i

TOPICS IN MODERN OPTICS

C. Fairoos

T. K . M. College of Arts and Science Kollam, Kerala, India-691005

Chapter 1

Introduction

1.1 The Wave Equation

1.2 Two-Beam Interference
The idea of interference is best illustrated by considering two plane waves with a relative phase difference. Consider
the following light waves,

E~A = E1 sin (kz − ωt)

E~B = E2 sin (kz − ωt + δ)

Both beams have wave vector k (same wavelength) and travel along z− direction. Also, their phase difference is a
constant, i.e., δ does not depend on time. Therefore, the given beams are coherent. The expression for resultant
wave is obtained from the superposition principle as,

E~R = E~A + E~B

= E1 sin (kz − ωt) + E2 sin (kz − ωt + δ)
= E1 sin (kz − ωt) + E2 sin (kz − ωt) cos δ + E2 cos (kz − ωt) sin δ
= (E1 + E2 cos δ) sin (kz − ωt) + E2 sin δ cos (kz − ωt) .
| {z } | {z }
E cos φ E sin φ

Here, we have re-labelled the constants. Now, the expression for the resultant wave becomes,

E~R = E cos φ sin (kz − ωt) + E sin φ cos (kz − ωt)

= E sin (kz − ωt + φ)

The amplitude of the resultant wave can be expressed in terms of the amplitudes of individual waves as,

E cos φ = E1 + E2 cos δ; E sin φ = E2 sin δ

E 2 = E12 + E22 + E1 E2 cos δ

| {z }
Interference term

Finally, the intensity of the resultant wave is obtained using I = 12 0 cE 2 i.e., I ∝ E 2 as,
p
I = I1 + I2 + 2 I1 I2 cos δ

Maximum intensity is obtained whenever δ = 2N π and minimum intensity is obtained when δ = (2N + 1) π, where
N = 0, 1, 2...

ii
CHAPTER 1. INTRODUCTION iii

Figure 1.1: A schematic diagram showing incident, reflected, and transmitted waves

1.3 Reflection and Transmission Coefficients

Let’s revisit the familiar phenomena of reflection and transmission that a light exhibits when it propagate from
one media to another. Our goal is to quantitatively estimate the percent of incident light that gets reflected or
transmitted. To this extend, consider a plane wave travelling in the z-direction from one medium of magnetic
permeability µ1 and permittivity 1 to another medium of permeability µ2 and permittivity 2 .
In Fig. ??, EI , ER , and ET denote incident, reflected, and transmitted electric fields respectively. • Incident
wave

E~I = î EI0 ei(k1 z−ωt)

B~I = ĵ BI0 ei(k1 z−ωt)
2π
k1 is the wave vector in medium-1, and k1 = λ1
• Reflected wave

E~R = î ER0 e−i(k1 z+ωt)

B~R = −ĵ BR0 e−i(k1 z+ωt)

• Transmitted wave

E~T = î ET 0 ei(k2 z−ωt)

B~T = ĵ BT 0 ei(k2 z−ωt)
2π
k2 is the wave vector in medium-2, and k2 = λ2
~ that are tangent to the surface the boundary must be continuous
• The components of the electric field vector E,
(Gauss’ law)

E~It + E~Rt = E~T t

⇒ EI0 + ER0 = ET 0

• Generally the tangential components of the magnetic field intensity (H = B

µ ) is discontinuous across the boundary.
• But the boundary condition at the interface between two dielectric materials, where there is no free charge or

free current is,

~I + H~R = H~T
H
BI0 BR0 BT 0
− =
µ1 µ1 µ2
CHAPTER 1. INTRODUCTION iv

BI0 BR0 BT 0
− =
µ1 µ1 µ2
E √
Now B = = E µ. Remember v = √1
v µ

√ EI0 √ ER0 √ ET 0
µ1 1 − µ1 1 = µ2 2
µ1 µ1 µ2
1 2
r r
(EI0 − ER0 ) = ET 0
µ1 µ2

Now we have two boundary conditions

EI0 + ER0 = ET 0
1 2
r r
(EI0 − ER0 ) = ET 0
µ1 µ2

Exercise:- Obtain an expression for the amplitude of the transmitted wave in terms of the amplitude of the incident
wave.
ET 0 = (?) EI0

EI0 + ER0 = ET 0 (1.1)

1 2
r r
(EI0 − ER0 ) = ET 0 (1.2)
µ1 µ2

We substitute for ER0 from Eq.(1) into Eq.(2),

1 2
r r
(EI0 − ET 0 + EI0 ) = ET 0
µ1 µ2
r r 
1 1 2
r
2EI0 = + ET 0
µ1 µ1 µ2
⇒ q
2 µ11
E T 0 = q q  EI0
1 2
µ1 + µ2

Exercise:- Obtain an expression for the amplitude of the reflected wave in terms of the amplitude of the incident
wave, i.e., ER0 = (?) EI0

EI0 + ER0 = ET 0

From the first equation,

ER0 = ET 0 − EI0

But, q
1
2 µ1
E T 0 = q q  EI0
1 2
µ1 + µ2

Now,
 q 
2 µ11
ER0 = EI0  q q  − 1
 
1 2
µ1 + µ2

q q 
1 2
µ1 − µ2
⇒ ER0 = q q  EI0
1 2
µ1 + µ2
CHAPTER 1. INTRODUCTION v

Refractive index of a medium is defined as,

c µ
r
n= =
v µ0 0
One can safely take µ ≈ µ0 , This implies, √
n∝ 
Now, the amplitude of the reflection becomes,

(n1 − n2 )
ER0 = EI0
(n1 + n2 )

If the refractive index of the second medium is grater than that of the first, i.e., n2 > n1

n2 − n1
 
ER0 = − EI0 Phase change of π!
n1 + n2

We have,
(n1 − n2 )
ER0 = EI0
(n1 + n2 )
Or, the amplitude of the reflection coefficient
ER0 n1 − n2
r= =
EI0 n1 + n2
Also,
2n1
ET 0 = EI0
n1 + n2
Or, the amplitude of the transmission coefficient
ET 0 2n1
t= =
EI0 n1 + n2
The reflectivity (R) and transmittivity (T ) are defined as,

R = |r|2 = r∗ r
T = |t|2 = t∗ t (1.3)

At this point we are equipped with the necessary tools to discuss the theory of multiple-beam interference.
Chapter 2

Multiple-Beam Interference

As promised before, we extend our knowledge of two-wave interference to the case where a bunch of waves superpose1 .
The interfering waves have to be coherent to have a consistence interference pattern. A common method to
produce coherent multiple beam is division of amplitude via multiple reflections between two parallel reflecting
surfaces(mirror, film, etc.). A model set-up of two parallel surfaces separated by a distance d is given in Fig. ??.

Figure 2.1: The transmitted waves (E0 t2 , E0 t2 r2 , E0 t2 r4 .. etc.,) after multiple reflections superpose to give an interference
pattern.

Now, we need to find the path differences between consecutive transmitted waves. In Fig. ??, θ represents angle
between any of the internally reflected ray and the normal to the surface. The path difference between two waves
shown in the figure is obtained by drawing a perpendicular from B to the point D, and is given by BC + CD. As
the triangles ABC, and BCD are congruent, one can choose the path difference to be AB + BC.

From ∆BCE,
d d
cos θ = ⇒ BC =
BC cos θ
Similarly, from ∆ABC,
AB d cos 2θ
cos 2θ = ⇒ AB =
BC cos θ
The path difference between the highlighted rays is given as,
d d cos 2θ d
BC + AB = + = (1 + 2 cos 2θ) = 2d cos θ
cos θ cos θ cos θ
The corresponding phase difference is obtained as,
1 This chapter is prepared based on [?]

vi
CHAPTER 2. MULTIPLE-BEAM INTERFERENCE vii

Figure 2.2: The path difference between two consecutive transmitted wave is BC + CD. The line BE is the normal drawn
at the second reflecting surface

2π 4π
δ= × 2d cos θ = d cos θ
λ λ
The phase difference between two adjacent waves can be expressed in terms of vacuum wavelength (λ0 = nλ) as,
4π
δ= nd cos θ (2.1)
λ0
Now, we calculate the resultant amplitude at the screen due to superposition of multiple beams as shown in Fig.
??. A phase difference of δ changes the amplitude by a factor of eiδ . The expression for total amplitude is given by
a geometric series with common ration r2 eiδ , and is given by,

ET ot = E0 t2 + E0 t2 r2 eiδ + E0 t2 r4 e2iδ + ...

h i
= E0 t2 1 + r2 eiδ + r4 e2iδ + ...
E0 t2
=
1 − r2 eiδ
Corresponding expression for the internsity is given as,

|t4 |
I = |ET ot |2 = I0 , (2.2)
|1 − r2 eiδ |2

where I0 = E02 is the intensity of incident beam. Note that while deriving Eq. ??, we have not considered the effect
of sudden phase change that occurs in certain cases. In the previous section, we have found out that there will be
a sudden phase shift of π if a wave traveling from a dielectric of lower refractive index gets reflected at an interface
with a medium having higher refractive index. In the case of a dielectric medium the possible phase changes that
can occur is either zero or π depending on the relative value of refractive index. However, the value of phase-shift
can be any value for the case of metals. In general, we can express the coefficient of reflection as,
δr
r = |r|ei 2 ,
δr
where 2 is the phase change caused by reflection. The modified expression for the total intensity becomes,

T2
I = I0 (2.3)
| (1 − R ei∆ ) |2

Here, R and T denote the reflectance and transmittance respectively. We have to work more to obtain a condition
CHAPTER 2. MULTIPLE-BEAM INTERFERENCE viii

for maxima or minima. Let’s begin by expanding the denominator of Eq. ??.
| 1 − Rei∆ |2 = 1 − R ei∆ 1 − R e−i∆

= 1 − R ei∆ + e−i∆ + R2

= 1 − 2R cos ∆ + R2
 
∆
= 1 − 2R 1 − 2 sin2 + R2
2
2
h 4R 2 ∆
i
= (1 − R) 1 + 2 sin
(1 − R) 2
Now, the expression for total intensity obtained after multiple-beam interference becomes,
T2 1
I = I0 2 . (2.4)
(1 − R2 ) 1 + F sin2 ∆
2
Here, F is called the coefficient of finesse and it is a measure of the sharpness of the interference fringes,
4R
F = 2.
(1 − R)
The term,
1
,
1 + F sin2 ∆ 2
is called Airy function. The behaviour of Airy function for different values of R is plotted in Fig. ?? The maximum

Figure 2.3: Plot of Airy function for different values of R. As R increases the sharpness of the fringe increases.

intensity is obtained for ∆

2 = N π, where N is some integer (1, 2, 3...) called the order of interference. Using Eq. ??,
the condition for maximum intensity becomes,
4π
2N π = nd cos θ + δr (2.5)
λ0

2.1 The Fabry-Perot Interferometer

Based on the theory of multiple-beam interference, C. Fabry and A. Perot designed an interferometer in 1899. The
Fabry-Perot interferometer is used to measure the wavelength of an unknown source with high precision. Also, this
instrument is used to study spectral lines. The set-up consists of two optically flat partially reflecting plates of glass
or quartz. This devise is used in two different ways depending on the separation between the parallel surfaces. If
the distance between them is kept constant, it is called Fabry-Perot etalon. If the distance between the partially
reflecting surfaces is varied by mechanically or optically, then the set-up is called Fabry-Perot interferometer.

A schematic diagram of Fabry-Perot etalon is given in Fig. ??. In this set-up, a broad source of light is directed
to the parallel surfaces through a collimating lens. The set of transmitted beams are focused on the screen to obtain
concentric dark and bright rings called fringes of equal inclination. A given ring corresponds to a set of rays with
same angle θ.
In the second method shown if Fig. ??, called the scanning method a point source is used and it allows to vary
the separation between the parallel surfaces. The intensity at the ring centre will be a sum of Airy patterns and
usually recorded photoelectrically.
 The Fabry-Perot works by having light reflect back and forth many times in a space bounded
by two very flat, highly reflecting mirrors. The set of mirrors is called an “etalon” (see Fig. 8).
The intensity of the light exiting the etalon is maximum when the phase shift of the light is an
integer multiple of 2π after reflecting from both mirrors (this is the requirement for constructive
interference). Light of differing wavelengths satisfies this condition at different exit angles from the
CHAPTER 2. MULTIPLE-BEAM INTERFERENCE etalon resulting in visible rings of differing circumference for the different wavelengths. The output
of the Fabry-Perot is not just one visible ring for each unique wavelength, but rather a repeating
ix
pattern of rings, successive patterns representing an additional phase shift of 2π.

INTER 1 EPOS 111 TER 91

Figure 8: Schematic illustration of a Fabry-Perot interferometer and the ring pattern it produces
on a screen when used with a monochromatic light source.
Figure 2.4: Schematic illustration of a Fabry-Perot etalon and the ring pattern it produces on a screen when used with a
The strength of the Fabry-Perot is its ability to resolve lines with closely spaced wavelengths. If
monochromatic light source ∆λmin is the minimum resolvable wavelength difference at wavelength λ, then for our instrument,
λ/∆λmin ≈ 106 .
A limitation of the Fabry-Perot is that it can only work with wavelength or frequency differences
up to a certain maximum which turns out to be

λ2
∆λmax = , (18)
2d
c
∆νmax = , (19)
2d
where λ is the nominal wavelength of the light entering the interferometer and d is the separation
between the mirror surfaces. This maximum difference in frequency or wavelength is called the
free spectral range (FSR) of the interferometer. At this wavelength difference, the ring due to light
with wavelength λ ± ∆λmax will just overlap the adjacent ring (remember, the pattern repeats!)
due to light with wavelength λ. This kind of overlap will result in a ring pattern that is difficult to
interpret, a situation that can usually be avoided by a judicious choice of mirror spacing.

11
(b

Figure 2.5: Arrangement of ent f 4r° (g:a; interferometer

-ige scanning t lc 94abry-Perot etalan;
with (b) a
thetypical
;canning in-
recorded output.
tc w' lncter.

2.1.1 The Free Spectral Range

the interferroancter° is usually mounted between a colli-
en zind fog-using lens as shovvin in Figure 4.4(a). If a broad
The free spectral range of of Fabry-Perot interferometer
cat, interferenceisfringes
a. defined
in theas
ormthe separation between adjacent orders of
of concentric
appearin
interference. In other words, it is the spacing in optical
the 1" cal frequency/wavelength
plane of the focusing lens (Figure
between two successive maxima or
mg x n,:a be ohser-t ed v sually or photographed. A given
minima. This idea can be illustrated by considering a source
nstant in which
value of()., and the cthe frequency
rculaar fringes are can be varied. The free spectral
, wl nation. ,Another
range would be the difference in frequency for two successive maxima wayorof minima.
using the inter-
Suppose we start with a frequency
thong iiictliod employs point source or a
ω, the condition for N th order maxima is, i c,ed so 4hart only one the center of the
l'Xi L a . ca.`t plane as sli as Figure 4.4(h).
4πnd cos θ by clan 2ω
plished N nd cos θ eitl
∆N = δr + = δr + in, ring
= 2N π
λN by changing the c re. The in t. :dty
,corded hotoele h.. I -ally. This gives a
"-d is essentially a The g rfaphic
Now, we change the frequency to ω 0 such that N th maxima becomes10° (N + a 1).
sum In terms ω 0 , the condition for maxima
of such
now reads, shown in
0
2ω(N +1) nd cos θ
∆0N +1 = δr + = 2(N + 1)π
c
The free spectral range is defined as the separation between these two path differences, i.e.,

∆0N +1 − ∆N = 2π (2.6)

Equivalently,
0
2ω(N +1) nd cos θ
2ωN nd cos θ πc
− = 2π ⇒ ω0 − ω =
c c nd cos θ
The free spectral range can be expressed in terms of ν as,
0 c
ν(N +1) − νN =
2nd cos θ
For small values of θ, the free spectral range becomes,
0 c
ν(N +1) − νN = (2.7)
2nd
The physical significance of free spectral range is that if you shine a spectrum of light into the instrument, then as
long as the range of frequencies in the spectrum is less than the free spectral range then the interference patten will
be distinct from one another. However, if the spectrum is wider, then the overlapping interference orders make it
harder to interpret the output.
 For small tl a .r, e in frequency is approximately
C
VN+ 1 (4.16)
2nd

CHAPTER 2. MULTIPLE-BEAM INTERFERENCE x

4.3 Resolution of Fabry-Perot Instruments
Suppose a spectrum consisting of two closely spaced frequencies
2.1.2 Resolution of Fabry-Perot
and w' is to beInterferometer
analyzed with a Fabry-Perot interferometer. The in-
tensity distribution will be a superposition of two fringe systems as in-
The Fabry-Perot interferometer is frequently used
dicated in Figure 4.6.to Here
examine
the twothe detailed
components arestructure of ofclosely spaced spectral lines.
assumed to be
Consider a spectrum consisting of equal intensity.
closely spaced The frequencies
fringe pattern isωgiven
andbyωthe
0 sum of two Airy
. The corresponding path differences are
0 functions:2ω 0 d
denoted by ∆ = δr + 2ωd c and ∆ = δr + c , respectively. Now, the resultant intensity profile will be a partially
A )-1
overlapping distribution of individual intensities as+ Fshown sing
2
in +Fig. ??.
10 1 + Therefore, (4.17)
F sin2 a certain degree of ambiguity
exists in deciding if two spectra arein distinct, i.e., if the closely spaced
which F is defined by Equation (4.9), and where
lines are said to be resolved. Consequently,
there are many criterions that characterises a resolved spectra.
2cwd

Figure 4.6. Graph of intensity distribution for two monochromatic lines in

Fabry-F'erot
Figure 2.6: Overlapping interferometry.
intensities of two closely spaced spectral lines.

According to Lord Rayleigh’s the fringes are just resolved when the combined irradiance of both fringes at the
center, or saddle point of the resultant broad fringe is π82 times the maximum irradiance. Another condition which
we are interested is called Taylor criteria says two equal lines are considered to be resolved if the individual curves
cross at half intensity point so that the total intensity at the saddle point is equal to the maximum intensity of the
individual curve. From Eq. ??, the N th maximum intensity corresponding to ω and ω 0 occur at ∆ = 2N π and
0
∆0 = 2N π respectively. Also, the maximum intensity for the combined curve will be at 2N π + ∆ 2−∆ . According
to Taylor’s criteria the maximum intensity at this point will be equal to the maximum intensity of the individual
curve, i.e., we have the condition
T2 1 T2
I0 2 (∆0 −∆)
= I0 2.
(1 − R2 ) 1 + F sin2 4
(1 − R2 )
Here, the L. H. S. corresponds to the intensity of the combined curve at the saddle point. The above condition lead
us to,
(∆0 − ∆)
F sin2 =1
4
2
∆0 − ∆

1
⇒ F ≈1 or ∆0 − ∆ = 4F − 2 , (2.8)
4
The approximation is justified since the individual frequencies are closely spaced. In terms of frequencies, the
resolution of Fabry-Perot interferometer can be expressed as,
c 1−R
 
0 2c − 1
δω = |ω − ω| = F 2 = √ (2.9)
d d R
δω is the fringe width at half-intensity or the resolution width of Fabry-Perot interferometer with plate separation
d and reflectivity R. Another parameter that gives a measure of how narrow the peak is in the intensity-frequency
plot is called the reflecting finesse (F). It is defined as the ratio of free spectral range (peak separation) to the peak
width, i.e.,
∆0 − ∆N
F = N +1 0
|∆ − ∆|
Using Eq. ?? and Eq. ??, the expression for reflecting finesse becomes,
√
π√ π R
F= F = (2.10)
2 (1 − R)
Note that the reflecting finesse is independent of the separation between the surfaces of the interferometer.
CHAPTER 2. MULTIPLE-BEAM INTERFERENCE xi

2.1.3 Resolving Power

The resolving power (RP) of an interferometer is defined as the reciprocal of fractional resolution, i.e.,

ω ν λ
RP = = =
δω δν |δλ|

Here, ω is the frequency corresponding to maximum intensity and δω denotes the resolution of the interferometer.
As we discussed before, the maximum intensity occur when the phase difference become,
4π 4πd 2dω N πc
δ= d cos θ ≈ = = 2N π ⇒ ω=
λ λ c d
Using Eq. (??) and Eq. (??), we obtain an expression for resolving power in terms of reflectance as,
√
R
R P = NF = Nπ (2.11)
(1 − R)

The resolving power can be increased by increasing the order of interference. This is achieved by increasing the
ω
plate separation since N = πc d. However, as d increases the free spectral range decreases according to Eq. ??.
This is not desired because free spectral range characterises the separation between two nearby peaks. Therefore,
one cannot arbitrarily increase the plate separation. Another way to get a high resolution is to use materials with
higher reflectance. Using multi-layer dielectrics the reflectance can be made close to unity. One can control the
reflectance and transmittance of an optical medium by specially designed layers of different dielectrics called films.
In the following section, we develop the theory of multi-layer films.

2.2 Theory of Multi-layer Films

There are many situations in our day-to-day life where the reflectance of a given material needs to be increased or
decreased. For example, we want minimum reflectance for the contact glasses so that a maximum amount of light
enters the eyes. On the other hand, you may also want to stop high-energy lights like ultraviolet rays entering your
eyes that might harm your vision. Depending on the requirements, one can reduce or enhance the reflectance of a
given material by selectively coating one or more layers of thin films. Here, we develop the theory of multi-layer
films for normal incidence.
To begin with, we consider a single layer of film with length l and refractive index n1 coated on a substrate of
refractive index nT . The substrate-film system is in air having refractive index n0 ≈ 1. Now, suppose a light ray
described by its electric field component E = E0 ei(kz−ωt) is travelling through the given system along z direction.
Let the air-film interface is at z = 0, see Fig. ??.
4.4 . THEORY OF MULTILAYER FILMS 97

Figure 4.7. Wave vectors and their dv ssociated electric fields for the case of
normal incidence on a single dielectric layer.
Figure 2.7: The incident, reflected, and transmitted fields and corresponding wave vectors are shown in different media.
cidence are easily made. The amplitude of the electric vector of the
Our goal is to find out theincident beam
relation is Ep. That
between theofamplitudes
the reflected of
beam is 'p', and
incident andthat of the
transmitted rays using electromagnetic
transmitted beam is ET. The electric-field amplitudes in the film are E1
boundary conditions at the and interfaces. The electric field vectors are parallel (directed
E1' for the forward and backward traveling waves, respectively, as up in the diagram) to the
surface at the first interface,indicated
thus the filed
in the is continuous, i.e.,
figure.
The boundary conditions require that the electric and magnetic
fields be continuous at each + E00 = These
E0 interface. E10 .
E1 +conditions are expressed (2.12)
as follows:

First Interface Second Interface

Electric: E0 + E0 = E1 + El E1 eikl + IJ e-xkl E

Magnetic: Ho - 11'==11, - H' H1e ikl - Hle-ikl = HT
CHAPTER 2. MULTIPLE-BEAM INTERFERENCE xii

Here, E0 and E00 are the amplitudes of incident and reflected rays in air, and E1 and E10 are the incident and
reflected rays in the film. These fields are equated at the interface, i.e., z = 0. The magnetic field (H = B/µ )
associated with the wave is also parallel to the the film. The parallel components of magnetic fields are generally
discontinuous across the boundary. However, since there is no free current flowing along the interface, the magnetic
fields are also continuous.

H0 − H00 = H1 − H10
B0 B0 B1 B0
− 0 = − 1
µ0 µ0 µ1 µ1
In the air, the magnetic field associated with the incident wave is H0 , and the reflected wave is H00 . Similarly, the
magnetic fields H1 and H10 correspond to forward and backward moving rays in the film. Note that the negative
sign before the reflected rays is because the magnetic field direction gets reversed under reflection to maintain the
transverse nature
√ of the wave. Since the electric and magnetic filed amplitudes of a given light wave is related by
B = E/v ≈ E = nC, where n is the refractive index and v denotes the velocity of the wave in the medium, one
can express the above continuity relation as,

n0 E0 − n0 E00 = n1 E1 − n1 E10 (2.13)

One can obtain similar relations between amplitudes at the second interface as well, i.e.,

E1 eik1 l + E10 e−ik1 l = ET (2.14)

n1 E1 e ik1 l
− n1 E10 e−ik1 l = nT ET . (2.15)

Note that the term e±ikl appears as the waves have traveled through a distance l to reach the second interface at
z = l. Also, k~1 = k~10 . One has to solve four relations given by Eq. ??, ??, ??, and ?? to find out the relation
between incident and transmitted amplitudes. This is achieved by replacing E1 and E10 in Eq. ?? and ??. First we
solve for E1 and E10 using Eq. ?? and ?? as,
 
1 nT
E1 = 1+ e−ikl ET
2 n1
 
1 nT
E10 = 1− eikl ET (2.16)
2 n1
Now, Eq ?? becomes,
E00
 
nT ET
1+ = cos kl − i sin kl . (2.17)
E0 n1 E0
Similarly, Eq. ?? becomes,
E00 ET
n0 − n0 = (−in1 sin kl + nT cos kl) (2.18)
E0 E0
Eq. ?? and Eq. ?? relates the incident amplitude with reflected and transmitted amplitudes. Here, the fractional
amplitudes can be written in terms of familiar coefficients as:
E00
Coefficient of reflection r =
E0
ET
Coefficient of transmission t =
E0
An useful representation of Eq. ?? and Eq. ?? is given in terms of matrices as,

− ni1 sin kl
      
1 1 cos kl 1
+ r= t (2.19)
n0 −n0 −in1 sin kl cos kl nT

It is easy to see that the effect of film is represented by a square matrix operation, and is called transfer matrix
([m]). Symbolically we express,
     
1 1 1
+ r = [m] t (2.20)
n0 −n0 nT
CHAPTER 2. MULTIPLE-BEAM INTERFERENCE xiii

The advantage of expressing the relation between the coefficients of reflection and transmission in terms of matrix
equation is that it allow us to generalize the expression for the case of multiple films. Suppose there are N films of
refractive indices n1 , n2 , .....nN with thickness l1 , l2 , ..., lN . Now, Eq. ?? is generalized to,
     
1 1 1
+ r = [m1 ] [m2 ].....[mN ] t (2.21)
n0 −n0 nT

The product of N square matrices is another square matrix and we denote it by M , i.e.,
 
A B
[m1 ] [m2 ] ...[mN ] = [M ] =
C D

The matrix elements A, B, C, and D are in general complex numbers. The couples relations in Eq. ?? is solved for
r and t to obtain,
An0 + BnT n0 − C − DnT
r =
An0 + Bn0 nT + C + DnT
2n0
t = (2.22)
An0 + Bn0 nT + C + DnT

Now, the reflectance R = |r|2 and transmittance T = |t|2 . As mentioned before, one of the application using thin
films is to reduce the reflection, these films are called anti-reflecting films.

2.2.1 Anti-reflecting Films

Consider a single film of refractive index n and thickness l coated on a substrate of refractive index nT . The transfer
matrix corresponding to the action of the film is,

− ni sin kl
 
cos kl
M=
−in sin kl cos kl

Now, we want to reduce the reflection for a light beam of wavelength λ. Therefore, we choose the the thick ness of
the film to be l = λ/4(orkl = π/2). Now the elements of the above transfer matrix becomes,

i
A = 0; B=− ; C = −in; D = 0;
n
Suppose the refractive index of air is unity n0 = 1, using Eq. ?? the reflectance becomes,

2
nT − n2
R= 2 . (2.23)
(nT + n2 )
√
It is straight forward to see from the above expression that the reflectance becomes zero if n = nT . A thin film
of Magnesium fluoride (n = 1.35) is commonly used to reduce the reflections of a glass (nT = 1.5) where the above
condition is approximately satisfied. It can be shown that the reflectance can be further reduced by using more
number of such films.

2.2.2 High-reflectance Films

The reflectance is enhanced by coating alternative layers of high and low refractive indices with same thickness Fig.
??. Consider a pair of thin films having refractive indices nH and nL placed on a given substrate of refractive index
nT . The thick ness of these films satisfies l = λ/4. Also, let M1 and M2 are the transfer matrices corresponding to
the film with refractive index nH and nL respectively. We have,

− niH
 
0
M1 =
−inH 0

− niL
 
0
M2 =
−inL 0
CHAPTER 2. MULTIPLE-BEAM INTERFERENCE xiv

The resultant transfer matrix of two layers is given by,

 nL
− nH

0
M2 = M1 M2 = (2.24)
0 − nnHL

One can generalize this expression for the case of N pairs of such films (i.e., a total of 2N layers) as,
 N 
N nL
− 0
 nL
− nH 0  nH
M2N = M = = (2.25)

− nnHL
N 
0

nH
0 − nL

Now, the
100reflectance is obtained
MULTIPI _ tIEAusing Eq.
Ii ?? as,

Figure 4.9. ulti' , .... foi

Figure 2.8: A stack of films having alternative high and low refractive indices on a substrate.
. g_ .c n-
si is of ter'i ate guar ndIO'-Yn . ma-
(lie e: A.. is th e j

"  nL 2N #2
nH − nT
2
R = |r| =  2N
, (2.26)
igh-Re ectance fflitrs In order to n 11 l e o- r

n
+ nT
L - ance h°
t .

n H
in a multllayc `elm. I acl os. !., G i i' ° l ; j', an d ¢

where we have assumed the refractive

low index, indicesi n.ls
n L, mate ,
t approaches unity as N →r ∞.
0 =is1. The reflectance ,,
1 leg
wavelength, 4.9. r c t.. 11 of the x E .

same form.,,f.w Fd L e pi-oduc

2.3 Problems
2.1 Show 0 0 0
√ that
√ √ the radii
√ of the interference fringes
ni
of a plane-parallel Fabry-Perot interferometer
0
are proportional
to 0, 1, 2...., N 0 , provided there is a zero-radius fringe. Here N is an integer. [Hint: use Eq.(4.35)
?? with the
θ2
approximation cos θ ≈ 1 − 2 ]
2.2 Calculate the reflectance of an 8 layer-stack (N=4) of zinc sulfide (nH = 2.3) and magnesium fluoride (nL = 1.35)
f the
on a substrate of stack
refractive consist,;
index 1.5. What ofwould be
layers,
the value of R if N = 10?sfer of the
complete multilayel: fill

(4.36)
0

Assuming, for siiiity, that P-- jot unit\r, the r ii ;. ance

of a multilayer stack is "O',, en by- n (4.30'
(nil/
2
(4.37)

The reflectance tti L r s approach

Chapter 3

Theory of Diffraction

3.1 Introduction
Naively, one understands the phenomenon of diffraction as the encroachment of light into the geometrical shadow
of an object. Note that this idea is not different from interference discussed in the previous chapter, and there is no
physical difference between these two concepts. Both interference and diffraction have their roots in the principle of
superposition. Usually, one distinguishes these two phenomena according to the size of the slit (aperture) and the
intensity pattern on the screen. A typical arrangement of light source-aperture-screen results a diffraction pattern
if the given slit is large enough so that it can be considered to be a bunch of many point sources. Consequently,
the intensity pattern will be a big bright spot at the symmetry-centre of the set-up followed by less bright maximas
on either sides of the symmetry line Fig. ??. On the other hand, interference is the superposition of a fewer point
sources that results an intensity distribution of equally spaced fringes of same width. Also, interference and diffrac-
tion are a particular case of a general phenomenon called scattering in which the obstacle need not be apertures or
pin holes instead it can be something more general like a water drop.

Figure 3.1: A typical diffraction set-up and its intensity distribution pattern.

The phenomenon of diffraction is formally explained by Christiaan Huygens and Augustin-Jean Fresnel. Ac-
cording to Huygens-Fresnel principle, the propagation of a light wave can be predicted by assuming that each
point of the wavefront acts the source of a secondary wave that spreads out in all directions. The envelope of all
the secondary waves becomes the new wavefront and more and more newer wavefronts are generated. Now, the
waves that originates from the portion of those parts of a wavefront that are not blocked by the obstacle forms the
diffraction pattern on the scree. The diffraction pattern caused by a straight edge is shown in Fig. ??.

In this chapter, we will learn. the theory of diffraction using Fresnel-Kirchhoff formula, a mathematical formu-
lation of Huygens principle.

xv
CHAPTER 3. THEORY OF DIFFRACTION xvi

WAVEFRONT

Incident lightT

GEOMETRICal SHADOW

STRIGHT EDGE SCREEN

Figure 3.2: Diffraction due to a straight edge. Light reaches within the geometrical shadow is due to the waves coming
from the wavefront.

3.2 Kirchhoff Integral Theorem

The underlying principle behind diffraction is the superposition principle which deals with the resultant amplitude
of multiple waves at a given point in space. Therefore, our first goal is to find a mathematical description of the
optical disturbance reaching a receiving point from a source point. This is achieved by Kirchhoff’s integral theorem.
The mathematical expression for Kirchhoff’s integral theorem relates the value of any scalar wave function at any
given point inside an arbitrary closed surface to the value of the wave function at the surface.

The primary step is to derive Green’s theorem. In order to do so, let us consider a vector field F~ constructed
using two continuous and integrable scalar functions U (r, t) and V (r, t) as,

F~ = U ∇V − V ∇U (3.1)

The divergence of F~ is given by,

∇ · F~ = ∇ · (U ∇V − V ∇U ) = U ∇2 V − V ∇2 U, (3.2)

where we have used the vector identity,

∇ · (U ∇V ) = U ∇2 V − (∇U ) · (∇V ) (3.3)

The divergence theorem , I Z  

F~ · n̂ dA = ∇ · F~ dV. (3.4)
S
Here, the volume V is the region enclosed by the surface S having outward normal along n̂. dA, dV denote the area
and volume elements respectively. Substituting Eq. ?? and Eq. ?? into Eq. ??, we obtain the Green’s theorem,
I Z
V ∇2 U − U ∇2 V dV,


(V ∇n U − U ∇n V ) dA = (3.5)
S

where ∇n denotes the normal component of the gradient at the surface of integration. As we are dealing with light
waves which are described by harmonic functions, we can safely assume the scalar functions U and V satisfy wave
equation, i.e.,
1 ∂2U 1 ∂2V
∇2 U = 2 2 ; ∇2 V = 2 2 ,
u ∂t u ∂t
where, u is the propagation speed of U and V . Now, we can also assume that these functions have harmonic time
dependance of the form e±iωt . Hence the volume integral in Eq. ?? vanishes, giving,
I
(V ∇n U − U ∇n V ) dA = 0 (3.6)
S
CHAPTER 3. THEORY OF DIFFRACTION xvii

Now, we will represent the optical disturbance by the function U . To obtain Kirchhoff’s integral formula we take
the function V to be a spherical wave given by,
7.7. KIRCHHOFF THEORY† ei(kr+ωt)
181
V (r, t) = V0 (3.7)
r
Now suppose that we are interested in evaluating the disturbance at the point P ,
In figure ?? we have considered a volume of arbitrary shape in which the location P represents the converging point
which is enclosed by the surface S (Fig. 7.16). We set ψ1 = be
of the spherical wave (i.e., r = 0). It is clear that the divergence theorem cannot
the amplitude
ψ,applied of the
because the function V
harmonic
blows up at P .disturbance. Wemodify
Therefore, one must are free to choose
the divergence anybyfunction
theorem forpoint
removing the ψ2 ,P provided it method
. The standard also
satisfies (7.39). It will prove useful to define ψ2 as
∂
is to carve out a small sphere of radius ρ around P creating an inner surface Σ. At Σ, r = ρ and ∇ n = − ∂r . Now,
the Eq. ?? becomes,
Z  ikr
e eikr

ee ikr∂U
Z " ikr
∂ eikr
 #
∇n U − U ∇n dAψ−2 = , − U ∂r r ρ2 dΩ = 0, (7.44)(3.8)
S r r Σ rr ∂r
r=ρ

wheredΩr=issinthe
where, θ dθ distance from
dφ is the solid . the
anglePon
iωt
This function
sphere centred athas a singularity
P and at r = 0area
ρ2 dΩ is the corresponding and so
element
on Σ. Note that the common factor V0 e has been canceled out. To obtain the desired result, we must evaluate
theintegral
the pointand must
P put be excluded
the limit ρ → 0. from the volume integral for Green’s theorem to be
valid. We can do this by placing a small sphere Σ around P . Green’s theorem may

Figure 7.16: A point P surrounded by a surface S of arbitrary shape. Kirchhoff’s formula

Figure 3.3: A point P surrounded by a surface S of arbitrary shape. Kirchhoff’s formula is derived by applying Green’s
is derived by applying Green’s theorem to the volume between S and an infinitesimal sphere
theorem to the volume between S and an infinitesimal sphere Σ surrounding P .
Σ surrounding P .
The integration on Σ can be simplified as,
now be applied
Z " ikrto the volume between
# Σ Zand S; these surfaces, together, make up
∂ eikr
 ikr
eikr
  
e ∂U e ∂U U ikr
the integration rsurface. On the surface of Σthe rsmall sphere
r the r2 outward normal to
2 2
− U ρ dΩ = − ikU + e ρ dΩ (3.9)
Σ ∂r ∂r r ∂r r=ρ
this volume is opposite to the direction r=ρ
of r and thus ∂/∂n can be replaced with
ρ → 0, and
−∂/∂r
As the esurface
the factor ikr
integral
→ 1 and over
the optical Σ mayapproaches
disturbance be expressed
the valueas
U has at P , namely UP . Therefore,
the integration over Σ reduces to,
! " # $
 ikr! #
% &'
∂ψ Z e∂ψ
"
ikr −e ikr ∂ψ ∂ eikr
2∂U ∂ e
Z
ψ2 − ψ r ∂r − dSU ∂r= r ρ2 dΩ = + UPψsin θ dθ dφ = 4πUdS
P (3.10)
Σ ∂n Σ ∂n Σ r=ρ r ∂r Σ ∂r r
$ % &'
Finally, Eq. ?? becomes, Z!  −eikr ikr ∂ψikr 
−eikr ikeikr
1 e e
UP = −= U ∇n − +∇nψU dA 2 + dS. (3.11)
4π SΣ rr ∂rr r r
This relation is known as the Kirchhoff ’s integral theorem. It relates the value of any scalar function at any point
(7.45)to
P inside an arbitrary closed surface tot the value of the wave function at the surface. We will use this theorem
derive Fresnel-Kirchhoff formula in the following section.
Now let us change this to an integral over solid angle Ω from the point P , in which
dS on Σ subtends dΩ and dS = r2 dΩ. Then
! ! " #
ikr ∂ψ
= −re − ψe ikr
+ rikψe ikr
dΩ. (7.46)
Σ around P ∂r
CHAPTER 3. THEORY OF DIFFRACTION xviii

3.3 Fresnel-Kirchhoff Formula

Fresnel-Kirchhoff formula is the mathematical statement corresponding to the Huygens principle. We consider a
typical diffraction set-up as shown in Fig. ??. We would like to estimate the total amount of optical disturbance
reaching at point P due to the waves originated from a monochromatic source S. We function U in Eq. ?? now
represents the light waves from S. It should be noted that the function U being a scalar cannot fully represent the
light wave. However, under scalar approximation, the square of absolute value of U may be regarded as a measure
of irradiance at a given point. The Fresnel-Kirchhoff formula is derived from Kirchhoff integral theorem using the
following two assumptions.
1. The wave function U and its gradient contributes negligibly to the integral except at the aperture opening
itself.
2. The value of U and its gradient at the aperture are the same as they would be in the absence of the partition.
We represent the light coming from S as spherical monochromatic waves as,
0
ei(kr −ωt)
U = U0 , (3.12)
r0
where r0 denotes the radial distance from the source to the aperture. Now, one can use the above expression in the
Kirchhoff integral theorem, i.e., Eq. ?? to obtain,
Z " ikr0 0
#
U0 e−iωt e eikr eikr eikr
UP = − ∇n − ∇n 0 dA (3.13)
4π r0 r r r

The normal component of the gradient is calculated as, 5.2 ° FUNDAMENTAL THEORY 109

Figure 5.2. Geometry for the Fresnel-Kirchhoff formula.

Figure 3.4: Huygens principle can be explained using Kirchhoff integral theorem by considering a closed surface in which
the the aperture lies on it. The surface encloses the point P where the resultant intensity is to be calculated.
Our task is to determine the optical disturbance reaching the
receiving point P from
 theikrsource
 S. In applyingthe Kirchhoff integral
we chooseeikra surface of
∂ integration such that it encloses
ikeikr theeikr
receiving

e
∇
point and = r̂ · n̂ = cos(n̂, r̂) −
n includes, as a part of it, the aperture opening as indicated in
r ∂r r r r2
the figure. 0 0
! 0 0
!
ikr
Twoebasic ∂ eikr
simplifying assumptions
0 are introduced:
0 ikeikr eikr
∇n 0 = r̂ · n̂ = cos(n̂, r̂ ) − 02 (3.14)
r ∂r0 r0 r0 r
(1) The wave function U and its gradient contribute negligible
Now, Eq. ?? becomes, amounts to the integral except at the aperture opening itself.
(2) The values of U and
ikU0 e−iωt
Z grad U 0at the aperture are the same
eik(r+r )
as
they would 0
− be in the absence of the partition.


UP = 0
cos(n̂, r̂) − cos(n̂, r̂ ) dA (3.15)
4π rr
Although
This expression is called the validity integral
Fresnel-Kirchhoff of theseformula.
assumptions
One iscan
open to considerable
understand the Huygens principle using the
debate, the results
following example of diffraction set-up. are generally in good agreement with experimental
observations.
If r' denotes the position of a point on the aperture relative to the
source S, then the wave function at the aperture is given by the
expression
ei(kr' -cot)
CHAPTER 3. THEORY OF DIFFRACTION xix

3.3.1 Diffraction by a Circular Aperture

Consider a circular aperture with a monochromatic light source as shown in Fig. ??. Every point on the aperture
is positioned at equal distance from the source so that cos(n̂, r̂0 ) = −1. Now, the value of the optical disturbance
110
at the point P is, D FFRACTION

The operations
ikU0 e−iωtindicated
Z ik(r+rin
e
0
) the ntegrand are carried out as
0


UP = −
follows: cos(n̂, r̂) − cos(n̂, r̂ ) dA
4π rr0
grade
eikr
r - cos (n,r
eikr f i .,ikr
Cos ln,r) 1 r - 9
(5.9)
(3.16)

For the sake of interpretation, let eikr'

us express the valueeof ikr'the wave function at P ,ikr" in the following format,
grade
r,
- cos
ik
(n,r'
Z r,
UA eai(kr−ωt)
-- cos (n,r')
ikeikr'

r' - r'2
(5.10)

UP = − (cos(n̂, r̂) + 1) dA, (3.17)

where (n,r') and (n,r')
4π denoter the angles between the vectors and the
normal to the surface of integration. Now, in Equations (5.9) and
where, (5.10) the second terms in parentheses are negligibly small compared
0

to the first terms in the normal eikr

UA =situation
U0 0 where both r and r' are much
r
larger than the wavelength of the radiation because k = 2i -/X. Con-
Here, UA is the complexsequently
amplitude Equation (5.8) gives
of the incident wave at the aperture and is called the primary wave according
to Huygens. From this primary wave, each area element aA of the aperture gives rise to a secondary spherical wave
eik(r+r')
given by, P
frr' [cos (n,r) -- cos (n,r' )] d (5.11)
ei(kr−ωt)
This equation is knownUas dA.Kirchhoff integral formula.
A the Fresnel
r
It is, in effect,
Now, the total optical disturbance a mathematical
at the statement
receiving point of Huygens'
P is obtained principle.
by summing allThis
the secondary waves from
is most easily seen by applying the formula to
each area element of the aperture.
This is essentially Huygens principle! a specific case, namely,
that of a circular aperture with the source symmetrically located as
The term cos(n̂, r̂) − cos(n̂, r̂0 ) in the Fresnel-Kirchhoff formula is called the obliquity factor. In example
shown in Figure 5.3. The surface of integration is taken to be a spher-

Figure 5.3. Diagram to show how Huygens' principle follows from the
Kirchhoff'
Figure 3.5: A circular aperture is usedintegral formula.
to explain Huygens principle Fresnel-Kirchhoff integral theorem.
ical cap bounded by;.., the aperture opening. In this case r' is constant
discussed above, one obtains maximum value of 2 for the obliquity factor if the point P is along the forward
and cos (n,r') = -1, The Fresnel-Kirchhoff formula then reduces to
direction from the aperture wavefront, ie., cos(n̂, r̂) = 1. However, the obliquity factor becomes zero if P is such
that cos(n̂, r̂) = −1. This point corresponds
iik
backward
U",
direction from the aperture wavefront. Therefore, the
ei(kr-wt)

it'ff
absence of backward propagation canUpbe- explained using r
[cos (n,r) + 1 integral
Fresnel-Kirchhoff ]. d formula.
(5.12)

where
3.4 Fraunhofer and Fresnel Diffraction
U0 eikr'
r'
In the previous section, we have discussed a mathematical statement of Huygens principle of diffraction. According
to this principle, the diffraction pattern is produced due to the those waves that originate from the secondary
wavefront at the aperture. The resultant amplitude on the screen is obtained by solving Fresnel-Kirchhoff’s integral
formula. The integration can be performed in two different ways based on the nature the wavefront at the aperture.
If the curvature of the wavefront of both incoming and diffracted waves can be neglected, then the Fresnel-Kirchhoff
integral formula becomes simplified and is called Fraunhofer diffraction. This is possible if the light source and the
screen are placed at infinitely large distance from the aperture. On the other hand, if the light source or the screen
is at a finite distance from the aperture, then the curvature of the corresponding wavefront for both incoming and
CHAPTER 3. THEORY OF DIFFRACTION xx

diffracted waves become significant. Thus, one cannot neglect the curvature of the wavefront, and such patterns
are called Fresnel diffraction. Now, let’s have a quantitative criterion that differentiates Fresnel and Fraunhofer
diffraction patterns.

Consider a diffraction set-up as shown in Fig. ??. The radial distance between the source (S) and the aperture
is denoted by r0 , and the distance between the aperture and the receiving point (P ) is labeled by r. Now, recall
the theory of diffraction detailed in the previous section using AND
5.3 ° FRAUNHOFER Fresnel Kirchhoff’s
FRESNEL integral theorem.
DIFFRACTION 113 According to
theorem, the total optical disturbance at the point P is obtained by integrating Eq. ?? along the wavefront on the
aperture surface. Therefore, the occurs
fraction curvaturewhenof both
the wavefront will
the incident andplay a crucialwaves
diffracted role in
arethis calculation. To understand
effec-
if the diffraction set-up tively
leads plane. This will
to a curved be the case
wavefront when
at the the distances
aperture site, wefrom the variable ∆ = r + r0 . We
the source
consider
express the variation of to∆ the diffracting
from aperture
the top edge andaperture
of the from the toaperture to the end
the bottom receiving
as, point
are both largep enough for the curvatures
p of the incident
p and diffracted p 
∆max − waves
∆minto=be neglected
d02 + (h0 +[Figure
δ)2 + 5.5d2(a)].
+ (h + δ)2 − d02 + h02 − d2 + h2 (3.18)
If either the source or the receiving point is close enough to the
Now, we simplify this expression
diffractingas,aperture so that the curvature of the wave front is signifi-
cant,
 then 0
one has
− 21Fresnel
 diffraction2 [Figure
− 12 5.5(b)]. There
− 12 is, of − 21
δ)2 line h02 However, h2

0course,(hno +sharp of (h +between
distinction δ) the two
0 cases.
∆max − ∆min = d 1 + +d 1+ − d 1 + 02 −d 1+ 2 (3.19)
d02 criterion can be obtained
a quantitative d2 d
as follows. Consider Figure d
5.6, which shows the general geometry of the diffraction problem.

fi
I Aperture
8

d'

Figure 5.6. Geometry to show distinction between Fraunhofer diffraction

and Fresnelofdiffraction.
Figure 3.6: Geometry diffraction set-up showing various length elements.

One can expand the The receiving

above point
expression in Pδ is
as,located a distance d from the plane of the
diffracting aperture, and the
 0sourceS is a distance
 d'from this plane.
One edge of the aperture ishlocatedh a distance
1 1 h from
1 the foot of the
∆max − ∆min = + δ+ + δ 2 + ... (3.20)
perpendicular drawn from Pd0to thed plane 2of the
d0 aperture.
d The corres-
ponding distance for the source is h' as shown. The size of the aper-
If the wavefront at the ture
aperture is plane,
opening then the
is 8. From the figure
variation
it is of ∆ will
seen that be
thelinear in δ.d of
variation Therefore,
the the quadratic term
in the expansion is a measure of the curvature of the wave front. Accordingly, the
quantity r + r' from one edge of the aperture to the other is given by curvature of the wave front can
be neglected if, -- d' (h ° T 8Y + 1  1+ (h 1  )2 _ 2 + h' -- -d+ h2
2
+ δ << λ (3.21)
2 d0 d (5.14)
2 (d, + d) Sz +
And the corresponding set-up gives Fraunhofer diffraction pattern. Similarly, if the quadratic term in Eq. ?? is
The quadratic
comparable to the wavelength used, theterm in the expansion
corresponding above
set-up willisbeessentially a measure of
Fresnel diffraction.
the curvature of the wave front. The wave is effectively plane over
the aperture if this term is negligibly small compared to the wave-
3.5 Fraunhofer Diffraction
length of the light, thatPattern
is, if

As explained before, the Fraunhofer diffraction pattern is due 2 <<to

X the optical disturbances
(5.15) coming from a planar
wave front at the aperture. This is true if the light source is placed at infinity. However, one can construct the
set-up using a near by point source but collimated by a lens as shown in Fig. ??. Also, one has to place a focusing
lens after the aperture to collect the diffracted rays on the screen.

To obtain the optical disturbance at the receiving point P , all we need to do is to calculate Fresnel- Kirchhoffs
formula given in Eq. ??. To make our life easy, let’s make the following approximations,
 hofer case is, in general, mathematically simpler than the Fresnel
case, Fraunhofer diffraction will be discussed first.

5.4 Fraunhofer Diffraction Patterns

CHAPTER 3. THEORY OF DIFFRACTION
The usual experimental arrangement for observing Fraunhofer dif- xxi
fraction is shown in Figure 5.7. Here the aperture is coherently

Figure 5.7. Arrangement for observing Fraunhofer diffraction.

Figure 3.7: Experimental set-up for Fraunhofer diffraction pattern.

illuminated by means of a point monochromatic source and a colli-
mating lens. A second lens is placed behind the aperture as shown.
The incident and diffracted wave fronts are therefore strictly plane,
cos(n̂, r̂) − cos(n̂, r̂0 ) does


1. The angular spread and the Fraunhofer
of the diffracted light case is rigorously
is small. valid. In applying
Therefore, the Fresnel-
the obliquity factor
Kirchhoff formula (:Equation (5.11)] to the calculation of the diffrac-
not vary much over the aperture and is taken outside the integral.
tion patterns, the following simplifying approximations are taken to be
0 valid:
2. The quantity eikr /r0 is nearly(1)aThe
constant and of
angular, spread is the
taken outside
diffracted light isthe
smallintegral
enough for
the obliquity factor [cos (n,r) -- cos (n,r')] not to vary appre-
3. The variation of the remaining factor eikrthe
ciably over /raperture
over aperture comes
and to be taken principally
outside the integral. from the exponential part, so the
(2)
factor 1/r is replaced by its mean The quantity
value and is is very nearly
taken constant and can be taken
outside.
outside the ;integral.
5.4 - FRAUNHOFER DIFFRACTION PATTERNS 115
Now, the Fresnel-Kirchhoff integral becomes,
(3) The variation of the remaining Z factor eikr/rr over the aperture
comes principally from the exponential part, so the factor l 1r
UP = C eikr dA, (3.22)
can be replaced by its mean value and taken outside the in-
tegral.
where all those constants pulled out of the integration constitutes the symbol C. Let’s exploit the above expression
Consequently, the Fresnel-Kirchhof formula reduces to the very
to obtain the intensity distribution due to a single slit.
simple equation

3.5.1 Single Slit Diffraction Pattern Up eikr dam' (5.16)

Consider the situation where monochromatic

where all light
constant factors is been
have passed through
lumped intoaone
narrow slit C.
constant as The
shown in Fig. ??. According
to Huygene-Fresnel theory of diffraction, the rays coming from various parts of the wavefront
formula above states that the distribution of the diffracted light is ob- at the slit interfere
to produce the diffraction patters
tained simplyatby
theintegrating
receiving the
plane.
phaseHere, the
factor incident
eikr waves
over the are plane, and the diffracted
aperture.
rays are focused using a lens tot the receiving plane. Therefore, the set-up falls under the category of Fraunhofer
diffraction. The Single Slit The case of diffraction by a single narrow slit is
treated here as a one-dimensional problem. Let the slit be of length L

+ b/2

Figure 5.8. Definition of the variables for Fraunhofer diffraction by a single

slit.
Figure 3.8: Fraunhofer diffraction due to single slit

and of width
To understand the nature b. The element
of diffraction of area
pattern, is then Fresnel-Kirchhoff
we exploit ddd = L dy as indicated in i.e., Eq. ??. Note that
formula,
Figure 5.8. Furthermore, we can express r as
the integration is carried out by taking the area element dA = Ldy, where, y varies from −b/2 to +b/2. To
r=r0+yatsin
proceed further, we consider all rays that are diffracted an 0angle θ with respect to(5.17)
the symmetry axis as shown
where rQ is the value of r for y = 0, and where 0 ris the angle shown.
The diffraction formula (5.16) then yields
r+ b12
U = Ceikro eiky sin O L dy
J - b/2
(5.18)
,ikro L sin ( kb sin ®) -,, s1
CHAPTER 3. THEORY OF DIFFRACTION xxii

in the figure. Fresnel-Kirchhoff formula basically integrate the quantity eikr over the slit area, where r denotes the
separation between areal element Ldy and the receiving point. One can express r using the following relation,

r = r0 + y sin θ

where r0 is the value of r for y = 0. The term y sin θ gives the path difference between waves coming from an area
element located at some y and the middle point. Now, Eq. ?? becomes,
Z + 2b
U = Ceikr0 eiky sin θ Ldy
− 2b
1


sin 2 kb sin θ
 
sin β
= 2Ceikr0
L = C0 ,
k sin θ β

where β = 12 kb, and C 0 = eikr0 CbL.

3.5.2 Double Slit Diffraction Pattern

3.6 Fresnel Diffraction Patters

3.6.1 Diffraction Due to Rectangular Aperture

3.7 Applications of Fourier Transform to Diffraction

3.8 Reconstruction of Wave Front using Diffraction (Holography)
3.9 Problems
Chapter 4

Propagation of Light in Crystals

Crystals are solids in which the atoms are arranged in a regular, repeating pattern. This ordered structure gives
crystals unique optical properties, such as the ability to refract and reflect light in specific directions. Crystals
are also used in many optical devices, such as lenses, prisms, and lasers. Out of many, birefringence is an optical
property of anisotropic crystals and is understood as the splitting of light into two rays as it passes through a
crystal. This is because the crystal has different refractive indices for light polarized in different directions. Such
crystals are called birefringent or doubly refracting crystals. Calcite is an example, see Fig. (??).

Figure 4.1: A calcite crystal showing birefringence.

Consider the situation where electromagnetic light is passed through a crystal. Due to the electric field, the
positive and negative charges in the crystal are displaced, resulting in a net polarization of the crystal. The
magnitude of the polarization depends on the strength of the electric field and the polarizability of the crystal. For
an isotropic crystal, if the electric field is E, the polarization vector is given by,

P~ = 0 χE,
~

where χ is the susceptibility scalar. However, if the crystal is anisotropic, i.e., the optical properties of the crystal in
different directions are different, then the polarization changes with the direction of E in relation to the orientation
of the crystal. Therefore, the expression for polarization becomes,

    
Px χ11 χ12 χ13 Ex
Py  = 0 χ21 χ22 χ23  Ey  (4.1)
Pz χ31 χ32 χ33 Ez

Note that the x- component of the polarization vector depends not only the x component of the electric field, but
y and z components as well. Here, χ which is expressed in matrix form is called the susceptibility tensor. In this
chapter, we intend to understand the theory of birefringence. The first step towards this goal is to solve Maxwell’s
equations!

xxiii
CHAPTER 4. PROPAGATION OF LIGHT IN CRYSTALS xxiv

4.1 General Wave Equation

Maxwell’s equations are a set of four partial differential equations that describe how electric and magnetic fields
are generated and interact with each other. Maxwell’s equations inside matter are,
∂B
∇×E = −
∂t
∂D
∇×H = +J
∂t
∇·D = ρ
∇·B = 0

Here, magnetization (M), magnetizing field (H), and magnetic field (B) are related as,

B = µ0 (H + M )

The displacement field D is given by,

D = 0 E + P
For the rest of the course, we consider only non-magnetic (M = 0) and electrically neutral (ρ = 0) materials.
Maxwell’s equations now reduces to,
∂H
∇×E = −µ0
∂t
∂E ∂P
∇×H = 0 + +J
∂t ∂t
1
∇·E = − ∇·P
0
∇·H = 0

One obtains the wave equation by taking curl of the first and time derivative of second, and add them together,
yielding,
1 ∂2E ∂2P ∂J
∇ × (∇ × E) + 2 2 = −µ0 2 − µ0 , (4.2)
c ∂t ∂t ∂t
where we have used the relation µ0 0 = c12 . Note that the terms on R. H. S. are the source terms for the electric
field. The solutions of above equation tell us how the source terms effects the propagation of electromagnetic wave
within the medium. However, depending on the nature of the medium, certain source term pays important role in
2
explaining certain physical phenomena. In the case of crystals, the term −µ0 ∂∂tP2 plays a crucial role in explaining
phenomena like dispersion, double refraction e. t. c. In the case of metals, it is the conduction term −µ0 ∂J ∂t that
is important, and the resulting solutions explain the large opacity and high reflectance of metals. Note that for
semi-conductors, both source terms must be taken into account.

4.2 Theory of Birefringence

We will restrict out discussion on crystals for the rest of the chapter. The wave equation inside a non-magnetic,
neutral crystal now takes the form,

1 ∂2E 1 ∂2E
∇ × (∇ × E) + 2 2
= − 2χ 2 , (4.3)
c ∂t c ∂t
where we have used the relation P = 0 χE. To proceed further, we exploit the technique of Fourier transforms.
The idea is the (x, y, x) (position) coordinates will be transformed to (kx , ky , kz ) (momentum) coordinate, and the
t (time) coordinate will be transformed to ω (frequency) coordinate. This transformation helps to solve Eq. (??)
easily with physical insights. In the Fourier space, Eq. (??) becomes,

ω2 ω2
k × (k × k) + 2
E = − 2 χE (4.4)
c c
CHAPTER 4. PROPAGATION OF LIGHT IN CRYSTALS xxv

Let ~k = (kx , ky , kz ), where kx , ky and kz denote the momentum components along x, y, z directions. In terms of
these components, above equation can be written a set of three equations as,
ω2 ω2
 
Ex −ky2 − kz2 + 2 + kx ky Ey + kx kz Ez = − χ11 Ex (4.5)
c c
ω2 ω2
 
Ey −kx2 − kz2 + 2 + ky kx Ex + ky kz Ez = − χ22 Ey
c c
2
ω2
 
ω
Ez −kx2 − ky2 + 2 + kz kx Ex + kz ky Ey = − χ33 Ez
c c
To understand the physical meaning of these relations we consider a special case in which the light propagates along
x direction only, i.e., kx = k and ky = kz = 0. The above set of relations now reduces to,

ω2 ω2
2
Ex = − χ11 Ex (4.6)
c c2
ω2 ω2

−k 2 + 2 Ey = − 2 χ22 Ey (4.7)
c c
2
ω2
 
ω
−k 2 + 2 Ez = − 2 χ33 Ez (4.8)
c c
Let us analyse each of these relations one by one. Eq. (??) can be expressed as,

ω2
(1 + χ11 ) Ex = 0
c2
As the susceptibility is always positive, the only non-trivial solution to the above relation is Ex = 0. This is not
surprising, by the way. We have chosen the direction of propagation along x-direction. Since electromagnetic waves
are transverse in nature, the electric or magnetic fields cannot have non-zero components along the direction of
propagation. Now, let us look at Eq. (??) which can be recast into the following form,
h ω2 i
− k2 + (1 + χ22 ) Ey = 0
c2
One can solve this equation immediately for Ey 6= 0 to obtain,
ωp ωp
k= 1 + χ22 = K22 , (4.9)
c c
√
where, K22 = 1 + χ22 . Now, the corresponding phase velocity (= ω/k) of the EM wave is given by,
c
v1 = √ (4.10)
K22
Similarly the solution to Eq. (??) for Ez 6= 0 is,
ωp ωp
k= 1 + χ33 = K33 , (4.11)
c c
√
where, K33 = 1 + χ33 . Now, the corresponding phase velocity is given by,
c
v2 = √ (4.12)
K33
What we have seen so far is the following: when an electromagnetic wave propagates along x-direction through a
crystal, the wave will have two different values of phase velocities if the crystal’s susceptibility is not isotropic (here,
χ22 6= χ33 ). Therefore, the material exhibits two values for the refractive indices corresponding to these different
phase velocities. This is the phenomenon of birefringence. Note that this conclusion is drawn from a special case
that the light travels in a particular direction. In general, for any direction of propagation vector ~k, one can show
that there can be two different values for the phase velocities for an anisotropic crystal. Before proceeding to do
that, let us fix some notations first. The relation between the electrical permittivity and the susceptibility is given
by,
0 (1 + χ) = 
CHAPTER 4. PROPAGATION OF LIGHT IN CRYSTALS xxvi

One can relate the susceptibility to the index of refraction as,

 µ c2
1+χ= ≈ = 2 = n2 (4.13)
0 0 µ0 v
Note that we have made a reasonable assumption that µ ≈ µ0 . Now, we identify the following relations.
p p
n1 = 1 + χ11 = K11
p p
n1 = 1 + χ22 = K22
p p
n1 = 1 + χ33 = K33

Here, n1 , n2 , and n3 are called principal indices of refraction. To obtain the possible values for the wave vector (and
hence the phase velocity) one should solve Eq. (??) for arbitrary ~k. One can express Eq. (??) in a matrix form,
and the nontrivial solutions are obtained by solving the following determinant equation.
n1 ω 2


− ky2 − kz2 k k kx kz
c
x y2
n2 ω 2 2
ky kx − k − k k k =0 (4.14)
c x z

y z2
n3 ω 2 2
kz kx kz ky c − kx − ky

One solves these equations to obtain possible values of k. We can have a pictorial representation of the solutions of
the above equations as a 3 dimensional surface in k space. To do so, we solve Eq. (??) on xy, yz, and xz planes in
the k space. On the xy plane, we have kz = 0. Therefore, Eq. (??) becomes,
n1 ω 2


c − ky2 kx ky 0
n2 ω 2


ky kx c − kx2 0 =0 (4.15)
n3 ω 2

2 2
0 0 c − kx − ky

Finding the determinant and equating to zero, we get,

h  n ω 2 ih  n ω 2  
n 2 ω 2
 i
3 2 2 1 2
− kx − ky − ky 2
− kx − kx2 ky2 = 0 (4.16)
c c c
As the product of two terms is zero, either one of the terms or both terms must be zero. When the first factor is
zero, we obtain,
 n ω 2
3
kx2 + ky2 =
c
This represents a circle with radius n3 ω/c. When the second factor is zero, we obtain an equation of ellipse given
as,
kx2 ky2
2 +
=1
n2 ω n1 ω 2


c c

In the same way, one can solve Eq. (??) on the xz plane by putting ky = 0. The corresponding solution surface on
xz plane is again a circle and an ellipse. Equation of the circle is,
 n ω 2
2
kx2 + kz2 =
c
And the equation for the ellipse is,
kx2 kz2

+
n3 ω 2 n1 ω 2

=1
c c

Finally, solving Eq. (??) on yz plane, we obtain equations for a circle and an ellipse. Equation for circle is,
 n ω 2
1
ky2 + kz2 =
c
And the equation for the ellipse is,
ky2 kz2

+
n3 ω 2 n2 ω 2

=1
c c
 172 OPTICS OF SOLIDS

CHAPTER 4. PROPAGATION
NowOF LIGHT(6.92),
in Equation IN CRYSTALS
in order for a nontrivial solution for E, xxvii
EY1

and E to exist, the determinant of the coefficients must vanish,

namely,
Now, we combine these solutions and plot the wave-vector surface, see Fig. ??. Note that the complete solution to
Eq. (??) will be two interconnected
(n1w/c)surfaces
2 _ ky2 - k (sheets)
2' inkxky
3 dimensional k kxkz space. It is straightforward that any vector
drawn from the origin would touch both k ykx these surfaces
(n20l c) 2 (corresponding
-- kx2 -- kz' to
kyk,two values for 0 k vector) except along the
direction in which these surfaces arekzkxjoined. The two k,values ky
for the
(n3WIkc)vector '
2 - kx2 indicates
1;
two phase velocities. The
(6.97)
axis in which these sheets overlap is called the optic axis of the crystal along
where we have used Equation (6.96). The above equation can be which both phase velocities become
equal. represented by a three-dimensional surface in k space. The form of
this k surface, or wave-vector surface, is shown in Figure 6.8. To see

Figure 6.8. The wave-vector surface.

Figure 4.2: The wave vector surface describing the solutions of wave equation on k space
how the surface is constructed, consider any one of the coordinate
planes,
Earlier we have shown that for say the xypropagating
a wave plane. In this plane
alongkzx= direction,
0, and the determinant
two phase velocities corresponds to
reduces to the product of the two factors
two mutually orthogonal directions of polarization. Now, we conclude that the same is true for any direction of
[(nw)2 kx 2 -light
propagation. Therefore, when unpolarized [()2
ky2 propagates
nx [()2
ky2 through---akx2crystal,
-- kx2ky2it= C}
will split into two independent
C
waves with mutually orthogonal polarization and different phase velocities. (6.98)
Since the product must vanish, either or both of the factors must be

4.2.1 Notes on Birefringence

1. The direction along which the inner and outer sheets meet corresponds to the optic axis of the crystal. Along
the optic axes, there will be a single value for ~k and the double refraction will be absent along the optic axis.

2. The theory of optical activity developed in the previous section considers distinct indices of refraction, i.e.,
n1 , n2 , and n3 are different. In such crystals, there will be two optic axes, and are called biaxial crystals Fig.
??.

3. In many crystals, two out of three indices of refraction are equal. In such cases there will be a single optic
axis and these are called uniaxial crystals. The ~k surface of such crystals consist of a sphere and an ellipsoid
of revolution.

4. If all the indices of refraction are equal, then the ~k degenerates to a single sphere, and the crystal will not
be birefringent. In other words, if all the three indices of refraction are equal the crystal will be optically
isotropic.
CHAPTER 4. PROPAGATION OF LIGHT IN CRYSTALS xxviii

174 OPTICS OF SOLIDS

Optic axis Optic axis

174 OPTICS OF SOLIDS

Optic axis Optic axis

(a)

(a)

Figure 4.3: Intercepts of the wave-vector surfaces in the xy plane for a biaxial crystal
174 OPTICS OF SOLIDS

Optic axis Optic axis

5. In uniaxial crystals the index of refraction corresponds to the two equal components of the susceptibility,
i.e., χ11 = χ22 , is called the ordinary index no . The remaining index, which corresponds to χ33 is called
extraordinary index nE .

6. If n0 < nE , the crystal is called uniaxial positive crystal. On the other hand, if n0 > nE it is called a negative
(a)
crystal Fig. ??.

(c)

(c)
Figure 6.9. Intercepts of the wave-vector surfaces in the xz plane for (a)
biaxial crystals; (b) uniaxial positive crystals; (c) uniaxial nega-
tive crystals.
Figure 6.9. Intercepts of the wave-vector surfaces in the xz plane for (a)
biaxial crystals;
Figure 4.4: Uniaxial positive (b) uniaxial
and negative positive
crystals crystals; (c) uniaxial nega-
respectively
tive crystals.

4.3 Optical Activity

(c)

4.4 Faraday Rotation

Figure 6.9. Intercepts of the wave-vector surfaces in the xz plane for (a)
4.5 Problemsbiaxial crystals; (b) uniaxial positive crystals; (c) uniaxial nega-
tive crystals.
Bibliography

[1] G. R. Fowles, “Introduction to Modern Optics,” Holt, Rinehart and Winston, 1968,

xxix