Electropolymerization

Vol. 2

ELECTROOPTICAL APPLICATIONS

283

ELECTROOPTICAL APPLICATIONS

Introduction

The term electrooptic can have widely different meanings. In the broadest sense,
it can imply any change produced in an optical beam by application of an electrical
ﬁeld from either a low frequency (dc to millimeter wave) ﬁeld or another optical
(micrometer wave) ﬁeld. Change of the index of refraction of the material through
which an optical beam is transiting can inﬂuence the (velocity of the) propagating
optical beam. Such change of index of refraction can be accomplished by altering
either the electron distribution or the nuclear conﬁguration of a material. The
associated optical phenomena are referred to as Kerr and Pockels effects (1–15)
and have been known since the nineteenth century.

However, altering an optical beam by electrically controlled mechanical

movement of a mirror (particularly as in micro-electro-mechanical-systems,
MEMS) is also referred to as an electrooptic operation. Indeed, even greater con-
fusion arises by usage of the terms electrooptic and all-optical signal processing
in the telecommunications industry. Electrooptic is taken to refer to converting
optical signals to electrical signals, carrying out signal processing in the electri-
cal (electronic) domain, and then converting back to the optical domain. Thus,
techniques such as bubble jet (Hewlett Packard/Agilent) and MEMS (Texas In-
struments) optical switching at network nodes are referred to as all-optical signal
processing by telecommunication engineers while an optical physicist would re-
serve the term all-optical switching for the dynamic Kerr phenomenon (switching
of one optical beam by another).

A review of electrooptical applications in the broadest sense is out of the

question. It would involve treatment of topics ranging from modulated lasers
(direct modulation) to light-emitting diodes (LEDs), to liquid crystalline materials,

284

ELECTROOPTICAL APPLICATIONS

Vol. 2

not to mention MEMS, thermooptic, and bubble jet technologies currently used
in telecommunications. Moreover, the literature is enormous. Over 1000 reviews
and texts have appeared on the subject of liquid crystalline electrooptic materials
alone. In this article, a narrow view of electrooptic phenomena, materials, and ap-
plications is taken and focus is upon the “electronic” Pockels effect, with emphasis
on the phenomenon realized using polymeric or dendritic materials rather than
inorganic or organic single crystals. The term electronic refers to changes in refrac-
tive index associated purely with electric-ﬁeld-induced change of the electron dis-
tribution of a material (in contrast to changes in nuclear conﬁguration associated
with molecular reorientation). Moreover, there is a focus upon

π-electron organic

chromophores rigidly embedded in polymeric or dendritic supramolecular matri-
ces. Currently, commercially utilized materials exploiting the electronic Pockels
effect are inorganic crystalline materials such as lithium niobate. Although the
most commercially viable organic electrooptic (deﬁned in the broader sense) mate-
rials are liquid crystalline materials (16–58), electronic Pockels effect organic ma-
terials are receiving increased attention because of the possibility of exceptional
(greater than 100 GHz) bandwidth performance. Although potential applications
of such materials range from telecommunications to information processing, to
defense, to transportation, and to display industries, the telecommunications and
defense industries are particularly important drivers of development of electroop-
tic materials with greater bandwidths. Electronic Kerr materials (59–67) have the
potential for even greater bandwidth and the possibility of convenient all-optical
signal processing; however, to the present time electronic Kerr coefﬁcients have
not been large enough to be commercially viable for telecommunication appli-
cations. Mott–Hubbard transition-metal oxide and halide systems with charge-
transfer gaps and low hydrogen-density organic materials such as the fullerenes
and chromophore-containing dendrimers may hold future promise for all-optical
switching at telecommunication wavelengths (eg, 1.55

µm) because of their large

two-photon absorption coefﬁcients (third-order optical nonlinearities) and their
low optical loss. Inorganic Kerr materials such as titanium sapphire are currently
used for mode-locked lasers and other applications.

Other reasons for heightened interest in organic electronic electrooptic mate-

rials include large electrooptic coefﬁcients (greater than 50 picometers/volt (pm/V)
at telecommunication wavelengths). Large electrooptic coefﬁcients permit volt-
ages from semiconductor electronics to drive electrooptic devices without the in-
troduction of ampliﬁers. Ampliﬁers are undesirable because of their bandwidth
limitations, noise ﬁgures, weight, bulk, and cost. Indeed, a major reason for in-
creased commercial interest in organic electrooptic materials is that drive voltages
of less than 1 V can be realized with simple broadband device structures (68,69).
Such drive voltages permit lossless telecommunication links to be achieved and
even permit realization of gain in electrical-to-optical-to-electrical signal trans-
duction. Another attractive feature of organic electrooptic materials is facile in-
tegration with semiconductor electronics and silica ﬁber optics (70–74). In this
article, both the state of development of organic electrooptic materials and the
principles that deﬁne material performance are reviewed.

The reader is referred elsewhere for reviews of organic photorefractive mate-

rials (75–86) and of photochromic and optomechanical materials (87). Electronic
electrooptic chromophores (such as charge transfer azobenzenes) have frequently

Vol. 2

ELECTROOPTICAL APPLICATIONS

285

been used for such applications, but the response times of these phenomena and
their conceptual basis are quite different from the phenomenon and applications
discussed here. (See also C

HROMOGENIC

OLYMERS

; L

IGHT

-E

MITTING

IODES

General Theoretical Principles

At the molecular level, the change in electron distribution (polarization) with
application of an electrical ﬁeld is given by

= p

+ α

i j

+ β

i jk

+ γ

i jkl

+ · · ·

(1a)

where p

is the permanent polarization of the molecular site “i” and E stands for

the local electric ﬁeld strength. The tensors

α, β, and γ are the molecular polar-

izability and molecular ﬁrst and second hyperpolarizabilities. For macroscopic or
bulk materials, the polarization equation is

= P

+ χ

(1)

i j

+ χ

(2)

i jk

+ χ

(3)

i jkl

+ · · ·

(1b)

It is assumed that the electric ﬁelds in equations (1a) and (1b) have frequen-

cies that are outside the domain of both electronic and lattice resonances of the
system. If only one electric ﬁeld is applied and it is in the form of a square wave
pulse (of amplitude E

), or an a-c ﬁeld (E

cos

ωt), or a d-c ﬁeld (of amplitude

), then equations (1a) and (1b) lead to the following expression for refractive

index n:

)

= (1/n

2
0

)

+ rE

+ RE

+ · · ·

(2)

where n

is the refractive index in the absence of the electric ﬁeld. The linear

and quadratic terms correspond to the Pockels and Kerr effects, respectively. The
Pockels effect is observable only for noncentrosymmetric materials; the electroop-
tic coefﬁcient r is related to the molecular ﬁrst hyperpolarizability

β and the

noncentrosymmetric order parameter

cos

θ by

= 2Nf(ω)β cos

θ /n

(3)

where N is the chromophore number density and f (

ω) is a local ﬁeld factor ac-

counting for the dielectric nature of the medium (host material). The electrooptic
coefﬁcient r is related to the drive voltage required to effect electrical-to-optical
signal transduction (with a Mach Zehnder interferometer) by

= λd/(n

)

(4)

is the voltage required for a

π phase shift of light passing through the arm

of the Mach Zehnder modulator to which an electric ﬁeld has been applied. The
optical wavelength of the transiting light is given by

λ, d is the gap between

the electrodes used to apply the electric ﬁeld, L is the interaction length of the

286

ELECTROOPTICAL APPLICATIONS

Vol. 2

electrical and optical ﬁelds (electrode length in the simplest case), and

is a modal

overlap integral. Equation (3) is greatly oversimpliﬁed as we have neglected the
vectorial nature of the interaction of the ﬁelds, but this equation is frequently
applicable to poled polymer materials.

From equations (2) and (3), it is readily seen that optimizing the drive volt-

age performance of electrooptic materials is a matter of optimizing

β, N, and

cos

θ . We shall proceed to discuss the optimization of these parameters and

the status of organic electrooptic materials, but before we do this, it is important
to indicate what additional material properties must be achieved for commercial
viability. These can be brieﬂy summarized as follows: (1) fast response time or
high operational bandwidths, (2) low optical loss including both material loss and
insertion (optical mode mismatch) loss, (3) good stability including thermal and
photochemical stability, (4) good mechanical properties, (5) good processability
and ease of integration with disparate materials such as silica ﬁbers, semiconduc-
tors, and metals, and (6) low production cost. For polymers to displace alternative
technologies such as modulated semiconductor lasers, lithium niobate electrooptic
modulators, or gallium arsenide (GaAs) electroabsorptive modulators, they must
exhibit competitive or superior performance characteristics in all material cate-
gories. Thus, in the following paragraphs, we consider auxiliary properties as well
as electrooptic activity.

Electrooptic Activity

It is clear that optimizing electrooptic activity involves optimizing the molecular
hyperpolarizability

β of organic chromophores. Quantum mechanics has provided

effective guidance for the improvement of molecular nonlinear optical (electroop-
tic) activity by several orders of magnitude in the past 10 years (88–97). If chro-
mophores are aligned into a noncentrosymmetric (acentric) lattice by electric ﬁeld
poling and if the interactions among chromophores are neglected, then the order
parameter becomes (in the high temperature approximation)

cos

θ = µF/5kT,

where

µ is the chromophore dipole moment, F is the applied electric poling ﬁeld

felt by the chromophores, k is the Boltzmann constant, and T is the poling tem-
perature (Kelvin). In this approximation, macroscopic electrooptic activity would
be expected to increase as

µβ. As can be seen from Figure 1, µβ values have been

systematically improved over the past decade from the value of the 1990 state-
of-the-art material Disperse Red (DR). Indeed, if current chromophores could be
organized into a pure chromophore lattice with perfect acentric order,

cos

θ =

1.0, then electrooptic coefﬁcients in excess of 1000 pm/V could be achieved. Such
electrooptic coefﬁcients would permit true millivolt drive voltages to be routinely
employed. Such materials would greatly exceed the performance of lithium niobate
(r

= 31 pm/V, V

= 5–6 V) and would likely displace liquid crystalline materials

in applications such as spatial light modulation and optical switching. However,
while the DR chromophore of Figure 1 exhibits an electrooptic coefﬁcient of 10–
15 pm/V and V

voltages of 12–20 V, the FTC chromophore exhibits a surprising

low electrooptic coefﬁcient of 35–50 pm/V and V

voltages of 3–5 V. Clearly, a prob-

lem existed/exists in translating large molecular electrooptic activity into usably
large macroscopic electrooptic activity.

Vol. 2

ELECTROOPTICAL APPLICATIONS

287

Fig. 1.

The variation of

µβ (dipole moment-ﬁrst hyperpolarizability product) with chro-

mophore structure is shown.

As is illustrated in Figure 2, the problem is now understood to derive from in-

termolecular electrostatic interactions among chromophores opposing the poling
ﬁeld in introducing the acentric order required for electrooptic activity (68,98–
102). The same interactions oppose introduction of acentric order by sequential
synthesis, self-assembly techniques where surface forces are used in place of a pol-
ing ﬁeld to guide the assembly of chromophores into an acentric lattice (103). In
this latter case, intermolecular electrostatic interactions cause the chromophores
to tip away from the normal to the surface in their assembly. The problem has been
successfully modeled using both equilibrium statistical mechanical methods and
kinetic Monte Carlo methods, with both approaches speciﬁcally treating the many-
body, long-range, spatially anisotropic interactions among chromophores. The in-
termolecular electrostatic interaction potential functions of equilibrium statistical
mechanical and Monte Carlo methods are found to be very similar and approxi-
mate the potential function introduced by Piekara in the 1930s (104). Theoretical
calculations explain a great many additional experimentally observed features
including the relatively weak dependence of maximum achievable electrooptic ac-
tivity upon electric poling ﬁeld strength and phenomena such as concentration
and electric-ﬁeld-dependent phase separation. The theoretical rationalization of
the dependence of achievable electrooptic activity upon the electronic and molec-
ular structure of chromophores and the processing conditions (applied poling ﬁeld

288

ELECTROOPTICAL APPLICATIONS

Vol. 2

Number Density, 10

/cm

Electrooptic Coefficient, pm

100

150

200

Fig. 2.

The experimental and theoretically predicted dependence of electrooptic coefﬁ-

cient for the FTC chromophore (see Fig. 1) in poly(methyl methacrylate) (PMMA) upon
chromophore number density (concentration in PMMA) is shown. Theoretical results are
shown for various shape approximations and for the neglect of intermolecular electro-
static interactions (the ideal gas model). - - - - - - Gas model; – – Sphere; —Prolate ellipsoid;

◦ Experimental.

strength, dielectric constant of the host polymer matrix, chromophore concentra-
tion, temperature, etc) has provided both good and bad news. The bad news is
that it will always be impossible to realize more than a fraction of the potential
electrooptic activity represented by a given chromophore structure. Electrooptic
coefﬁcients of a few hundred pm/V may be obtainable, but values of 1000 pm/V
or greater will be unobtainable with high dipole moment chromophores unless
some other force (eg, ionic) is used to overwhelm the effect of dipole–dipole in-
teractions. The good news is that theory provides guidance to quick optimization
of electrooptic activity for a given chromophore/polymer system. The concentra-
tion (number density) of chromophores leading to maximum electrooptic activity
is predicted by theory with high accuracy. Moreover, theoretical results argue
that the shape of chromophores should be altered from their normal prolate el-
lipsoidal shape to a more spherical shape. This is readily accomplished by adding
inert substituent (alkyl or alicyclic) groups to the waist of chromophores (68). An
even more promising route is that of dendrimer synthesis including dendrimer
structures where more than one chromophore is incorporated into the dendrimer.
Very simply, steric forces are used to prevent chromophores from achieving centric
order driven by dipole–dipole interactions. An example of an early version mul-
tichromophore dendrimer, which has led to electrooptic coefﬁcients of 60 pm/V at
1.55

µm telecommunication wavelength, is shown in Figure 3 (105).

Theoretical guidance has permitted polymeric electrooptic materials to be

fabricated that exhibit electrooptic coefﬁcients in excess of 100 pm/V (68,106)
at telecommunication wavelengths. With these materials, V

voltages of less

than 1 V have been achieved. This has been, without question, a very important

Vol. 2

ELECTROOPTICAL APPLICATIONS

289

Fig. 3.

(a) An early version multichromophore dendrimer (105) is shown. This cross-

linkable dendrimer yields an electrooptic coefﬁcient of 60 pm/V at 1.55

µm wavelength

that is stable under testing at 85

◦

C for 1000 h. (b) Schematic representation. M

= 4664;

Chromophore density: 33 wt%.

290

ELECTROOPTICAL APPLICATIONS

Vol. 2

result stimulating considerable interest from both large and small companies in
the telecommunications and defense ﬁelds. V

voltages on the order of 1 V permit

low cost complementary metal oxide semiconductor (CMOS) and high speed elec-
trochemical luminescent circuits to be used without ampliﬁers. Moreover, such low
voltages are important for the realization of transparent (low loss) communication
links. Indeed, the realization of deep optical modulation with low electrical drive
voltage can lead to gain in electrical-to-optical-to-electrical signal transduction. In
satellite telecommunication applications, the elimination of low noise ampliﬁers
is important for the reduction in launch weight as well as for reducing noise and
cost. The practical (commercial) importance of the advances in electrooptic activ-
ity rests upon two issues: (1) Can electrooptic activity be further improved and by
how much? (2) Can large electrooptic activity be achieved in conjunction with all
of the other required properties? The answer to the ﬁrst question is very likely
positive. Indeed, it is clear that electrooptic activity will be further improved by
at least another factor of 2. The answer to the second question will become clear
from the following paragraphs.

Bandwidth

The bandwidth (response time or highest frequency component of data that can
be handled) of devices fabricated from electrooptic materials is determined by
the index of refraction n and dielectric constant

ε of the materials from which

the devices are fabricated. For polymeric organic electrooptic materials,

ε = n

approximately. This means that low frequency electrical and high frequency op-
tical waves propagate with the same velocities in a material. This in turn means
that long interaction lengths can be employed, which leads to low drive voltage
requirements. For example, 3-cm electrodes were used to achieve sub 1-V V

val-

ues for polymeric modulators (68). For organic materials, voltage-length products
are currently of the order of 1–3 V

·cm. The typical index of refraction values

= 1.55–1.70) and dielectric constant values (ε = 2.4–3.0) for polymeric organic

materials have permitted electrooptic device bandwidths of 113 GHz to be demon-
strated using simple device structures (107–110). Pulsed measurements estab-
lish that the intrinsic material bandwidth response is of the order of 350 GHz
(essentially deﬁned by the phase relaxation time of the

π-electron system). De-

vice bandwidth-length product values could conceivably be as high as 350 GHz

·cm

with polymeric organic electrooptic materials; however, the difﬁculty of transmit-
ting electrical signals with frequency components above 120 GHz through metal
electrodes makes this a difﬁcult objective to achieve. The low dielectric constants
of organic materials also lead to very favorable power requirements (eg, typi-
cal P

values are on the order of 4 dB). Crystalline organic materials achieve

acentric order by exploiting ionic interactions to overwhelm the chromophore

π-

electron dipole interactions that favor centric order. These ionic interactions lead
to higher index of refraction and dielectric permittivity values, eg, 2.5 and 7.0
respectively.

For lithium niobate, dielectric constants are of the order of 28 while index

of refraction is of the order of 2.2. The bandwidth-length product is 10 GHz-cm
and the voltage-length product is 5 V-cm. By clever engineering (minimizing the

Vol. 2

ELECTROOPTICAL APPLICATIONS

291

time spent by electrical waves in high dielectric lithium niobate), lithium nio-
bate devices have been demonstrated to approximately 90 GHz. Commercial de-
vices are available from Lucent with bandwidths of 30 GHz, V

voltages of 6 V

(1550 nm), insertion loss of 6 dB, and temperature stability to 70

◦

C. The high

dielectric constant of lithium niobate results in high power dissipation, ie, P

32 dB currently. While high frequency operation is possible with devices fabri-
cated from lithium niobate, more sophisticated device engineering is required.
Moreover, such bandwidth is accomplished with sacriﬁce of drive voltage. The
high intrinsic bandwidth afforded by polymeric organic materials was early on
recognized as one of the greatest advantages of such materials. The fact that the
high bandwidth of organic electrooptic materials is now accompanied by low drive
voltage requirements argues for the serious commercial consideration of these
materials.

GaAs electroabsorptive modulators offer the possibility of high bandwidth

(50 GHz) and low drive voltage (1–2 V). However, these materials exhibit sub-
stantial chirp. It is impossible to change absorption without also changing index
of refraction. Moreover, GaAs devices are plagued by high insertion loss (9–12
dB). Lithium niobate modulators seem to be the currently preferred alternative
to GaAs modulators for high frequency (10–30 GHz) applications.

Modulated lasers, thermooptic switches, liquid crystalline modulators, bub-

ble jet switches, and MEMS devices suffer even greater bandwidth limitations. The
introduction of wavelength division multiplexing (WDM) has taken some of the
bandwidth pressure off of purely time division multiplexing (TDM) approaches to
handling telecommunication throughput. However, bandwidth is likely to remain
a serious issue for future telecommunication systems, and thus the bandwidth
performance organic and inorganic electrooptic materials will remain an issue of
active concern.

Optical Loss

One of the most attractive advantages of lithium niobate is exceptionally low
optical loss (material loss of 0.2–0.4 dB/cm and device insertion loss of 3–6 dB
at telecommunication wavelengths). Optical loss has been a problem with poly-
mer materials until recently when a better understanding of factors leading to
both material and insertion optical loss has permitted signiﬁcant improvement
in loss values (to material loss values of as low as 0.2 dB/cm and insertion loss
values as low as 4 dB at telecommunication wavelengths). Material optical loss
can be associated with either absorption loss or loss due to scattering. Absorption
loss can arise from either (or a combination of) interband electronic absorption
or C H (or N H, O H) vibrational overtone absorption at 1.3 or 1.55

µm wave-

lengths. Wavelength-dependent interband absorption is easily assessed by tech-
niques such as photothermal deﬂection spectroscopy. Appropriate chromophore
design can keep optical loss due to interband absorption below 1 dB/cm even
at 1.3

µm wavelength. Such loss is probably not a signiﬁcant factor at 1.55 µm

telecommunication wavelength. Vibrational absorption at 1.3 and 1.55

µm wave-

lengths is a much more serious problem. As with the production of low loss silica
ﬁber, the answer to the problem of intrinsic optical loss is to get the hydrogen out of

292

ELECTROOPTICAL APPLICATIONS

Vol. 2

the material. Without such reduction in hydrogen content, optical loss in organic
materials is typically limited to approximately 1 dB/cm at 1.3

µm wavelength and

somewhat higher values at 1.55

µm wavelength. With organic materials, reduc-

tion in hydrogen content typically means using partially ﬂuorinated materials or
materials constructed with functionalities, such as the cyanurate ligand, that do
not contain hydrogen (when the carbon positions are substituted). Using cyanu-
rate and ﬂuorinated dendrons, optical loss values as low as 0.2 dB/cm have been
obtained for electrooptic dendrimer materials. Fluorination of polymer materials
frequently involves unwanted changes in processability; for example, changes in
solubility in spin-casting solvents. Dendrimer synthesis has been demonstrated
to provide a convenient means of reducing optical absorption loss without paying
a price in material processability.

Scattering losses induced by processing have been a much more insidious

problem in the fabrication of low loss organic waveguide structures. Scattering
losses can arise from several sources. One of the most obvious is material inhomo-
geneity associated with spin casting. Any phase separation leads to disastrously
high loss values. Strong dipole–dipole interactions among chromophores can drive
phase separation. Fortunately, such interactions can be attenuated by appropriate
derivatization of chromophores. Such derivatization also improves the entropy of
mixing (and hence the solubility) of chromophores in spin casting solvents. Electric
ﬁeld poling and the concurrent cross-linking of materials to lock-in poling-induced
electrooptic activity can also lead to optical loss (71,100,106,111–113). Two types of
optical loss are typically associated with electric ﬁeld poling. The ﬁrst arises from
electric-ﬁeld-induced surface damage. This is simply a form of dielectric break-
down and can be avoided by keeping corona poling ﬁeld strengths below damage
thresholds. It is important to keep in mind that the harder (more cross-linked) the
polymer lattice, the higher the damage threshold. Thus, electric ﬁeld poling and
lattice hardening are typically carried out using a stepped temperature/electric
ﬁeld protocol (71,100,112,114–116). The second type of poling-induced scattering
loss is associated with introduction of inhomogeneity in the material as the re-
sult of electric ﬁeld poling. Such inhomogeneity can arise from an electrophoretic
effect or from poling-induced birefringence. If defects (such as dust particles) ex-
ist that are not inﬂuenced by the poling ﬁeld, then poling can enhance the index
difference between these defect regions and the electrooptic material leading to en-
hanced scattering loss (117). Lattice hardening (cross-linking) chemical reactions
can also be a source of material inhomogeneity that leads to increased scatter-
ing loss. This is particularly true with thermosetting reactions such as urethane
chemistry where a variety of oligomeric structures can exist and grow at differ-
ent rates (111). Some of these structures can become incompatible and phase
separate. Moreover, some cross-linking chemistries, such as urethane chemistry,
are susceptible to atmospheric moisture. The reaction with water from the at-
mosphere can upset the one-to-one stoichiometry of lattice hardening reactions
leading to the buildup of incompatible components and phase separation. The
occasionally observed cloudy appearance of polyurethane electrooptic polymers
is typically indicative of problems with atmospheric moisture. However, when
sources of optical loss associated with electric ﬁeld poling and lattice hardening
are taken into account, optical loss values on the order of 1 dB/cm are readily
obtained (71,100,106,111–113).

Vol. 2

ELECTROOPTICAL APPLICATIONS

293

Optical loss can also arise from fabrication of buried channel waveguide

structures by reactive ion etching (RIE) techniques and by the deposition of
cladding layers where the solvent employed in spin casting the cladding layer
causes solvent pitting of the electrooptic ﬁlm layer. Both of these loss mecha-
nisms can be controlled by lattice hardening of the electrooptic material and by
control of processing conditions. Control of the kinetic energy of reactive ions in
RIE (118,119) can keep optical loss values due to waveguide wall roughness to
insigniﬁcant values (0.01 dB/cm). Cladding layers pose an unanticipated prob-
lem for the utilization of organic-electrooptic materials. While cladding materials
have been utilized that make no problematic contribution to optical loss, prob-
lems in device operation can frequently be traced to cladding materials rather
than to the electrooptic materials. Conducting cladding materials can be par-
ticularly problematic. In general, one wants the electrical conductivity of the
cladding material to be higher than that of the electrooptic material so as to
drop the poling ﬁeld across the cladding layer. However, if the cladding mate-
rial has signiﬁcant optical loss, the penetration of the propagating optical ﬁeld
into the cladding layer will result in unwanted attenuation of the transiting
beam. The cladding material can also undergo photochemical damage and may
do so even when the electrooptic material does not show damage. This can man-
ifest in time-dependent loss of mode conﬁnement when no change in electrooptic
modulation efﬁciency is observed. A more subtle problem that can arise with
conducting cladding materials is that of photoconductivity (or optical power-
dependent conductivity). Electrooptic chromophores are very polarizable mate-
rials and can inject charge into a conducting medium (71,112). Thus, an optical
power-dependent conductivity of the cladding material can result if that mate-
rial is susceptible to accepting charge and transporting that charge. Such ef-
fects are typically observed only at high power levels in devices such as high
frequency oscillators based on electrooptic modulator technology. Of course, opti-
cal ﬁelds must be kept away from lossy metal electrodes. Again, if care is taken
in the design and execution of device structures, loss from waveguide fabrica-
tion, cladding deposition, and electrode deposition can be kept to insigniﬁcant
values, and overall active waveguide loss values of 1 dB/cm can be routinely
achieved.

The greatest loss issue to be faced in the commercial implementation of or-

ganic electrooptic devices is that of mode mismatch loss when coupling silica ﬁber
and polymeric electrooptic modulator waveguides. For 1.3

µm telecommunication

wavelength operation, silica optical ﬁber waveguides are cylindrical with spheri-
cal cross-sectional core diameters on the order of 10

µm, while the cross-sectional

shapes of polymeric electrooptic modulator waveguides are elliptical with heights
of 1–2

µm. Butt coupling of such active and passive waveguides leads to optical

loss on the order of 4 dB per transition. The coupling problem is readily addressed
using spherical lens or tapered transitions (70,71,112,120). A tapered transition
such as shown in Figure 4 reduces coupling loss to a few tenths of a decibel. Thus,
total insertion losses of 4–5 dB have been obtained for polymeric materials ex-
hibiting material loss on the order of 1 dB/cm. Tapered transition structures can
be fabricated by employing offset, gray scale, and shadow ion masking techniques
(70), which are to be discussed later. Such lithographic techniques are also useful
for constructing sophisticated 3-D circuits (70).

294

ELECTROOPTICAL APPLICATIONS

Vol. 2

Length

Constant 4 dB/cm

Hybrid core

Loss

, dB

Taper Length,

1000

2000

0.5

1.0

1.5

(active)

(passive)

Large length

Small length

material loss

radiation loss

Fig. 4.

A low loss polymeric electrooptic modulator transition structure is shown. The

optimum taper length is deﬁned by the trade-off between acceptable radiation loss and
material loss.

Stability

For electrooptic materials (eg, liquid crystalline materials) based on index of re-
fraction changes associated with molecular reorientation, one wants facile reori-
entation to permit fast switching times. Organic photorefractive materials also
exhibit enhanced response when molecular reorientation contributes to observed
index changes. In contrast, reorientation destroys the electric-ﬁeld-poling-induced
electronic electrooptic activity of chromophore-containing polymeric materials.
When one speaks of thermal instability of electronic electrooptic activity for or-
ganic materials, one is referring to reorientational (molecular rotational) dynam-
ics that destroy poling-induced acentric order. There are two general approaches
to improving the stability of poling-induced order. One is to place chromophores
in polymer materials, such as polyimides or polyquinolines, characterized by high
glass-transition temperatures. This may be accomplished either by physically dis-
solving chromophores in polymers to produce composite materials or by covalently
coupling chromophores to polymer lattices (71,112,121–134). The stability will be
deﬁned by both the segmental ﬂexibility of the polymer backbone (the glass tran-
sition temperature) and the number (and positioning) of covalent bonds coupling
the chromophore to the polymer lattice. The stability of poling-induced acentric
chromophore order (electrooptic activity) can be readily assayed by monitoring
second harmonic generation (another phenomena related to

(2)

) as a function of

heating the sample (see Fig. 5) (135). This is a nonlinear optical analogue of ther-
mal gravimetric analysis except that molecular reorientation rather than mate-
rial decomposition accounts for the loss of second-order nonlinear optical activity.
As shown in Figure 5, thermal stability of electrooptic activity is increased with
cross-linking. When dynamic thermal stability is increased to temperatures such
as 170

◦

C as shown in Figure 5, then long-term stability is typically observed for

periods of several thousand hours even when devices are operated at 100

◦

C. Seg-

mental ﬂexibility of polymers is one of the most serious problems with relaxation

Vol. 2

ELECTROOPTICAL APPLICATIONS

295

Temperature,

°C

Nor

maliz

ed d (2

)

100

150

200

0.0

0.2

0.4

0.6

0.8

1.0

Uncross-linked

Cross-linked

Fig. 5.

Second harmonic generation is shown as a function of temperature for heating

at a rate of 10

◦

C per minute for two samples. The uncross-linked sample (chromophore is

attached covalently by one end to a PMMA backbone) starts to lose second-order optical
nonlinearity (acentric chromophore order) before 100

◦

C. The cross-linked sample corre-

sponds to both ends of the chromophore being covalently coupled to the polymer lattice.
For this sample, thermally stable second-order optical nonlinearity is observed until nearly
170

◦

of poling-induced acentric order. Dendrimer materials appear to have distinct
advantages for enhancing the thermal stability of poling-induced electrooptic ac-
tivity. Not only can the covalent bonds and steric interactions of dendrimers hold
chromophores away from each other (inhibiting centric ordering) but dendrimer
structures can also oppose thermal relaxation. The dendrimer material of Figure
3 is observed to exhibit stable electrooptic activity for 1000 h at 85

◦

C. Greater ther-

mal stability should be achieved by greater attention to the ﬂexibility of dendrons
and by better control of cross-link density. In comparison, the thermal stability of
lithium niobate devices is typically listed as 90

◦

C while gallium arsenide devices

are speciﬁed as stable to 80

◦

Preliminary studies of the stability of organic electrooptic materials to high

energy (gamma ray) radiation and particles (protons) carried out by Lockheed
Martin (Palo Alto) and by researchers at the Air Force Research Laboratory
(Wright Patterson Air Force Base) suggest that space application of polymeric
electrooptic materials should proceed without problems.

Photochemical stability is more difﬁcult to assess. Lithium niobate is

speciﬁed as having optical power handing capabilities to 250 mW while gal-
lium arsenide is speciﬁed as having a maximum power handling capability
of 30 mW. However, such power handling capability is typically speciﬁed for
packaged materials where oxygen has been excluded. The photochemical sta-
bility of organic materials is dramatically affected by the presence of oxygen.
Exclusion of oxygen typically improves photochemical stability by several or-
ders of magnitude. The addition of oxygen (radical) scavengers can also im-
prove photochemical stability (136). Photochemical stability can also improve

296

ELECTROOPTICAL APPLICATIONS

Vol. 2

with lattice hardening (100) although this will depend upon the porosity of
the polymer lattice, which may not always correlate simply with cross-link
density. Unfortunately, controlled evaluation of photochemical stability under
conditions appropriate for commercial applications have not been carried out
for polymeric electrooptic materials; thus, photochemical stability remains the
most serious question concerning the commercialization of organic electrooptic
materials.

Mechanical Properties

Polymeric electrooptic materials have the advantage of being conformal and of
exhibiting effective bonding to a variety of surfaces. Moreover, by control of
cross-link density, the mechanical properties of polymeric electrooptic materials
can be controlled. Electrooptic device operation can be carried out over a wide
range of temperatures. Eventually, different coefﬁcients of thermal expansion
may cause delamination problems upon thermal cycling over wide temperature
ranges, but such problems have not been noted to the present. A unique fea-
ture of polymeric systems is that good mechanical properties are obtained for
lightweight materials. Weight is an issue in satellite telecommunication and re-
connaissance systems. Regarding weight issues in satellite telecommunications
and reconnaissance operations, the low drive voltage requirements of polymeric
electrooptic devices permit the elimination of low noise ampliﬁers, which in turn
results in substantial weight reduction. Thus, the weight issue is not the rela-
tive weights of lithium niobate and polymeric electrooptic materials but the rel-
ative weights of total systems fabricated from these materials. Also, the critical
weight saving comes from using photonic (optical ﬁber) versus electrical (coaxial)
cables.

Processing and Integration

Polymeric materials are amenable to facile processing by RIE (70,71,112,119,120)
and photolithographic techniques (71,112,137). Use of offset, gray scale, and
shadow ion lithographic masks and RIE has permitted implementation of ver-
tical waveguide transitions and the fabrication of 3-D circuits permitting the
ready integration of passive and active optical waveguide components (70,71).
Polymeric electrooptic materials are also very compatible with semiconductor pro-
cessing, and electrooptic circuitry has been fabricated on top of very large-scale
integration (VLSI) semiconductor chips (71,112). Indeed, through use of planariz-
ing polymers, electrooptic modulators have been vertically integrated (in high
density) with VLSI chips. Integration has been accomplished without the pertur-
bation of the performance characteristics of either electronic or optical (electroop-
tic) circuitry. One of the advantages of polymeric electrooptic materials is this
ready integration with semiconductor electronics and the ability to fabricate a
large number of modulator waveguides on a single chip. With lithium niobate,
such integration is not possible and connection to electronic circuitry is typi-
cally made by wires and ﬂip chip bonding. By use of tapered transitions and

Vol. 2

ELECTROOPTICAL APPLICATIONS

297

hologram photonic crystla

prism electrode

domain inverted prism

ground

electrode

EO polymer waveguide

layer

cladding
polymer

Light

out

silicon substrate

(b)

(a)

Propagation Ligh

−70°

−7°

+70°

+7°

Incident Light

Fig. 6.

Large-angle beam steering is shown for a device based on a polymeric electrooptic

material (a cross-linkable version of the CLD chromophore discussed in Ref. 68). (a) EO
waveguide prism introduces a small deﬂection angle to initialize the beam scanning. The
half-circle 2-D photonic crystal region is embedded into the waveguide so that the deﬂection
angle is “ampliﬁed” as the light passes through the crystal region. 3-D scanning can also be
provided if a 3-D structure is built. (b) The experimental observation of the angle sensitivity.

spherical lens, electrooptic waveguides have been coupled to silica ﬁber optics.
Again, coupling can be achieved in a more straightforward manner than with
lithium niobate.

An example of sophisticated device development is shown in Figure 6. Large-

angle (

−70

◦

+70

◦

) beam steering is demonstrated by implementing a series of

cascaded electrooptic prisms, which initiate beam steering and feed light into a
photonic band gap lattice (138,139). The photonic band gap lattice is fabricated by
holography in a photopolymer material (see Fig. 7). Beam steering is accomplished
with more than an order of magnitude lower drive voltage than required using

298

ELECTROOPTICAL APPLICATIONS

Vol. 2

Recording beam #1 Recording beam #

−1

hologram pattern

being formed

cladding polymer

silicon substrate

waveguide layer formed by

photopolymer

hexagonal wavevector lattice of the

combination of recording beams

−1′′

−1

−1′

′′

′

Fig. 7.

The holographic production of the polymeric photonic band gap lattice used in the

beam steering device of Figure 6 is shown.

lithium niobate materials. Through the cascaded prism electrodes, the deﬂection
angle is given for steering of a TM polarized beam by

θ = n

2
0

/d)(L/h)

(5)

where L and h are the length and width of the prism array, respectively. V is the
applied driving voltage and d is the electrode spacing. The electrooptic coefﬁcient
in the direction of the driving ﬁeld is r

, while n

is the voltage independent index

of refraction. In the present example, L/h is approximately 80 with L

= 16 mm

and h

= 200 µm. The model used to arrive at equation (4) is based only upon the

total phase retardation across the wave front. The deﬂection angle thus depends
only on the overall dimension of the prisms, since the accumulated phase differ-
ence across the wave front is independent of how the deﬂector is subdivided into
individual prisms, as long as the interfaces between adjacent prisms are straight
lines.

Devices

The most simple device conﬁgurations utilizing electrooptic materials are shown
in Figure 8 (122). These include the Mach Zehnder and birefringent modula-
tors and the directional coupler. As conﬁgured in Figure 8, the ﬁrst two de-
vices are amplitude modulators. A somewhat higher drive voltage is required
for the birefringent modulator than for the Mach Zehnder modulator. This is
because the V

voltage of the Mach Zehnder relates directly to the principle

component (r

) of the electrooptic coefﬁcient. The V

voltage of the birefringent

modulator relates to the difference between the major and minor components

Vol. 2

ELECTROOPTICAL APPLICATIONS

299

OUTPUT 1

OUTPUT 2

INPUT 1

INPUT 2

Polarizer

° Polarized

Input

Fig. 8.

Mach Zehnder interferometer (top), birefringent modulator (middle), and direc-

tional coupler (bottom) device conﬁgurations are shown.

3
3

− n

3
1

). A still higher drive voltage is required for the directional coupler,

and the relationship to material electrooptic activity is somewhat more complex
(122).

We have already shown a large-angle spatial light modulator based upon

organic electrooptic and photonic band gap materials (Figs. 6 and 7). Such beam
steering without moving parts is relevant for holographic optical data storage,
switching in all optical networks, advanced laser radar (ladar), photonic phased
array antennas, optical sensors, and laser printers. Spatial light modulation ex-
ploiting electronic electrooptic activity has not only the advantage of greater speed
but also avoids instability due to overshoot associated with the momentum of
mechanical beam steering (MEMS) devices. The fact that such beam steering
can be accomplished with application of driving voltages in the range of volts
rather than kilovolts (as is the case for lithium niobate) is also a very attrac-
tive feature. For ﬁber optic switching, spatial light modulation has the advan-
tage that a large number of ﬁber channels can be selectively addressed using
only one electrode. Overall, electrooptic beam steering offers a simpliﬁed op-
erating scheme, low driving voltage, high switching speed, small size, and low
cost.

Ultra high bandwidth analogue-to-digital (A/D) conversion has been accom-

plished utilizing polymeric electrooptic materials in several different device con-
ﬁgurations. Analogue-to-digital conversion employing a cascaded series of Mach
Zehnder modulators has been discussed elsewhere (112). Analogue-to-digital con-
version can also be accomplished by time stretching (140). In this application, a
femtosecond pulse is stretched (by a dispersive medium) to match the length of an
incoming data stream containing high frequency information. The data stream is
transduced as an amplitude modulation (by a polymeric electrooptic modulator)

300

ELECTROOPTICAL APPLICATIONS

Vol. 2

YIG Tuned

Bandpass Filter

Optical

Spectrum

Analyzer

Polarizing

Beamsplitting

Cube

Diode Pumped

Nd:YAG Laser

(1.3

µm)

Optical

Isolator

Low Noise

Amplifier

Collimating

Lens

Spectrum

Analyzer

20 dB

Coupler

Plate

Coupler

Fig. 9.

A broadband, ultrastable oscillator based on polymeric electrooptic materials is

shown.

onto the optical pulse. The pulse is then further time stretched to a length compat-
ible with conventional A/D converters. The pulse is then converted to an electrical
signal by a diode detector and A/D conversion is effected. Such time stretching
permits A/D conversion at rates of 100 Gbit/s.

Another prototype device, which has been demonstrated by researchers at

Paciﬁc Wave, is that of a high frequency oscillator (see Fig. 9). In this oscillator,
the stability of the oscillator is determined by the length (Q) of the optical circuit
(loop) and can thus be very high indeed.

A variety of phased array systems have been fabricated based on polymeric

electrooptic materials (141,142). One conﬁguration is based on the photonic phase
shifter shown in Figure 10. This provides a very linear phase shift as a function
of control d-c voltage. Optical signals of controlled phase are thus sent to various
radiating antenna elements. The optical signals are converted to radiofrequency
signals by diode detectors.

High frequency (including secured communication by frequency hopping)

telecommunication applications are receiving considerable attention (143). As can
be seen in Table 1, the bandwidth limitation of current electronic components is
the limiting factor to achieving increased bandwidth performance with polymeric
electrooptic materials. In Figure 11, a device conﬁguration capable of 130-GHz
bandwidth operation is shown.

Other demonstrated prototype devices include optical gyroscopes (144),

broadband acoustic spectrum analyzers (145), 1

× 2 Y-fed directional couplers

(146) and polarization-insensitive electrooptic modulators (147,148). For long-
haul telecommunication applications, it is difﬁcult to maintain polarization con-
trol; thus, a need exists for polarization-insensitive modulators. Indeed, polariza-
tion insensitivity was one of the advantages claimed for gallium arsenide elec-
troabsorptive modulators. By using different poling schemes, overall polarization
insensitivity has been achieved for polymeric modulators (147,148).

Vol. 2

ELECTROOPTICAL APPLICATIONS

301

Table 1. The Bandwidth Characteristics of Various Electrical Interfaces

Electrical

Bandwidth,

Technology

Modulator type

interface

GHz

availability

Reliability

Cost

Traditional

K-coax

0–40

Simple, commercial

Proven

Low

M-Z modulators

V-coax

0–65

Simple, commercial

Proven

Low

W-coax

0–110

Ongoing at TACAN

R&D

Low

WR-15

50–75

Complicated, customer

R&D

High

WR-10

75–110

Complicated, customer

R&D

High

Frequency shifter

RF in

Phase control

out

sin(

= E

cos(

Fig. 10.

A schematic representation of a photonic radiofrequency phase shifter is shown.

LO signal for
65

−130 GHz

Low
band

− 65 GHz

out2

out1

DC2

DC1

Fig. 11.

A 130-GHz telecommunications device is shown.

Commercialization and Cost

As with lithium niobate, the cost of devices based on polymeric electrooptic modu-
lators is in the packaging rather than in the materials. A Lucent lithium niobate
modulator sells for approximately $6000. Polymer modulator vendors such as Pa-
ciﬁc Wave Industries (Los Angeles, California) are quoting comparable prices. The
low material cost is associated with the fact that modulator devices are thin-ﬁlm

302

ELECTROOPTICAL APPLICATIONS

Vol. 2

devices and a chip containing 6–100 modulators requires very little polymeric
electrooptic material.

A number of small companies such as Lumera (Bothel, Washington), Paciﬁc

Wave Industries (Los Angeles, California), Radiant Research (Austin, Texas), IP-
ITEK/TACAN (Carlsbad, California), KVH Industries (Tinley Park, Illinois), etc
are exploring the commercialization of electrooptic devices based on polymeric ma-
terials. Lockheed Martin has pursued research in polymeric electrooptic materials
for more than a decade and continues to be a leader in the ﬁeld. Lucent has recently
resumed research on polymeric electrooptic materials. A number of telecommu-
nication companies including Nortel, Agilent, Cisco, and JDS-Uniphase are cur-
rently evaluating polymeric electrooptic technology to determine the viability of
R&D and commercial activity in this area.

The market for inorganic electrooptic and electroabsorptive modulators is

currently in the hundreds of millions of dollars per year with wait times of several
months for delivery of lithium niobate modulators. The perceived need for greater
bandwidth in the telecommunications industry is likely to drive demand even
higher.

Future Prognosis

Provided that no unexpected problems arise that would inhibit the commercializa-
tion of polymeric electrooptic technology, the prognosis is very good. The greatest
uncertainty involves whether or not organic electrooptic materials will exhibit ad-
equate photochemical stability over many years of in-ﬁeld operation. Preliminary
studies suggest that exclusion of oxygen and hermetic sealing will yield devices ca-
pable of exhibiting photochemical stability over many years for exposure to optical
ﬁelds at telecommunication wavelengths and powers. Almost certainly, electroop-
tic coefﬁcients will continue to be improved. New and improved chromophores are
being synthesized with great regularity (149). Devices operating with sub 1-V V

values should become commonplace. Utilization of photonic band gap materials
and controlled coupling to resonated structures should permit further reduction in
drive voltage requirements (possibly even to microvolt levels); such reduction is, of
course, at the expense of some reduction in response time (bandwidth). Low drive
voltage requirements, together with high bandwidth capabilities and the ease of
integration with semiconductor electronics and silica transmission ﬁbers, should
be a powerful driver for the deployment of polymeric electrooptic technology in
telecommunication, defense, transportation, and display industries. However, the
extent of utilization of polymer electrooptic technology in the telecommunication
industries will depend upon perceived bandwidth needs. If low cost alternatives
to high information throughput with reduced bandwidth requirements (such as
expanded use of WDM technology) become popular, limited bandwidth technolo-
gies such as thermooptic, liquid crystalline, MEMS, and modulated laser technolo-
gies may continue to dominate sales in the telecommunications area. However,
even if deployment is slow in telecommunications, defense industries should ﬁnd
use for the exceptional capabilities of polymeric electrooptic materials for applica-
tions such as radar, electronic counter measure, sensor, and display systems. Other

Vol. 2

ELECTROOPTICAL APPLICATIONS

303

niche applications including biomedical sensors, analytical instrumentation, and
high frequency test equipment are likely to develop.

The main limitation to general deployment of organic electrooptic materials

is that they have been optimized to this point in time for applications at telecom-
munication wavelengths. Inorganic materials such as lithium niobate are still the
materials of choice for applications at visible wavelengths. Such inorganic modula-
tors are likely going to continue to dominate the laboratory market for modulators
to be used with research laser systems.

Nomenclature

ith component of the molecular polarization vector in the presence of
electrical ﬁelds

ith component of the permanent polarization of the molecule

the local electric ﬁeld strength

the local electric poling ﬁeld strength

molecular polarizability tensor

molecular ﬁrst hyperpolarizability tensor

molecular second hyperpolarizability tensor

ith component of the bulk polarization vector in the presence of electrical
ﬁelds

ith component of the bulk polarization vector in the absence of external
ﬁelds

applied voltage

(1)

ﬁrst-order electric susceptibility tensor

(2)

ijk

second-order electric susceptibility tensor

(3)

ijkl

third-order electric susceptibility tensor

refractive index in the presence of an applied electric ﬁeld

refractive index in the absence of an applied electric ﬁeld

Pockels coefﬁcient or electrooptic coefﬁcient

electrooptic coefﬁcient in the direction of the applied ﬁeld (principal ele-
ment)

electrooptic coefﬁcient orthogonal to the applied ﬁeld direction (minor
element)

Kerr factor

applied electric ﬁeld amplitude

number density

f (

ω)

local ﬁeld factor

cos

θ acentric order parameter

applied voltage required to produce a phase shift of

wavelength of light

electrode gap

interaction length of electrical and optical ﬁelds in material

modal overlap integral

beam deﬂection angle

304

ELECTROOPTICAL APPLICATIONS

Vol. 2

BIBLIOGRAPHY

“Electrooptical Applications” in EPSE 2nd ed., Vol. 5, pp. 756–771, by M. Thakur and S.
Tripathy, GTE Laboratories, Inc.

1. J. Kerr, Phil. Mag. 50, 337 (1875).
2. F. Pockels, Lehrbuch der Kristalloptic, Teubner, Leipzig, Germany, 1906.
3. G. T. O’Konski, in N. M. Bikales, ed., Encyclopedia of Polymer Science and Technology,

1st ed., Wiley-Interscience, New York, Vol. 9, 1968, p. 551.

4. B. R. Jennings, IEE Proc. 130, 212 (1983).
5. C. T. O’Konski, ed., Molecular Electrooptics, Marcel Dekker, Inc., New York, 1976.
6. S. Krause, ed., Molecular Electro-Optics, Plenum Press, New York, 1980.
7. I. P. Kaminow, An Introduction to Electrooptic Devices, Academic Press, Inc.,

New York, 1974.

8. J. Wilson and J. F. B. Hawkes, Opto-electronics, An Introduction, Prentice-Hall,

Englewood Cliffs, N.J., 1983.

9. N. Bloembergen, Nonlinear Optics, Benjamin, New York, 1965.

10. P. N. Prasad and D. J. Williams, Introduction to Nonlinear Optical Effects in Molecules

and Polymers, John Wiley & Sons, Inc., New York, 1991.

11. L. R. Dalton, in J. I. Kroschwitz and M. Howe-Grant, eds., Kirk-Othmer Encyclopedia

of Chemical Technology, Vol. 17, John Wiley & Sons, Inc., New York, 1996, p. 287.

12. D. L. Wise, G. E. Wnek, D. J. Trantolo, T. M. Cooper, and J. D. Gresser, eds., Electrical

and Optical Polymer Systems, Marcel Dekker, Inc., New York, 1998.

13. H. Nalwa and S. Miyata, eds., Nonlinear Optics of Organic Molecules and Polymers,

CRC Press, Inc., Boca Raton, Fla., 1997.

14. Y. R. Shen, Nonlinear Optics, John Wiley & Sons, Inc., New York, 1984.
15. R. W. Boyd, Nonlinear Optics, Academic Press, Inc., New York, 1993.
16. T. J. Bunning and co-workers, Annu. Rev. Mater. Sci. 30, 83 (2000).
17. R. Dabrowski, Ferroelectrics 243, 1 (2000).
18. S.-M. Chang and co-workers, Mater. Sci. Eng. C 12(1/2), 111 (2000).
19. T. Ikeda and A. Kanazawa, Bull. Chem. Soc. Jpn. 73, 1715 (2000).
20. T. Fuhrmann and J. H. Wendroff, Int. J. Polym. Mater. 45, 621 (2000).
21. K. M. Blackwood, Int. J. Polym. Mater. 45, 585 (2000).
22. N. Boiko and V. Shibaev, Int. J. Polym. Mater. 45, 533 (2000).
23. F. Simoni and O. Francescangeli, Int. J. Polym. Mater. 45, 381 (2000).
24. V. Shibaev, Int. J. Polym. Mater. 45, 307 (2000).
25. J. Zubia and J. E. Ecenarro, Int. J. Polym. Mater. 45, 503 (2000).
26. E. Lizuka, Int. J. Polym. Mater. 45, 191 (2000).
27. G. Abbate and J. M. Oton, Adv. Mater. 12, 459 (2000).
28. D. A. Higgins, Adv. Mater. 12, 251 (2000).
29. D. J. Broer and co-workers, Macromol. Symp. 154, 1 (2000).
30. T. J. Sluckin, Contemp. Phys. 41, 37 (2000).
31. R. L. Wisnieff and J. J. Ritsko, IBM J. Res. Dev. 44, 409 (2000).
32. M. O’Neill and S. M. Kelly, J. Phys. D: Appl. Phys. 33, R67 (2000).
33. S. R. Collinson and D. W. Duncan, Perspect. Supramol. Chem. 285 (1999).
34. W. A. Crossland and co-workers, Proc. SPIE 3955, 70 (2000).
35. M. Hird, J. W. Goodby, and K. J. Toyne, Proc. SPIE 3955, 15 (2000).
36. J. W. Goodby and co-workers, Proc. SPIE 3955, 2 (2000).
37. B. L. Feringa and co-workers, Chem. Rev. 100, 1789 (2000).
38. K. Ichimura, Chem. Rev. 100, 1847 (2000).
39. S. Susumu, Opt. Rev. 6, 471 (1999).
40. F. Bloisi and L. Vicari, Recent Res. Dev. Macromol. Res. 4, 113 (1999).
41. T. Tsukada, Springer Ser. Mater. Sci. 7 (2000).

Vol. 2

ELECTROOPTICAL APPLICATIONS

305

42. J. W. Goodby, Curr. Opin. Solid State Mater. Sci. 4, 361 (1999).
43. I. C. Khoo, J. Nonlinear Opt. Phys. Mater. 8, 305 (1999).
44. G. Scalia and co-workers, Proc. SPIE 3800, 112 (1999).
45. R. A. Herke, M. H. Anderson, and T. G. Baur, Proc. SPIE 3800, 45 (1999).
46. R. W. Sabnis, Displays 20, 119 (1999).
47. S. Naemura, Mater. Res. Soc. Symp. Proc. 559, 263 (1999).
48. C. C. Bowley and co-workers, Mater. Res. Soc. Symp. Proc. 559, 97 (1999).
49. W. Haase, D. Ganzke, and E. P. Pozhidaev, Mater. Res. Soc. Symp. Proc. 559, 15 (1999).
50. G. W. Gray and S. M. Kelly, J. Mater. Chem. 9, 2037 (1999).
51. W. M. Gibbons and co-workers, Proc. SPIE 3635, 32 (1999).
52. P. Palffy-Muhoray, Phys. Prop. Liq. Cryst. 467 (1999).
53. L. M. Blinov, Phys. Prop. Liq. Cryst. 375 (1999).
54. D. Dunmur and K. Toriyama, Phys. Prop. Liq. Cryst. 113 (1999).
55. D. Dunmur and K. Kazuhisa, Phys. Prop. Liq. Cryst. 102 (1999).
56. G. W. Gray, Phys. Prop. Liq. Cryst. 1 (1999).
57. E. Lueder, Proc. Electrochem. Soc. 98 (1999).
58. A. Sinha, T. A. P. Rao, and V. R. K. Murthy, Liq. Cryst. 27, 191 (2000).
59. Ch. Bosshard, Springer Ser. Opt. Sci. 7 (2000).
60. M. Samoc, A. Samoc, and B. Luther-Davies, Synth. Met. 109, 79 (2000).
61. M. Samoc and co-workers, MCLC S&T, Sect. B.: Nonlinear Opt. 20, 183 (1999).
62. S. R. Marder and co-workers, Science 276, 1233 (1997).
63. R. L. Nelson and R. W. Boyd, Appl. Phys. Lett. 74, 2417 (1999).
64. D. R. Greve and co-workers, Synth. Met. 102, 1533 (1999).
65. F. Cherioux and co-workers, Chem. Commun. 20, 2083 (1999).
66. M. G. Kuzyk, Opt. Lett. 25, 1183 (2000).
67. K. A. Drenser and co-workers, J. Phys. Chem. 103, 2290 (1999).
68. Y. Shi and co-workers, Science 288, 119 (2000).
69. Y. Shi and co-workers, Appl. Phys. Lett. 77, 1 (2000).
70. S. M. Garner and co-workers, IEEE J. Quantum Electron. 35, 1146 (1999).
71. L. Dalton and co-workers, Ind. Eng. Chem. Res. 38, 8 (1999).
72. W. H. Steier and co-workers, Chem. Phys. 245, 487 (1999).
73. R. A. Norwood and co-workers, Proc. SPIE 3281, 2 (1998).
74. S. Kalluri and co-workers, IEEE Photon. Technol. Lett. 8, 644 (1996).
75. B. Kipplen and co-workers, ACS Symp. Ser. 72, 204 (1999).
76. R. J. Twieg and co-workers, Mater. Res. Soc. Symp. Proc. 561, 119 (1999).
77. S. R. Marder and co-workers, Nature 388, 845 (1997).
78. B. Kippelen and co-workers, J. Am. Chem. Soc. 119, 4559 (1997).
79. B. Kippelen and co-workers, Proc. SPIE 3144, 176 (1997).
80. B. Kippelen and co-workers, Science 279, 54 (1998).
81. E. Hendrickx and co-workers, Appl. Phys. Lett. 72, 1670 (1998).
82. Z. Peng, A. Gharavi, and L. Yu, J. Am. Chem. Soc. 119, 4622 (1997).
83. W. Li and co-workers, Adv. Mater. 10, 927 (1998).
84. W. E. Moerner, A. Grunnet-Jepsen, and C. L. Thompson, Annu. Rev. Mater. Sci. 27,

585 (1997).

85. A. Grunnet-Jepsen and co-workers, J. Opt. Soc. Am. B 15, 901 (1998).
86. A. Grunnet-Jepsen, C. L. Thompson, and W. E. Moerner, Opt. Commun. 145, 145

(1998).

87. J. A. Delaire and K. Nakatani, Chem. Rev. 100, 1817 (2000).
88. S. R. Marder, D. N. Beratan, and L. T. Chang, Science 252, 103 (1991).
89. S. R. Marder and J. W. Perry, Science 263, 1706 (1993).
90. D. R. Kanis, M. A. Ratner, and T. J. Marks, Chem. Rev. 94, 195 (1994).
91. I. D. L. Albert, T. J. Marks, and M. A. Ratner, J. Am. Chem. Soc. 119, 6575 (1997).

306

ELECTROOPTICAL APPLICATIONS

Vol. 2

92. B. Champage and B. Kirtman, Chem. Phys. 245, 211 (1999).
93. J. Lipinski and J. Bartkowiak, Chem. Phys. 245, 263 (1999).
94. A. Painelli, Chem. Phys. 245, 185 (1999).
95. A. K.-Y. Jen and co-workers, Adv. Mater. 9, 132 (1997).
96. S. R. Marder and co-workers, Science 265, 632 (1994).
97. J. L. Br´edas and co-workers, Acc. Chem. Res. 32, 267 (1999).
98. L. R. Dalton, A. W. Harper, and B. H. Robinson, Proc. Natl. Acad. Sci. USA 94, 4842

(1997).

99. A. W. Harper and co-workers, J. Opt. Soc. Am. B 15, 329 (1998).

100. L. R. Dalton and co-workers, J. Chem. Mater. 9, 1905 (1999).
101. B. H. Robinson and co-workers, Chem. Phys. 245, 35 (1999).
102. B. H. Robinson and L. R. Dalton, J. Phys. Chem. 104, 4785 (2000).
103. W. Lin and co-workers, J. Am. Chem. Soc. 119, 2205 (1997).
104. A. Piekara, Proc. R. Soc. London A 149, 1 (1939).
105. H. Ma and co-workers, J. Am. Chem. Soc. 123, 986 (2001).
106. L. R. Dalton, Advances in Polymer Science, Vol. 154, Springer-Verlag, Heidelberg,

2001.

107. D. Chen and co-workers, Proc. SPIE 3006, 314 (1997).
108. D. Chen and co-workers, Appl. Phys. Lett. 70, 3335 (1997).
109. D. Chen and co-workers, IEEE Photon. Tech. Lett. 11, 54 (1999).
110. A. H. Udupa and co-workers, Electron. Lett. 35, 1702 (1999).
111. S. S. H. Mao and co-workers, Chem. Mater. 10, 146 (1998).
112. L. R. Dalton, in D. L. Wise, G. E. Wnek, D. J. Trantolo, T. M. Cooper, and J. D. Gresser,

eds., Electrical and Optical Polymer Systems, Marcel Dekker, Inc., New York, 1998,
p. 609.

113. L. R. Dalton, Opt. Eng. 39, 589 (2000).
114. S. Kalluri and co-workers, Proc. SPIE 2285, 67 (1994).
115. S. Kalluri and co-workers, Appl. Phys. Lett. 65, 2651 (1994).
116. H. W. Oviatt Jr. and co-workers, Chem. Mater. 7, 493 (1995).
117. C. C. Teng, M. A. Mortazavi, and G. K. Boughoughian, Appl. Phys. Lett. 66, 667 (1995).
118. A. Chen and co-workers, J. Electrochem. Soc. 143, 3648 (1996).
119. W. H. Steier and co-workers, Mat. Res. Soc. Symp. Proc. 413, 147 (1996).
120. A. Chen and co-workers, Proc. SPIE 3005, 65 (1997).
121. D. M. Burland, R. D. Miller, and C. A. Walsh, Chem. Rev. 94, 31 (1994).
122. L. R. Dalton and co-workers, Adv. Mater. 7, 519 (1995).
123. L. R. Dalton and co-workers, Chem. Mater. 7, 1060 (1995).
124. Z. Liang and co-workers, Chem. Mater. 7, 941 (1995).
125. Z. Liang and co-workers, Chem. Mater. 7, 1756 (1995).
126. Z. Liang and co-workers, Chem. Mater. 8, 2681 (1996).
127. C. Xu and co-workers, Macromolecules 26, 5303 (1993).
128. M. A. Hubbard and co-workers, Chem. Mater. 4, 965 (1992).
129. J. E. Becker and co-workers, J. Am. Chem. Soc. 115, 12216 (1993).
130. R. E. Harelstad and co-workers, Mater. Res. Soc. Symp. Proc. 328, 637 (1994).
131. D. Yu, A. Gharavi, and L. Yu, Macromolecules 28, 784 (1995).
132. D. Yu and co-workers, ACS Symp. Ser. 672, 123 (1997).
133. H. Saadeh and co-workers, Macromolecules 30, 5403 (1997).
134. H. Ma and co-workers, Chem. Mater. 11, 2218 (1999).
135. M. W. Becker and co-workers, Chem. Mater. 6, 104 (1994).
136. A. Galvan-Gonzalez and co-workers, Appl. Phys. Lett. 77, 2083 (2000).
137. B. H. Cumpston and co-workers, Nature 398, 51 (1999).
138. L. Sun and co-workers, Opt. Eng. in press; L. Sun and co-workers, Proc. SPIE 3950,

98 (2000).

Vol. 2

ENGINEERING THERMOPLASTICS, OVERVIEW

307

139. L. Sun and co-workers, Proc. SPIE 3632, 134 (1999).
140. D. H. Chang and co-workers, IEEE Photon. Tech. Lett. 12, 537 (2000).
141. S. S. Lee and co-workers, IEEE Microwave and Guided Wave Lett. 9, 357 (1999).
142. R. Chen and co-workers, Proc. PSAA, in press.
143. J. H. Bechtel and co-workers, Proc. SPIE 4114, 58 (2001);