Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

4. POLARISATION.
Plane polarised light
Earlier, we considered one of the most useful solutions to the electromagnetic wave
equation, the plane wave, travelling in the z direction, say. That is a transverse wave
whose constant phase fronts are planes perpendicular to z, ie. the xy plane. We write the
plane wave in one of two forms;

   z  
E(z, t ) = A 0 cos 2π − νt  = A 0 cos[kz − ωt ] (4.1a)
 λ 

or equivalently

 
E(z, t ) = A 0 exp j[kz − ωt ] (4.1b)

We noted one important property of a wave, its phase and in the above case the phase is
simply the argument of the cosinusoid or exponential,

φ = kz − ωt

and this phase does not depend on x or y once z has been fixed, ie. the phase is the same
at any value of x and y for a given value of z and t.

The plane wave is usually represented with a diagram such as that shown below where
horizontally either position (time constant) or time (position constant) is plotted and
vertically the magnitude of the electric field is plotted.

λ or ν-1

z or t
or φ

90
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

So far we have ignored one very important property of the transverse plane wave. In

equations 4.1 the amplitude of the wave is correctly written as a vector, A 0 . ie. the

electric field possesses not only magnitude but direction. The vector E will be lying
somewhere in the xy plane for a wave propagating in the z direction ie. it has a
polarisation. Furthermore, in general that direction may alter with time or position. In
general a light source such as the sun or an incandescent light bulb continuously emits
separate uncorrelated wave trains (in quantum transitions) that are independent of one
another and the light produced is overall unpolarised.
NB. For the light to acquire an overall polarisation the involvement of some
polarising element or process is required.
eg. It has already been noted in discussions around the Fresnel equations that the
reflectivity or transmissivity depends on the polarisation of the light and two types of
plane polarisation were introduced in that discussion in order to arrive at the equations,
that is the Transverse Electric or TE polarisation and the Transverse Magnetic or TM
polarisation which are mutually orthogonal and defined with respect to an interface
(plane of incidence) between two dielectrics. It was noted that in an extreme case, at the
Brewster angle the reflectivity of the TM polarisation went to zero leaving only the TE
polarisation in the reflected wave. A dielectric orientated at the Brewster angle is an
example of a polarising element. This is also the reason why light reflected from the
surface of water at certain angles is polarised.

In some special circumstances the polarisation state can be specified. We shall list these
beginning with the simplest case.

1. Plane Polarised electromagnetic waves where the polarisation

direction is uniquely specified and is independent of time and position.
For a wave travelling in a direction k the plane of polarisation is that defined by k and
E.
Recalling from earlier electromagnetism courses the superposition principle which
states that if electric fields from two or more sources are present in the same region of
space at the same time there is a net field that is found by the simple addition of the
fields as vectors.

91
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

We have several simple situations of electrostatic fields to consider.

2EX
⇒ EX
+ (EX , EY )
EX EX ⇒
+
EY
The examples shown are of static fields where the two fields may or may not be in the
same direction. This is simple vector addition but needs to be re-emphasised before
continuing to look at further aspects of polarisation and of examining the superposition
principle in general and how it determines the effects of interference and diffraction.
Of more interest is to do the same vector additions with electric fields that vary in space
or in time as a plane wave.

EX

z or t

TWO ORTHOGONAL
INPUTS
δ=0
EY

z or t

y y

Ey E y Ey E y
x x
Ex Ex
Ex Ex
x x

E Ey E Ey

RESULTANT
92
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

The first example to analyse is the case of two electromagnetic waves of the same
amplitude, frequency and wavelength propagating in the same direction, say z and
propagating in phase. One of the waves is plane polarised in the x direction and the
other plane polarised in the y direction. The situation is depicted in the above diagrams
along with the resultant field worked out at the peaks and troughs. We find that the
resultant is again a plane polarised wave whose polarisation direction is oscillating at
+450 to the x and y axes. The amplitude of the resultant is

E = E x2 + E y2 = 2 E x (4.2)

If the amplitude of the two fields had been different then the wave would have remained
Ey
plane polarised but at an angle θ = tan −1 to the x axis.
Ex

EX

z or t

TWO ORTHOGONAL
INPUTS
δ=π

EY

z or t

y
y y

Ex Ex
x y x y
Ey E Ey E
E Ey E Ey

x x
Ex Ex

RESULTANT
93
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

Another example, depicted above, similar to the first is that of two electromagnetic
waves of the same amplitude, frequency and wavelength but of mutually orthogonal

plane polarisation propagating in the same direction but with a phase lag δ = π = 180 0
with respect to one another. The resultant field is shown again at the peaks and troughs.
This is another plane polarised wave with its polarisation vector at -450 to the x axis.
We can also see this mathematically;

In the first case with zero phase shift


E x = E0 x x cos(kz − ωt )

E y = E0 y y cos(kz − ωt ) (4.3)
 
E = ( x + y )E0 cos(kz − ωt )

NB. In Hecht the notation φ is used for the phase lag, ie. φ ≡ δ, here I use φ for the
phase generally.
NB. In these 3 equations E , Ex and Ey are the instantaneous electric fields and E 0 ,

E 0 x and E 0 y are the amplitudes of those time varying fields. It is important to make the

distinction in what follows.

 
This is a field where the vector is ( x + y ) pointing in the direction at 450 to the x (or y )
2π
axis oscillating with a frequency ω and a wavelength λ = . Note that the vector
k
(x + y ) pointing in a direction +450 to the x axis is no longer a unit vector of magnitude
1 but has magnitude 12 + 12 = 2 and therefore the amplitude of the resultant field in
4.3 is 2E 0 where E0 = E0 x = E0 y as stated.

In the second case with phase shift δ = π = 1800 between x and y electric field
components


E x = E0 x x cos(kz − ωt )
  
E y = E0 y y cos(kz − ωt + δ) = E0 y y cos(kz − ωt + π) = − E0 y cos(kz − ωt )) (4.4)

94
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

 
E = (x − y )E0 cos(kz − ωt )

 
In this case the field vector is ( x − y ) and points in the direction at -450 to the x axis.

It is not inevitably the case that the two orthogonal plane polarised and co-
propagating electric fields of the same frequency with different angles of polarisation
will form a third plane polarised wave. We next look at other examples of resultant
fields, that are the result of superposition of two orthogonal plane polarised waves,
which are not themselves plane polarised.

2. Circular Polarised Light.

EX

z or t

TWO ORTHOGONAL
INPUTS
δ = π/2
EY

z or t

y
y
y
y

x Ey x
Ex Ex
x x

Ey

z or
95
t
⇒
RESULTANT
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

π
In the above example the y plane polarised wave has a phase lag of δ = − = −90 0 over
2
the x plane polarised wave but the same amplitude and frequency/wavelength.
Following the resultant field at each of the peaks and troughs leads to a direction of
polarisation that rotates. This polarisation just described is known as right hand circular
polarisation and is so named as it behaves like a right handed screw, looking back
towards the source an observer sees the E field circulating clockwise. If, on the other
π
hand, the y component had been leading the x component by 900 , δ = + , the
2
polarisation would have rotated in the counter clockwise sense looking back towards the
source, this being left hand circularly polarised light.

We can describe this in mathematical terms by writing the two orthogonal fields as

E x = E 0 x xˆ cos(kz − ωt )
(4.5a)
E y = E0 y yˆ cos(kz − ωt + δ )

π
Where in the case we have just discussed, E0 x = E0 y = E0 and δ = −
2
In this case
 π
E y = E0 y yˆ cos kz − ωt −  = E0 yˆ sin (kz − ωt ) (4.5b)
 2

The sum of the two fields is then

  
E = E0 {x cos(kz − ωt ) + y sin (kz − ωt )} (4.5c)

We can recognise the meaning of 4.5c by considering the vector

 
A{x cos φ + y sin φ } (4.7)

96
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

which can be represented in the xy plane as a vector of length A at an angle φ to the x

axis as shown below.

y
Asinφ A

φ
Acosφ x

With φ = kz − ωt we see by comparing 4.5c with 4.7 that 4.5c the resultant field
represents a field with constant amplitude and a direction that rotates in a circular
motion at a frequency ω.

If the y plane polarised component leads the x component by 900 the resultant field is

  
E = E0 {x cos(kz − ωt ) − y sin (kz − ωt )} (4.5d)

Which is the left hand circularly polarised wave.

Right hand circular polarisation Left hand circular polarisation

The wave is emerging from the paper

97
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

If we take a superposition of a right hand and a left hand circularly polarised wave we
have

    
E = E0 {x cos(kz − ωt ) + y sin (kz − ωt )} + E0 {x cos(kz − ωt ) − y sin (kz − ωt )}

 
E = 2 E 0 x cos(kz − ωt ) (4.8)

ie. we have a plane polarised wave once again..

3. Elliptically Polarised Light (the general case).

Ey

Ex

Of course the phase difference, δ, can take any value and the electric field amplitudes
of the two orthogonal components may be different. In this more general case the
resultant field will change amplitude and direction with time (or equivalently position)
as it propagates. If the electric field is observed at a fixed position the field vector traces
an ellipse with time that may or may not be aligned with the x and y axes. Equivalently
if the electric field is measured as a function of z at a fixed time the field vector traces
an ellipse with position (in the direction of propagation.
In other words the two orthogonal fields we may write in their most general form

E x = E 0 x xˆ cos(kz − ωt ) (4.9a)

E y = E0 y yˆ cos(kz − ωt + δ ) (4.9b)

98
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

NB. The values of amplitude, E0 x and E0 y and the phase δ are three constants

describing the polarisation state and we may relate the Ex and the Ey instantaneous fields
to one another at any instant in time or point in space in a straightforward manner as
follows;

The curve that the resultant E vector traces should not depend on either position or time
so we begin by eliminating the z and t dependences in the field descriptions as follows;
Using the trigonometric identity cos( A + B) = cos A cos B − sin A sin B

We expand the Ey field

E y = E0 y {cos(kz − ωt ) cos δ − sin( kz − ωt ) sin δ } (4.10)

The dimensionless y component is then;

Ey
= {cos(kz − ωt ) cos δ − sin( kz − ωt ) sin δ} (4.10a)
E0 y

Also the dimensionless x component is;

Ex
= cos(kz − ωt ) (4.11)
E0 x

Using 4.11 in the RHS of 4.10 and re-arranging

Ey Ex
− cos δ = − sin( kz − ωt ) sin δ (4.12)
E0 y Eox

Using 4.11 and the trigonometric identity sin 2 A + cos 2 A = 1 in 4.11

1
  E 
2 2
sin( kz − ωt ) = 1 −  x   (4.13)
  E0 x  


99
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

and using 4.13 in 4.12

2  2
 Ey E   E 
 − x
cos δ  = 1 −  x   sin 2 δ (4.14)
 E0 y E0 x    E0 x  
   

Or by re-arranging terms

2 2
 Ey   E y  E x
  +  E x 
 − 2  
 cos δ = sin 2 δ (4.15)
 E0 y  E  
   0x   0 y  E0 x
E 

The above equation does not involve z or t and is the equation of an ellipse that makes
an angle α with the (x,y) co-ordinate system with;
2 E0 x E0 y
tan 2α = cos δ (4.16)
E02x − E02y

i) We may show it describes an ellipse

We can more easily see that 4.15 is the equation of an ellipse if the phase shift δ = 900
and cosδ = 0. This removes the cross product term and sin2δ = 1

2 2
 Ey 
  +  E x 
 = 1 (4.17
 E0 y  E
   0x 

equation 4.17 is easily recognisable as an ellipse with axes, E0 x , E0 y

ii) We may show that for δ = π/2 and equal amplitudes that it describes a
circle
If the amplitudes of the two orthogonal plane polarised waves were equal,
E0 x = E0 y = E0 then 4.17 becomes;

E x2 + E y2 = E02 (4.18)

100
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

This is the equation of a circle as previously encountered in the description of the

circularly polarised light where δ = 900 and E0 x = E0 y = E0 .

iii) We may show that for δ = 0 and equal amplitudes that it describes linear
plane polarisation

The above analysis means that if δ = 0 in 4.15 we should recover a linear polarised
wave. In this case cos 0 = 1 and sin 0 = 0 and inserting these values into 4.15

2 2
 Ey   Ey  E x
  +  E x 
 − 2  
 = 0 (4.19)
 E0 y  E  E0 y  E0 x
   0x    

Or rewriting as a square

2
 Ey E 
 − x  =0 (4.20)
 E0 y E0 x 
 

Ey Ex Ey E0 y
= or = (4.21)
E0 y E0 x Ex E0 x

The ratio of y component to x component is a constant independent of t or z and 4.21

thus represents a linear plane polarisation with the angle of the electric field, θ with
respect to the x axis given by

Eoy
tan θ = (4.22)
E0 x

Thus the elliptical polarisation state is the most general polarisation state with plane
polarised and circularly polarised light being special examples of elliptically polarised
light.

101
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

Birefringence
In the earlier parts of the course we have considered the propagation of light in simple
media after a consideration of propagation in vacuum and saw how the description
needed modification. The modification required the introduction of optical properties
for the medium namely; χE , ε and n, electronic susceptibility, dielectric constant and
refractive index respectively. We chose simple media, that is isotropic and
homogeneous media and we found relationships amongst these material properties;

 
P = ε0χ E E
   
D = εε 0 E = ε 0 E + P (4.23)

ε = 1+ χE

n = εµ = ε = 1 + χ

All of the above relationships apply for simple media but many materials are
anisotropic, particularly crystalline materials where there is some ordering amongst the
constituent atoms and the possibility arises that there is some anisotropy in the material
properties, ie. their value depends on the orientation of the electric field associated with
the light wave.
To be able to observe the effects of polarisation as described to their fullest extent we
need the electromagnetic field to be propagating in these more complex media where
there may be for example anisotropy in the optical properties that were not encountered
earlier.
So far, in all that has preceded, it has been assumed that the refractive index of a medium
is one scalar constant that is independent of the direction of polarisation of the
electromagnetic wave whose velocity it modifies. In actuality n will be dependent on the
direction of polarisation in most classes of solid except those that are amorphous or
possess cubic symmetry. This property of solids is known as birefringence. A simple
way of thinking about birefringence is to appreciate that the polarisation P induced by an
electric field E (and consequently D = (ε0E + P ) will not necessarily be in the same
direction as E. That is to say that we need to modify our view of the electronic
susceptibility, a measure of the ease with which a material is polarised (ie charge

102
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

seperated). Previously, in the above equations, 4.23, it is considered a scalar constant

and the polarisation per unit volume P is in the same direction as the electric field E.
More realistically we must treat χE (and consequently, ε and n) as a tensor reflecting the
varying ease of polarisability (charge separation) in different directions in a material.
This is most clearly demonstrated by an example where the effect is large.
Polydiacetylene is a conjugated polymer which may be obtained as macroscopic single
crystals where the polymer chains are all aligned. The chain is covalently bonded (as is
Si) but the individual chains are held together by weak van der Waals forces.
R
R Covalent
R
R C
Covalent C C
C C C
C C C C
C C C
C C RVan der Waals
C R R
R R
RVan der Waals R
R C
C C
C C C
C C C C
C C C
C C R
C R
R
R
Charge is relatively easily seperated by an electric field applied along the polymer chain
direction but far less readily seperated by a field perpendicular to the chain as electrons
are strongly restricted to the chain they find themselves on and cannot move from one
chain to another due to the interchain separation whereas they can readily move along
the covalently bonded polymer chain and it is the charge displacement that gives rise to
the polarisation. A little thought will show that a field, E. polarised at 450 to the chains
will have a large component of polarisation in the chain direction, due to the component
of E in that direction, and virtually zero polarisation perpendicular. P is therefore almost
entirely along the chain whilst E is at 450 to the chain. The situation is depicted
schematically below.

103
 Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

Electric Field

+δq Polymer chain

δp=δqdz dz Polarisation
-δq -Q +Q Polymer chain
∆x

P=Q∆x

The above gives a physical example and description of the origin of birefringence. There
is a small amount of charge displacement, ±δq from one chain to another and a much
larger charge displacement ±Q on the chain as it is here that the charge is mobile. This
results in a very small interchain polarisation δp and a large intrachain polarisation P.

More formally the relation between polarisation and applied field must be written in
tensor form :
 χ xx χ xy χ xz  E x 
  
(Px Py )
Pz = ε 0  χ yx χ yy χ yz  E y  (4.24)
 
 χ zx χ zy χ zz  E z 

(
Px = ε 0 χ xx E x + χ xy E y + χ xz E z )
Py = ε 0 (χ yx E x + χ yy E y + χ yz E z ) (4.24a)

Pz = ε 0 (χ zx E x + χ zy E y + χ zz E z )

where the χij are the components of the susceptibility tensor.

104
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

There exists a set of axes in any crystal called "the principle dielectric axes", x, y, z for
which all but the diagonal components, χii , are zero.

 χ xx 0 0  E x 
  
(Px Py )
Pz = ε 0  0 χ yy 0  E y 
 0 χ zz  E z 
 0

(4.24b)
However for all but amorphous materials such as glasses and certain of those with cubic
symmetry, the three remaining components χxx , χyy , and χzz are not equal. This
implies, as we have seen that P, E, and D are not in the same direction. It also has
implications for propagation of an electromagnetic wave in such a medium.
We noted earlier that there is a relationship between χΕ and n,

n = 1+ χE (4.25)

In other words, for these anisotropic media the refractive index must also be represented
by a tensor. If we have chosen the cartesian axes, x, y, z to be in the same direction as the
principle dielectric axes there will be three principle values of refractive, nx , ny , and nz
all of which may be different. Such behaviour is called birefringence and the implication
is that light propagating within a birefringent medium will propagate at a velocity that
depends on the direction of polarisation. For example plane polarised light, polarised in
c
the x direction will have a phase velocity v = etc.
nx
Recalling from earlier discussion that any polarisation state (circular, elliptical, plane)
can be described as the sum of two orthogonal plane polarised states, we can now see the
significance of that statement! We now discuss birefringence in more detail and see that
it is necessary to consider any light wave to be split into orthogonal plane polarised
components in order to describe its propagation in a birefringent medium.
In general birefringent materials may be subdivided into two classes as follows.

(i) Uniaxial materials where only one index, nz say, is different and nx = ny
In this case nx = ny = no the ordinary refractive index and nz = ne the
extraordinary refractive index.

105
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

Such materials may be further subdivided into;

positive uniaxial materials where no < ne , and the ordinary ray propagates at a
higher velocity as a result
and
negative uniaxial materials where no > ne and it is the extraordinary ray that
propagates at the higher velocity.
Two classic uniaxial birefringent materials are the minerals quartz where no = 1.5443
and ne = 1.5534 with a birefringence value ∆n = ne − no = +0.0910 and calcite which is
a negative uniaxial crystal with no = 1.6584 and ne = 1.4864 and its birefringence
∆n = ne − no = −0.1720 .

The z direction in the above example ( n x = n y = no and n z = ne ≠ no ) is called the

optic axis and is important because any wave travelling along the optic axis can only
have field components which see an identical refractive index, nX = nY . It will then
travel through the medium unchanged as far as its polarisation state is concerned.

(ii) Biaxial materials where nX ≠ nY ≠ nZ . These will have two optic axes
which are not the principle dielectric axes. They are a more difficult case to
consider and are less frequent.

We will be satisfied to limit ourselves to the analysis of uniaxial crystals in what follows.

106
 Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

Uniaxial birefringence.
Uniaxial materials have two defining refractive indices, nx = ny and nz known as the
ordinary refractive index, no and the extraordinary refractive index, ne respectively.Any
wave not propagating in the direction of the optic axis, ie. the z direction in this example,
will split into two mutually orthogonal plane polarised waves with differing phase
velocities known as the ordinary and the extraordinary waves. For any arbitrary
propagation direction a construction called the index ellipsoid may be used in order to
discover what happens. This is an ellipsoid constructed such that its semi-major/minor
axes have lengths nx , ny and nz and they are drawn below for positive and negative
uniaxial materials.
The electric field of a wave travelling in an arbitrary direction may for simplicity of
analysis be considered as split into two plane polarised components; an ordinary wave
polarised perpendicular to the optic axis, z, and an extraordinary wave polarised
orthogonal to the ordinary wave. In a positive material the ordinary wave moves at a
faster phase velocity than the extraordinary wave and vice versa for a negative material.

nZ

nZ
nX nY nY
nX
z
y
Index ellipsoid for x Index ellipsoid for
positive uniaxial material negative uniaxial material

107
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

nZ nZ
θ
OA ne

nY nY
nX nX n0

Propagation along the Propagation off axis

optic axis

Consider first propagation along the optic axis of a positive uniaxial crystal as
represented in the diagram. The large arrowed vector, OA, in the diagram indicates the
direction of propagation and the polarisation components see a refractive index
discovered by constructing a plane normal to that direction (the x-y plane) containing the
origin. This makes a circular section through the ellipsoid whose radius is defined by n0 .
Because that section of the ellipsoid is a circle any two orthogonal components will see a
refractive index n x = n y = no and will travel with no change in phase. Such a situation,

ie. propagation along the optic axis, gives rise to no unusual effects. Ie. our two
orthogonal plane polarised components (that form any polarisation state) will retain any
phase relationship they had possessed before entering the medium and the polarisation
state will be left unaltered. Now consider the off axis propagation at an angle θ to the
optic axis as represented in the second diagram. The bold arrowed vector indicates the
direction of propagation. To use the ellipsoid, a plane normal to this direction containing
the origin is constructed. It intersects the ellipsoid to form an ellipse whose minor axis is
n0 and whose major axis is ne(θ), a function of θ. These two refractive indices will be
the refractive indices of the ordinary and extraordinary wave into which the original

108
 Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

wave will split. The two orthogonal plane polarised components that describe the initial
polarisation state will become those resolved along and perpendicular to the axes of the
ellipse. In other words a point source radiating into the medium will split into an
ordinary wave whose velocity is independent of direction and therefore spreads out as a
circular wavefront, and an extraordinary wave whose velocity depends on θ as shown in
the diagrams below. The ordinary and extraordinary wavefronts touch in the direction of
the optic axis as shown. NB the O ray is polarised perpendicular to the optic axis.

OA

OA

Propagation from a point source into a Propagation from a point source into a
positive uniaxial crystal. The optic positive uniaxial crystal. The optic
axis, OA, is perpendicular to the axis, OA, is parallel to the surface.
surface.

109
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

EX
EX Birefringent
medium
nx = ne
ny = n o
∆n = nx - ny EY
EY L
λ0
= 2π/k0
x Optic Axis
y δ = ∆nk0L
z = Direction of propagation

Uniaxial birefringent materials may be useful in optical instruments as a means of

altering the polarisation state of the light. One example of this will suffice to
demonstrate the potential usefulness.
The above diagram shows two orthogonal plane polarised electromagnetic waves
incident on a birefringent medium from the left. The optic axis is in the x direction and
therefore n x = ne and therefore the orthogonal polarisation n y = no . Before entering

the medium the two components are in phase and the resultant field will also be plane
polarised at an angle of 450 to the x and y axes as found in an earlier analysis. Once
entered within the crystal however they each propagate at a different velocity and when
emerging from the crystal after a path length L they will no longer be in phase. To
understand this it is important to be clear about what the phase is. Recall that for a plane
wave described by

E = E0 cos(kz − ωt ) (4.26)

The phase, φ , is

φ = (kz − ωt )
(4.27)

110
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

In free space, light travels at a velocity c and there is a relation between λ0 and ν

λ 0ν = c (4.28a)

Once within the medium this velocity is altered as described by the refractive index and
the relationship is now

c λ 0ν
λν = = (4.28b)
n n

The frequency on the LHS has not changed but the velocity on the RHS has become
smaller. In fact the wavelength, λ , is no longer the same. Writing the wavelength in
free space as λ0 once the plane wave is propagating in the medium the wavelength
becomes
λ
λ= 0 (4.28c)
n
2π
Recalling the definition of the magnitude of the wavevector, k = , the wavevector in
λ
the medium is also altered when compared to the wavevector in free space, k0 .

k = nk 0 (4.29)

In the birefringent medium we have different wavevectors for the x polarised wave and
the y polarised wave and also different phases as a result

k x = nx k0 k y = n y k0 (4.30)

The phases are then

φ x = n x k 0 z − ωt φ y = n y k 0 z − ωt (4.31)

The phase difference is then

111
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

( )
δ = φ x − φ y = n x − n y k 0 z = ∆nk 0 z (4.32)

Thus on emerging from the crystal of length L the phase difference is

δ = ∆nk 0 L (4.33)

As we have seen in an earlier section the light wave will in general be elliptically
polarised when the two orthogonal plane polarised waves have a phase difference δ as
described by 4.15.
There are interesting possibilities as we noted earlier. If the two orthogonal polarisations
have a phase difference of δ = (2m + 1)π where m = 0, ±1, ±2…. the plane of
polarisation of the resultant will be rotated by 900 compared to where they are in phase.
They are in phase at the input in the above example and therefore if
δ = (2m + 1)π = ∆nk 0 L the plane of polarisation will be rotated through 900. For a given

wavelength (or wavevector) and a material with a given birefringence, ∆n, the length of
the crystal can be adjusted to satisfy the condition, ie.

(2m + 1)π (2m + 1)λ 0

Lλ = = (4.34)
2 ∆nk 0 2∆n

Such an arrangement is known as a half wave plate. It is important to note that this will
only be a half wave plate for certain wavelengths that satisfy 4.34 and that the
birefringent medium is able to transmit.
Altering the thickness of the wave plate may introduce a phase difference

δ = (2m + 1) π where m = 0, ±1, ±2 and this will cause a plane polarised input to
2
become circularly polarised (and vice verca) when L is such that

(2m + 1) π
Lλ = 2 = (2m + 1)λ (4.35)
4 ∆nk 0 4∆n

This is called a quarter wave plate. Again it is only a quarter wave plate for the set of
wavelengths which satisfy 4.35 and for which the medium is transparent.

112
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

Transmission of polarised light through polariser

It is useful to be able to calculate the amount of light in a given polarisation state that
will be transmitted by a second polarising element (the analyser) set to transmit plane
polarised light polarised in a particular sense wrt the plane of polarisation of the input
light.

Birefringent
L plate

E0 y
x
ET

Polariser, P1 z Analyser, P2 Detector, D

We approach this problem by considering the particular arrangement shown above. This
arrangement consists of ;

i) A polarising element P1 that prepares light in a state of plane polarisation

with the plane of polarisation in a particular direction.

ii) A plate of birefringent material of thickness L will alter the polarisation state
into a new polarisation state, after the light has traversed the plate, by
introducing a phase lag, δ, between the two orthogonal plane polarised waves
that constitute the original wave.

iii) An analyser, P2 , is a second polarising element that will only pass light
polarised in a plane orthogonal to that of the original polariser P1.

We wish to know the fraction of the light intensity incident on the polariser P1 that is
detected after it has traversed the whole system including P2.

113
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

With no birefringent plate present the plane polarised light from P1 will be unchanged as
far as its polarisaton state is concerned and hence no light will be transmitted by P2
which is set orthogonal to the polariser, P1. The introduction of the birefringent plate will
introduce a phase difference between the two orthogonal components that make up the
original plane polarised wave, these two components being the ordinary and
extraordinary waves plane polarised in the y and x directions respectively.

y x
EI EIy EIx
EIy EIx
ET
Before transmission through cell After transmission through P2

The above diagrams are invaluable in enabling us to visualise the components of the
input and output electric fields, EI and ET in the x and y directions. With this pictorial
representation the problem is straightforward to solve.
Examination of the reference frame as shown above shows that to find ET after passage
through the analyser requires addition of +ETx and -ETy . (in contrast to the EI which
requires +EIx and +EIy). Thus
At the entry into the plate the plane waves for each component are

EI
E Ix = cos(k 0 z − ωt )
2

EI
E Iy = cos(k 0 z − ωt )
2

As the input wave is plane polarised with an amplitude EI at 450 to both x and y axes
EI
there are equal amplitudes for x and y components, .
2

If we arbitrarily choose z = 0 at the input to the birefringent crystal.

Upon exiting the analyser, P2, the fields of each component are given by

114
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

E Ix EI
ETx = cos(ne k 0 L − ωt ) = cos(ne k 0 L − ωt )
2 2
E Iy E
ETy = − cos(no k 0 L − ωt ) = − I cos(no k 0 L − ωt )
2 2
Where each component has suffered a phase change φ x = ne k 0 L and φ y = n0 k 0 L . The

resultant transmitted field is thus

ET =
EI
[cos(ne k 0 L − ωt ) − cos(no k 0 L − ωt )]
2
NB. The x axis has been chosen to be the optic axis and light polarised in the x
direction “sees” the extraordinary refractive index.
Ie. after traversing the plate there is a phase difference given by

2π
δ = φ x − φ y = ∆nk 0 L = (ne − no ) L
λ0
Using the trigonometric identity

 A+ B  A− B
cos A − cos B = 2 sin   sin  
 2   2 

EI n + no ωt n − no
ET = 2 sin( e k 0 L − ) sin( e k 0 L)
2 2 2 2

Which we can rewrite more compactly using our expression for δ as

n + no ωt δ
ET = E I sin( e k 0 L − ) sin
2 2 2

We find the output intensity IT by finding the time average of the square of the electric
field divided by impedance of free space, η0 giving;

ET2 ET2 E2  n + no
δ ωt  δ
IT = = = I sin 2   sin 2  e k 0 L −  = I i sin 2  
η0 2η 0 η 0 2  2 2  2

115
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

NB. only one of the sinusoids (with the ωt in its argument) in the expression for ET has

any time dependence and has time averaged to 1 for the square of a sinusoid over
2
many cycles as usual leaving the simple final expression. and the transmission of the
system is then

I δ
T = 0 = sin 2  
Ii 2

Dichroism.
a) Linear Dichroism
After the recent discussion of anisotropic media and how this leads to birefringence; the
possession of two refractive indices that describe propagation in the anisotropic medium
for orthogonal plane polarised fields, it is possible to extend the description of
anisotropy to describe materials where one plane polarisation state may be more
strongly absorbed than its orthogonal partner. This is the origin of dichroism and the
way that Polaroid plastic works. In the preceding discussion of birefringence we were
interested in the light propagating in the medium where refraction is the dominant effect
of the anisotropic medium away from frequencies that are strongly absorbed by the
medium. However another interesting phenomenon of light propagating in a medium
(solid, liquid or gas) is that at certain frequencies the light may be absorbed and the
intensity fall as the light propagates through the medium.
Generally, in a simple medium, as light propagates through it at frequencies that may be
absorbed the intensity lost (amount of light energy absorbed) is proportional to the
distance travelled, and the intensity available to be lost, ie. After travelling an
infinitesimal distance dz the drop in intensity is proportional to dz and to the amount of
intensity originally present and is thus given by

δI = −αIdz

where α is the constant of proportionality. We may re-arrange this equation to give

dI
= −αdz
I

116
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

Integrating from z = 0 to L

I ( z) L
log e I I = −α z 0
0

Evaluating, taking exponentials and rearranging

I = I 0 exp(−αL)

This is known as Beers Law and α is the absorption coefficient.

It is simple now to see that for anisotropic media the absorption coefficient will depend
on the direction in which the electric field of the light wave is pointing, ie. on the plane
of polarisation. The act of absorption, while a quantum event, can be thought of as the
electrons in the medium moving in response to the electric field. They will find it easier
to move in one direction rather than another, eg. reverting to our earlier example, along
a polymer chain rather than perpendicular to it. This leads to preferential absorption for
different plane polarisation states and the necessity to identify different absorption
coefficients for different polarisations. This is known as linear Dichroism and is the
basis for the most common polariser, the polaroid sheet, in which the material has been
subject to stretching forces allowing the polymer chains that compose the plastic to
align in a preferred direction ie the direction of the stress. Light that is plane polarised in
this direction will be strongly absorbed whilst light polarised orthogonal to this
alignment direction will be weakly absorbed , we thus define α// and α⊥ light plane
polarised parallel to the direction of aligmment and perpendicular to the direction of
alignment and α// >> α⊥.

b) Circular Dichroism
It can also be the case that the left and right hand circularly polarised waves are
absorbed to a different extent providing the molecule/substance doing the absorbing is
able to exist in right handed and left handed forms known as enantiomers. Such
molecules are inevitably possessing of chirality and their structures allow a left handed
and a right handed form of the molecule to be identified, This is common among
biological molecules such as sugars and amino acids (and the proteins formed of amino

117
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

acids). If a plane polarised wave enters a sugar solution, provided it is a pure enantiomer
and made up of only one “handed” sugar, the two oppositely rotating circular
polarisations that we have already seen can be thought of as composing the linearly
polarised beam will be absorbed to differing extents and the right hand polarised beam
being the more weakly absorbed will come to dominate the light wave changing from a
plane polarised state to an elliptically polarised state. The polarisation is said to be
rotated clockwise and the solution of sugar is dextrorotary (from the Latin for right) and
if rotated counter clockwise the solution is laevo-rotatry. Hence the other name for
glucose, dextrose). We have seen how this occurs where the amplitudes of the two plane
polarisations are different in for example 4.15.

118
Electromagnetic Waves & Optics: Lecture Notes ©Kevin Donovan

In Summary the most important things we need to know

about polarisation states are;
i) There are three polarisation states of importance
a) Plane polarised light where the plane of polarisation is independent of
position and time i.e. the electric field vector always points in the same
direction as the wave propagates
b) Circularly polarised light, where the plane of polarisation rotates about
the axis of propagation at an angular frequency, ω, the same as the
frequency of the light wave but the amplitude is fixed independent of time
and position. We may speak of left and right circularly polarised light.
c) Elliptically polarised light is the most general polarisation state and the
others are limiting forms of the elliptically polarised state where both
amplitude and direction may vary with time and position as the wave
propagates.
ii) All polarisation states may be described as composed of two orthogonal
plane polarised waves.
iii) One can change between polarisation states by changing the phase
relationship between the orthogonal plane polarised states.
iv) Birefringent crystals/materials will naturally separate the two orthogonal
polarisations of an arbitrary polarisation states into an ordinary and an
extraordinary plane polarised wave each of which travels with a different
velocity.
v) The ordinary refractive index, n0 , of a birefringent crystal is independent
of the angle at which the wave propagates through the crystal wrt the
optic axis whereas the extraordinary refractive index will depend on the
angle ne(θ)

119