This document summarizes research on creating monolayer composite films of alumina particles and epoxy using a self-assembly process called fluid forming. Key points: - Fluid forming uses a high-spreading fluid to propel ceramic particles across an air-water interface, compressing them into a dense, ordered monolayer on a substrate. - Initial attempts using epoxy instead of mineral oil in fluid forming with alumina particles resulted in nonuniform, low density films. - Replacing the solvent 2-propanol with a 4:1 mixture of 2-propanol and 1-butanol produced uniform, high density alumina/epoxy composite films via fluid forming.

1. Self-Assembly of Monolayer-Thick Alumina Particle-Epoxy
Composite Films
Bryan R. Jackson,† Xiangyuan Liu,‡ Elizabeth F. McCandlish,§ and Richard E. Riman*,†
Department of Materials Science and Engineering, Rutgers, The State UniVersity of New Jersey, 607
Taylor Road, Piscataway, New Jersey 08854, JH Research USA, LLC, 10901 Elm AVenue,
Fontana, California 92337, and Ceramare´ Corporation, 12-D Jules Lane,
New Brunswick, New Jersey 08901
ReceiVed January 17, 2007. In Final Form: August 14, 2007
Monolayer-thick composite films composed of R-alumina and Spurr’s epoxy were prepared via a self-assembly
process known as fluid forming. The process makes use of a high-spreading-tension fluid composed of volatile and
nonvolatile components to propel particles across the air-water interface within a water bath. Continuous addition
of the particle suspension builds a 2D particle film at the air-water interface. The spreading fluid compresses the
film into a densely packed array against a submerged substrate. The assembled monolayer is deposited onto the
substrate by removing the substrate from the bath. A dispersion containing a narrow size distribution, 10 µm R-alumina
particles, light mineral oil, and 2-propanol was spread at the air-water interface and the alumina particles were
assembled into densely packed arrays with an aerial packing fraction (APF) of 0.88. However, when mineral oil was
replaced by Spurr’s epoxy nonuniform films with low packing density resulted. It was found that replacing 2-propanol
with a mixture of 2-propanol and 1-butanol with a volume ratio of 4:1 produced uniform, densely packed alumina/
epoxy composite films. The role of the solvent mixture will be discussed.
Introduction
Organic-inorganic composite materials have received con-
siderable attention because of their improved mechanical, optical,
and/or chemical properties. Monolayer and multilayer film
structures can be used as pervaporation membranes,1 ion
conducting films for batteries,2-4 high capacitance films,5,6
pyroelectric transducers,7-9 piezoelectric transducers,10,11 humid-
ity sensors,12 biosensors,13,14 and low loss optical media15,16 A
variety of methods have been developed to produce composite
films with tailored composition and properties. These include
the sol-gel process,13,17,18 spin coating,8,18 the roll-coating
method,6 electrophoretic deposition,19-22 and electrostatic layer-
by-layer assembly.23 These methods have proven to be very
effective for making composite films but do not make highly
ordered arrays of particles. Thin composite films are difficult to
produceusingparticlesuspensionswithhighsolidsconcentrations
because of the various interactions in suspension that lead to
high viscosity and hence low particle mobility. Low particle
mobility prevents the assembly of dense particle arrays. Slow
evaporation of solvent leads to ordered packing but requires a
long time for complete packing.24
Thin films composed of particle monolayer structures are of
interest because they possess several advantages over multilayer
structures for a wide range of functional and structural
materials. For example, monolayers can be used to functionalize
surfaces of another material to give it behavior characteristic of
the particle layer rather than the underlying bulk material. Thus,
particles that are expensive or not abundant can be conserved.
More specifically, the integrated optoelectronics community can
benefit from monolayer structures because they can push the
integration length scale to lower limits bounded by the diameter
of the particles used to assemble the structures.
The Langmuir-Blodgett (LB) method is a classical process
to make particle monolayers. The LB method applies surface
pressure by mechanical means via a movable barrier to compress
particles at the air-water interface and typically requires long
* To whom correspondence should be addressed.
† Rutgers, The State University of New Jersey.
‡ JH Research USA.
§ Ceramare´ Corporation.
(1) Nunes, S. P.; Ruffmann, B.; Rikowski, E.; Vetter, S.; Richau, K. J. Membr.
Sci. 2002, 203, 215-225.
(2) Kumar, B.; Rodrigues, S. J.; Spry, R. J. Electrochim. Acta 2002, 47, 1275-
1281.
(3) Leo, C. J.; Rao, G. V. S.; Chowdari, B. V. R. Solid State Ionics 2002, 148,
159-171.
(4) Croce, F.; Scrosati, B. AdV. Membr. Technol. 2003, 984, 194-207.
(5) Bhattacharya, S. K.; Tummala, R. R. J. Mater. Sci.: Mater. Electron.
2000, 11, 253-268.
(6) Cho, S. D.; Lee, J. Y.; Hyun, J. G.; Paik, K. W. Mater. Sci. Eng., B 2004,
110, 233-239.
(7) Sakamoto,W.K.;Kanda,D.H.F.;Das-Gupta,D.K.Mater.Res.InnoVations
2002, 5, 257-260.
(8) Sakamoto,W.K.;Shibatta-Kagesawa,S.T.;Melo,W.L.B. Sens.Actuators,
A 1999, 77, 28-33.
(9) Guggilla, P.; Batra, A. K.; Aggarwal, M. D.; Lal, R. B. In InVestigation
on Nanocomposites for Pyroelectric Infrared Sensors; Organic Photonic Materials
and Devices VII; Grote, J. G., Kaino, T., Kajzar, F., Eds.; SPIE: Bellingham,
WA, 2005; pp 295-300.
(10) White, J. R.; De Poumeyrol, B.; Hale, J. M.; Stephenson, R. J. Mater.
Sci. 2004, 39, 3105-3114.
(11) Dias, C. J.; Dasgupta, D. K. Ferroelectr. Polym. Ceram. Polym. Compos.
1994, 92-9, 217-247.
(12) Hirano, S.; Kishimoto, A.; Miyayama, M. Jpn J. Appl. Phys., Part 2 1998,
37, L1158-L1161.
(13) Sampath, S.; Lev, O. Anal. Chem. 1996, 68, 2015-2021.
(14) Rea, S. M.; Best, S. M.; Bonfield, W. J. Mater. Sci.: Mater. Med. 2004,
15, 997-1005.
(15) Carotenuto, G.; Nicolais, L.; Kuang, X. Appl. Compos. Mater. 1996, 3,
103-116.
(16) Almaral-Sanchez, J. L.; Rubio, E.; Mendoza-Galvan, A.; Ramirez-Bon,
R. J. Phys. Chem. Solids 2005, 66, 1660.
(17) Rouse, J. H.; MacNeill, B. A.; Ferguson, G. S. Chem. Mater. 2000, 12,
2502-2507.
(18) Hashizume, M.; Kunitake, T. Langmuir 2003, 19, 10172-10178.
(19) Peng, X. G.; Manna, L.; Yang, W. D.; Wickham, J.; Scher, E.; Kadavanich,
A.; Alivisatos, A. P. Nature 2000, 404, 59-61.
(20) Hamagami, J.; Kanamura, K.; Umegaki, T. Asian Ceram. Sci. Electron.
II,Proc.2ndAsianMeet.Electroceram.;Electroceram.Jpn.V,Proc.21stElectron.
DiV. Meet. Ceram. Soc. Jpn., 2001 2002, 228-2, 255-258.
(21) Zhitomirsky, I. Electrodeposition, of lanthanum hydroxide-polyethylen-
imine films. Mater. Lett. 2003, 57, 3761-3766.
(22) Zhitomirsky, I. Surf. Eng. 2004, 20, 43-47.
(23) Decher, G. Science 1997, 277, 1232-1237.
(24) Yu, K.; Hurd, A. J.; Eisenberg, A.; Brinker, C. J. Langmuir 2001, 17,
7961-7965.
11399Langmuir 2007, 23, 11399-11403
10.1021/la070138w CCC: $37.00 © 2007 American Chemical Society
Published on Web 10/06/2007

2. times to achieve equilibrium and complete particle packing.25 In
the present work, fluid forming, a self-assembly process that is
a variant of the LB method, was used to synthesize organic-
inorganic particle composite films. The general procedure for
fluid forming has been described previously.26 The process makes
use of a high-spreading-tension fluid composed of volatile (e.g.,
2-propanol) and nonvolatile (mineral oil) components to propel
particles across the air-water interface within a water bath.
Continuousadditionoftheparticlesuspensionbuildsa2Dparticle
film at the air-water interface. The spreading fluid compresses
thefilmintoadenselypackedarrayagainstasubmergedsubstrate,
and the assembled monolayer is deposited onto the substrate by
removing the substrate from the bath.
Fluid forming is capable of forming densely packed
monolayer arrays in several minutes that cover areas greater
than 10 cm2. The ability to produce expansive films in a short
time is an attractive feature of this process. Relative to the LB
method, fluid forming can be used on a continuous basis whereas
LB is a batch process. In addition, fluid forming is also easier
to perform and does not require expensive equipment.
Our previous work examined the assembly of ∼2 µm
monodisperse uniform PZT powders onto a substrate and then
incorporated a solvent-soluble polymer in a second processing
step to make a polymer-ceramic composite film that has a high
volume fraction of the ceramic phase. This paper demonstrates
recent improvements in fluid forming that make the process
more practical for making 2D ceramic particle-polymer
composite films. First, we will examine the utility of commercial
ceramic powders. Second, we will look at assembling particles
that are not uniform in shape. Third, we will examine the
possibility of integrating fluid forming with a polymerization
process and show that we can do this with a commercial polymer.
Alumina particles (∼10 µm) and Spurr’s epoxy were selected
as a model system for the primary study of this process. Alumina
is commercially available as a narrow-size-distribution powder
and in a wide range of mean sizes. Spurr’s epoxy has a well-
documented history as an embedding media27 and has well-
established methodologies for low-temperature polymerization
of monomer species. Because the reacting components of
Spurr’s epoxy are soluble in alcohol, they can be incorporated
into fluid forming by substituting for the mineral oil constituent.
Because Spurr’s epoxy is optically transparent, an optical
microscope can be used to characterize the packing of the alumina
in the assembled film. Thus, the objective of this work is to
investigate how fluid forming can be used to make monolayer
alumina powder-epoxy composite films with a high ceramic
packing fraction. Such films could be utilized to produce hard,
scratch-resistant coatings.28
Experimental Section
Materials. Sumicorundum (10 µm R-alumina powder) was
obtained from Sumitomo Chemical Co. (Tsukuba, Japan) and used
as received. HPLC-grade 2-propanol, light mineral oil (NF/FCC),
andACS-certified1-butanolwereallpurchasedfromFisherScientific
(FairLawn,NJ)andusedasreceived.SPI-Chemlowviscosity“spurr”
kits were purchased from Structure Probe, Inc. (West Chester, PA).
Preparation of Monolayer Alumina Particle Arrays. Light
mineral oil (∼0.0150 g) was added to 5 mL of 2-propanol, and the
mixture was sonicated in an ultrasonication bath (FS30, Fisher
Scientific, Pittsburgh, PA) for 1 min. Alumina powder (0.04-0.06
g) was then added to the solution, which was sonicated for 5 min
followed by mechanical shaking (multi-wrist shaker, Lab-line,
Melrose Park, IL) for 30 min. All of the alumina suspension (5 mL)
was spread at the air-water interface in a 350 mL water bath using
a narrow-tip pipet. The array was then transferred onto a microscope
slide substrate by vertically removing a previously submerged
substrate. The particle array was held at 500 °C for 2 h to remove
organicsfromtheassembledfilm.Auniformaluminafilmisproduced
as shown in Figure 1.
Preparation of Monolayer Alumina-Epoxy Composite Films.
Spurr’s epoxy components were mixed according to the product
instructions. Approximately 0.02 g of the epoxy was then
dissolved in 2.5 mL of 2-propanol. About 0.05 g of alumina
powder was added to 2.5 mL of 2-propanol and sonicated for 2 min
and then shaken on a mechanical shaker for 30 min to make a
suspension. The epoxy/2-propanol solution and alumina/2-propanol
suspension were combined and shaken for another 30 min.
Spreading of the dispersion at the air-water interface produced a
floating particle array. The particle array was transferred to a glass
microscope slide and cured in an oven overnight at 60 °C. In another
experiment, 2-propanol was replaced by a mixture of 2-propanol
and 1-butanol with 4:1 v/v 2-propanol/1-butanol. Film assembly
and curing of the epoxy were done as described for alumina/epoxy
composite films.
Characterization Methods. Following film formation, optical
microscopy initially confirmed that only a single monolayer was
formed.Monolayerfilmswereeasilydistinguishablefrommultilayers
because the particles were transparent to visible light. Thus, the
opticalabsorptionoftwoormorelayersortheoverlappingofparticles
(25) Sastry, M. Nanoparticle Organization at the Air-Water Interface and in
Langmuir-Blodgett Films. In Colloids and Colloid Assemblies: Synthesis,
Modification, Organization and Utilization of Colloid Particles; Caruso, F., Ed.;
Wiley-VCH: Weinheim, Germany, 2004; pp 369-397.
(26) Liu, X. Y.; McCandlish, E. F.; McCandlish, L. E.; Mikulka-Bolen, K.;
Ramesh, R.; Cosandey, F.; Rossetti, G. A.; Riman, R. E. Langmuir 2005, 21,
3207-3212.
(27) Spurr, A. R. J. Ultrastruct. Res. 1969, 26, 31-43.
(28) Bauer, F.; Glasel, H.-J.; Decker, U.; Ernst, H.; Freyer, A.; Hartmann, E.;
Sauerland, V.; Mehnert, R. Prog. Org. Coat. 2003, 47, 147-153.
Figure 1. (a) Photograph of an assembled particle array on a
microscope slide. The array (shown in white) is approximately 2.5
× 5 cm2. (b) Optical micrograph of a transparent alumina film
prepared by fluid forming.
11400 Langmuir, Vol. 23, No. 23, 2007 Letters

3. was easy to see because of the refractive index contrast. Optical
micrographs were collected with an Olympus BH-2 microscope
equipped with an Olympus DP12 digital camera with a resolution
of 1024 pixels × 668 pixels (Olympus America Inc., Center Valley,
PA). Electron micrographs were collected with an Amray 1830
scanning electron microscope (SEM, Amray Inc., Bedford, MA) at
an acceleration voltage of 20 kV. The electron micrographs were
convertedintoblackandwhite(binary)images,andtheaerialpacking
fraction (APF) was determined using ImageJ29 by integration of the
particle coverage.
Results and Discussion
An optical micrograph of a dense alumina film made by the
fluid-formingself-assemblyprocessisshownFigure1b,whereby
the formation of a monolayer is clearly observed. The optical
transparency of alumina to visible radiation enabled us to
distinguish between mono- and multilayer films, even though
the optical anisotropy of alumina leads to slight variations in the
amountoftransmittedlight.SEMmicrographsofaself-assembled
Al2O3 particle array and the corresponding binary image of the
array are shown in Figure 2. The APF for the film was 0.88.
Theoreticalpackingbasedonahexagonallyclose-packedstructure
of spherical particles is 0.91, indicating that the packing density
is approaching a theoretical maximum.
During the self-assembly process, the 2-propanol dispersion
spreads rapidly on the water surface because of the high spreading
tension. The surface forces involved in this process are illustrated
in Figure 3, which shows the force balance of water-vapor,
fluid-vapor, and fluid-water surface tensions. Depending on
the relative forces, a fluid droplet placed on the surface of another
fluid will either take the shape of a lens or will spread across
the surface. For spontaneous spreading to occur, the spreading
tension of a fluid at the air-water interface must be greater than
zero (ΠFWV > 0) and is given as follows
where γWV, γFV, and γFW are the surface tension values at the
air-water, air-fluid, and fluid-water interfaces, respectively.30
The values of the surface and spreading tension for the fluids
used in the experiments are shown in Table 1. The spreading
tension of 2-propanol at 20 °C is 51.1 mN/m, providing fast
spreading at the air-water interface and compression of the
alumina particles against the substrate into a monolayer. Because
2-propanol is both miscible with water and has a high vapor
pressure, it is depleted rapidly from the water surface by both
dissolution and evaporation. As the 2-propanol was removed,
particle rafts formed, which gradually coalesced into a particle
array. Continuous addition/spreading of the particle suspension
at the air-water interface served to increase the density of the
film by compressing it against the substrate. Mineral oil at the
air-particle interface provided a hydrophobic surface, which
kept the particles together in a densely packed array floating on
the water surface. Upon removal of the particle array, the film
was further compressed slightly by moving the substrate in a
diagonal direction to maintain the dense packing of the particle
array.
When mineral oil was replaced with Spurr’s epoxy, a particle
array was also formed by fluid forming, but the microstructure
oftheself-assembledfilmschangedsignificantly.Figure4ashows
an electron micrograph of an alumina/epoxy particle array on a
glass slide after curing the epoxy at 60 °C for 12 h. The resulting
(29) http://rsb.info.nih.gov/ij/.
(30) Adamson, A. W. Physical Chemistry of Surfaces, 5th ed.; John Wiley &
Sons: New York, 1990; p 777.
Figure 2. (a) SEM micrograph and (b) corresponding binary image
of a self-assembled film with APF ) 0.88.
Figure 3. Surface forces responsible for either the spreading or
lensing of a fluid on the surface of water are the fluid-air (γFV),
water-air (γWV), and fluid-water (γFW) interfacial tensions.
Table 1. Interfacial Energies Involved in a Spreading
Monolayer Film at the Air-Water Interface
fluid
fluid-vapor
interfacial tension
(mN/m)
water-fluid
interfacial tension
(mN/m)
spreading
tension
(mN/m)a
2-propanol 21.7 0 (miscible) 51.05
1-butanol 24.6 1.6 46.55
mineral oil 35 40 -2.25
a
γWV
) 72.8 mN/m.
ΠFWV
) γWV
- (γFV
+ γFW
)
Letters Langmuir, Vol. 23, No. 23, 2007 11401

4. film is a porous interconnected network of particles linked by
epoxy, which contrasts significantly with the dense array formed
when mineral oil was used. Considering that the diffusion of
2-propanol spreading liquid into the water subphase might be
responsible for this behavior, a solution composed of 2-propanol
and 1-butanol was used as a substitute, where the volumetric
ratio of 2-propanol to 1-butanol was varied from 3:2 to 19:1.
Figure 4b shows an alumina/epoxy composite film with a
2-propanol/1-butanol ratio of 4:1. The dense packing of the film
demonstrates that the film quality was significantly improved
compared with that of the composite film made using only
2-propanol as the spreading fluid (Figure 4a). Experiments using
alcohol ratios greater than 4:1 produced films with agglomerates
and voids. Alcohol ratios of less than 4:1 resulted in the
accumulation of 1-butanol on the water surface, which hindered
fluid spreading. Finally, when the water surface was saturated
with 1-butanol, no spreading was observed. Thus, from the above
results, a 4:1 volume ratio of 2-propanol/1-butanol is the optimum
solution composition for the spreading fluid.
With these results, it is clear that 1-butanol plays a critical role
when a polymerizing phase is active during self-assembly. Fluid
forming was performed in a closed trough, which confined fluid
spreading and particle assembly to a small area. Therefore, it is
important to consider the depletion rate of the spreading agent
from the spreading droplets containing the self-assembling
particles. When the dispersion containing 2-propanol, Spurr’s
epoxy,andaluminaparticleswasspreadattheair-waterinterface,
depletion of 2-propanol from the water surface resulted in a
rapid increase in the Spurr’s epoxy concentration in the spreading
layer. The concentrating solution of Spurr’s epoxy in alcohol
can increase in viscosity rapidly, which can both hinder the
rearrangement of particles and reduce the spreading velocity to
very small values. In addition, the loss of solvent also may speed
up the polymerization rate. As the epoxy concentrates, clustering
particles are frozen into a network that cannot rearrange in a
dense fashion. The optimal condition regarding the spreading
agent would be nearly complete depletion of the spreading agent
as the particles begin to form a close-packed planar array. In this
way, the viscosity of the suspension remains low so that particles
can rearrange into a dense particle array before complete removal
of the spreading fluid. Thus, finding ways to control the depletion
rate and hence the viscosity of the spreading liquid would enable
these optimum conditions to arise.
Alcohols with reduced diffusivity in water and lower vapor
pressures are good candidates for controlling the depletion rate
via dissolution and evaporation mechanisms, respectively. An
alcohol such as 1-butanol has a smaller diffusion coefficient
than 2-propanol in water, with values of 7.7 ×10-6 and 8.7
×10-6 cm2/s at 15 °C, respectively. In addition, 1-butanol has
a lower vapor pressure than 2-propanol, with respective values
of 6.1 and 1.7 kPa at 25 °C.31 Thus, it is expected that 1-butanol
would be depleted much more slowly than 2-propanol. However,
there are bounds to what depletion rate is appropriate. If the
depletion rate is too slow, then the droplet will spread (IIFWV for
1-butanolonwateris46.3mN/m)butneverdisappear,asobserved
when the ratio of 2-propanol/1-butanol was less than 4. When
the depletion rate is too fast, such as with pure 2-propanol or a
2-propanol/1-butanol ratio of greater than 4, the fluid will be
fully depleted too quickly to effect dense self-assembly of the
particles.
A final factor to consider is solubility. Alcohols such as
2-propanol are fully miscible with water, but 1-butanol is soluble
on the order of 9.1 wt % at 25 °C.32 In all of the experiments
performed in this study, the solubility of 1-butanol in water was
never exceeded, assuming a uniform distribution of 1-butanol in
water. However, within the time frame of these experiments it
is conceivable that diffusion would not affect such compositional
uniformity. In this work, because the dispersion containing
1-butanol was added directly to the water surface, slow diffusion
induced a boundary layer to form that is governed by the
solubility of 1-butanol in water. Thus, a local equilibrium was
established, which slowed 1-butanol diffusion into the water
subphase,leavingevaporationastheonlyothermeansof1-butanol
depletion. Therefore, we consider solubility, in addition to
diffusion, a factor that limited spreading experiments involving
high 1-butanol concentrations. Finally, a continuous assembly
of epoxy/particle composite films by fluid forming is possible
only under conditions of high, continuous spreading of the
particles and controlled slow diffusion of the spreading fluid
into the water subphase, which a 4:1 volume ratio of 2-propanol/
1-butanol helps to control.
(31) Lide, D. R. CRC Handbook of Chemistry and Physics, 73 ed.; CRC Press:
Boca Raton, FL, 1992.
(32) Budavari, S.; O’Neill, M.; Smith, A.; Heckelman, P.; Kinneary, J. Merck
Index, 12 ed.; Chapman & Hall: Whitehouse Station, NJ, 1996.
Figure 4. Electron micrograph of an alumina/Spurr’s epoxy
composite film using only 2-propanol as the spreading fluid. (a)
Electron micrograph of an alumina/Spurr’s epoxy monolayer
composite film formed using a 4:1 volume ratio of 2-propanol/1-
butanol. (Inset) SEM micrograph showing particles coated with
Spurr’s epoxy.
11402 Langmuir, Vol. 23, No. 23, 2007 Letters

5. Conclusions
Fluidforminghasbeenadaptedfortheself-assemblyofdensely
packedinorganicparticle(R-alumina)arraysfilledwithanorganic
matrix (Spurr’s epoxy). Particle dispersions containing 2-pro-
panol, mineral oil, and R-alumina particles were spread onto a
water substrate and self-assembled into monolayer arrays.
Replacing mineral oil with Spurr’s epoxy in the dispersion
produced particle arrays containing epoxy in interparticle voids.
Proper selection of the spreading agent components and their
respective concentration is critical in the process to ensure a
densely packed structure. Factors influencing the depletion of
the spreading agent via evaporation and diffusion into the water
bath are critical. Controlling the depletion of spreading fluid
allows for sufficient particle rearrangement, epoxy cross-linking,
and film densification.
Acknowledgment. We thank Dr. S. Miyazaki and Mr. Y.
UchidaofSumitomoChemicalCo.,Ltd.forprovidingthealumina
powder samples. Support for this work from the Office of Naval
Research and Corning Inc. is gratefully acknowledged.
LA070138W
Letters Langmuir, Vol. 23, No. 23, 2007 11403