Atomic Structure and Spectra: Selection Rules

182 5 Atomic structure and spectra
Visible the selection rules are derived by identifying the transitions

2000 Wavelength, λ/nm that conserve angular momentum when a photon (a boson of
1000
800
600
500
400
300
200
150
120
100
spin 1) is emitted or absorbed:
Hydrogenic Selection
Δl = ±1 Δml = 0, ± 1 atoms rules
(21.2)
The principal quantum number n can change by any amount

Balmer Lyman
consistent with the Δl for the transition, because it does not relate
Analysis
directly to the angular momentum. The mathematical basis of

Paschen the selection rules is developed in the following Justification.
Brackett
Brief illustration 21.1 Selection rules
Figure 21.1 The spectrum of atomic hydrogen. Both the To identify the orbitals to which a 4d electron may make radi-
observed spectrum and its resolution into overlapping series ative transitions, we first identify the value of l and then apply
are shown. Note that the Balmer series lies in the visible region. the selection rule for this quantum number. Because l = 2, the
final orbital must have l = 1 or 3. Thus, an electron may make
The form of eqn 21.1 strongly suggests that each spectral line a transition from a 4d orbital to any np orbital (subject to
arises from a transition, a jump from one state to another, each Δm l = 0, ±1) and to any nf orbital (subject to the same rule).
with an energy proportional to R H / n2 , with the difference in However, it cannot undergo a transition to any other orbital,
energy discarded as electromagnetic radiation of frequency so a transition to any ns orbital or to another nd orbital is
(and wavenumber) given by the Bohr frequency condition forbidden.
(ΔE = hν). As is shown in Topic 17, this is precisely the form of Self-test 21.2 To what orbitals may an electron in a 4s orbital
the energy levels of a hydrogen atom, and its explanation was make electric-dipole allowed radiative transitions?
an early triumph of quantum mechanics. Answer: to np orbitals only
Example 21.1 Calculating the shortest and longest

wavelength lines in a series Justification 21.1 The identification of selection rules
Determine the shortest and longest wavelength lines in the
To determine the selection rules for atoms, we need to iden-
Balmer series.
tify the conditions for which the transition dipole moment, μ fi,
Method Identify the value of n1 for the Balmer series. The (Topic 16) connecting the final state ψ f and the initial state ψ i
shortest-wavelength line corresponds to the largest wave- is nonzero:
number; from eqn 21.1, recognize that this line will arise from
n 2 = ∞. The longest wavelength corresponds to the smallest
wavenumber, which will arise from n2 = n1 + 1.
∫
μq ,n0 = ψ n* μ q ψ 0dτ q = x, y, z
Answer The Balmer series corresponds to n1 = 2. The larg- where μ q = −eq . We consider each component in turn. To evalu-
est wavenumber (use n 2 = ∞), calculated from eqn 21.1, is ate the integral, we note from Table 14.1 that z = (4π/3)1/2rY1,0, so
27 419 cm−1, which corresponds to a wavelength of 365 nm. The
z
smallest wavenumber (use n2 = 3) is 15 233 cm−1, which corre- 2π ψ f*
1/2 ψ i

∞ π
⎛4π ⎞
sponds to a wavelength of 656 nm.
∫ψ *
n μ z ψ 0dτ = ∫∫∫
0 0 0
Rnf lf Ylf*ml ,f ⎜
⎝ ⎟
3 ⎠
rY1,0 Rni li Yli ml ,i
Self-test 21.1 Calculate the shortest and longest wavelength dτ

lines in the Paschen series. r 2dr sinθ dθ dφ
Answer: 821 nm, 1876 nm
This multiple integral is the product of three factors, an inte-
gral over r and two integrals over the angles, so the factor on
the right can be grouped as follows:
(b) Selection rules
1/2
⎛ 4π ⎞
Although Topic 17 establishes the allowed energy levels of a
hydrogenic atom as −Z 2 hcR H /n2 , not all conceivable transi-
∫ ψ f* μzψ i dτ = ⎜ ⎟
⎝ 3 ⎠
∞ π 2π
tions are allowed. As explained in Topic 16, it is necessary to
identify and apply the selection rules. For hydrogenic atoms,
∫0
Rnf lf rRni li r 2dr ∫∫
0 0
Ylf*ml ,f Y1,0Yli ml ,i sinθ dθ dφ
www.ebook3000.com
Atkins09819.indb 182 9/11/2013 11:16:28 AM

21 Atomic spectroscopy 183
to identify these states and find a way to label them with a term
We now use the property of spherical harmonics that
symbol, a symbol that specifies the state.
π 2π
The key to identifying the various states that can arise from
∫∫
0 0
Yl ′′ml′′ (θ , φ )* Yl ′ml′ (θ , φ )Ylml (θ , φ ) sinθ dθ dφ = 0
a configuration and attaching a term symbol is the angular
momentum of the electrons: that includes their orbital angu-
unless l, l′, and l″ are integers denoting lengths of lines that lar momentum, their spin, and their total angular momentum.
can form the sides of a triangle (such as 1, 2, and 3, or 1, 1, Our first job is to identify the allowed values of these angular
and 1, but not 1, 2, and 4) and m l + m l ′ + m l ″ = 0. It follows momenta for atoms with more than one electron.
that the angular integral (blue) is zero unless l f = l i ± 1 and
ml,f = ml,i + m. Because m = 0 in the present case, the angular
integral, and hence the z-component of the transition dipole (a) The total orbital angular momentum
moment, is zero unless Δl = ±1 and Δm1 = 0, which is a part of
Consider an atom with two electrons outside a closed core, so
the set of selection rules. The same procedure, but considering
the x- and y-components (Problem 21.7), results in the com-
there are two sources of orbital angular momentum. We sup-
plete set of rules. pose that the orbital angular momentum quantum numbers
of the two electrons are l1 and l2. If the electron configuration
of the atoms we are considering is p2, both electrons are in
s p d p orbitals and l1 = l2 = 1. The total orbital angular momenta that
Paschen
can arise from a configuration depends on the magnitudes
Balmer
and the relative orientation of these individual momenta, and
15 328 (Hα) is described by the total orbital angular momentum quan-
20 571 (Hβ) tum number, L, a non-negative integer obtained by using the
23 039 (Hγ)
24 380 (Hδ) Clebsch–Gordan series:
82 259 L = l1 + l2 , l1 + l2 −1, …, l1 − l2 Clebsch–Gordan series (21.3)

97 492
Lyman 102 824
For instance, if l1 = l2 = 1, then L = 2, 1, and 0. Once we know the
value of the L we can calculate the magnitude of the total orbital
angular momentum from {L(L + 1)}1/2ħ. As for any angular
Figure 21.2 A Grotrian diagram that summarizes the
momentum, the total orbital angular momentum has 2L + 1
appearance and analysis of the spectrum of atomic hydrogen.
orientations, distinguished by the quantum number ML, which
The thicker the line, the more intense the transition. The
can take the values L, L – 1, …, –L.
wavenumbers of the transitions (in cm−1) are indicated.
Just as we use lowercase letters to tell us the value of l, so we
The selection rules and the atomic energy levels jointly use an uppercase letter to tell us the value of L. The code for
account for the structure of a Grotrian diagram (Fig. 21.2), converting the value of L into a letter is the same as for the s, p,
which summarizes the energies of the states and the transitions d, f,… designation of orbitals, but uses uppercase letters:
between them. In some versions, lines of various thicknesses Thus a p2 configuration can give rise to D, P, and S terms. A
are used to denote their relative intensities in the spectrum
obtained by evaluating the transition dipole moments. L: 0 1 2 3 4 5 6…
S P D F G H I…
closed shell has zero orbital angular momentum because all the
21.2 Term symbols individual orbital angular momenta sum to zero. Therefore,
when working out term symbols, we need consider only the
The spectra of many-electron atoms are considerably richer electrons of the unfilled shell. In the case of a single electron
than those of hydrogenic atoms. Although Topic 19 explains outside a closed shell, the value of L is the same as the value of l;
how to account for the ground-state configurations of atoms, so the configuration [Ne]3s1 has only an S term.
to understand their spectra we need to examine their states
in more detail, especially their excited states, and to consider Example 21.2Deriving the total orbital angular
the transitions between them. A complication we immedi-
momentum of a configuration
ately encounter is that a single configuration of an atom, such
as the excited configuration of He, 1s12s1 for instance, or the Find the terms that can arise from the configurations (a) d 2 ,
ground state of C, [He]2s22p2, can give rise to a number of dif- (b) p3.
ferent individual states with various energies. Our first task is
Atkins09819.indb 183 9/11/2013 11:16:32 AM

When there are several electrons to be taken into account,

Method Use the Clebsch–Gordan series and begin by finding
the minimum value of L (so that we know where the series ter- we must assess their total spin angular momentum quantum
minates). When there are more than two electrons to couple number, S (a non-negative integer or half integer). To do so, we
together, use two series in succession: first couple two elec- use the Clebsch–Gordan series, this time in the form
trons, and then couple the third to each combined state, and
so on. S = s1 + s2 , s1 + s2 − 1, …, s1 − s2 (21.4)
Answer (a) The minimum value is |l 1 – l 2 | = |2 – 2| = 0. noting that each electron has s = 12 , which gives S = 1, 0 for two
Therefore, electrons (Fig. 21.3). If there are three electrons, the total spin
L = 2 + 2, 2 + 2 − 1, …, 0 = 4, 3, 2, 1, 0 angular momentum is obtained by coupling the third spin to
each of the values of S for the first two spins, which results in
corresponding to G, F, D, P, S terms, respectively. (b) Coupling S = 23 and S = 12 .
two electrons gives a minimum value of |1 – 1| = 0. Therefore, The value of S for a term is expressed by giving the multi-
plicity of a term, the value of 2S + 1, as a left-superscript on the
L ′ = 1 + 1, 1 + 1 − 1, …, 0 = 2, 1, 0
term symbol. Thus, 1P is a ‘singlet’ term (S = 0, 2S + 1 = 1) and 3P
is a ‘triplet’ term (S = 1, 2S + 1 = 3). The multiplicity actually tells
Now couple l 3 = 1 with L′ = 2, to give L = 3, 2, 1; with L′ = 1, to
give L = 2, 1, 0; and with L′ = 0, to give L = 1. The overall result is
us the number of permitted values of MS = S, S – 1,…, –S for the
given value of S, and hence the number of orientations in space
L = 3, 2, 2, 1, 1, 1, 0 that the total spin can adopt. We shall see the importance of
this information shortly.
giving one F, two D, three P, and one S term.
Self-test 21.3 Repeat the question for the configurations
A note on good practice Throughout our discussion of atomic
(a) f1d1 and (b) d3.
spectroscopy, distinguish italic S, the total spin quantum
Answer: (a) H, G, F, D, P; (b) I, 2H, 3G, 4F, 5D, 3P, S
number, from Roman S, the term label. Thus, 3S is a triplet
term with S = 1 (and L = 0). All state symbols are upright;
all quantum numbers and physical observables are oblique
(sloping).
The terms that arise from a given configuration differ in
energy due to the Coulombic interaction between the elec-
trons. For example, to achieve a D (L = 2) term from a 2p13p1
configuration, both electrons need to be circulating in the same
direction around the nucleus, but to achieve an S (L = 0) term,
they would need to be circulating in opposite directions. In the
former arrangement, they do not meet; in the latter they do. MS = +1
On the basis of this classical picture, we can suspect that the
repulsion between them will be higher if they meet, and there-
fore that the S term will lie higher in energy than the D term.
The quantum mechanical analysis of the problem supports this
interpretation.
MS = –1
MS = 0
(b) The total spin angular momentum (a) S = 0 (b) S = 1
Figure 21.3 (a) Electrons with paired spins have zero resultant
The energy of a term also depends on the relative orientation of spin angular momentum (S = 0). They can be represented
the electron spins. Topic 20 provides a hint of this dependence, by two vectors that lie at an indeterminate position on the
where it explains that spin correlation results in states with par- cones shown here, but wherever one lies on its cone, the
allel spins having a lower energy than states with antiparallel other points in the opposite direction; their resultant is zero.
spins. Parallel (unpaired) and antiparallel (paired) spins differ (b) When two electrons have parallel spins, they have a
in their overall spin angular momentum. In the paired case, the nonzero total spin angular momentum (S = 1). There are three
two spin momenta cancel each other, and there is zero net spin ways of achieving this resultant, which are shown by these
(as is depicted in Fig. 19.2), and so once again we are brought to vector representations. Note that, whereas two paired spins
a correlation between an angular momentum, in this case spin, are precisely antiparallel, two ‘parallel’ spins are not strictly
and an energy. parallel.
www.ebook3000.com
Atkins09819.indb 184 9/11/2013 11:16:39 AM

Brief illustration 21.2 combine these angular momenta into a total angular momen-
The multiplicities of terms
tum and—perhaps—for the energy of the atom to depend on
When S = 0 (as for a closed shell, like 1s2), MS = 0, the electron its value. The total angular momentum quantum number, J
spins are all paired, and there is no net spin: this arrange- (a non-negative integer or half integer), takes the values
ment gives a singlet term, 1S. A single electron has S = s = 12
(MS = ms = ± 12 ), so a configuration such as [Ne]3s1 can give rise J = L + S, L + S −1, …, L − S Total angular momentum (21.5)
to a doublet term, 2S. Likewise, the configuration [Ne]3p1 is a
doublet, 2P. When there are two electrons with unpaired spins, and, as usual, we can calculate the magnitude of the total angu-
S = 1 (MS = ±1, 0), so 2S + 1 = 3, giving a triplet term, such as 3D. lar momentum from {J(J + 1)}1/2ħ. The specific value of J is given
Self-test 21.4 What terms can arise from scandium in the as a right-subscript on the term symbol; for example, a 3P term
excited configuration [Ar]3s23p14p1? with J = 2 is fully dressed as 3P2.
Answer: 1,3D,1,3P,1,3S If S ≤ L, there are 2S + 1 values of J for a given L, so the num-
ber of values of J is the same as the multiplicity of the term.
Each possible value of J designates a level of a term, so provided
S ≤ L, the multiplicity tells us the number of levels. For example,
It is explained in Topic 20 that the energies of two states, the [Ne]3p1 configuration of sodium (an excited state) has L = 1
one with paired spins and one with unpaired spins, differ on and S = 12 and a multiplicity of 2; the two levels are J = 23 and 12
account of the different effects of spin correlation. The fact that and the 2P term therefore has two levels, 2P3/2 and 2P1/2.
the parallel arrangement of spins, as in the 1s12s1 configura- Before moving on, we should note that there is a hidden
tion of the He atom, lies lower in energy than the antiparallel assumption in eqn 21.5. We have assumed that the orbital angular
arrangement can now be expressed by saying that the triplet momenta of the electrons all combine to give a total orbital angu-
state of the 1s12s1 configuration of He lies lower in energy lar momentum, that their spins all combine to give a total spin,
than the singlet state. This is a general conclusion that applies and that only then do these two totals combine to give the overall
to other atoms (and molecules), and for states arising from the total angular momentum of the atom. This procedure is called
same configuration, the triplet state generally lies lower in energy Russell–Saunders coupling. An alternative is that the orbital and
than the singlet state. The latter is an example of Hund’s rule of spin angular momenta of each electron combine separately into a
maximum multiplicity (Topic 20) which can be restated as: resultant for each one (with quantum number j), and then those
resultants combine to give an overall total. We shall not deal with
For a given conf iguration, the term of greatest
this so-called jj-coupling case: Russell–Saunders coupling turns
multiplicity lies lowest in energy.
out to be reasonably accurate for light atoms.
Because the Coulombic interaction between electrons in an
atom is strong, the difference in energies between singlet and
triplet states of the same configuration can be large. The triplet
Example 21.3 Deriving term symbols
and singlet terms of He1s12s1, for instance, differ by 6421 cm−1
(corresponding to 0.80 eV). Write the term symbols arising from the ground-state config-
Hund’s rule of maximum multiplicity is the first of three urations of (a) Na and (b) F, and (c) the excited-state configu-
rules devised by Friedrich Hund to identify the lowest energy ration 1s22s22p13p1 of C.
term of a configuration with the minimum of calculation. The
Method Begin by writing the configurations, but ignore inner
second rule is
closed shells. Then couple the orbital momenta to find L and the
For a given multiplicity, the term with the highest value spins to find S. Next, couple L and S to find J. Finally, express
of L lies lowest in energy. the term as 2S+1{L}J, where {L} is the appropriate letter. For F, for
which the valence configuration is 2p5, treat the single gap in
Therefore, as discussed above, a D term is expected to lie lower the closed-shell 2p6 configuration as a single particle.
in energy than an S term of the same multiplicity. The third rule
is introduced below after a discussion of spin–orbit coupling. Answer (a) For Na, the configuration is [Ne]3s1, and we con-
The three rules are reliable only for the ground-state configura- sider the single 3s electron. Because L = l = 0 and S = s = 12 , it is
tion of an atom. possible for J = j = s = 12 only. Hence the term symbol is 2S1/2. (b)
For F, the configuration is [He]2s22p5, which we can treat as
[Ne]2p −1 (where the notation 2p −1 signifies the absence of a
(c) The total angular momentum 2p electron). Hence L = 1, and S = s = 12 . Two values of J = j are
When there is a net orbital angular momentum and a net spin allowed: J = 23 , 12 . Hence, the term symbols for the two levels
angular momentum in an atom, we can expect to be able to
Atkins09819.indb 185 9/11/2013 11:16:41 AM

However, the relative orientation of the two momenta also

are 2P3/2 , 2P1/2. (c) We are treating an excited configuration of
carbon because, in the ground configuration, 2p2 , the Pauli determines the electron’s total angular momentum, so there is
principle (Topic 19) forbids some terms, and deciding which a correlation between the energy of interaction and the value
survive (1D, 3P, 1S, in fact) is quite complicated.1 That is, there of J. Magnetic dipole moments that are antiparallel (that is, lie
is a distinction between ‘equivalent electrons’, which are elec- in opposite directions) are lower in energy than when they are
trons that occupy the same orbitals, and ‘inequivalent elec- parallel; therefore, a lower energy is achieved when the orbital
trons’, which are electrons that occupy different orbitals; we and spin angular momenta are antiparallel, corresponding to a
consider only the latter here. The excited configuration of C lower value of J. In the case of the 2P term, we predict that the
2P 2
under consideration is effectively 2p13p1. This is a two-electron 1/2 level is lower in energy than the P3/2 level. For a system
problem, and l1 = l 2 = 1, s1 = s2 = 12 . It follows that L = 2, 1, 0 and with many electrons, a detailed analysis yields the following
S = 1, 0. The terms are therefore 3D and 1D, 3P and 1P, and 3S general statement, which is the third of Hund's rules:
and 1S. For 3D, L = 2 and S = 1; hence J = 3, 2, 1 and the levels are
3D , 3D , and 3D . For 1D, L = 2 and S = 0, so the single level is
3 2 1 Type of configuration Order of levels
1D . The triplet of levels of 3P is 3P , 3P , and 3P , and the singlet
2 2 1 0
Less than half-full shell Lowest value of J lies lowest in energy
is 1P1. For the 3S term there is only one level, 3S1 (because J = 1
only), and the singlet term is 1S0. More than half-full shell Highest value of J lies lowest in energy
Self-test 21.5 Write down the terms arising from the configu-
rations (a) 2s12p1, (b) 2p13d1. For a quantitative treatment of spin–orbit coupling, we
Answer: (a) 3P2, 3P1, 3P0, 1P1; need to include in the hamiltonian a term that depends on the
(b) 3F4, 3F3, 3F2, 1F3, 3D3, 3D2, 3D1, 1D2, 3P2, 3P1, 3P0, 1P1 relative orientation of the vectors that represent the spin and
orbital angular momenta. The simplest procedure is to write
the contribution as
(d) Spin–orbit coupling H so = λL ⋅ S λ = hcA / 2 Spin–orbit coupling (21.6)
The different levels of a term, such as 2P1/2 and 2P3/2, have dif- where λ is a measure of the strength of the coupling, and for
ferent energies due to spin–orbit coupling, a magnetic inter- practical purposes best expressed as a wavenumber by intro-
action between angular momenta. To see the origin of this ducing the parameter A. The quantity L⋅S is the scalar product
coupling, we need to note that a circulating current gives rise to of the vectors L and S (as we see in Mathematical background 4,
a magnetic moment (Fig. 21.4). The spin of an electron is one L⋅S is proportional to cos θ, where θ is the angle between the
source of magnetic moment and its orbital angular momentum two vectors, so the expression models the fact that the energy of
is another. Two magnetic dipole moments close to each other interaction depends on the relative orientation of the two mag-
interact to an extent that depends on their relative orientation. netic moments). To use this expression, we note that the total
angular momentum is J = L + S, so
l
l J ⋅ J = (L + S)⋅(L + S) = L2 + S 2 + 2L ⋅ S
High j
Low j and therefore (because J⋅J = J2)
s
λL ⋅ S = 1
2
λ ( J 2 − L2 − S 2 )
We now treat the J2, L2, and S2 as operators with eigenvalues

J(J + 1)ħ2, L(L + 1)ħ2, and S(S + 1)ħ2, respectively. It then follows
that the eigenvalues of H so are
High Low s
(a) energy (b) energy
Spin–orbit
Figure 21.4 Spin–orbit coupling is a magnetic interaction Eso = 12 hcA { J ( J + 1) − L(L + 1) − S(S + 1)} coupling (21.7)
between spin and orbital magnetic moments. When the energy
angular momenta are parallel, as in (a), the magnetic moments
are aligned unfavourably; when they are opposed, as in (b), the
interaction is favourable. This magnetic coupling is the cause of 1 For details, see our Inorganic chemistry, Oxford University Press and
the splitting of a configuration into levels. W. H. Freeman & Co. (2014).
www.ebook3000.com
Atkins09819.indb 186 9/11/2013 11:16:48 AM

Brief illustration 21.3 Spin–orbit coupling energy Brief illustration 21.4 Fine structure
When L = 1 and S = 1 2
2 , as in a P term,
Fine structure can be seen in the emission spectrum from
sodium vapour excited by an electric discharge (for exam-
{ }
Eso = 12 hcA J ( J + 1) − 2 − 34 = 12 hcA J ( J + 1) − 114 { } ple, in one kind of street lighting). The yellow line at 589 nm
(close to 17 000 cm−1) is actually a doublet composed of one
Therefore, for a level with J = 23 , Eso = 12 hcA , and for a level with line at 589.76 nm (16 956.2 cm−1) and another at 589.16 nm
J = 12 from the same configuration, Eso = −hcA . The separation (16 973.4 cm−1); the components of this doublet are the ‘D lines’
of the two levels is therefore ΔEso = 23 hcA . of the spectrum (Fig. 21.5). Therefore, in Na, the spin–orbit
Self-test 21.6 Confirm that the spin–orbit interaction leaves coupling affects the energies by about 17 cm−1.
the mean energy of the 2P term unchanged. Hint: Take account Self-test 21.7 In the emission spectrum of potassium lines are
of the degeneracies of the two levels. observed at 766.70 nm and 770.11 nm. What is the spin–orbit
Answer: 4( 12 hcA ) + 2(−hcA ) = 0 coupling constant of potassium?
Answer: 57.75 cm−1
The strength of the spin–orbit coupling as measured by A

depends on the nuclear charge. To understand why this is so,
imagine riding on the orbiting electron and seeing a charged
nucleus apparently orbiting around us (like the Sun rising and set- Selection rules of many-electron
21.3
ting). As a result, we find ourselves at the centre of a ring of cur- atoms
rent. The greater the nuclear charge, the greater this current, and
therefore the stronger the magnetic field we detect. Because the Any state of the atom, and any spectral transition, can be speci-
spin magnetic moment of the electron interacts with this orbital fied by using term symbols. For example, the transitions giving
magnetic field, it follows that the greater the nuclear charge, the rise to the yellow sodium doublet (which were shown in Fig.
stronger the spin–orbit interaction. The coupling increases sharply 21.5) are
with atomic number (as Z4 in hydrogenic atoms). Whereas it is
only small in H (giving rise to shifts of energy levels of no more
than about 0.4 cm−1), in heavy atoms like Pb it is very large (giving 3p1 2 P3/2 → 3s1 2 S1/2 3p1 2 P1/2 → 3s1 2 S1/2
shifts of the order of thousands of reciprocal centimetres).
Two spectral lines are observed when the p electron of an By convention, the upper term precedes the lower. The corre-
electronically excited alkali metal atom undergoes a transition sponding absorptions are therefore denoted 2P3/2 ← 2S1/2 and
and falls into a lower s orbital. The higher-frequency line is due 2P ← 2S
1/2 1/2 (the configurations have been omitted).
to a transition starting in a 2P3/2 level and the other line is due to We have seen (in Justification 21.1) that selection rules arise
a transition starting in the 2P1/2 level of the same configuration. from the conservation of angular momentum during a transi-
The presence of these two lines is an example of fine structure, tion and from the fact that a photon has a spin of 1. They can
the structure in a spectrum due to spin–orbit coupling. therefore be expressed in terms of the term symbols, because
the latter carry information about angular momentum. A
2
P3/2 detailed analysis leads to the following rules:
17 cm–1 16 973
Wavenumber, ν/cm–1
2
P1/2
16 956
~
D1 D2 ΔS = 0 ΔL = 0, ± 1 Δl = ±1 Many-
Selection
electron (21.8)
ΔJ = 0, ± 1, but J = 0 ← → J =0 atoms
rules
589.16 nm
589.76 nm
2
S1/2 where the symbol ←|→ denotes a forbidden transition. The
rule about ΔS (no change of overall spin) stems from the fact
0
Wavenumber, ν~ that the light does not affect the spin directly. The rules about
ΔL and Δl express the fact that the orbital angular momentum
Figure 21.5 The energy–level diagram for the formation of the of an individual electron must change (so Δl = ±1), but whether
sodium D lines. The splitting of the spectral lines (by 17 cm−1) or not this results in an overall change of orbital momentum
reflects the splitting of the levels of the 2P term. depends on the coupling.
Atkins09819.indb 187 9/11/2013 11:16:57 AM

Brief illustration 21.5 The selection rules given above apply when Russell–Saunders
Selection rules
coupling is valid (in light atoms). If we insist on labelling the
If we were presented with the following possible transitions terms of heavy atoms with symbols like 3D, then we shall find
in the emission spectrum of a many-electron atom, namely that the selection rules progressively fail as the atomic number
3D → 3P , 3P → 1S , and 3F → 3D we could decide which are
2 1 2 0 4 3 increases because the quantum numbers S and L become ill
allowed by constructing the following table and referring to defined as jj-coupling becomes more appropriate. As explained
the rules in eqn 21.8. Forbidden values are in red. above, Russell–Saunders term symbols are only a convenient
way of labelling the terms of heavy atoms: they do not bear any
ΔS ΔL ΔJ
direct relation to the actual angular momenta of the electrons
2→ P1
3D 3 0 –1 –1 Allowed in a heavy atom. For this reason, transitions between singlet
3P
2 → 1S0 –1 –1 –2 Forbidden and triplet states (for which ΔS = ±1), while forbidden in light
4 → D3 atoms, are allowed in heavy atoms.
3F 3 0 –1 –1 Allowed
Self-test 21.8 Which of the transitions (a) 2 P 3/2 → 2 S1/2 ,

(b) 3P0 → 3S1, (c) 3D3 → 1P1 are allowed?
Answer: (a), (b)
Checklist of concepts
☐ 1. The Lyman, Balmer, and Paschen series in the spec- ☐ 8. The allowed values of the total orbital angular momen-
trum of atomic hydrogen arise, respectively, from the tum L of a configuration are obtained by using the
transitions n → 1, n → 2, and n → 3. Clebsch–Gordan series L = l1 + l2, l1 + l2 –1, ..., |l1 – l2|.
☐ 2. The wavenumbers of all the spectral lines of a hydrogen ☐ 9. The allowed values of the total spin angular momen-
atom can be expressed in terms of transitions between tum S are obtained by using the Clebsch–Gordan series
allowed energy levels. S = s1 + s2, s1 + s2 –1, ..., |s1 – s2|.
☐ 3. A Grotrian diagram summarizes the energies of the ☐ 10. Spin–orbit coupling is the interaction of the spin mag-
states and the transitions between them. netic moment with the magnetic field arising from the
☐ 4. A level is a group of states with a common value of J. orbital angular momentum.
☐ 5. The multiplicity of a term is the value of 2S + 1; pro- ☐ 11. Russell–Saunders coupling is a coupling scheme based
vided L ≥ S, the multiplicity is the number of levels of on the view that if spin–orbit coupling is weak, then it is
the term. effective only when all the orbital momenta are operat-
☐ 6. A term symbol is a symbolic specification of the state of ing cooperatively.
an atom, 2S+1{L}J. ☐ 12. The total angular momentum J, in the Russell–
☐ 7. Hund’s rules, which allow the identification of the low- Saunders coupling scheme, has possible values J = L + S,
est energy term of a configuration, can be expressed as: L + S –1, ..., |L – S|.
r The term with the maximum multiplicity lies lowest ☐ 13. Fine structure is the structure in a spectrum due to
in energy. spin–orbit coupling.
r For a given multiplicity, the term with the highest ☐ 14. The selection rules for spectroscopic transitions in
value of L lies lowest in energy. polyelectronic atoms are set out in the Checklist of
r For atoms with less than half-filled shells, the level equations below. They apply when Russell–Saunders
with the lowest value of J lies lowest in energy; for coupling is valid.
more than half-filled shells, the highest value of J.
www.ebook3000.com
Atkins09819.indb 188 9/11/2013 11:16:58 AM

Checklist of equations
Property Equation Comment Equation number
Rydberg formula ␯ = R H (1/n12 −1/n22 ) n2 = n1 + 1, n1 + 2, … 21.1

Selection rules Δl = ±1, Δml = 0, ±1 Hydrogenic atoms 21.2
Clebsch–Gordan series J = j1 + j2, j1 + j2 – 1, …, |j1 – j2| j an angular momentum quantum number 21.3
Spin–orbit coupling energy Eso = hcA { J ( J +1) − L(L +1) − S(S +1)}
1
2
Russell–Saunders coupling 21.7
Selection rules ΔS = 0 , ΔL = 0, ±1, Δl = ±1, ΔJ = 0, ±1, but Many-electron atoms and Russell–Saunders 21.8
J = 0 ←|→ J = 0 coupling
Atkins09819.indb 189 9/11/2013 11:17:00 AM

Focus 5 on Atomic structure and spectra
Topic 17 Hydrogenic atoms

Discussion questions
17.1 Discuss the separation of variables procedure as it is applied to simplify 17.2 List and discuss the significance of the quantum numbers needed to
the description of a hydrogenic atom free to move through space. specify the internal state of a hydrogenic atom.
Exercises
17.1(a) Compute the ionization energy of the He+ ion. 17.3(b) The wavefunction for the 2s orbital of a hydrogen atom is
17.1(b) Compute the ionization energy of the Li2+ ion. N (2 − r /a0 )e −r /2a0 . Determine the normalization constant N.
17.2(a) When ultraviolet radiation of wavelength 58.4 nm from a helium lamp 17.4(a) By differentiation of the 2s radial wavefunction, show that it has two
is directed on to a sample of krypton, electrons are ejected with a speed of extrema in its amplitude and locate them.
1.59 × 106 m s−1. Calculate the ionization energy of krypton. 17.4(b) By differentiation of the 3s radial wavefunction, show that it has three
17.2(b) When ultraviolet radiation of wavelength 58.4 nm from a helium lamp extrema in its amplitude and locate them.
is directed on to a sample of xenon, electrons are ejected with a speed of
17.5(a) Locate the radial nodes in the 3p orbital of an H atom.
1.79 × 106 m s−1. Calculate the ionization energy of xenon.
17.5(b) Locate the radial nodes in the 3d orbital of an H atom.
17.3(a) The wavefunction for the ground state of a hydrogen atom is Ne − r /a0 .
Determine the normalization constant N.
Problems
17.1 The Humphreys series is a group of lines in the spectrum of atomic instead, with analogous definitions of units of length and energy, what would
hydrogen. It begins at 12 368 nm and has been traced to 3281.4 nm. What be the relation between these two sets of atomic units?
are the transitions involved? What are the wavelengths of the intermediate
17.4 Show that the radial wave equation (eqn 17.5) can be written in the form
transitions?
of eqn 17.6 by introducing the function u = rR.
17.2 A series of lines in the spectrum of atomic hydrogen lies at 656.46 nm,
17.5 Hydrogen is the most abundant element in all stars. However, neither
486.27 nm, 434.17 nm, and 410.29 nm. What is the wavelength of the next line
absorption nor emission lines due to neutral hydrogen are found in the
in the series? What is the ionization energy of the atom when it is in the lower
spectra of stars with effective temperatures higher than 25 000 K. Account for
state of the transitions?
this observation.
17.3 Atomic units of length and energy may be based on the properties of a
17.6 The initial value of the principal quantum number n was not specified in
particular atom. The usual choice is that of a hydrogen atom, with the unit of
Example 17.3. Show that the correct value of n can be determined by making
length being the Bohr radius, a0, and the unit of energy being the (negative
several choices and selecting the one that leads to a straight line.
of the) energy of the 1s orbital. If the positronium atom (e+,e−) were used
Topic 18 Hydrogenic atomic orbitals

18.1 Describe how the presence of orbital angular momentum affects the 18.2 Discuss the significance of (a) a boundary surface and (b) the radial
shape of the atomic orbital. distribution function for hydrogenic orbitals.
Exercises
18.1(a) What is the orbital angular momentum of an electron in the orbitals 18.1(b) What is the orbital angular momentum of an electron in the orbitals
(a) 2s, (b) 3p, (c) 5f? Give the numbers of angular and radial nodes in each case. (a) 3d, (b) 4f, (c) 3s? Give the numbers of angular and radial nodes in each case.
www.ebook3000.com
Atkins09819.indb 190 9/11/2013 11:17:03 AM

Exercises and problems 191
18.2(a) What is the degeneracy of an energy level in the L shell of a hydrogenic 18.5(b) Compute the mean radius and the most probable radius for a 2p
atom? electron in a hydrogenic atom of atomic number Z.
18.2(b) What is the degeneracy of an energy level in the N shell of a
18.6(a) Write down the expression for the radial distribution function of a 3s
hydrogenic atom?
electron in a hydrogenic atom and determine the radius at which the electron
18.3(a) State the orbital degeneracy of the levels in a hydrogen atom that have is most likely to be found.
energy: (a) −hcR̃ H; (b) − 14 hcR H ; (c) − 16
1
hcR H . 18.6(b) Write down the expression for the radial distribution function of a 3p
18.3(b) State the orbital degeneracy of the levels in a hydrogenic atom (Z electron in a hydrogenic atom and determine the radius at which the electron
in parentheses) that have energy: (a) –hcR̃ atom (2); (b) − 14 hcR atom (4); is most likely to be found.
(c) − 16
25
hcR atom (5).
18.7(a) Locate the angular nodes and nodal planes of each of the 2p orbitals of
18.4(a) Calculate the average kinetic and potential energies of an electron in a hydrogenic atom of atomic number Z. To locate the angular nodes, give the
the ground state of a He+ ion. angle that the plane makes with the z-axis.
18.4(b) Calculate the average kinetic and potential energies of a 3s electron in 18.7(b) Locate the angular nodes and nodal planes of each of the 3d orbitals of
an H atom. a hydrogenic atom of atomic number Z. To locate the angular nodes, give the
angle that the plane makes with the z-axis.
18.5(a) Compute the mean radius and the most probable radius for a 2s
electron in a hydrogenic atom of atomic number Z.
Problems
18.1 In 1976 it was mistakenly believed that the first of the ‘superheavy’ plane polar plots and boundary-surface plots for these orbitals. Construct
elements had been discovered in a sample of mica. Its atomic number was the boundary plots so that the distance from the origin to the surface is
believed to be 126. What is the most probable distance of the innermost the absolute value of the angular part of the wavefunction. Compare the
electrons from the nucleus of an atom of this element? (In such elements, s, p, and d boundary surface plots with that of an f orbital; for example,
relativistic effects are very important, but ignore them here.) ψ f ∝ x(5z 2 − r 2 ) ∝ sinθ (5cos2θ −1) cos φ .
18.2 (a) Calculate the probability of the electron being found anywhere within 18.6 Show that d orbitals with opposite values of ml may be combined in pairs
a sphere of radius 53 pm for a hydrogenic atom. (b) If the radius of the atom to give real standing waves with boundary surfaces, as shown in Fig. 18.7, and
is defined as the radius of the sphere inside which there is a 90 per cent with forms that are given in Table 18.1.
probability of finding the electron, what is the atom's radius?
18.7 As in Problem 18.2, the ‘size’ of an atom is sometimes considered to be
18.3 At what point in the hydrogen atom is there maximum probability of measured by the radius of a sphere that contains 90 per cent of the probability
finding a (a) 2pz electron, (b) 3pz electron? How do these most probable density of the electrons in the outermost occupied orbital. Explore how the
points compare to the most probable radii for the locations of 2pz and 3pz ‘size’ of a ground-state hydrogenic atom varies as the definition is changed to
electrons? other percentages, and plot your conclusion.
18.4 Show by explicit integration that hydrogenic (a) 1s and 2s orbitals are 18.8 A quantity important in some branches of spectroscopy is the probability
mutually orthogonal, (b) 2px and 2pz orbitals are mutually orthogonal. density of an electron being found at the same location as the nucleus.
Evaluate this probability density for an electron in the 1s, 2s, and 3s orbitals of
18.5‡ Explicit expressions for hydrogenic orbitals are given in Table 18.1.
a hydrogenic atom. What happens to the probability density when an orbital
(a) Verify that the 3px orbital is normalized and that 3px and 3dxy are
other than s is considered?
mutually orthogonal. (b) Determine the positions of both the radial nodes
and nodal planes of the 3s, 3px, and 3dxy orbitals. (c) Determine the mean 18.9 Some atomic properties depend on the average value of 1/r rather than
radius of the 3s orbital. (d) Draw a graph of the radial distribution function the average value of r itself. Evaluate the expectation value of 1/r for (a) a
for the three orbitals (of part (b)) and discuss the significance of the graphs hydrogen 1s orbital, (b) a hydrogenic 2s orbital, (c) a hydrogenic 2p orbital.
for interpreting the properties of many-electron atoms. (e) Create both xy-
Topic 19 Many-electron atoms

19.1 Distinguish between a fermion and a boson. Provide examples of each 19.3 Compare and contrast the properties of spin angular momentum and the
type of particle. properties of angular momentum arising from rotational motion in two and
three dimensions.
19.2 Describe the orbital approximation for the wavefunction of a
many-electron atom. What are the limitations of the approximation?
‡ These problems were supplied by Charles Trapp and Carmen Giunta.
Atkins09819.indb 191 9/11/2013 11:17:08 AM

Exercises
19.1(a) The classical picture of an electron is that of a sphere of radius re = 2.82 19.1(b) A proton has a spin angular momentum with I = 1 . Suppose it is a
2
fm. On the basis of this model, how fast is a point on the equator of the sphere of radius 1 fm. On the basis of this model, how fast is a point on the
electron moving? Is this answer plausible? equator of the proton moving?
Problems
19.1 Derive an expression in terms of l and ml for the half-angle of the apex deflection is given by x = ±(μBL2/4Ek)dB/dz, where μB = e /2me = 9.274 × 10−24 J
of the cone used to represent an angular momentum according to the vector T−1 is known as the Bohr magneton (see inside front cover), L is the length of
model. Evaluate the expression for an α spin. Show that the minimum the magnet, Ek is the average kinetic energy of the atoms in the beam, and
possible angle approaches 0 as l → ∞. dB/dz is the magnetic field gradient across the beam. (a) Given that the
average translational kinetic energy of the atoms emerging as a beam from
19.2‡ Stern–Gerlach splittings of atomic beams are small and require either
a pinhole in an oven at temperature T is ½kT, calculate the magnetic field
large magnetic field gradients or long magnets for their observation. For a
gradient required to produce a splitting of 2.00 mm in a beam of Ag atoms
beam of atoms with zero orbital angular momentum, such as H or Ag, the
from an oven at 1200 K with a magnet of length 80 cm.
Topic 20 Periodicity
20.1 Discuss the relationship between the location of a many-electron atom in 20.2 Describe and account for the variation of first ionization energies
the periodic table and its electron configuration. along Period 2 of the periodic table. Would you expect the same variation in
Period 3?
Exercises
20.1(a) What are the values of the quantum numbers n, l, ml, s, and ms for each 20.2(a) Write the ground-state electron configurations of the d metals from
of the valence electrons in the ground state of a carbon atom? scandium to zinc.
20.1(b) What are the values of the quantum numbers n, l, ml, s, and ms for each 20.2(b) Write the ground-state electron configurations of the d metals from
of the valence electrons in the ground state of a nitrogen atom? yttrium to cadmium.
Problems
20.1 The d metals iron, copper, and manganese form cations with different Aluminium, which causes anaemia and dementia, is also a member of the
oxidation states. For this reason, they are found in many oxidoreductases and group but its chemical properties are dominated by the +3 oxidation state.
in several proteins of oxidative phosphorylation and photosynthesis. Explain Examine this issue by plotting the first, second, and third ionization energies
why many d metals form cations with different oxidation states. for the Group 13 elements against atomic number. Explain the trends you
observe. Hint: The third ionization energy, I3, is the minimum energy needed
20.2 Thallium, a neurotoxin, is the heaviest member of Group 13 of
to remove an electron from the doubly charged cation: E2+(g) → E3+(g) + e−(g),
the periodic table and is found most usually in the +1 oxidation state.
I3 = E(E3+) − E(E2+). Consult the printed or online literature for sources of data.
Topic 21 Atomic spectroscopy

21.1 Discuss the origin of the series of lines in the emission spectra of 21.3 Explain the origin of spin–orbit coupling and how it affects the
hydrogen. What region of the electromagnetic spectrum is associated with appearance of a spectrum.
each of the series shown in Fig. 21.1?
21.2 Specify and account for the selection rules for transitions in hydrogenic
atoms. Are they strictly valid?
www.ebook3000.com
Atkins09819.indb 192 9/11/2013 11:17:09 AM

Exercises and problems 193
Exercises
21.1(a) Determine the shortest-and longest-wavelength lines in the Lyman 21.7(b) What are the allowed total angular momentum quantum numbers of a
series. composite system in which j1 = 4 and j2 = 2?
21.1(b) The Pfund series has n1 = 5. Determine the shortest-and longest-
21.8(a) What information does the term symbol 3P2 provide about the angular
wavelength lines in the Pfund series.
momentum of an atom?
21.2(a) Compute the wavelength, frequency, and wavenumber of the 21.8(b) What information does the term symbol 2D3/2 provide about the
n = 2 → n = 1 transition in He+. angular momentum of an atom?
21.2(b) Compute the wavelength, frequency, and wavenumber of the
21.9(a) Suppose that an atom has (a) 2, (b) 3 electrons in different orbitals.
n = 5 → n = 4 transition in Li+2.
What are the possible values of the total spin quantum number S? What is the
21.3(a) Which of the following transitions are allowed in the normal electronic multiplicity in each case?
emission spectrum of an atom: (a) 3s → 1s, (b) 3p → 2s, (c) 5d → 2p? 21.9(b) Suppose that an atom has (a) 4, (b) 5, electrons in different orbitals.
21.3(b) Which of the following transitions are allowed in the normal electronic What are the possible values of the total spin quantum number S? What is the
emission spectrum of an atom: (a) 5d → 3s, (b) 5s → 3p, (c) 6f → 4p? multiplicity in each case?
21.4(a) (i) Write the electron configuration of the Pd2+ ion. (ii) What are the 21.10(a) What atomic terms are possible for the electron configuration ns1nd1?
possible values of the total spin quantum numbers S and MS for this ion? Which term is likely to lie lowest in energy?
21.4(b) (i) Write the electron configuration of the Nb2+ ion. (ii) What are the 21.10(b) What atomic terms are possible for the electron configuration np1nd1?
possible values of the total spin quantum numbers S and MS for this ion? Which term is likely to lie lowest in energy?
21.5(a) Calculate the permitted values of j for (a) a d electron, (b) an f electron. 21.11(a) What values of J may occur in the terms (a) 3S, (b) 2D, (c) 1P? How
21.5(b) Calculate the permitted values of j for (a) a p electron, (b) an h electron. many states (distinguished by the quantum number MJ) belong to each level?
21.11(b) What values of J may occur in the terms (a) 3F, (b) 4G, (c) 2P? How
21.6(a) An electron in two different states of an atom is known to have j = 5
2 many states (distinguished by the quantum number MJ) belong to each level?
and 12 . What is its orbital angular momentum quantum number in each
case? 21.12(a) Give the possible term symbols for (a) Na [Ne]3s1, (b) K [Ar]3d1.
21.6(b) An electron in two different states of an atom is known to have j = 72 21.12(b) Give the possible term symbols for (a) Y [Kr]4d15s2,
and 23 . What is its orbital angular momentum quantum number in each case? (b) I [Kr]4d105s25p5.
21.7(a) What are the allowed total angular momentum quantum numbers of a
composite system in which j1 = 1 and j2 = 2?
Problems
21.1 The Li2+ ion is hydrogenic and has a Lyman series at 740 747 cm−1, 21.5 Calculate the mass of the deuteron given that the first line in the Lyman
877 924 cm−1, 925 933 cm−1, and beyond. Show that the energy levels are of series of H lies at 82 259.098 cm−1 whereas that of D lies at 82 281.476 cm−1.
the form –hcR̃ /n2 and find the value of R̃ for this ion. Go on to predict the Calculate the ratio of the ionization energies of H and D.
wavenumbers of the two longest-wavelength transitions of the Balmer series
21.6 Positronium consists of an electron and a positron (same mass, opposite
of the ion and find the ionization energy of the ion.
charge) orbiting round their common centre of mass. The broad features
21.2 A series of lines in the spectrum of neutral Li atoms rise from of the spectrum are therefore expected to be hydrogen-like, the differences
combinations of 1s22p1 2P with 1s2nd1 2D and occur at 610.36 nm, 460.29 nm, arising largely from the mass differences. Predict the wavenumbers of the first
and 413.23 nm. The d orbitals are hydrogenic. It is known that the 2P term three lines of the Balmer series of positronium. What is the binding energy of
lies at 670.78 nm above the ground state, which is 1s22s1 2S. Calculate the the ground state of positronium?
ionization energy of the ground-state atom.
21.7 Some of the selection rules for hydrogenic atoms were derived in
21.3‡ Wijesundera, et al. (Phys. Rev. A 51, 278 (1995)) attempted to determine Justification 21.1. Complete the derivation by considering the x- and y-
the electron configuration of the ground state of lawrencium, element 103. components of the electric dipole moment operator.
The two contending configurations are [Rn]5f147s27p1 and [Rn]5f146d17s2.
21.8 The distribution of isotopes of an element may yield clues about the
Write down the term symbols for each of these configurations, and identify
nuclear reactions that occur in the interior of a star. Show that it is possible
the lowest level within each configuration. Which level would be lowest
to use spectroscopy to confirm the presence of both 4He+ and 3He+ in a star
according to a simple estimate of spin–orbit coupling?
by calculating the wavenumbers of the n = 3 → n = 2 and of the n = 2 → n = 1
21.4 The characteristic emission from K atoms when heated is purple and lies transitions for each isotope.
at 770 nm. On close inspection, the line is found to have two closely spaced
components, one at 766.70 nm and the other at 770.11 nm. Account for this
observation, and deduce what information you can.
Integrated activities
F5.1 The Zeeman effect is the modification of an atomic spectrum by the applied magnetic fields and the magnetic moments due to orbital and spin
application of a strong magnetic field. It arises from the interaction between angular momenta (recall the evidence provided for electron spin by the
Atkins09819.indb 193 9/11/2013 11:17:12 AM

Stern–Gerlach experiment). To gain some appreciation for the so-called of Li2+ allowed and, if so, what are the frequency and wavenumber of the
normal Zeeman effect, which is observed in transitions involving singlet states, transition(s)?
consider a p electron, with l = 1 and ml = 0, ±1. In the absence of a magnetic
F5.4‡ Highly excited atoms have electrons with large principal quantum
field, these three states are degenerate. When a field of magnitude B is present,
numbers. Such Rydberg atoms have unique properties and are of interest
the degeneracy is removed and it is observed that the state with ml = +1
to astrophysicists. Derive a relation for the separation of energy levels for
moves up in energy by μBB, the state with ml = 0 is unchanged, and the state
hydrogen atoms with large n. Calculate this separation for n = 100; also
with ml = –1 moves down in energy by μBB, where μB = e /2me = 9.274 × 10−24
calculate the average radius, the geometric cross-section, and the ionization
J T−1 is known as the Bohr magneton. Therefore, a transition between a
1S term and a 1P term consists of three spectral lines in the presence of a energy. Could a thermal collision with another hydrogen atom ionize this
0 1
Rydberg atom? What minimum velocity of the second atom is required?
magnetic field where, in the absence of the magnetic field, there is only one.
Could a normal sized neutral H atom simply pass through the Rydberg atom
(a) Calculate the splitting in reciprocal centimetres between the three spectral
leaving it undisturbed? What might the radial wavefunction for a 100s orbital
lines of a transition between a 1S0 term and a 1P1 term in the presence of a
be like?
magnetic field of 2 T (where 1 T = 1 kg s−2 A−1). (b) Compare the value you
calculated in (a) with typical optical transition wavenumbers, such as those F5.5 Use mathematical software, a spreadsheet, or the Living graphs (labelled
for the Balmer series of the H atom. Is the line splitting caused by the normal LG) on the website of this book for the following exercises:
Zeeman effect relatively small or relatively large?
(a)LG Plot the effective potential energy of an electron in the hydrogen atom
F5.2 An electron in the ground-state He+ ion undergoes a transition to a state against r for several nonzero values of the orbital angular momentum l. How
described by the wavefunction R4,1(r)Y1,1(θ,ϕ)(a) Describe the transition using does the location of the minimum in the effective potential energy vary
term symbols. (b) Compute the wavelength, frequency, and wavenumber of with l?
the transition. (c) By how much does the mean radius of the electron change (b) Find the locations of the radial nodes in hydrogenic wavefunctions with n
due to the transition? up to 3.
(c) Plot the boundary surfaces of the real parts of the spherical harmonics
F5.3 The electron in a Li2+ ion is prepared in a state that is the following
Yl ,ml (θ , φ ) for l = 1. The resulting plots are not strictly the p orbital boundary
superposition of hydrogenic atomic orbitals:
surfaces, but sufficiently close to be reasonable representations of the shapes
ψ (r , θ , φ ) = −( 13 )1/2 R4,2 (r )Y2, −1 (θ , φ ) + 23 iR3,2 (r )Y2,1 (θ , φ ) − ( 29 )1/2 R1,0 (r )Y0,0 (θ , φ ) of hydrogenic orbitals.
(d) To gain insight into the shapes of the f orbitals, plot the boundary surfaces
(a) If the total energy of different Li2+ ions in this state is measured, what
of the real parts of the spherical harmonics Yl ,ml (θ , φ ) for l = 3.
(e) Calculate and plot the radial distribution functions for the hydrogenic 4s,
values will be found? If more than one value is found, what is the probability
of obtaining each result and what is the average value? 4p, 4d, and 4f orbitals. How does the degree of shielding experienced by an
(b) After the energy is measured, the electron is in a state described by
electron vary with l?
an eigenfunction of the hamiltonian. Are transitions to the ground state
www.ebook3000.com
Atkins09819.indb 194 9/11/2013 11:17:15 AM

Vectors 195
Mathematical background 4 Vectors

A scalar physical property (such as temperature) in general
Brief illustration MB4.1 Vector orientation
varies through space and is represented by a single value at
each point of space. A vector physical property (such as the The vector v = 2i + 3j – k has magnitude
electric field strength) also varies through space, but in general
has a different direction as well as a different magnitude at each v = {22 + 32 + ( - 1)2 }½ = 141/2 = 3.74
point. Its direction is given by
θ = arccos( - 1/141/2 ) = 105.5° φ = arctan(3 / 2) = 56.3°

MB4.1 Definitions
A vector v has the general form (in three dimensions):
v = v x i + v y j + vz k (MB4.1)
MB4.2 Operations
where i, j, and k are unit vectors, vectors of magnitude 1, point-
ing along the positive directions on the x, y, and z axes and vx, Consider the two vectors
vy , and vz are the components of the vector on each axis (Fig.
MB4.1). The magnitude of the vector is denoted v or |v| and is u = ux i + u y j + uz k v = v x i + v y j + vz k
given by
The operations of addition, subtraction, and multiplication are
v = (vx2 + v 2y + vz2 )1/2 Magnitude (MB4.2) as follows:
1. Addition:
The vector makes an angle θ with the z–axis and an angle φ to
the x-axis in the xy-plane. It follows that v + u = (vx + ux )i + (v y + u y ) j + (vz + uz )k (MB4.4a)
vx = v sin θ cos φ v y = v sin θ sin φ 2. Subtraction:

Orientation (MB4.3a)
vz = v cos θ v - u = (vx - ux )i + (v y - u y ) j + (vz - uz )k (MB4.4b)
and therefore that

Brief illustration MB4.2 Addition and subtraction
θ = arccos(vz / v) φ = arctan(v y / vx ) (MB4.3b)
Consider the vectors u = i – 4j + k (of magnitude 4.24) and
v = –4i + 2j + 3k (of magnitude 5.39). Their sum is
u + v = (1 − 4)i + (−4 + 2) j + (1 + 3)k = −3i − 2 j + 4k
vz
The magnitude of the resultant vector is 291/2 = 5.39. The differ-
k z ence of the two vectors is
v
u − v = (1 + 4)i + (−4 − 2) j + (1 − 3)k = 5i − 6 j − 2k
j θ v
i
The magnitude of this resultant is 8.06. Note that in this case
the difference is longer than either individual vector.
vy
vx φ y
x
3. Multiplication:
Figure MB4.1 The vector v has components vx, vy, and vz on (a) The scalar product, or dot product, of the two vectors
the x, y, and z axes, respectively. It has a magnitude v and makes u and v is
an angle θ with the z-axis and an angle φ to the x-axis in the
xy-plane. u ⋅ v = ux vx + u y v y + uz vz Scalar product (MB4.4c)
Atkins09819.indb 195 9/11/2013 11:17:30 AM

196 Mathematical background 4
and is itself a scalar quantity. We can always choose a new

u×v
coordinate system—we shall write it X, Y, Z—in which
the Z-axis lies parallel to u, so u = uK, where K is the unit
vector parallel to u. It then follows from eqn MB4.4c that
u
u⋅v = uvZ . Then, with vZ = v cos θ, where θ is the angle θ v
between u and v, we find
u ⋅ v = uv cos θ Scalar product (MB4.4d)
(b) The vector product, or cross product, of two vectors is

Figure MB4.2 A depiction of the ‘right-hand rule’. When the
i j k fingers of the right hand rotate u into v, the thumb points in the
direction of u × v.
u × v = ux uy uz
vx vy vz Vector
(MB4.4e)
= (u y vz − uz v y )i − (ux vz − uz vx ) j
product
MB4.3The graphical representation of vector
+ (ux v y − u y vx )k
operations
Consider two vectors v and u making an angle θ (Fig. MB4.3).
(Determinants are discussed in Mathematical background 5.) The first step in the addition of u to v consists of joining the
Once again, choosing the coordinate system so that u = uK tip (the ‘head’) of u to the starting point (the ‘tail’) of v. In the
leads to the simple expression second step, we draw a vector vres, the resultant vector, origi-
nating from the tail of u to the head of v. Reversing the order of
u × v = (uv sin θ )l Vector product (MB4.4f) addition leads to the same result; that is, we obtain the same vres
whether we add u to v or v to u. To calculate the magnitude of
where θ is the angle between the two vectors and l is a unit vec- vres, we note that
tor perpendicular to both u and v, with a direction determined
by the ‘right-hand rule’ as in Fig. MB4.2. A special case is when vres
2
= (u + v ) ⋅ (u + v ) = u ⋅ u + v ⋅ v + 2u ⋅ v = u2 + v 2 + 2uv cos θ
each vector is a unit vector, for then
where θ is the angle between u and v. In terms of the angle
i× j =k j ×k = i k ×i = j (MB4.5) θ ′ = π – θ shown in the figure, and cos (π – θ) = –cosθ, we obtain
the law of cosines:
It is important to note that the order of vector multiplication is
important and that u × v = –v × u. vres
2
= u2 + v 2 − 2uv cos θ ′ Law of cosines (MB4.6)
for the relation between the lengths of the sides of a triangle.

Brief illustration MB4.3 Scalar and vector products Subtraction of v from u amounts to addition of –v to u. It fol-
The scalar and vector products of the two vectors in Brief lows that in the first step of subtraction we draw –v by reversing
illustration MB4.2, u = i – 4j + k (of magnitude 4.24) and
v = –4i + 2j + 3k (of magnitude 5.39) are θ v θ v
u+v
u ⋅ v = {1 × (−4)} + {(−4) × 2} + {1 × 3} = −9
i j k u u u
π−θ
u × v = 1 −4 1
−4 2 3
= {(−4)(3) − (1))(2)}i − {(1)(3) − (1)(−4)} j θ v
+ {(1)(2) − (−4)(−4)}k (a) (b) (c)

= −14i − 7 j − 14k
Figure MB4.3 (a) The vectors v and u make an angle θ. (b) To
The vector product is a vector of magnitude 21.00 pointing in add u to v, we first join the head of u to the tail of v, making
a direction perpendicular to the plane defined by the two indi- sure that the angle θ between the vectors remains unchanged.
vidual vectors. (c) To finish the process, we draw the resultant vector by joining
the tail of u to the head of v.
www.ebook3000.com
Atkins09819.indb 196 9/11/2013 11:17:44 AM

Vectors 197
u u The derivatives of scalar and vector products are obtained using

–v –v the rules of differentiating a product:
θ
u−v du ⋅ v ⎛ du ⎞ ⎛ dv ⎞
=⎜ ⎟ ⋅ v +u ⋅ ⎜ ⎟ (MB4.8a)
dt ⎝ dt ⎠ ⎝ dt ⎠
(a) (b)
du × v ⎛ du ⎞ ⎛ dv ⎞
Figure MB4.4 The graphical method for subtraction of the = ⎜ ⎟ × v +u×⎜ ⎟ (MB4.8b)
dt ⎝ dt ⎠ ⎝ dt ⎠
vector v from the vector u (shown in Fig. MB4.3a) consists
of two steps: (a) reversing the direction of v to form –v, and In the latter, note the importance of preserving the order of
(b) adding –v to u. vectors.
The gradient of a scalar function f(x,y,z), denoted grad f or
the direction of v (Fig. MB4.4). Then, the second step consists ∇f, is
of adding –v to u by using the strategy shown in the figure: we
draw a resultant vector vres originating from the tail of u to the ⎛ ∂f ⎞ ⎛ ∂f ⎞ ⎛ ∂f ⎞
∇f = ⎜ ⎟ i + ⎜ ⎟ j + ⎜ ⎟ k Gradient (MB4.9)
head of –v. ⎝ ∂x ⎠ ⎝ ∂y ⎠ ⎝ ∂z ⎠
Vector multiplication is represented graphically by drawing
a vector (using the right-hand rule) perpendicular to the plane where partial derivatives are mentioned in Mathematical back-
defined by the vectors u and v, as shown in Fig. MB4.5. Its ground 1 and are treated at length in Mathematical background
length is equal to uv sin θ, where θ is the angle between u and v. 8. Note that the gradient of a scalar function is a vector. We can
treat ∇ as a vector operator (in the sense that it operates on a
function and results in a vector), and write
MB4.4 Vector differentiation
The derivative dv/dt, where the components vx, vy, and vz are ∂ ∂ ∂
∇=i + j +k (MB4.10)
themselves functions of t, is ∂x ∂y ∂z
dv ⎛ dv x ⎞ ⎛ dv y ⎞ ⎛ dv z ⎞ The scalar product of ∇ and ∇f, using eqns MB4.9 and MB4.10,
= i+ j+ k Derivative (MB4.7)
dt ⎜⎝ dt ⎟⎠ ⎜⎝ dt ⎟⎠ ⎜⎝ dt ⎟⎠ is
⎛ ∂ ∂ ∂⎞ ⎛ ∂ ∂ ∂⎞
z z ∇ ⋅ ∇f = ⎜ i + j + k ⎟ ⋅ ⎜ i + j + k ⎟ f
⎝ ∂x ∂y ∂z ⎠ ⎝ ∂x ∂y ∂z ⎠ Laplacian
∂f ∂f ∂f
2 2 2
(MB4.11)
v v = + +
u u ∂x 2 ∂y 2 ∂z 2
u×v
θ θ
Equation MB4.11 defines the laplacian (∇2 = ∇ ⋅ ∇) of a function.
uv sin θ
y y
v×u
uv sin θ
x x
(a) (b)
Figure MB4.5 The direction of the cross products of two

vectors u and v with an angle θ between them: (a) u × v and
(b) v × u. Note that the cross product, and the unit vector l of
eqn MB4.4f, are perpendicular to both u and v but the direction
depends on the order in which the product is taken. The
magnitude of the cross product, in either case, is uv sin θ.
Atkins09819.indb 197 9/11/2013 11:17:55 AM

this page left intentionally blank
www.ebook3000.com
FOCUS 6 ON Molecular structure
Topic 22 Topic 23 Topic 24 Topic 25 Topic 26
The Homo- Hetero-

Valence-
principles of nuclear nuclear Polyatomic
bond
molecular diatomic diatomic molecules
theory
orbital molecules molecules
theory
Topic 27
Self-
consistent
fields Topic 28 Topic 29 Topic 30
Semi- Density
Focus 5 Focus 4 Ab initio
empirical functional
methods
Atomic methods theory
Approximation
structure
methods
and spectra
Focus 2 Focus 8 Focus 9 Focus 17

The principles
Molecular Chemical
of quantum Interactions
spectroscopy kinetics
mechanics
Molecular structure lies at the heart of chemistry and is a hugely important aspect of the subject, as
it underlies the discussion of the properties of materials and the reactions they undergo. Primitive
versions of bonding theories emerged at the beginning of the twentieth century when G.N. Lewis
identified the crucial role of the electron pair. That role was clarified by the quantum mechanical
descriptions of bonding that are now pervasive in chemistry. All the descriptions can be regarded as
extensions of the quantum mechanical discussion of Atomic structure and spectra.
One of the earliest applications of The principles of quantum mechanics to the description of the
chemical bond was ‘valence-bond theory’ (Topic 22). It focused on the role of the electron pair and
introduced many widely used concepts into chemistry, such as hybridization.
At about the same time, ‘molecular orbital theory’ was introduced (Topic 23), and has become the
description of choice for quantitative computation. Molecular orbital theory extends the concept
of atomic orbital to wavefunctions that are delocalized over the entire molecule. It introduces the
concepts of bonding and antibonding orbital. We introduce the concepts of molecular orbital theory
with the simplest of all molecules, H2, and then progressively extend the discussion to homonuclear
diatomic molecules (Topic 24), heteronuclear diatomic molecules (Topic 25), and—the ultimate tar-
get of this group of Topics—polyatomic molecules (Topic 26).
All those Topics are essentially qualitative. In the remaining Topics of this group we show how
molecular orbital theory is used in computational chemistry, the computation of electron wave-
functions and their energies. All spring from the numerical procedure called the ‘self-consistent
field’ method (Topic 27). Three versions of this method are commonly encountered. In one, the
Atkins09819.indb 199 9/11/2013 11:17:59 AM

‘semi-empirical method’, the integrals that appear in the calculation are replaced by parameters that
lead to good agreement with certain experimentally determined properties (Topic 28). In another,
the ‘ab initio method’, attempts are made to evaluate the integrals from first principles (Topic 29);
one ab initio method has its foundations in perturbation theory described in Approximation methods.
A third and currently very popular procedure, ‘density functional theory’, takes a different route: it
seeks to calculate the electron density itself rather than the wavefunction (Topic 30).
www.ebook3000.com
Atkins09819.indb 200 9/11/2013 11:17:59 AM

TOPIC 22
Valence-bond theory
Here we summarize essential topics of valence-bond theory
Contents (VB theory) that should be familiar from introductory chem-
22.1 Diatomic molecules 202 istry and set the stage for the development of molecular orbital
(a) The basic formalism 202 theory (Topic 23). However, there is an important preliminary
Brief illustration 22.1: A valence-bond point. All theories of molecular structure make the same sim-
wavefunction 202 plification at the outset. Whereas the Schrödinger equation
(b) Resonance 203 for a hydrogen atom can be solved exactly, an exact solution
Brief illustration 22.2: Resonance hybrids 204 is not possible for any molecule because even the simplest
22.2 Polyatomic molecules 204 molecule consists of three particles (two nuclei and one elec-
Brief illustration 22.3: A polyatomic molecule 204
tron). We therefore adopt the Born–Oppenheimer approxi-
(a) Promotion 205 mation in which it is supposed that the nuclei, being so much
Brief illustration 22.4: Promotion 205
heavier than an electron, move relatively slowly and may be
(b) Hybridization 205 treated as stationary while the electrons move in their field.
Brief illustration 22.5: Hybrid structures 207
That is, we think of the nuclei as fixed at arbitrary locations,
Checklist of concepts 208
and then solve the Schrödinger equation for the wavefunction
Checklist of equations 208 of the electrons alone. The approximation is quite good for
ground-state molecules, for calculations suggest that the nuclei
in H2 move through only about 1 pm while the electron speeds
through 1000 pm.
The Born–Oppenheimer approximation allows us to select
an internuclear separation in a diatomic molecule and then to
➤ Why do you need to know this material? solve the Schrödinger equation for the electrons at that nuclear
Valence-bond theory was the first quantum mechanical separation. Then we choose a different separation and repeat
theory of bonding to be developed. The language it the calculation, and so on. In this way we can explore how the
introduced, which includes concepts such as spin pairing, σ energy of the molecule varies with bond length and obtain a
and π bonds, and hybridization, is widely used throughout molecular potential energy curve (Fig. 22.1). It is called a
chemistry, especially in the description of the properties potential energy curve because the kinetic energy of the sta-
and reactions of organic compounds. tionary nuclei is zero. Once the curve has been calculated or
determined experimentally (by using the spectroscopic tech-
➤ What is the key idea? niques described in Topics 40 − 46), we can identify the equi-
A bond forms when an electron in an atomic orbital on librium bond length, Re, the internuclear separation at the
one atom pairs its spin with that of an electron in an minimum of the curve, and the bond dissociation energy, D0,
atomic orbital on another atom. which is closely related to the depth, De, of the minimum below
the energy of the infinitely widely separated and stationary
➤ What do you need to know already? atoms. When more than one molecular parameter is changed
You need to know about atomic orbitals (Topic 18) and the con- in a polyatomic molecule, such as its various bond lengths and
cepts of normalization (Topic 5) and orthogonality (Topic 7). angles, we obtain a potential energy surface; the overall equi-
This Topic also makes use of the Pauli principle (Topic 19). librium shape of the molecule corresponds to the global mini-
mum of the surface.
Atkins09819.indb 201 9/11/2013 11:17:59 AM

202 6 Molecular structure
Energy
A(1)B(2) A(2)B(1)
Re
0
Internuclear
separation, R A(1)B(2) + A(2)B(1) Enhanced
electron density
–De
Figure 22.2 It is very difficult to represent valence-

Figure 22.1 A molecular potential energy curve. The
bond wavefunctions because they refer to two electrons
equilibrium bond length corresponds to the energy minimum.
simultaneously. However, this illustration is an attempt. The
atomic orbital for electron 1 is represented by the purple
shading, and that of electron 2 is represented by the green
22.1 Diatomic molecules shading. The left illustration represents A(1)B(2), and the right
illustration represents the contribution A(2)B(1). When the
We begin the account of VB theory by considering the simplest two contributions are superimposed, there is interference
possible chemical bond, the one in molecular hydrogen, H2. between the purple contributions and between the green
contributions, resulting in an enhanced (two-electron) density
in the internuclear region.
(a) The basic formalism
The spatial wavefunction for an electron on each of two widely Brief illustration 22.1 A valence-bond wavefunction
separated H atoms is
The wavefunction in eqn 22.2 might look abstract, but in fact
ψ = χ H1s (r1 ) χ H1s (r2 )
A B
(22.1) it can be expressed in terms of simple exponential functions.
Thus, if we use the wavefunction for an H1s orbital (Z = 1)
if electron 1 is on atom A and electron 2 is on atom B; in this given in Topic 18, then, with the radii measured from their
Topic, and as is common in the chemical literature, we use χ respective nuclei,
(chi) to denote atomic orbitals. For simplicity, we shall write A
(1)
B
(2)
A
(2)

this wavefunction as ψ = A(1)B(2). When the atoms are close, 1 1 1
it is not possible to know whether it is electron 1 or elec- ψ= e − rA 1 /a0
× e − rB 2 /a0
+ e − rA 2 /a0
(πa03 )1/2 (πa03 )1/2 (πa03 )1/2
tron 2 that is on A. An equally valid description is therefore B (1)

ψ = A(2)B(1), in which electron 2 is on A and electron 1 is on B. 1
× e − rB1 /a0
When two outcomes are equally probable, quantum mechanics (πa03 )1/2
instructs us to describe the true state of the system as a super- 1
position of the wavefunctions for each possibility (Topic 7), = 3 {e −(rA1 +rB 2 )/a0 + e −(rA 2 +rB1 )/a0 }
πa0
so a better description of the molecule than either wavefunc-
tion alone is one of the (unnormalized) linear combinations
Self-test 22.1 Express this wavefunction in terms of the
ψ = A(1)B(2) ± A(2)B(1). The combination with lower energy is
Cartesian coordinates of each electron given that the internu-
the one with a + sign, so the valence-bond wavefunction of the
clear separation (along the z-axis) is R.
electrons in an H2 molecule is
Answer: rAi = (xi2 + yi2 + zi2 )1/2 , rBi = (xi2 + yi2 + (zi − R)2 )1/2
ψ = A(1)B(2) + A(2)B(1) A valence-bond wavefunction (22.2)

The electron distribution described by the wavefunction in
The reason why this linear combination has a lower energy eqn 22.2 is called a σ bond. A σ bond has cylindrical symme-
than either the separate atoms or the linear combination try around the internuclear axis, and is so called because, when
with a negative sign can be traced to the constructive inter- viewed along the internuclear axis, it resembles a pair of elec-
ference between the wave patterns represented by the terms trons in an s orbital (and σ is the Greek equivalent of s).
A(1)B(2) and A(2)B(1), and the resulting enhancement of A chemist’s picture of a covalent bond is one in which the
the probability density of the electrons in the internuclear spins of two electrons pair as the atomic orbitals overlap. The ori-
region (Fig. 22.2). gin of the role of spin, as we show in the following Justification,
www.ebook3000.com
Atkins09819.indb 202 9/11/2013 11:18:07 AM

22 Valence-bond theory 203
is that the wavefunction in eqn 22.2 can be formed only by a pair

of spin-paired electrons. Spin pairing is not an end in itself: it is a
means of achieving a wavefunction and the probability distribu-
tion it implies that corresponds to a low energy.
Justification 22.1
Electron pairing in VB theory
The Pauli principle requires the overall wavefunction of two
electrons, the wavefunction including spin, to change sign
when the labels of the electrons are interchanged (Topic 19). Figure 22.3 The orbital overlap and spin pairing between
The overall VB wavefunction for two electrons is electrons in two collinear p orbitals that results in the formation
of a σ bond.
ψ (1, 2) = {A(1)B(2) + A(2)B(1)}σ (1, 2)
where σ represents the spin component of the wavefunction.

When the labels 1 and 2 are interchanged, this wavefunction +
becomes Nodal plane
+
Internuclear axis
ψ (2,1) = {A(2)B(1) + A(1)B(2)}σ (2,1)
= {A(1)B(2) + A(2)B(1)}σ (2,1)
–
The Pauli principle requires that ψ(2,1) = −ψ(1,2), which is sat- –
isfied only if σ(2,1) = −σ(1,2). The combination of two spins
that has this property is
Figure 22.4 A π bond results from orbital overlap and
σ − (1, 2) = (1/ 21/2 ){α (1)β (2) − β (1)α (2)} spin pairing between electrons in p orbitals with their axes
perpendicular to the internuclear axis. The bond has two lobes
which corresponds to paired electron spins (Topic 19). of electron density separated by a nodal plane.
Therefore, we conclude that the state of lower energy (and
hence the formation of a chemical bond) is achieved if the
+
electron spins are paired.
– –
+ + +
The VB description of H2 can be applied to other homonu- –
clear diatomic molecules. For N2, for instance, we consider
+
the valence electron configuration of each atom, which is –
2s2 2p1x 2p1y 2p1z . It is conventional to take the z-axis to be the –
internuclear axis, so we can imagine each atom as having a 2pz –
orbital pointing towards a 2pz orbital on the other atom (Fig.
22.3), with the 2px and 2py orbitals perpendicular to the axis. Figure 22.5 The structure of bonds in a nitrogen molecule,
A σ bond is then formed by spin pairing between the two elec- with one σ bond and two π bonds.
trons in the two 2pz orbitals. Its spatial wavefunction is given by
eqn 22.2, but now A and B stand for the two 2pz orbitals. in N2 is therefore a σ bond plus two π bonds (Fig. 22.5), which
The remaining N2p orbitals cannot merge to give σ bonds is consistent with the Lewis structure :N ≡ N: for nitrogen.
as they do not have cylindrical symmetry around the inter-
nuclear axis. Instead, they merge to form two π bonds. A π
bond arises from the spin pairing of electrons in two p orbitals
(b) Resonance
that approach side-by-side (Fig. 22.4). It is so called because, Another term introduced into chemistry by VB theory is reso-
viewed along the internuclear axis, a π bond resembles a pair nance, the superposition of the wavefunctions representing
of electrons in a p orbital (and π is the Greek equivalent of p). different electron distributions in the same nuclear framework.
There are two π bonds in N2, one formed by spin pairing To understand what this means, consider the VB description
in two neighbouring 2px orbitals and the other by spin pairing of a purely covalently bonded HCl molecule, which could be
in two neighbouring 2py orbitals. The overall bonding pattern written as ψ = A(1)B(2) + A(2)B(1), with A now a H1s orbital
Atkins09819.indb 203 9/11/2013 11:18:14 AM

and B a Cl2p orbital. However, there is something wrong

its covalent and ionic forms are in the ratio 100:1 (because
with this description: it allows electron 1 to be on the H atom 0.12 = 0.01).
when electron 2 is on the Cl atom, and vice versa, but it does
not allow for the possibility that both electrons are on the Cl Self-test 22.2 If a normalized wavefunction has the form ψ =
atom (ψ = B(1)B(2), representing H+Cl−) or even on the H 0.889ψ covalent + 0.458ψ ionic, what is the percentage probability
atom (ψ = A(1)A(2), representing the much less likely H−Cl+). of finding both electrons of the bond on one atom?
A better description of the wavefunction for the molecule is as Answer: 21.0 per cent
a superposition of the covalent and ionic descriptions, and we
write (with a slightly simplified notation, and ignoring the less
likely H− Cl+ possibility) ψ HCl =ψ H−Cl + λψ H Cl with λ (lambda)
+ − 22.2 Polyatomic molecules
some numerical coefficient. In general, we write
Each σ bond in a polyatomic molecule is formed by the spin
ψ =ψ covalent + λψ ionic (22.3) pairing of electrons in atomic orbitals with cylindrical symme-
try around the relevant internuclear axis. Likewise, π bonds are
where ψcovalent is the wavefunction for the purely covalent form formed by pairing electrons that occupy atomic orbitals of the
of the bond and ψionic is the wavefunction for the ionic form of appropriate symmetry.
the bond. The approach summarized by eqn 22.3 is an example
of resonance. In this case, where one structure is pure covalent
and the other pure ionic, it is called ionic–covalent resonance. Brief illustration 22.3 A polyatomic molecule
The interpretation of the wavefunction, which is called a reso-
The VB description of H2O will make this approach clear. The
nance hybrid, is that if we were to inspect the molecule, then valence-electron configuration of an O atom is 2s2 2p2x 2p1y 2p1z .
the probability that it would be found with an ionic structure is The two unpaired electrons in the O2p orbitals can each pair
proportional to λ2. If λ2 is very small, the covalent description with an electron in an H1s orbital, and each combination
is dominant. If λ2 is very large, the ionic description is domi- results in the formation of a σ bond (each bond has cylindri-
nant. Resonance is not a flickering between the contributing cal symmetry about the respective O eH internuclear axis).
states: it is a blending of their characteristics, much as a mule is Because the 2py and 2pz orbitals lie at 90° to each other, the two
a blend of a horse and a donkey. It is only a mathematical device σ bonds also lie at 90° to each other (Fig. 22.6). We predict,
for achieving a closer approximation to the true wavefunction therefore, that H2O should be an angular molecule, which it is.
of the molecule than that represented by any single contribut- However, the theory predicts a bond angle of 90°, whereas the
ing structure alone. actual bond angle is 104.5°.
A systematic way of calculating the value of λ is provided by
the variation principle which is proved in Topic 25:
H1s
If an arbitrary wavefunction is used to calculate
Variation
the energy, then the value calculated is never less principle H1s
than the true energy.
O2pz
The arbitrary wavefunction is called the trial wavefunction.
The principle implies that, if we vary the parameter λ in the
trial wavefunction until the lowest energy is achieved (by eval- O2py
uating the expectation value of the hamiltonian for the wave-
function), then that value of λ will be the best and through λ2 Figure 22.6 In a primitive view of the structure of an H2O
represents the appropriate contribution of the ionic wavefunc- molecule, each bond is formed by the overlap and spin
tion to the resonance hybrid. pairing of an H1s electron and an O2p electron.
Self-test 22.3 Use VB theory to suggest a shape for the ammo-

Brief illustration 22.2 Resonance hybrids nia molecule, NH3.
Consider a bond described by eqn 22.3. We might find that the Answer: Trigonal pyramidal with HNH bond angle 90°;
lowest energy is reached when λ = 0.1, so the best description experimental: 107°
of the bond in the molecule is a resonance structure described
by the wavefunction ψ = ψ covalent + 0.1ψ ionic . This wavefunc-
tion implies that the probabilities of finding the molecule in Resonance plays an important role in the valence-bond
description of polyatomic molecules. One of the most famous
www.ebook3000.com
Atkins09819.indb 204 9/11/2013 11:18:17 AM

examples of resonance is in the VB description of benzene, not a ‘real’ process in which an atom somehow becomes excited
where the wavefunction of the molecule is written as a super- and then forms bonds: it is a notional contribution to the over-
position of the wavefunctions of the two covalent Kekulé all energy change that occurs when bonds form.
structures:
ψ =ψ ( ) +ψ ( ) (22.4) Brief illustration 22.4 Promotion

Sulfur can form six bonds (an ‘expanded octet’), as in the mol-
The two contributing structures have identical energies, so
ecule SF6. Because the ground-state electron configuration of
they contribute equally to the superposition. The effect of
sulfur is [Ne]3s23p 4 , this bonding pattern requires the pro-
resonance (which is represented by a double-headed arrow,
motion of a 3s electron and a 3p electron to two different 3d
) in this case is to distribute double-bond charac-
orbitals, which are nearby in energy, to produce the notional
ter around the ring and to make the lengths and strengths configuration [Ne]3s13p33d 2 with all six of the valence elec-
of all the carbon–carbon bonds identical. The wavefunc- trons in different orbitals and capable of bond formation with
tion is improved by allowing resonance because it allows for six electrons provided by six F atoms.
a more accurate description of the location of the electrons,
and in particular the distribution can adjust into a state of Self-test 22.4 Account for the ability of phosphorus to form
lower energy. This lowering is called the resonance stabi- five bonds, as in PF5.
lization of the molecule and, in the context of VB theory, is Answer: Promotion of a 3s electron from [Ne]3s23p3 to
largely responsible for the unusual stability of aromatic rings. [Ne]3s13p33d1
Resonance always lowers the energy, and the lowering is

greatest when the contributing structures have similar ener-
gies. The wavefunction of benzene is improved still further,
and the calculated energy of the molecule is lowered further
still, if we allow ionic–covalent resonance too, by allowing a
(b) Hybridization
+
–
small admixture of structures such as . The description of the bonding in CH4 (and other alkanes)
is still incomplete because it implies the presence of three σ
bonds of one type (formed from H1s and C2p orbitals) and a
fourth σ bond of a distinctly different character (formed from
H1s and C2s). This problem is overcome by realizing that the
(a) Promotion electron density distribution in the promoted atom is equiva-
Another deficiency of this initial formulation of VB theory is its lent to the electron density in which each electron occupies a
inability to account for carbon’s tetravalence (its ability to form hybrid orbital formed by interference between the C2s and
four bonds). The ground-state configuration of C is 2s2 2p1x 2p1y , C2p orbitals of the same atom. The origin of the hybridiza-
which suggests that a carbon atom should be capable of form- tion can be appreciated by thinking of the four atomic orbitals
ing only two bonds, not four. centred on a nucleus as waves that interfere destructively and
This deficiency is overcome by allowing for promotion, the constructively in different regions, and give rise to four new
excitation of an electron to an orbital of higher energy. In car- shapes.
bon, for example, the promotion of a 2s electron to a 2p orbital As we show in the following Justification, the specific linear
can be thought of as leading to the configuration 2s1 2p1x 2p1y 2p1z , combinations that give rise to four equivalent hybrid orbitals
with four unpaired electrons in separate orbitals. These elec- are
trons may pair with four electrons in orbitals provided by four
other atoms (such as four H1s orbitals if the molecule is CH4), h1 = s + p x + p y + pz h2 = s − p x − p y + pz
and hence form four σ bonds. Although energy was required sp3 hybrid
(22.5)
to promote the electron, it is more than recovered by the pro- h3 = s − p x + p y − pz h4 = s + p x − p y − pz orbitals
moted atom’s ability to form four bonds in place of the two

bonds of the unpromoted atom. As a result of the interference between the component orbit-
Promotion, and the formation of four bonds, is a character- als, each hybrid orbital consists of a large lobe pointing in the
istic feature of carbon because the promotion energy is quite direction of one corner of a regular tetrahedron (Fig. 22.7). The
small: the promoted electron leaves a doubly occupied 2s angle between the axes of the hybrid orbitals is the tetrahe-
orbital and enters a vacant 2p orbital, hence significantly reliev- dral angle, arccos( − 13 ) = 109.47°. Because each hybrid is built
ing the electron–electron repulsion it experiences in the for- from one s orbital and three p orbitals, it is called an sp3 hybrid
mer. However, it is important to remember that promotion is orbital.
Atkins09819.indb 205 9/11/2013 11:18:24 AM

109.47°
Figure 22.7 An sp3 hybrid orbital formed from the

superposition of s and p orbitals on the same atom. There Figure 22.8 Each sp3 hybrid orbital forms a σ bond by overlap
are four such hybrids: each one points towards the corner of with an H1s orbital located at the corner of the tetrahedron.
a regular tetrahedron. The overall electron density remains This model accounts for the equivalence of the four bonds
spherically symmetrical. in CH4.
Justification 22.2
wavefunction for the bond formed by the hybrid orbital h1 and
Determining the form of tetrahedral the 1sA orbital (with wavefunction that we shall denote A) is
hybrid orbitals
We begin by supposing that each hybrid can be written in ψ = h1 (1)A(2) + h1 (2)A(1) (22.6)
the form h = as + bx px + by py + bz pz . The hybrid h1 that points
to the (1,1,1) corner of a cube must have equal contributions As for H2, to achieve this wavefunction, the two electrons it
from all three p orbitals, so we can set the three b coefficients describes must be paired. Because each sp3 hybrid orbital has
equal to each other and write h1 = as + b(px + py + pz). The other the same composition, all four σ bonds are identical apart from
three hybrids have the same composition (they are equivalent, their orientation in space.
apart from their direction in space), but are orthogonal to h1. A hybrid orbital has enhanced amplitude in the internu-
This orthogonality is achieved by choosing different signs for clear region, which arises from the constructive interference
the p orbitals but the same overall composition. For instance, between the s orbital and the positive lobes of the p orbitals.
we might choose h 2 = as + b(−px − p y + pz), in which case the As a result, the bond strength is greater than for a bond formed
orthogonality condition is from an s or p orbital alone. This increased bond strength is
another factor that helps to repay the promotion energy.
∫h h dτ = ∫{as + b(p + p
1 2 x y + pz )}{as + b(− p x − p y + pz )}dτ
Hybridization is used to describe the structure of an ethene

1
1
0
molecule, H2C=CH2, and the torsional rigidity of double bonds.
=a 2
∫ s dτ − b ∫
2 2
p2x dτ ∫
− − ab sp x dτ − An ethene molecule is planar, with HCH and HCC bond angles
close to 120°. To reproduce the σ bonding structure, each

0

C atom is regarded as promoted to a 2s12p3 configuration.
∫
− b 2 p x p y dτ + = a 2 − b 2 − b 2 + b 2 = a 2 − b 2 = 0 However, instead of using all four orbitals to form hybrids, we
form sp2 hybrid orbitals:
We conclude that a solution is a = b (the alternative solu-
tion, a = −b, simply corresponds to choosing different abso- h1 = s + 21/2 p y
lute phases for the p orbitals) and the two hybrid orbitals
h2 = s + ( 23 )1/2 p x − ( 12 )1/2 p y sp2 hybrid orbitals (22.7)
are the h1 and h 2 in eqn 22.5. A similar argument but with
h 3 = as + b(−px + py − pz) or h4 = as + b(px − py − pz) leads to the h3 = s − ( ) p x − ( ) p y
3 1/2
2
1 1/2
2
other two hybrids in eqn 22.5.

These hybrids lie in a plane and point towards the corners of an
It is now easy to see how the valence-bond description of the equilateral triangle at 120° to each other (Fig. 22.9 and Problem
CH4 molecule leads to a tetrahedral molecule containing four 22.2). The third 2p orbital (2pz) is not included in the hybridiza-
equivalent CeH bonds. Each hybrid orbital of the promoted C tion; its axis is perpendicular to the plane in which the hybrids
atom contains a single unpaired electron; an H1s electron can lie. The different signs of the coefficients, as well as ensuring
pair with each one, giving rise to a σ bond pointing in a tetra- that the hybrids are mutually orthogonal, also ensure that con-
hedral direction (Fig. 22.8). For example, the (unnormalized) structive interference takes place in different regions of space,
www.ebook3000.com
31_Atkins_Ch22.indd 206 9/11/2013 1:38:18 PM

120°
Figure 22.11 A representation of the structure of the triple
(a) (b) bond in ethyne; only the π bonds are shown explicitly.
Figure 22.9 (a) An s orbital and two p orbitals can be
hybridized to form three equivalent orbitals that point towards Table 22.1 Some hybridization schemes
the corners of an equilateral triangle. (b) The remaining,
unhybridized p orbital is perpendicular to the plane. Coordination number Arrangement Composition
2 Linear sp, pd, sd

so giving the patterns in the illustration. The sp2-hybridized Angular sd
C atoms each form three σ bonds by spin pairing with either 3 Trigonal planar sp2, p2d
the h1 hybrid of the other C atom or with H1s orbitals. The σ Trigonal pyramidal pd2
framework therefore consists of CeH and CeC σ bonds at 4 Tetrahedral sp3, sd3
120° to each other. When the two CH2 groups lie in the same Square planar p2d2, sp2d
plane, the two electrons in the unhybridized p orbitals can pair 5 Trigonal bipyramidal sp3d, spd3
Pentagonal planar p2d3
and form a π bond (Fig. 22.10). The formation of this π bond
6 Octahedral sp3d2
locks the framework into the planar arrangement, for any rota-
tion of one CH2 group relative to the other leads to a weakening
of the π bond (and consequently an increase in energy of the Other hybridization schemes, particularly those involving d
molecule). orbitals, are often invoked in elementary descriptions of molec-
A similar description applies to ethyne, HC ≡ CH, a linear ular structure to be consistent with other molecular geometries
molecule. Now the C atoms are sp hybridized, and the σ bonds (Table 22.1). The hybridization of N atomic orbitals always
are formed using hybrid atomic orbitals of the form results in the formation of N hybrid orbitals, which may either
form bonds or contain lone pairs of electrons.
h1 = s + pz h2 = s − pz sp hybrid orbitals (22.8)
Brief illustration 22.5 Hybrid structures

These two hybrids lie along the internuclear axis. The elec-
trons in them pair either with an electron in the corresponding For example, sp3d 2 hybridization results in six equivalent
hybrid orbital on the other C atom or with an electron in one of hybrid orbitals pointing towards the corners of a regular octa-
the H1s orbitals. Electrons in the two remaining p orbitals on hedron; it is sometimes invoked to account for the structure
each atom, which are perpendicular to the molecular axis, pair of octahedral molecules, such as SF6 (recall the promotion of
to form two perpendicular π bonds (Fig. 22.11). sulfur’s electrons in Brief illustration 22.4). Hybrid orbitals
do not always form bonds: they may also contain lone pairs
of electrons. For example, in the hydrogen peroxide molecule,
H 2O2 , each O atom can be regarded as sp3 hybridized. Two
of the hybrid orbitals form bonds, one O e O bond and one
O eH bond at approximately 109° (the experimental value is
much less, at 94.8°). The remaining two hybrids on each atom
accommodate lone pairs of electrons. Rotation around the
O e O bond is possible, so the molecule is conformationally
mobile.
Self-test 22.5 Account for the structure of methylamine,
CH3NH2.
Figure 22.10 A representation of the structure of the double Answer: C, N both sp3 hybridized; a lone pair on N
bond in ethene; only the π bond is shown explicitly.
31_Atkins_Ch22.indd 207 9/11/2013 1:38:24 PM

☐ 1. The Born–Oppenheimer approximation treats the ☐ 6. Resonance refers to the superposition of the wavefunc-
nuclei as stationary while the electrons move in their tions representing different electron distributions in the
field. same nuclear framework. The wavefunction resulting
☐ 2. A molecular potential energy curve depicts the varia- from the superposition is called a resonance hybrid.
tion of the energy of the molecule as a function of bond ☐ 7. To accommodate the shapes of polyatomic molecules,
length. VB theory introduces the concepts of promotion and
☐ 3. The equilibrium bond length is the internuclear sepa- hybridization.
ration at the minimum of the curve. ☐ 8. A σ bond has cylindrical symmetry around the inter-
☐ 4. The bond dissociation energy is the minimum energy nuclear axis.
needed to separate the two atoms of a molecule. ☐ 9. A π bond has symmetry like that of a p orbital perpen-
☐ 5. A bond forms when an electron in an atomic orbital on dicular to the internuclear axis.
one atom pairs its spin with that of an electron in an
atomic orbital on another atom.
Valence-bond wavefunction ψ = A(1)B(2) + A(2)B(1) A,B are atomic orbitals 22.2
∑c χ
Hybridization All atomic orbitals on the same atom; specific forms 22.5 (sp3)
h= i i in the text 22.7 (sp2)
i 22.8 (sp)
www.ebook3000.com
Atkins09819.indb 208 9/11/2013 11:18:36 AM

TOPIC 23
The principles of molecular

orbital theory
In molecular orbital theory (MO theory), electrons do not
Contents
belong to particular bonds but spread throughout the entire
23.1 Linear combinations of atomic orbitals 209 molecule. This theory has been more fully developed than VB
(a) The construction of linear combinations 210 theory (Topic 22) and provides the language that is widely used
Example 23.1: Normalizing a molecular orbital 210 in modern discussions of bonding. To introduce it, we fol-
Brief illustration 23.1: A molecular orbital 210 low the same strategy as in Topic 19, where the one-electron
(b) Bonding orbitals 211 H atom was taken as the fundamental species for discuss-
Brief illustration 23.2: Molecular integrals 212 ing atomic structure and then developed into a description of
(c) Antibonding orbitals 213 many-electron atoms. In this Topic we use the simplest molec-
Brief illustration 23.3: Antibonding energies 214 ular species of all, the hydrogen molecule-ion, H2+ , to introduce
23.2 Orbital notation 214 the essential features of bonding and then use it to describe the
Brief illustration 23.4: Inversion symmetry 214 structures of more complex systems.
Checklist of concepts 214
Checklist of equations 215
Linear combinations of
23.1
atomic orbitals
➤ Why do you need to know this material? The hamiltonian for the single electron in H2+ is
Molecular orbital theory is the basis of almost all
descriptions of chemical bonding, including that in
2 2 e2 ⎛ 1 1 1 ⎞
individual molecules and of solids. It is the foundation of H = − ∇ +V V =− + −
4 πε 0 ⎜⎝ rA1 rB1 R ⎟⎠
(23.1)
most computational techniques for the prediction and 2me 1
analysis of the properties of molecules.
where rA1 and rB1 are the distances of the electron from the two
➤ What is the key idea? nuclei A and B (1) and R is the distance between the two nuclei.
Molecular orbitals are wavefunctions that spread over all In the expression for V, the e
the atoms in a molecule. first two terms in parentheses rA1
are the attractive contribution 1 rB1
➤ What do you need to know already? from the interaction between A R
You need to be familiar with the shapes of atomic orbitals the electron and the nuclei; the
B
(Topic 18) and how an energy and probability density remaining term is the repulsive
are calculated from a wavefunction (Topics 5 and 7). The interaction between the nuclei.
entire discussion is within the framework of the Born– The collection of fundamental constants e2/4πε0 occurs widely
Oppenheimer approximation (Topic 22). throughout these Topics, and we shall denote it j0. The one-
electron wavefunctions obtained by solving the Schrödinger
Atkins09819.indb 209 9/11/2013 11:18:39 AM

equation Hψ = Εψ with this hamiltonian and its analogues for

other molecules are called molecular orbitals (MOs). A molec-
ular orbital is like an atomic orbital, but spreads throughout the
molecule. It gives, through the value of |ψ|2, the probability dis-
tribution of the electron in the molecule.
(a) The construction of linear combinations

The Schrödinger equation can be solved analytically for H2+ (a) (b)
(within the Born–Oppenheimer approximation, Topic 22),
Figure 23.1 (a) The amplitude of the bonding molecular orbital
but the wavefunctions are very complicated functions; more- in a hydrogen molecule-ion in a plane containing the two
over, the solution cannot be extended to polyatomic systems. nuclei and (b) a contour representation of the amplitude.
Therefore, we adopt a simpler procedure that, while more
approximate, can be extended readily to other molecules.
If an electron can be found in an atomic orbital belonging to
In H2+ , S ≈ 0.59, so N = 0.56.
atom A and also in an atomic orbital belonging to atom B, then
the overall wavefunction is a superposition of the two atomic Self-test 23.1 Normalize the orbital ψ− in eqn 23.2.
orbitals: Answer: N = 1/{2(1 − S)}1/2; if S ≈ 0.59, then N = 1.10
ψ ± = N ( A ± B) Linear combination of atomic orbitals (23.2)
where, for H2+ , A denotes an H1s atomic orbital on atom A, Figure 23.1 shows the contours of constant amplitude for
which we denote (as in Topic 22) χ H1s , B likewise denotes
A the molecular orbital ψ+ in eqn 23.2. Plots like these are readily
χ H1s , and N is a normalization factor. The technical term for
B obtained using commercially available software. The calcula-
the superposition in eqn 23.2 is a linear combination of atomic tion is quite straightforward, because all we need do is feed in
orbitals (LCAO). An approximate molecular orbital formed the mathematical forms of the two atomic orbitals and then let
from a linear combination of atomic orbitals is called an the program do the rest.
LCAO-MO. A molecular orbital that has cylindrical symmetry
around the internuclear axis, such as the one we are discuss-
ing, is called a σ orbital because it resembles an s orbital when Brief illustration 23.1 A molecular orbital
viewed along the axis and, more precisely, because it has zero
We can use the following two H1s orbitals
orbital angular momentum around the internuclear axis.
1 1
A= e − rA1 /a0 B= e − rB1 /a0
Example 23.1 (πa03 )1/2 (πa03 )1/2
Normalizing a molecular orbital
Normalize the molecular orbital ψ + in eqn 23.2. Note that rA1 and rB1 are not independent, but when expressed
Method We need to find the factor N such that ∫ψ*ψdτ = 1. To in Cartesian coordinates based on atom A (2) are related by
proceed, substitute the LCAO into this integral, and make use rA1 = {x 2 + y2 + z2}1/2 and rB1 = {x 2 + y2 + (z − R)2}1/2, where R is the
of the fact that the atomic orbitals are individually normalized. bond length.
Answer Substitution of the wavefunction gives e

rA1
y
rB1
⎧
1 1
S⎫
2⎪ ⎪ A
∫ *
⎪
2
∫ 2
∫ ⎪
∫
ψ ψ dτ = N ⎨ A dτ + B dτ + 2 ABdτ ⎬ = 2(1 + S)N 2 2
z x
R
B
⎩ ⎭ R–z
where S = ∫ABdτ and has a value that depends on the nuclear

separation (this ‘overlap integral’ will play a significant role
later). For the integral to be equal to 1, we require
1 The resulting surfaces of constant amplitude are shown in
N=
{2(1+ S)}1/2 Fig. 23.2.
www.ebook3000.com
Atkins09819.indb 210 9/11/2013 11:18:48 AM

23 The principles of molecular orbital theory 211
Figure 23.4 The electron density calculated by forming the

square of the wavefunction used to construct Fig. 23.1. Note
Figure 23.2 Surfaces of constant amplitude of the the accumulation of electron density in the internuclear
wavefunction ψ+ of the hydrogen molecule-ion. region.
Self-test 23.2 Repeat the analysis for ψ−.
Answer: The last contribution, the overlap density, is crucial, because
it represents an enhancement of the probability of finding the
electron in the internuclear region. The enhancement can be
traced to the constructive interference of the two atomic orbit-
als: each has a positive amplitude in the internuclear region,
so the total amplitude is greater there than if the electron were
confined to a single atomic orbital.
We shall frequently make use of the observation that bonds
form when electrons accumulate in regions where atomic orbitals
overlap and interfere constructively. The conventional explana-
tion of this observation is based on the notion that accumula-
Figure 23.3 Surfaces of constant amplitude of the tion of electron density between the nuclei puts the electron in
wavefunction ψ− of the hydrogen molecule-ion. a position where it interacts strongly with both nuclei. Hence,
the energy of the molecule is lower than that of the separate
atoms, where each electron can interact strongly with only
one nucleus. This conventional explanation, however, has
(b) Bonding orbitals
been called into question, because shifting an electron away
According to the Born interpretation, the probability density of from a nucleus into the internuclear region raises its poten-
the electron at each point in H2+ is proportional to the square tial energy. The modern (and still controversial) explanation
modulus of its wavefunction at that point. The probability den- does not emerge from the simple LCAO treatment given here.
sity corresponding to the (real) wavefunction ψ+ in eqn 23.2 is It seems that, at the same time as the electron shifts into the
internuclear region, the atomic orbitals shrink. This orbital
shrinkage improves the electron–nucleus attraction more than
ψ +2 = N 2 ( A2 + B 2 + 2AB) Bonding probability density (23.3)
it is decreased by the migration to the internuclear region, so
there is a net lowering of potential energy. The kinetic energy
This probability density is plotted in Fig. 23.4. An important of the electron is also modified because the curvature of the
feature becomes apparent when we examine the internuclear wavefunction is changed, but the change in kinetic energy is
region, where both atomic orbitals have similar amplitudes. dominated by the change in potential energy. Throughout the
According to eqn 23.3, the total probability density is propor- following discussion we ascribe the strength of chemical bonds
tional to the sum of to the accumulation of electron density in the internuclear
region. We leave open the question whether in molecules more
r A2, the probability density if the electron were
complicated than H2+ the true source of energy lowering is that
interpretation
confined to the atomic orbital A;

accumulation itself or some indirect but related effect.
r B2, the probability density if the electron were The σ orbital we have described is an example of a bond-
Physical
confined to the atomic orbital B; ing orbital, an orbital which, if occupied, helps to bind two
r 2AB, an extra contribution to the density from atoms together by lowering its energy below that of the separate
both atomic orbitals. atoms. Specifically, we label it 1σ as it is the σ orbital of lowest
Atkins09819.indb 211 9/11/2013 11:18:54 AM

energy. An electron that occupies a σ orbital is called a σ elec- Brief illustration 23.2 Molecular integrals
tron, and if that is the only electron present in the molecule (as
in the ground state of H2+ ), then we report the configuration of It turns out (see below) that the minimum value of E1σ occurs
the molecule as 1σ1. at R = 2.45a0. At this separation
The energy E1σ of the 1σ orbital is (see Problem 23.3)
⎧ 2.452 ⎫ −2.45
S = ⎨1+ 2.45 + ⎬e = 0.47
⎩ 3 ⎭
j0 j + k
E1σ = EH1s + − (23.4) j /a
j = 0 0 {1− 3.45e −4.90 } = 0.40 j0 /a0
Energy of bonding orbital
R 1+ S
2.45
j
where EH1s is the energy of a H1s orbital, j0/R is the potential k = 0 (1 + 2.45) e −2.45 = 0.30 j0 /a0
energy of repulsion between the two nuclei (remember that j0 is a0
shorthand for e2/4πε0), and
To express j0/a0 = e2/4πε0a0 in electronvolts, divide it by e, and
⎪⎧ R 1 ⎛ R ⎞ ⎫⎪
2
then find
∫
S = AB dτ = ⎨1+ + ⎜ ⎟ ⎬ e − R/a
⎩⎪
a0 3 ⎝ a0 ⎠ ⎪
⎭
0
(23.5a)
j0 e e π mee 2 mee 3
= = × = 2 2 = 27.211…V
A 2
j ⎧ ⎛ R⎞ ⎫ ea0 4 π ε 0a0 4 π ε 0 ε 0h2 4ε 0 h
j = j0
rB ∫
dτ = 0 ⎨1− ⎜ 1+ ⎟ e −2 R/a ⎬
R ⎩ ⎝ a0 ⎠ ⎭
0 (23.5b)
(The value should be recognized as 2hcR ∞ /e, Topic 17.)

AB j0 ⎛ R ⎞ − R/a Therefore, j = 11 eV and k = 8.2 eV.
k = j0 ∫r B
dτ = 1+
a0 ⎜⎝ a0 ⎟⎠
e 0
(23.5c)
Self-test 23.3 Evaluate the integrals when the internuclear
separation is twice its value at the minimum.
The integrals are plotted in Fig. 23.5. We can interpret them as Answer: 0.10, 5.5 eV, 1.2 eV
follows:
r All three integrals are positive and decline
towards zero at large internuclear separations
interpretation
Figure 23.6 shows a plot of E1σ against R relative to the energy

(S and k on account of the exponential term, j on
of the separated atoms. The energy of the 1σ orbital decreases as
account of the factor 1/R). The integral S is
Physical
the internuclear separation decreases from large values because

discussed in more detail in Topic 24.
electron density accumulates in the internuclear region as the
r The integral j is a measure of the interaction constructive interference between the atomic orbitals increases
between a nucleus and electron density centred on (Fig. 23.7). However, at small separations there is too little space
the other nucleus.
r The integral k is a measure of the interaction between a
0.2
nucleus and the excess electron density in the
internuclear region arising from overlap. 0.15
Energy, (E± – EH1s)/2hcRH
~
2σ (1σu)
0.1
1 1
0.05
0.8 0.8 1σ (1σg)
0
j/j0 and k/j0
0.6 0.6
S
–0.05
0.4 0.4
j –0.1
0 2 4 6 8 10
0.2 0.2 Internuclear distance, R/a0
k
0 0 Figure 23.6 The calculated molecular potential energy curves
0 2 4 6 8 10 0 2 4 6 8 10
(a) R/a0 (b) R/a0 for a hydrogen molecule-ion showing the variation of the
energies of the bonding and antibonding orbitals as the bond
Figure 23.5 The integrals (a) S and (b) j and k calculated for H2+ length is changed. The alternative notation of the orbitals is
as a function of internuclear distance. explained later.
www.ebook3000.com
Atkins09819.indb 212 9/11/2013 11:19:09 AM

Region of
constructive
interference
(a) (b)
Figure 23.9 (a) The amplitude of the antibonding molecular

Figure 23.7 A representation of the constructive interference orbital in a hydrogen molecule-ion in a plane containing the
that occurs when two H1s orbitals overlap and form a bonding two nuclei and (b) a contour representation of the amplitude.
σ orbital. Note the internuclear node.
between the nuclei for significant accumulation of electron

density there. In addition, the nucleus–nucleus repulsion (which
is proportional to 1/R) becomes large. As a result, the energy of
the molecule rises at short distances, and there is a minimum in
the potential energy curve. Calculations on H2+ give Re = 130 pm
and De = 1.76 eV (171 kJ mol−1); the experimental values are
106 pm and 2.6 eV, so this simple LCAO-MO description of the
molecule, while inaccurate, is not absurdly wrong.
(c) Antibonding orbitals

The linear combination ψ− in eqn 23.2 corresponds to a higher
Figure 23.10 The electron density calculated by forming the
energy than that of ψ+. Because it is also a σ orbital we label it
square of the wavefunction used to construct Fig. 23.9. Note
2σ. This orbital has an internuclear nodal plane where A and
the reduction of electron density in the internuclear region.
B cancel exactly (Figs 23.8 and 23.9; compare Fig. 23.1). The
probability density is orbital that, if occupied, contributes to a reduction in the cohe-
sion between two atoms and helps to raise the energy of the
ψ −2 = N 2 ( A2 + B 2 − 2AB) Antibonding probability density (23.6)
molecule relative to that of the separated atoms.
The energy E2σ of the 2σ antibonding orbital is given by (see
There is a reduction in probability density between the nuclei Problem 23.3)
due to the −2AB term (Fig. 23.10); in physical terms, there is
j0 j − k
destructive interference where the two atomic orbitals over- E2 σ = EH1s + − (23.7)
R 1− S
lap. The 2σ orbital is an example of an antibonding orbital, an
where the integrals S, j, and k are the same as before (eqn 23.5).
The variation of E2σ with R is shown in Fig. 23.6, where we see the
destabilizing effect of an antibonding electron. The effect is partly
Region of
due to the fact that an antibonding electron is excluded from the
destructive internuclear region and hence is distributed largely outside the
interference bonding region. In effect, whereas a bonding electron pulls two
nuclei together, an antibonding electron pulls the nuclei apart
(Fig. 23.11). Figure 23.6 also shows another feature that we draw
on later: |E− – EH1s| > |E+ – EH1s|, which indicates that the antibond-
ing orbital is more antibonding than the bonding orbital is bond-
ing. This important conclusion stems in part from the presence
of the nucleus–nucleus repulsion (j0/R): this contribution raises
Figure 23.8 A representation of the destructive interference the energy of both molecular orbitals. Antibonding orbitals are
that occurs when two H1s orbitals overlap and form an often labelled with an asterisk (*), so the 2σ orbital could also be
antibonding 2σ orbital. denoted 2σ * (and read ‘2 sigma star’).
Atkins09819.indb 213 9/11/2013 11:19:15 AM

Centre of
inversion
+ –
(a)
+ +
σg σu
(b)
Figure 23.12 The parity of an orbital is even (g) if its
wavefunction is unchanged under inversion through the centre
Figure 23.11 A partial explanation of the origin of bonding
of symmetry of the molecule, but odd (u) if the wavefunction
and antibonding effects. (a) In a bonding orbital, the nuclei
changes sign. Heteronuclear diatomic molecules do not
are attracted to the accumulation of electron density in the
have a centre of inversion, so for them the g,u classification is
internuclear region. (b) In an antibonding orbital, the nuclei are
irrelevant.
attracted to an accumulation of electron density outside the
internuclear region.
Brief illustration 23.3 g, as in σg. The same procedure applied to the antibonding 2σ
Antibonding energies
orbital results in the same amplitude but opposite sign of the
At the minimum of the bonding orbital energy we have seen wavefunction. This ungerade symmetry (‘odd symmetry’) is
that R = 2.45a0, and from Brief illustration 23.2 we know that denoted by a subscript u, as in σu.
S = 0.47, j = 11 eV, and k = 8.2 eV. It follows that at that separa-
tion, the energy of the antibonding orbital relative to that of a
hydrogen atom 1s orbital is Brief illustration 23.4 Inversion symmetry
27.2 11 − 8.2 Consider the 1σ orbital given in eqn 23.2 by N(A + B)
(E2 σ − EH1s )/ eV = − = 5.8
2.45 1 − 0.47 where the atomic orbitals A and B are specified in Brief
illustration 23.1. At the location of nucleus A, r A1 = 0
That is, the antibonding orbital lies above the bonding orbital and rB1 = R. The wavefunction at that point has the value
at this internuclear separation. N ( χ H1sA(0) + χ H1sB(R)) = N (1/πa03 )1/2 (1 + e − R/a ). Upon inversion
Self-test 23.4 What is the separation of the antibonding and through the centre of the molecule, which takes that point to
bonding orbital energies at twice that internuclear distance? rA1 = R and rB1 = 0, the wavefunction has the value correspond-
Answer: 1.4 eV ing to a point on nucleus B, namely, N (1/ π a03 )1/2 (e − R/a0 + 1). The
wavefunction has the same value and the σ orbital is gerade.
Self-test 23.5 Consider the antibonding 2σ orbital and show
23.2 Orbital notation in a similar way that it has ungerade symmetry.

Answer: 1− e − R/a0 → e − R/a0 −1
For homonuclear diatomic molecules and ions such as H2+ and
analogous many-electron species it proves helpful (for exam-
ple, in electronic spectroscopy, Topic 45) to label a molecular When using the g,u notation, each set of orbitals of the same
orbital according to its inversion symmetry, the behaviour of inversion symmetry are labelled separately so, whereas 1σ
the wavefunction when it is inverted through the centre (more becomes 1σg, its antibonding partner, which so far we have called
formally, the centre of inversion) of the molecule. Thus, if we 2σ, is the first orbital of a different symmetry, and is denoted 1σu.
consider any point on the bonding σ orbital, and then project The general rule is that each set of orbitals of the same symmetry
it through the centre of the molecule and out an equal dis- designation is labelled separately. This point is developed in Topic
tance on the other side, we arrive at an identical value of the 24. The inversion symmetry classification is not applicable to
wavefunction (Fig. 23.12). This so-called gerade symmetry the discussion of heteronuclear diatomic molecules in Topic 25
(from the German word for ‘even’) is denoted by a subscript because these molecules do not have a centre of inversion.
☐ 1. A molecular orbital is constructed as a linear combi- ☐ 2. A bonding orbital arises from the constructive overlap
nation of atomic orbitals. of neighbouring atomic orbitals.
www.ebook3000.com
Atkins09819.indb 214 9/11/2013 11:19:22 AM

☐ 3. An antibonding orbital arises from the destructive ☐ 5. A molecular orbital in a homonuclear diatomic mol-
overlap of neighbouring atomic orbitals. ecule is labelled ‘gerade’ or ‘ungerade’ according to its
☐ 4. σ Orbitals have cylindrical symmetry and zero orbital behaviour under inversion symmetry.
angular momentum around the internuclear axis.
Linear combination of atomic orbitals ψ ± = N ( A ± B) Homonuclear diatomic molecule 23.2
Energies of σ orbitals E1σ = EH1s + j0 /R − ( j + k )/(1+ S)

∫
S = ABdτ , 23.4
E2σ = EH1s + j0 /R − ( j − k)/(1− S) ∫

j = j0 ( A2 /rB)dτ 23.7
0
∫
k = j ( AB /r )dτ
B
Atkins09819.indb 215 9/11/2013 11:19:26 AM

TOPIC 24
Homonuclear diatomic molecules

In Topic 19 the hydrogenic atomic orbitals and the building-
Contents up principle are used as a basis for the discussion and predic-
24.1 Electron configurations 216 tion of the ground electronic configurations of many-electron
(a) σ Orbitals and π orbitals 216 atoms. We now do the same for many-electron diatomic mol-
Brief illustration 24.1: Ground-state configurations 218 ecules by using the H2+ molecular orbitals as a basis for their
(b) The overlap integral 218 discussion.
Brief illustration 24.2: Overlap integrals 219
(c) Period 2 diatomic molecules 219
Brief illustration 24.3: Bond order 220
Example 24.1: Judging the relative bond strengths 24.1 Electron configurations
of molecules and ions 221
24.2 Photoelectron spectroscopy 221 The starting point of the building-up principle for diatomic
Brief illustration 24.4: A photoelectron spectrum 222 molecules is the construction of molecular orbitals by combin-
Checklist of concepts 223 ing the available atomic orbitals. Once they are available, we
Checklist of equations 223 adopt the following procedure, which is essentially the same as
the building-up principle for atoms (Topic 19):
r The electrons supplied by the atoms are accommodated in
the orbitals so as to achieve the lowest overall energy
subject to the constraint of the Pauli exclusion principle,
➤ Why do you need to know this material? that no more than two electrons may occupy a
principle for
single orbital (and then must be paired).
Building-up
molecules
Although the hydrogen molecule-ion establishes the basic
approach to the construction of molecular orbitals, almost r If several degenerate molecular orbitals are
all chemically significant molecules have more than one available, electrons are added singly to each
electron, and we need to see how to construct their individual orbital before doubly occupying any
electron configurations. Homonuclear diatomic molecules one orbital (because that minimizes electron–
are a good starting point, not only because they are simple electron repulsions).
to describe but because they include such important r According to Hund's maximum multiplicity rule (Topics
species as H2, N2, O2, and the dihalogens. 20 and 21), if two electrons do occupy different degenerate
orbitals, then a lower energy is obtained if they do so with
➤ What is the key idea? parallel spins.
Each molecular orbital can accommodate up to two
electrons.
(a) σ Orbitals and π orbitals
➤ What do you need to know already? Consider H2, the simplest many-electron diatomic molecule.
You need to be familiar with the discussion of the bonding Each H atom contributes a 1s orbital (as in H2+ ), so we can form
and antibonding linear combinations of atomic orbitals the 1σg and 1σu orbitals from them, as explained in Topic 23.
in Topic 23 and the building-up principle for atoms At the experimental internuclear separation these orbitals will
(Topic 19). have the energies shown in Fig. 24.1, which is called a molecu-
lar orbital energy level diagram. Note that from two atomic
www.ebook3000.com
Atkins09819.indb 216 9/11/2013 11:32:15 AM

24 Homonuclear diatomic molecules 217
1σu We shall now see how the concepts we have introduced apply
to homonuclear diatomic molecules in general. In elementary
treatments, only the orbitals of the valence shell are used to
H1s H1s form molecular orbitals so, for molecules formed with atoms
from Period 2 elements, only the 2s and 2p atomic orbitals are
considered. We shall make that approximation here too.
1σg A general principle of molecular orbital theory is that all
orbitals of the appropriate symmetry contribute to a molecular
Figure 24.1 A molecular orbital energy level diagram for orbital. Thus, to build σ orbitals, we form linear combinations
orbitals constructed from the overlap of H1s orbitals; the of all atomic orbitals that have cylindrical symmetry about the
separation of the levels corresponds to that found at the internuclear axis. These orbitals include the 2s orbitals on each
equilibrium bond length. The ground electronic configuration atom and the 2pz orbitals on the two atoms (Fig. 24.3). The gen-
of H2 is obtained by accommodating the two electrons in the eral form of the σ orbitals that may be formed is therefore
lowest available orbital (the bonding orbital).
ψ = cA2s χ A2s + cB2s χ B2s + cA2pz χ A2pz + cB2pz χ B2pz (24.1)
orbitals we can build two molecular orbitals. In general, from N From these four atomic orbitals we can form four molecular
atomic orbitals we can build N molecular orbitals. orbitals of σ symmetry by an appropriate choice of the coef-
There are two electrons to accommodate, and both can enter ficients c.
1σg by pairing their spins, as required by the Pauli principle The procedure for calculating the coefficients is described
(just as for atoms, Topic 19). The ground-state configuration is in Topic 25. Here we adopt a simpler route, and suppose that,
therefore 1σ 2g and the atoms are joined by a bond consisting because the 2s and 2pz orbitals have distinctly different ener-
of an electron pair in a bonding σ orbital. This approach shows gies, they may be treated separately. That is, the four σ orbitals
that an electron pair, which was the focus of Lewis's account fall approximately into two sets, one consisting of two molecu-
of chemical bonding, represents the maximum number of elec- lar orbitals of the form
trons that can enter a bonding molecular orbital.
ψ =cA2s χ A2s + cB2s χ B2s (24.2a)
The same argument explains why He does not form diatomic
molecules. Each He atom contributes a 1s orbital, so 1σg and and another consisting of two orbitals of the form
1σu molecular orbitals can be constructed. Although these
ψ = cA2pz χ A2pz + cB2pz χ B2pz (24.2b)
orbitals differ in detail from those in H2, their general shapes
are the same and we can use the same qualitative energy level Because atoms A and B are identical, the energies of their 2s
diagram in the discussion. There are four electrons to accom- orbitals are the same, so the coefficients are equal (apart from a
modate. Two can enter the 1σg orbital, but then it is full, and possible difference in sign); the same is true of the 2pz orbitals.
the next two must enter the 1σu orbital (Fig. 24.2). The ground Therefore, the two sets of orbitals have the form χA2s ± χB2s and
electronic configuration of He2 is therefore 1 σ2g 1σ 2u . We see χA2pz ± χB2pz.
that there is one bond and one antibond. Because 1σu is raised The 2s orbitals on the two atoms overlap to give a bonding
in energy relative to the separate atoms more than 1σg is low- and an antibonding σ orbital (1σg and 1σu, respectively) in
ered, an He2 molecule has a higher energy than the separated exactly the same way as we have already seen for 1s orbitals.
atoms, so it is unstable relative to them. The two 2pz orbitals directed along the internuclear axis overlap
strongly. They may interfere either constructively or destruc-
tively, and give a bonding or antibonding σ orbital (Fig. 24.4).
1σu
2s 2s
2pz 2pz
He1s He1s
A B
1σg
Figure 24.3 According to molecular orbital theory, σ orbitals
Figure 24.2 The ground electronic configuration of the are built from all orbitals that have the appropriate symmetry.
hypothetical four-electron molecule He2 has two bonding In homonuclear diatomic molecules of Period 2, that means
electrons and two antibonding electrons. It has a higher that two 2s and two 2pz orbitals should be used. From these
energy than the separated atoms, and so is unstable. four orbitals, four molecular orbitals can be built.
Atkins09819.indb 217 9/11/2013 11:32:21 AM

in atoms, for when viewed along the axis of the molecule, a π

– + – +
orbital looks like a p orbital and has one unit of orbital angular
momentum around the internuclear axis. More formally, the
2σu orbital has a single nodal plane that includes the internuclear
axis. The two neighbouring 2px orbitals overlap to give a bond-
– + + –
ing and an antibonding πx orbital, and the two 2py orbitals
overlap to give two πy orbitals. The πx and πy bonding orbit-
2σg als are degenerate; so too are their antibonding partners. We
also see from Fig. 24.5 that a bonding π orbital has odd parity
Figure 24.4 A representation of the composition of bonding
and is denoted πu and an antibonding π orbital has even parity,
and antibonding σ orbitals built from the overlap of p orbitals.
denoted πg.
These illustrations are schematic.
These two σ orbitals are labelled 2σg and 2σu, respectively. In

(b) The overlap integral
general, note how the numbering follows the order of increas- The extent to which two atomic orbitals on different atoms
ing energy. We number only the molecular orbitals formed overlap is measured by the overlap integral, S:
from atomic orbitals in the valence shell.
∫
S = χ A* χ B dτ Definition Overlap integral (24.3)
Brief illustration 24.1 Ground-state configurations This integral also appears in Example 23.1 of Topic 23 and eqn
23.5a. If the atomic orbital χA on A is small wherever the orbital
The valence configuration of a sodium atom is [Ne]3s1, so 3s χB on B is large, or vice versa, then the product of their ampli-
and 3p orbitals are used to construct molecular orbitals. At
tudes is everywhere small and the integral—the sum of these
this level of approximation, we consider (3s,3s) and (3p,3p)
products—is small (Fig. 24.6). If χA and χB are both large in
overlap separately. In fact, because there are only two electrons
some region of space, then S may be large. If the two normal-
to accommodate (one from each 3s orbital), we need consider
ized atomic orbitals are identical (for instance, 1s orbitals on
only the former. That overlap results in 1σg and 1σu molecular
the same nucleus), then S = 1. In some cases, simple formulas
orbitals. The only two valence electrons occupy the former, so
the ground-state configuration of Na 2 is 1σ 2g . can be given for overlap integrals. For instance, the variation
of S with internuclear separation for hydrogenic 1s orbitals on
Self-test 24.1 Identify the ground-state configuration of Be2. atoms of atomic number Z is given by
Answer: 1σ 2g 1σ 2u built from Be2s orbitals
⎧⎪ ZR 1 ⎛ ZR ⎞ 2 ⎫⎪
S(1s,1s) = ⎨1+ + ⎜ ⎟⎠ ⎬ e
− ZR/a0 (1s,1s)-overlap integral (24.4)
a 3 ⎝ a
⎩⎪ 0 0
⎭⎪
Now consider the 2px and 2py orbitals of each atom.
These orbitals are perpendicular to the internuclear axis and and is plotted in Fig. 24.7 (eqn 24.4 is a generalization of eqn
may overlap broadside-on. This overlap may be construc- 23.5a, which is for H1s orbitals).
tive or destructive and results in a bonding or an antibond-
ing π orbital (Fig. 24.5). The notation π is the analogue of p
+ + – + + –
Centre of inversion
+ – + + (a) (b)
Figure 24.6 (a) When two orbitals are on atoms that are far apart,
– + – – the wavefunctions are small where they overlap, so S is small.
πg πu (b) When the atoms are closer, both orbitals have significant
amplitudes where they overlap, and S may approach 1. Note that
Figure 24.5 A schematic representation of the structure of π S will decrease again as the two atoms approach more closely
bonding and antibonding molecular orbitals. The figure also than shown here, because the region of negative amplitude
shows that the bonding π orbital has odd parity, whereas the of the p orbital starts to overlap the positive amplitude of the
antibonding π orbital has even parity. s orbital. When the centres of the atoms coincide, S = 0.
www.ebook3000.com
Atkins09819.indb 218 9/11/2013 11:32:27 AM

1 Constructive
+
0.8
+
Overlap integral, S
0.6
–
0.4
Destructive
0.2
Figure 24.9 A p orbital in the orientation shown here has
zero net overlap (S = 0) with the s orbital at all internuclear
0
0 2 4 6 separations.
Internuclear separation, R/a0
Now consider the arrangement in which an s orbital is super-

Figure 24.7 The overlap integral, S, between two H1s orbitals
imposed on a px orbital of a different atom (Fig. 24.9). The inte-
as a function of their separation R.
gral over the region where the product of orbitals is positive
exactly cancels the integral over the region where the product
Brief illustration 24.2 Overlap integrals of orbitals is negative, so overall S = 0 exactly. Therefore, there is
Familiarity with the magnitudes of overlap integrals is useful no net overlap between the s and p orbitals in this arrangement.
when considering bonding abilities of atoms, and hydrogenic
orbitals give an indication of their values. The overlap integral (c) Period 2 diatomic molecules
between two hydrogenic 2s orbitals (see Problem 24.5) is
To construct the molecular orbital energy level diagram for
⎧⎪ ZR 1 ⎛ ZR ⎞ 2 1 ⎛ ZR ⎞ 4 ⎫⎪ Period 2 homonuclear diatomic molecules, we form eight
S(2s, 2s) = ⎨1 + + + e − ZR/2a0
2a0 12 ⎜⎝ a0 ⎟⎠ 240 ⎜⎝ a0 ⎟⎠ ⎬⎪ molecular orbitals from the eight valence shell orbitals (four
⎩⎪ ⎭
from each atom). In some cases, π orbitals are less strongly
This expression is plotted in Fig. 24.8. For an internuclear bonding than σ orbitals because their maximum overlap occurs
distance of 8a0/Z, S(2s,2s) = 0.50. off-axis. This relative weakness suggests that the molecular
1 orbital energy level diagram ought to be as shown in Fig. 24.10.
However, we must remember that we have assumed that 2s and
0.8 2pz orbitals contribute to different sets of molecular orbitals
whereas in fact all four atomic orbitals have the same symme-
0.6 try around the internuclear axis and contribute jointly to the
(2s,2s)
S four σ orbitals. Hence, there is no guarantee that this order
0.4 of energies should prevail, and it is found experimentally (by
(2p,2p)
0.2
Atom Molecule Atom
2σu
0
0 5 10 15 20 1πg
ZR/a0
2p 2p
1πu
Figure 24.8 The overlap integral, S, between two
hydrogenic 2s orbitals and between two side-by-side 2p
orbitals as a function of their separation R. 2σg
1σu
Self-test 24.2 The side-by-side overlap of two 2p orbitals of 2s 2s
atoms of atomic number Z is 1σg
⎧⎪ ZR 1 ⎛ ZR ⎞ 2 1 ⎛ ZR ⎞ 3 ⎫⎪
S(2p, 2p) = ⎨1 + + ⎜ ⎟⎠ + 120 ⎜⎝ a ⎟⎠ ⎬ e
− ZR/2 a0
2a 10 ⎝ a Figure 24.10 The molecular orbital energy level diagram for
⎩⎪ 0 0 0
⎭⎪
homonuclear diatomic molecules. The lines in the middle are an
Evaluate this overlap integral for R = 8a0/Z. indication of the energies of the molecular orbitals that can be
Answer: See Fig. 24.8; 0.29 formed by overlap of atomic orbitals. As remarked in the text, this
diagram should be used for O2 (the configuration shown) and F2.
Atkins09819.indb 219 9/11/2013 11:32:31 AM

Li2 Be2 B2 C2 N2 O2 F2 asterisk to denote an antibonding orbital, in which case this

2σu configuration would be denoted 1σ 2g 1σ *u2 1π u4 2σ 2g .
1πg A measure of the net bonding in a diatomic molecule is its
2σu
2σg bond order, b:
1πg
b = 12 (N − N *)
Energy
Definition Bond order (24.5)

1πu
1πu
2σg where N is the number of electrons in bonding orbitals and
1σu
N* is the number of electrons in antibonding orbitals.
1σg 1σu
1σg Brief illustration 24.3 Bond order

Figure 24.11 The variation of the orbital energies of Period 2 Each electron pair in a bonding orbital increases the bond
homonuclear diatomics. order by 1 and each pair in an antibonding orbital decreases
b by 1. For H 2 , b = 1, corresponding to a single bond, H–H,
Atom Molecule Atom between the two atoms. In He2 , b = 0, and there is no bond.
2σu In N2 , b = 12 (8 − 2) = 3. This bond order accords with the Lewis
1πg structure of the molecule (:N≡N:).
2p 2σg 2p Self-test 24.3 Evaluate the bond orders of O2, O2+ , and O2− .
5 3
1πu Answer: 2, ,
2 2
1σu The ground-state electron configuration of O2, with 12 valence

2s 2s electrons, is based on Fig. 24.10, and is 1σ 2g 1σ 2u 2σ 2g 1π u4 1π 2g (or
1σg 1σ 2g 1σ *u2 2σ 2g 1π u4 1π *g 2 ). Its bond order is 2. According to the
building-up principle, however, the two 1πg electrons occupy
different orbitals: one will enter 1πg,x and the other will enter
Figure 24.12 An alternative molecular orbital energy level
1πg,y . Because the electrons are in different orbitals, they will
diagram for homonuclear diatomic molecules. As remarked in
have parallel spins. Therefore, we can predict that an O2 mol-
the text, this diagram should be used for diatomics up to and
ecule will have a net spin angular momentum S = 1 and, in the
including N2 (the configuration shown).
language introduced in Topic 21, be in a triplet state. As electron
spin is the source of a magnetic moment, we can go on to pre-
spectroscopy) and by detailed calculation that the order varies dict that oxygen should be paramagnetic, a substance that tends
along Period 2 (Fig. 24.11). The order shown in Fig. 24.12 is to move into a magnetic field (see Topic 39). This prediction,
appropriate as far as N2, and Fig. 24.10 is appropriate for O2 and which VB theory does not make, is confirmed by experiment.
F2. The relative order is controlled by the separation of the 2s An F2 molecule has two more electrons than an O2 mol-
and 2p orbitals in the atoms, which increases across the group: ecule. Its configuration is therefore 1σ 2g 1σ *u2 2σ 2g 1π u4 1π *g4 and
as the atomic number increases, the 2s electrons are pulled b = 1. We conclude that F2 is a singly bonded molecule, in
closer to the nucleus and thus increasingly shield the 2p elec- agreement with its Lewis structure. The hypothetical molecule
trons. The consequent switch in order occurs at about N2. dineon, Ne2, has two additional electrons: its configuration is
With the molecular orbital energy level diagram established, 1σ 2g 1σ *u2 2σ 2g 1π u4 1π g*4 2σ *u2 and b = 0. The zero bond order is con-
we can deduce the probable ground-state configurations of the sistent with the monatomic nature of Ne.
molecules by adding the appropriate number of electrons to The bond order is a useful parameter for discussing the charac-
the orbitals and following the building-up rules. Anionic species teristics of bonds, because it correlates with bond length and bond
(such as the peroxide ion, O2− 2 ) need more electrons than the par- strength. For bonds between atoms of a given pair of elements:
ent neutral molecules; cationic species (such as O2+ ) need fewer.
Consider N2, which has 10 valence electrons. Two electrons r The greater the bond order, the shorter the bond.
pair, occupy, and fill the 1σg orbital; the next two occupy and fill r The greater the bond order, the greater the bond strength.
the 1σu orbital. Six electrons remain. There are two 1πu orbitals,
so four electrons can be accommodated in them. The last two Table 24.1 lists some typical bond lengths in diatomic and
enter the 2σg orbital. Therefore, the ground-state configuration polyatomic molecules. The strength of a bond is measured
of N2 is 1σ g2 1σ u2 1π u4 2σ g2. It is sometimes helpful to include an by its bond dissociation energy, hcD 0, the energy required to
www.ebook3000.com
Atkins09819.indb 220 9/11/2013 11:32:41 AM

Table 24.1* Bond lengths
Bond Order Re/pm

24.2 Photoelectron spectroscopy
HH 1 74.14 So far we have treated molecular orbitals as purely theoretical
NN 3 109.76 constructs, but is there experimental evidence for their exist-
HCl 1 127.45 ence? Photoelectron spectroscopy (PES) measures the ioni-
CH 1 114 zation energies of molecules when electrons are ejected from
CC 1 154
different orbitals by absorption of a photon of known energy,
and uses the information to infer the energies of molecular
CC 2 134
orbitals. The technique is also used to study solids, and in Topic
CC 3 120
95 we see the important information that it gives about species
*More values are given in the Resource section. Numbers in italics are mean values for at or on surfaces.
polyatomic molecules.
Because energy is conserved when a photon ionizes a sam-
ple, the sum of the ionization energy, I, of the sample and
Table 24.2* Bond dissociation energies the kinetic energy of the photoelectron, the ejected elec-
tron, must be equal to the energy of the incident photon h␯
Bond Order hcD 0 (kJ mol −1 )
(Fig. 24.13):
HH 1 432.1
NN 3 941.7 h␯ = 12 me v 2 + I (24.6)
HCl 1 427.7
CH 1 435 This equation (which is like the one used for the photoelec-
CC 1 368
tric effect, eqn 4.5 written as h␯ = 12 me v 2 + Φ ) can be refined
in two ways. First, photoelectrons may originate from one of
CC 2 720
a number of different orbitals, and each one has a different
CC 3 962
ionization energy. Hence, a series of different kinetic ener-
*More values are given in the Resource section. Numbers in italics are mean values for gies of the photoelectrons will be obtained, each one satisfying
polyatomic molecules.
h␯ = 12 me v 2 + Ii , where Ii is the ionization energy for ejection
of an electron from an orbital i. Therefore, by measuring
separate the atoms to infinity or by the well depth, hcD e , with the kinetic energies of the photoelectrons, and knowing ␯,
hcD 0 = hcD e − 12 ω . Table 24.2 lists some experimental values these ionization energies can be determined. Photoelectron
of hcD 0 . spectra are interpreted in terms of an approximation called
Koopmans’ theorem, which states that the ionization energy Ii
Example 24.1 is equal to the orbital energy of the ejected electron (formally,
Judging the relative bond strengths
Ii = –εi). That is, we can identify the ionization energy with the
of molecules and ions
energy of the orbital from which it is ejected. The theorem
Predict whether N2+ is likely to have a larger or smaller bond is only an approximation because it ignores the fact that the
dissociation energy than N2. remaining electrons adjust their distributions when ionization
Method Because the molecule with the higher bond order is occurs.
likely to have the higher dissociation energy, compare their
electronic configurations and assess their bond orders.
X+ + e–(moving, Ek)
Answer From Fig. 24.12, the electron configurations and bond
orders are hν – Ii
N2 1σ 2g 1σ *2
u 1πu 2 σg
4 2
b=3
X+ + e–(stationary)
N2 1σ g 1σ u 1 π u 2 σ g
+ 2 *2 4 1
b = 2 12
Ii
Because the cation has the smaller bond order, we expect it to hν
have the smaller bond dissociation energy. The experimental
dissociation energies are 942 kJ mol−1 for N2 and 842 kJ mol−1
for N2+ .
Orbital i X
Self-test 24.4 Which can be expected to have the higher dis-
sociation energy, F2 or F2+ ? Figure 24.13 An incoming photon carries an energy h␯; an
Answer: F2+ energy Ii is needed to remove an electron from an orbital i, and
the difference appears as the kinetic energy of the electron.
Atkins09819.indb 221 9/11/2013 11:32:51 AM

Sample
to remove an electron from the occupied molecular orbital
with the highest energy of the N2 molecule, the 2σg bonding
Lamp
orbital.
Detector
Electrostatic
analyser
+
Signal
–
Figure 24.14 A photoelectron spectrometer consists of a

source of ionizing radiation (such as a helium discharge lamp
for UPS and an X-ray source for XPS), an electrostatic analyser,
and an electron detector. The deflection of the electron path 16 17 18 19 20
caused by the analyser depends on the speed of the electrons. Ionization energy, I/eV
Figure 24.15 The photoelectron spectrum of N2.

The ionization energies of molecules are several electron-
volts even for valence electrons, so it is essential to work in Self-test 24.5 Under the same circumstances, photoelectrons
at least the ultraviolet region of the spectrum and with wave- are also detected at 4.53 eV. To what ionization energy does
lengths of less than about 200 nm. Much work has been done that correspond? Suggest an origin.
with radiation generated by a discharge through helium: the Answer: 16.7 eV, 1πu
He(I) line (1s12p1 → 1s2) lies at 58.43 nm, corresponding to a
photon energy of 21.22 eV. Its use gives rise to the technique
of ultraviolet photoelectron spectroscopy (UPS). When core It is often observed that photoejection results in cations that
electrons are being studied, photons of even higher energy are are excited vibrationally. Because different energies are needed
needed to expel them: X-rays are used, and the technique is to excite different vibrational states of the ion, the photoelec-
denoted XPS. trons appear with different kinetic energies. The result is vibra-
The kinetic energies of the photoelectrons are measured tional fine structure, a progression of lines with a frequency
using an electrostatic deflector that produces different deflec- spacing that corresponds to the vibrational frequency of the
tions in the paths of the photoelectrons as they pass between molecule. Figure 24.16 shows an example of vibrational fine
charged plates (Fig. 24.14). As the field strength is increased, structure in the photoelectron spectrum of HBr.
electrons of different speeds, and therefore kinetic energies,
reach the detector. The electron flux can be recorded and
plotted against kinetic energy to obtain the photoelectron
spectrum. 2
Π3/2 2
Π1/2
Signal
Brief illustration 24.4 A photoelectron spectrum 10.5 11.0
Photoelectrons ejected from N2 with He(I) radiation have

kinetic energies of 5.63 eV (1 eV = 8065.5 cm −1, Fig. 24.15).
Helium(I) radiation of wavelength 58.43 nm has wave-
number 1.711 × 10 5 cm −1 and therefore corresponds to an
energy of 21.22 eV. Then, from eqn 24.6, 21.22 eV = 5.63 eV + Ii, 10 15
Ionization energy, I/eV
so Ii = 15.59 eV. This ionization energy is the energy needed
Figure 24.16 The photoelectron spectrum of HBr.
www.ebook3000.com
Atkins09819.indb 222 9/11/2013 11:32:54 AM

☐ 1. Electrons are added to available molecular orbitals in a they have a nodal plane that includes the internuclear
manner that achieves the lowest overall energy. axis.
☐ 2. As a first approximation, σ orbitals are constructed ☐ 5. An overlap integral is a measure of the extent of orbital
separately from valence s and p orbitals. overlap.
☐ 3. σ Orbitals have cylindrical symmetry and zero orbital ☐ 6. Photoelectron spectroscopy is a technique for deter-
angular momentum around the internuclear axis. mining the energies of electrons in molecular orbitals.
☐ 4. π Orbitals are constructed from the side-by-side ☐ 7. The greater the bond order of a molecule, the shorter
overlap of orbitals of the appropriate symmetry; and stronger the bond.
Overlap integral S = ∫ χ A* χ B dτ 24.3
Bond order b = 12 (N − N *) 24.5
Photoelectron spectroscopy h␯ = 12 me v2 + Ii Ii is the ionization energy from orbital i 24.6
Atkins09819.indb 223 9/11/2013 11:32:57 AM

TOPIC 25
Heteronuclear diatomic
molecules
Contents ➤ What do you need to know already?

25.1 Polar bonds 225 You need to know about the molecular orbitals of homo-
(a) The molecular orbital formulation 225 nuclear diatomic molecules (Topic 24) and the concepts
Brief illustration 25.1: Heteronuclear diatomic of normalization and orthogonality (Topics 5 and 7). This
molecules 1 225 Topic makes use of determinants (Mathematical back-
(b) Electronegativity 225 ground 5) and rules about differentiation (Mathematical
Brief illustration 25.2 Electronegativity 226 background 1).
25.2 The variation principle 226
(a) The procedure 227
Brief illustration 25.3: Heteronuclear diatomic The electron distribution in a covalent bond in a heteronuclear
molecules 2 228 diatomic molecule is not shared equally by the atoms because
(b) The features of the solutions 228 it is energetically favourable for the electron pair to be found
Brief illustration 25.4: Heteronuclear diatomic closer to one atom than the other. This imbalance results in a
molecules 3 229 polar bond, a covalent bond in which the electron pair is shared
Checklist of concepts 230 unequally by the two atoms. The bond in HF, for instance, is
Checklist of equations 230 polar, with the electron pair closer to the F atom. The accumu-
lation of the electron pair near the F atom results in that atom
having a net negative charge, which is called a partial negative
charge and denoted δ−. There is a matching partial positive
charge, δ+, on the H atom (Fig. 25.1).
➤ Why do you need to know this material?

Most molecules are heteronuclear, so you need to
δ+ H F δ–
appreciate the differences in their electronic structure
from homonuclear species, and how to treat those
differences quantitatively.
➤ What is the key idea? Figure 25.1 The electron density of the molecule HF,
The bonding molecular orbital of a heteronuclear diatomic computed with one of the methods described in Topic
molecule is composed mostly of the atomic orbital of the 29. Different colours show the distribution of electrostatic
more electronegative atom; the opposite is true of the potential and hence net charge, with blue representing the
antibonding orbital. region with largest partial positive charge, and red the region
with largest partial negative charge.
www.ebook3000.com
Atkins09819.indb 224 9/11/2013 11:33:00 AM

25 Heteronuclear diatomic molecules 225
the more negative value of − 12 {I (X ) + Eea (X )} contributes the

25.1 Polar bonds greater amount to the bonding orbital. As we shall see shortly,
the quantity 12 {I (X ) + Eea (X )} also has a further significance.
The description of polar bonds in terms of molecular orbital theory
is a straightforward extension of that for homonuclear diatomic
molecules, the central difference being that the atomic orbitals on Brief illustration 25.1 Heteronuclear diatomic molecules 1
the two atoms have different energies and spatial extensions.
These points can be illustrated by considering HF. The general
form of the molecular orbital is ψ = cHχH+cFχF, where χH is an
(a) The molecular orbital formulation H1s orbital and χF is an F2pz orbital (with z along the internu-
clear axis, the convention for linear molecules). The relevant
A polar bond consists of two electrons in a bonding molecular data are as follows:
orbital of the form
I/eV Eea/eV 1
{I + Eea }/ eV
ψ = c A A + cB B
2
Wavefunction of a polar bond (25.1)
H 13.6 0.75 7.2
with unequal coefficients. The proportion of the atomic orbital F 17.4 3.34 10.4
A in the bond is |cA|2 and that of B is |cB|2. A nonpolar bond
We see that the electron distribution in HF is likely to be pre-
has |cA|2 = |cB|2 and a pure ionic bond has one coefficient zero
dominantly on the F atom. We take the calculation further
(so the species A+B− would have cA = 0 and cB = 1). The atomic
below (in Brief illustrations 25.3 and 25.4).
orbital with the lower energy makes the larger contribution
to the bonding molecular orbital. The opposite is true of the Self-test 25.1 Which atomic orbital, H1s or N2p z makes the
antibonding orbital, for which the dominant component comes dominant contribution to the bonding σ orbital in the HN
from the atomic orbital with higher energy. molecular radical? For data, see Tables 20.2 and 20.3.
Deciding what values to use for the energies of the atomic Answer: N2pz
orbitals in eqn 25.1 presents a dilemma because they are known
only after a complicated calculation of the kind described in
Topic 29 has been performed. An alternative, one that gives (b) Electronegativity
some insight into the origin of the energies, is to estimate
them from ionization energies and electron affinities. Thus, the The charge distribution in bonds is commonly discussed in
extreme cases of an atom X in a molecule are X+ if it has lost terms of the electronegativity, χ (chi), of the elements involved
control of the electron it supplied, X if it is sharing the electron (there should be little danger of confusing this use of χ with
pair equally with its bonded partner, and X− if it has gained its use to denote an atomic orbital, which is another common
control of both electrons in the bond. If X+ is taken as defining convention). The electronegativity is a parameter introduced
the energy 0, then X lies at −I(X) and X− lies at −{I(X) + Eea(X)}, by Linus Pauling as a measure of the power of an atom to attract
where I is the ionization energy and Eea the electron affinity (Fig. electrons to itself when it is part of a compound. Pauling used
25.2). The actual energy of the orbital lies at an intermediate valence-bond arguments to suggest that an appropriate numer-
value, and in the absence of further information, we shall esti- ical scale of electronegativities could be defined in terms of
mate it as halfway down to the lowest of these values, namely bond dissociation energies, D0, and proposed that the differ-
− 12 {I (X ) + Eea (X )}. Then, to establish the MO composition and ence in electronegativities could be expressed as
energies, we form linear combinations of atomic orbitals with
these values of the energy and anticipate that the atom with | χA − χB | Pauling
= {D0 (AB) − 12 [D0 (AA) + D0 (BB)]}1/2 Definition electro- (25.2)
negativity
0 X+ + e–
I(X) where D0(AA) and D0(BB) are the dissociation energies of A–A
Energy
–½{I(X) + Eea(X)}
and B–B bonds and D0(AB) is the dissociation energy of an
A–B bond, all in electronvolts. (In later work Pauling used the
–I(X) X geometrical mean of dissociation energies in place of the arith-
Eea(X)
–I(X) – Eea(X) X– metic mean.) This expression gives differences of electronega-
tivities; to establish an absolute scale Pauling chose individual
Figure 25.2 The procedure for estimating the energy of an values that gave the best match to the values obtained from
atomic orbital in a molecule. eqn 25.2. Electronegativities based on this definition are called
Atkins09819.indb 225 9/11/2013 11:33:08 AM

Table 25.1* Pauling electronegativities scales are approximately in line with each other. A reasonably
reliable conversion relation between the two is
Element XPauling
H 2.2 χ Pauling =1.35 χ Mulliken

1/2
−1.37 (25.4)
C 2.6
N 3.0
O 3.4
F 4.0
25.2 The variation principle
Cl 3.2
A more systematic way of discussing bond polarity and finding
Cs 0.79
the coefficients in the linear combinations used to build molec-
* More values are given in the Resource section. ular orbitals is provided by the variation principle, which is
proved in the following Justification:
Pauling electronegativities (Table 25.1). The most electronega-
If an arbitrary wavefunction is used to calculate Variation
tive elements are those close to F (excluding the noble gases); principle
the energy, the value calculated is never less than
the least are those close to Cs. It is found that the greater the
the true energy.
difference in electronegativities, the greater the polar character
of the bond. The difference for HF, for instance, is 1.78; a C–H This principle, which is also described briefly in Topic 22, is the
bond, which is commonly regarded as almost nonpolar, has an basis of all modern molecular structure calculations (Topics
electronegativity difference of 0.35. 27−30). The arbitrary wavefunction is called the trial wave-
function. The principle implies that, if we vary the coefficients
Brief illustration 25.2 Electronegativity in the trial wavefunction until the lowest energy is achieved
(by evaluating the expectation value of the hamiltonian for
The bond dissociation energies of hydrogen, chlorine, and each wavefunction), then those coefficients will be the best. We
hydrogen chloride are 4.52 eV, 2.51 eV, and 4.47 eV, respec- might get a lower energy if we use a more complicated wave-
tively. From eqn 25.2 we find function (for example, by taking a linear combination of several
| χ Pauling (H) − χ Pauling (Cl )| = {4.47 − 12 (4.52 + 2.51)}1/2 = 0.98 ≈ 1.0 atomic orbitals on each atom), but we shall have the optimum
(minimum energy) molecular orbital that can be built from the
Self-test 25.2 Repeat the analysis for HBr. Use data from Table chosen basis set, the given set of atomic orbitals.
24.2.
Answer: |χPauling(H) – χPauling(Br)| = 0.73 Justification 25.1 The variation principle
To justify the variation principle, consider a trial (normalized)
The spectroscopist Robert Mulliken proposed an alterna- wavefunction written as a linear combination ψ trial = ∑n cnψ n
tive definition of electronegativity. He argued that an element of the true (but unknown), normalized, and orthogonal eigen-
functions of the hamiltonian H. The energy associated with
is likely to be highly electronegative if it has a high ionization
energy (so it will not release electrons readily) and a high elec- this trial function is the expectation value
tron affinity (so it is energetically favourable to acquire elec-
trons). The Mulliken electronegativity scale is therefore based ∫
E = ψ trial
*
Ηψ trial dτ
on the definition
The true lowest energy of the system is E0, the eigenvalue cor-
χ = (I + Eea )
1
Definition Mulliken electronegativity (25.3) responding to ψ 0. Consider the following difference:
2

(1)

where I is the ionization energy of the element and Eea is its elec-
tron affinity (both in electronvolts). It will be recognized that ∫
E − E0 = ψ trial
*
∫
H ψ trial dτ − E0 ψ trial
*
ψ trial dτ
this combination of energies is precisely the one we have used = ∫ ψ Hψ dτ − ∫ ψ E ψ

*
trial trial
*
trial 0 trial dτ
to estimate the energy of an atomic orbital in a molecule, and
we can therefore see that the greater the value of the Mulliken = ∫ ψ (H − E )ψ dτ
*
trial 0 trial
electronegativity the greater is the contribution of that atom to ⎛ ⎞ ⎛ ⎞

the electron distribution in the bond. There is one word of cau- = ∫ ⎜ ∑c ψ ⎟ (H − E ) ⎜ ∑c ψ
*
n
*
n 0 n′ n′⎟ dτ
tion: the values of I and Eea in eqn 25.3 are strictly those for a ⎝ ⎠ n ⎝ n′ ⎠
special ‘valence state’ of the atom, not a true spectroscopic state.
We ignore that complication here. The Mulliken and Pauling
= ∑c c ∫ψ (H − E )ψ
n ,n ′
*
n n′
*
n 0 n ′ dτ
www.ebook3000.com
Atkins09819.indb 226 9/11/2013 11:33:13 AM

B e caus e
∫ψ Hψ *
n n ′ dτ ∫
= En′ ψ n*ψ n′ dτ and
∫ψ E ψ
*
n 0 n ′ dτ = Justification 25.2The variation principle applied
∫
E0 ψ n*ψ n′ dτ , we write
to a heteronuclear diatomic molecule
The trial wavefunction in eqn 25.1 is real but not normalized
∫ψ (H − E )ψ
*
n 0 n ′ dτ ∫
= (En′ − E0 ) ψ n*ψ n′ dτ because at this stage the coefficients can take arbitrary val-
ues. Therefore, we can write ψ * = ψ but we do not assume that
and ∫ψ 2 dτ = 1. When a wavefunction is not normalized, we write
0 unless n ′ = n the expression for the energy (Topic 7) as

n ,n ′ ∫
E − E0 = ∑ cn*cn′ (En′ − E0 ) ψ n*ψ n′ dτ
∫ψ Ηψ dτ ⎯⎯⎯→ ∫ψΗψ dτ
*
ψ real
E= Energy (25.6)
The eigenfunctions are orthogonal, so only n′ = n contributes ∫ ψ ψ dτ*
∫ ψ dτ 2
to this sum, and as each eigenfunction is normalized, each

integral is 1. Consequently We now search for values of the coefficients in the trial func-
tion that minimize the value of E. This is a standard problem

≥0 ≥0

in calculus, and is solved by finding the coefficients for which
E − E0 = ∑ cn*cn (En − E0 ) ≥ 0
n
∂E ∂E
=0 =0
That is, E ≥ E0, as we set out to prove. ∂c A ∂c B
The first step is to express the two integrals in eqn 25.6 in

terms of the coefficients. The denominator is
(a) The procedure
The method can be illustrated by the trial wavefunction in
∫ψ 2
∫
dτ = (cA A + cB B)2 dτ

1
1
S

eqn 25.1. We show in the following Justification that the
coefficients are given by the solutions of the two secular
= cA2 ∫ A dτ ∫ B dτ + 2c c ∫ AB dτ = c
2
+ cB2 2
A B
2
A + cB2 + 2cA cBS
equations1
because the individual atomic orbitals are normalized and the
(α A − E )cA + (β − ES)cB =0 (25.5a) third integral is the overlap integral S (eqn 24.3). The numerator is
(β − ES)cA + (α B − E )cB =0 (25.5b) ∫ψHψ dτ = ∫ (c A + c B)H (c A + c B)dτ

A B A B
αA αB β β

where
= cA2 ∫
∫
AHA dτ + cB BHB dτ + cA cB AHB dτ + cA cB BHA dτ
2
∫ ∫
∫
dτ α = BHB
α A = AHA B
dτ
∫ Coulomb integrals (25.5c)
With the integrals written as shown (the two β integrals are
equal by hermiticity, Topic 6), the numerator is
∫ dτ = BHA
β = AHB dτ
∫ Resonance integral (25.5d)
∫ψHψ dτ = c α 2
A A + cB2α B + 2cA cB β
The parameter α is called a Coulomb integral. It is negative At this point we can write the complete expression for E as
and can be interpreted as the energy of the electron when it
occupies A (for αA) or B (for αB). In a homonuclear diatomic cA2 α A + cB2α B + 2cA cB β
E=
molecule, αA = αB. The parameter β is called a resonance inte- cA2 + cB2 + 2cA cBS
gral (for classical reasons). It vanishes when the orbitals do
not overlap, and at equilibrium bond lengths it is normally Its minimum is found by differentiation with respect to the
negative. two coefficients and setting the results equal to 0. After some
straightforward work we obtain
∂E 2{(α A − E )cA + (β − SE )cB }

=
1
∂c A cA2 + cB2 + 2cA cBS
The name ‘secular’ is derived from the Latin word for age or generation.
∂E 2{(α B − E )cB + (β − SE )cA }
The term comes from astronomy, where the same equations appear in con- =
nection with slowly accumulating modifications of planetary orbits. ∂c B cA2 + cB2 + 2cA cBS
Atkins09819.indb 227 9/11/2013 11:33:32 AM

For the derivatives to be equal to 0, the numerators of these Brief illustration 25.3
expressions must vanish. That is, we must find values of cA and
Heteronuclear diatomic molecules 2
cB that satisfy the conditions In Brief illustration 25.1 we estimated the H1s and F2p orbital
energies in HF as −7.2 eV and −10.4 eV, respectively. Therefore
(α A − E )cA + (β − SE )cB = 0 we set αH = −7.2 eV and αF = −10.4 eV. We take β = −1.0 eV as a
(α B − E )cB + (β − SE )cA = 0 typical value and S = 0. Substituting these values into eqn 25.8c
gives
which are the secular equations (eqn 25.5). 1/2
⎪ ⎛ ⎧ −2.0 ⎞
2⎫
⎪
E± /eV = 1
(−7.2 −10.4) ± 12 (−7.2 + 10.4) ⎨1 + ⎜ ⎬
2
⎝ −7.2 + 10.4 ⎟⎠
To solve the secular equations for the coefficients we need to ⎩⎪ ⎭⎪
know the energy E of the orbital. As for any set of simultaneous = −8.8 ± 1.9 = −10.7 and − 6.9
equations (Mathematical background 5), the secular equations
These values, representing a bonding orbital at −10.7 eV and
have a solution if the secular determinant, the determinant of
an antibonding orbital at −6.9 eV, are shown in Fig. 25.3.
the coefficients, is zero; that is, if
α A − E β − SE Ionization limit
= (α A − E )(α B − E ) − (β − SE )2
β − SE α B − E
10.7 eV
10.4 eV
6.9 eV
7.2 eV
= (1− S 2 )E 2 + {2βS − (α A + α B )}E (25.7)
+ (α Aα B − β ) 2
0.96χH – 0.28χF
=0
H1s
This quadratic equation has two roots that give the energies of
the bonding and antibonding molecular orbitals formed from
F2p
the atomic orbitals: 0.28χH + 0.96χF
α A + α B − 2 βS ± {(α A + α B − 2 βS)2 − 4(1− S 2 )(α Aα B − β 2 )}1/2 Figure 25.3 The estimated energies of the atomic orbitals
E± =
2(1− S 2 ) in HF and the molecular orbitals they form.
(25.8a) Self-test 25.3 Use S = 0.20 (a typical value), to find the two
energies.
This expression becomes more transparent in two cases. For a
Answer: E + = −10.8 eV, E − = −7.1 eV
homonuclear diatomic molecule we can set αA = αB = α and obtain
1/2
⎧ ⎫
(2 β −2αS ) 2
⎪ ⎪
2α − 2 βS ± ⎨(2α − 2 βS) − 4(1− S 2 )(α 2 − β 2 )⎬
2
(b) The features of the solutions
⎪⎩ ⎪⎭
E± = An important feature of eqn 25.8c is that as the energy dif-
(1− S
2

2
)
(1+ S )(1−S )
ference |αA − αB| between the interacting atomic orbitals
α − βS ± (β − αS) (α ± β )(1∓ S) increases, the bonding and antibonding effects decrease (Fig.
= = 25.4). Thus, when |αA − αB| 2|β| we can make the approxima-
(1+ S)(1− S) (1+ S)(1− S)
tion (1+ x )1/2 ≈ 1+ 12 x and obtain from eqn 25.8c
and therefore
β2 β2
α +β α −β Homonuclear E+ ≈ α A + E− ≈ α B − (25.9)
E+ = E− = diatomic (25.8b) α A −α B α A −α B
1+ S 1− S molecules
As these expressions show, and as can be seen from the graph,
For β < 0, E+ is the lower-energy solution. For heteronuclear when the energy difference is very large, the energies of the
diatomic molecules we can make the approximation that S = 0 resulting molecular orbitals differ only slightly from those of
(simply to get a more transparent expression), and find the atomic orbitals, which implies in turn that the bonding and
antibonding effects are small. That is:
E± = 12 (α A + α B )
1/2 Zero overlap The strongest bonding and antibonding effects
⎪⎧ ⎛ 2 β ⎞ ⎪⎫
2 (25.8c) Orbital
approximation are obtained when the two contributing
± (α A − α B ) ⎨1+ ⎜
1
⎬
⎝ α A − α B ⎟⎠ ⎪
2
contribution
⎩⎪ ⎭ orbitals have closely similar energies. criterion
www.ebook3000.com
Atkins09819.indb 228 9/11/2013 11:33:42 AM

6 E = E± given in eqn 25.8a. As before, this expression becomes

more transparent in two cases. First, for a homonuclear dia-
4
tomic molecule, with αA = αB = α and E± given in eqn 25.8b we
E/|αA + αB| – 1/2
2 E– find
α +β 1
0 E+ = cA = cB = c A (25.13a)
1+ S {2(1+ S)}1/2
E+ Homonuclear
–2
diatomic
molecules
–4 α −β 1
E− = cA = cB = −c A (25.13b)
1− S {2(1− S)}1/2
–6
0 2 4 6 8 10
|αA – αB|/|αA + αB|
For a heteronuclear diatomic molecule with S = 0, the coeffi-
Figure 25.4 The variation of the energies of the molecular cients for the orbital with energy E+ are given by
orbitals as the energy difference of the contributing atomic
1 1
orbitals is changed. The plots are for β = −1; the blue lines are cA = 1/2
cB = 1/2
for the energies in the absence of mixing (that is, β = 0). ⎧⎪ ⎛ α − E ⎞ 2
⎫⎪ ⎧⎪ ⎛ β ⎞ 2 ⎫⎪
+
⎨1+ ⎜ ⎨1+ ⎜
A
⎬ ⎬
⎝ β ⎟⎠ ⎝ α A − E+ ⎟⎠ ⎪
⎩⎪ ⎭⎪ ⎩⎪ ⎭
The difference in energy between core and valence orbitals is
the justification for neglecting the contribution of core orbitals Zero overlap approximation (25.14a)
to bonding. The core orbitals of one atom have a similar energy
to the core orbitals of the other atom; but core–core interaction and those for the energy E− are
is largely negligible because the overlap between them (and
hence the value of β) is so small. 1 −1
cA = 1/2
cB = 1/2
The values of the coefficients in the linear combination in ⎧⎪ ⎛ α − E ⎞ 2 ⎫⎪ ⎧⎪ ⎛ β ⎞ 2 ⎪⎫
−
⎨1+ ⎜ ⎨1+ ⎜
A
eqn 25.5 are obtained by solving the secular equations using the ⎟ ⎬ ⎟ ⎬
two energies obtained from the secular determinant. The lower ⎪⎩ ⎝ β ⎠ ⎪⎭ ⎪⎩ ⎝ α A − E− ⎠ ⎪⎭
energy, E+, gives the coefficients for the bonding molecular Zero overlap approximation (25.14b)
orbital, and the upper energy, E−, the coefficients for the anti-
bonding molecular orbital. The secular equations give expres- with the values of E± taken from eqn 25.8c.
sions for the ratio of the coefficients. Thus, the first of the two
secular equations in eqn 25.5a, (αA – E)cA+(β – ES)cB = 0, gives
Brief illustration 25.4 Heteronuclear diatomic molecules 3
⎛α −E⎞
cB = − ⎜ A c
⎝ β − ES ⎟⎠ A
(25.10) Here we continue Brief illustration 25.3 using HF. With
αH = −7.2 eV, αF = −10.4 eV, β = −1.0 eV, and S = 0 the two orbital
energies were found to be E + = −10.7 eV and E − = −6.9 eV. When
The wavefunction should also be normalized. This condition
these values are substituted into eqn 25.14 we find the follow-
means that the term cA2 + cB2 + 2cA cB S established in Justification ing coefficients:
25.2 must satisfy
E + = −10.7 eV ψ+ = 0.28χH + 0.96χF
cA2 + cB2 + 2cA cB S = 1 (25.11) E − = −6.9 eV ψ− = 0.96χH − 0.28χF
When the preceding relation is substituted into this expression, Notice that the lower-energy orbital (the one with energy
we find −10.7 eV) has a composition that is more F2p orbital than H1s,
and that the opposite is true of the higher-energy, antibonding
1 orbital.
cA =
⎧⎪ ⎛ α − E ⎞
1/2
Self-test 25.4 Find the energies and forms of the σ orbitals in
⎛ α − E ⎞ ⎫⎪
2
(25.12)
⎨1+ ⎜ − 2S ⎜ A the HCl molecule using β = −1.0 eV and S = 0. Use data from
A
⎟ ⎬
⎝ β − ES ⎠ ⎝ β − ES ⎟⎠ ⎪
⎩⎪ ⎭ Tables 20.2 and 20.3.
Answer: E + = −8.9 eV, E − = −6.6 eV; ψ− = 0.86χH − 0.51χ Cl;
which, together with eqn 25.10, gives explicit expressions for ψ + = 0.51χH+0.86χ Cl
the coefficients once we substitute the appropriate values of
Atkins09819.indb 229 9/11/2013 11:33:54 AM

☐ 1. A polar bond can be regarded as arising from a molec- ☐ 3. The variation principle provides a criterion of accept-
ular orbital that is concentrated more on one atom than ability of an approximate wavefunction.
on its partner. ☐ 4. A basis set refers to the given set of atomic orbitals from
☐ 2. The electronegativity of an element is a measure of the which the molecular orbitals are constructed.
power of an atom to attract electrons to itself when it is ☐ 5. The bonding and antibonding effects are strong-
part of a compound. est when contributing atomic orbitals have similar
energies.
Molecular orbital ψ = cAA+cBB 25.1
Pauling electronegativity | χ A − χ B | = {D0 (AB) − 12 [D0 (AA) + D0 (BB)]}1/2 25.2
Mulliken electronegativity χ = 12 (I + Eea ) 25.3
Coulomb integral
∫
dτ
α A = AHA 25.5c
∫ ∫
Resonance integral dτ = BHA
dτ
β = AHB 25.5d
Energy
∫
E = ψH ψ dτ
∫ψ 2 dτ Unnormalized real wavefunction 25.6
www.ebook3000.com
Atkins09819.indb 230 9/11/2013 11:33:58 AM

TOPIC 26
Polyatomic molecules
The molecular orbitals of polyatomic molecules are built in the
Contents same way as in diatomic molecules (Topics 24 and 25), the only
26.1 The Hückel approximation 232 difference being that more atomic orbitals are used to construct
(a) An introduction to the method 232 them. As for diatomic molecules, polyatomic molecular orbit-
Brief illustration 26.1: Ethene 232 als spread over the entire molecule. A molecular orbital has the
(b) The matrix formulation of the method 232 general form
Example 26.1: Finding molecular orbitals by matrix
26.2
diagonalization
Applications
233
234
ψ= ∑c χ
o
o o General form of LCAO (26.1)
(a) Butadiene and π-electron binding energy 234

Example 26.2: Estimating the delocalization where χo is an atomic orbital and the sum extends over all the
energy 235 valence orbitals of all the atoms in the molecule. To find the
(b) Benzene and aromatic stability 235 coefficients, we set up the secular equations and the secular
Example 26.3: Judging the aromatic character determinant, just as for diatomic molecules, solve the latter for
of a molecule 236 the energies, and then use these energies in the secular equa-
Checklist of concepts 237 tions to find the coefficients of the atomic orbitals for each
Checklist of equations 237 molecular orbital.
The principal difference between diatomic and polyatomic
molecules lies in the greater range of shapes that are possi-
ble: a diatomic molecule is necessarily linear, but a triatomic
molecule, for instance, may be either linear or angular (bent)
➤ Why do you need to know this material? with a characteristic bond angle. The shape of a polyatomic
Most molecules of interest in chemistry are polyatomic, so it molecule—the specification of its bond lengths and its bond
is important to be able to discuss their electronic structure. angles—can be predicted by calculating the total energy of
Although sophisticated computational procedures are the molecule for a variety of nuclear positions, and then
now widely available, to understand them it is helpful to identifying the conformation that corresponds to the low-
see how they emerged from the more primitive approach est energy. Such calculations are best done using the latest
described here. software, but a more primitive approach gives useful insight
for conjugated polyenes, in which there is an alternation of
➤ What is the key idea? single and double bonds along a chain of carbon atoms and
Molecular orbitals can be expressed as linear combinations on which we focus here, and sets the scene for more sophisti-
of all the atomic orbitals of the appropriate symmetry. cated approaches.
The planarity of conjugated polyenes is an aspect of their
➤ What do you need to know already? symmetry, and considerations of molecular symmetry play a
This Topic extends the approach used for heteronuclear vital role in setting up molecular orbitals. In the present case,
diatomic molecules in Topic 25, particularly the concepts planarity provides a distinction between the σ and π orbitals of
of secular determinants and secular equations. The the molecule, and in elementary approaches such molecules are
principal mathematical technique used is matrix algebra commonly discussed in terms of the characteristics of their π
(Mathematical background 5); you should be, or become, orbitals, with the σ bonds providing an unchanging underlying
familiar with the use of mathematical software to scaffolding, which forms a rigid framework that determines the
manipulate matrices numerically. general shape of the molecule.
Atkins09819.indb 231 9/11/2013 11:34:00 AM

2π α–β
26.1 The Hückel approximation
The π molecular orbital energy level diagrams of conjugated C2p C2p
molecules can be constructed using a set of approxima-
tions suggested by Erich Hückel in 1931. All the C atoms are
treated identically, so all the Coulomb integrals α (eqn 25.5c, 1π α+β
α = AHAd
A ∫
τ ) for the atomic orbitals that contribute to the
Figure 26.1 The Hückel molecular orbital energy levels of
π orbitals are set equal. For example, in ethene, which we use
ethene. Two electrons occupy the lower π orbital.
to introduce the method, we take the σ bonds as fixed, and
concentrate on finding the energies of the single π bond and its
These approximations convert eqn 26.3 to
companion antibond.
α −E β
(a) An introduction to the method = (α − E )2 − β 2 = (α − E + β )(α − E − β ) = 0
β α −E
We express the π orbitals as LCAOs of the C2p orbitals that lie
(26.4)
perpendicular to the molecular plane. In ethene, for instance,
we would write The roots of the equation are E± = α ± β. The + sign corresponds
to the bonding combination (β is negative) and the – sign cor-
ψ = c A A + cB B (26.2) responds to the antibonding combination (Fig. 26.1).
The building-up principle leads to the configuration 1π2,
where the A is a C2p orbital on atom A, and so on. Next, the because each carbon atom supplies one electron to the π system.
optimum coefficients and energies are found by the variation The highest occupied molecular orbital in ethene, its HOMO,
principle as explained in Topic 25. That is, we solve the secu- is the 1π orbital; the lowest unoccupied molecular orbital, its
lar determinant, which in the case of ethene is eqn 25.7 with LUMO, is the 2π orbital (or, as it is sometimes denoted, the 2π*
αA = αB = α: orbital). These two orbitals jointly form the frontier orbitals of
the molecule. The frontier orbitals are important because they
are largely responsible for many of the chemical and spectro-
α − E β − ES
=0 (26.3) scopic properties of the molecule.
β − ES α − E
Brief illustration 26.1 Ethene
∫ dτ )
and where β is the resonance integral (eqn 25.5d, β = AHB
We can estimate that the π* ← π excitation energy of ethene is
∫
and S is the overlap integral (eqn 24.3, S = AB dτ ). In a mod-
2|β|, the energy required to excite an electron from the 1π to
ern computation all the resonance integrals and overlap inte- the 2π orbital. This transition occurs at close to 40 000 cm−1,
grals would be included, but an indication of the molecular corresponding to 4.8 eV. It follows that a plausible value of β is
orbital energy level diagram can be obtained very readily if we about –2.4 eV (–230 kJ mol−1).
make the following additional Hückel approximations:
Self-test 26.1 The ionization energy of ethane is 10.5 eV.
r All overlap integrals are set equal to zero. Estimate α.
approximation
r All resonance integrals between non-neighbours Answer: –8.1 eV

are set equal to zero.
Hückel
r All remaining resonance integrals are set equal

(to β). (b) The matrix formulation of the method
These approximations are obviously very severe, but they let In preparation for making Hückel theory more sophisticated
us calculate at least a general picture of the molecular orbital and readily applicable to bigger molecules, we need to reformu-
energy levels with very little work. The assumptions result in late it in terms of matrices and vectors (see Mathematical back-
the following structure of the secular determinant: ground 5). Our starting point is the pair of secular equations
developed for a heteronuclear diatomic molecule in Topic 25:
r All diagonal elements: α – E.
r Off-diagonal elements between neighbouring atoms: β. (α A − E )cA + (β − ES)cB = 0
r All other elements: 0. (β − ES)cA + (α B − E )cB = 0
www.ebook3000.com
Atkins09819.indb 232 9/11/2013 11:34:09 AM

Atomic Structure and Spectra: Selection Rules

Uploaded by

Copyright:

Available Formats

Atomic Structure and Spectra: Selection Rules

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 and Spectra: Selection Rules

Uploaded by

Copyright:

Available Formats

182 5 Atomic structure and spectra

Visible the selection rules are derived by identifying the transitions

The principal quantum number n can change by any amount

directly to the angular momentum. The mathematical basis of

Example 21.1 Calculating the shortest and longest

Atkins09819.indb 182 9/11/2013 11:16:28 AM

82 259 L = l1 + l2 , l1 + l2 −1, …, l1 − l2 Clebsch–Gordan series (21.3)

Atkins09819.indb 183 9/11/2013 11:16:32 AM

When there are several electrons to be taken into account,

(b) The total spin angular momentum (a) S = 0 (b) S = 1

Atkins09819.indb 184 9/11/2013 11:16:39 AM

Atkins09819.indb 185 9/11/2013 11:16:41 AM

However, the relative orientation of the two momenta also

(d) Spin–orbit coupling H so = λL ⋅ S λ = hcA / 2 Spin–orbit coupling (21.6)

We now treat the J2, L2, and S2 as operators with eigenvalues

the splitting of a configuration into levels. W. H. Freeman & Co. (2014).

Atkins09819.indb 186 9/11/2013 11:16:48 AM

The strength of the spin–orbit coupling as measured by A

Atkins09819.indb 187 9/11/2013 11:16:57 AM

Self-test 21.8 Which of the transitions (a) 2 P 3/2 → 2 S1/2 ,

Atkins09819.indb 188 9/11/2013 11:16:58 AM

Rydberg formula ␯ = R H (1/n12 −1/n22 ) n2 = n1 + 1, n1 + 2, … 21.1

Atkins09819.indb 189 9/11/2013 11:17:00 AM

Focus 5 on Atomic structure and spectra

Topic 17 Hydrogenic atoms

Topic 18 Hydrogenic atomic orbitals

Atkins09819.indb 190 9/11/2013 11:17:03 AM

Topic 19 Many-electron atoms

‡ These problems were supplied by Charles Trapp and Carmen Giunta.

Atkins09819.indb 191 9/11/2013 11:17:08 AM

Topic 21 Atomic spectroscopy

Atkins09819.indb 192 9/11/2013 11:17:09 AM

Atkins09819.indb 193 9/11/2013 11:17:12 AM

Atkins09819.indb 194 9/11/2013 11:17:15 AM

Mathematical background 4 Vectors

θ = arccos( - 1/141/2 ) = 105.5° φ = arctan(3 / 2) = 56.3°

vx = v sin θ cos φ v y = v sin θ sin φ 2. Subtraction:

and therefore that

Atkins09819.indb 195 9/11/2013 11:17:30 AM

and is itself a scalar quantity. We can always choose a new

u ⋅ v = uv cos θ Scalar product (MB4.4d)

(b) The vector product, or cross product, of two vectors is

for the relation between the lengths of the sides of a triangle.

+ {(1)(2) − (−4)(−4)}k (a) (b) (c)

Atkins09819.indb 196 9/11/2013 11:17:44 AM

u u The derivatives of scalar and vector products are obtained using

Figure MB4.5 The direction of the cross products of two

Atkins09819.indb 197 9/11/2013 11:17:55 AM

Topic 22 Topic 23 Topic 24 Topic 25 Topic 26

The Homo- Hetero-

Focus 2 Focus 8 Focus 9 Focus 17

Atkins09819.indb 199 9/11/2013 11:17:59 AM

Atkins09819.indb 200 9/11/2013 11:17:59 AM

Atkins09819.indb 201 9/11/2013 11:17:59 AM

Figure 22.2 It is very difficult to represent valence-

ψ = A(1)B(2) + A(2)B(1) A valence-bond wavefunction (22.2)

Atkins09819.indb 202 9/11/2013 11:18:07 AM

is that the wavefunction in eqn 22.2 can be formed only by a pair

(d) Spin–orbit coupling H so = λL ⋅ S λ = hcA / 2 Spin–orbit coupling (21.6)

equation Hψ = Εψ with this hamiltonian and its analogues for