Field Guide to

Polarization

Edward Collett

SPIE Field Guides

Volume FG05

John E. Greivenkamp, Series Editor

Bellingham, Washington USA

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Library of Congress Cataloging-in-Publication Data

Collett, Edward, 1934-

Field guide to polarized light / Edward Collett.
p. cm. – (SPIE field guides ; v. FG05)
Includes bibliographical references and index.
ISBN 0-8194-5868-6
1. Polarization (Light) I. Title. II. Series.
QC441.C63 2005
535.5'2--dc22
2005006346

Published by

SPIE—The International Society for Optical Engineering

P.O. Box 10
Bellingham, Washington 98227-0010 USA
Phone: +1 360 676 3290
Fax: +1 360 647 1445
Email: [email protected]
Web: http://spie.org

Copyright © 2005 The Society of Photo-Optical Instrumentation

Engineers

All rights reserved. No part of this publication may be

reproduced or distributed in any form or by any means without
written permission of the publisher.

The content of this book reflects the work and thought of the
author. Every effort has been made to publish reliable and
accurate information herein, but the publisher is not responsible
for the validity of the information or for any outcomes resulting
from reliance thereon.

Printed in the United States of America.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Introduction to the Series

Welcome to the SPIE Field Guides! This volume is one of

the first in a new series of publications written directly for
the practicing engineer or scientist. Many textbooks and
professional reference books cover optical principles and
techniques in depth. The aim of the SPIE Field Guides is
to distill this information, providing readers with a handy
desk or briefcase reference that provides basic, essential
information about optical principles, techniques, or
phenomena, including definitions and descriptions, key
equations, illustrations, application examples, design
considerations, and additional resources. A significant
effort will be made to provide a consistent notation and
style between volumes in the series.

Each SPIE Field Guide addresses a major field of optical

science and technology. The concept of these Field Guides
is a format-intensive presentation based on figures and
equations supplemented by concise explanations. In most
cases, this modular approach places a single topic on a
page, and provides full coverage of that topic on that page.
Highlights, insights and rules of thumb are displayed in
sidebars to the main text. The appendices at the end of
each Field Guide provide additional information such as
related material outside the main scope of the volume, key
mathematical relationships and alternative methods.
While complete in their coverage, the concise presentation
may not be appropriate for those new to the field.

The SPIE Field Guides are intended to be living

documents. The modular page-based presentation format
allows them to be easily updated and expanded. We are
interested in your suggestions for new Field Guide topics
as well as what material should be added to an individual
volume to make these Field Guides more useful to you.
Please contact us at [email protected].

John E. Greivenkamp, Series Editor

Optical Sciences Center
The University of Arizona

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Field Guide to Polarized Light

The polarization of light is one of the most remarkable

phenomena in nature and has led to numerous discoveries
and applications. Today it continues to play a vital role in
optics. Before the 1950s there was very little activity on
the foundations of polarized light. For example, answers
to questions such as the nature and mathematical
formulation of unpolarized light and partially polarized
light were not readily forthcoming. Today there is a very
good understanding of polarized light. In particular, the
mathematical difficulties that had hindered complex
polarization calculations were finally overcome with the
introduction of the Mueller-Stokes matrix calculus and
the Jones matrix calculus. Research in polarized light
continues with much vigor as witnessed by the continued
appearance of numerous publications and conferences.

The primary objective of this Guide is to provide an

introduction to the developments in polarized light that
have taken place over the past half-century. In this Guide
I have tried to present the most salient topics on the
subject. Hopefully, this Field Guide will enable the reader
to have a good grasp of the material and most of all to
allow him or her to be comfortable and even delighted
with the beauty and subject of polarized light.

Finally, this Field Guide is dedicated to my wife, Mary

Ann, and my children Ron and Greg. Their
encouragement and support greatly simplified the task of
writing this Guide.

Edward Collett
Georgian Court University
Lakewood, New Jersey

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Table of Contents

Glossary x

The Foundations of Polarized Light 1

The Ray Theory of Light 1
The Polarization of Light 2
Malus’s Law 3
Brewster’s Law 4

The Wave Theory of Light 5

Fresnel’s Wave Theory 5
The Polarization Ellipse 7
Degenerate Polarization States 8
The Parameters of the Polarization Ellipse 9
The Poincaré Sphere 10
Degenerate States on the Poincaré Sphere 11

The Observables of Polarized Light 12

The Stokes Polarization Parameters 12
Stokes Parameter Relations 14
Classical Measurement of the Stokes Parameters 16
The Mueller Matrices for Polarizing Components 17
Polarizers 18
Wave Plates 20
Rotators 22
Mueller Matrices for Rotated Components 23
Mueller Matrix Applications–Malus’s Law 25
Mueller Matrix Applications–The Optical Shutter 26
Mueller Matrix Applications–Stokes Parameters 27

The Observable Polarization Sphere 28

The Observable Polarization Sphere 28
Plotting the Quarter-Wave Plate on the OPS 32
The Rotating Quarter-Wave Plate 34
The Babinet-Soleil Compensator 35
Linear and Circular Polarizers 36
The Generation of Elliptically Polarized Light 37
Measurement Methods of the Stokes Parameters 38
The Rotating Quarter-Wave Plate Measurement 39

vii

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Table of Contents (cont’d)

Birefringent Crystals and Wave Plates 40

Multiple and Zero-Order Wave Plates 41

Reflection and Transmission 42

Mueller Matrices for Reflection and Transmission 42
Reflection and Transmission Stokes Parameters 43
Reflection and Transmission Mueller Matrices 47
Total Internal Reflection 48
The Fresnel Rhomb 49
Single and Multiple Dielectric Plates 50
Pile of Polarizing Dielectric Plates 52
Fresnel’s Reflection and Transmission Coefficients 55

Other Polarization Matrix Calculi 57

The Jones Matrix Calculus 57
Wolf’s Coherency Matrix Calculus 62

Optical Activity and Optical Rotation 63

Optical Activity and Optical Rotation 63
Faraday Rotation 64
Optical Isolators 66

Depolarizers 72
Wave Plate Depolarizers 72
The Lyot Crystal Depolarizer 74

Polarizing Materials 75
Polarizers 75
Polarizing Prisms 76
Characterizing Polarizers 78
Wave Plate Materials 81

Superposition and Decomposition of Polarized

Beams 82
Incoherent Superposition and Decomposition 82
Incoherent Decomposition–Ellipses 83
Coherent Superposition and Decomposition 84

viii

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Table of Contents (cont’d)

The Electro-Optical Effect 85

The Electro-Optical Effect - Modulators 85
The Pockels Cell 87

Refractive Index Measurements 88

Incidence Refractive Index Measurement 88

The Radiation Field 91

Maxwell’s Equations 91
The Radiation Equation and the Stokes Parameters 92
The Linear Oscillating Bound Charge 93
The Randomly Oscillating Bound Charge 94
A Charge Moving in a Circle 95
A Charge Moving in a Magnetic Field 96
The Classical Zeeman Effect 98
Optical Scattering 101

The Optics of Metals and Semiconductors 105

The Optics of Metals and Semiconductors 105
Refractive Index and Absorption Coefficient 106
Incidence Angle Reflectivity 107
Complex Reflection Coefficients 109
The Principal Angle of Incidence Measurement 110

Appendix
Equation Summary 114
Notes 124

Bibliography 128
Index 130

ix

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Glossary

Frequently used variables and symbols:

Β birefringence
B(r,t) magnetic induction vector
c speed of light in a vacuum
cp circularly polarized
db decibels
D(r,t) electric displacement vector
e- extraordinary ray
ε permittivity constant
E0x maximum amplitude in the x direction
E0y maximum amplitude in the y direction
Ex(r,t) x component of the optical field
Ey(r,t) y component of the optical field
E Jones vector
E(r,t) electric field vector
F force vector
H0 Transmission of two parallel polarizers
H90 Transmission of two crossed polarizers
H(r,t) magnetic field vector
HWP half wave plate
i angle of incidence
iB Brewster angle
i,j,k Cartesian unit vectors
J Jones matrix
j(r,t) electric current density vector,
JPOL Jones matrix for a polarizer
JWP Jones matrix for a wave plate
JROT Jones matrix for a rotator
JQWP Jones matrix for a quarter-wave plate
JHWP Jones matrix for a half-wave plate

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Glossary (cont’d)

J(θ) Jones matrix for a rotated polarizing

element
k wave number
k wave vector
k1 major transmittance of a polarizer
k2 minor transmittance of a polarizer
KDP potassium dihydrogen phosphate
L–45P linear –45 polarization
L+45P linear +45 polarization
LCP Left circular polarization
LHP linear horizontal polarization
LVP linear vertical polarization
μ permeability constant
M Mueller matrix
MHWP Mueller matrix of a half-wave plate
MLP Mueller matrix of a linear polarizer
MPOL Mueller matrix of a polarizer
MQWP Mueller matrix of a quarter-wave plate
MR Mueller matrix for reflection
MROT Mueller matrix of a rotator
MT Mueller matrix for transmission
MWP Mueller matrix of a wave plate
M(θ) Mueller matrix of a rotated polarizing
element
n complex refractive index
ne refractive index of the extraordinary ray
no refractive index of the ordinary ray
np parallel refractive index
ns perpendicular refractive index
nL levo-rotary refractive index

xi

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Glossary (cont’d)

nR dextro-rotary refractive index

n refractive index
o- ordinary ray
OPS observable polarization sphere
p- parallel polarization state
px polarizer transmission coefficient (x)
py polarizer transmission coefficient (y)
P degree of polarization
PBS polarizing beam splitter
QWP quarter wave plate
r angle of refraction
r radius vector
RCP right circular polarization
s- perpendicular polarization state
S Stokes vector
S0 first Stokes parameter
S1 second Stokes parameter
S2 third Stokes parameter
S3 fourth Stokes parameter
SR Stokes vector for reflection
ST Stokes vector for transmission
TIR total internal reflection
UNP unpolarized
v(r,t) velocity vector
vx,vy.vz principal velocities
V Verdet’s constant
Vπ half-wave voltage
Vm maximum modulation voltage
WP wave plate
x,y,z Cartesian coordinate system

xii

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Glossary (cont’d)

2α coordinate on the observable polarization

sphere
2ψ coordinate angle on the Poincaré sphere
2χ coordinate angle on the Poincaré sphere
α auxiliary angle
ε complex dielectric constant
εx, εy, εz principal dielectric constants
δ coordinate on the observable polarization
sphere
δ phase difference
δx phase of the wave (x)
δy phase of the wave (y)
θ angle of rotation
κ absorption coefficient
ψ orientation angle
χ ellipticity angle
ρs,p Fresnel reflection coefficients
ρ(r,t) electric charge density
σ conductivity
τs,p Fresnel transmission coefficients
φ phase shift
φx phase shift (x)
φy phase shift (y)
ω angular frequency
ωc cyclotron frequency
ωL Larmor’s frequency
ωm modulation frequency
∇ spatial vector operator

xiii

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Foundations of Polarized Light 1
The Ray Theory of Light

Polarized light has its origins in the ray theory of light.

In the 11th century Al-Hazen examined a tower through a
very small hole in a darkened room. On the back wall the
tower appeared as an inverted image. This
demonstrated, along with the law of reflection later
defined by Snell, that
A B′ light proceeds from
SUN
B A‫׳‬ points A and B to A′
Tower Image and B′, respectively, in
Darkened Room the form of rays.

The ray theory appears when light rays are reflected from
a concave spherical surface as shown below.
Here C is the center
of the circle, R is the 1
f (θ) = 1 −
radius of the 2 cos θ
spherical surface, θ
θ
and f(θ) is the focal C R f(θ)
point for a circle of
1
f (0°) = = 0.5000
unit radius. 2
The focal point
moves from f(0°) = 0.5000 to f(60°) = 0.0000 as the in-
cident ray moves away from the axis of symmetry.
Plotting the ray paths for f(θ) from θ = 0° to 60° yields the
figure.
θ° f(θ)
0 0.5000
10 0.4925
Incident
20 0.4680
rays
30 0.4227
40 0.3475
50 0.2220
60 0.0000

The reflected light rays do not come to a single focus and

the rays describe the locus of a cardioid, which is also
called a caustic. The image is readily observed in a
drinking cup.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 2 Polarization

The Polarization of Light

In 1670 Bartholinus discovered that when a single ray of

natural incident light propagated through a
rhombohedral calcite crystal, two rays emerged,
demonstrating that a single ray of light actually consists
of two rays called the ordinary or o-ray (no), and the
extraordinary or e-ray (ne). Because the two rays
refract at different angles, the calcite crystal is said to be
doubly refractive or birefringent. Both rays obey Snell’s
law of refraction but experience different refractive
indices.

Reflected
ray

109° 71°
Incident e
ray 71° 109° o

Further investigation by Huygens, showed that by

rotating a second (analyzing) crystal the intensity of one
ray was maximized and the other ray vanished. A further
rotation of 90° then showed that the first ray reappeared
with a maximum intensity and the second ray vanished.
At a rotation angle of 45° the intensities of the two rays
were equal. Because of this opposite behavior of intensity,
the two rays were said to be polarized. Thus, a single ray
of natural light actually consists of two independent
oppositely polarized rays. The two rays are said to
represent the s- and p-polarization states.

Double refraction occurs in calcite because it is an

anisotropic crystal and the o-ray wave front
propagates as a sphere, whereas the e-ray propagates
as an ellipsoid. Furthermore, the refractive indices of
the o- and e-rays at λ = 5893 (sodium D line) are
no = 1.6853 and ne = 1.4864.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Foundations of Polarized Light 3
Malus’s Law

At the beginning of the nineteenth century the only

known way to generate polarized light was with a calcite
crystal. In 1808, using a calcite crystal, Malus discovered
that natural incident light became polarized when it was
reflected by a glass surface, and that the light reflected
close to an angle of incidence of 57° could be extinguished
when viewed through the crystal. He then proposed that
natural light consisted of the s- and p-polarizations, which
were perpendicular to each other.

Incident light Reflected light

p

s • i i •s
Calcite
crystal

Glass Plate

Since the intensity of the reflected light varied from a

maximum to a minimum as the crystal was rotated,
Malus proposed that the amplitude of the reflected beam
must be A = A0cosθ. However, in order to obtain the
intensity, Malus squared the amplitude relation so that
the intensity equation I(θ) of the reflected polarized light
was

I ( θ) = I 0 cos2 θ,

where I0 = A02; this equation is known as Malus’s Law. A

normalized plot of Malus’s Law is shown below.
Intensity

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 4 Polarization

Brewster’s Law
Around 1812 Brewster discovered that for different
glasses the p-polarized ray (in the plane of the paper)
vanished completely at a particular angle of incidence i.
Furthermore, by rotating the analyzing calcite crystal
through 90°, the s-polarized ray (directed out of the plane
of the paper) became extinguished. He then discovered
that the refracted ray angle r was simply related to the
incident ray angle i by
i + r = 90°.

Incident light Reflected light

p

s • i i •s
Calcite
crystal

r Refracted
Glass plate
ray

Snell’s law of refraction between the two media is

n1 sin i = n2 sin r,

where n1 and n2 are the refractive indices of the media,

respectively. For air n1 = 1 and for glass n2 = n.
Substituting the first equation into Snell’s law leads to
tan i = n2 / n1 = n.

This equation is known as Brewster’s law. Its

immediate practical use was that it enabled the refractive
index of glass to be determined by reflection rather than
by refraction; the measurement of the refractive angle r
by transmission is difficult. The result of this discovery
was that it led to the rapid development and
measurement of new optical glasses.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Wave Theory of Light 5

Fresnel’s Wave Theory

Around 1820 Fresnel proposed a theory of light now

known as Fresnel’s wave theory that completely
explained the three major phenomena of light:
interference, diffraction, and polarization. Furthermore,
Fresnel and Arago experimentally showed that the
optical field consisted of only two orthogonal
components in the plane transverse to the direction of
propagation.
Fresnel’s wave theory postulated that the orthogonal
components were u1(r,t) and u2(r,t), which he called
optical disturbances. We now know, however, that the
optical disturbances can be represented by the electric
field components of the electromagnetic field. Fresnel
hypothesized that the field components are described by
two equations known as the wave equations,

1 ∂ 2 E x ( r, t )
∇ 2 E x ( r, t ) = ,
v2 ∂t 2
2
1 ∂ E y ( r, t )
∇ 2 E y ( r, t ) = ,
v2 ∂t 2

where Ex(r,t) and Ey(r,t) are the optical-field components,

r is the radius vector to a point in space measured from
the origin of a coordinate system, t is the time, v is the
velocity of the waves, and ∇ 2 is the Laplacian operator.
The two components of the field and the direction k form
an orthogonal system as shown.

Ey(r,t) r

Ex(r,t)

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 6 Polarization

Fresnel’s Wave Theory (cont’d)

The solutions of the wave equations are

Ex (r,t ) = E0 x cos( ωt − k ⋅ r + δx )

and
E y (r,t ) = E0 y cos( ωt − k ⋅ r + δ y ) ,

where k is the (vector) wave number and describes the

direction of the propagation, and r is a point in the field.
In practice, the field is taken to be directed along the z-
axis. The two components, above, can then be written as
Ex ( z ,t ) = E0 x cos( ωt − kz + δ x )

and
E y ( z ,t ) = E0 y cos( ωt − kz + δ y ) ,

where ω = 2πf is the angular frequency, k = 2π/λ is the

wave number magnitude, E0x and E0y are the maximum
amplitudes and δx and δy are arbitrary phases; the term
ωt – kz is called the propagator.
Ey The propagation
of these two waves
(equations) can be
graphically repre-
Ex z sented as shown
in the figure.
The Ey component
is in the plane of the paper (also called the plane of
incidence) and is called the p-polarization component.
Similarly, the Ex component is perpendicular to the plane
of the paper and is called the s-polarization component.

The p and s notation come from the German

words for parallel (paralelle) and perpendicular
(senkrecht).

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Wave Theory of Light 7

The Polarization Ellipse

According to Fresnel’s theory, Ex(z,t) and Ey(z,t) describe

sinusoidal oscillations in the x-z and y-z planes,
respectively (see the figure on p. 6) By themselves, these
equations are not particularly revealing. However,
eliminating the time-space propagator ωt – kz between
the two equations leads to the equation of an ellipse,
namely,
2
Ex ( z ,t )2 E y ( z ,t ) 2E x ( z , t ) E y ( z , t )
2
+ 2
− cos δ = sin 2 δ ,
E0 x E0 y E0 x E0 y

where δ = δy – δx. The above equation describes an ellipse

in its nonstandard form. Because the equation refers to
polarized light, the equation is called the polarization
ellipse. In the equation, the time-space propagator has
been explicitly eliminated. Nevertheless, the field
components Ex(z,t) and Ey(z,t) continue to be time-space
dependent. A plot of the nonstandard polarization ellipse
is shown below.
The figure also shows the rotated ξ-η coordinate system.
Because of the amplitudes E0x and E0y and the phase δ are
constant, the polarization ellipse remains fixed as the
polarized beam propagates.

y
η
ξ

2E0y ψ x
O

2E0x

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 8 Polarization

Degenerate Polarization States

In general, the optical field is elliptically polarized, but

there are several combinations of amplitude and phase
that are especially important. These are called
degenerate polarization states: (1) linearly
horizontal/vertical polarized light (LHP/LVP), (2) linear
±45° polarized light (L+45P/L–45P), and (3) right/left
circularly polarized light (RCP/LCP). The polarization
states along with the mathematical conditions and
corresponding figures (polarization ellipses) are as
follows.

LHP: E0 y = 0 LVP: E0 x = 0

L+45P: L-45P:
E0 x = E0 y = E0 , δ = 0 E0 x = E0 y = E0 , δ = π

RCP: LCP:
E0 x = E0 y = E0 , δ = π/2 E0 x = E0 y = E0 , δ = −π/2

RCP light rotates clockwise and LCP rotates counter-

clockwise when propagating toward the observer.
These polarization states are important because (1) they
are relatively easy to create in a laboratory using linear
and circular polarizers, and (2) polarization
measurements as well as many polarization calculations
are greatly simplified using these specific polarization
states. This is especially true when a polarized beam
propagates through numerous polarizing elements.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Wave Theory of Light 9

The Parameters of the Polarization Ellipse

The polarization ellipse can be expressed in terms of two

angular parameters: the orientation angle
ψ(0≤ψ≤π) and the ellipticity angle χ( –π /4 <χ≤π /4 ) .
These angles can be defined in terms of the parameters
of the polarization ellipse:
2E0 x E0 y
tan 2ψ = cos δ, 0 ≤ ψ ≤ π,
E02x − E02y

2E0 x E0 y sin δ
sin 2χ = , −π / 4 < χ ≤ π / 4.
E02x + E02y

The right-hand side of both of these equations consists of

algebraic and trigonometric terms. The two equations can
be rewritten completely in trigonometric terms by
introducing an angle known as the auxiliary angle α
defined by
E0 y
tan α = , 0 ≤ α ≤ π /2.
E0 x

This leads to purely trigonometric equations

tan 2ψ = (tan 2α ) cos δ,
sin 2χ = (sin 2α )sin δ.

The conditions on the angles are 0 ≤α≤π /2 and 0 ≤δ<2π .

Example. We determine the orientation and the

ellipticity angles ψ and χ for RCP light. We have for
RCP light that E 0 y =E 0 x =E 0 and δ = π / 2. Then, tan
α yields α = 45° and
tan 2ψ = tan 90° cos90° = 0

sin 2χ = sin 90° sin 90° = +1

Thus, the angles for RCP light are ψ=0° and

χ =+ 45°.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 10 Polarization

The Poincaré Sphere

By itself, the polarization ellipse is an excellent way to

visualize polarized light. However, except for the
degenerate polarization states, it is practically impossible
to determine the orientation and ellipticity angles viewing
the polarization ellipse. Furthermore, the calculations
required to determine the new angles of a polarized beam
that propagates through one or more polarizing elements
are difficult and tedious.
In order to overcome these difficulties Poincaré (1892)
suggested using a sphere now known as the Poincaré
sphere to represent polarized light. The following figure
shows the Poincaré sphere and its spherical and
Cartesian coordinates.
z

O 2χ
y
2ψ

Here x, y, and z are Cartesian coordinate axes, ψ and χ

are the spherical orientation and ellipticity angles (of the
polarization ellipse), and P is a point on the surface of the
sphere. Note that on the sphere the angles are expressed
as 2ψ and 2χ. For a unit sphere the Cartesian coordinates
are related to the spherical coordinates by the equation.
x = cos(2χ ) cos(2ψ ), 0 ≤ ψ < π,

y = cos(2χ )sin(2ψ ) , −π/4 < χ ≤ π/4

z = sin(2χ ),
where x2+y2+z2=1 for a sphere of unit radius.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Wave Theory of Light 11

Degenerate States on the Poincaré Sphere

From the previous equations any polarization state can be

represented by the coordinate pair (2ψ,2χ). The
degenerate polarization states on the Poincaré sphere are
for LHP(0°,0°), for L+45P(+90°,0°), for LVP(180°,0°), for
L-45P(270°,0°), for
RCP
RCP(0°,+90°), and
for LCP(0°,–90°). LVP
The degenerate
states on the x, y,
and z axes are L-45P L+45
shown below.
All linear polar-
ization states lie LHP
on the equator and LCP
right and left
circular polarization states are at the north and south
poles, respectively. Elliptically polarized states are
represented everywhere else on the surface of the sphere.

The following figure shows polarization states plotted at

every intersection of the 7.5° latitude and 15° longitude
lines.

RCP

LHP
L-45P L+45P

LCP

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 12 Polarization

The Stokes Polarization Parameters

The most serious limitation to the Poincaré sphere and

the polarization ellipse are (1) the polarization ellipse is
an instantaneous representation of polarized light, and (2)
neither the rotation angle ψ nor the ellipticity angle χ is
directly measurable. In order to overcome these
limitations it is necessary to determine the measurables
of the polarized field. This can be done by taking a time
average of the polarization ellipse:
2
Ex ( z ,t )2 E y ( z ,t ) 2E x ( z , t )E y ( z , t )
2
+ 2
− cos δ = sin2 δ.
E0 x E0 y E0 x E0 y

The time average Ei ( z ,t )E j ( z ,t ) is defined by

T
1
Ei ( z ,t )E j ( z ,t ) = lim
T →∞ T ∫0 Ei ( z,t )E j ( z,t ) dt , i , j = x , y,

where T is total averaging time. Applying the time

average definition to the polarization ellipse then yields
the following equation:

S02 = S12 + S22 + S32 ,

where

S0 = E02x + E02y ,

S1 = E02x − E02y

S2 = 2E0 x E0 y cos δ,

S3 = 2E0 x E0 y sin δ, δ = δ y − δ x .

The quantities S0, S1, S2, and S3 are the observables of

the polarized field. They were introduced by Stokes (1852)
and are called the Stokes polarization parameters.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observables of Polarized Light 13

The Stokes Polarization Parameters (cont’d)

The first Stokes parameter S0 describes the total intensity

of the optical beam; the second parameter S1 describes the
preponderance of LHP light over LVP light; the third
parameter S2 describes the preponderance of L+45P light
over L-45P light and, finally, S3 describes the
preponderance of RCP light over LCP light.
The Stokes parameters can be expressed in complex
notation (in order to bypass formally the time
integration) by suppressing the propagator and writing
Ex (t ) = E0 x exp(i δx ),

E y (t ) = E0 y exp(i δ y ).

The Stokes parameters are then defined in complex

notation by the following equations.
S0 = Ex Ex * + E y E y* ,

S1 = Ex Ex * − E y E y * ,

S2 = Ex E y * + E y Ex * ,

(
S3 = i Ex E y * − E y Ex * , )
where i = −1 and * represents the complex conjugate.
It is convenient to arrange the Stokes parameters as a
column matrix, which is referred to as the Stokes vector
for elliptically polarized light:

⎛ S0 ⎞ ⎛ E0 x + E0 y ⎞
2 2

⎜ ⎟ ⎜ ⎟
⎜ S1 ⎟ ⎜ E02x − E02y ⎟
S= = .
⎜ S2 ⎟ ⎜ 2E0 x E0 y cos δ ⎟
⎜⎜ ⎟⎟ ⎜⎜ ⎟
⎟
⎝ S3 ⎠ ⎝ 2E0 x E0 y sin δ ⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 14 Polarization

Stokes Parameter Relations

The Stokes vectors for the degenerate polarization

states are readily found using the previous definitions
and equations:
⎛1 ⎞ ⎛1 ⎞ ⎛1 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
1 −1 0
SLHP = I0 ⎜ ⎟ , SLVP = I0 ⎜ ⎟ , SL +45P = I0 ⎜ ⎟ ,
⎜0⎟ ⎜0⎟ ⎜1 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝0⎠ ⎝0⎠ ⎝0⎠

⎛1 ⎞ ⎛1 ⎞ ⎛1 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
0 0 0
SL −45P = I 0 ⎜ ⎟ , SRCP = I 0 ⎜ ⎟ , SLCP = I0 ⎜ ⎟ ,
⎜ −1 ⎟ ⎜0⎟ ⎜0⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝0⎠ ⎝1 ⎠ ⎝ −1 ⎠

where I0 is the intensity and is very often normalized to

unity.
The Stokes parameters can be shown to be related to the
orientation and ellipticity angles, ψ and χ, associated with
the Poincaré sphere as follows:
S1 = S0 cos(2χ) cos(2ψ ),

S2 = S0 cos(2χ )sin(2ψ ),

S3 = S0 sin(2χ),
and

1 ⎛S ⎞
ψ= tan −1 ⎜ 2 ⎟ , 0 ≤ ψ ≤ π,
2 ⎝ S1 ⎠

1 ⎛S ⎞ π π
χ= sin −1 ⎜ 3 ⎟, − ≤χ≤ .
2 ⎝ S0 ⎠ 4 4

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observables of Polarized Light 15

Stokes Parameter Relations (cont’d)

The Stokes parameters describe not only completely

polarized light but unpolarized and partially
polarized light as well. The Stokes vector for
unpolarized light is
⎛1 ⎞
⎜ ⎟
0
Sunp = S0 ⎜ ⎟ ,
⎜0⎟
⎜ ⎟
⎝0⎠
where S0 is the first Stokes parameter (total intensity).
Since there are no amplitude or phase relations between
the orthogonal components, S1, S2, and S3 are 0. Partially
polarized light is a mixture of completely polarized light
and unpolarized light, and is represented by
⎛ S0 ⎞ ⎛ S0 ⎞ ⎛ S0 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ S1 ⎟ 0⎟ S
S= = (1 − P ) ⎜ + P ⎜ 1 ⎟, 0 ≤ P ≤ 1,
⎜ S2 ⎟ ⎜0⎟ ⎜ S2 ⎟
⎜⎜ S ⎟⎟ ⎜ ⎟ ⎜⎜ S ⎟⎟
⎝ 3⎠ ⎝0⎠ ⎝ 3⎠
where P is called the degree of polarization (DOP).
For completely polarized light, P = 1, and the above
equation reduces to the Stokes vector for elliptically
polarized light. Similarly, for unpolarized light, P = 0, and
the above equation reduces to the Stokes vector for
unpolarized light.
The DOP P is defined by the equation
I pol S12 + S22 + S32
P= = , 0 ≤ P ≤ 1,
I tot S0
where Itot is the total intensity. These results show that
the relation between the Stokes parameters must be
broadened to
S02 ≥ S12 + S22 + S32 ,
where the = and > sign indicate completely and
unpolarized/partially polarized light, respectively.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 16 Polarization

Classical Measurement of the Stokes Parameters

The four Stokes parameters of a polarized beam can

be measured by passing a beam sequentially through two
polarizing elements known as a wave plate and a
polarizer. The emerging beam is then incident on an
optical detector. The measurement configuration is shown
as below.
Y
Y

θ
Optical
source X
X

Wave plate Detector

Polarizer
In the measurement the wave plate introduces a phase
shift φ between the orthogonal components of the incident
optical beam. The polarizer then transmits the resultant
field along its transmission axis at an angle θ and the
intensity I(θ, φ) on the detector is then found to be
1
I ( θ, φ) =
S0 + S1 cos 2θ + S2 sin 2θ cos φ − S3 sin 2θ sin φ.
2
The Stokes parameters in the above equation are the
parameters of the incident beam. By rotating the
polarizer to θ = 0, π/4, and π/2 without the quarter
waveplate and then inserting the waveplate in the final
measurement (a total of four measurements) the Stokes
polarization parameters of the incident beam are found
from the above equation to be
S0 = I (0,0) + I ( π/2,0),

S1 = I (0,0) − I ( π/2,0),

S2 = 2 I ( π/4,0) − S0 ,

and
S3 = S0 − 2 I ( π/4, π/2).

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observables of Polarized Light 17

The Mueller Matrices for Polarizing Components

In general, the polarization ellipse is in a nonstandard

form. The polarization ellipse (polarization state) can be
changed by changing the amplitude(s), the phase, or by
rotating the ellipse. Polarizing materials are available
to do this. A polarized beam with a given polarization
state propagates through one or more polarizing
elements, where the beam acquires a new polarization
state. This process is represented in the following figure.

Input Polarizing Output

beam element(s) beam
The input beam is characterized by a Stokes vector S and
the output beam by S′. The assumption is made that S
and S′ are linearly related by a 4 × 4 transformation
matrix known as the Mueller matrix, which represents
the polarizing element(s):
⎛S ′ ⎞
⎜ 0 ⎟ ⎛ m00 m01 m02 m03 ⎞ ⎛ S0 ⎞
⎜S ′ ⎟ ⎜m m11 m12
⎟⎜ ⎟
m13 ⎟ ⎜ S1 ⎟
⎜ 1 ⎟ = ⎜ 10 .
⎜ S ′ ⎟ ⎜ m20 m21 m22 m23 ⎟ ⎜ S2 ⎟
⎜ 2 ⎟ ⎜⎜ m m31 m32
⎟⎜ ⎟
m33 ⎠⎟ ⎝⎜ S3 ⎠⎟
⎜ S ′ ⎟ ⎝ 30
⎝ 3 ⎠
All the elements in the 4 × 4 Mueller matrix are real
quantities. The above matrix relation can be written as a
matrix equation,
S′ = M ⋅ S
Only two polarizing elements are needed to change the
three parameters of the ellipse (the orthogonal
amplitudes and phase). The amplitude can be changed by
using a polarizing element known as a polarizer.
Similarly, the phase of an optical beam can be changed by
a wave plate (also called a retarder or phase shifter).
Finally, the polarization ellipse can be changed by
rotation using a component called a rotator. Using these
three polarizing elements, any elliptical polarization state
can be obtained.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 18 Polarization

Polarizers

The polarizing element that changes the amplitude is a

linear polarizer, which is characterized by two
absorption coefficients that differ along the x- and y-axes,
respectively. The absorption coefficients in the amplitude
domain are defined by px and py and are
0 ≤ px ≤ 1, 0 ≤ py ≤ 1.

The value of 0 represents total absorption (no

transmission) and the value of 1 represents total
transmission (no absorption).
py

Incident
beam
z Output
px beam

The Mueller matrix for a linear polarizer (the

polarization matrix form used in the intensity domain) is
found to be
⎛ px 2 + py2 px 2 − py2 0 0 ⎞
⎜ 2 2 2 2
⎟
1 ⎜ px − py px + py 0 0 ⎟
M POL ( px , py ) = ⎜ ⎟.
2⎜ 0 0 2 px py 0 ⎟
⎜ 0 0 0 2 px py ⎟⎠
⎝
For an ideal linear polarizer there is complete
transmission along one axis and no transmission along
the orthogonal axis. The Mueller matrix for an ideal
linear polarizer with its transmission axis along the x-
axis is px = 1 and py = 0. Therefore,

⎛1 1 0 0⎞
⎜ ⎟
1 1 1 0 0⎟
M POL = ⎜ .
2 ⎜0 0 0 0⎟
⎜ ⎟
⎝0 0 0 0⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observables of Polarized Light 19

Polarizers (cont’d)

For an ideal linear polarizer in which the transmission

axis is along the y-axis, the Mueller matrix is
 1 −1 0 0
 
1 −1 1 0 0
M POL =  .
2 0 0 0 0
 
0 0 0 0

The following figure shows a pair of crossed polarizers

(The transmission axes are orthogonal to each other.)

Incident Output
beam beam
x y

MPOL(px) MPOL (py)

The Mueller matrix for the pair is M = MPOL(py)⋅ MPOL(px)

0 0 0 0
 
1 0 0 0 0
M = M POL ( px ) ⋅ M POL py ( ) =
4 0 0 0 0
.
 
0 0 0 0

This is a null matrix and shows that no light emerges

when the linear polarizers are crossed. Another
interesting application occurs for equal absorption along
both axes [a neutral density (ND) filter]. In this case,
px = py = p, therefore

S ′ 
 0  1 0 0 0   S0   S0 
S ′      
0 1 0 0   S1  S
 1  = p2  = p2  1  .
 S2′  0 0 1 0   S2   S2 
       
S ′  0 0 0 1   S3   S3 
 3 
The polarization state of the incident beam remains
unchanged, but the intensity is reduced by a factor of p2.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 20 Polarization

Wave Plates

Wave plates have the property that along the x-axis

(called the fast axis) the x component of the field
experiences a phase shift of +φ/2 and, similarly, along the
y-axis (called the slow axis)the y component experiences
a phase shift of –φ/2. The configuration for the wave plate
−φ / 2
is seen in the
figure.
x The Mueller
Incident Output matrix for a
beam beam wave plate is
then found to
+φ / 2 be

⎛1 0 0 0 ⎞
⎜ ⎟ φ = φx − φ y ,
0 1 0 0 ⎟
M WP ( φ) = ⎜ , φ ⎛ φ⎞
⎜0 0 cos φ − sin φ ⎟ =+ − ⎜ − ⎟,
⎜ ⎟ 2 ⎝ 2⎠
⎝0 0 sin φ cos φ ⎠

where φ is the total phase shift between the orthogonal

components of the beam. Two forms of the Mueller
matrices that are very important in polarization are the
quarter-wave plate (QWP) (φ = π/2) and the half-wave
plate (HWP) (φ = π):
⎛1 0 0
0⎞
⎜ ⎟
0 1 0 0⎟
M QWP ( φ = π/2) = ⎜ ,
⎜0 0 0 −1 ⎟
⎜ ⎟
⎝0 0 1 0⎠

⎛1 0 0⎞0
⎜ ⎟
0 1 0 0⎟
M HWP ( φ = π) = ⎜ .
⎜0 0 −1 0 ⎟
⎜ ⎟
⎝0 0 0 −1 ⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observables of Polarized Light 21

Wave Plates (cont’d)

The QWP has the interesting property that it transforms

L+45P light to RCP light. This is shown in the following
Stokes vector calculation:
⎛1 0 0 0 ⎞⎛ 1 ⎞ ⎛ 1 ⎞
⎜ ⎟⎜ ⎟ ⎜ ⎟
0 1 0 0 ⎟⎜ 0 ⎟ ⎜ 0 ⎟
SOUT =⎜ = .
⎜0 0 0 −1 ⎟⎜ 1 ⎟ ⎜ 0 ⎟
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎝0 0 1 0 ⎠⎝ 0 ⎠ ⎝ 1 ⎠

Similarly, RCP light is transformed to L-45P light:

⎛1 0 0 0 ⎞⎛ 1 ⎞ ⎛ 1 ⎞
⎜ ⎟⎜ ⎟ ⎜ ⎟
0 1 0 0 ⎟⎜ 0 ⎟ ⎜ 0 ⎟
SOUT =⎜ = .
⎜0 0 0 −1 ⎟⎜ 0 ⎟ ⎜ −1 ⎟
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎝0 0 1 0 ⎠⎝ 1 ⎠ ⎝ 0 ⎠

The HWP has the unique property in that it reverses the

ellipticity and orientation angles of the polarization
ellipse:
⎛1 0 0
0 ⎞⎛ 1 ⎞ ⎛ 1 ⎞
⎜ ⎟⎜ ⎟ ⎜ ⎟
0 1 0 0 ⎟ ⎜ cos 2χ cos 2ψ ⎟ ⎜ cos 2χ cos 2ψ ⎟
SOUT =⎜ = .
⎜0 0 −1 0 ⎟ ⎜ cos 2χ sin 2ψ ⎟ ⎜ − cos2χ sin 2ψ ⎟
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎝0 0 0 −1 ⎠ ⎝ sin 2χ ⎠ ⎝ − sin 2χ ⎠
The orientation and ellipticity angles are expressed,
respectively, for the output beam by

1 ⎛S ⎞ 1 ⎛S ⎞
ψ= sin −1 ⎜ 2 ⎟⎟ , χ= sin −1 ⎜ 3 ⎟⎟ .
2 ⎜S 2 ⎜S
⎝ 1 ⎠ ⎝ 0 ⎠
Comparing the elements of the Stokes vector shows that
π π
ψ′ = − ψ, χ′ = χ − .
2 2

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 22 Polarization

Rotators

The final method for changing the polarization state of an

optical beam is to allow the beam to propagate through a
rotator. The Mueller matrix for a rotator is
⎛1 0 0 0⎞
⎜ ⎟
0 cos 2θ sin 2θ 0⎟
M ROT ( θ) = ⎜ ,
⎜ 0 − sin 2θ cos2θ 0 ⎟
⎜ ⎟
⎝0 0 0 1⎠

where θ is the angle of rotation. Rotators only rotate the

polarization ellipse; they do not affect the ellipticity. This
can be seen by the propagation of a beam through a
rotator:
⎛S ′ ⎞
⎜ 0 ⎟ ⎛1 0 0 0⎞⎛ 1 ⎞
⎜ S ′ ⎟ ⎜ 0 cos 2θ sin 2θ ⎟⎜
0 ⎟ ⎜ cos 2χ cos2ψ ⎟
⎟
⎜ 1 ⎟=⎜ ,
⎜ S2′ ⎟ ⎜ 0 − sin 2θ cos 2θ 0 ⎟ ⎜ cos 2χ sin 2ψ ⎟
⎜ ⎟ ⎜ ⎟⎜
sin 2χ
⎟
⎜S ′ ⎟ ⎝0 0 0 1 ⎠⎝ ⎠
⎝ 3 ⎠
which yields
⎛S ′ ⎞
⎜ 0 ⎟ ⎛ 1 ⎞
⎜ S ′ ⎟ ⎜ cos 2χ cos(2ψ + 2θ) ⎟
S′ = ⎜ 1 ⎟ = ⎜ ⎟.
⎜ S2′ ⎟ ⎜ cos2χ sin(2ψ + 2θ) ⎟
⎜ ⎟ ⎜ sin 2χ
⎟
⎜S ′ ⎟ ⎝ ⎠
⎝ 3 ⎠
The above equation shows that the ellipticity (χ) is not
affected by the rotator.
The polarizing elements given above are defined with
respect to the x-axis. If either the polarizer or the wave
plate M is rotated through an angle θ from the x-axis, the
Mueller matrix equation M(θ) for the rotated component
is
M ( θ) = M ROT ( −2θ) ⋅ M ⋅ M ROT (2θ),

where the angle θ is measured from the x-axis.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observables of Polarized Light 23

Mueller Matrices for Rotated Components

We now consider the Mueller matrices for several

polarizing components. The first is the Mueller matrix
for a rotated ideal linear polarizer which is
⎛ 1 cos 2θ sin 2θ 0⎞
⎜ 2 ⎟
1 ⎜ cos 2θ cos 2θ sin 2θ cos 2θ 0⎟
M POL ( θ) = ⎜ ⎟.
2 sin 2θ sin 2θ cos 2θ
⎜ sin 2 2θ 0⎟
⎜ 0 0 0 0 ⎟⎠
⎝
In particular, the Mueller matrix reduces to the following
special forms for θ = 0°, 45°, 90°, and 135°:
⎛1 1 0 0⎞ ⎛1 0 1 0⎞
⎜ ⎟ ⎜ ⎟
1 ⎜1 1 0 0⎟ 1 ⎜0 0 0 0⎟
M LHP = , M L+45P = ,
2 ⎜0 0 0 0⎟ 2 ⎜1 0 1 0⎟
⎜ ⎟ ⎜ ⎟
⎝0 0 0 0⎠ ⎝0 0 0 0⎠

⎛ 1 −1 0 0⎞ ⎛1 0 −1 0⎞
⎜ ⎟ ⎜ ⎟
1 −1 1 0 0⎟ 1 0 0 0 0⎟
M LVP = ⎜ , M L-45P = ⎜ .
2⎜ 0 0 0 0⎟ 2 ⎜ −1 0 1 0⎟
⎜ ⎟ ⎜ ⎟
⎝0 0 0 0⎠ ⎝0 0 0 0⎠

Another important rotated Mueller matrix is the Mueller

matrix of a rotated wave plate, which is
⎛1 0 0 0 ⎞
⎜ 2 2 ⎟
0 cos 2θ + cos φ sin 2θ (1 − cos φ)sin2 θ cos2θ sin φ sin2θ
MWP (θ) = ⎜ ⎟.
⎜ 0 (1 − cos φ)sin2θ cos2θ sin 2θ + cos φ cos 2θ − sin φ cos2θ ⎟
2 2
⎜ ⎟
⎝0 − sin φ sin2θ sin φ cos2θ cos φ ⎠
Of special interest is the Mueller matrix for the
rotated HWP, written as MHWP(θ), which is
⎛1 0 0 0⎞
⎜ ⎟
0 cos 4θ sin 4θ 0⎟
M HWP ( θ) = ⎜ .
⎜ 0 sin 4θ − cos4θ 0 ⎟
⎜ ⎟
⎝0 0 0 −1 ⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 24 Polarization

Mueller Matrices for Rotated Components (cont’d)

This Mueller matrix looks very similar to the Mueller

matrix for rotation. The difference is that the angle θ is
doubled and the negative signs show that the orientation
and ellipticity angles are reversed. Consequently, this
matrix is called a pseudo-rotator.
The Mueller matrix of a rotated QWP is also of
interest and is found to be
⎛1 0 0 0 ⎞
⎜ 2 ⎟
⎜0 cos 2θ sin 2θ cos 2θ sin 2θ ⎟
M QWP ( θ ) = ⎜ ⎟.
2
⎜ 0 sin 2θ cos2θ sin 2θ − cos2θ ⎟
⎜0 − sin 2θ cos 2θ 0 ⎟⎠
⎝
Multiplying this matrix by the Stokes vector of the output
beam for input L+45P light yields
⎛ 1 ⎞
⎜ ⎟
cos 2θ sin 2θ
S =⎜ ⎟.
⎜ sin 22θ ⎟
⎜⎜ ⎟⎟
⎝ cos2θ ⎠

Rotating the QWP from 0° in steps of 45° increments

generates the following Stokes vectors:
⎛1 ⎞ ⎛1 ⎞ ⎛1 ⎞ ⎛1 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
0 0 0 0
S (0°) = ⎜ ⎟ , S (45°) = ⎜ ⎟ , S (90°) = ⎜ ⎟ , S (135°) = ⎜ ⎟ .
⎜0⎟ ⎜1 ⎟ ⎜0⎟ ⎜1 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝1 ⎠ ⎝0⎠ ⎝ −1 ⎠ ⎝0⎠
Rotating the QWP generates RCP, L+45P, and LCP light
but not LVP, LHP, or L–45P polarized light.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observables of Polarized Light 25

Mueller Matrix Applications: Malus’s Law

The first of several applications of the Mueller matrices is

the propagation of a LHP beam through a polarizer
rotating through an angle.

Transmission axis

θ Output
Incident
LHP beam beam

Rotating linear polarizer

The Stokes vector of the output beam is

 1 cos 2θ sin 2θ 0  1 
 2  
1 cos 2θ cos 2θ cos 2θ sin 2θ 0  1 
S′ = 

2 sin 2θ cos 2θ sin 2θ sin 22θ 00
  
 0 0 0 00
and
 1 
 
1 + cos 2θ  cos2θ 
S′ = .
2  sin 2θ 
 
 0 
The output beam continues to be linearly polarized.
However, its intensity is
1 + cos 2θ
I ( θ) = = cos2 θ,
2
which is recognized as Malus’s Law.
An important application of polarizers and wave plates is
the optical shutter. A polarized incident beam is
completely blocked by a pair of crossed polarizers.
However, by placing a wave plate with its fast axis at
+45° between a pair of crossed polarizers and varying the
phase of the wave plate the intensity of the light that
emerges from the second polarizer can be controlled.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 26 Polarization

Mueller Matrix Applications – The Optical Shutter

45°
Input Output
beam beam

Horizontal Wave Vertical

polarizer plate polarizer
The shutter configuration is shown in the following figure.
The Mueller matrix for the variable-phase wave plate
with its fast axis at +45° is
⎛1 0 0 0 ⎞
⎜ ⎟
0 cos φ 0 sin φ ⎟
M WP (45°) = ⎜ .
⎜0 0 1 0 ⎟
⎜ ⎟
⎝ 0 − sin φ 0 cos φ ⎠

For an input beam of arbitrary polarization the Stokes

vector of the output beam is then
S′ = M ⋅ M WP ( φ, 45° ) ⋅ M ⋅ S,
polv polh

⎛1 ⎞

S′ =
(1 − cos φ ) (S0 + S1 ) ⎜⎜ −1 ⎟⎟
.
4 ⎜0⎟
⎜ ⎟
⎝0⎠
When φ = 0° (the wave plate is not present) the intensity,
as expected, is zero. We also see that the output beam
intensity will be zero if the input beam is LVP. The
intensity of the beam can be written, in general, as

I ( φ) = I 0 (1 − cos φ ) .

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observables of Polarized Light 27

Mueller Matrix Applications – Stokes Parameters

A plot of the output intensity of the optical shutter is

shown in the following figure.
Intensity

Phase Angle (Degrees)

The maximum intensity occurs at 180°. Thus, by varying

the phase from 0° to 180° there is complete control of the
output intensity. This behavior is the basis of optical
shutters used to prevent “flash blindness.”
The final application is of the Mueller calculus to the
classical measurement of the Stokes parameters. The
measurement requires a wave plate with its fast axis
fixed at 0° and a linear polarizer rotated through θ.
The input beam is characterized by its four Stokes
parameters S0, S1, S2, and S3. The Mueller matrix
M = Mpol(θ) ⋅ MWP(φ) and the output intensity are then

⎛ 1 cos2θ sin 2θ cos φ − sin 2θ sin φ ⎞

⎜ cos 2θ 2
cos 2θ sin 2θ cos 2θ cos φ − sin 2θ cos 2θ sin φ
⎟
1
M = ⎜ ⎟,
2 2
2 ⎜ sin 2θ sin 2θ cos2θ sin 2θ cos φ − sin 2θ sin φ ⎟
⎜ ⎟
⎝ 0 0 0 0 ⎠
I ( θ, φ) = (1 / 2)( S0 + S1 cos 2θ + S2 sin 2θ cos φ − S3 sin 2θ sin φ ).

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 28 Polarization

The Observable Polarization Sphere

When an incident beam propagates through several

polarizing elements, the mathematical manipulations to
determine the polarization state of the output beam are
tedious. Poincaré suggested that a graphical solution to
this problem could be obtained by plotting the
polarization state(s) in terms of the orientation and
ellipticity angles, ψ and χ, on the surface of a sphere,
where the input and output beams are represented by the
points P(ψ, χ) and P(ψ′, χ′). A difficulty arises because
measurements of polarized light are made in the intensity
(observable) domain, but the Poincaré sphere and its
spherical coordinates ψ and χ apply to the amplitude
domain. This problem can be overcome by reformulating
the Poincaré sphere so that the Stokes parameters S1, S2,
and S3, are plotted on the Cartesian axes and the
amplitude products E0x2, E0y2, and E0yE0x, and the phase δ
are plotted on the surface of the sphere. This requires
that the amplitude products be transformed to a spherical
angle. The reformulated sphere is called the observable
polarization sphere (OPS).
The Stokes parameters can be expressed in terms of
spherical coordinate angles by first dividing each of the
four equations representing the Stokes parameters by S0,
and then using the auxiliary relation equation
tanα = E0y / E0x transform S1, S2, and S3 to trigonometric
terms:
S0 = E0 x 2 + E0 y2 → 1,

S1 = E0 x 2 − E0 y2 → cos 2α,

S2 = 2E0 x E0 y cos δ → sin 2α cos δ,

S3 = 2E0 x E0 y sin δ → sin 2α sin δ, δ = δ y − δx .

On the OPS, the Stokes polarization parameters S1, S2,

and S3 represent the Cartesian axes, and 2α and δ are the
spherical coordinate angles.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observable Polarization Sphere 29

The Observable Polarization Sphere (cont’d)

The angles 2α and δ on the OPS are related to the Stokes

parameters by
⎛S ⎞
2α = cos−1 ⎜ 1 ⎟ , 0 ≤ 2α < π ,
⎝ S0 ⎠

⎛S ⎞
δ = tan −1 ⎜ 3 ⎟ , 0 ≤ δ < 2π.
⎝ S2 ⎠
The orientation and ellipticity angle s, ψ and χ, can be
expressed in terms of 2α and δ by
tan(2ψ ) = tan(2α ) cos δ,

sin(2χ) = sin(2α )sin δ.

Similarly, the above

S1
equations can be inverted
P
so 2α and δ can be
expressed in terms ψ and
2α
cos(2α ) = cos(2χ ) cos(2ψ ),
S3 cot δ = cot(2χ)sin(2ψ ).
δ
χ:
S2
The coordinates axes and
angles of the OPS are now arranged as shown. S2, S3, and
S1 correspond to the x-, y- and z-axes. The angles 2α and δ
are measured from the
positive S1 and S2, axes, S1
respectively.
The S1-S2 is the prime
meridian plane and
S2-S3 is the equatorial
plane. The latitude
lines represent the
phase δ and the S3
S2
longitude lines
represent 2α. OPS and the Cartesian axes

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 30 Polarization

The Observable Polarization Sphere (cont’d)

The Stokes vectors for LHP

beams emerging from a
wave plate, a rotator, and
an ideal linear polarizer
on the OPS are now LCP RCP
L+45P
plotted. For a wave plate
we have
S ′ = M WP ( φ) ⋅ S ,
LVP

⎛S ′ ⎞
⎜ 0 ⎟ ⎛ 1 ⎞ ⎛ 1 ⎞
⎜ S ′ ⎟ ⎜ cos 2α′ ⎟ ⎜ cos 2α
⎟
S′ = ⎜ 1 ⎟ = ⎜ ⎟=⎜ ⎟.
⎜ S2′ ⎟ ⎜ sin 2α ′ cos δ′ ⎟ ⎜ sin 2α cos ( δ + φ ) ⎟
⎜ ⎟ ⎜ ⎟ ⎜⎜ ⎟⎟
⎜ S ′ ⎟ ⎝ sin 2α ′ sin δ′ ⎠ ⎝ sin 2α sin ( δ + φ ) ⎠
⎝ 3 ⎠

This equation shows

S1 that the phase
increases along small
circles around the S1
axis. The equation is
plotted in increments
of φ = 15° on the OPS
S3 at a fixed latitude
S2 angle of 2α = 75°.
The phase behavior
on the sphere is restricted to the latitude lines. Now
consider the rotator:
S ′ = M ROT ( θ) ⋅ S ,
⎛S ′ ⎞
⎜ 0 ⎟ ⎛ 1 ⎞ ⎛ 1 ⎞
⎜ S ′ ⎟ ⎜ cos2α′ ⎟ ⎜ cos2θ cos2α − sin 2θ sin 2α cos δ ⎟
S′ = ⎜ 1 ⎟ = ⎜ ⎟=⎜ ⎟.
⎜ S2′ ⎟ ⎜ sin 2α′ cos δ′ ⎟ ⎜ sin 2θ cos2α + sin 2α cos2θ cos δ ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ S ′ ⎟ ⎝ sin 2α′ sin δ′ ⎠ ⎝ sin 2α sin δ ⎠
⎝ 3 ⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observable Polarization Sphere 31

The Observable Polarization Sphere (cont’d)

The equation for

rotation is plotted in S1
increments of θ = 10°
for an input beam with
α = 45° and δ = 30°.
On the OPS rotation
appears as small
circles around the S3
S3
axis. Finally, for a S2
rotated ideal linear
polarizer we have
S ′ = M POL ( θ) ⋅ S ,

⎛ 1 ⎞
⎜ ⎟
1 cos2θ ⎟
S ′ = ( S0 + S1 cos 2θ + S2 sin 2θ ) ⎜ .
2 ⎜ sin 2θ ⎟
⎜ ⎟
⎝ 0 ⎠
The output beam is linearly polarized and is always
restricted to the prime meridian as shown below on the
OPS.
S1

S3

S2

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 32 Polarization

Plotting the Quarter-Wave Plate on the OPS

We now consider several plotting applications on the OPS.

The first is rotation of a QWP.

Incident Rotating Output

Beam QWP Beam

It is easiest to understand the behavior of the rotating

QWP using an incident L+45P beam. For this input
polarization state the output beam is then
S′ = MQWP(θ) ⋅ SL+45P and the Stokes vector of the S′
output beam is
⎛ 1 ⎞
⎜ ⎟
cos 2θ sin 2θ ⎟
S′ = ⎜ ,
⎜ sin 2 2θ ⎟
⎜ ⎟
⎝ cos 2θ ⎠
where S′ is now plotted over 180° on the OPS and yields
S1

S3

The boldface dot is the initial L+45P point. A complete

rotation through 180° leads to a “figure 8.” The rotation
of the QWP in increments of 45° on the next page.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observable Polarization Sphere 33

Plotting the Quarter-Wave Plate on the OPS

(cont’d)

LCP L+45P RCP

L+45P

(a) (b)

LCP RCP
L+45P L+45P

(c) (d)

The input L+45P beam is immediately transformed to

RCP prior to any rotation and appears at the RCP point;
this can be confirmed by evaluating the Stokes vector S′
for θ = 0. Rotating the QWP through 45° the polarization
returns to (a) the L+45P point. Another 45° of rotation
moves the beam to (b) the LCP point; an additional 45°
rotation brings the polarization back again to (c) the
L+45P point, and another rotation through 45° brings the
polarization state back to the original starting point (d)
RCP light.
It was shown that the only degenerate polarization states
generated by rotating the QWP are RCP, L+45P, and
LCP. The rotation of the QWP cannot generate LHP,
LVP, or L-45P polarization states. This is clearly seen by
plotting the rotating QWP on the OPS. We also note that
rotation of the QWP appears in the form of phase shifts.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 34 Polarization

The Rotating Quarter-Wave Plate

The configuration for arbitrary linear polarization input

states that are rotated through 180°:

S Rotating LP Rotating QWP S′

The Stokes vector of the output beam S′ is

⎛ 1 ⎞
⎜ ⎟
1 1
⎜ cos(4θ − 2α ) + cos 2α ⎟
⎜2 2 ⎟
S' = ⎜
1 1 ⎟,
⎜ sin(4θ − 2α ) + sin 2α ⎟
⎜2 2 ⎟
⎜ − sin(2θ − 2α ) ⎟
⎝ ⎠
where θ and 2α refer to the rotation of the QWP and
orientation of the incident LP beam on the QWP. The
equation for S′ is now plotted for different LP
orientations.

(a) (b)
The views in (a) and (b) are along the positive S2 axis and
S3 axes, respectively. The figures show that using a
rotating linear polarizer and a rotating QWP any
polarization state on the polarization sphere can be
obtained.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observable Polarization Sphere 35

The Babinet-Soleil Compensator

A device that is used to generate and analyze polarized

light is the Babinet-Soleil compensator. This device is
a variable-phase wave plate that can be rotated through
360°. The device is described by the Mueller matrix for a
rotated wave plate.

Rotating
Incident Output
Variable-phase
beam beam
wave plate

For an incident L+45P beam the Stokes vector S′ is

⎛ 1 ⎞
⎜ cos 2θ sin 2θ(1 − cos φ) ⎟
S′ = ⎜ ⎟,
⎜ sin22θ + cos22θ cos φ ⎟
⎜⎜ ⎟⎟
⎝ cos 2θ sin φ ⎠
where θ is the rotation angle of the compensator and φ is
the phase shift. The equation for S′ is plotted on the OPS.

S1 S1

S3

S3

S2

(a) (b)

View (a) is along the S2 axis, and (b) is a nonaxial view.

The views show that the Babinet-Soleil compensator can
generate any polarization state using incident L+45P
light.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 36 Polarization

Linear and Circular Polarizers

Two polarizing components that are used extensively in

an optics laboratory are linear and circular polarizers. It
is worthwhile to understand the origins of these names.
The Mueller matrix for a rotated linear polarizer is
⎛ 1 cos 2θ sin 2θ 0⎞
⎜ cos 2θ 2
cos 2θ cos 2θ sin 2θ 0 ⎟⎟
1
M LP ( θ) = ⎜ .
2 ⎜ sin 2θ cos 2θ sin 2θ sin 2 2θ 0⎟
⎜ ⎟
⎝ 0 0 0 0⎠

The Stokes vector of an output beam using the standard

form S = (S0 , S1 , S2 , S3 ) for the input Stokes vector is then

⎛ 1 ⎞
⎜ ⎟
1 cos2θ ⎟
SLP = ( S0 + S1 cos2θ + S2 sin 2θ ) ⎜ .
2 ⎜ sin 2θ ⎟
⎜ ⎟
⎝ 0 ⎠
Thus, regardless of the polarization state of the input
beam the output beam is always linearly polarized.
A circular polarizer is constructed from an L+45P
polarizer and a QWP. The Mueller matrix of a circular
polarizer and the output beam are, respectively,
⎛1 0 1 0⎞
⎜ ⎟
0 0 0 0⎟
M CP = M QWP ⋅ M L+45P =⎜ ,
⎜0 0 0 0⎟
⎜ ⎟
⎝1 0 1 0⎠

⎛1 ⎞
⎜ ⎟
S
1 0
= S + S ⎜ ⎟.
CP 2 0 2 ⎜0⎟( )
⎜ ⎟
⎝1 ⎠
The output beam is always circularly polarized regardless
of the polarization state of the input beam.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observable Polarization Sphere 37

The Generation of Elliptically Polarized Light

As a final application of the OPS we consider the problem

of generating elliptically polarized light of any orientation
and ellipticity. To do this, consider the following figure:

SL+45P Wave plate Rotator S′,

where SL+45P is Stokes vector for L+45P light. The matrix

equation for the above diagram and the output S′ are

S ′ = M ROT ( θ) ⋅ M WP ( φ) ⋅ SL + 45P ,

⎛ 1 ⎞
⎜ ⎟
sin 2θ cos φ ⎟
S′ = ⎜ .
⎜ cos 2θ cos φ ⎟
⎜ ⎟
⎝ sin φ ⎠

The above equation is plotted on the OPS.

S1 S1

S3

S2 S3

(a) (b)

The figures show that S′ generates every polarization

state very uniformly on the OPS unlike, for example, the
Babinet-Soleil compensator. In addition, the Stokes vector
S′ shows that the ellipticity and the orientation angles are
directly obtained from χ= φ/2 and ψ = π/4 – θ.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 38 Polarization

Measurement Methods of the Stokes Parameters

In addition to the classical measurement to determine the

Stokes parameters, several other measurement methods
have been developed. In the classical measurement a
QWP, which is also absorbing, must be used in the final
measurement. In order to avoid this problem as well as
axial alignment problems of the linear polarizer and the

Input
beam Output
beam

Circular
polarizer
wave plate, a circular polarizer can be used.

A circular polarizer has the property that on one side it

behaves as a circular polarizer, whereas on the other side
it behaves as a linear polarizer. For an incident beam
with an arbitrary polarization, the intensity IC ( θ) of the
output beam on the circular side is
1
I C ( θ) = (S0 − S1 sin 2θ + S2 cos 2θ).
2
The circular polarizer is then flipped to the linear side.
The intensity IL ( θ) of the output beam is found to be
independent of the rotation of the circular polarizer and is
1
I L( θ) = (S0 + S3 ).
2
From the above two equations the intensities are
measured and the Stokes parameters are then found to be
S0 = I C (0°) + IC (90°), S1 = S0 − 2 IC (90°),

S2 = IC (0°) − IC (90°), S3 = 2 I L (0°) − S0 .

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observable Polarization Sphere 39

The Rotating Quarter-Wave Plate Measurement

Another method for measuring the Stokes parameters of a

polarized beam is to allow the beam to propagate through
a rotating QWP followed by a linear horizontal polarizer;
the wave plate rotates at an angular frequency of ω. This
arrangement is shown in the following figure.

Rotating Linear horizontal

S S′
QWP polarizer

The Mueller matrix equation for the above

configuration is
S ′ = M LHP ⋅ M QWP ( θ ) ⋅ S ,

where θ = ωt . Carrying out the matrix multiplication and

using the trigonometric half-angle formulas, the intensity
of the output beam is found have the form
1
I ( θ) = ( A + B sin 2θ + C cos 4θ + D sin 4θ )
2
where
S1 S1 S2
A = S0 + , B = S3 , C= , D= .
2 2 2
The Stokes parameters are then
S0 = A − C , S1 = 2C , S2 = 2D, S3 = B.

The terms A, B, C, and D can be determined from the

equation I(θ), which is a truncated Fourier series. It
shows that there is a dc term (A), a double frequency term
(B) and two quadruple frequency terms (C and D). Using
Fourier analysis the coefficients A, B, C, and D are then
2π 2π
1 2
A=
π ∫ I (θ)dθ,
0
B=
π ∫ I (θ)sin 2θdθ,
0

2π 2π
2 2
C=
π ∫ I (θ) cos 4θdθ,
0
D=
π ∫ I (θ)sin 4θdθ.
0

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 40 Polarization

Birefringent Crystals and Wave Plates

Quartz is a birefringent crystal used extensively as a

wave plate. A crystal is generally characterized by three
orthogonal axes, the x-, y-, and z-axis. Along these axes
the propagation velocities of a wave are called the
principal velocities vx, vy, and vz , and are expressed by
vx = c / ε x , vy = c / ε y , vz = c / ε z ,

where εx, εy, and εz are the principal dielectric constants

and c is the speed of light. The velocity of propagation v of
a wave in a crystal is governed by Fresnel’s equation of
wave normals sx, sy, and sz:
2
sx 2 sy sz 2
+ + = 0.
v2 − vx 2 v2 − vy 2 v2 − vz 2
There are two solutions for v and are found to correspond
to a spherical wave and an ellipsoidal wave. In the
direction of the optic axis, however, the ellipsoidal wave
becomes spherical.
For a uniaxial crystal with its optic axis in the z
direction, v x = v y = v o [the ordinary wave (o)] and v z = v e
[the extraordinary wave (e)]. Substituting these conditions
into Fresnel’s equation, above, yields two solutions:
v′2 = vo 2 , and v′′2 = vo 2 cos2 θ + ve 2 sin2 θ,
where the angle θ is measured from the optics axis. Along
the optic axis v ″ 2 = v o 2 , and perpendicular to the optic
axis v ″ 2 = v e 2 . The maximum phase shift occurs when the
propagation is perpendicular to the optic axis. The phase
shift φ of the wave propagating through a wave plate is
then
2π
φ= (no − ne )d,
λ
where no and ne are the refractive indices of the ordinary
and extraordinary axes, respectively, and d is the path
length.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Observable Polarization Sphere 41

Multiple and Zero-Order Wave Plates

For a QWP and a HWP, the optical thickness

( no − ne )d can be expressed as

λ λ
(no − ne )d = ( 4m + 1 ) = mλ + = mλ + k1λ,
4 4
λ λ
(no − ne )d = ( 2m + 1 ) = mλ + = mλ + k2 λ,
2 2
where m = 0 , 1 , 2 ⋅ ⋅ ⋅ , and k 1 = 1 / 4 and k 2 = 1 / 2 ,
respectively. For m = 0 , 1 , 2 ⋅ ⋅ ⋅ , the wave plates behave
like QWPs and HWPs. Because multiple wavelengths can
propagate through the wave plate, it is called a multiple-
order wave plate.
It is difficult to control the optical thickness in the
fabrication of wave plates. However, by using two wave
plates with their fast axes at +45° and
–45° this can be overcome. The Mueller matrix for this
combination is
⎛1 0 0 0 ⎞
⎜ ⎟
0 cos( φ1 − φ2 ) 0 sin( φ1 − φ2 ) ⎟
M WP12 =⎜ .
⎜0 0 1 0 ⎟
⎜⎜ ⎟
⎝ 0 − sin( φ1 − φ2 ) 0 cos( φ1 − φ2 ) ⎟⎠
Then, using the above relations the Mueller matrix
becomes
⎛1 0 0 0 ⎞
⎜ ⎟
⎜ 0 cos ⎡⎣ 2π ( k1 − k )2 ⎤⎦ 0 sin ⎡⎣ 2π ( k1 − k2 ) ⎤⎦ ⎟
M WP12 =⎜ ⎟.
⎜0 0 1 0 ⎟
⎜ 0 − sin ⎡ 2π ( k − k ) ⎤ 0 cos ⎣ 2π ( k1 − k2 ) ⎦ ⎟⎠
⎡ ⎤
⎝ ⎣ 1 2 ⎦

Thus, the multiplicity parameter m cancels out. Because

m does not appear, the two-wave-plate combination is
commonly called a zero-order wave plate. The parameters
k1 and k2 appear as k 1 – k 2 = k . Since there are now two
variables (k1, k2), greater control can be exercised when
polishing in order to obtain the correct phase shift,φ.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 42 Polarization

Mueller Matrices for Reflection and Transmission

The reflection and transmission of light at an air-

dielectric (glass) interface is shown.

RP
EP

Es Rs
i i

n
r Tp

Ts

Here, Es,p, Rs,p, and Ts,p are the incident, reflected, and
transmitted field components, s and p refer to the
components perpendicular and parallel to the plane of the
paper, i and r are the incident and refracted angles, and n
is the refractive index of the medium (glass). The
reflection and transmission is governed by Fresnel’s
equations:
tan(i − r ) sin(i − r )
Rp = Ep , Rs = − Es ,
tan(i + r ) sin(i + r )

2sin r cos i 2sin r cos i

Tp = Ep , Ts = Es .
sin(i + r ) cos(i − r ) sin(i + r )

The Stokes parameters for the reflected field are

defined by
S0 R = cos i( Rs Rs * + Rp Rp *), S1 R = cos i( Rs Rs * − Rp Rp *),

S2 R = cos i( Rs Rp * + Rp Rs *), S3 R = j cos i( Rs Rp * − Rp Rs *),

where the factor j = −1 and * is the complex conjugate.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Reflection and Transmission 43
Reflection and Transmission Stokes Parameters

The Mueller-Stokes matrix equation for reflection is

given by
SR = M R ⋅ S ,
where MR is the Mueller matrix for reflection. The
individual Stokes polarization parameters are
S0 R = fR [(cos2 α − + cos2 α + )S0 + (cos2 α − − cos2 α + )S1 ],
S1 R = fR [(cos2 α − − cos2 α + )S0 + (cos2 α − + cos2 α + )S1 ],
S2 R = − fR (2cos α − cos α + )S2 , S3 R = − fR (2 cos α − cos α + )S3 ,
2
1 ⎛ tan α − ⎞
fR = ⎜ ⎟ ,
2 ⎝ sin α + ⎠

where equations α ± = i ± r . The Stokes parameters for

reflection correspond to reflection by a polarizing element.
Similarly, the Mueller-Stokes equation for
transmission is
ST = M T ⋅ S.
The Stokes parameters for the transmitted components
are
S0T = n cos i(TsTs * +TpTp *), S1T = n cos i(TsTs * −TpTp *),

S 2T = n cos i (TsTp* +TpTs*), S3T = j n cos i(TsTp * −TpTs *).

The Stokes parameters for the transmitted beam
are then
S0T = fT [(cos2 α − + 1)S0 + (cos2 α − − 1)S1 ],

S1T = fT [(cos2 α − − 1)S0 + (cos2 α − + 1)S1 ],

S2T = − fT (2cos α − )S2 ,
1 sin 2i sin 2r
S3T = − fT (2cos α − )S3 , fT = .
2 (sin α + cos α − )2

The Stokes parameters for transmission correspond to

transmission through a polarizer.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 44 Polarization

Reflection and Transmission Stokes Parameters

(cont’d)

From the previous relations S 0 = S 0R+ S 0T; that is, the

total intensity (energy) is conserved. The previous
relations are now used to consider the behavior of incident
unpolarized light, S = I 0 { 1,0,0,0} , that is reflected by a
dielectric surface. The Stokes vector SR is

⎛ S0 R ⎞ ⎛ cos2 α − + cos2 α + ⎞
⎜ ⎟ 2 ⎜ ⎟
⎜ S1 R ⎟ 1 ⎛ tan α − ⎞ ⎜ cos2 α − − cos2 α + ⎟
SR = = .
⎜ S2 R ⎟ 2 ⎝⎜ sin α + ⎠⎟ ⎜ 0 ⎟
⎜⎜ ⎟⎟ ⎜⎜ ⎟⎟
⎝ S3 R ⎠ ⎝ 0 ⎠
The DOP P of the reflected light is

S1 R cos2 α − − cos2 α +
P= = .
S0 R cos2 α − + cos2 α +

In general, P is less than 1. If, however, cos α + = 0 then

P = 1 ( = 100%) and i + r = π / 2 . This is the Brewster’s
angle condition and SR then reduces to

⎛ S0 R ⎞ ⎛1 ⎞
⎜ ⎟ ⎜ ⎟
S1R ⎟ 1 1
SR = ⎜ = cos 2iB ⎜ ⎟ .
2
⎜ S2 R ⎟ 2 ⎜0⎟
⎜⎜ ⎟⎟ ⎜ ⎟
⎝ S3 R ⎠ ⎝0⎠
Thus, at the Brewster’s angle, iB, the reflected light is
LHP and the LVP component vanishes. When using a
LVP polarizer as an analyzer, the intensity of the beam
that emerges is 0. The angle of incidence iB has now been
found so that from Brewster’s law (n = taniB) the
refractive index of the glass can be determined.
On the following page a figure is shown for P versus the
incident angle i for a glass with a refractive index of
n =1.50.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Reflection and Transmission 45
Reflection and Transmission Stokes Parameters
(cont’d)

Polarization
Deg of

Incident Angle i

Here, the vertical line represents the Brewster angle. The

DOP is 0 at an incident angle of 0°, rises to a maximum of
1 at the Brewster angle (~57.4°) and then returns to 0 at
90°. The above figure also shows that changing the
incident angle i allows any DOP to be obtained, a property
of reflection that is often overlooked.
The Stokes vector of a transmitted beam through a
single surface of a dielectric (glass) for incident
unpolarized light is
⎛ S0T ⎞ ⎛ cos2 2α − + 1 ⎞
⎜ ⎟ ⎜ 2 ⎟
⎜ S1T ⎟ = 1 sin 2i sin 2r ⎜ cos 2α − − 1 ⎟ ,
⎜ S2T ⎟ 2 (sin α cos α )2 ⎜ 0 ⎟
+ −
⎜⎜ ⎟⎟ ⎜⎜ ⎟⎟
⎝ S3T ⎠ ⎝ 0 ⎠
and the DOP of the transmitted beam is

S1 cos2 α − − 1
P= = .
S0 cos2 α − + 1

n = 3.5
DOP

n = 2.5

n = 1.5

Incident Angle i

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 46 Polarization

Reflection and Transmission Stokes Parameters

(cont’d)

The transmitted light is always partially polarized, and

the incident unpolarized light never becomes completely
polarized using a single surface. However, by increasing
the refractive index, the DOP increases, and for an
incident angle of 90°, the DOP reduces to

n2 − 1
P= .
n2 + 1
DOP

n = 1…20

Refractive Index n

The Mueller matrices for reflection and transmission

assume simplified forms:
Normal Incidence (i = 0°):
⎛1 0 0
0⎞ ⎛1 0 0 0⎞
⎜ 2 ⎟ ⎜ ⎟
⎛ n −1 ⎞ ⎜0 1 0 0⎟ 4n ⎜0 1 0 0⎟
MR = ⎜ ⎟ , MT = ,
⎝ n +1 ⎠ ⎜0 0 −1 0 ⎟ ( + 1) ⎜0 0 1 0⎟
2
n
⎜ ⎟ ⎜ ⎟
⎝0 0 0 −1 ⎠ ⎝0 0 0 1⎠

where n is the refractive index of the medium. For

incident unpolarized light of intensity I0 the reflected and
transmitted intensities are
2
⎛ n −1 ⎞ 4n
IR = ⎜ ⎟ I 0 , IT = I ,
( n + 1)
2 0
⎝ n +1 ⎠

and IR + IT = I0.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 48 Polarization

Total Internal Reflection (TIR)

An interesting phenomenon takes place when an optical

beam propagates from the medium into air.

EP RP

n Es Rs
i i

Tp
r
Ts

In this arrangement Snell’s law becomes n sin i = sin r.

The maximum value of r is π /2; therefore, the beam
cannot propagate from the medium to air if n sin i >1. For
this condition the incident beam is reflected back into the
medium and there is total internal reflection (TIR).
Fresnel’s reflection equations for the parallel (p) and
perpendicular (s) components become
− iδ p
Rp = e E p , Rs = e −iδs Es ,

which leads to
2 2
δ cos i n sin i − 1
tan = , δ = δ p − δs .
2 n sin 2 i
The Mueller matrix for TIR then becomes
1 0 0 0

 
0 1 0 0
MR =  .
0 0 cos δ − sin δ 
 
0 0 sin δ cos δ 

This matrix is seen to correspond to a wave plate.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Reflection and Transmission 49
The Fresnel Rhomb

The phenomenon of TIR was first used by Fresnel to

create circularly polarized light from linearly polarized
light. In order to use TIR, Fresnel constructed a glass
rhomb.
Input
Beam

i i
i Output
Beam

Using input L+45P light, circularly polarized light can be

obtained by creating a phase shift (δ) of 45° at the lower
and upper surfaces of the glass rhomb, respectively. The
Mueller matrix for two reflections is

1 0 0  0
 
0 1 0 0
M = ,
0 0 cos( δU + δL ) − sin( δU + δ L ) 
 
0 0 sin( δU + δ L ) cos( δU + δL ) 

where δU,L are the phase shifts at the upper and lower
surfaces, respectively. For glass such as BK7 the
refractive index is n = 1.5151 at 6328 Å (HeNe). From the
equation for δ on the previous page the angle of incidence
of I = 55°05′ yields a phase shift of δ = 45.0°. Then
δU + δL = 90°, so the Mueller matrix for the Fresnel
rhomb is
1 0 0 0
 
0 1 0 0
M = ,
0 0 0 −1 
 
0 0 1 0

For an input L+45P beam SL+45P = {1, 0, 1, 0} and the

Stokes vector of the output beam is then SRCP = {1, 0, 0, 1}.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 50 Polarization

Single and Multiple Dielectric Plates

For a single dielectric plate, there are an infinite number

of internal reflections between the two surfaces; however,
most of these can usually be ignored.

S
i i

n
r

i
S′
A single dielectric plate.

For transmission through the upper and lower surface of

the dielectric plate the Mueller matrix equation for the
plate is
ST = M T2 ⋅ S.

In terms of the Stokes parameters of the incident beam,

the stokes parameters of the transmitted beam are
S0T = fT [(cos4 α − + 1)S0 + (cos4 α − − 1)S1 ],

S1T = fT [(cos4 α − − 1)S0 + (cos4 α − + 1)S1 ],

S2T = − fT (2cos2 α − )S2 ,

2
1  sin 2i sin 2r 
S3T = − fT (2 cos2 α − )S3 , fT =   ,
2  (sin α + cos α − )2 

where equations α ± = i ± r = sin–1[sin( i ) / n ]. For incident

unpolarized light, the equation for the intensity of the
output beam is
2
1  sin 2i sin 2r 
IT = 
4
(
 cos α − + 1 .
2  (sin α + cos α − )2 
)

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Reflection and Transmission 51
Single and Multiple Dielectric Plates (cont’d)

n = 1.40
Transmission intensity

n = 1.60

Incident angle i
A plot of the output beam intensity equation for a refractive index of n =
1.4, 1.5, and 1.6 from i = 0° to 90°.

The DOP is
S1 1 − cos4 α −
P= = .
S0 1 + cos4 α −

This equation is plotted for n = 1.40, 1.45…1.60 from

i = 0° to 90°.

n = 1.60
DOP

n = 1.40

Incident angle

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 52 Polarization

Pile of Polarizing Dielectric Plates

The transmitted beam is always partially polarized. An

increase in the DOP can be obtained using a configuration
known as a pile of polarizing plates.

Input
The plates
are sep- Output
beam beam
arated in
A configuration of 6 dielectric plates. order to
reduce the
effects of reflection. The Mueller matrix equation and the
Stokes vector for a beam transmitted through m
dielectric plates are
ST m = M T 2m ⋅ S ,

⎛ (cos4 m α − + 1)S0 + (cos4m α − − 1)S1 ⎞

⎜ 2m ⎟
m 1 ⎡ sin 2i sin 2r ⎤ ⎜ (cos4 m α − − 1)S0 + (cos4 m α − + 1)S1 ⎟
ST = ⎢ ⎥ .
2 ⎣ (sin α + cos α − )2 ⎦ ⎜ 2cos2m α − S2 ⎟
⎜⎜ ⎟
⎟
⎝ 2cos2m α − S2 ⎠
For input unpolarized light the transmitted intensity is
2m
1 ⎡ sin 2i sin 2r ⎤
IT = ⎢ ⎥ (cos4 m α − + 1).
2 ⎣ (sin α + cos α − )2 ⎦
Transmitted intensity

m=2

m = 10

Incident angle i
m = 2, 4…10 and n = 1.5

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Reflection and Transmission 53
Pile of Polarizing Dielectric Plates (cont’d)

The DOP for m dielectric plates is

S1 cos4 m(i − r ) − 1
P= = .
S0 cos4m(i + r ) − 1

m = 10
DOP

m =2

Incident Angle i
m = 2, 4…10 plates and n = 1.50

As the number of plates increase the DOP increases. At

least 4 plates are required to obtain a DOP of 100%. Also,
from the previous page, as the number of plates increase,
the intensity decreases and the DOP increases. Thus,
there is a trade-off when using a pile of polarizers to
obtain linearly polarized light from unpolarized light.
For unpolarized incident light ST m is

⎛ (cos2m α − + 1) ⎞
2m ⎜ ⎟
1 ⎡ sin 2i sin 2r ⎤ 2m
⎜ (cos α − − 1) ⎟ .
ST m = ⎢ ⎥
2 ⎣ (sin α + cos α − )2 ⎦ ⎜ 0 ⎟
⎜⎜ ⎟⎟
⎝ 0 ⎠
In the limit as m→∞ the term cos2mα_ in the Stokes vector
vanishes and LVP light STm →{1,–1,0,0} is obtained.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 54 Polarization

Pile of Polarizing Dielectric Plates (cont’d)

The transmission of unpolarized light through m

dielectric plates at the Brewster angle iB yields the Stokes
vector
⎛ sin 4m 2iB + 1 ⎞
⎜ ⎟
1 sin 4m 2iB − 1 ⎟
ST ,B = ⎜ .
2⎜ 0 ⎟
⎜⎜ ⎟⎟
⎝ 0 ⎠
The transmitted intensity is
1
IT ,B =
2
(
1 + sin 4m 2iB ,)
where iB = tan–1 (n). A plot of the above intensity is made
for refractive indices of n = 1.5…1.8.
Transmitted intensity

n = 1.5

n = 1.8

Number of plates m

Similarly, the degree of polarization (DOP) and its

plot are

S1 1 − sin 4m 2iB
P= = .
S0 1 + sin 4m 2iB

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Reflection and Transmission 55
Fresnel’s Reflection and Transmission Coefficients

A simpler notation for the Mueller matrices for reflection

and transmission is to use the Fresnel reflection and
transmission coefficients. These are defined to be
for reflection:
2 2 2 2
⎛R ⎞ ⎛ sin α − ⎞ ⎛ Rp ⎞ ⎛ tan α − ⎞
ρs = ⎜ s ⎟ = ⎜ ⎟ , ρp = ⎜ ⎟⎟ = ⎜ ⎟ ,
⎜E
E
⎝ s⎠ ⎝ sin α + ⎠ ⎝ p ⎠ ⎝ tan α + ⎠
and for transmission:
2
n cos r ⎛ Ts ⎞ sin2i sin 2r
τs = ⎜ ⎟ = ,
cos i ⎝ Es ⎠ sin2 α+
2
n cos r ⎛ Tp ⎞ sin 2i sin 2r
τp= ⎜ ⎟= 2 ,
cos i ⎝⎜ E p ⎟ sin α+ cos2 α−
⎠
where
ρs + ρ p = 1, τs + τ p = 1.
The Fresnel coefficients are between 0 and 1 and can be
easily expressed as percentages. At the Brewster angle the
Fresnel coefficients are
ρs,B = cos2 2iB , ρ p,B = 0, τs,B = sin 2 2iB , τ p,B = 1.

The Mueller matrices for reflection and transmission

are then
⎛ ρs + ρp ρs − ρ p 0 0 ⎞
⎜ ⎟
ρ − ρp ρs + ρ p 0 0
1⎜ s ⎟
Mρ = ⎜ ,
2⎜ 0 0 2 ρs ρ p 0 ⎟
⎟
⎜ 0 0 0 2 ρs ρ p ⎟⎠
⎝

⎛ τs + τp τs − τ p 0 0 ⎞
⎜ ⎟
τ − τp τs + τ p 0 0
1⎜ s ⎟
Mτ = ⎜ .
2⎜ 0 0 2 τs τ p 0 ⎟
⎟
⎜ 0 0 0 2 τs τ p ⎟⎠
⎝

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 56 Polarization

Fresnel’s Reflection and Transmission Coefficients

(cont’d)

The following four plots are made for the coefficients for
refractive indices of n = 1.5, 2.0, and 2.5.
Rho - S

2.5
2.0
1.5

Incident angle i
Rho - P

2.5
2.0
1.5
Incident angle i

2.5
2.0
Tau - S

1.5

Incident angle i

2.5
2.0
Tau - P

1.5

Incident angle i

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Other Polarization Matrix Calculi 57

The Jones Matrix Calculus

The Jones matrix calculus is a matrix formulation of

polarized light that consists of 2 × 1 Jones vectors to
describe the field components and 2 × 2 Jones matrices
to describe polarizing components. While a 2 × 2
formulation is “simpler” than the Mueller matrix
formulation the Jones formulation is limited to treating
only completely polarized light; it cannot describe
unpolarized or partially polarized light. The Jones
formulation is used when treating interference
phenomena or in problems where field amplitudes must
be superposed. A polarized beam propagating through a
polarizing element is shown below.

Y
Ey(z,t) Ey(z,t)′

Input Output
beam beam

Ex(z,t) Ex(z,t)′
X
Polarizing element

The 2 × 1 Jones column matrix or vector for the field is

⎛ Ex ⎞ ⎛ E0 x e x ⎞
iδ

E = ⎜⎜ ⎟⎟ = ⎜ ⎟,
⎝ E y ⎠ ⎜⎝ E0 y e ⎟
iδy
⎠
where E0x and E0y are the amplitudes, δx and δy are the
phases, and i = −1. The components Ex and Ey are
complex quantities. An important operation in the Jones
calculus is to determine the intensity I:
⎛ Ex ⎞
( )
I = Ex * E y * ⎜⎜ ⎟⎟ = Ex Ex * + E y E y * .
⎝ Ey ⎠
The row matrix is the complex transpose † of the column
matrix, so I can be written formally as
I = E † ⋅ E.
It is customary to normalize I to 1.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 58 Polarization

The Jones Matrix Calculus (cont’d)

The Jones vectors for the degenerate polarization

states are:
1  0 1 1 
ELHP =   , ELVP =   , EL + 45P =  ,
0 1  2 1 
1 1  1 1  1 1
EL − 45P =  , ERCP =  , ELCP =  .
2  −1  2 i 2  −i 
The Jones vectors are orthonormal and satisfy the
relation Ei†⋅Ej = δij, where δij(i = j , 1 , i ≠ j,0) is the
Kronecker delta.
The superposition of two orthogonal Jones vectors leads to
another Jones vector. For example,
 1   0  1 
E = ELHP + ELVP =   +   =   ,
 0   1  1 
which, aside from the normalizing factor of 1 / 2 , is
L+45P light. Similarly, the superposition of RCP and LCP
yields
1 1  1  1  2 1 
ERCP + ELCP =  +  =  ,
2 i 2  −i  2 0
which, again, aside from the normalizing factor is seen to
be LHP light. Finally, in its most general form, LHP and
LVP light are
 E0 x e i δ x   0 
ELHP = , ELVP =  .
 E ei δ y 
 0   0y 
Superposing ELHP and ELVP yields
 E0 x eiδx 
E = ELHP + ELVP =  .
 E0 y eiδ y 
 
This shows that two orthogonal oscillations of arbitrary
amplitude and phase can yield elliptically polarized
light.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Other Polarization Matrix Calculi 59

The Jones Matrix Calculus (cont’d)

A polarizing element is represented by a 2 × 2 Jones

matrix
⎛ jxx jxy ⎞
J = ⎜⎜ ⎟.
⎝ jyx jyy ⎟⎠

It is related to the 2 × 1 output and input Jones vectors by

E′ = J ⋅ E. For a linear polarizer the Jones matrix is
⎛ px 0⎞
JPOL = ⎜⎜ ⎟, 0 ≤ px , py ≤ 1.
⎝0 py ⎟⎠

For an ideal linear horizontal and linear vertical polarizer

the Jones matrices take the form, respectively,
⎛1 0 ⎞ ⎛0 0⎞
JLHP = ⎜ ⎟, JLVP = ⎜ ⎟.
⎝0 0⎠ ⎝0 1⎠
The Jones matrices for a wave plate (E0x = E0y = 1) with a
phase shift of φ/2 along the x-axis (fast) and φ/ 2 along the
y-axis (slow) are (i = −1 )
⎛ e i (φ / 2) 0 ⎞ ⎛1 0 ⎞
J WP = ⎜ − i (φ / 2) ⎟
→⎜ −i φ ⎟
.
⎝ 0 e ⎠ ⎝0 e ⎠
The Jones matrices for a QWP φ = π / 2 and HWP φ = π
are, respectively,
⎛1 0 ⎞ ⎛1 0 ⎞
JQWP = ⎜ ⎟, JHWP = ⎜ ⎟.
⎝ 0 −i ⎠ ⎝ 0 −1 ⎠
For an incident beam that is L-45P the output beam from
a QWP aside from a normalizing factor is
⎛ 1 0 ⎞ ⎛ 1 ⎞ ⎛1 ⎞
E′ = ⎜ ⎟⎜ ⎟ = ⎜ ⎟,
⎝ 0 −i ⎠ ⎝ −1 ⎠ ⎝ i ⎠
which is the Jones vector for RCP light. Finally, the
Jones matrix for a rotator is
⎛ cos θ sin θ ⎞
JROT ( θ) = ⎜ ⎟.
⎝ − sin θ cos θ ⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 60 Polarization

The Jones Matrix Calculus (cont’d)

For a rotated polarizing element the Jones matrix is given

by
J( θ) = JROT ( −θ) ⋅ J ⋅ JROT ( θ).

The Jones matrix for a rotated ideal LHP is

⎛ cos2 θ sin θ cos θ ⎞
JLHP ( θ) = ⎜ ⎟.
⎝ sin θ cos θ sin 2 θ ⎠

Similarly, the Jones matrix for a rotated wave plate

is

⎛ φ φ φ ⎞
⎜ cos 2 + i sin 2 cos 2θ i sin sin 2θ
2 ⎟
J WP ( φ, θ) = ⎜ ⎟.
⎜ φ φ φ ⎟
⎜ i sin sin 2θ cos − i sin cos 2θ ⎟
⎝ 2 2 2 ⎠
For a HWP φ = π the matrix reduces to
⎛ cos2θ sin 2θ ⎞
JWP ( θ) = ⎜ ⎟.
⎝ sin 2θ − cos2θ ⎠
The matrix is almost identical to the matrix for a rotator
except that the presence of the negative sign with cosθ
rather than with sinθ along with the factor of 2 shows
that the matrix is a pseudo-rotator; a rotating HWP
reverses the polarization ellipse and doubles the rotation
angle.
An application of the Jones matrix calculus is to
determine the intensity of an output beam when a
rotating polarizer is placed between two crossed
pol
ariz
ers.
Input Output
beam beam

LHP LP(θ) LVP

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Other Polarization Matrix Calculi 61

The Jones Matrix Calculus (cont’d)

The Jones vector for the output beam is E′ = J ⋅ E and the

Jones matrix for the three polarizer configuration is
 0 0
J = JLVP ⋅ JLP ( θ) ⋅ JLHP =  .
 sin θ cos θ 0 
For input LHP light the intensity of the output beam is
I = E′† ⋅ E′ = [1 – cos(4θ)]/8.
Intensity

Rotation angle θ

An important optical device is an optical isolator.

Incident Reflecting
beam surface

Circular polarizer

The Jones matrix equation and its expansion is

( ) (
J = JL-45P ⋅ JQWP ⋅ JREFL ⋅ JQWP ⋅ JL+45P , )
1  1 −1   1 0   1 0   1 0  1 1   0 0 
J=  ⋅ ⋅ ⋅ ⋅ = .
4  −1 1   0 i   0 −1   0 i  1 1   0 0 

Thus, no light is returned to the optical source and the

circular polarizer acts as an ideal optical isolator.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 62 Polarization

Wolf’s Coherency Matrix Calculus

Another 2 × 2 polarization matrix calculus is Wolf’s

coherency matrix calculus. This matrix calculus serves
as a useful bridge between the Mueller and Jones matrix
calculi. The coherency matrix C is defined in terms of the
complex products of the optical field:
⎛ Ex Ex * Ex E y * ⎞ ⎛ Cxx Cxy ⎞
C=⎜ ⎟=⎜
⎜ ⎟.
⎜ E E * E y E y * ⎟⎠ ⎝ C yx C yy ⎟⎠
⎝ y x
The matrix elements are related to the Stokes parameters
by
⎛ Ex Ex * Ex E y * ⎞ 1 ⎛ S0 + S1 S2 + iS3 ⎞
C=⎜ ⎟= ⎜ ⎟.
⎜ E E * E y E y * ⎟⎠ 2 ⎝ S2 − iS3 S0 − S1 ⎠
⎝ y x
The coherency matrix can be expanded as a linear
superposition of the Stokes parameters,
1 3
C = ∑ σi Si ,
2 i =0
where
1 ⎛1 0 ⎞ 1 ⎛1 0 ⎞
σ0 = ⎜ ⎟ , σ1 = ⎜ ⎟,
2 ⎝0 1⎠ 2 ⎝ 0 −1 ⎠
1 ⎛0 1⎞ 1⎛0 i⎞
⎜ σ2 =
⎟ , σ3 = ⎜ ⎟,
2 ⎝1 0 ⎠ 2 ⎝ −i 0 ⎠
are the Pauli spin matrices. The common states of
polarized light expressed in terms of the coherency matrix
are
1 ⎛1 0 ⎞ 1 ⎛0 0⎞ 1 ⎛1 1 ⎞
CLHP = ⎜ ⎟ , CLVP = ⎜ ⎟ , CL+45P = ⎜ ⎟
2 ⎝0 0⎠ 2 ⎝0 1⎠ 2 ⎝1 1 ⎠

1 ⎛ 1 −1 ⎞ 1 ⎛ 1 i⎞ 1 ⎛1 −i ⎞
CL-45P = ⎜ ⎟ , CRCP = ⎜ ⎟ , CLCP = ⎜ ⎟,
2 ⎝ −1 1 ⎠ 2 ⎝ −i 1 ⎠ 2 ⎝i 1 ⎠
1 ⎛1 0⎞ 1 ⎛ 1 + cos2α sin 2αei δ ⎞
CUNP = ⎜ ⎟, CELP = ⎜ ⎟.
2 ⎝0 1⎠ 2 ⎝ sin 2αe −i δ 1 − cos 2α ⎠
where α = arctan(E0y/E0x) and δ is the phase for ELP light.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Optical Activity and Optical Rotation 63

Optical Activity and Optical Rotation

When polarized light propagates through a quartz crystal

the polarization ellipse rotates. This behavior also takes
place in many liquids such as sugars and fruit acids. This
is called natural optical activity. For liquids, the angle
of rotation θ is given by
γρl
θ= ,
10
where γ is the rotary power of the liquid, ρ is the density
(gms/cc), and l is the path length (cm). Natural optical
activity can be described by the Mueller matrix for
rotation
⎛1 0 0 0⎞
⎜ ⎟
0 cos 2θ sin 2θ 0⎟
M ROT ( θ) = ⎜ .
⎜ 0 − sin 2θ cos 2θ 0⎟
⎜ ⎟
⎝0 0 0 1⎠

A beam propagates through an optically active medium,

which is then reflected by an ideal reflecting surface and
propagates back through the medium.
The Mueller
Incident Reflecting
beam matrix for this
surface
configuration is
Optically active described by
medium
M = M ROT ( θ) ⋅ M REFL ⋅ M ROT ( θ).

The rotation angle θ in the reflected path remains

unchanged since the optically active medium is
independent of direction (The rotation angle is θ = kz and
on the return path k →–k, z→–z so (–k)(–z) → kz; k is the
propagator and z is the distance.) Carrying out the matrix
multiplication yields
⎛1 0 0
0⎞
⎜ ⎟
0 1 0
0⎟
M = M ROT ( θ) ⋅ M REFL ⋅ M ROT ( θ) = ⎜ .
⎜0 0 −1 0 ⎟
⎜ ⎟
⎝0 0 0 −1 ⎠
Thus, the effect of natural rotation is cancelled.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 64 Polarization

Faraday Rotation

When a polarized beam propagates through a block of

glass that is subjected to a very strong magnetic field, the
direction of the beam’s propagation is parallel to the
direction of the magnetic field and the polarization ellipse
rotates. Materials that exhibit this behavior are called
Faraday, magneto-optical media, or, more commonly,
Faraday rotators. The rotation angle is given by
θ = V H l,

where V is Verdet’s constant, H is magnetic field

intensity, and l is the propagation distance. Faraday
rotation is described by a Mueller rotation matrix that is
identical to that of optical activity except the magnetic
field intensity is directional: H is +H when propagation is
left to right, and –H when the propagation is right to left.
From the relation the rotation angle becomes –θ on the
return path. The configuration for propagation of a beam
that is reflected back through a magneto-optical medium
is shown.

Incident beam Reflecting

surface
Reflected beam
Magneto-optical medium

The Mueller matrix for this configuration is

⎡1 0 0 0⎤
⎢ ⎥
0 cos(4θ) sin(4θ) 0 ⎥
M = M ROT ( −θ) ⋅ M REFL ⋅ M ROT (θ) = ⎢ .
⎢0 sin(4θ) − cos(4θ) 0 ⎥
⎢ ⎥
⎣0 0 0 −1⎦

The above matrix is the Mueller matrix of a pseudo-

rotation matrix; that is, in a single trip the rotation angle
doubles and the ellipticity is reversed.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Optical Activity and Optical Rotation 65

Faraday Rotation (cont’d)

If an additional reflector is placed before the Faraday

medium (the medium is now in an optical cavity,) then for
N trips the Mueller rotation matrix becomes
⎛1 0 0 0⎞
⎜ ⎟
0 cos 4 N θ sin 4 N θ 0⎟
MN =⎜ .
⎜ 0 − sin 4 N θ cos 4 N θ 0⎟
⎜ ⎟
⎝0 0 0 1⎠

By slightly tilting the reflecting surface on the right side

of the optical cavity, the effect of allowing the beam to
make N passes can be used to measure the small rotation
angle θ.
For propagation in a Faraday medium the field can be
expressed as a superposition of two circularly
polarized waves propagating with different wave
numbers:
k′ = nL k0 , k′′ = nR k0 ,

where k0 is the propagation constant in free space and nL

and nR are the refractive indices associated with each of
the circular field components, respectively. The circular
birefringence is then defined to be
β
nL − nR = ,
n
where β is a parameter and n is the mean refractive
index. The refractive indices are also called dextro-
rotary and levo-rotatory or, simply, R- and L-rotatory,
and indicates that LCP and RCP waves propagate with
different phase velocities.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 66 Polarization

Optical Isolators

The advent of coherent lasers required the development of

optical isolators in order to prevent reflections from
returning to the laser source. The classical optical
isolator is the circular polarizer constructed from a QWP
and a L+45P polarizer. If the axial alignment between the
wave plate and the polarizer is not exact, then the
intensity of the beam returned to the laser is
I ( θ) = 2sin2 2θ(1 − cos2θ),

where θ is the angle of alignment. This equation shows

that at θ = 0, the reflected intensity is zero so there is
total (ideal) isolation.
Isolation (db)

Alignment angle θ

The isolation intensity can range over a very wide range

so it is useful to express the isolation in terms of db
(decibels).
db = −10 log10 ⎡⎣ I ( θ) ⎤⎦ .
Isolation – normalized

Alignment angle θ

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Optical Activity and Optical Rotation 67

Optical Isolators (cont’d)

Reflectors are imperfect due to phase shifts other than

180° and attenuation (α); ideally the reflected intensity
should be 0. For nonideal conditions the phase shift for
the isolation are plotted directly and in db.
Intensity
Isolation

α = 0.1

α = 0.9

Phase angle (degrees)

Intensity db
Isolation

α = 0.1
α = 0.9
Phase angle (degrees)

There is maximum isolation for a phase shift of 180°; the

plots are for attenuations (α) from 10% to 90%. The actual
performances of optical isolators attenuate the reflected
beam by ~ 30 to 40 db.
The classical optical isolator is limited to operate at a
single wavelength. This can be overcome by using a
Faraday isolator configuration.

PBS
Laser Faraday Reflecting
rotator surface

A polarizing beam splitter (PBS) transmits LHP light

directly through the prism and reflects LVP light out
through the sides of the prism. The dots and the vertical
lines represent LHP and LVP light, respectively.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 68 Polarization

Optical Isolators (cont’d)

The Mueller matrix for the Faraday rotator-

reflecting surface combination is
⎛1 0 0 0⎞
⎜ ⎟
0 cos 4θ sin 4θ 0⎟
M FARADAY = ⎜ .
⎜ 0 sin 4θ − cos 4θ 0 ⎟
⎜ ⎟
⎝0 0 0 −1 ⎠

In the forward direction the PBS behaves as a LHP

polarizer, so if the return beam propagates along this
path back to the laser:
M LASER = M LHP ⋅ M FARADAY ⋅ M LHP

⎛1 1 0 0⎞
⎜ ⎟
1 + cos 4θ ⎜ 1 1 0 0⎟
= .
4 ⎜0 0 0 0⎟
⎜ ⎟
⎝0 0 0 0⎠

Similarly, for a beam that is reflected by the prism so that

it propagates out of the side of the prism:
M SIDE = M LVP ⋅ M FARADAY ⋅ M LHP

⎛1 1 0 0⎞
⎜ ⎟
1 − cos 4θ ⎜ −1 −1 0 0⎟
= .
4 ⎜0 0 0 0⎟
⎜ ⎟
⎝0 0 0 0⎠

If the fast axis of the Faraday rotator is rotated to θ = 45°,

the above two matrices are reduced to
⎛0 0 0 0⎞ ⎛1 1 0 0⎞
⎜ ⎟ ⎜ ⎟
0 0 0 0⎟ 1 −1 −1 0 0⎟
M LASER =⎜ , and M SIDE = ⎜ .
⎜0 0 0 0⎟ 2⎜ 0 0 0 0⎟
⎜ ⎟ ⎜ ⎟
⎝0 0 0 0⎠ ⎝0 0 0 0⎠

No energy is returned to the optical source (laser), and all

of the reflected energy emerges out of the lower side of the
PBS.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Optical Activity and Optical Rotation 69

Optical Isolators (cont’d)

The optical isolation can be expressed as a function of the

rotation angle of the Faraday rotator. The intensity of the
beam returned to a laser that emits LHP light of unit
intensity is
1 + cos4θ
I ( θ) = .
2
normalized
Isolation -

Alignment angle of the fast axis θ

When the fast axis of the Faraday rotator is at an angle of

θ = 45° the isolation (intensity) is 0. However, if the fast
axis is at 90°, the intensity returned to the laser is unity;
that is, there is no isolation.
The isolation can also be expressed in terms of db:
db = −10 log10 [(1 + cos4θ) /2].
Isolation

Alignment angle of the fast axis θ

These two plots represent an ideal isolator. In practice,

linear polarizers are not ideal and Mueller matrices for
imperfect polarizers must be used.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 70 Polarization

Optical Isolators (cont’d)

Linear polarizers can also be expressed as major and

minor transmittances k1 and k2, respectively, where
they are related to the elements of the polarizer matrix by
k1 = px2 and k2 = px2.
Isolation db

k2 = 0.1

k2 = 0.9

Alignment of the fast axis θ

An analysis of Faraday isolators is readily made using

imperfect polarizers. The following two plots show the
behavior of the isolation in terms of intensity and
expressed also in db for k1 = 0.95 and k2 = 0.9, 0.7…0.1;
the smaller the value of k2, the better the polarizer
(k2 = k1 = k corresponds to a neutral density filter). The
best isolation is for k2 = 0.1.

k2 = 0.9
Isolation db

k2 = 0.1

Alignment of the fast axis - θ

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Optical Activity and Optical Rotation 71

Optical Isolators (cont’d)

The maximum isolation is approximately 30 db. Unlike

classical optical isolators that are fixed frequency QWP,
Faraday rotators are tunable over frequency ω. This is
due to the fact that Verdet’s constant V actually varies
over ω. The rotation angle of the Faraday rotator is θ =
VHz. Verdet’s constant actually is described by the
Lorentz spectrum and is given by
2ω
V =f ,
( ω − ω0 ) + (2ωL ω)2
2 2 2

where ω is the frequency, ω0 is the resonance frequency,

ωL is the Larmor frequency, and f is a constant.
Verdet’s constant

Frequency ω
ω0 = 1000 Hz and ωL = 10 Hz

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 72 Polarization

Wave Plate Depolarizers

An optical depolarizer is an element that transforms

polarized light to depolarized light. In order to obtain
depolarized light the Mueller matrix and the Stokes
vector relation must be

⎛ m00 0 0 0 ⎞ ⎛ S0 ⎞ ⎛1 ⎞
⎜ ⎟⎜ ⎟ ⎜ ⎟
0 0 0 0 ⎟ ⎜ S1 ⎟ 0
S =⎜
⎜ 0 0 0 0 ⎟ ⎜ S2 ⎟
= I0 ⎜ ⎟ ,
⎜0⎟ ( I0 = m00S0 ) .
⎜ ⎟⎜ ⎟ ⎜ ⎟
⎝ 0 0 0 0 ⎠ ⎜⎝ S3 ⎟⎠ ⎝0⎠
Thus, all the elements except for m00 must be 0 in the
matrix of a depolarizer. Depolarizers can be created by
spatial or temporal averaging of the matrix elements. One
method is to use two variable-phase wave plates whose
fast axes are oriented at +45° from one another as shown
below, where n is an integer.

Wave plate Wave plate Output

Input
0°, δ 45°, nδ

If the phase of the first wave plate is taken to be δ and the

second is nδ the matrix equation is
M = M WP (n δ) ⋅ M WP ( δ),

⎛1 0 0 0 ⎞
⎜ ⎟
⎜ 0 cos n δ sin nδ sin δ sin nδ cos δ ⎟
M = .
⎜0 0 cos δ − sin δ sin δ ⎟
⎜ ⎟
⎝ 0 − sin nδ cos nδ sin δ cos nδ cos δ ⎠
Averaging over the phase δ that varies linearly in time
the Mueller matrix for a depolarizer is obtained for
n = 2:

⎛1 0 0 0⎞
⎜ ⎟
0 0 0 0⎟
M Depolarizer =⎜ .
⎜0 0 0 0⎟
⎜ ⎟
⎝0 0 0 0⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Depolarizers 73

Wave Plate Depolarizers (cont’d)

Another method to depolarize polarized light is to rotate a

quarter- and half-wave plate at rotation angles of θ1 and
θ2.

Input QWP θ1 HWP θ2 Output

In this method the previous variable phase shifting (an

electro-optical effect) is replaced by a mechanical
rotation of each of the wave plates; the rotation angles are
expressed in terms of angular frequencies θ1 = ωt and θ2 =
nωt, where n is an integer. For an input beam that is
LHP, the Stokes vector of the output beam is then found
to be

⎛ 1 ⎞
⎜ ⎟
1 1
⎜ cos ⎡4(n − 1)ω t ⎤ + cos(4n ω t ) ⎟
⎜2 ⎣ ⎦ 2 ⎟
Sout =⎜ ⎟,
1 1
⎜ sin ⎣⎡4(n − 1)ω t ⎦⎤ + sin(4n ωt ) ⎟
⎜2 2 ⎟
⎜ sin(2ω t ) ⎟
⎝ ⎠
where ω = 2π/T and T is the period of rotation. Taking the
time average of each Stokes parameter over a period T,
the Stokes vector (Sout) is found to be for
n = 2,
⎛1 ⎞
⎜ ⎟
0
Sout = ⎜ ⎟,
⎜0⎟
⎜ ⎟
⎝0⎠
which is the Stokes vector for depolarized (unpolarized)
light. Thus, a QWP and a HWP in which the rotation rate
of the HWP is twice that of the QWP can be used as a
depolarizer. For the mechanical method to work requires,
however, that the rotation rate significantly exceed the
response time of the optical detector.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 74 Polarization

The Lyot Crystal Depolarizer

The Lyot crystal depolarizer operates in the

wavelength domain. This depolarizer consists of two
plates of quartz whose thicknesses have a ratio of 2:1
(corresponding to a QWP and a HWP); their x-axes differ
by an angle of 45°.

Input QWP HWP Output

The Mueller matrix for the Lyot crystal depolarizer

is
⎛1 0 0 0 ⎞
⎜ ⎟
0 cos n φ sin n φ sin φ sin n φ cos φ
M LYOT ( φ, θ = π /4) = ⎜ ⎟.
⎜0 0 cos φ − sin φ ⎟
⎜ ⎟
⎝ 0 − sin n φ cos n φ sin φ cos n φ sin φ ⎠
The phase shift is assumed to vary linearly with
wavelength:
2π( λ − λ1 )
φ( λ ) = a λ + b = ,
λ 2 − λ1

where a and b are evaluated at λ1 where φ(λ1) = 0 and at

λ2 where φ(λ2) = 2π. The average value of each matrix
element mij is determined by integrating from λ1 to λ2
using
λ2
1
mij =
λ 2 − λ1 ∫m
λ1
ij ( λ ) dλ.

Evaluating all the matrix elements yields 0 only for n = 2,

and so the Mueller matrix reduces to a depolarizer:
⎛1 0 0 0⎞
⎜ ⎟
0 0 0 0⎟
M LYOT =⎜ .
⎜0 0 0 0⎟
⎜ ⎟
⎝0 0 0 0⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Polarizing Materials 75

Polarizers

For polarizers the most commonly used materials are

naturally occurring calcite and synthetically produced
polaroid (sheets) and PolaCor. For wave plates the
most common materials are quartz and synthetic
materials. Finally, rotators are usually made using
quartz.
The best polarizing material for use in the visible
spectrum is calcite. Materials can be characterized as
isotropic, e.g., glass, uniaxial, e.g., calcite, and biaxial,
e.g., mica. For each of these structures an orthogonal
triad axis system can be constructed.
z, no z, ne z, nz

y, no y, no y, ny
x, no x, no x, nx
Isotropic Uniaxial Biaxial
medium medium medium

Both calcite and quartz are uniaxial crystals. In order to

understand the behavior of the propagation of e- and o-
rays we restrict the discussion to isotropic and uniaxial
media. In an isotropic medium all three axes have the
same refractive index whereas in a uniaxial medium the
refractive index along the z-axis is different from the
other two.
In an isotropic medium for an unpolarized beam
propagating along the z-axis, the transverse components
are directed along the x- and y-axes and the refractive
indices are both no. Thus, as the beam propagates along
the z-axis the transverse waves both experience the same
refractive index, so they propagate at the same velocity.
Similarly, the same behavior is observed if the
propagation is along the y- and z-axes. Thus, the
polarization state of an optical beam is unaffected when
propagating in an isotropic medium.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 76 Polarization

Polarizing Prisms

For a uniaxial crystal the transverse components of a

propagating beam along the z-axis both experience the
same refractive index. The z-axis is the optic axis where
the e- and o-rays do not divide and the beam propagates
through the crystal unaffected. However, when the beam
propagates along the x- or y-axis, one component
experiences a refractive index
The velocity of the
of no and the other experiences
wave is ν = c / n ,
ne. Thus, each component where c is the speed
travels at different velocities of light and n is the
and divides according to refractive index.
Snell’s law.
In the Glan-Foucault po-
larizer, the refractive indices of calcite are no = 1.658 and
ne = 1.486 (NaD line). The vertical double arrow and the
dot show the direction of the optic axis. The incident beam
is perpendicular to the optic axis so the unpolarized beam
divides into o- and e-rays. By choosing an apex angle of
θ = 38.5° the o-ray is internally reflected and e-ray is
refracted into the
air gap. The second θ
prism bends the e-
ray back to the e ray
original path and
emerges
perpendicular to Side view Top view
Air gap
the exit face; the
ray is 100% polarized.
In the Glan-Foucault prism only one ray appears at the
exit of the polarizer.
Other polarizing
prisms have been
designed in which both
rays can be used, and
o
o e include the Rochon,
e o the Sénarmont, and
e the Wollaston.
Rochon Sénarmont Wollaston

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Polarizing Materials 77

Polarizing Prisms (cont’d)

In most polarizing prisms, the optic axes are parallel in

both halves, but in the Rochon, Sénarmont, and Wollaston
prisms they are at right angles. Polarizing beam splitters
can also be made using quartz for the ultraviolet regime.
In most applications, however, the preferred material is
calcite because it yields a greater angular separation of
the beams, typically 10° as compared to 0.5° for quartz.
The Wollaston prism is especially useful because it
generates two orthogonal linear polarized beams of equal
intensity with a large angle of separation.
Calcite is expensive, so
synthetic polarizers have It is the iodine that gives
been developed. The most Polaroid its color.
common is H-sheet
Polaroid, where polarization occurs because of unequal
absorption along the orthogonal transmission axes.
Polaroid consists of polyvinyl-alcohol that is stretched,
cemented to a sheet of plastic, and dipped into a liquid
solution of iodine, fixing the long-chain molecules in the
plastic. Three of the most common polarizers and their
transmissions are HN-38 (38%), HN-32 (32%), and HN-22
(22%).
.9
HN-38 K1
Principal Transmittances – k1, k2

.8 K1
HN-32
.7
.6 HN-22 K1
.5
.4
.3
.2
.1

HN-38 K2
.01
.00 HN-32 K2
10–4
10–6 HN-22 K2
400 500 600 700

Wavelength µm

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 78 Polarization

Characterizing Polarizers
Commercial linear polarizers use a different form of
the Mueller matrix:
⎛ k1 + k2 k1 + k2 0 0 ⎞
⎜ ⎟
1 ⎜ k1 + k2 k1 + k2 0 0 ⎟
M POL = ⎜ , 0 ≤ k1, 2 ≤ 1,
2⎜ 0 0 2 k1k2 0 ⎟
⎟
⎜ 0 0 0 2 k1k2 ⎟⎠
⎝

where k1 = px2 is called the major principal

transmittance and k2 = py2 is called the minor
principal transmittance. For an ideal linear horizontal
polarizer k1 = 1 and k2 = 0. The Stokes vector for the
output beams are
⎛1 ⎞ ⎛1 ⎞ ⎛1 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
′ k1 + k2 ⎜ 1 ⎟ ′ k1 ⎜ 1 ⎟ ′ k2 ⎜ −1 ⎟
SUNP = , SLHP = , SLVP = .
2 ⎜0⎟ 2 ⎜0⎟ 2⎜0⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝0⎠ ⎝0⎠ ⎝0⎠
The transmittance, T, of the linear polarizer is
k1 + k2 k k
TUNP = , TLHP = 1 , TLVP = 2 .
2 2 2
The major and minor principal transmittances k1 and k2
are obtained using LHP and LVP light, respectively. The
transmittance of the linear polarizer shown on the
previous page using input unpolarized light is
k1 + k2
TUNP = ,
2
The unpolarized light condition can be produced by
sequentially using incident LHP and LVP light and then
adding the respective intensities; superposing the
transmittances TLHP and TLVP forms the transmittance
TUNP.
Polarizer characteristics are usually evaluated using
pairs of polarizers. The transmittances occur when the
horizontal trans-mission axes of both polarizers are in the
x-direction (parallel).

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Polarizing Materials 79

Characterizing Polarizers (cont’d)

Input
Output
beam
beam
Linear Linear
polarizer polarizer
The light the transmittances are measured when the
transmission axes of both polarizers are in the x-direction:
1 2 k2 k2
TUNP = H 0 =
2
k1 + k22 , ( TLHP = 1 ,
2
) TLVP = 2 ,
2
where H0 refers to the transmittances for parallel
polarizers. Similarly, for polarizers that are crossed:
TUNP = H 90 = k1 k2 .
For an ideal polarizer k2 = 0, so H90 = 0.
The transmittance for a single rotated polarizer
with its transmission axis at an angle θ for input UNP
light is
T ( θ) = k1 sin 2 θ + k2 cos2 θ.
For a polarizer pair consisting of a single fixed polarizer
followed by a rotated polarizer:
1
T ( θ) = k1 k2 sin 2 θ + ( k12 + k22 ) cos2 θ
2
= H 90 + ( H 0 − H 90 ) cos2 θ.
Because transmittance values vary so widely it is common
to express them as logarithms:
D0 = − log 10 H 0 , D90 = − log 10 H 90 .
Finally, an important means of characterizing linear
polarizers is by means of the contrast ratio.
The transmission for a single linear polarizer for the
values of k1 = 0.9 and k2 = 0.1 is
T ( θ) = 0.9sin 2 θ + 0.1cos2 θ.
The contrast ratio is a measure of the maximum to the
minimum transmittance and is defined by
T − Tmin k1 − k2
C = max = .
Tmax + Tmin k1 + k2

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 80 Polarization

Characterizing Polarizers (cont’d)

For a neutral density filter k1 = k2, C = 0. Similarly, for

k1 >> k2 the contrast ratio approaches unity (an ideal
linear polarizer). Thus, the greater the contrast ratio is
the better the polarizer is. The ratio of the minor principal
transmittance to the major principal transmittance is
defined to be k = k2 / k1, so C = (1–k)/(1+k) The contrast
ratio for a rotated linear polarizer is:
T ( θ) − Tmin ⎛ 1 − k ⎞
C ( θ) = = sin 2 θ =C(0)sin 2 θ.
Tmax + Tmin ⎜⎝ 1 + k ⎟⎠

The contrast ratio can be changed by rotating the

polarizer.
Contrast ratio

k = 0.001

k = 0.5

Rotation angle θ

The graph clusters around a contrast ratio of 1, where

most polarizers perform. Because the contrast ratio varies
so widely, it is also defined in terms of db:
C(db) = −10 log 10C( θ).
Contrast ratio

k = 0.5
k = 0.001

Rotation angle θ

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Polarizing Materials 81

Wave Plate Materials

The wave plate (retarder) changes the Optic axis

polarization state of an optical beam by
introducing a phase shift between the
orthogonal components of a polarized
beam. Wave plates materials include ne
calcite, quartz, or synthetic retarders.
For a calcite crystal the ordinary wave
expands as a circle whereas the no
A wave-vector
extraordinary wave expands as an surface for a
ellipse. For propagation parallel to the negative uniaxial
optic axis the refractive indices are crystal such as
calcite.
identical and the phases of the beam
components are unaffected. However, for
beam propagation perpendicular to the optic axis the phase
y-axis (slow axis) difference φ for a
wavelength λ and a
optic axis path length of d is
Incident 2π
beam z-axis φ= (ne − no )d.
λ
Calcite wave plates
x-axis (fast axis)
are rarely used
even though their birefringence is large; the wave plate
would be extremely thin and fragile. Quartz is a positive
uniaxial crystal and has refractive indices, no = 1.544 and
ne = 1.553; its birefringence is about 1/20th of that of
calcite, making it a much more suitable wave plate
material. Also, inexpensive QWPs and HWPs are made
from birefringent polyvinyl are available.
When an unpolarized beam propagates through a wave
plate, the emerging beam remains unpolarized. This
shows that the o- and e-rays of the unpolarized beam are
independent of each other. Wave plates, therefore, can
only affect completely or partially polarized light.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 82 Polarization

Incoherent Superposition and Decomposition

The superposition and decomposition of optical

beams in the intensity and amplitude domains are very
different. In the intensity domain a beam is described by
the Stokes vector and in the amplitude domain by the
Jones vector. The Stokes vector for unpolarized light
is
⎛1 ⎞
⎜ ⎟
0
SUNP = I 0 ⎜ ⎟ ,
⎜0⎟
⎜ ⎟
⎝0⎠
where I0 is the intensity of the optical beam. The first
Stokes parameter is the intensity and the remaining
three parameters describe the polarization. Because the
polarization parameters are 0 there is no amplitude or
phase relation between the orthogonal components of the
unpolarized light. Thus, the Stokes vector for unpolarized
light consists of two independent beams of equal
intensity; furthermore, the beams are orthogonal to each
other. This allows the Stokes vector for unpolarized light
to be decomposed into orthogonal polarization states in
the following ways:
⎛1 ⎞ ⎛1 ⎞ ⎛1 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
0 I 1 I −1
SUNP = I0 ⎜ ⎟ = 0 ⎜ ⎟ + 0 ⎜ ⎟ ,
⎜0⎟ 2 ⎜0⎟ 2 ⎜ 0 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝0⎠ ⎝0⎠ ⎝0⎠
⎛1 ⎞ ⎛1 ⎞ ⎛1 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
0 I 0 I 0
SUNP = I0 ⎜ ⎟ = 0 ⎜ ⎟ + 0 ⎜ ⎟ ,
⎜ 0 ⎟ 2 ⎜ 1 ⎟ 2 ⎜ −1 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝0⎠ ⎝0⎠ ⎝0⎠

⎛1 ⎞ ⎛1 ⎞ ⎛1 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
0 I 0 I 0
SUNP = I0 ⎜ ⎟ = 0 ⎜ ⎟ + 0 ⎜ ⎟ .
⎜0⎟ 2 ⎜0⎟ 2 ⎜ 0 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝0⎠ ⎝1 ⎠ ⎝ −1 ⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Superposition and Decomposition of Polarized Beams 83

Incoherent Decomposition––Ellipses

Unpolarized light can be decomposed into the Stokes

vectors for LHP and LVP light, L+45P and L-45P light
and RCP and LCP light, respectively. Conversely, two
orthogonal independent beams can be superposed to form
unpolarized light. It is this property that is used to
determine the transmission coefficients of linear
polarizers, using two independent beams of LHP and LVP
light to simulate unpolarized light.
Unpolarized light can, in general, be decomposed
(incoherent decomposition) into two orthogonal
ellipses,
⎛1 ⎞ ⎛ 1 ⎞ ⎛ 1 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
0 1 cos 2ψ cos 2χ ⎟ 1 ⎜ cos 2ψ cos(2χ − π) ⎟
SUNP =⎜ ⎟= ⎜ + .
⎜ 0 ⎟ 2 ⎜ sin 2ψ cos 2χ ⎟ 2 ⎜ sin 2ψ cos(2χ − π) ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝0⎠ ⎝ sin 2χ ⎠ ⎝ sin(2χ − π) ⎠

The orthogonal polarization ellipses for ψ = 30° and χ = 15°.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 84 Polarization

Coherent Superposition and Decomposition

Coherent superposition and decomposition applies

only when there is an amplitude and phase relation
between the orthogonal components. Consider the vector
representation E(z,t) of the optical field in the x-y plane:
E( z ,t ) = Ex ( z ,t ) i + E y ( z ,t ) j
= E0 x cos( kz − ωt ) i + E0 y cos( kz − ωt + δ) j,
where i and j are the unit vectors along the x- and y-axes,
respectively, and δ = δy − δx is the phase shift between the
orthogonal components. The above equation is the
vectorial representation for elliptically polarized
light. Where amplitudes E0x = E0y = E0 and the phase
shift is δ = ±90°:
E1 ( z ,t ) = E0 [cos( kz − ωt ) i − sin( kz − ωt ) j],
E2 ( z ,t ) = E0 [cos( kz − ωt ) i + sin( kz − ωt ) j],
describing RCP and LCP light, respectively. Together
they form the vectorial equation for LHP light:
E( z ,t ) = 2E0 cos( kz − ωt ) i.
Thus, the coherent superposition of RCP and LCP light
yields LHP light, and the coherent decomposition of LHP
yields RCP and LCP light. In terms of the Jones vectors,
LHP light can also be decomposed for RCP and LCP light:
⎛ 1 ⎞ 1 ⎛1 ⎞ 1 ⎛ 1 ⎞
E = ⎜ ⎟ = ⎜ ⎟ + ⎜ ⎟.
⎝ 0 ⎠ 2 ⎝ i ⎠ 2 ⎝ −i ⎠
Another example of coherent superposition is of L+45P
and L-45P light to obtain LHP light:
1 ⎛1 ⎞ 1 ⎛ 1 ⎞ 2 ⎛1 ⎞
⎜ ⎟+ ⎜ ⎟= ⎜ ⎟.
2 ⎝1 ⎠ 2 ⎝ −1 ⎠ 2 ⎝0⎠
In the intensity domain only unpolarized light can be
decomposed into two independent elliptically polarized
beams. In the amplitude domain, elliptically or
degenerate polarized light can be decomposed or
superposed to form other polarization states.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Electro-Optical Effect 85

The Electro-Optical Effect––Modulators

Certain liquids and crystals exhibit a phenomenon known

as the electro-optical effect. The Kerr effect is
proportional to the square of the applied electric field and
and the Pockels effect is linearly proportional to the
electric field. In both phenomena the electro-optical effect
causes a phase shift between the orthogonal field
components. This indicates that the Mueller matrix for
the effect can be represented by a wave plate. In the Kerr
effect the field is applied perpendicular to the incident
light.

Output
Birefringent beam
medium

Input L_ 45
beam polarizer
L+45
polarizer Modulating
voltage

For the Kerr quadratic effect, the field is applied

transversely to the direction of the incident field. When an
isotropic liquid is placed in the electric field it behaves
like a uniaxial crystal with the optic axis in the direction
of the propagation. The birefringence is given by

n p − ns = λBE 2 ,

where np and ns are the refractive indices parallel and

perpendicular to the plane of the paper, λ is the
wavelength, B is Kerr’s constant and E is the applied field.
Kerr cells have serious drawbacks: they require 5 to 10
times the voltage needed by a Pockels cell and many
liquids are toxic. As a result crystals that exhibit the
Pockels effect are used much more frequently.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 86 Polarization

The Electro-Optical Effect––Modulators (cont’d)

For the Pockels effect the phase shift varies with the
applied linear field. The most used crystal is potassium
dihydrogen phosphate (KDP) in its deuterated form.

Birefringent Output
medium beam

Input L+45
beam polarizer
L-45 Modulating
polarizer voltage

In the Pockels cell the applied field is longitudinal and in

the direction of the crystal optics axis; the incident light
propagates along the same axis. The phase shift between
the transverse field components is
ωn03 r63
φ = φ y − φx = V,
c
where r63 is a constant of the crystal and V is the applied
modulating voltage; the phase φ is often called the
retardation. The constants can be eliminated and the
equation rewritten as φ = π ( V / V π ) , where Vπ is the “half-
wave voltage”, which occurs at φ = π.
The Mueller matrix for Pockels electro-optical
modulator is
M PEOM = M L+45P ⋅ M ( φ) ⋅ M L-45P

⎛1 0 0⎞1
⎜ ⎟
1 − cos φ ⎜ 0 0 0 0⎟
M PEOM = .
4 ⎜ −1 0 −1 0 ⎟
⎜ ⎟
⎝0 0 0 0⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Electro-Optical Effect 87

The Pockels Cell

The intensity of the Pockels cell output beam is

I0 I
(1 − cos φ ) = 0 ⎡⎣1 − cos ( πV / Vπ )⎤⎦ ,
I ( φ) =
4 4
where V is the applied voltage. Vπ is the half-wave voltage
and the maximum intensity also occurs at this voltage.
Intensity

Voltage ratio πV / Vπ
The applied amplitude modulation voltage is
V
V = π + Vm sin ( ωm t ),
2
where Vm is the maximum modulation voltage and ωm is
the modulation frequency. Assuming Vm << Vπ and the
small angle approximation (sinθ ≈ θ ), the intensity and its
plot is

⎛I ⎞⎡ ⎛ Vm ⎞ ⎤
I (V ) = ⎜ 0 ⎟ ⎢1 + π ⎜ ⎟ sin( ωm t ) ⎥ ,
⎝ 2 ⎠ ⎣⎢ ⎝ Vπ ⎠ ⎦⎥

V
π m = 0.1
Intensity

Vπ

Modulation angle (deg)

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 88 Polarization

Refractive Index Measurement

The refractive index n is the most important optical

characteristic of glass. To measure n, the most accurate
transmission method is minimum deviation where a
transparent, extremely homogeneous prism of glass must
be made, and accurately cut and polished. An alternate
measurement method is to use reflection from an optical
surface, though it not as accurate as transmission
methods. The most common reflection methods are:
(1) The normal incidence reflection
Polarizing The Mueller matrix for the
beam reflected beam is
Optical ⎛1 0 0 0⎞
Source
Glass ⎜ 2 ⎟
⎛ n −1 ⎞ ⎜0 1 0 0⎟
M REFL =⎜ ⎟ .
⎝ n +1 ⎠ ⎜0 0 −1 0 ⎟
Optical ⎜ ⎟
detector ⎝0 0 0 −1 ⎠

The beam intensity of the reflecting surface is

2
⎛ n −1 ⎞
I REFL = ⎜ ⎟ I0 ,
⎝ n +1 ⎠
where I0 is the intensity of the beam incident on the glass.
The beam splitter prevents direct measurement of the
refractive index of the glass and this reduces the
measurement accuracy.
Incident Reflected (2) The Brewster angle
Laser beam beam D
measurement
Linear
i Analyze The Mueller matrix
polarizer i
for reflection at the
Brewster angle iB is
Glass
⎛1 1 0 0⎞
⎜ ⎟
1 1 1 0 0⎟
M B = cos 2 2iB ⎜ ,
2 ⎜0 0 0 0⎟
⎜ ⎟
⎝0 0 0 0⎠

which is the matrix for a linear horizontal polarizer LHP.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Refractive Index Measurements 89

Refractive Index Measurement (cont’d)

The Mueller matrix shows that the Stokes vector S′ of the

reflected beam is
⎛1 ⎞
⎜ ⎟
1 1
S ′ = ( S0 + S1 ) cos 2iB ⎜ ⎟ ,
2
2 ⎜0⎟
⎜ ⎟
⎝0⎠
where S0 and S1 are the Stokes parameters associated
with the incident beam. If the incident beam is LVP light
(S1 = –1), then the Stokes vector S′ becomes 0; that is, the
LVP component vanishes and only the LHP component of
the incident beam remains. At the Brewster angle the
analyzing linear
polarizer is rotated
Refractive

until a null con-

index n

dition is reached
and the refractive
index n is found
using Brewster’s
law n = taniB. Brewster angle iB (deg)

The slow rise in

this curve indi-
Refractive
index n

cates that a very

good mechanical
mount is re-
quired to find
the location of Brewster angle iB (deg)
the Brewster
angle. Then, even when this angle or surrounding region
is found, the refractive index varies very slowly, and is
practically linear over a range of small angles, making the
refractive index measurement accurate to about three
decimal places.
(3) The 45° incidence angle measurement
This method is to irradiate the glass surface with an
incident beam at 45° form the surface normal.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 90 Polarization

Refractive Index Measurement (cont’d)

At this angle the Mueller matrix for reflection can be

expressed solely in terms of the angle of refraction:
⎛ 1 sin 2r 0 0 ⎞
⎜ ⎟
1 − sin 2r ⎜ sin 2r 1 0 0 ⎟
M (i = 45°) = .
(1 + sin 2r )2 ⎜ 0 0 − cos2r 0 ⎟
⎜ ⎟
⎝ 0 0 0 − cos 2r ⎠
If the surface of the glass is irradiated with LHP and then
with LVP light the intensities are found to be maximum
and minimum intensities, respectively, and are
1 − sin 2r
Imax = I 0 ,
1 + sin 2r

(1 − sin 2r )2
Imin = I 0 ,
(1 + sin 2r )2
which yields
Imax − Imin
sin 2r = .
Imax + Imin

Calling the intensity ratio (Imax – Imin) / (Imax + Imin) = f the

refractive index n can then be determined:

(1 + )
1/ 2
1− f2
n= .
f
Using this measurement method the expensive
mechanical mount is eliminated and there is no
mechanical motion other than flipping the polarizer from
LHP to LVP.

Sample
Ideal linear
Laser polarizer 45°
45°

Digital Transimpedance Optical

voltmeter amplifier detector

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Radiation Field 91

Maxwell’s Equations

A complete description of polarized light requires that the

nature of the optical source and its relation to the
propagating wave and polarization be included. This
requires the use of Maxwell’s electrodynamic equations
(expressed here in Gaussian units):
1 ∂ B(r,t )
∇ × E(r,t ) = − ,
c ∂t

4π 1 ∂ D(r,t )
∇ × H(r,t ) = j(r,t ) + ,
c c ∂t

∇ ⋅ D(r,t ) = 4 πρ(r,t ),

∇ ⋅ B(r,t ) = 0.
E(r,t) and H(r,t) are the electric and magnetic field
vectors, D(r,t) and B(r,t) are the displacement and
magnetic induction vectors, j(r,t) is the electric current
density vector, and ρ(r,t) is the electric charge density.
∇ is the (spatial) “del” operator, and c is the speed of
light. The field vectors are related by the constitutive
equations for the medium, D(r,t) = εE(r,t) and
B(r,t) = µH(r,t,), where ε and µ are permittivity and
permeability constants of the medium.
Maxwell’s equations can be recast only in terms of
E(r, t ) and j(r, t ) ,

µε ∂ 2 E(r,t ) 4 πµ ∂ j(r,t )
∇2 E(r,t ) − = 2 .
c2 ∂t 2 c ∂t

The current density is j(r,t) = ev(r,t), where e is the

electric charge and v(r,t) is the velocity. The above
equation is then
µ ε ∂ 2 E(r,t ) 4 πµ e ∂ v(r,t )
∇2 E(r,t ) − = .
c2 ∂t 2 c2 ∂t

The above equation shows that the radiated electric field

E(r,t) arises from accelerating charges, e∂v(r,t)/ ∂t.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 92 Polarization

The Radiation Equation and the Stokes Parameters

The above equation is known as the radiation equation.

We note that in free space there are no current sources so
the radiation equation reduces to the classical wave
equation. The radiation equation can be solved for its
transverse field components. In spherical coordinates
these are:
e
Eθ = [ x cos θ − z sin θ],
c2 R
e
Eφ = 2
[ 
y ],
c R
where Eθ and Eφ are the transverse field components, R is
the distance of the observation point P from the origin of
the coordinate system, and x ,  y , and z are the
accelerations of the charged particle in a Cartesian
coordinate system.

O R

y
x

The Stokes parameters for the radiation field

expressed in spherical coordinates are defined by
S0 = Eθ Eθ * + Eφ Eφ *,

S1 = Eθ Eθ * − Eφ Eφ *,

S2 = Eθ Eφ * + Eφ Eθ *,

S3 = i( Eθ Eφ * − Eφ Eθ *).

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Radiation Field 93

The Linear Oscillating Bound Charge

We now determine the Stokes vectors for several

radiation systems. The first is the Stokes vector for a
bound charge oscillating along the z-axis.
The motion of the Eθ Eφ
bound charge along z z
is described by P

d 2 z (t ) -e
+ ω02 z(t ) = 0, θ R
dt 2 O
where ω0 is the
φ y
frequency of oscillation
of the charge. In x
complex notation the
solution of this differential equation is

z(t ) = z0 ei ( ω 0 t +α ) ,

where z0 is the amplitude. Differentiating the above

equation and substituting into the radiation field
components yields
e z0 ω0 2
Eθ = sin θ ei( ω0 t +α ) , Eφ = 0.
c2 R
The Stokes vector is
⎛1 ⎞
2 ⎜ ⎟
⎛ ez ⎞ 1
S = ⎜ 2 0 ⎟ sin 2 θ ω04 ⎜ ⎟ .
⎝c R⎠ ⎜0⎟
⎜ ⎟
⎝0⎠
Thus, the radiation corresponds to LHP light and the
frequency appears as a fourth power. The intensity of the
radiation is

I ( θ) = I 0 sin2 θ.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 94 Polarization

The Randomly Oscillating Bound Charge

z The z-axis angle is θ = 0

and the intensity is zero.
θ Perpendicular to the z-
I(θ)
y axis the angle is θ = 90°
and the intensity is
maximum. The oscil-
lating bound charge
emits a radiation pattern that corresponds to a dipole
radiator.
To determine the Stokes vector
for an ensemble of randomly y
O α -e
oriented bound oscillating
charges of amplitude A, the A
vector equation is β
r(t ) + ω02 r(t ) = 0,

x
where the charge oscillates z
along A through the origin O.
The equations of motion along the x-, y-, and z-axes are
x (t ) = A sin α cos β ei ω 0 t , y(t ) = A sin α sin β ei ω 0 t , and
z(t ) = A cos α ei ω 0 t .
The accelerating charges for x(t), y(t), and z(t) are
determined, substituted into the field equations, and the
Stokes parameters are then found. The random behavior
of the oscillating charge is described by taking an
ensemble average over the angles α and β of the Stokes
parameters, leading to a Stokes vector for unpolarized
light:
z ⎛1 ⎞
2 ⎜ ⎟
8π ⎛ eA ⎞ 4⎜ 0 ⎟
S= ω .
3 ⎜⎝ c 2 R ⎟⎠ ⎜0⎟
0
O ⎜ ⎟
a -e
ω0 t ⎝0⎠
x y The Stokes vector for a
charge moving in a
circular path of radius a in
the x-y plane with an angular frequency of ω0 is shown.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Radiation Field 95

A Charge Moving in a Circle

The equations of motion of the charge are

x (t ) = a cos( ω0 t ),

y(t ) = a sin( ω0 t ), and

z(t ) = 0.
The accelerations are determined along with the field
equations, which then lead to the following Stokes vector
for elliptically polarized light:
⎛1 + cos 2 θ ⎞
⎜2 2 ⎟
⎛ eA ⎞ 4 ⎜ sin θ ⎟
S = ⎜ 2 ⎟ ω0 ,
⎝c R⎠ ⎜ 0 ⎟
⎜⎜ ⎟⎟
⎝ 2 cos θ ⎠
which corresponds to the polarization ellipse is in its
standard form. The angle θ refers to the observer’s
viewing angle measured from the z-axis. This result
shows that the observed polarization states depend on the
position of the observer. Along the z-axis at the angles θ =
0° or 180° the Stokes vector reduce, respectively, to

2
⎛1 ⎞ ⎛1 ⎞
⎛ eA ⎞ ⎜ ⎟ 2 ⎜ ⎟
0 ⎛ eA ⎞ 0
SRCP = 2⎜ ⎟ ω04 ⎜ ⎟ , and SLCP = 2 ⎜ 2 ⎟ ω04 ⎜ ⎟ .
⎜ 2 ⎟ ⎜0⎟ ⎝c R⎠ ⎜0⎟
⎝c R⎠ ⎜ ⎟ ⎜ ⎟
⎝1 ⎠ ⎝ −1 ⎠
Similarly, viewing the radiation at θ = 90°or 270° the
Stokes vector for LHP light is
⎛1 ⎞
⎜ ⎟
2
⎛ eA ⎞ 1
SLHP = ⎜ 2 ⎟ ω04 ⎜ ⎟ .
⎝ c R ⎠ ⎜ 0⎟
⎜ ⎟
⎝0⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 96 Polarization

A Charge Moving in a Magnetic Field

The motion of a charged particle moving in a

constant magnetic field is directed along the z-axis so
H = Hk, where k is the unit vector.

H
. -e

O y
x

In the magnetic field the motion of the charge is governed

by the Lorentz force equation,
e
F = mr
( v × H).  =
c
The components for the motion along the three axes are
eH
x(t ) = y (t ),
mc
eH
y(t ) = −
 x (t )
mc
z(t ) = 0,
where the “over dots” represent differentiation with
respect to time. The equations show that the charge
moves only in the x-y plane. The differential equation pair
must be solved simultaneously. The equations are solved
with the initial conditions taken for the charge at the
origin (0, 0) moving with a velocity v0 along the x-axis and
0 along the y-axis (v0, 0):
x (0) = 0, y(0) = 0

and
x (0) = v0 , y (0) = 0.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Radiation Field 97

A Charge Moving in Magnetic Field (cont’d)

The solutions are

v0
x (t ) = sin( ωc t )
ωc
and
v0
y(t ) = ⎡1 − cos( ωc t ) ⎤⎦ ,
ωc ⎣

where ωc = eH/mc is the cyclotron frequency. The

equations are added and squared:
2 2
⎛ v ⎞ ⎛v ⎞
x (t )2 + ⎜ y(t ) − 0 ⎟ = ⎜ 0 ⎟ ,
⎝ ωc ⎠ ⎝ ωc ⎠
which is the equation of a circle with a radius v0 / ωc and
center at x = 0 and y = v0 / ωc. From the above equations
for x(t) and y(t) the Stokes vector is
⎛1 + cos2 θ ⎞
2 ⎜ ⎟
⎛ ea ⎞ 2
4 ⎜ − sin θ ⎟
S = ⎜ 2 ⎟ ωc ⎜ ⎟.
⎝c R⎠ ⎜ 0 ⎟
⎜ −2 cos θ ⎟
⎝ ⎠
Thus, in a constant magnetic field the emitted radiation is
elliptically polarized and the intensity is proportional to
ωc4. The radiation pattern for the intensity, S0 is referred
to as “peanut” radiation.

θ I(θ)

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 98 Polarization

The Classical Zeeman Effect

When a sodium flame is examined with a spectroscope a

very brilliant pair of yellow lines known as the D lines
appear. If this same source is then placed between the
poles of a very powerful magnet with its magnetic field
directed parallel to the z-axis a single line splits into two
lines if the radiation is observed parallel to the z-axis, or
three lines if the radiation is observed perpendicular to
the z-axis. This phenomenon is known as the classical
z Zeeman effect. To analyze
this effect the Stokes vectors of
H α these lines must be deter-
A mined. The model of the
O
sodium atom is assumed to
β y behave as a bound oscillator.
x
The double-headed arrow
describes the motion of the
charged particle, where A is the maximum oscillation
distance and H is the magnetic field. The equation of
motion of the bound charged particle is described by
the Lorentz force equation
e
r(t ) + k r(t ) = [ v × H].
m 
c
The components of this equation are
⎛ eH ⎞
x(t ) + ω02 x (t ) = ⎜ ⎟ y (t ),
⎝ mc ⎠
⎛ eH ⎞
y(t ) + ω0 2 y(t ) = − ⎜
 ⎟ x (t ),
⎝ mc ⎠
z(t ) + ω02 z(t ) = 0,
where ω 0 = k / m is the natural frequency of oscillation
of the electron in the sodium atom.
The solutions to the differential equations are
x (t ) = A sin α cos( ωL t ) cos( ω0 t ),
y(t ) = A sin α sin( ωL t ) cos( ω0 t ),
z(t ) = A cos α cos( ω0 t ),

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Radiation Field 99

The Classical Zeeman Effect (cont’d)

where ωL = eH / 2mc is known as the Larmor precession

frequency, because in the magnetic field the oscillating
electron is found to precess around the z-axis. The
corresponding Stokes vector is
⎡2 4 4 2 4 4 2 ⎤
⎢ 3 ( ω+ + ω− )(1 + cos θ) + 3 ω0 sin θ ⎥
⎢ ⎥
2 ⎢ 2 4 4 2 4 4 2 ⎥
2 ⎛ eA ⎞ − ( ω + ω )sin θ + ω sin θ ⎥,
S= ⎜ 2 ⎟ ⎢ 3 + −
3
0

3 ⎝ 2c R ⎠ ⎢ ⎥
⎢ 0 ⎥
⎢ 4 4 4
⎥
⎢ ( ω+ − ω− ) cos θ ⎥
⎣ 3 ⎦

where ω± = ω0 ± ωL and θ is the spherical polar angle

measured from the z-axis. The equation can be
decomposed in terms of its frequencies ω−, ω0, and ω+:

⎡ ⎛1 + cos2 θ ⎞ ⎛ sin 2 θ ⎞ ⎛1 + cos2 θ ⎞ ⎤

⎢ ⎜ 2 ⎟ ⎜ 2 ⎟ ⎜ ⎟⎥
2 ⎛ eA ⎞ 2
− sin θ ⎟ sin θ ⎟ − sin 2 θ ⎟ ⎥
S = ⎜ 2 ⎟ ⎢ω− 4 ⎜ + 2ω04 ⎜ + ω+ 4 ⎜ .
3 ⎝ 2c R ⎠ ⎢ ⎜ 0 ⎟ ⎜ 0 ⎟ ⎜ 0 ⎟⎥
⎢ ⎜⎜ ⎟⎟ ⎜⎜ ⎟⎟ ⎜⎜ ⎟⎟ ⎥
⎢⎣ ⎝ −2cos θ ⎠ ⎝ 0 ⎠ ⎝ 2cos θ ⎠ ⎥⎦

The Lorentz-Zeeman model predicts that three

spectral lines will be observed at frequencies of
ω− = ω0 –ωL, ω0, and ω+ = ω0 – ωL. Furthermore, the spectral
lines are each polarized.
The line associated with ω−
is left elliptically polarized,
for ω0 the line is linear
horizontally polarized and
for ω+ the line is right
ω− ω0 ω+ elliptically polarized. In a
spectroscope the spectral pattern is observed. The central
spectral line ω0 is twice as bright as the left and right
spectral lines ω− and ω+.
When the radiation is viewed parallel and perpendicular
to the magnetic field the former the observation angle is
θ = 0° and the Stokes vector reduces to

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 100 Polarization

The Classical Zeeman Effect (cont’d)

⎡ ⎛1 ⎞ ⎛ 1 ⎞⎤
2⎢ ⎜ ⎟ ⎜ ⎟⎥
⎛
4 eA ⎞ 0 0
S = ⎜ 2 ⎟ ⎢ω−4 ⎜ ⎟ + ω+4 ⎜ ⎟ ⎥ .
3 ⎜⎝ 2 c R ⎟⎠ ⎢ ⎜0⎟ ⎜ 0 ⎟⎥
⎢ ⎜ ⎟ ⎜ ⎟⎥
⎢⎣ ⎝ −1 ⎠ ⎝ 1 ⎠ ⎥⎦

Thus, in a spectroscope
viewing the radiation
parallel to the magnetic
field only two spectral lines
ω− ω+ will be observed at ω− and
ω+ and the lines are left-
and right-circularly polarized.
Viewing the radiation perpendicular to the magnetic field
the observation angle is θ = 90° and the Stokes vector
reduces to
⎡ ⎛1 ⎞ ⎛1 ⎞ ⎛ 1 ⎞⎤
2 ⎢ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎥
⎛
2 eA ⎞ −1 1 −1
S = ⎜ 2 ⎟ ⎢ ω− 4 ⎜ ⎟ + 2ω0 4 ⎜ ⎟ + ω+ 4 ⎜ ⎟ ⎥ .
3 ⎝ 2c R ⎠ ⎢ ⎜0⎟ ⎜0⎟ ⎜ 0 ⎟⎥
⎢ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟⎥
⎣⎢ ⎝0⎠ ⎝0⎠ ⎝ 0 ⎠ ⎦⎥

The spectral lines are

vertically, horizon-
tally, and vertically
polarized. The central
spectral line is twice
as bright as the ω– ω0 ω+
adjacent linear ver-
tically polarized lines. When the magnetic field is
removed (ω±→ω0) the three Stokes vectors reduce to a
single Stokes vector:
⎛1 ⎞
2 ⎜ ⎟
8 ⎛ eA ⎞ 0
S = ⎜ 2 ⎟ ω04 ⎜ ⎟ .
3 ⎝ 2c R ⎠ ⎜ 0⎟
⎜ ⎟
⎝0⎠
This is the original Stokes vector (unpolarized) that is
observed in a spectroscope at the natural frequency of ω0.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Radiation Field 101

Optical Scattering

The basic idea behind the phenomenon of scattering is

that radiation incident on an electron is reradiated or
scattered. Scattering can occur with either free electrons
in Thomson scattering, or bound electrons in Rayleigh
scattering.
In Thomson scattering, the incident beam propagates
along the z-axis
x Scattered and is then
beam scattered by
θ
Incident the electron.
z The motion of a
beam
Scattering free electron is
y center
r (t ) = −e E(t ),
m
where E(t) describes the transverse field of the incident
beam. The vector components are
−e
x(t ) = Ex (t ),
m
−e
y(t ) =
 E y (t ),
m
and the spherical components of the scattered field are
−e 2
Eθ = cos θ Ex ,
m c2 R
−e2
Eφ = Ey.
m c2 R
The Stokes vector of the scattered beam S′ is related to
the Stokes vector S of the incident beam by the matrix
relation
S ′ = M ⋅ S,
where M is the Mueller matrix of the scatter and is
⎛1 + cos2 θ − sin 2 θ 0 0 ⎞
⎜ 2 ⎟
1⎛ e 2
⎞ ⎜ − sin θ 1 + cos θ
2 2
0 0 ⎟
M = ⎜ ⎟ .
2 ⎝ mc2 R ⎠ ⎜ 0 0 2cos θ 0 ⎟
⎜⎜ ⎟
⎝ 0 0 0 2cos θ ⎠⎟

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 102 Polarization

Optical Scattering (cont’d)

The term e2/mc2 (= 2.82 × 10–13 cm) is the classical

radius of an electron. Because this term appears as a
squared quantity the scattering is directly related to the
area of the electron, that is, scattering is proportional to
the area of the scatterer.
The scattering matrix M reduces to two special forms for
the viewing angles of θ = 0° and θ = 90°, respectively.
These are
⎛1 0 0 0⎞
⎛ e2 ⎞ ⎜⎜ 0 1 0 0 ⎟⎟
2
M (0°) = ⎜ ,
⎜ mc2 R ⎟⎟ ⎜ 0 0 1 0 ⎟
⎝ ⎠ ⎜ ⎟
⎝0 0 0 1⎠
⎛ 1 −1 0 0 ⎞
2⎜ ⎟
1⎛ e 2
⎞ ⎜ −1 1 0 0 ⎟
M (90°) = ⎜ ⎟ .
2 ⎝ mc2 R ⎠ ⎜ 0 0 0 0 ⎟
⎜ ⎟
⎝ 0 0 0 0⎠
Viewing the scattered radiation along the z-axis (θ = 0°)
the polarization state of the incident beam remains
unchanged. Viewing the scattered radiation perpendicular
to the z-axis (θ = 90°) the Mueller matrix takes the form of
a linear vertical polarizer.
At any viewing angle the observed intensity of the
scattered radiation is
1
[S0 (1 + cos2 θ) − S1 sin2 θ].
I ( θ) =
2
Here, the constant factor has been dropped. For incident
linearly vertically polarized light, LVP, unpolarized light,
UNP, and linearly horizontally polarized light, LHP, the
intensities of the scattered beam are
I LVP ( θ) = 1,
1
I ( θ)UNP = (1 + cos2 θ),
2
I ( θ)LHP = cos2 θ.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Radiation Field 103

Optical Scattering (cont’d)

In the figure, intensity polar

θ plots of the above equations are
shown. The innermost plot
corresponds to LHP light and
corresponds to dipole radiation.
The next plot is for unpolarized
light and corresponds to
“peanut” radiation. Finally, the
outermost plot corresponds to LVP radiation and is a
circle.
For incident unpolarized light the DOP for the scattered
light is
S′ sin 2 θ
P= 1 = .
S0′ 1 + cos2 θ
DOP

Thus, unpolarized light

becomes partially po-
larized for 0° < θ < 90°
Scattering angle θ and completely polar-
ized for θ = 90°.
The Rayleigh scattering phenomenon is scattering by a
bound charge. The vector equation of motion of the charge
and it components are
m 
r(t ) + kr(t ) = −e E(t ),

e
Transmittance

x(t ) + ω02 x (t ) = − Ex (t ),
m

e
y(t ) + ω02 y(t ) = −
 E y (t )
Rotation angle θ m
z(t ) + ω0 2 z(t ) = 0,
where ω0 = k / m . The last equation is ignored since it does
not contribute to the reradiated field.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 104 Polarization

Optical Scattering (cont’d)

The solutions of the above equations are

−e ω2
x(t ) = Ex (t )
m( ω2 − ω 0 2 )
−e ω2
y(t ) =
 E y (t ).
m( ω2 − ω02 )

The Stokes vector S′ of the scattered beam for an incident

beam with an arbitrary Stokes vector for Rayleigh
scattering is
⎛ S0 (1 + cos2 θ) − S1 sin 2 θ ⎞
2 ⎜ ⎟
1⎡ e2 ⎤ 4 ⎜ −S0 sin 2 θ + S1 (1 + cos2 θ) ⎟
S′ = ⎢ ⎥ ω .
2 ⎣ mc 2 ( ω2 − ω 0 2 ) ⎦ ⎜ 2S2 cos θ ⎟
⎜⎜ ⎟⎟
⎝ 2S3 cos θ ⎠
For ω0 = 0 the Stokes vector reduces to scattering by a free
electron. However, for the bound electron model the
scattering intensity is inversely proportional to the fourth
power of the wavelength since ω4 = (2πc/λ)4. Thus, as the
wavelength of light decreases from the red region to the
blue region of the optical spectrum, the intensity of the
scattered light increases; this explains why the sky is
blue.
Sunlight is unpolarized and perpendicular to its direction.
The scattered light is linearly polarized. The observed
intensity varies by using a rotating.
Intensity

Analyzer rotation angle (deg)

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Optics of Metals and Semiconductors 105

The Optics of Metals and Semiconductors

Metals and semiconductors are absorbing media and are

very good conductors of electricity and heat (conducting
media), due to the relatively large number of free
electrons in metals whereas in dielectrics there are
relatively few free electrons. Consequently, most metals
and semiconductors are almost always highly reflective
and are practically opaque. Dielectrics almost always
have smaller reflectivities than metals. For metals the
reflection of light from the surface is one of the few ways
to investigate the optical constants ε and κ.
The interaction of polarized light (optical fields) with
metals and semiconductors is described by the relation
j(r,t) = σE(r,t), where j(r,t) is the electric current density,
E(r,t) is the electric field, and σ is the conductivity of the
medium. Maxwell’s equations for the electric field in the
metallic-semiconductor medium are found to be
µ ε ∂ 2 E(r,t ) 4 πµ σ ∂ E(r,t )
∇2 E(r,t ) = + .
c2 ∂ t2 c2 ∂t

The term ∂E(r,t)/∂t corresponds to an attenuated (or

damped) wave. The terms can be regrouped and lead to a
dielectric constant for metals:
⎛ 4 πσ ⎞
ε = ε − i⎜ ⎟,
⎝ ω ⎠
where ε is a complex dielectric constant, ε is a real
dielectric constant, and ω is the angular frequency of the
incident wave. The refractive index can be as ε = n2. This
relation can be written in terms of a real refractive index
n and attenuation or absorption coefficient κ by a complex
relation n = (1 – iκ).

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 106 Polarization

Refractive Index and Absorption Coefficient

Using ε = n2 the real and imaginary parts of this last

equation yield

⎛ 4 πσ ⎞
ε = n2 = n2 (1 − i κ)2 = ε − i ⎜ ⎟,
⎝ ω ⎠
where
n2 (1 − κ2 ) = ε,

2πσ σ
n2 κ = = ,
ω ν
and n u → ν is the (linear) frequency. These last two
equations can be solve simultaneously and yield

1 ⎡⎢ 2 ⎛ 2σ ⎞ ⎤ 1 ⎡⎢ 2 ⎛ 2σ ⎞ ⎤
2 2

n2 = ε +⎜ ⎟ + ε ⎥ , and n 2 2
κ = ε + ⎜ ⎟ − ε ⎥.
2⎢ ⎝ ν ⎠ ⎥ 2⎢ ⎝ ν ⎠ ⎥
⎣ ⎦ ⎣ ⎦
From these two parameters ε and σ can then be found
from the equations on the previous page, which also serve
as a check on the measurements.
The equations for the s- and p-components of the reflected
intensity are
2 2
R Rp
Rs = s , and Rp = .
Es Ep

At normal incidence the s- and p-components of the

reflected intensity are equal and are

⎡ (n − 1) 2 + (nκ ) 2 ⎤
Rs = Rp = ⎢ 2 ⎥
.
⎣ (n + 1) + ( nκ ) ⎦
2

For κ = 0, this equation reduces to the familiar equation

for the reflectivity from a dielectric at normal incidence.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Optics of Metals and Semiconductors 107

Incidence Angle Reflectivity

The previous equation for n = 1.0, 1.5, 2.0, and 2.5 plotted
over a range of κ from 0 to 10 is shown in the graph.
Normal reflectivity

n = 2.5

n = 1.0

Absorption coefficient κ

For absorbing media with increasing κ (metals) the

reflectivity approaches unity. Thus, highly reflecting
metals are characterized by high values of κ.
For non-normal incidence the s- and p-intensity
components are
2
Rs ⎡ (n − cos i )2 + (nκ)2 ⎤
Rs = =⎢ 2 2 ⎥
,
Es ⎣ (n + cos i ) + (nκ) ⎦
2
Rp ⎡ (n − 1/cos i )2 + (nκ )2 ⎤
Rp = =⎢ 2 2 ⎥
.
Ep ⎣ (n + 1/cos i ) + (nκ) ⎦

For i = 0° the above equations reduce to the normal

incidence equation shown above.
The values of n and κ are listed for λ = 0.60 µm.

Metal n κ
Gold (Au) 0.47 6.02
Silver (Ag) 0.20 17.2
Copper (Cu) 0.62 4.15
Aluminum (Al) 1.44 3.63

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 108 Polarization

Incidence Angle Reflectivity (cont’d)

The reflection equations for the s- and p-components are

plotted below using the values for n and κ in the table.

Ag
Reflectivity s

Al
Au

Cu

Incidence angle i

Ag
Reflectivity p

Al Au

Cu

Incidence angle i

For the s-component the reflectivity has a pseudo-

Brewster angle minimum value, because, unlike the
Brewster angle behavior for dielectrics, the intensity does
not go to zero for metals.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Optics of Metals and Semiconductors 109

Complex Reflection Coefficients

The effect of κ on the reflected p- and s-components are

shown for n = 1.5 and κ = 0, 1…4.

n = 1.5
κ = 4.0
Reflectivity Rp

κ = 0.0
Incidence angle i

κ = 4.0
Reflectivity Rs

n = 1.5

κ = 0.0
Incidence angle i

For the p-component, plot κ = 0 corresponds to the

Brewster angle null condition for a dielectric.
To deal with light of any polarization state it is necessary
to determine the Mueller matrix for metals. Since a
complex refractive index is required to describe metals
and semiconductors, the reflection coefficients (complex)
are defined to be
rp = Rp / E p = ρ p exp(i φ p ),

rs = Rs / Es = ρs exp(i φs ),

where ρs,p and φs,p are the absolute values of the reflection
coefficients and the phases, respectively.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 110 Polarization

The Principal Angle of Incidence Measurement

The reflection coefficients describing metals and

semiconductors can be immediately used to determine the
form of the Mueller matrix for reflection from
metals:
 ρs 2 + ρ p2 ρs 2 − ρ p 2 0 0 
 2 
1  ρs − ρ p2 ρs 2 + ρ p 2 0 0 ,
M = 
2 0 0 2ρs ρ p cos Δ −2ρs ρ p sin Δ 
 
 0 0 2ρs ρ p sin Δ 2ρs ρ p cos Δ 

where Δ = φs – φp.
A measurement method for measuring n and κ known as
the principal angle of incidence measurement.

D
Laser
Analyzer
L+45
polarizer λ/4 wave plate
Incident i i Reflected
beam beam

Metallic sample

The measurement configuration is identical to that for the

Brewster angle measurement except that a λ/4 wave plate
is now in the reflection arm. The measurement method is
very similar to that used in ellipsometry, a method that
determines not only the optical constants but the
thickness of thin films as well.
Incident L+45P polarized light is reflected from the metal
sample. In general, the reflected light is elliptically
polarized so the corresponding polarization ellipse is in a
nonstandard form. The angle of the incident beam is
changed until the phase becomes 90° and is observed in
the reflected beam. The incident angle where this takes
places is the principal angle of incidence i .

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Optics of Metals and Semiconductors 111

The Principal Angle of Incidence Measurement

(cont’d)

The polarization ellipse is in its standard form, and the

orthogonal field components are parallel and
perpendicular to the plane of incidence. The reflected
beam then proceeds through the QWP and the beam of
light that emerges is linearly polarized with its azimuthal
angle at an unknown angle. The beam is incident on an
analyzing polarizer which is rotated until a null intensity
is found. This null takes place at the principal azimuthal
angle ψ . From the measurement of the principal angle of
incidence and the principal azimuth angle the optical
constants n and κ are determined.
For the incident L+45P light that is reflected from the
metal surface and using the above Mueller matrix the
Stokes vector of the reflected beam is
⎡ 1 + P2 ⎤
⎢ ⎥
⎢ −(1 − P 2 ) ⎥
′
S = I0 ,
⎢ 2P cos ∆ ⎥
⎢ ⎥
⎣⎢ 2P sin ∆ ⎦⎥
where P = ρs / ρp = tan ψ , and ψ is the principal
azimuthal angle. The reflected beam passes through the
QWP where the elliptically polarized beam is transformed
to its standard form, by changing the incident angle until
∆ becomes 90°.
The Stokes vector becomes
⎡ 1 + P2 ⎤
⎢ ⎥
−(1 − P 2 ) ⎥
S ′ = I0 ⎢ ,
⎢ 0 ⎥
⎢ ⎥
⎢⎣ 2P ⎥⎦

which describes the polarization ellipse in its standard

form.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 112 Polarization

The Principal Angle of Incidence Measurement

(cont’d)

χ and ψ of the polarization ellipse are

1 ⎛S ′ ⎞ 1 ⎛ 2P ⎞
χ= sin −1 ⎜ 3 ⎟ = sin −1 ⎜ ⎟ , and
2 ⎜S ′ ⎟ 2 ⎝ 1 + p2 ⎠
⎝ 0 ⎠

1 ⎛S ′ ⎞
ψ= tan −1 ⎜ 2 ⎟ = 0,
2 ⎜S′ ⎟
⎝ 1 ⎠
so the orientation angle of the polarization ellipse is 0, as
required. The QWP now transforms this Stokes vector to
linear polarized light:
⎛1 0 0 0 ⎞⎛ 1 + P2 ⎞ ⎛ 1 + P2 ⎞
⎜ ⎟⎜ 2 ⎟ ⎜ ⎟
0 1 0 0 ⎟ ⎜ −(1 − P ) ⎟ −(1 − P 2 ) ⎟
S ′ = I0 ⎜ = I0 ⎜ .
⎜0 0 0 −1 ⎟ ⎜ 0 ⎟ ⎜ 2P ⎟
⎜ ⎟⎜ ⎟⎟ ⎜⎜ ⎟⎟
⎝0 0 1 0 ⎠ ⎜⎝ 2P ⎠ ⎝ 0 ⎠
The final step is to relate the principle angle of incidence
i (∆ = π / 2 ) and the principal azimuth angle ψ to the
optical constants n and κ. By rotating the analyzer until a
null intensity is obtained the value of P is determined and
is expressed as
1 + jP − sin i tan i
= ,
1 − jP n(1 − j κ )
where j = −1. Multiplying this equation by its complex
conjugate then yields
n 1 + κ2 = sin i tan i .

From this equation and using the relation P = tan ψ it is

straightforward to determine the refractive index n and
the absorption coefficient κ in terms of i and ψ . The
optical constants of the metal are then found to be

n = − sin i tan i cos2ψ,

and

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 The Optics of Metals and Semiconductors 113

The Principal Angle of Incidence Measurement

(cont’d)

κ = tan 2ψ,

where i and ψ are the principal angles of incidence and

azimuth, respectively. The above two equations can be
written as a single equation in terms of the complex
refractive index:

Example: The principal angles of incidence and

azimuth for Au are found to be i = 108.3° and
ψ = 40.3° at a wavelength of 6000 Å (1 Å = 10–8 cm).

n = − sin(108.3°) tan(108.3°) cos(80.6°) = 0.477,

and
κ = tan(80.6°) = 6.17.

The dielectric constant ε and conductivity σ for Au can

be determined at a frequency of ν = 0.5 × 1014 Hz
(which corresponds to 6000 Å):
ε = n2 (1 − κ2 ) = −8.46,
and
σ = n2 κν = 7.05 × 1014 sec-1.

For metals, negative values of ε can occur.

n = n (1 − j κ ) = − sin i tan i exp( − j 2 ψ ).

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 114 Polarization

Equation Summary

General equations
Malus’ Law and Brewster’s Law:
I ( θ) = I 0 cos2 θ tan i = n2 / n1 = n

Field components propagating in the z direction:

Ex ( z ,t ) = E0 x cos( ωt − kz + δ x )

E y ( z ,t ) = E0 y cos( ωt − kz + δ y )

Equation of the polarization ellipse:

2
E x ( z , t )2 E y ( z , t ) 2E x ( z , t ) E y ( z , t )
2
+ 2
− cos δ = sin 2 δ
E0 x E0 y E0 x E0 y

Parameters of the polarization ellipse:

2E 0 x E 0 y
tan 2ψ = cos δ 0 ≤ψ ≤ π
E02x − E02y

2E0 x E0 y sin δ
sin 2χ = −π / 4 < χ ≤ π / 4
E02x + E02y

Auxilliary angle definition and polarization

parameters:
E0 y
tan α = , 0 ≤ α ≤ π /2
E0 x

tan 2ψ = (tan 2α ) cos δ

sin 2χ = (sin 2α )sin δ 0 ≤ δ < 2 π

Stokes polarization parameters:

S0 = E02x + E02y

S1 = E02x − E02y

S2 = 2E0 x E0 y cos δ

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Appendix 115
Equation Summary

Stokes polarization parameters (cont’d):

S3 = 2E0 x E0 y sin δ , δ = δ y − δ x

S02 = S12 + S22 + S32

Stokes vector:

⎛ S0 ⎞ ⎛ E02x + E02y ⎞
⎜ ⎟ ⎜ ⎟
S E02x − E02y ⎟
S = ⎜ 1 ⎟ = ⎜⎜
⎜ S2 ⎟ 2E0 x E0 y cos δ ⎟
⎜⎜ ⎟⎟ ⎜⎜ ⎟
⎟
⎝ S3 ⎠ ⎝ 2E0 x E0 y sin δ ⎠

Stokes parameters in complex notation:

Ex (t ) = E0 x exp(i δx ) , E y (t ) = E0 y exp(i δ y )

S0 = Ex Ex * + E y E y *

S1 = Ex Ex * − E y E y *

S2 = Ex E y * + E y Ex *

(
S3 = i Ex E y * − E y Ex * )
Stokes vector on the Poincaré sphere:

⎛ S0 ⎞ ⎡ 1 ⎤
⎜ ⎟ ⎢ ⎥
S ⎢ cos(2χ ) cos(2ψ ) ⎥
S =⎜ 1⎟=
⎜ S2 ⎟ ⎢ cos(2χ)sin(2ψ ) ⎥
⎜⎜ S ⎟⎟ ⎢
sin(2χ )
⎥
⎝ 3⎠ ⎣ ⎦

1 ⎛S ⎞ 1 ⎛S ⎞ π π
ψ= tan −1 ⎜ 2 ⎟ 0 ≤ ψ ≤ π, χ = sin −1 ⎜ 3 ⎟ − ≤ χ ≤
2 S
⎝ 1⎠ 2 S
⎝ 0⎠ 4 4

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 116 Polarization

Equation Summary

Stokes vector on the observable polarization

sphere:
⎛ S0 ⎞ ⎛ 1 ⎞
⎜ ⎟ ⎜ ⎟
⎜ S1 ⎟ ⎜ cos 2α ⎟
S= =
⎜ S2 ⎟ ⎜ sin 2α cos δ ⎟
⎜⎜ ⎟⎟ ⎜ ⎟
⎝ S3 ⎠ ⎝ sin 2α sin δ ⎠
Angle relation of 2α and δ to the Stokes parameters:
⎛S ⎞
2α = cos−1 ⎜ 1 ⎟ 0 ≤ 2α < π
⎝ S0 ⎠
⎛S ⎞
δ = tan −1 ⎜ 3 ⎟ 0 ≤ δ < 2π
⎝ S2 ⎠
Ellipticity and orientation angles relations:
tan(2ψ ) = tan(2α ) cos δ sin(2χ ) = sin(2α )sin δ
cos(2α ) = cos(2χ) cos(2ψ ) cot δ = cot(2χ)sin(2ψ )
Stokes vectors for unpolarized and partially
polarized light:
⎛1 ⎞
⎜ ⎟
0
SUNP = S0 ⎜ ⎟
⎜0⎟
⎜ ⎟
⎝0⎠
⎛ S0 ⎞ ⎛ S0 ⎞ ⎛ S0 ⎞
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎜ S1 ⎟ 0⎟ S
S= = (1 − P ) ⎜ + P⎜ 1 ⎟ 0 ≤ P ≤1
⎜ S2 ⎟ ⎜0⎟ ⎜ S2 ⎟
⎜⎜ ⎟⎟ ⎜ ⎟ ⎜⎜ ⎟⎟
⎝ S3 ⎠ ⎝0⎠ ⎝ S3 ⎠
Definition of the degree of polarization P:
I pol S12 + S22 + S32
P= = 0 ≤ P ≤1
I tot S0

S02 ≥ S12 + S22 + S32

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Appendix 117
Equation Summary

Mueller matrices
Mathematical form of the Mueller matrix:
⎛ m00 m01 m02 m03 ⎞
⎜ ⎟
m m11 m12 m13 ⎟
M = ⎜ 10
⎜ m20 m21 m22 m23 ⎟
⎜⎜ ⎟
⎝ m30 m31 m32 m33 ⎟⎠

Mueller matrix for a polarizer:

⎛ px 2 + py 2 px 2 − py 2 0 0 ⎞
⎜ 2 ⎟
1 ⎜ px − py 2 px 2 + py2 0 0 ⎟
M POL ( px , py ) =
2⎜ 0 0 2 px py 0 ⎟
⎜ ⎟
⎜ 0 0 0 2 px py ⎟⎠
⎝
0 ≤ px ≤1 0 ≤ py ≤ 1
Mueller matrices for a linear horizontal and
vertical polarizer:
⎛1 1 0 0⎞ ⎛ 1 −1 0 0⎞
⎜ ⎟ ⎜ ⎟
1 1 1 0 0⎟ 1 −1 1 0 0⎟
M POL = ⎜ M POL = ⎜
2 ⎜0 0 0 0⎟ 2⎜ 0 0 0 0⎟
⎜ ⎟ ⎜ ⎟
⎝0 0 0 0⎠ ⎝0 0 0 0⎠
Mueller matrix for a linear polarizer with its
transmission axis at +45°:
⎛1 0 1 0⎞
⎜ ⎟
1 ⎜0 0 0 0⎟
M POL =
2 ⎜1 0 1 0⎟
⎜ ⎟
⎝0 0 0 0⎠
Mueller matrix for a wave plate:
⎛1 0 0 0 ⎞
⎜ ⎟
0 1 0 0 ⎟ 2π
M WP ( φ) = ⎜ φ= (no − ne )d
⎜0 0 cos φ − sin φ ⎟ λ
⎜ ⎟
⎝0 0 sin φ cos φ ⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 118 Polarization

Equation Summary

Mueller matrices for a quarter- and half-wave plate:

⎛1 0 0 0⎞ ⎛1 0 0 0⎞
⎜ ⎟ ⎜ ⎟
0 1 0 0⎟ 0 1 0 0⎟
M QWP =⎜ M HWP =⎜
⎜0 0 0 −1 ⎟ ⎜0 0 −1 0 ⎟
⎜ ⎟ ⎜ ⎟
⎝0 0 1 0⎠ ⎝0 0 0 −1 ⎠
Mueller matrix for a rotator:
⎛1 0 0 0⎞
⎜ ⎟
0 cos 2θ sin 2θ 0⎟
M ROT ( θ) = ⎜
⎜ 0 − sin 2θ cos 2θ 0⎟
⎜ ⎟
⎝0 0 0 1⎠
Mueller matrix for a rotated polarizing element:
M ( θ) = M ROT ( −2θ) ⋅ M ⋅ M ROT (2θ)
Mueller matrix for a rotated linear polarizer:
⎛ 1 cos2θ sin 2θ 0⎞
⎜ 2 ⎟
1 cos 2θ cos 2θ sin 2θ cos2θ 0⎟
M POL ( θ) = ⎜
2 ⎜ sin 2θ sin 2θ cos 2θ sin 2 2θ 0⎟
⎜ ⎟
⎝ 0 0 0 0⎠
Mueller matrix of a circular polarizer:
⎛1 0 1 0⎞
⎜ ⎟
0 0 0 0⎟
M CP =⎜
⎜0 0 0 0⎟
⎜ ⎟
⎝1 0 1 0⎠
Reflection and transmission
Fresnel’s Equations for reflection and refraction:
tan(i − r ) sin(i − r )
Rp = Ep Rs = − Es
tan(i + r ) sin(i + r )
2sin r cos i 2sin r cos i
Tp = E p Ts = Es
sin(i + r ) cos(i − r ) sin(i + r )

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Appendix 119
Equation Summary

Stokes parameters for reflection:

S0 R = fR [(cos2 α − + cos2 α + )S0 + (cos2 α − − cos2 α + )S1 ]
S1 R = fR [(cos2 α − − cos2 α + )S0 + (cos2 α − + cos2 α + )S1 ]
S2 R = − fR (2cos α − cos α + )S2
2
1 ⎛ tan α − ⎞
S3 R = − fR (2 cos α − cos α + )S3 fR = ⎜ ⎟
2 ⎝ sin α + ⎠
Stokes parameters for transmission (refraction):
S0T = fT [(cos2 α − + 1)S0 + (cos2 α − − 1)S1 ]

S1T = fT [(cos2 α − − 1)S0 + (cos2 α − + 1)S1 ]

S2T = − fT (2cos α − )S2
1 sin 2i sin 2r
S3T = − fT (2cos α − )S3 fT =
2 (sin α + cos α − )2
Mueller matrices for reflection and transmission:
⎛1 0 0 0⎞ ⎛1 0 0 0⎞
⎜ 2 ⎟ ⎜ ⎟
⎛ n −1 ⎞ ⎜0 1 0 0⎟ 4n ⎜0 1 0 0⎟
MR = ⎜ ⎟ MT =
⎝ n +1 ⎠ ⎜0 0 −1 0 ⎟ ( n + 1)
2 ⎜0 0 1 0⎟
⎜ ⎟ ⎜ ⎟
⎝0 0 0 −1 ⎠ ⎝0 0 0 1⎠
Mueller matrices for Brewster angle reflection and
transmission:
⎛1 1 0 0⎞
⎜ ⎟
1 1 1 0 0⎟
M R ,B = cos2 2iB ⎜
2 ⎜0 0 0 0⎟
⎜ ⎟
⎝0 0 0 0⎠

⎛ sin 2 2iB + 1 sin 2 2iB − 1 0 0 ⎞

⎜ 2 2 ⎟
1 sin 2iB − 1 sin 2iB + 1 0 0
M T ,B = ⎜ ⎟
2 ⎜ 0 0 2sin 2iB 0 ⎟
⎜⎜ ⎟
⎝ 0 0 0 2sin 2iB ⎟⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 120 Polarization

Equation Summary

Mueller matrix for total internal reflection:

⎛1 0 0 0 ⎞
⎜ ⎟ 2 2
0 1 0 0 ⎟ δ cos i n sin i − 1
MR = ⎜ tan =
⎜0 0 cos δ − sin δ ⎟ 2 n sin 2 i
⎜ ⎟
⎝0 0 sin δ cos δ ⎠

Mueller matrix of the Fresnel rhomb:

⎛1 0 0 0 ⎞
⎜ ⎟
0 1 0 0
M =⎜ ⎟
⎜0 0 cos( δU + δ L ) − sin( δU + δ L ) ⎟
⎜⎜ ⎟
⎝0 0 sin( δU + δL ) cos( δU + δL ) ⎟⎠

Jones matrix calculus

Jones vector:

⎛ E x ⎞ ⎛ E0 x e x ⎞
iδ

E = ⎜⎜ ⎟⎟ = ⎜ ⎟
⎝ E y ⎠ ⎜⎝ E0 y e ⎟
i δy
⎠
Jones matrix formulation of optical intensity:
⎛ Ex ⎞
I = E† ⋅ E I = ( Ex * E y * ) ⎜⎜ ⎟⎟ = Ex Ex * + E y E y *
⎝ Ey ⎠
Jones matrix definition:
⎛ jxx jxy ⎞
J = ⎜⎜ ⎟
⎝ jyx jyy ⎟⎠

Jones matrices for linear polarizers:

⎛ px 0⎞
JPOL = ⎜⎜ ⎟ 0 ≤ px , py ≤ 1
⎝0 py ⎟⎠

⎛1 0 ⎞ ⎛0 0⎞
JLHP = ⎜ ⎟ JLVP = ⎜ ⎟
⎝0 0⎠ ⎝0 1 ⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Appendix 121
Equation Summary

Jones matrices for a wave plate, quarter-wave

plate, and a half-wave plate:
⎛ e i ( φ / 2) 0 ⎞ ⎛1 0 ⎞
JWP = ⎜ ⎟→⎜ −i φ ⎟
⎠ ⎝0 e ⎠
− i ( φ / 2)
⎝ 0 e

⎛1 0 ⎞ ⎛1 0 ⎞
JQWP = ⎜ ⎟ JHWP = ⎜ ⎟
⎝ 0 −i ⎠ ⎝ 0 −1 ⎠
Jones matrix for a rotator:
⎛ cos θ sin θ ⎞
JROT ( θ) = ⎜ ⎟
⎝ − sin θ cos θ ⎠
Jones matrix for a rotated polarizing element:
J( θ) = JROT ( −θ) ⋅ J ⋅ JROT ( θ)

Polarizer characterization
Transmittance of a single linear polarizer:
k1 + k2 k1 k2
TUNP = TLHP = TLVP =
2 2 2
Transmittance of a pair of linear polarizers:
1 2 k12 k22
TUNP = H 0 = ( k1 + k22 ) TLHP = TLVP =
2 2 2
Transmittance of a pair of crossed polarizers:
TUNP = H 90 = k1 k2
Transmittance for a single rotated polarizer at an
angle θ:
T ( θ) = k1 sin 2 θ + k2 cos2 θ
Transmittance for a fixed and rotated polarizer at
an angle θ:
1
T ( θ) = k1 k2 sin 2 θ + ( k12 + k22 ) cos2 θ
2
= H 90 + ( H 0 − H 90 ) cos2 θ

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 122 Polarization

Equation Summary

Logarithmic definition of transmittance (density):

D0 = − log 10 H 0 D90 = − log 10 H 90
Contrast ratio of a polarizer:
Tmax − Tmin k1 − k2
C= =
Tmax + Tmin k1 + k2
Kerr and Pockel cell characterization
Birefringence of the Kerr effect:
np − ns = λBE 2

Pockel cell phase shift:

ωn03 r63
φ = φ y − φx = V
c
Stokes vectors for radiating charges
Stokes vector for a linearly oscillating charge:
⎛1 ⎞
2 ⎜ ⎟
⎛ e z0 ⎞ 1
S = ⎜ 2 ⎟ sin 2 θ ω0 4 ⎜ ⎟
⎝c R⎠ ⎜0⎟
⎜ ⎟
⎝0⎠
Stokes vector for a randomly oriented oscillating
charge:
⎛1 ⎞
⎜ ⎟2
8π ⎛ eA ⎞ 4⎜ 0 ⎟
S= ω
3 ⎜⎝ c2 R ⎟⎠ ⎜0⎟
0

⎜ ⎟
⎝0⎠
Stokes vector for a charge moving in a circle:
⎛1 + cos 2 θ ⎞
2 ⎜ 2 ⎟
⎛ eA ⎞ 4 ⎜ sin θ ⎟
S = ⎜ 2 ⎟ ω0
⎝c R⎠ ⎜ 0 ⎟
⎜⎜ ⎟⎟
⎝ 2 cos θ ⎠

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Appendix 123
Equation Summary

Stokes vector for the Zeeman effect:

⎡ ⎛1 + cos2 θ ⎞ ⎛ sin 2 θ ⎞ ⎛1 + cos2 θ ⎞ ⎤
⎢ ⎜ 2 ⎟ ⎜ 2 ⎟ ⎜ ⎟⎥
2 ⎛ eA ⎞ 2
− sin θ ⎟ sin θ ⎟ − sin 2 θ ⎟ ⎥
S = ⎜ 2 ⎟ ⎢ ω− 4 ⎜ + 2ω04 ⎜ + ω+ 4 ⎜
3 ⎝ 2c R ⎠ ⎢ ⎜ 0 ⎟ ⎜ 0 ⎟ ⎜ 0 ⎟⎥
⎢ ⎜⎜ ⎟⎟ ⎜⎜ ⎟⎟ ⎜⎜ ⎟⎟ ⎥
⎣⎢ ⎝ −2 cos θ ⎠ ⎝ 0 ⎠ ⎝ 2 cos θ ⎠ ⎦⎥
Mueller matrix for Thomson scattering:
⎛1 + cos2 θ − sin 2 θ 0 0 ⎞
⎜ 2 ⎟
1⎛ e ⎞ ⎜ − sin θ 1 + cos θ
22 2
0 0 ⎟
M = ⎜ ⎟
2 ⎝ mc2 R ⎠ ⎜ 0 0 2cos θ 0 ⎟
⎜⎜ ⎟
⎝ 0 0 0 2cos θ ⎟⎠
Stokes vector for Rayleigh scattering:
⎛ S0 (1 + cos2 θ) − S1 sin 2 θ ⎞
⎜ 2 ⎟
1⎡ e2 ⎤ 4 ⎜ −S0 sin2 θ + S1 (1 + cos2 θ) ⎟
S' = ⎢ 2 2 ⎥ ω
2 ⎣ mc ( ω − ω 02 ) ⎦ ⎜ 2S2 cos θ ⎟
⎜⎜ ⎟⎟
⎝ 2S3 cos θ ⎠
Absorbing media––semiconductors and metals
Complex dielectric constant and refractive index
for metals:
⎛ 4 πσ ⎞
ε = ε − i⎜ ⎟ n = n (1 − i κ )
⎝ ω ⎠
Reflectivity at normal incidence for absorbing
media:
⎡ (n − 1)2 + (nκ )2 ⎤
Rs = Rp = ⎢ 2 2 ⎥
⎣ (n + 1) + (nκ ) ⎦
s- and p- reflectivity at non-normal incidence:
2
Rs ⎡ (n − cos i )2 + (nκ)2 ⎤
Rs = =⎢ 2 2 ⎥
Es ⎣ (n + cos i ) + (nκ) ⎦
2
Rp ⎡ (n − 1/cos i )2 + (nκ )2 ⎤
Rp = =⎢ 2 2 ⎥
Ep ⎣ (n + 1/cos i ) + (nκ) ⎦

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 128 Polarization

Bibliography

Azzam, R.M.A., and Bashara, N.M., Ellipsometry and

Polarized Light, North-Holland, Amsterdam, 1977.

Billings, B.H., Selected Papers on Polarization, SPIE

Press, Bellingham, WA, 1990.

Billings, B.H., Selected Papers on Applications of

Polarized Light, SPIE Press, Bellingham, WA, 1992.

Born, M., and Wolf, E., Principles of Optics, 7th ed.,

Cambridge Univ. Press, New York, 1999.

Brosseau, C., Fundamentals of Polarized Light, Wiley

Interscience, New York, 1998.

Clarke, D., and Grainger, J.F., Polarized Light and

Optical Measurement, Pergamon Press, Oxford, 1971.

Collett, E., Polarized Light: Fundamentals and

Applications, Marcel Dekker, New York, 1992.

Collett, E., Polarized Light in Fiber Optics, The PolaWave

Group, Lincroft, NJ, and SPIE Press, Bellingham, WA,
2005.

Gerrard, A., and Burch, J.M., Introduction to Matrix

Methods in Optics, John Wiley & Sons, New York, 1975.

Hecht, E., Optics, 4th ed., Addison-Wesley Publishing Co.,

Reading, Mass., 2001.

Huard, S., Polarization of Light, J. Wiley, New York,

1997.

Jackson, J.D., Classical Electrodynamics, 3rd ed., J. Wiley

& Sons, New York, 1998.

Jenkins, F.A., and White, H.E., Fundamentals of Optics,

McGraw-Hill Book Co., New York, 1957.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Appendix 129
Bibliography

Kliger, D.S., Lewis, J. W., and Randall, C. E., Polarized

Light in Optics and Spectroscopy, Academic Press, New
York, 1990.

Lakhtakia, A., Selected Papers on Natural Optical Activity

SPIE Press, Bellingham, WA, 1990.

Pipes, L.A., Matrix Methods for Engineering, Prentice-

Hall, Englewood Cliffs, New Jersey, 1963.

Poincaré, H., Théorie Mathématique de la Lumière, Vol.2,

George Carré, Paris, 1892.

Shurcliff, W.A., Polarized Light: Production and Use,

Harvard University Press, Cambridge, Mass., 1962.

Shurcliff, W.A., and Ballard, S.S., Polarized Light, D. Van

Nostrand Co., Princeton, New Jersey, 1964.

Strong, J.M., Concepts in Classical Optics, Freeman, San

Francisco, 1958.

Stratton, J.A., Electromagnetic Theory, McGraw-Hill Book

Co., New York, 1941.

Stone, J., Radiation and Optics, McGraw-Hill Book Co.,

New York, 1963.

Taylor, A.E., Calculus with Analytic Geometry, Prentice

Hall, Englewood Cliffs, New Jersey, 1959.

Wahlstrom, E.E., Optical Crystallography, 3rd ed., John

Wiley & Sons, New York, 1960.

Wood, R.W., Physical Optics, 3rd ed., Optical Society of

America, Washington, DC, 1988.

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Index

45° incidence angle commercial linear

measurement, 89 polarizers, 78
complex dielectric constant,
accelerating charges, 91, 94 105
Al-Hazen, 1
amplitude modulation complex notation, 13, 93
voltage, 87 constitutive equations, 91
Arago, 5 contrast ratio, 80
auxiliary angle, 9 crossed polarizers, 19, 25,
60
Babinet-Soleil compensator, current density, 91, 105
35 cyclotron frequency, 97
Bartholinus, 2
biaxial material, 75 degenerate polarization
birefringence, 81, 85 states, 8, 10, 11, 33, 58
birefringent crystal, 40 degree of polarization
bound charge, 93, 94, 103 (DOP), 15, 44, 45, 46, 51,
Brewster, 4 54
Brewster angle, 89, 108, degree of polarization m
109, 110 dielectric plates, 53
Brewster angle degree of polarization of the
measurement, 88 transmitted beam, 45
Brewster’s angle condition, depolarizers, 72
44 dextro-rotary, 65
Brewster’s law, 44, 89 dipole radiator, 94

calcite crystal, 2, 3, 4, 81 electro-optical effect, 73, 85

caustic, 1 elliptically polarized light,
circular birefringence, 65 58, 95
circular polarizer, 8, 36, 38, ellipticity and orientation
61, 66 angles, 21, 37
classical measurement, 27, ellipticity angle, 12
38 equation of an ellipse, 7
classical optical isolator, 66, equation of motion of the
67, 71 bound charge, 98
classical radius of an extraordinary ray (e-ray), 2,
electron, 102 76, 81
classical wave equation, 92
classical Zeeman effect, 98 Faraday isolator, 67, 70
coherent superposition and Faraday rotation, 64
decomposition, 84 Faraday rotator, 64, 69, 71
fast axis, 20, 25, 26, 27, 69
figure 8, 32

130

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Index

focal point, 1 Laplacian operator, 5

Fourier series, 39 Larmor frequency, 71
free electrons, 105 Larmor precession
Fresnel, 5 frequency, 99
Fresnel reflection and levo-rotatory, 65
transmission coefficients, linear polarizer, 8, 18, 19,
55 27, 34, 38, 59, 70, 78, 79,
Fresnel rhomb, 49 83, 89
Fresnel’s equations, 42 liquid crystal displays, 27
Fresnel’s reflection logarithms of
equations, 48 transmittances, 79
Fresnel’s wave theory, 5, 7 Lorentz force equation, 96,
98
Glan-Foucault polarizer, 76 Lorentz-Zeeman model, 99
Glan-Foucault prism, 76
half-wave plate (HWP), 20, m dielectric plates, 54
21, 23, 41, 59, 60, 73, 74 magneto-optical media, 64
half-wave voltage, 86, 87 major principal
H-sheet Polaroid, 77 transmittance, 78, 80
Huygens, 2 Malus, 3
Malus’s Law, 3, 25
ideal linear polarizer, 18, Maxwell’s equations, 91,
19, 23, 30, 31, 80 105
incoherent decomposition, measurement of the
83 refractive index, 88
independent beams, 82, 83 minimum deviation
inverted image, 1 method, 88
isotropic material, 75 minor principal
isotropic medium, 75 transmittance, 78, 80
motion of a charged particle
Jones column matrix, 57 moving in a constant
Jones matrices, 57, 59 magnetic field, 96
Jones matrix calculus, 57, Mueller matrices, 17, 20,
60 22, 25, 26, 27, 46, 55
Jones matrix for a rotated Mueller matrices for
ideal LHP, 60 reflection and
Jones matrix for a rotated transmission, 55
wave plate, 60 Mueller matrix at 45°
Jones matrix for a rotator, incidence angle, 90
59 Mueller matrix equation,
Jones vectors, 57, 58, 59, 84 39, 50, 52
Mueller matrix for a
depolarizer, 72
Kerr effect, 85

131

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Index

Mueller matrix for a linear Mueller matrix of a rotated

polarizer, 18 QWP, 24
Mueller matrix for a Mueller matrix of a rotated
rotated ideal linear wave plate, 23
polarizer, 23 Mueller-Stokes equation for
Mueller matrix for a transmission, 43
rotated linear polarizer, multiple-order wave plate,
36 41
Mueller matrix for a wave
plate, 20 natural frequency of
Mueller matrix for an ideal oscillation of the electron
linear polarizer, 18 in the sodium atom, 98
Mueller matrix for m natural light, 2, 3
dielectric plates, 52 natural optical activity, 63
Mueller matrix for Pockel’s neutral density filter, 70,
electro-optical 80
modulator, 86 normal incidence, 46, 106,
Mueller matrix for 107
reflection, 43 normal incidence reflection,
Mueller matrix for 88
reflection at the
Brewster angle, 88 observable polarization
Mueller matrix for sphere (OPS), 28, 29,
reflection from metals, observables, 12
110 optic axis, 40, 76, 81, 85
Mueller matrix for rotation, optical activity, 64
63 optical isolator, 61, 66
Mueller matrix for the optical shutter, 25, 27
Faraday rotator, 68 optical thickness, 41
Mueller matrix for the ordinary ray (o-ray), 2, 76
Fresnel rhomb, 49 orientation angle, 9, 10, 14,
Mueller matrix for the Lyot 21, 24, 28, 29, 112
crystal depolarizer, 74 orthogonal components, 5,
Mueller matrix for the 15, 16, 20, 81, 82, 84
rotated HWP, 23 orthogonal ellipses, 83
Mueller matrix for orthogonal triad axis, 75
Thomson scattering, 101
Mueller matrix for total parameters of the
internal reflection, 48 polarization ellipse, 9
Mueller matrix for variable partially polarized light, 15,
phase wave plate, 35 57, 81
Mueller matrix of a circular Pauli spin matrices, 62
polarizer, 36

132

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Index

peanut radiation, 97, 103 Rayleigh scattering, 101,

phase shifter, 17 103
phase shifts, 33 rays, 1, 2, 75, 76
pile of polarizing plates, 52, refractive indices, 2, 4, 40,
53 54, 56, 65, 75, 76, 81, 85
Pockel’s effect, 85, 86 reradiated electron, 101
Pockel’s electro-optical retardation, 86
modulator, 86 retarder, 17, 81
Poincaré, 10 Rochon prism, 76
Poincaré sphere, 10, 11, 14, rotated HWP, 23
28 rotated ideal linear
PolaCor, 75 polarizer, 23, 31
polarization, 5 rotated linear polarizer, 36,
polarization ellipse, 7, 8, 9, 80
10, 17, 21, 22, 63, 64, 95, rotated QWP, 24
110, 111 rotating polarizer, 60
polarizer characteristics, 78 rotating QWP, 32, 33, 34,
polarizing materials, 17 39
polarizing prisms, 76 rotation of a QWP, 32, 33,
polaroid, 75, 77 34
polyvinyl waveplate, 81 rotator, 17, 22, 30, 60, 69
potassium dihydrogen scattering, 101
phosphate (KDP), 86 Sénarmont prism, 76
p-polarization state, 2 slow axis, 20
prime meridian, 30 Snell, 1
principal angle of incidence, Snell’s law of refraction, 2,
110 4, 47, 48, 76
principal angle of incidence spherical surface, 1
measurement, 110 s-polarization state, 2, 3
propagator, 6, 7, 13, 63 Stokes parameter, 27, 28,
pseudo-rotator, 24, 60 29
Stokes parameters, 13, 14,
quarter-wave plate (QWP), 15, 16, 27, 28, 38, 39, 43,
, 20, 21, 24, 32, 33, 34, 50, 62, 89, 94
36, 38, 39, 41, 59, 66, 71, Stokes parameters for the
73, 74, 81, 111, 112 radiation field, 92
quartz, 40, 63, 74, 75, 77, Stokes parameters for the
81 reflected field, 42
Stokes parameters for the
radiation equation, 92 transmitted beam, 43
randomly oriented bound Stokes parameters of a
oscillating charges, 94 polarized beam, 16
ray theory of light, 1 Stokes spectral lines for the
Zeeman effect, 99

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms
 Index

Stokes vector, 17, 21, 24, Thomson scattering, 101

25, 30, 32, 33, 34, 35, 36, time average of the
37, 44, 49, 52, 53, 72, 73, polarization ellipse, 12
78, 82, 83, 89, 93, 94, 95, total internal reflection
98, 99, 100, 101, 104, 111 (TIR), 48
Stokes vector for a bound transmittance for a single
chare, 93 rotated polarizer, 79
Stokes vector for a charge transmittance of a polarizer
moving in a circular pair, 79
path, 94 transmittances, 70
Stokes vector for a charge transmittances for parallel
moving in a constant polarizers, 79
magnetic field, 97 transverse field
Stokes vector for elliptically components, 86, 92
polarized light, 15
Stokes vector for elliptically uniaxial crystal, 40, 75, 76,
polarized light, 13 85
Stokes vector for Rayleigh uniaxial material, 75
scattering, 104 unpolarized light, 15, 44,
Stokes vector for the 45, 46, 50, 52, 53, 78, 82,
degenerate polarization 84, 94, 102, 103
states, 14 unpolarized light, 78, 83
Stokes vector for
unpolarized light, 15, 82, variable-phase wave plate,
94 26
Stokes vector of a variable-phase wave plates,
transmitted beam, 45 72
superposition and vectorial representation for
decomposition of optical elliptically polarized
beams, 82 light, 84
superposition of two wave equations, 5, 6
circularly polarized wave plates, 16, 17, 20, 25,
waves, 65 41, 72, 73, 81
synthetic polarizers, 77 Wolf’s coherency matrix
calculus, 62
Wollaston prism, 76

134

Downloaded from SPIE Digital Library on 11 Jul 2012 to 140.112.6.4. Terms of Use: http://spiedl.org/terms