13-Apr-20
OPTICAL PROPERTIES OF SOLIDS
BY
MARK FOX
LECTURE 9
BY
PROF. DR. SAFIA ANJUM
CHAPTER 2: CLASSICAL PROPAGATION
2.2.2, 2.2.4, 2.3
2
1
13-Apr-20
topiCs
2.2.2 muLtipLe resonanCes
2.2.4 LoCaL FieLd CorreCtion
2.3 the Kramers-Kronig reLationships
oscillator model
It was known that atoms emit and absorb at discrete frequencies.
The oscillator model of the atom is illustrated schematically in Fig. 2.1.
It is assumed that the electron is held in a stable orbit with respect to the
nucleus, and the spring represents the restoring force for small displacements
from the equilibrium.
We have to suppose that there are several dipoles within every atom, to
account for the fact that a given atom has many transition frequencies.
During the oscillations of the atomic dipole, the nucleus remains more or less
stationary due to its heavy mass, while the electron oscillates backwards and
4
forwards at frequency Ꙍo.
2
13-Apr-20
oscillator model
We can also use the dipole model to understand how the atom interacts with an external
electromagnetic wave at frequency Ꙍ.
The AC electric field exerts forces on the electron and the nucleus and drives oscillations
of the system at frequency Ꙍ. If Ꙍ coincides with one of the natural frequencies of the
atom, then we have a resonance phenomenon.
This induces very large amplitude oscillations, and transfers energy from the external
wave to the atom.
2.2.2 Multiple Resonance
In general, an optical medium will have many characteristic resonant frequencies.
We already discussed in previous section, we expect to observe separate resonances due to the
lattice vibrations and to the oscillations of the bound electrons within the atoms.
Furthermore, a particular medium may have many resonances of each type. We can treat
these multiple resonances without difficulty in our model provided they occur at distinct
frequencies.
The model assumes that there are several different types of oscillators within a medium, each
with their own characteristic resonant frequency.
In our discussion we will investigate the classical theory of optical propagation, in which the
light is treated as electromagnetic waves and the atoms or molecules are modelled as
classical dipole oscillators.
This gives us the frequency dependence of the absorption coefficient and
6
refractive index.
3
13-Apr-20
2.2.2 Multiple Resonance
Furthermore, a particular medium may have many resonances of each type.
In writing eq 2.12 we split the polarization of the medium into a resonant part
and a non-resonant part.
We then discussed the resonant part in detail, without specifying very
accurately what we meant by the non-resonant term. We simply stated that P
was proportional to € through the susceptibility .
D= ε ϵ + + .
D= ε ϵ + +
7
2.2.2 Multiple Resonance
In reality, the non-resonant polarization of the medium must originate from the
polarizability of the atoms in exactly the same way as the resonant part.
Equation 2.19:
tells us that the dielectric constant decreases each time we go through an
absorption line.
The contributions that enter the background electric susceptibility χ in eqn 2.12
thus arise from the polarization due to all the other oscillators at higher
frequencies.
4
13-Apr-20
2.2.2 Multiple Resonance
We can understand this point better by making it more quantitative.
The contribution to the polarization of a particular oscillator is given by eqn
2.10.
2.2.2 Multiple Resonance
In a medium with many electronic oscillators of different frequencies, the total
polarization will therefore be given by
P= ∑ ( )
ε ……..(2.22)
where ωj and γj are the angular frequency and damping coefficient of a
particular resonance line.
We then substitute this into eqn 2.11, and recall the definition of Єr given in
eqn 2.13, to obtain:
10
5
13-Apr-20
2.2.2 Multiple Resonance
This equation takes account of all the transitions in the
medium and can be used to calculate the full
frequency dependence of the dielectric constant.
The refractive index and absorption coefficient
calculated from eq 2.23 are plotted against frequency
in Fig. 2.6.
We can understand this figure by starting at the
highest frequencies and gradually down to the lower
frequencies. At the very highest frequencies, the
electrons are incapable of responding to the driving
11
field.
2.2.2 Multiple Resonance
The figure has been calculated for a hypothetical solid with three well-separated
resonances with ωj equal to 4 × 1013 rad/s, 4 × 1015 rad/s, and 1 × 1017 rad/s
respectively.
The width of each absorption line has been set to 10% of the centre frequency by
appropriate choice of the damping coefficient.
The resonance in the infrared is included to represent the vibrational
absorption. In a real solid, we would have to adapt the model appropriately to
account for the different reduced mass and effective charge of the vibrational
oscillator.
12
6
13-Apr-20
2.2.2 Multiple Resonance
We can understand this figure by starting at the highest frequencies and
gradually working our way down to the lower frequencies.
At the very highest frequencies, the electrons are too slow to respond to
the driving field.
The medium therefore has no polarization, and the dielectric constant is unity.
As we reduce the frequency, we first run into the transitions of the inner
electrons in the X-ray/vacuum-ultraviolet spectral region, and then the
transitions of the outer electrons in the ultraviolet and visible.
We then have a region with no transitions until we finally reach the vibrational
frequencies in the infrared. 13
2.2.2 Multiple Resonance
Figure: transitions of the outer electrons in the ultraviolet
and visible
Figure: transitions of the inner electrons in the X-
14
ray/vacuum-ultraviolet spectral region
7
13-Apr-20
2.2.2 Multiple Resonance
The medium therefore has no polarization, and the dielectric constant is unity.
Each time we go through one of these resonances, we see the characteristic frequency
dependence of the Lorentz oscillator, with a peak in the absorption spectrum and a
'wiggle' in the refractive index.
In between the resonances the medium is transparent: the absorption coefficient is
zero and the refractive index is almost constant.
However, experimental data shows that the absorption strength actually varies
considerably between different atomic transitions.
15
2.2.2 Multiple Resonance
The value of the refractive index in the transparent regions gradually
increases as we go through more and more resonance lines on decreasing
the frequency. This increase of the refractive index is caused by the fact that
Єst > Є∞ (cf. eqn 2.19), which implies that n is larger below an absorption line
than above it.
By reference to Fig. 2.6, we now see that we have to understand ‘static’ and
‘∞’ as relative to a particular resonance.
The variation of n with frequency due to the resonances is the origin of the
dispersion found in optical materials even when they are transparent
16
8
13-Apr-20
2.2.2 Multiple Resonance
The dipole oscillator model predicts that each oscillator contributes a
term given by eqn 2.10. This leads to a series of absorption lines of the
same strength. However, experimental data shows that the absorption
strength actually varies considerably between different atomic transitions.
We know that this is caused by the variation of the quantum mechanical
transition probability. In classical physics, however, there is no explanation,
and we just assign a phenomenological oscillator strength fj to each
transition,
rewriting eqn 2.23 as: 17
2.2.2 Multiple Resonance
It can be shown from quantum mechanics that we must have
fj = 1 for each electron.
Since the classical model predicts fj = 1 for each oscillator, we
then interpret this by saying that a particular electron is involved
in several transitions at the same time, and the absorption
strength is being divided between these transitions.
18
9
13-Apr-20
2.2.4 Local Field Correction
In a dense optical medium such as a solid, there is another factor that we
must consider.
The individual atomic dipoles respond to the local field that they
experience. This may not necessarily be the same as the external field,
because the dipoles themselves generate electric fields that will be felt
by all the other dipoles. The actual local field experienced by an atom
therefore takes the form:
where E and Eother dipoles represent the fields due to the external field
and the other dipoles respectively. 19
2.2.4 Local Field Correction
The calculation of the correction field due to the
other dipoles in the medium is actually a rather
complicated one.
An approximate solution due to Lorentz can be
derived if we assume that all the dipoles are
Fig. 2.8 Model used to calculate the
parallel to the applied field and are arranged on a local field by the Lorentz correction.
An imaginary spherical surface drawn
cubic lattice. around a particular atom divides the
medium into nearby dipoles and distant
The calculation works by separating the dipoles. The field at the centre of
the sphere due to the nearby dipoles
contribution from the nearby dipoles and that from is summed exactly, while the field due
to the distant dipoles is calculated by
the rest of the sample, as indicated in Fig. 2.8. 20
treating the material outside the sphere
as a uniformly polarized dielectric.
10
13-Apr-20
2.2.4 Local Field Correction
The division is effected by an imaginary spherical surface with a radius large
enough to make it sensible to average the material outside it. The problem is
then reduced to summing the field of the dipoles inside the sphere at the one
in the middle, and then calculating the effect of a uniformly polarized dielectric
outside the sphere. The final result is:
where P is the polarization of the dielectric outside the sphere. By using the
result of eqn 2.28 in eqn 2.27 we find that:
21
2.2.4 Local Field Correction
The macroscopic polarization P will be given by
where χa is the electric susceptibility per atom. χa is defined by:
p being the induced dipole moment per atom. This is analogous to the
usual definition of the macroscopic susceptibility, except that it is now applied
to individual atoms interacting with the local field. We can see from eqn 2.10
that χa is given by
22
11
13-Apr-20
2.2.4 Local Field Correction
if there is just a single resonance. This is modified to
if there are multiple resonances, We can combine eqns 2.29 and 2.30 with
eqns 2.11 and 2.13 by writing
We put all this together to find that:
This result is known as the Clausius–Mossotti relationship. The relationship
works well in gases and liquids also in cubic crystal in which Lorentz
correction occur. 23
2.3 The Kramers-Kronig relationships
The discussion of the dipole oscillator shows that the refractive index and the absorption
coefficient are not independent parameters but are related to each other.
This is a consequence of the fact that they are derived from the real and imaginary parts
of a single parameter, namely the complex refractive index. If we invoke the law of
causality (that an effect may not precede its cause) and apply complex number analysis,
we can derive general relationships between the real and imaginary parts of the
refractive index. These are known as the Kramers-Kronig relationships and may be stated
as follows:
24
12
13-Apr-20
2.3 The Kramers-Kronig relationships
where P indicates that we take the principal part of the integral. The Kramers-Kronig
relationships allow us to calculate n from K, and vice versa.
This can be very useful in practice, because it would allow us, for example, to
measure the frequency dependence of the optical absorption and then calculate
the dispersion without needing to make a separate measurement of n.
25
26
13
