Indian Institute of Technology, Kanpur

Department of Physics
PHY-224A: OPTICS
Lecture -14: Interference of Light
Date: 22 Sept. 2014, Time :1700 hours Instructor: S.A. Ramakrishna
1 Interference of two optical elds
The basic rule of electromangetic elds is the elds add and the corresponding intensities do
not. The intensity at a point corresponding to a propagating electromagnetic wave is given by
I = v
ph
E
2
= c

E
2

where the indicates time averaging over several time periods for the electric eld of the
radiation. For a vectorial wave (which em waves are), and using a phasor notation where the
time-harmonic elds that we talk of are given by real parts of the complex phasors

E(r, t) =
Re[

E(r) exp(it)], we have then

E
2
=
1
4
[

E
2
e
2it
+

E
2
e
2it
+ 2

E

E

].
Hence, we have

E
2
=
1
2

E

E

=
1
2
[E
2
x
+ E
2
y
+ E
2
z
].
Now suppose two monochromatic waves with the same frequency,

E
1
and

E
2
, are superposed
at the same point in space, P, then

E =

E
1
+

E
2
, (1)

E
2
=

E
2
1
+

E
2
2
+ 2

E
1

E
2
.
The intensity is given by
I

E
2
=

E
2
1
+

E
2
2
+ 2

E
1

E
2

= I
1
+ I
2
+ J
12
, (2)
where
J
12
2

E
1

E
2
=
1
2
(

E
1

2
+

E

1


E
2
) = (E
x1
E
x2
cos
x
+ E
y1
E
y2
cos
y
+ E
z1
E
z2
cos
z
),
1
where
x,y,z
are the phase dierence between the respective components of the two waves (Exam-
ple,
x
=
x2

x1
). We will assume that this phase dierence is maintained over the timescales
relevant for our measurements.
Let us analyse the simplest cases. If
x
=
y
=
z
= (say), and for transverse waves
polarized along the x-axis (say), we have J
12
E
x1
E
x2
cos , I
1
1/2 E
2
x1
and I
2
1/2E
2
x2
.
Hence,
I = I
1
+ I
2
+ 2

I
1
I
2
cos (3)
The intensities due to the two waves do not add. There are interference terms which make the
intensity at P very dierent from the sum of the intensities. The maximum possible intensity is
I
max
= (

I
1
+

I
2
)
2
for = 2m and the minimum possible intensity is I
min
= (

I
1

I
2
)
2
for = 2(m + 1), where m is an integer. Interference is a property unique to waves.
2 Interference of two beams: Wavefront division
We will discuss now the classic Youngs double slit interference. Consider a wave-front of a plane
wave that is incident on two slits. We will assume perfect coherence of the wave over the wave
front here. Coherence here refers to the idea that a denite relationship of the phase between
any two points on the wavefront exists, and does not change over the timescales relevant for the
experiment. Now consider a perfectly opaque screen containing two long slits placed apart by
a distance of 2d. At this stage, it will suce for us to declare that the size of the slits are very
small and that they almost act as point sources the size of the slits are assumed small with
respect to the wavelength. Now let us examine what intensity of light will be measured at a
point P, on a screen that is placed beyond the two slits. The plane of the screen is taken to be
the x-y plane.
Now the path lengths between the slits and the point, P, can be written as
s
2
1
= (x d)
2
+ y
2
+ z
2
0
, s
2
2
= (x + d)
2
+ y
2
+ z
2
0
,
and
s
2
2
s
2
1
= 2x(2d) s
2
s
1
=
2x (2d)
s
2
+ s
1

2x (2d)
2z
0
,
where it is assumed that z
0
2d: i.e, the screen is sucient far from the slits as well as the
distances on the screen (x, y) are small compared to z
0
. Thus, we obtain the dierence of the
two paths as
s =
x (2d)
z
0
(4)
and the corresponding optical phase dierence is
=
nkx (2d)
z
0
=
2nx (2d)
z
0
(5)
2
Figure 1: A schematic diagram of the Youngss double slit experiment and the associated path
lengths.
At the point P, the eld will be a maximum or minimum depending on the position x. It will
be a maximum if
2nx (2d)
z
0
= 0, 2, 4,
or x = m
z
0
n (2d)
(6)
Similarly, it will be a minimum if
x = (m + 1/2)
z
0
n (2d)
These intensity maxima and minima are called interference fringes. Note that there will be no
dependence of the intensity on the y-coordinates. We call this interference by wave-front division
because we take light from two dierent places on a wavefront and interfere them.
Now we will assume the two slits are of identical nature, transmit the same amounts of
the incident elds with the same phase - eectively they both act as point sources with the
same phase. The assumption of a plane wave incident is useful for this purpose. In that case,
the electric eld contributions due to slit-1 and slit-2 at the point P on the screen beyond
the slits will be of equal amplitude and will only have a phase dierence that depends on the
path dierence. We will assume that the electric elds involved are polarized only along the
x-direction justied due to the consideration z
0
2d. In that case, the intensity maxima will
be 4 times the intensity due to one slit and the intensity at the minima will be zero.
3
A serious point is that interference does not violate any conservation law, such as for energy.
Interference only redistributes the energy in the optical elds and does nothing more. One can
prove this (problem -1 of assignment) by integrating the energy available in the interference pat-
tern over the entire screen and checking if this is the sum of the energies contributed individually
by the two slits.
A second important point concerns the phase of the wave emitted by the two slits, which we
assumed to be equal in the above arguments. What if one of the slit (say, slit-1) transmits the
wave with an extra phase shift of
1
compared to the other. You are invited to prove that it will
merely lead to a spatial shift of all the fringes along the x-axis (problem-2 of assignment) . This
same eect would be seen if a plane wavewere to be incident on the two slits at a small oblique
angle with respect to the normal of the screen with the slits (problem -3 for assignment). I do
not suggest a very large angle as that would complicate the scattering from the two slits which
would confuse you at the moment.
A third question that arises is that what would happen if the sources were point sources
instead of slits? How would the interference pattern then appear would it continue to be
straight lines as in the earlier case? Note the y-invariance. Try to work out the pattern by
similar arguments as presented here (problem 4 of the assignment).
In much of the discussions from now onwards, we will assume that the interfering electro-
magnetic waves have a simple spatial structure and the same polarization. Then we can work
with scalars for the elds which considerably simplies the notation.
Read about alternative ways of accomplishing two-beam interferences by wavefront divi-
sion such as a Fresnels mirror, Lloyds mirror and the fresnels bi-prism (problem -5 of the
assignment).
4