Nanocomposite Materials for Optical Applications
Laura L. Beecroft and Christopher K. Ober*
Department of Materials Science and Engineering, Cornell University, Ithaca, New York 14850
Received August 20, 1996. Revised Manuscript Received May 6, 1997X
A substantial amount of work has been carried out in the area of nanocomposite materials
for optical applications. Composites are typically constructed by embedding an optically
functional phase into a processable, transparent matrix material. By doing so, the optical
properties can be utilized in more technologically important forms such as films and fibers.
This review covers many areas of optical compositeresearch todate. Composites with second-
and third-order nonlinearities and laser amplification properties arediscussed with examples
from the recent literature. Other composites, including transparent magnets, may be made
using similar structures. The principles used to construct these composites may have
important technological applications soon and are therefore summarized in this review.
As optical materials applications are expanding, the
need for novel optically functional and transparent
materials increases.These needs range from high
performance, all optical switches for futureusein optical
computing, to hard transparent coatings as protective
or barrier layers. In addition to optical needs such as
switching and amplification, the materials must be
integrated into existing structures such as waveguides
and optical fibers. As such, films and fibers are of great
interest as the final form of these novel materials.
Nanocomposite materials show great promise as they
can provide the necessary stability and processability
for these important applications.
The general principles in the construction of optical
composites involve the intimate mixing of optically
functional materials within a processable matrix. Fig-
ure 1 shows this type of composite schematically, where
the small particles possess the desirable optical proper-
ties and the enclosing matrix imparts processability in
film or fiber forms. Examples of incorporated phases
include quantum-confined semiconductors, solid-state
lasers, small molecules, and polymers. Matrix materials
can be polymers, copolymers, polymer blends, glasses,
or ceramics. Using such a composite structure, nano-
composites have been formed with nonlinear optical
(NLO) and laser amplification properties, among others.
Optical scattering must be avoided in these types of
composites, resulting in a tradeoff between particle size
and refractive index (RI) mismatch (the difference in
RI between the matrix and the particles). For very
small particles (typically <25 nm), the RI mismatch is
not important, but for larger particles the RI of the
matrix and the particles must be carefully matched to
avoid scattering. In our work, calculations based on
Rayleigh scattering have shown that particles as large
as 100 nm require matrix materials with closely matched
RI (within 0.02).1
Nanocomposite structures have been used to create
optically functional materials. By incorporating semi-
conductor nanoparticles intopolymer, glass, or ceramic
matrix materials, many of their interesting optical
properties including absorption, fluorescence, lumines-
cence, and nonlinearity may be studied.
systems, very small particle sizes enhance the optical
properties while the matrix materials act to stabilize
the particle size and growth. Ceramic nanoparticles of
solid-state laser materials can be incorporated into
polymer matrix materials resulting in laser-active com-
posites. This structure allows the formation of solid-
state laser amplifying films which would traditionally
be very difficult to make. Optically functional small
molecules and polymers may alsobeincluded in polymer
and glass matrixes and retain their optical properties.
Other applications of nanocomposite structures have
resulted in transparent materials with unusually high
RI, magnetic properties, and excellent mechanical prop-
Nanocomposite structures provide a new method to
improve the processability and stability of materials
with interesting optical properties. The applications of
such composites are extremely broad, ranging from
solid-stateamplifier films totransparent magnets. This
review focuses on recent developments in the synthesis
and applications of optical composite materials.
2. Nonlinear Optical Composites
Nonlinear optical materials can be useful for all
optical switching and wavelength manipulation. ?2, or
XAbstract published in Advance ACS Abstracts, J une 15, 1997.
F igure 1. Schematic of optical composite.
1302 Chem. Mater. 1997, 9, 1302-1317
S0897-4756(96)00441-3 CCC: $14.00 © 1997 American Chemical Society
second-order nonlinear optical materials, can be useful
for frequency doubling and optoelectronic switching. By
incorporating these materials into composites the sta-
bility of ?2processes can be improved as shown in
section 2.2.3 ?3, or third-order nonlinear optical materi-
als, can be useful for all optical switching and signal
processing. The strength of the nonlinearity can be
enhanced by using particles exhibiting quantum con-
finement effects which can be stabilized by a composite
structure. Much work in the area of colloidal semicon-
ductors has focused on taking advantage of this en-
The third-order nonlinear properties of solids are
typically reported as the nonlinear susceptibility (?3,
esu), the nonlinear refractive index (n2, cm2/W), or the
nonlinear absorption coefficient (R2/R0, cm2/W). Table
1 summarizes nonlinear optical properties for theoptical
composites that will bediscussed in this section. A wide
range of ?3values have been reported, and large
susceptibilities have been shown for several systems.
The nonlinear susceptibility (?3) is a complex number
(both real and imaginary parts), making comparisons
between different measurement techniques difficult.
The best strategy is tocompare similar measurements,
as the conversion between real and imaginary ?3can
be complicated. Additionally, the susceptibility will
changedepending on themeasurement wavelength. The
wavelength may be near or far from resonance, and it
will affect the amount of absorption (imaginary contri-
bution to ?3) in the sample.
Entries 1-3 in Table 1 can be compared toeach other
as they are all based on nonlinear absorption mea-
surements.2-4They can also be compared to the non-
linear absorption of bulk CdS (3.2 × 10-9cm/W at 610
nm), which is considered to have a strong nonlinear
response.5Entries 4-7 were all measured using de-
generate four-wave mixing experiments, which give the
susceptibility as the square sum of the real and imagi-
nary parts.6-8The experiment used to measure entry
8 resulted in the nonlinear refractive index, or real part
of ?3.9Finally, entries 9-11 in Table 1 show the real
part of ?3for several commonly studied nonlinear optical
materials.5,10Quartz has a weak nonlinear response,
while bulk CdS is considered to be strongly nonlinear.
All composite values in Table 1 show larger or
comparable nonlinear susceptibilities tobulk CdS. The
best results have been seen with quantum-confined
CdSxSe1-x(entries 1-4). CdS grown in situ in an ion-
exchange resin (section 2.1.2), CdSe grown ex situ using
a capping method (section 22.214.171.124), and CdSSe in com-
mercial glass (section 126.96.36.199) all show excellent non-
linear behavior. Composites containing nonlinear op-
tical polymer inclusions have also shown strong non-
linearity (entries 6 and 7, section 2.2.1). Other com-
posites studied show nonlinearities similar tobulk CdS.
Many reviews concerning the synthesis, nonlinear
optical properties, and applications of colloidal semi-
conductors have been written.11-17
wealth of information exists for colloidal semiconductor
materials, this review will focus only on examples where
composite structures are utilized to synthesize and
process quantum-confined semiconductors. Wang has
coauthored several reviews discussing optical compos-
ites and mentioning the potential of these materials as
Boyd and Sipe published several papers that showed
theoretically that the composite structure itself should
enhance nonlinear optical properties. Calculations for
both spherical inclusions21as well as layered compos-
ites22were presented and demonstrated that the en-
hancement can come when the NLO material is either
the inclusion or the host.
confirmed experimentally using a dense atomic potas-
sium vapor23as well as a layered TiO2/poly(p-phenyle-
A significant amount of experimental work has been
done with semiconductor containing composites in a
variety of matrix materials. Polymers, including ho-
mopolymers, block and random copolymers, and poly-
mer blends, as well as glasses and ceramics such as
porous glass, zeolite, and sol-gel glass, have improved
the stability and processability of quantum-confined
semiconductors. Small molecules and polymers with
nonlinear optical properties have also been found to be
useful for nonlinear optics in composite form. Compos-
ites have been characterized for their structure using
techniques such as X-ray diffraction and transmission
electron microscopy (TEM).
review, the optical properties of these nanocomposites
including absorption, fluorescence, luminescence, and
optical nonlinearity have been investigated.
2.1. Nanocomposites Containing Semiconduc-
tors. 2.1.1. Optical Absorption as a Measureof Particle
Sizeand Distribution. Currently, semiconductor nanoc-
rystals are primarily studied for their enhanced optical
properties. These optical properties are directly related
to the size and size distribution of the nanocrystals
which can be extracted from the optical absorption
spectrum. If crystallite sizes are below the Bohr radius
of both the holes and electrons in the semiconductor,
strong quantum confinement occurs.25Table 2 lists the
maximum diameter for strong confinement for several
Because such a
The theories have been
More relevant to this
T able 1. T hird-Order Nonlinear Optical Properties in Nanocomposite Materials
compositemeasd NLO strength
-6.1 × 10-7
-8.3 × 10-7
1.2 × 10-5
1.3 × 10-8
5 × 10-12
3 × 10-10
6 × 10-10
-5.6 × 10-12
part of ?3measda
CdS in Nafion
CdS in Nafion/NH3
capped CdSe in PMMA
CdS in sol-gel glass
PPV in SiO2
PPV in V2O5
GaAs in Vycor glass
Standard NLO Materials
8.5 × 10-14
8 × 10-13
-5 × 10-11
9 fused quartz
aR is the absorption coefficient.
ReviewsChem. Mater., Vol. 9, No. 6, 1997 1303
common semiconductors.26,27Weaker confinement ef-
fects can be seen at somewhat larger crystal diameters.
The confinement effect appears as a shift to lower
wavelengths in the absorption spectrum representing
a changing bandgap. Ideally this shift should be ac-
companied by exciton features in the spectrum, which
show that the particle size distribution is very narrow.
The absorption shift and spectral features can act as a
measure of particle size and size distribution. Control
of these parameters is extremely important in the
enhancement of ?3effects.
The shift in absorption is illustrated in Figure 2a,
which shows a blue-shift with decreasing PbS nanoc-
rystallite size as well as a steepening of the absorption
edge for a PbS/polymer composite.28Spectra of crystals
above 25 Å are featureless, but those below 13 Å show
some structure indicating a narrow size distribution in
the sample. The smallest particles exhibit a bandgap
of 2.3 eV, considerably shifted from the bulk value of
0.41 eV. Figure 2b displays an absorption spectrum
with strong exciton features of PbS in poly(vinyl alcohol)
made in our laboratories following a modified procedure
of Nenadovic et al.29
The strong exciton features
indicated a narrow particle size distribution.The
magnitudeof thebandgap shift has been correlated with
particlesizes of 4 nm,29although theshift is at a slightly
lower wavelength than that shown for 4.5 nm particles
in Figure 2a.
Alivisatos et al. studied the homogeneous and inho-
mogeneous contributions to the optical spectrum of
CdSe nanocrystals using optical hole burning.30The
CdSe clusters were synthesized using inverse micelles
and redispersed in polystyrene for the optical measure-
ments. The inhomogeneous contributions, caused by
even small variations in sizedistribution, dominated the
spectra. Becauseof this inhomogeneous broadening, the
size distribution must be carefully controlled to maxi-
mize quantum effects. As will be shown throughout the
semiconductor examples, size distribution control is
often quite difficult with these quantum-confined ma-
Early absorption measurements showed either little
shift in absorption edges31or no exciton features when
the edges were shifted.32Copolymers containing poly-
ethylene (85%) and poly(methacrylic acid) (15%) aided
the synthesis of PbS-containing composites prepared by
Wang et al.28,33The particle sizes could be altered by
changing the concentration of Pb2+in the films and by
subsequent heat treatments. Bandgap and absorption
measurements were made over a range of crystallite
sizes (Figure 2a), and theoretical models were used to
explain the trends of the data.
Work with poly(vinyl alcohol) as a stabilizer and
matrix material for PbS has consistently shown shifted
and featured absorption spectra (Figure 2b). This was
first observed by Gallardo et al. in 1989 in a study of
theabsorption and fluorescenceproperties.34Nenadovic
et al. also studied similarly prepared materials with
particular interest in the effects of surface properties
on the bleaching of PbS.29Our group has also carried
out some research in this area, improving the film-
forming nature of the composites.35
2.1.2. Semiconductor Composites with Polymer Ma-
trix Materials. Early work with quantum-confined
semiconductors was performed in colloidal solutions,36-38
which could often be stabilized with small amounts of
polymer.39After several years of work with polymeric
stabilizers, these stabilizers were discovered to be
excellent matrix materials yielding processablepolymer
films with semiconductor optical properties. The result-
ing films often were more stable than colloidal solutions
and were useful for many optical measurements.
The first group to formally recognize semiconductor
polymer composites such as thesein terms of engineered
optical media was Akimov et al. in 1992.40In this work,
CdS nanocrystals from 2 to50 nm in size were prepared
in poly(vinyl alcohol), poly(vinylpyridine), and photo-
graphic gelatin. Other polymers that did not stabilize
the particles were also mentioned. Remarkably high
CdS concentrations, up to 50 wt %, could be prepared
without agglomeration. The composites exhibited good
photosensitivity and photoconductivity. Particular ap-
plications of interest tothe authors included dispersive
optical elements, bandpass and cutoff filters, and elec-
trophotographic and photothermoplastic materials.
The earliest semiconductor composites were used for
their catalytic properties rather than their optical
properties.41Early in the catalysis work, in situ syn-
thesis methods were developed by Krishnan et al.42By
F igure 2. (a) Absorbance spectra shifts as a function of
particle size. Reprinted with permission from ref 28. (b)
Absorbance spectrum with exciton features from PbS in PVA.
T able 2. Maximum Crystallite Size for Strong
Confinement in Several Semiconductors
1304Chem. Mater., Vol. 9, No. 6, 1997 Reviews
ion exchanging a Nafion matrix (perfluoroethylene-
sulfonic acid ion-exchange polymer) with Cd2+, then
exposing to H2S, submicron particles were produced,
which could photocatalytically drive chemical reactions.
Mau et al. used a similar method that created 20 nm
size particles which were used for photocatalytic hydro-
gen generation in the presence of platinum catalyst.43
These in situ methods were later utilized for the novel
optical properties that could be generated.
The general scheme used to prepare polymer-
semiconductor composites in situ is shown in Figure 3.
In a typical experiment, the matrix material and metal
ions are mixed in solution and are then exposed to the
counterion (S2-, Se2-) in the form of gas or as ions
dissolved in solution. The composite may be cast as a
film before or after exposure to the counterion. For
example, if poly(vinyl alcohol) (PVA) were the matrix
material for PbS nanoparticles, a solution of PVA and
Pb2+could beprepared in water (step 1a). This solution
could then be exposed to H2S gas in order to form the
PbS and then cast as a film (step 2a). In this particular
reaction, the crystalline semiconductor forms extremely
quickly yielding a wine-red solution. Similarly, if a
block copolymer were to be used as the matrix, Pb2+
could be complexed with the copolymer in solution and
then cast as a film (step 1b). The phase-separated film,
which might require annealing, would then be exposed
toH2S gas in a closed container (step 2c). Over a period
of several hours, the film would turn brown, indicating
PbS formation. Many variations in this general syn-
thesis scheme can be imagined.
188.8.131.52. NLO Properties of Polymer Composites. While
most authors praise the enhanced nonlinear optical
properties of quantum confined semiconductors, few
actually have measured them. Wang et al. at DuPont
have been the leaders in making these measurements
on polymeric systems. In 1987, Wang and Mahler
reported the first study of NLO properties in polymer
stabilized quantum-confined semiconductors using the
degenerate four-wave mixing (DFWM) results of a CdS/
Nafion composite.44The particles studied were 50 Å in
size, which would cause moderate confinement (Table
2) and resulted in a bandgap of 2.55 eV (the bandgap of
bulk CdS is 2.5 eV). DFWM measurements were taken
at 505 nm (2.43 eV) as shown in Figure 4, where a slope
of 1.9 and saturation at about 2 MW/cm2were found,
indicating a third-order process. The signal strength
was about half that seen in semiconductor-doped glass
samples, and a 10 ns response time was measured. The
saturation indicated that a three-level saturable ab-
sorber model was the most likely mechanism for non-
linearity in these composites. A later paper reported
DFWM experiments and absorption saturation experi-
ments that confirmed the interesting nonlinear optical
properties of these composites.45
Hillinski et al. presented work concerning 5.5 nm CdS
clusters grown in Nafion, with a focus on the NLO
properties.2Their films showed large ?3nonlinearity
(R2/R0) -6.1 × 10-7cm2/W) which they attributed to
the bleaching of the exciton absorption. This bleaching
was enhanced due to the high concentration of surface
defects on the small particles. They cautioned that an
understanding of the surface chemistry of these par-
ticles will be very important in the interpretation of
Continuing this work, Wang et al. further discussed
bleaching of quantum-confined CdS particles in Nafion
films.3Crystallites in a range of sizes up to 40 Å were
synthesized and passivated with ammonia. The effects
of surface-trapped, electron-hole pairs were examined
using absorption, photoluminescence, and pump-probe
experiments. An even higher ?3nonlinearity (R2/R0)
-8.3 × 10-7cm2/W) was measured in the surface
passivated samples. A short review paper by the same
group reported the NLO properties for the CdS in
Nonlinear optical measurements were also made on
CdS grown in a swollen, cross-linked, copolymer ma-
trix.47A third harmonic generation experiment was
used, comparing intensities toa quartz reference. Near
resonance at 1.45 µm, an increase in signal as much as
11.2 times the reference was observed. Off resonance
at 1.06 and 1.5 µm small increases from 1.2 to1.6 times
the quartz standard were shown.
F igure 3. Schematic of in situ polymer synthesis methods.
F igure 4. Nonlinear optical properties in a CdS/polymer
nanocomposite. Reprinted with permission from ref 44.
Reviews Chem. Mater., Vol. 9, No. 6, 19971305
184.108.40.206. Using Polymer Architecture To Control Par-
ticleSize. As has been mentioned, particle size and size
distribution must becarefully controlled in order totake
advantage of NLO enhancement effects. The architec-
ture of the polymer matrix can be used to help in
controlling these parameters. Both block copolymers
and polymer blends exhibit phase separation, which
may help isolate the semiconductor clusters as they
form. In some cases this controlled phase separation
can result in superlattice structures. Most work in the
area of superlattices have been with colloidal solids;48,49
however, similar structures may be possible with com-
Researchers at MIT have done extensive work with
block copolymers prepared using the ring-opening me-
tathesis polymerization (ROMP) synthesis technique. By
taking advantage of the phase separation of the block
copolymers, good stabilization of the semiconductor was
found. STEM studies of PbS50as well as CdS and ZnS51
composites showed that small particles can be synthe-
sized utilizing spherical phase separation of the metal-
containing phase; however, the optical properties were
not measured. The bandgap of 30 Å ZnS particles
prepared by this technique was 5.7 eV, which is con-
siderably higher than the bulk bandgap of 3.5 eV.
Similar copolymers weresynthesized which wereused
to prepare ZnS- and Zn-containing composites in both
lamellar and spherical morphologies.52
Figure5, both lamellar (Figure5a) and spherical (Figure
5b) morphologies effectively isolated the ZnS particles
(dark regions) in a controlled manner. Both morphol-
ogies resulted in clusters less than 20 Å in size with a
bandgap of 6.3 eV.X-ray and X-ray photoelectron
spectroscopy were also used to characterize these ma-
terials. In 1994, members from the same research
As seen in
group again reported PbS grown in a ROMP block
These composites displayed a featured
absorption spectrum, indicating a narrow size distribu-
tion, but were primarily characterized by TEM and
Mo ¨ller reported the growth of nanocrystals of many
semiconductors, including CuS, CdS, and PbS, in block
copolymers of polystyrene and poly(vinylpyridine) with
Shifted absorption spectra
were shown indicating small particle sizes, but the
absorbancespectra werefeatureless, indicating that the
particles were not monodisperse. More recent work by
the same group on gold nanoparticles has shown that
the stabilization of the ionic block is very important if
a single particle per micelle is desired.55When a poly-
(ethylene oxide) block was used as a matrix for gold
particles, heat treatment caused the particles to come
together, resulting in a system where the particle size
could be controlled by the composition of the copolymer
and the concentration of ions. However, when poly-
(vinylpyridine) was used as thestabilizing block for gold
(and several semiconductors) the particles did not
coalesceduring heat treatment becausethestabilization
was much stronger. The strategy for forming one
particleper micelleusing a weakly stabilizing block and
heat treatment might be very important if it can be
extended to semiconductor particles.
The first work utilizing polymer blends as matrix
materials was reported by Yuan et al.56,57Quantum-
confined CdS was prepared in poly(styrenephosphonate
diethyl ester) (PSP) and cellulose acetate (CA) polymer
blends, resulting in structured absorption spectra. The
phosphonate ester chelates metal nitrates, isolating
them before reacting with H2S. The fluorescence re-
sults, shown in Figure 6 for several PSP:CA composi-
tions, showed a shift to lower energies with increasing
amounts of PSP. This behavior indicated increasing
particle sizes with increasing PSP composition. The
particle sizes measured by absorbance edge position
were 44 Å (Figure 6a, 2.81 eV), 58 Å (Figure 6b, 2.63
eV), and 82 Å (Figure 6d, 2.48 eV). The higher energy
peak in Figure 6a may be due to excitonic fluorescence
or due to crystalline cellulose acetate being present.
Continued work showed size quantization by the chang-
ing bandgap and structured absorption spectra.
2.1.3. Semiconductor Composites with Inorganic Ma-
trix Materials. A significant amount of work in thearea
F igure 5. TEM micrographs of ZnS in a block copolymer. (a)
Lamellar morphology, (b) spherical morphology. Reprinted
with permission from ref 52.
F igure 6. Fluorescence spectra of CdS in several polymer
blend compositions. PSP:CA ) (a) 1:4, (b) 1:1, (c) 2:1, and (d)
4:1. λex) 400 nm. Reprinted with permission from ref 57.
1306 Chem. Mater., Vol. 9, No. 6, 1997Reviews
of quantum-confined semiconductors has been carried
out with high-temperature glass, porous glass, zeolite,
and sol-gel derived matrix materials. The small and
regular pore sizes of these materials are useful in
controlling the particle sizes and distributions of the
semiconductor nanoparticles. Additionally, semicon-
ductor/semiconductor composites have been found to
have enhanced optical properties. Early work in this
area was done on commercially available CdSxSe1-x
glasses which are used as sharp-cutoff color filters.
Large size distributions and varying sulfur concentra-
tions made these materials difficult to study quantita-
tively. More recent work deals with precisely synthe-
sized compositions. Figure 7 summarizes the synthesis
of nanocomposites in various glass and ceramic struc-
220.127.116.11. High-Temperature Glasses.
semiconductors in glasses involves the incorporation of
the necessary ions in the glass at high temperature,
casting toform a monolith, annealing toremove stresses,
and heat treatment to crystallize the semiconductor as
shown in Figure 7a. Nonlinear optical measurements
were made very early in the work with glass nanocom-
posites. In 1983, J ain and Lind studied degeneratefour-
wave mixing in commercial CdSxSe1-xglass, and com-
pared the results to single-crystal CdS.6They showed
that the glass could be used as an aberration corrector,
with ?3values as high as 1.3 × 10-8esu and a fast
response time. Warnock and Awschalom showed con-
finement in these glasses using luminescence experi-
ments.58Cullen et al. fabricated directional couplers
using ion-exchanged waveguides from similar semicon-
Roussignol et al. continued experiments on theglasses
used by J ain and Lind.60,61
researchers at Corning Inc. reported results concerning
In an extensive study,
commercially availablefilter glasses (CdSexS1-x) as well
as experimental CdS and CdSe glasses.62The experi-
mental glasses they studied showed featured absorption
and luminescence spectra, indicating small monodis-
perse particle sizes, which were confirmed by TEM.
Other phases such as AgCl and CuCl were grown by
similar techniques and resulted in quantum-confined
More recently, the first observation of resonatorless
bistability and nonlinear switching due to increasing
absorption in a semiconductor-doped glass was re-
ported.63A nondegenerate pump-probe technique was
used tomeasure these effects in CdSxSe1-x-doped glass.
Perfect switching behavior could be demonstrated by
changing the pump intensity as shown in Figure 8.
Input power above 167 kW/cm2resulted in a region of
high absorption, whilebelow this valuethematerial was
very transparent, defining twodistinct regions of switch-
ing and dephasing. Absorption induced bistability does
not require extra resonators such as Fabry-Perot
resonators or phase matching elements to create non-
18.104.22.168. Porous Glasses. Porous glasses contain well-
defined pores which can assist in confinement of the
semiconductor clusters. They are attractive as host
materials because low-temperature solution and gas-
phase synthesis techniques can be used. Figure 7b
shows a typical synthesis procedure where the glass is
infiltrated with the metal ions and subsequently ex-
posed toH2S (or the appropriate counterion) toform the
semiconductor nanoparticles. The work of Kuczynski
and Thomas discussed CdS prepared in porous Vycor
glass.64The CdS was confined in only one dimension,
giving some excitonic structure to its absorption spec-
trum, but bandgap properties were similar tobulk CdS.
The effects of methylviologen and water on the spec-
troscopic properties including emission and quenching
F igure 7. Schematic of semiconductor nanocomposite syn-
thesis in glass and ceramic matrixes. (a) Traditional glass, (b)
porous glass or zeolite, (c) sol-gel glass.
F igure 8.
CdSSe glass. Reprinted from ref 63.
Switching behavior in commercially available
ReviewsChem. Mater., Vol. 9, No. 6, 1997 1307
Luong discussed thegrowth of Cd, Pb, and Zn sulfides
and selenides in Vycor glass.65For Zn- and Cd-contain-
ing crystallites, only a small shift from bulk properties
was seen, due to broad particle size distributions and
particles larger than theBohr radius of theexciton. PbS
and PbSe showed quantum confinement in the glass
with blue-shifted absorption spectra which became
structured upon heat treatment.
J ustus et al. measured the optical properties of 50 Å
GaAs particles grown in Vycor glass.9The nonlinear
measurements showed n2 of -5.6 × 10-12cm2/W, an
order of magnitude stronger than the bulk material.
This is shown in the Z-scan data presented in Figure 9,
where the nonlinear response increases with increasing
power. The valley which develops with increasing
intensity indicates nonlinear absorption. Thedifference
in transmission between the maxima and minima is
directly related to the change in refractive index.
Because the peak of transmission precedes the valley,
the response is negative. Measurements were taken in
the technologically important near IR region, although
this region is far from resonance for GaAs.
22.214.171.124. Zeolites. Zeolites are crystalline ceramics
with well-defined pores of uniform size and size distri-
bution. The synthesis of composites based on these
materials is quite similar to that of porous glass
composites shown in Figure 7b. Unlike porous glasses,
the zeolite pores are situated on a lattice so that
materials substituted in them might form a superstruc-
tureas well as havequantum semiconductor properties.
A survey of synthesis techniques is presented by Ozin
Wang and Herron presented absorption spectra from
CdS and PbS grown in two zeolites.66A blue-shifted
absorption was found for both semiconductors which
shifted tored as theconcentration of semiconductor was
increased, shown in Figure 10 for a CdS/Zeolite Y
composite. Above concentrations of about 4 wt %, the
optical shift was abruptly arrested because the percola-
tion threshold of the zeolite was reached, and nochange
was observed with loading up to 18 wt % and 100 °C
heating. No intermediate absorption levels were ob-
served between those shown in the figure. The three-
dimensional lattice allowed the synthesis of clusters
with solid-state behavior which differs from the bulk.
Later some of the same authors described work
concerning CdS clusters grown in other zeolite hosts.67
As had been seen earlier, the CdS particles remained
discreet within the host at low concentration, but above
concentrations of 4% a supercluster structuredeveloped.
The supercluster structure had optical properties in-
termediate between the individual clusters and the
bulk. The authors indicated the potential of controlling
the structure and electronic properties by the choice of
zeolite host material.
Wang and Herron reported the luminescence and
excited-state dynamics of CdS grown in zeolites X, Y,
Three emission bands were found in the
yellow-green, red, and blue, which could be attributed
to defects. The yellow-green emission was attributed
to Cd atoms, while the red and blue were attributed to
Later Mo ¨ller et al. compared the structure of PbS in
twozeolites toPbS in a polymer matrix using extended
X-ray absorption fine structure.69The PbS was found
to be more confined in the zeolite matrix than in the
polymer, with higher order at higher concentrations.
Considerably larger PbS particles weregenerated in the
126.96.36.199. Sol-Gel Derived Glasses. Sol-gel type syn-
theses allow for low-temperature processing, high pu-
rity, and more flexibility in the components of the glass.
Additionally, sol-gel precursors lend themselves tofilm
F igure 9. Z-scan data for GaAs-doped Vycor glass. Reprinted
with permission from ref 9.
F igure 10. Absorption spectra for CdS in zeolite Y. (a) 1.1
wt % CdS/Y, (b) 7.4 wt % CdS/Y, (c) micron-sized CdS.
Reprinted with permission from ref 66.
1308 Chem. Mater., Vol. 9, No. 6, 1997Reviews
and fiber applications morereadily than other glass and
ceramic composites that have been discussed. Figure
7c shows this typeof synthesis schematically. Typically
thesol-gel glass is prepared with themetal ions present
and exposed to S2-(or other counterions) after glass
formation. Levy and Esquivas give a nice review of the
use of sol-gel matrix materials todesign optical materi-
als.70They discuss CdSe as well as laser dyes and
liquid-crystalline materials as incorporated phases, all
taking advantage of the small pore sizes of the sol-gel
Several semiconductors were incorporated into sol-
gel silicate glasses by Rajh et al.71
involved preparing the colloidal semiconductor with a
stabilizer (20-40 Å) and then adding tetramethyl ortho-
silicate to form the matrix. The samples were dried,
but not heat treated, and showed shifted and featured
absorption for several compositions shown in Figure 11.
Somesolutions lost exciton features during drying which
could be recovered by exposure to H2S.
Structured and shifted absorbance spectra were seen
in CdS nanoparticles in sol-gel glass depending on heat
treatment of the glass.72,73The crystalline structure of
the nanoparticles was confirmed by X-ray diffraction.
Othmani et al. discussed the preparation of CdS in sol-
gel derived SiO2at concentrations up to 20%.74X-ray
and Raman studies were combined topredict a particle
sizeof about 7 nm. Nogami et al. reported thecontrolled
preparation of CdSxSe1-xin sol-gel derived SiO2glass.75
Shifted absorption edges and bandgaps varying with
particle size were shown, but the absorption spectra
were featureless due to large size distributions.
Glasses derived from sol gel precursors for SiO2and
1.4Na2O-20.8ZrO2-77.8SiO2were used to stabilize in
situ growth of CdS crystallites.7
showed much higher stability to CdS oxidation at high
temperatures than the simple SiO2glass. Absorption
edges were featureless and ?3parameters of about 5 ×
10-12esu were measured for several conditions. This
synthesis of a mixed oxide glass took advantage of the
compositional flexibility of the sol-gel route.
188.8.131.52. Semiconductor Matrix Composites. Thin-film
composites have been constructed which contain semi-
conductor nanoparticles in a semiconductor matrix.76
Thesynthesis involves preparation of quantum-confined
CdSe and CdSe coated with ZnSe by a colloidal method.
Thenanoparticles arethen dispersed by an electrospray
during the organometallic chemical vapor deposition of
The sodium glass
a ZnSe thin film. Because the bandgap of ZnSe is much
higher than CdSe, the nanoparticles are very well
isolated from each other. The photoluminescence be-
havior was measured and showed that nanoparticles
that were coated with ZnSe before film formation had
much higher efficiency. Composites such as these may
bemoreeasily integrated intoelectronic devices because
the conductivity of the matrix can be controlled by
2.1.4. Controlling Particle Size. In all of the semi-
conductor work discussed here, the size and size distri-
bution of the particles have been of great interest.
Particle size and size distribution are critical in maxi-
mizing enhancement in quantum-confined systems. The
particle size affects the magnitude of the shift in
absorption (change in bandgap), while the distribution
affects the strength of the quantum effect at a given
wavelength. These properties can be controlled by the
polymer matrix architecture (184.108.40.206), porous matrix
materials (2.1.3), additives and heat treatments (220.127.116.11),
and ex situ particle-capping methods (18.104.22.168). Polymer
architecture and porous matrix materials have been
discussed and typically result in a fixed particle size
determined by the matrix structure. Other methods
discussed in this section can result in tunable particle
sizes which might beparticularly useful in changing the
wavelength of operation in optical devices.
22.214.171.124. Heat Treatments and Additives. Heat treat-
ments and additives have been coupled with composite
synthesis strategies tocontrol particle size and distribu-
tion. In the early work of Wang et al., shown in Figure
2a, PbS particle size was controlled by varying the
concentration of thePb2+and heat treating thesamples.28
Similarly, work from the same group on CdS in Nafion
(section 2.1.2) also found that heat treatment could be
used tocontrol particlesize.3Sankaran et al. found that
ZnS cluster size was increased when the block copoly-
mer films were exposed to H2S at higher temperatures
or in the presence of solvent vapor (section 2.1.4).52
Annealing after ZnS formation, however, did not result
in particle growth probably because the matrix acted
as a diffusion barrier. Heat treatments have alsobeen
used in many of the glass and ceramic matrix materials
Kyprianidou-Leodidou et al. varied the size of PbS in
poly(ethylene oxide) from 4 to 80 nm using acid and
surfactant additives.77Without additives, particlesizes
were about 29 nm. Acetic acid increased the particle
sizes, while sodium dodecyl sulfate, a surfactant, low-
ered them. Although thecontrol of averageparticlesize
was demonstrated, the size distributions were quite
In our work with PbS synthesized in poly(vinyl
alcohol) (PVA), attempts were made to change particle
sizes using similar treatments.35In a standard reaction
of PbS in PVA, 4 nm particles were reproducibly
synthesized as shown in the absorbance spectrum in
Figure 2b. The concentration of the PVA was increased
from the standard 0.1% up to5% in water and found to
have noeffect on the particle sizes while improving the
film-forming abilities and increasing photostability.
Acetic acid was added to the synthesis and showed an
increase in particle size by X-ray line broadening seen
in Figure12. As shown, thestandard reaction produced
F igure 11. Absorption spectra for several semiconductors
prepared in sol-gel glasses. Reprinted with permission from
ReviewsChem. Mater., Vol. 9, No. 6, 1997 1309
a particle size (4 nm) which resulted in a featureless
spectrum (Figure 12a). The addition of acid clearly
caused an increase in particle size which made the PbS
peaks visible (Figure 12b-f), indicating crystallite sizes
between 10 and 15 nm (calculated using the Scherrer
equation). The absorption spectra lost their exciton
features upon addition of acid, probably due to a
broadening of the size distribution. Surfactants and
hydrochloric acid were also added to the synthesis, but
they did not result in particlesizechanges or absorption
126.96.36.199. Ex Situ Semiconductor Capping. In further
efforts to control semiconductor particle sizes and size
distributions, capping methods have been developed.
Capping agents are typically thiols or mercaptans,
which can compete with sulfur (or other counterions)
for Pb2+or Cd2+surfaces. In doing so, thecapping agent
to sulfur ratio chemically determines the particle size.
In addition to controlling the particle size, the capping
agents act as a surfacetreatment for theparticles which
may aid their dispersion intovarious matrix materials.
In these experiments, the particles are prepared ex situ
and then dispersed in a polymer matrix to form nano-
Figure 13 shows the formation of the capped particles
schematically. Typically, the sulfide (or other counte-
rions) and capping agent (RS-) ions are mixed in
solution. A solution containing the metal ions is added
and semiconductor particles form with sizes determined
from the S2-/RS-ratio. The resulting particles can be
studied in solution or can be dried to a powder. By
redispersing the capped particles in a polymer solution,
nanocomposites can be formed.
Early work in this area focused on colloidal syntheses
with no effort to form composites.78-82Several thiols
were studied and found to control particle size with
respect to the ratio of capping agent to counterion in
the system. Researchers at DuPont synthesized CdS
particles which werecapped with thiophenol.46,83Using
X-ray diffraction and absorbance measurements, they
showed that the ratio of capping agent to sulfur chemi-
cally controlled the particle size. The nonlinearity of
these clusters was measured by third harmonic genera-
tion.84?3increased from 4.7 × 10-12esu for a 7 Å
molecular cluster to 3.2 × 10-10esu for 30 Å clusters.
Further characterization of these particles included
capped precursor synthesis85and inelastic neutron-
Some work has been done incorporating the capped
particles in polymer matrix materials. Majetich and
Carter reported the synthesis of colloidal CdSe capped
with six different terminating ligands to study surface
effects.4Samples were prepared for optical measure-
ments by dispersing the colloids in solid epoxy or poly-
(methyl methacrylate) matrix materials. The bandgap,
absorption oscillator strength, and spectral hole width
and trapping time were unaffected by the various
surfaces tested. However, the optical hole-burning and
luminescenceexperiments wereaffected by thesurfaces
of the particles. Absorption spectra showed excitonic
features, indicating a small size distribution in the
particles. Nonlinear optical measurements indicated
very promising ?3values ranging from 6.5 × 10-8to2.9
× 10-7esu (R2, 8.7 × 10-6to 1.2 × 10-5cm/W).
Our group has studied capping syntheses for PbS
particles for composite applications.35Thiophenol and
several other thiols have been investigated as possible
capping agents to control particle size as well as tailor
the particle surfaces for dispersion into a matrix. All
capping systems show the expected Scherrer line broad-
ening in X-ray diffraction studies. Preliminary work
has begun on redispersing the capped particles in
polymer matrix materials for composite applications.
F igure 12. X-ray line broadening due toadded acetic acid in
the formation of PbS in PVA. (a) No added acid, (b) 0.015 M,
(c) 0.076 M, (d) 0.38 M, (e) 0.46 M, (f) 0.53 M.
F igure 13. Schematic of ex situ capped semiconductor (MS)
synthesis. RS-is the capping agent, S2-is the sulfur, and M2+
is the metal ion.
1310 Chem. Mater., Vol. 9, No. 6, 1997Reviews
2.2. Nanocomposites Containing Polymers and
Small Molecules. Many polymers and small molecules
exhibit interesting optical effects including second (?2)
and third (?3) order nonlinearities and laser properties.
Using thenanocompositestructure, polymers and small
molecules such as these can be incorporated intoglass,
ceramic, and polymer matrix materials as shown in
Figure 14. The matrix material can stabilize the
compounds, improvetheoptical properties, and improve
the processability as well.
2.2.1. Third-Order Nonlinear Optical Polymers in
Glasses. Prasad and co-workers havepresented several
papers concerning poly(p-phenylenevinylene) (PPV) in
silica and V2O5matrix materials. PPV is a conjugated
polymer with a fast ?3nonlinear optical response;
however, it shows high optical losses when used in the
bulk. Silica and V2O5 both show excellent photonic
properties, in particular low losses, and improve the
optical properties of the PPV when they are combined
in a composite. The composite preparation is shown
schematically in Figure 14a.
Initial work with silica glass matrixes showed that
composites exhibited good optical quality.87The UV-
vis spectrum of thePPV compositeshowed a slight blue-
shift compared to bulk PPV indicating that the conju-
gation length was reduced in the composite. The same
group performed a study of the NLO response in PPV-
silica composites using degenerate four-wave mixing
and optical Kerr gate techniques.88
Using similar composites, with V2O5instead of silica,
two-dimensional gratings were produced by introducing
refractive index changes or surface relief patterns with
a laser.89Gratings such as these could be useful in
laser-array systems and multichannel optical com-
munications. A later report on V2O5composites showed
that this matrix resulted in a longer conjugation length
than seen in the silica composites.8The measured ?3
values were 3 × 10-10esu for PPV/silica and 6 × 10-10
esu for PPV/V2O5, showing a slight improvement in the
2.2.2. Small Molecules in Glasses and Ceramics.
Glass and ceramic matrixes can improve the optical,
mechanical, and thermal properties of small molecules
as discussed in the review by Levy and Esquivas.70The
synthesis of small molecule glass composites is very
similar to that shown in 14a, except that the polymer
is replaced with small molecules. Early work in the
area of small molecules in glass was reported by Avnir
et al. for rhodamine 6G/silica composites.90Rhodamine
6G is a laser dye which displays fluorescence, absorp-
tion, and emission in thevisibleregion between 500 and
600 nm. This dye is typically used in solution and has
problems with concentration quenching and photosta-
bility. Many virtues of glass-based composites over
solutions include the removal of intermolecular interac-
tions between dye molecules, isolation of impurities,
isolation from surrounding atmosphere, good thermal
and photostability, and good optical properties among
others. Several other laser dyes have also been suc-
cessfully incorporated into sol-gel derived matrix ma-
terials and retained their optical properties.91Theeffect
of the dyes on the sol-gel synthesis has been studied.
A significant amount of work has been done incorpo-
rating enzymes and other proteins in sol-gel materi-
als.92,93Several papers have focused on optical appli-
cations of these composites. Bacteriorhodopsin is a
light-transducing protein which could be useful as an
active component in optically coupled devices. After
encapsulation in sol-gel silica glass, the optical and
photocycle behavior was retained.94This material may
alsobe useful as an optically based ion-sensor. Similar
composites were reported which were potentially useful
as a real-time holographic medium.95
Another light-transducing protein, phycoerythrin, has
also been incorporated into sol-gel silica glass.96Ab-
sorption and fluorescence measurements of the com-
posite showed that the protein retained its optical
properties in the matrix and even showed enhanced
stability toward photodegradation. Figure 15 shows
that the optical properties were retained in the com-
posite by comparing the absorption at several stages in
the sol-gel process (Figure 15a-c) tothe optical proper-
ties in solution (Figure 15d). Only a slight change in
the intensity of the peaks at 565 and 495 nm can be
identified. Thefluorescencespectra arealsounchanged
with the addition of the matrix, and the material can
exhibit two-photon fluorescence. Possible applications
of the material include biosensors, 3D biomolecular
imaging, and 3D optical storage.
Zeolites have been examined as potential hosts for
small molecules with second-order nonlinearities.97
Composites were prepared as shown in Figure 14b,
wherethesmall moleculewas vaporized toinfiltratethe
zeolite. These studies have shown that noncentrosym-
metric hosts were necessary to retain the ?2properties
in the small molecules and in some cases the ?2
properties were enhanced by the zeolite matrix. The
structure of the small molecules in the zeolite host was
discussed. The composites can be tuned by changing
F igure 14. Schematic of composite synthesis with optically
active polymers and small molecules. (a) Glass matrix, (b)
zeolite matrix, (c) polymer matrix.
Reviews Chem. Mater., Vol. 9, No. 6, 19971311
theloading level, changing thesmall-moleculestructure,
and changing the host framework.
2.2.3. Small Molecules with Second-Order Nonlin-
earities in Polymer Matrixes. Tripathy et al. have been
leaders in the study of small molecules with ?2nonlin-
earity embedded in various high-Tg, cross-linkable
polymer matrixes. Several variations of matrixes and
small molecules have been studied. As shown in Figure
14c, after mixing the polymer and small molecules, the
composites were poled during the cross-linking and
cooling procedures to lock in alignment of the ?2
molecules. In some cases the NLO chromophore could
react with the matrix to further lock in orientation.
While these hybrids are not true composites since they
are largely single-phase materials, they share many
features of optical composites.
One example of a system in which the matrix and
chromophore interact involves an epoxy-based NLO
polymer and a small-molecule NLO chromophore that
has been functionalized to be chemically reactive with
the epoxy.98By photo-cross-linking in the poled state,
the optical nonlinearity is stabilized, even at 100 °C.
Similar results were found for an alkoxysilane dye,
which was incorporated into a siloxane-based host
polymer.99,100In this case, the host polymer could be
cross-linked and could incorporate the dye using sol-
gel chemistry. The stability of the ?2response at 100
°C was discussed.
Polyimides were alsostudied as matrix materials for
NLO dyes because of their low dielectric constant, ease
of processing, and high-temperaturestability.101Figure
16 shows thestability of thesecond harmonic coefficient
of such a composite at both room temperature and 120
°C. In this case the NLO dye was incorporated intothe
polyimide using sol-gel chemistry. The composite was
quite stable at room temperature; however, it showed
a slight loss of ?2strength at 120 °C. This result is
similar toall of the composites discussed in this section.
Finally, NLO active chromophores have been incorpo-
rated into hexakis(methoxymethyl)melamine matrix
materials.102The melamine matrix shows good trans-
parency and high Tgand can be cross-linked using sol-
gel chemistry. Theresulting composites areslightly less
stable than those shown in Figure 16 but show low
3. Nanocomposites with L aser Amplification
Using the optical composite principles, our group has
developed a method touse the nanocomposite structure
tocreate films with solid-state laser properties.103,104By
synthesizing composites, the cumbersome process of
single-crystal growth typically used for solid-statelasers
can be avoided. Additionally, they can be processed as
technologically useful films or fibers instead of as
monolithic materials. This type of structure may be
very important for optical communications applications.
The solid-state laser material our group has focused
on is chromium-doped forsterite (Cr-Mg2SiO4). In the
form of a singlecrystal, this material has shown tunable
lasing in the technologically attractive near-IR region
(1167-1345 nm).1051300 nm is a particularly impor-
tant wavelength in optical communications because it
is a dispersion minimum for silica waveguide materials.
However, nosolid-statelaser amplifiers for 1300 nm are
currently available. Using thenanocompositestructure,
we have developed laser amplifying films that contain
Unlike the bulk of the work with semiconductors, our
composites are prepared ex situ, with much larger
particle sizes (∼100 nm), which require tailored refrac-
tive index matrix materials to avoid significant scat-
tering. Three steps used in this nanocomposite synthe-
sis are outlined in Figure 17. First, small particles of
forsteritewereprepared using a dispersion-polymerized
prepolymer.106The polymer was based on silicon- and
magnesium-containing methacrylate monomers, which
were randomly copolymerized resulting in 100-500 nm
size beads. The beads acted as size templates as they
were heated to 1000 °C to remove the organics and
crystallize the forsterite. The resulting forsterite par-
ticles were about 100 nm in size.
Second, a polymeric matrix material was prepared
with the average refractive index of forsterite (1.652 at
589.3 nm).107Aromatic and brominated monomers were
used toattain this high RI. Several copolymer systems
werestudied as potential matrix materials, but themost
successful was the tribromostyrene/naphthyl methacry-
late system. The refractive index could easily be fixed
F igure 15. Absorption spectra of phycoerythrin in sol-gel
glass at several stages of the sol-gel process (a-c) and in
solution (d). Reprinted with permission from ref 96.
F igure 16. Time behavior of the second harmonic coefficient
of a polyimide-small molecule composite. Reprinted with
permission from ref 101.
1312 Chem. Mater., Vol. 9, No. 6, 1997Reviews
to 1.65 by the composition of the copolymer, and the
material showed good near-IR transparency and good
film-forming properties. Finally, the matrix and par-
ticles were mixed, and a film was cast on a glass
substrate. Typical film dimensions were 2 cm long, 1
cm wide, and 5 µm thick.
Optical measurements were performed on the com-
posite films to examine amplification behavior.1The
experiment consisted of collinear beams of a Cr forster-
ite reference at 1.24 µm and a Nd-yittrium aluminum
garnet pump at 1.06 µm which were passed through the
film as shown in Figure 18a.
blocked independently from the reference, and was
removed from the signal using a monochromator. By
comparing the detected signal at the reference wave-
length before and during pumping of the sample, a
2-fold increase in signal was observed. Figure 18b
shows therelationship between amplification and pump
power for these composites. The gain of 300 dB/m
exceeds the gain of single-crystal Cr forsterite as well
as Er-doped silica fibers used in optical communications
at 1.55 µm, which show a gain of 3 dB/m. Other solid-
state laser materials and optically functional ceramics
are expected to benefit from composite structures such
Recently our group has extended the particle synthe-
sis and composite construction to another system.1Cr
diopside (Cr-CaMgSi2O6) is a material that fluoresces
in thenear-IR (700-1200 nm); however, a singlecrystal
of this material cannot be prepared due toincongruent
melting. Thus the potentially useful solid-state laser
properties of this material have not been studied. We
have synthesized this material using a method similar
to the Cr forsterite synthesis and embedded the par-
The pump could be
ticles in a high RI matrix. Optical amplification of 95
dB/m has been measured in these composites from 760
to 810 nm using a Ti sapphire laser as the reference
signal and an argon laser as a pump. This particular
composite system is an excellent example of how nano-
composites can be used to improve the processability
and functionality of a material which otherwise was
impossible to use.
4. Practical Applications and Issues
4.1. R eal Devices. A limited amount of work has
been published on devices based on polymer semicon-
ductor composites including two papers that discuss
electroluminescenceeffects. Colvin et al. prepared light-
emitting diodes from a layered CdSe/p-phenylenevi-
nylene(PPV) composite.108Theeffects from thepolymer
could beseparated from theCdSein thediodeoperation.
The very low threshold voltage of 4 V, lower than for
PPV alone, caused luminescence from the CdSe from
red to yellow (depending on particle size), while larger
voltages caused the PPV toluminesce in the green. This
is seen in Figure 19 as a change in the shape of the
electroluminescence curve with changing voltage. This
voltage-tunable light source could have important ap-
plications in display technology where the applied
voltage could determine the color of a pixel. Three
particle diameters were examined.
Dabbousi et al. showed electroluminescence using a
homogeneous composite of 5-10% CdSe in poly(vinyl-
carbazole) and an oxadiazolederivative.109Experiments
with threeparticlesizes wereperformed, with emissions
from 530 to 650 nm. The turn-on voltage for these
composites was a bit higher, at 13 V. These composites
also displayed a voltage dependence on light emission,
going from red to white with increased voltage.
F igure 17. Schematic of ex situ optical composite synthesis
for laser amplifying films.
F igure 18. (a) Schematic of the amplification experiment. (b)
Amplification in a Cr forsterite nanocomposite as a function
of pump power. Reprinted with permission from ref 104.
Reviews Chem. Mater., Vol. 9, No. 6, 19971313
4.2. Issues in Nonlinear Optical Composites.
Traditional nonlinear optical materials such as barium
titanatearedifficult toprocess and integrateintooptical
applications as they are typically used in single-crystal
form. Additionally, stronger nonlinear optical proper-
ties are desirable to make these materials technologi-
cally useful. Commonly studied semiconductors such
as CdS and PbS must be used in the nanocrystal form
in order to take advantage of their optical properties
as they are opaque in the bulk and the nonlinear optical
properties are enhanced at small crystal sizes. Using
nanocomposite structures the processability and stabil-
ity of these materials can be greatly improved. Several
issues remain to be solved before these materials will
become integrated into mainstream devices.
In quantum-confined semiconductor composites, the
particle size and size distributions still need greater
control. Enhancement in the optical properties can
come only when a very narrow distribution of controlled
particle size is used. Additionally, the ability to gener-
ate a range of particle sizes in one system is desirable
as this would allow operation at many wavelengths.
Current techniques are showing improvement in this
area, but more is needed to make these materials
technologically useful. In systems with polymers and
small molecules, stability and processability are key to
making useful devices.
4.3. Issues in Optical Amplification with Com-
posites. The area of optically amplifying composites
is quite new, and many issues need to be solved before
these composites can be integrated into real optical
systems. Thesizeof theparticles used is very important
to the scattering characteristics of the composites.
Current techniques discussed here provide laser par-
ticles that are larger than would be desirable to avoid
all scattering. Additionally the uniformity of the com-
posites must be studied toensure that the particles are
dispersed evenly and without agglomeration in the
composites. Finally, the stability of the composites,
particularly the polymer matrix materials, must be
evaluated. Both prolonged exposure to laser radiation
and heat evolution from the laser particles may be
5. Other Optical Applications
5.1. Composites for High R efractive Index Ap-
plications. High refractive index polymers have many
applications ranging from antireflection coatings for
solar cells tohigh RI lenses. Our group has synthesized
high RI copolymers (up to 1.7) as index-matched ma-
trixes for solid-statelaser containing composites (section
3).107Semiconductor polymer composites can exhibit
even higher refractive indices due to the high RI of the
In the work by Zimmerman et al., composites of PbS
and gelatin were synthesized and continuously in-
creased the refractive index of the gelatin from 1.5 to
2.5.110Thesematerials represented someof thehighest
refractive index polymers available and were targeted
for antireflection coatings for solar cell materials. Kyp-
rianidou-Leodidou et al., from the same research group,
varied the size of PbS in poly(ethylene oxide) from 4 to
80 nm by adding surfactants and acids.77Again optical
measurements were used to evaluate refractive index,
showing composite RI values as high as 3.9. As shown
in Figure 20, the extrapolated RI of particles greater
than 25 nm in size was found toagree with the bulk RI
of PbS (4.3). However, the RI was highly dependent on
particle size below 20 nm, with 4 nm particles showing
a RI of about 2.5. These effects may be due toquantum
confinement as strong confinement should become
evident at particle sizes near and below 20 nm (Table
2). The absorption of the composites at 632.8 nm
decreased as the particle size decreased, alsoindicating
confinement effects. TEM and X-ray were used to
characterize the composites and showed broad size
5.2. T ransparent Magnetic Composites. Trans-
parent magnetic materials have many applications
including microwave magnetooptical modulation of vis-
ible lasers with very low modulation power per unit
bandwidth, optical deflection and isolation, magnetoop-
tic displays, and holograms. Usable transparent mag-
nets must have high Curie temperatures, avoid bire-
fringence, and have fundamental absorption edges in
thenear UV region. Thecompositestructures discussed
here may make transparent magnetic materials a
Several groups have prepared magnetic composites
with both polymer and glass matrixes. Okada et al.
reported the production of sub-100-nm Fe2O3particles
in a multibilayer film.111Opaque composite films that
F igure 19. Electroluminescence from a CdSe/PPV composite.
Reprinted with permission from ref 108.
F igure 20. Effects of PbS particle size on the refractive index
at 632.8 nm. Nanocomposite values (O), extrapolated values
for PbS (b). Reprinted with permission from ref 77.
1314 Chem. Mater., Vol. 9, No. 6, 1997Reviews
contained magnetite behaved as ferromagnets with
coercivities of 20 Oe. Papaefthymiou et al. reported
magnetic studies of 9-20 nm Fe particles coated with
Fe2O3embedded in wax, but the optical properties were
not mentioned.112Shull et al. prepared Fe2+and Fe3+
confined to20 nm regions in a silica matrix using a sol-
Again the magnetic properties were
examined with nomention of transparency. Zhaoet al.
discussed work embedding colloidal Fe2O3 magnetic
particles in bilayer lipid membranes.114The samples
were thin enough to be studied in transmission, but
general absorption properties were not measured. Re-
flective properties and the Kerr and Faraday rotations
Zioloet al. reported the production of optically trans-
parent magnetic composites for the first time.115γ-Fe2O3
was grown in a polymeric ion-exchange resin by intro-
ducing the Fe3+and then exposing to H2O2 at 60 °C,
very similar tothe process for preparing CdS in Nafion
shown in Figure 3 (steps 1b, 2c). The particles created
ranged from 50 to 250 Å in size, which resulted in
superparamagnetic properties. If the ionomer resin
were loaded multiple times, the particle size did not
increase but the concentration of particles did. The
composites showed appreciable magnetization at and
below room temperature. The highest saturation mo-
ment observed was 46 emu/g for 250 Å particles, which
is considerably stronger than transparent crystals of
FeBO3(3 emu/g) and FeF3(1 emu/g). The absorption
coefficient was nearly an order of magnitude less than
the bulk crystal, and the composite had a refractive
index of 1.6.
Clement et al. presented very promising results for a
material which exhibits both ?2nonlinearity and mag-
netic properties.116This was accomplished by interca-
lating a ?2small molecule into a ferrimagnetic layered
material. The ?2response showed very high efficiency,
and the magnetic properties could be seen below 40 K.
Miller has alsooutlined many potential applications for
5.3. Hard T ransparent Coatings. By combining
the strength and hardness of ceramics with the pro-
cessability and ductility of polymers, novel transparent
materials have been synthesized. Typically these ma-
terials are prepared using an in situ synthesis like that
shown in Figure 14a. The polymer (or polymer precur-
sor) and silica precursor (or other alkoxide) are mixed
in solution and cast as a film or monolith which is
subsequently heated to convert the precursors to their
final form. Hard transparent coatings and barrier
layers are the primary applications of polymer nano-
composites prepared with reinforcing ceramic phases.
A significant amount of work in this area has been with
polyimide-silica hybrid materials which have high
thermal stability, toughness, and hardness.
Our group at Cornell has investigated silica/polyimide
materials that show improved hardness and modulus
while retaining transparency.118The hybrids are based
on two different silicon-containing poly(amic acids)
(PAA) which are mixed with the silica in situ using sol-
gel chemistry. The poly(amic acids) are tailored to be
chemically reactive with the silica network. One PAA
contains a short siloxane segment which can be cleaved
tochemically bond with thesilica as it forms. Theother
PAA is end functionalized with ethoxy groups which can
participate in the sol-gel chemistry.
The hybrids formed extremely small silica domains
as shown by SEM studies.119As expected, the hardness
of both types of hybrid increased with increasing silica
content as shown in Figure 21. A 17-fold increase in
hardness was observed for the siloxane-containing
hybrids. Themodulus of thehybrids was alsoimproved
with increasing silica content. Other work in the area
of polyimide-silica hybrids has confirmed that chemical
reactivity between phases improves homogeneity of the
6. Conclusions and Prospects
Nanocomposite structures based on embedded func-
tional materials in processable matrixes have many
optical applications. Particulatephases can beprepared
in situ or ex situ and may have properties ranging from
laser amplification to improved hardness. Matrix ma-
terials can be polymers, glasses, or ceramics. Compos-
ites such as these have been used to prepare materials
with nonlinear optical properties, laser properties,
magnetic properties, and those with high refractive
Semiconductor particles have a wealth of optical
functions and can be prepared in polymers, ceramics,
and glasses. Of particular interest are their third-order
nonlinear optical properties, which may be useful for
all optical switches. By processing in film or fiber forms,
utilizing a composite structure, these devices will be
more easily integrated intotargeted applications. Com-
posites containing optical polymers and small molecules
find advantages over the bulk materials in stability and
processability. Solid-state laser materials show great
promise as amplifying films in the near-IR and beyond.
Finally, novel transparent magnetic materials havealso
been synthesized using nanocomposite principles and
have prospects in printing, data storage, and magne-
Acknowledgment. L.L.B. is grateful totheNational
Science Foundation, Department of Education, and
Cornell Materials Science Center for support of this
research. Eric Kutcher participated in the PbS/PVA
F igure 21. Hardness of siloxane containing (b) and end
functionalized (2) hybrids as a function of silica content.
Reprinted with permission from ref 119.
Reviews Chem. Mater., Vol. 9, No. 6, 19971315
composite studies. Todd Krauss and Inuk Kang have Download full-text
contributed to useful discussions.
(1) Beecroft, L. L. Advanced Nanocomposite Materials for Optical
Applications, Ph.D. Thesis, Cornell University Department of
Materials Science and Engineering, 1996; Chapter 4.
(2) Hillinski, E. F.; Lucas, P. A.; Wang, Y. J . Chem. Phys. 1988, 89,
(3) Wang, Y.; Suna, A.; McHugh, J .; Hillinski, E. F.; Lucas, P. A.;
J ohnson, R. D. J . Chem. Phys. 1990, 92, 6927.
(4) Majetich, S. A.; Carter, A. C. J . Phys. Chem. 1993, 97, 8727.
(5) Krauss, T. D.; Wise, F. W. Appl. Phys. Lett. 1994, 65, 1739.
(6) J ain, R. K.; Lind, R. C. J . Opt. Soc. Am. 1983, 73, 647.
(7) Nogami, M.; Nakamura, A. Phys. Chem. Glasses 1993, 34, 109.
(8) Wung, C. J .; Wijekoon, W. M. K. P.; Prasad, P. N. Polymer 1993,
(9) J ustus, B. L.; Tonucci, R. J .; Berry, A. D. Appl. Phys. Lett. 1992,
(10) Adair, R.; Chase, L. L.; Payne, S. A. J . Opt. Soc. Am. B 1987, 4,
(11) Weller, H. Adv. Mater. 1993, 5, 88.
(12) Alivisatos, A. P. Science 1996, 271, 933.
(13) Pepper, D. M. Sci. Am. 1986, 254, 74.
(14) Stegeman, G. I.; Seaton, C. T. J . Appl. Phys. 1985, 58, R57.
(15) Brus, L. E. IEEE J . Quantum Electron. 1986, QE-22, 1909.
(16) Steigerwald, M. L.; Brus, L. E. Annu. Rev. Mater. Sci. 1989, 19,
(17) Steigerwald, M. L.; Brus, L. E. Acc. Chem. Res. 1990, 23, 183.
(18) Wang, Y. Acc. Chem. Res. 1991, 24, 133.
(19) Wang, Y.; Herron, N. Res. Chem. Intermediates 1991, 15, 17.
(20) Wang, Y.; Herron, N. J . Phys. Chem. 1991, 95, 525.
(21) Sipe, J . E.; Boyd, R. W. Phys. Rev. A 1992, 46, 1614.
(22) Boyd, R. W.; Sipe, J . E. J . Opt. Soc. Am. B 1994, 11, 297.
(23) Boyd, R. W.; Sipe, J . E. Non-Linear Optics and Optical Physics;
Khoo, I. C., Lam, J . F., Simoni, F., Eds.; World Scientific:
Singapore, 1994; p 104.
(24) Fischer, G. L.; Boyd, R. W.; Gehr, R. J .; J enekhe, S. A.; Osaheni,
J . A.; Sipe, J . E.; Weller-Brophy, L. A. Phys. Rev. Lett. 1994, 74,
(25) Banyai, L.; Hu, Y. Z.; Lindberg, M.; Koch, S. W. Phys. Rev. B
1988, 38, 8142.
(26) Ozin, G. A.; Kirkby, S.; Meszaros, M.; Ozkar, S.; Stein, A.;
Stucky, G. D. Materials for Nonlinear Optics; Marder, S. R.,
Sohn, J . E., Stucky, G. D., Eds.; American Chemical Society:
Washington, DC, 1991; Vol. 455, p 554.
(27) PbS, PbSe, and CdSe values calculated by Inuk Kang.
(28) Wang, Y.; Suna, A.; Mahler, W.; Kasowski, R. J . Chem. Phys.
1987, 87, 7315.
(29) Nenadovic, M. T.; Comor, M. I.; Vasic, V.; Micic, O. I. J . Phys.
Chem. 1990, 94, 6390.
(30) Alivisatos, A. P.; Harris, A. L.; Levinos, N. J .; Steigerwald, M.
L.; Brus, L. E. J . Chem. Phys. 1988, 89, 4001.
(31) Kuczynski, J . P.; Milosavljevic, B. H.; Thomas, J . K. J . Phys.
Chem. 1984, 88, 980.
(32) Ohtani, B.; Adzuma, S.; Nishimoto, S.; Kagiya, T. J . Polym. Sci.,
Part C: Polym. Lett. 1987, 25, 383.
(33) Mahler, W. Inorg. Chem. 1988, 27, 435.
(34) Gallardo, S.; Gutierrez, M.; Henglein, A.; J anata, E. Ber. Bunsen-
Ges. Phys. Chem. 1989, 93, 1080.
(35) Beecroft, L. L. Advanced Nanocomposite Materials for Optical
Applications, Ph.D. Thesis, Cornell University Department of
Materials Science and Engineering, 1996; Chapter 6.
(36) Berry, C. R. Phys. Rev. 1967, 161, 848.
(37) Heinglein, A. Ber. Bunsen-Ges. Phys. Chem. 1982, 86, 301.
(38) Ramsden, J . J .; Gratzel, M. J . Chem. Soc., Faraday Trans. 1
1984, 80, 919.
(39) Kalyanasundaram, K.; Borgarello, E.; Duonghong, D.; Gratzel,
M. Angew. Chem., Int. Ed. Engl. 1981, 20, 987.
(40) Akimov, I. A.; Denisyuk, I. Yu.; Meshkov, A. M. Opt. Spectrosc.
1992, 72, 558.
(41) Meissner, D.; Memming, R.; K astening, B. Chem. Phys. Lett.
1983, 96, 34.
(42) Krishnan, M.; White, J . R.; Fox, M. A.; Bard, A. J . J . Am. Chem.
Soc. 1983, 105, 7002.
(43) Mau, A. W.; Huang, C.; Kakuta, N.; Bard, A. J .; Campion, A.;
Fox, M. A.; White, J . M.; Webber, S. E. J . Am. Chem. Soc. 1984,
(44) Wang, Y.; Mahler, W. Opt. Commun. 1987, 61, 233.
(45) Wang, Y.; Suna, A.; Mahler, W. Mater. Res. Soc. Symp. Proc.
1988, 109, 187.
(46) Wang, Y.; Herron, N.; Mahler, W.; Suna, A. J . Opt. Soc. Am. B
1989, 6, 808.
(47) Ohashi, Y.; Ito, H.; Hayashi, T.; Nitta, A.; Matsuda, H.; Okada,
S.; Nakanishi, H.; Kato, M. Springer Proc. Phys. 1989, 36, 81.
(48) Weller, H. Angew. Chem., Int. Ed. Engl. 1996, 35, 1079.
(49) Murray, C. B.; Kagan, C. R.; Bawendi, M. G. Science1995, 270,
(50) Sankaran, V.; Cummins, C. C.; Schrock, R. R.; Cohen, R. E.;
Silbey, R. J . J . Am. Chem. Soc. 1990, 112, 6858.
(51) Cummins, C. C.; Schrock, R. R.; Cohen, R. E. Chem. Mater. 1992,
(52) Sankaran, V.; Yue, J .; Cohen, R. E.; Schrock, R. R.; Silbey, R. J .
Chem. Mater. 1993, 5, 1133.
(53) Tassoni, R.; Schrock, R. R. Chem. Mater. 1994, 6, 744.
(54) Mo ¨ller, M. Synth. Met. 1991, 41-43, 1159.
(55) Spatz, J . P.; Roescher, A.; Mo ¨ller, M. Adv. Mater. 1996, 8, 337.
(56) Yuan, Y.; Cabasso, I.; Fendler, J . Macromolecules 1990, 23, 3198.
(57) Yuan, Y.; Fendler, J .; Cabasso, I. Chem. Mater. 1992, 4, 312.
(58) Warnock, J .; Awschalom, D. D. Phys. Rev. B 1985, 32, 5529.
(59) Cullen, T. J .; Ironside, C. N.; Seaton, C. T.; Stegeman, G. I. Appl.
Phys. Lett. 1986, 49, 1403.
(60) Roussignol, P.; Ricard, D.; Rustagi, K. C.; Flytzanis, C. Opt.
Commun. 1985, 55, 143.
(61) Roussignol, P.; Ricard, D.; Lukasik, J .; Flytzanis, C. J . Opt. Soc.
Am. B 1987, 4, 5.
(62) Borelli, N. F.; Hall, D. W.; Holland, H. J .; Smith, D. W. J . Appl.
Phys. 1987, 61, 5399.
(63) Sombra, A. S. B. Solid State Commun. 1992, 82, 805.
(64) Kuczynski, J .; Thomas, J . K. J . Phys. Chem. 1985, 89, 2720.
(65) Luong, J . C. Superlattices Microstruct. 1988, 4, 385.
(66) Wang, Y.; Herron, N. J . Phys. Chem. 1987, 91, 257.
(67) Herron, N.; Wang, Y.; Eddy, M. M.; Stucky, G. D.; Cox, D. E.;
Moller, K.; Bein, T. J . Am. Chem. Soc. 1989, 111, 530.
(68) Wang, Y.; Herron, N. J . Phys. Chem. 1988, 92, 4988.
(69) Moller, K.; Bein, T.; Herron, N.; Mahler, W.; Wang, Y. Inorg.
Chem. 1989, 28, 2914.
(70) Levy, D.; Esquivias, L. Adv. Mater. 1995, 7, 120.
(71) Rajh, T.; Vucemilovic, M. I.; Dimitrijevic, N. M.; Micic, O. I.;
Nozik, A. J . Chem. Phys. Lett. 1988, 143, 305.
(72) Nogami, M.; Nagasaka, K.; Takata, M. J . Non-Cryst. Solids 1990,
(73) Nogami, M.; Nagaska, K.; Kato, E. J . Am. Ceram. Soc. 1990,
(74) Othmani, A.; Bovier, C.; Dumas, J .; Champagnon, B. J . Phys.
IV 1992, 2, C2-275.
(75) Nogami, M.; Kato, A.; Tanaka, Y. J . Mater. Sci. 1993, 28, 4129.
(76) Danek, M.; J ensen, K. F.; Murray, C. B.; Bawendi, M. G. J . Cryst.
Growth 1994, 145, 714.
(77) Kyprianidou-Leodidou, T.; Caseri, W.; Suter, U. J . Phys. Chem.
1994, 98, 8992.
(78) Nosaka, Y.; Yamaguchi, K.; Miyama, H.; Hayashi, H. Chem. Lett.
(79) Steigerwald, M. L.; Alivisatos, A. P.; Gibson, J . M.; Harris, T.
D.; Kortan, R.; Muller, A. J .; Thayer, A. M.; Duncan, T. M.;
Douglass, D. C.; Brus, L. E. J . Am. Chem. Soc. 1988, 110, 3046.
(80) Fischer, C.; Heinglein, A. J . Phys. Chem. 1989, 93, 5578.
(81) Swayambunathan, V.; Hayes, D.; Schmidt, K. H.; Liao, Y. X.;
Meisel, D. J . Am. Chem. Soc. 1990, 112, 3831.
(82) Lawless, D.; K apoor, S.; Meisel, D. J . Phys. Chem. 1995, 99,
(83) Herron, N.; Wang, Y.; Eckert, H. J . Am. Chem. Soc. 1990, 112,
(84) Cheng, L. T.; Herron, N.; Wang, Y. J . Appl. Phys. 1989, 66, 3417.
(85) Herron, N.; Suna, A.; Wang, Y. J . Chem. Soc., Dalton Trans.
(86) Markichev, I.; Sheka, E.; Natkaniec, I.; Muzychka, A.; Khavryutch-
enko, V.; Wang, Y.; Herron, N. Physica B 1994, 198, 197.
(87) Wung, C. J .; Pang, Y.; Prasad, P. N.; Karasz, F. E. Polymer 1991,
(88) Pang, Y.; Samoc, M.; Prasad, P. N. J . Chem. Phys. 1991, 94,
(89) He, G. S.; Wung, C. J .; Xu, G. C.; Prasad, P. N. Appl. Opt. 1991,
(90) Avnir, D.; Levy, D.; Reisfeld, R. J . Phys. Chem. 1984, 88, 5956.
(91) Haruvy, Y.; Webber, S. E. Chem. Mater. 1991, 3, 501.
(92) Avnir, D.; Braun, S.; Lev, O.; Ottolenghi, M. Chem. Mater. 1994,
(93) Ellerby, L. M.; Nishida, C. R.; Nishida, F.; Yamanaka, S. A.;
Dunn, B.; Valentine, J . S.; Zink, J . I. Science 1992, 255, 1113.
(94) Wu, S.; Ellerby, L. M.; Cohan, J . S.; Dunn, B.; El-Sayed, M. A.;
Valentine, J . S.; Zink, J . I. Chem. Mater. 1993, 5, 115.
(95) Weetall, H. H.; Robertson, B.; Cullin, D.; Brown, J .; Walch, M.
Biochim. Biophys. Acta 1993, 1142, 211.
(96) Chen, Z.; Samuelson, L. A.; Akkara, J .; Kaplan, D. L.; Gao, H.;
Kumar, J .; Marx, K. A.; Tripathy, S. K. Chem. Mater. 1995, 7,
(97) Cox, S. D.; Gier, T. E.; Stucky, G. D. Chem. Mater. 1990, 2, 609.
(98) J eng, R. J .; Chen, Y. M.; Kumar, J .; Tripathy, S. K. J . Macromol.
Sci.sPure Appl. Chem. 1992, A29, 1115.
(99) J eng, R. J .; Chen, Y. M.; J ain, A. K.; Kumar, J .; Tripathy, S. K.
Chem. Mater. 1992, 4, 972.
(100) J eng, R. J .; Chen, Y. M.; J ain, A. K.; Tripathy, S. K.; Kumar, J .
Opt. Commun. 1992, 89, 212.
(101) J eng, R. J .; Chen, Y. M.; J ain, A. K.; Kumar, J .; Tripathy, S. K.
Chem. Mater. 1992, 4, 1141.
1316 Chem. Mater., Vol. 9, No. 6, 1997Reviews