Focus 4 On: Approximation Methods

FOCUS 4 ON Approximation methods
Focus 5 Focus 9
Atomic
Topic 15 Topic 16 Molecular
structure and
Time- spectroscopy
spectra
independent
Transitions
perturbation
theory
Focus 6 Focus 8
Focus 2 Focus 3
The principles The quantum Molecular
Interactions
of quantum mechanics structure
mechanics of motion
Exact solutions of the Schrödinger equation can be found for only a small number of problems,
such as the model systems described in The quantum mechanics of motion. Almost all the problems
of interest in chemistry do not have exact solutions. To make progress with these problems, which
include many-electron atoms and molecules, we need to develop techniques of approximation.
There are three major approaches to finding approximate solutions. The first is to try to guess
the shape and mathematical form of the wavefunction. ‘Variation theory’ provides a criterion of suc-
cess with such an approach and, as it is most commonly encountered in the context of molecular
orbital theory, we consider it in Molecular structure. The second approach, an iterative method called
the ‘self-consistent field procedure’ which is often used alongside variation theory, is useful to find
numerical solutions of the Schrödinger equation for many-electron systems (see Atomic structure
and spectra and Molecular structure). The third approach takes the hamiltonian operator for the prob-
lem and separates it into a simple model hamiltonian for which the Schrödinger equation can be
solved and a ‘perturbation’, which is the difference between the true and model hamiltonians. The
aim of ‘perturbation theory’, which provides the mathematical tools for solving complex problems
by this approach, is to generate the wavefunction and energy of the perturbed system from knowl-
edge of the model hamiltonian and a systematic procedure for taking into account the presence of
the perturbation.
Time-independent perturbation theory (Topic 15) begins by identifying a model system that
resembles the system of interest, and then shows how to distort the wavefunctions and adjust the
energies to approach those of the actual system in the presence of a time-independent perturba-
tion. This analytical procedure is the basis for the many-body perturbation theory method described
in Molecular structure and is also useful in the discussion of electric and magnetic properties of mate-
rials (Interactions).
Time-dependent perturbation theory (Topic 16) is the basis of the discussion of transitions
between states when the system is subjected to a perturbation that varies with time. A very impor-
tant example of such a system is an atom or molecule exposed to an oscillating electromagnetic
field and is the basis of accounting for the transitions observed in Atomic structure and spectra and
Molecular spectroscopy.
Atkins09819.indb 131 9/11/2013 11:09:28 AM

TOPIC 15
Time-independent perturbation
theory
Contents the latter is then used to develop approximations to the

energies and wavefunctions of the system.
15.1 Perturbation expansions 132
Brief illustration 15.1: The corrections to the ➤ What do you need to know already?
energy and wavefunction 133
You need to be familiar with the concepts introduced in
15.2 The first-order correction
Topics 5–7, particularly the Schrödinger equation, expecta-
to the energy 134
tion values, hermiticity, normalization, and orthogonality.
Example 15.1: Evaluating the first-order
correction to the energy 134 The calculations in this Topic rely heavily on the description
of the particle in a one-dimensional box (Topic 9).
15.3 The first-order correction
to the wavefunction 135
Example 15.2: Evaluating the first-order
correction to the wavefunction 135 Perturbation theory comes in two formulations, depending on
15.4 The second-order correction to the whether or not the perturbation varies with time. In this Topic,
energy 136 we consider only time-independent perturbation theory,
Example 15.3: Evaluating the second-order which is the basis of our discussion of the electric and magnetic
correction to the energy 137 properties of molecules. In Topic 16, we discuss time-depend-
Checklist of concepts 137 ent perturbation theory, which is used to discuss the response
Checklist of equations 138 of atoms and molecules to time-dependent electromagnetic
fields and is central to the discussion of spectroscopy.
15.1 Perturbation expansions

➤ Why do you need to know this material? We suppose that the hamiltonian for the problem we are trying
Perturbation theory is used throughout chemistry to solve, H , can be expressed as a sum of a simple hamiltonian,
H , which has known eigenvalues, E(0), and eigenfunctions,
(0)
where exact solutions of the Schrödinger equation
ψ(0), and a contribution, H , which represents the extent to
(1)
are impossible to find. It is used to find approximate
solutions of complicated problems, such as the electric which the true hamiltonian differs from the ‘model’ hamiltonian:
and magnetic properties of matter (Topics 34 and 39) Partition
H = H + H Perturbation
(0) (1)
and the interaction between matter and electromagnetic theory
of the (15.1)
hamiltonian
radiation (Topic 16).
We seek the energies and eigenvalues of H . To find them we
➤ What is the key idea? suppose that the true energy of the system, E, differs from the
The hamiltonian of the system is expressed as a sum of a energy of the model system, E(0), and that we can write
simple model hamiltonian and a perturbation hamiltonian; Expansion
Perturbation
E = E (0) + E (1) + E (2) +… theory
of the (15.2)
energy
www.ebook3000.com
Atkins09819.indb 132 9/11/2013 11:09:33 AM

15 Time-independent perturbation theory 133
where E(1) is the ‘first-order’ correction to the energy, a contribu- This Topic is filled with derivations. If you do not require this
tion proportional to H (1) , and E(2) is the ‘second-order’ correc- level of detail and wish to proceed to the key results, they are as
tion to the energy, a contribution proportional to the square of follows for the ground state of the system.
H (1) , and so on. The true wavefunction, ψ, also differs from the
r The first-order correction to the energy is given by
‘simple’ wavefunction of the model system, ψ (0), and we write
E0(1) = ψ 0( )* H (1) ψ 0(0) dτ
∫
0
Expansion First-order energy correction (15.4)
Perturbation
ψ =ψ (0) +ψ (1) +ψ (2) +… theory
of the (15.3)
wavefunction
where ψ 0(0) is the ground-state wavefunction for the
where ψ (1) is the ‘first-order’ correction to the wavefunction and ‘model’ system with hamiltonian H (0).
so on. In practice, time-independent perturbation theory typi-
cally needs only first- and second-order energy corrections to See Example 15.1 for an illustration of how this expression
provide accurate estimates of energies of the perturbed system is used. The integral is an expectation value (Topic 7); in this
(as long as the perturbation is weak.) Likewise, it rarely needs case, it is the expectation value of the perturbation calculated
to proceed beyond first-order corrections to the wavefunction. using the unperturbed ground-state wavefunction. We can
therefore interpret E0(1) as the average value of the effect of the
Brief illustration 15.1 The corrections to the energy perturbation. An analogy is the shift in energy of vibration of
and wavefunction a violin string when small weights are hung along its length.
The weights hanging close to the nodes have little effect on its
Consider an electron in a one-dimensional metallic nano- energy of vibration. Those hanging at the locations of maxi-
particle of length L. A related simple problem is a particle in
mum amplitude, however, have a profound effect (Fig. 15.2a).
a one-dimensional box. The model hamiltonian H (0) is there-
The overall effect is the average of all the weights.
fore that of a particle in a one-dimensional box (Topic 9). The
effect of the varying potential energy of the electron inside r The wavefunction corrected to first order is given by
the nanoparticle can be modelled by supposing that it has the
form V (x ) = −ε sin(πx /L) (Fig. 15.1). The perturbation is there- ψ 0 =ψ 0(0) + ∑c ψ n
(0)
n
First-order
fore H (1) = −ε sin(πx /L). The corrections E(1) and ψ (1) are pro- n≠0
correction to the (15.5)
portional to ε and E(2) is proportional to ε2.
c =
n
∫ψ ( 0 )*
n H (1) ψ 0(0) dτ wavefunction
E0(0) − En(0)
∞ ∞
where ψ n(0) and En(0) are the eigenfunctions and
Potential energy, V
eigenvalues, respectively, of the model system.
See Example 15.2 for an illustration of how this expression is

used. In this case, the wavefunction is distorted by the pertur-
bation. In terms of the violin-string analogy, the weights distort
the shape of the vibrating string.
0 ε L
Location, x
Large effect
Figure 15.1 A model for the potential energy used to
illustrate the application of perturbation theory to a simple
system. The potential energy is infinite at the walls of the box Small effect
(a)
and varies as V across the floor of the box. No effect
Self-test 15.1 The vibrational motion of a certain diatomic

molecule is known to be described by a potential energy of the Perturbed
form V = 12 k f x 2 + ax3, where k f is the force constant, x is the dis- wavefunction
placement of the internuclear distance from the equilibrium (b)
bond length, and a is a constant. Identify the model hamilto-
nian and the perturbation and predict the dependencies of the Figure 15.2 (a) The first-order correction to the energy is an
corrections E(1), E(2), and ψ (1) on the constant a. average of the perturbation (represented by the hanging
Answer: H (0) is that of the harmonic oscillator (Topic 12); weights) over the unperturbed wavefunction. (b) The second-
H (1) = ax 3 ; E(1) and ψ (1) are proportional to a, E(2) to a 2 order correction is a similar average, but over the distortion
induced by the perturbation.
Atkins09819.indb 133 9/11/2013 11:09:47 AM

134 4 Approximation methods
r The second-order correction to the energy is given by By comparing powers of λ, we find

2
Terms in λ 0 : H (0) ψ 0(0) = E0(0)ψ 0(0)
∫ ψ ( 0 )*
n H (1) ψ (0) dτ
0 Second-
Terms in λ 1 : H (1) ψ 0(0) + H (0) ψ 0(1) = E0(0)ψ 0(1) + E0(1)ψ 0(0)
∑
order energy (15.6)
E0(2) = correction
E −E (0) (0)
n≠0 0 n
Terms in λ 2 : H (0) ψ 0(2) + H (1) ψ 0(1) = E0(2)ψ 0(0) + E0(1)ψ 0(1) + E0(0)ψ 0(2)
See Example 15.3 for an illustration of how this expression is
used. The correction also represents an average of the perturba- and so on. At this point, λ has served its purpose, and can now
tion similar to that for the first-order energy correction, but now be discarded.
it is an average over the perturbed wavefunctions. In terms of the The first of the three above equations is the Schrödinger
equation for the ground state of the unperturbed system,
violin analogy, the average is now taken over the distorted wave-
which we can assume we can solve. To solve the second of the
form of the vibrating string, in which the locations of minimum
equations (the one in λ1), we suppose that the first-order cor-
and maximum amplitudes are slightly shifted (Fig. 15.2b).
rection to the wavefunction can be expressed as a linear com-
bination of the wavefunctions of the unperturbed system, and
write
15.2The first-order correction
to the energy
ψ 0(1) = ∑c ψ
n
n
(0)
n
Although the equations of perturbation theory are applicable with the coefficients cn to be determined. We can isolate the
to any state of the perturbed system, we focus attention on the term in E0(1) by making use of the fact that the ψ n(0) form a
ground state for which the wavefunction of the ‘model’ system complete orthonormal set (eqn 7.3b of Topic 7) in the sense
is ψ 0(0) . We show in the following Justification that the first-order that
correction to the energy of the ground state is given by eqn 15.4. 1if n = 0 , 0if n ≠ 0

∫ ψ 0(0)*ψ n(0)dτ = δ 0n
Justification 15.1 The first-order correction to the energy where δij, the Kronecker delta (Topic 7), is 1 when i = j and
To develop expressions for the corrections to the ground-state 0 when i ≠ j. Therefore, when we multiply the equation in λ1
wavefunction and energy of a system subjected to a time- through by ψ 0(0)* and integrate over all space, we get
independent perturbation, we write E (0) if n = 0, 0 othe
erwise
0
ψ 0 = ψ 0(0) + λψ 0(1) + λ 2ψ 0(2) +
∫ψ (0 )* (1) (0 )
0 H ψ 0 dτ + ∑ ∫ cn ψ 0(0)* H (0) ψ n(0)dτ
where λ is a dummy variable that will help us keep track of δ0n
n
the order of the correction. At the end of the calculation, we

1

discard it. Likewise we write = ∑ n
cn E0(0) ∫ ψ 0(0)*ψ n(0)dτ + E0(1) ∫ψ 0 ψ 0 dτ
(0 )* (0 )
H = H (0) + λ H (1)
That is, because the middle two terms are equivalent,
and
E0 = E0(0) + λE0(1) + λ 2 E0(2) + ∫ψ (0 )* (0 ) (0 )

0 H ψ 0 dτ = E0(1)
When these expressions are inserted into the Schrödinger

which is eqn 15.4.
equation, H ψ = Eψ , we obtain
(H (0) + λ H (1) )(ψ 0(0) + λψ 0(1) + λ 2ψ 0(2) +) =

(E0(0) + λE0(1) + λ 2 E0(2) +)(ψ 0(0) + λψ 0(1) + λ 2ψ 0(2) +)
Example 15.1 Evaluating the first-order correction
which, by collecting powers of λ, we can rewrite as to the energy
H (0) ψ 0(0) + λ (H (1) ψ 0(0) + H (1) ψ 0(1) ) + λ 2 (H (0) ψ 0(2) + H (1) ψ 0(1) ) + Use the model described in Brief illustration 15.1 for the potential
energy of an electron in a one-dimensional nanoparticle to eval-
= E0(0)ψ 0(0) + λ (E0(0)ψ 0(1) + E0(1)ψ 0(0) ) + λ 2 (E0(2)ψ 0(0)
uate the first-order correction to the energy of the ground state.
+ E0(1)ψ 0(1) + E0(0)ψ 0(2) ) +
www.ebook3000.com
Atkins09819.indb 134 9/11/2013 11:10:07 AM

Method Identify the first-order perturbation hamiltonian and through by ψ k(0)* , where now k ≠0, and integrate over all space,
evaluate E0(1) from eqn 15.4. The ground-state wavefunction of we get
the particle in a one-dimensional box is given in Topic 9 (where (0)
it corresponds to n = 1). Use Integral T.3 listed in the Resource c k Ek

Ek if n = k , 0 otherrwise
(0)
section.
Answer The perturbation hamiltonian is H (1) = −ε sin(πx /L)

and the unperturbed ground-state (n = 1) wavefunction of the
∫ψ (0 )* (1) (0 )
k H ψ 0 dτ + ∑c ∫ψ
n
n
(0 )* (0 )
k H ψ n(0)dτ
(0)
particle in a one-dimensional box is ψ (0) = (2/L)1/2 sin(πx /L).
c
k E0

δ kn
Therefore, the first-order correction to the wavefunction is

0

4 π
L/3 = ∑c E ∫ψ(0)
n 0 k ψ n dτ
(0 )* (0 )
∫
+ E0(1) ψ k(0)*ψ 0(0)dτ
L
2ε L
π x 8ε n
0 ∫
E (1) = ψ (0) H (1) ψ (0)dx = −
L 0
sin3 ∫
L
dx = −
3 π
That is,
The energy is lowered by the perturbation, as would be
expected for the shape shown in Fig. 15.1.
∫ψ (0 )* (1) (0 )
k H ψ 0 dτ + ck Ek(0) = ck E0(0)
Self-test 15.2 Evaluate the first-order correction to the energy
of the ground state if, in the same model, V (x ) = −ε sin2 (πx /L). which we can rearrange into
Use Integral T.4 listed in the Resource section.
Answer: E (1) = − 34 ε
ck =
∫ψ (0 )* (1) (0 )
k H ψ 0 dτ
E0(0) − Ek(0)
Although this expression is for the coefficient c k, it applies to

15.3The first-order correction all coefficients: to get cn, simply change the index k wherever it
occurs to n. The result is eqn 15.5.
to the wavefunction
We show in the following Justification that the wavefunc-
tion corrected to first order in the perturbation is quoted in Example 15.2 Evaluating the first-order correction
eqn 15.5. The first-order correction to the wavefunction is a to the wavefunction
linear combination of the unperturbed wavefunctions of the
Once again, use the model described in Brief illustration 15.1,
system. Equation 15.5 shows that a particular state ψ n(0) this time to evaluate the contribution from the state with n = 3
r makes no contribution to the linear combination to the first-order correction to the wavefunction of the ground
interpretation
if ∫ψ n(0)* H (1) ψ 0(0) dτ = 0 state.

Physical
r makes a larger contribution the smaller the energy Method The first-order correction to the wavefunction is
difference |E0(0) − En(0)| (absolute values used because given by the sum in eqn 15.5. The ground state of the particle
En(0) is greater than the ground-state energy E0(0)). in the box (denoted ψ 0(0) in the equation) is ψ1; we seek the
contribution from ψ 3. The wavefunctions and energies of the
particle in a box are given in Topic 9. All the wavefunctions
are real, so the * can be ignored. Use Integral T.6 listed in the
Justification 15.2
The first-order correction Resource section.
to the wavefunction Answer The contribution from the n = 3 state to the first-order
To find the first-order correction to the wavefunction, we con- correction to the wavefunction is given by the coefficient
tinue the work of Justification 15.1 and seek the coefficients cn
of the first-order correction to the wavefunction:
c3 =
∫ψ 3H
(1) ψ dτ
1
E1 − E3
ψ 0(1) = ∑ n
cnψ n(0)
The energies of the particle (an electron) in a box are E n =
n 2h 2/8meL 2 , so the denominator is E1 − E3 = −h2 /me L2 . The
When we multiply the equation in λ1 (rewritten here)
numerator is
H (1) ψ 0(0) + H (0) ψ 0(1) = E0(0)ψ 0(1) + E0(1)ψ 0(0)
Atkins09819.indb 135 9/11/2013 11:10:27 AM

15.4The second-order correction

∫ ψ 3 H (1) ψ 1dτ
to the energy
ψ3 H (1) ψ
1
1/2 1/2
L
⎛ 2⎞ ⎛ 3πx ⎞ ⎧ ⎛ πx ⎞ ⎫ ⎛ 2 ⎞ ⎛ πx ⎞
∫
= ⎜ ⎟ sin ⎜ × ⎨−ε sin ⎜ ⎟ ⎬ ⎜ ⎟ sin ⎜ ⎟ dx
The second-order correction to the energy (eqn 15.6), which
0 ⎝ ⎠L ⎝ L ⎟
⎠ ⎩ ⎝ ⎠ ⎭⎝ ⎠
L L ⎝ L⎠ is derived in the following Justification, is rather more compli-
2ε L
⎛ 3πx ⎞ 2 ⎛ πx ⎞ cated, but we can note three important features:
=−
L ∫0
sin ⎜
⎝ L ⎟⎠
sin ⎜ ⎟ dx
⎝ L⎠ r Because En(0) > E0(0) , all the terms in the denominator are
Integral T.6

⎛ 2ε ⎞ ⎛ L ⎞ ⎧ 1 1 ⎫ negative, and because the numerators are all positive, E0(2)
=
⎩
1
{ 3
}
⎜⎝ − L ⎟⎠ × ⎜⎝ − 2π ⎟⎠ ⎨ 3 − 2(3 + 2) − 2(3 − 2) ⎬ (−1) − 1 is negative. That is, the second-order energy correction,
⎭
8ε but not necessarily the first-order correction, always
= lowers the energy of the ground state.
15π
interpretation
r The perturbation appears (as its square) in the
Therefore, the contribution from the n = 3 state is
Physical
numerator; so, the stronger the perturbation, the
8ε /15π 8εme L2 greater the lowering of the ground-state energy.
c3 = =−
−h /me L
2 2
15πh2 r If the energy levels of the system are widely
spaced, all the denominators are large, so the sum is
and the corrected ground-state wavefunction is likely to be small. In this case, the perturbation has little
1/2 1/2
effect on the second-order correction to the energy of the
⎛ 2⎞ ⎛ πx ⎞ 8εme L2 ⎛ 2 ⎞ ⎛ 3πx ⎞ system: the system is ‘stiff’, and unresponsive to
ψ 1 = ⎜ ⎟ sin ⎜ ⎟ − sin ⎜
⎝ L⎠ ⎝ L ⎠ 15πh2 ⎜⎝ L ⎟⎠ ⎝ L ⎟⎠ perturbations. The opposite is true when the energy
levels lie close together.
which, as Fig. 15.3 shows, corresponds to a greater accumula-
tion of amplitude in the middle of the well. Notice that there
are no contributions from even-n states because the integral
∫ ψ n H (1) ψ 1dτ vanishes unless n is odd. Justification 15.3 The second-order correction
to the energy
1.2
0.1 We continue the work on the previous two Justifications. To
0.05
Wavefunction, ψ1/(2/L)1/2
1 0 isolate the second-order energy correction E0(2) we multiply

both sides of the equation (with terms λ2):
0.8
0.6 H (0) ψ 0(2) + H (1) ψ 0(1) = E0(2)ψ 0(0) + E0(1)ψ 0(1) + E0(0)ψ 0(2)
0.4
by ψ 0(0)* and integrate over all space to obtain
0.2
0 ∫
E0(0) ψ 0(0)*ψ 0(2) dτ

0 0.2 0.4 0.6 0.8 1
∫
(0) (0) (2)
∫
ψ 0 *H ψ 0 dτ + ψ 0(0)*H (1) ψ 0(1)dτ
x/L
Figure 15.3 The ground-state wavefunction corrected to

1
∫ ∫ ∫
first order as evaluated in Example 15.2 for different values
= E0 ψ 0 * ψ 0 dτ + E0(1) ψ 0(0)* ψ 0(1)dτ + E0(0) ψ 0(0)* ψ 0(2)dτ
(2) (0) (0 )
of 8εmeL2/15πh2. Notice the progressive accumulation of
amplitude in the middle of the well as the perturbation
increases. In the first term, we have used the hermiticity (Topic 6) of H (0).
The first and last terms cancel, and we are left with
Self-test 15.3 Use the same model to evaluate the contribution
from the n = 5 state to the first-order correction to the wave- ∫ ∫
E0(2) = ψ 0(0)*H (1) ψ 0(1)dτ − E0(1) ψ 0(0)*ψ 0(1)dτ
function of the ground state.
Answer: c5 = −8εme L2 /315πh2 We have already found the first-order corrections to the energy
and wavefunction (eqns 15.4 and 15.5), so this expression
www.ebook3000.com
Atkins09819.indb 136 9/11/2013 11:10:44 AM

could be regarded as an explicit expression for the second- Method The second-order correction to the wavefunction is
order energy correction. However, we can go one step further given by eqn 15.6. The ground state of the particle in the box
by substituting ψ 0(1) = cnψ n(0) : ∑ n
(denoted ψ 0(0) in the equation) is ψ1; we seek the contribu-
tion from ψ 3. All the integrals needed have been evaluated in
(1) Example 15.2.
c 0 E0

δ0n
Answer Using the results in Example 15.2 gives
E0(2) = ∑∫ (0 )* (1) (0 )
∑
cn ψ 0 H ψ n dτ − cn E0 ψ 0 ψ n dτ
(1)
∫
(0 )* (0 )
2
∫ ψ 3 H (1) ψ 1dτ
n n
(8ε /15π)2 64ε 2me L2

The final term cancels the n = 0 term in the sum, and we are = =−
E1 − E3 −h /me L
2 2
225π2h2
left with
which corresponds to a lowering of the energy due to the per-
E0(2) = ∑c ∫
n≠0
n ψ 0(0)* H (1) ψ n(0)dτ turbation. We can now collect the first and second-order cor-
rections to the energy and write the perturbed energy of the
ground state as
Substitution of the expression for c n (see Justification 15.2)
produces the final result, eqn 15.6. h2 8ε 64ε 2me L2
E0 = − −
8me L 3π 225π 2h2
2
Self-test 15.4 Evaluate the contribution from the n = 5 state to

Example 15.3 the second-order correction to the energy of the ground state
Evaluating the second-order correction
in this model.
to the energy
Answer: −64ε 2me L2 /33075π2h2
Continue to use the model in previous Examples and evaluate
the contribution from the state with n = 3 to the second-order
correction to the energy of the ground state.
Checklist of concepts
☐ 1. In perturbation theory, the hamiltonian operator for the smaller its energy difference with the unperturbed
the problem is separated into a simple model hamilto- ground-state energy.
nian, for which the Schrödinger equation can be solved ☐ 5. The second-order correction to the energy is an average
exactly, and a ‘perturbation’, which is the difference of the perturbation over the perturbed wavefunctions.
between the true and model hamiltonians. ☐ 6. The second-order energy correction always lowers the
☐ 2. In time-independent perturbation theory, the pertur- energy of the ground state.
bation does not vary with time. ☐ 7. The stronger the perturbation and the smaller the spac-
☐ 3. The first-order correction to the energy can be ings of the energy levels in the unperturbed system, the
interpreted as the average value of the effect of the greater the lowering of the ground-state energy.
perturbation.
☐ 4. An unperturbed state makes a larger contribution to the
first-order correction to the ground-state wavefunction
Atkins09819.indb 137 9/11/2013 11:10:52 AM

Checklist of equations
Property Equation Comment Equation number
Division of the hamiltonian H = H (0) + H (1) H (1) is the perturbation 15.1
Expansion of the energy E = E (0) + E (1) + E (2) +… E(N) is the Nth-order energy correction 15.2
Expansion of the wavefunction ψ =ψ (0) +ψ (1) +ψ (2) +… ψ(N) is the Nth-order correction to the 15.3
wavefunction
First-order energy correction

∫
E0(1) = ψ 0(0)* H (1)ψ 0(0)dτ Ground-state energy 15.4
Wavefunction corrected to first order ψ 0 =ψ 0(0) +

∑c ψ
n≠0
(0 )
n n Ground state 15.5
⎛ ⎞
∫ (
cn = ⎜ ψ n(0)* H (1) ψ 0(0)dτ ⎟ / E0(0) − En(0)
⎜ ⎟
)
⎝ ⎠
15.6
2
Second-order energy correction E0(2) =
∑ ∫ψ (0) (1) (0)
n * H ψ 0 dτ (
/ E0(0) − En(0) ) Ground-state energy
n≠0
www.ebook3000.com
Atkins09819.indb 138 9/11/2013 11:10:58 AM

TOPIC 16
Transitions
When the potential energy that occurs in a hamiltonian is
Contents independent of time the wavefunctions are solutions of the
16.1 Time-dependent perturbation theory 140 time-independent Schrödinger equation (Topic 6). However,
(a) The general procedure 140 that doesn’t mean that they do not vary with time. In fact, if
Example 16.1: Using time-dependent perturbation the energy corresponding to a wavefunction ψ(r) is E, then the
theory for a slowly switched perturbation 140 time-dependent wavefunction is
(b) Oscillating perturbations and spectroscopic
transitions 142
Example 16.2: Analysing spectroscopic transitions ψ (r , t ) =ψ (r )e − iEt / Stationary-state wavefunction (16.1)
of a harmonic oscillator 142
(c) The energy of time-varying states 142
and, because e−iEt/ħ = e−i(E/ħ)t = cos (Et/ħ) − i sin (Et/ħ) (Euler’s
Brief illustration 16.1: Energy–time uncertainty 143
formula, Mathematical background 3), the time-depend-
16.2 The absorption and emission of radiation 143 ent wavefunction oscillates in time with a frequency E/ħ.
Brief illustration 16.2: The Einstein coefficients 144 Although the wavefunction oscillates, the physical observable
Checklist of concepts 145 it represents, the probability density, does not, for
Checklist of equations 145
Ψ (r , t )*Ψ (r , t ) = (ψ (r )* eiEt / )(ψ (r )e − iEt / ) =ψ (r )*ψ (r ).
➤ Why do you need to know this material?

One of the most important ways of exploring the which is independent of time. It is for this reason that Ψ(r,t)
structure of matter is spectroscopy, in which the radiation describes a stationary state, a state with a probability density
absorbed, emitted, or scattered by molecules is analysed. that persists unchanged in time.
To understand the processes involved, we need to know Matters are very different when the potential energy varies
how wavefunctions change when molecules are exposed with time, either because an external influence, a perturbation,
to perturbations of various kinds, such as an oscillating is switched on or because it is always present but changing with
electromagnetic field or the impact of a collision. time, as is the case when an atom or molecule is exposed to an
electromagnetic field. The states are no longer stationary, and a
➤ What is the key idea? molecule in one state might be forced into another state. That
A selection rule results from analysis of the expression for is, transitions, changes of state, occur in the presence of time-
the transition rate, which is proportional to the square of dependent perturbations.
the strength of the perturbation. Time-dependent perturbation theory provides a tech-
nique for analysing transitions between states, allowing us to
➤ What do you need to know already? identify the conditions under which transitions can occur and
You need to be familiar with the general principles of to develop expressions for their rates. Consequently, our dis-
quantum mechanics (Topics 5–7) and the Schrödinger cussion begins with time-dependent perturbation theory and
equation, as well as the basic concepts of perturbation then describes one of its most important applications, an initial
theory (Topic 15). An example draws on information about exploration of the absorption and emission of electromagnetic
the harmonic oscillator (Topic 12). radiation. This material is fundamental to the detailed account
of atomic and molecular spectra.
Atkins09819.indb 139 9/11/2013 11:11:01 AM

Time-dependent perturbation
16.1 The challenge is to find explicit expressions for the coef-
ficients. We do not give the details here,1 but it turns out that
theory if we confine our attention to the first-order effect of the per-
turbation, then the coefficients of all the wavefunctions other
To predict the outcome of the presence of a time-dependent than the one occupied initially, which we take to be ψ 0(0) (r ) (for
perturbation we have to solve the time-dependent Schrödinger instance, the 1s orbital in the previous example), is given by
equation:
First-order Coefficient
time-
(t )Ψ (r , t ) = i ∂Ψ (r , t )
1 t
∫H
of initially
H
Time-dependent
(16.2) cn (t ) = (1)
(t )e iω n 0 t
dt dependent unoccupied (16.4a)
∂t Schrödinger equation i 0
n0
perturbation state n
theory
where
(a) The general procedure
As in Topic 15 we use a perturbation procedure to find approxi-
∫
H n(10) (t ) = ψ n(0)* H (1) (t )ψ 0(0) dτ (16.4b)
mate solutions, and begin by supposing that the hamiltonian We have supposed that the perturbation is applied at t = 0 and
can be partitioned into a simple, solvable part and a perturba- have written ω n0 = (En(0) − E0(0) )/ . All it is necessary to do, there-
tion, which now is taken to be time-dependent: fore, is to introduce the form of the perturbation and evaluate
these integrals, one over space (in eqn 16.4b) and the other over
H = H (0) + H (1) (t ) (16.3a) time (after substituting the outcome into eqn 16.4a).
A perturbation that is suddenly applied is like an impact
For instance, when a molecule is exposed to an electromagnetic and can be expected to knock the system into a variety of
field of frequency ω and strength E parallel to the z-direction, states; one that is applied slowly and grows to its final value
the perturbation is produces a less violent effect and, well after it has been
Time- switched on, gives results equivalent to those obtained from
Oscillating
H (1) (t ) = −μzEcos ω t electric field
dependent
(16.3b) time-independent perturbation theory where the pertur-
perturbation bation is always present. Even oscillating perturbations are
often best treated as being switched on slowly, for a sudden
where μ z is the z-component of the dipole moment operator application will introduce transient responses that are hard
μ and is proportional to z. The underlying classical idea is that to treat mathematically.
for an atom or molecule to be able to interact with the electro-
magnetic field and absorb or create a photon of a specific fre-
quency, it must possess, at least transiently, a dipole oscillating Example 16.1 Using time-dependent perturbation
at that frequency.
theory for a slowly switched perturbation
As in the case of time-independent perturbation theory,
we suppose that the solutions of eqn 16.2 can be expressed as To e x p l o r e t h e s l o w l y s w i t c h e d c a s e , w r i t e
linear combinations of the eigenfunctions of the unperturbed (1) (t ) = H
H (1) (1− e −t /τ 0 ) where the time constant τ0 is very long
hamiltonian, the only difference being that the coefficients (Fig. 16.1). Show that at times t >> τ0 (at which the perturbation
has reached the constant value H (1)), the square modulus of
are allowed to vary with time. For instance, although initially
an electron in a hydrogenic atom might be in a 1s orbital, as the coefficient cn(t) is the same as we obtain from time-inde-
time goes on the probability increases that it will be found in pendent perturbation theory (Topic 15).
a 2p orbital, so the coefficient of that orbital grows with time Method Use eqn 16.4 to evaluate cn(t) and compare its square
while the coefficient of the 1s orbital decreases. In general, modulus to that of eqn 15.5 from time-independent perturba-
we write tion theory. We use τ0 for the time constant to avoid confusion
with the τ in the volume element, dτ.
Ψ (r , t ) = ∑c (t )Ψ (r , t ) = ∑c (t )ψ (r )e − iE
n t /
(0)
(0) (0)
n n n n
n n
where, as usual, the ψ n(0) (r ) are the eigenfunctions of H (0) and 1 For a full account of the solution, see our Molecular quantum mechan-
the En(0) are their energies. ics, Oxford University Press (2011).
www.ebook3000.com
Atkins09819.indb 140 9/11/2013 11:11:15 AM

16 Transitions 141
^ eiωn 0t (1)
H(1) cn (t ) = − H
ω n0 n0
τ0 short
Now we recognize that ω n0 = En(0) − E0(0) , which gives
eiωn 0t H n(10)
H(1)(t)
cn (t ) = −
En(0) − E0(0)
^
τ0 long When we form the square modulus of c n , the factor eiωn 0t

disappears (its square modulus is 1) and we obtain a
result identical to the square modulus of the coefficient in
eqn 15.5.
Time, t
Self-test 16.1 What is the form of cn(t) if a system is exposed
Figure 16.1 The time dependence of a slowly switched to t he osci l lat ing per turbat ion H (1) (t ) = 2H (1) cos ω t =
perturbation. A large value of τ 0 corresponds to very slow H (1) (eiω t + e − iω t )?
switching. Answer: (1/ i )H n(10) {(ei (ωn 0 +ω )t − 1)/[i(ω n0 + ω )]
(1) (t ) = H (1) (1 − e −t /τ 0 ) into eqn 16.4b +(ei(ωn 0 −ω )t −1)/[i(ω n0 −ω)]}

Answer Substitution of H
gives
The probability that at time t the perturbation has induced a

∫ (1) (1 − e −t /τ 0 )ψ (0)dτ
H n(10) (t ) = ψ n(0)* H 0
(1)
transition from the initial state ψ 0(0) to the state ψ n(0) is
H n0
2
(0 )* (1) (0 ) Pn (t ) = cn (t )
∫
= (1 − e ) ψ n H ψ 0 dτ
− t /τ 0 Transition probability (16.5)
= (1 − e −t /τ 0 )H n(10) The transition rate, w, is the rate of change of the probability

of being in the final state ψ n(0) due to transitions from the initial
Now substitute this expression into eqn 16.4a, recognizing state ψ 0(0) :
(1)
that H n0 is independent of time:
dPn
wn (t ) = Transition rate (16.6)
H n(10) t dt
cn (t ) =
i ∫ (1− e )e
0
− t /τ 0 iω n 0 t
dt
We can now develop an important relation between the tran-
H (1) t
= n0
i ∫ (e − e (
0
iω n 0 t − 1/τ 0 −iω n 0 )t
)dt sition rate and the strength of the perturbation. We consider
time-dependent perturbations of the form
Now use
H (1) (t ) = H (1)f (t )
t
eiωn 0t −1 t
e − (1/τ 0 −iωn 0 )t −1
∫ 0
eiωn 0t dt =
iω n0
and ∫ 0
e − (1/τ 0 −iωn 0 )t dt = −
1/τ 0 − iω n0
where all time dependence is contained in f(t), as we illus-
trated in a specific case in Example 16.1 and we have also
seen for the perturbation of eqn 16.3b, H (1) = −μ z E and f(t) =
At this point, we suppose that the perturbation is switched
cos ωt. Because the probability, Pn, is proportional to the square
slowly, in the sense that τ 0 >> 1/ω n0 so that the (blue) 1/τ 0
modulus of the coefficient, cn, of the state (eqn 16.5) and the
in the second denominator can be ignored. Note that the
assumption τ 0 >> 1/ω n0 provides a quantitative criterion of coefficient is itself proportional to the magnitude of the per-
‘slow’. We also suppose that we are interested in the coeffi- turbation (eqn 16.4), we conclude that the transition rate, a
cients long after the perturbation has settled down into its time derivative, is proportional to the square modulus of the
final value, when t >> τ 0, so that the (blue) exponential in the strength of the perturbation:
second numerator is close to zero and can be ignored. Under Oscill-
2
∫
wn (t ) ∝ H n(10) = ψ n(0)* H (1)ψ 0(0) dτ
2 Transition
these conditions, ating (16.7)
electric rate
field
Atkins09819.indb 141 9/11/2013 11:11:38 AM

(b)Oscillating perturbations and rule specifies the general features a molecule must have if it is
spectroscopic transitions to have a spectrum of a given kind. Selection rules, and oth-
ers like it for other types of transition, are explained in relevant
When the perturbation has the form given in eqn 16.3b, the Topics (see, for example, Topics 42–44).
transition rate is of the form
2 Example 16.2 Analysing spectroscopic transitions

∫
wn (t ) ∝ E 2 ψ n(0)* μzψ 0(0) dτ Transition rate (16.8a) of a harmonic oscillator
Suppose a hydrogen atom adsorbed on the surface of a gold
which includes a factor of the form nanoparticle undergoes harmonic motion perpendicular to
the surface. Can the absorption of electromagnetic radiation
∫ result in a transition from the ground vibrational state of the

Transition dipole moment
μz ,n0 = ψ n(0)* μ zψ 0(0) dz (z-component) (16.8b)
oscillator to the first excited vibrational state?
Method The harmonic oscillator wavefunctions are given in
Such an integral is called a transition dipole moment. It is a
Topic 12; they depend on the displacement, z, of the oscillator
measure of the electric dipole moment associated with the
from its equilibrium position. Consider the component of the
migration of charge from the initial state to the final state.
electromagnetic radiation along the axis perpendicular to the
The size of the transition dipole can be regarded as a measure
surface (that is, along z). Proceed to determine if the transition
of the charge redistribution that accompanies a transition: a dipole moment ∫ψ 1(0)* μzψ 0(0)dz is nonzero; μ z ∝z. Use Integral
transition is active (and generates or absorbs photons) only if G.3 listed in the Resource section.
the accompanying charge redistribution is dipolar (Fig. 16.2).
Clearly, if there is no such dipole moment, then wn = 0 and no Answer The ground-state and first excited-state harmonic
transition occurs. If the transition dipole moment is not zero, oscillator wavefunctions are the real functions
then wn ≠ 0 and we infer that a transition to n has occurred. A
transition is forbidden if the transition dipole moment is zero; ⎛ 2⎞
ψ 0(0) = N 0e − z /2α 2
ψ 1(0) = N1 ⎜ ⎟ ze − z /2α
2 2 2
it is allowed if any one of the components of the transition ⎝α ⎠

dipole moment (there are, in the most general case, also com-
ponents μx,n0 and μy,n0) is nonzero. where N 0 and N1 are normalization constants and α =
To summarize the discussion so far, we interpret eqn 16.8 as (ħ 2/μk f )1/4 with μ and k f the effective mass and force constant,
respectively, of the oscillator. The transition dipole moment is
follows: the intensity of absorption or emission of radiation in a
therefore
transition, which is proportional to the transition rate, is pro-
portional to the square modulus of the transition dipole moment. ∞ ∞
2 N 0 N1
μz ,10 ∝ ∫ ψ 1(0)zψ 0(0)dz = ∫ z 2e − z /α 2
2
A detailed study of the transition dipole moment leads to the dz
−∞ α −∞
specific selection rules that express the allowed transitions in Integral G.3
4 N 0 N1 ∞
terms of the changes in quantum numbers. A gross selection =

α ∫ 0
2 − z 2 /α 2
ze dz = N 0 N1α 2 π1/2
Because the integral is not zero, the spectroscopic transition

is allowed.
Self-test 16.2 Can the absorption of electromagnetic radia-
tion result in a transition from the ground state of the har-
monic oscillator to the second excited vibrational state? (Hint:
Consider the symmetry of the integrand in the transition
dipole moment.)
(0 ) (0 )
Answer: No; the integrand ψ 2 zψ 0 is an odd function of z
(a) (b)
Figure 16.2 (a) When a 1s electron becomes a 2s electron,

there is a spherical migration of charge. There is no dipole (c) The energy of time-varying states
moment associated with this migration of charge, so this
transition is electric-dipole forbidden. (b) In contrast, when a The energy of a time-varying state (not a stationary state) can-
1s electron becomes a 2p electron, there is a dipole associated not be specified exactly. To see why this is so, suppose that the
with the charge migration; this transition is allowed. wavefunction has the form
www.ebook3000.com
Atkins09819.indb 142 9/11/2013 11:11:45 AM

16 Transitions 143
Ψ (r , t ) =ψ (r )e − iEt / × e − t /2τ Brief illustration 16.1 Energy–time uncertainty

so that it oscillates and decays and the probability density Consider an excited electronic state which has a lifetime of
decays exponentially with a time constant τ. An exponen- 25 ns. The state corresponds to a spread of energies
tial function can be expressed as a superposition of sine and
1.055 ×10−34 Js
cosine functions, and therefore as a superposition of functions ΔE ≈ = = 4.2 ×10−27 J
τ 25 ×10−9 s
of the form e−iεt/ℏ with a wide range of values of ε. Therefore, the
decaying function has the form or, by dividing by hc, to the wavenumber 2.1 × 10−4 cm−1.
Self-test 16.3 Determine the lifetime of a short-lived
Ψ (r , t ) =ψ (r ) ∑c e ε
− i ( E + ε )t /
excited state that corresponds to a range of wavenumbers
ε 1.0 × 10−2 cm−1.
Answer: 530 ps
The probability that the energy E + ε is measured is propor-
tional to |cε|2, and it is no longer possible to say with certainty
that the energy of the decaying state is E. Indeed, the shorter
the time constant τ, the lifetime of the state, the wider the
spread of energies that must be included in the sum, as is illus- 16.2The absorption and emission
trated in Fig. 16.3, where we have shown how a decaying expo-
nential function with various decay rates can be expressed as of radiation
a superposition of cosine functions. The relation between the
lifetime and the spread of energies ΔE is summarized by the The rates of transitions between states induced by an elec-
expression tromagnetic field (or any other kind of perturbation) can be
calculated from eqns 16.4–16.6 by inserting the appropriate
perturbation. However, Einstein was able to carry out an anal-

ΔE ≈ Energy – time uncertainty (16.9) ysis that led to a surprising conclusion: transitions took place
τ
even in the apparent absence of a perturbation!
His argument went as follows. First, he recognized that the
This expression is reminiscent of the Heisenberg uncertainty transition from a low-energy state to one of higher energy that
principle (Topic 8), and consequently it is often referred to as is driven by the electromagnetic field oscillating at the transi-
‘the energy–time uncertainty principle’. However, its derivation frequency is called stimulated absorption. This transition
tion is quite different, and a better term is simply energy–time rate is proportional to E2 (eqn 16.8a) and therefore to the inten-
uncertainty. We see in Topic 40 that it plays an important role sity of the incident radiation. Therefore, the more intense the
in governing the widths of spectral lines. incident radiation, the stronger is the absorption by the sample
(Fig. 16.4). Einstein wrote this transition rate as
w f ←i = Bfi ρ Stimulated absorption Transition rate (16.10)

Amplitude of contribution
Probability density
The constant Bfi is the Einstein coefficient of stimulated

absorption and ρdν is the energy density of radiation in the
frequency range from ν to ν + dν, where ν is the frequency
of the transition. For instance, when the atom or molecule is
exposed to black-body radiation from a source of temperature
T, ρ is given by the Planck distribution (Topic 4):
(a) Time (b) Frequency 8πh␯3 /c 3

ρ= Planck distribution (16.11)
Figure 16.3 (a) Two wavefunctions that decay at different eh␯/kT −1
rates; the purple curve has a faster decay rate and shorter
lifetime than the blue curve. (b) The spread of frequencies (and At this stage Bfi as an empirical parameter that characterizes
therefore energies) that when superimposed re-creates the the transition: if it is large, then a given intensity of incident
decaying wavefunctions; the state with the shorter lifetime has radiation will induce transitions strongly and the sample will
a wider spread of contributions (purple). be strongly absorbing. The total rate of absorption, Wf←i, is the
Atkins09819.indb 143 9/11/2013 11:11:52 AM

N i Bfi ρ = N f ( A + Bif ρ ) Thermal equilibrium (16.16)
and therefore
Stimulated Stimulated Spontaneous
absorption emission emission
divide by Nf Bfi
Energy
Nf A A/Bfi A/Bfi
(B) (B) (A) ρ= = =
N i Bfi − N f Bif N i /N f − Bif /Bfi ehν /kT − Bif /Bfi
(16.17)
We have used the Boltzmann expression (Topic 2) for the ratio

of populations of the upper state (of energy Ef) and lower state
(of energy Ei):
Figure 16.4 The transitions analysed by Einstein in his theory h␯

of stimulated and spontaneous processes. Nf
= e − ( E −E )/kT
f i
Ni
transition rate of a single molecule multiplied by the number of
molecules Ni in the lower state: This result has the same form as the Planck distribution (eqn
16.11), which describes the radiation density at thermal equi-
Wf ←i = N i w f ←i = N i Bfi ρ Total absorption rate (16.12) librium. Indeed when we compare eqns 16.11 and 16.17, we
can conclude that Bif = Bfi and that
Einstein considered that the radiation was also able to
induce the molecule in the upper state to undergo a transition ⎛ 8πh␯3 ⎞
A=⎜ 3 ⎟ B (16.18)
to the lower state, and hence to generate a photon of frequency ⎝ c ⎠
ν. Thus, he wrote the rate of this stimulated emission as
The important point about eqn 16.18 is that it shows that
w f →i = Bif ρ Stimulated emission Transition rate (16.13) the relative importance of spontaneous emission increases
as the cube of the transition frequency and therefore that it
where Bif is the Einstein coefficient of stimulated emission. is therefore potentially of great importance at very high fre-
This coefficient is in fact equal to the coefficient of stimulated quencies. Conversely, spontaneous emission can be ignored at
absorption, as we shall see below. Moreover, only radiation of low transition frequencies, in which case intensities of those
the same frequency as the transition can stimulate an excited transitions can be discussed in terms of stimulated emission
state to fall to a lower state. and absorption.
At this point, it is tempting to suppose that the total rate
of emission is this individual rate multiplied by the num-
ber of molecules in the upper state, Nf, and therefore to write Brief illustration 16.2 The Einstein coefficients
Wf→i = NfBifρ. But here we encounter a problem: at equilibrium
(as in a black-body container), the rate of emission is equal to For a transition in the X-ray region of the electromagnetic
the rate of absorption, so NiBfiρ = NfBifρ and therefore, since spectrum (corresponding to an excitation of a core electron in
Bif = Bfi, Ni = Nf. The conclusion that the populations must be a molecule), a typical wavelength is 100 pm, corresponding to a
equal at equilibrium is in conflict with another very funda- frequency of 3.00 × 1018 s−1. The ratio of the Einstein coefficients
mental conclusion, that the ratio of populations is given by the of spontaneous and stimulated emission is
Boltzmann distribution (Topics 2 and 51).
A 8π × (6.626 × 10−34 Js) × (3.00 × 1018 s −1 )3
Einstein realized that to bring the analysis of transition rates into = = 1.67 × 10−2 kg m −1 s
B (2.998 × 108 ms −1 )3
alignment with the Boltzmann distribution there must be another
route for the upper state to decay into the lower state, and wrote
Self-test 16.4 Calculate the ratio of the Einstein coefficients
w f →i = A + Bif ρ Emission rate (16.14) of spontaneous and stimulated emission for a transition in
the microwave region of the electromagnetic spectrum with
The constant A is the Einstein coefficient of spontaneous emis-
wavelength 1.0 cm.
sion. The total rate of emission, Wf→i, is therefore
Answer: A/B = 1.7 × 10 −26 kg m−1 s
Wf →i = N f w f →i = N f ( A + Bif ρ ) Total emission rate (16.15)
At thermal equilibrium, Ni and Nf do not change over time. The presence of so-called spontaneous emission, in which,
This condition is reached when the total rate of emission and contrary to our earlier discussion, a transition occurs in the
absorption are equal: absence of a perturbation, might seem paradoxical. But in
www.ebook3000.com
Atkins09819.indb 144 9/11/2013 11:12:03 AM

16 Transitions 145
fact there is an unseen perturbation at work. The electromag- the radiation corresponds to the oscillator being in its ground
netic field can be modelled as a collection of harmonic oscil- state. But an oscillator has a zero-point energy (Topic 12), and
lators, one for each of the infinite range of frequencies that is never truly still. Therefore, even in the absence of observable
can be stimulated. This is the basis of the original derivation radiation, the electromagnetic fields are oscillating. It is these
of Planck’s expression for the density of states. Then, the pres- ‘unseen’ zero-point oscillations of the electromagnetic field
ence of radiation is equivalent to the excitation of oscillators of that act as a perturbation and drive the so-called ‘spontaneous’
the appropriate frequency, and the absence of that frequency in emission of radiation.
☐ 1. A stationary state corresponds to a wavefunction with a ☐ 7. The intensity of absorption or emission of radiation in a
probability density that does not change with time. transition is proportional to the square modulus of the
☐ 2. Transitions are changes in the state of the system; transition dipole moment.
they often occur in the presence of time-dependent ☐ 8. A specific selection rule expresses the allowed transi-
perturbations. tions in terms of the changes in quantum numbers.
☐ 3. In the classical picture, for an atom or molecule to be ☐ 9. A gross selection rule specifies the general features a
able to interact with the electromagnetic field and molecule must have if it is to have a spectrum of a given
absorb or create a photon of a specific frequency, it must kind.
possess, at least transiently, a dipole oscillating at that ☐ 10. The shorter the lifetime of a state, the wider is the
frequency. spread of energies to which the state corresponds.
☐ 4. The transition dipole moment is a measure of the elec- ☐ 11. A transition from a low-energy state to one of higher
tric dipole moment associated with the migration of energy that is driven by an oscillating electromagnetic
charge from the initial state to the final state. field is called stimulated absorption.
☐ 5. A transition is forbidden if the transition dipole ☐ 12. A transition driven from high energy to low energy is
moment is zero; it is allowed if any one of the compo- called stimulated emission.
nents of the transition dipole moment is nonzero. ☐ 13. The relative importance of spontaneous emission
☐ 6. The transition rate is proportional to the square modu- increases as the cube of the transition frequency.
lus of the strength of the perturbation.
Time-dependent Schrödinger equation H (t )Ψ (r ,t ) = i ∂Ψ (r ,t )/∂ t 16.2
Perturbation due to oscillating electric field H (1) (t ) = −μz E cosω t z is z-component of dipole moment
μ 16.3b
operator
t
Coefficient of initially unoccupied state n cn (t ) = (1/i )
∫H
0
(1)
n0 (t )e
iω n 0 t dt First-order time-dependent perturbation
theory
16.4
∫
H n(10) (t ) = ψ n(0)* H (1) (t )ψ 0(0)dτ
Transition probability Pn(t) = |cn(t)|2 16.5

Transition rate wn(t) = dPn/dt 16.6
Transition dipole moment

∫
μz ,n0 = ψ n(0)* μ˘ zψ 0(0)dz z-component 16.8b
Energy–time uncertainty ΔE ≈ ħ/τ τ is the lifetime of the state 16.9

Planck distribution ρ = (8πhν3/c3)/(ehν/kT−1) 16.11
Ratio of Einstein coefficients of spontaneous A/B = 8πhν3/c3 Einstein coefficients of stimulated 16.18
and stimulated emission emission and absorption are equal
Atkins09819.indb 145 9/11/2013 11:12:09 AM

Focus 4 on Approximation methods
Topic 15 Time-independent perturbation theory

Discussion questions
15.1 Identify the two different forms of perturbation theory and explain why 15.3 Provide a physical interpretation for the expression for the first-order
they are useful. correction in the wavefunction.
15.2 Describe what is meant by ‘first-order’ and ‘second-order’ corrections in 15.4 Provide a physical interpretation for the expression for the second-order
perturbation theory. correction in the energy.
Exercises
15.1(a) Calculate the first-order correction to the energy of an electron in 15.4(a) Does the vibrational frequency of an O–H bond depend on whether it
a one-dimensional nanoparticle modelled as a particle in a box when the is horizontal or vertical at the surface of the Earth? Suppose that a harmonic
perturbation is V(x) = −ε sin(2πx/L) and n = 1. oscillator of mass m is held vertically, so that it experiences a perturbation
15.1(b) Calculate the first-order correction to the energy of an electron in V(x) = mgx, where g is the acceleration of free fall. Calculate the first-order
a one-dimensional nanoparticle modelled as a particle in a box when the correction to the energy of the ground state.
perturbation is V(x) = −ε sin(3πx/L) and n = 1. 15.4(b) Repeat the previous exercise to find the change in excitation energy
from v = 1 to v = 2 in the presence of the perturbation.
15.2(a) Calculate the first-order correction to the energy of a particle in a box
when the perturbation is V(x) = −ε cos(2πx/L) and n = 1. 15.5(a) Evaluate the second-order correction to the energy of the harmonic
15.2(b) Calculate the first-order correction to the energy of a particle in a box oscillator for the perturbation described in Exercise 15.4(a). Hint: You will
when the perturbation is V(x) = −ε cos(3πx/L) and n = 1. find that there is only one term that contributes to the sum in eqn 15.6.
15.5(b) Evaluate the second-order correction to the energy of a particle
15.3(a) Suppose that the ‘floor’ of a one-dimensional box slopes up from x = 0
in a one-dimensional square well for the perturbation described in
to ε at x = L. Calculate the first-order effect on the energy of the state n = 1.
Exercise 15.2(a). Hint: The only term that contributes to the sum in
15.3(b) Suppose that the ‘floor’ of a one-dimensional box slopes up from x = 0
eqn 15.6 is n = 3.
to ε at x = L. Calculate the first-order effect on the energy of the state n = 2.
Problems
15.1 Suppose that the floor of a one-dimensional nanoparticle has an 15.2 We normally think of the one-dimensional well as being horizontal.
imperfection that can be represented by a small step in the potential Suppose it is vertical; then the potential energy of the particle depends on
energy, as in Fig. F4.1. (a) Write a general expression for the first-order x because of the presence of the gravitational field. Calculate the first-order
correction to the ground-state energy, E0(1). (b) Evaluate the energy correction to the zero-point energy, and evaluate it for an electron in a box
correction for a = L/10 (so the blip in the potential occupies the central on the surface of the Earth. Account for the result. Hint: The energy of the
10 per cent of the well), with n = 1. particle depends on its height as mgh, where g = 9.81 m s−2. Because g is so
small, the energy correction is small; but it would be significant if the box
were near a very massive star.
∞ ∞
15.3 Calculate the second-order correction to the energy for the system
described in Problem 15.2 and calculate the ground-state wavefunction.
Potential energy, V
Account for the shape of the distortion caused by the perturbation. Hint: The
integrals required are listed in the Resource section.
15.4 The vibrations of molecules are only approximately harmonic because
the energy of a bond is not exactly parabolic. Calculate the first-order
a correction to the energy of the ground state of a harmonic oscillator
subjected to an anharmonic potential of the form ax3 + bx4, where a and b
ε are small (anharmonicity) constants. Consider the three cases in which the
anharmonic perturbation is present: (a) during bond expansion (x ≥ 0)
0 L/2 L and compression (x ≤ 0); (b) during expansion only; (c) during
Location, x compression only.
Figure F4.1 The definition of the potential energy in

Problem 15.1.
www.ebook3000.com
Atkins09819.indb 146 9/11/2013 11:12:11 AM

Exercises and problems 147
Topic 16 Transitions
Discussion questions
16.1 What are likely to be the differences in outcome of the application of 16.3 Identify some perturbations that are likely to be encountered in the
sudden and slowly switched perturbations? discussion of the properties of molecules.
16.2 What is the physical interpretation of a selection rule?
Exercises
16.1(a) Calculate the ratio of the Einstein coefficients of spontaneous 16.1(b) Calculate the ratio of the Einstein coefficients of spontaneous
and stimulated emission, A and B, for transitions with the following and stimulated emission, A and B, for transitions with the following
characteristics: (a) 70.8 pm X-rays; (b) 500 nm visible light; (c) 3000 cm−1 characteristics: (a) 500 MHz radiofrequency radiation; (b) 3.0 cm microwave
infrared radiation. radiation.
Problem
16.1 The motion of a pendulum can be thought of as representing the
location of a wavepacket that migrates from one turning point to the other
periodically. Show that whatever superposition of harmonic oscillator states is
used to construct a wavepacket, it is localized at the same place at the times 0,
T, 2T, …, where T is the classical period of the oscillator.
Atkins09819.indb 147 9/11/2013 11:12:11 AM

this page left intentionally blank
www.ebook3000.com
FOCUS 5 ON Atomic structure and spectra
Topic 17 Topic 18 Topic 19 Topic 20 Topic 21
Hydrogenic Many-
Hydrogenic Atomic
atomic electron Periodicity
atoms spectroscopy
orbitals atoms
Focus 2 Focus 3 Focus 6 Focus 4

The principles The quantum
Molecular Approximation
of quantum mechanics
structure methods
mechanics of motion
Atoms are the currency of chemistry. They are the building blocks of all the forms of matter that
chemists consider and it is essential to understand their structure. To do so, we draw on The quantum
mechanics of motion, especially but not only rotational motion in three dimensions, and apply it to
progressively more complex atoms.
The simplest atom of all is a hydrogen atom with its single electron, and its generalization to ‘hydro-
genic atoms’ of general atomic number but still only one electron (Topic 17). When the Schrödinger
equation is solved for hydrogenic atoms we obtain the wavefunctions known as ‘atomic orbitals’ (Topic
18). These orbitals play a central role in the description of chemical bonding, as is explained in Molecular
structure, and, in combination with the ‘Pauli exclusion principle’, are the basis of the description of
the structures of many-electron atoms (Topic 19). With the structures of these atoms understood, the
structure of the periodic table and the ‘periodicity’ of the properties of the elements fall into place
(Topic 20).
The experimental investigation of the structure of atoms is largely through ‘atomic spectroscopy’
(Topic 21). The origin of the transitions that give rise to spectral lines is time-dependent perturbation
theory, which is described in Approximation methods. These transitions reveal details about the inter-
actions between electrons and the coupling of the various sources of angular momentum in an atom.
What is the impact of this material?

Atomic spectroscopy is widely used by astronomers to determine the chemical composition of stars
(Impact 5.1). Stellar material consists of neutral and ionized forms of atoms, such as hydrogen, helium,
carbon, and iron, and each element, and indeed each isotope of an element, has a characteristic
spectral signature that is transmitted through space by the star's light.
To read more about the impact of this material, scan the QR code, or go to
http://bcs.whfreeman.com/webpub/chemistry/qmc2e/impact/qchem_
impact5.html.
Atkins09819.indb 149 9/11/2013 11:12:13 AM

TOPIC 17
Hydrogenic atoms
Contents orbitals, that are defined by three quantum numbers, n, l,

and ml, and energies that depend only on n.
17.1 The structure of hydrogenic atoms 150
The separation of variables 151
(a)
➤ What do you need to know already?
Brief illustration 17.1: The angular part of the
hydrogenic wavefunction 153 You should be familiar with the form of the Schrödinger
(b) The radial solutions 153 equation (Topic 6), its solution by the separation of
variables technique (Mathematical background 2), and
17.2 The atomic orbitals and their energies 155
(a) The atomic orbitals 155
the quantum mechanical description of rotation in three
dimensions (Topic 14).
Brief illustration 17.2: The radial nodes of the
wavefunction 155
Brief illustration 17.3: The probability density for an
electron 156
(b) The energy levels 156 In this Topic we begin to describe the electronic structure of an
Example 17.1: Determining the energy for atom, the distribution of electrons around its nuclei. We need
an electron occupying an atomic orbital 157 to distinguish between two types of atoms. A hydrogenic atom
Example 17.2: Calculating the wavenumber of a is a one-electron atom or ion of general atomic number Z such
line in the emission spectrum of H 157
as H, He+, Li2+, O7+, and even U91+. A many-electron atom (or
(c) Ionization energies 158
polyelectronic atom) is an atom or ion with more than one elec-
Example 17.3: Measuring an ionization energy
tron, such as all neutral atoms other than H. This Topic, which
spectroscopically 158
focuses exclusively on hydrogenic atoms, provides a set of con-
Checklist of concepts 159
cepts that are used to describe the structure of many-electron
atoms (Topic 19).
17.1 The structure of hydrogenic atoms

To set the stage for your study of the structure of atoms To describe a hydrogenic atom using quantum theory,
and molecules, you need to know how the Schrödinger we need to write and solve the Schrödinger equation. The
equation is solved for the hydrogen atom. Moreover, potential-energy term in the Schrödinger equation arises from
the solutions for the hydrogen atom underlie the entire the Coulomb interaction between the electron and the nucleus
discussion of many-electron atoms, periodicity, and the of charge Ze (see Foundations, Topic 2):
formation of chemical bonds.
➤ What is the key idea? Ze 2
V (r ) = − Hydrogenic atom Potential energy (17.1)
Solution of the Schrödinger equation for the hydrogen
4πε 0 r
atom subject to the appropriate boundary conditions
results in one-electron wavefunctions, known as atomic where r is the distance of the electron from the nucleus
and ε 0 is the vacuum permittivity. Because this expression
www.ebook3000.com
Atkins09819.indb 150 9/11/2013 11:12:14 AM

17 Hydrogenic atoms 151
depends only on the distance from a single point (here, the z

nucleus), it is an example of a central potential. The hamil-
θ
tonian operator for the electron and a nucleus of mass mN is
φ
therefore
r
electronic nuclear potential
kinetic energy kinetic energy energy y

2
2
Ze 2 Hydro- Hamil-
H = − ∇2e − ∇2N − genic tonian (17.2) x
2me 2mN 4πε 0 r atom operator
The subscripts on ∇2 indicate differentiation with respect to

Figure 17.2 The coordinate system for describing the position
the electronic or nuclear coordinates (Fig. 17.1).
of the electron relative to the nucleus. The nucleus is placed
at the origin, and spherical polar coordinates are used for
describing the position of the electron relative to the nucleus.
(a) The separation of variables
As shown in the following Justification, the Schrödinger equa- Justification 17.1 The separation of internal and
tion for a hydrogenic atom can be separated into two equations,
external motion
one for the motion of the atom as a whole through space and
the other for the motion of the electron relative to the nucleus. We consider a one-dimensional system in which the poten-
The Schrödinger equation for the internal motion of the elec- tial energy depends only on the separation of the two par-
tron relative to the nucleus is ticles, and then generalize to three dimensions. The total
Internal classical energy is
2 2 Ze 2 Schrödinger
− ∇ψ− ψ = Eψ motion of
equation (17.3a)
2μ 4 πε 0 r electron p12 p2
E= + 2 +V
2m1 2m2
me mN
μ= Hydrogenic atom Reduced mass (17.3b) where p1 = m1x1 and p2 = m2 x 2 (where the dot indicates differ-
me + mN
entiation with respect to time) are the linear momenta of the
two particles. The centre of mass (Fig. 17.3) is located at
where differentiation is now with respect to the coordinates
of the electron relative to the nucleus (Fig. 17.2). The quantity m1 m
X= x + 2x
μ, which is called the reduced mass, is very similar in value m 1 m 2
to the electron mass because mN, the mass of the nucleus,
where m = m1 + m 2 . Because the separation of the particles is
is much larger than the mass of an electron, so μ ≈ me. In
x = x1 − x 2, it follows that
all except the most precise work, the reduced mass can be
replaced by me. m1 m
X= x + 2 (x − x )
m 1 m 1
giving
ze
m2 m
z x1 = X + x and x2 = x1 − x = X − 1 x
e m m
zN r
m2 m1
N
yN
xe ye x
y x2 x1
xN X
x
Figure 17.3 The coordinates used for discussing the separation
Figure 17.1 The coordinate system for describing the positions of the relative motion of two particles from the motion of the
of the electron and nucleus in a hydrogenic atom. centre of mass.
Atkins09819.indb 151 9/11/2013 11:12:25 AM

152 5 Atomic structure and spectra
The linear momenta of the particles in terms of the rates of where the Y are the spherical harmonics that occur for a particle
change of x and X are on a sphere and are labelled with the quantum numbers l and
mm ml that specify the angular momentum of the particle. The fac-
p1 = m1x1 = m1 X + 1 2 x tor R, the radial wavefunction, is a solution of the radial wave
m
m m equation
p2 = m2 x 2 = m2 X − 1 2 x
m
2 ⎛ 2 d 2 R dR ⎞
− r + 2r
Then it follows that 2 μR ⎜⎝ dr 2 dr ⎟⎠ Hydrogenic
Radial
wave
atom (17.5)
p12 p2 1 ⎛ m1m2
2
⎞ 1 ⎛ m1m2 ⎞
2
l(l + 1) 2 equation
+ 2 = m X+ x ⎟ + m X− x ⎟ + Vr 2 − Er 2 = −
2m1 2m2 2m1 ⎜⎝ 1 m ⎠ 2m2 ⎜⎝ 2 m ⎠ 2μ
1 1⎛mm ⎞
= mX 2 + ⎜ 1 2 ⎟ x 2
2 2⎝ m ⎠ Justification 17.2The solutions of the Schrödinger
By writing P = mX for the linear momentum of the system equation for a hydrogenic atom
as a whole and defining p as μx, where the reduced mass Substitution of the wavefunction ψ = RY into eqn 17.3 gives
μ = m1m2/m, we find
∇2 in spherical coordinates

P 2 p2 ⎛ ∂2 2 ∂ 1 2 ⎞
2
E= + +V − + + Λ RY + VRY = ERY
2m 2μ 2 μ ⎜⎝ ∂r 2 r ∂r r 2 ⎟⎠
The corresponding hamiltonian is therefore
Because R depends only on r and Y depends only on the angu-
P p 2
2
H = + +V lar coordinates, this equation becomes
2m 2μ
2 ⎛ d 2 R 2Y dR R 2 ⎞
− Y + + Λ Y ⎟ + VRY = ERY
which, generalized to three dimensions, is 2 μ ⎜⎝ dr 2 r dr r 2 ⎠
2 2 2 2
H = − ∇ cm − ∇ +V Multiplication through by r2/RY gives
2m 2μ
2 ⎛ 2 d 2 R dR ⎞ 2 2
− r + 2r + Vr 2 − Λ Y = Er 2
where the first term differentiates with respect to the centre of 2 μR ⎜⎝ dr 2 dr ⎟⎠ 2 μY
mass coordinates and the second with respect to the relative
coordinates. and therefore
Now we write the overall wavefunction as the product Depends on r alone on θ ,φ alone
ψtotal = ψcmψ, where ψcm is a function of only the centre of mass Depends
coordinates and ψ is a function of only the relative coordi- ⎛ 2d R
2 2
dR ⎞ 2
− r + 2r + (V − E )r =
2
Λ2Y
nates. The overall Schrödinger equation, H ψ total = Etotalψ total , 2 μR ⎜⎝ dr 2 dr ⎟⎠ 2μY
then separates by the argument presented in Mathematical
background 2 with Etotal = Ecm + E. At this point we employ the usual argument for the separation
of variables (Mathematical background 2), that each term on
either side of the equals sign must be a constant, and conclude
The centrosymmetric Coulomb potential energy has spheri- that the differential equation separates into two equations:
cal symmetry in the sense that at a given distance its value 2 2
is independent of angle. We can therefore suspect that the − Λ Y = constant
2 μY
Schrödinger equation will separate into an angular component
2 ⎛ 2 d 2 R dR ⎞
that mirrors the equation for a particle on a sphere treated in − r + 2r + (V − E )r 2 = − constant
2 μR ⎜⎝ dr 2 dr ⎟⎠
Topic 14 and, new to this system, an equation for the radial fac-
tor. That is, we can think of the electron as free to move around The first of these two equations is encountered in Topic 14,
the nucleus on a spherical surface with the additional freedom where it is noted that
to move between concentric surfaces of different radii. The fol-
lowing Justification confirms that the wavefunction can indeed Λ2Y = −l(l + 1)Y
be written as a product of functions
and that the solutions (subject to cyclic boundary conditions)
Hydrogenic Electronic are the spherical harmonics.
ψ (r ,θ , φ ) = R(r )Y (θ , φ ) atom wavefunction (17.4)
www.ebook3000.com
Atkins09819.indb 152 9/11/2013 11:12:42 AM

Brief illustration 17.1 The angular part of the hydrogenic 0

l≠0
Effective potential energy, Veff

wavefunction
Consider an electron in a hydrogen atom with the angu-
lar part of its wavefunction given by the spherical harmonic
Y2,−1(θ,φ). Because l = 2 and m l = −1, the magnitude of the
angular momentum is 61/2 and the z-component is − ; the l=0
negative sign indicates that, in classical terms, the electron
is circulating in an anticlockwise sense as seen from below.
Furthermore, because (Table 14.1) Y2,−1(θ,φ) ∝ cos θ sin θ,
there are angular nodes at θ = 0, π/2, π: the probability of find-
Radius, r
ing the electron anywhere on the z-axis and in the xy-plane is
therefore zero. Figure 17.4 The effective potential energy of an electron in a
Self-test 17.1 If the angular part of the electronic wavefunc- hydrogenic atom. When the electron has zero orbital angular
tion is given by Y1,+1(θ,φ), determine the magnitude and momentum, the effective potential energy is the Coulombic
z-component of the angular momentum as well as the loca- potential energy. When the electron has nonzero orbital
tions of any angular nodes. angular momentum, the centrifugal effect gives rise to a
Answer: 21/2 , + , θ = 0, π
positive contribution which is very large close to the nucleus.
The l = 0 and l ≠ 0 wavefunctions are therefore very different
near the nucleus.
proportional to 1/r, and the net result is an effective repulsion

(b) The radial solutions
of the electron from the nucleus. The two effective potential
The appearance of the radial wave equation (eqn 17.5) is sim- energies, the one for l = 0 and the one for l ≠ 0, are qualitatively
plified by writing R = u/r, with u a function of r, for it then very different close to the nucleus. However, they are similar
becomes (see Problem 17.4) at large distances because the centrifugal contribution tends to
zero more rapidly (as 1/r2) than the Coulombic contribution
Radial (as 1/r). Therefore, we can expect the solutions with l = 0 and
2 d 2 u Hydrogenic
− + Veff u = Eu atom
wave (17.6a) l ≈ 0 to be quite different near the nucleus but similar far away
2μ dr 2 equation
from it.
We show in the following Justification that:
where
r Close to the nucleus the radial wavefunction is
proportional to rl.
Ze 2 l(l + 1) 2 Hydrogenic
Effective
Veff = − + potential (17.6b) r Far from the nucleus all radial wavefunctions approach
4 πε 0 r 2 μr 2 atom
energy
zero exponentially.
Equation 17.6 describes the motion of a particle of mass μ in It follows that when l ≠ 0 (and rl = 0 when r = 0) there is a zero
a one-dimensional region 0 ≤ r < ∞ where the potential energy probability density for finding the electron at the nucleus, and
is Veff. the higher the orbital angular momentum, the less likely the
The first term in eqn 17.6b is the Coulomb potential energy electron is to be found near the nucleus (Fig. 17.5). However,
of the electron in the field of the nucleus. The second term, when l = 0 (and rl = 1 even when r = 0) there is a nonzero prob-
which depends on the angular momentum of the electron ability density of finding the electron at the nucleus. The con-
around the nucleus, stems from what in classical physics would trast in behaviour has profound implications for chemistry,
be called the ‘centrifugal effect’. When l = 0, the electron has no for it underlies the structure of the periodic table (Topic 20).
angular momentum, and the effective potential energy is purely The exponential decay of wavefunctions has a further impor-
Coulombic and attractive at all radii (Fig. 17.4). When l ≠ 0, tant implication: it means that atoms can, to a good approxi-
the centrifugal term gives a positive (repulsive) contribution mation, be represented by spheres with reasonably well
to the effective potential energy. When the electron is close to defined radii. This feature is especially important in the dis-
the nucleus (r ≈ 0), this repulsive term, which is proportional to cussion of solids (Topic 37), which are commonly modelled as
1/r2, dominates the attractive Coulombic component, which is aggregates of spheres representing their atoms and ions.
Atkins09819.indb 153 9/11/2013 11:12:45 AM

l=0 this equation has the form

2 d 2 R
− r ErR
Wavefunction, ψ
2μ dr 2
or, after cancelling the r,

1
2 2 d 2 R
3
− ER
2μ dr 2
Because R must be finite and the energy is negative for a

bound state, a state in which the energy of the electron is lower
Radius, r
than when the electron is infinitely distant and stationary, the
Figure 17.5 Close to the nucleus, orbitals with l = 1 are acceptable solution for r large is
proportional to r, orbitals with l = 2 are proportional to r2,
R e −(2 μ|E|/
2 1/2
) r
and orbitals with l = 3 are proportional to r3. Electrons are
progressively excluded from the neighbourhood of the as may be verified by substitution. That is, all the bound-
nucleus as l increases. An orbital with l = 0 has a finite, nonzero state wavefunctions decay exponentially towards zero as
value at the nucleus. r increases.
Justification 17.3The radial wavefunctions close to and We shall not go through the technical steps of solving the
far from the nucleus radial wave equation (eqn 17.6) for the full range of radii, and
see how the form rl close to the nucleus blends into the expo-
When r is very small (close to the nucleus), u = rR ≈ 0, so the
nentially decaying form at great distances.1 It is sufficient to
right-hand side of eqn 17.6a is zero. We can also ignore all but
know that:
the largest terms (those depending on 1/r2) in eqn 17.6b and
write r The two limits can be matched only for integral values of
d 2u l(l + 1) a new quantum number n.
− + 2 u≈0
dr 2 r r The allowed energies corresponding to the allowed
solutions are
The solution of this equation (for r ≈ 0) is
Z 2 μe 4 Hydro-
B Enlm = − n = 1, 2,… genic Energy (17.7)
u ≈ Ar l +1
+ l l
32π 2 ε 02 2n2 atom
r
as may be verified by substitution of the solution into the dif- r The allowed energies are independent of the values of l
ferential equation. Because R = u/r, and R cannot be infinite and ml; to avoid overburdening the notation, henceforth
anywhere and specifically at r = 0, we must set B = 0. Therefore we denote them simply En.
we obtain u ≈ Arl+1 or R ≈ Arl, as we wanted to show. r The radial wavefunctions depend on the values of both n
Far from the nucleus, when r is very large and we can ignore and l (but not on ml) with l restricted to values 0, 1, …,
all terms in 1/r and 1/r2, eqn 17.6a becomes n − 1.
r All radial wavefunctions have the form
2 d 2u
− Eu
2μ dr 2 R(r ) = r l × (polynomial in r ) × (decaying exponential in r )
where ⯝ means ‘asymptotically equal to’ (that is, the two sides
of the expression become equal as r → ∞). Because These functions are most simply written in terms of the dimen-
sionless quantity ρ (rho), where
d 2u d 2 (rR) d d d ⎛ dR ⎞
= = (rR) = ⎜ r + R⎟
dr 2 dr 2 dr dr dr ⎝ dr ⎠ 2Zr 4 πε 0 2
ρ= a= (17.8)
This factor is na μe 2
very large

d2R dR d2R
= r 2
+2 r 2
dr dr dr 1 For a full account of the solution, see our Molecular quantum mechan-
ics, Oxford University Press (2011).
www.ebook3000.com
Atkins09819.indb 154 9/11/2013 11:12:57 AM

For simplicity, and introducing negligible error, the μ in eqn Table 17.1 Hydrogenic radial wavefunctions
17.8 is often replaced by me, in which case ρ = 2Zr/na0 and
n l Rnl
1 0
4 πε 0 2 ⎛ Z⎞
3/2
a0 = (17.9)
2⎜ ⎟ e − ρ /2
me e 2 Definition Bohr radius ⎝ a⎠
2 0 3/2
The Bohr radius, a0, has the value 52.9 pm; it is so called 1 ⎛ Z⎞
81/2 ⎝⎜ a ⎠⎟
(2 − ρ ) e− ρ/2
because the same quantity appeared in Bohr’s early model of
the hydrogen atom as the radius of the electron orbit of low- 2 1 3/2
est energy. Specifically, the radial wavefunctions for an electron 1 ⎛ Z⎞
ρe − ρ /2
241/2 ⎜⎝ a ⎟⎠
with quantum numbers n and l are the (real) functions
3 0 3/2
Hydrogenic Radial 1 ⎛ Z⎞
Rnl (r ) = N nl ρ l L2nl++11 (ρ )e − ρ/2 (6 − 6 ρ + ρ 2 )e − ρ /2
2431/2 ⎜⎝ a ⎟⎠
atom wavefunctions (17.10)
where Nnl is a normalization constant (with a value that 3 1 3/2

1 ⎛ Z⎞
depends on n and l) and L is a polynomial in ρ called an associ- (4 − ρ )ρe − ρ /2
4861/2 ⎜⎝ a ⎟⎠
ated Laguerre polynomial. The polynomial L connects the r ≈ 0
3 2
solutions on its left (corresponding to R ∝ ρl) to the exponen- 1 ⎛ Z⎞
3/2
ρ 2 e − ρ /2
tially decaying function on its right. The notation might look 24301/2 ⎝⎜ a ⎠⎟
fearsome, but the polynomials have quite simple forms, such as
ρ = (2Z/na)r with a = 4πε 0 2/μe2. For an infinitely heavy nucleus (or one that may be
1, ρ, and 2 − ρ (they can be picked out in Table 17.1). assumed to be), μ = me and a = a0, the Bohr radius.
The components of eqn 17.10 can be interpreted as follows:
r The exponential factor ensures that the wavefunction

approaches zero far from the nucleus. Self-test 17.2 Find the radial nodes for an electron in Li 2+ ion
interpretation
with n = 3, l = 0, ml = 0.
r The factor ρl ensures that (provided l > 0) the
Answer: (3 ± 31/3)a0/2, 125 and 33.5 pm
Physical
wavefunction vanishes at the nucleus.

r The associated Laguerre polynomial is a function
that oscillates from positive to negative values and
accounts for the presence of radial nodes. The atomic orbitals and
17.2
Expressions for some radial wavefunctions are given in Table their energies
17.1 and illustrated in Fig. 17.6. Note that because r is never
negative, the zero in the radial wavefunctions at r = 0 (for The wavefunctions of hydrogenic atoms are called ‘atomic
l > 0) is not a node: the wavefunction does not pass through orbitals’. More precisely, an atomic orbital is a one-electron
zero there. wavefunction for an electron in an atom.
(a) The atomic orbitals

Brief illustration 17.2 The radial nodes of the
Each hydrogenic atomic orbital (eqn 17.4) is defined by three
wavefunction quantum numbers, designated n, l, and ml:
Consider an electron in a hydrogen atom with n = 2, l = 0,
Hydrogenic Atomic
ml = 0. Because the radial wavefunction R 2,0 ∝ (2 − ρ) (see Table ψnlm (r ,θ , φ ) = Rnl (r )Ylm (θ ,φ ) atom
(17.11)
l l orbital
17.1), the radial node is found at ρ = 2. Using eqns 17.8 and 17.9,
we find, with Z = 1 and n = 2, When an electron is described by one of these wavefunc-
tions, we say that it ‘occupies’ that orbital.
na0 ρ
r= = 2a0 The quantum number n is called the principal quantum
2Z
number; it can take the values n = 1, 2, 3, … and determines the
Therefore, there is zero probability of finding the electron in a energy of the electron, as in eqn 17.7. The two other quantum
small volume element at a distance 105.8 pm from the nucleus numbers, l and ml, come from the angular solutions and spec-
in the hydrogen atom. ify the angular momentum of the electron around the nucleus,
as explained in Topic 14. However, the various boundary
Atkins09819.indb 155 9/11/2013 11:13:12 AM

2
0.8 range of values of ml. For example, if n = 3, the allowed values of
l (and corresponding ml) are 0 (0), 1 (0, ±1), and 2 (0, ±1, ±2).
1.5 0.6
n = 2, l = 0
Brief illustration 17.3 The probability density for
R(r)/(Z/a0)3/2
R(r)/(Z/a0)3/2
0.4
1
n = 1, l = 0
an electron
0.2 To calculate the probability density at the nucleus for an elec-
0.5 tron occupying the n = 1, l = 0, ml = 0 orbital, we begin by evalu-
0 ating ψ1,0,0 at r = 0 (see Tables 14.1 and 17.1):
(a) (b) 3/2 1/2
0 –0.2 ⎛Z⎞ ⎛ 1 ⎞
0 1 Zr/a0 2 3 0 5 10 15 ψ 1,0,0 (0,θ , φ ) = R1,0 (0)Y0,0 (θ , φ ) = 2 ⎜ ⎟ ⎜⎝ 4 π ⎟⎠
Zr/a0 ⎝ a0 ⎠
0.4 0.15
The probability density is therefore
0.3 Z3
ψ 1,0,0 (0,θ , φ )2 =
πa03
R(r)/(Z/a0)3/2
0.1
R(r)/(Z/a0)3/2
0.2
which evaluates to 2.15 × 10 −6 pm−3 when Z = 1. Therefore, the
0.1 n = 2, l = 1 probability of finding the electron inside a region of volume
n = 3, l = 0
0.05 1.00 pm 3 located at the nucleus, ignoring the variation of ψ
0 within the very small region, is 2.15 × 10 −6 or about 1 part in
(c) (d) 5 × 105.
–0.1 0
0 7.5 15 22.5 0 5 10 15 Self-test 17.3 Evaluate the probability density at the nucleus
Zr/a0 Zr/a0
for an electron occupying the n = 2, l = 0, ml = 0 orbital. What
0.1 0.05 is the probability of finding the electron inside a region of vol-
ume 1.00 pm3 located at the nucleus of the H atom?
0.04 Answer: Z 3/(8πa03 ), 2.69 × 10 −7
0.05 n = 3, l = 2
R(r)/(Z/a0)3/2
R(r)/(Z/a0)3/2
0.03
n = 3, l = 1
0.02 (b) The energy levels
0
0.01 The energy levels predicted by eqn 17.7 are depicted in Fig.
(e) (f) 17.7. The energies, and also the separation of neighbouring lev-
–0.05
0 7.5 15 22.5
0 els, are proportional to Z2; therefore, the levels are four times
0 7.5 15 22.5
Zr/a0 Zr/a0 as wide apart in He+ (Z = 2) as in H (Z = 1) and the ground state
(n = 1) is four times deeper in energy. All the energies given by
Figure 17.6 The radial wavefunctions of the first few states of eqn 17.7 are negative because they refer to the bound states of
hydrogenic atoms of atomic number Z. Note that the orbitals with the atom, in which the energy of the atom is lower than that
l = 0 have a nonzero and finite value at the nucleus. The horizontal of the infinitely separated, stationary electron and nucleus (the
scales are different in each case: orbitals with high principal
n = ∞ limit). There are also solutions of the Schrödinger equa-
quantum numbers are relatively distant from the nucleus.
tion (eqn 17.3) with positive energies. These solutions corre-
conditions that the angular and radial wavefunctions must sat- spond to unbound states of the electron, the states to which
isfy put additional constraints on their values: an electron is raised when it is ejected from the atom by a
high-energy collision or photon. The energies of the unbound
An electron in an orbital with quantum number l has an
electron are not quantized (the boundary condition that the
angular momentum of magnitude {l(l + 1)}1/2 , with l = 0, 1,
wavefunction must vanish at infinity is no longer relevant) and
2, …, n − 1.
form the continuum states of the atom.
An electron in an orbital with quantum number ml has a For the hydrogen atom, the expression for the energy in eqn
z-component of angular momentum ml , with 17.7 can be written in the form
ml = 0, ± 1, ± 2,…, ± l.
hcR H Energy,
Note how the value of the principal quantum number, n, con- En = − n = 1, 2, … Alternative
hydrogen (17.12a)
n2 form
trols the maximum value of l and how, in turn, l restricts the atom
www.ebook3000.com
Atkins09819.indb 156 9/11/2013 11:13:22 AM

where R H is the Rydberg constant for hydrogen Therefore, for n = 3, the energy is

μ e4 Z
2 hcR ∞
R H = H2 3 Rydberg constant for H (17.12b) 4 × 2.179 87 × 10−18 J
8ε 0 h c E3 = −
9
with a value of 109 677 cm−1. The Rydberg constant itself, R ∞ , n2
is defined by the same expression as eqn 17.12b except for the = −9.688 31 × 10−19 J
replacement of μH by the mass of an electron, me, correspond-
ing to a nucleus of infinite mass: or −0.968 831 aJ (a, for atto, is the prefix that denotes 10 −18).
In some applications it is useful to express the energy
m e4 in electronvolts (1 eV = 1.602 176 × 10 −19 J); in this case,
R ∞ = 2e 3 Definition Rydberg constant (17.13)
8ε 0 h c E3 = −6.046 97 eV.
Self-test 17.4 What is the energy of the electron in an excited
state of a Li+2 ion (Z = 3) for which the wavefunction is R4,3(r) ×
Y3,−2(θ,φ)?
Example 17.1 Determining the energy for an electron Answer: −1.226 18 aJ, −7.653 22 eV
occupying an atomic orbital
The single electron in a certain excited state of a hydrogenic
Equation 17.12 is a convenient expression for analyses of the
He + ion (Z = 2) is described by the wavefunction R 3,2 (r) ×
Y2,−1(θ ,φ). What is the energy of its electron? emission spectrum of atomic hydrogen, as shown in the follow-
ing Example, and explored in more detail in Topic 21.
Method Because the energy of a hydrogenic atom depends on
n but is independent of the values of l and ml, we need iden-
tify only the quantum number n, which is done by noting the
form of the wavefunction given in eqn 17.11. Then use eqn 17.7 Example 17.2 Calculating the wavenumber of a line in
to calculate the energy. To a good approximation, the reduced the emission spectrum of H
mass in eqn 17.7 can be replaced by me and the energy can be
When an electric discharge is passed through gaseous hydro-
written in terms of the Rydberg constant of eqn 17.13. (For
gen, the H2 molecules are dissociated and energetically excited
greater accuracy, use the reduced mass of the electron and
H atoms are produced. If the electron in an excited H atom
helium nucleus.)
makes a transition from n = 2 to n = 1, calculate the wavenum-
Answer Replacing μ by me and using = h/2π, we can write ber of the corresponding line in the emission spectrum.
the expression for the energy (eqn 17.7) as
Method When an excited electron makes a transition from
2
Z me 4
Z hcR ∞
2 a state with quantum number n 2 to a lower energy state with
En = − 2 e2 2 = − quantum number n1, it loses an energy
8ε 0 h n n2
⎛ 1 1⎞
with ΔE = En2 − En1 = −hcR H ⎜ 2 − 2 ⎟
⎝ n2 n1 ⎠
me
e
4
9 . 109 38 × 10 −31
kg × (1 . 602 176 × 10 −19
C )4 The frequency of the emitted photon is ν = ΔE/h; the wave-
R ∞ = number is ␯ = ␯/c = ΔE /hc .
8 × (8.854 19 ×10 J C m ) × (6.626 08 ×10−34 Js)3
−12 −1 2 −1 2

ε 02 h3 Answer The wavenumber of the photon emitted when an elec-
× 2.997 926 ×1010 cms −1 tron makes a transition from n2 = 2 to n1 = 1 is given by

c
⎛ 1 1⎞
= 109 737 cm −1 ␯ = − R H ⎜ 2 − 2 ⎟
⎝ n2 n1 ⎠
⎛ 1 1⎞
and = −(109677cm −1 ) × ⎜ 2 − 2 ⎟
⎝2 1 ⎠
hcR ∞ = (6.626 08 × 10−34 Js) × (2.997 926 × 1010 cms −1 ) = 82 258 cm −1
× (109 737cm −1 )
= 2.179 87 × 10−18 J The emitted photon has a wavelength of 122 nm, correspond-
ing to ultraviolet radiation.
Atkins09819.indb 157 9/11/2013 11:13:38 AM

Self-test 17.5 Calculate the wavelength and wavenumber of then when the atom makes a transition to E lower, a photon of
the emitted photon if the electron in H makes a transition wavenumber
from n = 3 to n = 2.
Answer: 656 nm, 15 233 cm−1, visible R H Elower
␯ = − −
n2 hc
is emitted. However, because I = −Elower, it follows that

(c) Ionization energies
I R H
The ionization energy, I, of an element is the minimum energy ␯ = −
hc n2
required to remove an electron from the ground state, the state
of lowest energy, of one of its atoms in the gas phase. Because A plot of the wavenumbers against 1/n2 should give a straight
the ground state of hydrogen is the state with n = 1, with energy line of slope − R H and intercept I/hc. Use mathematical soft-
E1 = −hcR H , and the atom is ionized when the electron has been ware to make a least-squares fit of the data to get a result that
excited to the level corresponding to n = ∞ (see Fig.17.7), the reflects the precision of the data.
energy that must be supplied is
110
I = hcR H Hydrogen atom Ionization energy (17.14)
The value of I is 2.179 aJ, which corresponds to 13.60 eV.

100
ν/(103 cm–1)
Continuum
H+ + e– n
0 ∞ ~
~ 90
–hcRH/9 3
~
2
Energy, E
–hcRH/4
80
Classically 0 0.1 0.2
allowed 1/n2
energies
~
Figure 17.8 The plot of the data in Example 17.3 used to
–hcRH 1 determine the ionization energy of an atom (in this case,
of H).
Figure 17.7 The energy levels of a hydrogen atom. The values
are relative to an infinitely separated, stationary electron and a Answer The wavenumbers are plotted against 1/n2 in Fig. 17.8.
proton. The (least-squares) intercept lies at 109 679 cm−1, so the ioni-
zation energy is 2.1788 aJ (1317.1 kJ mol−1). The slope is, in this
instance, numerically the same, so the experimentally deter-
Example 17.3 Measuring an ionization energy mined value of the Rydberg constant is 109 679 cm−1, very close
spectroscopically to the actual value of 109 677 cm−1.
The emission spectrum of atomic hydrogen shows a series Self-test 17.6 The emission spectrum of atomic deuterium
of lines at 82 259, 97 492, 102 824, 105 292, 106 632, and shows lines at 15 238, 20 571, 23 039, and 24 380 cm−1, which
107 440 cm −1, which correspond to transitions to the same correspond to transitions to the same lower state. Determine
lower state. Determine (a) the ionization energy of the lower (a) the ionization energy of the lower state, (b) the ionization
state, (b) the value of the Rydberg constant. energy of the ground state, (c) the mass of the deuteron (by
expressing the Rydberg constant in terms of the reduced mass
Method The spectroscopic determination of ionization ener-
of the electron and the deuteron, and solving for the mass of
gies depends on the determination of the ‘series limit’, the
the deuteron).
wavenumber at which the series terminates and becomes
Answer: (a) 328.1 kJ mol−1, (b) 1317.4 kJ mol−1, (c) 2.8 × 10 −27 kg,
a continuum. If the upper state lies at an energy −hcR H / n2 , a result very sensitive to R D
www.ebook3000.com
Atkins09819.indb 158 9/11/2013 11:13:49 AM

☐ 1. A hydrogenic atom is a one-electron atom or ion of ☐ 5. The quantum numbers l and ml specify the magnitude
general atomic number Z. A many-electron atom is an (as {l(l + 1)}1/2 ) and the z-component (as ml ), respec-
atom or ion with more than one electron. tively, of the angular momentum of the electron around
☐ 2. The wavefunction of a hydrogenic atom is the product the nucleus. The allowed values are l = 0, 1, 2, …, n − 1;
of a radial wavefunction and an angular wavefunction ml = 0, ± 1, ± 2,…, ± l.
(spherical harmonic) and is labelled by the quantum ☐ 6. The energy of an infinitely separated, stationary electron
numbers n, l, and ml. and nucleus is zero. Electron energies which are negative
☐ 3. An atomic orbital is a one-electron wavefunction for an correspond to bound states of the atom. Positive energies
electron in an atom. correspond to unbound or continuum states.
☐ 4. The principa l quantum number n determines ☐ 7. The ionization energy of an element is the minimum
the energy of an electron in a hydrogenic atom; n = 1, energy required to remove an electron from the ground
2, …. state of one of its atoms in the gas phase.
Coulomb potential energy V(r) = −Ze2/4πε 0r 17.1
Reduced mass μ = memN/(me + mN) 17.3b
Radial wave equation −( 2/2μ)(d2u/dr2) + Veffu = Eu Hydrogenic atom 17.6a
Effective potential energy Veff = −Ze2/4πε0r + l(l + 1) 2/2μr2 Hydrogenic atom 17.6b
Electronic energy En = − Z 2 μe 4 / 32π2ε 02 2n2 Hydrogenic atom 17.7
n = 1, 2, …
Bohr radius a0 = 4πε0 2/mee2 52.9 pm 17.9
Atomic orbital ψnlml(r , θ , φ ) = Rnl (r )Ylml (θ , φ ) 17.11
Rydberg constant for hydrogen R H = μHe 4 /8ε 02h3c 109 677 cm−1 17.12b
Rydberg constant R ∞ = mee 4 /8ε 02h3c 109 737 cm−1 17.13
Ionization energy of hydrogen I = hcR H 2.179 aJ, 13.60 eV 17.14
Atkins09819.indb 159 9/11/2013 11:13:54 AM

TOPIC 18
Hydrogenic atomic orbitals
Contents (Topic 9). This Topic draws heavily on the forms of the
radial wavefunctions given in Table 17.1 of Topic 17 and the
18.1 Shells and subshells 160
spherical harmonics given in Table 14.1 of Topic 14. Various
Brief illustration 18.1: The number of orbitals
in a shell 161 techniques of integration are used; they are reviewed in
(a) s Orbitals 161 Mathematical background 1.
Example 18.1: Calculating the mean radius of an
s orbital 162
(b) p Orbitals 163 Hydrogenic atomic orbitals, which are the electronic wavefunc-
Brief illustration 18.2: The 2px orbital 165 tions for one-electron atoms with atomic number Z, are defined
(c) d Orbitals 165 by the three quantum numbers n, l, and ml (Topic 17). They
Example 18.2: Finding the most probable have the form ψ nlm = RnlYlm where the radial wavefunction R is
l l
locations of an electron 165 given in Table 17.1 and the spherical harmonic Y in Table 14.1.
2
18.2 Radial distribution functions 165 The probability densities ψ nlm result in the shapes of atomic
l
Example 18.3: Identifying the most probable orbitals familiar from introductory chemistry courses. Here we
radius 166 explore the properties of hydrogenic atomic orbitals because
Checklist of concepts 168 they provide the basis for describing the electronic structure
and periodic properties of many-electron atoms (Topic 19)
and, by extension, molecules (Topics 22–30).
18.1 Shells and subshells

➤ Why do you need to know this material? All atomic orbitals of a given value of n are said to form a single
The properties of hydrogenic atomic orbitals figure
shell of the atom; in a hydrogenic atom, all orbitals of given n,
prominently in the discussion of electronic structure
and therefore belonging to the same shell, have the same energy.
of many-electron atoms and, by extension, molecules.
It is common to refer to successive shells by uppercase letters:
Therefore, to understand the form of the periodic table,
n= 1 2 3 4…
molecular structure, and chemical reactivity, you need a
firm understanding of atomic orbitals. K L M N…
➤ What is the key idea? Thus, all the orbitals of the shell with n = 2 form the L shell of
Atomic orbitals describe the probability density distribution
the atom, and so on. The orbitals with the same value of n but
of an electron in an atom.
different values of l (allowed values are 0, 1, …, n – 1) are said
to form a subshell of a given shell.The subshells are generally
➤ What do you need to know already? referred to by lowercase letters:
You need to know what is meant by the terms probability
density (Topic 5), expectation value (Topic 7), and node l=0 1 2 3 4 5 6…
s p d f g h i…
www.ebook3000.com
Atkins09819.indb 160 9/11/2013 11:13:59 AM

18 Hydrogenic atomic orbitals 161
n s
∞ p d f Brief illustration 18.1 The number of orbitals in a shell
4 4s[1] 4p[3] 4d [5] 4f [7]
3 When n = 3 there are three subshells, l = 0 (3s), l = 1 (3p), and
3s[1] 3p[3] 3d [5]
l = 2 (3d). Since the s, p, and d subshells contain one, three,
2
2s[1] 2p[3] and five orbitals, respectively, there are a total of 9 = 32 orbitals
Energy
in the M shell.
Self-test 18.1 Identify the orbitals in the N shell.
Answer: one 4s, three 4p, five 4d, seven 4f orbitals; 16 in total
1s [1]
1 (a) s Orbitals
Figure 18.1 The energy levels of a hydrogenic atom showing The orbital occupied in the ground state of a hydrogenic atom
the subshells and (in square brackets) the numbers of orbitals is the one with n = 1 and therefore with l = 0 and ml = 0. The
in each subshell. All orbitals of a given shell have the same wavefunction ψ1,0,0 is the product of the radial wavefunction
energy. R1,0 and the spherical harmonic Y0,0, so from Tables 14.1 and
17.1 it follows that
The letters then run alphabetically omitting j (because some 1/2
languages do not distinguish between i and j). Figure 18.1 ⎛ Z3 ⎞
ψ =⎜ 3⎟ e − Zr /a
0 Hydrogenic atom 1s wavefunction (18.1)
shows the energy levels of the subshells explicitly for a hydro- ⎝ πa0 ⎠
genic atom. Because l can range from 0 to n − 1, giving n values
in all, it follows that there are n subshells of a shell with princi- This wavefunction is independent of the angular location of
pal quantum number n. Thus, the electron and has the same value at all points of constant
radius; that is, the 1s orbital is ‘spherically symmetrical’. The
r when n = 1, there is only one subshell, the one with l = 0 wavefunction decays exponentially from a maximum value of
(the 1s subshell); (Z 3 /π a03 )1/2 at the nucleus (at r = 0), and therefore the greatest
r when n = 2, there are two subshells, the 2s subshell (with probability density is at the nucleus.
l = 0) and the 2p subshell (with l = 1). A note on good practice Always keep in mind the distinction
between the probability density (dimensions: 1/volume) at a
When n = 1 there is only one subshell, that with l = 0, and that point, ψ 2, and the probability (dimensionless) of the electron
subshell contains only one orbital, with ml = 0 (the only value being in an infinitesimal region dτ at that point, ψ 2dτ.
of ml permitted). When n = 2, there are four orbitals, one in
the s subshell with l = 0 and ml = 0, and three in the l = 1 sub- The general form of the ground-state wavefunction can be
shell with ml = + 1, 0, −1. In general, the number of orbitals in a understood by considering the contributions of the potential
shell of principal quantum number n is n2, so in a hydrogenic and kinetic energies to the total energy of the atom. The closer
atom each energy level is n2-fold degenerate. The organization the electron is to the nucleus, the lower (that is, more negative)
of orbitals in the shells is summarized in Fig. 18.2. its potential energy. This dependence suggests that the lowest
potential energy is obtained with a sharply peaked wavefunc-
Subshells tion that has a large amplitude at the nucleus and is very small
everywhere else (Fig. 18.3a). However, this shape implies a
s p d
high kinetic energy, because such a wavefunction has a very
M shell, n = 3
high average curvature. The electron would have very low
kinetic energy if its wavefunction had only a very low average
Orbitals curvature. However, such a wavefunction (Fig. 18.3b) spreads
L shell, n = 2 to great distances from the nucleus and the average potential
energy of the electron is correspondingly high (that is, less
negative). The actual ground-state wavefunction is a compro-
mise between these two extremes (Fig. 18.3c): the wavefunc-
Shells
K shell, n = 1
tion spreads away from the nucleus (so the expectation value
of the potential energy is not as low as in the first example, but
Figure 18.2 The organization of orbitals (white squares) nor is it very high) and has a reasonably low average curvature
into subshells (characterized by l) and shells (characterized (so the expectation value of the kinetic energy is not very low,
by n). but nor is it as high as in the first example). The contributions of
Atkins09819.indb 161 9/11/2013 11:14:04 AM

Low potential energy z

but z
high kinetic energy
Wavefunction, ψ
a
Lowest total energy
c Low kinetic energy
but
high potential energy x y
b
x y
Radius, r
(a) 1s (b) 2s
Figure 18.3 The balance of kinetic and potential energies that
accounts for the structure of the ground state of hydrogenic
atoms. (a) The sharply curved but localized orbital has high Figure 18.4 Representations of cross-sections through the (a) 1s
mean kinetic energy, but low mean potential energy; (b) the and (b) 2s hydrogenic atomic orbitals in terms of their electron
mean kinetic energy is low, but the potential energy is not very probability densities (as represented by the density of shading).
favourable; (c) the compromise of moderate kinetic energy and
moderately favourable potential energy. z
the kinetic energy and the potential energy to the ground-state

energy can also be understood by using the virial theorem, as
shown in the following Justification.
One way of depicting the probability density of the elec-
tron is to represent |ψ|2 by the density of shading (Fig. 18.4). x
y
Justification 18.1
The virial theorem and the energies
of hydrogenic atomic orbitals Figure 18.5 The boundary surface of a 1s orbital, within which
The virial theorem (Topic 12) states that if the potential energy there is a 90 per cent probability of finding the electron. All s
of the system is of the form V = ax b, where a and b are con- orbitals have spherical boundary surfaces.
stants, then the average kinetic and potential energies are
related by 2〈E k 〉 = 〈V〉. For the Coulomb potential energy
(V ∝ −1/r), b = −1; therefore Example 18.1 Calculating the mean radius of an s orbital
1 Evaluate the mean radius of a 1s orbital of a hydrogenic atom
〈 Ek 〉 = − 〈V 〉
2 of atomic number Z.
As the average distance of the electron from the nucleus Method The mean radius is the expectation value
increases, 〈V〉 increases (becomes less negative) so 〈E k 〉
∫ ∫
2
decreases (becomes less positive), as described in the text. 〈r 〉 = ψ ∗ rψ dτ = r ψ dτ
We therefore need to evaluate the integral using the hydro-

A simpler procedure is to show only the boundary surface, the genic 1s wavefunction (Tables 14.1 and 17.1); the volume ele-
surface that captures a high proportion (typically about 90 per ment in spherical polar coordinates is dτ = r2dr sin θ dθ dφ (see
cent) of the electron probability density. For the 1s orbital, the The chemist’s toolkit 14.1). The angular parts of the wavefunc-
boundary surface is a sphere centred on the nucleus (Fig. 18.5). tion are normalized in the sense that
All s orbitals are spherically symmetric, but differ in the π 2π
∫∫
2
number of radial nodes. For example, the 1s, 2s, and 3s orbitals, Ylml sin θ dθ dφ = 1
0 0
which are collected in Table 18.1, have 0, 1, and 2 radial nodes,
respectively (see Fig. 17.6 and note the number of zeroes in the where the limits on the first integral sign refer to θ, and
radial wavefunction). The radial nodes for the 2s and 3s orbitals those on the second to φ (recall the procedure for multiple
are calculated in Brief illustration 17.2. In general, an ns orbital
has n−1 radial nodes.
www.ebook3000.com
Atkins09819.indb 162 9/11/2013 11:14:16 AM

The hydrogenic 2p orbital with ml = 0 is

integration in Mathematical background 1). Use Integral E.1
listed in the Resource section. 1 ⎛Z⎞
5/2
ψ 2 p = R2,1Y1,0 = r cosθ e − Zr /2a 2pz atomic

4(2π )1/2 ⎜⎝ a0 ⎟⎠
0
0
(18.2a)
Answer With the wavefunction written in the form orbital
ψ nlml = RnlYlml , the expectation value 〈r〉 is = r cosθ f (r )
2
ψ

dτ
where f(r) is a function only of r. Because in spherical polar
∞ π 2π
coordinates z = r cos θ, this wavefunction may also be written
∫∫∫
2
〈r 〉 = rRnl2 Ylml r dr sinθ dθ dφ
2
0 0 0

1
ψ 2p = zf (r ) (18.2b)
∞ π 2π z
∫ ∫∫
2
= r 3
Rnl2 dr × Ylml sinθ dθ dφ
0
∞
0 0
All p orbitals with ml = 0 have wavefunctions of this form
=∫ r 3Rnl2 dr regardless of the value of n (with different forms of f for dif-
0
ferent values of n). This way of writing the orbital is the ori-
gin of the name ‘pz orbital’. The boundary surface (for n = 2) is
For a 1s orbital (Table 17.1),
shown in Fig. 18.6. The wavefunction is zero everywhere in the
3/2 xy-plane, where z = 0, so the xy-plane is a nodal plane of the
⎛Z⎞
R1,0 = 2 ⎜ ⎟ e − Zr /a0 orbital: the wavefunction changes sign on going from one side
⎝ a0 ⎠
of the nodal plane to the other.
Hence, by using Integral E.1 given in the Resource section with The hydrogenic 2p orbitals with ml = ± 1 are
n = 3 and a = 2Z/a0,
ψ 2 p = R2,1Y1, ±1
±1
Integral E.1
⎛Z⎞
3 ∞
⎛Z⎞
3
3! 1 ⎛Z⎞
5/2
〈r 〉 = 4 ⎜ ⎟
⎝ a0 ⎠ ∫0
r 3e −2 Zr /a0 dr = 4⎜ ⎟ ×
⎝ 0 ⎠ (2Z / a0 )4
a =∓
8π 1/2 ⎜⎝ a0 ⎟⎠
r e − Zr /2a sinθ e ± iφ
0
3a0 1
= = ∓ 1/2 r sinθ e ± iφ f (r ) (18.3)
2Z 2
For H, 〈r〉 = 79.4 pm, and for He + , 〈r〉 = 39.7 pm. In general, the with f(r) a function only of r (but not the same function as in
higher the nuclear charge, the closer the electron is drawn to
eqn 18.2). These functions correspond to nonzero angular
the nucleus.
Self-test 18.2 Evaluate the mean radius of (a) a 3s orbital and
θ = 90°
(b) a 3p orbital for a hydrogenic atom of atomic number Z.
Answer: (a) (27/2)a0/Z; (b) (25/2)a0/Z
φ = 90° φ=0
+
–
(b) p Orbitals +
– +
The three p orbitals are distinguished by the three different

–
values that ml can take when l = 1. Because the quantum num-
ber ml tells us the projection of the orbital angular momen- pz z px py
θ
tum onto a particular axis (by convention, the z-axis), these y
different values of ml denote orbitals in which the electron
x φ
has different orbital angular momenta around an arbitrary
z-axis. The orbital with ml = 0, for instance, has zero angular Figure 18.6 The boundary surfaces of 2p orbitals. A nodal
momentum around the z-axis. Its angular variation is pro- plane passes through the nucleus and separates the two lobes
portional to cos θ (Table 14.1), so the probability density, of each orbital. The dark and light areas denote regions of
which is proportional to cos2 θ, has its maximum value on opposite sign of the wavefunction. The angles of the spherical
either side of the nucleus along the z-axis (at θ = 0 and 180°, polar coordinate system are also shown. All p orbitals have
where cos2 θ = 1). boundary surfaces like those shown here.
Atkins09819.indb 163 9/11/2013 11:14:30 AM

momentum about the z-axis: e +iφ corresponds to clockwise The 2px orbital has the same shape as a 2pz orbital, but it is
rotation when viewed from below, and e−iφ corresponds to anti- directed along the x-axis (see Fig. 18.6); the 2py orbital is simi-
clockwise rotation (from the same viewpoint). They have zero larly directed along the y-axis. The wavefunction of any p orbital
amplitude where θ = 0 and 180° (along the z-axis) and maxi- of a given shell can be written as a product of x, y, or z and the
mum amplitude at θ = 90°, which is in the xy-plane. same radial function (which depends on the value of n). The 2p
To draw the functions in eqn 18.3, it is customary to con- hydrogenic wavefunctions presented in eqns 18.2–18.4 are col-
struct two real wavefunctions that are linear combinations lected in Table 18.1 as well as the 3p hydrogenic wavefunctions.
of the degenerate functions ψ 2 p . We show in the following
±1
All 2p orbitals have no radial nodes; 3p orbitals have one radial
Justification that it is permissible to take linear combinations of node, and np orbitals have n – 2 radial nodes. These numbers,
degenerate orbitals. In particular, we form the real linear com- including that for s orbitals given earlier, are special cases of
binations (using e + iφ = cos φ ± i sin φ)
x Table 18.1 Hydrogenic atomic orbitals

1
ψ 2p = − 1/2
(ψ 2p − ψ 2p ) = r sin θ cos φ f (r ) 2px atomic
(18.4a)
x
2 +1 −1
orbital s orbitals
= xf (r ) 1s 1/2
⎛ Z3 ⎞
ψ 1s = ⎜ 3 ⎟ e −Zr /a0
⎝ π a0 ⎠
i y

ψ 2p = (ψ + ψ 2 p ) = r sinθ sin φ f (r ) 2py atomic
(18.4b)
2s 1/2
y
21/2 2 p +1 −1 ⎛ Z3 ⎞ ⎛ Zr ⎞ −Zr / 2a0
orbital ψ 2s = ⎜ ⎟ ⎜⎝ 2 − a ⎟⎠ e
= yf (r ) ⎝ 32π a03 ⎠ 0
3s 1/2
Because the functions in eqn 18.4 are superpositions of states ⎛ Z3 ⎞ ⎛ 4 Zr 4 Z 2r 2 ⎞ −Zr /3a
ψ 3s = ⎜ ⎟ ⎜ 6 − a + 9a2 ⎟ e 0
with equal and opposite values of ml (namely, + 1 and −1), the ⎝ 972π a03 ⎠ ⎝ 0 0 ⎠
linear combinations are standing waves with no net orbital p orbitals

angular momentum around the z-axis.
2p 5/2
1 ⎛Z⎞
With f (r )= e − Zr / 2a0
(32π )1/2 ⎜⎝ a0 ⎟⎠
Justification 18.2 The linear combination of degenerate Polar form Cartesian form
wavefunctions ψ 2p0 = r cos θ f (r ) ψ 2px = xf (r )
The freedom to take linear combinations of degenerate 1 ψ 2p y = yf (r )
ψ 2p±1 = ∓ 1/2 r sin θ e ± iφ f (r )
functions rests on the fact that whenever two or more wave- 2 ψ 2pz = zf (r )
functions correspond to the same energy (as is the case with
3p
ψ 2 p+1 and ψ 2 p−1), any linear combination of them (such as ⎛ 2 ⎞
1/2
⎛Z⎞
5/2
⎛ Zr ⎞ − Zr /3a0
ψ 2px or ψ 2p y ) is an equally valid solution of the Schrödinger
With f (r )= ⎜
⎝ 729π ⎟⎠ ⎜⎝ a ⎟⎠ ⎜⎝ 2 − 3a ⎟⎠ e
0 0
equation.
ψ 3p0 = r cos θ f (r ) ψ 3px = xf (r )
Suppose ψ 1 and ψ 2 are both solutions of the Schrödinger
equation with energy E; then we know that Hψ = Eψ ψ 3p±1
1
= ∓ 1/2 r sin θ e ± iφ f (r )
ψ 3p y = yf (r )
1 1

and H ψ 2 = Eψ 2 . Now consider the linear combination 2 ψ 3pz = zf (r )
ψ = c1ψ1 + c2ψ 2, where c1 and c2 are arbitrary coefficients. Then it d orbitals
follows that
3d 1/2 7/2
⎛ 2 ⎞ ⎛Z⎞
With f (r )= ⎜ e − Zr /3a0
⎝ 6561π ⎟⎠ ⎜⎝ a ⎟⎠
0
H ψ = H (c1ψ 1 + c2ψ 2 ) = c1 H ψ 1 + c2 H ψ 2
1/2 1/2
= c1Eψ 1 + c2 Eψ 2 = E(c1ψ 1 + c2ψ 2 ) = Eψ ⎛ 1⎞ ⎛ 1⎞
ψ 3d0 = ⎜ ⎟ r 2 (3 cos2 θ −1) f (r ) ψ 3d 2 = ⎜ ⎟ (3z 2 − r 2 ) f (r )
⎝ 12 ⎠ z ⎝ 12 ⎠
1
Hence, the linear combination is also a solution correspond- ψ 3d 2 2 = (x 2 − y 2 ) f (r )
x −y 2
ing to the same energy E. The result that any linear combi- 1/2
⎛ 1⎞ ψ 3d xy = xyf (r )
ψ 3d ±1 = ∓ ⎜ ⎟ r 2cos θ sin θ e ± iφ f (r )
nation of eigenfunctions of an operator all having the same ⎝ 2⎠ ψ 3d yz = yzf (r )
eigenvalue is also an eigenfunction of the operator with the ψ 3d zx = zxf (r )
1/2
same eigenvalue applies to all quantum mechanical operators, ⎛ 1⎞
ψ 3d ±2 =⎜ ⎟ r 2sin2 θ e ±2iφ f (r )
not just the hamiltonian. ⎝ 8⎠
www.ebook3000.com
Atkins09819.indb 164 9/11/2013 11:15:01 AM

the general expression that a hydrogenic orbital with quantum Example 18.2 Finding the most probable locations of
numbers n and l has n – l –1 radial nodes and l nodal planes.
an electron
Find the most probable location of an electron occupying the
Brief illustration 18.2 The 2px orbital 3d z 2 orbital in the hydrogen atom.
The angular variation of the 2px orbital is proportional to Method The 3d z 2 orbital corresponds to n = 3, l = 2, ml = 0. To
sin θ cos φ, so the probability density, which is proportional find the most probable location of the electron, construct the
2
to sin 2 θ cos 2 φ, has its maximum values at θ = 90° (where probability density ψ nlml by using Table 18.1 and then find
sin2 θ = 1) and φ = 0 and 180° (where cos2 φ = 1). Therefore, the its maxima by finding where the first derivative of the prob-
electron is most likely to be found on either side of the nucleus ability density vanishes.
along the x-axis (Fig. 18.6). In addition, the wavefunction is
Answer The 3d z 2 hydrogen atomic orbital is
zero (and passes through zero) at θ = 0, 180° and φ = 90°, 270°,
so the yz-plane is a nodal plane.
1/2 2
Self-test 18.3 Identify the nodal plane of the 2py orbital. ⎛ 1 ⎞ ⎛ 2r ⎞ − r /3a0
ψ 3d = R3,2 (r )Y2,0 (θ , φ ) = ⎜ e (3 cos2 θ −1)
Answer: xz-plane
z 2
⎝ 7776π a03 ⎟⎠ ⎜⎝ 3a0 ⎟⎠
1/2
⎛ 1 ⎞
= Nr 2e − r /3a0 (3 cos2θ − 1) N =⎜
⎝ 39366π a07 ⎟⎠
(c) d Orbitals Therefore,

When n = 3, l can be 0, 1, or 2. As a result, this shell consists of ψ 3d2 = N 2r 4 e −2r /3a0 (3cos2θ − 1)2
z2
one 3s orbital, three 3p orbitals, and five 3d orbitals. The five d
orbitals have ml = +2, +1, 0, −1, −2 and correspond to five differ- The probability density, which is independent of φ, has maxima
ent components of the angular momenta around the z-axis (but at θ = 0, 180° (where cos θ = ± 1, cos2 θ = 1, and 3 cos2 θ−1 = 2),
the same magnitude of angular momentum, because l = 2 in each so the electron is most likely to be found on either side of the
case). As for the p orbitals, d orbitals with opposite values of ml z-axis. To find the most probable distance r from the electron
(and hence opposite senses of motion around the z-axis) may be to the nucleus, we need to find, by differentiation, the maxi-
combined in pairs to give real standing waves (the d orbital with mum in the function r 4 e −2r /3a0 :
ml = 0 is real and designated d z ) . The boundary surfaces of the
2
⎛ f
f (dg /dr )

g
⎞
(df /dr ) g
d ⎜
4 −2r /3a0 ⎟
resulting 3d orbitals are shown in Fig. 18.7. The 3d orbitals and
2
their real combinations are collected in Table 18.1. r e = 4r e
3 −2 r /3a0
− 4 −2 r /3a0
r e
dr ⎜ ⎟ 3a0
⎝ ⎠
= 2r 3e −2r /3a0 (2 − r / 3a0 )
+
where we have used the product rule of differentiation

z +
–
(dfg = fdg + gdf; Mathematical background 1). This function is
y –
– zero where the term in parenthesis is zero (ignore r = 0 which
+
x corresponds to a minimum in the function), which is at r = 6a0.
+
Therefore the most probable locations of the electron are at a
dz2 dx2–y2 distance of 6a0 from the nucleus on either side of the z-axis.
–
Self-test 18.4 Identify the nodal planes of the 3d xy orbital for
– + +
+ –
the hydrogen atom.
–
–
+
+ Answer: xz- and yz-planes
– +
+ +
+
– –
Radial distribution functions

dxy dyz dzx
18.2
Figure 18.7 The boundary surfaces of 3d orbitals. Two nodal
planes in each orbital intersect at the nucleus and separate the The wavefunction tells us, through the value of |ψ|2, the prob-
lobes of each orbital. The dark and light areas denote regions of ability of finding an electron in any region in space. Imagine a
opposite sign of the wavefunction. All d orbitals have boundary probe with a volume dτ that is sensitive to electrons and can
surfaces like those shown here. move around near the nucleus of a hydrogenic atom. Because
Atkins09819.indb 165 9/11/2013 11:15:22 AM

2
ψ
π 2π
dτ
∫∫
2
Probability, ψ *ψdτ
P (r )dr = R(r )2 Y (θ , φ r 2dr sinθ dθ dφ

0 0
r
1
π 2π
∫∫
2
= r 2 R(r )2 dr Y (θ , φ ) sinθ dθ dφ
0 0
= r 2 R(r )2 dr
The last equality follows from the fact that the spherical har-
monics are normalized to 1 (see Example 18.1). It follows that
Radius, r
P(r) = r2R(r)2, as stated in the text.
Figure 18.8 A constant-volume electron-sensitive detector
(the small cube) gives its greatest reading at the nucleus, and
The radial distribution function, P(r), is a probability den-
a smaller reading elsewhere. The same reading is obtained
anywhere on a circle of given radius: the s orbital is spherically sity in the sense that when it is multiplied by dr, it gives the
symmetrical. probability of finding the electron anywhere between the two
walls of a spherical shell of thickness dr at the radius r. For a 1s
orbital, with R1,0 given in Table 17.1,
the probability density in the ground state of the atom is
4 Z 3 2 −2 Zr /a Ground-state Radial
ψ 1s 2∝ e − 2Zr /a , the reading from the probe decreases exponen-
0
P (r ) = r e 0
hydrogenic distribution
a03 (18.6)
tially as the probe is moved out along any radius but is constant orbital function
if the probe is moved on a circle of constant radius (Fig. 18.8).
Now consider the probability of finding the ground-state We can interpret this expression as follows:
electron anywhere between the two walls of a spherical shell
r Because r2 = 0 at the nucleus, P(0) = 0. Although the
of thickness dr at a radius r. The sensitive volume of the probe
probability density itself is a maximum at the nucleus,
is now the volume of the shell (Fig. 18.9), which is 4πr2dr
the radial distribution function is zero at r = 0 on
(the product of its surface area, 4πr2, and its thickness, dr).
interpretation
account of the r2 factor.
The probability that the 1s electron will be found between the
r As r→∞, P(r)→0 on account of the exponential
Physical
inner and outer surfaces of this shell is the probability den-
sity at the radius r multiplied by the volume of the probe, or term.
|ψ1s|2 × 4πr2 dr. This expression has the form P(r)dr, where r The increase in r2 and the decrease in the
exponential factor means that P(r) passes through
2 Ground-state Radial a maximum at an intermediate radius (see Fig. 18.9).
P (r ) = 4 πr 2 ψ 1s hydrogenic orbital distribution (18.5a)
function
The maximum of P(r), which can be found by differentiation,
In the following Justification we show that the more general marks the most probable distance from the nucleus at which
expression, which applies to any orbital, is the electron will be found.
The radial distribution functions of the hydrogen atom for
Radial n = 1, 2, 3 are shown in Fig.18.10. Note how the most probable
P (r ) = r 2 R(r )2 Definition distribution (18.5b)
function distance of the electron from the nucleus increases with n but it
shifts to lower values as the quantum number l increases within
where R(r) is the radial wavefunction for the orbital in question. a given shell. The small secondary maxima close to the nucleus
might seem insignificant, but in Topic 20 we see that they have
Justification 18.3 great significance for the structure of the periodic table.
The general form of the radial
distribution function
Example 18.3 Identifying the most probable radius
The probability of finding an electron in a volume element dτ
when its wavefunction is ψ = RY is R 2|Y|2 dτ with dτ = r2 dr sin θ Identify the most probable radius, r*, at which an electron will
dθ dφ (recall that R is real). The total probability of finding the be found when it occupies a 1s orbital of a hydrogenic atom of
electron at any angle at a constant radius is the integral of this atomic number Z, and tabulate the values for the one-electron
probability over the surface of a sphere of radius r, and is writ- species from H to Ne9 +.
ten P(r)dr; so
www.ebook3000.com
Atkins09819.indb 166 9/11/2013 11:15:31 AM

Radial distribution function, P/(Z/a0)

0.6
Method Find the radius at which the radial distribution func-
tion of the hydrogenic 1s orbital has a maximum value by
solving dP/dr = 0. You will need to use the rule for differen-
tiating a product of functions (Mathematical background 1: 0.4 r
d(fg)/dx = f(dg/dx) + g(df/dx)).
Answer The radia l distribution function is given in
eqn 18.6. It follows (by using the product rule again) that 0.2
⎛ (df /dr )g

f (dg /dr )
⎞
⎛
f
3
g
⎞ 3 ⎜ 2 ⎟
dP 4 Z d ⎜ 2 −2 Zr /a0 ⎟ 4 Z 2 Zr
= 3 r e = ⎜ 2re −2 Zr /a0 − a e −2 Zr /a0 ⎟ 0
dr a0 dr ⎜ ⎟ a03 ⎜ 0 ⎟ 0 1 2 3 4
⎝ ⎠ Radius, Zr/a0
⎝ ⎠
8Z 3 ⎛ Zr ⎞ −2 Zr /a0 Figure 18.9 The radial distribution function P is the probability
= r 1− e
a03 ⎜⎝ a0 ⎟⎠ density that the electron will be found anywhere in a shell of
radius r; the probability itself is Pdr, where dr is the thickness of
This function is zero where the term in parentheses is zero the shell. For a 1s electron in hydrogen, P is a maximum when
(reject r = 0, which corresponds to a minimum in P), which is at r is equal to the Bohr radius a0. The value of Pdr is equivalent
a0 to the reading that a detector shaped like a spherical shell of
r∗ = thickness dr would give as its radius is varied.
Z
That is, the most probable distance of the electron from the
nucleus in a hydrogen atom (Z = 1) is the Bohr radius itself,
a0 = 52.9 pm. For other hydrogenic species, the most probable
radius lies at 0.25
1s
Notice how the 1s orbital is drawn towards the nucleus as
0.2 2p 2s
He + Li2 + Be3 + B4 + C5 + N6 + O7 + F8 + Ne9 +

0.15
r*/pm 26.5 17.6 18.2 10.6 8.82 7.56 6.61 5.88 5.29
P(r)a0
3d 3p
0.1 3s
the nuclear charge increases. At uranium the most probable
radius is only 0.58 pm, almost 100 times closer than for hydro-
0.05
gen. (On a scale where r* = 10 cm for H, r* = 1 mm for U91+ .)
The electron then experiences strong accelerations and rela-
tivistic effects are important. 0
0 5 10 15 20
r/a0
Self-test 18.5 Find the most probable distance of an electron from
the nucleus in a hydrogenic atom when it occupies a 2s orbital.
Figure 18.10 The radial distribution functions of the hydrogen
Answer: [(3 + 51/2)a0/Z; for H: 277 pm
atom for n = 1, 2, 3 and the allowed values of l in each case.
We have taken only the first steps in our exploration of atoms (Topic 19), as well as aspects of atomic spectroscopy
atomic structure. The preceding discussion of hydrogenic (Topic 21). Furthermore, an elaboration of the concept of an
atomic orbitals provides the basis for understanding the elec- orbital is used in the description of the electronic structure of
tronic structure and periodic properties of many-electron molecules (Topics 22−30).
Atkins09819.indb 167 9/11/2013 11:15:39 AM

☐ 1. A shell of an atom consists of all the orbitals of a given ☐ 5. A boundary surface is the surface that captures a high
value of n: K(n = 1); L (n = 2); M (n = 3); N (n = 4); … proportion (typically about 90 per cent) of the electron
☐ 2. A subshell of a shell of an atom consists of all the orbit- probability density.
als with the same value of n but different values of l: ☐ 6. The radial distribution function is the probability
s (l = 0); p (l = 1); d (l = 2); f (l = 3); g (l = 4); … density of finding the electron anywhere at a distance r
☐ 3. Each shell consists of n2 orbitals. from the nucleus.
☐ 4. Each subshell consists of 2l + 1 orbitals.
Ground-state (1s) wavefunction ψ = (Z 3 /π a03 )1/2 e −Zr /a0 Hydrogenic atom 18.1
Radial distribution function P(r) = 4πr2|ψ1s|2 Ground state of hydrogenic atom 18.5a
Radial distribution function P(r) = r2R(r)2 Definition 18.5b
www.ebook3000.com
Atkins09819.indb 168 9/11/2013 11:15:40 AM

TOPIC 19
Many-electron atoms
The Schrödinger equation for a many-electron atom is highly
Contents complicated because all the electrons interact with one another.
19.1 The orbital approximation 169 Even for a helium atom, with its two electrons, no analytical
Brief illustration 19.1: Atomic configurations 170 expression for the wavefunctions and energies can be given, and
19.2 Factors affecting electronic structure 170 we are forced to make approximations. We shall adopt a simple
(a) Spin 170 approach, called the ‘orbital approximation’, based on the struc-
Brief illustration 19.2: Spin 171 ture of hydrogenic atoms and the energies of orbitals.
(b) The Pauli principle 171
Brief illustration 19.3: The Pauli principle 172
(c) Penetration and shielding 172
Example 19.1: Analysing the extent 19.1 The orbital approximation
of penetration 173
19.3 Self-consistent field calculations 174 The wavefunction of a many-electron atom is a very compli-
Checklist of concepts 175 cated function of the coordinates of all the electrons, and we
Checklist of equations 175 should write it ψ(r1,r2,…), where ri is the vector from the
nucleus to electron i. However, in the orbital approximation
we suppose that a reasonable first approximation to this exact
wavefunction is obtained by thinking of each electron as occu-
pying its ‘own’ orbital, and write the product
➤ Why do you need to know this material? ψ (r1 , r2 ,…) = ψ (r1 )ψ (r2 )… Orbital approximation (19.1)
Atoms are the currency of chemistry, and although
hydrogenic atoms provide an excellent introduction to We can think of the individual orbitals as resembling the hydro-
atomic orbitals, it is essential to see how to adapt those genic orbitals of Topic 18, but corresponding to nuclear charges
concepts to the description of the electronic structure modified by the presence of all the other electrons in the atom.
of many-electron atoms as a basis for understanding This description is only approximate, as explained in the fol-
chemical periodicity, which is treated in Topic 20. lowing Justification, but it is a useful model for discussing the
chemical properties of atoms, and is the starting point for more
➤ What is the key idea? sophisticated descriptions of atomic structure.
Electrons occupy orbitals in such a way as to achieve the
lowest total energy subject to the requirements of the Justification 19.1 The orbital approximation
Pauli principle.
The orbital approximation would be exact if there were no
➤ What do you need to know already? interactions between electrons. To demonstrate the validity of
You need to be familiar with the concept of atomic orbitals this remark for a two-electron atom, we need to consider a sys-
(Topic 18), which are the basis of the discussion in this tem in which the hamiltonian for the energy is the sum of two
Topic. The introduction to electron spin makes use of contributions, one for electron 1 and the other for electron 2:
some of the conclusions about angular momentum in
Topic 14. H = H 1 + H 2
Atkins09819.indb 169 9/11/2013 11:15:42 AM

In an actual atom (such as helium), there is an additional term (a) Spin

corresponding to the interaction of the two electrons, but we
The quantum mechanical property of electron spin, the pos-
are ignoring that term. We now show that if ψ(r1) is an eigen-
session of an intrinsic angular momentum, was identified by
function of H 1 with energy E1, and ψ(r2) is an eigenfunction
of H 2 with energy E2 , then the product ψ(r1,r2) = ψ(r1)ψ(r2) is the experiment performed in 1921 by Otto Stern and Walther
Gerlach, who shot a beam of silver atoms through an inhomo-
an eigenfunction of the combined hamiltonian H . To do so
geneous magnetic field (Topic 14). Stern and Gerlach observed
we write
two bands of Ag atoms in their experiment. This observa-
H ψ (r1 , r2 ) = (H 1 + H 2 )ψ (r1 )ψ (r2 ) = H 1ψ (r1 )ψ (r2 ) +ψ (r1 )H 2ψ (r2 ) tion seems to conflict with one of the predictions of quantum
= E1ψ (r1 )ψ (r2 ) +ψ (r1 )E2ψ (r2 ) = (E1 + E2 ) ψ (r1 )ψ (r2 ) mechanics, because an angular momentum l gives rise to 2l + 1
= Eψ (r1 , r2 ) orientations, which is equal to 2 only if l = 12 , contrary to the
conclusion that l must be an integer. The conflict was resolved by
where E = E1 + E2. This is the result we need to prove. However, the suggestion that the angular momentum they were observ-
if the electrons interact (as they do in fact), then the proof ing was not due to orbital angular momentum (the motion of an
fails; nevertheless, it remains a reasonable and almost uni- electron around the atomic nucleus) but arose instead from an
versally used starting point for the discussion of atomic intrinsic angular momentum of the electron, which classically
structure. can be thought of as the rotation of the electron on its own axis.
This intrinsic angular momentum, or ‘spin’, also emerged when
Dirac combined quantum mechanics with special relativity and
The orbital approximation allows us to express the electronic established the theory of relativistic quantum mechanics.
structure of an atom by reporting its configuration, the list of The spin of an electron does not have to satisfy the same
occupied orbitals (usually, but not necessarily, in its ground boundary conditions as those for a particle circulating around a
state). central point, so the quantum number for spin angular momen-
tum is subject to different restrictions. To distinguish this spin
Brief illustration 19.1
angular momentum from orbital angular momentum we use the
Atomic configurations spin quantum number s (in place of the l in Topic 14; like l, s
As the ground state of a hydrogenic atom consists of the single is a non-negative number) and ms, the spin magnetic quantum
electron in a 1s orbital, we report its configuration as 1s1 (read number, for the projection on the z–axis. The magnitude of the
‘one s one’). The He atom has two electrons. We can imagine spin angular momentum is {s(s + 1)}1/2ħ and the component msħ
forming the atom by adding the electrons in succession to the is restricted to the 2s + 1 values ms = s, s – 1, …, –s. To account for
orbitals of the bare nucleus (of charge 2e). The first electron Stern and Gerlach’s observation, s = 12 and ms = ± 12 .
occupies a 1s hydrogenic orbital, but because Z = 2 that orbital
A note on good practice You will sometimes see the quantum
is more compact than in H itself. The second electron joins
number s used in place of ms, and written s = ± 12 . That is
the first in the 1s orbital, so the electron configuration of the
wrong: like l, s is never negative and denotes the magnitude
ground state of He is 1s2 (‘one s two’).
of the spin angular momentum. For the z-component, use m s.
Self-test 19.1 Show that for the actual hamiltonian for a
helium atom, with electron–electron repulsion included, that The picture of spin as an actual spinning motion can be
the proof in Justification 19.1 fails. very useful when used with care. However, that is a classi-
Answer: e2/r12 term interferes with the argument cal picture of a quantum mechanical property, and it is better
to regard spin as an intrinsic property of the electron, like its
rest mass and its charge, with every electron having exactly the
same unchangeable, characteristic value. On the vector model
19.2Factors affecting electronic of angular momentum (Topic 14), the spin may lie in two dif-
ferent orientations (Fig. 19.1). One orientation corresponds to
structure ms = + 12 (this state is often denoted α or ↑); the other orienta-
tion corresponds to ms = – 12 (this state is denoted β or ↓).
It is tempting to suppose that the electronic configurations of the Other elementary particles have characteristic spin. For
atoms of successive elements with atomic numbers Z = 3, 4, …, example, protons and neutrons are spin- 12 particles (that is,
and therefore with Z electrons, are simply 1sZ. That, however, is s = 12 ) and invariably spin with the same angular momentum.
not the case. The reason lies in two aspects of nature: that elec- Because the masses of a proton and a neutron are so much
trons possess ‘spin’ and must obey the very fundamental ‘Pauli greater than the mass of an electron, yet they all have the same
principle’. spin angular momentum, the classical picture would be of these
www.ebook3000.com
Atkins09819.indb 170 9/11/2013 11:15:48 AM

19 Many-electron atoms 171
ms = –½
ms = +½
ms = +½
Figure 19.1 The vector representation of the spin of an ms = –½

electron. The length of the side of the cone is 31/2/2 units and
the projections are ± 21 units.
two particles spinning much more slowly than an electron. Figure 19.2 Electrons with paired spins have zero resultant
Some mesons are spin-1 particles (that is, s = 1, as are some spin angular momentum. They can be represented by two
atomic nuclei), but for our purposes the most important spin-1 vectors that lie at an indeterminate position on the cones
particle is the photon. The importance of photon spin in spec- shown here, but wherever one lies on its cone, the other points
troscopy is explained in Topic 40; proton spin is the basis of in the opposite direction; their resultant is zero.
Topic 47 (Magnetic resonance).
Electrons with paired spins, denoted ↑↓, have zero net spin
Brief illustration 19.2 angular momentum because the spin of one electron is can-
Spin
celled by the spin of the other. Specifically, one electron has
The magnitude of the spin angular momentum, like any angu- ms = + 12 , the other has ms = – 12 , and they are orientated on their
lar momentum, is {s(s + 1)}1/2ħ. For any spin- 12 particle, not respective cones so that the resultant spin is zero (Fig. 19.2).
only electrons, this angular momentum is ( 34 )1/2ħ = 0.866ħ, or The exclusion principle is the key to the structure of complex
9.13 × 10−35 J s. The component on the z-axis is msħ, which for a atoms, to chemical periodicity, and to molecular structure. It
spin- 12 particle is ± 12 ħ, or ±5.27 × 10−35 J s. was proposed by Wolfgang Pauli in 1924 when he was trying
Self-test 19.2 Evaluate the spin angular momentum of a to account for the absence of some lines in the spectrum of
photon. helium. Later he was able to derive a very general form of the
Answer: 21/2ħ = 1.49 × 10 −34 J s principle from theoretical considerations.
The Pauli exclusion principle in fact applies to any pair of
identical fermions. Thus it applies to protons, neutrons, and 13C
Particles with half–integral spin are called fermions and nuclei (all of which have spin 12 ) and to 35Cl nuclei (which have
those with integral spin (including 0) are called bosons. Thus, spin 23 ). It does not apply to identical bosons, which include
electrons and protons are fermions and photons are bosons. It photons (spin 1) and 12C nuclei (spin 0). Any number of identi-
is a very deep feature of nature that all the elementary particles cal bosons may occupy the same state.
that constitute matter are fermions whereas the elementary The Pauli exclusion principle is a special case of a general
particles that are responsible for the forces that bind fermions statement called the Pauli principle:
together are all bosons. Photons, for example, transmit the
electromagnetic force that binds together electrically charged When the labels of any two identical fermions are
particles. Matter, therefore, is an assembly of fermions held exchanged, the total wavefunction changes sign; when the
together by forces conveyed by bosons. labels of any two identical bosons are exchanged, the total
wavefunction retains the same sign.
(b) The Pauli principle By ‘total wavefunction’ is meant the entire wavefunction,
The role of spin in determining electronic structure becomes including the spin of the particles; that is, the total wave-
apparent as soon as we consider lithium, Z = 3, and its three function must be a function of the positions as well as spins
electrons. The first two occupy a 1s orbital drawn even more of the particles. To see that the Pauli principle implies the
closely than in He around the more highly charged nucleus. Pauli exclusion principle, we consider the (total) wavefunc-
The third electron, however, does not join the first two in the tion for two electrons, ψ(1,2). The Pauli principle implies
1s orbital because that configuration is forbidden by the Pauli that it is a fact of nature (which has its roots in the theory
exclusion principle: of relativity) that the wavefunction must change sign if we
interchange the labels 1 and 2 wherever they occur in the
No more than two electrons may occupy any given function: ψ(2,1) = –ψ(1,2). That the Pauli principle implies
orbital, and if two do occupy one orbital, then their spins the Pauli exclusion principle is shown in the following
must be paired. Justification.
Atkins09819.indb 171 9/11/2013 11:15:51 AM

Now we see that only one of the four possible states is

Justification 19.2 The Pauli exclusion principle allowed by the Pauli principle, and the one that survives has
To derive the Pauli exclusion principle from the more funda- paired α and β spins. This is the content of the Pauli exclu-
mental Pauli principle we need to infer what spin states are sion principle. The exclusion principle is irrelevant when the
allowed when two electrons in an atom occupy the same orbital orbitals occupied by the electrons are different, and both elec-
ψ. According to the orbital approximation, the overall spatial trons may then have (but need not have) the same spin state.
wavefunction is ψ (1)ψ (2). There are several possibilities for Nevertheless, even then the overall wavefunction must still be
two spins: both electrons can be in state α, denoted α(1)α(2); antisymmetric, and must still satisfy the more general Pauli
both β, denoted β(1)β(2); and one α the other β, denoted either principle itself.
α(1)β(2) or α(2)β(1). Because we cannot tell which electron is
α and which is β, in the last case it is appropriate to express the
spin states as the (normalized) linear combinations Brief illustration 19.3 The Pauli principle
σ + (1, 2) = (1/ 21/2 ){α (1)β (2) + β (1)α (2)} (19.2a) An excited state of He has the configuration 1s12s1. One accept-
able overall wavefunction that acknowledges that we can-
σ − (1, 2) = (1/ 21/2 ){α (1)β (2) − β (1)α (2)} (19.2b) not know which electron is in which orbital is {ψ1s(1)ψ 2s(2) +
because these combinations allow one spin to be α and the ψ1s(2)ψ 2s(1)}σ − (1,2), which is antisymmetric with respect to
other β with equal probability. A stronger justification for tak- interchange of the two electrons. This wavefunction corre-
ing these two linear combinations is that they correspond to sponds to the two electrons being paired and is a so-called
eigenfunctions of the total spin operators S2 and Sz, with S = 1, ‘singlet state’ of the atom. The Self-test explores the possibili-
Ms = 0 for σ + and S = 0, MS = 0 for σ −. A crucial point is that the ties of the electrons having parallel spins.
latter combination, with zero net spin angular momentum, Self-test 19.3 Show that the 1s12s1 configuration may also give
corresponds to the two electrons being paired (↑↓). rise to a ‘triplet’ state in which the spins are parallel.
The total wavefunction of the system is the product of the Answer: ψ1s(1)ψ 2s(2) − ψ1s(2)ψ 2s(1), which is antisymmetric, also
orbital part and one of the four spin states: acknowledges that we cannot know which electron is in which orbital,
and it may be combined with α(1)α(2), β(1)β(2), or σ +(1,2)
ψ (1)ψ (2)α (1)α (2) ψ (1)ψ (2)β (1)β (2)
ψ (1)ψ (2)σ + (1, 2) ψ (1)ψ (2)σ − (1, 2)
Now we can return to lithium. In Li (Z = 3), the third elec-

The Pauli principle implies that for a total wavefunction to be
tron cannot enter the 1s orbital because that orbital is already
acceptable (for electrons), it must change sign when the elec-
full: we say the K shell is complete and that the two electrons
trons are exchanged. In each case, exchanging the labels 1
form a closed shell. Because a similar closed shell is charac-
and 2 converts the factor ψ(1)ψ(2) into ψ(2)ψ(1), which is the
same, because the order of multiplying the functions does not teristic of the He atom, we denote it [He]. The third electron
change the value of the product. The same is true of α(1)α(2) is excluded from the K shell and must occupy the next avail-
and β(1)β(2). Therefore, the first two overall products (of the able orbital, which is one with n = 2 and hence belonging to
four listed above) are not allowed, because they do not change the L shell. However, we now have to decide whether the next
sign. The combination σ +(1,2) changes to available orbital is the 2s orbital or a 2p orbital, and therefore
whether the lowest energy configuration of the atom is [He]2s1
or [He]2p1.
σ + (2,1) = (1/ 21/2 ){α (2)β (1) + β (2)α (1)} = σ + (1, 2)
because it is simply the original function written in a different (c) Penetration and shielding
order. The third overall product is therefore also disallowed.
All the terms in grey are therefore disallowed. Finally, con- In hydrogenic atoms all orbitals of a given shell are degener-
sider σ −(1,2): ate. In many-electron atoms, although orbitals of a given sub-
shell remain degenerate, the subshells themselves have different
σ − (2,1) = (1/ 21/2 ){α (2)β (1) − β (2)α (1)} energies. The difference can be traced to the fact that an elec-
tron in a many-electron atom experiences a Coulombic repul-
= −(1/ 21/2 ){α (1)β (2) − β (1)α (2)} = − σ − (1, 2)
sion from all the other electrons present. If it is at a distance r
This combination does change sign (it is ‘antisymmetric’). from the nucleus, it experiences an average repulsion that can
Therefore the (blue) product ψ (1)ψ (2)σ − (1,2) also changes be represented by a point negative charge located at the nucleus
sign under particle exchange and is acceptable. and equal in magnitude to the total charge of the electrons
www.ebook3000.com
Atkins09819.indb 172 9/11/2013 11:15:56 AM

No net effect of
Radial distribution function, P

these electrons
r
3p
Net effect equivalent 3s
to a point charge at
the nucleus
Figure 19.3 An electron at a distance r from the nucleus

experiences a Coulombic repulsion from all the electrons
within a sphere of radius r and which is equivalent to a point 0 4 8 12 16 20
Radius, Zr/a0
negative charge located on the nucleus. The negative charge
reduces the effective nuclear charge of the nucleus from Ze Figure 19.4 An electron in an s orbital (here a 3s orbital) is
to Zeffe. more likely to be found close to the nucleus than an electron
in a p orbital of the same shell (note the closeness of the
innermost peak of the 3s orbital to the nucleus at r = 0). Hence
within a sphere of radius r (Fig. 19.3). The effect of this point an s electron experiences less shielding and is more tightly
negative charge, when averaged over all the locations of the bound than a p electron.
electron, is to reduce the full charge of the nucleus from Ze to
Zeff e, the effective nuclear charge. In everyday parlance, Zeff
itself is commonly referred to as the ‘effective nuclear charge’. Example 19.1 Analysing the extent of penetration
We say that the electron experiences a shielded nuclear charge,
and the difference between Z and Zeff is called the shielding Although hydrogenic orbitals are only approximations to the
constant, σ : orbitals of many-electron atoms, their properties give some
insight into the extent of penetration. Explore the probability
Z eff = Z − σ Effective nuclear charge (19.3) of finding an electron at a distance R from the nucleus for 3s,
3p, and 3d orbitals.
The electrons do not actually ‘block’ the full Coulombic attrac- Method Use the radial distribution function (Topic 18) to
tion of the nucleus: the shielding constant is simply a way of calculate the total probability of finding the electron within
expressing the net outcome of the nuclear attraction and the a sphere of radius R by integration from r = 0 to R. The radial
electronic repulsions in terms of a single equivalent charge at wavefunctions are given in Table 17.1.
the centre of the atom.
Answer The radial distribution function is Pnl(r) = r2Rnl(r)2, so
The shielding constant is different for s and p electrons
we need to evaluate the total probability, Pnl (R), of being in a
because they have different radial distribution functions (Fig.
sphere of radius R:
19.4; also see Fig. 18.10). An s electron has a greater penetra-
R
tion through inner shells than a p electron, in the sense that it
is more likely to be found close to the nucleus than a p electron
Pnl (R) = ∫ 0
r 2 Rnl2 (r )dr
of the same shell (the wavefunction of a p orbital, remember, is for the various radial wavefunctions. For instance, the 3s
zero at the nucleus). Because only electrons inside the sphere radial wavefunction is
defined by the location of the electron contribute to shield- 3/2
1 ⎛Z⎞ ⎛ 4 Zr 4 Z 2r 2 ⎞ − Zr /3a0
ing, an s electron experiences less shielding than a p electron. R3,0 (r ) =
2431/2 ⎜⎝ a0 ⎟⎠ ⎜⎝ 6 − a + 9a2 ⎟⎠ e
0 0
Consequently, by the combined effects of penetration and
shielding, an s electron is more tightly bound than a p electron In each case the integral is best evaluated by using mathemati-
of the same shell. Similarly, a d electron penetrates less than a cal software, although hand integration is feasible (but tire-
p electron of the same shell (recall that the wavefunctions of some). The results are plotted in Fig. 19.5. If, arbitrarily, we ask
orbitals are proportional to rl close to the nucleus and there- for the probability of finding the electron within a sphere of
fore that a d orbital varies as r2 close to the nucleus, whereas a radius a0/Z, we find
p orbital varies as r), and therefore experiences more shielding.
The consequence of penetration and shielding is that the ener- 3s (l = 0) 3p (l = 1) 3d (l = 2)
gies of subshells of a shell in a many-electron atom in general Pnl (a0 / Z ) 0.0098 0.0013 0.000 006
lie in the order s < p < d < f.
Atkins09819.indb 173 9/11/2013 11:16:04 AM

electron occupies the 2s orbital. This occupation results in the

These figures show that an electron in a 3s orbital is much
more likely to be found close to the nucleus than one in a 3p ground-state configuration 1s22s1, with the central nucleus sur-
orbital, which in turn is much more likely to be found close to rounded by a complete helium-like shell of two 1s electrons,
the nucleus than one in a 3d orbital. and around that a more diffuse 2s electron. The electrons in
the outermost shell of an atom in its ground state are called the
0.020 valence electrons because they are largely responsible for the
chemical bonds that the atom forms. Thus, the valence electron
0.015
in Li is a 2s electron and its other two electrons belong to its
core, the inner electrons of the atoms.
P3l(r)
0.010
–
3s 19.3Self-consistent field
0.005
3p calculations
3d
0 The treatment we have given to the electronic configuration
0 0.5 1 1.5 2
R/Za0 of many-electron species is only approximate because it is
hopeless to expect to find exact solutions of a Schrödinger
Figure 19.5 The results obtained in Example 19.1. The graphs equation that takes into account the interaction of all the elec-
show the total probability of an electron being inside a trons with one another. However, computational techniques
sphere of radius R when it occupies a 3s, 3p, and 3d orbital of are available that give very detailed and reliable approximate
a hydrogenic atom of atomic number Z. solutions for the wavefunctions and energies. The techniques
were originally introduced by D.R. Hartree (before comput-
Self-test 19.4 Repeat the analysis for orbitals of the L shell (n = 2). ers were available) and then modified by V. Fock to take into
Answer: P2s (a0 /Z ) = 0.0343, P2 p (a0 /Z ) = 0.0036 account the Pauli principle correctly. These techniques are of
great interest to chemists when applied to molecules, and are
explained in detail in Topic 27; however, we should be aware
of the general principles at this stage too. In broad outline, the
Shielding constants for different types of electrons in atoms Hartree–Fock self-consistent field (HF-SCF) procedure is as
have been calculated from their wavefunctions obtained follows.
by numerical solution of the Schrödinger equation (Table Imagine that we have an approximate idea of the structure
19.1). We see that, in general, valence-shell s electrons do of the atom. In the Ne atom, for instance, the orbital approxi-
experience higher effective nuclear charges than p electrons, mation suggests the configuration 1s22s22p6 with the orbitals
although there are some discrepancies which are considered approximated by hydrogenic atomic orbitals. Now consider
in Topic 20. one of the 2p electrons. A Schrödinger equation can be writ-
We can now complete the Li story. Because the shell with ten for this electron by ascribing to it a potential energy due to
n = 2 consists of two non-degenerate subshells, with the 2s the nuclear attraction and the repulsion from the other elec-
orbital lower in energy than the three 2p orbitals, the third trons. Although the equation is for the 2p orbital, it depends
on the wavefunctions of all the other occupied orbitals in the
atom. To solve the equation, we guess an approximate form of
Table 19.1* Effective nuclear charge, Zeff = Z − σ
the wavefunctions of all the orbitals except 2p and then solve
Element Z Orbital Zeff the Schrödinger equation for the 2p orbital. The procedure is
He 2 1s 1.6875 then repeated for the 1s and 2s orbitals. This sequence of cal-
C 6 1s 5.6727
culations gives the form of the 2p, 2s, and 1s orbitals, and in
general they will differ from the set used initially to start the
2s 3.2166
calculation. These improved orbitals can be used in another
2p 3.1358
cycle of calculation, and a second improved set of orbitals and
*More values are given in the Resource section. a better energy are obtained. The recycling continues until
www.ebook3000.com
Atkins09819.indb 174 9/11/2013 11:16:06 AM

the orbitals and energies obtained are insignificantly differ- 1s

2p
ent from those used at the start of the current cycle. The solu- 2s
Radial distribution function, P

tions are then self-consistent and accepted as solutions of the
problem.
Figure 19.6 shows plots of some of the HF-SCF radial distri-
bution functions for sodium. They show the grouping of elec-
tron density into shells, as was anticipated by the early chemists,
K L M
and the differences of penetration as discussed above. These
SCF calculations therefore support the qualitative discussions
3s
that are used to explain chemical periodicity (Topic 20). They
also considerably extend that discussion by providing detailed
wavefunctions and precise energies.
0 2 4 6 8
Radius, r/a0
Figure 19.6 The radial distribution functions for the orbitals of

Na based on SCF calculations. Note the shell-like structure, with
the 3s orbital outside the inner concentric K and L shells.
☐ 1. In the orbital approximation it is supposed that each ☐ 7. The effective nuclear charge, Z effe, is the net charge
electron occupies its ‘own’ orbital. experienced by an electron allowing for electron–elec-
☐ 2. The configuration of an atom is the list of occupied tron repulsions.
orbitals. ☐ 8. Shielding is the effective reduction in charge of a
☐ 3. The Pauli exclusion principle states that no more than nucleus by surrounding electrons.
two electrons may occupy any given orbital, and if two ☐ 9. Penetration is the ability of an electron to be found
do occupy one orbital, then their spins must be paired. inside inner shells and close to the nucleus.
☐ 4. A fermion is a particle with half-integral spin quantum ☐ 10. The outermost electrons of an atom are called its
number; a boson is a particle with integral spin quan- valence electrons; its inner electrons form the atom's
tum number. core.
☐ 5. An electron is a fermion with s = 12 . ☐ 11. In the Hartree–Fock self-consistent field (HF-SCF)
☐ 6. The Pauli principle states that when the labels of any procedure the Schrödinger equation is solved numeri-
two identical fermions are exchanged, the total wave- cally and iteratively until the solutions no longer change
function changes sign; when the labels of any two iden- (to within certain criteria).
tical bosons are exchanged, the total wavefunction
retains the same sign.
Orbital approximation ψ(r1,r2,…) = ψ(r1)ψ(r2)… Valid if electron–electron interactions ignored 19.1
Effective nuclear charge Zeff = Z – σ Charge is actually Zeffe 19.3
Atkins09819.indb 175 9/11/2013 11:16:07 AM

TOPIC 20
Periodicity
and ionization energies, reflect the configurations and vary
Contents periodically with increasing atomic number.
20.1 The building-up principle 176
Brief illustration 20.1: The building-up principle 176
20.2 The configurations of the elements
Brief illustration 20.2: The configurations of ions
177
178
20.1 The building-up principle
20.3 The periodicity of atomic properties 178 The extension of the argument presented for lithium is called
(a) Atomic radii 178 the building-up principle, or the Aufbau principle, from the
(b) Ionization energies 179
German word for building up. We presume that it is famil-
(c) Electron affinities 180
iar from introductory courses. In brief, we imagine the bare
nucleus of atomic number Z, and then feed into the orbitals Z
electrons in succession. The order of occupation, which reflects
the consequences of shielding and penetration as explained in
Topic 19, is
1s 2s 2p 3s 3p 4s 3d 4p 5s 4d 5p 6s
The periodic table lies at the heart of chemistry, and it is According to the Pauli exclusion principle (Topic 19), each
essential for a chemist to understand the origins of the orbital may accommodate up to two electrons.
periodicity of the properties of the elements that the table
summarizes. Brief illustration 20.1 The building-up principle
➤ What is the key idea? Consider the carbon atom, for which Z = 6 and there are six
The periodic repetition of analogous configurations as electrons to accommodate. Two electrons enter and fill the 1s
electrons are added to nuclei accounts for the periodicity orbital, two enter and fill the 2s orbital, leaving two electrons
of the properties of the elements and the structure of the to occupy the orbitals of the 2p subshell. Hence the ground-
periodic table. state configuration of C is 1s 2 2s 2 2p2 , or more succinctly
[He]2s22p2, with [He] the helium-like 1s2 core.
➤ What do you need to know already?
Self-test 20.1 Identify the ground-state configuration of a sili-
You need to be familiar with the concepts developed con atom.
in Topic 19 concerning the features that govern the Answer: [Ne]3s23p2 with [Ne] = [He]2s22p6 = 1s22s22p6
occupation of atomic orbitals.
We can be more precise than merely specifying the subshell:

Topic 19 introduces the considerations that allow us to predict we can expect the last two electrons of carbon, as treated in
the ground-state electron configurations of atoms. That Topic Brief illustration 20.1, to occupy different 2p orbitals because
takes the story as far as lithium. This Topic extends the discus- they will then be further apart on average and repel each other
sion to all the remaining elements. It also shows how certain less than if they were in the same orbital. Thus, one electron
chemically important atomic properties, such as atomic radii can be thought of as occupying the 2px orbital and the other the
www.ebook3000.com
Atkins09819.indb 176 9/11/2013 11:16:07 AM

20 Periodicity 177
2py orbital (the x, y, z designation is arbitrary, and it would be total spin operator S2 with S = 1 as mentioned in Topic 19 and
equally valid to use the complex forms of these orbitals), and as more fully discussed in Topic 21).
the lowest energy configuration of the atom is [He]2s2 2p1x 2p1y . Now consider the values of the two combinations ψ ± when
The same rule applies whenever degenerate orbitals of a sub- one electron approaches another, and eventually r1 = r2. We see
shell are available for occupation. Thus, another rule of the that ψ− vanishes, which means that there is zero probability of
building-up principle is: finding the two electrons at the same point in space when they
have parallel spins. The decreasing probability that the elec-
Electrons occupy different orbitals of a given subshell trons approach one another in the state ψ− is called a Fermi
before doubly occupying any one of them. hole. The other combination does not vanish when the two
For instance,nitrogen (Z = 7) has the configuration electrons are at the same point in space. Because the two elec-
2
, and only when we get to oxygen (Z = 8) is trons have different relative spatial distributions depending
[He]2s 2p1x 2p1y 2p1z
on whether their spins are parallel or not, it follows that their
a 2p orbital doubly occupied, giving [He]2s2 2p2x 2p1y 2p1z . When
Coulombic interaction is different, and hence that the two
electrons occupy orbitals singly we invoke Hund's maximum
states have different energies, with the state corresponding to
multiplicity rule:
parallel spins being lower in energy.
An atom in its ground state adopts a configuration with However, we have to be cautious with this explanation, for
the greatest number of electrons with unpaired spins. it supposes that the original wavefunctions are unchanged.
Detailed numerical calculations have shown that in the spe-
The explanation of Hund's rule is subtle, but it reflects the cific case of a helium atom electrons with parallel spins are
quantum mechanical property of spin correlation, that elec- actually closer together than those with antiparallel spins.
trons with parallel spins behave as if they have a tendency to The explanation in this case is that spin correlation between
stay well apart (see the following Justification), and hence repel electrons with parallel spins allows the entire atom to shrink.
each other less. In essence, the effect of spin correlation is to Therefore, although the average separation is reduced, the
allow the atom to shrink slightly, so the electron–nucleus inter- electrons are found closer to the nucleus, which lowers their
action is improved when the spins are parallel. We can now potential energy.
conclude that in the ground state of the carbon atom, the two
2p electrons have the same spin, that all three 2p electrons in
the N atom have the same spin, and that the two 2p electrons in
different orbitals in the O atom have the same spin (the two in
the 2px orbital are necessarily paired).
The configurations of the
20.2
elements
Neon, with Z = 10, has the configuration [He]2s22p6, which
Justification 20.1 completes the L shell. This closed-shell configuration is
Spin correlation
denoted [Ne], and acts as a core for subsequent elements. The
Suppose electron 1 is described by a spatial wavefunction ψa(r1) next electron must enter the 3s orbital and begin a new shell,
and electron 2 is described by a wavefunction ψ b(r2); then, in so an Na atom, with Z = 11, has the configuration [Ne]3s1. Like
the orbital approximation, the joint wavefunction of the elec-
lithium with the configuration [He]2s1, sodium has a single s
trons is the product ψ = ψ a(r1)ψ b(r2). However, this wavefunc-
electron outside a complete core. This analysis has brought us
tion is not acceptable, because it suggests that we know which
to the origin of chemical periodicity. The L shell is completed
electron is in which orbital, whereas we cannot keep track
by eight electrons, so the element with Z = 3 (Li) should have
of electrons. According to quantum mechanics, the correct
description is either of the two following wavefunctions:
similar properties to the element with Z = 11 (Na). Likewise, Be
(Z = 4) should be similar to Z = 12 (Mg), and so on, up to the
ψ ± = (1/ 2)1/2 {ψ a (r1 )ψ b (r2 ) ±ψ b (r1 )ψ a (r2 )} noble gases He (Z = 2), Ne (Z = 10), and Ar (Z = 18).
Ten electrons can be accommodated in the five 3d orbitals,
According to the Pauli principle, because ψ + is symmetri- which accounts for the electron configurations of scandium to
cal under particle interchange, it must be multiplied by an zinc. Calculations of the type discussed in Topic 27 show that
antisymmetric spin function (the one denoted σ − in eqn for these atoms the energies of the 3d orbitals are always lower
19.2b). That combination corresponds to a spin-paired state. than the energy of the 4s orbital. However, spectroscopic results
Conversely, ψ− is antisymmetric, so it must be multiplied by show that Sc has the ground-state configuration [Ar]3d14s2,
one of the three symmetric spin states(α(1)α(2), β(1)β(2), or instead of [Ar]3d3 or [Ar]3d24s1. To understand this observa-
σ +(1,2)). These three symmetric states correspond to electrons tion, we have to consider the nature of electron–electron repul-
with parallel spins (they correspond to eigenfunctions of the sions in 3d and 4s orbitals, where the effects are particularly
Atkins09819.indb 177 9/11/2013 11:16:13 AM

achieve the specified charge. The configurations of anions of

the p-block elements are derived by continuing the building-up
procedure and adding electrons to the neutral atom until the
configuration of the next noble gas has been reached.
Energy
Brief illustration 20.2 The configurations of ions

Because the configuration of V is [Ar]3d 34s2 , the V2+ cation
has the configuration [Ar]3d3. It is reasonable that we remove
Figure 20.1 Strong electron–electron repulsions in the 3d the more energetic 4s electrons in order to form the cation, but
orbitals are minimized in the ground state of Sc if the atom it is not obvious why the [Ar]3d3 configuration is preferred in
has the configuration [Ar]3d14s2 (shown on the left) instead of V2+ over the [Ar]3d14s2 configuration, which is found in the
[Ar]3d24s1 (shown on the right). The total energy of the atom is isoelectronic Sc atom. Calculations show that the energy dif-
lower when it has the [Ar]3d14s2 configuration despite the cost ference between [Ar]3d 3 and [Ar]3d14s2 depends on Z eff. As
of populating the high energy 4s orbital. Zeff increases, transfer of a 4s electron to a 3d orbital becomes
more favourable because the electron–electron repulsions are
compensated by attractive interactions between the nucleus
finely balanced because there is only a small difference in and the electrons in the spatially compact 3d orbital (see the
energy between the orbitals. The most probable distance of a 3d radial distribution function in Fig. 18.10). Indeed, calcula-
3d electron (with no radial nodes) from the nucleus is less than tions reveal that for a sufficiently large Zeff, [Ar]3d3 is lower in
that for a 4s electron (with three radial nodes), so two 3d elec- energy than [Ar]3d14s2. This conclusion explains why V2+ has
trons repel each other more strongly than two 4s electrons. As a an [Ar]3d 3 configuration and also accounts for the observed
result, Sc has the configuration [Ar]3d14s2 rather than the two [Ar]4s03dn configurations of the M2+ cations of Sc through Zn.
alternatives, for then the strong electron–electron repulsions in Self-test 20.2 Identify the configuration of the O2– ion.
the 3d orbitals are minimized. The total energy of the atom is Answer: [He]2s22p6 = [Ne]
least despite the cost of allowing electrons to populate the high
energy 4s orbital (Fig. 20.1).
The effect just described is generally true for scandium
through zinc, so their electron configurations are of the form The periodicity of atomic
20.3
[Ar]3dn4s2, where n = 1 for scandium and n = 10 for zinc. Two
notable exceptions, which are observed experimentally, are Cr,
properties
with electron configuration [Ar]3d54s1, and Cu, with electron
configuration [Ar]3d104s1. The theoretical basis of the exceptions Three atomic properties are of considerable importance for
represented by Cr and Cu is the additional energy lowering char- determining the chemical properties of the elements, namely
acteristic of half-filled and completely filled d subshells. atomic radii, ionization energies, and electron affinities. All
At gallium, the building-up principle is used in the same way three show periodic variation with increasing atomic number.
as in preceding periods. Now the 4s and 4p subshells consti-
tute the valence shell, and the period terminates with krypton.
Because 18 electrons have intervened since argon, this period is
(a) Atomic radii
the first ‘long period’ of the periodic table. The existence of the An atom does not have a precise radius because far from the
d-block elements (the ‘transition metals’) reflects the stepwise nucleus the electron density falls off only exponentially (but
occupation of the 3d orbitals, and the subtle shades of energy sharply). However, we can expect atoms with numerous elec-
differences and effects of electron–electron repulsion along trons to be larger, in some sense, than atoms that have only a
this series gives rise to the rich complexity of inorganic d-metal few electrons. The atomic radius of an element is defined as
chemistry. A similar intrusion of the f orbitals in Periods 6 and half the distance of neighboring atoms in a solid (such as Cu)
7 accounts for the existence of the f block of the periodic table or, for non-metals, in a homonuclear molecule (such as H2 or
(the lanthanoids and actinoids). S8). The data in Table 20.1 show that atomic radii increase down
We derive the configurations of cations of elements in the s, a group, and that they decrease from left to right across a period.
p, and d blocks of the periodic table by removing electrons from These trends are readily interpreted in terms of the electronic
the ground-state configuration of the neutral atom in a spe- structure of the atoms. On descending a group, the valence
cific order. First, we remove valence p electrons, then valence electrons are found in orbitals of successively higher principal
s electrons, and then as many d electrons as are necessary to quantum number. The atoms within the group have a greater
www.ebook3000.com
Atkins09819.indb 178 9/11/2013 11:16:13 AM

20 Periodicity 179
Table 20.1 Atomic radii of main-group elements, r/pm and an increase in Zeff. That loss often leaves behind only the
much more compact closed shells of electrons. Once these gross
Li 157 Be 112 B 88 C 77 N 74 O 66 F 64 differences are taken into account, the variation in ionic radii
Na 191 Mg 160 Al 143 Si 118 P 110 S 104 Cl 99 through the periodic table mirrors that of the atoms.
K 235 Ca 197 Ga 153 Ge 122 As 121 Se 117 Br 114
Rb 250 Sr 215 In 167 Sn 158 Sb 141 Te 137 I 133
(b) Ionization energies
Cs 272 Ba 224 Tl 171 Pb 175 Bi 182 Po 167
The minimum energy necessary to remove an electron from
a many-electron atom in the gas phase is the first ionization
number of completed shells of electrons in successive periods energy, I1, of the element. The second ionization energy, I2,
and hence their radii increase down the group. Across a period, is the minimum energy needed to remove a second electron
the valence electrons enter orbitals of the same shell; how- (from the singly charged cation). Some numerical values are
ever, the increase in effective nuclear charge across the period given in Table 20.2.
draws in the electrons and results in progressively more com- As will be familiar from introductory chemistry, ioniza-
pact atoms. The general increase in radius down a group and tion energies show periodicities (Fig. 20.2). Lithium has a low
decrease across a period should be remembered as they corre- first ionization energy because its outermost electron is well
late well with trends in many chemical properties. shielded from the nucleus by the core electrons (Zeff = 1.3, com-
Period 6 shows an interesting and important modification to pared with Z = 3). The ionization energy of beryllium (Z = 4) is
these otherwise general trends. The metallic radii in the third greater but that of boron is lower because in the latter the out-
row of the d block are very similar to those in the second row, ermost electron occupies a 2p orbital and is less strongly bound
and not significantly larger as might be expected given their than if it had been a 2s electron.
considerably greater numbers of electrons. For example, the The ionization energy increases from boron to nitrogen on
atomic radii of Mo (Z = 42) and W (Z = 74) are 140 and 141 pm, account of the increasing nuclear charge. However, the ioniza-
respectively, despite the latter having many more electrons. tion energy of oxygen is less than would be expected by sim-
The reduction of radius below that expected on the basis of a ple extrapolation. The explanation is that at oxygen a 2p orbital
simple extrapolation down the group is called the lanthanide must become doubly occupied, and the electron–electron
contraction. The name points to the origin of the effect. The
elements in the third row of the d block (Period 6) are preceded Table 20.2* First and second ionization energies
by the elements of the first row of the f block, the lanthanoids,
Element I1/(kJ mol−1) I2/(kJ mol−1)
in which the 4f orbitals are being occupied. These orbitals have
poor shielding properties and so the valence electrons expe- H 1312
rience more attraction from the nuclear charge than might He 2372 5250
be expected. The repulsions between electrons being added Mg 738 1451
on crossing the f block fail to compensate for the increas- Na 496 4562
ing nuclear charge, so Zeff increases from left to right across a
*More values are given in the Resource section.
period. The dominating effect of the latter is to draw in all the
electrons and hence to result in a more compact atom. A simi-
30
lar contraction is found in the elements that follow the d block
for the same reasons. For example, although there is a substan- He
tial increase in atomic radius between C and Si (77 and 118 pm, Ne
Ionization energy, I/eV
respectively), the atomic radius of Ge (122 pm) is only slightly 20

greater than that of Si. Ar
Kr
All monatomic anions are larger than their parent atoms and Xe
Hg Rn
all monatomic cations are smaller than their parent atoms (in
10
some cases, markedly so). The increase in radius of an atom on
anion formation is a result of the greater electron–electron repul-
sions that occur when an additional electron is added to form an Li Na Rb Cs Tl
anion. There is also an associated decrease in the value of Zeff. The 0
0 20 40 60 80 100
smaller radius of a cation compared with its parent atom is a con-
Atomic number, Z
sequence not only of the reduction in electron–electron repul-
sions that follows electron loss but also of the fact that cation Figure 20.2 The first ionization energies of the elements
formation typically results in the loss of the valence electrons plotted against atomic number.
Atkins09819.indb 179 9/11/2013 11:16:15 AM

repulsions are increased above what would be expected by Table 20.3* Electron affinities, Eea/(kJ mol−1)
simple extrapolation along the row. In addition, the loss of a
2p electron results in a configuration with a half-filled sub- Cl 349
shell (like that of N), which is an arrangement of low energy, so F 322
the energy of O+ + e− is lower than might be expected, and the H 73
ionization energy is correspondingly low too. (The kink is less O 141 O− −844
pronounced in the next row, between phosphorus and sulfur *More values are given in the Resource section.
because their orbitals are more diffuse.) The values for oxygen,
fluorine, and neon fall roughly on the same line, the increase The incoming electron is repelled by the charge already present.
of their ionization energies reflecting the increasing attraction Electron affinities are also small, and may be negative, when an
of the more highly charged nuclei for the outermost electrons. electron enters an orbital that is far from the nucleus (as in the
The outermost electron in sodium is 3s. It is far from the heavier alkali metal atoms) or is forced by the Pauli principle to
nucleus, and the latter's charge is shielded by the compact, occupy a new shell (as in the noble gas atoms).
complete neon-like core. As a result, the ionization energy of The values of ionization energies and electron affinities can
sodium is substantially lower than that of neon. The periodic help us to understand a great deal of chemistry and, through
cycle starts again along this row, and the variation of the ioni- chemistry, biology. We can now begin to see why carbon is
zation energy can be traced to similar reasons. an essential building block of complex biological structures.
Among the elements in Period 2, carbon has intermediate val-
ues of the ionization energy and electron affinity, so it can share
(c) Electron affinities electrons (that is, form covalent bonds) with many other ele-
The electron affinity, Eea, is the energy released when an elec- ments, such as hydrogen, nitrogen, oxygen, sulfur, and, more
tron attaches to a gas-phase atom (Table 20.3). In a common, importantly, other carbon atoms. As a consequence, such net-
logical, but not universal convention (which we adopt), the works as long carbon–carbon chains (as in lipids) and chains of
electron affinity is positive if energy is released when the elec- peptide links can form readily. Because the ionization energy
tron attaches to the atom. and electron affinity of carbon are neither too high nor too low,
Electron affinities are greatest close to fluorine, for the the bonds in these covalent networks are neither too strong nor
incoming electron enters a vacancy in a compact valence shell too weak. As a result, biological molecules are sufficiently stable
and can interact strongly with the nucleus. The attachment of to form viable organisms but are still susceptible to dissociation
an electron to an anion (as in the formation of O2– from O−) and rearrangement.
invariably requires the absorption of energy, so Eea is negative.
☐ 1. The building-up (Aufbau) principle is the procedure ☐ 4. The lanthanide contraction is the reduction in atomic
for filling atomic orbitals that leads to the ground-state radius of elements following the lanthanoids.
configuration of an atom. ☐ 5. The ionization energy is the minimum energy nec-
☐ 2. Hund’s maximum multiplicity rule states that an atom essary to remove an electron from an atom in the gas
in its ground state adopts a configuration with the phase.
greatest number of electrons with unpaired spins. ☐ 6. The electron affinity is the energy released when an
☐ 3. Atomic radii typically decrease across a period and electron attaches to an atom in the gas phase.
increase down a group.
www.ebook3000.com
Atkins09819.indb 180 9/11/2013 11:16:15 AM

TOPIC 21
Atomic spectroscopy
Contents ➤ What do you need to know already?

21.1 The spectrum of hydrogen 181 You need to be aware of the discussion of the energy levels
(a) The spectral series 181 of hydrogenic atoms (Topic 17) and how that discussion is
Example 21.1: Calculating the shortest and extended to account for the structures of many-electron
longest wavelength lines in a series 182 atoms (Topic 19). The discussion of selection rules makes
(b) Selection rules 182 use of the discussion of transitions (Topic 16).
Brief illustration 21.1: Selection rules 182
21.2 Term symbols 183
(a) The total orbital angular momentum 183 The general idea behind atomic spectroscopy is straightfor-
Example 21.2: Deriving the total orbital angular ward: lines in the spectrum (in either emission or absorption)
momentum of a configuration 183 occur when the atom undergoes a transition with a change
(b) The total spin angular momentum 184 of energy |ΔE|, and emits or absorbs a photon of frequency
Brief illustration 21.2: The multiplicities of terms 185 ν = |ΔE|/h and wavenumber ␯ = ΔE /hc. Hence, we can expect
(c) The total angular momentum 185 the spectrum to give information about the energies of elec-
Example 21.3: Deriving term symbols 185 trons in atoms. However, in many-electron atoms the actual
(d) Spin–orbit coupling 186 energy levels are not given solely by the energies of the orbit-
Brief illustration 21.3: Spin–orbit coupling energy 187 als, because the electrons interact with one another in various
Brief illustration 21.4: Fine structure 187 ways, and it is necessary to consider contributions to the energy
21.3 Selection rules of many-electron atoms 187 beyond those of the orbital approximation.
Brief illustration 21.5: Selection rules 188
Checklist of equations 189 21.1 The spectrum of hydrogen
When an electric discharge is passed through gaseous hydro-
gen, the H2 molecules are dissociated and the energetically
excited H atoms that are generated emit light of discrete fre-
➤ Why do you need to know this material? quencies, producing a spectrum of a series of ‘lines’ (Fig. 21.1).
Atomic spectroscopy not only inspired the development
of quantum mechanics but provides detailed information
about the energies of electrons in atoms. The labels that
(a) The spectral series
specify atomic states, and their molecular counterparts, The Swedish spectroscopist Johannes Rydberg noted (in 1890)
play a crucial role in spectroscopy, in the discussion that all the lines are described by the expression
of magnetic properties, in photochemistry, and in the
description of the operation of lasers. ⎛ 1 1⎞
␯ = R H ⎜ 2 − 2 ⎟ R H = 109 677cm −1 Rydberg expression (21.1)
⎝ n1 n2 ⎠
➤ What is the key idea?
Transitions take place between allowed energy states subject with n1 = 1 (the Lyman series), 2 (the Balmer series), 3 (the
to selection rules that stem from the angular momentum of Paschen series), and 4 (the Brackett series), and that in each case
the photon and the conservation of angular momentum. n2 = n1 + 1, n1 + 2, …. The constant R H is now called the Rydberg
constant for the hydrogen atom (Topic 17).
Atkins09819.indb 181 9/11/2013 11:16:18 AM

Focus 4 On: Approximation Methods

Uploaded by

Copyright:

Available Formats

Focus 4 On: Approximation Methods

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

Focus 4 On: Approximation Methods

Uploaded by

Copyright:

Available Formats

FOCUS 4 ON Approximation methods

Atkins09819.indb 131 9/11/2013 11:09:28 AM

Contents the latter is then used to develop approximations to the

15.1 Perturbation expansions

Atkins09819.indb 132 9/11/2013 11:09:33 AM

eigenvalues, respectively, of the model system.

See Example 15.2 for an illustration of how this expression is

Self-test 15.1 The vibrational motion of a certain diatomic

Atkins09819.indb 133 9/11/2013 11:09:47 AM

r The second-order correction to the energy is given by By comparing powers of λ, we find

the order of the correction. At the end of the calculation, we 

E0 = E0(0) + λE0(1) + λ 2 E0(2) + ∫ψ (0 )*  (0 ) (0 )

When these expressions are inserted into the Schrödinger

(H (0) + λ H (1) )(ψ 0(0) + λψ 0(1) + λ 2ψ 0(2) + ) =

Atkins09819.indb 134 9/11/2013 11:10:07 AM

it corresponds to n = 1). Use Integral T.3 listed in the Resource c k Ek

Answer The perturbation hamiltonian is H (1) = −ε sin(πx /L)

Although this expression is for the coefficient c k, it applies to

if ∫ψ n(0)* H (1) ψ 0(0) dτ = 0 state.

Atkins09819.indb 135 9/11/2013 11:10:27 AM

15.4The second-order correction

1 0 isolate the second-order energy correction E0(2) we multiply

Figure 15.3 The ground-state wavefunction corrected to 

Atkins09819.indb 136 9/11/2013 11:10:44 AM

(8ε /15π)2 64ε 2me L2

Self-test 15.4 Evaluate the contribution from the n = 5 state to

Atkins09819.indb 137 9/11/2013 11:10:52 AM

Property Equation Comment Equation number

Division of the hamiltonian H = H (0) + H (1) H (1) is the perturbation 15.1

First-order energy correction

Wavefunction corrected to first order ψ 0 =ψ 0(0) +

Atkins09819.indb 138 9/11/2013 11:10:58 AM

Ψ (r , t )*Ψ (r , t ) = (ψ (r )* eiEt / )(ψ (r )e − iEt / ) =ψ (r )*ψ (r ).

➤ Why do you need to know this material?

Atkins09819.indb 139 9/11/2013 11:11:01 AM

Atkins09819.indb 140 9/11/2013 11:11:15 AM

τ0 long When we form the square modulus of c n , the factor eiωn 0t

 (1) (t ) = H (1) (1 − e −t /τ 0 ) into eqn 16.4b +(ei(ωn 0 −ω )t −1)/[i(ω n0 −ω)]}

The probability that at time t the perturbation has induced a

= (1 − e −t /τ 0 )H n(10) The transition rate, w, is the rate of change of the probability

Atkins09819.indb 141 9/11/2013 11:11:38 AM

2 Example 16.2 Analysing spectroscopic transitions

∫ result in a transition from the ground vibrational state of the

it is allowed if any one of the components of the transition ⎝α ⎠

terms of the changes in quantum numbers. A gross selection =

Because the integral is not zero, the spectroscopic transition

Figure 16.2 (a) When a 1s electron becomes a 2s electron,

Atkins09819.indb 142 9/11/2013 11:11:45 AM

Ψ (r , t ) =ψ (r )e − iEt / × e − t /2τ Brief illustration 16.1 Energy–time uncertainty

w f ←i = Bfi ρ Stimulated absorption Transition rate (16.10)

The constant Bfi is the Einstein coefficient of stimulated

(a) Time (b) Frequency 8πh␯3 /c 3

Atkins09819.indb 143 9/11/2013 11:11:52 AM

N i Bfi ρ = N f ( A + Bif ρ ) Thermal equilibrium (16.16)

We have used the Boltzmann expression (Topic 2) for the ratio

Wf →i = N f w f →i = N f ( A + Bif ρ ) Total emission rate (16.15)

Atkins09819.indb 144 9/11/2013 11:12:03 AM

Time-dependent Schrödinger equation H (t )Ψ (r ,t ) = i ∂Ψ (r ,t )/∂ t 16.2

the order of the correction. At the end of the calculation, we

E0 = E0(0) + λE0(1) + λ 2 E0(2) + ∫ψ (0 )* (0 ) (0 )

(H (0) + λ H (1) )(ψ 0(0) + λψ 0(1) + λ 2ψ 0(2) +) =

it corresponds to n = 1). Use Integral T.3 listed in the Resource c k Ek

Answer The perturbation hamiltonian is H (1) = −ε sin(πx /L)

if ∫ψ n(0)* H (1) ψ 0(0) dτ = 0 state.

Figure 15.3 The ground-state wavefunction corrected to

Division of the hamiltonian H = H (0) + H (1) H (1) is the perturbation 15.1

Ψ (r , t )Ψ (r , t ) = (ψ (r ) eiEt / )(ψ (r )e − iEt / ) =ψ (r )*ψ (r ).

(1) (t ) = H (1) (1 − e −t /τ 0 ) into eqn 16.4b +(ei(ωn 0 −ω )t −1)/[i(ω n0 −ω)]}

Time-dependent Schrödinger equation H (t )Ψ (r ,t ) = i ∂Ψ (r ,t )/∂ t 16.2

x Table 18.1 Hydrogenic atomic orbitals

⎛ (df /dr )g