Solutions:, The Exponential

Solutions
Chapter 2
2.1 In a diatomic molecule, there exists an interatomic distance r0 at which the

potential energies of the ionic and the covalent bonding types are equal, i. e. at which
the potential curves intersect:
Pi (r0 ) = Pn (r0 ) . (1)
At shorter distances, the ionic molecule is more stable. For r , the exponential
term for the repulsion in the two given potential functions can be neglected, so that
in the case of the ionic molecule, essentially only the Coulomb interaction remains:
e2
Pi (r) = − + IE . (2)
4π ε0 r
With the simplification C = 0, it follows that Pn (r) = 0, and thus from (1) and
(2) we obtain the determining equation for the distance r0 :
e2
− + IE = 0 , (3)
4π ε0 r
e2
r0 = . (4)
4π ε0 I E
With I E = 1.42 eV, one finds
r0 = 1.01 · 10−9 m .
2.2 In order to obtain the best possible resolution in the transmission electron
microscope, i. e. the smallest possible value of dp , one chooses an electron beam
intensity (dose I) which approaches the threshold for damage of the molecules,
(I = Ncr ). Furthermore, the necessary signal/noise ratio S/N is minimised. If these
two factors have been optimised, the remaining possibility is to increase the contrast
Cn by stacking many monolayers upon each other. Cn is proportional to the square
root of the number n of layers, so that with the contrast C1 = C for a single
monolayer, we find:
√
Cn = n C . (1)
2 Solutions
The theoretical point resolution can be calculated using the formula of Freyer:
S/N
dp = √ √ , (2)
nC f I
where f is the electron yield factor (0.25). At the required resolution of dp = 1 Å,
this yields the number n of monolayers which will generate the necessary contrast:
(S/N)2
n= . (3)
dp2 C 2 f I
With the given data for the microscope, one thus finds
n = 100
which in the case of hexadecachloro copper phthalocyanine corresponds to overall

layer thicknesses of 340 to 380 Å.
2.3 Due to the potential difference U d − Us between the detector of the time-of-
flight mass spectrometer and the substrate, which is at a distance d from the detector,
an electric field E acts and accelerates the ionised molecules. If q = n e is the charge
of an n-fold ionised molecule and M is its molar mass, then for the acceleration a
we find:
qE q (U d − Us ) n e (U d − Us )
a= = = . (1)
M Md Md
The time T for the uniformly accelerated motion is given by

2d
T= (2)
a
and therefore

2d 2 M
T= . (3)
n e (U d − Us )
If we choose the time of the ionisation as the zero point of the time scale,
then the apparatus registers a maximum at the time T (in the ideal case as a δ-
function), which indicates the arrival of the large molecule. An accelerating voltage
of U d − Us = 2 kV, on the assumption that every molecule is only singly ionised
(n = 1), yields
T = 5.1 · 10−5 s .
The time dependence of the current from the detector in fact exhibits a number
of maxima for t < T , which are caused by solvent molecules as well as by fragments
of the large molecules.
Chapter 2 3
2.4 The number of C60 molecules which form a Langmuir film with a surface
area A can be computed from the volume V of the solution with a molar particle
concentration of c:
n = cV . (1)
If the film consists of a single monomolecular layer, the average area AM per
molecule is found to be
A A
AM = = . (2)
n cV
With the experimental values given, this means that
AM = 2.16 · 10−19 m2 = 21.6 Å2 .
The area of a C60 molecule defined by the van der Waals interaction is 78.5 Å2 .
Considering that even for spherical closest packing, the space per molecule required
is somewhat more than this area, one arrives at a value which is about four times
greater per molecule than the available area AM . This means that the film consists
of four stacked monolayers of C60 molecules.
2.5 In the experimental method given, the osmotic pressure is compensated by the
hydrostatic pressure p of a solution of density = 0.867 g/cm3 and height h:
p = gh. (1)
In equilibrium, π = p, we thus have
g h = c NA k T (1 + α c) . (2)
Calculating the mass concentration c from the molar particle concentration c is

carried out using the molecular mass M according to
n c
c= = (3)
V M
Inserting (3) into (2) leads to the equation
h NA k T NA k T α

= + c (4)
c gM g M2
h
= a + b c . (5)
c
Plotting h/c against c yields a straight line with a y-intercept of a =
0.325 cm4 /mg and a slope of b = 0.202 cm7 /mg2 .
From a, using (4) and (5), we obtain the molecular mass M of the polymer:
g
M = 89606 .
mole
4 Solutions
The slope b yields the virial coefficient
α = 5.57 · 107 cm3 /mole
and thus via α = 1

2 NA vp the occupied volume
vp = 1.85 · 10−16 cm3 .
The volume occupied by a spherical molecule of radius r is a sphere with a

radius 2r around the centre of the molecule, since the centre of every neighbouring
molecule has to be at least at the distance 2r from the centre of the given molecule.
It follows from this consideration that
4 32
vp = π (2r)3 = π r3 . (6)
3 3
From this, we calculate a molecular radius of
r = 1.77 · 10−6 cm = 177 Å .
2.6 The specific heat C V of a material at constant volume is defined as the derivative
of its internal energy U with respect to the temperature:

dU
CV = . (1)
dT dV =0
Using the differential expression for the internal energy,
dU = T dS − p dV + µ dN , (2)
and dN = 0, we obtain from (1) and from the entropy S per degree of freedom,
1
S= k ln T (3)
2
the specific heat per molecule and degree of freedom:
∂S k
CV = T = . (4)
∂T 2
For a molecule with f degrees of freedom, we thus find
k
CV = f · . (5)
2
In analogy to the derivation of C V from the internal energy U, one finds the
specific heat C P at constant pressure from the enthalpy H = U + pV . Per molecule
we obtain
C P = CV + k . (6)
Chapter 2 5
In an ideal gas, the velocity of sound cS is given by the equation

γ RT
cS = (7)
M
with γ = C P /C V and R = NA k. To compute cS at a temperature T , we thus require
in addition to the molecular mass M the value of the ratio γ , which from (5) and (6)
follows directly from the number f ∗ of excited degrees of freedom of translation,
rotation, and vibration:
k ∗
CV = f + f R∗ + f S∗ . (8)
2 T
If the temperature is too low to allow the population of the vibrational levels,
i. e. f ∗ = 0, then from (5), (6) and (8) we obtain for γ the relation
2
γ =1+ . (9)
f T∗ + f R∗
a) In an ideal diatomic gas, three degrees of freedom of translation and two
degrees of freedom of rotation are excited. We thus obtain γ = 75 and therefore for
the velocity of sound

7RT
cS = . (10)
5M
b) For an ideal linear triatomic gas, the same result holds, so that in this case,
also,

7RT
cS = . (11)
5M
c) In an ideal nonlinear triatomic gas, the third degree of freedom of rotation
can also be excited, so that γ = 43 and

4RT
cS = . (12)
3M
According to the CRC Handbook of Chemistry and Physics, air is composed
of 78.084% N2 , 20.946% O2 , 0.934% Ar, and 0.033% CO2 . For an estimate of the
velocity of sound in air, only the two main components N2 and O2 need be considered,
since both are diatomic molecules, so that (10) can be used with the weighted mean
of their molecular masses. With MN2 = 28.014 g/mole and MO2 = 31.998 g/mole,
it follows that
g(N2 ) MN2 + g(O2 ) MO2 g
M= = 28.857 (13)
g(N2 ) + g(O2 ) mole
and, from (7) with T = 298 K,
m
cS, Luft = 342.9 .
s
(For comparison: the measured value is 346.7 m/s.)
6 Solutions
Chapter 3
3.1 a) The electrostatic energy of the system of N independent rigid dipoles due
to their interaction with the externally applied field E is given by

N
H=− p E cos θi . (1)
i=1
The distribution density of the angles θ between the dipoles and the direction of
the field, using the notation β = 1/k T , is assumed to be given by
1 −β H(θ)
f(θ) = e (2)
Z
with the density of states (partition function)
+1
Z= e−β H(θ) d N (cos θ) . (3)
−1
The mean value P of the macroscopic polarisation

N
P= p cos θi (4)
i=1
along the direction of the external field E can be found by using the distribution
function f from (2); it is
+1 N

P= f(θ) p cos θi d N (cos θ) . (5)
−1 i=1
With (1), (2) and (3) and employing the substitutions u i = cos θi and x = β p E,
this becomes
+1 +1 −1
P=Np xu
u e du xu
e du . (6)
−1 −1
After integration, one obtains the expression

1
P = N p coth x − = N p L(x) . (7)
x
Writing P as
P = N p cos θ , (8)
we find from (7) the thermal average of the angle θ between the dipoles and the field
direction:
1
cos θ = coth x − = L(x) . (9)
x
Chapter 3 7
b) The orientation of a dipole precisely along the field direction corresponds to

a thermal average of the angle θ between the dipole p and the field E of
cos θ = L(x) = 1 ; (10)
compare (9). With the limiting value
lim L(x) = 1 (11)

x→∞
this corresponds for x = pE/kT and T = 293 K to
E → ∞.
In order to orient the dipole precisely along the field direction, one would thus have
to apply an infinitely strong electric field!
c) The function coth can be expanded in a series according to
1 x x3 2 x5
coth(x) = + − + + . . . for 0 < |x| < π . (12)
x 3 45 945
For sufficiently small |x|, the series can be terminated after the first-order term
in x, thereby simplifying the Langevin function L(x) in (9) to
x
L(x) → . (13)
3
If the orientation polarisation reaches 50% of its theoretical maximum value,
then from (8) we find
P = N p cos θ = 0.5 N p (14)
and thus with (9) and (13)

x 3
L(x) = = 0.5 ⇒ x= . (15)
3 2
The simplification of L(x) is a poor approximation and allows at most a qualita-
tive estimate. The required electric field
xkT
E= (16)
p
for the given values of p and T is found to be
E = 9.8 · 108 V/m .
d) With E = 107 V/m and T = 4.2 K, one obtains
L(x) = 0.355
and thus an orientation of ca. 36%.

8 Solutions
3.2 The water molecule represents an electric dipole ( p = 1.85 D = 6.17 · 10−30
Cm) with a negative partial charge on the O atom and a positive charge between the
H atoms. In the presence of a (negative) anion, the molecule will therefore orient
itself with its H atoms pointing towards the ion.
The molecule is considered in the following as a dipole of length d, consisting
of a positive and a negative elementary charge. The electrostatic potential of a point
charge q at a distance r is
q
φ= (1)
4π ε0r
and produces a field
dφ q
E=− = . (2)
dr 4π ε0 r 2
The distance from the two point-like elementary charges q = ±e to the (also
point-like) anion is r ∓ d/2. At the position of the anion, the fields produced by the
two point charges then superpose to give
e e
E(r) = − (3)
4π ε0 (r − d/2) 2
4π ε0 (r + d/2)2
pr
E(r) = 2 (4)
2π ε0 r 2 − d 2 /4
with p = d r. From the given dipole moment, one finds a dipole length of d =
3.85 · 10−11 m, so that d 2 /4 can be neglected in comparison to r 2 :
p
E(r) = . (5)
2π ε0 r 3
For the distances given, one finds the following field strengths at the position of
the anion:
a) at r = 10−9 m : E = 1.11 · 108 V/m
b) at r = 3 · 10−10 m : E = 4.11 · 109 V/m
c) at r = 3 · 10−8 m : E = 4.11 · 103 V/m
3.3 a) From α/4π ε0 = 10.3 · 10−30 m3 , it follows that:

pind = 1.15 · 10−33 Cm = 3.4 · 10−4 D .
b) From the power density P/A = W · c and the energy density W = 12 ε0 E 2 , the
electric field strength at the laser focus is found to be
V
E = 8.7 · 109
m
and thus
pind = 2.98 D .
Chapter 3 9
3.4 The behaviour of a permanent magnetic moment µp in a magnetic field B

can be described in precise analogue to the case of an electric dipole in a field E
(cf. Problem 3.1). From the competition between the orientation in the external field
and the thermal disorder at the temperature T , the average of the angle θ between
µp and B takes on the value
1
cos θ = L(x) = coth x − (1)
x
with x = µp B/k T . The macroscopic magnetisation along the field direction is then
M = N µp cos θ . (2)
In the high-temperature approximation, we find
x
L(x) → , (3)
3
so that for a required orientation of 1% at room temperature, we obtain
µp B
L(x) = = 0.01 . (4)
3k T
With the given values, one derives the required magnetic field
B = 7.14 T .
Only at low temperatures can one work with lower magnetic field strengths.
3.5 Making use of the polarisability volume α = α/4π ε0 instead of the polaris-
ability α, the Debye equation (3.19) becomes
4π NA α NA p20
Pm = + (1)
3 9ε0 kT
1
= a + b . (2)
T
Plotting Pm against 1/T thus yields a straight line with a y-intercept a and slope b.
The temperature-dependent relative dielectric constant εr = ε and therefore Pm (T)
are found from the measured values of the capacity C of a condenser (having C0
without a dielectric) quoted in the problem by applying
C
εr = , (3)
C0
so that from (1), a plot of Pm (T −1 ) is possible (M = 152.23 g/cm3 ). A straight-line
fit yields the intercept a = 82.63 cm3 /mole and the slope b = 10 833 cm3 K/mole.
From b, we an compute the dipole moment
p0 = 4.45 · 10−30 Cm = 1.33 D
and from a, the polarisability volume
α = 32.75 · 10−30 m3 .
10 Solutions
3.6 At optical frequencies, the reorientation of the molecules can no longer follow
the rapid changes in the applied field, so that the orientation polarisation becomes
negligibly small. The Debye equation (cf. Problem 3.4) can then be simplified to the
Clausius-Mosotti equation with ε = n 2 :
n2 − 1 M NA α
Pm = = . (1)
n +2
2 3ε0
Solving for α and using α = α/4π ε0 , one obtains from the given index of
refraction n = 1.333 with the molecular mass M = 18.015 g/mole and the density
= 0.9982 g/cm3 the polarisability volume
α = 1.47 · 10−30 m3 .
3.7 The dipole moment pH2 O of the water molecule can be treated as the vector sum
of the dipole moments pOH of two OH bonds. The dipole moments perpendicular
to the plane of symmetry compensate each other, while the parallel components
superpose constructively:
φ
pH2 O = 2 pOH cos , (1)
2
so that with pH2 O = 1.85 D and φ = 104.25◦ , it follows that:
pOH = 1.51 D . (2)
If the hydrogen peroxide molecule, H−O−O−H, is projected onto the plane

of the page, its geometrical structure is similar to that of the water molecule, if a
bonding angle of 90◦ is assumed for O−O−H. The O−O bond is then perpendicular
to the plane. Since it does not contribute to the dipole moment, the latter can be
represented for this molecule also as pH2 O2 as the vector sum of the two moments
of the pOH bonds:
φ
pH2 O2 = 2 pOH cos , (3)
2
Chapter 3 11
where φ is the angle between the two OH groups in the two-dimensional projection.
With pH2 O2 = 2.13 D, we obtain
φ = 90.3◦ .
3.8 The density of water vapour at a pressure of 1 bar is

18 g/mole
Dampf (1 bar) = = 8.0 · 10−4 g/cm3 .
22.4 l/mole
From εfl = 1 + χfl = 78.5 for liquid water of density 1 g/cm3 , one thus finds for
water vapour at 1 bar
χDampf (1 bar) = 77.5 · 8.0 · 10−4 = 6.2 · 10−2 .
This means that at 50 mbar
χDampf (50 mbar) = 3.1 · 10−3
and thus
εDampf (50 mbar) = 1 + 3.1 · 10−3
19% relative humidity at 40 ◦ C already exhibits the same effect on the value of ε as
does dry air alone.
3.9 The Debye equation, (3.19), for the molar polarisability Pm makes the connex-
ion between a macroscopically measurable quantity, the dielectric constant ε, and the
microscopic quantities dipole moment p0 and polarisability volume α = α/4π ε0
of chloroform.
Fig. S.1. A graph of the molar polarisability Pm of CHCl3 vs. the inverse temperature, 1/T :
data points and fitted line
12 Solutions
If one plots Pm against 1/T , then for temperatures above the melting point TS =
−64 ◦ C of chloroform, one obtains approximately a straight line with a y-intercept
a and a slope b. Below TS , the orientation polarisation is frozen in, so that its
contribution to Pm (i. e. the term containing the permanent dipole moment p0 )
vanishes. For 1/T > 1/TS , Pm is therefore practically independent of temperature and
lies on a straight line parallel to the x-axis in the plot. For T → ∞ the contribution
from the orientation polarisation likewise vanishes, so that the y-intercept a in the
plot should have the value of the constant ε for 1/T > 1/TS . For fitting the line to
the data points in the plot, we thus use only the first three points and the required
intercept on the y-axis.
From the slope b = 4508.5 cm3 K/mole, we can compute the dipole moment of
chloroform
p0 = 2.87 · 10−30 Cm = 0.86 D
and from the intercept a = 30.0 cm3 /mole its polarisability volume
α = 11.89 · 10−30 m3 .
3.10 From the Debye equation for the molar polarisability Pm

ε−1 M NA p20
Pm = = α+ (1)
ε+2 3ε0 3 kT
with a given dipole moment p0 and polarisability volume α = α/4π ε0 one can
calculate the dielectric constant ε of chlorobenzene. To this end, we insert in (1)

ε−1 NA p20
= α+ ≡x (2)
ε+2 3ε0 M 3 kT
and solve for ε:
2x + 1
ε= . (3)
1−x
With the given data, we obtain x = 0.801 and thus
ε = 13.08 .
3.11 The magnetic polarisability β is a microscopic quantity which relates the

induced magnetic moment m to the externally applied magnetic field B:
mind = β B (1)
(N: particle density). The moment m can, on the other hand, also be expressed in
terms of the macroscopic magnetic susceptibility:
1 µ−1 κ
mind = J= B= B (2)
µ0 N µ0 N µ0 N
Chapter 3 13
(J = macroscopic magnetisation). From this it follows that

κ
β= . (3)
µ0 N
In the classical ring-current model for the π electron system of benzene, one
considers the six delocalised electrons to be forced onto a circular orbit by the
Lorentz force in the field B. The Lorentz force thus plays the role of a centripetal
force:
mω×v =qv× B (4)
(q = −e = electronic charge; m = electron mass; ω = circular frequency; v =
orbital velocity). With ω = 2π/T , we obtain the orbital period T :
2m π
T= . (5)
eB
The magnetic moment can be calculated classically according to
m ind = I A , (6)
where A is the area enclosed by the ring current I. For a ring current of radius r
carried by n electrons, we thus have
−ne
m ind = π r2 (7)
T
and, with (5), we find
ne2r 2 B
m ind = (8)
2m
(8) and (1) yield
ne2r 2
β= . (9)
2m
According to this model, the induced ring current gives rise to the anisotropy in
the susceptibility and to the magnetic polarisability β. The expression found in (9)
for β thus corresponds to the difference between the components perpendicular and
parallel to the plane of the ring:
β = β⊥ − β (10)
so that the radius r of the ring current can be computed from (9) and the data given
in the problem:
2m
r2 = (µ0 β⊥ − µ0 β ) (11)
ne2 µ0
and finally
r = 2.91 · 10−11 m .
This result shows that the diamagnetism of benzene is only approximately de-
scribed by the ring-current model for the π electron system. The radius of the
electronic orbits gives at best only the order of magnitude.
14 Solutions
Fig. S.2. Dielectric function of the NaCl ionic crystal
3.12 The dielectric function of the NaCl crystal (Fig. S.2) is based on the model
of a coupled photon-phonon field (polaritons). From the given values, one finds
ΩLO = 4.9 · 1013 s−1
and then obtains the following frequency dependence for the dielectric constant ε:
An effect of the photon-phonon coupling is the formation of a frequency√ gap:
for ΩTO < ω < ΩLO , ε is negative and thus the wave vector k = ω ε(ω)/c is a
purely imaginary quantity, so that in this frequency range, the wave is damped in a
NaCl crystal as e−|k|x . In the neighbourhood of ΩTO , |k| is large, the wave penetrates
only a very short distance into the crystal, and it is completely reflected. For ω near
ΩLO , it is less strongly damped and can pass through a thin NaCl sheet (although
attenuated). The transmitted rays are the so called residual radiation. More about
this subject can be found in textbooks on solid-state physics.
Chapter 4
4.1 a) The time-independent Schrödinger equation in regions I, II, and III can be
written in the following ways:
2m
I) ∂ 2 ψ (1) /∂x 2 − κ 2 ψ (1) = 0, κ2 = (V0 − E) , (1)
2
2m
II) ∂ 2 ψ (2) /∂x 2 − k2 ψ (2) = 0, k2 = 2 E , (2)

2m
III) ∂ 2 ψ (3) /∂x 2 − κ 2 ψ (3) = 0, κ = 2 (V0 − E) .
2
(3)

It follows from (1) and (3) that
2m
k2 + κ 2 = V0 = C . (4)
2
Chapter 4 15
V(x)
Ι ΙΙ ΙΙΙ
x
–a a
Fig. S.3. The “square-well” potential V(x)
b) Boundary conditions:
lim ψ (1) (x) = 0 , therefore B (1) = 0 , (5)
x→−∞
lim ψ (3) (x) = 0 , therefore A(3) = 0 . (6)

x→+∞
Continuity condition at the boundary:

ψ (1) (−a) = ψ (2) (−a)
yields
A(1) e−κa = A(2) e−ika + B (2) eika . (7)
(1) (2)
∂ψ (−a)/∂x = ∂ψ (−a)/∂x
leading to
−κA(1) e−κa = −ik A(2) e−ika + ikB (2) eika . (8)
(3) (2)
ψ (a) = ψ (a)
yields
B (3) e−κa = A(2) eika + B (2) e−ika . (9)
∂ψ (3) (a)/∂x = ∂ψ (2) (a)/∂x
and, finally,
−κB (3) e−κa = ik A(3) eika − ikB (3) e−ika . (10)
Here, ∂ψ(±a)/∂x means, more precisely, ∂ψ(x)/∂x |x=±a . It follows from (7)
through (10) that

k k
1∓i A(2) e∓ika + 1 ± i B (2) e±ika = 0 . (11)
κ κ
16 Solutions
These are two homogeneous linear equations for A(2) , B (2) . The condition that
they have a non-trivial solution is that the determinant of the coefficients vanishes,
i. e.
κ 2 sin(2ka) + 2kκ cos(2ka) − k2 sin(2ka) = 0 . (12)
We find two types of solutions:
α) symmetric, κ = k tan(ka) , (13)
N
A(2) = B (2) = ; A(1) = B (3) = Neκa cos(ka) , (14)
2
and thus

 κ(x+a)
e cos ka (I)
ψs (x) = N cos kx (II) . (15)

 κ(−x+a)
e cos ka (III)
β) antisymmetric, κ = −k cot(ka) , (16)
N
A(2) = −B (2) = , A(1) = −B (3) = −Neκa sin(ka) (17)
2
and thus

 κ(x+a)
 −e sin ka (I)
ψa (x) = N sin kx (II) . (18)

 κ(−x+a)
e sin ka (III)
3) Normalisation:
+∞ −a
!
1 = |(x)| dx = ψ (1) (x) 2 dx
2
−∞ −∞
+a ∞
(2) 2
+ ψ dx + ψ (3) (x) 2 dx . (19)
−a a
The evaluation of the integrals is elementary. For the symmetric wavefunction,

we find
−1/2
1
N = a + cos(ka) [sin(ka) + cos(ka)] , (20)
a
Chapter 4 17
15 Fig. S.4. The graphical determi-
nation of the energy values for the
symmetric bound states, a = 1,
C 2 = 10. There are four energy val-
ues. (Note that the vertical lines are
10 added only to guide the eye)
0
0 5 10
k
15
10
Fig. S.5. As in Fig. S.4, but for the

0 antisymmetric bound states
0 5 10
k
and for the antisymmetric wavefunction,

−1/2
1
N = a + sin(ka) [sin(ka) − cos(ka)] . (21)
a
c) Graphical determination of the energy eigenvalues. For this purpose, we plot
the curves (4), (13), i. e. (22) or (4), (16), i. e. (23) in the (k, κ)-plane and find the
respective points of intersection of (22) and (23), Fig. S.4
α) κ = k tan(ka) , κ 2 + k2 = C 2 . (22)
For the example a = 1, C 2 = 10, we find four eigenvalues.
β) κ = −k cot(ka) , κ 2 + k2 = C 2 . (23)
In the same example, we obtain three eigenvalues, cf. Fig. S.5.
18 Solutions
In the symmetric case, α), there is at least one intercept, independently of a; the
ground state is symmetric.
d) The limiting case V0 → ∞:
For V0 → ∞, C → ∞ also holds, and due to (1) and (3), we have also
κ → ∞. (24)
For finite values of k, we thus have α) tan(ka) = ∞, i. e.

2n − 1 π
kn = , n∈N (25)
2 a
or
πn
kn = , n odd , (26)
2a
1 cos kn x (II)
ψs = √ . (27)
a 0 (I, III)
β) cot(ka) = ∞, i. e.
π
kn = n , n ∈ N or (28)
a
πn
kn = , n even , (29)
2a
1 sin kn x (II)
ψa = √ (30)
a 0 (I, III) .
The energy values are:
2 kn2 2 π 2 2
En = = n , n ∈ N. (31)
2m 8ma2
Matrix elements: the integrals are elementary.
a
xnm = xψn ψm dx = 0 (32)
−a
for n and m even or n and m odd. This result also follows from symmetry consider-
ations.
n odd, m even:
n+m−1 16anm
xnm = (−1) 2 , (33)
π 2 (n 2 − m 2 )2
(x 2 )nm = 0 (34)
for n even, m odd (or vice versa).

Chapter 4 19
n, m odd, n = m:
n−m 32a2 nm
(x 2 )nm = (−1) 2 , (35)
(n 2 − m 2 )2
n, m even, n = m:
n−m 32a2 nm
(x 2 )nm = (−1) 2 , (36)
π 2 (n 2 − m 2 )2
n = m:
a2 2a2
(x 2 )nn = − 2 2. (37)
3 n π
Momentum matrix elements:
a
d
pnm = ψn ψm dx = 0 (38)
i dx
−a
for n and m even or n and m odd.

n odd, m even:
n+m−1 2nm
pnm = (−1) 2 = p∗mn . (39)
ia n2 − m2
e) Wave packet:
1
ψ(x, t) = √ ψ1 (x)e−i E1 t/~ + ψ2 (x)e−iE2 t/~ , (40)
2
where
2 π 2
E 2 − E 1 = 3ω0 with ω0 = . (41)
8ma2
The integrals are elementary.
32a
x = cos(3ω0 t) , (42)
9π 2
2 a3 3a2
x = − 2, (43)
3 4π
4
p = − sin(3ω0 t) . (44)
3a
Wave packet:
1
ψ(x, t) = √ ψ1 (x)e−iE1 t/~ + ψ3 (x)e−iE3 t/~ , (45)
2
20 Solutions
where
E 3 − E 1 = 8ω0 , (46)
x = p = 0 , (47)

2 2 1 8 3
x =a + − cos(8ω0 t) . (48)
3 9π 2 2π 2
4.2 Ground-state wavefunction

1
ψ100 = 2a−3/2 e−r/a √ with a ≡ a0 . (1)
π
The definition of spherical polar coordinates:
x = r cos φ sin θ , (2)

y = r sin φ sin θ , (3)
z = r cos θ . (4)
Expectation values in real space:
ψ100 | x |ψ100 = ψ100 | y |ψ100

= ψ100 | z |ψ100 = 0 . (5)
This follows from symmetry considerations.

2
x = ψ100 | x 2 |ψ100 , (6)
∞ π 2π
1 −3
= a dr dθ r 4 cos2 φ sin3 θe−2r/a dφ (7)
π
0 0 0
∞ π
1 −3
= a dr dθr 4 e−2r/a sin3 θ π (8)
π
0 0
∞
4
= a−3 drr 4 e−2r/a · (9)
3
0
= a2 . (10)
Correspondingly,
2
y = a2 (11)
and
2
z = a2 . (12)
Chapter 4 21
(11) and (12) also follow directly from (10) due to the spherical symmetry of (1).
Momentum expectation values:

p x = p y = pz = 0 (13)
from symmetry considerations.

For the calculation of p2x , etc. we take
∂ 1x
ψ100 = − ψ100 , (14)
∂x ar

∂2 1 1 x2 x2
ψ100 = − − − ψ100 , (15)
∂x 2 a r r3 ar 2

p2x = ψ100 | p2x |ψ100
∞ π 2π
2
= dr dθ dφe−2r/a r 2 sin θ
πa4
0 0 0
 
 1 cos2 φ sin2 θ cos2 φ sin2 θ 
×
 r − − 
 (16)
! r ! a !
(1) (2) (3)
(1) yields
∞
42 2
e−2r/a rdr = ; (17)
a4 a2
0
(2) yields
∞ π
2 −2r/a
− 4 dre r sin3 θdθ , (18)
a
0 0
∞
42 1 2
=− 4 dre−2r/a r = − ; (19)
3a 3 a2
0
and (3) yields

∞
2 r2 4 1 2
− 4 dre−2r/a =− 2. (20)
a a 3 3a
0
22 Solutions
All together, we find

2 1
p2x = 2 . (21)
a 3
Owing to the spherical symmetry of the wavefunction (1), we find the corre-
sponding result
2 1
p2y = p2z = 2 . (22)
a 3
Because of the definition
1/2
∆ px = p2x − px 2 (23)
and
1/2
∆x = x 2 − x 2 , (24)
as well as (5, 10, 13, 21), it follows that

1
∆ px = √ , ∆x = a (25)
3a
and we thus find for Heisenberg’s uncertainty relation
1
∆x ∆ p = √ > . (26)
3 2
The expectation value for the kinetic energy is
2
E kin = ψ100 | − ∆ |ψ100 . (27)
2m
The Laplace operator in spherical polar coordinates is given by
d2 2 d l(l + 1)
∆= + − , (28)
dr 2 r dr r2
2
E kin = . (29)
2ma2
e2 1 e2
E pot = ψ100 | − |ψ100 = − . (30)
4πε0r 4πε0 a
Expectation values for the angular momentum:
L 2 ψklm = l(l + 1)2 ψklm , (31)

L z ψklm = mψklm . (32)
Chapter 4 23
From the commutation relations for the angular-momentum operators, it follows

that
[L z , L ± ] = ±L ± , (33)
where
L ± = L x ± iL y . (34)
Applying (33), we write the equation
L z L + |lm = L + (L z + ) |lm = (m + 1)L + |lm . (35)
Thus, ψ+ ≡ L + |lm is an eigenfunction of L z (32) with the eigenvalue (m + 1)
up to a normalisation factor, i. e.
ψ+ ≡ L + |lm = N |lm + 1 . (36)
In order to determine N, we set
ψ+ | ψ+ ≡ lm| L − L + |lm = N 2 lm + 1 | lm + 1 . (37)
Using (34), we write
L − L + = L 2x + L 2y + iL x L y − iL y L x
= L2 − L 2z − L z . (38)
With (31) and (32) and the normalisation of the wavefunctions, it follows from
(37) that:

2 l 2 + l − m 2 − m = N 2 (39)
or
%
N= (l − m)(l + m + 1) (40)
so that
%
L + |lm = (l − m)(l + m + 1) |lm + 1 (41)
and, correspondingly,
%
L − |lm = (l + m)(l − m + 1) |lm − 1 . (42)
From (31), (32) and the normalisation of the wavefunctions, we have
ψklm | L 2 |ψklm = l(l + 1)2 , (43)
ψklm | L z |ψklm = m . (44)
With (41) and (42), we find
1 1
L x = (L + + L − ) , L y = (L + − L − ) (45)
2 2i
and the orthogonality of the wavefunctions follows:
ψklm | L x |ψklm = ψklm | L y |ψklm = 0 . (46)
24 Solutions
4.3 The trial solution is

ψ± = c(φa ± φb ) with dV |φa | 2 = dV |φb | 2 = 1 (1)
a) Normalisation:

∗
|ψ± | dV = c 2 ± 2 dVφa φb
2 2
!
= c2 (2 ± 2S) = 1, and thus
−1/2
c = [2(1 ± S)] . (2)
b) Expectation value of the Hamiltonian:

E = dVψ±∗ Hψ± . (3)
Insertion of (1) and (2) into (3):

2 ∗ 2 e2 e2
E= dVφa − ∆− − φa (4)
2(1 ± S) 2m 4πε0ra 4πε0rb
2
2 e2 e2
± dVφb∗ − ∆− − φa . (5)
2(1 ± S) 2m 4πε0ra 4πε0rb
The factors of 2 in (4) and (5) result from the fact that in E (3), there were originally
two further integrals in which the indices a and b were exchanged relative to (4) and
(5), but which otherwise were identical to (4) and (5) from symmetry considerations.
Using the abbreviations E 0 , C, D, S as in the text, we find
1 0
E= E + C ± S · E0 + D . (6)
1±S
This can be brought into the form
C±D
E = E0 + . (7)
1±S
c) with the trial solution
ψ = c1 φa + c2 φb (8)
we obtain for the normalisation

!
dVφ∗ φ = c21 + c22 + 2c1 c2 S = 1 (9)
and thus
Chapter 4 25
1
ψ=& (c1 φa + c2 φb ) . (10)
c21 + c22 + 2c1 c2 S
!
:=N
Then the energy expectation value is found to be

Ẽ = dVψ ∗ Hψ = N 2 c21 + c22 [E 0 + C] + 2c1 c2 [SE 0 + D] , (11)

Ẽ = E 0 + N 2 c21 + c22 C + 2c1 c2 D . (12)
Variation of Ẽ (12) with respect to c1 :
∂ Ẽ
(i) = 2N 2 (c1 C + c2 D)
∂c1
!
− 2N 4 (c1 + c2 S)[(c21 + c22 )C + 2c1 c2 D] = 0 , (13)
and with respect to c2 :
∂ Ẽ
(ii) = 2N 2 (c2 C + c1 D)
∂c2
!
− 2N 4 (c2 + c1 S) [ c21 + c22 C + 2c1 c2 D] = 0 . (14)
We take
(i)c2 − (ii)c1
4
= 0 = c21 − c22 S[ c21 + c22 C + 2c1 c2 D] (15)
2N
and thereby obtain
c21 = c22 c1 = ±c2 (16)
in agreement with the original trial solution (1).

d) As in Fig. S.6, we have
1/2
1 2
ra/b = r + R ∓ rR cos θ
2
. (17)
4
Prolate elliptical coordinates are defined by

ra + rb ra − rb
ξ= , η= . (18)
R R
To compute the volume element dV , which in spherical polar coordinates is
given by
dV = r 2 sin θdθdφ , (19)

26 Solutions
Fig. S.6. The hydrogen molecule-ion
A
ra
R/2
ϑ
(ξ,η,ϕ)
R γ
R/2
rb
we write the determinant of the derivatives:

∂ξ ∂ξ

∂r ∂θ ,
D = ∂η ∂η (20)
∂r ∂θ
so that
dξdη = Ddrdθ , (21)
whereby, due to (17–19):

' (
1
1 2ra (2r − R cos θ)+ 2rb (2r + R cos θ); ' 2ra − 2rb ( + R sin θ
1 1 1
D= 2 1 . (22)
R (2r − R cos θ)− 1 (2r + R cos θ); 1 + 1 + R sin θ
2ra 2rb 2ra 2rb
Calculation of D yields
2 r 2 sin θ
D= (23)
R ra rb
or, due to (18),
r 2 sin θ
D= . (24)
(R/2)3 (ξ + η)(ξ − η)
The volume element dV (19), applying (21, 24), is given by
3
R 2
dV = ξ − η2 dξdηdφ , (25)
2
Chapter 4 27
where
0 ≤ φ < 2π ,
−1 ≤ η < 1 ,
1 ≤ ξ < ∞. (26)
As trial wavefunctions, it is advisable to choose
3/2
1 α αr
φ(r; α) = √ exp − . (27)
π a0 a0
We first set β = α/a0 .
Computation of the overlap integral S:

β3
S = φa φb dV = dV exp [−β (ra + rb )] (28)
π
or, in prolate elliptical coordinates,
3 2π 1 ∞
β3 R
S= dφ dη dξ ξ 2 − η2 exp (−βRξ) . (29)
π 2
0 −1 1
Integrating over dφ and dη immediately yields

∞
1 2
S = (βR)3 dξ 2ξ − exp (−βRξ) .
2
(30)
4 3
1
Making use of the ancillary formulas from Problem 4.2, we can carry out the
integration over ξ after combining terms,
) *
1 αR 2 αR αR
S= + + 1 exp − . (31)
3 a0 a0 a0
The Coulomb integral C:

e2 1
C=− φa φa dV
4πε0 rb

e2 β 3 1
=− dV exp (−2βra ) . (32)
4πε0 π rb
Or, in prolate elliptical coordinates,
3 2π +1 ∞
e2 β 3 R
C=− 2 dφ dη dξ(ξ 2 − η2 )
4π ε0 2
0 −1 1
2
× exp [−βR(ξ + η)] . (33)
R(ξ − η)
28 Solutions
Integration over φ and cancellation of (ξ − η) yields:

+1 ∞
e2 β 3 R2
C=− dη dξ(ξ + η) exp [−βR(ξ + η)] . (34)
8πε0
−1 1
Integration over η and introduction of a new variable u = ξ + η gives:

∞ u=ξ+1
e2 β 3 R2 1 1
C=− dξ u+ exp(−βRu) (35)
8πε0 βR (βR)2
1 ! u=ξ−1
=F(u)
∞ 
∞
e2 β 3 R2 
=− du F(u) + du F(u) . (36)
8πε0
2 0
Evaluation of the integrals using the ancillary formulas of Problem 4.2 and
combining terms, with β = α/a0 , gives:

e2 1 α 1 αR
C= − + + exp −2 . (37)
4πε0 R a0 R a0
The exchange integral D is:

e2 1
D=− φa φb dV
4πε0 ra

e2 β 3 1
=− dV exp [−β(ra + rb )] . (38)
4πε0 π ra
Or, in prolate elliptical coordinates:
3 2π +1 ∞
e2 βR
D=− 2 dφ dη dξ
4π ε0 2
0 −1 1
2 2
× ξ − η2 exp (−βRξ) . (39)
R(ξ + η)
Integration over φ and cancellation of (ξ + η) yields:
1 ∞
e2 β 3 R2
D=− dη dξ(ξ − η) exp (−βRξ) . (40)
8πε0
−1 1
Integration over η:
∞
e2 β 3 R2
D=− dξ ξ exp (−βRξ) . (41)
4πε0
1
Chapter 4 29
The final result is:

e2 α αR
D=− 1+ exp (−βR) . (42)
4πε0 a0 a0
For the ground-state energy E 0 , we find:

2 e2 1
E = φa −
0
∆− · φa dV . (43)
2m 4πε0 ra
Evaluation in spherical polar coordinates: since the wavefunction does not de-
pend on θ and φ, we use
1 ∂ 2∂
∆ = ∆r = r (44)
r 2 ∂r ∂r
and dV = 4πr 2 dr on integrating only over r. Then

r 2 ∆r φa ∝ −2rβ + (βr)2 exp(−βr) . (45)
For (43), it follows that

∞
β3
E =
0
4π dr
π
0

2 e2
× − −2βr + (βr)2 − r exp (−2βr) , (46)
2m 4πε0
and, after integrating,

2
β e2 1
E 0 = 4β 3 − . (47)
2m 4 4πε0 4β 2
Using β = α/a0 and the definition of the Bohr radius a0 , we obtain
e2 α ' α (
E0 = −1 . (48)
4πε0 a0 2
e) The total energy is given by:
' (
e2 α α
E mol = −1
4πε0 a0 2

−R−1+(α/a0 + R−1 )e−2αR/a0 ∓(α/a0 )(1 + αR/a0 )e−αR/a0
+
[(1/3)(αR/a0 )2 + αR/a0 + 1]e−αR/a0
2
e 1
+ . (49)
4πε0 R
30 Solutions
E mol
2
1
0
–1
0,8
1
1,2
1,4
R 1,6 0,8
0,9
1,8 1
1,1
2 1,2 α
1,3
Fig. S.7. E mol (49) as a function of α and R (in units of a0 ) for the symmetric state
E mol
10
9
8 antisymmetric
7
6
5
4
3
2
1
0 R
0,5 1 1,5 2 2,5 3 3,5 4 4,5 5
–1
–2 symmetric
Fig. S.8. E mol (49) for fixed α = 1 as a function of R in units of a0 for the symmetric and
the antisymmetric states, resp.
The evaluation must now be carried out numerically, which can be easily done
using a PC (cf. Figs. S.7–9).
f) Computation of the expectation values x, p, x 2 .
In order to express x, y, z in prolate elliptical coordinate, we proceed as follows:
It follows from (18) that
R R
(ξ + η) = ra , (ξ − η) = rb (50)
2 2
Chapter 4 31
E mol
10
9
8 antisymmetric
7
6
5
4
3
2 symmetric
1
0 α
0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2
–1
–2
Fig. S.9. E mol (49) for fixed R = 1.36 a0 as a function of α for the symmetric and the
antisymmetric states
and thus, using (17),
R2 1
(ξ ± η)2 = r 2 + R2 ∓ rR cos θ . (51)
4 4
From this, we find
R2 1
(ξ + η)2 + (ξ − η)2 = 2r 2 + R2 (52)
4 2
and
R2
(ξ + η)2 − (ξ − η2 ) = −2rR cos θ . (53)
4 !
=4ξη
It follows from (53) that

1
− Rξη = r cos θ = z . (54)
2
Due to (52), we have
R2 2
ξ + η2 − 1 = r 2 = x 2 + y2 + z 2 (55)
4
and, in addition, due to (54),
R2 2
ξ + η2 − ξ 2 η2 − 1 = x 2 + y2 ≡ r 2 sin2 θ . (56)
4
32 Solutions
With x = r cos θ cos φ, y = r cos θ sin φ, it follows that

R 2 1/2
ξ + η2 − ξ 2 η2 − 1 cos φ = x , (57)
2
 1/2
R 2 
ξ + η − ξ 2 η2 − 1 sin φ = y .
2
(58)
2 !
(ξ 2 −1)(1−η2 )
We write

x = dV xψ 2 . (59)
Now, the wavefunction ψ = N(φa + φb ) as well as the expressions (54, 57, 58)
for (x, y, z) and relation (25) for the volume element dV can be inserted into (59). Its
evaluation is, as above, elementary and yields zero for all components. 1 This can also
be concluded directly from symmetry considerations, e. g. for x = dVxψ 2 . Since
ψ(−x, y, z) = ψ(x, y, z), ψ 2 is also invariant with respect to (x, y, z) → (−x, y, z).
On the other hand, the value of the integral must remain unchanged, from which
it follows that x = −x and thus x = 0. The same reasoning also holds for
∂/∂x, ∂/∂y, and ∂/∂z, since e. g. in the transformation (x, y, z) → (−x, y, z),
∂/∂x → −∂/∂x also holds.
In order to obtain a measure of the spatial extension of the wavefunction ψ, one
must compute x 2 , y2 , and z 2 . We illustrate this using as an example z 2 , going
immediately to prolate elliptical coordinates. With
1 X
cosh X ≡ e + e−X , (60)
2
and inserting the expressions for z 2 , ψ, and dV , we obtain
5 2π 1 ∞
2 β3 R
z =2 dφ dη dξ ξ 2 − η2 ξ 2 η2
π 2
0 −1 1
× [cosh(βRη) ± 1] exp (−βRξ) . (61)
Due to (60), all the integrations can be carried out explicitly (first over φ, then
over η, finally over ξ). With β = α/a0 , the result is
' a (2
0
z 2 =
α

1 αR 4 7 αR 3 9 αR 2 αR
+ + +2
30 a0 30 a0 10 a0 a0

1 αR
× e−αR/a0 + +2 . (62)
2 a0
Chapter 4 33
4.4 Conjecture:
1 ∗
Ψ HΨdV1 ...dVn
1 ≥ E0 . (1)
Ψ ∗ ΨdV1 ...dVn
Let E 0 be the smallest energy eigenvalue of the Schrödinger equation
HΦ0 = E 0 Φ0 (2)
and Φ0 the corresponding eigenfunction, with Ψ the approximation for Φ0 .

Since H is hermitian, the eigenfunctions Φ j , j = 0, 1, . . . can be chosen in such
a way that they form a complete orthonormal basis, i. e.
Φi | Φ j = δij . (3)
Every vector of the space of states can thus be expressed as a linear combination
of the Φ j :
∞

Ψ = ajΦj . (4)
j=0
Inserting (4) into the left side of (1), we obtain using (3)
∞
∞

a2j E j / a2j (5)
j=0 j=0
and due to E j ≥ E 0
∞
∞
∞
∞

a2j E j / a2j ≥ E 0 a2j / a2j = E 0 , (6)
j=0 j=0 j=0 j=0
i. e. the inequality (1).

4.5 Let the Hamiltonian H of the Schrödinger equation
HΨ = EΨ (1)
be given by

N
H= Hi (2)
i=1
with
Hi φk(i) = E k(i) φk(i) (3)

34 Solutions
and let
2 3
( j)
Hi , φk = 0 , i = j . (4)
We insert
4
N
( j)
Ψk1 ,k2 ...k N = φk j , (5)
j=1
where k1 , . . . k N are a selection from the quantum numbers of (3), into (1) and obtain
with (4)

N 4
N
N 4
i−1 ' ( 4
N
( j) ( j) ( j)
Hi φk j = φk j · Hi φk(i)i φk j , (6)
i=1 j=1 i=1 j=1 j=i+1
which, due to (3), reduces to

N 4
N
( j)
E k(i)i φk j = EΨ (7)
i=1 j=1
with

N
E= E k(i)i . (8)
i=1
4.6 The functions

1
Ψ1 = √ [α(1)β(2) + β(1)α(2)] Ψu (1)
2
1
Ψ2 = √ [α(1)β(2) − β(1)α(2)] Ψg (2)
2
are given, where α and β are spin wavefunctions for spin ↑ and spin ↓, resp.
The arguments 1, 2 refer to particles 1, 2, and Ψu (Ψg ) is the even (odd) spatial
wavefunction. We make use of the Pauli matrices

01 0 −i 1 0
σx = , σy = , σz = , (3)
10 i 0 0 −1
where the connexion to the spin operator is given by:

S= σ. (4)
2
For the components of the total spin, we find
Si = S1i + S2i , i = x, y, z, (5)

Chapter 4 35
where the indices 1, 2 refer to the two particles. Then
S2 = Sx2 + S2y + Sz2

= S12 + S22 + 2S1z S2z + 2S1x S2x + 2S1y S2y . (6)
As usual, we define the ladder operators
S1± = S1x ± iS1y , S2± = S2x ± iS2y , (7)
whereby the following relations hold:
S1+ α(1) = 0, S1+ β(1) = α(1),

S1− α(1) = β(1), S1− β(1) = 0 , (8)
and correspondingly for S2± . Then (6) can be cast in the form
S2 = S12 + S22 + 2S1z S2z + S1+ S2− + S1− S2+ . (9)
For the basis vectors
α(1)α(2) , α(1)β(2) , β(1)α(2) , β(1)β(2) ,
it follows using (8) and (3), i. e.
σx2 = σ y2 = σz2 = 1 (10)
and owing to (4):
S2 α(1)α(2) = 22 α(1)α(2) (11)

S α(1)β(2) = (α(1)β(2) + β(1)α(2))
2 2
(12)
S β(1)α(2) = (β(1)α(2) + α(1)β(2))
2 2
(13)
S β(1)β(2) = 2 β(1)β(2) .
2 2
(14)
From (11) and (12), α(1)α(2) and β(1)β(2) are already eigenfunctions belonging
to S2 with the eigenvalue 22 . To determine the remaining eigenfunction, we add or
subtract equations (12, 13) to or from one another, and obtain
S2 (α(1)β(2) + β(1)α(2)) = 22 [α(1)β(2) + β(1)α(2)] , (15)
i. e. the spin part of (1) is an eigenfunction of S2 with the eigenvalue 22 . Corre-
spondingly, we find
S2 (α(1)β(2) − β(1)α(2)) = 0 , (16)
i. e. the spin part of (2) is an eigenfunction of S2 with the eigenvalue 0. Due to Sz α =

~ α, S β = − ~ β, we immediately obtain the result that α(1)β(2) ± α(2)β(1) are
2 z 2
eigenfunction belonging to the z-component of the total spin, Sz = S1z + S2z , with
the eigenvalue 0.
36 Solutions
4.7 We make use of the atomic orbitals

1 1 ' r ( −r/(2a) 1
ψ2s = √ R2,0 (r) = a−3/2 √ 1 − e √ (1)
4π 2 2a 4π

3
ψ2,1,1 = R2,1 (r) sin θeiφ (2)
8π

3
ψ2,1,0 = R2,1 (r) cos θ (3)
4π

3
ψ2,1,−1 = R2,1 (r) sin θe−iφ (4)
8π
with
1 r −r/(2a)
R2,1 (r) = a−3/2 √ e . (5)
6 2a
In order to identify them with wavefunctions of the px , p y , pz orbitals, we
superpose the wavefunctions:

1 3 x
ψ2 px = √ ψ2,1,1 + ψ2,1,−1 = R2,1 (6)
2 4π r

1 1 3 y
ψ2 p y = √ ψ2,1,1 − ψ2,1,−1 = R2,1 (r) (7)
i 2 4π r

3 z
ψ2 pz = ψ2,1,0 = R2,1 (r) . (8)
4π r
a) The points of maximum (charge) density
Both in the case of sp hybridisation and in the case of tetragonal sp3 hybridisation
(see part c of the problem), the wavefunctions can be written in the form
) *
3
ψ = α(1 − γr) + γ bi xi e−γr , (9)
i=1
where x1 = x, x2 = y, x3 = z.
To find the maximum of | ψ |2 , we take:
) *5
∂ψi xj xj 3
= −γ +γb j − γ α(1 − γr)+γ bi xi e−γr . (10)
∂x j r r i=1
From the requirement that ∂|ψi |2/∂x j = 2ψi (∂ψi /∂x j ) = 0, it follows that

3
a(1 − γr) + γ bi xi = 0 (11)
i=1
Chapter 4 37
or
xj = 0 (12)
or

3
x j 1 + a(1 − γr) + γ bi xi = b j r . (13)
i=1
We specialise to the case of sp hybridisation, with
ψ± = ψ2s ± ψ2 px (14)
and thus in (9)
a = 1, b1 = ±1, b2 = b3 = 0, xi = x, y, z . (15)
We then find
x(2 − γr ± γx) = ±r (16)

y(2 − γr ± γx) = 0 (17)
z(2 − γr ± γx) = 0 . (18)
From this, it follows for the maxima of the charge density that
3
y = z = 0, x = ∓ . (19)
2γ
b) Centre of gravity of the charge distribution
Since in (9) ψ2s and
√ψ2 px are normalised and orthogonal to one another, (9) is
to be normalised by 1/ 2:
1
x± = ψ± xψ± , (20)
2
where . . . denotes the integration in spherical polar coordinates r, θ, φ, i. e.
∞ π 2π
2
r dr sin θdθ dφψ± xψ± . (21)
0 0 0
The integrals can be evaluated in an elementary manner, and the result is

27
x+ = − πa, y+ = 0, z+ = 0 (22)
32
27
x− = πa, y− = 0, z− = 0 . (23)
32
38 Solutions
c) Maximum charge densities in the case of tetragonal hybridisation (sp3 ). We

make use of the result (13) from part (a) and obtain for
1) a = 1, b1 = b2 = b3 = 1:
   
x r
 y [2 − γr + γ(x + y + z)] = r  . (24)
z r
From this, it follows immediately that x = y = z and then

√
2+ 3
x=− √ . (25)
(3 + 3)γ
2) a = 1, b1 = 1, b2 = b3 = −1: owing to (13), we have
   
x r
 y [2 − γr + γ(x − y − z)] = −r  . (26)
z −r
From this, it follows that y = z = −x and

√
2+ 3
x=− √ . (27)
(3 + 3)γ
3) a = 1, b1 = −1, b2 = 1, b3 = −1: in analogy to 1) and 2), we obtain
√
2+ 3
x= √ , x = z = −y . (28)
(3 + 3)γ
4) a = 1, b1 = −1, b2 = −1, b3 = 1:
√
2+ 3
x= √ , x = y = −z . (29)
(3 + 3)γ
Denoting x in (28) by x0 , then the four coordinates of the maxima of the (charge)
densities are given by:
x0 (−1, −1, −1); x0 (−1, 1, 1); x0 (1, −1, 1); x0 (1, 1, −1), (30)
√
which corresponds to the vertices of a tetrahedron with the edge length x0 2 2.
d) With the wavefunctions (1–8), the superpositions corresponding to the hy-
bridisation can be written immediately. In spherical polar coordinates, the integrals
can be evaluated in elementary fashion, and they demonstrate the orthogonality of
the wavefunctions.
Chapter 5 39
Chapter 5
5.1 The electronic structure of polyenes (Fig. S.10)
Fig. S.10. The π-electron system of the polyenes
Model for π-electrons: motion in a quantum wire. Equation of motion:

2 d 2
φ(x) + (E − Vab ) φ(x) = 0 . (1)
2m dx 2
a) Boundary condition for simple connectivity (Fig. S.11, upper part):
φba (a) = φab (a) , (2)

φba (a) = φab (a) . (3)
b'
b a
~
b e e
Fig. S.11.
b) Boundary condition at the end of the molecule (Fig. S.11, lower part): a π-
electron can also be located with a certain probability outside the last atom e. (ẽ is
a fictitious atom or “empty cell” at which φ can decay to zero).
φbe (e) = φeẽ (e) , (4)
φeẽ (ẽ ) = 0 , (5)

φbe (e) = φe ẽ (e) . (6)
c) The butadiene molecule:
2 d 2
Heff = − + V(x)
2m dx 2
0 for 0 ≤ x ≤ 5d
with V(x) = . (7)
∞ otherwise
40 Solutions
Heff φ(x) = E (8)

k 2 2
with E= , (9)
2m
π
k=n . (10)
5d

2
φ(x) = sin kx with 0 ≤ x ≤ 5 d . (11)
5d
5.2 The complete solution can be found in 6.
Chapter 6
6.1 a) Symmetry operations:

1. Triangle
∧
Rotation around the centroid by 120◦ , 240◦ (= −120◦ ).
Reflection on each of the lines passing through the centroid and a vertex.
2. Square
Rotation around the centre point by 90◦ , 180◦ , 270◦ or in the reverse direction.
Reflection through a diagonal (2)
Reflection on each line passing through the centre point and the middle of each
side (2).
3. Three collinear and equidistant points.
Reflection through the midpoint.
1 a 2 1 2 Fig. S.12. A straight line in a square.

Case 1
7 7
b 6 Fig. S.13. Line in a square. Case 2
6
5
5
4 3 4 3
1 a 2 1 2
Fig. S.14. Triangle in a square. Case 3
6 6
b Fig. S.15. Triangle in a square. Case 4
7
5 7 5
4 3 4 3
Chapter 6 41
The orientations in which the number of symmetry operations is maximal: square

and straight line; (Fig. S.12) (Fig. S.13)
Square and triangle: (Fig. S.14) (Fig. S.15)
b) When two symmetry operations are carried out in succession, the result again
corresponds to a symmetry operation, as one can show by examining all the possible
combinations. The inverse of a symmetry operation is clearly itself a symmetry
operation. One can directly demonstrate the validity of the associative law by first
taking A · B and then (A · B) · C, or first B · C and then A · (B · C), noting that
(A · B) · C = A(B · C), where A, B, C are symmetry operations. It is sufficient to
examine a few examples.
The identification as a group can be carried out systematically, simply and in an
elegant manner by considering that each symmetry operation can be represented by
a permutation of the vertices and that the permutations form a group.
c) Interaction potentials
In Fig. S.12: the spacings are:
√
d1(1) ≡ r12 = a, d2(1) ≡ r24 = 2a, d3(1) ≡ r56 = b ,
a
d4(1) ≡ r57 = 2b, d5(1) ≡ r16 = √ ,
2
%
(1)
d6 ≡ r45 = a − 2ab + 2b ,
2 2
%
d7(1) ≡ r15 = a2 + 2ab + 2b2 .
If we write V(dk ) = V(k), k = 1, 2, . . . , 7, then, taking the symmetries into

account, we obtain
V (1) = 4V(1) + 2V(2) + 2V(3) + V(4) + 4V(5)

+ 4V(6) + 4V(7) .
In the case that V = r 2 , this reduces to
V (1) = 14(a2 + b2 ) .
In Fig. S.13:
√
d1(2) ≡ r12 = a, d2(2) ≡ r24 = 2a, d3(2) ≡ r56 = b ,

a a
d4(2) ≡ r57 (2) (2)
= 2b, d5 ≡ r16 = √ , d6 ≡ r45 = b − √ ,
2 2

a a + 2b
2 2
d7(2) ≡ r25 = b + √ , d8(2) ≡ r17 = .
2 2
V (2) = 4V(1) + 2V(2) + 2V(3) + V(4) + 4V(5)

+ 2V(6) + 2V(7) + 4V(8) .
42 Solutions
For the case V = r 2 , this reduces to
V (2) = 14(a2 + b2 ) .
In Fig. S.14:
√
d1(3) ≡ r12 = a, d2(3) ≡ r24 = 2a, d3(3) ≡ r56 = b ,
& √
1
d4(3) ≡ r45 = √ 3a2 − (3 + 3)ab + 2b2 ,
6
&
1 √
d5(3) ≡ r46 = √ 3a2 + 2 3ab + 2b2 ,
6
& √
1
d6(3) ≡ r47 = √ 3a2 + (3 − 3)ab + 2b2 ,
6
&
1 √
d7(3) ≡ r16 = √ 3a2 − 2 3ab + 2b2 ,
6
& √
1
d8(3) ≡ r15 = √ 3a2 − (3 − 3)ab + 2b2 ,
6
& √
1
d9(3) ≡ r17 = √ 3a2 + (3 + 3)ab + 2b2 .
6
(3)
V = 4V(1) + 2V(2) + 3V(3) + 2V(4) + 2V(5)
+ 2V(6) + 2V(7) + 2V(8) + 2V(9) .
In the case V = r 2 , this reduces to

b2
V (3) = 14 a2 + .
2
In fig. S.15:
√
d1(4) ≡ r12 = a, d2(4) ≡ r24 = 2a, d3(4) ≡ r56 = b, d4(4)

b a

≡ r45 = √ − √ ,
3 2
& &
a2 ab √ √ b2
d5(4) ≡ r46 = +√ 2+ 3− 2− 3 − ,
2 12 3
b a
d6(4) ≡ r25 = √ + √ ,
3 2
& &
(4) a 2 ab √ √ b2
d7 ≡ r26 = +√ 2− 3− 2+ 3 − ,
2 12 3

(4) a2 b2
d8 ≡ r15 = − ,
2 3
Chapter 6 43
& &
a2 ab √ √ b2
d9(4) ≡ r16 = −√ 2− 3+ 2+ 3 − ,
2 12 3
& &
(4) a2 ab √ √ b2
d10 ≡ r17 = +√ 2− 3+ 2+ 3 − .
2 12 3
V (4) = 4V(1) + 2V(2) + 3V(3) + V(4) + 2V(5)

+ V(6) + 2V(7) + 2V(8) + 2V(9) + 2V(10) .
In the case that V = r 2 , this reduces to

(4) b2
V = 14 a + 2
.
2
d) The numerical results can be found in Figs. S.16–S.19. From them, for ex-
ample, it can be read out whether in the “square environment” the triangle can be
stretched into a straight line.
1,6
1,2
V1 – V2
0,8
0,4
0
0 0
1 1
2 2
~ 3 3 ~
b 4 4 a
5 5
Fig. S.16. The potential difference V 1 − V 2, where V 1 = V (1) with V(r) = 1 − e(−r)
6.2 The orthogonality properties of group characters

If the group g is finite, then every representation R → U(R) is equivalent to a
unitary representation, where U(R) is thus a unitary matrix.
Proof : Take any positive definite Hermitian form, e. g. the unit form, apply all
the transformations U(R) to it and sum over R. In this manner, we obtain a positive
definite Hermitian form H, which is invariant under every such transformation U(R).
Now choose the coordinate system in such a way that H assumes the unit form. Then
U(R), expressed in these coordinates, is unitary.
44 Solutions
0,2
0
–0,2
–0,4
V2 – V4
–0,6
–0,8
–1
–1,2
–1,4
0
0 1
0,5 1 2
1,5 3
2 2,5 4 ~
~ 3 3,5 5 a
b
Fig. S.17. The potential difference V 2 − V 4
2
V4 – V2
–1
0 1 2 1 0
3 4 3 2
~ 5 5 4 ~
b a
The same method of summation over the elements of the group is the basis of
the orthogonality relations which we now discuss.
Let
R → U(R) g dimensional , (1)

R → V(R) g dimensional (2)
be two inequivalent, irreducible representations of the finite group G. We express

the matrices U, V in terms of their elements:
Chapter 6 45
0,3
V3 – V4
0,2
0,1
0
0 0
1 1
2 2
~ 3 3 ~
b 4 4 a
5 5
U(R) = (u ik (R)) , (3)

V(R) = (vlm (R)) (4)
and write
V −1 (R) = (v̂lm (R)) . (5)
Let A be an arbitrary matrix with g rows and g columns. We take

U(R)AV −1 (R) = B , (6)
R
where the sum runs over all the elements R of the group. Then B remains unchanged
under the following transformation:
U(R0 )BV −1 (R0 ) = B , (7)
where R0 is an arbitrary element of the group. For a unitary representation, the

following relation naturally holds:
∗
v̂lm = vml . (8)
Proof :
U(R0 )U(R) = U(RR0 ) , (9)

−1 −1 −1
V (R)V (R0 ) = V (RR0 ) (10)
due to the properties of the representations. If the sum in (6) runs over all the elements
R, so then so does R = RR0 , with R0 fixed. From I of the general theorem, it follows
that the matrix B = 0, i. e.
46 Solutions

u ik (R)akl v̂ lm (R) = 0 . (11)
R kl

Since the akl are arbitrary, the following g2 · g 2 equations result:

u ik (R)v̂ lm (R) = 0 (12)
R
or, for unitary transformations,

∗
u ik (R)vml (R) = 0 . (13)
R
Using only one irreducible representation R → U(R), we find in a similar manner
U(R)AU −1 (R) = B , (14)
i. e.
U(R)B = BU(R) . (15)
According to the theorem, B must be a multiple of the unit matrix, i. e.

u ik (R)akl û lm (R) = αδim , (16)
R kl
where the constant α still depends on the matrix A. We set all the elements in the
matrix A equal to zero except for one, akl = 1 for a particular pair of values k, l. It
then follows that

u ik (R)û lm (R) = αkl δim . (17)
R
Since (û lm (R)) is the inverse matrix of (u lm (R)), it holds that

û lm (R)u mk (R) = δlk . (18)
R,m
In (17), we set m = i and sum over m = 1, 2, ..., g, obtaining then for the left
side of (17)

u mk (R)û lm (R) (19)
R m
which, owing to (18), is

δlk (20)
R
Chapter 6 47
and, since the group has h elements, finally
hδlk . (21)
Setting i = m in (17) and summing over the right-hand side with m = 1, 2, . . . g,

we obtain
gαkl . (22)
Comparing the final results for both sides, we thus find
hδl j = gαkl , (23)
where then αkl is determined by the order h of the group and the dimension g of the
matrix representation. We thus obtain the first main result:
1
1 for i = m, k = l
u ik (R)û lm (R) = g (24)
h 0 otherwise
R
for every irreducible representation R → U(R) and

1
u ik (R)v̂ lm (R) = 0 (25)
h R
for two arbitrary irreducible representations R → U(R) and R → V(R). In order

to obtain the fundamental orthogonality relations of group characters, we set i = k
and l = m and sum over i or l. We set

u ii (R) = χ(R) (26)
i
and

û ll (R) = χ̂(R) (27)
l
or

v̂ ll (R) = χ̂ (R) (28)
l
and obtain
1
χ(R)χ̂(R) = 1 . (29)
h R
For any two inequivalent irreducible representations, it holds correspondingly

that
1
χ(R)χ̂ (R) = 0 . (30)
h R
48 Solutions
The expressions χ̂(R) or χ̂ (R) can be cast in a different form. If the represen-
∗
tation matrices are unitary, it holds that û lm = vml and χ̂ = χ ∗ follows, where
the
asterisk denotes complex conjugate quantities. Furthermore, U −1 (R) = U R−1
and therefore

û lm (R) = u lm R−1 . (31)
It then also follows that

χ̂(R) = χ R−1 . (32)
If we distinguish the irreducible representations by indices i, j, then we can

combine the Eqns. (29) and (30) into a single orthogonality relation, wherein we can
write either
1 −1
χj R χi (R) = δij (33)
h R
due to (32), or, in the case of unitary representations,

1 ∗
χ (R)χi (R) = δij . (34)
h R j
6 6
6.3 We apply Pk = |kmkm| to |v = cln |ln. Due to the orthogonality
m ln
relation (37) from problem 6.4, we find immediately

Pk |v = ckn |kn .
n
6.4 We show that

1
Pj = χ̂ j (R)R (35)
h R
is a projection operator, which projects the space of the basis vectors of the repre-
sentation onto the subspace of the irreducible representation j with the character χ j .
Making use of Dirac’s bra and ket notation, we denote the basis vectors by
|k, m , (36)
where k refers to the subspace of the irreducible representation k, and m to the vector
considered within this space. k thus distinguishes the blocks in Fig. 6.14. We first
treat the case that all the blocks belong to different irreducible representations. We
consider unitary representations and therefore can assume the orthonormalisation
k, m | l, n = δkl δmn (37)

Chapter 6 49
and the representation of the identity I in the form

I= |k, m k, m| . (38)
k,m
We now insert I (38) into (35) before and after R and obtain
1
χ̂ j (R) |k, m k, m| R |l, n l, n| . (39)
k R k,m l,n
Keeping the bra and ket notation in mind, we obtain in (39) the matrix element
k, m| R |l, n = δkl u km,n (R) , (40)
where the Kronecker delta δkl results from the zeroes outside the blocks in Fig. 6.14.
Making use of the definition of the group character, χ̂ j (R), we write
j
χ̂ j (R) = δµν û µν (R) . (41)
µν
Employing (40) and (41), (39) can be brought into the form:

1 j
δµν û µν (R)u km,n |k, m k, n| . (42)
µν k,m,n
h R
According to (24) and (25) (with u → u j , v → u k ), the expression in brackets in

(42) reduces to
1
δµn δνm δ jk . (43)
g
Using the Kronecker δ notation, we obtain successively
1
δµν δµn δνm | j, m j, n| , (44)
µν m,n
g
1
δµn δµm | j, m j, n| , (45)
µ m,n
g
1
δµn | j, µ j, n| (46)
µ n
g
and thereby finally for (35)

1 1
Pj = | j, n j, n| = I j . (47)
g n g
50 Solutions
Due to (37), P j is clearly a projection operator onto the subspace j. Here, we use
χ̂ j (R), which is the same for equivalent representations.If several such representa-
tions occur in the representation of a group, then P j projects onto the sum of the
corresponding subspaces, i. e. the definition (35) now leads to
1
Pj = | j, n n, j| , (48)
g j,n
where the sum j extends only over the representations which are equivalent to each
other. The most important aspect of (48) is that subspaces with a particular, identical
symmetry behaviour are sorted out.
The above proof is also interesting in the sense that we do not at all need to know
the explicit form of | j, n, |k, m or to determine it; only general properties enter the
proof.
6.5 The result can be found in the text of 6.
Chapter 7
7.1 a) Let the solutions of the single-particle Schrödinger equation be known:

H( j)χk ( j) = E k χk ( j) , (1)
where we can without loss of generality enumerate the quantum numbers by k =
1, 2 . . . . Since only N particles are present, it suffices to let k run up to N. The
determinant det(χ) is defined by

det(χ) = (−1) P χk1 (1)χk2 (2) . . . χk N (N) , (2)
P(k)
where the sum runs over all permutations k1 , k2 , . . . , k N of 1, 2, . . . , N. For an

even permutation, (−1) P = 1; otherwise = −1. Exchange of two indices (quantum
numbers or particle index) reverses the sign of (2), so that the Pauli principle is
fulfilled. Every term in the sum
4
N
Ψk = χkl (l) (3)
l=1
in (2) obeys

N
N
H( j)Ψk = EΨk with E= Ek , (4)
j=1 k=1
since

N 4
N
N 4
N
H( j) χkl (l) = χkl (l)H( j)χk j ( j) (5)
j=1 l=1 j=1 l=1
l= j
Chapter 7 51
and due to (1),

N 4
N
χkl (l)E k j = EΨk.6) (6)
j=1 l=1
b) Inserting (2) and Ω = V(l, m) into

7 8
1 ∗
Ω= det(χ )V(l, m) det(χ)dV1 . . . dVN (7)
N!
yields
1
Ω=
N!
7 8
4 ∗
× (−1) P+P χk j ( j)V(l, m)χk ( j )dV1 . . . dVN . (8)
j
P(k) P (k ) j, j
Due to
7 8
∗
χ j χk dV = δ jk (9)
the permutations P(k), P (k ) in (8) must be identical up to the factors in which j, j
are the same as l, m, i. e. where
kl = kl
and km = km (10)
or
kl = km
and km = kl . (11)

For (10), the permutations P(k), P (k ) are the same and (−1) P+P = 1, for (11)

they differ by an odd permutation and (−1) P+P = −1. We thus find
 
1
V(l, m) = (N − 2)!  Vkk ,kk − Vkk ,k k  (12)
N! k,k
with
7 8
Vij,kl = χi∗ (1)χ ∗j (2)V(1, 2)χk (1)χl (2)dV1 dV2 . (13)
The factor (N − 2)! comes from the number of possible permutations in (8).
Note that the indices on the variable within the integral play no role. For this reason,
to obtain
52 Solutions
1
Ω= V(l, m) (14)
2 l,m
l=m
it is necessary only to count

1 1
Ω= N(N − 1)V(l, m) = Vkk ,kk − Vkk ,k k . (15)
2 2 k=k
k,k
6 In an analogous but still more simple manner, the expectation value of Ω =

W( j) is found, where W( j) is a single-particle operator, to be given by
j

Ω= Wkk , (16)
k
with
7 8
Wkk = χk∗ (1)W(1)χk (1)dV1 . (17)
7.2 We have
7 8
E= Ψ ∗ HΨdV1 . . . dVN (18)
with
1
H= H( j) + V(l, m) . (19)
j
2 l,m
l=m
From problem 7.1, we have

1
E= Hkk + Vkk ,kk − Vkk ,k k (20)
k
2
k,k
k=k
with
7 8
Hkk = χk∗ Hχk dV , (21)
and the definition (13). W separate the wavefunction χk into its spatial part ψq and
its spin part sm
χk = ψq · sm (22)
Chapter 7 53
with
sm sm = δmm . (23)
Since the electronic
6 shell
6 is closed, there are just as many spin-↑ as spin-↓
electrons, so that → 2 , but in the case of Vkk ,k k , the spins of k and k are the
k q
same:

E=2 Hq,q + 2Vqq ,qq − Vqq ,q q , (24)
q qq
q=q
where

Vq1 q2 ,q3 q4 = ψq∗1 (1)ψq∗2 (2)V(1, 2)ψq3 (1)ψq4 (2)dV1 dV2 . (25)
We vary E (18) with the condition

2
ψq dV = 1 (26)
using the Lagrange parameter 2 εq , with respect to ψq∗ :

2 !
δψ ∗ E − 2 εq
ψq dV = 0 . (27)
q
q
For the individual terms in (27) with (24), we obtain:

∗
A) δψq∗ (1) 2 ψq (1)H(1)ψq (1)dV1 = 2H(1)ψq (1) , (28)
q

∗ ∗
B) δψq∗ (1) 2 ψq (1)ψq (2)V(1, 2)ψq (1)ψq (2)dV1 dV2
q q
q =q

=4 dV2 ψq (2) 2 V(1, 2)ψq (1) (29)
q
q =q

C) δψq∗ (1) − ψq∗ (1)ψq∗ (2)V(1, 2)ψq (1)ψq (2)dV1 dV2
q q
q =q

= −2 dV2 ψq∗ (2)ψq (2)V(1, 2) ψq (1) (30)
q
q =q
54 Solutions
 

D) δψq∗ (1) −2 εq | ψq |2 dV  = −2εq ψq (1) . (31)
q
The Hartree-Fock equations are found from
2 (A + B + C + D) = 0 .
1
(32)
An iterative solution with a step denoted by i means that we can write (32) in
the form
 

(H − εq )ψqi+1 (1)+2 |ψqi (2)|2 V(1, 2)dV2 ψqi+1(1)
q =q
∗

− ψqi (2)ψqi (2)V(1, 2)dV2 ψqi+1
(1) = 0 . (33)
q =q
If the (i + 1)-th step again leads to ψqi , then clearly (33) is fulfilled for ψqi+1 = ψqi =
ψ.
7.3 We obtain −13 Ψmn with the aid of
+
an↓ am↑ Ψ , n = m, k . (1)
With this, we find

4
g
+ + + +
−1 Ψm = ak↓ · an↓ am↓ Φ0 .
3 n
ak↑ (2)
k=1
To apply the spin raising operator

S+ = a+j↑ a j↓ , (3)
j
we not that
a+j↑ a j↓ ak↑
+ + + + +
ak↓ = ak↑ ak↓ a j↑ a j↓ for k = j (4)
and
+ + + + + +
ak↑ ak↓ ak↑ ak↓ = −ak↑ ak↑ ak↓ ak↓ = 0, (5)
as well as
' (
a+j↑ a j↓ an↓
+ +
am↓ = a+j↑ δ jn − an↓+
a j↓ am↓+
' (
= a+j↑ am↓
+
δ jn − a+j↑ an↓
+ +
δm j − am↓ a j↓ . (6)
Chapter 7 55
From (4–6) and aΦ0 = 0 (a is some annihilation operator), it follows almost

immediately that
4
g
+ + + +
0 Ψm = S+− 31 Ψmn = ak↓ · an↑ am↓ Φ0
3 n
ak↑
k=1
4
g
+ + + +
− ak↑ ak↓ am↑ an↓ Φ0 , (7)
k=1
i. e. up to an (unimportant) common sign, we obtain (7.35). In a corresponding

manner, it follows that
4
g
+ + + +
1 Ψm = S+ 30 Ψmn =− ak↓ am↑ · an↑ Φ0 ,
3 n
ak↑ (8)
k=1
i. e. up to an unimportant sign, we obtain (7.36).

7.4 We choose from k = 1, 2, . . . , ∞ a selection of quantum numbers k1 < k2 <
. . . < k N . The possible wavefunctions are
Ψk1 k2 ...k N = ak+1 ak+2 . . . ak+N Φ0 , (1)
where for the vacuum state Φ0 , we have
ak Φ0 = 0 (2)
for all k = 1, 2, . . . , ∞. The energies belonging to (1) are:
E = E k1 + E k2 + . . . + E k N . (3)
7.5 To calculate
9 :

Ψk1 Ω Ψk1 (1)
with
Ψk j = ak+j Φ0 , (2)
where Φ0 is the vacuum state, we make use of the rule

Ψk | = ak+ Φ0 = Φ0 | ak . (3)
Using (2) and
Ω = E k ak+ ak (4)
we obtain with the aid of (3)
9 : 9 :

E k ak+ Φ0 ak+ ak ak+1 Φ0 = E k Φ0 ak1 ak+ ak ak+1 Φ0 . (5)
1
56 Solutions
We apply the exchange relations
ak ak+ + ak+ ak = δkk , (6)
and
ak Φ0 = 0 for all k , (7)
obtaining
9 : 9 :

E k ak+ Φ0 ak+ ak ak+1 Φ0 = E k Φ0 ak1 ak+ ak ak+1 Φ0
1
= E k Φ0 | ak1 ak+ δkk1 |Φ0

= E k Φ0 | Φ0 δkk1 δkk1 , (8)
where
Φ0 | Φ0 = 1 . (9)
The final result for (1) with (2) and (4) is thus
9 :

Ψk1 Ω Ψk1 = E k δkk1 δkk1 . (10)
For the case that

Ω= E k ak+ ak (11)
k
we must carry out the sum.

9 :

Ψk1 Ω Ψk1 = E k δkk1 δkk1 (12)
k
= E k1 δk1 k1 . (13)
Since on the right in
Ω = a+j a+j V j1 j2 ; j1 j2 a j1 a j2 (14)

1 2
there are two annihilation operators, we have

Ω Ψk1 = 0 , (15)
so that (1) vanishes for (14). The computation of (1) and later also of (9) with (4) or
(14) can e carried out in a much more obvious and efficient manner if we interpret the
effect of the annihilation operator ak in (5) as follows: it can annihilate an electron in
the state k only when it was previously created in the same state from the vacuum Φ0 ,
i. e. we must have k = k1 and it follows that
ak ak+ Φ0 = Φ0 . (16)
Chapter 7 57
For k = k1 , on the other hand, we find
ak ak+1 Φ0 = 0 . (17)
Now applying the creation operator ak to (16), we generate the new wavefunction
Ψk = ak+ Φ0 . The bracket in in (5) can now be interpreted as
9 :

ak+ Φ0 ak+ Φ0 ≡ Ψk1 Ψk = δk1 k , (18)
1
which, due to the orthogonality of the wavefunctions is just equal to the Kronecker
delta, as given. The orthogonality can be readily demonstrated from (3), (6), and (7).
Later, we shall explain the procedure for many-particle wavefunctions.
We investigate
9

Ψk1 k2 Ω Ψk1 k2 (19)
for
Ψk1 k2 = ak+1 ak+2 Φ0 (20)
and (4), and initially obtain

9 :

ak+ ak+ Φ0 E k ak+ ak ak+1 ak+2 Φ0 ; (21)
1 2
we then apply the rule

9 9

ak+ ak+ Φ0 = Φ0 ak2 ak1 . (22)
1 2
We obtain from (21)

9 :

E k Φ0 ak2 ak1 ak+ ak ak+1 ak+2 Φ0 . (23)
Using (6), we transfer all the annihilation operators to the right-hand side and
apply (7):
9 :

E k Φ0 ak2 ak1 ak+ ak ak+1 ak+2 Φ0
9 ' ( :

= E k Φ0 ak2 ak1 ak+ δkk1 − ak+1 ak ak+2 Φ0
9 :

= E k Φ0 ak2 ak1 ak+ ak+2 Φ0 δkk1
9 :

− E k Φ0 ak2 ak1 ak+ ak+1 Φ0 δkk2 . (24)
Application of the exchange relations can now be continued. The final result is:
58 Solutions
9 :

E k Φ0 ak2 ak1 ak+ ak ak+1 ak+2 Φ0
= E k (δkk1 δkk1 δk2 k2 − δkk1 δk1 k2 δkk2
− δkk2 δkk1 δk1 k2 + δkk2 δkk2 δk1 k1 ) . (25)
6
For k (13), we find

E k1 + E k2 δk1 k1 δk2 k2 − E k1 + E k2 δk1 k2 δk2 k1 . (26)
The evaluation of (23) and of the following expression (27) can be carried out
much more efficiently (see below). We finally treat
9 :

Ψk1 k2 a+j a+j V j1 j2 ; j1 j2 a j1 a j2 Ψk1 k2 , (27)
1 2
which can be cast in the form
V j1 j2 ; j1 j2 Φ0 | ak2 ak1 a+j a+j a j1 a j2 ak+1 ak+2 |Φ0 . (28)
1 2
Making use of the exchange relations, we transfer all the annihilation operators to
the right-hand side. The somewhat lengthy but elementary computation yields:
9 :

Ψk1 k2 a+j a+j V j1 j2 ; j1 j2 a j1 a j2 Ψk1 k2
1
2
= V j1 j2 ; j1 j2 δk1 j1 δk2 j2 − δ j1 k2 δ j2 k1 δk1 j2 δk2 j1 − δk2 j2 δk1 j1 . (29)
Finally, we introduce the more elegant method for calculating the matrix elements
of many-body functions, which we will explain using the (4) as an example. The
many-body functions have the form
Ψk1 k2 ...k N = ak+1 ak+2 . . . ak+N Φ0 , (30)
where all the k s are different, since otherwise, due to ak+ ak+ = 0, Ψk1 ...k N would
vanish. Furthermore, due to ak+ ak+ = −ak+ ak+ , we can bring the indices (quantum
numbers) into a particular ordering which we are free to determine, e. g. k1 <
k2 . . . < k N . We consider
9 :

Ψk1 ...kN ak+ ak Ψk1 ...k N
9 :

= ak+ . . . ak+ Φ0 ak+ ak ak+1 ak+2 . . . ak+N Φ0 . (31)
1 N
Applying ak to the right, we remember that only an electron already in the state k
can be annihilated. We thus have k = k j for some j = 1, . . . , N. Otherwise, we
obtain zero. To carry out the “annihilation”, we have to bring ak into the j-th position
in ak+1 . . . .ak+N , i. e. ak must be exchanged ( j − 1) times with ak+l , l < j. Due to (6)
with k = k , this yields a factor of (−1) j−1 . We then transfer the creation operator ak+
in (31) to the now empty position. This again yields a factor of (−1) j−1 . All together,
Chapter 8 59
we thus have
ak+ ak Ψk1 ...k N = Ψk1 ...k N if k = k j , j = 1, . . . , N , (32)

otherwise = 0 . (33)
In the case (32), the orthogonality relation

9 :

Ψk1 k2 ...kN Ψk1 k2 ...k N = δk1 k1 . . . δk N kN (34)
therefore is all that remains of (31). It can readily be proved with the aid of (6) and
(7). The procedure can be applied in an analogous way to
Ω = ak+ ak (35)
and to (27) with the corresponding functions (30).

7.6 Instead of the indices i, j, k, l in (7.55), we introduce double indices k, s, where
k = 1, 2, . . . refers to the quantum number(s) of the spatial part of the electronic
wavefunctions and s = 1/2, −1/2 to the spin quantum numbers, which we can
denote more simply by ↑, ↓. The wavefunction −13 Ψmn (7.33) in 2nd quantisation is
then given by
4
g
+ + + +
−1 Ψm = ak↓ · am↓ an↓ Φ0 .
3 n
ak↑ (1)
k=1
To evaluate (7.37), we must compute the expectation value of (7.55) with the
corresponding new indices with respect to (1). For the computation of the expectation
values, we can use the fact that the annihilation operators ak+ ↓ or ak+ ↑ act in (1) where
k = k and the spins are the same. Here, however, we must keep in mind the sign
change due to the exchange operations. The final result is given in (7.37).
Chapter 8
8.1 According to Heisenberg’s uncertainty relations, the lifetime τ = ∆t and the

energy uncertainty ∆E of a state is given by
∆E · ∆t ≈ h . (1)
Since the ground state has an infinite lifetime, its energy is precisely defined.
The lifetime broadening of a transition line is thus due only to the energy width,
i. e. the uncertainty, of the excited state:
∆E 1
∆ν = = . (2)
hc c · ∆t
60 Solutions
For the energy uncertainties of the two excited states a and b and the linewidths
of their transitions to the ground state, we thus have
a) ∆t = 10 s : ∆E = 6.6 · 10−35 J
ν = 3.3 · 10−12 cm−1
b) ∆t = 1 ns : ∆E = 6.6 · 10−25 J
ν = 3.3 · 10−2 cm−1 .
8.2 The thermal population ratio N2 /N1 of two states with the energies E 1 and E 2
is given in equilibrium by the Boltzmann distribution
N2
= e−∆E/k T (1)
N1
(∆E = E 2 − E 1 ; k = 1.38 · 10−23 KJ : Boltzmann’s constant; T : temperature;

the differing statistical weights of the states have been neglected here). For a given
number N = N1 + N2 of molecules in an ensemble, we then find for the thermal
population of state 2:
N N
N2 = = . (2)
1 + N1 /N2 1 + e∆E/k T
With the value given, N = 105 , one obtains for ∆E = 20.15 cm−1 :
a) T = 29 K : N2 = 2.69 · 104
b) T = 290 K : N2 = 4.75 · 104
c) T = 2900 K : N2 = 4.97 · 104
and for ∆E = 20 150 cm−1 :
a) T = 29 K : N2 = 0
b) T = 290 K : N2 = 0
c) T = 2900 K : N2 = 5
While neighbouring rotational terms are already nearly equally populated at

room temperature, electronically excited states are hardly populated thermally even
at high temperatures.
Chapter 9
9.1 Only molecules with permanent electric dipole moments exhibit a pure mi-
crowave rotational spectrum: H2 O, H2 O2 , CH3 Cl, CH2 Cl2 , NH3 , NH4 Cl, HCl, HBr.
For molecules such as CH4 and CS2 , rotational spectra become observable only
when IR-active vibrations are excited.
Chapter 9 61
9.2 a) The spectral position ν of the transition between the rotational levels J and
J + 1 is (from (9.15)) given by
ν J→J+1 = F(J + 1) − F(J) = 2B (J + 1) . (1)
The spectral spacing ∆ν between each two neighbouring lines is, in this simpli-
fied model, independent of J:
∆ν = ν J+1→J+2 − ν J→J+1 = 2 B . (2)
With the given value ∆ν = 0.71433 cm−1 , one finds for the rotational constant
of Br 19 F:
79
B = 0.357 cm−1
and for the moment of inertia
Θ = 7.84 · 10−46 kg m2 .
The moment of inertia of a diatomic molecule with the atomic masses m 1 and m 2
and the bond length (i. e. the internuclear distance) R can be calculated from
Θ = m 1 R12 + m 2 R22 = m r R2 (3)
where
m1m2
mr = (4)
m1 + m2
denotes the reduced mass in the centre-of-mass system. With (3) and (4), one obtains
from Θ the bond lengths
R = 1.76 · 10−10 m = 1.76 Å .
The position of the transition J = 9 → (J + 1) = 10 can be computed from

(1); the result is
ν9→10 = 7.14 cm−1 .
The transition moment between two rotational levels of a linear molecule depends
to a good approximation only on the electric dipole moment of the molecule and
can therefore be considered to be constant for all purely rotational transitions. The
intensity of a line in a pure rotational spectrum can thus be explained in terms of
the population N J of the absorbing state alone; it in turn is given by the degeneracy
factor g J of the level and its thermal population at the temperature T (Boltzmann
distribution):
NJ g J −(E J −E0 )/kT
= e (5)
N0 g0
62 Solutions
and thus, due to g J = 2J + 1,

N J = N0 (2J + 1) e−hcB J(J+1)/kT . (6)
From the derivative of this function with respect to J, the most strongly populated
state Jmax is found to be given by the value of

kT 1
J= − (7)
2hcB 2
rounded to an integer.
It then follows for 79 Br 19 F at T = 298 K that J = 16.53 and thus
Jmax = 17 .
The most intense transition is therefore observed for J = 17 → J = 18.
b) Classical mechanics gives for the energy of a rigid rotor
1
E= Θω2 = 2π 2 Θ ν2 , (8)
2
where ν is the rotational frequency of the molecule. Setting this expression equal
to the quantum-mechanically calculated energy of the state, E = hcB J(J + 1), we
find
h % %
ν= J(J + 1) = 2cB J(J + 1) . (9)
4π Θ
2
From this, one obtains

J = 0: ν = 0
J = 1 : ν = 3.0 · 1010 s−1
J = 10 : ν = 2.2 · 1011 s−1 .
9.3 From the atomic masses

m 1 H = 1.673 · 10−27 kg
m 2 D = 3.344 · 10−27 kg
m 35 Cl = 58.06 · 10−27 kg
m 37 Cl = 61.38 · 10−27 kg
one can compute the reduced masses of the diatomic molecules, finding
m r (1 H 35 Cl) = 1.626 · 10−27 kg
m r (1 H 37 Cl) = 1.629 · 10−27 kg
m r (2 D 35 Cl) = 3.162 · 10−27 kg .
From the formula for the rotational constant,
h h
B= 2
= (1)
8π cΘ 8π c m r R2
2
Chapter 9 63
and under the assumption that the bond length is the same for different isotopes, it
follows that
B1 m r,2
= (2)
B2 m r,1
and thus with B1 H 35 Cl = 10.5909 cm−1 :
B1 H 37 Cl = 10.5714 cm−1 and B2 D 35 Cl = 5.446 cm−1 .
9.4 The energy of a transition line J → J + 1 in a pure rotational spectrum is

given in terms of the rotational constant B by
ν = 2B (J + 1) (1)
With B = 10.6 cm−1 , one obtains
for ν1 = 106.0 cm−1 : J1 = 4

for ν2 = 233.2 cm−1 : J2 = 10 .
Under the assumption of a constant transition moment, the line intensity is
proportional to the population N(J) of the absorbing rotational level J. According to
(9.16), it depends on the degeneracy factor (2J + 1) and on the thermal population
from the Boltzmann factor:
N J = N0 (2J + 1) e−hcB J(J+1)/kT . (2)
The transitions J1 and J2 have the same intensity for N J1 = N J2 and therefore
for
N0 (2J1 +1) e−hcB J1 (J1 +1)/kT = N0 (2J2 +1)e−hcB J2 (J2 +1)/kT . (3)
With the values computed above for J1 and J2 , this condition is met at a temper-
ature of
90 hcB
T= . (4)
k ln 73
From this, one computes
T = 1620.8 K .
9.5 The spectral position of a line in a purely rotational spectrum of a rigid rotor
can be calculated from (9.15):
ν J→J+1 = F(J + 1) − F(J) = 2B (J + 1) . (1)

64 Solutions
From this, a constant spacing of ∆ν between any two neighbouring lines is

found:
∆ν = 2B = const. (2)
Since the centrifugal stretching constant D is in general several orders of mag-
nitude smaller than the rotational constant B, this first approximation is sufficient to
associate the transitions ν J→J+1 with the corresponding levels J. The lines given in
the problem
J1 → J2 : ν1 = F(J2 ) − F(J1 ) = 84.544 cm−1
J2 → J3 : ν2 = F(J3 ) − F(J2 ) = 101.355 cm−1
J3 → J4 : ν3 = F(J4 ) − F(J3 ) = 118.112 cm−1
have a spacing of 2B ≈ 16.8 cm−1 , which, together with (1), permits the following
identification:
J1 = 4; J2 = 5 ; J3 = 6 ; J4 = 7 .
The model of a rigid rotor is greatly simplified. On replacing it by a nonrigid
rotor, one takes the effect of centrifugal stretching into account and thus obtains for
the energy (in cm−1 ) of a pure rotational state of quantum number J
F(J) = B J(J + 1) − D J 2 (J + 1)2 , (3)
where D is the centrifugal stretching constant; for a diatomic molecule, we have
3
D= (4)
4π k Θ 2 Re2 c
(Θ: moment of inertia of the molecule; Re : equilibrium internuclear distance without
rotation; k: force constant of the elastic valence oscillation along the internuclear
axis). The line belonging to the transition J → J + 1 has the energy
ν J→J+1 = F(J + 1) − F(J) = 2B (J + 1) − 4D (J + 1)3 . (5)
Inserting the corresponding J values for the first two transitions given, ν1 and
ν2 , we obtain the system of linear equations
J1 = 4 → J2 = 5 : ν1 = 10 B − 500 D (6)
J2 = 5 → J3 = 6 : ν2 = 12 B − 864 D (7)
and thus
6ν1 − 5ν2 1
D= and B= ν1 + 50 D . (8)
1320 10
Inserting the numerical values yields
D = 3.70 · 10−4 cm−1 and B = 8.47 cm−1 .
Chapter 9 65
The resulting exact value of B agrees sufficiently well with the value used above,
so that the correspondence of the J values to the lines is confirmed. Making use of
the well-known formula for the rotational constant,
h
B= (9)
8π 2 c m r Re2
and inserting the atomic masses m 1 H = 1.673 ·10−27 kg and m 79 Br = 131.03 ·10−27 kg
as well as
m 1 H m 79 Br
mr = , (10)
m 1 H + m 79 Br
the equilibrium distance between the nuclei and thus the average bond length of the
molecule at rest can be found:
Re = 1.41 · 10−10 m = 1.41 Å .
The model of the nonrigid diatomic rotor is based on the assumption of an elastic
coupling of the nuclei with the force constant k. The resulting harmonic oscillation
along the bond axis has the fundamental frequency

k
ω= . (11)
mr
Solving this equation for k and inserting into (4), and making use of (9) with
some elementary rearrangements, we find for the oscillation frequency

B3
ω = 4π c . (12)
D
With the values of D and B obtained above, this yields
∧
ω = 4.83 · 1014 s−1 = ν = 2564 cm−1 .
9.6 The moment of inertia Θ of the hydrogen cyanide or Prussic acid molecule
H−C≡N in its centre-of-mass system is
Θ = m C RC2 + m N RN
2
+ m H RH
2
. (1)
Here, m C , m N and m H are the relative atomic masses of carbon, nitrogen and
hydrogen, and RC , RN , RH are the distances of each of the atoms from the centre of
mass.
In the centre-of-mass system, we have
m N RN = m C RC + m H RH (2)
so that, with
RCN = RC + RN (3)
66 Solutions
and
RCH = RH − RC (4)
we have three equations at our disposal with which the distances in the centre-of-
mass system in (1) can be eliminated. With the total molecular mass
M = mC + mN + mH , (5)
equation (1) for the moment of inertia can be rearranged to give

1
Θ = m N RCN
2
+ m H RCH
2
− (m N RCN − m H RCH )2 . (6)
M
Inserting the values given, we obtain
Θ = 1.885 · 10−46 kg m2
and from this, using

h
B=
8π 2 c Θ
we arrive at the rotational constant
B = 1.485 cm−1 .
Fig. S.20. A linear triatomic molecule HCN: bond lengths and distances from the atoms to
the centre of mass S
9.7 The rotational energy of a generalised three-dimensional body transformed

to the principal axis system with the remaining diagonal elements Θxx , Θ yy and
Θzz of the moment of inertia tensor and the components L x , L y and L z of angular
momentum L is given by
L 2x L 2y L 2z
E= + + . (1)
2 Θxx 2 Θ yy 2 Θzz
Chapter 9 67
The ammonia molecule, NH3 , is a symmetric top with an axis of rotation which
in the following is taken to be the z axis of the coordinate system. Then the following
relation holds:
Θzz ≡ Θ . (2)
Owing to the symmetry of NH3 , the two remaining principal components of the
moment of inertia are equal; the associated principal axes are perpendicular to the
rotational axis:
Θxx = Θ yy ≡ Θ⊥ . (3)
Inserting this in (1) and making use of L 2 = L 2x + L 2y + L 2z , we obtain
L 2x + L 2y L 2z
E= + (4)
2 Θ⊥ 2 Θ
2
L 1 1
= + L 2z − . (5)
2 Θ⊥ 2 Θ 2 Θ⊥
In addition to the quantisation of the total angular momentum L of the molecule

according to
L 2 = 2 J(J + 1) , (6)
in a symmetric-top molecule there exists a quantisation of the component L z of

the angular momentum along the rotational axis, since the symmetry of the charge
distribution makes one direction in space unique:
Lz = K with K = 0, ±1, ±2, . . . , ±J . (7)
Then we obtain from (5)

2
2 2
E= J(J + 1) + − K2 (8)
2 Θ⊥ 2 Θ 2 Θ⊥
and finally
E
F(J, K) = = B J(J + 1) + C K 2 (9)
hc
with the rotational constants

h h 1 1
B= and C = − . (10)
8π 2 c Θ⊥ 8π 2 c Θ Θ⊥
From the data and formulas given, we can compute the two principal elements
of the moment of inertia tensor of the symmetric-top molecule, obtaining
68 Solutions
Θ⊥ = 2.86 · 10−47 kg m2
Θ = 4.38 · 10−47 kg m2
and from this,
B = 9.78 cm−1
C = −3.39 cm−1 .
Due to Θ > Θ⊥ , we require C < 0, so that in the case of NH3 we are dealing with
a “cushion-shaped” molecule. The rotational spectrum of the molecule is obtained
from (9) taking the selection rules
∆J = ±1, ∆K = 0 (11)
into account, so that we find for the transition energies:
ν J→J+1 = F(J + 1, K) − F(J, K) = 2B (J + 1) . (12)
9.8 As derived in problem 9.5 (Eq. 12), the frequency of the valence oscillation of
a diatomic molecule is determined from the rotational constant B and the centrifugal
stretching constant D according to the relation

B3
ω = 4π c . (1)
D
With the values given, we find
∧
ω = 5.64 · 1014 s−1 = ν = 2995 cm−1 .
The experimental value is ν = 2991 cm−1 . The discrepancy results from the
use of an insufficiently precise model as the basis of the calculation: the diatomic
molecule is treated as a harmonic oscillator by approximating the bond as an elastic
spring and applying the linear force relation (Hooke’s Law), F = −k (R− Re ). Mak-
ing use of an anharmonic oscillator potential for the bond leads to an improvement
of the calculated values.
9.9 The reduced mass of 12 C 16 O is 1.139 · 10−26 kg, and the bond length
is 112.82 · 10−12 m. From these values, we can compute the rotational constant
h h
B= = , (1)
8π 2cΘ 8π c m r R2
2
obtaining
B = 1.931 cm−1 .
The transition frequencies in the microwave rotational spectrum are given by
ν J→J+1 = 2B (J + 1) . (2)
Chapter 9 69
For the first four lines, one obtains from this formula the microwave frequencies:
∧
ν0→1 = 2B = 3.862 cm−1 = 115.86 GHz
∧
ν1→2 = 4B = 7.724 cm−1 = 231.72 GHz
∧
ν2→3 = 6B = 11.586 cm−1 = 347.58 GHz and
∧
ν3→4 = 8B = 15.448 cm−1 = 463.44 GHz .
An analogous calculation leads for the isotopes 13 C 16 O to m r = 1.191 · 10−26 kg
and B = 1.847 cm−1 , and thus to
∧
ν0→1 = 3.694 cm−1 = 110.82 GHz ,
In order to obtain a spectral separation of the two lines belonging to the 0 → 1
transitions, one would need a resolution defined by
ν ν0→1
A= = . (3)
∆ν ν0→1 − ν0→1
With ∆ν = 0.168 cm−1 or ∆ν = 5.04 GHz, this corresponds to a resolution of
A = 23.0 .
9.10 The vector of total angular momentum L of a molecule has 2J +1 orientational

components with respect to a particular direction in space. Without an external
electric field, the 2J + 1 states are energetically degenerate. For a molecule with
a permanent electric dipole moment, this degeneracy is lifted in an electric field E
by the Stark effect, since the different spatial orientations of the angular momentum
relative to the direction of the field correspond to different energy shifts ∆E:
p2 E 2
∆E J = f(J, M 2 ) (1)
B
with M = 0, ±1, ±2, . . . , ±J ( p: electric dipole moment). For M = 0, a twofold
degeneracy remains, and therefore each rotational level J exhibits (J + 1)-fold
splitting:
J = 2 : 3-fold splitting (M = 0, ±1, ±2)
J = 3 : 4-fold splitting (M = 0, ±1, ±2, ±3) .
As a result of the selection rules ∆J = ±1 and ∆M = 0, each line in the
rotational spectrum thus shows a (J + 1)-fold splitting. Then e. g. the transition
J = 2 → (J + 1) = 3 leads to three spectral lines:
J = 2, M = 0 → J = 3, M = 0
J = 2, M = ±1 → J = 3, M = ±1
J = 2, M = ±2 → J = 3, M = ±2 .
The number of lines into which a transition is split in an E-field gives a unique
method of identifying the J values in the rotational spectrum.
70 Solutions
9.11 Setting the classical energy of a rigid rotor

1
E rot,klass =Θω2 (1)
2
equal to the term calculated quantum mechanically:
2
E rot,QM = hcB J(J + 1) = J(J + 1) , (2)
2Θ
one obtains the dependence of the circular frequency ω (which is defined only
classically) on the rotational quantum number J (which is defined only quantum
mechanically):
%
ω(J) = J(J + 1) . (3)
Θ
Projected onto a plane, the rotating dipole moment can be identified within the
framework of classical electrodynamics as a moment which oscillates harmonically
at the frequency ω. It radiates electromagnetic waves whose frequency (in cm−1 )
corresponds to the rotational frequency of the rotor:
ω(J) %
νklass = = J(J + 1) . (4)
2πc 2πcΘ
In contrast, the quantum-mechanical result for the frequency of a (very improb-
able) radiative transition from J → J − 1 is

νQM = 2B J = J. (5)
2πcΘ
The ratio of the two frequencies is

νQM J
= . (6)
νklass J +1
In the classical limit J → ∞, the two theories yield the same result.
Chapter 10
10.1 All molecules whose dipole moment changes during oscillation also absorb
radiation in the infrared spectral region (vibrational absorption spectra). In the case
of symmetric molecules, in particular bending vibrations and asymmetric stretching
vibrations are important, since the symmetric stretch vibrations are IR inactive. Thus
of the molecules given, all exhibit IR absorption with the exception of H2 and N2 .
In the case of the linear azide ion, N−3 , the bending vibration is IR active, as can be
seen from the mesomeric representation:
|N−N≡N| − ←→ N=N=N − ←→ |N≡N−N| − .
Chapter 10 71
10.2 The total number of all the thermodynamic degrees of freedom of an N-atomic
molecule is f = 3N. The balance for translation, rotation, and vibration is as follows:
Total number of degrees of freedom: f = 3N

Number of translational degrees of freedom: f T = 3
Number of rotational de-
2 for a linear molecule
grees of freedom: fR =
3 otherwise
Number of vibrational
3N − 5 for a linear molecule
degrees of freedom: fV =
3N − 6 otherwise .
From this it follows that the number of vibrational degrees of freedom f V for
a) HBr (linear) : fV =1
b) OCS (linear) : fV =4
c) SO2 (nonlinear) : fV =3
d) H−O−O−H (nonlinear) : fV =6
e) H−C≡C−H (linear) : fV =7
f) C6 H6 : fV = 30 .
10.3 A quantum-mechanical calculation yields for the energy terms G v of a har-

monic oscillator

1 1 1 ω
Gv = ω v + = νe v + with νe = . (1)
hc 2 2 2π c
The vibrational frequency of the harmonic oscillator is

k
ω= , (2)
mr
so that for a transition between neighbouring states v and v + 1, it follows that:
ω
νv → v+1 = G v+1 − G v = νe = . (3)
2π c
With (2), one therefore obtains the wavenumber of a photon which can stimulate
transitions between neighbouring levels:

1 k
νv → v+1 = (4)
2π c m r
as well as the wavelength of the photon:
1
λ= . (5)
νv → v+1
72 Solutions
From the given values, k = 855 N/m and m r = m Proton = 1.6726 · 10−27 kg,
one obtains
νv → v+1 = 3795.6 cm−1 and λ = 2.60 m . (6)
10.4 Since the energy levels of an harmonic oscillator obey the equation

1
E v = ω v + with v = 0, 1, 2, . . . , (1)
2
the minimum excitation energy E min is given by

k
E min = ω = , (2)
mr
where for the vibrational frequency, the relation

k
ω= (3)
mr
(m r : reduced mass) is employed.

a) For the limiting case of small displacements, the equation of motion of a
frictionless pendulum in a gravitational field can be used as the equation of motion
of the harmonic oscillator. The force constant k corresponds in this case to
mg
k= (4)
l
with m = m r : pendulum mass; g = 9.81 m/s2 : acceleration of gravity; l = 1 m:
length of the pendulum. Then the oscillation frequency is

k g
ω= = (5)
mr l
so that it follows from (2) that:
E min = 3.3 · 10−34 J = 2.1 · 10−15 eV = 1.7 · 10−11 cm−1 .
b) At the oscillation frequency ν = 5 Hz and ω = 2πν, one finds for the movement
of a mechanical clock the minimum excitation energy
E min = 3.3 · 10−33 J = 2.1 · 10−14 eV = 1.7 · 10−10 cm−1 .
c) Correspondingly, for the quartz oscillator of a clock with ν = 33 kHz
E min = 2.2 · 10−29 J = 1.4 · 10−10 eV = 1.1 · 10−6 cm−1 .

Chapter 10 73
d) The minimum excitation energy of the valence oscillation of O2 with a force

constant of k = 1177 N/m and m r = 13.28 · 10−27 kg is given by
E min = 3.1 · 10−20 J = 0.2 eV = 1580 cm−1 .
It can be seen that in the cases (a) to (c), the minimum excitation energies are so
small that one can assume a practically continuous energy spectrum. In the oxygen
molecule in (d), the effect of quantisation is, in contrast, clearly observable, and the
system can no longer be described classically.
10.5 Treating the N≡O molecule as an anharmonic oscillator, we find the vibronic
energy levels to be given by

1 1 2
E v = ωe v + − xe ωe v + (1)
2 2
and with
ωe = 2πc ν e , (2)
the vibrational energy terms are:

Ev 1 1 2
Gv = = νe v + − xe νe v + . (3)
hc 2 2
The transition energies ν0→v from the ground state (v = 0) to an excited vibra-
tional level (v > 0) are found from this to be
ν0→v = G v − G 0 = νe v [1 − xe (v + 1)] . (4)
Then for the fundamental frequency (transition from v = 0 → v = 1), we find
ν0→1 = νe (1 − 2xe ) , (5)
and for the first harmonic (transition from v = 0 → v = 2)
ν0→2 = 2 ν e (1 − 3xe ) . (6)
Equations (5) and (6) form a linear system of two equations for two unknowns (νe ,
xe ). From them, we obtain the vibrational constant
νe = 3ν0→1 − ν0→2 (7)
and the anharmonicity constant
2ν0→1 − ν0→2
xe = . (8)
6ν0→1 − 2ν0→2
With the given values for the central frequencies of the rotational-vibrational
transitions (∆J = 0), one obtains
νe = 1903.98 cm−1 and xe = 0.007332 .
74 Solutions
In the vibronic ground state, v = 0, the anharmonic oscillator has according to

(3) the zero-point energy
1 ' xe (
G0 = νe 1 − (9)
2 2
or
G 0 = 948.5 cm−1 .
Treating the molecule approximately as an harmonic oscillator with the reduced

mass
m 14 N · m 16 O
mr = , (10)
m 14 N + m 16 O
one assumes a linear force law and computes the force constant k from
k = m r ω2e . (11)
The energy levels of the harmonic oscillator are

1
E v = ωe v + , (12)
2
so that in going from the harmonic (12) to the anharmonic oscillator (1), we must
make th substitution

1
ωe −→ ωe 1 − xe v + . (13)
2
The constant vibrational frequency ω e of the harmonic oscillator thus becomes a
frequency ν that depends on the vibrational quantum number. In the anharmonic
oscillator, the equilibrium vibrational frequency defined in (1) and used in (11), ωe ,
thus corresponds to the frequency of an infinitesimally small oscillation about the
equilibrium internuclear distance in the hypothetical state v = − 12 .
With ωe = 3.59 · 1014 s−1 and m r = 1.24 · 10−27 kg, one finds for the force
constant
kg
k = 1594.9 .
s2
Treating the vibrational quantum number v as a continuous variable, the expres-
sion (3) for the vibrational terms G v can be differentiated with respect to v:

1 1 2
G(v) = νe v+ − xe νe v + (14)
2 2

dG(v) 1
= νe − 2xe νe v + . (15)
dv 2
Chapter 10 75
From this, we can find at the frequency

1 1
vmax = − = 67.69 ≈ 68 (16)
2 xe 2
the maximum of the function G(v):
G(vmax ) = 64 919 cm−1 .
This maximum vibrational energy can be considered as the lower limit of the
energy continuum, so that after subtracting the zero-point energy, the dissociation
energy D0 of the molecule if found to be:
D0 = G(vmax ) − G 0 (17)
or
kJ
D0 = 63 970 cm−1 = 7.93 eV = 765.3 .
mole
This is not an especially good estimate. Its deviation from the experimental
value
0 = 5.91 eV is due in particular to the neglect of terms of higher order in
D
v + 12 in the expression for the energies. The cubic, quadratic and higher terms
are important especially for large values of v; in the approximation used, we have
namely G(v) → −∞ for v → ∞.
10.6 The given data permit a calculation only in the harmonic oscillator approxi-
mation. For the energy terms, we thus find the expression

1
G v = νe v + , (1)
2
and thus we have for the zero-point energy of the molecule
1
G0 = νe . (2)
2
Under the assumption that the molecule is in its vibrational ground state v = 0
both before and immediately after the reaction, the energy difference ∆E between
the reactants and the products is due solely to their different zero-point energies:
HCl + D2 −→ DCl + HD (+∆E) (3)

∆E = G 0,HCl + G 0,D2 − (G 0,DCl + G 0,HD ) . (4)
With the given values for the νe of the individual molecules, we obtain
kJ
∆E = 129 cm−1 = 1.54 .
mole
In the reaction of one mole each of reactant molecules, an energy of 1.54 kJ is
thus set free – the reaction is exothermic.
76 Solutions
10.7 The force constant k of a molecular bond can be calculated on the model of
an harmonic oscillator from
k = m r ω2 = 4π 2 c2 m r ν2 . (1)
Here, m r is the reduced mass of the molecule and ν is the frequency of the
fundamental oscillation. For the hydrogen halides, the following table can be written:
1
H 19 F 1
H 35 Cl 1
H 81 Br 1
H 127 I
ν/cm−1 4141.3 2988.9 2649.7 2309.5
m r /10−27 kg 1.598 1.624 1.653 1.661
k/N/m 968.0 515.7 411.7 314.4
10.8 The iodine molecule is treated as an anharmonic oscillator, so that the vibronic
energy terms are given by

1 1 2
G v = νe v+ − xe νe v + . (1)
2 2
From this, we find the spacing ν0→1 of the first two vibrational levels
ν0→1 = νe (1 − 2xe ) (2)
and with the given values for the vibrational constant νe and the anharmonicity
constant xe
ν0→1 = 213.71 cm−1 .
The intensity of the vibrational transitions with ∆v = 1 is proportional to the

population Nv of the absorbing state v, since the excited states are nearly unpopulated
at low temperatures. From the Boltzmann distribution, we then find for the ratio of
the intensities I0→1 of the fundamental vibrational band and I1→2 of the “hot band”
at T = 300 K:
I1→2 N1
= = e−hcν0→1 /kT (3)
I0→1 N0
(more precisely, this holds for every pair of equivalent rotational-vibrational transi-
tions), and thus
I1→2
= 0.359 .
I0→1
Chapter 10 77
10.9 The diatomic functional groups which contribute are treated as a first approx-
imation as harmonic oscillators. Their vibrational frequency is then

k
ω = 2πc ν = . (1)
mr
As an additional approximation, we assume that on exchange of two isotopes,

e. g. H ↔ D, or two atoms which are similar in their bonding properties and have
the same electron configurations in their outer shells, e. g. O ↔ S, the bond strength
remains unchanged. While the reduced mass m r thus is different after the exchange,
the force constant k remains the same. The relation between the wavenumber ν of
the oscillation and the reduced mass m r of the functional groups for the same force
constant is then according to (1)
1
ν∝√ . (2)
mr
The frequency of the O−D stretching vibration can be estimated from the value
for the O−H vibration:

m r,OH
νO–D ≈ νO–H (3)
m r,OD
and with νO–H = 3600 cm−1 :
νO–D ≈ 2620 cm−1 .
Since the force constant of the C−S group is comparable at most with that of
the C−O single bond, but not with that of the C=S double bond (different bonding
character), they can be estimated as:

m r,CO
νC–S ≈ νC–O (4)
m r,CS
and with νC–O = 1150 cm−1 :
νC–S ≈ 1020 cm−1 .
10.10 The problem can be solved only in the harmonic-oscillator approximation

with the given data.
a) The energy levels G v of the harmonic oscillator are

1
G v = νe v + . (1)
2
78 Solutions
The vibrational constant νe corresponds here to the energy of the fundamental

oscillation of the harmonic oscillator and can be computed from the force constant
k and the reduced mass m r = 1.6266 · 10−27 kg of the HCl molecule from

ωe 1 k
νe = = , (2)
2πc 2πc m r
leading to
νe = 2991.3 cm−1 .
For v = 0, the zero-point energy is found to be

1
G0 = νe (3)
2
or
G 0 = 1495.7 cm−1 .
b) To calculate the rotational-vibrational spectrum, we need the rotational con-

stant B. It can be determined from the equilibrium bond length Re and the reduced
mass m r according to
h
B= , (4)
8π 2 cm r Re2
leading to the result
B = 10.59 cm−1 .
c) Without taking centrifugal stretching and vibrational stretching into account,

the energy terms of the states with vibrational quantum number v and rotational
quantum number J are given by

1
G(v, J) = νe v + + B J(J + 1) . (5)
2
The lines of the rotational-vibrational spectrum are found from the transition
energies between the states (v , J ) and (v , J ), with the convention v > v , to be
ν(v , v ; J , J ) = G(v , J ) − G(v , J ) (6)

= νe (v − v ) + B J (J + 1) − J (J + 1) . (7)
For the fundamental oscillation, v = 1 and v = 0, so that Eq. (7) can be
simplified to

ν(J , J ) = νe + B J (J + 1) − J (J + 1) . (8)
Chapter 10 79
From this we obtain

for the P branch (J = J − 1) : ν(J ) = νe − 2B J
for the R branch (J = J + 1) : ν(J ) = νe + 2B (J + 1) .
Finally, one computes from these expressions the innermost three lines of the P
and R branches:
P(1) : J = 1 ⇒ ν = 2970.12 cm−1

P(2) : J = 2 ⇒ ν = 2948.94 cm−1
P(3) : J = 3 ⇒ ν = 2927.76 cm−1
R(0) : J = 0 ⇒ ν = 3012.48 cm−1
R(1) : J = 1 ⇒ ν = 3033.66 cm−1
R(2) : J = 2 ⇒ ν = 3054.84 cm−1 .
d) Two qualitative differences are found between the spectrum calculated in (c)
and the experimentally-determined spectrum.
On the one hand, chlorine is found in nature as an isotopic mixture of 35 Cl and
37
Cl in the ratio 3:1. Due to the different reduced masses of the two isotopically
different HCl molecules, one observes two spectra which are shifted relative to each
other and exhibit an intensity ratio of 3:1.
Secondly, another difference is due to vibrational stretching. It leads to a loss
of validity of the Born-Oppenheimer approximation, since the internuclear distance
now depends on the vibrational state. The rotational constants Bv and Bv are then
different, and the transition energies from (7) are
ν(v , v ; J , J ) = νe (v − v ) + Bv J (J + 1)

− Bv J (J + 1) . (9)
For the P branch, (J = J − 1) follows from a simple rearrangement
ν P = νe (v − v ) − 2Bv (J + 1)

− (Bv − Bv ) J (J + 1) . (10)
Since the internuclear distance is greater in the higher state v than in the state
with v , and therefore Bv > Bv , the lines of the P branch shift further apart with
increasing J (i. e. in the direction towards the outside of the vibrational band).
For the R branch (J = J + 1), from (9) we find
ν R = νe (v − v ) + 2Bv (J + 1)

− (Bv − Bv ) (J + 1)(J + 2) . (11)
Since Bv > Bv , the lines in the R branch move closer together with increasing
J , until they finally reverse along the wavenumber axis and shift towards the centre
of the band. The R branch thus exhibits a sharp edge.
80 Solutions
10.11 The selection rules for the rotational-vibrational transition (v , J , M J ) −→
(v , J , M J ) with M J = −J, . . . , +J are given by
∆v = 0, ±1, ±2, . . .
∆J = ±1
and ∆M J = 0, ±1 .
In the P branch, (J = J − 1), the degree of degeneracy of the absorbing state

is 2J + 1, since the M J values run from −J to +J . In contrast, in the state
J , the quantum number M J can take on only values from −(J − 1) to (J − 1).
For ∆M J = −1, 0 or + 1 there are thus exactly 2J − 1 subtransitions, so that all
together, 6J − 3 = 6J + 3 subtransitions are possible.
In the R branch, (J = J + 1), M J can take on values from −(J + 1) to
(J + 1), and there are in each case 2J + 1 subtransitions for ∆M J = −1, 0, +1.
All together, one thus obtains 6J + 3 subtransitions.
With the definition J = min (J , J ), for both branches there are 3 · (2J + 1)
subtransitions with ∆M J = 0, ±1. If each subtransition makes the same contribution
to the line intensity, the latter is proportional to the product of the number of
subtransitions times the thermal population of the absorbing state. The ratio of a line
intensity to that of the first line of the R-branch is then given by
Iv ,J →v ,J 3 (2J + 1) − hc Bv [ J (J +1)−0]
= e kT
Iv ,0→v ,1 3
hc B J (J +1)
= (2J + 1) e− kT v .
10.12 As derived in problem (10.10d), the positions of the lines in the R branch
(J = J + 1) of the rotational-vibrational spectrum of a diatomic molecule such as
HCl are found from
ν R = νe (v − v ) + 2Bv (J + 1)

− (Bv − Bv ) (J + 1)(J + 2) . (1)
With increasing J , the lines move closer together, until finally at JK a direction
reversal along the wavenumber axis occurs. JK is the level which forms the band
edge of the R branch.
The spacing between two neighbouring lines in the R branch with J = JK and

J = JK + 1 is
∆ν R = . . . = 2Bv − 2 (Bv − Bv )(JK + 2) . (2)
At the band edge, ∆ν R becomes negative, so that with Bv > Bv , the value of
the rotational quantum number JK can be calculated:
Bv
JK > − 2. (3)
Bv − Bv
Chapter 10 81
From Bv = B0 and Bv = B1 , one finds
JK > 32.46 ⇒ JK = 33 .
Beyond JK + 1 = 34, the line of the R branch move again towards the centre
of the band. The line corresponding to JK forms the band edge. From (1), it follows
for the fundamental vibrational band that
ν R (JK ) = ν(1, 0) + 2B0 (JK + 1)

− (B0 − B1 ) (JK + 1)(JK + 2) (4)
and, with the given value of ν(1, 0), the band edge is at
ν R (JK ) = 3235.25 cm−1 .
(For comparison: ν R (JK − 1) = 3234.97 cm−1 and ν R (JK + 1) = 3234.92 cm−1 .)

10.13 If we wish to describe a diatomic molecule using the model of an anharmonic
oscillator, the Morse potential is often used:
2
V(R) = De 1 − e−a(R−Re ) (1)
(R: internuclear distance; Re : equilibrium internuclear distance or equilibrium bond

length). It reproduces an approximately parabolic potential curve in the neighbour-
hood of the equilibrium internuclear distance, and takes the effect of nuclear repulsion
for small R and the transition to a continuum for large R into account.
The prefactor De determines the gap between the equilibrium potential and the
edge of the continuum and can be determined from the anharmonicity constant
ωe
xe = , (2)
4De
yielding
hcν2e
De = . (3)
4xe νe
With the given values for νe and xe νe , one obtains
De = 15 505 cm−1 .
The parameter a is related to De and the reduced mass m r of the molecule via

mr mr
a = ωe = 2πcνe . (4)
2De 2De
Using the atomic masses MH = 1.008 g/mole and MRb = 85.47 g/mole, one
finds m r = 1.654 · 10−27 kg and thus
a = 9.144 · 107 cm−1 .

82 Solutions
The rotational constant B a the equilibrium distance Re is found from

h
B= (5)
8π 2 cm r Re2
with the given values to be
B = 3.02 cm−1
and the rotational constant B(R), which depends on the internuclear distance, is
Re2
B(R) = B. (6)
R2
As a result, the rotational energy (in cm−1 ) is given by
1
E rot (R) = B(R) J(J + 1) = B Re2 J(J + 1) · . (7)
R2
The effective potential of the rotating anharmonic oscillator RbH is finally ob-
tained as the sum of a pure Morse potential and the rotational energy:
Veff (R) = V(R) + E rot (R) . (8)
As the rotational quantum number J increases, the bond is weakened, until it fi-
nally becomes unstable. This phenomenon is based upon the fact that with decreasing
internuclear distance, the rotational energy increases, while for R > Re , the potential
(Morse) energy decreases. Beyond a certain value of J, the effective potential based
on the sum of these two energies no longer has a minimum; cf. Fig. S.21.
Fig. S.21. A graph of the (effective) potential of the anharmonic oscillator RbH vs. internuclear
distance - - - J = 100 ..... J = 40 —- J = 0
Chapter 10 83
10.14 a) The positions of the vibrational levels of the anharmonic oscillator 1 H 35 Cl

are described by

1 1 2
G v = νe v+ − xe νe v+ . (1)
2 2
From this we find the transition energies (selection rule ∆v = 1) to be
νv→v+1 = G v+1 − G v = νe [1 − 2xe (v + 1)] . (2)
A numerical fit of the function G(v) to the given energy values yields as fit
parameters the vibrational constant νe and the anharmonicity constant xe , Figs. S.22
and S.23.
The function G v computed in this way shows clearly that with increasing vi-
brational quantum number v, the spacing between neighbouring levels becomes
smaller.
Fig. S.22. A graph of the energy values G v : measured points and fit curve (fit parameters:
vibrational constant ν e = 2989.74 cm−1 ; anharmonicity constant x e = 0.0174)
Fig. S.23. Extrapolation of the fitted function G v out to its maximum

84 Solutions
The values for νe and xe can just as well be obtained from a plot of the data for
νv→v+1 with a straight-line fit. A purely algebraic solution of the over-determined
system of equations is also readily possible.
The anharmonicity constant xe is related to the distance De between the energy
minimum of the Morse potential and the edge of the continuum as well as the
vibrational constant νe through
ωe hcνe
xe = = . (3)
4De 4De
From this, the dissociation energy D0 can be computed by subtracting the zero-
point energy of the oscillator, G 0 from De :
hcνe
D0 = De − hc G 0 = − hc G 0 ; (4)
4xe
with the values obtained above, we find
D0 = 8.239 · 10−19 J = 5.14 eV = 41 474 cm−1 .
b) Since the Morse potential curves of isotopic molecules are to a first approxima-
tion identical, the values of the potential minimum De , the equilibrium internuclear
distance Re , and the force constant k associated with the potential parameter a are
the same for the bond in the deuterated molecule 2 D 35 Cl as those for 1 H 35 Cl. The re-
duced masses of the two isotopic molecules are however different (in the following,
HCl will be denoted by the index 1, and DCl by the index 2):
m r,1 H 35 Cl = 1.6266 · 10−27 kg ≡ m 1

m r,2 D 35 Cl = 3.1623 · 10−27 kg ≡ m 2 .
Due to the identical force constants, one finds
k = m 1 ω2e,1 = m 2 ω2e,2 . (5)
From this, we obtain the ratio of the vibrational constants:

νe,2 ωe,2 m1
= = ≡γ, (6)
νe,1 ωe,1 m2
and with
γ = 0.7172
the vibrational constant νe,2 for DCl:
νe,2 = 2144.23 cm−1 .

Chapter 10 85
The heavy isotope thus vibrates more slowly than the light one. The anharmonic-
ity constant is obtained from (3) with De,2 = De,1 , so that
xe,2 ωe,2
= =γ (7)
xe,1 ωe,1
and thus for DCl
xe,2 = 0.0125 .
The heavy isotope remains closer to its equilibrium position during the vibration,
it does not move as far onto the flat portion of the Morse potential as the light isotope;
so it thus experiences less anharmonicity.
The rotational constants B are inversely proportional to the moments of inertia
and, due to the identical equilibrium internuclear distance, also to the reduced masses
of the isotopic molecules:
B2 m1
= = γ2 . (8)
B1 m2
Using the value of B1 given for HCl, we obtain B2 for DCl:
B2 = 5.45 cm−1 .
The rotational structure of the heavier molecule is thus more finely spaced than
that of the light molecule.
c) From (1), for v = 0 we can derive the zero-point energy G 0,2 of the anharmonic
oscillator DCl:

1 1
G 0,2 = νe,2 1 − xe,2 . (9)
2 2
Making use of the values calculated in part (b), we find
G 0,2 = 1065.41 cm−1
and thus at constant De the dissociation energy D0,2 of DCl:
D0,2 = De − G 0,2 (10)

−19 −1
D0,2 = 8.32 · 10 J = 5.19 eV = 41 891 cm .
The difference of the dissociation energies of the two isotopic molecules can
be used for their laser photochemical separation. An isotopic mixture HCl/DCl is
∧
excited by intense laser light at ca. 41 700 cm−1 = 240 nm; this is possible e. g. with
a frequency-doubled dye laser. Since the energy of the photons lies between the
dissociation energies of the two isotopes, it is sufficient to break the HCl bonds
while the DCl molecule remains stable. With additional separation procedures one
can then sort out the remaining DCl molecules.
86 Solutions
10.15 The SO2 molecule exhibits three normal modes of vibration, which are sim-
ilar to those of the water molecule. Since in both cases the dipole moment of the
molecule undergoes a change, all of the vibrational modes are IR active. In the sym-
metrical vibrational mode ν1 , only the bond lengths change to first approximation,
while the bond angle Θ remains constant. In contrast, in the ν2 mode, mainly the nor-
mal coordinate Θ varies, while the bond lengths change only a little. In polyatomic
molecules, vibrations can change the symmetry of the electronic wavefunction. The
electronic transition dipole moments thus contain coupled electronic/vibronic matrix
elements, and a large value of the vibronic overlap integral leads to intense lines.
In the case of SO2 , the overlap of the “bending oscillation” ν2 with the geometric
variation between the ground and the excited state is greatest (the bond angle Θ
changes and the bond lengths remain constant). The 0−0 transition is thus followed
by an intense vibrational series with line spacing ν2 . Although the other vibrational
modes are also IR active, they are observed with much lower intensities.
Chapter 11
11.1 Angular momentum and the Born-Oppenheimer Approximation

We consider a triatomic molecule:
2
H=− ∆ j + V (R31 , R32 ) ; Rij = Ri − R j . (1)
j
2m j
Let V be invariant with respect to common rotations of the Rij .

Total angular momentum:

L= Li = nuclear angular momenta . (2)
i
Total momentum:

P= Pi = momenta of the nuclei (3)
i
with
Li = Ri × Pi (4)
and
P2
H= i
+ V = H0 + V . (5)
i
2m i
Consider the commutation relation:

1
[H0 , L] = Pi2 , L j . (6)
ij
2m i
Chapter 12 87
For i = j and considering the example of L z , we find (without index)

2
∂ ∂2 ∂2 ∂ ∂
+ + , x − y f(x, y, z) = 0 . (7)
∂x 2 ∂y2 ∂z 2 ∂y ∂x
For i = j, the following results hold:

Pi , L j = Pi L j − L j Pi = 0 due to different nuclei . (8)
[H0 , L] = 0 . (9)
Consider
[V, L] . (10)
Due to rotational invariance, V commutes with the rotation operators D(α, β, γ) =
Dz (α)D y (β)Dx (γ), where
i i i
D(φ) = e− ~ L z α · e− ~ L y β · e− ~ L x γ . (11)
Treating α, β, γ as infinitesimal quantities, we find also that
[V, L] = 0 , (12)
as one can show through a series expansion of (11).
Chapter 12
12.1 Since the H2 molecule has no permanent electric dipole moment, the pure
rotational spectrum cannot be observed as a microwave absorption spectrum. Instead,
one employs Raman spectroscopy.
The condition for Raman activity of this molecule is a change in its polarisability
on rotation. Since the length of each principal axis i of the polarisability ellipsoid
√
is proportional to 1/ αi (αi : polarisability in the direction i), its sections perpen-
dicular to the molecular axis are circular. The section in the plane containing the
molecular axis is, in contrast, for non-spherically-symmetric molecules always a
non-degenerate ellipse. Upon rotation of the molecule about an axis which is not
the molecular symmetry axis, the polarisability in the plane of polarisation of the
exciting light thus changes, so that the pure rotational spectrum can be observed by
applying Raman spectroscopy.
The rotational states of the molecule are given in first approximation by
FJ = B J(J + 1) . (1)
In the Raman spectrum, the selection rule for optical transitions is ∆J = ±2;
the distance of the lines from the band origin is therefore:
ν J = ν J → J+2 = B (4J + 6) (2)
and the spacing of two neighbouring rotational Raman lines is:
∆ν = ν J+1 − ν J = 4B . (3)
88 Solutions
The rotational constant

h h
B= = (4)
8π 2 c Θ 8π 2 c m 2
r Re
can be calculated using the atomic mass m H = 1.673 · 10−27 kg, the reduced mass
m r = 12 m H , and the equilibrium bond length Re = 0.7414 Å; the result is
B = 60.829 cm−1 .
From (3), we find the spacing of neighbouring spectral lines:
∆ν = 243.3 cm−1 .
12.2 Every molecule whose polarisability ellipsoid is non-spherical in the ground

state (i. e. not spherical tops) exhibits a pure rotational Raman spectrum, since its
polarisability in the plane of polarisation of the exciting light changes when it rotates.
Except for the molecules CH4 and SF6 , which each have spherical symmetry, all
the molecules listed exhibit a pure rotational Raman spectrum.
12.3 If the electric dipole moment of a molecule changes on oscillation, then it is
IR active. A mode of oscillation is, on the other hand, Raman active, if the oscillation
produces a periodic change of the polarisability.
Molecules with a centre of inversion exhibit a complementary IR activity and
Raman activity of their oscillation modes: an oscillation is either IR active or Raman
active.
We thus observe the following behaviour:
N2 : only Raman active.
C2 H4 : planar structure with a centre of inversion: Raman and IR
complementarity.
CH3 OH: no centre of inversion, dipole moment and polarisability
change, thus both Raman and IR activity.
HD: only different isotopes, behaves like a homonuclear, di-
atomic molecule: only Raman activity.
CCl4 : tetrahedral structure, no centre of inversion: Raman active;
asymmetric bending and stretch vibrations are also IR
active.
CS2 : linear, centre of inversion: Raman and IR complementar-
ity.
SO2 : bent, no centre of inversion: both Raman and IR.
NH3 : pyramidal: both Raman and IR activity.
BeCl2 : exists in the gas phase at T > 750 ◦ C as a linear molecule
with a centre of inversion: Raman and IR complementar-
ity.
CH3 COCH3 : no centre of inversion: both Raman and IR activity.
(CO)5 Re−Re(CO)5 : centre of inversion: Raman and IR complementarity.
Chapter 12 89
12.4 The ammonia molecule, NH3 , is a symmetric top, whose symmetry axis
passes through the N atom. The two principal moments of inertia perpendicular and
parallel to the axis of symmetry have the values
Θ⊥ = 2.86 · 10−47 kg m2 and Θ = 4.38 · 10−47 kg m2 .
The rotational constants B and C can be calculated from these values:

h h 1 1
B= and C = − , (1)
8π 2 cΘ⊥ 8π 2 c Θ Θ ⊥
leading to B = 9.78 cm−1 and C = −3.39 cm−1 . The positions of the rotational
levels are given by
F(J, K) = B J(J + 1) + C K 2
with K = 0 , ±1 , ±2 , . . . , ±J . (2)
Here, the quantum number J denotes the total angular momentum of the
molecule, while K refers to rotation around the axis of symmetry:
|L | = K . (3)
The rotation around the symmetry axis is not Raman active, since the polaris-
ability of the molecule remains unchanged during this rotation. The first selection
rule for Raman transitions in symmetric-top molecules follows from this:
∆K = 0 . (4)
For a change in the quantum number J of the total angular momentum, in

contrast, we find
∆J = 0 , ±1 , ±2 for K = 0 , (5)
∆J = ±2 for K = 0. (6)
In the special case of the symmetric top, we thus observe that for all rotational
states with J = 0, Raman transitions with ∆J = ±1 are allowed. The Raman
spectrum consists of a superposition of the two subspectra with ∆J = ±1 (P and R
branches) and ∆J = ±2 (O and S branches).
If J and J are the rotational levels of the molecule before and after the Raman
scattering process, with ∆J = J − J , then we find from (2), taking the selection
rules (4–6) into account, the pure rotational Raman shift ∆ν = FJ − FJ of the
scattered light relative to the primary light for the Stokes lines:
S branch:
∆J = +2 ⇒ ∆ν = −B (4J + 6) (J = 0, 1, 2, . . . )
R branch:
∆J = +1 ⇒ ∆ν = −2B (J + 1) (J = 1, 2, 3, . . . )
90 Solutions
Anti-Stokes- lines:
O branch:
∆J = −2 ⇒ ∆ν = B (4J − 2) (J = 2, 3, 4, . . . )
P branch:
∆J = −1 ⇒ ∆ν = 2B J (J = 2, 3, 4, . . . ).
∧
The wavelength of the primary light (N2 -Laser) lies at 336.732 nm = 29 697.2
−1
cm . For the innermost Raman lines, we obtain the following positions:
J = 0 1 2 3
Stokes: S branch 29 638.5 cm−1 29 599.4 cm−1 29 560.3 cm−1 29 521.2 cm−1
337.39 nm 337.84 nm 338.29 nm 338.73 nm
R branch 29 658.1 cm−1 29 638.5 cm−1 29 619.0 cm−1
337.17 nm 337.39 nm 337.62 nm
Anti-Stokes: O branch 29 755.9 cm−1 29 795.0 cm−1
336.06 nm 335.62 nm
P branch 29 736.3 cm−1 29 755.9 cm−1
336.28 nm 336.06 nm
See also Fig. S.24.
12.5 a) IR and Raman lines do not occur together, they are complementary
(cf. problem 12.3). From this we can conclude that the molecule A2 B2 has a centre
of inversion. Since an IR line with PR structure is observed, it must also be a linear
molecule. The general structure formula is thus
A−B−B−A .
The number of normal vibrations is f = 3N − 5 = 7, where however twice

two bending vibrations are degenerate; there remain five non-equivalent vibrations.
Since the two most energetic modes lie in the range of the C−H stretching vibration,
the given frequencies can be associated to th normal vibrations of the molecule
H−C≡C−H
as follows:
1)3374 cm−1 ←−→←−→ Raman active symmetric stretching vibration C−H
2)3287 cm−1 ←−→→←− IR active (PR) asymmetric stretching vibration C−H
3)1973 cm−1 ←←→→ Raman active symmetric stretching vibration C≡C
4)729 cm−1 ↑↓↓↑ IR active (PQR) bending vibration
5)612 cm−1 ↑↓↑↓ Raman active bending vibration
(weak)
b) Since one of the vibrations is both IR and also Raman active, the molecule
AB2 cannot have a centre of inversion. Owing to the complex structure of the infrared
bands, it must be a bent molecule:
BB A
A or B B.
Chapter 12 91
Fig. S.24. Rotational Raman spectrum of the symmetric-top molecule NH3 : superposition
of the two series with ∆J = ±1 (above: R and P branch) and ∆J = ±2 (centre: S and
O branch). below: superposed spectrum; every second line of the R and P branches falls
together with a line of the S and O branches. This causes the alternating line intensities
Both structures lead to three vibrational modes. From data on the polarisation
of Raman lines, one obtains information about whether the vibrational mode is
symmetric or asymmetric: in the case of a symmetric molecule, a symmetric vibration
leads to an isotropic deformation of the polarisabiity ellipsoid, so that polarised
primary light remains polarised after Raman scattering. Asymmetrical vibrations
produce an anisotropic deformation of the polarisability ellipsoid with the result that
the molecules are arbitrarily oriented and lead to a depolarisation of the scattered
light.
The two most energetic vibrations lie in the range of the O−H stretching vibra-
tion, so that the observed modes can be attributed to the molecule
92 Solutions
H2 O : (1)
←−
1) 3756 cm−1 − 1 IR active asymmetric stretching
vibration
2) 3652 cm−1 ↑
IR and Raman symmetric stretching
active (polarised) vibration
3) 1595 cm−1 ↑
IR active bending vibration
The frequencies of the bending vibrations are in general lower than those of the
stretching vibrations; also even if they are in principle Raman active, the intensity
of the lines is often so weak that they are not observed.
12.6 Since both molecules have vibrational modes which are both IR and also
Raman active (i. e. the IR and Raman active modes are not complementary), neither
N2 O nor NO2 can have a centre of inversion symmetry. Due to the PR structure of
its IR bands, N2 O must be linear. The only remaining possible structure is thus
N−N−O .
In contrast, the complex rotational structure of the IR bands of NO2 indicate that
this molecule is not linear. On the basis of the available information, we can give
two possible molecular structures:
OO N
N or O O.
12.7 In diatomic homonuclear molecules, the nuclear spin I of the two identical
nuclei A and B (I = I A = I B ) has an influence on the molecular statistics, i. e. on
the probability that a particular molecular state will occur. Under the assumption of
decoupled partial motions, the total wavefunction of the molecule can be written in
terms of a product state:
Ψ(r A , σ A , r B , σ B , ri , σi ) (1)
−1
=Ψ KS
(σ A , σ B ) Ψ ES
(σi )Ψ (R , ri ) R
el
Ψ vib
(R) Ψ (θ, φ)
rot
=Ψ KS
(σ A , σ B ) Ψ (r A , r B , ri , σi )
R
(2)
with r A , r B , ri : coordinates of the nuclei A and B and of the electrons i in the

centre-of-mass system;
R: internuclear distance;
Ψ KS : nuclear spin state;
Ψ ES : electronic spin state;
Ψ el : solution of the Schrödinger equation for the two-centre problem with
fixed nuclei (Born-Oppenheimer approximation);
Ψ vib : state of the harmonic oscillator;
Ψ rot : state of the rigid rotor.
Chapter 12 93
Upon exchange of the identical nuclei A and B, Ψ KS is multiplied by a factor

(−1)κ , and the “remainder wavefunction” Ψ R by a factor (−1) . The parity of the
total wavefunction, i. e. its behaviour under exchange of the nuclei, A ↔ B, can be
mathematically described by
Ψ(r B , σ B , r A , σ A , ri , σi ) = (−1)κ+ Ψ(r A , σ A , r B , σ B , ri , σi ) . (3)
Since the parity of the wavefunction is independent of internal interactions

(couplings), it can be determined from the product state (1) [if coupling is present:
series expansion of the wavefunction in terms of product states with the same
(−1)κ+ ].
The behaviour of Ψ with respect to inversion of the spatial vectors of all the
particles is described by the parity (−1)P (upper index of the term symbol: + for
P = 0 or − for P = 1). Because spins are axial vectors, they are unaffected by
spatial inversion, so that the spin wavefunctions Ψ KS and Ψ ES have positive parity.
While the vibrational wavefunction R−1 Ψ vib depends only on the magnitude R of the
internuclear distance and therefore remains unchanged upon inversion, the rotational
function Ψ rot in the state J is multiplied by (−1) J . Finally, one imagines the inversion
for the electronic eigenfunction Ψ el to consist of two symmetry operations which are
carried out one after the other: the first operation, a rotation of the molecule by 180◦
about an axis perpendicular to the molecular symmetry axis which passes through
the centre of gravity, leaves Ψ el unchanged; the following reflection through a plane
perpendicular to the axis of rotation and containing the nuclei changes the relative
coordinates and causes a modification of Ψ el by a factor of (−1)s .
The parity of the overall state under spatial inversion is thus
(−1) P = (−1) J+s . (4)
Since the indices + and − of the term symbol refer to the rotational ground state
J = 0, we have:
(−1)s = +1 ⇐⇒ s = 0 for + , (5)

(−1)s = −1 ⇐⇒ s = 1 for − . (6)
In order to understand the behaviour of the “remainder wavefunction” Ψ R

[without nuclear spins, cf. (1),(2)] under exchange of the nuclear positions, we
likewise decompose this operation into two sequential symmetry operations: first,
Ψ R is multiplied by (−1) P on inversion of all the particles. Under the following
inversion of the electrons alone, an additional factor of (−1)π occurs; it is denoted in
the term symbol by the lower index g (for π = 0) or u (for π = 1). The two individual
operations together amount to an exchange of the nuclei, while the electrons remain
unaffected. The parity of Ψ R under nuclear exchange is thus, taking account of (4),
given by
(−1) = (−1) P+π = (−1) J+s+π . (7)

94 Solutions
Finally, we need to investigate the behaviour of the nuclear spin wavefunction

Ψ KS under exchange of the identical nuclei A and B with nuclear spins I A = I B = I
and magnetic nuclear spin quantum numbers M I = −I , . . . , I. The total nuclear
spin of the molecule is found by vector addition of the two individual spins and can
take on the values
Iges = 2I , 2I − 1 , . . . , 0 . (8)
From a quantum-mechanical analysis (cf. problem 12.8), we obtain the symmetric
and antisymmetric states with positive or negative parity, (−1)κ :
for integral I:
Iges = 2I , 2I − 2 , . . . , 0 symmetric
Iges = 2I − 1 , 2I − 3 , . . . , 1 antisymmetric
for half-integral I:
Iges = 2I , 2I − 2 , . . . , 1 symmetric
Iges = 2I − 1 , 2I − 3 , . . . , 0 antisymmetric.
Each state Iges is (2Iges +1)-fold degenerate; from this fact, one derives the ratio
of the statistical weights of the antisymmetric to the symmetric nuclear spin states
(cf. problem 12.10):
ga I
= . (9)
gs I +1
a) nuclear spin I = 0
Because nuclei with integer spins are Bosons, the total wavefunction Ψ must be
symmetric under exchange of the nuclei, i. e. with (3) and (7):
!
(−1)κ+ = (−1)κ+J+s+π = +1 . (10)
Only the symmetric nuclear spin state Iges = 0 exists, so that
(−1)κ = +1 . (11)
From (10) and (11), it follows that:
!
(−1) = (−1) J+s+π = +1 . (12)
Now with (5) and (6) we have the result
– for even electronic terms (index g):
!
π = 0 ⇒ (−1) J+s = +1
!
⇒ for terms with positive parity (index +): s = 0 ⇒ (−1) J = +1
only rotational states with even J exist.
example: 1 Σg+ , 1 Πg+
⇒ for terms with negative parity (index −):
!
s = 1 ⇒ (−1) J = −1
only rotational states with odd J exist.
example: 1 Σg− , 3 Σg− , 1 Πg−
Chapter 12 95
– for odd electronic terms (index u):

!
π = 1 ⇒ (−1) J+s = −1
⇒ for terms with positive parity (index +):
!
s = 0 ⇒ (−1) J = −1
only rotational states with odd J exist.
example: 1 Σu+ , 1 Πu+
⇒ for terms with negative parity (index −):
!
s = 1 ⇒ (−1) J = +1
only rotational states with even J exist.
example: 1 Σu− , 1 Πu− .
For terms with Λ > 0 (Π terms), there are in each case two states with + and −
parity and the same J which are degenerate; thus, all the J values occur.
b) nuclear spin I = 1
The nuclei are again Bosons, so that here also, the total wavefunction must be
symmetric:
!
(−1)κ+ = (−1)κ+J+s+π = +1 . (13)
We now must distinguish between symmetric and antisymmetric nuclear spin

states:
!
– Iges = 2, 0: Ψ KS symmetric, (−1)κ = +1 ⇒ (−1) = +1 [cf. (12)].
degree of degeneracy: 5 + 1 = 6.
!
– Iges = 1: Ψ KS antisymmetric, (−1)κ = −1 ⇒ (−1) = −1
degree of degeneracy: 3.
The ratio of the statistical weights of the antisymmetric to the symmetric nuclear
spin states is, as can be expected from (9), given by
ga 1
= . (14)
gs 2
For the individual electronic terms, the result is just as in (a); however, the
population ratios between rotational states with even and odd J are not 1:0 or 0:1
(allowed/forbidden states), but instead 2:1 or 1:2. The states which can exclusively
exist in (a) are here doubly populated in comparison to the states which are missing
in (a).
12.8 The space of angular momenta ε( j1 , j2 ) is spanned by two angular momenta
j1 and j2 . The states |J, M are eigenvectors of the operators J 2 and Jz :
J 2 |J, M = J(J + 1) 2 |J, M (1)

Jz |J, M = M |J, M
2
(2)
96 Solutions
and in addition, like all the vectors in the space ε( j1 , j2 ), they are eigenvectors of J12
and J22 with the eigenvalues j1 ( j1 + 1) 2 and j2 ( j2 + 1) 2 .
The angular-momentum space ε( j1 , j2 ) can be represented as the direct sum of
invariant subspaces ε(J):
ε( j1 , j2 ) = ε( j1 + j2 ) ⊕ ε( j1 + j2 − 1) ⊕ . . . ⊕ ε(| j1 − j2 |) (3)
The vectors |J, M, which span the subspaces, will be determined in general in
the following.
The subspace ε(J = j1 + j2 )
The ket vector | j1 , j2 ; m 1 = j1 , m 2 = j2 is the only eigenvector belonging to
the eigenvalue M = j1 + j2 and, since J 2 and Jz commute, it is also an eigenvector
of J 2 with the eigenvalue J = j1 + j2 . With a suitable choice of the phase factor,
we then have
| J = j1 + j2 , M = j1 + j2 = | j1 , j2 ; m 1 = j1 , m 2 = j2 . (4)
In order to find the remaining vectors | J, M of the subspace ε( j1 , j2 ), we make

use of the “ladder operator” J− with
%
J− | J, M = J(J + 1) − M(M − 1) | J, M − 1 . (5)
Application to the vector (4) yields

%
J− | j1 + j2 , j1 + j2 = 2( j1 + j2 ) | j1 + j2 , j1 + j2 − 1 . (6)
Due to J− = J1 − + J2 − , it follows with (4) that
| j1 + j2 , j1 + j2 − 1
1
= √ J− | j1 + j2 , j1 + j2 (7)
2( j1 + j2 )
1
= √ (J1 − + J2 − ) | j1 , j2 ; j1 , j2 (8)
2( j1 + j2 )
1 2 %
= √ 2 j1 | j1 , j2 ; j1 − 1 , j2
2( j1 + j2 )
% 3
+ 2 j2 | j1 , j2 ; j1 , j2 − 1 (9)
and thus

j1
| j1 + j2 , j1 + j2 − 1 = | j1 , j2 ; j1 − 1 , j2
j1 + j2

j2
+ | j1 , j2 ; j1 , j2 − 1 . (10)
j1 + j2
Chapter 12 97
This vector, which is already normalised, is a linear combination of the two basis
vectors of ε( j1 , j2 ) which belong to M = j1 + j2 − 1. By repeated application of
the operator J− , we obtain the vectors of the subspace up to
|J = j1 + j2 , M = −( j1 + j2 ) = | j1 , j2 ; − j1 , − j2 . (11)
In this way, all the 2( j1 + j2 ) + 1 basis vectors |J, M of the subspace ε( j1 + j2 )

can be found.
The remaining subspaces ε(J)
In the subspace ε( j1 + j2 −1), M = j1 + j2 −1 is the largest eigenvalue of Jz , and
the associated eigenvector is |J = j1 + j2 − 1 , M = j1 + j2 − 1. It can be expanded
in terms of the basis vectors of ε( j1 , j2 ) with the eigenvalue M = j1 + j2 − 1:
| j1 + j2 − 1 , j1 + j2 − 1 = α | j1 , j2 ; j1 , j2 − 1
+ β | j1 , j2 ; j1 − 1 , j2 . (12)
The normalisation of this vector is determined by the condition
|α| 2 + |β| 2 = 1 . (13)
From its orthogonality to the vector | j1 + j2 , j1 + j2 − 1, which belongs to the

subspace ε( j1 + j2 ), it also follows with (10) that

j2 j1
α +β = 0. (14)
j1 + j2 j1 + j2
With a suitable choice of the phase factors, we obtain from (13) and (14) the
relations

j1
| j1 + j2 − 1 , j1 + j2 − 1 = | j1 , j2 ; j1 , j2 − 1
j1 + j2

j2
− | j1 , j2 ; j1 − 1 , j2 . (15)
j1 + j2
By repeated application of the operator J− , one finds the 2( j1 + j2 − 1) + 1 basis

vectors which span the subspace ε( j1 + j2 − 1).
Applying a corresponding procedure, one can finally generate all the basis vectors
of all the subspaces. The eigenstates | j1 , j2 ; m 1 , m 2 of the operators J 21 , J 22 , J1z
and J2z are transformed thereby into eigenstates |J, M of J 21 , J 22 , J 2 and Jz . The
coefficients which occur in the series expansion of the new basis vectors in terms of
the old ones are called “Clebsch-Gordan coefficients”.
98 Solutions
a) The above formalism, developed for the general case, will now be applied
to the special case of two nuclear spins j1 = I A = 1 and j2 = I B = 1. The total
nuclear spin I = I A + I B of the molecule can take on the values
I = 2, 1, 0. (16)
The angular-momentum space ε(1, 1) is thus decomposed into the subspaces ε(I =
2), ε(I = 1), and ε(I = 0) with dimensions 5, 3 and 1.
The subspace ε(I = 2)
According to (4), the vector |J = 2 , M = 2 can be written as
|2, 2 = |1, 1; 1, 1 . (17)
Application of the ladder operator J− leads –analogously to (10)– to
1
|2, 1 = √ [ |1, 1; 1, 0 + |1, 1; 0, 1 ] . (18)
2
Furthermore, one finds
1
|2, 0 = √ [ |1, 1; 1, −1+ 2 |1, 1; 0, 0+ |1, 1; −1, 1] (19)
6
1
|2, −1 = √ [ |1, 1; 0, −1 + |1, 1; −1, 0 ] (20)
2
|2, −2 = |1, 1; −1, −1 (21)
The general result (15) now corresponds to
1
|1, 1 = √ [ |1, 1; 1, 0 − |1, 1; 0, 1 ] (22)
2
and therefore
1
|1, 0 = √ [ |1, 1; −1, 1 − |1, 1; 1, −1 ] , (23)
2
1
|1, −1 = √ [ |1, 1; 0, −1 − |1, 1; −1, 0 ] . (24)
2
It is notable that the vector |0, 1 contains no contribution from the components
|1, 1; 0, 0. The corresponding Clebsch-Gordan coefficient is 0.
The choice of suitable real phase factors leads to the basis vector of the one-
dimensional subspace:
1
|0, 0 = √ [ |1, 1; 1, −1 − |1, 1; 0, 0 + |1, 1; −1, 1 ] . (25)
3
b) One can derive from the results of (a) the fact that the states belonging to I = 2
and I = 0, |J, M, are symmetric on nuclear exchange, while the states belonging
to I = 1 change their sign on nuclear exchange, and are thus antisymmetric.
Chapter 12 99
12.9 The (unstable) isotope 19 F is a Fermion due to its nuclear spin of 1/2; the total
wavefunction of the molecule 19 F2 is thus antisymmetric under nuclear exchange.
With the notation from problem 12.7, this yields
!
(−1)κ+J+s+π = −1 . (1)
The electronic term for the ground state is even (π = 0) and has positive parity
(s = 0). From this, one can conclude that
!
(−1)κ+J = −1 . (2)
The maximum total spin I which can result from vector addition of the two
nuclear spins I A = I B = 12 is I = 1 (ortho fluorine). The three degenerate states
with M = 0 , ±1 are symmetric (κ = 0), so that due to (2), only rotational states of
odd quantum number J can occur.
In contrast, the state belonging to a total spin of I = 0 (para fluorine) is anti-
symmetric (κ = 1), and its rotational states have even values of J.
The statistical weights of the two types of molecule behave as
gPara ga 1
= = . (3)
gOrtho gs 3
12.10 In order to derive the ratio of the statistical weights of the antisymmetric
and the symmetric states of a diatomic homonuclear molecule in general, we need
to investigate the following cases for the nuclear spins I A = I B = I and the total
spin Iges :
• integer I:
– symmetric states:
Iges = 2I , 2I − 2 , . . . , 0 (I + 1 states)
degree of degeneracy in each case 2Iges + 1
overall abundance:
gs = (2 · 2I + 1) + [2 · (2I − 2) + 1] + . . . + 1 (1)

I
=4 n + I +1 (2)
n=0
= (I + 1) (2I + 1) ; (3)
– antisymmetric states:
Iges = 2I − 1 , 2I − 3 , . . . , 1 (I states)
overall abundance:
100 Solutions
ga = [2 · (2I − 1) + 1] + [2 · (2I − 3) + 1]
+ . . . + (2 · 1 + 1) (4)

I−1
=4 n + 3I (5)
n=0
= I (2I + 1) . (6)
For integer values of I, one thus obtains the ratio

ga I (2I + 1) I
= = .
gs (I + 1)(2I + 1) I +1
• half-integer I:
– symmetric states:
Iges = 2I , 2I − 2 , . . . , 1 (I + 12 states)
overall abundance:
gs = (2 · 2I + 1) + [2 · (2I − 2) + 1]
+ . . . + (2 · 1 + 1) (7)

I
5
=4 n +I+ (8)
2
n= 12
= (I + 1) (2I + 1) ; (9)
– antisymmetric states:
Iges = 2I − 1 , 2I − 3 , . . . , 0 (I + 12 states)
overall abundance:
ga = [2 · (2I − 1) + 1] + [2 · (2I − 3) + 1] + . . . + 1 (10)

1
I− 2
1
=4 n +I+ (11)
n=0
2
= I (2I + 1) . (12)
For half-integral I also, the ratio is given by

ga I (2I + 1) I
= = . (13)
gs (I + 1)(2I + 1) I +1
Chapter 13 101
Chapter 13
13.1 The information contained in the term symbols of the form 2S+1
Λ±
(g,u) Ω is
collected in the table below. There, we find for a diatomic molecule
– Λ, the z-component of the total angular momentum of all the electrons in the
molecule (the z direction is the direction of the molecular symmetry axis);
– S, the total spin of the shell electrons (2S + 1: multiplicity; z-component: Σ);
– Ω = | Λ + Σ |, the total angular momentum of the electronic shell in the z
direction;
– σ, the symmetry of the electronic wavefunction on reflection in a plane containing
the nuclei;
– P, the parity of the electronic wavefunction of a homonuclear molecule (with
identical nuclear charges and therefore an inversion-symmetric nuclear field),
i. e. the symmetry on inversion of all the particles (gerade/ungerade ≡ even/odd)
(see the table below).
13.2 From the data given, one can extract the following information about the
states of the separated atoms A and B:
A: 1s ⇒ L A = 0 ; m L,A = 0 ;
S A = 12 ; m S,A = ± 21 ;
B: 1s2 2s2 2 p; 1s2 and 2s2 are closed shells with S = L = 0
⇒ L B = 1 ; m L,B = 0, ±1 ;
S B = 12 ; m S,A = ± 12 .
Since AB is a heteronuclear molecule, the possible exclusion of terms due to the
Pauli principle need not be considered. For the molecule AB we thus find:
AB:m L = 0, ±1 ⇒ Λ = 0, 1
⇒ one Σ and one Π term ; S = 0, 1 ⇒ Σ = 0, ±1 .
From this we find the possible terms
Λ = 0: 1
Σ0+ , 1 Σ0− , 3 Σ0+ , 3 Σ0− , 3 Σ1+ , 3 Σ1−
Λ = 1: 1
Π1 , 3 Π0 , 3 Π1 , 3 Π2 .
Λ S Ω Σ σ P Remark
1
Σ+ 0 0 0 0 + 1
Σ + and 1 Σ − not degenerate!
1 +
Σu 0 0 0 0 + u identical nuclear charges; molecular or-
bitals ug
3 +
3
Φ3 3 1 3 0 Φ3 and 3 Φ3− degenerate
3
Φ3− 3 1 3 0 −
3 −
Φg3 3 1 3 0 − g identical nuclear charges; molecular or-
bitals gg or uu
102 Solutions
13.3 a) The electron configuration of a diatomic homonuclear molecule can be

found by successively filling the orbitals of the known term scheme (cf. Fig. 13.4),
taking into account the Pauli principle and Hund’s rules. For the oxygen molecules
under discussion here, one then obtains
O+ 2 ∗ 2 2 ∗ 2 4 2 ∗
2 : (σg 1s) (σu 1s) (σg 2s) (σu 2s) (πu 2 p) (σg 2 p) (πg 2 p)
O2 : (σg 1s)2 (σu∗ 1s)2 (σg 2s)2 (σu∗ 2s)2 (πu 2 p)4 (σg 2 p)2 (πg∗ 2 p)2
O− ∗ 2 2 ∗ 2 4 2 ∗
2 3
∗ ∗ ∗
O2− 2 2 2 2 4 2 4
b) From the electron configurations in (a), the term symbols 2S+1 Λ± (g,u) Ω of
the ground states can be derived. The problem can be simplified by remembering
that closed electronic shells have neither orbital angular momentum nor spin. For
computing the total orbital angular momentum Λ = L z and spin S of the electronic
shells therefore only the outermost unfilled shell needs to be considered.
The homonuclear diatomic molecules have a nuclear field with inversion sym-
metry, so that the electron terms exhibit an even/odd symmetry; with an even number
of electrons in u orbitals, one has a g term, with an odd number, a u term.
O+
2 : Λ = 1 ; S = 12 ⇒ 2 Πg
O2 : Λ = 0 ; S = 0 ⇒ 1Σg+
Λ = 0 ; S = 1 ⇒ 3 Σg−
Λ = 2 ; S = 0 ⇒ 1 ∆g
Λ = 2 ; S = 2 Pauli forbidden
according to Hund’s rule, the state of highest multiplicity lies energetically
lowest; the ground state is thus 3 Σg−
−
O2 : Λ = 1 ; S = 12 ⇒ 2 Πg (π 3 behaves like π 1 )
O2−
2 : all shells are closed, therefore no orbital angular momentum or spin; com-
pletely symmetric state ⇒ 1 Σg+
c) The energetically lowest allowed transition of O2 changes the electronic
configuration of the molecule according to
(σg 1s)2 (σu∗ 1s)2 (σg 2s)2 (σu∗ 2s)2 (πu 2 p)4 (σg 2 p)2 (πg∗ 2 p)2
−→ . . . (πu 2 p)3 (σg 2 p)2 (πg∗ 2 p)3
and its terms according to

3
Σg− −→ 3 Σu− .
Here, we have taken the selection rules for radiative transitions, namely ∆Λ =
0 , ±1, ∆S = 0 and in particular also g ↔ u into account. We are thus dealing with
a π → π ∗ transition.
Chapter 14 103
13.4 From the term symbol 3 Πu , one can conclude that we are dealing with S + 1
(triplet state) and Λ = 1. In order to yield an H atom in its ground state upon
dissociation, an electron must occupy the molecular orbital (1sσg ); due to Λ = 1,
the other orbital must then be a π state. Since the molecular state is stable in
comparison to the dissociation products, we can assume that both electrons occupy
bonding orbitals. The odd electron term requires finally one g and one u state for the
participating electrons.
The energetically lowest electronic configuration which fulfills the conditions
listed above is then
(1sσg ) (2 pπu ) .
Chapter 14
14.1 a) For the position G v of the vibronic levels above the minimum of the
potential curve of the given electronic state, the following approximate formula
holds:

1 1 2
G v = νe v + − xe νe v + . (1)
2 2
The spacing between two neighbouring vibrational levels v and v + 1 is then
∆νv = G v+1 − G v = νe [1 − 2xe (v + 1)] . (2)
The point of convergence of the limiting continuum is at the vibrational level vC

at which the difference to the level vC + 1 becomes negative or zero:
!
∆νvC = 0 . (3)
With (2), this yields the relation between the position of the point of convergence
vC and the anharmonicity constant xe :
1
vC = − 1. (4)
2xe
From this relation, we can compute the points of convergence of the limiting
continuum and thus the number of vibrational levels up to the particular dissociation
limit, namely vC for the ground state and vC for the excited state, with the result
vC = 70 and vC = 54 .
The dissociation energy D0 of an electronic state is the difference between the

energy G vC at the point of convergence and the zero-point energy G 0 :
D0 = G vC − G 0 . (5)
One calculates D0 for the ground state and D0 for the excited state:
104 Solutions
kJ
D0 = 56 896 cm−1 = 1.13 · 10−18 J = 680.6 = 7.1 eV,
mole
kJ
D0 = 47 755 cm−1 = 9.49 · 10−19 J = 571.3 = 5.9 eV .
mole
b) The dissociation energies of the ground state and of the electronically excited
state can also be found from spectroscopic data. To this end, we require knowledge
of
– the energy hc ν00 of the 0−0 transition line (from the v = 0 level of the ground
state into the same level of the excited state),
– the energy νC of the point of convergence (distance of the dissociation limit of
the excited state from the v = 0 level of the ground state; the absorption begins
at the latter)
– and the sum of the excitation energies E At of the atomic dissociation products
of the electronically excited state (the difference between the continuum edges
of the ground state and the excited state).
We then find for the dissociation energy D0 of the ground state:
D0 = hc ν C − E At
⇒ D0 = 28 923 cm−1 = 5.75 · 10−19 J
kJ
= 346.0 = 3.6 eV
mole
and for the dissociation energy D0 of the excited state
D0 = hc νC − hc ν00
⇒ D0 = 19 853 cm−1 = 3.94 · 10−19 J
kJ
= 237.5 = 2.5 eV .
mole
Fig. S.25. Visualisation of the calculation of the dissociation energies (explained in the text)
Chapter 14 105
The discrepancy between the results in (a) and in (b) is due to the linear extrap-
olation according to (2) which is employed in (a): the linear relation between ∆νv
and v is extrapolated to vC with ∆νvC = 0. This approximation however becomes
worse and worse as the vibrational quantum number v increases, since for large
values of v in the calculation of the level energies from (1), terms of third and fourth
order in v become significant. This results in a value of vC calculated from (3) which
is larger than the true value and thus an overestimate of the number of vibrational
levels below the point of convergence.
14.2 The energy spacings ∆νv between neighbouring vibrational levels v and v +1
are, according to Eq. (2) in problem 14.1, given by
∆νv = G v+1 − G v = νe [1 − 2xe (v + 1)] . (1)
In the plot of ∆E = ˆ ∆νv against v, one however finds a strong deviation from
the linear behaviour for large values of the quantum number v, so that the linear
extrapolation from small v to the point of convergence vC with ∆νvC = 0, the so
called Birge-Sponer extrapolation, no longer seems reasonable. The function ∆νv (v)
decreases faster than the extrapolation line; cf. Fig. S.26. The reason for this is the
increasing contribution of the terms of third and fourth order in the equation for the
energetic positions of the vibronic levels for increasing v:

1 1 2 1 3
G v = νe v + − xe νe v + + ye νe v +
2 2 2
4
1
− z e νe v + + ... . (2)
2
Extrapolation of ∆νv (v) using a power series of third or fourth order thus leads
to a better result for vC than the linear Birge-Sponer extrapolation.
The dissociation energy D is obtained as the area under the function ∆νv (v),
i. e. the integral
Fig. S.26. A Birge-Sponer plot for determining the point of convergence of the limiting
continuum of I2 (extrapolation to ∆E = 0: vC = 97)
106 Solutions
vC
D= ∆νv (v) dv . (3)
0
Applying this procedure with the given data to the molecule I2 leads to the point
of convergence vC = 97 and the dissociation energy
kJ
D = 12 293 cm−1 = 2.44 · 10−19 J = 147.1 = 1.5 eV .
mole
14.3 a) The given data for the energetic positions of the vibronic levels of the
electronically-excited state of O2 can be evaluated in terms of a Birge-Sponer plot,
cf. Fig. S.27. From the nonlinear extrapolation of the function ∆νv (v) out to the point
of convergence of the limiting continuum vC with ∆νvC = 0, one obtains vC = 18
and, from the area under the curve, the dissociation energy D0 of the excited state
kJ
D0 = 6897 cm−1 = 1.37 · 10−19 J = 82.5 = 0.86 eV .
mole
Fig. S.27. A Birge-Sponer plot for determining the point of convergence of the limiting
continuum of O2 (Extrapolation to ∆E = 0: vC = 18)
b) From the energy ν00 of the 0−0 transition, the dissociation energy D0 of the
excited state, and the total excitation energy E At of the atomic dissociation products,
we can calculate the dissociation energy D0 of the ground state:
D0 = hc ν 00 + D0 − E At
= (50 062.6 + 6897 − 15875) cm−1
⇒ D0 = 41 085 cm−1 = 8.16 · 10−19 J
kJ
= 491.5 = 5.1 eV .
mole
Chapter 14 107
14.4 a) The data given for the oscillation frequency νe of the CO molecule in its
ground stand and the band intervals νe of the vibrational states in the electronically
excited state permit us –within the harmonic-oscillator model– to calculate the force
constants k and k of the molecular bonding in the ground state and in the excited
state:
k = m r ω2e = m r 4π 2 c2 ν2e (1)
and, with the reduced mass m r = 1.14 · 10−26 kg

kg
k = 1855.5 ,
s2
kg
k = 885.0 2 .
s
Relative to the ground state, the bonding in the excited state is thus weaker.
The relation between the rotational constants B in the ground state and B in
the excited state as well as the equilibrium bond lengths Re and Re in the rigid-rotor
approximation is given by
h h
B= = . (2)
8π 2 c Θ 8π 2 c m r Re2
Due to B < B , we therefore find Re > Re , i. e. the average internuclear
distance is larger in the excited state than in the ground state. Calculation with the
given data yields
Re = 1.13 Å ,
Re = 1.24 Å .
For radiative electronic transitions between different rotational levels, the selec-
tion rule ∆J = 0 , ±1 applies. In a diatomic molecule, the angular momentum of
the electronic shells Ω and the angular momentum N of the rotational motion of the
whole molecule are coupled to the total angular momentum J:
J = N+Ω. (3)
If we are dealing with a Π electron term, then Ω = 1, and the state J = 0

does not exist. The origins of the branches of the rotational bands thus permit us to
draw conclusions about the terms which contribute to the electronic transition. In
the present case, the band origins are P(2), Q(1), and R(0), i. e.:
P branch (J = J − 1) : first line 2 → 1,

Q branch (J = J ) : first line 1 → 1 ,

R branch(J = J + 1) : first line 0 → 1 .
108 Solutions
This means that the rotational level J = 0 is lacking in the excited state, while
it exists in the excited state. We can therefore assume that we are dealing with a
transition of the type 1 Σ −→ 1 Π.
b) With the information obtained in part (a), a simplified term scheme can
be sketched, if one retains the harmonic-oscillator approximation and thus draws
parabolic potential curves U(R):
k
U(r) = (R − Re )2 . (4)
2
This term scheme contains the ground state and the excited electronic state
including its vibronic levels. According to the Frank-Condon principle, only vertical
transitions are allowed, since in the adiabatic approximation the internuclear distance
remains constant during the transition.
The absorption begins in the J = 0 level of the ground state for which the prob-
ability density at the equilibrium internuclear distance Re it greatest. The transition
ends at the “turning point” of the J = 1 level of the excited state. Since the greatest
probability density of this state lies in the neighbourhood of the turning point, the
overlap integral of the wavefunctions in the ground state and the excited state is
maximal for the 0–1 transition. For this transition, we can thus expect the highest
intensity.
The equilibrium internuclear distance Re in the excited state is further away from
Re than the turning point of the vibronic levels J = 2, 3, . . . . The intensity of the

0–0 transition should therefore be less than that of the first three or four lines of the
vibrational series.
c) The molecule CO is isoelectronic with N2 , in its ground state it thus has the
same electron configuration:
(σ1 s)2 (σ ∗ 1s)2 (σ2s)2 (σ ∗ 2s)2 (π2 p)4 (σ2 p)2 −→ 1
Σ.
Chapter 14 109
Since we are dealing here with a heteronuclear molecule, no g/u symmetry

occurs. On excitation, an electron from the bonding σ state is raised into the non-
bonding π state. This determines the configuration of the excited state:
. . . (π2 p)4 (σ2 p) (π ∗ 2 p) −→ 1

Π.
This verifies the assumption derived in (a) for the term symbols of the two states.
14.5 The continuous or diffuse spectra of polyatomic molecules in the short wave-
length region, λ < 200 nm (VUV) have several causes:
– owing to the numerous possibilities of absorption by different chromophore
groups, several band systems overlap;
– the energy is not sufficient to excite the molecule into the dissociation continuum,
whereby large molecules often have flatter potential curves than smaller ones
and are thus less stable;
– in the region of the excitation spectra considered, there is an intersection of
the potential curves of several different electronic terms and thus a tendency to
predissociation, which causes a shorter lifetime and therefore a greater linewidth
in the excited states.
14.6 The molecular orbital which is delocalised over the whole polyene can be
modeled in terms of a potential well of length L = 10 Å with infinitely high walls.
The energy of an electron in this potential well is then quantised according to
n2 h2
En = with n = 1, 2, 3, . . . (1)
8 me L 2
(m e : electron mass). The spacings between neighbouring energy levels become
greater as n increases. Each state which is associated with the quantum number n is
described by the wavefunction ψn :
' nπx (
2
ψn = sin . (2)
L L
The minimum excitation energy of the electron which is in the ground state
n = 1 is, from (1)
h2
∆E min = E 2 − E 1 = 3 , (3)
8m e L 2
and, with the given data,
E min = 1.83 · 10−19 J = 1.14 eV = 9098 cm−1 .
The minimum excitation energy of the electron lies in the near infrared.
The probability density is given by the squared magnitude of the wavefunction ψ.
The occupation probability W0≤x≤l of the particle in the state n in the range 0 to l is
calculated by integration:
110 Solutions
l l ' nπx (2
2
W0≤x≤l = ψn ψn∗ dx = sin dx (4)
0 L 0 L

l 1 2nπl
= − sin . (5)
L 2nπ L
With n = 1 and l = L5 , we obtain from this
W0≤x≤l = 0.05 .
The electron thus can be found with a probability of 5% at a distance of 0 to 2 Å

from the edge of the potential well.
Chapter 15
15.1 a) For the transmission T of a weakly absorbing homogeneous (i. e. stirred)

sample, in the first approximation, the Lambert-Beers law applies:
I
T= = e−αcd (1)
I0
(I0 : intensity of the incident light; I: intensity of the transmitted light; α: absorption
coefficient; c: molar concentration of the sample; d: thickness of the sample).
With the extinction coefficient (decadic absorption coefficient) ε = α log e , this
relation becomes
I
T= = 10−εcd . (2)
I0
From the definition of the optical density
I0
OD = log = εcd , (3)
I
we find the relation between the optical density and the transmission:
OD = − log T , (4)
and, with the given value T = 16% = 0.16 for λ = 256 nm,
OD (256) = 0.796 .
From (3) we obtain the extinction coefficient

OD
ε= (5)
cx
and finally
l
ε (256) = 159.2 .
mole cm
Chapter 15 111
For a sample thickness of 0.1 cm, we calculate using (2) that the transmission is
T = 83% .
b) The integral A giving the area under an inhomogeneously broadened absorp-

tion band of Gaussian profile
2
g(ν) = e−aν (6)
can be evaluated simply using

∞
−aν 2 π
A= e dν = . (7)
−∞ a
The full width at half maximum, (FWHM) ∆ν of the Gaussian curve is

ln 2
∆ν = 2 . (8)
a
Approximating the line shape as a rectangle with the same height as the Gaussian
gmax = 1 and the width ∆ν, we obtain for the area

ln 2
A =1·2 . (9)
a
An equivalent approximation is to treat the area as an equilateral triangle of
height gmax and base width 2 ∆ν. The areas determined from (7) and (9) differ by a
factor
√
π
√ ≈ 1.06 . (10)
2 ln 2
Calculation of the integral absorption from the linewidth ∆ν using the rectangle
approximation (9) is thus a sufficiently good approximation.
The definition of the integral absorption used for the computation of the oscillator
strength is

A= ε(ν) dν . (11)
In changing the variable from ν to ν, we must still take a factor of c (velocity of

light) into account:

A=c ε(ν) dν , (12)
so that the approximation (9) becomes
A = c εmax ∆ν . (13)
112 Solutions
With the data given in the problem, one finds the integral absorption of the band
to be
l cm2
A = 1.91 · 1016 = 1.91 · 1016 .
mole cm s mmole s
The oscillator strength f of the transition is found from the value of A thus
calculated by applying the simple relation
f = 1.44 · 10−19 · A (14)

with A in units of cm2 /mmole s . This leads to
f = 2.75 · 10−3 .
One thus obtains f # 1; the transition is in fact forbidden.

15.2 From the density = 1 g/cm3 and the molecular mass M = 18 g/mole, we
can calculate the approximate molarity c of seawater to be
mole
c= = 55.6 . (1)
M l
From (3) and (4) from problem 15.1, for a given extinction coefficient ε, we find
the relation between the layer thickness d and the transmission T to be
− log T
d= . (2)
εc
Using the given transmissions T , we then obtain the following water depths d:
for T = 50% : d = 873 cm ,

for T = 10% : d = 290.1 cm .
15.3 In a one-dimensional potential well of length L with infinitely high walls,

we find the energy eigenvalues E n and eigenfunctions ψn from problem 14.6. The
strength of an electrical dipole transition from the initial state ψn into the final state
ψm depends on the value of the transition dipole moment Θmn , i. e. on the transition
matrix element of the dipole operator er, which is a measure of the change in the
dipole moment during the transition:

Θmn = ψm∗ er ψn dV (1)
or, in the one-dimensional case,

L
Θmn = ψm∗ ex ψn dx . (2)
0
Chapter 15 113
With the wavefunctions ψn , we find

' mπx ( ' nπx (
2e L
Θmn = sin x sin dx . (3)
L 0 L L
Applying the trigonometric addition theorems, we obtain
eL
Θmn = {cos [(m − n)π] − 1}
π2(m − n)2
eL
− 2 {cos [(m + n)π] − 1} (4)
π (m + n)2
and thus for the two special cases

2eL 1
a) n → m = n + 1 : Θmn = 2 −1 , (5)
π (2n + 1)2
b) n → m = n + 2 : Θmn = 0. (6)
The oscillator strength f of the electronic transition n → m at the frequency ν

is given by
8π 2 m e ν
f = |Θmn |2 (7)
3 h e2
and with (5) and (6)
2
32 m e ν L 2 1
a) n → m = n + 1 : f a = − 1 (8)
3π 2 h (2n + 1)2
b) n → m = n + 2 : f b = 0 . (9)
From the bond energies E n , we determine the transition frequencies ν:
h2
a) n → m = n + 1 : ∆E a = hν = (2n + 1) , (10)
8m e L 2
h2
b) n → m = n + 2 : ∆E b = hν = (n + 1) . (11)
2m e L 2
Inserting into (8) and (9), this gives finally the oscillator strengths of the two
transitions
64 n 2 (n + 1)2
a) n → m = n + 1 : f a = · , (12)
3π 2 (2n + 1)3
b) n → m = n + 2 : f b = 0 . (13)
In β carotene, the conjugated π electron system of the polyene chain can be

modeled as a potential well. It consists of eleven double and ten single bonds and
extends over 22 C atoms, i. e. each wavefunction in the potential well is delocalised
114 Solutions
over 22 C atoms. With an average bond length of R = 140 pm, the length L of the
potential well is then
L = 22 R = 3.08 · 10−9 m . (14)
The levels in the potential well are to be filled with the 22 delocalised π electrons
from the eleven double bonds from the lowest upwards, so that up to n = 11, all the
levels are occupied with two electrons, each with antiparallel spins. The excitation
of the molecule thus starts from the state n = 11. According to (10) and (11), we
can now calculate the excitation energies:
h2
a) n → m = n + 1 : ∆E a = · 23
8m e L 2
∧
= 1,46 · 10−19 J = 7353 cm−1 , (15)
2
h
b) n → m = n + 2 : ∆E b = · 12
2m e L 2
∧
= 3.05 · 10−19 J = 15 346 cm−1 . (16)
These results correctly describe the order of magnitude of the measured transition
energies. Finally, using (12), the oscillator strength of the transition 11 → 12 can be
determined:
f a = 3.1
The oscillator strength gives the ratio between the actual intensity of the transition
and that of an ideal harmonic oscillator. This follows from its definition: for an
electron which is harmonically oscillating in three dimensions, we find f = 1; in
the one-dimensional case, we have f = 13 . The above calculated value of f = 3.1
can thus not be particularly precise. To be sure, we should not expect it to be, since
we are comparing two essentially different models (potential well and harmonic
oscillator).
15.4 a) On cooling of the H2 gas, the populations of the excited rotational and
vibronic states is reduced. The ratio of the population numbers N1 of the first excited
state of energy E 1 and N0 in the ground state is, for a given temperature T , given by:
N1 g1 −E1 /kT
= e (1)
N0 g0
(g1 , g0 : degree of degeneracy of the excited and the ground states; k: Boltzmann’s
constant). As a first criterion for the freezing out of the rotational and vibronic
degrees of freedom, we can apply the condition
!
1st criterion: E 1 = kT . (2)
The population of the first excited state is then (neglecting the different degrees of
degeneracy)
Chapter 15 115
1
N1 = · N0 ≈ 0.37 · N0 . (3)
e
Freezing out of the vibrational degree of freedom thus occurs at the temperature
hc ν0→1
TS,1 = , (4)
k
where ν0→1 = 4159.2 cm−1 is the wavenumber of the first vibrational transition of
hydrogen in its ground state:
TS,1 = 5987 K .
With the rotational constant B = 60.8 cm−1 of the H2 ground state, the temper-
ature for freezing out of the rotational degrees of freedom follows:
hcB J(J + 1) 2hcB
TR,1 = = (5)
k k
with J = 1 for the first excited rotational level and thus
TR,1 = 175 K .
As the second criterion for freezing out of the rotational and vibrational degrees
of freedom, we now require that the population of the corresponding first excited
state decrease to 1% of the population of th ground state:
N1 !
2nd criterion: = 1% = 0.01 . (6)
N0
For the vibrational states, the two degrees of degeneracy are g1 = g0 = 1, so
that with (1), we find
hc ν0→1
TS,2 = . (7)
k ln 100
At a temperature TS,2 , the n = 1 vibrational state is practically unoccupied:
TS,2 = 1300 K .
The degree of degeneracy of the rotational states of quantum number J are

g J = 2J + 1. With g0 = 1 and g1 = 3, one finds from (1) the temperature for
freezing out of the rotational degrees of freedom:
2hcB
TR,2 = (8)
k ln 300
and therefore for H2
TR,2 = 30.7 K .
116 Solutions
b) For HCl, we use B = 10.39 cm−1 and ν0→1 = 2885.6 cm−1 . From (4), (5),
(7), and (8) we can then compute the temperatures at which, according to the above
criteria, the rotational and vibrational degrees of freedom of HCl gas freeze out:
TS,1 = 4154 K ,
TS,2 = 902 K ,
TR,1 = 29.9 K ,
TR,2 = 5.2 K .
c) Hydrogen gas can be cooled with liquid helium in a cryostat, since it does not
yet condense at the calculated temperatures.
In contrast, HCl is already solid at the temperatures which are necessary to freeze
out the rotational degrees of freedom (TF = 158.3 K). The effective temperature
for rotation and vibration can however be reduced to the required value in a jet
(supersonic beam).
15.5 The relativistic red shift of the light emitted by the star if it is moving away
from the terrestrial observer at the velocity v F , is given by

λvF c + vF
= . (1)
λ0 c − vF
Here, λvF and λ0 are the wavelengths of the light with and without the Doppler
shift, and c is the velocity of light in vacuum. From this, we obtain the proper velocity
of the star:
λ2vF − λ20
vF = c · , (2)
λ2vF + λ20
and, inserting the values given in the problem text,

m
vF = 2.3 · 107 .
s
To compute the inhomogeneous linewidth due to the Doppler effect, we consider
the velocity distribution of the 48 Ti 8+ ions in the rest system of the star. This velocity
distribution has a Gaussian profile
m v2
N = N0 e− 2kT , (3)
where N and N0 are the numbers of ions which are moving with the velocities v
or 0 relative to the observer, and m is the mass of the ions. The full width at half
maximum (FWHM), ∆v, of the velocity distribution is

2kT ln 2
∆v = 2 . (4)
m
Chapter 15 117
Since the velocity v of the ions is much smaller than the velocity of light
c, the Doppler-shifted frequency νv of a moving atom can be determined non-
relativistically from the unshifted frequency ν0 of an ion at rest:
' v(
νv = ν0 1 − . (5)
c
For the wavelength shift of the moving source relative to that of the source at
rest, we find from this the relation
v
δλ = λ0 . (6)
c
The Doppler shift of the radiation emitted by the moving ions is proportional
to their velocity. We thus find from the Gaussian distribution of the velocities a
Gaussian absorption profile which, owing to 4) and (6), has a linewidth (FWHM) of

λ0 2λ0 2kT ln 2
∆λ = ∆v = . (7)
c c m
For a given width ∆λ of the spectral line, one can extract the temperature:
2
c ∆λ m
T= . (8)
2λ0 2k ln 2
With ∆λ = 0.618 Å, we calculate the surface temperature of the star to be
T = 8.4 · 105 K .
15.6 The energy uncertainty ∆E of a state with the lifetime τ is given by the
uncertainty relation
∆E · τ ≈ . (1)
Since the ground state has an infinitely long lifetime, its energy is arbitrarily
sharp. The lifetime broadening of an absorption line from the ground state into an
excited state is therefore determined only by the uncertainty of the excited state; it
is inversely proportional to its lifetime.
a) Due to the depopulation of the S2 state through rapid radiationless processes
(internal conversion, vibrational relaxation), its lifetime is of the order of 10−12 s,
while that of the S1 state lies near 10−8 s. Since the S1 state is thus sharper than the
S2 state, the S0 − S1 absorption line is narrower than the S0 − S2 line.
b) Radiative transitions from the triplet state T1 into the ground state S0 are forbid-
den, so that the T1 state is metastable and exhibits a long lifetime τ = 10−8 − 102 s.
The S0 –T1 line is thus in general very narrow.
118 Solutions
15.7 The CO molecule is isoelectronic with N2 ; its electron configuration in the

neutral ground state is
(1s σ)2 (1s σ ∗ )2 (2s σ)2 (2s σ ∗ )2 (2 p π)4 (2 p σ)2 .
The three observable bands in the photoelectron spectrum belong to electrons
from the three energetically highest occupied molecular orbitals, since these have
the lowest binding or ionisation energies E ion . The binding energy can to a first
approximation be determined from the difference of the energy of the ionising He-
lamp radiation (E He = 58.4 nm = 21.21 eV) and the kinetic energy E kin of the
photoelectrons; cf. Fig. S.28:
Fig. S.28. The photoelectron spectrum of CO after excitation at 21.21 eV. For problem 15.7
E ion = E He − E kin .
E kin /eV E ion /eV Molecular Vibrational constant

orbital νe /cm−1
7.2 14.0 2p σ
4.9 16.3 2p π 1610
1.7 19.5 2s σ ∗ 2420
The νe values calculated from the vibronic maxima of the bands correspond
approximately to the vibrational constants in the corresponding excited states of
CO+ . For the ground state of neutral CO, one finds νe = 2143 cm−1 . Removing a
2 p π electron weakens the bond, so that the vibrational frequency decreases. 2 p π
can thus be identified as a bonding orbital, 2s σ ∗ in contrast as an antibonding orbital.
Chapter 18 119
15.8 From the spacings of the vibronic maxima in the 9 eV band of the photoelec-
tron spectrum of H2 O, the vibrational constant νe of the corresponding molecular
orbital can be determined:
∆E = 0.41 eV ⇒ νe = 3306 cm−1 .
This value is comparable with the vibrational constant ν1 = 3657 cm−1 of the
symmetric stretching vibration of neutral H2 O. For the 7 eV band, one calculates a
value according to
∆E = 0.125 eV ⇒ νe = 1008 cm−1 ,
which is more strongly shifted compared to ν2 = 1595 cm−1 for the bending vibra-
tion. This suggests that the 7 eV electron originates from a more strongly bonding
orbital, since its excitation has a stronger effect on the vibrational behaviour of the
molecule.
Chapter 18
18.1 a) The nuclear spin can be treated as a classical angular momentum I. It

corresponds to a magnetic moment µ:
µ= γ I, (1)
where we have introduced as proportionality constant the gyromagnetic ratio γ . In a

constant external magnetic field B, a torque therefore acts on the spin and it is equal
to the rate of change of the angular momentum:
d
I = µ × B. (2)
dt
Together with (1), this yields the equation of motion of the magnetic moments:
d
µ = µ × (γ B) (3)
dt
as well as that of the nuclear spins:
d
I = I × (γ B) . (4)
dt
In order to solve this equation of motion, one introduces a rotating coordinate
system with the unit vectors i, j, k, which rotates in the laboratory (rest) system
according to
d
i =Ω ×i (5)
dt
120 Solutions
at the frequency Ω. The total differential (3) thus becomes

∂
µ + Ω × µ = µ × γ B. (6)
∂t
The second term in the sum on the left side describes the rotation of the coordinate
system, while the first term gives the rate of change of µ in the rotating system:
∂
µ = µ × (γ B + Ω) . (7)
∂t
In the rest system of µ, its time derivative is zero, so that we find from Eq. (7)
for the Larmor frequency ΩL with which the rest system and thus the vector µ in
the laboratory system rotates:
ΩL = −γ B . (8)
Thus in a static magnetic field, µ and –from (1)– also the nuclear spins I carry
out a precessional motion around the direction of the field at the frequency ΩL .
b) The gyromagnetic ratio γ for a particular nucleus can be computed from the
corresponding nuclear g-factor g I (5.58569 for 1 H), the nuclear magneton µN =
0.505 · 10−26 A m2 , and Planck’s constant according to
g I µN
γ = . (9)

For the Larmor frequency νL = 1
2π ΩL it follows using (8) that
g I µN
νL = B, (10)
h
and, with the given value for the Earth’s magnetic field B,
νL = 553.4 Hz .
18.2 a) Inserting the fixed magnetic field B = 2.166 T into expression (10) from
problem 18.1 for the Larmor frequency of a proton, one finds
νL = 9.22 · 107 Hz = 92.2 MHz .
This is, however, the frequency of the radio station SDR 3, which broadcasts from
the Stuttgart Television Tower at high power. The student is worried that therefore
his measurements will be disturbed by the radio signals (or that reception of SDR3
in the laboratory will be disturbed by the NMR apparatus).
b) Solving Eq. (10) from problem 18.1 for the magnetic field B, we find
hν
B= . (1)
µN g I
Chapter 18 121
With the g-factor g I = 5.58569 for protons, 1 H, and the given frequency ν, we
calculate the field strength
BH = 0.71 T .
On the other hand, with g I = 1.40482 for 13 C, we obtain
BC = 2.83 T .
18.3 Denoting the resonance magnetic field for the standard, TMS, as BT and the
resonance magnetic field for the sample substance as BS , the chemical shift δ obeys
the equation
BT − BS
δ = 106 . (1)
BT
There is thus a linear relationship between δ and the field BS , so that the calcu-
lation of the centre of gravity of the lines of a multiplet can be carried out equally
well on the B axis or on the δ axis:
1
B = (B + BS ) (2)
2 S
1
or δ = (δ + δ ) (3)
2
(for a doublet of lines). The chemical shift of the nuclei A and B is defined as the
centre of gravity of the doublets, since in the special case of weak coupling J # δ,
the splitting due to the dipole-dipole interaction is symmetric around the position of
the unsplit line:
δ A = 5.75 ppm and δ B = 1.05 ppm
(the assignment of the nuclei A and B is arbitrary). The doublet splitting of the two
lines is caused by the dipole-dipole coupling between two nuclei with spins 12 and
the coupling constant J AB . It corresponds to the line spacing in the two doublets:
J AB = 0.1 ppm .
Since from Eq. (10) in problem 18.1, the resonance frequency

g I µN
ν= B (4)
h
is proportional to the resonance field, the chemical shift can be expressed in terms
of a frequency using (1):
νT − νS
δ = 106 . (5)
νT
122 Solutions
The coupling constant, i. e. the line spacing in the doublet, is then
J AB = ∆νS = (−) 10−6 νT ∆δ , (6)
and with νT = 100 MHz and ∆δ = 0.1,
J AB = 10 Hz .
18.4 An electronegative neighbour atom withdraws electrons from the vicinity of

a spin and thereby reduces the magnetic shielding of the electrons at the nuclei.
Resonance then occurs at a lower B field, so that the chemical shift δ increases.
From the quantitative values of the electronegativities EN of the neighbouring
atoms, X = O, N, C,
ENO > ENN > ENC ,
it follows for the chemical shifts of the methyl protons that
δ(O−CH3 ) > δ(N−CH3 ) > δ(C−CH3 ) ,
or, more precisely: δ(O−CH3 ) ≈ 3.3 ppm, δ(N−CH3 ) ≈ 2.3 ppm and δ(C−CH3 ) ≈
0.9 ppm.
18.5 For the NMR spectrum of the fluorine nucleus, the coupling constants JHH
between the protons are irrelevant when, as in the present case, the chemical shifts
δ of the three inequivalent protons are very different: ∆δ JHi H j (the protons are
so to speak “very inequivalent”). The dipole-dipole interaction of the F nucleus with
the proton Ha then causes a doublet splitting of the F resonance line with a line
spacing of JHa F ; the interaction with Hb gives rise to an additional splitting of both
lines into pairs with a spacing of JHb F ; and finally, the interaction with Hc causes a
splitting of each line into two new lines at a spacing of JHc F . The final result is a
splitting into 23 = 8 lines of equal intensity.
Chapter 18 123
Ha Hb Hc ∆ν relative to the centre of the octet

↑ ↑ ↑ 2 (85 + 50 + 20) Hz = 77.5 Hz
1
↑ ↑ ↓ 2 (85 + 50 − 20) Hz = 57.5 Hz

1
↑ ↓ ↑ 2 (85 − 50 + 20) Hz = 27.5 Hz

1
↑ ↓ ↓ 2 (85 − 50 − 20) Hz = 7.5 Hz

1
↓ ↑ ↑ 2 (−85 + 50 + 20) Hz = −7.5 Hz

1
↓ ↑ ↓ 2 (−85 + 50 − 20) Hz = −27.5 Hz

1
↓ ↓ ↑ 2 (−85 − 50 + 20) Hz = −57.5 Hz

1
↓ ↓ ↓ 2 (−85 − 50 − 20) Hz = −77.5 Hz

1
18.6 In 1,1,2 trichloroethane, CHa2 Cl−CHb Cl2 , there are two types a and b of
inequivalent protons. Since the Hb is in the immediate neighbourhood of two elec-
tronegative chlorine atoms, its chemical shift is larger than that of the two Ha , whose
C atom is substituted with only one Cl. Due to the dipole-dipole interaction with Hb ,
the lines of the two Ha protons split into two lines with a spacing of JHa Hb and equal
intensity; the Hb resonance, in contrast, splits into three lines due to the coupling
with the two Ha , with a line spacing of JHa Hb and relative intensities in the ratio
1:2:1. The integrated intensity ratio of the resonance signals of Ha (two protons) to
Hb (one proton) is 2:1.
In the measured spectrum, one observes a doublet at δ = 3.95 ppm (Ha ) and
a triplet at δ = 5.77 ppm (Hb ). The line splitting is in each case JHa Hb = 6.0 Hz.
Furthermore, one observes a smaller solvent peak at δ = 7.28 ppm, which is caused
by CHCl3 impurities in the CDCl3 solvent.
18.7 In the spectrum of H3 a C−CH2 b −CH2 c −NO2 , the assignment of the signals
to the inequivalent proton species can be carried out noting the inductive effect of the
electron-attracting NO2 group: because of the lowest electron density and nuclear
shielding at the positions of the Hc protons, their resonance occurs at the lowest B
field, so that their chemical shift has the largest value:
δc = 4.38 ppm .
Taking into account the long-distance effects on the two other groups of protons,
one finds
δb = 2.07 ppm and δa = 1.03 ppm .
Through the dipole-dipole interaction with the two Hb , the Hc line splits into a
triplet. Since there is no measurable additional splitting due to interactions with the
Ha , one concludes that the coupling constant Jac is negligible:
Jac ≈ 0 .
124 Solutions
The same result can be derived from the threefold splitting of the Ha signal.
It is valid only within the precision of the measurement, i. e. the resolution of the
spectrometer.
For the Hb line, due to the interactions with three protons of type Ha and two
protons of type Hc , one would expect a 4·3 = 12-fold splitting. From the multiplicity
of 6, however, it can be concluded that there is an interaction with five (for this
interaction) equivalent protons (5 + 1 = 6-fold splitting) and thus an equally strong
coupling of Hb to Ha and Hc :
Jba = Jbc .
In fact, small deviations from a simple sextet pattern are observed, since Jba = Jbc
is not precisely fulfilled.
18.8 a) In the molecule CH3 a CHb O, there are two types a and b of inequiva-
lent protons. Taking the inductive effect of the electronegative oxygen atom in the
aldehyde group into account, one sees that for the chemical shifts of the two protons,
δb > δa .
The precise numerical values can be taken from part b):
δa = 2.20 ppm and δb = 9.80 ppm .
The ratio of the integrated intensities of the signals of three nuclei of type Ha to that
from one nucleus of type Hb is approximately 3:1.
The Ha line splits into a doublet of equally intense lines with a spacing of
J = 2.9 Hz; the Hb resonance, in contrast, splits into a quartet with the same line
spacing and the relative intensity ratios 1:3:3:1.
b) The chemical shift δ is computed from the resonance-field strengths BT for
the frequency standard TMS and BS for the sample, according to the formula
BT − BS
δ = 106 · . (1)
BT
The linear relationship between δ and BS allows us to extract the difference ∆BS
of the magnetic resonance field strengths directly from the difference ∆δ between
the chemical shifts of two types of protons. Since for the standard, the chemical shift
is defined to be δT = 0, it corresponds to an external B0 field at a value of BT , so
that it follows from (1) that:
∆BS = (−) 10−6 BT ∆δ . (2)
With ∆δ = 7.6 ppm, one obtains
for B0 = 1.5 T : ∆BS = 1.14 · 10−5 T ,
for B0 = 7.0 T : ∆BS = 5.32 · 10−5 T .
The differences ∆BP of the resonance field strengths correspond to the differ-
ences between the local magnetic fields in the different regions of the molecule.
Chapter 18 125
18.9 To predict the NMR spectrum of incompletely deuterated acetone,

CD3 COCD2 H, we need merely take into account the coupling of the proton to
the two deuterium nuclei of the CD2 H group. Since the spin of the deuterium nuclei
is I = 1, it has three possible orientations relative to the direction of the external
magnetic field: m I = −1, 0, 1. The nine possible combinations for two spins are
then
(+1, +1) ; (+1, 0) ; (+1, −1) ; (0, +1) ;

(0, 0) ; (0, −1) ; (−1, +1) ; (−1, 0) ; (−1, −1) .
From the coupling of the proton spin to the two deuterium spins, we thus find a
multiplet consisting of five equidistant lines (pentet; the line spacing is JHD ) with the
relative line intensities 1:2:3:2:1, as one can see from the number of possible combi-
nations for the total (combined) spin of the two D nuclei (m Iges = −2, −1, 0, 1, 2).
18.10 Due to diffusion, rotation, or other processes, the magnetic environment of
a spin is constantly changing, e. g. due to changes in its distance and angle to a
neighbouring dipole. In the model, this is simulated by superposing a statistically
fluctuating field onto the constant external magnetic field B0 , which takes on the
values +Bi and −Bi . The shift of the precession frequency of the spin relative to the
frequency ω0 in the B0 field alone is then given by
∆ω0 = ± γBi . (1)
If τ is the average time during which the magnetic environment remains constant,
then it follows for the average phase shift δφ which the precessing spin undergoes
between two successive field jumps, relative to a spin in the field B0 , that
δφ = ± τ γBi (2)
with δφ # 1. The time-dependent change of the phase angle φ can be compared

with a “random walk”, in which a drunken man staggers alternately forward and
backward along a straight path; cf. Fig. S.29. The average distance L which he will
move after n steps of length s, expressed as the mean squared distance, is
2
L = n s2 . (3)
Applied to the model of a spin in a fluctuating B field, this means that
φ2 = n (δφ)2 = n τ 2 γ 2 Bi2 . (4)
Here, φ2 is the mean squared phase shift relative to a spin in the constant B0 field.
The spin contributes to the resonance signal approximately so long a its phase shift
is less than a value of 1 rad. The number n of steps required to reach this limit is
found from (4) to be
1
n= . (5)
τ2 γ 2 Bi2
126 Solutions
[!t]
Fig. S.29. The phase ϕ of a spin in a fluctuating magnetic field (· · · magnetic field B0 ± Bi ,
—– Phase shift ϕ relative to spins in only the B0 field)
The time required for n fluctuations of an average length τ, i. e. the time for a
sufficient phase loss of a spin relative to the spin in only the B0 field, is denoted as
the transverse relaxation time T2 in Bloch’s theory of magnetic resonance:
1
T2 = n τ = . (6)
τ γ 2 Bi2
In contrast, in a rigid lattice, in which merely the unchanging magnetic environ-

ments B0 and B0 ± Bi are present, the result is
1
T2 = . (7)
γBi
The linewidth in the case of rapid fluctuations with the characteristic time τ
follows from (6):
1
∆ω = = τ γ 2 Bi2 . (8)
T2
With (7) and (8), we obtain an expression for the line narrowing caused by
fluctuations relative to the width ∆ω0 in a rigid lattice :
∆ω = (∆ω0 )2 · τ . (9)
The requirement δφ # 1 means that ∆ω0 · τ # 1, so that
∆ω # ∆ω0 . (10)
Motional narrowing thus becomes more and more effective the shorter the mean
fluctuation time τ.
Chapter 18 127
18.11 In liquids, the fluctuations of the local magnetic field at the location of a spin
are often much more rapid than in a solid, in which neither diffusion nor rotation of
the molecules are effective. According to Eq. (6) from problem 18.10, the transverse
relaxation time T2 depends on the mean fluctuation time τ and the strength Bi of the
field fluctuations:
1
T2 = n τ = .
τ γ 2 Bi2
It is therefore longer in liquids (T2 ≈ 10−4 − 10 s) than in solids (T2 ≈ 10−4 s).
The shorter lifetime of the phase-coherent spin states in solids leads –due to the
uncertainty relation– to broader NMR lines (T2 : phase-loss time).
18.12 a) At a lower exchange frequency, one finds two distinct resonance signals,
which are separated by ∆ν = 39 Hz. The two lines begin to merge when their widths
∆ω L , determined by the lifetimes, correspond to their separation on the frequency
axis. A single line is accordingly observed when
∆ωL > 2π ∆ν . (1)
The linewidth ∆ωL is coupled to the lifetime τ of the states (i. e. of the magnetic
environments):
1
∆ωL = . (2)
τ
From (1) and (2), we obtain
1
τ< , (3)
2π ∆ν
and thus for the exchange frequency νA = 1
τ between the different environments:
1
νA = = 2π ∆ν . (4)
τ
With the given value for ∆ν, one obtains
νA = 245 Hz .
b) The cyclohexane ring occurs in three different conformations, of which two

are termed “chair” forms and the third a “boat” form. On inversion between the
two chair conformations and the boat conformation, the methyl substituent jumps
between an axial and an equatorial position and thus between different magnetic
environments. The protons of the methyl group exhibit different chemical shifts,
depending on their position, so that one observes two groups of NMR lines at
low temperatures. On raising the sample temperature, one increases the inversion
frequency, until according to (a), line broadening and merging and finally, at high
frequencies, exchange narrowing occurs (compare problem 18.10).
128 Solutions
18.13 a) For the manipulation of an individual experiment, the maximum time

available is the time up to complete loss of phase coherence, i. e. the transverse
relaxation time T2 :
1 1 1
= + .
T2 T2 2T1
Here, T2 is composed of the inhomogeneous contributions (e. g. due to different B
fields at the locations of different nuclei) and the homogeneous contributions due to
spin-spin interactions. T1 denotes the longitudinal or spin-lattice relaxation time.
b) The pulse sequence π2 −τ −π −τ is used in the spin-echo method to switch off
the inhomogeneous contributions to the phase-loss time T2 (local variations in the B
field, chemical shifts, etc.). Irradiation at a second radiofrequency leads to frequent
and rapid flipping of the spins of those nuclear species which are in resonance at this
second frequency. In this way, the anisotropic dipole-dipole interaction with these
nuclei is switched off.
18.14 Instead of carrying out numerous individual measurements with different τ
as in the spin-echo technique, one can also follow the decrease of the echo height
in a single experiment by applying the method of Carr and Purcell; here, several π
pulses are applied at times τ, 3τ, 5τ, . . . and thus echoes are obtained at times of
2τ, 4τ, 6τ, . . . ; cf. Fig. S.30. The pulse sequence has the following form:
π
− τ − π − τ −Echo− τ − π − τ −Echo− τ − π − . . . .
2
The nuclear magnetisation is first rotated into the x − y-plane, then it decreases due
to loss of phase coherence of the spins (“spreading out” of spin directions); and then,
following inversion of the spins around the y axis by the π pulse, it again increases
up to the echo. Thereafter, the spins again spread out in the reverse sense, and the
process is repeated.
The decrease of the echo height is due here (besides longitudinal relaxation) only
to homogeneous relaxation, caused by spin-spin interactions, so that the relaxation
time T2, hom can be simply determined. The echo height decreases according to
Fig. S.30. Pulse sequence and echoes in the Carr-Purcell method. After M. Mehring: Princi-
ples of High Resolution NMR in Solids; Chap. 3 (Springer, Berlin, Heidelberg 1983)
Chapter 18 129
Mxy (t) = Mxy (0) · e−1/T2, hom
with t = n · 2τ.
18.15 We write the Schrödinger equation
iψ̇ = Hψ (1)
in the form
i
ψ̇ = − Hψ . (2)

Its formal solution is
i
ψ = e− ~ Ht ψ(0) . (3)
Inserting (3) into

ψ ∗ D ψdx , (4)
where D is an arbitrary operator, we obtain

2 3
i i
= e ~ Ht ψ ∗ (0) De− ~ Ht ψ(0)dx . (5)
Since H is an hermitian operator, we have

Hψ ∗ φdx = ψ ∗ Hφdx , (6)
and likewise

n ∗
H ψ φdx = ψ ∗ H n φdx . (7)
Making use of (7) and

∞

i (it)n
e ~ Ht = Hn , (8)
n=0
n!
we obtain from (5)

i − ~i Ht
ψ ∗ (0) e ~ Ht D
e ! ψ(0)dx . (9)
D H (t)
Using the bra and ket notation, we can write with (4) and (9)
ψ(t)| D |ψ(t) = ψ(0)| DH (t) |ψ(0) , (10)

130 Solutions
where
i i
DH = e ~ Ht D e− ~ Ht . (11)
Differentiating both sides of (11) with respect to time, we find

i i − ~i Ht i i − ~i Ht
Ḋ H (t) = H e ~ Ht De
! − e ~ Ht De
!H, (12)

DH DH
i
= (HDH − DH H) , (13)

or, finally, using the commutator notation,
i
Ḋ H = [H, DH ] . (14)

18.16 We represent the spin function as a superposition
|φ = a |↑ + b |↓ , (15)
whereby

1 0
|↑ = , |↓ = . (16)
0 1
In units of , the operator for the x component of the angular momentum is

1 01
Ix = . (17)
2 10
We compute its expectation value:

φ| Ix φ| = a∗ ↑| + b∗ ↑| Ix (a |↑ + b |↓) , (18)
where the following relations hold
↑| Ix |↑ = 0 , (19)

↓| Ix |↓ = 0 , (20)
1
↓| Ix |↑ = , (21)
2
1
↑| Ix |↓ = . (22)
2
With (19–22), we find for (18)
1 ∗
(a b + b∗ a) . (23)
2
Chapter 18 131
If we set a = ceiφ/2 and b = ce−iφ/2 in this expression, we find for the expectation
value of the x component:
1 2 −iφ
|c| e + eiφ = |c| 2 cos φ , (24)
2
and correspondingly for the y component:
φ| I y |φ = − |c| 2 sin φ (25)
as well as
φ| Iz |φ = 0 . (26)
The expectation value of the spin vector thus undergoes a coherent rotational
motion.
18.17 We begin with the usual commutation relations,
Ix I y − I y Ix = iIz , I y Iz − Iz I y = iIx , Iz Ix − Ix Iz = iI y (27)
and Heisenberg’s equation of motion

i
I˙kx = [H, Ikx ] . (28)

Taking
H = JIkz Ilz , l = k (29)
and (27), we obtain
I˙kx = i2J [Ikz Ilz , Ikx ]

= i2JIlz [Ikz , Ikx ] (30)
!
iIky
and thus
I˙kx = −2JIlz Iky . (31)
Since this equation is not closed, we set up another Heisenberg equation

·
Iky Ilz = i2J Ikz Ilz , Iky Ilz

= i2J Ilz2 Ikz , Iky , (32)
! !
1 −iIkx
4
obtaining
· 1
Iky Ilz = JIkx . (33)
2
132 Solutions
From (31) and (33), we find by standard methods the result

I¨kx = −J 2 Ikx . (34)
Its solution is
Ikx (t) = a cos Jt + b sin Jt , (35)
where
a = Ikx (0) (36)
and, due to

I˙kx (0) = Jb = −2J Ilz Iky t=0 , (37)
b is found to be given by

b = −2 Ilz Iky t=0 . (38)
The solution we are seeking is therefore

Ikx (t) = Ikx (0) cos Jt − 2 Ilz Iky t=0 sin Jt . (39)
The solution for Iky Ilz = Ilz Iky follows immediately from (31) and (39).
18.18 We write

ψ| I± (0) |ψ = α∗ ↑| + β ∗ ↓| Ix (0) ± iI y (0)
× (α |↑ + β |↓) . (40)
On multiplying through, we obtain expressions of the form α∗ α↑ |Ix | ↑ etc. These
can be evaluated with the aid of (19–22) and corresponding relations for I y . We find
for (40)
eiΩt I+ = eiΩt α∗ β (41)
and
e−iΩt I− = e−iΩt αβ ∗ . (42)
We set
α∗ β =| α β | e−iφ (43)
and obtain
1
Ix (t) = [I+ (t) + I− (t)] =| α β | cos (Ωt − φ) (44)
2
and
1
I y (t) = [I+ (t) − I− (t)] =| α β | sin (Ωt − φ) , (45)
2i
from which we can see the precessional motion directly.
Chapter 18 133
18.19 The magnetic moment of the proton is

e
µ= s, (46)
m0
and its energy in a magnetic field:
Vs = −µ B . (47)
The following relations hold:

1
B = rot A = Nλ bλ i (kλ × eλ ) eikλ r + c.c. (48)
ωλ
with

ωλ
Nλ = . (49)
2ε0 V
In the usual experimental arrangement (see also problem 18.20), the spin is subjected
to a constant field B along the z direction and an oscillating field perpendicular to it
(48). Correspondingly, we choose kλ = (0, 0, k z ) and eλ = (0, e y , 0).
We then obtain

e 1 ωλ
Vs = ikλ eikλ z sx bλ + h.c. (50)
m 0 ωλ 2ε0 V
Setting sx = 12 (s+ + s− ) and s = I, we find

e
Vs = · kλ i eikλ z (I+ + I− )bλ + h.c. (51)
2m 0 2ε0 ωλ V
From this result, we can immediately read off the coupling constant that we are
seeking. We arrive at the “rotating-wave” approximation by investigating the motion
of the operators in the Heisenberg representation (without interactions between I and
B (48)). Then bλ ∼ e−iωλ t , I± ∼ eiΩt is found. In the neighbourhood of resonance,
Ω ≈ ωλ , the terms I+ bλ , I− b+ +
λ vary slowly, while the terms I+ b ∼ e
iΩt+iωλ t
and
I− b oscillate rapidly and make practically no contribution to the interaction energy.
They can therefore be neglected in (51).
18.20 We start with
i
ḃ+ = [H, b+ ] . (1)

Due to the commutation relations
[b, b+ ] = 1, [b+ , Ix ] = [b+ , I y ] = [b+ , Iz ] = 0 , (2)

134 Solutions
it follows that
ḃ+ = iωb+ − gI+ . (3)
A precessing spin is described by:
I+ (t) = I+ (0) eiΩt . (4)
The equation of motion with dampíng is:
ḃ+ = iωb+ − κb+ − g I+ (0) eiΩt . (5)
It has the solution:
b+ (t) = −g I+ (0) eiΩt (iΩ − iω + κ)−1 . (6)
Chapter 19
19.1 The dipole-dipole interaction between an electron spin and the nuclear spins
leads to a hyperfine splitting of the ESR lines. The number of lines is given by the
number 2I + 1 of different possible orientations of the nuclear spin I relative to the
direction of the electronic spin. The spins of the participating nuclei are I = 0 for
12
C; I = 12 for 13 C, 1 H and 19 F; I = 1 for 2 D; and I = 32 for 35 Cl and 37 Cl.
[12 CF2 H]• 6 lines (2-fold splitting by H, 3-fold by F2 with Iges ,

F2 = 1);
[13 CF2 H]• 12 lines (additional 2-fold splitting by 13 C);
[12 CF2 D]• 9 lines (3-fold splitting by D, 3-fold by F2 );
[12 CClH2 ]• 12 lines (4-fold splitting by Cl, 3-fold by H2 ).
19.2 a) The hyperfine structure of the naphthalene radical is determined by the
interaction of its electronic spin with two groups α and β, each having four equivalent
protons. The coupling constants are aα and aβ .
For the orientations of the four equivalent α proton spins, I = 12 , relative to
the electronic spin, there are five possibilities (M I = 0 , ±1 , ±2), so that the ESR
signal splits into five lines with a spacing of aα and with relative intensities of
1:4:6:4:1 (Pascal’s triangle!). Due to the interaction with the four β proton spins,
each of these five lines again splits into five lines with a spacing of aβ and with
relative intensities of 1:4:6:4:1. Thus, five groups of five lines each are observed,
whereby the intensity ratios and spacings of the group maxima are determined by
the α interaction (1:4:6:4:1; aα ).
The 25 lines can thus be associated with the orientations (Mα , Mβ ) of the total
spins Iα and Iβ of the two groups of equivalent protons. With increasing B field,
i. e. for the lines of the given spectrum going from left to right, one obtains
Chapter 19 135
(−2, −2)(−2, −1)(−2, 0)(−1, −2)(−2, 1)(−1, −1)

(−2, 2)(−1, 0)(0, −2)(−1, 1)(0, −1)(−1, 2)(0, 0)
(1, −2)(0, 1)(1, −1)(0, 2)(1, 0)(2, −2)(1, 1)(2, −1)
(1, 2)(2, 0)(2, 1)(2, 2).
b) In agreement with the notation of the H locations in the problem text (structural
formula), in 1-iodo naphthalene, the iodine substituent is located at an α position.
The interaction of the electronic spin with the remaining three α proton spins and
the four β spins then leads to a splitting into 20 lines which can be grouped into four
pentets. The intensity ratio of the group maxima is 1:3:3:1, and their spacing is aα .
In 2-iodo naphthalene, on the other hand, the iodine atom replaces a β proton,
so that a splitting into five quartets and thus likewise a total of 20 lines is observed.
The intensity ratios of the group maxima are 1:4:6:4:1, and their spacing is again aα .
The interaction with the spin of the 127I nucleus (nuclear spin 52 ) was neglected in
the above considerations. It would further split each of the lines into six sublines.
c) The substitution of one hydrogen by a methyl group leads in analogy to (b) to
a splitting of the ESR signal in each case into 20 lines. In addition, the interactions
with the three methyl protons produces a further quartet splitting of each line with
the intensity ratio (1:3:3:1) – the hyperfine structure thus includes all together 80
lines.
19.3 a) In the [NH2 ]• radical, the dipole-dipole interaction of the electronic spin
with the spin of the nitrogen nucleus 14 N (I = 1) leads to a splitting of the signal
into three lines of equal intensity, with a spacing of aN = 1.03 mT. The coupling
with the two proton spins I = 12 splits each of the three lines into a triplet (line
spacing aH = 0.35 mT, intensity ratio 1:2:1).
b) The ESR spectrum of the azide radical [14 N 3 ]• exhibits a hyperfine structure
with seven lines, which is caused by the spin-spin interaction of the free electron with
the total spin of the three nitrogen nuclei. The lines of the septet have the relative
intensities 1:3:6:7:6:3:1; cf. Figs. S.31 and S.32.
Fig. S.31. Hyperfine structure in the ESR

spectrum of the [14 NH2 ]• radical. One ob-
serves nine lines, which can be grouped into
three triplets
136 Solutions
Fig. S.32. Since the nuclear spin is I = 1,

the relative line intensities cannot be obtained
using Pascal’s triangle. Instead, one uses the
“family-tree method” (counting the various
possibilities for going rom the tip to a par-
ticular end point)
19.4 a) From the hyperfine-structure coupling constants aα and aβ of naphthalene,

using the McConnell equation one can calculate the spin density at the C atoms α
and β; cf. Fig. S.33:
aα = (−) 0.495 mT ⇒ α = 0.22

aβ = (−) 0.186 mT ⇒ β = 0.08 .
Fig. S.33. The spin-density distribution in the naphthalene radical

anion. The free electron spends 22% of its time at each of the α-C
atoms, 8% at each of the β-C atoms
The π electron is localised

6 to the extent of 88% at the four α-C atoms. 6 The
normalisation condition i = 1 is not well fulfilled; instead, one obtains i =
1.2. The McConnell equation is merely a rough approximation. No information can
be obtained about the occupation probability at the two central C atoms (which have
no C−H bonds).
b) The assignment of the given values for the spin density, i , to the inequivalent
C positions in anthracene (see above) can be carried out by means of a comparison
with naphthalene. The spin density , and with it the coupling strength a, decrease
on going from the centre of the molecule outwards, and we find
α = 0.193 , β = 0.097 , γ = 0.048 .

6
The normalisation condition is well fulfilled, i = 0.966. Applying the
McConnell equation, one can obtain the hyperfine-structure coupling constants for
Chapter 19 137
the interaction of the electronic spin with the proton spins, starting from the spin
densities:
aα = Q α = (−) 0.434 mT ,
aβ = Q β = (−) 0.218 mT ,
aγ = Q γ = (−) 0.108 mT .
The interaction of the electronic spin with the three types of protons leads to the
following splittings (number of lines, relative intensity ratios, and line spacings):
Hα : 3 lines, 1:2:1, aα ,
Hβ : 5 lines, 1:4:6:4:1, aβ ,
Hγ : 5 lines, 1:4:6:4:1, aγ .
One thus expects a hyperfine-structure splitting of the ESR signal of anthracene
into 75 lines.
19.5 The resonance frequency ν of an electron in a magnetic field of strength B is
given by the relation
g µB
ν= B (1)
h
(µB = 9.274 · 10−24 Am2 : Bohr magneton). From it, with the given values for the
frequency and the B field, we can compute the g-factor of the electron:
hν
g= . (2)
µB B
Since the hyperfine-structure splitting of the signal is symmetric around the
centre of gravity of the lines, we take as the B field the average value of the two line
positions (B = 303.32 mT). From this, we obtain
g = 2.0022
for an electron in atomic hydrogen. The strength of the dipole-dipole interaction is
independent of the applied B field, so that we find the coupling constant directly
from the spacing of the two hyperfine-structure lines:
a = ∆B = 50.70 mT .
19.6 In the CH3 radical, the hyperfine structure of the ESR lines is due to the
interactions of the free electron spin with the three equivalent protons of nuclear spin
I = 12 . One observes a quartet, for which the relative line intensities of 1:3:3:1 can be
derived using Pascal’s triangle. In the triply-deuterated CD3 radical, an interaction
with three equivalent D nuclear spins of I = 1 occurs. The line intensities in the
resulting septet are determined by the family-tree method described in problem 19.3
b) to be 1:3:6:7:6:3:1.
138 Solutions
19.7 The six protons in the benzene molecule are chemically and magnetically
equivalent. This has two consequences: on the one hand, spin diffusion takes place
between the nuclei, i. e. there is a rapid exchange of their spin orientations. Any
differences in the interactions of the electron with the individual nuclear spins is
averaged out, so that one must treat the equivalent protons all together, i. e. as a total
spin. This picture also enters into the description of the dipole-dipole interaction for
the elucidation of the fine structure of NMR spectra.
On the other hand, the free electron of the benzene radical anions is localised to
the extent of 16 on each of the equivalent C atoms, so that its spin couples equally
strongly to each of the six protons. It is therefore irrelevant how the spin orientations
which contribute to a given total spin are distributed among the individual protons.
The rotation of the molecule as a whole has no influence at all upon these two
intramolecular effects. It therefore causes no wiping out of the hyperfine structure
owing to exchange phenomena.
In order to calculate the hypothetical exchange narrowing of the ESR signal,
we first briefly ignore the counterarguments given. The hyperfine-structure coupling
constant for the interaction of the free electron spin with the proton spins in benzene
is a = 0.375 mT. The hyperfine structure of the ESR signals vanishes when the
lifetime broadening νL of the individual lines becomes as large as the spacing ∆νHfs
of two neighbouring lines. The lifetime τ of a spin state is obtained from Heisenberg’s
uncertainty relation according to:
∆E · τ ≡ h ∆ν · τ = , (1)
and from it, the lifetime broadening of a hyperfine-structure line:
1
νL = ∆ν = = . (2)
hτ 2π τ
In the case of a benzene molecule rotating about its sixfold symmetry axis, the
lifetime of a particular spin state is limited by the rotational frequency νR . The
rotation leads to a rapid exchange of the nuclear spin positions at an exchange
frequency of
1
νA = = 6 νR . (3)
τ
It follows from this that
6νR
νL = , (4)
2π
and with
g µB
νL = ∆νHfs = a, (5)
h
one finally obtains the rotational frequency at which, according to this model, the
hyperfine structure should be averaged out:
π g µB
νR = a. (6)
3h
Chapter 19 139
Wit the value given for a, we calculate

νR = 1.10 · 107 Hz .
The rotational frequency corresponds classically to a rotational energy of
1 ∧
ER = Θ (2π νR )2 = 7.1 · 10−30 J = 3.6 · 10−7 cm−1 .
2
The rotational energy levels of the symmetric-top molecule are given by
F(J, K) = B J(J + 1) + C K 2 , (7)
with K = −J . . . + J and the rotational constants
h
B= = 0.19 cm−1 and
8π 2 c Θ⊥

h 1 1
C
= − = −0.10 cm−1 .
8π 2 c Θ Θ⊥
This shows that for benzene in the gas phase, even at the lowest temperatures,
the thermal population of the rotational excitation states would be sufficient to
average out the hyperfine structure. In measurements in the gas phase, however,
the hyperfine structure is indeed observed. The assumption that the rotation of the
benzene molecule leads to an exchange narrowing of the ESR signal must therefore
be incorrect.
19.8 Excitons are delocalised excitation states which can move due to the interac-
tions between neighbouring molecules in a crystal. Since the nitrogen nuclear spins
are not uniquely oriented, the different dipole-dipole interactions between the elec-
tronic spin and the nuclear spins are spatially averaged out, as long as the hopping
rate of the triplet excitons, and with it the exchange frequency between different envi-
ronments of the electronic spins, is high enough. There is thus a motional narrowing
of the lines which are broadened by the hyperfine structure.
If one dopes NaNO2 crystals with 1% of KNO2 , then the guest molecules act
as traps for the triplet excitons. Since the triplet excitation state is localised on the
KNO2 molecules, one can now observe the hyperfine structure.
19.9 The fine-structure constant D of pyrazine represents a quantum-mechanical
expectation value and implies an integration over the distance to the electron, r12 ,
and its component z 12 , over all of space. To carry this out, knowledge of the exact
triplet wavefunction 3 Ψ is required. In the SI unit system and with the correct units
for D, we have
2
µ0 3 2 2 3 ∗ r12 − 3z 212 3
hc D = ge µB Ψ 5
Ψ dV1 dV2 . (1)
4π 4 r12
One can, however, determine an averaged distance r of the two electrons which
are localised in the triplet state, if they are treated as localised point dipoles. We thus
require
!
r12 = r = const . (2)
140 Solutions
Since pyrazine has an essentially planar molecular structure, the z-component of

the distance r can be neglected (z 12 ≈ 0), and we obtain

µ0 3 2 2 1 3 ∗ 3
hc D = g µ Ψ Ψ dV1 dV2 . (3)
4π 4 e B r 3
Owing to the normalisation condition for the wavefunction, the integral has a
value of 1. In the point-dipole model, one thus obtains for the average distance r of
the two triplet electronic spins:
13
3µ0 ge2 µ2B
r= . (4)
16πD hc
In the present case, we can calculate
r = 1.63 Å .
19.10

3 1 2 2 2
r12 − 3z 212 3
3 ∗
D= g µ Ψ ΨdV1 dV2 (1)
4 2 e B 5
r12
2 2
3 1 2 2 3 ∗ y12 − x12
E= g µ Ψ 3
ΨdV1 dV2 . (2)
4 2 e B 5
r12
19.11 Proceed as described in the text.

19.12 a) The size of the molecules and thus the extent of their π electron system
increases in the series
benzene < naphthalene < anthracene < tetracene . (1)
Since, owing to the planarity of the molecules, the z component of the distance r12
of the triplet electrons can be neglected, the fine-structure parameter D is directly
correlated with r12 and thereby with the molecular size:
1
D∝ 3
. (2)
r12
Here, r12 can already be understood as the averaged distance. The larger the
molecule, the larger r12 , and the smaller the fine-structure constant D. We thus find
that
Dbenzene > Dnaphthalene > Danthracene > Dtetracene . (3)
The experimental values are 0.15, 0.10, 0.07, and 0.06 cm−1 .
Chapter 19 141
Fig. S.34. A schematic ADMR spectrum
The fine-structure parameter E, in contrast, contains the expectation value of

2
the difference y12 − x12
2
of the components of the distance of the electrons in the
molecular plane. Since the longer molecular axis is denoted as x, for all molecules
we find E ≤ 0. In particular, in the rotationally-symmetric benzene molecule, it
follows that E = 0. The magnitude of E increases with the x dimension of the
molecules:
|E| benzene < |E| naphthalene < |E| anthracene < |E| tetracene . (4)
The experimental values for these energies are 0, −0.013, −0.015, and −0.030 cm−1 .
b) Corresponding to the results in (a), one attributes a D value of 0.15 cm−1 to
benzene, while tetracene has the other extreme value of 0.06 cm−1 . With the aid
of Eq. (4) from problem 19.9, one can calculate from this the averaged electron
distances in the two molecules to be
rbenzene = 2.05 Å and rtetracene = 2.79 Å .
19.13 a) in the case that irradiation with microwaves causes a redistribution of a

slowly-decaying sublevel into a rapidly-decaying sublevel of the triplet state, the
overall population of the triplet state decreases, while that of the singlet ground state
increases. The transmission of the sample, which is determined by the ground-state
population, therefore decreases in an ADMR experiment:
∆I < 0 . (1)
The ADMR spectrum yields the transmitted intensity I at a fixed optical fre-
quency as a function of the wavelength of the applied microwave radiation. One
thus observes three signals, when the microwave frequency corresponds to one of
the transition frequencies between the triplet sublevels 0, D − E and D + E:
νsignal = 2E, D − E, D + E . (2)

142 Solutions
b) The transmission of a sample is given by the Lambert-Beers law according to:
I = I0 · 10−ε S0 c d with (3)
I: transmitted intensity;
I0 : irradiation intensity;
ε: extinction coefficient;
d: optical path length;
c: sample concentration;
S0 : fraction of the molecules in the S0 state.
Irradiation with microwaves leads to a change in the ground-state population by
∆S0 ; the transmitted light intensity thus becomes
I = I0 · 10−ε (S0 +∆S0 ) c d . (4)
For the relative change in transmission, it follows that
∆I I − I
= = 10−ε ∆S0 c d − 1 = e−ε ln 10 ∆S0 c d − 1 . (5)
I I
This expression can be expanded up to the first-order term in the case of low
optical densities
OD = εc d , (6)
leading to the result:

∆I
= −ε ln 10 ∆S0 c d , (7)
I
or finally,
∆I
= −∆S0 OD ln 10 . (8)
I
Here, OD is the optical density (absorption) of the sample without microwave
irradiation.
c) With the values given for Rb. sphaeroides GA, one can compute the change in
the ground-state population, finding
∆S0 = + 7.6 · 10−4 .
The fraction of molecules in the S0 ground state, relative to the overall number
of molecules, thus increases by 0.076%. This shows that ADMR is an extremely
sensitive method of measurement, which however also requires a correspondingly
sensitive detection apparatus.
Chapter 20 143
Chapter 20
20.1 a) The interaction energy E WW between the two dipoles µD and µA at a

distance of r = rA − rD is
(µD · µA ) r 2 − 3 (r · µD ) (r · µA )
E WW = const ·
r5
µD µA
= const · κ (1)
r3
with
κ = cos φDA − 3 cos φD cos φA (2)
(φDA is the angle between the dipole orientations; φD , φA is the angle between the
dipoles and the vector connecting them). The transfer rate kET is proportional to the
square of the interaction energy:
µ2D µ2A
kET ∝ E WW
2
∝ κ2 . (3)
r6
If one identifies µ2D and µ2A with the oscillator strengths of the transitions of the
donor and the acceptor, then one finally arrives at a qualitative relation:

2 1
kET ∝ κ 6 f D (ν) εA (ν) , (4)
r
which contains the overlap integral of the donor fluorescence and the acceptor
absorption, the orientation factor κ 2 , and the distance term 1/r 6 . (In contrast, for the
exchange interaction, an exponential distance dependence is obtained.)
b) Inserting the data into the formula given in the problem text, we can calculate
the Förster radius of fluorescein, obtaining
r0 = 50.2 Å
c) At low concentrations, c # c0 , the yield φ for Förster transfer is proportional
to the probability that an acceptor molecule is located at a distance r ≤ r0 from the
donor and is therefore within the volume 4π 3
3 r0 around the position of the donor. It
is thus proportional to the concentration c of the molecules:
φ ∝ c for c # c0 . (5)
For the case of high concentrations, c c0 , near each excited molecule there
are several acceptor molecules within the critical radius r0 . The yield φ is then
proportional to the rate kET of energy transfer to the next neighbours and thus to the
square of the concentration:
1
φ ∝ kET ∝ ∝ c2 for c c0 . (6)
r6
144 Solutions
20.2 The distance dependence of the energy transfer rate kET is given by
1
kET ∝ , (1)
r6
so that for the (critical) Förster radius and an arbitrary distance r1 , we can write the
relation
1
k1 6
k0 r06 = k1 r16 ⇐⇒ r0 = r1 . (2)
k0
In the case of twelve chain links, we find for the chain length and thus the distance
r1 between the donor and the acceptor:
r1 = 46.1 Å ,
and from it, using (2), the Förster radius
r0 = 37.7 Å .
This corresponds to a chain with ca. n = 9 links.

20.3 a) With the techniques of Second Harmonic Generation (SHG) and ellip-
sometry, one measures different physical properties of the system consisting of the
substrate and the overlayer.
The non-linear optical process of frequency doubling (SHG) takes place on the
gold surface. The decrease in SHG intensity during layer formation is due to a change
in the electronic structure of the surface and is thus a measure of the coverage with
sulphur atoms.
In contrast, ellipsometry, i. e. the polarisation-dependent measurement of the
reflection coefficient of the interface, is sensitive to structural changes in the layer
itself. From the longer time for layer formation measured with this technique, one can
derive the fact that following the relatively rapid process of chemisorption, a slower
process of reorientation and restructuring of the molecules occurs, until finally all
of the alkyl chains are stretched and oriented parallel to each other.
b) The parallel orientation of the alkyl chains is stabilised by lateral intermolec-
ular van der Waals forces. These increase with increasing length of the molecules.
In the solution, the molecules are often bent (cis-conformations) and are nearly
spherical in overall shape. The loss of entropy on stretching out the molecules is
compensated by the energy gain from their chemisorption and orientation.
c) The depth of the holes (2.5 Å) just corresponds to the lattice constant of the
substrate [Au(111) surface] and can thus be identified as a defect in the uppermost
atomic layer of the gold substrate. The cause of this is corrosion of the surface,
i. e. removal of gold atoms into the solution. There, they react with the alkane-thiols
to give gold-thiolate compounds.
Chapter 22 145
Chapter 22
22.1 One can, at least theoretically, make use of the entire inhomogeneous line-
width Γ for photochemical hole-burning. The width of the hole corresponds to twice
the homogeneous linewidth, which is due only to the lifetime limit:
2
γhole = 2 γh = . (1)
2π τ
With Γ = 700 cm−1 and γhole = 0.55 cm−1 , we calculate the maximum num-
ber n max of the holes to be burned to be
Γ
n max = = 1273 . (2)
γhole
The information density D of the optical data-storage medium can thus be
increased by a factor of the order of 103 :
bits bits
D = 108 2
· n max ≈ 1011 . (3)
cm cm2
We mention a few of the many practical problems which crop up in trying to
construct this type of optical storage medium:
– Cooling of the material to a low temperature, (4 K).
– Tuning of the laser to the individual address bits, both spatially and in its fre-
quency.
– The velocity of the photochemical reactions limits the writing speed.
– A satisfactory signal/noise ratio for reading requires a minimum readout time.
– Laser systems for the required wavelength range (diffraction limit λ2 !) are very
impractical. A fourth dimension could be included by making use of an holo-
graphic storage scheme.
22.2 a) Organic materials which are to be employed as data-storage media must

fulfill at least the following criteria:
– Thermal and chemical stability. There should be no thermally-induced reactions
even during longer storage periods, since they would erase the stored information.
Likewise, the material in the form used must be completely chemically inert and
must not for example by oxidised by oxygen from the air.
– Lack of material and chemical fatigue. The photochemical reaction must be
completely reversible, i. e. there must be no side or decomposition reactions.
After a cycle of forward and backward reactions, exactly the original number of
molecules must be present in the initial state. If the yield of a side reaction, which
removes molecules from the cycle, is only 0.1%, then the number of photoactive
molecules is reduced to 37% of its original value after 1000 cycles.
146 Solutions
– Sensitivity in the wavelength range of semiconductor lasers (700 – 800 nm). In

order to design miniaturised devices, in particular the irradiation apparatus must
be small. This requirement is satisfied by semiconductor lasers.
– High velocities of the photochemical reactions, in order to allow short times for
writing and erasing operations.
– Destruction-free readout of the information. On reading the data bits, i. e. essen-
tially registering an absorption spectrum, which may consist of only one point,
the back reaction is always also induced. Thus the readout process erases a por-
tion of the stored information. There are various suggestions for the solution of
this problem (two-photon processes, switching the photochromism on and off
by protonation of the molecules, etc.).
b) In the course of the photoreaction A → B, a portion of the A molecules
is converted into the form B. The absorption spectrum of the mixture represents
at every point in time an additive superposition of the spectra of both of the pure
species. At a certain wavenumber ν, the time-dependent optical density OD of the
sample is given by
OD(t) = ε A (ν) c A (t) d + ε B (ν) c B (t) d . (1)
Here, ε A (ν) and ε B (ν) are the extinction coefficients of the two species at the
wavenumber ν, and c A (t) and c B (t) are their concentrations, depending on the total
irradiation time; d is the optical path length. If only two forms, A and B, are involved
in the reaction, then due to conservation of the overall number of both molecules,
we have
c0 = c A (t) + c B (t) ⇐⇒ c B (t) = c0 − c A (t) , (2)
where c0 is the initial concentration of the starting material. >From (1), it follows
that
OD(t) = [ε A (ν) − ε B (ν)] c A (t) d + ε B (ν) c0 d . (3)
At the wavenumbers ν at which the two absorption spectra of the pure species
overlap, i. e. for ε A (ν) = ε B (ν), OD(t) remains constant in time:
OD(t) = const(t) = ε B (ν) c0 d for ε A (ν) = ε B (ν) . (4)
At these wavenumbers ν, all the spectra thus overlap –isosbestic points exist.
If only two species with differing absorption behaviour take part in the photore-
action, then the existence of isosbestic points is certain. >From the fact that none are
present in the case shown, one must conclude that in the course of the photoreaction,
still (at least) a third form is generated. If its spectrum differs from that of the other
two forms, then one will not observe any isosbestic points.
In the case of the fulgides, an undesirable side reaction in the form of a rotation
of the aromatic rings around the exocyclic double bond occurs. It can, however, be
suppressed by a suitable substitution of the molecule, so that the fulgide becomes
truly bistable.
http://www.springer.com/978-3-540-40792-8

Solutions:, The Exponential

Uploaded by

Copyright:

Available Formats

Solutions:, The Exponential

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

Solutions:, The Exponential

Uploaded by

Copyright:

Available Formats

Solutions

2.1 In a diatomic molecule, there exists an interatomic distance r0 at which the

which in the case of hexadecachloro copper phthalocyanine corresponds to overall

AM = 2.16 · 10−19 m2 = 21.6 Å2 .

In equilibrium, π = p, we thus have

Calculating the mass concentration c from the molar particle concentration c is

The slope b yields the virial coefficient

α = 5.57 · 107 cm3 /mole

and thus via α = 1

vp = 1.85 · 10−16 cm3 .

The volume occupied by a spherical molecule of radius r is a sphere with a

r = 1.77 · 10−6 cm = 177 Å .

In an ideal gas, the velocity of sound cS is given by the equation

The mean value P of the macroscopic polarisation

After integration, one obtains the expression

b) The orientation of a dipole precisely along the field direction corresponds to

cos θ = L(x) = 1 ; (10)

compare (9). With the limiting value

lim L(x) = 1 (11)

this corresponds for x = pE/kT and T = 293 K to

P = N p cos θ = 0.5 N p (14)

and thus with (9) and (13)

E = 9.8 · 108 V/m .

d) With E = 107 V/m and T = 4.2 K, one obtains

and thus an orientation of ca. 36%.

3.3 a) From α/4π ε0 = 10.3 · 10−30 m3 , it follows that:

3.4 The behaviour of a permanent magnetic moment µp in a magnetic field B

pOH = 1.51 D . (2)

If the hydrogen peroxide molecule, H−O−O−H, is projected onto the plane

3.8 The density of water vapour at a pressure of 1 bar is

χDampf (1 bar) = 77.5 · 8.0 · 10−4 = 6.2 · 10−2 .

This means that at 50 mbar

χDampf (50 mbar) = 3.1 · 10−3

εDampf (50 mbar) = 1 + 3.1 · 10−3

3.10 From the Debye equation for the molar polarisability Pm

3.11 The magnetic polarisability β is a microscopic quantity which relates the

(J = macroscopic magnetisation). From this it follows that

Fig. S.2. Dielectric function of the NaCl ionic crystal

ΩLO = 4.9 · 1013 s−1

Fig. S.3. The “square-well” potential V(x)

lim ψ (3) (x) = 0 , therefore A(3) = 0 . (6)

Continuity condition at the boundary:

κ 2 sin(2ka) + 2kκ cos(2ka) − k2 sin(2ka) = 0 . (12)

We find two types of solutions:

α) symmetric, κ = k tan(ka) , (13)

β) antisymmetric, κ = −k cot(ka) , (16)

The evaluation of the integrals is elementary. For the symmetric wavefunction,

Fig. S.5. As in Fig. S.4, but for the

and for the antisymmetric wavefunction,

For finite values of k, we thus have α) tan(ka) = ∞, i. e.

The energy values are:

for n even, m odd (or vice versa).

for n and m even or n and m odd.

4.2 Ground-state wavefunction

x = r cos φ sin θ , (2)

Expectation values in real space:

ψ100 | x |ψ100  = ψ100 | y |ψ100 

This follows from symmetry considerations.

ψ100 | x |ψ100 = ψ100 | y |ψ100

Φi | Φ j = δij . (3)