Peter Atkins Julio de Paula Ron Friedman Physical Chemistry Quanta (0460-0510)

Download as pdf or txt
Download as pdf or txt
You are on page 1of 51

46 Decay of excited states 

437

Table 46.1 Characteristics of laser radiation and their chemical I


applications
A
Characteristic Advantage Application

Laser action
High power Multiphoton process Spectroscopy
Low detector noise Improved sensitivity Pump

High scattering intensity Raman spectroscopy


(Topics 40, 42 − 44)
Monochromatic High resolution Spectroscopy
X
State selection Photochemical studies
(Topic 93)
Figure 46.9 The transitions involved in one kind of three-level
Collimated beam Long path lengths Improved sensitivity laser. The pumping pulse populates the intermediate state I,
Forward scattering Raman spectroscopy which in turn populates the laser state A. The laser transition is
observable (Topics 40, 42 − 44) the stimulated emission A → X.
Coherent Interference between Coherent anti-Stokes
separate beams Raman spectroscopy*
involved overall, this arrangement leads to a three-level laser.
Pulsed Precise timing of Fast reactions (Topic 93)
excitation In practice, I consists of many states, all of which can convert
Relaxation (Topic 84) to the upper of the two laser states A. The I ← X transition is
Energy transfer (Topic 93) stimulated with an intense flash of light in the process called
*See our other Physical chemistry (2014). pumping. The pumping is often achieved with an electric dis-
charge through xenon or with the light of another laser. The
conversion of I to A should be rapid, and the laser transitions
radiation of the same frequency: the more photons that are pre- from A to X should be relatively slow.
sent, the greater the probability of the emission. The essential The disadvantage of the three-level arrangement is that it
feature of laser action is positive feedback: the more photons is difficult to achieve population inversion, because so many
present of the appropriate frequency, the more photons of that ground-state molecules must be converted to the excited state
frequency that will be stimulated to form. by the pumping action. The arrangement adopted in a four-
Laser radiation has a number of striking characteristics level laser simplifies this task by having the laser transition
(Table 46.1). Each of them (sometimes in combination with the terminate in a state A′ other than the ground state (Fig. 46.10).
others) opens up interesting opportunities in physical chem- Because A′ is unpopulated initially, any population in A cor-
istry. Raman spectroscopy has flourished on account of the responds to a population inversion and we can expect laser
high-intensity monochromatic radiation available from lasers action if A is sufficiently metastable. Moreover, this popula-
(Topics 40 and 42 − 44), and the ultrashort pulses that lasers can tion inversion can be maintained if the A′ → X transitions are
generate make possible the study of light-initiated reactions on rapid, for these transitions will deplete any population in A′
timescales of femtoseconds and even attoseconds. that stems from the laser transition, and keep the state A′ rela-
tively empty.
(a) Population inversion
One requirement of laser action is the existence of a metasta- I
ble excited state, an excited state with a long enough lifetime
for it to participate in stimulated emission. Another require- A
Laser action

ment is the existence of a greater population in the metastable


state than in the lower state where the transition terminates, for
Pump
then there will be a net emission of radiation. Because at ther-
mal equilibrium the opposite is true, it is necessary to achieve a
A’
population inversion in which there are more molecules in the
upper state than in the lower.
X
One way of achieving population inversion is illustrated
in Fig. 46.9. The molecule is excited to an intermediate state I, Figure 46.10 The transitions involved in a four-level laser.
which then gives up some of its energy nonradiatively and Because the laser transition terminates in an excited state (A′),
changes into a lower state A; the laser transition is the return the population inversion between A and A′ is much easier to
of A to the ground state X. Because three energy levels are achieve.

Atkins09819.indb 437 9/11/2013 12:03:40 PM


438 9 Molecular spectroscopy

Brief illustration 46.3 essentially a region between two mirrors, which reflect the light
Simple lasers
back and forth. This arrangement can be regarded as a version
The ruby laser is an example of a three-level laser (Fig. 46.11). of the particle in a box, with the particle now being a photon.
Ruby is Al2O3 containing a small proportion of Cr3+ ions. The As in the treatment of a particle in a box (Topic 9), the only
lower level of the laser transition is the 4A2 ground state of the wavelengths that can be sustained satisfy
Cr3+ ion. The process of pumping a majority of the Cr3+ ions
into the 4T 2 and 4T1 excited states is followed by a radiation- n × 12 λ = L Resonant modes (46.1)
less transition to the 2E excited state. The laser transition is
2E → 4 A , and gives rise to red 694 nm radiation.
2 where n is an integer and L is the length of the cavity. That is,
only an integral number of half-wavelengths fit into the cavity;
4
T1 all other waves undergo destructive interference with them-
4
T2 Fast
selves. In addition, not all wavelengths that can be sustained by
2
E the cavity are amplified by the laser medium (many fall outside
the range of frequencies of the laser transitions), so only a few
Laser action

contribute to the laser radiation. These wavelengths are the res-


Pump onant modes of the laser.
694.3 nm

Brief illustration 46.4 Resonant modes


4
A2
It follows from eqn 46.1 that the frequencies of the resonant
modes are ␯ = c / λ = (c / 2L) × n. For a laser cavity of length
Figure 46.11 The transitions involved in a ruby laser.
30.0 cm, the allowed frequencies are

The neodymium laser is an example of a four–level laser  c 

(Fig. 46.12). In one form it consists of Nd 3+ ions at low con- 2.998 ×108 ms −1
␯= × n = (5.00 ×108 s −1 ) × n
centration in yttrium aluminium garnet (YAG, specifically 2 × (0.300 m)

Y3Al5O12), and is then known as a Nd:YAG laser. A neodym- L

ium laser operates at a number of wavelengths in the infrared, = (500MHz) ×n


the band at 1064 nm being most common.
with n = 1, 2, …, and therefore ν = 500 MHz, 1000 MHz, ….
Self-test 46.4 Consider a laser cavity of length 1.0 m. What is
4
the frequency difference between successive resonant modes?
F
Answer: 150 MHz
Laser action

1.06 μm
Pump
Photons with the correct wavelength for the resonant modes
4
I of the cavity and the correct frequency to stimulate the laser
Thermal
decay transition are highly amplified. One photon might be generated
spontaneously and travel through the medium. It stimulates
Figure 46.12 The transitions involved in a neodymium laser. the emission of another photon, which in turn stimulates more
(Fig. 46.13). The cascade of energy builds up rapidly, and soon
Self-test 46.3 In the arrangement discussed above, does a ruby the cavity is an intense reservoir of radiation at all the resonant
laser more easily generate pulses of light or a continuous beam modes it can sustain. Some of this radiation can be withdrawn
of light? if one of the mirrors is partially transmitting.
Answer: Pulses of light The resonant modes of the cavity have various natural char-
acteristics, and to some extent may be selected. Only pho-
tons that are travelling strictly parallel to the axis of the cavity
undergo more than a couple of reflections, so only they are
(b) Cavity and mode characteristics
amplified, all others simply vanishing into the surround-
The laser medium is confined to a cavity that ensures that only ings. Hence, laser light generally forms a beam with very low
certain photons of a particular frequency, direction of travel, divergence. It may also be polarized, with its electric vector in
and state of polarization are generated abundantly. The cavity is a particular plane (or in some other state of polarization), by

www.ebook3000.com

Atkins09819.indb 438 9/11/2013 12:03:48 PM


46 Decay of excited states  439

(a)
Thermal (c) Pulsed lasers
equilibrium
A laser can generate radiation for as long as the population
Pump inversion is maintained. A laser can operate continuously
Population when heat is easily dissipated, for then the population of the
(b)
inversion upper level can be replenished by pumping. When overheating
is a problem, the laser can be operated only in pulses, perhaps
Laser
of microsecond or millisecond duration, so that the medium
(c) action has a chance to cool or the lower state discard its population.
However, it is sometimes desirable to have pulses of radiation
rather than a continuous output, with a lot of power concen-
Figure 46.13 A schematic illustration of the steps leading to trated into a brief pulse. One way of achieving pulses is by
laser action. (a) The Boltzmann population of states, with more Q-switching, the modification of the resonance characteristics
atoms in the ground state. (b) When the initial state absorbs, of the laser cavity. The name comes from the ‘Q-factor’ used
the populations are inverted (the atoms are pumped to the as a measure of the quality of a resonance cavity in microwave
excited state). (c) A cascade of radiation then occurs, as one
engineering.
emitted photon stimulates another atom to emit, and so on.
The radiation is coherent (phases in step).
Example 46.1 Relating the power and energy of a laser
including a polarizing filter into the cavity or by making use of A laser rated at 0.10 J can generate radiation in 3.0 ns pulses at
polarized transitions in a solid medium. a pulse repetition rate of 10 Hz. Assuming that the pulses are
Laser radiation is coherent in the sense that the electromag- rectangular, calculate the peak power output and the average
netic waves are all in step. In spatial coherence the waves are power output of this laser.
in step across the cross-section of the beam emerging from
Method The power output is the energy released in an inter-
the cavity. In temporal coherence the waves remain in step
val divided by the duration of the interval, and is expressed
along the beam. The former is normally expressed in terms of a
in watts (1 W = 1 J s −1). To calculate the peak power output,
coherence length, lC, the distance over which the waves remain P peak, we divide the energy released during the pulse by the
coherent, and is related to the range of wavelengths, Δλ, present duration of the pulse. The average power output, Paverage, is the
in the beam: total energy released by a large number of pulses divided by
the duration of the time interval over which the total energy
λ2
lC = Coherence length (46.2) was measured. So, the average power is simply the energy
2Δ λ released by one pulse multiplied by the pulse repetition rate.

When many wavelengths are present, and Δλ is large, the waves Answer From the data,
get out of step in a short distance and the coherence length is
0.10 J
small. Ppeak = = 3.3 × 107 J s −1 = 33 MJ s −1 = 33 MW
3.0 × 10−9 s

The pulse repetition rate is 10 Hz, so ten pulses are emitted by


Brief illustration 46.5 Coherence length the laser in every second of operation. It follows that the aver-
age power output is
A typical light bulb gives out light with a coherence length of
only about 400 nm. By contrast, a He–Ne laser with λ = 633 nm Paverage = 0.10 J × 10 s −1 = 1.0 J s −1 =1.0 W
and Δλ = 2.0 pm has a coherence length of
 λ 

2
The peak power is much higher than the average power
(633 nm)2 because this laser emits light for only 30 ns during each second
lC = = 1.0 × 108 nm = 0.10 m =10 cm
2 × (0.0020 nm) of operation.

Δλ
Self-test 46.6 Calculate the peak power and average power
Self-test 46.5 What is the condition that would lead to an infi- output of a laser with a pulse energy of 2.0 mJ, a pulse duration
nite coherence length? of 30 ps, and a pulse repetition rate of 38 MHz.
Answer: A perfectly monochromatic beam, or Δλ = 0 Answer: Ppeak = 67 MW, Paverage = 76 kW

Atkins09819.indb 439 9/11/2013 12:03:56 PM


440 9 Molecular spectroscopy

Pump 1 ps
1 ns

Cavity nonresonant

(a)
Switch

Cavity resonant

(b) Pulse Time, t

Figure 46.14 The principle of Q-switching. (a) The excited Figure 46.15 The output of a mode-locked laser consists of a
state is populated while the cavity is nonresonant. (b) Then stream of very narrow pulses (here 1 ps in duration) separated
the resonance characteristics are suddenly restored, and the by an interval equal to the time it takes for light to make a
stimulated emission emerges in a giant pulse. round trip inside the cavity (here 1 ns).

The aim of Q-switching is to achieve a healthy population


inversion in the absence of the resonant cavity, then to plunge
the population-inverted medium into a cavity and hence to
N=5
obtain a sudden pulse of radiation. The switching may be
achieved by impairing the resonance characteristics of the cav- N = 10

Intensity, I
ity in some way while the pumping pulse is active and then
suddenly improving them (Fig. 46.14). One technique is to use
the ability of some crystals, such as those of potassium dihy-
drogenphosphate (KH2PO4), to change their optical properties
when an electrical potential difference is applied. Switching the
potential on and off can store and then release energy in a laser 0 1 2 3 4 5
cavity, resulting in an intense pulse of stimulated emission. Time, ct/2L
The technique of mode locking can produce pulses of pico-
second duration and less. A laser radiates at a number of dif- Figure 46.16 The structure of the pulses generated by a mode-
locked laser for N = 5 and 10. Note that the peaks become
ferent frequencies, depending on the precise details of the
sharper as N increases. In each case the pulses have been
resonance characteristics of the cavity and in particular on the
adjusted to have the same maximum amplitude.
number of half-wavelengths of radiation that can be trapped
between the mirrors (the cavity modes). The resonant modes
differ in frequency by multiples of c/2L (Brief illustration 46.4).
Normally, these modes have random phases relative to each Justification 46.1 The origin of mode locking
other. However, it is possible to lock their phases together. As The general expression for a (complex) wave of amplitude E0
we show in the following Justification, interference then occurs and frequency ω is E0eiωt . Therefore, each wave that can be
to give a series of sharp peaks, and the energy of the laser is supported by a cavity of length L has the form
obtained in short bursts (Fig. 46.15). More specifically, the
intensity, I, of the radiation varies with time as En (t ) = E0e2 πi(␯+nc/2 L )t

sin2 (N πct / 2L) where ν is the lowest frequency. A wave formed by superim-
I (t ) ∝E02 Mode-locked laser output (46.3)
posing N modes with n = 0, 1, …, N – 1 has the form
sin2 (πct / 2L)
N −1 N −1
where E0 is the amplitude of the electromagnetic wave describ-
ing the laser beam and N is the number of locked modes. This
E (t ) = ∑n=0
En (t ) = E0e2 πit ∑en=0
iπnct /L

function is shown in Fig. 46.16. We see that it is a series of peaks


with maxima separated by t = 2L/c, the round-trip transit time The sum simplifies to:
of the light in the cavity, and that the peaks become sharper as N −1
N is increased. In a laser with a cavity of length 30 cm, the peaks
are separated by 2 ns. If 1000 modes contribute, the width of the
S= ∑e
n=0
iπnct /L
=1 + eiπct /L + e2iπct /L + …+ e( N −1)iπct /L

pulses is 4 ps.

www.ebook3000.com

Atkins09819.indb 440 9/11/2013 12:04:07 PM


46 Decay of excited states  441

Kerr medium
At this point we recognize that the sum is a geometrical series:
Aperture
e Nx − 1
1 + e x + e2 x + …+ e( N −1) x = Laser
e x −1
beam

with x = eiπct/L , so

e Ni π ct /L − 1
S=
ei π ct /L −1
Figure 46.17 An illustration of the optical Kerr effect. An
The term on the right can be written intense laser beam is focused inside a Kerr medium and passes
through a small aperture in the laser cavity. This effect may be
e Ni π ct /L −1 e Ni π ct /2 L − e − Ni π ct /2 L ( N −1)i π ct /2 L used to mode-lock a laser, as explained in the text.
= i π ct /2 L − i π ct /2 L × e
ei π ct /L −1 e −e
when it is exposed to intense laser pulses. Because a beam of
light changes direction when it passes from a region of one
Finally, we use the relation sin x = (1/ 2i)(eix − e − ix ) to conclude
that
refractive index to a region with a different refractive index,
changes in refractive index result in the self-focusing of an
sin(N π ct / 2L) ( N −1)i π ct /2 L intense laser pulse as it travels through the Kerr medium
S= ×e
sin(π ct / 2L) (Fig. 46.17).
To bring about mode locking, a Kerr medium is included in
The intensity, I(t), of the radiation is proportional to the square
the laser cavity and next to it is a small aperture. The procedure
modulus of the total amplitude, so
makes use of the fact that the gain, the growth in intensity, of a
E =E0 S

sin2 (N πct / 2L)
frequency component of the radiation in the cavity is very sen-
I (t ) ∝E E = E02
*
sitive to amplification, and once a particular frequency begins
sin2 (πct / 2L)
to grow, it can quickly dominate. When the power inside the
which is eqn 46.3. cavity is low, a portion of the photons will be blocked by the
aperture, creating a significant loss. A spontaneous fluctuation
in intensity—a bunching of photons—may begin to turn on the
Mode locking is achieved by varying the Q-factor of the cavity optical Kerr effect and the changes in the refractive index of the
periodically at the frequency c/2L. The modulation can be pic- Kerr medium will result in a Kerr lens, which is the self-focus-
tured as the opening of a shutter in synchrony with the round- ing of the laser beam. The bunch of photons can pass through
trip travel time of the photons in the cavity, so only photons and travel to the far end of the cavity, amplifying as it goes.
making the journey in that time are amplified. The modulation The Kerr lens immediately disappears (if the medium is well
can be achieved by linking a prism in the cavity to a transducer chosen), but is re-created when the intense pulse returns from
driven by a radiofrequency source at a frequency c/2L. The the mirror at the far end. In this way, that particular bunch of
transducer sets up standing-wave vibrations in the prism and photons may grow to considerable intensity because it alone is
modulates the loss it introduces into the cavity. stimulating emission in the cavity. Sapphire is an example of a
Another mechanism for mode-locking lasers is based on Kerr medium that facilitates the mode locking of titanium sap-
the optical Kerr effect, which arises from a change in refrac- phire lasers, resulting in very short laser pulses of duration in
tive index of a well-chosen medium, the Kerr medium, the femtosecond range.

Checklist of concepts
☐ 1. Fluorescence is radiative decay between states of the ☐ 4. A Jablonski diagram is a schematic diagram of the
same multiplicity; it ceases as soon as the exciting types of nonradiative and radiative transitions that can
source is removed. occur in molecules.
☐ 2. Phosphorescence is radiative decay between states of ☐ 5. An additional fate of an electronically excited species is
different multiplicity; it persists after the exciting radia- dissociation.
tion is removed. ☐ 6. Internal conversion is a nonradiative conversion to a
☐ 3. Intersystem crossing is the nonradiative conversion to state of the same multiplicity.
a state of different multiplicity.

Atkins09819.indb 441 9/11/2013 12:04:15 PM


442 9 Molecular spectroscopy

☐ 7. Predissociation is the observation of the effects of dis- ☐ 10. The resonant modes of a laser are the wavelengths of
sociation before the dissociation limit is reached. radiation sustained inside a laser cavity.
☐ 8. Laser action is the stimulated emission of coher- ☐ 11. Laser pulses are generated by the techniques of
ent radiation between states related by a population Q-switching and mode locking.
inversion.
☐ 9. A population inversion is a condition in which the
population of an upper state is greater than that of a rel-
evant lower state.

Checklist of equations
Property Equation Comment Equation number

Resonant modes n × 12 λ = L Laser cavity of length L 46.1

Coherence length lC = λ 2 / 2Δλ 46.2

Mode-locked laser output I (t ) ∝E02 {sin2 (N πct / 2L)/ sin2 (πct / 2L)} N locked modes 46.3

www.ebook3000.com

Atkins09819.indb 442 9/11/2013 12:04:19 PM


Exercises and problems  443

Focus 9 on Molecular spectroscopy


Note: The masses of nuclides are listed in Table 0.2 of the Resource section.

Topic 40 General features


Discussion questions
40.1 Distinguish between the basic experimental arrangements commonly ~ = ε {1 – κ(ν~ – ν~ )2}
εmax ε(ν)
used for absorption, emission, and Raman spectroscopy. max max

Molar absorption coefficient, ε


40.2 Describe the physical origins of linewidths in absorption and emission
spectra. Do you expect the same contributions for species in condensed and
gas phases?

Exercises
40.1(a) The molar absorption coefficient of a substance dissolved in hexane ~
νmax
is known to be 723 dm3 mol−1 cm−1 at 260 nm. Calculate the percentage
reduction in intensity when light of that wavelength passes through 2.50 mm
Wavenumber, ~
ν
of a solution of concentration 4.25 mmol dm−3.
40.1(b) The molar absorption coefficient of a substance dissolved in hexane
Figure F9.1 A model parabolic absorption lineshape.
is known to be 227 dm3 mol−1 cm−1 at 290 nm. Calculate the percentage
reduction in intensity when light of that wavelength passes through 2.00 mm
of a solution of concentration 2.52 mmol dm−3. 40.5(a) The following data were obtained for the absorption by Br2 in carbon
tetrachloride using a 2.0 mm cell. Calculate the molar absorption coefficient of
40.2(a) A solution of an unknown component of a biological sample when
bromine at the wavelength employed:
placed in an absorption cell of path length 1.00 cm transmits 18.1 per cent
of light of 320 nm incident upon it. If the concentration of the component is
[Br2]/(mol dm−3) 0.0010 0.0050 0.0100 0.0500
0.139 mmol dm−3, what is the molar absorption coefficient?
40.2(b) When light of wavelength 400 nm passes through 2.5 mm of a T/(%) 81.4 35.6 12.7 3.0 × 10−3
solution of an absorbing substance at a concentration 0.717 mmol dm−3,
the transmission is 61.5 per cent. Calculate the molar absorption 40.5(b) The following data were obtained for the absorption by a dye dissolved
coefficient of the solute at this wavelength and express the answer in in methylbenzene using a 2.50 mm cell. Calculate the molar absorption
cm2 mol−1. coefficient of the dye at the wavelength employed:
40.3(a) The molar absorption coefficient of a solute at 540 nm is
[dye]/(mol dm−3) 0.0010 0.0050 0.0100 0.0500
386 dm3 mol−1 cm−1. When light of that wavelength passes through a 5.00 mm
cell containing a solution of the solute, 38.5 per cent of the light was absorbed. T/(%) 68 18 3.7 1.03 × 10−5
What is the concentration of the solution?
40.3(b) The molar absorption coefficient of a solute at 440 nm is 40.6(a) A 2.0 mm cell was filled with a solution of benzene in a non-absorbing
423 dm3 mol−1 cm−1. When light of that wavelength passes through a 6.50 mm solvent. The concentration of the benzene was 0.010 mol dm−3 and the
cell containing a solution of the solute, 48.3 per cent of the light was absorbed. wavelength of the radiation was 256 nm (where there is a maximum in
What is the concentration of the solution? the absorption). Calculate the molar absorption coefficient of benzene at
this wavelength given that the transmission was 48 per cent. What will the
40.4(a) The absorption associated with a particular transition begins at
transmittance be in a 4.0 mm cell at the same wavelength?
220 nm, peaks sharply at 270 nm, and ends at 300 nm. The maximum value
40.6(b) A 5.00 mm cell was filled with a solution of a dye. The concentration of
of the molar absorption coefficient is 2.21 × 104 dm3 mol−1 cm−1. Estimate
the dye was 18.5 mmol dm−3. Calculate the molar absorption coefficient of the
the integrated absorption coefficient of the transition assuming a triangular
dye at this wavelength given that the transmission was 29 per cent. What will
lineshape.
the transmittance be in a 2.50 mm cell at the same wavelength?
40.4(b) The absorption associated with a certain transition begins at 156 nm,
peaks sharply at 210 nm, and ends at 275 nm. The maximum value of the 40.7(a) A swimmer enters a gloomier world (in one sense) on diving to greater
molar absorption coefficient is 3.35 × 104 dm3 mol−1 cm−1. Estimate the depths. Given that the mean molar absorption coefficient of seawater in the
integrated absorption coefficient of the transition assuming an inverted visible region is 6.2 × 10−3 dm3 mol−1 cm−1, calculate the depth at which a diver will
parabolic lineshape (Fig. F9.1). experience (a) half the surface intensity of light, (b) one-tenth the surface intensity.

Atkins09819.indb 443 9/11/2013 12:04:31 PM


444 9 Molecular spectroscopy

40.7(b) Given that the maximum molar absorption coefficient of a molecule 40.9(b) Estimate the lifetime of a state that gives rise to a line of width
containing a carbonyl group is 30 dm3 mol−1 cm−1 near 280 nm, calculate (a) 200 MHz, (b) 2.45 cm−1.
the thickness of a sample that will result in (a) half the initial intensity of
40.10(a) A molecule in a liquid undergoes about 1.0 × 1013 collisions in
radiation, (b) one-tenth the initial intensity.
each second. Suppose that (i) every collision is effective in deactivating the
40.8(a) What is the Doppler-shifted wavelength of a red (680 nm) traffic light molecule vibrationally and (ii) that one collision in 100 is effective. Calculate
approached at 60 km h−1? the width (in cm−1) of vibrational transitions in the molecule.
40.8(b) At what speed of approach would a red (680 nm) traffic light appear 40.10(b) A molecule in a gas undergoes about 1.0 × 109 collisions in each
green (530 nm)? second. Suppose that (i) every collision is effective in deactivating the
molecule rotationally and (ii) that one collision in 10 is effective. Calculate the
40.9(a) Estimate the lifetime of a state that gives rise to a line of width (a)
width (in hertz) of rotational transitions in the molecule.
0.20 cm−1, (b) 2.0 cm−1.

Problems
40.1 Refer to Fig. 40.4 describing a Michelson interferometer. The mirror M1 40.8‡ Ozone absorbs ultraviolet radiation in a part of the electromagnetic
moves in finite distance increments, so the path difference p is also incremented spectrum energetic enough to disrupt DNA in biological organisms and
in finite steps. Explore the effect of increasing the step size on the shape of the that is absorbed by no other abundant atmospheric constituent. This
interferogram for a monochromatic beam of wavenumber ␯ and intensity I0. spectral range, denoted UV-B, spans the wavelengths of about 290 nm to
That is, draw plots of I(p)/I0 against ␯ p each with a different number of data 320 nm. The molar extinction coefficient of ozone over this range is given in
points spanning the same total distance path taken by the movable mirror M1. the table below (DeMore, et al., Chemical kinetics and photochemical data for
use in stratospheric modeling: Evaluation Number 11, JPL Publication 94–26
40.2 Using mathematical software, elaborate on the results of Example 40.1
(1994)).
by (a) exploring the effect of varying the wavenumbers and intensities of the
three components of the radiation on the shape of the interferogram, and (b)
calculating the Fourier transforms of the functions you generated in part (a). λ/nm 292.0 296.3 300.8 305.4 310.1 315.0 320.0

40.3 The flux of visible photons reaching Earth from the North Star is about ε/(dm3 mol−1 cm−1) 1512 865 477 257 135.9 69.5 34.5
4 × 103 mm−2 s−1. Of these photons, 30 per cent are absorbed or scattered
Compute the integrated absorption coefficient of ozone over the wavelength
by the atmosphere and 25 per cent of the surviving photons are scattered
range 290–320 nm. (Hint: ε(␯ ) can be fitted to an exponential function quite
by the surface of the cornea of the eye. A further 9 per cent are absorbed
well.)
inside the cornea. The area of the pupil at night is about 40 mm2 and the
response time of the eye is about 0.1 s. Of the photons passing through the 40.9 In many cases it is possible to assume that an absorption band has a
Gaussian lineshape (one proportional to e −x ) centred on the band maximum.
2
pupil, about 43 per cent are absorbed in the ocular medium. How many
photons from the North Star are focused onto the retina in 0.1 s? For a Assume such a lineshape, and show that A = ∫ ε (␯ )d␯ ≈ 1.0645ε max Δ␯ 1/2 ,
continuation of this story, see R.W. Rodieck, The first steps in seeing, Sinauer where Δ␯ 1/2 is the width at half-height. The absorption spectrum of azoethane
(1998). (CH3CH2N2) between 24 000 cm−1 and 34 000 cm−1 is shown in Fig. F9.2.
First, estimate A for the band by assuming that it is Gaussian. Then use
40.4 A Dubosq colorimeter consists of a cell of fixed path length and a
mathematical software to fit a polynomial to the absorption band (or a
cell of variable path length. By adjusting the length of the latter until the
transmission through the two cells is the same, the concentration of the
second solution can be inferred from that of the former. Suppose that a plant 10
dye of concentration 25 μg dm−3 is added to the fixed cell, the length of which
is 1.55 cm. Then a solution of the same dye, but of unknown concentration, is
added to the second cell. It is found that the same transmittance is obtained 8
ε/(dm3 mol–1 cm–1)

when the length of the second cell is adjusted to 1.18 cm. What is the
concentration of the second solution? 6
40.5 The Beer–Lambert law is derived on the basis that the concentration of
absorbing species is uniform. Suppose, instead, that the concentration falls
4
exponentially as [J] = [J]0e−x/λ. Derive an expression for the variation of I with
sample length; suppose that L ≫ λ.
2
40.6 It is common to make measurements of absorbance at two wavelengths
and use them to find the individual concentrations of two components A and
B in a mixture. Show that the molar concentrations of A and B are 0
22 26 30 34
~/(103 cm–1)
Wavenumber, ν
ε B2 A1 − ε B1 A2 ε A1 A2 − ε A2 A1
[A] = [B] =
(ε A1ε B2 − ε A2ε B1 )L (ε A1ε B2 − ε A2ε B1 )L
Figure F9.2 The absorption spectrum of azoethane.
where A1 and A2 are absorbances of the mixture at wavelengths λ1 and λ2, and
the molar extinction coefficients of A (and B) at these wavelengths are εA1 and Gaussian), and integrate the result analytically.
εA2 (and εB1 and εB2). 40.10‡ Wachewsky, et al. (J. Phys. Chem. 100, 11559 (1996)) examined
the UV absorption spectrum of CH3I, a species of interest in connection
40.7 When pyridine is added to a solution of iodine in carbon tetrachloride
with stratospheric ozone chemistry. They found the integrated absorption
the 520 nm band of absorption shifts toward 450 nm. However, the absorbance
coefficient to be dependent on temperature and pressure to an extent
of the solution at 490 nm remains constant: this feature is called an isosbestic
inconsistent with internal structural changes in isolated CH3I molecules;
point. Show that an isosbestic point should occur when two absorbing species
are in equilibrium. ‡ These problems were supplied by Charles Trapp and Carmen Giunta.

www.ebook3000.com

Atkins09819.indb 444 9/11/2013 12:04:43 PM


Exercises and problems  445

they explained the changes as due to dimerization of a substantial fraction Detector


of the CH3I, a process which would naturally be pressure- and temperature-
dependent. (a) Compute the integrated absorption coefficient over a
triangular lineshape in the range 31 250 to 34 483 cm−1 and a maximal molar Incident
absorption coefficient of 150 dm3 mol−1 cm−1 at 31 250 cm−1. (b) Suppose ray
1 per cent of the CH3I units in a sample at 2.4 Torr and 373 K exist as dimers. θ
Compute the absorbance expected at 31 250 cm−1 in a sample cell of length
12.0 cm. (c) Suppose 18 per cent of the CH3I units in a sample at 100 Torr and Monochromatic
373 K exists as dimers. Compute the absorbance expected at 31 250 cm−1 in a source
sample cell of length 12.0 cm; compute the molar absorption coefficient which Scattering intensity, I Sample
would be inferred from this absorbance if dimerization was not considered.
40.11 Laser light scattering is a technique that uses the fact that the intensity Figure F9.3 A typical experimental arrangement of a laser light scattering
of light scattered—by Rayleigh scattering—by a particle is proportional to measurement.
the molar mass of the particle and to λ−4, so shorter-wavelength radiation is
scattered more intensely than longer wavelengths. Consider the experimental expression for the collision-limited lifetime of an excited state assuming that
arrangement shown in Fig. F9.3 for the measurement of light scattering every collision is effective. Estimate the width of rotational transition in HCl
from solutions of macromolecules. Typically, the sample is irradiated with (σ = 0.30 nm2) at 25 °C and 1.0 atm. To what value must the pressure of the gas
monochromatic light from a laser. The intensity of scattered light is then be reduced in order to ensure that collision broadening is less important than
measured as a function of the angle θ that the line of propagation of the Doppler broadening?
laser beam makes with a line from the sample to the detector. For dilute 40.13 The spectrum of a star is used to measure its radial velocity with respect
solutions of a spherical macromolecule with a diameter much smaller than to the Sun, the component of the star’s velocity vector that is parallel to a
the wavelength of incident radiation, the intensity, Iθ, of light scattered by a vector connecting the star’s centre to the centre of the Sun. The measurement
sample of mass concentration cM (units: kg m−3) is given by relies on the Doppler effect. When a star emitting electromagnetic radiation of
frequency ν moves with a speed s relative to an observer, the observer detects
I0 1 16π2 R2 ⎛ I0 1 ⎞ radiation of frequency νreceding = νf or νapproaching = ν/f, where f = {(1 − s/c)/
= + sin2 θ ⎟
Iθ KcM M 5λ 2 ⎜⎝ Iθ 2 ⎠ (1 + s/c)}1/2 and c is the speed of light. (a) Three Fe I lines of the star HDE
271 182, which belongs to the Large Magellanic Cloud, occur at 438.882 nm,
where I0 is the intensity of the incident laser radiation, M is the molar mass, R is 441.000 nm, and 442.020 nm. The same lines occur at 438.392 nm,
the radius of the particle, and K is a parameter that depends on the refractive 440.510 nm, and 441.510 nm in the spectrum of an Earth-bound iron arc.
index of the solution, the incident wavelength, and the distance between Determine whether HDE 271 182 is receding from or approaching the
the detector and the sample, which is held constant during the experiment. Earth and estimate the star’s radial speed with respect to the Earth. (b) What
It follows that structural properties, such as size and the molar mass of a additional information would you need to calculate the radial velocity of HDE
macromolecule, can be obtained from measurements of light scattering by a 271 182 with respect to the Sun?
sample at several angles θ relative to the direction of propagation of an incident 40.14 In Problem 40.13, we saw that Doppler shifts of atomic spectral lines
beam. The following data for an aqueous solution of a macromolecule with cM = are used to estimate the speed of recession or approach of a star. A spectral
2.0 kg m−3 were obtained at 20 °C with laser light at λ = 532 nm. In a separate line of 48Ti8+ (of mass 47.95mu) in a distant star was found to be shifted from
654.2 nm to 706.5 nm and to be broadened to 61.8 pm. What are the speed of
θ/o 15.0 45.0 70.0 85.0 90.0
recession and the surface temperature of the star?
102 × I0/Iθ 4.20 4.37 4.63 4.83 4.90
40.15 The Gaussian shape of a Doppler-broadened spectral line reflects
experiment, it was determined that K = 2.40 × 10−2 mol m3 kg−2. From this the Maxwell distribution of speeds in the sample at the temperature of the
information, calculate R and M for the macromolecule. experiment. In a spectrometer that makes use of phase-sensitive detection
the output signal is proportional to the first derivative of the signal intensity,
40.12 The collision frequency z of a molecule of mass m in a gas at a pressure dI/dν. Plot the resulting lineshape for various temperatures. How is the
p is z = 4σ(kT/πm)1/2p/kT, where σ is the collision cross-section. Find an separation of the peaks related to the temperature?

Topic 41 Molecular rotation


Discussion questions
41.1 Account for the rotational degeneracy of the various types of rigid rotor. 41.2 Discuss the differences between an oblate and a prolate symmetric rotor
Would their lack of rigidity affect your conclusions? and give several examples of each.

Exercises
41.1(a) Calculate the moment of inertia around the C2 axis (the bisector of the 41.1(b) Calculate the moment of inertia around the C3 axis (the threefold
OOO angle) and the corresponding rotational constant of an 16O3 molecule symmetry axis) and the corresponding rotational constant of a 31P1H3
(bond angle 117°; OO bond length 128 pm). molecule (bond angle 93.5°; PH bond length 142 pm).

Atkins09819.indb 445 9/11/2013 12:04:48 PM


446 9 Molecular spectroscopy

41.2(a) Plot the expressions for the two moments of inertia of a symmetric-top 41.3(b) Classify the following rotors: (a) CH2 = CH2, (b) SO3, (c) ClF3,
version of an AB4 molecule (Table 41.1) with equal bond lengths but with the (d) N2O.
angle θ increasing from 90° to the tetrahedral angle.
41.4(a) Determine the HC and CN bond lengths in HCN from the rotational
41.2(b) Plot the expressions for the two moments of inertia of a symmetric-
constants B(1H12C14N) = 44.316 GHz and B(2H12C14N) = 36.208 GHz.
top version of an AB4 molecule (Table 41.1) with θ equal to the tetrahedral
41.4(b) Determine the CO and CS bond lengths in OCS from the rotational
angle but with one A–B bond varying. Hint: Write ρ = RAB ′ / RAB , and
constants B(16O12C32S) = 6081.5 MHz, B(16O12C34S) = 5932.8 MHz.
allow ρ to vary from 2 to 1.
41.3(a) Classify the following rotors: (a) O3, (b) CH3CH3, (c) XeO4, (d) FeCp2
(Cp denotes the cyclopentadienyl group, C5H5).

Problems
41.1 Show that the moment of inertia of a diatomic molecule composed 41.2 Confirm the expression given in Table 41.1 for the moment of inertia of a
of atoms of masses mA and mB and bond length R is equal to meffR2, where linear ABC molecule. Hint: Begin by locating the centre of mass.
meff = mAmB/(mA + mB).

Topic 42 Rotational spectroscopy


Discussion questions
42.1 Account for the existence of a rotational zero-point energy in molecular 42.3 Discuss the physical origins of the gross selection rules for rotational
hydrogen. Raman spectroscopy.
42.2 Discuss the physical origins of the gross selection rules for microwave 42.4 Discuss the role of nuclear statistics in the occupation of energy levels in
spectroscopy. 1H12C ≡ 12C1H, 1H13C ≡ 13C1H, and 2H12C ≡ 12C2H. For nuclear spin data, see

Table 47.2.

Exercises
42.1(a) Which of the following molecules may show a pure rotational 42.5(b) Which of the following molecules may show a pure rotational Raman
microwave absorption spectrum: (a) H2, (b) HCl, (c) CH4, (d) CH3Cl, (e) spectrum: (a) CH2Cl2, (b) CH3CH3, (c) SF6, (d) N2O?
CH2Cl2?
42.6(a) The wavenumber of the incident radiation in a Raman spectrometer is
42.1(b) Which of the following molecules may show a pure rotational
20 487 cm−1. What is the wavenumber of the scattered Stokes radiation for the
microwave absorption spectrum: (a) H2O, (b) H2O2, (c) NH3, (d) N2O?
J = 2 ← 0 transition of 14N2?
42.2(a) Calculate the frequency and wavenumber of the J = 3 ← 2 transition in 42.6(b) The wavenumber of the incident radiation in a Raman spectrometer is
the pure rotational spectrum of 14N16O. The equilibrium bond length is 115 pm. 20 623 cm−1. What is the wavenumber of the scattered Stokes radiation for the
Does the frequency increase or decrease if centrifugal distortion is considered? J = 4 ← 2 transition of 16O2?
42.2(b) Calculate the frequency and wavenumber of the J = 2 ← 1 transition in the
42.7(a) The rotational Raman spectrum of 35Cl2 shows a series of Stokes lines
pure rotational spectrum of 12C16O. The equilibrium bond length is 112.81 pm.
separated by 0.9752 cm−1 and a similar series of anti-Stokes lines. Calculate
Does the frequency increase or decrease if centrifugal distortion is considered?
the bond length of the molecule.
42.3(a) The wavenumber of the J = 3 ← 2 rotational transition of 1H35Cl 42.7(b) The rotational Raman spectrum of 19F2 shows a series of Stokes lines
considered as a rigid rotor is 63.56 cm−1; what is the HeCl bond length? separated by 3.5312 cm−1 and a similar series of anti-Stokes lines. Calculate
42.3(b) The wavenumber of the J = 1 ← 0 rotational transition of 1H81Br the bond length of the molecule.
considered as a rigid rotor is 16.93 cm−1; what is the HeBr bond length?
42.8(a) What is the ratio of weights of populations due to the effects of nuclear
42.4(a) The spacing of lines in the microwave spectrum of 27Al1H is statistics for 35Cl2?
12.604 cm−1; calculate the moment of inertia and bond length of the molecule. 42.8(b) What is the ratio of weights of populations due to the effects of nuclear
42.4(b) The spacing of lines in the microwave spectrum of 35Cl19F is statistics for 12C32S2? What effect would be observed when 12C is replaced by
1.033 cm−1; calculate the moment of inertia and bond length of the molecule. 13C? For nuclear spin data, see Table 47.2.

42.5(a) Which of the following molecules may show a pure rotational Raman
spectrum: (a) H2, (b) HCl, (c) CH4, (d) CH3Cl?

Problems
42.1 The rotational constant of NH3 is 298 GHz. Compute the separation of cm−1, and a wavelength in mm, and show that the value of B is consistent with
the pure rotational spectrum lines as a frequency in GHz, a wavenumber in an N–H bond length of 101.4 pm and a bond angle of 106.78°.

www.ebook3000.com

Atkins09819.indb 446 9/11/2013 12:04:51 PM


Exercises and problems  447

42.2 Rotational absorption lines from 1H35Cl gas were found at the following Evaluate the rotational constant of the molecule. Also, estimate the value of J
wavenumbers (R.L. Hausler and R.A. Oetjen, J. Chem. Phys. 21, 1340 (1953)): for the most highly populated rotational energy level at 298 K and at 100 K.
83.32, 104.13, 124.73, 145.37, 165.89, 186.23, 206.60, 226.86 cm−1. Calculate
42.8 The rotational terms of a symmetric top, allowing for centrifugal
the moment of inertia and the bond length of the molecule. Predict the
distortion, are commonly written
positions of the corresponding lines in 2H35Cl.
42.3 Is the bond length in HCl the same as that in DCl? The wavenumbers  ( J +1) + ( A − B )K 2 − D J 2 ( J +1)2 − D J ( J +1)K 2 − D K 4
F ( J , K ) = BJ J JK K
of the J = 1 ← 0 rotational transitions for H35Cl and 2H35Cl are 20.8784 and
10.7840 cm−1, respectively. Accurate atomic masses are 1.007 825mu and
Derive an expression for the wavenumbers of the allowed rotational
2.0140mu for 1H and 2H, respectively. The mass of 35Cl is 34.968 85mu. Based
transitions. The following transition frequencies (in gigahertz, GHz) were
on this information alone, can you conclude that the bond lengths are the
observed for CH3F:
same or different in the two molecules?
51.0718 102.1426 102.1408 153.2103 153.2076
42.4 Thermodynamic considerations suggest that the copper monohalides CuX
should exist mainly as polymers in the gas phase, and indeed it proved difficult
Determine the values of as many constants in the expression for the rotational
to obtain the monomers in sufficient abundance to detect spectroscopically. This
terms as these values permit.
problem was overcome by flowing the halogen gas over copper heated to 1100 K
(Manson, et al. J. Chem. Phys. 63, 2724 (1975)). For CuBr the J = 13–14, 14–15, 42.9 Derive an expression for the value of J corresponding to the most
and 15–16 transitions occurred at 84 421.34, 90 449.25, and 96 476.72 MHz, highly populated rotational energy level of a diatomic rotor at a temperature
respectively. Calculate the rotational constant and bond length of CuBr. T, remembering that the degeneracy of each level is 2J + 1. Evaluate the
expression for ICl (for which B = 0.1142 cm−1) at 25 °C. Repeat the problem
42.5 The microwave spectrum of 16O12CS gave absorption lines (in GHz) as
for the most highly populated level of a spherical rotor, taking note of the fact
follows:
that each level is (2J + 1)2-fold degenerate. Evaluate the expression for CH4 (for
which B = 5.24 cm −1) at 25 °C.
J 1 2 3 4
42.10 A. Dalgarno, in Chemistry in the interstellar medium, Frontiers of
32S 24.325 92 36.488 82 48.651 64 60.814 08
Astrophysics, ed. E.H. Avrett, Harvard University Press, Cambridge (1976),
34S 23.732 33 47.462 40 notes that although both CH and CN spectra show up strongly in the
interstellar medium in the constellation Ophiuchus, the CN spectrum has
become the standard for the determination of the temperature of the cosmic
Using the expressions for moments of inertia in Table 41.1 and assuming that
microwave background radiation. Demonstrate through a calculation why
CH would not be as useful for this purpose as CN. The rotational constant B
the bond lengths are unchanged by substitution, calculate the CO and CS
0
bond lengths in OCS.
for CH is 14.190 cm−1.
42.6 Equation 42.8b may be rearranged into
42.11 The space immediately surrounding stars, the circumstellar space, is
significantly warmer because stars are very intense black-body emitters with
 J +1 ← J )/ 2( J +1) = B − 2D J
␯( ( J +1)2 temperatures of several thousand kelvin. Discuss how such factors as cloud
temperature, particle density, and particle velocity may affect the rotational
which is the equation of a straight line when the left-hand side is plotted spectrum of CO in an interstellar cloud. What new features in the spectrum
against (J + 1)2. The following wavenumbers of transitions (in cm−1) were of CO can be observed in gas ejected from and still near a star with
observed for 12C16O: temperatures of about 1000 K, relative to gas in a cloud with temperature
of about 10 K? Explain how these features may be used to distinguish
J: 0 1 2 3 4 between circumstellar and interstellar material on the basis of the rotational
spectrum of CO.
3.845 033 7.689 919 11.534 510 15.378 662 19.222 223
42.12 Pure rotational Raman spectra of gaseous C6H6 and C6D6
Determine B , D J , and the equilibrium bond length of CO.  C H ) = 0.189 60 cm −1 ,
yield the following rotational constants: B( 6 6
 −1
B(C 6 D6 ) = 0.156 81 cm . The moments of inertia of the molecules about
42.7‡ In a study of the rotational spectrum of the linear FeCO radical,
any axis perpendicular to the C6 axis were calculated from these data as
Tanaka, et al. (J. Chem. Phys. 106, 6820 (1997)) reported the following
I(C6H6) = 1.4759 × 10−45 kg m2, I(C6D6) = 1.7845 × 10−45 kg m2. Calculate the
J + 1 ← J transitions:
CC, CH, and CD bond lengths.
J 24 25 26 27 28 29
␯ /m −1 214 777.7 223 379.0 231 981.2 240 584.4 249 188.5 257 793.5

Topic 43 Vibrational spectroscopy: diatomic molecules


Discussion questions
43.1 Discuss the strengths and limitations of the parabolic and Morse 43.3 How is the method of combination differences used in rotation–vibration
functions as descriptors of the potential energy curve of a diatomic molecule. spectroscopy to determine rotational constants?
43.2 Discuss the effect of vibrational excitation on the rotational constant of a
diatomic molecule.

Atkins09819.indb 447 9/11/2013 12:05:01 PM


448 9 Molecular spectroscopy

Exercises
43.1(a) An object of mass 100 g suspended from the end of a rubber band has a 2309.5 cm−1 (H127I). Calculate the force constants of the hydrogen–halogen
vibrational frequency of 2.0 Hz. Calculate the force constant of the rubber band. bonds.
43.1(b) An object of mass 1.0 g suspended from the end of a spring has a 43.4(b) From the data in Exercise 43.4(a), predict the fundamental vibrational
vibrational frequency of 10.0 Hz. Calculate the force constant of the spring. wavenumbers of the deuterium halides.
43.2(a) Calculate the percentage difference in the fundamental vibrational 43.5(a) For 16O2, ΔG values for the transitions v = 1 ← 0, 2 ← 0, and 3 ← 0 are,
wavenumbers of 23Na35Cl and 23Na37Cl on the assumption that their force respectively, 1556.22, 3088.28, and 4596.21 cm−1. Calculate ␯ and xe. Assume
constants are the same. ye to be zero.
43.2(b) Calculate the percentage difference in the fundamental vibrational 43.5(b) For 14N2, ΔG values for the transitions ␯ = 1 ← 0, 2 ← 0, and 3 ← 0 are,
wavenumbers of 1H35Cl and 2H37Cl on the assumption that their force respectively, 2329.91, 4631.20, and 6903.69 cm−1. Calculate ␯ and xe. Assume
constants are the same. ye to be zero.
43.3(a) The wavenumber of the fundamental vibrational transition of 35Cl2 is 43.6(a) Which of the following molecules may show infrared absorption
564.9 cm−1. Calculate the force constant of the bond. spectra: (a) H2, (b) HCl, (c) CO2, (d) H2O?
43.3(b) The wavenumber of the fundamental vibrational transition of 79Br81Br 43.6(b) Which of the following molecules may show infrared absorption
is 323.2 cm−1. Calculate the force constant of the bond. spectra: (a) CH3CH3, (b) CH4, (c) CH3Cl, (d) N2?
43.4(a) The hydrogen halides have the following fundamental vibrational
wavenumbers: 4141.3 cm−1 (HF); 2988.9 cm−1 (H35Cl); 2649.7 cm−1 (H81Br);

Problems
43.1 The vibrational levels of NaI lie at the wavenumbers 142.81, 43.7 The rotational constant for CO is 1.9314 cm−1 and 1.6116 cm−1 in the
427.31, 710.31, and 991.81 cm−1. Show that they fit the expression ground and first excited vibrational states, respectively. By how much does the
(v+ 12 )␯ − (v + 12 )2 x e ␯ , and deduce the force constant, zero-point energy, and internuclear distance change as a result of this transition?
dissociation energy of the molecule.
43.8 The average spacing between the rotational lines of the P and R branches
43.2 The HCl molecule is quite well described by the Morse potential with of 12C21H2 and 12C22H2 is 2.352 cm−1 and 1.696 cm−1, respectively. Estimate the
De = 5.33 eV, ␯ = 2989.7 cm −1 , and x e ␯ = 52.05 cm −1 . Assuming that the CC and CH bond lengths.
potential is unchanged on deuteration, predict the dissociation energies (D0)
43.9 Absorptions in the v = 1 ← 0 vibration–rotation spectrum of 1H35Cl were
of (a) HCl, (b) DCl.
observed at the following wavenumbers (in cm−1):
43.3 The Morse potential (eqn 43.14) is very useful as a simple representation
of the actual molecular potential energy. When RbH was studied, it was found 2998.05 2981.05 2963.35 2944.99 2925.92
that ␯ = 936.8 cm −1 and x e ␯ =14.15 cm −1 . Plot the potential energy curve 2906.25 2865.14 2843.63 2821.59 2799.00
from 50 pm to 800 pm around Re = 236.7 pm. Then go on to explore how the
Assign the rotational quantum numbers and use the method of combination
rotation of a molecule may weaken its bond by allowing for the kinetic energy
 ( J +1), with B = / 4 πc μR2 . differences to determine the rotational constants of the two vibrational levels.
of rotation of a molecule and plotting V * = V + hc BJ
Plot these curves on the same diagram for J = 40, 80, and 100, and observe how 43.10 Suppose that the internuclear distance may be written R = Re + x where
the dissociation energy is affected by the rotation. (Taking B = 3.020 cm −1 at the Re is the equilibrium bond length. Also suppose that the potential well is
equilibrium bond length will greatly simplify the calculation.) symmetrical and confines the oscillator to small displacements. Deduce
expressions for 1/〈R〉2, 1/〈R2〉, and 〈1/R2〉 to the lowest nonzero power of
43.4‡ Luo, et al. (J. Chem. Phys. 98, 3564 (1993)) reported experimental
〈 x 2 〉 / Re2 and confirm that the values are not the same.
observation of the He2 complex, a species which had escaped detection
for a long time. The fact that the observation required temperatures in the 43.11 Continue the development of Problem 43.10 by using the virial
neighbourhood of 1 mK is consistent with computational studies which expression to relate 〈x2〉 to the vibrational quantum number. Does your result
~
suggest that hcDe for He2 is about 1.51 × 10−23 J, hcD 0 ≈ 2 × 10−26 J, and Re imply that the rotational constant increases or decreases as the oscillator
about 297 pm. (See Problem 29.1.) (a) Estimate the fundamental vibrational becomes excited to higher quantum states? What would be the effect of
wavenumber, force constant, moment of inertia, and rotational constant anharmonicity?
based on the harmonic-oscillator and rigid-rotor approximations. (b) Such
43.12 The rotational constant for a diatomic molecule in the vibrational
a weakly bound complex is hardly likely to be rigid. Estimate the vibrational
state v typically fits the expression B v = B e − a(v + 12 ). For the interhalogen
wavenumber and anharmonicity constant based on the Morse potential.
molecule IF it is found that B e = 0.279 71 cm −1 and a = 0.187 m−1 (note the
43.5 Confirm that a Morse oscillator has a finite number of bound states, the change of units). Calculate B 0 and B1 and use these values to calculate the
states with V < hcD e . Determine the value of vmax for the highest bound state. wavenumbers of the J′ → 3 transitions of the P and R branches. You will need
the following additional information: ␯ = 610.258 cm −1 and x e ␯ = 3.141 cm −1 .
43.6 Provided higher-order terms are neglected, eqn 43.17 for the vibrational
Estimate the dissociation energy of the IF molecule.
wavenumbers of an anharmonic oscillator, ΔG v+(1/2) = ␯ − 2(v +1)x e ␯ +…, is
the equation of a straight line when the left-hand side is plotted against v + 1. 43.13 At low resolution, the strongest absorption band in the infrared absorption
Use the following data on CO to determine the values of ␯ and x e ␯ for CO: spectrum of 12C16O is centred at 2150 cm−1. Upon closer examination at higher
resolution, this band is observed to be split into two sets of closely spaced peaks,
v 0 1 2 3 4 one on each side of the centre of the spectrum at 2143.26 cm−1. The separation
between the peaks immediately to the right and left of the centre is 7.655 cm−1.
ΔG v+ 1 /cm −1 2143.1 2116.1 2088.9 2061.3 2033.5
2 Make the harmonic oscillator and rigid rotor approximations and calculate from

www.ebook3000.com

Atkins09819.indb 448 9/11/2013 12:05:27 PM


Exercises and problems  449

these data: (a) the vibrational wavenumber of a CO molecule, (b) its molar zero- 43.14 The analysis of combination differences summarized in Section 43.4b
point vibrational energy, (c) the force constant of the CO bond, (d) the rotational considered the R and P branches. Extend the analysis to the O and S branches
 and (e) the bond length of CO.
constant B, of a Raman spectrum.

Topic 44 Vibrational spectroscopy: polyatomic molecules


Discussion questions
44.1 Discuss the physical origins of the gross selection rules for infrared 44.3 Suppose that you wish to characterize the normal modes of benzene in
spectroscopy. the gas phase. Why is it important to obtain both infrared absorption and
Raman spectra of your sample?
44.2 Discuss the physical origins of the gross selection rules for vibrational
Raman spectroscopy.

Exercises
44.1(a) Which of the following molecules may show infrared absorption 44.5(a) Which of the three vibrations of an AB2 molecule are infrared or
spectra: (a) H2, (b) HCl, (c) CO2, (d) H2O? Raman active when it is (a) angular, (b) linear?
44.1(b) Which of the following molecules may show infrared absorption 44.5(b) Which of the vibrations of an AB3 molecule are infrared or Raman
spectra: (a) CH3CH3, (b) CH4, (c) CH3Cl, (d) N2? active when it is (a) trigonal planar, (b) trigonal pyramidal?
44.2(a) How many normal modes of vibration are there for the following 44.6(a) Consider the vibrational mode that corresponds to the uniform
molecules? (a) H2O, (b) H2O2, (c) C2H4. expansion of the benzene ring. Is it (a) Raman, (b) infrared active?
44.2(b) How many normal modes of vibration are there for the following 44.6(b) Consider the vibrational mode that corresponds to the boat-like
molecules? (a) C6H6, (b) C6H5CH3, (c) HC≡C–C≡C–H. bending of a benzene ring. Is it (a) Raman, (b) infrared active?
44.3(a) How many vibrational modes are there for the molecule 44.7(a) The molecule CH2Cl2 belongs to the point group C2v. The
NC–(C≡C–C≡C–)10CN detected in an interstellar cloud? displacements of the atoms span 5A1 + 2A2 + 4B1 + 4B2. What are the
44.3(b) How many vibrational modes are there for the molecule symmetries of the normal modes of vibration?
NC–(C≡C–C≡C–)8CN detected in an interstellar cloud? 44.7(b) A carbon disulfide molecule belongs to the point group D∞h. The nine
displacements of the three atoms span A1g + 2A1u + 2E1u + E1g. What are the
44.4(a) Write an expression for the vibrational term for the ground vibrational
symmetries of the normal modes of vibration?
state of H2O in terms of the wavenumbers of the normal modes. Neglect
anharmonicities as in eqn 44.1. 44.8(a) Which of the normal modes of CH2Cl2 (Exercise 44.7(a)) are infrared
44.4(b) Write an expression for the vibrational term for the ground vibrational active? Which are Raman active?
state of SO2 in terms of the wavenumbers of the normal modes. Neglect 44.8(b) Which of the normal modes of carbon disulfide (Exercise 44.7(b)) are
anharmonicities as in eqn 44.1. infrared active? Which are Raman active?

Problems
44.1 Suppose that the out-of-plane distortion of a planar molecule could be bands at 870, 1370, 2869, and 3417 cm−1. The Raman spectrum of the same
described by a potential energy V = V0 (1− e −bh ), where h is the distance by
4
sample has bands at 877, 1408, 1435, and 3407 cm−1. All bands correspond to
which the central atom is displaced. Sketch this potential energy as a function fundamental vibrational wavenumbers and you may assume that (a) the 870
of h (allow h to be both negative and positive). What could be said about (a) and 877 cm−1 bands arise from the same normal mode, and (b) the 3417 and
the force constant, (b) the vibrations? Sketch the form of the ground-state 3407 cm−1 bands arise from the same normal mode. (i) If H2O2 were linear,
wavefunction. how many normal modes of vibration would it have? (ii) Give the symmetry
+ point group of each of the three proposed conformations of nonlinear H2O2.
44.2 Predict the shape of the nitronium ion, NO2 , from its Lewis structure and
(iii) Determine which of the proposed conformations is inconsistent with the
the VSEPR model. It has one Raman active vibrational mode at 1400 cm−1,
spectroscopic data. Explain your reasoning.
two strong IR active modes at 2360 and 540 cm−1, and one weak IR mode at
3735 cm−1. Are these data consistent with the predicted shape of the molecule?
Assign the vibrational wavenumbers to the modes from which they arise.
44.3 Consider the molecule CH3Cl. (a) To what point group does the molecule
belong? (b) How many normal modes of vibration does the molecule have?
(c) What are the symmetries of the normal modes of vibration for this
molecule? (d) Which of the vibrational modes of this molecule are infrared
active? (e) Which of the vibrational modes of this molecule are Raman active?
1 2 3
44.4 Suppose that three conformations are proposed for the nonlinear molecule
H2O2 (1, 2, and 3). The infrared absorption spectrum of gaseous H2O2 has

Atkins09819.indb 449 9/11/2013 12:05:33 PM


450 9 Molecular spectroscopy

Topic 45 Electronic spectroscopy


Discussion questions
45.1 Explain the origin of the term symbol 3 Σ g− for the ground state of 45.4 Explain how colour can arise from molecules.
dioxygen.
45.5 Suppose that you are a colour chemist and have been asked to intensify
45.2 Explain the basis of the Franck–Condon principle and how it leads to the the colour of a dye without changing the type of compound, and that the dye
formation of a vibrational progression. in question was a polyene. (a) Would you choose to lengthen or to shorten the
chain? (b) Would the modification to the length shift the apparent colour of
45.3 How do the band heads in P and R branches arise? Could the Q branch
the dye towards the red or the blue?
show a head?

Exercises
45.1(a) One of the excited states of the C2 molecule has the valence electron the rotational constant of the upper state? Has the bond length increased or
configuration 1σ 2g 1σ 2u 1π3u 1π1g . Give the multiplicity and parity of the term. decreased in the transition?
45.1(b) One of the excited states of the C2 molecule has the valence electron
45.8(a) The complex ion [Fe(OH2)6]3+ has an electronic absorption spectrum
configuration 1σ 2g 1σ 2u 1π2u 1π2g . Give the multiplicity and parity of the term.
with a maximum at 700 nm. Estimate a value of ΔO for the complex.
45.2(a) Which of the following transitions are electric-dipole allowed? 45.8(b) The complex ion [Fe(CN)6]3− has an electronic absorption spectrum
(a) 2Π ↔ 2Π, (b) 1Σ ↔ 1Σ, (c) Σ ↔ Δ, (d) Σ+ ↔ Σ−, (e) Σ+ ↔ Σ+. with a maximum at 305 nm. Estimate a value of ΔO for the complex.
45.2(b) Which of the following transitions are electric-dipole allowed?
45.9(a) Suppose that we can model a charge-transfer transition in a one-
(a) 1Σ g+ ↔1Σ u+ , (b) 3 Σ g+ ↔ 3 Σ u+ , (c) π* ↔ n. dimensional system as a process in which a rectangular wavefunction that
45.3(a) The ground-state wavefunction of a certain molecule is described by is nonzero in the range 0 ≤ x ≤ a makes a transition to another rectangular
the vibrational wavefunction ψ 0 = N 0 e −ax . Calculate the Franck–Condon
2
wavefunction that is nonzero in the range 12 a ≤ x ≤ b. Evaluate the transition
factor for a transition to a vibrational state described by the wavefunction moment ∫ ψ f xψ i dx. (Assume a < b.)
ψ v = N v e −b( x − x0 ) with b = a/2.
2
45.9(b) Suppose that we can model a charge-transfer transition in a one-
45.3(b) The ground-state wavefunction of a certain molecule is described by dimensional system as a process in which an electron described by a
the vibrational wavefunction ψ 0 = N 0 e −ax . Calculate the Franck–Condon rectangular wavefunction that is nonzero in the range 0 ≤ x ≤ a makes a
2

factor for a transition to a vibrational state described by the wavefunction transition to another rectangular wavefunction that is nonzero in the range
ψ v = N v xe −b( x − x0 ) , with b = a/2.
2
ca ≤ x ≤ a where 0 ≤ c ≤ 1. Evaluate the transition moment ∫ ψ f xψ i dx and
explore its dependence on c.
45.4(a) Suppose that the ground vibrational state of a molecule is modelled by
using the particle-in-a-box wavefunction, ψ0 = (2/L)1/2 sin(πx/L) for 0 ≤ x ≤ L 45.10(a) Suppose that we can model a charge-transfer transition in a one-
and 0 elsewhere. Calculate the Franck–Condon factor for a transition to a dimensional system as a process in which a Gaussian wavefunction centred on
vibrational state described by the wavefunction ψv = (2/L)1/2sin{π(x −L/2)/L} x = 0 and width a makes a transition to another Gaussian wavefunction of the
for L/4 ≤ x ≤ 5L/4 and 0 elsewhere. same width centred on x = 12 a. Evaluate the transition moment ∫ ψ f xψ i dx.
45.4(b) Suppose that the ground vibrational state of a molecule is modelled by 45.10(b) Suppose that we can model a charge-transfer transition in a one-
using the particle-in-a-box wavefunction ψ0 = (2/L)1/2 sin(πx/L) for 0 ≤ x ≤ L dimensional system as a process in which an electron described by a Gaussian
and 0 elsewhere. Calculate the Franck–Condon factor for a transition to a wavefunction centred on x = 0 and width a makes a transition to another
vibrational state described by the wavefunction ψv = (2/L)1/2 sin{π(x −L/4)/L} Gaussian wavefunction of width a/2 and centred on x = 0. Evaluate the
for L/2 ≤ x ≤ 3L/2 and 0 elsewhere. transition moment ∫ ψ f xψ i dx .
45.11(a) The two compounds 2,3-dimethyl-2-butene (4) and 2,5-dimethyl-
45.5(a) Use eqn 45.8a to infer the value of J corresponding to the location of
2,4-hexadiene (5) are to be distinguished by their ultraviolet absorption
the band head of the P branch of a transition.
spectra. The maximum absorption in one compound occurs at 192 nm and
45.5(b) Use eqn 45.8c to infer the value of J corresponding to the location of
in the other at 243 nm. Match the maxima to the compounds and justify the
the band head of the R branch of a transition.
assignment.
45.6(a) The following parameters describe the electronic ground state and an
excited electronic state of SnO: B = 0.3540 cm −1 , B ′ = 0.3101cm −1 . Which
branch of the transition between them shows a head? At what value of J will
it occur?
45.6(b) The following parameters describe the electronic ground state and an
excited electronic state of BeH: B =10.308 cm −1 , B ′ = 10.470 cm −1 . Which 4 2,3-Dimethyl-2-butene 5 2,5-Dimethyl-2,4-hexadiene
branch of the transition between them shows a head? At what value of J will 45.11(b) 3-Buten-2-one (6) has a strong absorption at 213 nm and a weaker
it occur? absorption at 320 nm. Justify these features and assign the ultraviolet
45.7(a) The R branch of the 1 Π u ←1Σ g+ transition of H2 shows a band head absorption transitions.
at the very low value of J = 1. The rotational constant of the ground state is O
60.80 cm−1. What is the rotational constant of the upper state? Has the bond
length increased or decreased in the transition?
45.7(b) The P branch of the 2 Π ←2 Σ + transition of CdH shows a band head
at J = 25. The rotational constant of the ground state is 5.437 cm−1. What is 6 3-Butene-2-one

www.ebook3000.com

Atkins09819.indb 450 9/11/2013 12:05:54 PM


Exercises and problems  451

Problems
45.1 The term symbol for the first excited state of N2+ is 2Πg. Use the building- 3652 cm−1. (a) What conclusions can be drawn from the nature of the orbital
up principle to find the excited-state configuration to which this term symbol from which the electron is ejected? (b) In the same spectrum of H2O, the band
corresponds. near 7.0 eV shows a long vibrational series with spacing 0.125 eV. The bending
mode of H2O lies at 1596 cm−1. What conclusions can you draw about the
45.2‡ Dojahn, et al. (J. Phys. Chem. 100, 9649 (1996)) characterized the
characteristics of the orbital occupied by the photoelectron?
potential energy curves of the ground and electronic states of homonuclear
diatomic halogen anions. These anions have a 2 Σ u+ ground state and 2Πg, 45.6 A lot of information about the energy levels and wavefunctions of
2Π , and 2 Σ + excited states. To which of the excited states are electric-dipole small inorganic molecules can be obtained from their ultraviolet spectra.
u g
transitions allowed? Explain your conclusion. An example of a spectrum with considerable vibrational structure, that of
gaseous SO2 at 25 °C, is shown in Fig. 45.5. Estimate the integrated absorption
45.3 The vibrational wavenumber of the oxygen molecule in its electronic
coefficient for the transition. What electronic states are accessible from the A1
ground state is 1580 cm−1, whereas that in the excited state (B 3 Σ u− ), to which
ground state of this C2v molecule by electric dipole transitions?
there is an allowed electronic transition, is 700 cm−1. Given that the separation
in energy between the minima in their respective potential energy curves 45.7 Assume that the electronic states of the π electrons of a conjugated
of these two electronic states is 6.175 eV, what is the wavenumber of the molecule can be approximated by the wavefunctions of a particle in a
lowest energy transition in the band of transitions originating from the v = 0 one-dimensional box, and that the magnitude of the dipole moment can
vibrational state of the electronic ground state to this excited state? Ignore any be related to the displacement along this length by μ = −ex. Show that the
rotational structure or anharmonicity. transition probability for the transition n = 1 → n = 2 is nonzero, whereas
that for n = 1 → n = 3 is zero. Hints: (a) The following relation will be useful:
45.4 We are now ready to understand more deeply the features of
sin x sin y = 12 cos(x − y ) − 12 cos(x + y ). (b) Relevant integrals are found in the
photoelectron spectra (Topic 24). Figure F9.4 shows the photoelectron
Resource section.
spectrum of HBr. Disregarding for now the fine structure, the HBr lines fall
into two main groups. The least tightly bound electrons (with the lowest 45.8 1,3,5-Hexatriene (a kind of ‘linear’ benzene) was converted into benzene
ionization energies and hence highest kinetic energies when ejected) are those itself. On the basis of a free-electron molecular orbital model (in which
in the lone pairs of the Br atom. The next ionization energy lies at 15.2 eV, hexatriene is treated as a linear box and benzene as a ring), would you expect
and corresponds to the removal of an electron from the HBr σ bond. (a) The the lowest-energy absorption to rise or fall in energy?
spectrum shows that ejection of a σ electron is accompanied by a considerable
45.9 Estimate the magnitude of the transition dipole moment of a charge-
amount of vibrational excitation. Use the Franck–Condon principle to
transfer transition modelled as the migration of an electron from an H1s
account for this observation. (b) Go on to explain why the lack of much
orbital on one atom to another H1s orbital on an atom a distance R away.
vibrational structure in the other band is consistent with the nonbonding role
Approximate the transition moment by −eRS, where S is the overlap integral
of the Br4px and Br4py lone-pair electrons.
of the two orbitals. Sketch the transition moment as a function of R using
45.5 The highest kinetic energy electrons in the photoelectron spectrum of H2O the curve for S given in Fig. 24.7. Why does the intensity of a charge-transfer
using 21.22 eV radiation are at about 9 eV and show a large vibrational spacing transition fall to zero as R approaches zero and infinity?
of 0.41 eV. The symmetric stretching mode of the neutral H2O molecule lies at
45.10 Figure F9.5 shows the UV–visible absorption spectra of a selection of
amino acids. Suggest reasons for their different appearances in terms of the
structures of the molecules.

Trp
Cys
Absorbance

Tyr
Signal

Magnification
Gly

220 280
Wavelength, λ/nm
17 15 13 11 9
Ionization energy, I/eV
Figure F9.5 Electronic absorption spectra of selected amino acids.
Figure F9.4 The photoelectron spectrum of HBr.

Topic 46 Decay of excited states


Discussion questions
46.1 Describe the mechanism of fluorescence. In what respects is a 46.3 Describe the principles of (a) continuous-wave and (b) pulsed laser
fluorescence spectrum not the exact mirror image of the corresponding action.
absorption spectrum?
46.4 How might you use a Q-switched or mode-locked laser in the study of a
46.2 What is the evidence for the correctness of the mechanism of very fast chemical reaction that can be initiated by absorption of light?
fluorescence?

Atkins09819.indb 451 9/11/2013 12:06:00 PM


452 9 Molecular spectroscopy

Exercises

emission intensity
46.1(a) The line marked A in Fig. F9.6 is the fluorescence spectrum of
benzophenone in solid solution in ethanol at low temperatures observed
B
when the sample is illuminated with 360 nm light. What can be said about
the vibrational energy levels of the carbonyl group in (a) its ground electronic
state and (b) its excited electronic state?
A
46.1(b) When naphthalene is illuminated with 360 nm light it does not absorb,
but the line marked B in Fig F9.6 is the phosphorescence spectrum of a solid
solution of a mixture of naphthalene and benzophenone in ethanol. Now a
component of fluorescence from naphthalene can be detected. Account for 15 20 25
~
ν/(1000 cm–1)
this observation.
46.2(a) The oxygen molecule absorbs ultraviolet radiation in a transition Figure F9.6 The fluorescence and phosphorescence spectra of two
from its 3 Σ g− ground electronic state to an excited state that is energetically solutions.
close to a dissociative 5Πu state. The absorption band has a relatively large
experimental linewidth. Account for this observation.
46.3(a) A pulsed laser rated at 0.10 mJ can generate radiation with peak power
46.2(b) The hydrogen molecule absorbs ultraviolet radiation in a
output of 5.0 MW and average power output of 7.0 kW. What are the pulse
transition from its 1 Σ g+ ground electronic state to an excited state that
duration and repetition rate?
is  energetically close to a dissociative 1 Σ u+ state. The absorption band
46.3(b) A pulsed laser rated at 20.0 μJ can generate radiation with peak power
has a relatively large experimental linewidth. Account for this
output of 100 kW and average power output of 0.40 mW. What are the pulse
observation.
duration and repetition rate?

Problems
46.1 The fluorescence spectrum of anthracene vapour shows a series of peaks mass spectrometer, the macromolecule is first embedded in a solid matrix that
of increasing intensity with individual maxima at 440 nm, 410 nm, 390 nm, often consists of an organic acid such as 2,5-dihydroxybenzoic acid, nicotinic
and 370 nm followed by a sharp cut-off at shorter wavelengths. The absorption acid, or α-cyanocarboxylic acid. This sample is then irradiated with a laser
spectrum rises sharply from zero to a maximum at 360 nm with a trail of pulse. The pulse of electromagnetic energy ejects matrix ions, cations, and
peaks of lessening intensity at 345 nm, 330 nm, and 305 nm. Account for these neutral macromolecules, thus creating a dense gas plume above the sample
observations. surface. The macromolecule is ionized by collisions and complexation with
H+ cations, resulting in molecular ions of varying charges. The spectrum of a
46.2 Use mathematical software or an electronic spreadsheet to simulate the
mixture of polymers consists of multiple peaks arising from molecules with
output of a mode-locked laser (that is, plots such as that shown in Fig. 46.16)
different molar masses. A MALDI-TOF mass spectrum consists of two intense
for L = 30 cm and N = 100 and 1000.
features at m/z = 9912 and 4554 g mol−1. Does the sample contain one or two
46.3 Matrix-assisted laser desorption/ionization (MALDI) is a type of mass distinct biopolymers? Explain your answer.
spectrometry, a technique in which the sample is first ionized in the gas phase
46.4 A certain molecule fluoresces at a wavelength of 400 nm with a half-life
and then the mass-to-charge number ratios (m/z) of all ions are measured.
of 1.0 ns. It phosphoresces at 500 nm. If the ratio of the transition probabilities
MALDI-TOF mass spectrometry, so called because the MALDI technique
for stimulated emission for the S* → S to the T → S transitions is 1.0 × 105,
is coupled to a time-of-flight (TOF) ion detector, is used widely in the
what is the half-life of the phosphorescent state?
determination of the molar masses of macromolecules. In a MALDI-TOF

Integrated activities
F9.1 In the group theoretical language developed in Topics 31–33, a spherical F9.3‡ The H3+ ion has recently been found in the interstellar medium and
rotor is a molecule that belongs to a cubic or icosahedral point group, a in the atmospheres of Jupiter, Saturn, and Uranus. The rotational energy
symmetric rotor is a molecule with at least a threefold axis of symmetry, levels of H3+ , an oblate symmetric rotor, are given by eqn 41.13, with
and an asymmetric rotor is a molecule without a threefold (or higher) axis. C  replacing A, when centrifugal distortion and other complications are
Linear molecules are linear rotors. Classify each of the following molecules as ignored. Experimental values for vibrational–rotational constants are
a spherical, symmetric, linear, or asymmetric rotor and justify your answers ␯ (E′)=2521.6 cm −1 , B = 43.55 cm −1 , and C = 20.71cm −1 . (a) Show that, for a
with group theoretical arguments: (a) CH4, (b) CH3CN, (c) CO2, (d) CH3OH, nonlinear planar molecule (such as H3+ ), IC = 2IB. The rather large discrepancy
(e) benzene, (f) pyridine. with the experimental values is due to factors ignored in eqn 41.13. (b)
 = 4 B 3 /␯ 2 ) for the centrifugal distortion constant Calculate an approximate value of the HeH bond length in H3+ .
F9.2 Derive eqn 41.17 (D J
~ (c) The value of Re obtained from the best quantum mechanical calculations
D J of a diatomic molecule of effective mass meff. Treat the bond as an elastic
by J.B. Anderson (J. Chem. Phys. 96, 3702 (1991)) is 87.32 pm. Use this result
spring with force constant kf and equilibrium length re that is subjected to a  (d) Assuming
to calculate the values of the rotational constants B and C.
centrifugal distortion to a new length rc. Begin the derivation by letting the
that the geometry and force constants are the same in D3+ and H3+ , calculate
particles experience a restoring force of magnitude kf(rc − re) that is countered
the spectroscopic constants of D3+ . The molecular ion D3+ was first produced
perfectly by a centrifugal force meffω2rc, where ω is the angular velocity of the
by Shy, et al. (Phys. Rev. Lett. 45, 535 (1980)), who observed the ν2(E′) band in
rotating molecule. Then introduce quantum mechanical effects by writing
the infrared.
the angular momentum as {J(J + 1)}1/2 ħ. Finally, write an expression for the
energy of the rotating molecule, compare it with eqn 41.16, and infer an F9.4 Use molecular modelling software and the computational method
expression for D J . of your choice (semi-empirical, ab initio, or DFT method) to construct

www.ebook3000.com

Atkins09819.indb 452 9/11/2013 12:06:15 PM


Exercises and problems  453

molecular potential energy curves like the one shown in Fig. 43.1. Consider
the hydrogen halides (HF, HCl, HBr, and HI): (a) plot the calculated energy
of each molecule against the bond length, and (b) identify the order of force
constants of the H–Hal bonds. 10
F9.5 The semi-empirical, ab initio, and DFT methods discussed in Topics
F9.8‡ There is a gaseous interstellar cloud in the constellation Ophiuchus
28–30 can be used to simulate the vibrational spectrum of a molecule, and
which is illuminated from behind by the star ζ-Ophiuci. Analysis of the
it is then possible to determine the correspondence between a vibrational
electronic–vibrational–rotational absorption lines shows the presence
frequency and the atomic displacements that give rise to a normal mode.
of CN molecules in the interstellar medium. A strong absorption line in
(a) Using molecular modelling software and the computational method of your
the ultraviolet region at λ = 387.5 nm was observed, corresponding to the
choice (semi-empirical, ab initio, or DFT method), calculate the fundamental
transition J = 0–1. Unexpectedly, a second strong absorption line with 25
vibrational wavenumbers and visualize the vibrational normal modes of SO2
per cent of the intensity of the first was found at a slightly longer wavelength
in the gas phase. (b) The experimental values of the fundamental vibrational
(Δλ = 0.061 nm), corresponding to the transition J = 1–1 (here allowed).
wavenumbers of SO2 in the gas phase are 525 cm−1, 1151 cm−1, and 1336 cm−1.
Calculate the temperature of the CN molecules. Gerhard Herzberg, who
Compare the calculated and experimental values. Even if agreement is poor,
was later to receive the Nobel Prize for his contributions to spectroscopy,
is it possible to establish a correlation between an experimental value of the
calculated the temperature as 2.3 K. Although puzzled by this result, he did
vibrational wavenumber with a specific vibrational normal mode?
not realize its full significance. If he had, his prize might have been for the
F9.6 Use appropriate electronic structure software to perform calculations discovery of the cosmic microwave background radiation.
on H2O and CO2 using basis sets of your or your instructor’s choosing.
F9.9‡ One of the principal methods for obtaining the electronic spectra of
(a) Compute ground-state energies, equilibrium geometries, and vibrational
unstable radicals is to study the spectra of comets, which are almost entirely
frequencies for each molecule. (b) Compute the magnitude of the dipole
due to radicals. Many radical spectra have been found in comets, including
moment of H2O; the experimental value is 1.854 D. (c) Compare computed
that due to CN. These radicals are produced in comets by the absorption
values to experiment and suggest reasons for any discrepancies.
of far-ultraviolet solar radiation by their parent compounds. Subsequently,
F9.7 The protein haemerythrin is responsible for binding and carrying O2 their fluorescence is excited by sunlight of longer wavelength. The spectra of
in some invertebrates. Each protein molecule has two Fe2+ ions that are in comet Hale–Bopp (C/1995 O1) have been the subject of many recent studies.
very close proximity and work together to bind one molecule of O2. The One such study is that of the fluorescence spectrum of CN in the comet at
Fe2O2 group of oxygenated haemerythrin is coloured and has an electronic large heliocentric distances by R.M. Wagner and D.G. Schleicher (Science
absorption band at 500 nm. The resonance Raman spectrum of oxygenated 275, 1918 (1997)), in which the authors determine the spatial distribution
haemerythrin obtained with laser excitation at 500 nm has a band at 844 cm−1 and rate of production of CN in the coma. The (0–0) vibrational band is
that has been attributed to the O–O stretching mode of bound 16O2. (a) Why centred on 387.6 nm and the weaker (1–1) band with relative intensity 0.1 is
is resonance Raman spectroscopy and not infrared spectroscopy the method centred on 386.4 nm. The band heads for (0–0) and (0–1) are known to be
of choice for the study of the binding of O2 to haemerythrin? (b) Proof 388.3 and 421.6 nm, respectively. From these data, calculate the energy of the
that the 844 cm−1 band arises from a bound O2 species may be obtained by excited S1 state relative to the ground S0 state, the vibrational wavenumbers
conducting experiments on samples of haemerythrin that have been mixed and the difference in the vibrational wavenumbers of the two states, and the
with 18O2, instead of 16O2. Predict the fundamental vibrational wavenumber relative populations of the v = 0 and v = 1 vibrational levels of the S1 state.
of the 18O–18O stretching mode in a sample of haemerythrin that has been Also estimate the effective temperature of the molecule in the excited S1 state.
treated with 18O2. (c) The fundamental vibrational wavenumbers for the O–O Only eight rotational levels of the S1 state are thought to be populated. Is that
stretching modes of O2, O2− (superoxide anion), and O2− 2 (peroxide anion) observation consistent with the effective temperature of the S1 state?
are 1555, 1107, and 878 cm−1, respectively. Explain this trend in terms of the
F9.10 The moments of inertia of the linear mercury(II) halides are very
electronic structures of O2, O2− , and O2− 2 . Hint: Review Topic 24. What are the
large, so the O and S branches of their vibrational Raman spectra show
bond orders of O2, O2− , and O2− 2 ? (d) Based on the data given above, which
little rotational structure. Nevertheless, the peaks of both branches can be
of the following species best describes the Fe2O2 group of haemerythrin:
identified and have been used to measure the rotational constants of the
Fe22+ O2 , Fe2+ Fe3+ O2− , or Fe32+ O22−? Explain your reasoning. (e) The resonance
molecules (R.J.H. Clark and D.M. Rippon, J. Chem. Soc. Faraday Soc. II 69,
Raman spectrum of haemerythrin mixed with 16O18O has two bands that
1496 (1973)). Show, from a knowledge of the value of J corresponding to the
can be attributed to the O–O stretching mode of bound oxygen. Discuss how
intensity maximum, that the separation of the peaks of the O and S branches
this observation may be used to exclude one or more of the four proposed  /hc)1/2 . The following widths
is given by the Placzek–Teller relation, δ = (32 BkT
schemes (7–10) for binding of O2 to the Fe2 site of haemerythrin.
were obtained at the temperatures stated:

HgCl2 HgBr2 HgI2


θ/ °C 282 292 292
δ/cm−1 23.8 15.2 11.4

7 8 Calculate the bond lengths in the three molecules.


F9.11‡ A mixture of carbon dioxide (2.1 per cent) and helium, at 1.00 bar and
298 K in a gas cell of length 10 cm, has an infrared absorption band centred at
2349 cm−1 with absorbances, A(␯ ), described by

a1 a4
A(␯ ) = +
1+ a2 (␯ − a3 )2 1+ a5 (␯ − a6 )2

where the coefficients are a1 = 0.932, a2 = 0.005050 cm2, a3 = 2333 cm−1,


9 a4 = 1.504, a5 =0.01521 cm2, a6 = 2362 cm−1. (a) Draw graphs of A(␯ ) and

Atkins09819.indb 453 9/11/2013 12:06:32 PM


454 9 Molecular spectroscopy

ε(␯ ) . What is the origin of both the band and the bandwidth? What are the C 11
allowed and forbidden transitions of this band? (b) Calculate the transition +
wavenumbers and absorbances of the band with a simple harmonic N
oscillator–rigid rotor model and compare the result with the experimental H
spectra. The CO bond length is 116.2 pm. (c) Within what height, h, is 11
basically all the infrared emission from the Earth in this band absorbed by
atmospheric carbon dioxide? The mole fraction of CO2 in the atmosphere is F9.14 Aromatic hydrocarbons and I2 form complexes from which charge-
3.3 ×10−4 and T/K = 288 – 0.0065(h/m) below 10 km. Draw a surface plot of transfer electronic transitions are observed. The hydrocarbon acts as an
the atmospheric transmittance of the band as a function of both height and electron donor and I2 as an electron acceptor. The energies hνmax of the charge
wavenumber. transfer transitions for a number of hydrocarbon–I2 complexes are given
below:
F9.12 Use a group theoretical arguments to decide which of the following
transitions are electric-dipole allowed: (a) the π* ← π transition in ethene, Hydrocarbon Benzene Biphenyl Naphthalene Phenanthrene Pyrene Anthracene
(b)  the π* ← n transition in a carbonyl group in a C2v environment. 4.184 3.654 3.452 3.288 2.989 2.890
hνmax/eV
F9.13 Use molecule (11) as a model of the trans conformation of the
Investigate the hypothesis that there is a correlation between the energy of the
chromophore found in rhodopsin. In this model, the methyl group bound
HOMO of the hydrocarbon (from which the electron comes in the charge-
to the nitrogen atom of the protonated Schiff ’s base replaces the protein.
transfer transition) and hνmax. Use one of the molecular electronic structure
(a) Using molecular modelling software and the computational method
methods discussed in Topics 28–30 to determine the energy of the HOMO of
of your instructor’s choice, calculate the energy separation between the
each hydrocarbon in the data set.
HOMO and LUMO of (11). (b) Repeat the calculation for the 11-cis form
of (11). (c) Based on your results from parts (a) and (b), do you expect the F9.15 Spin angular momentum is conserved when a molecule dissociates
experimental frequency for the π* ← π visible absorption of the trans form into atoms. What atom multiplicities are permitted when (a) an O2 molecule,
of (11) to be higher or lower than that for the 11-cis form of (11)? (b) an N2 molecule dissociates into atoms?

www.ebook3000.com

Atkins09819.indb 454 9/11/2013 12:06:34 PM


FOCUS 10 ON Magnetic resonance

Topic 47 Topic 48 Topic 49 Topic 50


Pulse Electron
General Features of
techniques paramagnetic
principles NMR spectra
in NMR resonance

Focus 3 Focus 8 Focus 5


The quantum Atomic
mechanics Interactions structure
of motion and spectra

In Atomic structure and spectra we see that electrons possess a property called ‘spin’. The techniques
of ‘magnetic resonance’ which are treated in this set of Topics probe transitions between spin states
of electrons and nuclei in molecules. ‘Nuclear magnetic resonance’ (NMR) spectroscopy, the main
focus of this group of Topics, is one of the most widely used techniques in chemistry for the explo-
ration of structural and dynamical properties of molecules as large as biopolymers. In Topic 47 we
begin our discussion of magnetic resonance with an account of the principles that govern spectro-
scopic transitions between spin states of nuclei and electrons in molecules.
Topic 48 is a discussion of conventional NMR, showing how the properties of a magnetic nucleus
are affected by its electronic environment and the presence of magnetic nuclei in its vicinity. These
concepts lead to understanding of how molecular structure governs the appearance of NMR spec-
tra. Modern versions of NMR are based on the use of pulses of electromagnetic radiation and the
processing of the resulting signal by ‘Fourier transform’ techniques (Topic 49); they also utilize con-
cepts introduced in Interactions and The quantum mechanics of motion. It is through the application
of these pulse techniques that NMR spectroscopy can probe a vast array of small and large molecules
in a variety of environments.
The experimental techniques for electron paramagnetic resonance (EPR) resemble those used in
the early days of NMR. The information obtained is used to investigate species with unpaired elec-
trons, using concepts from Atomic structure and spectra. Topic 50 is a brief survey of the applications
of EPR to the study of organic radicals and d-metal complexes.

What is the impact of this material?


One of the most striking applications of nuclear magnetic resonance is in medicine. ‘Magnetic reso-
nance imaging’ (MRI) is a portrayal of the concentrations of protons in a solid object (Impact 10.1). The
technique is particularly useful for diagnosing disease. In Impact 10.2 we highlight an application of
electron paramagnetic resonance in materials science and biochemistry: the use of a ‘spin probe’, a
radical that interacts with a biopolymer or a nanostructure and has an EPR spectrum that reveals its
structural and dynamical properties.

Atkins09819.indb 455 9/11/2013 12:06:44 PM


To read more about the impact of this material, scan the QR code or go to
http://bcs.whfreeman.com/webpub/chemistry/qmc2e/impact/qchem_
impact10.html.

www.ebook3000.com

Atkins09819.indb 456 9/11/2013 12:06:44 PM


TOPIC 47

General principles
the motion of the common axle. As a result, energy flows
Contents between the two pendulums. The energy transfer occurs
47.1 Nuclear magnetic resonance 457 most efficiently when the frequencies of the two pendulums
(a) The energies of nuclei in magnetic fields 458 are identical. The condition of strong effective coupling
Brief illustration 47.1: The resonance condition when the frequencies of two oscillators are identical is called
in NMR 459 resonance. Resonance is the basis of a number of everyday
(b) The NMR spectrometer 459 phenomena, including the response of radios to the weak
Brief illustration 47.2: Nuclear spin populations 460 oscillations of the electromagnetic field generated by a distant
47.2 Electron paramagnetic resonance 461 transmitter. Historically, spectroscopic techniques that meas-
(a) The energies of electrons in magnetic fields 461 ure transitions between nuclear and electron spin states have
Brief illustration 47.3: The resonance condition carried the term ‘resonance’ in their names because they have
in EPR 461 depended on matching a set of energy levels to a source of
(b) The EPR spectrometer 462 monochromatic radiation and observing the strong absorp-
Brief illustration 47.4: Electron spin populations 462 tion that occurs at resonance. In fact, all spectroscopy is a
Checklist of concepts 463 form of resonant coupling between the electromagnetic field
Checklist of equations 463 and the molecules; what distinguishes magnetic resonance is
that the energy levels themselves are modified by the applica-
tion of a magnetic field.
The Stern–Gerlach experiment (Topic 19) provided evidence
for electron spin. It turns out that many nuclei also possess spin
angular momentum. Orbital and spin angular momenta give
➤ Why do you need to know this material? rise to magnetic moments, and to say that electrons and nuclei
Nuclear magnetic resonance spectroscopy is used widely have magnetic moments means that, to some extent, they
in chemistry and medicine. To understand the power behave like small bar magnets with energies that depend on
of magnetic resonance, you need to understand the their orientation in an applied magnetic field. Here we estab-
principles that govern spectroscopic transitions between lish how the energies of electrons and nuclei depend on the
spin states of electrons and nuclei in molecules. applied field. This material sets the stage for the exploration of
the structure and dynamics of complex molecules by magnetic
➤ What is the key idea? resonance spectroscopy (Topics 48–50).
Resonant absorption occurs when the separation between
the energy levels of spins in a magnetic field matches the
energy of incident photons.
47.1 Nuclear magnetic resonance
➤ What do you need to know already?
You need to be familiar with the quantum mechanical The application of resonance that we describe here depends
concept of spin (Topic 19) and the Boltzmann distribution on the fact that many nuclei possess spin angular momentum
(Foundations, Topic 2, and Topic 51). characterized by a nuclear spin quantum number I (the ana-
logue of s for electrons). To understand the nuclear magnetic
resonance (NMR) experiment we must describe the behaviour
When two pendulums share a slightly flexible support and of nuclei in magnetic fields, and then the basic techniques for
one is set in motion, the other is forced into oscillation by detecting spectroscopic transitions.

Atkins09819.indb 457 9/11/2013 12:06:46 PM


458 10 Magnetic resonance

(a) The energies of nuclei in magnetic fields occasionally used: 1 T = 104 G. Quantum mechanically, we write
the hamiltonian as
The nuclear spin quantum number, I, is a fixed characteristic
property of a nucleus and, depending on the nuclide, is either  = −μ ⋅B
H (47.2)
an integer or a half-integer (Table 47.1). A nucleus with spin
quantum number I has the following properties: To write an expression for μ we use the fact that, just as for
electrons (Topic 21), the magnetic moment of a nucleus is pro-
r An angular momentum of magnitude {I(I + 1)}1/2ħ. portional to its angular momentum. The operators in eqn 47.2

Physical interpretation
r A component of angular momentum m Iħ on a are then
specified axis (‘the z-axis’), where m I = I, I – 1, …, –I.
μ = γ N I and  = −γ B ⋅ I
H
r If I > 0, a magnetic moment with a constant N (47.3a)
magnitude and an orientation that is determined by
the value of m I. where γN is the nuclear magnetogyric ratio of the specified
nucleus, an empirically determined characteristic arising from its
According to the second property, the spin, and hence the mag- internal structure (Table 47.2). For a magnetic field of magnitude
netic moment, of the nucleus may lie in 2I + 1 different orienta- B0 along the z-direction, the hamiltonian in eqn 47.3a becomes
tions relative to an axis. A proton has I = 12 and its spin may
adopt either of two orientations; a 14N nucleus has I = 1 and its H = −γ N B0 Iz (47.3b)
spin may adopt any of three orientations; both 12C and 16O have
I = 0 and hence zero magnetic moment. Because the eigenvalues of the operator Iz are mIħ, the eigen-
Classically, the energy of a magnetic moment μ in a mag- values of this hamiltonian are
netic field B is equal to the scalar product (Mathematical back-
Energies of a nuclear
ground 4) Em = −γ N B0 mI
I spin in a magnetic field (47.4a)

E = −μ⋅B (47.1) The expression for the energy is often written in terms of the
nuclear magneton, μN,
More formally, B is the magnetic induction and is measured
in tesla, T; 1 T = 1 kg s−2 A−1. The (non-SI) unit gauss, G, is also e
μN = Nuclear magneton (47.4b)
2mp
Table 47.1 Nuclear constitution and the nuclear spin quantum
number* where mp is the mass of the proton and μN = 5.051 × 10−2 J T−1,
and in terms of an empirical constant called the nuclear
Number of protons Number of neutrons I
g-factor, gI, when eqn 47.4a becomes
Even Even 0
Odd Odd Integer (1, 2, 3, …) γ N Energies of a nuclear
Em = − g I μN B0 mI gI = (47.4c)
Even Odd Half-integer ( 1 3 5
, ,
2 2 2
,…) I
μN spin in a magnetic field

Odd Even Half-integer ( 1 3 5


, , ,…)
2 2 2 Nuclear g-factors are experimentally determined dimen-
* The spin of a nucleus may be different if it is in an excited state; throughout this
sionless quantities with values typically between –6 and +6
Topic we deal only with the ground state of nuclei. (Table 47.2). Positive values of gI and γN denote a magnetic

Table 47.2* Nuclear spin properties

Nuclide Natural abundance/% Spin I g-factor, gI Magnetogyric ratio, NMR frequency


γN/(107 T−1 s−1) at 1 T, ν/MHz

1n 1 –3.826 –18.32 29.164


2

1H 99.98 1 5.586 26.75 42.576


2
2H 0.02 1 0.857 4.11 6.536

13C 1.11 1 1.405 6.73 10.708


2
14N 99.64 1 0.404 1.93 3.078
* More values are given in the Resource section.

www.ebook3000.com

Atkins09819.indb 458 9/11/2013 12:07:01 PM


47 General principles  459

moment that lies in the same direction as the spin angular


β, mI = –½
momentum vector; negative values indicate that the mag-
netic moment and spin lie in opposite directions. A nuclear
magnet is about 2000 times weaker than the magnet associ- Magnetic Magnetic
field off field on
ated with electron spin.
γNhB0
For the remainder of our discussion of nuclear magnetic
resonance we assume that γN is positive, as is the case for the
majority of nuclei. In such cases, it follows from eqn 47.4 that
states with mI < 0 lie above states with mI > 0. It follows that
the energy separation between the lower mI = + 12 (α) and α, mI = +½

upper mI = − 12 (β) states of a spin- 21 nucleus, a nucleus with


Figure 47.1 The nuclear spin energy levels of a spin- 21 nucleus
I = 12 , is
with positive magnetogyric ratio (for example, 1H or 13C) in a
magnetic field. Resonance occurs when the energy separation
1 ⎛ 1 ⎞
ΔE = E−1/2 − E+1/2 = γ N B0 − ⎜ − γ N B0 ⎟ = γ N B0 (47.5) of the levels matches the energy of the photons in the
2 ⎝ 2 ⎠ electromagnetic field.

and resonant absorption occurs when the resonance condition It follows that resonance absorption by spin- 21 nuclei occurs
(Fig. 47.1) when the Larmor precession frequency is the same as the
γ N B0 Resonance frequency of the applied electromagnetic field.
h␯ = γ N B0 or ␯ = Spin-1/2 nuclei
condition (47.6)

(b) The NMR spectrometer


is fulfilled. At resonance there is strong coupling between the
spins and the radiation, and absorption occurs as the spins In its simplest form, NMR is the study of the properties of
flip from the lower energy state to the upper state. The electro- molecules containing magnetic nuclei by applying a magnetic
magnetic field also stimulates the spins to flip from the upper field and observing the frequency of the resonant electromag-
state to the lower state, with the emission of radiation. The net netic field. Larmor frequencies of nuclei at the fields normally
absorption is the difference of these two processes, as we dis- employed (about 12 T) typically lie in the radiofrequency region
cuss in more detail shortly. of the electromagnetic spectrum (close to 500 MHz), so NMR
is a radiofrequency technique. For much of our discussion we
Brief illustration 47.1 consider spin- 21 nuclei, but NMR is applicable to nuclei with any
The resonance condition in NMR
nonzero spin. As well as protons, which are the most common
To calculate the frequency at which radiation comes into reso- nuclei studied by NMR, spin- 21 nuclei include 13C, 19F, and 31P.
nance with proton (s = 12 ) spins in a 12.0 T magnetic field we An NMR spectrometer consists of an appropriate source of
use eqn 47.6 as follows: radiofrequency radiation and a magnet that can produce a uni-
 γN
   form, intense field. Most modern instruments use a supercon-
B0
(2.6752 × 108 T−1s −1 ) × (12.0 T) ducting magnet capable of producing fields of the order of 10 T
␯= = 5.11 × 108 s −1 = 511MHz and more (Fig. 47.3). The sample is rotated rapidly to average

out magnetic inhomogeneities; however, although sample spin-
Self-test 47.1 Determine the resonance frequency for 31 P
ning is essential for the investigation of small molecules, for
nuclei, for which γ N = 1.0841 × 10 8 T−1 s −1, under the same large molecules it can lead to irreproducible results and is often
conditions.
Answer: 207 MHz
z

mI = +½
It is sometimes useful to compare the quantum mechani-
cal and classical pictures of magnetic nuclei pictured as tiny
bar magnets. A bar magnet in an externally applied magnetic mI = –½
field undergoes the motion called precession as it twists round
the direction of the field (Fig. 47.2). The rate of precession νL is
called the Larmor precession frequency: Figure 47.2 The interactions between the mI states of a spin- 21
nucleus and an external magnetic field may be visualized as the
γ N B0
␯L = Definition Larmor frequency of a nucleus (47.7) precession of the vectors representing the angular momentum.

Atkins09819.indb 459 9/11/2013 12:07:16 PM


460 10 Magnetic resonance

Justification 47.1 Intensities in NMR spectra


Superconducting
magnet As indicated in the text, at resonance both absorption and
emission are stimulated. From the general considerations
Computer of transition intensities in Topic 16, we know that the rate of
Probe
absorption of electromagnetic radiation is proportional to
the population of the lower-energy state (N α in the case of a
Preamplifier Receiver Detector Transmitter proton NMR transition) and the rate of stimulated emission is
proportional to the population of the upper state (Nβ). At the
low frequencies typical of magnetic resonance, spontaneous
Figure 47.3 The layout of a typical NMR spectrometer. The link emission can be neglected as it is very slow. Therefore, the net
from the transmitter to the detector indicates that the high rate of absorption is proportional to the difference in popula-
frequency of the transmitter is subtracted from the high-frequency tions, and we can write
received signal to give a low-frequency signal for processing.
Rate of absorption ∝ N α − N β

avoided. Although a superconducting magnet (Topic 39) oper- The intensity of absorption, the rate at which energy is
ates at the temperature of liquid helium (4 K), the sample itself is absorbed, is proportional to the product of the rate of absorp-
normally at room temperature or held in a variable-temperature tion (the rate at which photons are absorbed) and the energy
enclosure between, typically, –150 and +100 °C. of each photon, and the latter is proportional to the frequency
Modern NMR spectroscopy uses pulses of radiofrequency ν of the incident radiation (through E = hν). At resonance,
radiation. These techniques of Fourier-transform (FT) NMR this frequency is proportional to the applied magnetic field
make possible the determination of structures of very large (through ν = ν L = γ NB0/2π), so we can write
molecules in solution and in solids (Topic 49). Rate of absorption ∝ (N α − N β )B0
The intensity of an NMR transition depends on a number of
factors. We show in the following Justification that as in eqn 47.8a. To write an expression for the population dif-
Intensity ∝(N α − N β )B0 (47.8a) ference, we use the Boltzmann distribution (Foundations,
Topic 2, and Topic 51) to write the ratio of populations as
where ΔE e − x =1

− x +…

Nβ ≈ γ N B0
= e −γ N B0 /kT 1−
Nγ N B0 Nα kT
Nα − Nβ ≈ Nuclei Population difference (47.8b)
2kT
The expansion of the exponential term is appropriate for
with N the total number of spins (N = Nα + Nβ). It follows that ΔE = γ N B0 kT , a condition usually met for nuclear spins.
decreasing the temperature increases the intensity by increas- It follows that
ing the population difference. 1−γ N B0 /kT

N α − N β N α (1 − N β /N α ) 1 − N β /N α
= =
Brief illustration 47.2 Nuclear spin populations N α + N β N α (1 + N β /N α ) 1 + N β /N α

  
N 1−γ N B0 /kT
For protons γ N = 2.675 × 10 8 T−1 s −1. Therefore, for 1 000 000 1 − (1 − γ N B0 /kT ) γ N B0 /kT
protons in a field of 10 T at 20 °C ≈ =
1 + (1 − γ N B0 /kT ) 2

γN
  
N       

B0 ≈1

1000000 × (2.675 ×10 T s ) × (1.055 ×10 Js) × (10T)


8 −1 −1 −34
Nα − Nβ ≈ which is eqn 47.8b.
2 × (
1.381
 ×10

−23

−1
JK ) × (
293K )
k T
≈35 By combining eqns 47.8a and 47.8b we see that the intensity
is proportional to B02 , so NMR transitions can be enhanced
Even in such a strong field there is only a tiny imbalance of significantly by increasing the strength of the applied mag-
population of about 35 in a million. netic field. The use of high magnetic fields also simplifies
Self-test 47.2 For 13C nuclei, γ N = 6.7283 × 107 T−1 s −1. Determine the appearance of spectra (Topic 48) and so allows them
the magnetic field necessary to induce the same imbalance in to be interpreted more readily. We can also conclude that
the distribution of 13C spins at 20 °C. absorptions of nuclei with large magnetogyric ratios (1H, for
Answer: 40 T, an unrealistically high field for an NMR spectrometer instance) are more intense than those with small magneto-
gyric ratios (13C, for instance).

www.ebook3000.com

Atkins09819.indb 460 9/11/2013 12:07:34 PM


47 General principles  461

47.2 Electron paramagnetic resonance degeneracy is removed: the state with ms = + 12 moves up in
energy by 12 g e μBB0 . and the state with ms = − 12 moves down by
Electron paramagnetic resonance (EPR), or electron spin 1
g μ B From eqn 47.11b, the separation between the (upper)
2 e B 0
resonance (ESR), is the study of molecules and ions contain- ms = + 12 and (lower) ms = − 12 levels of an electron spin in a
ing unpaired electrons by observing the magnetic field at which magnetic field of magnitude B0 in the z-direction is
they come into resonance with radiation of known frequency.
As we have done for NMR, we write expressions for the reso- 1 ⎛ 1 ⎞
ΔE = E+1/2 − E−1/2 = g e μBB0 − ⎜ − g e μBB0 ⎟ = g e μBB0 (47.12a)
nance condition in EPR and then describe the general features 2 ⎝ 2 ⎠
of EPR spectrometers.
If the sample is exposed to radiation of frequency ν, the energy
separations come into resonance with the radiation when the
(a) The energies of electrons in magnetic fields frequency satisfies the resonance condition (Fig. 47.4)
The spin magnetic moment of an electron, which has a spin
quantum number s = 12 (Topic 19), is proportional to its spin h␯ = g e μBB0 Electrons Resonance condition (47.12b)
angular momentum. The spin magnetic moment and hamilto-
nian operators are, respectively, At resonance there is strong coupling between the electron
spins and the radiation, and strong absorption occurs as the
μ = γ e s and  = −γ B ⋅ s
H (47.9a) spins make the transition α ← β. As for NMR, at resonance the
e
opposite transition also occurs, and the detected signal is the net
where s is the spin angular momentum operator, and γe is the outcome taking into account the relative populations of the levels,
magnetogyric ratio of the electron: as we explore in more detail shortly.
g ee
γe =− Electrons Magnetogyric ratio (47.9b) Brief illustration 47.3 The resonance condition in EPR
2me
Magnetic fields of about 0.30 T (the value used in most com-
with ge = 2.002 319… as the g-value of the electron. Dirac’s
mercial EPR spectrometers) correspond to resonance at
relativistic theory (his modification of the Schrödinger equa-
μB
tion to make it consistent with Einstein’s special relativity)   
g e     
B0 

gives ge = 2; the additional 0.002 319… arises from interactions (2.0023) × (9.274 ×10 JT ) × (0.30 T)
−24 −1
␯=
of the electron with the electromagnetic fluctuations of the 6.626 ×10−34 Js

vacuum that surrounds the electron. The negative sign of γe h
(arising from the sign of the electron’s charge) shows that the = 8.4 ×109 s −1 = 8.4 GHz
orbital moment is opposite in direction to the orbital angular
momentum vector. which corresponds to a wavelength of 3.6 cm.
For a magnetic field of magnitude B0 in the z-direction,
Self-test 47.3 Determine the magnetic field for EPR transi-
 = −γ eB0 sz tions occurring at λ = 0.88 cm.
H (47.10)
Answer: 1.2 T

Because the eigenvalues of the operator sz are msħ with


ms = + 12 (α) and ms = − 12 (β), it follows that the energies of an α, ms = +½
electron spin in a magnetic field are
Magnetic Magnetic
Em = −γ e B0 ms Energies of an electron
(47.11a) field off field on
s
spin in a magnetic field
geμBB0
and can also be expressed in terms of the Bohr magneton, μB, as

e Energies of an electron
Em = g e μBB0 ms μB = spin in a magnetic field (47.11b)
s
2me β, ms = –½

where me is the mass of the electron and μB = 9.274 × 10−24JT−1. Figure 47.4 Electron spin levels in a magnetic field. Note that
The Bohr magneton, a positive quantity, is often regarded as the the β state is lower in energy than the α state (because the
fundamental quantum of magnetic moment. magnetogyric ratio of an electron is negative). Resonance is
In the absence of a magnetic field, the states with differ- achieved when the frequency of the incident radiation matches
ent values of ms are degenerate. When a field is present, the the frequency corresponding to the energy separation.

Atkins09819.indb 461 9/11/2013 12:07:56 PM


462 10 Magnetic resonance

(b) The EPR spectrometer


Absorption, A
It follows from Brief illustration 47.3 that most commercial EPR
spectrometers operate at wavelengths of approximately 3 cm.
Because 3 cm radiation falls in the microwave region of the

Signal
Slope
electromagnetic spectrum, EPR is a microwave technique. Slope
Both Fourier-transform (FT) and continuous wave (CW)
EPR spectrometers are available. The FT-EPR instrument is
Field, B
based on the concepts developed in Topic 49 for NMR spec-
Derivative
troscopy, except that pulses of microwaves are used to excite
of absorption, dA/dB
electron spins in the sample. The layout of the more common
CW-EPR spectrometer is shown in Fig. 47.5. It consists of a Figure 47.7 When phase-sensitive detection is used, the signal
microwave source (a klystron or a Gunn oscillator), a cavity in is the first derivative of the absorption intensity. Note that the
which the sample is inserted in a glass or quartz container, a peak of the absorption corresponds to the point where the
microwave detector, and an electromagnet with a field that can derivative passes through zero.
be varied in the region of 0.3 T. The EPR spectrum is obtained
by monitoring the microwave absorption as the field is changed,
and a typical spectrum (of the benzene radical anion, C 6 H6− ) is which is in fact the first derivative of the absorption, arises
shown in Fig. 47.6. The peculiar appearance of the spectrum, from the detection technique, which is sensitive to the slope of
the absorption curve (Fig. 47.7).
As we have indicated, the intensities of spectral lines depend
Microwave Detector on the difference in populations between the ground and
source
excited states. For an electron, the β state lies below the α state
Sample in energy and, by a similar argument to that for nuclei,
cavity
Ng e μBB0
Phase- Nβ − Nα ≈ Electrons Population difference (47.13)
sensitive 2kT
detector
where N is the total number of spins.

Electromagnet
Brief illustration 47.4 Electron spin populations
Modulation
When 1000 electron spins are exposed to a 1.0 T magnetic field
unit
at 20 °C (293 K),
Figure 47.5 The layout of a continuous-wave EPR μB
spectrometer. A typical magnetic field is 0.3 T, which requires
N  ge
    
B0

1000 × 2.0023 × (9.274 ×10−24 JT−1 ) × (1.0 T)


9 GHz (3 cm) microwaves for resonance. N β − Nα ≈
2 × (1.381 × 10−23 JK −1 ) × (293 K )
  
k T
≈ 2.3
There is an imbalance of populations of only about two elec-
trons in a thousand. However the imbalance is much larger for
electron spins than for nuclear spins (Brief illustration 47.2)
because the energy separation between the spin states of elec-
trons is larger than that for nuclear spins, even at the lower
magnetic field strengths normally employed.
Field strength
Self-test 47.4 It is common to conduct EPR experiments at very
low temperatures. At what temperature would the imbalance
in spin populations be 5 electrons in 100, with B0 = 0.30 T?
Figure 47.6 The EPR spectrum of the benzene radical anion, Answer: 4 K
C6H6− , in fluid solution.

www.ebook3000.com

Atkins09819.indb 462 9/11/2013 12:08:04 PM


47 General principles  463

Checklist of concepts
☐ 1. The nuclear spin quantum number, I, of a nucleus is a ☐ 5. The resonance absorption intensity increases with the
non-negative integer or half-integer. strength of the applied magnetic field (as B02 ).
☐ 2. Nuclei with different values of m I have different ener- ☐ 6. Electrons with different values of m s have different
gies in the presence of a magnetic field. energies in the presence of a magnetic field.
☐ 3. Nuclear magnetic resonance (NMR) is the observation ☐ 7. Electron paramagnetic resonance (EPR) is the observa-
of the resonant absorption of electromagnetic radiation tion of the resonant absorption of electromagnetic radi-
by nuclei in a magnetic field. ation by unpaired electrons in a magnetic field.
☐ 4. NMR spectrometers consist of a source of radiofre- ☐ 8. EPR spectrometers consist of a microwave source, a
quency radiation and a magnet that provides a strong, cavity in which the sample is inserted, a microwave
uniform field. detector, and an electromagnet.

Checklist of equations
Property Equation Comment Equation number

Nuclear magneton μN = e /2mp 47.4b

Energies of a nuclear spin in a magnetic field EmI = −γ N B0mI 47.4c


= − g I μNB0mI

Resonance condition (spin- 12 nuclei) h␯ =γ N B0 γN > 0 47.6

Larmor frequency ␯L =γ NB0 /2π γN > 0 47.7

Population difference (nuclei) N α − N β ≈ Nγ N B0 /2kT 47.8b

Magnetogyric ratio (electron) γ e = − g ee /2me ge = 2.002 319 47.9b

Energies of an electron spin in a magnetic field Ems = −γ e B0ms 47.11a


= g e μBB0ms

Bohr magneton μB = e /2me 47.11b

Resonance condition (electrons) h␯ = g e μBB0 47.12

Population difference (electrons) N β − N α ≈ Ng e μBB0 /2kT 47.13

Atkins09819.indb 463 9/11/2013 12:08:14 PM


TOPIC 48

Features of NMR spectra

Contents ➤ What is the key idea?


48.1 The chemical shift 464 The resonance frequency of a magnetic nucleus is
Brief illustration 48.1: The δ scale 465 affected by its electronic environment and the presence
Example 48.1: Interpreting the NMR spectrum of magnetic nuclei in its vicinity.
of ethanol 465
➤ What do you need to know already?
48.2 The origin of shielding constants 466
(a) The local contribution 466 You need to be familiar with the general principles of
Example 48.2: Using the Lamb formula 467 magnetic resonance (Topic 47).
(b) Neighbouring group contributions 467
Brief illustration 48.2: Ring currents 468
(c) The solvent contribution 469 Nuclear magnetic moments interact with the local magnetic
Brief illustration 48.3: The effect of aromatic field. The local field may differ from the applied field because
solvents 469 the latter induces electronic orbital angular momentum (that
48.3 The fine structure 469 is, the circulation of electronic currents) which gives rise to a
(a) The appearance of the spectrum 469 small additional magnetic field δB at the nuclei. This additional
Example 48.3: Accounting for the fine structure field is proportional to the applied field, and it is conventional
in a spectrum 471 to write
(b) The magnitudes of coupling constants 472
Brief illustration 48.4: The Karplus equation 472 δB = −σ B0 Definition Shielding constant (48.1)
(c) The origin of spin–spin coupling 472
Brief illustration 48.5: Magnetic fields from nuclei 472 where the dimensionless quantity σ is called the shielding con-
48.4 Conformational conversion and exchange stant of the nucleus (σ is usually positive but may be negative).
processes 474 The ability of the applied field to induce an electronic current
Brief illustration 48.6: The effect of chemical in the molecule, and hence affect the strength of the resulting
exchange on NMR spectra 474 local magnetic field experienced by the nucleus, depends on the
Checklist of concepts 475 details of the electronic structure near the magnetic nucleus
Checklist of equations 475 of interest, so nuclei in different chemical groups have differ-
ent shielding constants. The calculation of reliable values of the
shielding constant is very difficult, but trends in it are quite well
understood and we concentrate on them.

➤ Why do you need to know this material?


To make progress with the analysis of NMR spectra and 48.1 The chemical shift
extract the wealth of information they contain you need to
understand how the appearance of a spectrum correlates Because the total local field Bloc is
with molecular structure.
Bloc = B0 + δB = (1− σ )B0 (48.2)

www.ebook3000.com

Atkins09819.indb 464 9/11/2013 12:08:22 PM


48 Features of NMR spectra  465

the nuclear Larmor frequency is


Self-test 48.1 What is the shift of the resonance from TMS of
a group of nuclei with δ = 3.50 and an operating frequency of
γ N Bloc γ N B0
␯L = = (1− σ ) (48.3) 350 MHz?
2π 2π
Answer: 1.23 kHz

This frequency is different for nuclei in different environments.


Hence, different nuclei, even of the same element, come into The relation between δ and σ is obtained by substituting eqn
resonance at different frequencies if they are in different molec- 48.3 into eqn 48.4:
ular environments.
The chemical shift of a nucleus is the difference between (1− σ )B0 − (1− σ °)B0
δ= ×106
its resonance frequency and that of a reference standard. The (1− σ °)B0 Relation between
(48.6)
standard for protons is the proton resonance in tetramethylsi- σ ° −σ δ and σ
= ×106 ≈ (σ ° − σ ) ×106
lane, Si(CH3)4, commonly referred to as TMS, which bristles 1− σ °
with protons and dissolves without reaction in many solutions.
For 13C, the reference frequency is the 13C resonance in TMS, The last line follows from σ ° 1. As the shielding σ gets
and for 31P it is the 31P resonance in 85 per cent H3PO4(aq). smaller, δ increases. Therefore, we speak of nuclei with large
Other references are used for other nuclei. The separation of chemical shifts as being strongly deshielded. Some typical
the resonance of a particular group of nuclei from the stand- chemical shifts are given in Fig. 48.1. As can be seen from the
ard increases with the strength of the applied magnetic field illustration, the nuclei of different elements have very differ-
because the induced field is proportional to the applied field, ent ranges of chemical shifts. The ranges exhibit the variety of
and the stronger the latter, the greater the shift. electronic environments of the nuclei in molecules: the higher
Chemical shifts are reported on the δ scale, which is the atomic number of the element, the greater the number of
defined as electrons around the nucleus and hence the greater the range of
the extent of shielding. By convention, NMR spectra are plotted
␯ − ␯° with δ increasing from right to left.
δ= ×106 Definition δ scale (48.4)
␯°

where ν° is the resonance frequency of the standard. The advan- RCH3


tage of the δ scale is that shifts reported on it are independent of –CH –
R–NH 2
the applied field (because both numerator and denominator are RC–CH3 –CH– 2
ArC–CH3
proportional to the applied field). The resonance frequencies –CO–CH3
–C=CH– ROH
themselves, however, do depend on the applied field through (a) ArOH
Ar–H
–COOH –CHO
␯ = ␯° + (␯° /106 )δ (48.5) 12 10 8 6 4 2 0
δ
R–C–H
Brief illustration 48.1 The δ scale >C=C< R3C–
X
–C=C– –C=C<
A nucleus with δ = 1.00 in a spectrometer where ν° = 500 MHz C–X in ArX

R–C=N
(a ‘500 MHz NMR spectrometer’) will have a shift relative to R–CHO R–COOH
R2C=O
the reference equal to R=C=R
(b) R3C+
300 200 100 0
␯ − ␯° = (500 MHz /10 ) × 1.00 = (500 Hz) × 1.00 = 500 Hz
6 δ

because 1 MHz = 10 6 Hz. In a spectrometer operating at Figure 48.1 The range of typical chemical shifts for (a) 1H
ν° = 100 MHz, the shift relative to the reference would be only resonances and (b) 13C resonances.
100 Hz.
A note on good practice In much of the literature, chemi- Example 48.1 Interpreting the NMR spectrum
cal shifts are reported in parts per million, ppm, in of ethanol
recognition of the factor of 106 in the definition; this is
unnecessary. If you see ‘δ = 10 ppm’, interpret it, and use it Figure 48.2 shows the NMR spectrum of ethanol. Account for
in eqn 48.5, as δ = 10. the observed chemical shifts.

Atkins09819.indb 465 9/11/2013 12:08:29 PM


466 10 Magnetic resonance

electron density in the ground and excited states and the exci-
tation energies of the molecule. Nevertheless, considerable
success has been achieved with small molecules such as H2O
CH3CH2OH CH3CH2OH CH3CH2OH and CH4 and even large molecules, such as proteins, are within
the scope of some types of calculation. However, it is easier to
understand the different contributions to chemical shifts by
studying the large body of empirical information now available.
The empirical approach supposes that the observed shielding
constant is the sum of three contributions:

σ = σ (local) + σ (neighbour) + σ (solvent) (48.7)


δ
4.0 3.6 1.2
The local contribution, σ(local), is essentially the contribution
Figure 48.2 The 1H-NMR spectrum of ethanol. The bold of the electrons of the atom that contains the nucleus in ques-
letters denote the protons giving rise to the resonance peak, tion. The neighbouring group contribution, σ(neighbour), is
and the step-like curve is the integrated signal. the contribution from the groups of atoms that form the rest of
the molecule. The solvent contribution, σ(solvent), is the con-
Method Consider the effect of an electron-withdrawing atom:
tribution from the solvent molecules.
it deshields strongly those protons to which it is bound, and
has a smaller effect on distant protons.
Answer The spectrum is consistent with the following (a) The local contribution
assignments: It is convenient to regard the local contribution to the shielding
r The CH3 protons form one group of nuclei with δ = 1.2. constant as the sum of a diamagnetic contribution, σd, and a
paramagnetic contribution, σp:
r The two CH2 protons are in a different part of the
molecule, experience a different local magnetic field,
Local contribution
and resonate at δ = 3.6. σ (local) = σ d + σ p to the shielding (48.8)
r The OH proton is in another environment, and has a constant
chemical shift of δ = 4.0.
A diamagnetic contribution to σ(local) opposes the applied
The increasing value of δ (that is, the decrease in shielding) magnetic field and shields the nucleus in question. A paramag-
is consistent with the electron-withdrawing power of the O netic contribution to σ(local) reinforces the applied magnetic
atom: it reduces the electron density of the OH proton most, field and deshields the nucleus in question. Therefore, σd > 0
and that proton is strongly deshielded. It reduces the electron and σp < 0. The total local contribution is positive if the diamag-
density of the distant methyl protons least, and those nuclei netic contribution dominates, and is negative if the paramag-
are least deshielded. netic contribution dominates.
The relative intensities of the signals are commonly repre- The diamagnetic contribution arises from the ability of the
sented as the height of step-like curves superimposed on the applied field to generate a circulation of charge in the ground-
spectrum, as in Fig. 48.2. In ethanol the group intensities are state electron distribution of the atom. The circulation gener-
in the ratio 3:2:1 because there are three CH3 protons, two CH2 ates a magnetic field that opposes the applied field and hence
protons, and one OH proton in each molecule. shields the nucleus. The magnitude of σd depends on the elec-
Self-test 48.2 The NMR spectrum of acetaldehyde (ethanal) has tron density close to the nucleus and can be calculated from the
lines at δ = 2.20 and δ = 9.80. Which feature can be assigned to Lamb formula:1
the CHO proton?
Answer: δ = 9.80 e 2 μ0 1
σd = 〈 〉 Lamb formula (48.9)
12πme r

where μ0 is the vacuum permeability (a fundamental con-


48.2 The origin of shielding constants stant, see inside the front cover) and r is the electron–nucleus
distance.
The calculation of shielding constants is difficult, even for
small molecules, for it requires detailed information (using the 1 For a derivation, see our Molecular quantum mechanics, Oxford Univer-

techniques outlined in Topics 28–30) about the distribution of sity Press, Oxford (2011).

www.ebook3000.com

Atkins09819.indb 466 9/11/2013 12:08:33 PM


48 Features of NMR spectra  467

1
Example 48.2 Using the Lamb formula

Chemical shift relative to CH4, δ


Calculate the shielding constant for the proton in a free CH3CH2X
H atom. 2
Method To calculate σ d from the Lamb formula, calcu-
late the expectation value of 1/r for a hydrogen 1s orbital. CH3CH2X
Wavefunctions are given in Table 18.1. 3

Answer The wavefunction for a hydrogen 1s orbital is


1/2
⎛ 1 ⎞ 4
I Br Cl F
ψ =⎜ 3 ⎟ e − r /a0 2 2.5 3 3.5 4
⎝ πa0 ⎠ Electronegativity of halogen, χ

so, because dτ = r2dr sinθ dθdφ, the expectation value of 1/r is Figure 48.3 The variation of chemical shielding with
written as electronegativity. The shifts for the methyl protons agree with
the trend expected with increasing electronegativity. However,
〈 1r 〉 = ∫ ψ rψ dτ =
* 2π π ∞
1
πa03 ∫
0
dφ ∫
0
sinθ dθ ∫
0
re −2r /a0 dr to emphasize that chemical shifts are subtle phenomena,
notice that the trend for the methylene protons is opposite
a02 /4 ( Integral E.1)
 to that expected. For these protons another contribution (the

4 1

magnetic anisotropy of CeH and CeX bonds) is dominant.
= 3 re −2r /a0 dr =
a0 0 a0

where we used the integral listed in the Resource section. the molecule by making use of orbitals that are unoccupied in
Therefore, the ground state. It is zero in free atoms and around the axes of
linear molecules (such as ethyne, HCbCH) where the electrons
⎛ kg m2 s −2

⎞ can circulate freely and a field applied along the internuclear
(1.602 ×10−19 C)2 × ⎜ 4 π ×10−7 J s2 C −2 m −1 ⎟ axis is unable to force them into other orbitals. We can expect
⎜ ⎟
e 2 μ0 ⎝ ⎠ large paramagnetic contributions from small atoms (because
σd = =
12πmea0 12π × (9.109 ×10−31 kg ) × (5.292 ×10−11 m) the induced currents are then close to the nucleus) in molecules
= 1.775 ×10−5 with low-lying excited states (because an applied field can then
induce significant currents). In fact, the paramagnetic contri-
Self-test 48.3 Derive a general expression for σ d that applies to bution is the dominant local contribution for atoms other than
all hydrogenic atoms. hydrogen.
Answer: Ze2μ0/12πmea0

(b) Neighbouring group contributions


The diamagnetic contribution is the only contribution in The neighbouring group contribution arises from the currents
closed-shell free atoms. It is also the only contribution to the induced in nearby groups of atoms. Consider the influence
local shielding for electron distributions that have spherical of the neighbouring group X on the proton H in a molecule
or cylindrical symmetry. Thus, it is the only contribution to such as HeX. The applied field generates currents in the elec-
the local shielding from inner cores of atoms, for cores remain tron distribution of X and gives rise to an induced magnetic
nearly spherical even though the atom may be a component of a moment proportional to the applied field; the constant of pro-
molecule and its valence electron distribution highly distorted. portionality is the magnetic susceptibility, χ (chi), of the group
The diamagnetic contribution is broadly proportional to the X: μinduced = χB0 (Topic 39). The susceptibility is negative for a
electron density of the atom containing the nucleus of interest. diamagnetic group because the induced moment is opposite to
It follows that the shielding is decreased if the electron density the direction of the applied field. The induced moment gives
on the atom is reduced by the influence of an electronegative rise to a magnetic field with a component parallel to the applied
atom nearby. That reduction in shielding as the electronegativ- field and at a distance r and angle θ (1) that has the form (The
ity of a neighbouring atom increases translates into an increase chemist’s toolkit 48.1)
in the chemical shift δ (Fig. 48.3).
The local paramagnetic contribution, σp, arises from the abil- μinduced
Blocal ∝ (1− 3 cos2θ ) Local dipolar field (48.10a)
ity of the applied field to force electrons to circulate through r3

Atkins09819.indb 467 9/11/2013 12:08:43 PM


468 10 Magnetic resonance

We see that the strength of the additional magnetic field expe-


with z = r cos θ, the z-component of the distance vector r. If the
rienced by the proton is inversely proportional to the cube of magnetic dipole is also parallel to the z-direction, it follows
the distance r between H and X. If the magnetic susceptibility is that
independent of the orientation of the molecule (is ‘isotropic’),
because 1 − 3 cos2 θ is zero when averaged over a sphere (see ⎛μ  μ⋅r

 
z  ⎞
μ0 ⎜
z
3 ( μr cosθ )(r cosθ ) ⎟ μμ0
Problem 48.6) the local field averages to zero. To a good Bz = μ− ⎟ = 4 πr 3 (1 − 3 cos θ )
2

approximation, the shielding constant σ(neighbour) depends 4 πr 3 ⎜⎜ r2 ⎟


on the distance r as ⎝ ⎠

Neigh-
⎛ 1− 3 cos2 θ ⎞
σ (neighbour)∝ (χ  − χ ⊥ ) ⎜ ⎟⎠
bouring
(48.10b)
⎝ r3 group
contribution –

where χ and χ⊥ are, respectively, the parallel and perpendicu-


lar components of the magnetic susceptibility, and θ is the angle
μ +
between the XeH axis and the symmetry axis of the neigh- +
bouring group (2). Equation 48.10 shows that the neighbour-
ing group contribution may be positive or negative according
to the relative magnitudes of the two magnetic susceptibilities

and the relative orientation of the nucleus with respect to X. If
54.7° < θ < 125.3°, then 1 − 3 cos2 θ is positive, but it is negative
otherwise (Figs 48.4 and 48.5).
Figure 48.4 A depiction of the field arising from a point
θ magnetic dipole. The three shades of colour represent the
θ
r r strength of field declining with distance (as 1/r3), and each
H surface shows the angle dependence of the z-component of
μinduced χ||
χ⊥
the field for each distance.
C

1 2 1

0
The chemist’s toolkit 48.1
1 – 3 cos2 θ

Dipolar fields
Standard electromagnetic theory gives the magnetic field at a
point r from a point magnetic dipole μ as

μ0 ⎛ 3( μ ⋅ r )r ⎞
B= μ−
4 πr 3 ⎜⎝ r 2 ⎟⎠
–2
0 ½π π
where μ0 is the vacuum permeability (a fundamental constant θ
with the defined value 4π × 10−7 T2 J−1 m3). The electric field due
to a point electric dipole is given by a similar expression: Figure 48.5 The variation of the function 1 − 3 cos2 θ with the
angle θ.
1 ⎛ 3( μ ⋅ r )r ⎞
E= μ−
4 πε 0r 3 ⎜⎝ r 2 ⎟⎠ Brief illustration 48.2 Ring currents

where ε 0 is the vacuum permittivity, which is related to μ0 by A special case of a neighbouring group effect is found in aro-
ε0 = 1/μ0c2. The component of magnetic field in the z-direction is matic compounds. The strong anisotropy of the magnetic sus-
ceptibility of the benzene ring is ascribed to the ability of the
μ0 ⎛ 3( μ ⋅ r )z ⎞ field to induce a ring current, a circulation of electrons around
Bz = μ −
3 ⎜ z
r 2 ⎟⎠
the ring, when it is applied perpendicular to the molecular
4 πr ⎝
plane. Protons in the plane are deshielded (Fig. 48.6), but any

www.ebook3000.com

Atkins09819.indb 468 9/11/2013 12:09:01 PM


48 Features of NMR spectra  469

B
Ring
current

Magnetic
field

Figure 48.7 An aromatic solvent (benzene here) can give


rise to local currents that shield or deshield a proton in a
Figure 48.6 The shielding and deshielding effects of the solute molecule. In this relative orientation of the solvent
ring current induced in the benzene ring by the applied and solute, the proton on the solute molecule is shielded.
field. Protons attached to the ring are deshielded but a
proton attached to a substituent that projects above the Answer: Proton on the solute molecule coplanar
ring is shielded. with the benzene ring

that happen to lie above or below the plane (as members of


substituents of the ring) are shielded.
Self-test 48.4 Consider ethyne, HC b CH. Are its protons
shielded or deshielded by currents induced by the triple bond?
48.3 The fine structure
Answer: Shielded
The splitting of resonances into individual lines by spin–spin
coupling in Fig. 48.2 is called the fine structure of the spec-
trum. It arises because each magnetic nucleus may contribute
to the local field experienced by the other nuclei and so modify
(c) The solvent contribution their resonance frequencies. The strength of the interaction
is expressed in terms of the scalar coupling constant, J, and
A solvent can influence the local magnetic field experienced by reported in hertz (Hz). The scalar coupling constant is so called
a nucleus in a variety of ways. Some of these effects arise from because the energy of interaction it describes is proportional
specific interactions between the solute and the solvent (such as to the scalar product of the two interacting spins: E ∝ I1⋅I2.
hydrogen-bond formation and other forms of Lewis acid–base As explained in Mathematical background 4, a scalar product
complex formation). The anisotropy of the magnetic suscepti- depends on the angle between the two vectors, so writing the
bility of the solvent molecules, especially if they are aromatic, energy in this way is simply a way of saying that the energy of
can also be the source of a local magnetic field. Moreover, if interaction between two spins depends on their relative ori-
there are steric interactions that result in a loose but specific entation. The constant of proportionality in this expression is
interaction between a solute molecule and a solvent molecule, written hJ/ 2 (so E = (hJ/ 2)I1⋅I2); because each spin angular
then protons in the solute molecule may experience shielding momentum is proportional to , E is then proportional to hJ
or deshielding effects according to their location relative to the and J is a frequency (with units hertz). For nuclei that are con-
solvent molecule. strained to align with the applied field in the z-direction, the
only contribution to I1⋅I2 is I1zI2z, with eigenvalues m1m2 2, so
in that case the energy due to spin–spin coupling is
Brief illustration 48.3 The effect of aromatic solvents
Em m = hJm1m2 Spin–spin coupling energy (48.11)
An aromatic solvent like benzene can give rise to local cur- 1 2

rents that shield or deshield a proton in a solute molecule. The


arrangement shown in Fig. 48.7 leads to shielding of a proton
on the solute molecule. (a) The appearance of the spectrum
Self-test 48.5 Refer to Fig. 48.7 and suggest an arrangement In NMR, letters far apart in the alphabet (typically A and X)
that leads to deshielding of a proton on the solute molecule. are used to indicate nuclei with very different chemical shifts;
letters close together (such as A and B) are used for nuclei with

Atkins09819.indb 469 9/11/2013 12:09:05 PM


470 10 Magnetic resonance

similar chemical shifts. We shall consider first an AX system, a βAβX


molecule that contains two spin- 12 nuclei A and X with very dif-
ferent chemical shifts in the sense that the difference in chemi-
cal shift corresponds to a frequency that is large compared to J.
βAαX
For a spin- 12 AX system there are four spin states: αAαX,
αAβX, βAαX, βAβX. The energy depends on the orientation of αAβX
the spins in the external magnetic field, and if spin–spin cou-
pling is neglected αAαX

Em = −γ N (1− σ A )B0 mA − γ N (1− σ X )B0 mX


A mX
(48.12a) Figure 48.9 An alternative depiction of the energy levels
= −h␯ A mA − h␯ X mX and transitions shown in Fig. 48.8. Once again, we have
exaggerated the effect of spin–spin coupling.
where νA and νX are the Larmor frequencies of A and X and mA
and mX are their quantum numbers (mA = ± 12 , mX = ± 12 ). This
expression gives the four lines on the left of Fig. 48.8. When When a transition of nucleus A occurs, nucleus X remains
spin–spin coupling is included (by using eqn 48.11), the energy unchanged. Therefore, the A resonance is a transition for which
levels are ΔmA = +1 and ΔmX = 0. There are two such transitions, one in
which βA ← αA occurs when the X nucleus is α, and the other
Em A mX
= −h␯ A mA − h␯ X mX + hJmA mX (48.12b) in which βA ← αA occurs when the X nucleus is β. They are
shown in Fig. 48.8 and in a slightly different form in Fig. 48.9.
If J > 0, a lower energy is obtained when mAmX < 0, which is The energies of the transitions are
the case if one spin is α and the other is β. A higher energy
is obtained if both spins are α or both spins are β. The oppo- ΔE = h␯ A ± 12 hJ (48.13a)
site is true if J < 0. The resulting energy level diagram (for J > 0)
is shown on the right of Fig. 48.8. We see that the αα and ββ Therefore, the A resonance consists of a doublet of separation J
states are both raised by 14 hJ and that the αβ and βα states are centred on the chemical shift of A (Fig. 48.10). Similar remarks
both lowered by 14 hJ. apply to the X resonance, which consists of two transitions
according to whether the A nucleus is α or β (as shown in Fig.
48.9). The transition energies are
βAβX ¼hJ βAβX
½hνA + ½hνX ΔE = h␯ X ± 12 hJ (48.13b)
Energy

It follows that the X resonance also consists of two lines of the


½hνA – ½hνX same separation J, but they are centred on the chemical shift of
No spin–spin βAαX ¼hJ βAαX With X (as shown in Fig. 48.10).
coupling spin–spin
–½hνA + ½hνX coupling
αAβX αAβX
¼hJ
A resonance

X resonance

αAαX

αAαX ¼hJ
–½hνA – ½hνX

Figure 48.8 The energy levels of an AX system. The four


J J
levels on the left are those of the two spins in the absence of
spin–spin coupling. The four levels on the right show how a
positive spin–spin coupling constant affects the energies. The δA δX
transitions shown are for β ← α of A or X, the other nucleus (X
or A, respectively) remaining unchanged. We have exaggerated Figure 48.10 The effect of spin–spin coupling on an AX
the effect for clarity. In practice, the splitting caused by spin– spectrum. Each resonance is split into two lines separated by J.
spin coupling is much smaller than that caused by the applied The pairs of resonances are centred on the chemical shifts of
field. the protons in the absence of spin–spin coupling.

www.ebook3000.com

Atkins09819.indb 470 9/11/2013 12:09:26 PM


48 Features of NMR spectra  471

X resonance
in AX2

X resonance
in AX

δX δA

Figure 48.11 The X resonance of an AX2 species is also a Figure 48.13 The origin of the 1:3:3:1 quartet in the A
doublet, because the two equivalent X nuclei behave like a resonance of an AX3 species. The third X nucleus splits each of
single nucleus; however, the overall absorption is twice as the lines shown in Fig. 48.12 for an AX2 species into a doublet,
intense as that of an AX species. and the intensity distribution reflects the number of transitions
that have the same energy.

Three equivalent X nuclei (an AX3 species) split the reso-


nance of A into four lines of intensity ratio 1:3:3:1 (Fig. 48.13).
The X resonance remains a doublet as a result of the splitting
caused by A. In general, N equivalent spin- 12 nuclei split the
resonance of a nearby spin or group of equivalent spins into
N + 1 lines with an intensity distribution given by Pascal’s tri-
angle (3). Successive rows of this triangle are formed by adding
together the two adjacent numbers in the line above.

δA Example 48.3 Accounting for the fine structure


in a spectrum
Figure 48.12 The origin of the 1:2:1 triplet in the A resonance of
an AX2 species. The resonance of A is split into two by coupling Account for the fine structure in the NMR spectrum of the
with one X nucleus (as shown in the inset), and then each of CeH protons of ethanol.
those two lines is split into two by coupling to the second X
Method Consider how each group of equivalent protons (for
nucleus. Because each X nucleus causes the same splitting,
instance, three methyl protons) split the resonances of the
the two central transitions are coincident and give rise to an
other groups of protons. There is no splitting within groups of
absorption line of double the intensity of the outer lines.
equivalent protons. Each splitting pattern can be decided by
referring to Pascal’s triangle.
If there is another X nucleus in the molecule with the same
chemical shift as the first X (giving an AX2 species), the X reso- Answer The three protons of the CH 3 group split the reso-
nance of the AX2 species is split into a doublet by A, as in the nance of the CH 2 protons into a 1:3:3:1 quartet with a split-
AX case discussed above (Fig. 48.11). The resonance of A is ting J. Likewise, the two protons of the CH 2 group split the
split into a doublet by one X, and each line of the doublet is resonance of the CH3 protons into a 1:2:1 triplet with the same
split again by the same amount by the second X (Fig. 48.12). splitting J. The OH resonance is not split because the OH pro-
This splitting results in three lines in the intensity ratio 1:2:1 tons migrate rapidly from molecule to molecule (including
(because the central frequency can be obtained in two ways). molecules of impurities in the sample) and their effect aver-
ages to zero. In gaseous ethanol, where this migration does not
1
occur, the OH resonance appears as a triplet, showing that the
1 1 CH2 protons interact with the OH proton.
1 2 1
1 3 3 1 Self-test 48.6 What fine structure can be expected for the pro-
1 4 6 4 1
1 5 10 10 5 1 tons in 14 NH 4+? The spin quantum number of nitrogen-14 is 1.
Answer: 1:1:1 triplet from N
3

Atkins09819.indb 471 9/11/2013 12:09:32 PM


472 10 Magnetic resonance

(b) The magnitudes of coupling constants C = +4.3 Hz. In an investigation of the polypeptide flavodoxin,
the 3J HH coupling constant for such a grouping was deter-
The scalar coupling constant of two nuclei separated by N bonds mined to be 2.1 Hz. Is this value consistent with a helical or
is denoted NJ, with subscripts for the types of nuclei involved. sheet conformation?
Thus, 1JCH is the coupling constant for a proton joined directly Answer: Helical conformation
to a 13C atom, and 2JCH is the coupling constant when the same
two nuclei are separated by two bonds (as in 13CeCeH). A typi-
cal value of 1JCH is in the range 120 to 250 Hz; 2JCH is between 10 14

Spin–spin coupling constant, J/Hz


and 20 Hz. Both 3J and 4J can give detectable effects in a spec- 12 HCCH
trum, but couplings over larger numbers of bonds can generally
10
be ignored. One of the longest-range couplings that has been
detected is 9JHH = 0.4 Hz between the CH3 and CH2 protons in 8
CH3CbCeCbCeCbCeCH2OH.
6
As remarked (in the discussion following eqn 48.12b), the HNCH
sign of JXY indicates whether the energy of two spins is lower 4

when they are parallel (J < 0) or when they are antiparallel 2


(J > 0). It is found that 1JCH is often positive, 2JHH is often nega-
0
tive, 3JHH is often positive, and so on. An additional point is that 0 π/2 π
J varies with the angle between the bonds (Fig. 48.14). Thus, a Angle, φ
3J
HH coupling constant is often found to depend on the dihedral Figure 48.14 The variation of the spin–spin coupling constant
angle φ (4) according to the Karplus equation: with angle predicted by the Karplus equation for an HCCH
group and an HNCH group.
3
J HH = A + B cosφ + C cos 2φ Karplus equation (48.14)

H φ (c) The origin of spin–spin coupling


H
Spin–spin coupling is a very subtle phenomenon and it is better
to treat J as an empirical parameter than to use calculated val-
ues. However, we can get some insight into its origins, if not its
4 precise magnitude—or always reliably its sign—by considering
the magnetic interactions within molecules.
with A, B, and C empirical constants with values close to +7 Hz, A nucleus with spin projection mI gives rise to a magnetic
−1 Hz, and +5 Hz, respectively, for an HCCH fragment. It fol- field with z-component Bnuc at a distance R, where, to a good
lows that the measurement of 3JHH in a series of related com- approximation,
pounds can be used to determine their conformations. The
coupling constant 1JCH also depends on the hybridization of γ N μ 0
Bnuc = − (1− 3 cos2 θ )mI (48.15)
the C atom, as the following values indicate: 4 πR3

sp sp2 sp3 The angle θ is defined in 1; we saw a version of this expression


1J
in eqn 48.10a.
CH/Hz 250 160 125

Brief illustration 48.5 Magnetic fields from nuclei


Brief illustration 48.4 The Karplus equation
The z-component of the magnetic field arising from a proton
The investigation of HeNe CeH couplings in polypeptides (mI = 12 ) at R = 0.30 nm, with its magnetic moment parallel to
can help reveal their conformation. For 3JHH coupling in such the z-axis (θ = 0) is
a group, A = +5.1 Hz, B = −1.4 Hz, and C = +3.2 Hz. For a helical
polymer, φ is close to 120°, which would give 3JHH≈ 4 Hz. For  γN
   μ0
  (1−3 cos2 θ )m
(2.821 ×10−26 JT−1 ) × (4 π ×10−7 T2 J −1 m3 )
I
the sheet-like conformation, φ is close to 180°, which would Bnuc = − × ( −1)
give 3JHH ≈ 10 Hz. 4 π × (3.0 ×10−10 m)3

R
Self-test 48.7 NMR experiments reveal that for HeCeCeH
= 1.0 ×10−4 T = 0.10mT
coupling in polypeptides, A = +3.5 Hz, B = −1.6 Hz, and

www.ebook3000.com

Atkins09819.indb 472 9/11/2013 12:09:43 PM


48 Features of NMR spectra  473

Fermi Pauli Fermi


A field of this magnitude can give rise to the splitting of
resonance signals in solid samples. In a liquid, the angle θ
sweeps over all values as the molecule tumbles, and the factor X Y
1 − 3 cos2θ averages to zero. Hence the direct dipolar interac-
tion between spins cannot account for the fine structure of the
spectra of rapidly tumbling molecules.
X Y
Self-test 48.8 In gypsum the splitting in the H 2O resonance
can be interpreted in terms of a magnetic field of 0.715 mT
generated by one proton and experienced by the other. Figure 48.16 The polarization mechanism for spin–spin
With θ = 0, what is the separation of the protons in the H 2O coupling (1JHH). The two arrangements have slightly different
molecule? energies. In this case, J is positive, corresponding to a lower
Answer: 158 pm energy when the nuclear spins are antiparallel.

the failure of the point-dipole approximation—depends on the


Spin–spin coupling in molecules in solution can be explained very close approach of an electron to the nucleus and hence can
in terms of the polarization mechanism, in which the inter- occur only if the electron occupies an s orbital (which is the
action is transmitted through the bonds. The simplest case to reason why 1JCH depends on the hybridization ratio). We shall
consider is that of 1JXY, where X and Y are spin- 12 nuclei joined suppose that it is energetically favourable for an electron spin
by an electron-pair bond. The coupling mechanism depends on and a nuclear spin to be antiparallel (as is the case for a proton
the fact that the energy depends on the relative orientation of and an electron in a hydrogen atom).
the bonding electron and nuclear spins. This electron–nucleus If the X nucleus is α, a β electron of the bonding pair will
coupling is magnetic in origin, and may be either a dipolar tend to be found nearby, because that is an energetically favour-
interaction or a Fermi contact interaction. A pictorial descrip- able arrangement (Fig. 48.16). The second electron in the bond,
tion of the latter is as follows. First, we regard the magnetic which must have α spin if the other is β (by the Pauli princi-
moment of the nucleus as arising from the circulation of a cur- ple, Topic 19), will be found mainly at the far end of the bond
rent in a tiny loop with a radius similar to that of the nucleus because electrons tend to stay apart to reduce their mutual
(Fig. 48.15). Far from the nucleus the field generated by this repulsion. Because it is energetically favourable for the spin
loop is indistinguishable from the field generated by a point of Y to be antiparallel to an electron spin, a Y nucleus with β
magnetic dipole. Close to the loop, however, the field differs spin has a lower energy than when it has α spin. The opposite is
from that of a point dipole. The magnetic interaction between true when X is β, for now the α spin of Y has the lower energy.
this non-dipolar field and the electron’s magnetic moment is In other words, the antiparallel arrangement of nuclear spins
the contact interaction. The contact interaction—essentially lies lower in energy than the parallel arrangement as a result of

Hund

Pauli C Pauli

Fermi Fermi
X
Y

Figure 48.15 The origin of the Fermi contact interaction.


From far away, the magnetic field pattern arising from a ring
of current (representing the rotating charge of the nucleus, Figure 48.17 The polarization mechanism for 2 JHH spin–spin
the pale grey sphere) is that of a point dipole. However, if an coupling. The spin information is transmitted from one bond
electron can sample the field close to the region indicated by to the next by a version of the mechanism that accounts for the
the sphere, the field distribution differs significantly from that lower energy of electrons with parallel spins in different atomic
of a point dipole. For example, if the electron can penetrate the orbitals (Hund’s rule of maximum multiplicity). In this case, J < 0,
sphere, then the spherical average of the field it experiences is corresponding to a lower energy when the nuclear spins are
not zero. parallel.

Atkins09819.indb 473 9/11/2013 12:09:50 PM


474 10 Magnetic resonance

their magnetic coupling with the bond electrons. That is, 1JCH is broadening occurs when the lifetime, τ, of a conformation
positive. gives rise to a linewidth that is comparable to the difference
To account for the value of 2JXY, as in HeCeH, we need a of resonance frequencies, δν, and both broadened lines blend
mechanism that can transmit the spin alignments through together into a very broad line. Coalescence of the two lines
the central C atom (which may be 12C, with no nuclear spin occurs when
of its own). In this case (Fig. 48.17), an X nucleus with α spin
polarizes the electrons in its bond, and the α electron is likely 21/2 Condition for coalescence
τ= (48.16)
to be found closer to the C nucleus. The more favourable πδ␯ of two NMR lines
arrangement of two electrons on the same atom is with their
spins parallel (Hund’s rule, Topic 19), so the more favourable
Brief illustration 48.6 The effect of chemical exchange
arrangement is for the α electron of the neighbouring bond
to be close to the C nucleus. Consequently, the β electron of on NMR spectra
that bond is more likely to be found close to the Y nucleus, and The NO group in N,N-dimethylnitrosamine, (CH3)2NeNO
therefore that nucleus will have a lower energy if it is α. Hence, (5), rotates about the NeN bond and, as a result, the magnetic
according to this mechanism, the lower energy will be obtained environments of the two CH 3 groups are interchanged. The
if the Y spin is parallel to that of X. That is, 2JHH is negative. two CH 3 resonances are separated by 390 Hz in a 600 MHz
The coupling of nuclear spin to electron spin by the Fermi spectrometer. According to eqn 48.16,
contact interaction is most important for proton spins, but it is
not necessarily the most important mechanism for other nuclei. 21/2
τ= = 1.2 ms
These nuclei may also interact by a dipolar mechanism with the π × (390 s −1 )
electron magnetic moments and with their orbital motion, and
there is no simple way of specifying whether J will be positive It follows that the signal will collapse to a single line when the
interconversion rate exceeds about 1/τ = 830 s−1.
or negative.

Conformational conversion and


48.4
exchange processes
5 N,N-Dimethylnitrosamine
The appearance of an NMR spectrum is changed if magnetic
nuclei can jump rapidly between different environments. Self-test 48.9 What would you deduce from the observa-
Consider a molecule, such as N,N-dimethylformamide, that tion of a single line from the same molecule in a 300 MHz
can jump between conformations; in this case, the methyl shifts spectrometer?
depend on whether they are cis or trans to the carbonyl group Answer: Conformation lifetime less than 2.3 ms
(Fig. 48.18). When the jumping rate is low, the spectrum shows
two sets of lines, one each from molecules in each conforma-
tion. When the interconversion is fast, the spectrum shows A similar explanation accounts for the loss of fine struc-
a single line at the mean of the two chemical shifts. At inter- ture in solvents able to exchange protons with the sample. For
mediate inversion rates, the line is very broad. This maximum example, hydroxyl protons are able to exchange with water pro-
tons. When this chemical exchange occurs, a molecule ROH
with an α-spin proton (we write this ROHα) rapidly converts
O to ROHβ and then perhaps to ROHα again because the protons
H
C provided by the solvent molecules in successive exchanges have
random spin orientations. Therefore, instead of seeing a spec-
N trum composed of contributions from both ROHα and ROHβ
molecules (that is, a spectrum showing a doublet structure due
to the OH proton), we see a spectrum that shows no splitting
caused by coupling of the OH proton (as in Fig. 48.2 and as dis-
Figure 48.18 When a molecule changes from one cussed in Example 48.3). The effect is observed when the life-
conformation to another, the positions of its protons are time of a molecule due to this chemical exchange is so short
interchanged and the protons jump between magnetically that the lifetime broadening is greater than the doublet split-
distinct environments. ting. Because this splitting is often very small (a few hertz), a

www.ebook3000.com

Atkins09819.indb 474 9/11/2013 12:09:59 PM


48 Features of NMR spectra  475

proton must remain attached to the same molecule for longer no splitting from the OH protons. In dry dimethylsulfoxide
than about 0.1 s for the splitting to be observable. In water, (DMSO), the exchange rate may be slow enough for the split-
the exchange rate is much faster than that, so alcohols show ting to be detected.

Checklist of concepts
☐ 1. The chemical shift of a nucleus is the difference ☐ 6. Fine structure is the splitting of resonances into indi-
between its resonance frequency and that of a reference vidual lines by spin–spin coupling.
standard. ☐ 7. Spin–spin coupling is expressed in terms of the spin–
☐ 2. The shielding constant is the sum of a local contribu- spin coupling constant J and depends on the relative
tion, a neighbouring group contribution, and a solvent orientation of two nuclear spins.
contribution. ☐ 8. The coupling constant decreases as the number of
☐ 3. The local contribution is the sum of a diamagnetic bonds separating two nuclei increases.
contribution and a paramagnetic contribution. ☐ 9. Spin–spin coupling can be explained in terms of the
☐ 4. The neighbouring group contribution arises from the polarization mechanism and the Fermi contact
currents induced in nearby groups of atoms. interaction.
☐ 5. The solvent contribution can arise from specific molec- ☐ 10. Coalescence of two NMR lines occurs when a confor-
ular interactions between the solute and the solvent. mational interchange or chemical exchange of nuclei
is fast.

Checklist of equations
Property Equation Comment Equation
number

δ-Scale of chemical shifts δ = {(ν − ν°)/ν°} × 106 48.4

Relation between chemical shift and shielding δ ≈ (σ° − σ) × 106 48.6


constant

Local contribution to the shielding constant σ(local) = σd + σp 48.8

Lamb formula σd = (e2μ0/12πme)〈1/r 〉 48.9

Neighbouring group contribution to the σ(neighbour) ∝(χ − χ⊥)(1 − 3 cos2 θ)/r3 The angle θ is defined in 1 48.10b
shielding constant

HH = A + B cos φ + C cos 2φ
Karplus equation 3J A, B, and C are empirical constants 48.14

Condition for coalescence of two NMR lines Conformational conversions and 48.16
τ = 21/2 /πδ␯
exchange processes

Atkins09819.indb 475 9/11/2013 12:10:01 PM


TOPIC 49

Pulse techniques in NMR

Contents (Topics 48), the vector model of angular momentum


49.1 The magnetization vector 476 (Topic 14), the magnetic properties of molecules (Topic
(a) The effect of the radiofrequency field 477 39), and Fourier transforms (Topic 40 and Mathematical
Brief illustration 49.1: Radiofrequency pulses 478 background 6).
(b) Time- and frequency-domain signals 478
Brief illustration 49.2: Fourier analysis 479
49.2 Spin relaxation 479
Brief illustration 49.3: Inhomogeneous broadening 481 The common method of detecting the energy separation
49.3 The nuclear Overhauser effect 481 between nuclear spin states is more sophisticated than simply
Brief illustration 49.4: NOE enhancement 483 looking for the frequency at which resonance occurs. One of
the best analogies that has been suggested to illustrate the pre-
49.4 Two-dimensional NMR 483
ferred way of observing an NMR spectrum is that of detecting
Example 49.1: Interpreting a two-dimensional
NMR spectrum 483
the spectrum of vibrations of a bell. We could stimulate the bell
with a gentle vibration at a gradually increasing frequency, and
49.5 Solid-state NMR 484
note the frequencies at which it resonated with the stimula-
Brief illustration 49.5: Dipolar fields in solids 485
tion. A lot of time would be spent getting zero response when
Checklist of concepts 485
the stimulating frequency was between the bell’s vibrational
Checklist of equations 486
modes. However, if we were simply to hit the bell with a ham-
mer, we would immediately obtain a clang composed of all the
frequencies that the bell can produce. The equivalent in NMR
is to monitor the radiation nuclear spins emit as they return
➤ Why do you need to know this material? to equilibrium after the appropriate stimulation. The result-
To understand how nuclear magnetic resonance spectro- ing Fourier-transform NMR (FT-NMR) spectroscopy gives
scopy is used to study large molecules and even diagnose greatly increased sensitivity, so opening up much of the peri-
disease, you need to understand how spectral information odic table to the technique. Moreover, multiple-pulse FT-NMR
is obtained by analysing the response of nuclei to the gives chemists unparalleled control over the information con-
application of strong pulses of radiofrequency radiation. tent and display of spectra.

➤ What is the key idea?


Fourier-transform NMR spectroscopy is the analysis of
the radiation emitted by nuclear spins as they return to 49.1 The magnetization vector
equilibrium after stimulation by one or more pulses of
radiofrequency radiation. Consider a sample composed of many identical spin- 12 nuclei.
By analogy with the discussion of angular momenta in Topic
➤ What do you need to know already? 14, a nuclear spin can be represented by a vector of length {I(I+
You need to be familiar with the general principles of 1)}1/2 units with a component of length mI units along the z-axis.
magnetic resonance (Topic 47), the features of NMR spectra As the uncertainty principle does not allow us to specify the x-
and y-components of the angular momentum, all we know is

www.ebook3000.com

Atkins09819.indb 476 9/11/2013 12:10:12 PM


49 Pulse techniques in NMR  477

that the vector lies somewhere on a cone around the z-axis. For actual representation of reality but is inspired by the motion
I = 12 , the length of the vector is 12 31/2 and it makes an angle of of a classical bar magnet in a magnetic field). As the field is
55° to the z-axis (Fig. 49.1). increased, the Larmor frequency increases and the precession
In the absence of a magnetic field, the sample consists of becomes faster.
equal numbers of α and β nuclear spins with their vectors lying r The populations of the two spin states (the numbers of α
at random angles on the cones. These angles are unpredict- and β spins) at thermal equilibrium change, with slightly
able, and at this stage we picture the spin vectors as stationary. more α spins than β spins (see Topic 47).
The magnetization, M, of the sample, its net nuclear magnetic
moment, is zero (Fig. 49.2a). Despite its smallness, this imbalance means that there is a net
Two changes occur in the magnetization when a mag- magnetization that we can represent by a vector M pointing in
netic field of magnitude B0 is present and aligned in the the z-direction and with a length proportional to the popula-
z-direction: tion difference (Fig. 49.2b).
r The energies of the two orientations change, the α spins
moving to low energy and the β spins to high energy (a) The effect of the radiofrequency field
(provided γ N > 0).
Now consider the effect of a radiofrequency field circularly
At 10 T, the Larmor frequency for protons is 427 MHz, and in polarized in the xy-plane, so that the magnetic component of
the vector model the individual vectors are pictured as precess- the electromagnetic field (the only component we need to con-
ing at this rate (Topic 48). This motion is a pictorial representa- sider) is rotating around the z-direction in the same sense as
tion of the difference in energy of the spin states (it is not an the Larmor precession of the nuclei. The strength of the rotat-
ing magnetic field is B1.
To interpret the effects of radiofrequency pulses on the
z magnetization, it is useful to imagine stepping on to a plat-
form, a so-called rotating frame, that rotates around the direc-
tion of the applied field. Suppose we choose the frequency of
½ the radiofrequency field to be equal to the Larmor frequency
½√3
of the spins, νL = γNB0/2π; this choice is equivalent to selecting
the resonance condition in the conventional experiment. The
rotating magnetic field is in step with the precessing spins, and
Figure 49.1 The vector model of angular momentum for the nuclei experience a steady B1 field and precess about it at a
a single spin- 12 nucleus. The angle around the z–axis is frequency γNB1/2π (Fig. 49.3). Now suppose that the B1 field
indeterminate. is applied in a pulse of duration Δτ = 14 × 2π / γ N B1. The mag-
netization tips through an angle of 14 × 2π = π/2 (90°) in the

M
M
M

α
νL
β

(a) (b) ν = νL
(a) B1 (b) B1

Figure 49.2 The magnetization of a sample of spin- nuclei is


1
2
the resultant of all their magnetic moments. (a) In the absence Figure 49.3 (a) In a resonance experiment, a circularly
of an externally applied field, there are equal numbers of polarized radiofrequency magnetic field B1 is applied in the
α and β spins at random angles around the z-axis (the field xy-plane (the magnetization vector lies along the z-axis). (b) If
direction) and the magnetization is zero. (b) In the presence we step into a frame rotating at the radiofrequency, B1 appears
of a field, the spins precess around their cones (that is, there to be stationary, as does the magnetization M if the Larmor
is an energy difference between the α and β states) and there frequency is equal to the radiofrequency. When the two
are slightly more α spins than β spins. As a result, there is a net frequencies coincide, the magnetization vector of the sample
magnetization along the z-axis. rotates around the direction of the B1 field.

Atkins09819.indb 477 9/11/2013 12:10:22 PM


478 10 Magnetic resonance

90° pulse B0
Detecting
coil

Signal
M
νL Time, t
B1
(a) (b)

Figure 49.4 (a) If the radiofrequency field is applied for a


certain time, the magnetization vector is rotated into the
xy-plane. (b) To an external stationary observer (the coil), the Figure 49.5 A simple free-induction decay of a sample of spins
magnetization vector is rotating at the Larmor frequency, and with a single resonance frequency.
can induce a signal in the coil.
We have considered the effect of a pulse applied at exactly the
rotating frame and we say that we have applied a 90° pulse, or Larmor frequency. However, virtually the same effect is obtained
a ‘π/2 pulse’ (Fig. 49.4a). off resonance, provided that the pulse is applied close to νL. If the
difference in frequency is small compared to the inverse of the
duration of the 90° pulse, the magnetization will end up in the xy-
Brief illustration 49.1 Radiofrequency pulses plane. Note that we do not need to know the Larmor frequency
beforehand: the short pulse is the analogue of the hammer blow
The duration of a radiofrequency pulse depends on the
on the bell, exciting a range of frequencies. The detected signal
strength of the B1 field. If a 90° pulse requires 10 μs, then for
shows that a particular resonant frequency is present.
protons
π
B1 =
2 × (
2. × 10
675  s ) × (
−8 −1 −1
T 1.0× −5
10 s)
= 5.9 ×10−4 T (b) Time- and frequency-domain signals

γN ⌬τ We can think of the magnetization vector of a homonuclear AX
spin system with spin–spin coupling constant J = 0 as consisting
Self-test 49.1 How long would a 180° pulse require for protons?
of two parts, one formed by the A spins and the other by the X
Answer: 20 μs
spins. When the 90° pulse is applied, both magnetization vec-
tors are rotated into the xy-plane. However, because the A and
X nuclei precess at different frequencies, they induce two sig-
Now imagine stepping out of the rotating frame. To a nals in the detector coils, and the overall FID curve may resem-
fixed external observer (the role played by a radiofrequency ble that in Fig. 49.6a. The composite FID curve is the analogue
coil), the magnetization vector is rotating at the Larmor fre- of the struck bell emitting a rich tone composed of all the fre-
quency in the xy-plane (Fig. 49.4b). The rotating magnetiza- quencies (in this case, just the two resonance frequencies of the
tion induces in the coil a signal that oscillates at the Larmor uncoupled A and X nuclei) at which it can vibrate.
frequency and that can be amplified and processed. In prac-
tice, the processing takes place after subtraction of a constant
high-frequency component (the radiofrequency used for B1),
so that all the signal manipulation takes place at frequencies (a)
of a few kilohertz.
Signal

As time passes, the individual spins move out of step (partly


because they are precessing at slightly different rates, as we
Time, t
explain later), so the magnetization vector shrinks exponen-
tially with a time constant T2 and induces an ever weaker signal
in the detector coil. The form of the signal that we can expect is
therefore the oscillating-decaying free-induction decay (FID)
shown in Fig. 49.5. The y-component of the magnetization (b)
Frequency, ν
varies as
Figure 49.6 (a) A free-induction decay signal of a sample of AX
M y (t ) = M0 cos(2π␯L t )e − t /T 2
Free-induction decay (49.1) species and (b) analysis into its frequency components.

www.ebook3000.com

Atkins09819.indb 478 9/11/2013 12:10:31 PM


49 Pulse techniques in NMR  479

The problem we must address is how to recover the reso- where ‘Re’ means take the real part of the following expres-
nance frequencies present in a free-induction decay. We know sion. This integral gives a nonzero value if S(t) contains a
that the FID curve is a sum of decaying oscillating functions, component that matches the oscillating function e2iπ␯t . The
so the problem is to analyse it into its components by carrying integration is carried out at a series of frequencies ␯ on a com-
out a Fourier transformation (see the following Justification). puter that is built into the spectrometer.
When the signal in Fig. 49.6a is transformed in this way, we get
the frequency-domain spectrum shown in Fig. 49.6b. One line
represents the Larmor frequency of the A nuclei and the other
that of the X nuclei.

Signal
Brief illustration 49.2 Fourier analysis Time
Fourier analysis is a common feature of most mathemati-
cal software packages, but one simple example is the Fourier
transform of the function
Figure 49.7 A free-induction decay signal of a sample of
S(t ) = S(0)cos(2π␯Lt )e −t /T2
ethanol. Its Fourier transform is the frequency-domain
spectrum shown in Fig. 48.2. The total length of the image
which describes the behaviour of the FID signal in eqn 49.1.
corresponds to about 1 s.
The result is (Problem 49.4)
S(0)T2
I (␯) =
1 + (␯ L − ␯ ) (2πT2 )2
2
The FID curve in Fig. 49.7 is obtained from a sample of
ethanol. The frequency-domain spectrum obtained from it by
which has the so-called ‘Lorentzian’ shape, with a maximum Fourier transformation is the one discussed in Topic 48 (see
intensity at I(␯L) = S(0)T2. Fig. 48.2). We can now see why the FID curve in Fig. 49.7 is so
complex: it arises from the precession of a magnetization vector
Self-test 49.2 What is the width at half-height, Δν1/2 , of the
that is composed of eight components, each with a characteris-
Lorentzian function above?
tic frequency.
Answer: Δν1/2 = 1/πT2

49.2 Spin relaxation


Justification 49.1 Fourier transformation of the There are two reasons why the component of the magnetiza-
FID curve tion vector in the xy-plane shrinks. Both reflect the fact that
The analysis of the FID curve is achieved by the standard the nuclear spins are not in thermal equilibrium with their
mathematical technique of Fourier transformation, which is surroundings (for then M lies parallel to z). At thermal equi-
explained more fully in Mathematical background 6. We start librium the spins have a Boltzmann distribution, with more α
by noting that the signal S(t) in the time domain, the total FID spins than β spins, and lie at random orientations on their pre-
curve, is the sum (more precisely, the integral) over all the cessional cones. The return to equilibrium is the process called
contributing frequencies: spin relaxation.
Consider the effect of a 180° pulse, which may be visualized in
∞ the rotating frame as a flip of the net magnetization vector from
S(t ) = ∫−∞
I (␯)e −2πi␯t d␯ (49.2)
one direction along the z-axis (with more α spins than β spins)
to the opposite direction (with more β spins than α spins). After
Because e2πi␯t = cos(2π␯t) + i sin(2π␯t), this expression is a the pulse, the populations revert to their thermal equilibrium
sum over harmonically oscillating functions, with each one values exponentially. As they do so, the z-component of mag-
weighted by the intensity I(␯). netization reverts to its equilibrium value M0 with a time con-
We need I(␯), the spectrum in the frequency domain; it is stant called the longitudinal relaxation time, T1 (Fig. 49.8):
obtained by evaluating the integral

∞ Longitudinal
I (␯) = 2Re ∫ S(t )e2 πi␯t dt (49.3) M z (t ) − M0 ∝ e −t /T Definition relaxation (49.4)
1

0 time

Atkins09819.indb 479 9/11/2013 12:10:40 PM


480 10 Magnetic resonance

t + T1 at about the resonance frequency will the fluctuating magnetic


t
field be able to induce spin changes effectively, and only then
will T1 be short. The rate of molecular tumbling increases with
temperature and with decreasing viscosity of the solvent, so
β β we can expect a dependence like that shown in Fig. 49.9. The
quantitative treatment of relaxation times depends on setting
up models of molecular motion and using, for instance, the dif-
α α fusion equation (Topic 81) adapted for rotational motion.
Now consider the events following a 90° pulse. The mag-
netization vector in the xy-plane is large when the spins are
bunched together immediately after the pulse. However, this
orderly bunching of spins is not at equilibrium and, even if
Figure 49.8 In longitudinal relaxation the spins relax back there were no spin–lattice relaxation, we would expect the indi-
towards their thermal equilibrium populations. On the left vidual spins to spread out until they were uniformly distrib-
we see the precessional cones representing spin- 12 angular uted with all possible angles around the z-axis (Fig. 49.10). At
momenta, and they do not have their thermal equilibrium that stage, the component of magnetization vector in the plane
populations (there are more β spins than α spins). On the would be zero. The randomization of the spin directions occurs
right, which represents the sample a long time after a time exponentially with a time constant called the transverse relaxa-
T1 has elapsed, the populations are those characteristic of a tion time, T2:
Boltzmann distribution. In actuality, T1 is the time constant for
Transverse
relaxation to the arrangement on the right and T1 ln 2 is the M y (t ) ∝ e −t /T
2
Definition relaxation (49.5)
half-life of the arrangement on the left. time

Because the relaxation involves the relative orientation of the


Because this relaxation process involves giving up energy to spins around their respective cones, T2 is also known as the
the surroundings (the ‘lattice’) as β spins revert to α spins, the spin–spin relaxation time. Any relaxation process that changes
time constant T1 is also called the spin–lattice relaxation time. the balance between α and β spins will also contribute to this
Spin–lattice relaxation is caused by local magnetic fields that randomization, so the time constant T2 is almost always less
fluctuate at a frequency close to the resonance frequency of than or equal to T1.
the β → α transition. Such fields can arise from the tumbling Local magnetic fields also affect spin–spin relaxation. When
motion of molecules in a fluid sample. If molecular tumbling the fluctuations are slow, each molecule lingers in its local
is too slow or too fast compared to the resonance frequency, it
will give rise to a fluctuating magnetic field with a frequency
t t +T2
that is either too low or too high to stimulate a spin change
from β to α, so T1 will be long. Only if the molecule tumbles

β β

T1
α α
Relaxation time

T2
Low High
Figure 49.10 The transverse relaxation time, T2, is the time
temperature, temperature,
high viscosity low viscosity
constant for the phases of the spins to become randomized
Rate of motion
(another condition for equilibrium) and to change from the
orderly arrangement shown on the left to the disorderly
Figure 49.9 The variation of the two relaxation times with arrangement on the right (long after a time T2 has elapsed).
the rate at which the molecules move (either by tumbling or Note that the populations of the states remain the same; only
migrating through the solution). The horizontal axis can be the relative phase of the spins relaxes. In actuality, T2 is the time
interpreted as representing temperature or viscosity. Note that constant for relaxation to the arrangement on the right and T2
at rapid rates of motion, the two relaxation times coincide. ln 2 is the half-life of the arrangement on the left.

www.ebook3000.com

Atkins09819.indb 480 9/11/2013 12:10:47 PM


49 Pulse techniques in NMR  481

magnetic environment and the spin orientations randomize where Δ␯1/2 is the observed width at half-height of a line with a
quickly around their cones. If the molecules move rapidly from Lorentzian shape of the form I ∝ 1/(1+ ␯2).
one magnetic environment to another, the effects of differences
in local magnetic field average to zero: individual spins do not
Brief illustration 49.3 Inhomogeneous broadening
precess at very different rates, they can remain bunched for
longer, and spin–spin relaxation does not take place as quickly. Consider a line in a spectrum with a width of 10 Hz. It follows
In other words, slow molecular motion corresponds to short T2 from eqn 49.7 that the effective transverse relaxation time is
and fast motion corresponds to long T2 (as shown in Fig. 49.9).
1
Calculations show that, when the motion is fast, the main rand- T2∗ = = 32 ms
omizing effect arises from β → α transitions rather than differ- π × (10 s −1 )
ent precession rates on the cones, and then T2 ≈ T1.
Self-test 49.3 Name two processes that could contribute to
If the y-component of magnetization decays with a time
further broadening of the NMR line.
constant T2, the spectral line is broadened (Fig. 49.11), and its Answer: Conformational conversion or
width at half-height becomes (see Self-test 49.2) chemical exchange (see Topic 48)

1
Δ␯1/2 = Width at half-height of an NMR line (49.6)
πT2
Typical values of T2 in proton NMR are of the order of seconds,
so linewidths of around 0.1 Hz can be anticipated, in broad 49.3 The nuclear Overhauser effect
agreement with observation.
So far, we have assumed that the equipment, and in particu- One advantage of protons in NMR is their high magnetogyric
lar the magnet, is perfect, and that the differences in Larmor ratio, which results in relatively large Boltzmann population
frequencies arise solely from interactions within the sample. In differences and strong coupling to the radiofrequency field,
practice, the magnet is not perfect, and the field is different at and hence greater resonance intensities than for most other
different locations in the sample. The inhomogeneity broadens nuclei. In the steady-state nuclear Overhauser effect (NOE),
the resonance, and in most cases this inhomogeneous broad- spin relaxation processes involving internuclear dipole–dipole
ening dominates the broadening we have discussed so far. It interactions are used to transfer this population advantage to
is common to express the extent of inhomogeneous broaden- another nucleus (such as 13C or another proton), so that the
ing in terms of an effective transverse relaxation time, T2∗ , by latter’s resonances are modified. In a dipole–dipole interaction
using a relation like eqn 49.6, but writing between two nuclei, one nucleus influences the behaviour of
another nucleus in much the same way that the orientation of a
1 bar magnet is influenced by the presence of another bar magnet
T2∗ = Effective transverse relaxation time (49.7) nearby.
πΔ␯1/2
To understand the effect, consider the populations of the
four levels of a homonuclear (for instance, proton) AX system;
these levels were shown in Fig. 48.8. At thermal equilibrium,
1 the population of the αAαX level is the greatest and that of
the βAβX level is the least; the other two levels have the same
0.8 energy and an intermediate population. The thermal equilib-
rium absorption intensities reflect these populations, as shown
Δν1/2 = 1/πT2
Signal height

0.6 in Fig. 49.12. Now consider the combined effect of spin relaxa-
tion and keeping the X spins saturated. When we saturate
0.4 the X transition, the populations of the X levels are equalized
(N α = N β ) and all transitions involving αX ↔ βX spin flips are
x x

0.2 no longer observed. At this stage there is no change in the pop-


ulations of the A levels. If that were all there were to happen,
0 all we would see would be the loss of the X resonance and no
0
Frequency offset, ν – ν0 effect on the A resonance.
Now consider the effect of spin relaxation. Relaxation
Figure 49.11 A Lorentzian absorption line. The width at half- can occur in a variety of ways if there is a dipolar interaction
height is inversely proportional to the parameter T2 and the between the A and X spins. One possibility is for the magnetic
longer the transverse relaxation time, the narrower the line. field acting between the two spins to cause them both to flip

Atkins09819.indb 481 9/11/2013 12:10:54 PM


482 10 Magnetic resonance

βAβX the illustration that the population differences in the states


involved in the A transitions are decreased, so the resonance
A X absorption is diminished.
Energy
Which effect wins? Does the NOE enhance the A absorp-
αAβX βAαX
tion or does it diminish it? As in the discussion of relaxation
times in Section 49.2, the efficiency of the intensity-enhancing
X A
βAβX ↔ αAαX relaxation is high if the dipole field oscillates at a
αAαX frequency close to the transition frequency, which in this case
is about 2␯; likewise, the efficiency of the intensity-diminishing
Figure 49.12 The energy levels of an AX system and an αAβX ↔ βAαX relaxation is high if the dipole field is stationary
indication of their relative populations. Each green square (as there is no frequency difference between the initial and final
above the line represents an excess population and each white states). A large molecule rotates so slowly that there is very little
square below the line represents a population deficit. The motion at 2␯, so we expect an intensity decrease (Fig. 49.14). A
transitions of A and X are marked. small molecule rotating rapidly can be expected to have substan-
tial motion at 2␯, and a consequent enhancement of the signal.
simultaneously from β to α, so the αAαX and βAβX states regain In practice, the enhancement lies somewhere between the two
their thermal equilibrium populations. However, the popu- extremes and is reported in terms of the parameter η (eta), where
lations of the αAβX and βAαX levels remain unchanged at the
I A − I A°
values characteristic of saturation. As we see from Fig. 49.13, η= NOE enhancement parameter (49.8)
I A°
the population difference between the states joined by transi-
tions of A is now greater than at equilibrium, so the resonance Here I A° and IA are the intensities of the NMR signals due to
absorption is enhanced. Another possibility is for the dipolar nucleus A before and after application of the long (> T1) radio-
interaction between the two spins to cause αA to flip to βA and frequency pulse that saturates transitions due to the X nucleus.
simultaneously βX to flip to αX (or vice versa). This transition When A and X are nuclei of the same species, such as pro-
equilibrates the populations of αAβX and βAαX but leaves the tons, η lies between –1 (diminution) and + 12 (enhancement).
αAαX and βAβX populations unchanged. Now we see from However, η also depends on the values of the magnetogyric

βAβX βAβX
Saturate βAβX βAβX
Saturate
A X A X A X A X
Energy

Relax

Energy

αAβX βAαX αAβX βAαX αAβX Relax βAαX


αAβX βAαX

X A X A X A X A
Saturate
(a) αAαX (b) αAαX Saturate
αAαX αAαX
(a) (b)
βAβX βAβX
Enhanced Diminished

A X A X

αAβX βAαX αAβX βAαX

Enhanced Diminished
X A X A

(c) αAαX (c) αAαX

Figure 49.13 (a) When the X transition is saturated, the Figure 49.14 (a) When the X transition is saturated, just as
populations of its two states are equalized and the population in Fig. 49.13 the populations of its two states are equalized
excess and deficit become as shown (using the same symbols and the population excess and deficit become as shown.
as in Fig. 49.12). (b) Dipole − dipole relaxation relaxes the (b) Dipole–dipole relaxation relaxes the populations of the two
populations of the highest and lowest states, and they regain intermediate states, and they regain their original populations.
their original populations. (c) The A transitions reflect the (c) The A transitions reflect the difference in populations
difference in populations resulting from the preceding changes, resulting from the preceding changes, and are diminished
and are enhanced compared with those shown in Fig. 49.12. compared with those shown in Fig. 49.12.

www.ebook3000.com

Atkins09819.indb 482 9/11/2013 12:11:05 PM


49 Pulse techniques in NMR  483

ratios of A and X. In the case of maximal enhancement it is pos-


sible to show that
γX
η= (49.9)
2γ A
where γA and γX are the magnetogyric ratios of nuclei A and X,
respectively.

Brief illustration 49.4 NOE enhancement


δ
From eqn 49.9 and the data in Table 47.2, the NOE enhance-
ment parameter for 13C close to a saturated proton is δ
γ 1H
 Figure 49.15 An idealization of the COSY spectrum of an AX
2.675 ×108 T−1 s −1
η= = 1.99 spin system.
2 × (6.73 ×107 T−1 s −1 )
 
γ 13C resonances belonging to different groups lying at different loca-
tions on the second axis. This separation is essentially what is
which shows that an enhancement of about a factor of 2 can achieved in two-dimensional NMR.
be achieved. Much modern NMR work makes use of correlation spec-
troscopy (COSY), in which a clever choice of pulses and Fourier
Self-test 49.4 Interpret the following features of the NMR
transformation techniques makes it possible to determine all
spectra of a protein: (a) saturation of a proton resonance
assigned to the side chain of a methionine residue changes the
spin–spin couplings in a molecule. A typical outcome for an AX
intensities of proton resonances assigned to the side chains of system is shown in Fig. 49.15. The diagram shows contours of
a tryptophan and a tyrosine residue; (b) saturation of proton equal signal intensity on a plot of intensity against the frequency
resonances assigned to the tryptophan residue did not affect coordinates ␯1 and ␯2. The diagonal peaks are signals centred on
the spectrum of the tyrosine residue. (δA,δA) and (δX,δX) and lie along the diagonal where ␯1 = ␯2. That
Answer: The tryptophan and tyrosine residues are close
is, the spectrum along the diagonal is equivalent to the one-
to the methionine residue, but are far from each other dimensional spectrum obtained with the conventional NMR
technique (as in Fig. 48.10). The cross-peaks (or off-diagonal
peaks) are signals centred on (δA,δX) and (δX,δA) and owe their
The NOE is also used to determine interproton distances. existence to the coupling between the A and X nuclei.
The Overhauser enhancement of a proton A generated by satu- Although information from two-dimensional NMR spec-
rating a spin X depends on the fraction of A’s spin–lattice relax- troscopy is trivial in an AX system, it can be of enormous help
ation that is caused by its dipolar interaction with X. Because in the interpretation of more complex spectra, leading to a map
the dipolar field is proportional to r −3, where r is the internu- of the couplings between spins and to the determination of the
clear distance, and the relaxation effect is proportional to the bonding network in complex molecules. Indeed, the spectrum
square of the field, and therefore to r −6, the NOE may be used to of a synthetic or biological polymer that would be impossible
determine the geometries of molecules in solution. The deter- to interpret in one-dimensional NMR can often be interpreted
mination of the structure of a small protein in solution involves reasonably rapidly by two-dimensional NMR.
the use of several hundred NOE measurements, effectively cast-
ing a net over the protons present. The enormous importance Example 49.1 Interpreting a two-dimensional NMR
of this procedure is that we can determine the conformation of spectrum
biological macromolecules in an aqueous environment and do
Figure 49.16 is a portion of the COSY spectrum of the amino
not need to try to make the single crystals that are essential for
acid isoleucine (1), showing the resonances associated with the
an X-ray diffraction investigation (Topic 37).
protons bound to the carbon atoms.
c O
49.4 Two-dimensional NMR
b a
d OH
An NMR spectrum contains a great deal of information and, if
many protons are present, is very complex when the fine struc- e NH2
tures of different groups of lines overlap. The complexity would 1 Isoleucine
be reduced if we could use two axes to display the data, with

Atkins09819.indb 483 9/11/2013 12:11:11 PM


484 10 Magnetic resonance

We have seen that the nuclear Overhauser effect can pro-


vide information about internuclear distances through analy-
sis of enhancement patterns in the NMR spectrum before and
after saturation of selected resonances. In nuclear Overhauser
effect spectroscopy (NOESY) a map of all possible NOE inter-
1
actions is obtained by again using a proper choice of radio-
frequency pulses and Fourier transformation techniques. Like
2
δ a COSY spectrum, a NOESY spectrum consists of a series of
diagonal peaks that correspond to the one-dimensional NMR
3
spectrum of the sample. The off-diagonal peaks indicate which
nuclei are close enough to each other to give rise to a nuclear
4 Overhauser effect. NOESY data reveal internuclear distances
4 3 2 1
δ up to about 0.5 nm.
Figure 49.16 Proton COSY spectrum of isoleucine. (Example
49.1 and the corresponding spectrum are adapted from K.E.

Solid-state NMR
van Holde, et al., Principles of physical biochemistry, Prentice
Hall (1998).) 49.5

The principal difficulty with the application of NMR to sol-


Method From the molecular structure, we expect that: (i) the ids is the low resolution characteristic of solid samples.
C aeH proton is coupled only to the C b eH proton; (ii) the Nevertheless, there are good reasons for seeking to overcome
CbeH protons are coupled to the C aeH, C ceH, and C deH these difficulties. They include the possibility that a com-
protons; and (iii) the inequivalent CdeH protons are coupled to
pound of interest is unstable in solution or that it is insolu-
the CbeH and CeeH protons.
ble, so conventional solution NMR cannot be employed.
Answer We note that: Moreover, many species, such as polymers and nanomateri-
als, are intrinsically interesting as solids, and it is important to
r The resonance with δ = 3.6 shares a cross-peak with
determine their structures and dynamics when X-ray diffrac-
only one other resonance at δ = 1.9, which in turn
shares cross-peaks with resonances at δ = 1.4, 1.2, and tion techniques fail.
0.9. We conclude that the resonances at δ = 3.6 and 1.9 There are three principal contributions to the linewidths of
correspond to the CaeH and CbeH protons, solids. One is the direct magnetic dipolar interaction between
respectively. nuclear spins. As we saw in the discussion of spin–spin cou-
pling, a nuclear magnetic moment will give rise to a local
r The proton with resonance at δ = 0.8 is not coupled to magnetic field, which points in different directions at different
the CbeH protons, so we assign the resonance at
locations around the nucleus. If we are interested only in the
δ = 0.8 to the CeeH protons.
component parallel to the direction of the applied magnetic
r The resonances at δ = 1.4 and 1.2 do not share field (because only this component has a significant effect),
cross-peaks with the resonance at δ =0.9. then, provided certain subtle effects arising from transforma-
r In the light of the expected couplings, we assign the tion from the static to the rotating frame are neglected, we can
resonance at δ = 0.9 to the CceH protons and the use a classical expression in The chemist's toolkit 48.1 to write
resonances at δ = 1.4 and 1.2 to the inequivalent CdeH the magnitude of the local magnetic field as
protons.
γ N μ0 mI
Bloc = − (1− 3 cos2 θ ) (49.10)
Self-test 49.5 The proton chemical shifts for the NH, C α H, 4 πR3
and C β H groups of alanine (H 2NCH(CH3)COOH) are 8.25,
4.35, and 1.39, respectively. Describe the COSY spectrum of Unlike in solution, in a solid this field is not motionally aver-
alanine between δ = 1.00 and 8.50. aged to zero. Many nuclei may contribute to the total local field
experienced by a nucleus of interest, and different nuclei in a
Answer: Only the NH and C α H protons and the C α H and C β H protons
are expected to show coupling, so the spectrum has only two off-diagonal sample may experience a wide range of fields. Typical dipole
peaks, one at (8.25, 4.35) and the other at (4.35, 1.39) fields are of the order of 1 mT, which corresponds to splittings
and linewidths of the order of 10 kHz.

www.ebook3000.com

Atkins09819.indb 484 9/11/2013 12:11:15 PM


49 Pulse techniques in NMR  485

Brief illustration 49.5 gradient, such as may arise from a non-spherical distribution
Dipolar fields in solids
of charge around the nucleus. This interaction also varies as
When the angle θ can vary only between 0 and θmax, eqn 49.10 1 – 3 cos2 θ.
becomes Fortunately, there are techniques available for reducing the
γ N μ0mI linewidths of solid samples. One technique, magic-angle spin-
Bloc = (cos2 θ max + cos θ max ) ning (MAS), takes note of the 1 – 3 cos2θ dependence of the
4 πR3
dipole–dipole interaction, the chemical shift anisotropy, and
When θmax = 30° and R = 160 pm, the local field generated by a the electric quadrupole interaction. The ‘magic angle’ is the
proton is angle at which 1 – 3 cos2θ = 0, and corresponds to 54.74°. In
γN the technique, the sample is spun at high speed at the magic
   

2.675 ×10 T s × (1.055 × 10−34 Js)


8 −1 −1 angle to the applied field (Fig. 49.17). All the dipolar interac-
cos2 θ max +
tions and the anisotropies average to the value they would have
μ0
mI
cos θ max at the magic angle, but at that angle they are zero. The difficulty
 
⎛ 1 ⎞  with MAS is that the spinning frequency must not be less than
× (4 π ×10 T J m ) × ⎜ ⎟ × (1.616)
−7 2 −1 3
⎝ 2⎠ the width of the spectrum, which is of the order of kilohertz.
Bloc =
4 π × (1.60 ×10 m)
−10
However, gas-driven sample spinners that can be rotated at
 
R 25 kHz or higher frequencies are now routinely available, and a
= 5.57 ×10−4 T = 0.557 mT considerable body of work has been done.
Pulsed techniques similar to those described in the previous
Self-test 49.6 Calculate the distance at which the local field section may also be used to reduce linewidths. Elaborate pulse
from a 13C nucleus is 0.50 mT, with θmax = 40°. sequences have also been devised that reduce linewidths by
Answer: R = 99 pm averaging procedures that make use of twisting the magnetiza-
tion vector through a series of angles.

A second source of linewidth is the anisotropy of the chemi-


54.74°
cal shift. Chemical shifts arise from the ability of the applied
field to generate electron currents in molecules. In general, this
ability depends on the orientation of the molecule relative to
the applied field. In solution, when the molecule is tumbling
rapidly, only the average value of the chemical shift is relevant.
However, the anisotropy is not averaged to zero for stationary
molecules in a solid, and molecules in different orientations
have resonances at different frequencies. The chemical-shift
anisotropy also varies with the angle between the applied field
Magnetic field
and the principal axis of the molecule, as 1 – 3 cos2 θ.
The third contribution is the electric quadrupole interaction.
Nuclei with I > 12 have a distribution of charge that gives rise to
an electric quadrupole moment, a measure of the non-spherical Figure 49.17 In magic-angle spinning, the sample spins at
distribution of charge over the nucleus (for instance, the posi- 54.74° (that is, arccos 1/31/2) to the applied magnetic field. Rapid
tive charge may be concentrated around the equator or at the motion at this angle averages dipole − dipole interactions and
poles). An electric quadrupole interacts with an electric field chemical-shift anisotropies to zero.

Checklist of concepts
☐ 1. Free-induction decay (FID) is the decay of the magnet- ☐ 3. During longitudinal (or spin–lattice) relaxation, β
ization after the application of a radiofrequency pulse. spins revert to α spins.
☐ 2. Fourier transformation of the FID curve gives the NMR ☐ 4. Transverse (or spin–spin) relaxation is the randomiza-
spectrum. tion of spin directions around the z-axis.

Atkins09819.indb 485 9/11/2013 12:11:22 PM


486 10 Magnetic resonance

☐ 5. The nuclear Overhauser effect is the modification ☐ 7. Magic-angle spinning (MAS) is technique in which the
of the intensity of one resonance by the saturation of NMR linewidths in a solid sample are reduced by spin-
another. ning at an angle of 54.74° to the applied magnetic field.
☐ 6. In two-dimensional NMR, spectra are displayed in
two axes, with resonances belonging to different groups
lying at different locations on the second axis.

Checklist of equations
Property Equation Comment Equation number

Free-induction decay M y (t ) = M0 cos(2π␯Lt )e −t /T2 T2 is the transverse relaxation time 49.1

Longitudinal relaxation M z (t ) − M0 ∝ e −t /T1 T1 is the spin–lattice relaxation time 49.4

Transverse relaxation M y (t ) ∝ e −t /T2 49.5

Width at half-height of an NMR line Δ␯1/2 = 1/πT2 Inhomogeneous broadening is treated by using T2∗ 49.6

NOE enhancement parameter η = (I A − I A° )/ I A° Definition 49.8

www.ebook3000.com

Atkins09819.indb 486 9/11/2013 12:11:26 PM


TOPIC 50

Electron paramagnetic resonance


Electron paramagnetic resonance (EPR), which is also known
Contents as electron spin resonance (ESR), is used to study radicals
formed during chemical reactions or by radiation, radicals that
50.1 The g-value 487
act as probes of biological structure, many d-metal complexes,
Brief illustration 50.1: The g-value of a radical 487
and molecules in triplet states (such as those involved in phos-
50.2 Hyperfine structure 488 phorescence, Topic 46). The sample may be a gas, a liquid, or a
(a) The effects of nuclear spin 488
solid, but the free rotation of molecules in the gas phase gives
Example 50.1: Predicting the hyperfine structure rise to complications.
of an EPR spectrum 489
(b) The McConnell equation 490
Brief illustration 50.2: The McConnell equation 490
(c) The origin of the hyperfine interaction

Brief illustration 50.3: The composition of a


490
50.1 The g-value
molecular orbital from analysis of the hyperfine
structure 491 The resonance frequency for a transition between the ms = − 12
Checklist of concepts 491 and the ms = + 12 levels of an electron is
Checklist of equations 491
h␯ = g e μBB0 Free electron Resonance condition (50.1)

where ge ≈ 2.0023 (Topic 47). The magnetic moment of an


unpaired electron in a radical also interacts with an external
➤ Why do you need to know this material? field, but the field it experiences differs from the applied field
Some chemical reactions generate intermediates or due to the presence of local magnetic fields arising from elec-
products containing unpaired electrons, and you need to tron currents induced in the molecular framework. This differ-
know how to characterize the structures of such species ence is taken into account by replacing ge by g and expressing
with special spectroscopic techniques. the resonance condition as
h␯ = g μBB0 EPR resonance condition (50.2)
➤ What is the key idea?
The electron paramagnetic resonance spectrum of a where g is the g-value of the radical.
radical arises from the ability of the applied magnetic
field to induce local electron currents and the magnetic
interaction between the unpaired electron and nuclei Brief illustration 50.1 The g-value of a radical
with spin.
The centre of the EPR spectrum of the methyl radical occurred
at 329.40 mT in a spectrometer operating at 9.2330 GHz (radia-
➤ What do you need to know already?
tion belonging to the X band of the microwave region). Its
You need to be familiar with the concepts of electron spin g-value is therefore
(Topic 19), the general principles of magnetic resonance
 h    ␯ 
(Topic 47), and the magnetic properties of molecules
(6.626 08 × 10 Js) × (9.2330 × 109 s −1 )
−34
(Topic 39). The discussion refers to spin–orbit coupling g= = 2.0027
(9.2740 × 10−24 JT−1 ) × (0.329 40 T)
in atoms (Topic 21) and the Fermi contact interaction in  
μB B0
molecules (Topic 48).

Atkins09819.indb 487 9/11/2013 12:37:02 PM

You might also like