electrons are

shared between the nuclei

However, unlike H+2

126
 e-1
r12
e-2
rA1
rB2
rA2
rB1

p+ R p+
A B
Figure 5.6: Coordinate system describing the neutral hydrogen molecule,
H2 . The two protons are labelled by the letters A and B, while the two
electrons are labelled by the numbers 1 and 2.

- Because H2 contains two electrons, the total wavefunctions must sat-

isfy the Pauli principle and be antisymmetric with respect to exchange
of the electrons.

Using the system of coordinates indicated in Figure 5.6, the Hamiltonian

for H2 can be expressed in atomic units as
∇21 ∇22
1 1 1 1 1 1
Ĥelec = − − − − − − + + (5.68)
2 2 rA1 rA2 rB1 rB2 r12 R
| {z } | {z } | {z }
e− K. E. p+ − e− attraction repulsion

5.4.1 H2 electronic wavefunctions

In H2 , the electronic wavefunctions must account for the spin of the elec-
trons. Therefore the total electronic wavefunction, ψS, T , takes the form
~ ~ri ) χS, T ,
ψS, T = ψ(R; (5.69)
~ ~ri ) is the part of the wavefunction without spin, and χ S, T are the
where ψ(R;
singlet (S⇒ S = 0) and triplet (T⇒ S = 1) two-electron spin wavefunctions
(see Section 3.4.1).
To satisfy the Pauli principle the total electronic wavefunctions, ψ S, T ,
must be antisymmetric with respect to exchange of the two electrons. There-
fore

127
 ~ ~ri ) is antisymmetric with respect to exchange, the spin part of
• If ψ(R;
the wavefunction must be symmetric, i.e., χT

~ ~ri ) is symmetric with respect to exchange, the spin part of the

• If ψ(R;
wavefunction must be antisymmetric, i.e., χ S

As in the case of H+2 , the electronic wavefunctions associated with the

lowest-lying eigenstates of H2 can be constructed from atomic orbitals. Con-
sidering the 1s orbitals of the H atom, φ1s (~r), the electronic wavefunctions
of the two lowest energy states of H2 are

1 h i
ψT = √ φ1s (~rA1 ) φ1s (~rB2 ) − φ1s (~rA2 ) φ1s (~rB1 ) χT (5.70)
2
and
1 h i
ψS = √ φ1s (~rA1 ) φ1s (~rB2 ) + φ1s (~rA2 ) φ1s (~rB1 ) χS , (5.71)
2

where φ1s (~rA1 ) corresponds to electron 1 in the 1s orbital around proton A

√
etc. . . and the factor of 1/ 2 in these expressions ensure that the wavefunc-
tions are normalised at large internuclear separations, i.e., when R → ∞. At
smaller internuclear separations the wavefunctions must be renormalised.
The spatial part of the function ψT is antisymmetric with respect to exchange
of the two electrons, and the spin part of the wavefunction is therefore sym-
metric. On the other hand the spatial part of the function ψ S is symmetric
with respect to exchange and the spin part is therefore antisymmetric.
The determine the energies, E± (R), of these lowest energy eigenstates of
H2 , where

E+ (R) ⇒ ψ S ≡ ψ+ (5.72)
E− (R) ⇒ ψT ≡ ψ− (5.73)

the expectation value of the normalised Hamiltonian must be calculated for

128
a range of internuclear separations, i.e.,
R ∗
ψ S, T Ĥelec ψ S, T dτ
E± (R) = R ∗
(5.74)
ψ S, T ψ S, T dτ

A± (R)
= (5.75)
N± (R)

expectation value of Ĥelec

≡ . (5.76)
normalisation factor
The expectation value, A± , of the Hamiltonian, Ĥelec , can be written explic-
itly as
Z h
1 i∗
A± = φ1s (~rA1 ) φ1s (~rB2 ) ± φ1s (~rA2 ) φ1s (~rB1 ) Ĥel ×
2
h i
φ1s (~rA1 ) φ1s (~rB2 ) ± φ1s (~rA2 ) φ1s (~rB1 ) dτ (5.77)

"Z
1
= φ∗1s (~rA1 ) φ∗1s (~rB2 ) Ĥelec φ1s (~rA1 ) φ1s (~rB2 ) dτ
2
Z
+ φ∗1s (~rA2 ) φ∗1s (~rB1 ) Ĥelec φ1s (~rA2 ) φ1s (~rB1 ) dτ
Z
± φ∗1s (~rA1 ) φ∗1s (~rB2 ) Ĥelec φ1s (~rA2 ) φ1s (~rB1 ) dτ
Z #
∗ ∗
± φ1s (~rA2 ) φ1s (~rB1 ) Ĥelec φ1s (~rA1 ) φ1s (~rB2 ) dτ . (5.78)

The integrals in Equation 5.78 can be solved one line at a time for the
Hamiltonian in Equation 5.68 in the following way:
Line 1:

Z  2 
 ∇1 ∇22 1 1 1 1 1 1 
φ∗1s (~rA1 ) φ∗1s (~rB2 ) −

− − − − − + +  φ1s (~rA1 ) φ1s (~rB2 ) dτ
2 2 rA1 rB1 rA2 rB2 R r12 

(5.79)
∇21 1 ∇22 1
The − 2 − rA1 and − 2 − rB2 terms highlighted in red and blue in Equation 5.79
correspond to the Hamiltonian of a H atom. As a result each of these pairs
of terms contribute an energy
Z
E1s φ∗1s (~rA1 ) φ∗1s (~rB2 ) φ1s (~rA1 ) φ1s (~rB2 ) dτ = E1s , (5.80)

129
where E1s is the energy of a H atom in the 1s orbital, to the H2 molecule.
Therefore Equation 5.79 can be simplified to
Z  
1 1 1 1
⇒ E1s + E1s + φ∗1s (~rA1 ) φ∗1s (~rB2 ) − − + + φ1s (~rA1 ) φ1s (~rB2 ) dτ
rB1 rA2 R r12
(5.81)
Z
1
= E1s + E1s + φ∗1s (~rA1 ) φ1s (~rA1 ) φ1s (~rB2 )∗ φ1s (~rB2 ) dτ
R
Z  
1 1 1
+ φ∗1s (~rA1 ) φ∗1s (~rB2 ) − − + φ1s (~rA1 ) φ1s (~rB2 ) dτ (5.82)
rB1 rA2 r12

1
= 2E1s + + J(R) (5.83)
R
where
Z  
1 1 1
J(R) = φ∗1s (~rA1 ) φ∗1s (~rB2 ) − − + φ1s (~rA1 ) φ1s (~rB2 ) dτ
rB1 rA2 r12
(5.84)

is a Coulomb integral representing the interaction of electron 1, in the 1s

orbital around proton A, with proton B and with electron 2 in the 1s orbital
around it.
Line 2:
Z  2 
 ∇1 ∇22 1 1 1 1 1 1 
φ∗1s (~rA2 ) φ∗1s (~rB1 ) −

− − − − − + +  φ1s (~rA2 ) φ1s (~rB1 ) dτ
2 2 rA1 rB1 rA2 rB2 R r12 

(5.85)

is equivalent to Equation 5.79 but with the electron labels changed. It

therefore results in a equivalent contribution to the electronic energy of the
molecule
1
= 2E1s + + J(R). (5.86)
R

Line 3:
Z  2 
 ∇1 ∇22 1 1 1 1 1 1 
φ∗1s (~rA1 ) φ∗1s (~rB2 ) −

− − − − − + +  φ1s (~rA2 ) φ1s (~rB1 ) dτ
2 2 rA1 rB1 rA2 rB2 R r12 

(5.87)

130
 ∇21 1 ∇22 1
As in Equation 5.79, the − 2 − rB1 and − 2 − rA2 terms highlighted in red and
blue in Equation 5.87 correspond to the Hamiltonian of a H atom. As a
result each of these pairs of terms contribute an energy
Z
E1s φ∗1s (~rA1 ) φ∗1s (~rB2 ) φ1s (~rA2 ) φ1s (~rB1 ) dτ = E1s I(R)2 (5.88)

where
Z
I(R) = φ∗1s (~rA1 ) φ1s (~rB1 ) dr1 (5.89)
Z
= φ∗1s (~rB2 ) φ1s (~rA2 ) dr2 (5.90)

are overlap integrals, and E1s is the energy of a H atom in the 1s state.
Therefore Equation 5.87 can be simplified to

⇒ ± E1s I(R)2 + E1s I(R)2
Z   
1 1 1 1
+ φ∗1s (~rA1 ) φ∗1s (~rB2 ) − − + + φ1s (~rA2 ) φ1s (~rB1 ) dτ
rA1 rB2 R r12
(5.91)
 Z
1
= ± E1s I(R)2 + E1s I(R)2 + φ∗1s (~rA1 ) φ1s (~rB1 ) φ1s (~rB2 )∗ φ1s (~rA2 ) dτ
R
Z   
1 1 1
+ φ∗1s (~rA1 ) φ∗1s (~rB2 ) − − + φ1s (~rA2 ) φ1s (~rB1 ) dτ
rA1 rB2 r12
(5.92)
 
1
= ± 2E1s + I(R)2 ± K(R), (5.93)
R
where
Z  
1 1 1
K(R) = φ∗1s (~rA1 ) φ∗1s (~rB2 ) − − + φ1s (~rA2 ) φ1s (~rB1 ) dτ
rA1 rB2 r12
(5.94)

is an exchange integral representing the exchange of the electrons between

the two protons.
Line 4:
Z  2 
 ∇1 ∇22 1 1 1 1 1 1 
φ∗1s (~rA2 ) φ∗1s (~rB1 ) −

− − − − − + +  φ1s (~rA1 ) φ1s (~rB2 ) dτ
2 2 rA1 rB1 rA2 rB2 R r12 

(5.95)

131
is equivalent to Equation 5.87 but with the electron labels changed. It
therefore results in a equivalent contribution to the electronic energy of the
molecule
 
1
= ± 2E1s + I(R)2 ± K(R). (5.96)
R

Combining each of these results, Equation 5.78 becomes


1 1
A± = 2E1s + + J(R)
2 R
1
+2E1s + + J(R)
R
1
±(2E1s + ) I(R)2 ± K(R)
R

1
±(2E1s + ) I(R)2 ± K(R) (5.97)
R
   
1 1
= 2E1s + ± 2E1s + I(R)2 + J(R) ± K(R) (5.98)
R R
  
1
= 2E1s + 1 ± I(R)2 + J(R) ± K(R). (5.99)
R

Returning to Equation 5.75, the normalisation factor, N± , can be written

explicitly as
Z h
1 i∗
N± = φ1s (~rA1 ) φ1s (~rB2 ) ± φ1s (~rA2 ) φ1s (~rB1 ) ×
2
h i
φ1s (~rA1 ) φ1s (~rB2 ) ± φ1s (~rA2 ) φ1s (~rB1 ) dτ (5.100)

"Z
1
= φ∗1s (~rA1 ) φ∗1s (~rB2 ) φ1s (~rA1 ) φ1s (~rB2 ) dr1 dr2
2
Z
+ φ∗1s (~rA2 ) φ∗1s (~rB1 ) φ1s (~rA2 ) φ1s (~rB1 ) dr1 dr2
Z
± φ∗1s (~rA1 ) φ∗1s (~rB2 ) φ1s (~rA2 ) φ1s (~rB1 ) dr1 dr2
Z #
∗ ∗
± φ1s (~rA2 ) φ1s (~rB1 ) φ1s (~rA1 ) φ1s (~rB2 ) dr1 dr2 (5.101)

132
This can be rearranged to give
"Z Z
1 ∗
N± = φ1s (~rA1 ) φ1s (~rA1 ) dr1 φ∗1s (~rB2 ) φ1s (~rB2 ) dr2
2
Z Z
∗
+ φ1s (~rA2 ) φ1s (~rA2 ) dr2 φ∗1s (~rB1 ) φ1s (~rB1 ) dr1
Z Z
± φ∗1s (~rA1 ) φ1s (~rB1 ) dr1 φ∗1s (~rB2 ) φ1s (~rA2 ) dr2
Z Z #
∗ ∗
± φ1s (~rA2 ) φ1s (~rB2 ) dr2 φ1s (~rB1 ) φ1s (~rA1 ) dr1 (5.102)

which contains integrals over the squares of individual wavefunctions

Z
φ∗1s (~rN,i ) φ1s (~rN,i ) dri = 1 (5.103)

and overlap integrals

Z
φ∗1s (~rN,i ) φ1s (~rM,i ) dri = I(R) (5.104)

with the nuclei A and B labelled M and N and the electrons i = 1, 2. Inserting
these values into Equation (5.102) leads to

1h i
N± = 1 · 1 + 1 · 1 ± I(R) · I(R) ± I(R) · I(R) (5.105)
2

= 1 ± I(R)2 (5.106)

Finally, inserting the results in Equation 5.106 and Equation 5.99 into Equa-
tion 5.75 leads to
  
2E1s + R1 1 ± I(R)2 + J(R) ± K(R)
E± = (5.107)
1 ± I(R)2

1 J(R) K(R)
= 2E1s + + ± (5.108)
R 1 ± I(R)2 1 ± I(R)2

The overlap, I(R), Coulomb, J(R), and exchange, K(R), integrals in this
expression have the following dependence on the internuclear distance R:

133
 • Overlap integral, I(R): Because the wavefunctions associated with the
φ1s (~r ) atomic orbital are normalised, the overlap integral
Z
I(R) = φ∗1s (~rN,i ) φ∗1s (~rN,i ) dτ (5.109)
< 1 (5.110)

for all values of R. Therefore in Equation 5.108

1 ± I(R)2 > 0. (5.111)

• Coulomb integral, J(R): In general

J(R) > 0 (5.112)

for all values of R.

• Exchange integral, K(R): In general

K(R) < 0 (5.113)

for all values of R.

Therefore based upon these considerations and Equation 5.108,

- The state ψ S with energy E+ is shifted to lower energy by the correc-

tions associated with the overlap, Coulomb and exchange integrals.

- The state ψT with energy E− is shifted to higher energy by the correc-

tions associated with the overlap, Coulomb and exchange integrals.

The dependence of the energies of these two states on the internuclear

distance can be seen in Figure 5.7.
The most important factor for the stability of a molecule is that the
bonding electrons neutralise the mutual Coulomb repulsion of the nuclei,
i.e., cancel the effective of the repulsive 1/R term in the Hamiltonian (Equa-
tion 5.68). The molecules formed are more stable if there is a significant
probability that the electrons are located in the region between the two
nuclei.
In the triplet state, ψT , the two electrons keep away from each other
because the spatial part of the electronic wavefunction is antisymmetric

134
 T
Energy - 2 E1s (a.u.)

R (a.u.)

Figure 5.7: Gerade, ψT , and ungerade, ψT , electronic potential energy

functions for the ground states of the neutral hydrogen molecule H2 .
The dashed curves represent the results obtained using the procedure
described above, while the continuous curves are the results of exact cal-
culations.

135
with respect to exchange. Consequently, a stable molecule cannot be formed
in this state.
In the singlet state, ψ S , the probability of finding the electrons located
between the two nuclei is non-zero because the spatial part of the electronic
wavefunction is symmetric with respect to exchange of the two electrons.
As can be seen in Figure 5.7 this leads to a minimum in the electronic energy
at an internuclear separation of R ≃ 1.4 a0 about which a stable molecule is
formed.

5.4.2 Molecular orbitals

The singlet, ψ S , and triplet, ψT , wavefunctions (Equation 5.71 and Equa-

tion 5.71) associated with the two lowest energy eigenstates of H2 cannot be
expressed directly as a product of one-electron atomic orbitals. However,
they can be defined in terms of the one-electron molecular orbitals of H+2
(see Section 5.3.1). The wavefunctions associated with these one-electron
molecular orbitals are
1 h i
1 σg = ψ+ = √ φ1s (~rA ) + φ1s (~rB ) (5.114)
2

1 h i
1 σ u = ψ− = √ φ1s (~rA ) − φ1s (~rB ) , (5.115)
2
where 1 σg and 1 σu denote the symmetric ‘gerade’ and antisymmetric
’ungerade‘, one-electron orbitals with principal quantum number n = 1,
and zero electron orbital angular momentum. This is the molecular analog
of the labelling of atomic orbitals as 1s, 2p, . . . . Using this notation, the
ground ψ S state of H2 is denoted

1 σ2g (5.116)

while the excited ψ T state is denoted

1 σg 1 σu . (5.117)

A simplified schematic diagram of the formation of these orbitals at large

internuclear separation and at R = Re can be seen in Figure 5.8. Because the
singlet state, which is in general shifted to lower energy in the molecular

136
 T
Antibonding orbital
-

(rA) (rB)
Energy
1s 1s

S
+
Bonding orbital

Figure 5.8: Schematic diagram of the singlet (bonding), and triplet (anti-
bonding) orbitals of H2 at large large internuclear separation (outer energy
levels) and at R = Re .

bond, corresponds to the stable molecular orbital, ψ S is known as a bonding

orbital. On the other hand, because the triplet state is shifted to higher
energy for all internuclear distances and corresponds to a purely repulsive
molecular potential, ψT is known as an antibonding orbital.
The above approach to constructing molecular orbitals as linear combi-
nations of atomic orbitals can also be extended to heteronuclear diatomic
molecules with two different nuclei A and B. In these cases the molecular
orbitals

ψi = CA φa (~rAi ) + CB φb (~rBi ) (5.118)

differ from those of homonuclear diatomic molecules such as H2 in that there

is no reflection symmetry in the plane perpendicular to the molecular bond
and therefore no designation of ‘gerade’ and ‘ungerade’. Consequently

|CA |2 , |CB |2 . (5.119)

In these heteronuclear diatomic molecules the energy levels of the two

atoms forming the molecular bond are generally not degenerate at large
internuclear separation. Therefore in addition, to mixing of levels with the
same values of ℓ, the molecular bond can also be composed of hybrid orbitals,
e.g., mixtures of s- and p-orbitals.

137
5.5 Nuclear motion

Within the Born-Oppenheimer approximation the total wavefunction of a

molecule with N nuclei and i electrons can be expressed as

~ N ; ~ri ) = ν(R
Ψ(R ~ N ) ψ(R
~ N ; ~ri ), (5.120)

and the solution to the electronic Schrödinger equation gives the potential
in which the nuclear motion takes place.

2.0

1.5
Energy (arb. units)

1.0

Vibrational states
v=3
v=2 D0
0.5
De
v=1
Rotational states
v=0
0.0

0 Re 5 10 15
R (arb. units)

Figure 5.9: Schematic diagram of a typical molecular potential energy

curve in which stable nuclear motion can take place. The energies of the
vibrational (blue - continuous lines) and rotational (red - dashed lines)
states are indicated along with the equilibrium internuclear distance, Re ,
the equilibrium dissociation energy, De , and the dissociation energy from
the lowest vibrational state with v = 0, i.e., D0 .

In general a molecule with N nuclei has 3N nuclear degrees of freedom

arising form three types of nuclear motion.

138
 1. Translational motion: The free motion of the entire molecule in three-
dimensional space, i.e., 3 degrees of freedom.

2. Vibrational motion: The vibrational motion of the nuclei about their

equilibrium position. For a molecule with N nuclei there are 3N − 6
vibrational modes (or 3N − 5 if it is a linear molecule, e.g., a diatomic
molecule).

3. Rotational motion: The rotational motion of the molecule about its

center of mass. In general there are three axes about which this motion
can take place leading to 3 rotational degrees of freedom. However,
for a linear molecule, e.g., a diatomic molecule, there are only two
distinct rotational axes and therefore only two rotational degrees of
freedom.

For a diatomic molecule with nuclei A and B, reduced mass µAB , and an
~ the nuclear Schrödinger equation has the
internuclear separation R = |R|,
form
" #
~2 2
− ∇ + V(R) ν(R) = E ν(R). (5.121)
2µAB R

Writing ∇2R in spherical polar coordinates leads to

Radial
z }| {
" !
~2 2 ~2 1 ∂ 2 ∂
− ∇ = − R +...
2µAB R 2µAB R2 ∂R ∂R
! !#
1 ∂ ∂ 1 ∂2
+ sin θ + . (5.122)
sin θ ∂θ ∂θ sin2 θ ∂φ2
| {z }
Angular

ˆ
Since the square of the angular momentum operator, ~J 2 , in spherical polar
coordinates is
" ! !#
~Jˆ 2 = −~2 1 ∂ ∂ 1 ∂2
sin θ + . (5.123)
sin θ ∂θ ∂θ sin2 θ ∂φ2
The expression in Equation 5.122 can be simplified to

~Jˆ 2
" !#
~2 2 ~2 1 ∂ 2 ∂
− ∇R = − R + . (5.124)
2µAB 2µAB R2 ∂R ∂R 2µAB R2

139
Separating the nuclear wavefunction into radial, F(R), and angular, Y(θ, φ),
parts leads to

F(R)
ν(R) = Y(θ, φ) (5.125)
R
|{z} | {z }
Rot.
Vib.

corresponding to a separation of the vibrational motion and the rotational

motion as indicated. Therefore in terms of the vibrational energy, Evib , and
the rotational energy, Erot , Equation 5.121 becomes
 
~Jˆ 2  F(R)
" !#
 ~2 1 ∂ ∂
−
 R2 + V(R) + 
 Y(θ, φ)
2µAB R2 ∂R ∂R 2µAB R2  R
F(R)
= (Evib + Erot ) Y(θ, φ).
R
(5.126)

140
 A B
x

CM

MA MB

Figure 5.10: Schematic diagram of a rigid rotor with nuclei A and B of

mass, MA and MB , respectively, separated by a fixed internuclear distance
R. The position of nucleus B with respect to the center-of-mass of the
two-body system, CM, is denoted x.

5.5.1 Rotation

The rotational motion of a diatomic molecule is a free rotation and, as in

the case of the orbital angular momentum of the electron in the H atom, the
eigenfunctions that satisfy the rotational part of the nuclear Schrödinger
equation are the spherical harmonic functions, Y J MJ (θ, φ). Therefore

~Jˆ 2
Y J MJ (θ, φ) = Erot Y J MJ (θ, φ), (5.127)
2µAB R2

and the rotational energy of the molecule is

~2
Erot = J(J + 1), (5.128)
2µAB R2

where J is the rotational quantum number which takes values J = 0, 1, 2, . . . .

If the molecule is assumed to behave as a rigid rotor, i.e., the internuclear
distance remains constant and is not affected by the molecule’s rotation, its
moment of inertial, I, can be expressed as

I = MA (R − x)2 + MB x2 , (5.129)

where MA and MB are the masses of the two nuclei, R is the internuclear
distance, and x is the distance of nucleus B from the center of mass of the

141
system (see Figure 5.10). If R = Re , in the frame of reference associated with
the center of mass this becomes

Ie = µAB R2e . (5.130)

and therefore
~2
Erot = J(J + 1) (5.131)
2 Ie

= B J(J + 1), (5.132)

where
~2 ~2
B = = (5.133)
2µAB R2e 2 Ie

is known as the rotational constant. In the form in Equation 5.133, the

rotational constant is given in S.I. units of energy, i.e., joules. However, it is
most common to refer to rotational constants in units of cm−1 . In this case,

1 ~2
B = (5.134)
hc 2µAB R2e
~
= (5.135)
4πc µAB R2e
h
= (5.136)
8π2 c µAB R2e

From this expressions in Equation 5.133 and Equation 5.133 it can be seen
that
1
Erot ∝ (5.137)
µAB

∝ ωrot , (5.138)

where for a fixed rotational quantum number ωrot is the angular rotational
frequency of the molecule. If J = 0 the molecule does not rotate and Erot = 0.

142
5.5.2 Vibration

Using the above solution for the rotational energy of the molecule, the
nuclear Schrödinger equation in Equation 5.126 can be rewritten in the
form
! " #
~2 1 ∂ 2 ∂ F(R) ~2 F(R) F(R)
− R + J(J + 1) + V(R) = E
2µAB R2 ∂R ∂R R 2µAB R2e R R
(5.139)

However,
!
~2 1 ∂ 2 ∂ F(R) ~2 1 ∂2 F(R)
− R = − , (5.140)
2µAB R ∂R
2 ∂R R 2µAB R ∂R2
so Equation 5.141 can be reduced to
" #
~2 1 ∂2 F(R) ~2 F(R) F(R)
− 2
+ 2
J(J + 1) + V(R) = E
2µAB R ∂R 2µAB Re R R
(5.141)

which, by multiplying across by R to remove the factors indicated in red in

the denominator of each term leads to
" #
~ 2 ∂2 ~2
− + J(J + 1) + V(R) F(R) = E F(R),
2µAB ∂R2 2µAB R2e
(5.142)

where
" #
~2 ∂2
− + V(R) F(R) = Evib F(R),
2µAB ∂R2
(5.143)

gives the vibrational energy.

For a stable molecule V(R) has a minimum at R = Re (see Figure 5.9), and
this minimum contains the quantised motional (vibrational and rotational)
states. Because the nuclear motion of the stable molecule is confined around
this equilibrium internuclear position, a first approximation to the form of
V(R) can be obtained considering the molecular potential very close to Re .
To do this the potential V(R) can be expanded in a Taylor series about
R = Re , such that for a general function f (x) expanded about x = a,
(x − a) (x − a)2
f (x) = f (a) + f ′ (a) + f ′′ (a) + .... (5.144)
1! 2!

143
Therefore
 
(R − Re ) dV  (R − Re )2 d2 V 
V(R) = = V(Re ) +  +  + ...
1 dR R=R 2 dR2 R=Re
e
(5.145)

with higher order terms neglected provided the value of (R − Re ) remains

small. However, at the minimum of the potential

dV 
 = 0, (5.146)
dR  R=Re

therefore to second order

1
V(R) = = V(Re ) + k (R − Re )2 , (5.147)
2
where k is a constant. Since the second term of this equation corresponds to
the potential associated with a harmonic oscillator, k is the spring constant.
The Schrödinger equation for the vibrational motion ( Equation 5.148) there-
fore becomes
" #
~2 ∂2 1 2
− + k (R − Re ) +V(Re ) F(R) = Evib F(R),
2µAB ∂R2 2
| {z }
Quantum harmonic oscillator
(5.148)

where V(Re ) is the electronic energy at R = Re .

For a quantum harmonic oscillator the eigenvalues, Ev , are given by

1
Ev = ~ω0 (v + ), (5.149)
2
where v = 0, 1, 2, . . . is the vibrational quantum number. For a harmonic
oscillator with a fundamental vibrational frequency, ω0 ,
s
k
ω0 = , (5.150)
µAB

therefore
s
k 1
Ev = ~ (v + ). (5.151)
µAB 2

144
Form this expression it can be seen that for v = 0,
s
1 k
Ev=0 = ~ , 0. (5.152)
2 µAB

This non-zero energy associated with the vibrational ground state is known
as the zero-point energy of the oscillator.

5.6 Total energy of a diatomic molecule

From above, the total energy, EAB , of a diatomic molecule (rigid rotor) can
be expressed in the harmonic approximation as
s
k 1
EAB = V(Re ) + B J(J + 1) + ~ (v + ) . (5.153)
|{z} | {z } µAB 2
Electronic Rotational | {z }
Vibrational

This expression is valid when:

1. v is small; i.e., near the bottom of the potential well where the harmonic
approximation is reasonable

2. J is small; i.e., when centrifugal distortion, which can occur when the
molecule is rapidly rotating and can cause an increase in the bond
length, does not play a significant role.

However, corrections can be made to the expression in Equation 5.153 to

account for effects of high rotational and vibrational excitation. To improve
the precision with which the rotational energy can be determined:

i. To account for changes in the mean internuclear separation with high

vibrational excitation (high v), the rotational constant, B, can be ad-
justed to be dependent on the vibrational quantum number such that
Z ∞
~2
Bv = F∗ (ν) F(R) dR. (5.154)
0 2µAB R2

ii. For high J the effects of centrifugal distortion of the molecular bond
(2)
can be accounted for be introducing a second order correction, E J ,

145
 1
Harmonic oscillator potential
0
Energy (arb. units)

-1
Morse potential
-2

-3

v=3
-4 v=2
v=1
v=0
-5
0 2 4 6 8
R (arb. units)

Figure 5.11: Schematic diagram of the Morse (continuous lines/curve) and

harmonic oscillator (dashed lines/curve) potentials with their associated
energy level structure.

to the rotational energy which is dependent upon Dv the centrifugal

distortion constant,

(2)
EJ = −Dv J2 (J + 1)2 . (5.155)

Note: Dv is several orders of magnitude smaller than Bv , e.g., in the

ground state of the CO molecule Bv ∼ 2 cm−1 while Dv ∼ 0.000 006 cm−1 .

Combining these two corrections leads to

E J = Bv J(J + 1) − Dv J2 (J + 1)2 . (5.156)

To describe the vibrational energies more precisely it is necessary to

consider how the real molecular potential energy curve deviates form that
of a harmonic oscillator. The main deviations occur for the more highly
excited vibrational states (see Figure 5.11), where the anharmonicity of the

146
true molecular potential becomes significant. As shown in Figure 5.11 a
more appropriate analytic potential energy function which can be used to
model that of a diatomic molecule is the Morse potential, VMorse (R), where
h i2
VMorse (R) = −De + De 1 − e−α(R−Re ) (5.157)

h i
= De e−2α(R−Re ) − 2e−α(R−Re ) , (5.158)

where the constant α is related to the force constant of the harmonic oscil-
lator and can be determined by expanding VMorse (R) in powers of (R − Re ),
i.e.,
h i
VMorse (R) ≃ De −1 + α2 (R − Re ) + . . . . (5.159)

Hence,
r
k
α = . (5.160)
2 De
The Morse potential supports a finite number of vibrational states and as
can be seen in Figure 5.11 these become more closely spaced in energy as
they approach the dissociation limit. In the Morse potential the vibrational
energy of each state can be expressed as
h i
1 ~ ω0 (v + 21 )
Ev = ~ ω0 (v + ) − , (5.161)
2 4 De
where
s
2 De
ω0 = α . (5.162)
µAB

147
5.7 Spectra of diatomic molecules

Molecular spectra are composed of three main kinds of transitions.

1. Pure rotational transitions

- These are transitions which do not involve any change in the

electronic state or the vibrational state of the molecule.
- They are only strong for molecules with non-zero electric dipole
moments with which an oscillating electromagnetic field can in-
teract. Examples of such molecules are LiF and HCl.
- Homonuclear diatomic molecules, e.g, H2 , D2 or H+2 do not un-
dergo pure rotational transitions within the electric dipole ap-
proximation.
- Pure rotational transitions occur at frequencies on the order of
the rotational energy level spacings within a molecule, i.e, on the
same order of magnitude as the rotational constant B. They there-
fore lie in the microwave region of the electromagnetic spectrum
(ν = 1 − 100 GHz).
- The selection rule for pure rotational transitions within the elec-
tric dipole approximation is that

∆J = ±1 (5.163)

for linear molecules, including diatomics. (Note: non-linear

polyatomic molecules can undergo pure rotational transition for
which ∆J = 0).

Since the energy, E J , of rotational state with rotational quantum num-

ber J in a diatomic molecule is

E J = B J(J + 1), (5.164)

the energy difference between the states, J and J′ = J + 1, involved in

a pure rotational transition is

∆E J = E J+1 − E J (5.165)
= B (J + 1)(J + 2) − B J(J + 1) (5.166)
= 2B (J + 1). (5.167)

148
 Therefore as J increases the differences in energy between rotational
states J and J′ , which correspond to the frequencies of the pure rota-
tional transitions, increase by 2B. Pure rotational spectra are therefore
composed of series of sharp spectral lines separated in energy by 2B
(see Figure 5.12). From such a spectrum the rotational constant can be
accurately determined and hence if the reduced mass of the systems
is known the internuclear distance can be deduced.

(a) (b)
....

J=3 EJ = 12B J=1 J’ = 2 J=2 J’ = 3 J=3 J’ = 4

Intensity
Energy

∆E = 2B ....
J=2 EJ = 6B
µ-wave photon energy
J=1 EJ = 2B
J=0 EJ = 0

Figure 5.12: Schematic diagram of (a) the rotational energy level structure,
and (b) a pure rotational spectrum of a heteronuclear diatomic molecule.

2. Vibration-rotation transitions

- These are transitions which do not involve a change in the elec-

tronic state of the molecule, but do involve changes in the rota-
tional and vibrational states.
- Vibration-rotation transitions occur at frequencies on the order
of the vibrational energy level spacings within a molecule, i.e., on
the same order of magnitude as the spring constant k associated
with the molecular bond. They therefore lie in the infrared region
of the electromagnetic spectrum.
- The selection rules for vibration-rotation transitions, within the
electric dipole approximation, are that

∆v = ±1 (5.168)

149
(a) (b)

....
J=4

J=3
J=2
v=1
P-branch R-branch
J=1
J=0 ∆v = +1 ∆v = +1
∆ J = -1 ∆ J = +1
Energy

∆E = 2B

Intensity
.... ....
....

J=4 E= 0

J=3 IR photon energy

J=2
v=0
J=1
J=0

Figure 5.13: Schematic diagram of (a) the rotational and vibrational en-
ergy level structure, and (b) a typical rotation-vibration spectrum of a
heteronuclear diatomic molecule.

and

∆J = ±1. (5.169)

This vibrational selection rule is that associated with a quantum

harmonic oscillator. Along with these transitions in which the
vibrational quantum number changes by ±1, overtone transitions
for which ∆v = ±2, ±3, . . . can also occur but are much weaker.

Rotation-vibration spectra of diatomic molecules typically exhibit two

groupings of lines relating to each allowed change in the rotational
quantum number. One group associated with ∆J = −1 transitions,
and the other associated with ∆J = +1 transitions. These are known
as the P-branch and the R-branch, respectively.

The P-branch transitions, for which ∆J = −1, involve states separated

150
 in energy by, ∆Ero−vib , where

∆Ero−vib = E(v + 1, J − 1) − E(v, J) (5.170)

= −2B J + ~ω0 . (5.171)

where the first term in Equation 5.173 corresponds to the change in

the rotational energy and the second term corresponds to the change
in the vibrational energy.
On the other hand, the R-branch transitions, for which ∆J = +1,
involve states separated in energy by

∆Ero−vib = E(v + 1, J + 1) − E(v, J) (5.172)

= 2B (J + 1) + ~ω0 , (5.173)

The P- and R-branch transitions in rotation-vibration spectra are there-

fore composed of sets spectral lines equally spaced by 2B and located
on either side of the energy difference corresponding to the pure vi-
brational interval, ~ω0 , as indicated in Figure 5.13. In such spectra
the rotational constant of the molecule, B, can be determined from the
spacing between the spectral lines in each of the branches. This then
allows the moment of inertia of the molecule and the equilibrium
internuclear separation to be calculated. The position, ~ω0 , of the
missing line that would correspond to the forbidden ∆J = 0 transition
gives the spring constant.

3. Electronic transitions
Diatomic molecules have axial symmetry and are not spherically sym-
metric as atoms are. As a result, the molecular axis defines a fixed
direction in space onto which the angular parts of the electronic wave-
functions have a constant projection, Lz (where the molecular axis is
defined as the z axis). Consequently, the electronic eigenfunctions of
the molecule are simultaneous eigenstates of Ĥelec and L̂z . In general,
when acting on an angular wavefunction Y(θ, φ), the operator L̂z has
eigenvalues ~ML , i.e.,

L̂z Y(θ, φ) = ML ~ Y(θ, φ), (5.174)

151
where ML = 0, ±1, ±2, . . . , ±L. In this case, where the z axis is chosen
the be the axis of the molecule, the quantum number ML is denoted
Λ such that

L̂z Y(θ, φ) = Λ ~ Y(θ, φ), (5.175)

with Λ defined as the projection of the total electronic orbital angular

momentum onto the internuclear axis.

This value of Λ can then be used in constructing the term symbol

describing the electronic state of a diatomic molecule which has the
form

2S+1
Λ+, −
g, u (5.176)

where S is the total electron spin quantum number, g (gerade) and u

(ungerade) represent the symmetry of the electronic wavefunctions
in the case of homonuclear diatomic molecules, and + and − indicate
the reflection symmetry of the electronic wavefunction in a plane
containing the internuclear axis. In such a term symbol Λ ≡ |Λ|.

In the same way that the letters S, P, D, F,. . . are used to denote the
total electronic orbital angular momentum in the term symbols of
atoms, the value of Λ, is indicated in the molecular term symbol
in Equation 5.176 by the greek letters Σ, Π, ∆, Φ, . . . .

|Λ| = 0 1 2 3 ...
Σ Π ∆ Φ ...

For one-electron molecular orbitals, these labels are written as lower

case characters (as for one-electron atomic orbitals), i.e.,

λ= 0 1 2 3 ...
σ π δ φ ...

The selection rules for allowed electric dipole transitions between

152
 different electronic states are then

∆Λ = 0, ±1 (5.177)
∆S = 0 (5.178)
g → u or u → g (i.e., g 9 g and u 9 u) (5.179)
+ → + or − → − (i.e., + 9 − and − 9 +) (5.180)
∆v = any value (5.181)
∆J = 0, ±1 (5.182)

The energy differences between the electronic states of a molecule

are much larger than those between rotational or vibrational states.
Electronic transitions therefore typically occur in the ultraviolet or
vacuum ultraviolet regions of the electromagnetic spectrum. Because
the equilibrium internuclear distance and rotational constants of dif-
ferent electronic states can be very different, the different rotational
transitions do not necessarily give rise to series of equally spaced
lines as seen in pure rotational spectra or vibration-rotation spectra.
Instead, each electronic transition is split into bands of closely spaced
rotational lines associated with each initial and final vibrational state.

5.7.1 Franck-Condon principle

If it is assumed that the time required for the electronic wavefunction of a

molecule to change when a transition occurs is so fast that the nuclei do
not move in this time, the spectral intensities of the transitions between
each pair of vibrational states of the initial and final electronic states can be
determined from the overlap of their vibrational wavefunctions.
From Fermi’s golden rule (see Section 4.1, Equation 4.4) the probability
of a transition occurring between an initial state, i, and a final state, f, is
Z  2
 ∗ 
Tif =  Ψf Vdip Ψi dτ  (5.183)
Z  2
∗ ˆ

∝  Ψf µ ~ elec Ψi dτ  (5.184)
Z  2
∗ ˆ
 ~
∝  Ψf r Ψi dτ  .

(5.185)

153
For a diatomic molecule with a total wavefunction

~ ~r ) = ν(R)
Ψ(R; ~ ψ(R;
~ ~r ), (5.186)

considering only the vibrational part, F(R ~ ), of the nuclear wavefunction,

~ ), (see Section 5.5) the transition probability can therefore be written as
ν(R
Z  2
Tif ∗ ~ ∗ ~ ˆ ~ ~
∝  Ff (R) ψf (R; ~r ) ~r Fi (R) ψi (R; ~r ) dτ 
 
(5.187)

Z  2 Z  2
 ∗ ~
~ ˆ
~ ~ ~ 
=  ψf (R; r ) r ψi (R; r ) dr  ×
 ∗ ~ ~ 
 Ff (R) Fi (R) dR  (5.188)
| {z } | {z }
Electronic transition moment Franck-Condon factor

where,
Z
~ Fi (R)
Ff∗ (R) ~ dR (5.189)

is known as the Franck-Condon overlap integral (see Figure 5.14), and, as

indicated,
Z  2
 ∗ ~ ~ 
 Ff (R) Fi (R) dR  (5.190)

is the Franck-Condon factor.

5.8 Ionically bound molecules

The homonuclear diatomic molecules discussed above are covalently bound,

with electrons shared equally between the two nuclei. They therefore do not
exhibit any electric dipole moments, i.e, the electron charge is distributed
equally between the two nuclei, or pure rotational spectra. On the other
hand, molecules that exhibit strong pure rotational spectra must have a
~ elec , i.e., an uneven distribution of elec-
significant electric dipole moment, µ
tron charge between the two nuclei with one end of the molecular bond
more negatively charged than the other [see for example Figure 5.15(a)].
The interaction of this electric dipole moment with an electromagnetic field
gives rise to a torque on the molecule and a change in its rotational state.

154
 The heteronuclear isotopomers of the molecular hydrogen cation and
neutral hydrogen molecule, e.g., HD+ , HT+ , HD, and HT have small elec-
tric dipole moments, because of the slightly different masses of the two
nuclei, and very weak pure rotational spectra. However, ionically bound
molecules, e.g, HCl, HF, or LiF have significant electric dipole moments (on
the order of 3.33 × 10−30 Cm ≡ 1 debye, or 1 D) and therefore strong pure
rotational spectra.
Ionic bonds are most readily formed between atoms one of which has
one electron outside an otherwise closed shell, e.g., the alkali metal atoms
Li, Ns, Rb, Cs, and the other of which is one electron short of a closed shell,
e.g, the halogen atoms F, Cl, Br, I.

5.8.1 Dissociation energy of alkali metal halides

The form and strength of an ionic bond can be seen by considering the
electron affinity, A, of the atoms involved. This is the binding energy of a
single electron attached to the neutral atom to form a negative ion (anion).
In the case of Na and Cl, e.g,

Na + e− → Na− (5.191)
Cl + e− → Cl− (5.192)

the electron affinities are

ANa
= 0.54 eV (5.193)
e
ACl
= 3.62 eV, (5.194)
e
indicating that an additional electron would be bound more tightly to the
Cl atom than to the Na atom. On the other hand, the ionisation energies, I,
of the two atoms are
INa
= 5.14 eV (5.195)
e
ICl
= 12.97 eV. (5.196)
e
From these values of the electron affinities and ionisation energies it can be
seen that energy is required to remove an electron from the Na atom and

155
attach it to the Cl atom, but this energy is less than that required to remove
an electron from the Cl atom and attach it to the Na atom. This energy, E, is

E = INa − ACl (5.197)

≡ 1.52 eV. (5.198)

However, if this electron transfer is achieved the resulting Na+ and Cl−
ions will interact via an attractive Coulomb interaction. The interatomic
(or internuclear) distance, R, for which the this Coulomb energy matches
the energy loss associated with the transfer of the electron between the two
atoms can therefore be determined to be
e2
EC = (5.199)
4πǫ0 R
≡ 1.52 eV. (5.200)

Therefore,

R = 9.47 × 10−10 m (5.201)

= 17.9 a0 , (5.202)

and this is approximately the equilibrium internuclear distance at which a

stable, ionically bound, NaCl molecule will form.
The true equilibrium internuclear distance in the ground electronic state
of NaCl, determined from measurements of the rotational constant, is Re =
4.5 a0 , and at this separation the energy, EC , associated with the Coulomb
attraction between the Na+ and Cl− ions is
EC (Re )
= 6 eV. (5.203)
e
This is therefore the energy required to separate the two nuclei and produce
two atomic ions. However, considering the energy saved by transferring
the electron from the Cl− back to the Na+ in the dissociation process, the
dissociation energy, De , of NaCl to from two neutral ground state atoms is

De = EC (Re ) − (INa − ACl ) (5.204)

≡ 4.48 eV. (5.205)

This simple argument leads to a dissociation energy for NaCl that is very
close to the true value of De ≡ 4.26 eV. The causes of the discrepancy
between these values are associated with

156
 1. The assumption that the Na+ and Cl− ions are point charges, while
in a more complete description of the molecule it is necessary to take
into account their finite sizes.

2. Not accounting for the effects of the overlap of the electronic wave-
functions of the two ions on the energy of the system. At small
internuclear distances this will raise the energy of the molecule.

5.8.2 Ionic character of a bond

The ionic bonds in molecules are not purely ionic, in that the bonding
electron is not completely localised at of the nuclei. If the bond was purely
ionic the electric dipole moment of the molecule would be

µelec = e Re . (5.206)

In NaCl, with an equilibrium internuclear distance of Re = 4.54 a0 = 2.40 ×

10−10 m, a purely ionic bond would lead to a dipole moment of

µelec = 3.8 × 10−29 Cm (5.207)

= 11.5 debye. (5.208)

However, the measured electric dipole moment is

µexpt = 9.0 debye. (5.209)

Therefore there must be a non-zero probability of finding the electron lo-

cated close to the Na+ atom. In such a situation the fractional ionic character,
fionic , of the bond can be defined as
µexpt
fionic = (5.210)
e Re

= 0.78 for NaCl. (5.211)

In this way:

If fionic = 1 ⇒ Completely ionic bond (5.212)

If fionic = 0 ⇒ Completely covalent bond (5.213)

The ionic character of the bonds in the ground electronic states of a selection
of alkali metal halide molecules are listed in Table 5.1

157
 Re (10−10 m) µelec = e Re (debye) µexpt (debye) fionic
LiF 1.56 7.49 6.32 0.84
NaCl 2.40 11.52 9.00 0.78
KBr 2.82 13.54 10.62 0.78

Table 5.1: Ionic character of the bonds in the ground electronic states of
the alkali metal halide molecules LiF, NaCl and KBr.

Electronic state B

v' = 3
v' = 2
v' = 1
v' = 0
Energy

Electronic state A
v=3
v=2
v=1
v=0

0.6 0.8 1 1.2 1.4 1.6

Internuclear distance (arb. units)
Figure 5.14: The overlap (shaded regions) of the v = 0 and v′ = 0 vi-
~ ), of two electronic states, labelled A and B,
brational wavefunctions, F(R
respectively.

158
 (a) (b)
A B

-q +q

Nucleus µelec = q d

Electron charge
distribution

Figure 5.15: (a) Schematic diagram of an ionically bound molecule AB in

which the electron charge (shaded blue regions) is localised more closely
toward one end of the molecular bond (close to nucleus B) than the other.
~ elec , associated with two charges, −q and
(b) The electric dipole moment, µ
~
+q, separated by a distance d.

159
 Part III

Atoms in electric and magnetic

fields

160
Chapter 6

Introduction

Atoms can be very sensitive to electric and magnetic fields. These fields
cause energy shifts and energy splittings of the field-free electronic struc-
ture. As a result, samples of atoms can be used as very sensitive probes of
electric and magnetic fields. However, in atomic clocks, and when atoms
are used in tests of fundamental physics, this sensitivity means that it is
essential to control and minimise these fields. In addition, electric and
magnetic fields can be used to accelerate, decelerate and trap neutral atoms
and antiatoms.

6.1 Hamiltonian operator

The Hamiltonian operator, Ĥ, for an isolated atom with many electrons in
the absence of external electric or magnetic fields can be expressed as

Ĥ = Ĥ0 (~r1 , ~r2 , . . . , ~ri ) + Ĥ SO , (6.1)

where Ĥ SO accounts for the spin-orbit interaction in the atom.

In the presence of an external magnetic field B, ~ and electric field F,
~ the
Hamiltonian operator becomes

Ĥ = Ĥ0 + Ĥ SO + VB + VF , (6.2)

where VB and VF represent the interactions with the magnetic and electric
fields, respectively. The interaction of the atom with the magnetic field

161
is known as the Zeeman interaction as it was first observed by the Dutch
physicist Pieter Zeeman in 1896. The interaction with the electric fields is
known as the Stark interaction after the early experimental work by Johannes
Stark on its observation.

6.1.1 First order perturbation theory

When contributions from the applied electric and magnetic fields are small
compared to the field-free Hamiltonian, i.e., when

Ĥ0 ≫ Ĥ SO , VB and VF , (6.3)

the fields can be considered to only perturb the field-free electronic struc-
ture. In this situation the energy shifts arising from the interactions with the
external fields are small compared to the energy spacings between states
with different values of n in the atom, and the energy shift, ∆E, of a state ψ0
is given by the expectation value of the perturbing potential, V, i.e,
Z
∆E ≃ ψ∗0 V ψ0 dτ. (6.4)

The total energy, Etot , of the state in the presence of the field is then

Etot = E0 + ∆E, (6.5)

where

Ĥ0 ψ0 = E0 ψ0 , (6.6)

and

E0 ≫ ∆E. (6.7)

162
Chapter 7

Atoms in magnetic fields

7.1 Introduction

The Zeeman effects arises from the interaction of the magnetic dipole mo-
~ The interaction
~ mag , with an external magnetic field, B.
ment of an atom, µ
potential, VB , is therefore

VB = −~ ~
µmag · B. (7.1)

Within the atom there are two sources of magnetic dipole moments. The
~ S , and
magnetic dipole moment associated with the total electron spin, µ
the magnetic dipole moment associated with the total electron orbital an-
~ L . As in the treatment of the spin-orbit interaction
gular momentum, µ
in Section 3.7,
~
S
~ S = −ge µB ,
µ (7.2)
~
where ge ≃ 2 is the electron spin g-factor, and µB is the Bohr magneton, and
~L
~ L = −gL µB ,
µ (7.3)
~
where gL = 1 is the orbital g-factor. In the following the Zeeman effect will
be considered in three different situations:

1. For states in which the total electron spin is zero, i.e., S = 0.

2. For states in which S , 0 and the Zeeman interaction (energy shift) is

small compared to the spin-orbit energy splitting.

163
 3. For states in which S , 0 and the Zeeman interaction (energy shift) is
large compared to the spin-orbit energy splitting.

7.2 Zeeman effect when S = 0

In terms for which S = 0, an atom exhibits no magnetic dipole moment

~ S = 0, therefore
associated with its total electron spin, i.e, µ

VB = −~ ~
µL · B. (7.4)

~ = (0, 0, Bz ) acting in the z-direction

For a magnetic field B
µB
VB = ~L · B
~ (7.5)
~

µB
= Lz Bz . (7.6)
~
Therefore the resulting Zeeman energy shift, ∆EB , is
Z
∆EB = ψ∗0 VB ψ0 dτ (7.7)

Z
µB
= ψ∗0 L̂z Bz ψ0 dτ (7.8)
~
Z
µB
= Bz ψ∗0 L̂z ψ0 dτ. (7.9)
~
However, from the properties of the angular momentum operators

L̂z ψ0 = ~ML ψ0 , (7.10)

so
Z
∆EB = µB ML Bz ψ∗0 ψ0 dτ (7.11)
= µB ML Bz , (7.12)

provided that the field-free wavefunctions, ψ0 , are normalised.

In this situation, when S = 0, each term with total orbital angular mo-
mentum quantum number L is therefore split into 2L + 1 equally spaced

164
 (a) Helium n 1P (b) Helium n 1D
5 5
4 4 ML = +2
3 3
Energy / hc (cm-1)

Energy / hc (cm-1)
ML = +1 ML = +1
2 2
1 ML = 0 1 ML = 0
0 0
-1 -1
-2 -2
-3 ML = -1 -3 ML = -1
-4 -4 ML = -2
-5 -5
0 1 2 3 4 5 0 1 2 3 4 5
Magnetic field strength (T) Magnetic field strength (T)

Figure 7.1: The Zeeman effect in the absence of spin, i.e., when S = 0, in
(a) the n 1 P term, and (b) the n 1 D term in He.

sublevels by the magnetic field. This can be seen in Figure 7.1 in the case of
the singlet-P and singlet-D (S = 0) terms in He.
In the absence of electron spin the electric dipole selection rules for
transitions between Zeeman sublevels require that

∆ML = 0, ±1. (7.13)

Therefore each spectral line is split into three components, one associated
with the ∆ML = −1 transitions, one associated with the ∆ML = 0 transitions,
and one associated with the ∆ML = +1 transitions, as indicated in Figure 7.2.
This splitting of a single spectral line into three components when S = 0 is
often referred to as the normal Zeeman effect.

7.3 Zeeman effect when S , 0 (weak-field regime)

For terms in which S , 0, the magnetic dipole moments associated with the
electron spin and orbital angular momentum must both be accounted for
when treating the interaction of the atom with a magnetic field. In the weak-
field limit, i.e., when the applied magnetic field is weaker than the internal
magnetic fields in the atom and the Zeeman energy shift is therefore smaller

165
 (a) (b)
S=0 “Normal Zeeman effect”
ML = +1
2 1P 1 ML = 0
ML = -1

ML = 0
Energy

Intensity
ML = +1

ML = -1
ML = 0 ML = +1 ML = -1

Wavenumber

1 1S 0 ML = 0
Zero field B-field

Figure 7.2: Schematic diagrams (a) indicating the electric dipole allowed
transitions between Zeeman sublevels when S = 0, and (b) the correspond-
ing spectral intensity distributions.

than the spin-orbit energy splitting, the spin-orbit interaction couples ~L and
~ so that
S

~
~J = ~L + S. (7.14)

~ vectors,
If the external magnetic field is too weak to uncouple the ~L and S
ML and MS are not good quantum numbers with which to describe the
system. This must be taken into account when calculating the effect of the
magnetic field on the atomic energy level structure.

166
 In this case

~ −µ
µS · B
VB = −~ ~
~L · B (7.15)

~ + gL~L ) µB
= (ge S ~
·B (7.16)
~

µB ~)
= ~ · (~L + 2S
B (7.17)
~

µB ~ ).
= ~ · (~J + S
B (7.18)
~
~ = (0, 0, Bz ), the energy shift resulting from the Zeeman interaction is
If B
given by the expectation value of VB , so
Z Z
µB µB ~ ψ0 dτ
∆EB = ψ∗0 Bz Ĵz ψ0 dτ + ψ∗0 ~ ·S
B (7.19)
~ ~
Z
µB ~ ψ0 dτ.
= µB M J Bz + ψ∗0 ~ ·S
B (7.20)
~
~ precesses about the total
In the LS-coupling regime the total spin vector S
orbital angular momentum vector ~J with a constant projection S J along the
vector ~J. However, it does not have a constant projection along B.
~ This can
be seen schematically in Figure 7.3. In this situation the projection S J is give
by1
~ · ~J
S
SJ = . (7.21)
|~J|
This can be written in vector form by multiplying by the unit vector in the
direction defined by ~J so that
~ · ~J
S ~J
~J =
S , (7.22)
|~J | |~J |
and when written in operator form

~ˆ · ~Jˆ ˆ
S
~ˆ J =
S ~J. (7.23)
~Jˆ 2
1
Note: For two vectors ~a and ~b with an angle θ between them, ~a · ~b = | ~a | | ~b | cos θ, so the
projection of ~a onto ~b is | ~a | cos θ = ~a · ~b / | ~b |.

167
 B

L J
SJ

Figure 7.3: Schematic diagram illustrating the precession of the total an-
gular momentum vector ~J, in a weak magnetic field B.
~ In this situation, the
~ precesses about ~J with a constant projection S J indicated
total spin vector S
by the dashed red line.

ˆ ˆ ~ˆ
Then using the fact that ~J = ~L + S can be rearranged such that

~ˆ
~Lˆ = ~Jˆ − S, (7.24)

and therefore

~ˆ 2 − 2 ~Jˆ · S,
~Lˆ 2 = ~Jˆ 2 + S ~ˆ (7.25)

and
 
~ˆ =
~Jˆ · S 1 ~ˆ 2 ~ˆ 2 ~ˆ 2
J +S −L . (7.26)
2
~ J has a constant projection on the external magnetic field B,
Since the vector S ~
the total energy shift resulting from the Zeeman interaction (Equation 7.20)

168
can be expressed as
Z
µB ~ J ψ0 dτ
∆EB = µB M J Bz + ~ ·S
ψ∗0 B (7.27)
~

µB
Z ~ · ~J
S
= µB M J Bz + ~·
ψ∗0 B ~J ψ0 dτ (7.28)
~ |~J |2
 
µB
Z ~ˆ 2 − ~Lˆ 2 
 ~Jˆ 2 + S
= µB M J Bz + ψ∗0 Bz   J ψ dτ
 z 0 (7.29)
~ ˆ2
~
2J
" #
J(J + 1) + S(S + 1) − L(L + 1)
= µB M J Bz + µB M J Bz (7.30)
2J(J + 1)

= µB M J g J Bz , (7.31)

where
" #
J(J + 1) + S(S + 1) − L(L + 1)
gJ = 1 + (7.32)
2J(J + 1)

is known as the Landé g-factor.

An example of the weak-field Zeeman effect when S , 0 can be seen
in Figure 7.4 for the case of the 2 2 P term in the H atom. For the range of
magnetic fields considered in this figure, the Zeeman energy shifts of all of
the M J sublevels are smaller than the spin-orbit energy splitting, ∆ESO .
For atomic levels with non-zero total electron spin the electric dipole
selection rules for transitions between Zeeman sublevels require that

∆M J = 0, ±1. (7.33)

This leads to sets of allowed electric dipole transitions such as those in-
dicated in Figure 7.5(a) in the case of the 3 2 P1/2 and 3 2 P3/2 levels in the
Na atom. The transitions indicated in this figure between levels for which
∆J = 0 give rise to spectra of the kind depicted in Figure 7.5(b). The form of
such spectra, with two pairs of equally spaced lines separated by a larger
interval, has led to this weak-field Zeeman effect when S , 0 being known
as the anomalous Zeeman effect.

169
 0.3 ∆EB = B
MJ gJ Bz
MJ = +3/2
0.2 2
2 P3/2
MJ = +1/2
Energy / hc (cm-1)

0.1 MJ = -1/2

0 MJ = -3/2
∆ESO
-0.1

-0.2 MJ = +1/2

-0.3 2 2P1/2 MJ = -1/2

0 0.02 0.04 0.06 0.08 0.1

Magnetic field strength (T)

Figure 7.4: The weak-field Zeeman effect in the 2 2 P1/2 and 2 2 P3/2 levels of
the H atom. In zero magnetic field the two levels are separated in energy
by the spin-orbit splitting, ∆ESO .

(a) (b)
S 0 “Anomalous Zeeman effect”
Na MJ = +3/2
2P MJ = +1/2
3 3/2 MJ = -1/2
MJ = -3/2

3 2P1/2 MJ = +1/2
MJ = -1/2
MJ = 0 MJ = 0
Energy

Intensity

MJ = +1 MJ = -1
MJ = 0, ±1

Wavenumber

MJ = +1/2
3 2S1/2 MJ = -1/2
Zero field Weak B-field

Figure 7.5: Schematic diagrams (a) indicating the electric dipole allowed
transitions between Zeeman sublevels when S , 0 in the Na atom, and (b)
the corresponding spectral intensity distributions.

170
7.4 Zeeman effect when S , 0 (strong-field regime)

In strong magnetic fields, when

VB ≫ ĤSO (7.34)

and therefore

∆EB ≫ ∆ESO , (7.35)

the external magnetic field can be considered larger than the internal mag-
netic fields of the atom and the total orbital angular momentum vector ~L
~ decouple from each other and precess
and the total electron spin vector S
independently about the magnetic field. Therefore in this case
~ −µ
µS · B
VB = −~ ~
~L · B (7.36)

µB ~ µ
= ~ + B gL ~L · B
ge S · B ~ (7.37)
~ ~

µB ~
= ~ · (~L + 2S).
B (7.38)
~
~ = (0, 0, Bz ),
As above, if B
µB
VB = Bz (Lz + 2 Sz ), (7.39)
~
and therefore
Z
µB
∆EB = ψ∗0 Bz (L̂z + 2 Ŝz ) ψ0 dτ (7.40)
~

= µB Bz (ML + 2MS ). (7.41)

~ is
This strong-field regime, in which the magnetic field decouples ~L and S,
known as the Paschen-Back regime.
A schematic diagram of the energy level structure of the lowest lying
energy levels in the Na atom in the Paschen-Back regime can be seen in Fig-
ure 7.6 (right hand side). In this regime the electric dipole selection rules
for transitions between Zeeman sublevels require that

∆ML = 0, ±1 (7.42)
∆S = 0, (7.43)

171
 S 0 ML + 2MS
+2 EB = + 2 µB Bz

+1 EB = + µB Bz
3 2P3/2
0 EB = 0

3 2P1/2 -1 EB = - µB Bz
Energy

-2 EB = - 2 µB Bz

+1 EB = + µB Bz

3 2S1/2
Zero field Weak B-field -1 EB = - µB Bz
Strong B-field
Paschen-Back regime

Figure 7.6: Schematic diagrams of the changes in atomic energy level

structure when S , 0 in weak and strong magnetic fields. The red arrows
indicate the electric dipole allowed transitions in the Paschen-Back regime.

and, for the case of the ground and first excited states of Na, are indicated
by the red arrows in Figure 7.6.

7.5 The Stern-Gerlach experiment

The seminal experiments of Gerlach and Stern in the early 1920s in which
space-quantisation was demonstrated exploited beams of ground-state sil-
ver (Ag) atoms. The ground level of Ag is a 2 S1/2 level, with S = 1/2, L = 0
and J = 1/2. This term exhibits no spin-orbit splitting and so there is no
distinction between the Zeeman effect in strong and weak fields with the

172
 (a) (b)
Oven

N
Inhomogeneous
S magnetic field

Photographic plate

Figure 7.7: (a) Schematic diagram of the apparatus used by Stern and
Gerlach to deflect/split beams of silver atoms into their different Zeeman
components using inhomogeneous magnetic fields. (b) Images recorded
by Gerlach and Stern of the undeflected beam (bottom panel) and the
deflected/split beam (bottom panel)2 .

result that

∆EB = µB g J M J Bz (7.44)
= µB ge MS Bz (7.45)
= 2µB MS Bz . (7.46)

In Gerlach and Stern’s experiments, forces were exerted on the atoms in the
M J = MS = ±1/2 Zeeman sublevels by generating inhomogeneous mag-
netic fields with strong gradients transverse to the direction of propagation
of the atomic beam (see Figure 7.7).
~ its potential energy,
When an atom is subjected to a magnetic field B,
V, in the field is given by the Zeeman energy shift, ∆EB , of the sublevel in
which it is prepared. If the magnetic field is inhomogeneous, the potential
energy of the atom is therefore position dependent and at any position in
the field the atom experiences a force, f~, equal to the negative gradient of
2
W. Gerlach and O. Stern, Der experimentelle Nachweis des magnetischen Moments des
Silberatoms, Z. Phys. 8, 110 (1921); W. Gerlach and O. Stern, Der experimentelle Nachweis
der Richtungsquantelung im Magnetfeld, Z. Phys. 9, 349 (1922); W. Gerlach and O. Stern,
Das magnetische Moment des Silberatoms, Z. Phys. 9, 353 (1922)

173
the potential, i.e.,

f~ = −∇V. (7.47)

Considering only the z-component of the field gradient, the force exerted
on the atom in the z-dimension, fz , is therefore

∂V
fz = − (7.48)
∂z

∂Bz
= −2µB MS . (7.49)
∂z
Hence, an equal but opposite force will be exerted on the MS = +1/2
and MS = −1/2 Zeeman sublevels causing the beam to split as seen in
the experiments of Gerlach and Stern (Figure 7.7). In the case depicted
in Figure 7.7(a) where the magnetic field is very strong close to the upper
magnet pole piece, and weak close to the lower magnet pole piece the
magnetic field gradient is greatest in the vertical direction. As a result
atoms in the MS = +1/2 sublevel will be forced downwards, and atoms in
the MS = −1/2 sublevel will be forced upwards splitting the beam into its
two Zeeman components as seen in the lower panel in Figure 7.7(b).

7.6 Multistage Zeeman deceleration of atomic beams

Forces of the same origin as those in the experiments of Gerlach and Stern
can also be exploited to decelerate beams of neutral atoms if the magnetic
field gradients are generated along the axis of propagation of the beams.
This can be achieved if a beam of atoms in a sublevel with M J > 0 enters
an electromagnet travelling along the axis of the coil (see Figure 7.8). As an
atom in such a state travels into the coil, it experiences a positive magnetic
field gradient on the axis of the coil and therefore, if the axis of the coil is
chosen to lie along the z-axis, a decelerating force as given by Equation 7.49.
Because the energy of the Zeeman sublevel of the atom increases as the
atom enters the field, the atoms must lose kinetic energy. If the coil is left on
the atoms would then accelerate out the other side when they experience
the negative magnetic field gradient. However, if the current flowing in

174
the coil is switched off rapidly when the atoms have reached the magnetic
field maximum, they will only experience the decelerating force and will
therefore be slowed down.
Typically in experiments of the kind depicted in Figure 7.8 beams of H
atom can be generated with initial speeds of 500 m/s and magnetic fields of
up to 2 T can be generated in the coils of the decelerator with gradients on
the order of 2 T/cm. At this speed the atoms have an initial kinetic energy
of
1
Ekin = MH v2 (7.50)
2
≃ 2.0 × 10−22 J. (7.51)

However, the maximum Zeeman shift of the M J = MS = +1/2 sublevel in

the magnetic field of one decelerator coil is

∆EB = 2µB MS Bz (7.52)

≃ 2.0 × 10−23 J. (7.53)

Therefore all of the kinetic energy of an individual atom cannot be removed

in one coil (stage) of the decelerator. Therefore

Ekin
= 10 (7.54)
∆EB
coils (stages) are required to completely decelerate the atoms from 500 m/s
to zero velocity in the laboratory-fixed frame of reference. A device of the
kind depicted in Figure 7.8 can be employed to achieve this. The decelerated
atoms can then be trapped in a magnetic quadrupole trap at the end of the
decelerator (see Section 4.4).

3
N. Vanhaecke, U. Meier, M. Andrist, B. H. Meier, and F. Merkt, Multistage Zeeman
deceleration of hydrogen atoms, Phys. Rev. A 75, 031402(R) (2007); S. D. Hogan, A. W.
Wiederkehr, H. Schmutz, and F. Merkt, Magnetic Trapping of Hydrogen after Multistage
Zeeman Deceleration, Phys. Rev. Lett. 101, 143001 (2008).

175
Figure 7.8: Schematic diagram of a multistage Zeeman decelerator used
for decelerating beams of H and D atoms3 .

176
Chapter 8

Atoms in electric fields

8.1 Introduction
~ on the
The Stark effect, i.e., the effect of an externally applied electric field, F,
energy level structure of an atom, arises from the interaction of the electric
~ elec with the field.
dipole moment of the atom µ
For two equal but opposite charges, q1 = +e and q2 = −e, separated by a
distance ~r (see Figure 8.1), the corresponding electric dipole moment is

~ elec = −e ~r.
µ (8.1)

In the presence of the field the potential energy, VF , of this dipole is

VF = −~ ~
µelec · F. (8.2)

q1 = +e q2 = -e

elec
= -e r
~ elec , associated with two charges
Figure 8.1: The electric dipole moment, µ
separated by a distance ~r.

177
 ~ = (0, 0, Fz ), considering
If the electric field is acts in the z-direction, i.e., if F
the expression in Equation 8.1, this then becomes

VF = e Fz z. (8.3)

In the following we will consider two regimes. one in which the effect
of the electric field on the atom is weak and the Stark shifts are small
compared to the energy differences between the unperturbed states with
different values of n, i.e.,

e2
VF ≪ . (8.4)
4πǫ0 r
In this case, results from perturbation theory can be applied to determine
the effects of the field on the energy level structure. The second regime will
be that in which the electric field strongly modifies the potential experienced
by the electron in the atom leading to ionisation, i.e.,

e2
VF ≫ . (8.5)
4πǫ0 r
This regime is reached for low values of n in very strong electric fields, or
for high values of n in weaker fields.

8.2 Stark effect in weak fields

In general weak electric fields give rise to shifts and splittings of atomic
energy levels. Energy levels with different orbital angular momentum
quantum numbers that are degenerate in energy exhibit linear Stark shifts,
i.e., ∆EF ∝ F, while non-degenerate levels exhibit quadratic Stark shifts, i.e.,
∆EF ∝ F 2 .

8.2.1 Quadratic Stark effect

In situations where an atom has no intrinsic electric dipole moment, i.e., in

states with no ℓ-degeneracy, linear Stark energy shifts do not occur. In such
cases quadratic Stark shifts occur. This can be understood by calculating the
expectation value of the Stark Hamiltonian operator, VF (see Equation 8.3),

178
for a non-degenerate state. Consider, for example, the non-degenerate
1 2 S1/2 ground state of the H atom with the wavefunction ψnℓm = ψ100 ,
Z
∆EF = ψ∗100 e Fz ẑ ψ100 dτ (8.6)
Z
= e Fz ψ∗100 ẑ ψ100 dτ (8.7)
|{z} |{z} |{z}
even odd even
= 0. (8.8)

This indicates that the part of the operator associated with the electric dipole
moment, µ̂elec = eẑ, is zero and in this state the atom has no intrinsic electric
dipole moment.
However, the electric field can polarise the atom by redistributing the
charge of the electron. This leads to an induced electric dipole moment the
magnitude and direction of which depends on the magnitude and direction
of the electric field vector,

µ ~
~ elec ∝ F. (8.9)

In this case, the energy shift of a non-degenerate state in an electric field

can be determined by applying second order perturbation theory. For a
perturbing potential V, the energy shift of a state, ψnℓm , is given to second
order by perturbation theory as
R 2
Z X  ψ∗n′ ℓ′ m′ V ψnℓm dτ
∆EV = ψ∗nℓm V ψnℓm dτ + . (8.10)
Enℓm − En′ ℓ′ m′
n′ ℓ′ m′ , nℓm

Therefore in this case

R 
X X  ψ∗ ′ ′ ′ ẑ ψnℓm dτ 2
n ℓ m
∆EV = e2 F2z . (8.11)
n′
Enℓm − En′ ℓ′ m′
ℓ=ℓ±1

The restrictions on the range of values of ℓ and ℓ′ in the sum arises from the
form of the ẑ operator. Since

z = r cos θ (8.12)
r
4π
= r Y1 0 , (8.13)
3

179
 -109676
H atom
Energy / hc (cm-1) -109677

-109678
1s

∆EF = - –12 F
2
-109679

-109680
0 2 4 6 8 10
Electric field strength (108 V/m)

Figure 8.2: The quadratic Stark effect in the 1 2 S1/2 ground state of the H
atom.

the integral in the sum only has non-zero values when the values of ℓ and
ℓ′ differ by 1 and the values of m and m′ are equal.
In the ground state of the H atom
R 
X  ψ∗ ẑ ψ100 dτ 2
n10
∆EF = e2 Fz 2
(8.14)
n′
E100 − En′ 10
1
= − α F2z , (8.15)
2
where
R 
X  ψ∗ ẑ ψ100 dτ 2
n10
α = −2e2 (8.16)
n′
E 100 − En′ 10

is the electric dipole polarizability. Since

1
∆EF = − α F2z (8.17)
2
 
1
= − αFz Fz , (8.18)
2
comparison with Equation 8.2 indicates that αFz /2 is the induced electric
dipole moment of the state in the presence of the field. For the ground state

180
of the H atom

α = 7.42 × 10−41 C m2 V−1 , (8.19)

and the corresponding dependence of the ground state energy on the

strength of the applied electric field can be seen in Figure 8.2.
The quadratic Stark shift generally lowers the energy of an isolated
non-degenerate energy level. This is because the electric dipole moment
induced by the field is always oriented parallel to the field vector.

8.2.2 Linear Stark effect

If an atom possesses an intrinsic electric dipole moment, which is the case

when the electron is in an orbital with a value of the orbital angular mo-
mentum quantum number ℓ that is degenerate in energy with other orbitals
with different values of ℓ, linear Stark energy shifts occur. For example, in
excited states of the H atom for all n > 1, levels with all allowed values of ℓ
are degenerate, as a result excited states of the H atom always exhibit linear
Stark shifts. As in the case of the quadratic Stark shift, the Stark Hamilto-
nian operator (see Equation 8.3) mixes states according to the electric dipole
selection rules that require that ∆ℓ = ±1 and ∆m = 0.
Consider the states with n = 2 in the H atom (see Figure 8.3). Because
of the selection rules associated with the Stark Hamiltonian operator, the
field can only mix the 2s state for which mℓ = 0 (i.e., the state ψ200 ) with the
2p state for which mℓ = 0 (i.e., the state ψ210 ). But it will not mix (affect)
the 2p states for which mℓ = ±1. Therefore the eigenstates of the complete
Hamiltonian

Ĥ = Ĥ0 + VF , (8.20)

expressed in terms of the field-free wavefunctions, must take the form

1

ψ1 = ψ+ = √ ψ200 + ψ210 (8.21)
2
1

ψ2 = ψ− = √ ψ200 − ψ210 (8.22)
2
ψ3 = ψ21 −1 (8.23)
ψ4 = ψ21 +1 . (8.24)

181
 2s 2p 2p 2p
ml = 0 ml = -1 ml = 0 ml = +1

Energy

1s
ml = 0

Figure 8.3: Energy levels in the H atom for which n = 1 and n = 2.

Energy shifts

To determine the energy shifts of the ψ± states in the field first order pertur-
bation theory can be applied and the expectation value of VF determined,
i.e.,
Z
∆EF = ψ∗± VF ψ± dτ (8.25)

Z
1
∗

= ψ200 ± ψ210 eFz ẑ ψ200 ± ψ210 dτ (8.26)
2
"Z Z
1
= eFz ψ∗200 ẑ ψ200 dτψ∗210 ẑ ψ210 dτ + . . .
+
2
Z Z #
∗ ∗
± ψ200 ẑ ψ210 dτ ± ψ210 ẑ ψ200 dτ . (8.27)

Expressing each of the wavefunctions as a product of a radial wavefunction

and a spherical harmonic function, such that

ψ200 = R20 (r) Y00 (θ, φ), (8.28)

ψ210 = R21 (r) Y10 (θ, φ), (8.29)

182
 15
1 ( - )
10 �2 2,0,0 2,1,0

Energy / hc (cm-1)
5
n=2
2,1,+1
0
2,1,-1

-5

-10 1 ( + )
�2 2,0,0 2,1,0

-15
0 20000 40000 60000 80000 100000
Electric field strength (kV/cm)

Figure 8.4: The linear Stark effect in the n = 2 states in the H atom.

using the analytic forms of these functions in Table 2.2 and in Table 2.1, and
expressing the ẑ operator in terms of a spherical harmonic function as

z = r cos θ (8.30)
r
4π
= r Y10 (θ, φ), (8.31)
3
leads to the result that

∆E±F = ∓ 3 eaH Fz , (8.32)

where aH is the Bohr radius of the H atom. Comparing this expression

to Equation 8.2 indicates that for these states

−µ±elec = ∓ 3 eaH , (8.33)

and therefore

µ±elec = ± 3 eaH . (8.34)

The energy shifts of all four states with n = 2 in the H atom in the
presence of an electric field are displayed in Figure 8.4. The states ψ3 = ψ21 −1

183
and ψ4 = ψ21 +1 are not affected by the field. They therefore do not possess
any intrinsic electric dipole moments and do not change their energy as the
field strength increases.
Because the wavefunctions ψ± are also eigenstates of the field-free
Hamiltonian, no energy is required to mix them. As a result an intrin-
sic electric dipole moment exists within the atom and it is not necessary for
the field to induce the dipole moment. Consequently, in this case

∆EF ∝ Fz . (8.35)

Electron charge distributions

The wavefunctions ψ± can be expressed in the form

1
ψ± = √ (R20 Y00 ± R21 Y10 ) (8.36)
2

1  √ 
= √ R20 ± R21 3 cos θ . (8.37)
8π

Using these expressions the electron probability distributions |ψ± |2 r 2 can

be calculated and are displayed in Figure 8.5. This shows that in the ψ−
state [Figure 8.5(a)] the electron charge is localised along the electric field
axis away from the positively charged proton. Therefore the electric dipole
moment of the atoms is antiparallel to the electric field vector leading to a
positive Stark energy shift. On the other hand in the ψ+ state [Figure 8.5(b)]
the electron charge is localised behind the proton with the result that the
electric dipole moment is oriented parallel to the electric field vector. This
leads to a negative Stark energy shift.

8.3 Ionisation in strong electric fields

In strong electric fields the Stark Hamiltonian, VF , significantly modifies

the Coulomb potential, VC , experienced by an electron within an atom and
can lead to electric field ionisation. The origin of this effect can be seen
by considering the potential energy part, V(z), of the total Hamiltonian

184
 (a) = 1( - )
- √2 2,0,0 2,1,0

| -| 2 r 2 F
+20 +

+10
0 +20
x / aH -10 0
+10
-20 -10
-20 z / aH

(b) = 1( + )
+ √2 2,0,0 2,1,0

| +|2 r2 F
+20
+10
0 +20
x / aH -10 0
+10
-20 -10
-20 z / aH

Figure 8.5: Electron charge distributions in (a) the ψ− , and (b) ψ+ n = 2

Stark states in the H atom. In each figure the proton is located at the origin
of the coordinate system indicated by the ‘⊕’ symbol in (a).

operator in the direction of an applied field

V(z) = VC (z) + VFz (8.38)

e2
= − + eFz z. (8.39)
4πǫ0 z
This can be seen in Figure 8.6 in which the pure Coulomb potential as-
sociated with the H atom and containing excited states with high values
of n is displayed, together with the potential in the presence of a field
~ = (0, 0, −Fz ). From this figure it can be seen that for the field considered,
F
states with values of n ≥ 28 lie higher in energy that the saddle point of the
continuous curve. Therefore if the electron is in one of these states when
the field is applied it will no longer remain bound to the proton and will
instead be ionised.

185
 0
Ionisation limit

Energy/hc (cm )
-1

-100

n = 28

-200
V(z) = VC(z)
V(z) = VC(z) - eFz z
-0.4 -0.2 0 0.2 0.4
PositionPosition ( m) (μm)
in z dimension

Figure 8.6: The pure Coulomb potential, VC , experienced by the electron in

the H atom (dashed black curve), with the positions of excited states with
high values of n indicated. The potential that results when this potential
~ = (0, 0, −Fz ) is indicated by the
is modified by an external electric field F
continuous red curve.

The minimum electric field required to ionise an electron in a state with

a particular value of n can be determined by equating the position of the
saddle point in the potential in the presence of the field with the binding
energy of the state. To calculate this field consider the situation in Figure 8.6
~ = (0, 0, −Fz ), in this case
in which F

e2
V(z) = − − eFz z. (8.40)
4πǫ0 z
To determine the value of n at the saddle point in the positive z direction

dV(z) e2
= − eFz (8.41)
dz 4πǫ0 z2

= 0, (8.42)

186
and so the position of the saddle point, zsaddle , is
r
e
zsaddle = . (8.43)
4πǫ0 Fz

Therefore, substituting this back into Equation 8.40, the potential energy,
Vsaddle (z), of the saddle point is
r r
e2 4πǫ0 Fz e
Vsaddle (z) = − − eFz (8.44)
4πǫ0 e 4πǫ0 Fz
r
e3 Fz
= −2 . (8.45)
4πǫ0
To find the state with its binding energy equal to the energy of the saddle
point, Vsaddle (z) can be set equal to the binding energy given by the Rydberg
formula such that
hc RH
Vsaddle (z) = − , (8.46)
n2
with the result that for any value of n the ionisation field, Fion , is

πǫ0 (hc RH )2
Fion = (8.47)
e3 n4
and therefore
1
Fion ∝ . (8.48)
n4
From this the electric field required to ionise a ground state H atom is

Fion (n = 1) ≃ 3 × 1010 V/m (8.49)

while the electric field required to ionise a H atom in a state with n = 100, is

Fion (n = 100) ≃ 300 V/m (≡ 3 V/cm). (8.50)

187
 Part IV

Scattering

188
Chapter 9

Low-energy particle scattering

9.1 Introduction

Collisions between atoms, molecules, ions and electrons are described in

general as scattering processes. In scattering processes that occur at low
collision energies the quantised energy-level structure of the particles in-
volved must be accounted for, together with their wave nature (de Broglie
wavelength).
Low-energy particle scattering is of importance in the production of
antihydrogen and positronium atoms, and in chemical processes at low
temperatures. Low energy scattering of positrons and positronium from
biomolecules also plays a role in medical imaging [positron emission to-
mography (PET)].
In solving the Schrödinger equation to obtain the energy level structure
of atoms and molecules some aspects of particle scattering have already
been encountered. For example, in systems with a single excited outer
electron described by the central field approximation, this bound electron
may be considered to scatter from the ion core at short range. The details of
this this quantum mechanical scattering process contribute to the position-
dependence of the electronic wavefunctions, and the values of the quantum
defects.
In free-particle scattering processes, for example, the collision of a free
electron with an atomic ion, the center-of-mass collision energy, Ecoll , is

189
greater than zero, i.e., the collision process takes place in the continuum
above the ionization limit of the neutral system in which the ion and elec-
tron would be bound to form an atom. In this case the time evolution of
the de Broglie wave during the collision as the particles interact must be
considered and depends strongly on the interaction potential between the
two colliding particles.

9.2 Scattering processes

In general collisions, or scattering processes, can be categorised as either

elastic or inelastic.
Elastic scattering: In elastic scattering processes the total kinetic energy
of the system, composed of the target particle and the projectile, remains
unchanged following the collision event, but kinetic energy maybe trans-
ferred from one particle to the other. An example of an elastic scattering
process is

e− + p+ → e− + p+ . (9.1)

Inelastic scattering: In inelastic scattering processes the total kinetic

energy of the system changes because some of the initial kinetic energy
is transferred to internal energy following the excitation (e.g., electronic
excitation, vibrational excitation, or rotational excitation) of either the target
particle or the projectile. An example of an inelastic scattering process is

e− + He (1s2 1 S0 ) → e− + He (1s2s 3 S1 ). (9.2)

This is an example of a non-reactive inelastic scattering process in which the

composition of the target and projectile does not change. However, reactive
inelastic scattering can also be studied in which the composition of the target
and projectiles do change in the collision, i.e., molecular bonds are broken
or formed. An example of a reactive inelastic scattering process is

F + H2 → HF + H. (9.3)

190
 Target particles B

x Beam of particles A
Flux Ji
z
Figure 9.1: Schematic diagram of a particle scattering experiment in which
a beam of projectiles, A, passes through a cloud of target particles, B. The
flux of the incident beam is denoted Ji , while the number density of target
particles in the cell is nt .

9.3 Scattering cross-sections

Considering a collision experiment in which a beam of projectile particles

collides with a cloud of target particles as depicted schematically in Fig-
ure 9.1. In this situation the flux of incident projectiles, i.e., the number of
particles passing through a unit area per unit time is Ji . The number density
of target particles is nt and the path length over which the incident particles
interact with the target particles is l. If a detector is placed at an angular
position (θ, φ) with respect to the collision center to count the projectile
particles scattered into a solid angle dΩ, the detected flux, i.e., the number
of particles per unit time reaching the detector, dJs (θ, φ), is

dJs (θ, φ) ∝ Ji nt l dΩ. (9.4)

In this situation the constant of proportionality, dσ(θ, φ)/dΩ, is known as

the differential cross section, and hence
dσ(θ, φ) 1 dJs (θ, φ)
= . (9.5)
dΩ Ji nt l dΩ
The differential cross section is therefore the ratio of the particle flux scat-
tered into the solid angle dΩ to the incident flux and depends on the ori-
entation of the detector with respect to the direction of propagation of the
incident particle beam.

191
 v
2r
Projectile
Target

Figure 9.2: The collision of a projectile with a target particle of diameter

2r leading to a total collision cross-section of σtot ≃ πr2 .

The total scattering cross section, or integral cross section, σtot , can be
obtained by integrating the differential cross section over all angles such
that
Z
dσ(θ, φ)
σtot = dΩ (9.6)
dΩ
Z π Z 2π
dσ(θ, φ)
= sin θ dθ dφ. (9.7)
0 0 dΩ
Since a collision between atoms, molecules, electrons, positrons or ions in
free space is cylindrically symmetric, under such conditions the differential
cross section must be independent of φ. Consequently,
Z π
dσ(θ)
σtot = 2π sin θ dθ, (9.8)
0 dΩ
with the result that the total flux of scattered particles, Js , can be expressed
as

Js = σtot Ji . (9.9)

In classical scattering processes between pairs of hard spheres, the target

can be considered to have hard edges with the total cross-section approx-
imately equal to the cross-sectional area of the target, i.e., σtot = πr2 as
indicated in Figure 9.2. However, in collisions in which charged particles
are involved, e.g., in collisions of electrons and protons, or collisions of
electrons and atoms, the total cross-section can be larger than that in the
classical case because of the Coulomb interactions between the charges. In

192
these situations the total collision cross section is better described as the
effective area, normal to the direction of propagation of incident projectiles,
provided by the target.

9.4 Quantum treatment of low energy particle scatter-

ing

To develop a quantum treatment of low energy particle scattering, con-

sider an elastic scattering process involving two structureless particles with
masses m1 and m2 that interact via a central potential V(R), where R = |R ~|
is the interparticle distance. The two-body problem can be solved in terms
of the motion of the center of mass using the reduced mass, µ12 , such that
m1 m2
µ12 = . (9.10)
m1 + m2
Because collisions between particles are time-dependent events a quantum
treatment of the scattering process requires solving the time-dependent
Schrödinger equation

~ t)
∂Ψ(R,
~ t) = i~
Ĥ Ψ(R, , (9.11)
∂t
where
~ 2
Ĥ = − ∇ + V(R), (9.12)
2µ12

~ t) is the wavefunction describing the evolution of the effective

and Ψ(R,
particle of reduced mass µ12 in space and time.
Because the potential V(R) is not time dependent, the wavefunction,
~ t) can be separated into time-independent and time-dependent parts
Ψ(R,
such that
 
~ ~ iEt
Ψ(R, t) = ψ(R) exp − , (9.13)
~
where

~ = E ψ(R).
Ĥ ψ(R) ~ (9.14)

193
 (a) +10
(b) +10

+8 k +8
Position in x dimension (arb. units)

Position in x dimension (arb. units)

+6 +6

+4 +4

+2 +2
Target
0 0

-2 -2

-4 -4

-6 -6

-8 -8

-10 -10
-10 -8 -6 -4 -2 0 +2 +4 +6 +8 +10 -10 -8 -6 -4 -2 0 +2 +4 +6 +8 +10
Position in z dimension (arb. units) Position in z dimension (arb. units)

(c) +10
(d) +10

+8 +8
Position in x dimension (arb. units)

Position in x dimension (arb. units)

+6 +6

+4 +4

+2 +2

0 0

-2 -2

-4 -4

-6 -6

-8 -8

-10 -10
-10 -8 -6 -4 -2 0 +2 +4 +6 +8 +10 -10 -8 -6 -4 -2 0 +2 +4 +6 +8 +10
Position in z dimension (arb. units) Position in z dimension (arb. units)

Figure 9.3: Inelastic quantum scattering. (a) A plane wave with wave
vector ~k propagating in the positive z dimension incoming toward a target
particle as indicated. (b) The continuation of the plane wave in (a) ac-
counting for the directly transmitted component. (c) A scattered spherical
wave. The total wavefunction of the system combining the incoming and
transmitted plane waves in (c) with the scattered spherical wave in (b).

The solutions to the time-dependent Schrödinger equation of interest in

this elastic scattering process are those for which the total energy, E > 0,
and has a constant value. Although there are in general an infinite number
~ for which E > 0, the solutions of interest must satisfy
of functions ψ(R)
the boundary conditions for the problem. They are therefore composed of
three components (i) an incident wave, (ii) a transmitted wave, and (iii) a
scattered wave.

194
 (i) Incident wave: The wavefunction of a free particle with a linear
momentum ~ p = ~~k, where ~k is the wavevector, such that

~2 k 2
E = , (9.15)
2µ12
corresponds to a plane wave. As a result, the incident wave in this
elastic scattering process has the form
 
~ i = exp i~k · R
ψ(R) ~ . (9.16)

Such a wavefunction is displayed in Figure 9.3(a).

(ii) Transmitted wave: The wavefunction associated with the directly

transmitted component of the incident wave has a similar form to that
of the incident wave, i.e., the corresponding function is that in Equa-
tion 9.16. The combination of the incoming and directly scattered
plane waves are displayed in Figure 9.3(b).

(ii) Scattered wave: At large interparticle distances, i.e., as R → ∞, the

wavefunction of the system takes the form of a plane wave with an
angular-dependent amplitude. This angular dependence, associated
with the scattered wave, may be expressed as
exp (i k R )
~ s =
ψ(R) fk (θ, φ) . (9.17)
R
The angular dependent scattering amplitude, fk (θ, φ), depends on the
collision energy, i.e., magnitude of the wavevector, while the r in the
denominator in Equation 9.17 ensures that this outgoing scattered
wave is square integrable.

The scattering amplitude in Equation 9.17 is related to the differential

scattering cross section such that
dσ(θ, φ)  2
=  fk (θ, φ) . (9.18)
dΩ

Combining these three components the total quantum mechanical wave-

function associated with this inelastic scattering process is expressed as
  exp (i k R )
~ s ≃ exp i~k · R
ψ(R) ~ + fk (θ, φ) . (9.19)
R

195
Considering the quantum mechanical flux density ~J,

~J = ~

ψ∗ ∇ψ − ψ∇ψ∗ , (9.20)
2µ12
where in general

~J = ~ ~k
. (9.21)
µ12
For the wavefunction in Equation 9.19, the flux density in the positive z
direction, Jz , is therefore
" #
~ ∂ ∂
Jz = exp(−ikz) exp(ikz) − exp(ikz) exp(−ikz) . (9.22)
2iµ12 ∂z ∂z
Neglecting interference terms that only appear for forward scattering when
θ = 0, the radial dependence of the flux density
" #
~ ∗ exp(−ikR) ∂ exp(ikR) exp(ikR) ∂ exp(−ikR)
JR = f f −
2iµ12 k k R ∂θ R R ∂R R
(9.23)
2
| fk | ~ k
= . (9.24)
µ12 R2

For the angular components J(θ, φ) ∝ R−3 therefore the radial part of the flux
density provides the dominant contribution at long range, i.e., as R → ∞.
In terms of this scattered flux density, the number of particles scattered
into the detector solid angle R2 dΩ per unit time is

dJs = JR R2 dΩ (9.25)

| fk | 2 ~ k
= dΩ. (9.26)
µ12
However, this quantity can also be expressed in terms of the incident particle
flux, Ji , such that
dσ
dJs = Ji dΩ. (9.27)
dΩ
Since the magnitude of the incident particle flux can be expressed in terms
of the incident particle wavevector as
~k
Ji = , (9.28)
µ12

196
from Equation 9.26 and Equation 9.27 it is seen that

| fk |2 ~ k dσ ~k
dΩ = dΩ, (9.29)
µ12 dΩ µ12
and hence
dσ
| fk | 2 = . (9.30)
dΩ
The total scattering cross section, σtot , is then, as in Equation 9.7,
Z π Z 2π
σtot = | fk (θ, φ)|2 sin θ dθ dφ. (9.31)
0 0

9.5 Angular momentum partial waves

The wavefunction describing the stationary scattering state in Equation 9.19

is associated with a central potential and an incident particle with a wavevec-
tor propagating in the z dimension. In elastic scattering from a central
potential the orbital angular momentum, ℓ, of the incident particle rela-
tive to the target is conserved. In this situation, to identify the individual
angular momentum components of the incident particle wavefunction it
can be decomposed into its constituent angular momentum partial waves.
~ and fk (θ, φ) in terms of known angular
This is achieved by expanding ψ(R)
functions. Given the cylindrical symmetry of the two-particle collision the
most appropriate functions to choose are the set of Legendre polynomials,
Pℓ (cos θ). Thus
X
~ ∝
ψ(R) Rℓ (R) Pℓ (cos θ), (9.32)
ℓ

and
X
fk (θ, φ) ∝ (2ℓ + 1) fℓ Pℓ (cos θ), (9.33)
ℓ

where Rℓ (R) is a radial function, and fℓ is a constant coefficient.

Because of the requirement to conserve angular momentum, the scat-
tering problem can then be simplified to a set of problems, one for each
individual partial wave component. The resulting partial wave solutions
ˆ
are simultaneous eigenfunctions of ~L2 and Lz .

197