Berreman4x4

Olivier Castany(a)
Department of Optics, Telecom Bretagne, 29238 Brest, France
June 16, 2013

Abstract
Electromagnetic plane wave propagation in stratified anisotropic media was described by Berreman with the
use of 44 matrices. Berreman4x4 is a numerical implementation of the method in Python. Examples of
applications are ellipsometry analysis, design of Bragg mirrors or study of twisted liquid crystal structures.

(a)

Electronic mail: olivier.castany@telecom-bretagne.eu

Description of Berreman4x4
Olivier Castany
Er

Electromagnetic wave propagation in stratified media is

important in several applications like ellipsometry analysis, Bragg mirrors or twisted liquid crystal structures.
Propagation of plane waves in isotropic media can be
solved with 22 propagation matrix methods 1 . In the
case of anisotropic layers, a 44 propagation matrix
method was developed 2,3 and is now known as Berremans method.
The present work is an open-source implementation of
the method in Python. This programming language was
chosen for ease of use, readability of the source code,
portability and availability of scientific libraries (NumPy
and SciPy). A drawback is slow speed, because the language is interpreted1 .
Calculations are based on articles from Berreman 3 and
Schubert 4 . Application to ellipsometry is based on Fujiwaras book 5 . General references for optics in anisotropic
media are found from Born and Wolf 1 , and Jackson 6 .

Front
half-space

Ei

Ex

Et

Back half-space,
anisotropic

s
Sample

FIG. 1: Geometry of the sample, with input and output

plane waves.

tion point towards the sample in the front and exit half
spaces2 . In our work, we used a different convention, in
which the base electric vectors point towards the x direction in both half-spaces. With this convention, the
base electric vectors for the incident, reflected and transmitted waves are all in the same direction in the case of
reflection with normal incidence.
Gaussian units are used in the detail of the calculation. This makes it easier to compare our results with
past literature, which uses Gaussian units for the most
part. Of course, reflection and transmission coefficients
have no unit and do not depend on the choice of unit.
Conversion between units is presented in the appendix on
electromagnetic units in reference 6. In Gaussian units,
Maxwells equations read

Presentation

As described on figure 1, we consider a stratified sample

with layers invariant in the (x, y) plane and stacked in
the z direction, starting from zf = 0. Because of the
translation invariance, modes with bounded fields can
be classified according to wave numbers (kx , ky ) R2 .
Without loss of generality, we consider plane waves in
the (x, z) direction, i.e. ky = 0.
The front half-space is isotropic and a plane wave
can be decomposed into s and p polarizations. The s
polarization is a wave with electric field perpendicular
(senkrecht) to the plane of incidence, i.e. along y. The
p polarization is a wave with electric field parallel to the
plane of incidence. A plane wave i is incident from the
front half-space with incidence angle i and reflected into
a plane wave r with the same angle. Angles are oriented
by the y direction. In the general case, the back halfspace is anisotropic and two transmitted plane waves t
are induced, with the same kx , but different angles t .
The electric vector is not necessarily perpendicular to ~k
and s and p waves can not be considered.

E =

1 B
c t

and H =

1 D
.
c t

We consider linear materials with tensorial constitutive

relations D = E and B = H. In Gaussian units,
0 = 0 = 1 and the relation to S.I. units is
ES.I. =

E
40

and HS.I. =

H
,
40

where 0 and 0 are the usual S.I. values. The convention

for time-varying complex fields is taken as exp(it),
and Maxwells equations read

Conventions and units

E = ik0 H

and H = ik0 E,

Formulas are mostly based on Schuberts article , with

one difference being the orientation of the incident p polarization. Schubert takes the convention used in ellipsometry, where the base electric vectors for p polariza-

with k0 = /c. We consider non-magnetic materials, i.e.

= 1. The permittivity may be complex. For example,
isotropic materials with Im() > 0 are lossy.

1 High speed calculations are possible with NumPy and SciPy

if array operations are vectorized. However the structure of Berreman4x4 is difficult to vectorize with enough generality.

2 In Fujiwaras book, reference 5, p. 223, the base electric vectors for p polarization point towards the z direction, which leads
to the same definition as Schuberts.

Propagation inside the sample

In the stratified sample, Maxwells equations lead to

a propagation equation for the transverse components
(Ex , Ey , Hx , Hy ). The demonstration can be found in
reference 3 and leads to3

Ex
Ey

(1)
= ik0 (z)(z), with =
Hx ,
z
Hy
where (z) is a 4 4 matrix. For a general dielectric
tensor, the matrix is 4 (z) =

3,1
3,2
Kx2
Kx
0 1
Kx
3,3
3,3

3,3

0
0
1
0

3,1

2,3
3,2

2,3
2,1 Kx2 2,2 + 2,3
0
Kx

3,3
3,3
3,3

3,1
3,2
1,3
1,1 1,3
1,2 1,3
0 Kx
3,3
3,3
3,3
The reduced wave number Kx = kx /k0 is a constant
throughout the sample and depends only on the angle of
the incident wave.
For a homogeneous slab z1 < z < z2 , the matrix (z)
is constant and equation 1 can be integrated into (z2 ) =
Phs (z2 , z1 ) (z1 ), where the propagator is given by the
matrix exponential


Phs (z2 , z1 ) = exp i (z2 z1 ) k0
The numerical computation of a matrix exponential is
generally slow. Berreman suggested to diagonalize the
matrix by searching the eigenvalues and eigenvectors 3 .
However, the knowledge of the eigenvectors is not necessary and the result can be expressed based on the eigenvalues only4 . I. Abdulhalim et al. used the LagrangeSylvester interpolation polynomial 79 and Whler et al.
used Cayley-Hamiltons theorem 10,11 . Both approaches
lead to the same expression. The eigenvalues can be
calculated numerically as the roots of the characteristic
polynomial. Literal expressions can be found for specific
situations, like for a uniaxial material 10,11 or in the case
of normal incidence 7 . Also, in the case of a diagonal
tensor, a specific solution for Phs is available 8 .
If a part of the sample is inhomogeneous, it is subdivided into slices over which the variation of (z) is small.
For such a slice, the propagator P (z2 , z1 ) is approximated
by a homogeneous slab for which the matrix is evaluated in the middle of the interval,



z1 + z2
P (z2 , z1 ) ' exp i (z2 z1 ) k0
.
2
3 In Berremans article, there is a misprint in equation (3), it
should read C = MG.
4 Knowledge of the eigenvectors is necessary only for the exit
transition matrix if the back half-space is anisotropic.

The total propagator P (zN , z0 ) for N slices between z0

and zN is approximated the product
Pa (zN , z0 ) = Phs (zN , zN 1 ) Phs (z1 , z0 ).
The order of the error is
P (zN , z0 ) Pa (zN , z0 ) = O(1/N 2 )
and Z. Lu demonstrated 12 that this does not depend on
whether the thin slab propagator Phs is the exact propagator or an approximation, possibly to first order. Consequently, the simplest solution to equation 1 is to take
the first order expansion


z1 + z2
(z1 ),
(z2 ) ' i(z2 z1 )k0
2
which corresponds to the first order expansion of the exponential,


z1 + z2
Phs (z2 , z1 ) ' i(z2 z1 )k0
2
For improving convergence and efficiency, Z. Lu presented an extrapolation method to eliminate the leading
terms of the error 12 . If the propagator used for the thin
homogeneous slabs is the exact propagator Phs , the error
is reduced to O(1/N 4 ). Z. Lu presented another version
with a symplectic integrator that showed improved convergence 13 . In this version, the propagator for a thin slab
is simply the product of three homogeneous slab propagators evaluated for different thickness and position (see
equation (10) in the reference).

Transition to half-spaces

For isotropic half-spaces, the relation between the wave

coefficients and the vector at the boundary is given by
the transition matrices Lf and Lb with

Eis
Ets
Ers
0

(0) = Lf
and (zb ) = Lb
Eip
Etp
Erp z=z
0 z=z
f

The ordering of the components is (s , s , p , p ), where

the superscripts indicate the traveling direction along the
z axis. The direction of each wave is characterized by
~ = (Kx , 0, Kz ) satisfying the
a reduced wave vector K
2
~ = n2 = . An angle determines
Helmholtz relation K
~ = n (cos(), 0, sin()) and may be a complex number
K
when the material is lossy or when an evanescent wave
occurs. It is unique with 0 ] , ] and 00 R.
Considering the angle for the wave traveling in the z
direction, we have 0 [/2, /2] and

0
0
cos() cos()

1
1
0
0

L=
n cos() n cos()
0
0
0
0
n
n
+

3
When the back half-space is anisotropic, we can not decompose the transmitted wave on s and p polarizations,
but by analogy, we decompose (zb ) over the eigenvectors k of the b matrix,
(zb ) =

4
X

ck k

k=1

We sort the eigenvectors so that 1 and 3 correspond

to waves propagating in the z direction, i.e. the eigenvalues q1 and q3 have positive real part. This description incorporates the isotropic case for which (c1 , c3 ) =
(Ets , Etp )(zb ). We can write

1 1 0 3 1 0
c1
12 0 32 0
0

(zb ) = Lb
c3 with Lb = 13 0 33 0
1 4 0 3 4 0
0

Matrix assembling and Jones matrices

The global propagation matrix and the two transition

matrices are assembled in order to relate the coefficients
of the waves in the two half-spaces. We obtain the total
transfer matrix T with

Eis
c1
c1
Ers

0
1

0
Eip = Lf P (zf , zb ) Lb c3 = T c3 .
Erp z
0
0

It is possible to consider other bases, for example the

left and right circular polarizations. For the incident and
transmitted waves the Jones vectors are
 
 
1
1 1
1
and EiR , EtR =
.
EiL , EtL =
i
i
2
2
For the reflected wave, we have
 
 
1
1 1
1
ErL =
and ErR =
.
2 i
2 i
The transformation matrix from the (s, p) basis to the
(L, R) basis will be called C in the case of incident and
reflected waves, and D for the reflected wave. We have




1
1 1 1
1 1
and D =
.
C=
2 i i
2 i i
The relations between the Jones vectors in the two bases
are
 






Eip
EiL
Etp
EtL
=C
,
=C
,
Eis
EiR
Eis
EiR




Erp
ErL
and
=D
.
Eis
EiR
As a result, the Jones matrices for circularly polarized
c
light are Ttic = C 1 Tti C and Tri
= D1 Tri C.

The two useful relations

 

Eip
T33
=
Eis z
T13
f



Erp
T43
=
Ers z
T23
f

can be extracted,
 
 
T31
c3
c3
= Tit
T11
c1
c1
 
 
T41
c3
c3
= Trt
.
T21
c1
c1

Reflection of the incident wave can be described by a

Jones matrix Tri , and if the back half-space is isotropic, a
Jones matrix Tti for transmission can also be defined 5,14 ,



 
 
Erp
rpp rps
Eip
Eip
=
= Tri
Ers z
rsp rss
Eis z
Eis z
f
f
f

 
 


tpp tps
Eip
Eip
Etp
= Tti
=
Ets z
tsp tss
Eis z
Eis z
b

These matrices contain all the information on reflection

and transmission of the sample. They are obtained by
the relations
Tri = Trt Tit1

and Tti = Tit1 .

Circularly polarized light

When fields are decomposed over the s and p polarizations, the basis for the Jones vectors is (Ep , Es ) with
 
 
1
0
Ep =
and Es =
.
0
1

Ellipsometry parameters

Ellipsometry parameters describe the reflection of the

sample by the normalized reflection matrix Tri /rss , as
presented in Fujiwaras book, reference 5, p. 220. However, since we use an opposite orientation convention for
Erp , a change of sign is necessary for matching the convention of ellipsometry and we define
tan(pp )
tan(sp )

eipp
eisp

!
tan(ps ) eips
1


rpp /rss rps /rss
=
.
rsp /rss
1

The minus sign in front of is chosen due to the

exp(it) phase convention. The ellipsometry angles are
chosen with [0, /2] and ] , ]. If the sample
is isotropic in the (x, y) direction, the off-diagonal coefficients sp and ps vanish and only two parameters are
needed to describe the reflection. We have




rpp 0
tpp 0
Tri =
and Tti =
0 rss
0 tss
and we define the ellipsometry parameters and with
tan() ei = rpp /rss .

Installation and use

Berreman4x4 is offered as a Python module named

Berreman4x4.py, which can be imported and used in
Python scripts with the command
import Berreman4x4
The module file should either be present in the working
directory or be accessible in the module path. A convenient organization is to store Python modules in a special
directory pointed by the PYTHONPATH environment variable. For example, my .bashrc script contains
export PYTHONPATH="/home/castany/.python"
and the .python directory contains symbolic links to the
different Python modules I use.
Berreman4x4 depends on the standard Python modules NumPy 15 and SciPy 16 . Application examples need
module matplotlib 17 for plotting.
General workflow with Berreman4x4 consists of (1)
building the structure, (2) calculating Jones matrices, (3)
extracting the desired coefficients and plotting.

Code documentation and examples

The source code contains detailed information on the

classes and functions, incorporated as docstrings. They
can be conveniently displayed when working with a shell
like IPython 18 .
Commented examples are bundled with the program
and can be run directly from the command line. They
range from simple situations to more complex structures:
reflection on an interface, interferences in a thin layer,
reflection on a Bragg mirror, 90 twisted nematic liquid
crystal. Whenever possible, result from Berreman4x4 is
compared with exact analytical result, and show excellent match. The code is also validated against results
presented in reference 5.
The UML class diagram of Berreman4x4 is presented on figures 2 and 3. A Structure is built
from a list of Layer objects, between one front
and one back HalfSpace objects.
Material objects are created for defining MaterialLayer objects.
Several classes of non-dispersive materials are provided, but dispersive materials may also be created, like the class IsotropicDispersive. Additional
classes of materials can be created by deriving classes
Material and DispersionLaw. Layers with a continuous spatial variation of the permittivity tensor are described with an InhomogeneousMaterial and form an
InhomogeneousLayer object.

10

TODO

General:

Check Berremans equations again. Which are the

assumptions?
Verify efficiency of Lus method. It seems to be
worse than the midpoint method for the twisted nematic example. Strange.
Check code if back half-space is lossy or total internal reflection occurs. Complex ?
Provide function for displaying elementary plots
R, T = f () for a structure.
Source code:
Homogeneous layer: implement exact solution with
Cayley-Hamilton or Lagrange-Sylvester.
Calculation of the fields inside the structure (not just
t and r).
Class MaterialLayer may be a bad idea, because
classes Material and InhomogeneousMaterial are
not brothers.
Maintenant
que
jai
cr
getPermittivityProfile(),
cest
peuttre possible de lutiliser pour simplifier
getPropagationMatrix().
Se demander si
cest une bonne ide. Il faudra alors ajouter le choix
de direction inv getPermittivityProfile().

Structure
+frontHalfSpace
+backHalfSpace
+layers
-__init__(self,front=None,layers=None,back=None)
+setFrontHalfSpace(halfSpace)
+setBackHalfSpace(halfSpace)
+setLayers(layers)
+getPermittivityProfile(lbda)
+getPropagationMatrix(Kx,k0=1e6,inv=False)
+getStructureMatrix(Kx,k0=1e6)
+getJones(Kx,k0=1e6)
+getIndexProfile(lbda=1e-6,v=e_x)
+drawStructure(lbda,margin=0.15)

Material
+getTensor(lbda)

NonDispersiveMaterial
+epsilon

IsotropicMaterial
+getRefractiveIndex(lbda)

-__init__(self,epsilon=None)
+getTensor(lbda=None)
+rotated(R)

UniaxialNonDispersiveMaterial
+n
-__init__(self,no=1.5,ne=1.7)

BiaxialNonDispersiveMaterial
+n

IsotropicNonDispersiveMaterial

IsotropicDispersive

+n

+law

-__init__(self,n=1.5)
+getRefractiveIndex(lbda=None)

-__init__(self,law=None)
+getTensor(lbda)
+getRefractiveIndex(lbda)

-__init__(self,diag=(1.5,1.6,1.7))

DispersionLaw
+getValue(lbda)

Berreman4x4 module
+c, pi
+rotation_Euler(angles)
+rotation_V(V)
+rotation_v_theta(v,theta)
+buildDeltaMatrix(Kx,eps)
+hs_propagator(Delta,h,k0,method="linear",q=None)
+hs_propagator_lin(Delta,h,k0,q=None)
+hs_propagator_Pade(Delta,h,k0,q=7)
+hs_propagator_Taylor(Delta,h,k0,q=5)
+hs_propagator_eig(Delta,h,k0,q=None)
+extractCoefficient(Jones,coeff_name)
+circularJones(Jones)
+extractEllipsoParam(Jr)

SellmeierLaw
+A, B, lbda0
-__init__(self,coeff=(1.0, 1.0, 300e-9))
+getValue(lbda)

FIG. 2: Class diagram of Berreman4x4: structure, materials, and module functions.

Layer

RepeatedLayers

+getPermittivityProfile(lbda)
+getPropagationMatrix(Kx,k0,inv)

MaterialLayer
+material
+setMaterial(material)

HomogeneousLayer
+h
+material
+hs_propagator: function
+hs_order
-__init__(self,material=None,h=1e-6,
hs_method="eig",hs_order=2)
+setThickness(h)
+setMethod(hs_method,hs_order=2)
+getPermittivityProfile(lbda)
+getPropagationMatrix(Kx,k0=1e6,inv=False)
+getDeltaMatrix(Kx,k0=1e6)

+n
+before
+after
+layers
-__init__(self,layers=None,n=2,before=0,after=0)
+setRepetition(n,before=0,after=0)
+setLayers(layers)
+getPermittivityProfile(lbda)
+getPropagationMatrix(Kx,k0=1e6,inv=False)

InhomogeneousLayer
+material
+z
+getSlicePropagator: function
+hs_propagator: function
+hs_order
+__init__(self,material=None,evaluation="midpoint",
hs_method="Pad",q=2)
+setMethod(evaluation,hs_method,q)
+getPermittivityProfil(lbda)
+getPropagationMatrix(Kx,k0=1e6,inv=False)
+getSlicePropagator_mid(z2,z1,Kx,k0=1e6)
+getSlicePropagator_sym(z2,z1,Kx,k0=1e6)

HomogeneousIsotropicLayer
+setThickness(h)
+get_QWP_thickness(lbda=1e-6)

InhomogeneousMaterial
+getTensor(z,lbda)
+getSlices()

Berreman4x4.HalfSpace
+material
-__init__(self,material=None)
+setMaterial(material)
+getTransitionMatrix(Kx,k0=1e6)

Berreman4x4.IsotropicHalfSpace
+material
-__init__(self,material=None)
+get_Kx_from_Phi(Phi,k0=1e6)
+get_Phi_from_Kx(Kx,k0=1e6)
+getTransitionMatrix(Kx,k0=1e6,inv=False)

T
wistedMaterial
+material
+d
+angle
+div
-__init__(self,material=None,
d=4e-6,angle=pi/2,div=25)
+setMaterial(material)
+setThickness(d)
+setAngle(angle)
+setDivision(div)
+getTensor(z,lbda=None)
+getSlices(self)

FIG. 3: Class diagram of Berreman4x4: layers, inhomogeneous materials, and half-spaces.

Validation examples taken from Fujiwaras book

Olivier Castany
We reproduce the situation presented by Fujiwara in his
book Spectroscopic Ellipsometry 5 , section 6.4.2, p. 241
243 and section 6.4.1, p. 237239. He presents detailed
calculations and intermediate steps that are useful for
testing our code.

Example of section 6.4.2

The situation is depicted on figure 4. A uniaxial film is

formed on a silicon substrate. The incident medium is
air and the silicon substrate has refractive index nt =
3.898 + 0.016i. The thickness of the film is d = 100 nm
and the refractive indices are no = 2.0 and ne = 2.5. The
orientation of the film is given by Euler angles E = /4
and E = /4. Angle E is the first Euler angle, inducing
a rotation of axis x around axis z, leading to axis x0 .
Angle E is the second Euler angle, inducing a rotation
of the axis of the material around x0 .

Example of section 6.4.1

In validation-Fujiwara-641.py, light is reflected in

the air by an anisotropic substrate as depicted on figure 5. We reproduce figure 6.16, p. 238, when the angle
of incidence i is varied and figure 6.17, p. 239, when
the Euler angle E of the anisotropic substrate is varied.
The difference with section 6.4.2 is that the substrate is
anisotropic and the code uses a general HalfSpace instead of an IsotropicHalfSpace.

In validation-Fujiwara-642.py, we reproduce Fujiwaras results with our code and obtain the same values,
except for a few places where a sign is reversed. The reason is that Fujiwara uses the convention of ellipsometry
for orienting Erp , which is the opposite of our convention. We also reproduce figure 6.19, p. 242, when the
orientation of the anisotropic film is varied.

Er

Er
x

Ex

Et

Ei

s
Air

Et

i
Ei

Film

i
s
Air

Silicon substrate

FIG. 4: Geometry of the situation treated by Fujiwara

in his book, section 6.4.2.

Anisotropic substrate

FIG. 5: Geometry of the situation treated by Fujiwara

in his book, section 6.4.1.
7

Example of the frustrated total internal reflection

Olivier Castany, Cline Molinaro

For p polarization, we have

Frustrated total internal reflection is used as a validation

example. Analytical and numerical results are compared.

~ = H(x, z) ~y
H

This duality implies that expressions for p polarization

relate naturally to H. However, transmission and reflection coefficients are always defined with respect to E,

Presentation

We consider the situation on figure 6, where two halfspaces with indices nf and nb are separated by a medium
of index ns and thickness d. The three media are assumed to be lossless. The incoming plane wave defines
vector kx throughout the structure. The reduced wave
vector is Kx = kx /k0 = nf sin(i ). The scalar Helmholtz
equation ( + k02 ){E, H} = 0, holds separatley inside
the three media, and implies that there are at most two
waves in each medium, with wave vector kz , given by
kz2 = k02 n2 kx2 .

t=

2
d

Er

nb

t
z

p
i

Ei
s

z=0

z=d

FIG. 6: Frustrated total internal reflection with input

and output plane waves.

with

If we consider a p-polarized wave defined by

We consider the separation medium. Since kx is real,

kz is either purely real or purely complex. The first case
happens for incidence angles i smaller than the critical
angle ic , given by sin(ic ) = ns /nf . In this case, two
plane waves are present, with kz wavenumber. The second case corresponds to the total internal reflexion in the
front half-space medium and these is an evanescent wave
in the separation medium. If the third material is close
enough, a plane wave is transmitted from the evanescent
wave. This phenomenon is the frustrated total internal
reflexion. In this situation, two evanescent waves are
present in the medium, with kz00 exponential evolution.
Due to the (xz) mirror symmetry, s and p modes can
be considered separately. For s polarization, we have
~ = E(x, z) ~y
E

Er
, for both s and p polarizations.
Ei

E(x, y) = E0 eit+i(kx x+kz z) ,

Kz
~ = E(x, y) 0 ,
we deduce H
Kx

Kz

~ = E0 eiti(kx x+kz z) 0 ,
H
Kx

Kx
~ t = c |E0 |2 e2kz00 z 0 .
and hi
(2)
8
Kz0

t
E+

r=

Anatomy of a single wave

~ = E(x, y) ~y
E

p
Et

and

If we consider an s-polarized wave defined by

Back half-space

ns

nf

Et
Ei

In the next sections, we will study the anatomy of the

waves in detail. For that purpose, the Poynting vector
gives useful physical insight. In Gaussian units, the time
average of the Poynting vector is


~ t = c Re E
~ H
~ .
hi
8

x
Front half-space

z H
i
~ =

0 .
and E
k0
x H

~ = H(x, y) ~y
H
we deduce

with

H(x, y) = H0 eit+i(kx x+kz z) ,

Kz
H(x,
y)
~ =
0 ,
E

Kx

~ = H0 eiti(kx x+kz z) ~y ,
H

Kx
2
00
c
|H
|
0
~ t=
and hi
e2kz z 0 .
8
Kz0

(3)

(4)

In these expressions, we defined kz = kz0 + ikz00 . We

observe that the Poynting vector is parallel to the real
part of the wave vector. If kz is real (kz00 = 0), the wave
amplitude is constant along z. If kz is purely complex
(kz0 = 0), there is no energy flow in the z direction, and
the wave decays exponentially in the z direction. Figure 7
represents the two cases.

z E
1
~ =
0 .
and H
ik0
x E
8

9
x

kz real

(a) Plane wave

which exhibits an energy flow in the z direction, proportionnal to Kz00 rs00 .

For p polarisation, we consider
 +
 +
~ = H + ~y
H
H = H0+ eit+i(kx x+kz z)
with

~ = H ~y
H = H0 eit+i(kx xkz z) .
H

kz purely complex

(b) Evanescent wave

FIG. 7: Anatomy of a homogeneous plane wave and an

evanescent wave. The arrows show the real part of the
wave vector. The thickness of the arrow indicates the
intensity of the wave.

Reflection on an interface: fields and

Poynting vector

We consider an incident wave partially reflected at z = 0

by a structure that does not affect the parity of the wave
(s or p polarization). The complex reflection coefficients
are called rs and rp . The waves may either be plane or
evanescent waves. The incident and reflected waves are
named with + and subscripts. The total field is
~ =E
~+ + E
~ . For the s polarisation, we consider
E
 +
 +
~ = E + ~y
E
E = E0+ eit+i(kx x+kz z)
with

~ = E ~y
E = E0 eit+i(kx xkz z) .
E
The magnetic excitation is

Kz
Kz
= E+ 0 + E 0 .
Kx
Kx

~ =H
~++H
~
H

The amplitudes of the incident and reflected waves are

connected by rs = E0 /E0+ and the Poynting vector is

Kx
Kx

00
00
c
~ t=
|E + |2 e2kz z 0 + |rs |2 e2kz z 0 +
hi
8 0
Kz0
Kz0

Kx cos(s 2kz0 z)

0
+2 |rs |

0
00
Kz sin(s 2kz z)
where we defined rs = |rs | e . The first and second
terms correspond to the incident and reflected waves.
The third term arises from the interference of the two
waves. If Kz is real, the expression in curly braces becomes

Kx
Kx
Kx cos(s 2kz z)
0 + |rs |2 0 + 2 |rs |
.
0
Kz
Kz
0
is

If Kz is purely complex, it becomes

Kx
Kx
Kx rs0
00
00
e2kz z 0 + |rs |2 e2kz z 0 + 2 0 ,
0
0
Kz00 rs00

The electric field is

~ =E
~+ + E
~
E

Kz
Kz

H+
H
0 .
0 +
=

Kx
Kx

(5)

The amplitudes of the incident and reflected waves are

connected by rp = Ep /Ep+ = H0 /H0+ . The Poynting
vector has the same expression as for the s polarization,
with rs replaced by rp = |rp | eip and |E0+ |2 replaced by
|H0+ |2 /,

Kx
Kx
+ 2
00
00
|H
|
c
0
~ t=
e2kz z 0 + |rp |2 e2kz z 0 +
hi
8
Kz0
Kz0

Kx cos(p 2kz0 z)
.
0
+2 |rp |

00
0
Kz sin(p 2kz z)
If Kz is purely complex, there is an energy flow in the
z direction, proportionnal to Kz00 rp00 . Figure 8 represents
the variation of the Poynting vector along z, in the case
of evanescent waves.
x

~ t
hi

h
x it

Interface

h+
x it

hx,int it
hz,int it

z=0

FIG. 8: Refection of an evanescent wave on an interface

(kz is purely complex and r00 0). The Poynting vector
is decomposed in different terms.

Expression of the reflexion coefficients

for an interface between two media

We consider the reflexion on an interface between two

media 1 and 2, with light coming from medium 1.
The reflection and transmission coefficients for the s polarization are 19
rs =

kz1 kz2
kz1 + kz2

and

ts = 1 + rs =

2 kz1
.
kz1 + kz2

10
The expressions are valid for complex wave vectors and
we deduce
rs0 =

|kz1 |2 |kz2 |2
|kz1 + kz2 |2

and

rs00 =

00 0
0
00
2(kz1
kz2 kz1
kz2
)
.
2
|kz1 + kz2 |

In the case of an evanescent wave in region 1, the wave

00
number is kz1 = ikz1
and we deduce
rs0 =

002
kz1
|kz2 |2
00
|ikz1 + kz2 |2

and

rs00 =

00 0
2kz1
kz2
.
00
|ikz1 + kz2 |2

The last expression shows that rs00 0, which implies

that the energy flow is directed to the right, as expected.
For p polarization, the coefficients for the magnetic

+
+
+
and tH
field are rpH = H1p
/H1p
p = H2p /H1p with
rpH

2 kz1 1 kz2
=
2 kz1 + 1 kz2

and

tH
p

1+rpH

22 kz1
=
.
2 kz1 + 1 kz2

The direction and value of the electric field is deduced

from equations 3 and 5, which leads to rp = rpH and
19
tp = n1 /n2 tH
p and implies
1 kz2 2 kz1
rp =
2 kz1 + 1 kz2

and

2n1 n2 kz1
tp =
.
2 kz1 + 1 kz2

|1 kz2 |2 |2 kz1 |2

We deduce rp0 =
and
|2 kz1 + 1 kz2 |2
rp00 =

00 0
0
00
21 2 (kz1
kz2 kz1
kz2
)
.
2
|2 kz1 + 1 kz2 |

In the case of an evanescent wave in region 1, the wave

00
number is kz1 = ikz1
and we deduce
rp0 =

002
|1 kz2 |2 22 kz1
00
|2 ikz1 + 1 kz2 |2

and

rp00 =

00 0
21 2 kz1
kz2
.
00
|2 ikz1 + 1 kz2 |2

The last expression shows that rp00 0, which implies

that the energy flow is directed to the right, as expected.
The power flow along the z direction is deduced from
equations 2 and 4, leading to the expressions for the
power coefficients for both polarizations,
R=

h
z1 it
= |r|2
+
hz1 it

and

T =

h+
k0
2
z2 it
= z2
+
0 |t| .
kz1
hz1 it

We verify that R + T = 1 for both polarizations when

the materials are lossless5 . Also, if we consider the coefficients for the reverse directions, we show that for both
polarizations we have r12 = r21 , t12 /k2 = t21 /k1 and
2
=1.
t12 t21 + r12
5 The relation only has a meaning when k
z1 is real. The demonstration uses the fact that kz2 is either real or purely complex. The
two cases are considered separately and both verify R + T = 1.

Application to the frustrated total

internal reflection

We consider waves Ei and Er in the incident medium,

E + and E in the separation medium, and Et in the
back half-space. The relations at the interfaces can be
written in the same fashion for both polarizations. At
the z = 0 interface, we have
 +
E (0) = rf s E (0) + tsf Ei (0)
Er (0) = rsf Ei (0) + tf s E (0)
and at the z = d interface, we have

Et (d) = tbs E + (d)
E (d) = rbs E + (d).
Propagation between planes z = 0 and z = d implies
 +
E (d) = E + (0) eikz d
E (0) = E (d) eikz d ,
where kz is the wavenumber in the separation medium.
From these equations, we get6
E + (0)/Ei (0) =

tsf
1 rf s rbs ei2kz d

E (0)/Ei (0) =

rbs tsf ei2kz d

1 rf s rbs ei2kz d

r = Er (0)/Ei (0) =

rsf + rbs ei2kz d

1 rf s rbs ei2kz d

t = Et (d)/Ei (0) =

tbs tsf eikz d

1 rf s rbs ei2kz d

The power reflection and transmission coefficient are

R = |r|2

and

T =

hzt it
k0
= 0zb |t|2 .
hzi it
kzf

We verify that R + T = 1 when the materials are

lossless. To demonstrate this, different cases are separated, depending on the type of wave in the regions.
In the case when there are plane waves in both half0
0
spaces, we show that kzb
/kzf
|tbs tsf |2 exp(2kz00 d) =
00
|tf s tsf tbs tsb | exp(2kz d). When there is also a plane
wave in the separation region, the reflection coefficients are real, and the term simplifies to (1 rf2 s )(1
2
rbs
). When there is an evanescent wave in the separation medium, the reflection coefficients are complex with unitary modulus, and the term simplifies to
00 00
4 rbs
rf s exp(2kz00 d). In both cases, putting this term in
the expansion of R + T leads to the result R + T = 1.

6 We used the relation t t

f s sf rf s rsf = 1 for simplifying the
expression of the coefficient r.

Various examples
Cline Molinaro, Olivier Castany

and transmission are calculated, with linear or circular

bases, for normal incidence.
In interface-reflection.py, the power reflection
and transmission coefficients R and T are plotted for both
polarizations as a function of the reduced wave vector
Kx . We observe that R + T = 1. The coefficient |t2 | is
also plotted and is clearly different from T . The result
from Berreman4x4 and the analytic solution are plotted
together and are identical. The refractive indices can be
given from the command line and total internal reflection
is observed if n1 > n2 .

Glass layer

In Interferences.py, we consider the glass plate from

figure 9, where light arrives with a 30 incidence angle
(Kx = 0.5). The reflection and transmission coefficients
R and T are plotted as a function of the glass thickness,
for both polarizations. Interferences are visible.
x

We consider TiO2 /SiO2 Bragg mirrors as presented on

figure 11. In validation-Bragg.py, the reflection coefficient R is calculated with Berreman4x4 and compared to the analytical result 19 , for different incidence
angles and polarizations. The results are identical. In
Bragg-example.py, the reflection and transmission coefficients R and T are calculated for a Bragg mirror with
8.5 periods at normal incidence.

30
i
Air
n = 1.0

Glass
n = 1.5

Bragg Mirror

Air

FIG. 9: Glass plate with multiple reflections inside.

Reflection on an interface
t

A plane wave is reflected and refracted by an interface,

as represented on figure 10.

z
i

x
n1

n2

Air

TiO2

SiO2

TiO2

SiO2

Glass

period
t
t

FIG. 11: TiO2 /SiO2 Bragg mirror with two periods.

i
i

Twisted nematic liquid crystal

We consider a twisted nematic liquid crystal between

two glass substrates as represented on figure 12. In
twisted-nematic.py, the transmission coefficient T calculated with Berreman4x4 is compared with the GoochTarry law. Agreement is excellent when the twisted layer
is divided into 18 divisions. When only 7 divisions are
used, there is a slight discrepancy in the high wavenumber range.

FIG. 10: Reflection of a plane wave on an interface between two materials.

In interface-Jones.py, we consider an interface with
n1 = 1.0 and n2 = 1.5. The Jones matrix for reflection
11

12
x
Glass

LC

Glass
A
z

d
FIG. 12: Twisted nematic liquid crystal between glass
substrates with polarizer and analyzer.

Cholesteric liquid crystal

We consider the cholesteric liquid crystal from figure 13,

with right handedness.
x
p
2

Glass

Glass
z

FIG. 13: Cholesteric liquid crystal aligned along the x

direction on the two glass substrates. The pitch p corresponds to a 360 turn of the helix.
In validation-cholesteric.py, the reflection spectrum of a right-handed circular polarization is calculated
with Berreman4x4 and compared to the analytic expressions from references 20 and 21. Agreement is good and
the little differences may be due to the fact that the analytic expression does not take into account the reflection
at the glass boundaries. In Berreman4x4, the half-pitch
of the twisted material was divided into 25 parts, which
is more than enough for an accurate result.
In cholesteric-example.py, transmission and reflection of a cholesteric structure is simulated for right
handed circularly polarized light. Transmission of unpolarized light is also simulated, and analyzed with different
polarizers. Eigenvalues and eigenvectors of the transmission matrix for different cases are calculated.

Bibliography
[1]

M. Born and E. Wolf. Principles of Optics. Cambridge University Press, 1999.

[2]

S. Teitler and B. W. Henvis. Refraction in stratified, anisotropic media. J. Opt. Soc. Am., 60
(6):830834, Jun 1970. doi: 10.1364/JOSA.60.
000830. URL http://www.opticsinfobase.org/
abstract.cfm?URI=josa-60-6-830.

[3]

[4]

[13] Zhao Lu.

Efficient 4 4 propagation matrix
method using a fourth-order symplectic integrator
for the optics of one-dimensional continuous inhomogeneous materials. Progress In Electromagnetics
Research Letters, 14:19, Sep 2010. doi: 10.2528/
PIERL10031501.

Mathias Schubert. Polarization-dependent optical parameters of arbitrarily anisotropic homogeneous layered systems. Phys. Rev. B, 53:4265
4274, Feb 1996.
doi: 10.1103/PhysRevB.53.
4265. URL http://link.aps.org/doi/10.1103/
PhysRevB.53.4265.
Hiroyuki Fujiwara.
Spectroscopic Ellipsometry:
Principles and Applications. John Wiley & Sons,
March 2007. ISBN 9780470016084.

[6]

J.D. Jackson.
1998.

[7]

I. Abdulhalim, L. Benguigui, and R. Weil. Selective reflection by helicoidal liquid crystals. results of an exact calculation using the 4
4 characteristic matrix method.
J. Phys.
France, 46(5):815825, 1985. doi: 10.1051/jphys:
01985004605081500. URL http://dx.doi.org/10.
1051/jphys:01985004605081500.

[9]

[12] Zhao Lu. Accurate and efficient calculation of

light propagation in one-dimensional inhomogeneous
anisotropic media through extrapolation. J. Opt.
Soc. Am. A, 24(1):236242, Jan 2007. doi: 10.1364/
JOSAA.24.000236. URL http://josaa.osa.org/
abstract.cfm?URI=josaa-24-1-236.

Dwight W. Berreman.
Optics in stratified
and anisotropic media:
44-matrix formulation.
J. Opt. Soc. Am., 62(4):502510, Apr
1972.
doi: 10.1364/JOSA.62.000502.
URL
http://www.opticsinfobase.org/abstract.
cfm?URI=josa-62-4-502.

[5]

[8]

[11] H. Whler, M. Fritsch, G. Haas, and D. A. Mlynski. Characteristic matrix method for stratified
anisotropic media: optical properties of special
configurations. J. Opt. Soc. Am. A, 8(3):536
540, Mar 1991. doi: 10.1364/JOSAA.8.000536.
URL http://josaa.osa.org/abstract.cfm?URI=
josaa-8-3-536.

Classical Electrodynamics.

[14] R. Clark Jones. A new calculus for the treatment of optical systems. J. Opt. Soc. Am., 31
(7):488493, Jul 1941. doi: 10.1364/JOSA.31.
000488. URL http://www.opticsinfobase.org/
abstract.cfm?URI=josa-31-7-488.

Wiley,

[15] NumPy, 2012. http://numpy.scipy.org/.

[16] SciPy, 2012. http://www.scipy.org/.
[17] matplotlib plotting library, 2012.
matplotlib.sourceforge.net/.

http://

[18] IPython, 2012. http://ipython.org/.

[19] Wikipedia.
Coefficients de Fresnel, 2012.
http://fr.wikipedia.org/wiki/Coefficients_
de_Fresnel.

I. Abdulhalim. Analytic propagation matrix method

for linear optics of arbitrary biaxial layered media. Journal of Optics A: Pure and Applied Optics, 1(5):646, 1999. URL http://stacks.iop.org/
1464-4258/1/i=5/a=311.

[20] D.K. Yang and S.T. Wu. Fundamentals of Liquid

Crystal Devices. Wiley Sid Series in Display Technology. John Wiley, 2006. ISBN 9780470015421.
[21] S. Chandrasekhar. Liquid Crystals. Cambridge University Press, 1992. ISBN 9780521417471.

Feliks Ruvimovich Gantmakher. The Theory of Matrices. American Mathematical Soc., 1959. ISBN
9780821813768.

[10] H. Whler, G. Haas, M. Fritsch, and D. A. Mlynski. Faster 4 4 matrix method for uniaxial inhomogeneous media. J. Opt. Soc. Am. A, 5(9):1554
1557, Sep 1988. doi: 10.1364/JOSAA.5.001554.
URL http://josaa.osa.org/abstract.cfm?URI=
josaa-5-9-1554.
13