Electro-Optic Effects:

definition
Electro-optic effects refer to changes in the
refractive index of a material induced by the a
pplication of an external electric field
Which therefore modulates the optical
properties
The applied field is not the electric field of any light
wave but a separate external field

We can apply an external field by placing

electrodes on opposite faces of a crystal and c
onnecting these electrodes to a battery

The presence of the field distorts the electron

motions in the atoms/molecules of the substance
Distorts the crystal structure resulting in changes in
optical properties

Electro-optic effect
An applied external field can cause an
optically isotropic crystal such as GaAs to b
ecome birefringent

The field induces principal axes and an optic axis

Typically changes in the refractive index are
small
The frequency of the applied field has to be such
that the field appears static over the time scale
it takes for the medium to change its properties,
as well as for any light to cross the substance

The electro-optic effect are classified

according to first or second order effects

Field induced refractive

index
Take the refractive index n to be a function of the
applied E-field, that is n=n(E), we can expand this as
a Taylor series in E. The new refractive index n woul
d be:
n= n + a1E + a2E 2+.
where the coefficients are called the linear electro-optic
effect and second order electro-optic effect coefficients.

The change in n due to the first E term is called the

Pockels effect
n = a1E
The change in n due to the second E 2 term is called
the Kerr effect and a2=K where K is called the Kerr coeffi
cient

Pockels Effect
Suppose x, y and z are principal axes of a crystal
with refractive indices n1, n2 and n3 along these
directions

For an optically isotropic crystal, these would be the

same
For a uniaxial crystal n1= n2 n3

Apply a voltage across a crystal and thereby

apply an external dc field Ea along z-axis

In Pockels effect, the field will modify the optical

indicatrix.
The exact effect depends on the crystal structure
GaAs (isotropic) with a spherical indicatrix becomes
birefringent

Pockels Effect: KDP (KH2PO4)

The field Ea along z rotates the principal axes
by 45 about z
Changes the principal indices as shown in
Fig.10(b)

The new principal indices are now n1 & n2, which

means the cross section is now an ellipse
Propagation along the z-axis under an applied
field now occurs with different refractive indices n
1 & n2
The applied field induces new principal axes x &
y for this crystal

Fig 10: Pockels Effect

y

n 2 =n o

n 1 =n o

(a)

n 2

n 1

KDP,LiNbO 3

Ea

Ea

n 2

45

n 1

KDP

LiNbO 3
(b)

(c)

(a)Crosssectionoftheopticalindicatrixwithnoappliedfield,n1=n2=no(b)The
appliedexternalfieldmodifiestheopticalindicatrix.InaKDPcrystal,itrotatesthe
principalaxesby45toxandyandn1andn2changeton1andn2.(c)Applied
fieldalongyinLiNbO2modifiestheindicatrixandchanges n1andn2changeton1
andn2.
1999S.O.Kasap,Optoelectronics (PrenticeHall)

Pockels Effect: LiNbO3

In the case lithium niobate (uniaxial crystal),
a field Ea is applied along the y-direction

It does not significantly rotate the principal axes

changes the principal refractive indices n1 & n2
(both equal to no) to n1 & n2 as shown in Fig 10(c)

Consider a wave propagating along the zdirection (optic axis) in the crystal

Before a field Ea is applied, this wave experience

n1=n2=no whatever in the polarization as Fig 10(a)
In the presence of an applied field Ea, the light
propagates as two orthogonally polarized waves (
parallel to x and y) experiencing different refractiv
e indices n1 & n2

Pockels Effect: LiNbO3, cont

The applied field thus induces a
birefringence for light traveling along the z-a
xis.
The field induced rotation of principal axes
is neglected.
The Pockels effect gives the new refractive
indices n1 & n2 in the presence of Ea as
n1 n1 + n13 r22 Ea & n2 n2 n23 r22 Ea

where r22 is a constant, called a Pockels coeffient

that depends on the crystal structure and the ma

Phase modulator
It is clear that the control of the refractive
index by an external applied field is a disti
nct advantage that enables the phase cha
nge through a Pockels crystal to be control
led or modulated
Such a phase modulator is called a Pockels cell

In the longitudinal Pockels cell phase

modulator, the applied field is in the directi
on of light propagation -> Fig 10(b)
In the transverse phase modulator, the
applied field is transverse to the direction
of light propagation -> Fig 10(c)

Transverse Phase
Modulator

The applied field Ea = V/d is applied parallel to the y-direction

(normal to the direction of light propagation along z)
The incident beam is linearly polarized at 45 to the y-axis.
It is represented in terms of polarizations (Ex & Ey components)
along the x and y axes
Ex & Ey experience refractive indices nx & ny.

When Ex traverses the length distance L, its phase changes by

2n1
2L
V
1.
1 3
1
L
n

n
r
o 2 o 22

When Ey traverses the distance L, its phase changes by 2,

given by a similar expression except that r22 changes sign. Thus
2 3 L
, the phase
changes
1 2 between

no r22two
V components is

Polarization modulator
The applied voltage thus inserts an
adjustable phase different between the t
wo field components

The polarization state of the output wave can be

controlled by the applied voltage and the
Pockels cell is a polarization modulator.

The medium can be changed from a

quarter-wave to a half-wave plate by simply
adjusting V.
The voltage V = V/2 , the half-wave voltage and
generate a half-wave plate ()

Fig 11: Transverse Phase

modulator
Ey
Input
light

y
d

Ex

Ea x

Ey

Ex

z Output
light

TranversePockelscellphasemodulator.Alinearlypolarizedinputlight
intoanelectroopticcrystalemergesasacircularlypolarizedlight.
1999S.O.Kasap,Optoelectronics (PrenticeHall)

Transverse Intensity Modulator

From the polarization modulator in Fig.11, an
intensity modulator can be built as shown in Fig.12
by inserting a polarizer P and an analyzer A before and
after the phase modulator
P and A have their transmission axes at 90 to each other

The transmission axis of P is at 45 to the y-axis

The light entering the crystal has equal Ex and Ey
components

In the absence of applied voltage, two components

travel with the same refractive index and polarizatio
n output is the same as its input
There is no light detected at the detector as A and P are at
the right angle

Fig 12: Transverse Intensity

Modulator
QWP
Transmissionintensity

Io

A
Detector

Input
light

Crystal

x
z

Left:AtranversePockelscellintensitymodulator.Thepolarizer P andanalyzer A have

theirtransmissionaxisatrightanglesand Ppolarizesatanangle45 toyaxis.Right:
Transmissionintensityvs.appliedvoltagecharacteristics.Ifaquarterwaveplate
isinsertedafter P ,thecharacteristicisshiftedtothedashedcurve.
1999S.O.Kasap, Optoelectronics(PrenticeHall)

Transverse Intensity Modulator,

2
An applied voltage inserts a phase difference
between the two E-field components

The light leaving the crystal now has an elliptical

polarization and hence a field components along the trans
mission axis of A
A portion of this light will therefore pass through A to the
detector
The transmitted intensity now depends on the applied
voltage V.

Suppose that Eo is the amplitude of the wave

incident on the crystal face

The amplitude along x- and y-axis will be Eo/2 each.

Eothe x direction
Eo
is along
Ex E

x
cos t y
cos t

2
2
The total field
E at the analyzer is

Transverse Intensity Modulator,

3
A factor cos(45) of each component passes
through A.

We can resolve Ex and Ey along As transmission axis

Then add these components and use trigonometric
identity to obtain the field emerging from A
The final result is

E = Eo sin( ) sin(t + )

The intensity I of the detected beam is

I = Io sin2( )
or
I = Io sin2(/2 V/V )

where Io is the light intensity under full transmission

and V is an applied voltage needed to allow full
transmission

Example: Pockel Cell Modulator

What should be the aspect ratio d/L for
the transverse LiNiO3 phase modulator i
n Fig.11 that will operate at a free-spac
e wavelength of 1.3m and will provide
a phase shift of (half wavelength) b
etween the two field components propa
gating through the crystal for an applied
voltage of 24V?
( no = 2.2, r22 = 3.41012 m/V)

Solution
Substitute for the phase difference between the field components Ex and E y
Letting V V / 2
2 3 L
no r22 V / 2

d
d
1 2 3
1
2

no r22V / 2
2
.
2
3.4 10 12 24 1.3 10 3

6
L
1.3 10
This particular transverse phase modulator has the field applied along the y - direction
and light traveling along the z - direction. If we were to use the transverse arrangement
in which the field is applied along the z - axis, and the light travels along the y - axis,
the relevant Pockels coefficient would be greater and the corresponding aspect ratio d/L

would be ~ 10- 2. We cannot arbitrarily set d/L to any ratio we like for the simple reason
that when d becomes too small, the light will suffer diffraction effect that will prevent it
from passing through the device. d/L ratios 10-3 10- 2 in practice can be implemented by
fabricating an integrated optical device.

Integrated optical
modulators

Phase and polarization modulation

Integrated optics refers to the integration
of various optical devices and components o
n a single substrate such as lithium niobate.

In integrated electronics, all necessary devices

are integrated in the same semiconductor crystal
substrate

There is a distinct advantage to

implementing various optical communicated
devices on the same substrate

E.g. laser diodes, waveguides, splitters,

modulators, photodetectors etc in a miniature d
evice

Polarization modulator
Polarization modulator is shown Fig.13

An embedded waveguide has been fabricated

by implanting a LiNbO3 substrate with Ti atoms w
hich increase the refractive index
Two coplanar strip electrodes run along the
waveguide and enable the application of a transv
erse field Ea to light propagation direction z

The external modulating voltage V(t) is

applied between the coplanar to drive electr
odes

By virtue of the Pockels effect, induces a

change n in the refractive index and hence a vol

Fig 13: Polarization

modulator
V(t)

Coplanarstripelectrodes

Thinbufferlayer

Polarized
input
light

LiNbO 3

Ea

EOSubstrate

x
y

Waveguide

LiNbO 3

Crosssection

IntegratedtranversePockelscellphasemodulatorinwhichawaveguideisdiffused
intoanelectrooptic(EO)substrate.Coplanarstripelectrodesapplyatransverse
fieldE athroughthewaveguide.ThesubstrateisanxcutLiNbO3andtypicallythere
isathindielectricbufferlayer(e.g.~200nmthickSiO2)betweenthesurface
electrodesandthesubstratetoseparatetheelectrodesawayfromthewaveguide.
1999S.O.Kasap,Optoelectronics (PrenticeHall)

Polarization modulator, 2
The light propagation along the guide can be
represented in terms of two orthogonal modes, Ex al
ong x and Ey along y
These two modes experience symmetrically opposite phase
changes
The phase shift between the Ex and Ey polarized waves
would normally be given by Pockels effect

In the case of the applied field is not uniform

between the electrodes and further not all applied fi
eld lines lie inside the waveguide

The electro-optic effect takes place over the spatial

overlap region between the applied field & the optical fiel
ds
This spatial overlap efficiency is lumped into a
coefficient and the phase shift is written as

(no3r22)(L/d )V

Optical Switching:
Mach-Zehnder Modulator
One potential application of self-phase
modulation is in optical switching to switc
h the output from low to high intensity in fs
time scale.
In optical switching, induced phase shift by
applied voltage can be converted to an
amplitude variation by using an interfero
meter

Interferometer is a device that interferes two

waves of the same frequency but different phas
e

Fig 14: Optical Switching:

Mach-Zehnder Modulator
V(t)

Electrode
C
In

B
B

Out

Waveguide
LiNbO 3

EOSubstrate

AnintegratedMachZenderopticalintensitymodulator.Theinputlightis
splitintotwocoherentwavesAandB,whicharephaseshiftedbythe
appliedvoltage,andthenthetwoarecombinedagainattheoutput.
1999S.O.Kasap,Optoelectronics (PrenticeHall)

Mach-Zehnder Modulator, 2
Consider the structure shown in Fig.14, which has
implanted single mode waveguide in a LiNbO3 subst
rate in the geometry.
The waveguide at the input braches out at C to two arms A
and B
These arms are later combined at D to constitute the
output
The splitting at C and combining at D involve a simple Yjunction waveguides

In the ideal case, the power is equally split at C

so that the field is scaled by a factor 2 going into e
ach arm
The structure acts as an interferometer because the two
waves traveling through the arm A and B interfere at the o
utput port D

Mach-Zehnder Modulator, 3
Two back-to-back identical phase
modulators enable the phase changes in A
and B to be modulated.

The applied field in branch A is in opposite

direction to that in branch B
The refractive index changes are opposite and
phase changes in arm A and B are also
opposite

If applied voltage induces a phase change

of /2 in arm A, this will be /2 in arm B so t
hat A & B would be out of phase by .
These two waves will interfere destructively and
cancel each other at D.
The output intensity would be zero

Mach-Zehnder Modulator, 4
Since the applied voltage controls the phase
difference between the two interfering
waves A and B at the output

This voltage also control the output light intensity

(the relationship is not linear)

The relative phase difference between the

two waves A and B is doubled with respect
to a phase change in a single arm

The switching intensity can be predicted by

adding waves A and B at D with A as amplitude o
f wave A & B:

Eoutput A cos(t+) + A cos(t) = 2A cos cost

Mach-Zehnder Modulator,
5
The output power is proportional to E2output which
is maximum when = 0. Thus,

Pout
cos 2
Pout 0

The derivation represents approximately the right

relationship between the power transfer and the i
nduced phase change per modulating arm.
The power transfer is zero when /2.
In practice, the Y-junction losses and uneven
splitting results in less than ideal performance
A and B do not totally cancel out when /2