Atomic Structure

Chapter 9
Atomic structure
Previously, we have seen that the quantum mechanics of atomic hydrogen, and
hydrogen-like atoms is characterized by a large degeneracy with eigenvalues
separating into multiplets of n2 -fold degeneracy, where n denotes the principle
quantum number. However, although the idealized Schrödinger Hamiltonian,
p̂2 1 Ze2
Ĥ0 = + V (r), V (r) = − , (9.1)
2m 4π"0 r
provides a useful platform from which develop our intuition, there are several
important effects which mean that the formulation is a little too naı̈ve. These
“corrections”, which derive from several sources, are important as they lead to
physical ramifications which extend beyond the realm of atomic physics. Here
we outline some of the effects which need to be taken into account even for
atomic hydrogen, before moving on to discuss the quantum physics of multi-
electron atoms. In broad terms, the effects to be considered can be grouped
into those caused by the internal properties of the nucleus, and those which
derive from relativistic corrections.
To orient our discussion, it will be helpful to summarize some of key aspects
of the solutions of the non-relativisitic Schrödinger equation, Ĥ0 ψ = Eψ on
which we will draw:
Hydrogen atom revisited:
$ As with any centrally symmetric potential, the solutions of the Schrödinger

equation take the form ψ!m (r) = R(r)Y!m (θ, φ), where the spherical har-
monic functions Y!m (θ, φ) depend only on spherical polar coordinates,
and R(r) represents the radial component of the wavefunction. Solving
the radial wave equation introduces a radial quantum number, nr ≥ 0.
In the case of a Coulomb potential, the energy depends on the principal
quantum number n = nr + ' + 1 ≥ 1, and not on nr and ' separately.
$ For atomic hydrogen (Z = 1), the energy levels of the Hamiltonian (9.1)
are given by The fine structure constant is
! 2 " known to great accuracy and is
Ry e m e2 1 1 given by,
En = − 2 , Ry = = = mc2 α2 ,
n 4π"0 2! 2 4π"0 2a0 2
α = 7.297352570(5) × 10−3
4π#0 !2 e2 1
where a0 = is the Bohr radius, α = 4π#0 !cdenotes the fine 1
e2 m = .
structure constant, and strictly speaking m represents the reduced 137.035999070(9)
mass of the electron and proton. Applied to single electron ions with
higher atomic weight, such as He+ , Li2+ , etc., the Bohr radius is reduced
by a factor 1/Z, where Z denotes the nuclear charge, and the energy is
2
given by En = − Zn2 Ry = − 2n1 2 mc2 (Zα)2 .
Advanced Quantum Physics

9.1. THE “REAL” HYDROGEN ATOM 90
$ Since n ≥ 1 and nr ≥ 0, the allowed combinations of quantum num-

bers are shown on the right, where we have introduced the conventional
notation whereby values of ' = 0, 1, 2, 3, 4 · · · are represented by letters n ' Subshell(s)
s, p, d, f, g · · · respectively. 1 0 1s
2 0, 1 2s 2p
$ Since En depends only on n, this implies, for example, an exact degen- 3 0, 1, 2 3s 3p 3d
eracy of the 2s and 2p, and of the 3s, 3p and 3d levels. 4 0, 1, 2, 3 4s 4p 4d 4f
n 0 · · · (n − 1) ns · · ·
These results emerge from a treatment of the hydrogen atom which is
inherently non-relativistic. In fact, as we will see later in our discussion of the
Dirac equation in chapter 15, the Hamiltonian (9.1) represents only the leading
term in an expansion in v 2 /c2 $ (Zα)2 of the full relativistic Hamiltonian To see that v 2 /c2 $ (Zα)2 , we
(see below). Higher order terms provide relativistic corrections, which impact may invoke the virial theorem.
significantly in atomic and condensed matter physics, and lead to a lifting of The latter shows that the aver-
the degeneracy. In the following we will discuss and obtain the heirarchy of age kinetic energy is related to
the potential energy as %T & =
leading relativistic corrections.1 This discussion will provide a platform to
− 12 %V &. Therefore, the aver-
describe multi-electron atoms. age energy is given by %E& =
%T & + %V & = −%T & ≡ − 21 mv 2 .
We therefore have that 12 mv 2 =
9.1 The “real” hydrogen atom Ry ≡ 21 mc2 (Zα)2 from which
follows the relation v 2 /c2 $
The relativistic corrections (sometimes known as the fine-structure cor- (Zα)2 .
rections) to the spectrum of hydrogen-like atoms derive from three different
sources:
$ relativistic corrections to the kinetic energy;
$ coupling between spin and orbital degrees of freedom;
$ and a contribution known as the Darwin term.
In the following, we will discuss each of these corrections in turn.
9.1.1 Relativistic correction to the kinetic energy

Previously, we have taken the kinetic energy to have the familiar non-relativistic
p̂2
form, 2m . However, from the expression for the relativistic energy-momentum
invariant, we can already anticipate that the leading correction to the non-
relativistic Hamiltonian appears at order p4 ,
# p2 1 (p2 )2
E = (pµ pµ )1/2 = p2 c2 + m2 c4 = mc2 + − + ···
2m 8 m3 c2
As a result, we can infer the following perturbation to the kinetic energy of
the electron,
1 (p̂2 )2
Ĥ1 = − .
8 m3 c2
When compared with the non-relativistic kinetic energy, p2 /2m, one can see
that the perturbation is smaller by a factor of p2 /m2 c2 = v 2 /c2 $ (Zα)2 , i.e.
Ĥ1 is only a small perturbation for small atomic number, Z ( 1/α $ 137.
1
It may seem odd to discuss relativistic corrections in advance of the Dirac equation
and the relativistic formulation of quantum mechanics. However, such a discussion would
present a lengthy and unnecessarily complex digression which would not lead to further
illumination. We will therefore follow the normal practice of discussing relativistic corrections
as perturbations to the familiar non-relativistic theory.

We can therefore make use of a perturbative analysis to estimate the scale of

the correction.
In principle, the large-scale degeneracy of the hydrogen atom would de-
mand an analysis based on the degenerate perturbation theory. However,
fortunately, since the off-diagonal matrix elements vanish,2
%n'm|Ĥ1 |n'! m! & = 0 for ' )= '! or m )= m! ,
degenerate states are uncoupled and such an approach is unnecessary. Then

making use of the identity, Ĥ1 = − 2mc
1
2 [Ĥ0 − V (r)] , the scale of the resulting
2
energy shift can be obtained from first order perturbation theory, By making use of the form of the
radial wavefunction for the hy-
1 & 2 '2 drogen atom, one may obtain the
%Ĥ1 &n!m ≡ %n'm|Ĥ1 |n'm& = − En − 2En %V (r)&n! + %V 2 (r)&n! .
2mc2 identities,
$ %
Since the calculation of the resulting expectation values is not particularly 1 Z
=
illuminating, we refer to the literature for a detailed exposition3 and present r n! a0 n2
here only the required identities (right). From these considerations, we obtain $ %
1 Z2
the following expression for the first order energy shift, = 2 3 .
2
r n! a0 n (' + 1/2)
! " ! "
mc2 Zα 4 n 3
%Ĥ1 &n!m = − − . (9.2)
2 n ' + 1/2 4
From this term alone, we expect the degerenacy between states with different
values of total angular momentum ' to be lifted. However, as well will see,
this conclusion is a little hasty. We need to gather all terms of the same
order of perturbation theory before we can reach a definite conclusion. We
can, however, confirm that (as expected) the scale of the correction is of order
"Ĥ1 #n!m
En ∼ ( Zα
n ) . We now turn to the second important class of corrections.
2
9.1.2 Spin-orbit coupling

As well as revealing the existence of an internal spin degree of freedom, the
relativistic formulation of quantum mechanics shows that there is a further
relativistic correction to the Schrödinger operator which involves a coupling
between the spin and orbital degrees of freedom. For a general potential V (r),
this spin-orbit coupling takes the form,
1 1
Ĥ2 = (∂r V (r)) Ŝ · L̂ .
2m2 c2 r
2
For a hydrogen-like atom, V (r) = − 4π#
1 Ze
0 r
, and
1 Ze2 1
Ĥ2 = Ŝ · L̂ .
2m2 c2 4π"0 r3
$ Info. Physically, the origin of the spin-orbit interaction can be understoon

from the following considerations. As the electron is moving through the electric
field of the nucleus then, in its rest frame, it will experience this as a magnetic
field. There will be an additional energy term in the Hamiltonian associated with
the orientation of the spin magnetic moment with respect to this field. We can make
an estimate of the spin-orbit interaction energy as follows: If we have a central field
determined by an electrostatic potential V (r), the corresponding electric field is given
2
The proof runs as follows: Since [Ĥ1 , L̂2 ] = 0, !2 [!! (!! + 1) − !(! + 1)] "n!m|Ĥ1 |n!! m! # =
0. Similarly, since [Ĥ1 , L̂z ] = 0, !(m! − m)"n!m|Ĥ1 |n!! m! # = 0.
3
see, e.g., Ref [1].

by E = −∇V (r) = −er (∂r V ). For an electron moving at velocity v, this translates
to an effective magnetic field B = c12 v × E. The magnetic moment of the electron
q
associated with its spin is equal to µs = gs 2m S ≡ −m
e
S, and thus the interaction
energy is given by
e e e 1
−µs · B = S · (v × E) = − S · (p × er (∂r V )) = (∂r V ) L · S ,
mc 2 (mc)2 (mc)2 r
where we have used the relation p × er = p × rr = − Lr . In fact this isn’t quite correct;
there is a relativistic effect connected with the precession of axes under rotation, called
Thomas precession which multiplies the formula by a further factor of 12 .
Once again, we can estimate the effect of spin-orbit coupling by treating Ĥ2
as a perturbation. In the absence of spin-orbit interaction, one may express
the eigenstates of hydrogen-like atoms in the basis states of the mutually
commuting operators, Ĥ0 , L̂2 , L̂z , Ŝ2 , and Ŝz . However, in the presence
of spin-orbit coupling, the total Hamiltonian no longer commutes with L̂z
or Ŝz (exercise). It is therefore helpful to make use of the degeneracy of
the unperturbed Hamiltonian to switch to a new basis in which the angular
momentum components of the perturbed system are diagonal. This can be
achieved by turning to the basis of eigenstates of the operators, Ĥ0 , Ĵ2 , Jˆz ,
L̂2 , and Ŝ2 , where Ĵ = L̂ + Ŝ denotes the total angular momentum. (For a
discussion of the form of these basis states, we refer back to section 6.4.2.)
Making use of the relation, Ĵ2 = L̂2 + Ŝ2 + 2L̂ · Ŝ, in this basis, it follows
that,
1
L̂ · Ŝ = (Ĵ2 − L̂2 − Ŝ2 ) .
2
Combining the spin and angular momentum, the total angular momentum
takes values j = ' ± 1/2. The corresponding basis states |j = ' ± 1/2, mj , '&
(with s = 1/2 implicit) therefore diagonalize the operator,
! "
!2 '
Ŝ · L̂|j = ' ± 1/2, mj , '& = |' ± 1/2, mj , '& ,
2 −' − 1
where the brackets index j = ' + 1/2 (top) and j = ' − 1/2 (bottom). As for
the radial dependence of the perturbation, once again, the off-diagonal matrix
elements vanish circumventing the need to invoke degenerate perturbation
theory. As a result, at first order in perturbation theory, one obtains
! " $ %
1 !2 ' Ze2 1
%H2 &n,j=!±1/2,mj ,! = .
2m c 2
2 2 −' − 1 4π"0 r3 n!
Then making use of the identity (right),4 one obtains For ' > 0,
! " ( ) $ % ! "3
1 2 Zα 4 n 1
1 mcαZ 1
%Ĥ2 &n,j=!±1/2,mj ,! = mc j
1 . = .
4 n j + 1/2 − j+1 r3 n! !n '(' + 12 )(' + 1)
Note that, for ' = 0, there is no orbital angular momentum with which to
couple! Then, if we rewrite the expression for %Ĥ1 & (9.2) in the new basis,
! " ( 1 )
1 2 Zα 4
%Ĥ1 &n,j=!±1/2,mj ,! = − mc n j
1 ,
2 n j+1
and combining both of these expressions, for ' > 0, we obtain

! " ! "
1 2 Zα 4 3 n
%Ĥ1 + Ĥ2 &n,j=!±1/2,mj ,! = mc − ,
2 n 4 j + 1/2
while for ' = 0, we retain just the kinetic energy term (9.2).
4
For details see, e.g., Ref [1].

9.1.3 Darwin term

The final contribution to the Hamiltonian from relativistic effects is known as
the Darwin term and arises from the “Zitterbewegung” of the electron –
trembling motion – which smears the effective potential felt by the electron.
Such effects lead to a perturbation of the form,
! "
!2 !2 e Ze2 !2
Ĥ3 = ∇ 2
V = Qnuclear (r) = 4πδ (3) (r) ,
8m c
2 2 8m c
2 2 "0 4π"0 8(mc)2
where Qnuclear (r) = Zeδ (3) (r) denotes the nuclear charge density. Since the
perturbation acts only at the origin, it effects only states with ' = 0. As a
result, one finds that
! "4
Ze2 !2 1 Zα
%Ĥ3 &njmj ! = 4π|ψ!n (0)|2 = mc2 nδ!,0 .
4π"0 8(mc)2 2 n
Intriguingly, this term is formally identical to that which would be obtained

from %Ĥ2 & at ' = 0. As a result, combining all three contributions, the total
energy shift is given simply by
! "4 ! "
1 Zα 3 n
∆En,j=!±1/2,mj ,! = mc2 − , (9.3)
2 n 4 j + 1/2
a result that is independent of ' and mj .

To discuss the predicted energy shifts for particular states, it is helpful to
introduce some nomenclature from atomic physics. For a state with principle
quantum number n, total spin s, orbital angular momentum ', and total an-
gular momentum j, one may use spectroscopic notation n2s+1 Lj to define
the state. For a hydrogen-like atom, with just a single electron, 2s + 1 = 2.
In this case, the factor 2s + 1 is often just dropped for brevity.
If we apply our perturbative expression for the relativistic corrections (9.3),
how do we expect the levels to shift for hydrogen-like atoms? As we have seen,
for the non-relativistic Hamiltonian, each state of given n exhibits a 2n2 -fold
degeneracy. For a given multiplet specified by n, the relativistic corrections
depend only on j and n. For n = 1, we have ' = 0 and j = 1/2: Both 1S1/2
states, with mj = 1/2 and −1/2, experience a negative energy shift by an
amount ∆E1,1/2,mj ,0 = − 41 Z 4 α2 Ry. For n = 2, ' can take the values of 0 or 1.
With j = 1/2, both the former 2S1/2 state, and the latter 2P1/2 states share
the same negative shift in energy, ∆E2,1/2,mj ,0 = ∆E2,1/2,mj ,1 = − 64 Z α Ry,
5 4 2
while the 2P3/2 experiences a shift of ∆E2,3/2,mj ,1 = − 64 Z α Ry. Finally, for

1 4 2
n = 3, ' can take values of 0, 1 or 2. Here, the pairs of states 3S1/2 and 3P1/2 ,
and 3P3/2 and 2D3/2 each remain degenerate while the state 3D5/2 is unique.
These predicted shifts are summarized in Figure 9.1.
This completes our discussion of the relativistic corrections which develop Willis Eugene Lamb, 1913-2008
from the treatment of the Dirac theory for the hydrogen atom. However, A physicist who
won the Nobel
this does not complete our discription of the “real” hydrogen atom. Indeed, Prize in Physics
there are further corrections which derive from quantum electrodynamics and in 1955 “for his
discoveries con-
nuclear effects which we now turn to address. cerning the fine
structure of the
hydrogen spectrum”. Lamb and
9.1.4 Lamb shift Polykarp Kusch were able to pre-
cisely determine certain electromag-
According to the perturbation theory above, the relativistic corrections which netic properties of the electron.
follow from the Dirac theory for hydrogen leave the 2S1/2 and 2P1/2 states
degenerate. However, in 1947, a careful experimental study by Willis Lamb

Figure 9.1: Figure showing the

heirarchy of energy shifts of the spec-
tra of hydrogen-like atoms as a result
of relativistic corrections. The first
column shows the energy spectrum
predicted by the (non-relativistic)
Bohr theory. The second column
shows the predicted energy shifts
from relativistic corrections arising
from the Dirac theory. The third col-
umn includes corrections due quan-
tum electrodynamics and the fourth
column includes terms for coupling
to the nuclear spin degrees of free-
dom. The H-α line, particularly
important in the astronomy, corre-
sponds to the transition between the
levels with n = 2 and n = 3.
and Robert Retherford discovered that this was not in fact the case:5 2P1/2 Hans Albrecht Bethe 1906-2005
A German-
state is slightly lower in energy than the 2S1/2 state resulting in a small shift American
of the corresponding spectral line – the Lamb shift. It might seem that such a physicist, and
tiny effect would be deemed insignificant, but in this case, the observed shift Nobel laureate
in physics “for
(which was explained by Hans Bethe in the same year) provided considerable his work on the
insight into quantum electrodynamics. theory of stellar
nucleosynthesis.”
In quantum electrodynamics, a quantized radiation field has a zero-point A versatile theo-
energy equivalent to the mean-square electric field so that even in a vacuum retical physicist,
Bethe also made important contribu-
there are fluctuations. These fluctuations cause an electron to execute an tions to quantum electrodynamics,
oscillatory motion and its charge is therefore smeared. If the electron is bound nuclear physics, solid-state physics
by a non-uniform electric field (as in hydrogen), it experiences a different and particle astrophysics. During
World War II, he was head of the
potential from that appropriate to its mean position. Hence the atomic levels Theoretical Division at the secret
are shifted. In hydrogen-like atoms, the smearing occurs over a length scale, Los Alamos laboratory developing
the first atomic bombs. There he
! "2 played a key role in calculating the
2α ! 1 critical mass of the weapons, and
%(δr)2 & $ ln , did theoretical work on the implosion
π mc αZ method used in both the Trinity test
and the “Fat Man” weapon dropped
some five orders of magnitude smaller than the Bohr radius. This causes the on Nagasaki.
electron spin g-factor to be slightly different from 2,
! "
α α2
gs = 2 1 + − 0.328 2 + · · · .
2π π
There is also a slight weakening of the force on the electron when it is very
close to the nucleus, causing the 2S1/2 electron (which has penetrated all the
way to the nucleus) to be slightly higher in energy than the 2P1/2 electron.
Taking into account these corrections, one obtains a positive energy shift
! "4 ! "
Z 8 1
∆ELamb $ nα Ry ×
2
α ln δ!,0 ,
n 3π αZ
for states with ' = 0.

5 Hydrogen fine structure and hy-
W. E. Lamb and R. C. Retherfod, Fine Structure of the Hydrogen Atom by a Microwave
Method, Phys. Rev. 72, 241 (1947).
perfine structure for the n = 3 to
n = 2 transition (see Fig. 9.1).

9.1.5 Hyperfine structure

So far, we have considered the nucleus as simply a massive point charge respon-
sible for the large electrostatic interaction with the charged electrons which
surround it. However, the nucleus has a spin angular momentum which is
associated with a further set of hyperfine corrections to the atomic spec-
tra of atoms. As with electrons, the protons and neutrons that make up a
nucleus are fermions, each with intrinsic spin 1/2. This means that a nucleus
will have some total nuclear spin which is labelled by the quantum number,
I. The latter leads to a nuclear magnetic moment,
Ze
µN = gN I.
2MN
where MN denotes the mass of the nucleus, and gN denotes the gyromagnetic
ratio. Since the nucleus has internal structure, the nuclear gyromagnetic ratio
is not simply 2 as it (nearly) is for the electron. For the proton, the sole
nuclear constituent of atomic hydrogen, gP ≈ 5.56. Even though the neutron
is charge neutral, its gyromagnetic ratio is about −3.83. (The consitituent
quarks have gyromagnetic ratios of 2 (plus corrections) like the electron but
the problem is complicated by the strong interactions which make it hard to
define a quark’s mass.) We can compute (to some accuracy) the gyromagnetic
ratio of nuclei from that of protons and neutrons as we can compute the
proton’s gyromagnetic ratio from its quark constituents. Since the nuclear
mass is several orders of magnitude higher than that of the electron, the nuclear
magnetic moment provides only a small perturbation.
According to classical electromagnetism, the magnetic moment generates
a magnetic field
µ0 2µ0
B= (3(µN · er )er − µN ) + µ δ (3) (r) .
4πr3 3 N
To explore the effect of this field, let us consider just the s-electrons, i.e. ' = 0,
for simplicity.6 In this case, the interaction of the magnetic moment of the
electrons with the field generated by the nucleus, gives rise to the hyperfine
interaction,
e
Ĥhyp = −µe · B = Ŝ · B .
mc
For the ' = 0 state, the first contribution to B vanishes while second leads to
the first order correction,
! "4
Z 8 m 1
%Hhyp &n,1/2,0 = nα2 Ry × gN S · I.
n 3 MN ! 2
Once again, to evaluate the expectation values on the spin degrees of free-
dom, it is convenient to define the total spin F = I + S. We then have
1 1 1
S · I = 2 (F2 − S2 − I2 ) = (F (F + 1) − 3/4 − I(I + 1))
! *2! 2
2
1 I F = I + 1/2
=
2 −I − 1 F = I − 1/2
Therefore, the 1s state of Hydrogen is split into two, corresponding to the

two possible values F = 0 and 1. The transition between these two levels has
frequency 1420 Hz, or wavelength 21 cm, so lies in the radio waveband. It
6
For a full discussion of the influence of the orbital angular momentum, we refer to Ref. [6].

9.2. MULTI-ELECTRON ATOMS 96
is an important transition for radio astronomy. A further contribution to the

hyperfine structure arises if the nuclear shape is not spherical thus distorting
the Coulomb potential; this occurs for deuterium and for many other nuclei.
Finally, before leaving this section, we should note that the nucleus is not
point-like but has a small size. The effect of finite nuclear size can be A muon is a particle somewhat
estimated perturbatively. In doing so, one finds that the s (' = 0) levels are like an electron, but about 200
those most effected, because these have the largest probability of finding the times heavier. If a muon is
electron close to the nucleus; but the effect is still very small in hydrogen. It captured by an atom, the cor-
responding Bohr radius is 200
can be significant, however, in atoms of high nuclear charge Z, or for muonic
times smaller, thus enhancing
atoms. the nuclear size effect.
This completes our discussion of the “one-electron” theory. We now turn

to consider the properties of multi-electron atoms.
9.2 Multi-electron atoms

To address the electronic structure of a multi-electron atom, we might begin
with the hydrogenic energy levels for an atom of nuclear charge Z, and start
filling the lowest levels with electrons, accounting for the exclusion principle.
The degeneracy for quantum numbers (n, ') is 2 × (2' + 1), where (2' + 1) is
the number of available m! values, and the factor of 2 accounts for the spin
degeneracy. Hence, the number of electrons accommodated in shell, n, would
be 2 × n2 ,
n ' Degeneracy in shell Cumulative total
1 0 2 2
2 0, 1 (1 + 3) × 2 = 8 10
3 0, 1, 2 (1 + 3 + 5) × 2 = 18 28
4 0, 1, 2, 3 (1 + 3 + 5 + 7) × 2 = 32 60
We would therefore expect that atoms containing 2, 10, 28 or 60 electrons
would be especially stable, and that in atoms containing one more electron
than this, the outermost electron would be less tightly bound. In fact, if we
look at data (Fig. 9.2) recording the first ionization energy of atoms, i.e. the
minimum energy needed to remove one electron, we find that the noble gases,
having Z = 2, 10, 18, 36 · · · are especially tightly bound, and the elements
containing one more electron, the alkali metals, are significantly less tightly
bound.
The reason for the failure of this simple-minded approach is fairly obvious –
we have neglected the repulsion between electrons. In fact, the first ionization
energies of atoms show a relatively weak dependence on Z; this tells us that the
outermost electrons are almost completely shielded from the nuclear charge.7
Indeed, when we treated the Helium atom as an example of the variational
method in chapter 7, we found that the effect of electron-electron repulsion
was sizeable, and really too large to be treated accurately by perturbation
theory.
7
In fact, the shielding is not completely perfect. For a given energy shell, the effective
nuclear charge varies for an atomic number Z as Zeff ∼ (1 + α)Z where α > 0 characterizes
2
the ineffectiveness of screening. This implies that the ionization energy IZ = −EZ ∼ Zeff ∼
(1 + 2αZ). The near-linear dependence of IZ on Z is reflected in Fig. 9.2.

Figure 9.2: Ionization energies of the elements.
9.2.1 Central field approximation

Leaving aside for now the influence of spin or relativistic effects, the Hamilto-
nian for a multi-electron atom can be written as
+ , !2 1 Ze2
- +
1 e2
Ĥ = − 2
∇i − + ,
2m 4π"0 ri 4π"0 rij
i i<j
where rij ≡ |ri − rj |. The first term represents the “single-particle” contri-
bution to the Hamiltonian arising from interaction of each electron with the
nucleus, while the last term represents the mutual Coulomb interaction be-
tween the constituent electrons. It is this latter term that makes the generic
problem “many-body” in character and therefore very complicated. Yet, as we
have already seen in the perturbative analysis of the excited states of atomic
Helium, this term can have important physical consequences both on the over-
all energy of the problem and on the associated spin structure of the states.
The central field approximation is based upon the observation that
the electron interaction term contains a large central (spherically symmetric)
component arising from the “core electrons”. From the following relation,
!
+
|Ylm (θ, φ)|2 = const.
m=−!
it is apparent that a closed shell has an electron density distribution which

is isotropic (independent of θ and φ). We can therefore develop a perturbative
scheme by setting Ĥ = Ĥ0 + Ĥ1 , where
+ , !2 1 Ze2
- + 1 e2 +
Ĥ0 = − 2
∇i − + Ui (ri ) , Ĥ1 = − Ui (ri ) .
2m 4π"0 ri 4π"0 rij
i i<j i
Here the one-electron potential, Ui (r), which is assumed central (see below),
incorporates the “average” effect of the other electrons. Before discussing how
to choose the potentials Ui (r), let us note that Ĥ0 is separable into a sum of
terms for each electron, so that the total wavefunction can be factorized into
components for each electron. The basic idea is first to solve the Schrödinger
equation using Ĥ0 , and then to treat Ĥ1 as a small perturbation.
On general grounds, since the Hamiltonian Ĥ0 continues to commute with
the angular momentum operator, [Ĥ0 , L̂] = 0, we can see that the eigenfunc-
tions of Ĥ0 will be characterized by quantum numbers (n, ', m! , ms ). However,

since the effective potential is no longer Coulomb-like, the ' values for a given Douglas Rayner Hartree FRS
n need not be degenerate. Of course, the difficult part of this procedure is 1897-1958
An English
to estimate Ui (r); the potential energy experienced by each electron depends mathematician
on the wavefunction of all the other electrons, which is only known after the and physicist
most famous for
Schrödinger equation has been solved. This suggests that an iterative approach the development
to solving the problem will be required. of numerical
analysis and its
To understand how the potentials Ui (r) can be estimated – the self- application to
consistent field method – it is instructive to consider a variational ap- atomic physics.
proach due originally to Hartree. If electrons are considered independent, the He entered
St John’s College Cambridge in
wavefunction can be factorized into the product state, 1915 but World War I interrupted
his studies and he joined a team
Ψ({ri }) = ψi1 (r1 )ψi2 (r2 ) · · · ψiN (rN ) , studying anti-aircraft gunnery. He
returned to Cambridge after the war
where the quantum numbers, ik ≡ (n'm! ms )k , indicate the individual state and graduated in 1921 but, perhaps
because of his interrupted studies,
occupancies. Note that this product state is not a properly antisymmetrized he only obtained a second class
Slater determinant – the exclusion principle is taken into account only in so degree in Natural Sciences. In 1921,
a visit by Niels Bohr to Cambridge
far as the energy of the ground state is taken to be the lowest that is consistent inspired him to apply his knowledge
with the assignment of different quantum numbers, n'm! ms to each electron. of numerical analysis to the solution
of differential equations for the
Nevertheless, using this wavefunction as a trial state, the variational energy is calculation of atomic wavefunctions.
then given by
+. ! 2 2 "
∗ ! ∇ 1 Ze2
E = %Ψ|Ĥ|Ψ& = 3
d r ψi − − ψi
2m 4π"0 r
i
. .
1 + e2
+ 3
d r d3 r! ψi∗ (r)ψi∗ (r! ) ψj (r! )ψi (r) .
4π"0 |r − r! |
i<j
Now, according to the variational principle, we must minimize the energy

functional by varying E[{ψi }] with respect to the complex wavefunction, ψi ,
subject to the normalization condition, %ψi |ψi & = 1. The latter can be imposed
using a set of Lagrange multipliers, εi , i.e.
, !. "-
δ
E − εi d r|ψi (r)| − 1
3 2
= 0.
δψi∗
Following the variation,8 one obtains the Hartree equations,
! 2 2 " .
! ∇ 1 Ze2 1 + e2
− − ψi + d3 r! |ψj (r! )|2 ψi (r)
2m 4π"0 r 4π"0 |r − r! |
j&=i
= εi ψi (r) . (9.4)
Then according to the variational principle, amongst all possible trial functions
ψi , the set that minimizes the energy are determined by the effective potential,
.
1 + e2
Ui (r) = d3 r! |ψj (r! )|2 .
4π"0 |r − r! |
j&=i
Equation (9.4) has a simple interpretation: The first two terms relate to the
nuclear potential experienced by the individual electrons, while the third term
represents the electrostatic potential due to the other electrons. However, to
simplify the procedure, it is useful to engineer the radial symmetry of the
potential by replacing Ui (r) by its spherical average,
.
dΩ
Ui (r) -→ Ui (r) = Ui (r) .
4π
8
Note that, in applying the variation, the wavefunction ψi∗ can be considered independent
of ψi – you might like to think why.

Finally, to relate the Lagrange multipliers, εi (which have the appearance

of one-electron energies), to the total energy, we can multiply Eq. (9.4) by
ψi∗ (r) and integrate,
. ! 2 2 "
∗ ! ∇ 1 Ze2
"i = d r ψi −
3
− ψi
2m 4π"0 r
.
1 + e2
+ d3 r! d3 r |ψj (r! )|2 |ψi (r)|2 .
4π"0 |r − r! |
j&=i
If we compare this expression with the variational state energy, we find that
+ .
1 + e2
E= "i − d3 r! d3 r |ψj (r! )|2 |ψi (r)|2 . (9.5)
4π"0 |r − r! |
i i<j
To summarize, if we wish to implement the central field approximation to

determine the states of a multi-electron atom, we must follow the algorithm:
1. Firstly, one makes an initial “guess” for a (common) central potential,

U (r). As r → 0, screening becomes increasingly ineffective and we ex-
1 (Z−1)e2
pect U (r) → 0. As r → ∞, we anticipate that U (r) → 4π# 0 r ,
corresponding to perfect screening. So, as a starting point, we make
take some smooth function U (r) interpolating between these limits. For
this trial potential, we can solve (numerically) for the eigenstates of the
single-particle Hamiltonian. We can then use these states as a plat-
form to build the product wavefunction and in turn determine the self-
consistent potentials, Ui (r).
2. With these potentials, Ui (r), we can determine a new set of eigenstates

for the set of Schrödinger equations,
, -
!2 2 1 Ze2
− ∇ − + Ui (r) ψi = εi ψi .
2m 4π"0 r
3. An estimate for the ground state energy of an atom can be found by

filling up the energy levels, starting from the lowest, and taking account
of the exclusion principle.
4. Using these wavefunctions, one can then make an improved estimate of

the potentials Ui (ri ) and return to step 2 iterating until convergence.
Since the practical implemention of such an algorithm demands a large degree

of computational flair, if you remain curious, you may find it useful to refer to
the Mathematica code prepared by Ref. [4] where both the Hartree and the
Hartree-Fock procedures (described below) are illustrated.
$ Info. An improvement to this procedure, known the Hartree-Fock method,

takes account of exchange interactions. In order to do this, it is necessary to ensure
that the wavefunction, including spin, is antisymmetric under interchange of any pair
of electrons. This is achieved by introducing the Slater determinant. Writing the
individual electron wavefunction for the ith electron as ψk (ri ), where i = 1, 2 · · · N and
k is shorthand for the set of quantum numbers (n'm! ms ), the overall wavefunction
is given by
/ /
/ ψ1 (r1 ) ψ1 (r2 ) ψ1 (r3 ) · · · /
/ /
1 // ψ2 (r1 ) ψ2 (r2 ) ψ2 (r3 ) · · · //
Ψ = √ / ψ3 (r1 ) ψ3 (r2 ) ψ3 (r3 ) · · · / .
N ! // /
.. .. .. . . //
/ . . . .

Note that each of√the N ! terms in Ψ is a product of wavefunctions for each individual Vladimir Aleksandrovich Fock
electron. The 1/ N ! factor ensures the wavefunction is normalized. A determinant 1898-1974
A Soviet physi-
changes sign if any two columns are exchanged, corresponding to ri ↔ rj (say); this cist, who did
ensures that the wavefunction is antisymmetric under exchange of electrons i and j. foundational
Likewise, a determinant is zero if any two rows are identical; hence all the ψk s must work on quan-
be different and the Pauli exclusion principle is satisfied.9 In this approximation, a tum mechanics
and quantum
variational analysis leads to the Hartree-Fock equations (exercise), electrodynamics.
, - His primary
!2 2 Ze2
εi ψi (r) = − ∇i − ψi (r) scientific con-
2m 4π"ri tribution lies
+. 1 e2 0 1 in the development of quantum
+ d 3 rj ψ (r
∗ #
) ψ j (r#
)ψ i (r) − ψ j (r)ψ i (r#
)δ m ,m . physics, although he also contributed
4π"0 |r − r# | j s i s j
significantly to the fields of me-
j"=i
chanics, theoretical optics, theory of
The first term in the last set of brackets translates to the ordinary Hartree contribution gravitation, physics of continuous
above and describes the influence of the charge density of the other electrons, while medium. In 1926 he derived the
Klein-Gordon equation. He gave
the second term describes the non-local exchange contribution, a manifestation of his name to Fock space, the Fock
particle statistics. representation and Fock state, and
developed the HartreeFock method
in 1930. Fock made significant
The outcome of such calculations is that the eigenfunctions are, as for contributions to general relativity
theory, specifically for the many
hydrogen, characterized by quantum numbers n, ', m! , with ' < n, but that the body problems.
states with different ' for a given n are not degenerate, with the lower values
of ' lying lower. This is because, for the higher ' values, the electrons tend
to lie further from the nucleus on average, and are therefore more effectively
screened. The states corresponding to a particular value of n are generally
referred to as a shell, and those belonging to a particular pair of values of n, '
are usually referred to as a subshell. The energy levels are ordered as below
(with the lowest lying on the left): !
"
Subshell name 1s 2s 2p 3s 3p 4s 3d 4p 5s 4d ··· !
!
"
! !
7s 7p
! ! 7d
n= 1 2 2 3 3 4 3 4 5 4 ··· ···
!
"
!
'= 0 0 1 0 1 0 2 1 0 2 ··· 6s !
6p !
6d 6f · · ·
!
" ! ! !
Degeneracy 2 2 6 2 6 2 10 6 2 10 ··· ! 5p 5d
5s ! ! 5g
! 5f
Cumulative 2 4 10 12 18 20 30 36 38 48 ··· " ! ! ! !
!
4s 4p
! ! 4d
! 4f!
Note that the values of Z corresponding to the noble gases, 2, 10, 18, 36, at " ! ! ! !!
!
! 3p
3s ! 3d!
which the ionization energy is unusually high, now emerge naturally from this " ! ! !!
!
filling order, corresponding to the numbers of electrons just before a new shell 2s
! !2p
" ! !!
!
(n) is entered. There is a handy mnemonic to remember this filling order. By 1s
! !
writing the subshells down as shown right, the order of states can be read off !
along diagonals from lower right to upper left, starting at the bottom.
We can use this sequence of energy levels to predict the ground state
electron configuration of atoms. We simply fill up the levels starting from
the lowest, accounting for the exclusion principle, until the electrons are all
accommodated (the aufbau principle). Here are a few examples:
Z Element Configuration 2S+1
LJ Ioniz. Pot. (eV)
1 H (1s) 2
S1/2 13.6
2 He (1s)2 1
S0 24.6
3 Li He (2s) 2
S1/2 5.4
4 Be He (2s)2 1
S0 9.3
5 B He (2s)2 (2p) 2
P1/2 8.3
6 C He (2s)2 (2p)2 3
P0 11.3
7 N He (2s)2 (2p)3 4
S3/2 14.5
8 O He (2s)2 (2p)4 3
P2 13.6
9 F He (2s)2 (2p)5 2
P3/2 17.4
10 Ne He (2s)2 (2p)6 1
S0 21.6
11 Na Ne (3s) 2
S1/2 5.1
9
Note that for N = 2, the determinant reduces to the familiar antisymmetric wavefunc-
tion, √12 [ψ1 (r1 )ψ2 (r2 ) − ψ2 (r1 )ψ1 (r2 )].

Figure 9.3: Periodic table of elements.
Since it is generally the outermost electrons which are of most interest, con-
tributing to chemical activity or optical spectra, one often omits the inner
closed shells, and just writes O as (2p)4 , for example. However, the configura-
tion is not always correctly predicted, especially in the heavier elements, where
levels may be close together. It may be favourable to promote one or even
two electrons one level above that expected in this simple picture, in order to
achieve a filled shell. For example, Cu (Z = 29) would be expected to have con-
figuration · · · (4s)2 (3d)9 , and actually has configuration · · · (4s)1 (3d)10 . There
are several similar examples in the transition elements where the d subshells
are being filled, and many among the lanthanides (rare earths) and actinides
where f subshells are being filled.
$ Info. Since the assignment of an electron configuration requires only the

enumeration of the values of n and ' for all electrons, but not those of m! and ms ,
each configuration will be accompanied by a degeneracy g. If νn! denotes the number
of electrons occupying a given level En,! , and δ! = 2 × (2' + 1) is the degeneracy of
that level, there are
δ! !
dn! = (9.6)
νn! !(δ! − νn! )!
ways of distributing the νn! electrons among the δ! individual states. The total
degeneracy, g, is then obtained from the product.
This scheme provides a basis to understand the periodic table of el-

ements (see Fig. 9.3). We would expect that elements which have similar
electron configurations in their outermost shells (such as Li, Na, K, Rb, Cs, Fr
which all have (ns)1 or F, Cl, Br, I, which all have (np)5 ) would have similar
chemical properties, such as valency, since it is the unpaired outer electrons
which especially participate in chemical bonding. Therefore, if one arranges
the atoms in order of increasing atomic number Z (which equals the number
of electrons in the atom), periodic behaviour is seen whenever a new subshell
of a given ' is filled.

9.3. COUPLING SCHEMES 102
9.3 Coupling schemes

The procedure outlined above allows us to predict the occupation of subshells
in an atomic ground state. This is not in general sufficient to specify the ground
state fully. If there are several electrons in a partially filled subshell, then their
spins and orbital angular momenta can combine in several different ways, to
give different values of total angular momentum, with different energies. In
order to deal with this problem, it is necessary to consider the spin-orbit
interaction as well as the residual Coulomb interaction between the outer
electrons.
Schematically we can write the Hamiltonian for this system as follows:
+ 1 e2 + +
Ĥ ≈ Ĥ0 + − Ui (r) + ξi (ri )L̂i · Ŝi ,
4π"0 rij
i<j i i
2 34 5 2 34 5
Ĥ1 Ĥ 2
where Ĥ0 includes the kinetic energy and central field terms, Ĥ1 is the residual
Coulomb interaction, and (with ξi (ri ) = 2m12 c2 1r (∂r V (r))) Ĥ2 is the spin-orbit
interaction. We can then consider two possible scenarios:
Ĥ1 2 Ĥ2 : This tends to apply in the case of light atoms. In this situation,
one considers first the eigenstates of Ĥ0 + Ĥ1 , and then treats Ĥ2 as a
perturbation. This leads to a scheme called LS (or Russell-Saunders)
coupling.
Ĥ2 2 Ĥ1 : This can apply in very heavy atoms, or in heavily ionized light
atoms, in which the electrons are moving at higher velocities and rela-
tivistic effects such as the spin-orbit interaction are more important. In
this case, a scheme called jj coupling applies.
It is important to emphasise that both of these scenarios represent approxima-

tions; real atoms do not always conform to the comparatively simple picture
which emerges from these schemes, which we now discuss in detail.
9.3.1 LS coupling scheme

In this approximation, we start by considering the eigenstates of Ĥ0 + Ĥ1 . We
note that this Hamiltonian must commute with the total angular momentum
Ĵ2 (because of invariance under rotations in space), and also clearly commutes
with the total spin Ŝ2 . It also commutes with the total orbital angular momen-
tum L̂2 , since Ĥ1 only involves internal interactions, and must therefore be
invariant under global rotation of all the electrons. Therefore the energy levels
can be characterised by the corresponding total angular momentum quantum
numbers L, S, J. Their ordering in energy is given by Hund’s rules:
1. Combine the spins of the electrons to obtain possible values of total spin
S. The largest permitted value of S lies lowest in energy.
2. For each value of S, find the possible values of total angular momentum
L. The largest value of L lies lowest in energy.
3. Couple the values of L and S to obtain the values of J (hence the name
of the scheme). If the subshell is less than half full, the smallest value
of J lies lowest; otherwise, the largest value of J lies lowest.

In deciding on the permitted values of L and S, in addition to applying the

usual rules for adding angular momenta, one also has to ensure that the exclu-
sion principle is respected, as we will see later when considering some examples.
These rules are empirical; there are exceptions, especially to the L and J
rules (2 and 3). Nevertheless, Hund’s rules are a useful guide, and we should
try to understand their physical origin.
1. Maximising S makes the spin wavefunction as symmetric as possible.

This tends to make the spatial wavefunction antisymmetric, and hence
reduces the Coulomb repulsion, as we saw when discussing the exchange
interactions in Helium.
2. Maximising L also tends to keep the electrons apart. This is less obvious,
though a simple classical picture of electrons rotating round the nucleus
in the same or different senses makes it at least plausible.
3. The separation of energies for states of different J arises from treating

the spin-orbit term Ĥ2 as a perturbation (fine structure). It can be
shown (using the Wigner-Eckart theorem – beyond the scope of these
lectures) that
+
%|J, mJ , L, S| ξi (ri )L̂i · Ŝi |J, mJ , L, S&
i
= ζ(L, S)%J, mJ , L, S|L̂ · Ŝ|J, mJ , L, S&
ζ(L, S)
= [J(J + 1) − L(L + 1) − S(S + 1)] , (9.7)
2
where the matrix element ζ(L, S) depends on the total L and S values.
Since one may show that the sign of ζ(L, S) changes according to the
whether the subshell is more or less than half-filled, the third Hund’s
rule is established.
To understand the application of LS coupling, it is best to work through

some examples. Starting with the simplest multi-electron atom, helium, the
ground state has an electron configuration (1s)2 , and must therefore have
L = S = J = 0. In fact, for any completely filled subshell, we have L = S = 0
and hence J = 0, since the total mL and mS must equal zero if all substates
are occupied. Consider now an excited state of helium, e.g. (1s)1 (2p)1 , in
which one electron has been excited to the 2p level. We can now have S = 1
or S = 0, with the S = 1 state lying lower in energy according to Hund’s
rules. Combining the orbital angular momenta of the electrons yields L = 1
and thus, with S = 0, J = 1, while with S = 1, J = 0, 1, 2 with J = 0 lying
lowest in energy.
Once again, as with the hydrogen-like states, we may index the states of
multi-electron atoms by spectroscopic term notation, 2S+1 LJ . The superscript
2S + 1 gives the multiplicity of J values into which the level is split by the
spin-orbit interaction; the L value is represented by a capital letter, S, P , D,
etc., and J is represented by its numerical value. Thus, for the (1s)1 (2p)1
state of helium, there are four possible states, with terms:
3
P0 3 P1 3 P2 1
P1 ,
where the three 3 P states are separated by the spin-orbit interaction, and the
singlet 1 P state lies much higher in energy owing to the Coulomb interaction.
The separations between the 3 P2 and 3 P1 and the 3 P1 and 3 P0 should be in the

ratio 2:1. This is an example of the Landé interval rule, which states that
the separation between a pair of adjacent levels in a fine structure multiplet is
proportional to the larger of the two J values involved. This is easily shown
using Eq. (9.7) – the separation in energy between states J and J − 1 is
∝ J(J + 1) − (J − 1)J = 2J .
Actually in the case of helium the situation is a bit more complicated, because
it turns out that the spin-orbit interaction between different electrons makes
a non-negligible additional contribution to the fine structure. Other excited
states of helium, of the form (1s)1 (n')1 , can be handled similarly, and again
separate into singlet and triplet states.
$ Exercise. For the case of boron, with the electron configuration (1s)2 (2s)2 (2p),
use Hund’s rules to show that the ground state is 2 P1/2 .
We next consider the case of carbon, which has ground state electron
configuration (1s)2 (2s)2 (2p)2 . This introduces a further complication; we now
have two identical electrons in the same unfilled subshell, and we need to
ensure that their wavefunction is antisymmetric with respect to electron ex-
change. The total spin can either be the singlet S = 0 state, which has an
antisymmetric wavefunction √12 [| ↑1 & ⊗ | ↓2 & − | ↓1 & ⊗ | ↑2 &], or one of the triplet
S = 1 states, which are symmetric, √12 [| ↑1 & ⊗ | ↓2 & + | ↓1 & ⊗ | ↑2 &], | ↑1 & ⊗ | ↑2 &
(1) (2)
or | ↓1 & ⊗ | ↓2 &. We must therefore choose values of L with the appropriate m! m! mL
symmetry to partner each value of S. To form an antisymmetric state, the two 1 0 1
1 −1 0
electrons must have different values of m! , so the possibilities are as shown
0 −1 −1
right. Inspecting the values of mL we can deduce that L = 1.10 By contrast,
to form a symmetric total angular momentum state, the two electrons may (1) (2)
have any values of m! , leading to the possibilities shown right. Inspecting the m! m! mL
1 1 2
values of mL we infer that L = 2 or 0.
1 0 1
We must therefore take S = 1 with L = 1 and S = 0 with L = 2 or 0. 1 −1 0
Finally, to account for the fine structure, we note that the states with S = 1 0 0 0
and L = 1 can be combined into a single J = 0 state, three J = 1 states, and 0 −1 −1
five J = 2 states leading to the terms 3 P0 , 3 P1 , and 3 P2 respectively. Similarly −1 −1 −2
the S = 0, L = 2 state can be combined to give five J = 2 states, 1 D2 , while the
S = 0, L = 0 state gives the single J = 0 state, 1 S0 . Altogther we recover the E /cm−1
1 + 3 + 5 + 5 + 1 = 15 possible states (cf. Eq. (9.6) with the ordering in energy 1
S0 20649
given by Hund’s rules (shown to the right). The experimental energy values 1
D2 10195
are given using the conventional spectroscopic units of inverse wavelength.
3
P2 43
Note that the Landé interval rule is approximately obeyed by the fine structure
3
P1 16
3
P0 0
triplet, and that the separation between L and S values caused by the electron-
electron repulsion is much greater than the spin-orbit effect.
In an excited state of carbon, e.g. (2p)1 (3p)1 , the electrons are no longer
equivalent, because they have different radial wavefunctions. So now one can
combine any of S = 0, 1 with any of L = 0, 1, 2, yielding the following terms
(in order of increasing energy, according to Hund’s rules):
3
D1,2,3 3
P0,1,2 3
S1 1
D2 1
P1 1
S0 .
For nitrogen, the electron configuration is given by (1s)2 (2s)2 (2p)3 . The
maximal value of spin is S = 3/2 while L can take values 3, 2, 1 and 0. Since
10
This result would also be apparent if we recall the that angular momentum states are
eigenstates of the parity operator with eigenvalue (−1)L . Since there are just two electrons,
this result shows that both the L = 0 and L = 2 wavefunction must be symmetric under
exchange.

the spin wavefunction (being maximal) is symmetric, the spatial wavefunction

must be completely antisymmetric. This demands that all three states with
m! = 1, 0, −1 must be involved. We must therefore have L = 0, leading to
J = 3/2 and the term, 4 S3/2 .
$ Exercise. Construct the L = 0 state involving the addition of three ' = 1

angular momentum states. Hint: make use of the total antisymmetry condition.
As a final example, let us consider the ground state of oxygen, which has
electron configuration (2p)4 . Although there are four electrons in the (2p)
subshell, the maximum value of S = 1. This is because there are only three
available values of m! = ±1, 0, and therefore one of these must contain two
electrons with opposite spins. Therefore, the maximum value of mS = 1,
achieved by having electrons with ms = + 12 in both the other m! states. By
pursuing this argument, it is quite easy to see that the allowed values of L, S Level scheme of the carbon
atom (1s)2 (2s)2 (2p)2 . Draw-
and J are the same as for carbon (2p)2 . This is in fact a general result – the
ing is not to scale. On the
allowed quantum numbers for a subshell with n electrons are the same as for left the energy is shown with-
that of a subshell with n “holes”. Therefore, the energy levels for the oxygen out any two-particle interaction.
ground state configuration are the same as for carbon, except that the fine The electron-electron interaction
structure multiplet is inverted, in accordance with Hund’s third rule. leads to a three-fold energy split-
ting with L and S remaining
good quantum numbers. Spin-
9.3.2 jj coupling scheme
orbit coupling leads to a further
When relativitic effects take precedence over electron interaction splitting of the states with J re-
6 effects, we maining a good quantum num-
must start by considering the eigenstates of Ĥ0 + Ĥ2 = Ĥ0 + i ξi (ri )L̂i · Ŝi .
ber. Finally on the right, the lev-
These must be eigenstates of Ĵ2 as before, because of the overall rotational els show Zeeman splittings in an
invariance, and also of Ĵ2i for each electron. Therefore, in this case, the cou- external magnetic field. In this
pling procedure is to find the allowed j values of individual electrons, whose case, the full set of 15 levels be-
energies will be separated by the spin-orbit interaction. Then these individual come non-degenerate.
j values are combined to find the allowed values of total J. The effect of the
residual Coulomb interaction will be to split the J values for a given set of js.
Sadly, in this case, there are no simple rules to parallel those of Hund.
As an example, consider a configuration (np)2 in the jj coupling scheme, to
be compared with the example of carbon which we studied in the LS scheme.
Combining s = 1/2 with ' = 1, each electron can have j = 1/2 or 3/2. If
the electrons have the same j value, they are equivalent, so we have to take
care of the symmetry of the wavefunction. We therefore have the following
possibilities:
$ j1 = j2 = 3/2 ⇒ J = 3, 2, 1, 0, of which J = 2, 0 are antisymmetric.
$ j1 = j2 = 1/2 ⇒ J = 1, 0, of which J = 0 is antisymmetric.
$ j1 = 1/2, j2 = 3/2 ⇒ J = 2, 1.
In jj coupling, the term is written (j1 , j2 )J , so we have the following terms in

our example:
(1/2, 1/2)0 (3/2, 1/2)1 (3/2, 1/2)2 (3/2, 3/2)2 (3/2, 3/2)0
in order of increasing energy. Note that both LS and jj coupling give the same
values of J (in this case, two states with J = 0, two with J = 2 and one with
J = 1) and in the same order. However, the pattern of levels is different; in
LS coupling we found a triplet and two singlets, while in this ideal jj scenario,
we have two doublets and a singlet. The sets of states in the two coupling

9.4. ATOMIC SPECTRA 106
schemes must be expressible as linear combinations of one another, and the

physical states for a real atom are likely to differ from either approximation.
In fact, this idealized form of jj coupling is not seen in the heaviest such
atom in the periodic table, lead (6p)2 . However, it is seen in some highly
ionized states, for example in Cr18+ , which has the same electron configu-
ration (2p)2 as carbon, but where, because of the larger unscreened charge
on the nucleus, the electrons are moving more relativistically, enhancing the
spin-orbit effect. However, a classic example of the transition from LS to jj
coupling is seen in the series C-Si-Ge-Sn-Pb in the excited states (2p)(3s),
(3p)(4s), · · · (6p)(7s) (see figure right). Here, the electrons are not in the same
subshell, so their wavefunctions overlap less, and the Coulomb repulsion is
reduced compared to the spin-orbit interaction. Analysing this situation in
the LS coupling approximation, one expects a triplet and a singlet:
3
P0,1,2 1
P1 ,
while in the jj scheme one expects two doublets:
(1/2, 1/2)0,1 (1/2, 3/2)2,1 .
Experimentally, C and Si conform to the LS expectation and Pb to the jj

scheme, while Ge and Sn are intermediate.
9.4 Atomic spectra

Atomic spectra result from transitions between different electronic states of
an atom via emission or absorption of photons. In emission spectra, an
atom is excited by some means (e.g. thermally through collisions, or by an
electric discharge), and one observes discrete spectral lines in the light emitted
as the atoms relax. In absorption spectra, one illuminates atoms using a
broad waveband source, and observes dark absorption lines in the spectrum
of transmitted light. Of course the atoms excited in this process subsequently
decay by emitting photons in random directions; by observing in directions
away from the incident light this fluorescence radiation may be studied. The
observation of these spectral lines is an important way of probing the atomic
energy levels experimentally. In the case of optical spectra and the nearby
wavebands, the excited states responsible generally involve the excitation of
a single electron from the ground state to some higher level. In some cases,
it may simply involve a different coupling of the angular momenta within the
same electron configuration. These are the kinds of excitations which we are
about to discuss. However, other types of excitations are also possible. For
example, X-ray emission occurs when an electron has been removed from one
of the innermost shells of a heavy atom; as electrons cascade down to fill the
hole, high energy photons may be emitted.
The basic theory governing stimulated emission and absorption, and spon-
taneous emission of photons will be outlined in detail when we study radiative
transitions in chapter 13. Here we must anticipate some of the basic con-
clusions of that study. In the electric dipole approximation, the rate of
transitions is proportional to the square of the matrix element of the electric
dipole operator between the initial and final states, |%ψf |d̂|ψi &|2 . In addition,
the rate of spontaneous transitions is proportional to ω 3 , where ω = |Ef − Ei |
denotes the energy separation between the states.
The form of the dipole operator, d̂ means that the matrix elements may
vanish identically. This leads to a set of selection rules defining which tran-
sitions are allowed. Here we consider the simplest case of a single electron, but

the principles can be generalized. Referring to chapter 13 for a more detailed

discussion, one finds that, for a transition to take place:
$ Parity must change;
$ ∆J = ±1, 0 (but 0 → 0 is not allowed) and ∆MJ = ±1, 0 .
Atomic states are always eigenstates of parity and of total angular momentum,
J, so these selection rules can be regarded as absolutely valid in electric dipole
transitions. It should be emphasized again, though, that the electric dipole
approximation is an approximation, and higher order processes may occur,
albeit at a slower rate, and have their own selection rules.
In specific coupling schemes, further selection rules may apply. In the case
of ideal LS coupling, we also require:
$ ∆S = 0 and ∆MS = 0;
$ ∆L = ±1, 0 (but 0 → 0 is not allowed) and ∆ML = ±1, 0;
$ and ∆'i = ±1 if only electron i is involved in the transition.
In LS coupling, the states are eigenstates of total spin; since the dipole operator
does not operate on the spin part of the wavefunction, the rules on ∆S and
∆MS follow straightforwardly. This, and the absolute rules relating to J,
imply the rules for L and ML . The rule for ∆'i follows from the parity change
rule, since the parity of the atom is the product of the parities of the separate
electron wavefunctions, given by (−1)!i . However, since LS coupling is only
an approximation, these rules should themselves be regarded as approximate.
With this preparation, we now turn to the consequences of the selection rules
on the atomic spectra of atoms.
9.4.1 Single electron atoms

In this context, “single electron atoms” refer to atoms whose ground state
consists of a single electron in an s level, outside closed shells; it is this electron
which is active in optical spectroscopy. Our discussion therefore encompasses
the alkali metals, such as sodium, and also hydrogen. We take sodium,
whose ground state configuration is (3s)1 , as our example:
$ The ground state has term 2 S1/2 . The excited states are all doublets
with J = L ± 1/2, except for the s states, which are obviously restricted
to J = 1/2.
$ The parity is given by (−1)!i , so the allowed transitions involve changes

in ' by ±1 unit, i.e. s ↔ p, p ↔ d, d ↔ f , etc. Changes of more than
one unit in ' would fall foul of the ∆J rule.
$ The s ↔ p transitions are all doublets. All the doublets starting or

ending on a given p state have the same spacing in energy. The transition
3s ↔ 3p gives rise to the familiar yellow sodium “D-lines” at 589 nm
(see right).
$ The p ↔ d transitions involve two doublets, 2 P1/2,3/2 and 2 D3/2,5/2 .

However, the 2 P1/2 ↔2 D5/2 transition is forbidden by the ∆J rule, so
the line is actually a triplet. In practice, the spin-orbit interaction falls
quite rapidly with increasing ' (and with increasing n) as the effect of
screening increases, so that the effect of the 2 D3/2,5/2 splitting may not
be resolved experimentally.

$ As n increases, the energy levels approach (from below) those for hy-
drogen, because the nuclear charge is increasingly effectively screened
by the inner electrons. This happens sooner for the higher ' values, for
which the electron tends to lie further out from the nucleus.
$ In an absorption spectrum, the atoms will start from the ground state,
so only the 3s → np lines will be seen. In emission, the atoms are excited
into essentially all their excited levels, so many more lines will be seen
in the spectrum.
The comments above for sodium also apply for hydrogen, except that, in this
case, (2s, 2p), (3s, 3p, 3d), etc. are degenerate. One consequence is that the 2s
state in hydrogen is metastable – it cannot decay to the only lower lying level
(1s) by an electric dipole transition. In fact its favoured spontaneous decay
is by emission of two photons; a process which is described by second-order
perturbation theory. In practice, hydrogen atoms in a 2s state are more likely
to deexcite through collision processes. During an atomic collision, the atoms
are subject to strong electric fields, and we know from our discussion of the
Stark effect that this will mix the 2s and 2p states, and decay from 2p to 1s
is readily possible.
9.4.2 Helium and alkali earths

We next discuss atoms whose ground state consists of two electrons in an s
level. Our discussion therefore covers helium (1s)2 , and the alkali earths:
beryllium (2s)2 , magnesium (3s)2 , calcium (4s)2 , etc. We start with helium.
$ The ground state has term 1 S0 . The excited states are of the form
(1s)(n') (the energy required to excite both of the 1s electrons to higher
states is greater than the first ionization energy, and therefore these
form discrete states within a continuum of ionized He+ +e− states). The
excited states can have S = 0 or S = 1, with S = 1 lying lower in energy
(Hund).
$ The LS coupling approximation is pretty good for helium, so the ∆S = 0

selection rule implies that the S = 0 and S = 1 states form completely
independent systems as far as spectroscopy is concerned.
$ The lines in the S = 0 system are all singlets. They can be observed
in emission, and those starting from the ground state can be seen in
absorption.
$ The lines in the S = 1 system are all multiplets. They can be observed
in emission only. Transitions of the form 3 S1 ↔3 P2,1,0 are observed as
triplets, spaced according to the Landé interval rule. Transitions of the
form 3 P2,1,0 ↔3 D3,2,1 are observed as sextuplets, as is easily seen by
application of the ∆J = ±1, 0 rule. Actually, as mentioned above, the
fine structure is a little more subtle in the case of helium.
The alkali earths follow the same principles. In the case of calcium, the
triplet 4p state is the lowest lying triplet state, and therefore metastable. In
fact a faint emission line corresponding to the 3 P1 →1 S0 decay to the ground
state may be observed; this violates the ∆S = 0 rule, and indicates that the LS
coupling approximation is not so good in this case. A more extreme example is
seen in Mercury, ground state (6s)2 (5d)10 . Excited states involving promotion
of one of the 6s electrons to a higher level can be treated just like the alkali

9.5. ZEEMAN EFFECT 109
earths. In this case the “forbidden” 3 P1 →1 S0 is actually a prominent feature

of the emission spectrum in the visible, implying a significant breakdown of
the LS approximation.
9.4.3 Multi-electron atoms

Similar principles can be used to make sense of the spectra of more complicated
atoms, though unsurprisingly their structure is more complex. For example,
carbon, with ground state (2s)2 (2p)2 , corresponds to terms 3 P0,1,2 , 1 D2 and
1 S as discussed above. The excited states are of the form (2s)2 (2p)1 (n')1 ,
0
and can be separated into singlets and triplets, and in addition excitations
of the form (2s)1 (2p)3 can arise. Nitrogen, with ground state (2s)2 (2p)3 , has
three unpaired electrons, so the ground state and excited states form doublets
(S = 1/2) and quartets (S = 3/2) with correspondingly complex fine structure
to the spectral lines.
9.5 Zeeman effect

9.5.1 Single-electron atoms
Before leaving this chapter on atomic structure, we are now in a position to
revisit the question of how atomic spectra are influenced by a magnetic field.
To orient our discussion, let us begin with the study of hydrogen-like atoms
involving just a single electron. In a magnetic field, the Hamiltonian of such
a system is described by Ĥ = Ĥ0 + Ĥrel. + ĤZeeman , where Ĥ0 denotes the
non-relativistic Hamiltonian for the atom, Ĥrel. incorporates the relativistic
corrections considered earlier in the chapter, and
eB
ĤZeeman = − (L̂z + 2Ŝz ) ,
2mc
denotes the Zeeman energy associated with the coupling of the spin and orbital
angular momentum degrees of freedom to the magnetic field. Here, since
we are dealing with confined electrons, we have neglected the diamagnetic
contribution to the Hamiltonian. Depending on the scale of the magnetic field,
the spin-orbit term in Ĥrel. or the Zeeman term may dominate the spectrum
of the atom.
Previously, we have seen that, to leading order, the relativistic corrections
lead to a fine-structure energy shift of
! " ! "
1 2 Zα 4 3 n
∆En,j = mc
rel.
− ,
2 n 4 j + 1/2
for states |n, j = ' ± 1/2, mj , '&. For weak magnetic fields, we can also treat
the Zeeman energy in the framework of perturbation theory. Here, although
states with common j values (such as 2S1/2 and 2P1/2 ) are degenerate, the
two spatial wavefunctions have different parity (' = 0 and 1 in this case),
and the off-diagonal matrix element of ĤZeeman coupling these states vanishes.
We may therefore avoid using degenerate perturbation theory. Making use of
the relation (exercise – refer back to the discussion of the addition of angular
momenta and spin in section 6.4.2),
!mj
%n, j = ' ± 1/2, mj , '|Sz |n, j = ' ± 1/2, mj , '& = ± ,
2' + 1
we obtain the following expression for the first order energy shift,
! "
1
∆Ej=!±1,mj ,! = ' ± 1/2, mj , '& = µB Bmj 1 ±
Zeeman
,
2' + 1

Figure 9.4: The well known doublet which is responsible for the bright yellow light
from a sodium lamp may be used to demonstrate several of the influences which
cause splitting of the emission lines of atomic spectra. The transition which gives
rise to the doublet is from the 3p to the 3s level. The fact that the 3s state is lower
than the 3p state is a good example of the dependence of atomic energy levels on
orbital angular momentum. The 3s electron penetrates the 1s shell more and is less
effectively shielded than the 3p electron, so the 3s level is lower. The fact that there
is a doublet shows the smaller dependence of the atomic energy levels on the total
angular momentum. The 3p level is split into states with total angular momentum
J = 3/2 and J = 1/2 by the spin-orbit interaction. In the presence of an external
magnetic field, these levels are further split by the magnetic dipole energy, showing
dependence of the energies on the z-component of the total angular momentum.
where µB denotes the Bohr magneton. Therefore, we see that all degenerate
levels are split due to the magnetic field. In contrast to the “normal” Zeeman
effect, the magnitude of the splitting depends on '.
$ Info. If the field is strong, the Zeeman energy becomes large in comparison
with the spin-orbit contribution. In this case, we must work with the basis states
|n, ', m! , ms & = |n, ', m! & ⊗ |ms & in which both Ĥ0 and ĤZeeman are diagonal. Within
first order of perturbation theory, one then finds that (exercise)
! "4 ! "
1 Zα 3 n nm! ms
∆En,!,m! ,ms = µB (m! + ms ) + mc2 − − ,
2 n 4 ' + 1/2 '(' + 1/2)(' + 1)
the first term arising from the Zeeman energy and the remaining terms from Ĥrel. . At
intermediate values of the field, we have to apply degenerate perturbation theory to
the states involving the linear combination of |n, j = '±1/2, mj , '&. Such a calculation
reaches beyond the scope of these lectures and, for details, we refer to the literature
(see, e.g., Ref. [6]. Let us instead consider what happens in multi-electron atoms.
9.5.2 Multi-electron atoms

For a multi-electron atom in a weak magnetic field, the appropriate unper-
turbed states are given by |J, MJ , L, S&, where J, L, S refer to the total
angular momenta. To determine the Zeeman energy shift, we need to deter-
mine the matrix element of Ŝz . To do so, we can make use of the following

argument. Since 2L̂ · Ŝ = Ĵ2 − L̂2 − Ŝ2 , this operator is diagonal in the basis of
states, |J, MJ , L, S&. Therefore, the matrix element of the operator (exercise,
hint: recall that [Ŝi , Ŝj ] = i!"ijk Ŝk and [L̂i , Ŝk ] = 0),
−i!Ŝ × L̂ ≡ Ŝ(L̂ · Ŝ) − (L̂ · Ŝ)Ŝ
must vanish. Moreover, from the identity [L̂ · Ŝ, Ĵ] = 0, it follows that the
matrix element of the vector product,
−i!(Ŝ × L̂) × Ĵ = Ŝ × Ĵ(L̂ · Ŝ) − (L̂ · Ŝ)Ŝ × Ĵ ,
must also vanish. If we expand the left hand side, we thus find that the matrix
element of
L̂=Ĵ−Ŝ
(Ŝ × L̂) × Ĵ = L̂(Ŝ · Ĵ) − Ŝ(L̂ · Ĵ) = Ĵ(Ŝ · Ĵ) − ŜĴ2 ,
also vanishes. Therefore, it follows that %ŜĴ2 & = %Ĵ(Ŝ · Ĵ)&, where the expec-
tation value is taken over the basis states. Then, with Ŝ · Ĵ = 12 (Ĵ2 + Ŝ2 − L̂2 ),
we have that
J(J + 1) + S(S + 1) − L(L + 1) Figure 9.5: In the weak field
%Ŝz & = %Jˆz & . case, the vector model (top) im-
2J(J + 1)
plies that the coupling of the or-
As a result, we can deduce that, at first order in perturbation theory, the bital angular momentum L to
energy shift arising from the Zeeman term is given by the spin angular momentum S
is stronger than their coupling
∆EJ,MJ ,L,S = µB gJ MJ B , to the external field. In this
case where spin-orbit coupling is
where dominant, they can be visual-
ized as combining to form a to-
J(J + 1) + S(S + 1) − L(L + 1) tal angular momentum J which
gJ = 1 + , then precesses about the mag-
2J(J + 1) netic field direction. In the
strong field case, S and L cou-
denotes the effective Landé g-factor, which lies between 1 and 2. Note that, ple more strongly to the exter-
in the special case of hydrogen, where S = 1/2 and J = L ± 1/2, we recover nal magnetic field than to each
our previous result. The predicted Zeeman splitting for sodium is shown in other, and can be visualized as
figure 9.4. independently precessing about
the external field direction.
$ Info. In the strong field limit, where the influence of Zeeman term domi-
nates, the appropriate basis states are set by |L, ML , S, MS &, in which the operators
L̂2 , L̂z , Ŝ 2 , Ŝz , and ĤZeeman are diagonal. In this case, the energy splitting takes the
form
! "4
1 Zα nML MS
∆EL,ML ,S,MS = µB B(ML + 2MS ) + mc2 ,
2 n '(' + 1/2)(' + 1)
where the second term arises from the spin-orbit interaction.

Atomic Structure

Uploaded by

Copyright:

Available Formats

Atomic Structure

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Atomic Structure

Uploaded by

Copyright:

Available Formats

Chapter 9

Hydrogen atom revisited:

$ As with any centrally symmetric potential, the solutions of the Schrödinger

Advanced Quantum Physics

$ Since n ≥ 1 and nr ≥ 0, the allowed combinations of quantum num-

$ relativistic corrections to the kinetic energy;

$ coupling between spin and orbital degrees of freedom;

$ and a contribution known as the Darwin term.

In the following, we will discuss each of these corrections in turn.

9.1.1 Relativistic correction to the kinetic energy

Advanced Quantum Physics

We can therefore make use of a perturbative analysis to estimate the scale of

%n'm|Ĥ1 |n'! m! & = 0 for ' )= '! or m )= m! ,

degenerate states are uncoupled and such an approach is unnecessary. Then

9.1.2 Spin-orbit coupling

$ Info. Physically, the origin of the spin-orbit interaction can be understoon

Advanced Quantum Physics

and combining both of these expressions, for ' > 0, we obtain

Advanced Quantum Physics

9.1.3 Darwin term

Intriguingly, this term is formally identical to that which would be obtained

a result that is independent of ' and mj .

while the 2P3/2 experiences a shift of ∆E2,3/2,mj ,1 = − 64 Z α Ry. Finally, for

Advanced Quantum Physics

Figure 9.1: Figure showing the

for states with ' = 0.

Advanced Quantum Physics

9.1.5 Hyperfine structure

Therefore, the 1s state of Hydrogen is split into two, corresponding to the

Advanced Quantum Physics

is an important transition for radio astronomy. A further contribution to the

This completes our discussion of the “one-electron” theory. We now turn

9.2 Multi-electron atoms

Advanced Quantum Physics

Figure 9.2: Ionization energies of the elements.

9.2.1 Central field approximation

it is apparent that a closed shell has an electron density distribution which

Advanced Quantum Physics

Now, according to the variational principle, we must minimize the energy

Advanced Quantum Physics

Finally, to relate the Lagrange multipliers, εi (which have the appearance

To summarize, if we wish to implement the central field approximation to

1. Firstly, one makes an initial “guess” for a (common) central potential,

2. With these potentials, Ui (r), we can determine a new set of eigenstates

3. An estimate for the ground state energy of an atom can be found by

4. Using these wavefunctions, one can then make an improved estimate of

Since the practical implemention of such an algorithm demands a large degree

$ Info. An improvement to this procedure, known the Hartree-Fock method,

Advanced Quantum Physics

Advanced Quantum Physics

Figure 9.3: Periodic table of elements.

$ Info. Since the assignment of an electron configuration requires only the

This scheme provides a basis to understand the periodic table of el-

Advanced Quantum Physics

9.3 Coupling schemes

It is important to emphasise that both of these scenarios represent approxima-

9.3.1 LS coupling scheme

Advanced Quantum Physics