Full text of DOI:10.1111/j.1151-2916.2000.tb01400.x

J. Am. Ceram. Soc., 83 [6] 1399–401 (2000)
journal
Mimicking Nanometer Atomic Processes on a Micrometer Scale via
Electrophoretic Deposition
,
†
,‡
,§
,‡
Partho Sarkar,* Debnath De,* Kimihiro Yamashita,* Patrick S. Nicholson,** and
,
¶
Takao Umegaki*
Advanced Materials Group, Alberta Research Council, Edmonton, Alberta, Canada T6N 1E4
Department of Materials Science and Engineering, McMaster University, Hamilton, Ontario, Canada L8S 4L7
Institute for Medical and Dental Engineering, Tokyo Medical and Dental University, Tokyo 101-0062, Japan
Department of Industrial Chemistry, Tokyo Metropolitan University, Hachioji, Tokyo 192-03, Japan
The process of submonolayer formation during the electro-
phoretic deposition (EPD) of colloidal films of micrometer-
sized (diameter Ϸ 0.5 ␮m) silica particles on a silicon wafer has
been observed as a function of deposition time. The process of
nucleation and growth of the silica monolayer is compared
with that of atomic film growth (10000 times smaller scale) via
molecular-beam epitaxy (MBE), and for the first time, a
striking similarity between the two growth processes is ob-
served. Likewise in the atomic growth process via MBE, the
entire nucleation, growth, and aggregation process during
EPD of silica particles can be broadly classified into two
regions. At low surface coverage when silica particles are
deposited outside of clusters, diffuse randomly, and stick to a
cluster on touching them, the mechanism of growth in this
region follows diffusion-limited aggregation (DLA) and the
fractal dimension of the two-dimensional clusters is found to
be close to 1.65. Later on, as the clusters grow in size,
deposition of particles inside the clusters become important
and clusters become more and more compact, resulting in a
dense, close-packed, and homogeneous monolayer. This region
is termed a consolidation region, and a change in fractal
dimension from 1.65 toward 2 with increasing surface cover-
age has been observed.
latter, controlled numbers of atoms are deposited from the gas
phase onto a well-characterized, crystalline substrate under con-
trolled growth conditions. The deposition of atoms onto the
3
substrate disturbs the system equilibrium which is restored by
4
aggregation of the atoms into clusters.
Colloidal film processing from
a stable suspension of
nanometer- to micrometer-sized particles involves manipulation of
the particles by externally imposed fields, e.g., electrophoretic
deposition (EPD) by an electrical field, slip-casting by capillary
forces, centrifugal-casting by centrifugal forces, pressure-casting
by fluid pressure differences, and sedimentation by the gravita-
tional field. Demand for minimal defect-, close-packed-, dense-,
and homogeneous-microstructure-film is developing in the ad-
vanced microelectronic, optical, magnetic, and high-performance
structural device fields. The morphology and growth of a colloidal
film is influenced by the properties of two-dimensional arrays of
particles or clusters that evolve during submonolayer formation on
5
the substrate.
This paper describes visualization of the process of submono-
layer formation during the film growth of micrometer-sized
colloidal particles by EPD. The mechanism of colloidal film
growth process is investigated and compared with atomic-scale,
thin-film growth by MBE.
I. Introduction
II. Experimental Procedure
HE order–disorder transitions of monodispersed microspheres
T
from “fluids” to “crystals” in colloidal suspensions can serve
A colloidal suspension of monodispersed silica spheres (PCI,
Inc., diameter Ϸ 0.5 ␮m) in ethanol (0.05 g of silica/100 mL of
ethanol) was stabilized by tetramethylammonium hydroxide
as a direct analogue of the structural phase transitions of atoms in
1
crystalline materials. Unlike atoms in crystalline materials,
spheres in colloidal dispersion can be tracked and imaged by
digital video microscopy. The phase transition of colloidal spheres
from regular, ordered crystalline arrays to chaotic-fluid disorder
can be controlled by manipulating their concentration and chem-
(
TMAH) (pH Ϸ9.5). The SiO sphere zeta potential was Ϫ75 mV
2
(Coulter DELSA 440 Analyzer). EPD was conducted at room
temperature (25°C) (constant current ϭ 0.25 mA). The cathode
2
(substrate) was an optically polished silicon wafer (depositing
ical environment in the colloidal suspension. The mechanism of
2
area: 126 mm ). The anode was platinum and was separated from
colloidal film growth of micrometer-sized particles is poorly
understood, and insight could result from comparison with atomic-
scale, thin-film growth via molecular-beam epitaxy (MBE). In the
the cathode by 20 mm. Silica particles were electrophoretically
deposited on the substrate for different times. The substrate was
then taken out of suspension and dried. Figures 1(a) through (d)
show the SEM images of the substrate as a function of deposition
time (or surface coverage). Capillary-force-driven fusion of do-
mains at the periphery of the deposited layer during drying is
C. Randall—contributing editor
6,7
neglected. Figures 1(a) through (d) resemble the in situ particle-
clustering images during EPD of latex particles as reported by
6
B o¨ hmer. The present authors also visualized similar particle-
clustering images (Figs. 1(a) through (d)) in situ via an optical
microscope during the EPD of silica particles. SEM images were
digitized and a prefixed basis area encompassing a cluster of
particles chosen to calculate the fractional surface coverage (area
covered by the particles/the basis area) by the particles. The fractal
dimensions of the ramified clusters (Figs. 1(a) through (d)) were
Manuscript No. 189007. Received June 4, 1999; approved November 30, 1999.
*
*
Member, American Ceramic Society.
*Fellow, American Ceramic Society.
†
‡
§
Advanced Materials Group, Alberta Research Council.
Department of Materials Science and Engineering, McMaster University.
Institute for Medical and Dental Engineering, Tokyo Medical and Dental
University.
¶
Department of Industrial Chemistry, Tokyo Metropolitan University.
1
399

1
400
Journal of the American Ceramic Society—Sarkar et al.
Vol. 83, No. 6
Fig. 1. SEM microstructures of submonolayer of colloidal silica particles (diameter Ϸ 0.5 ␮m) as a function of deposition time: (a) at low deposition time
10 s), (b) at 120 s, (c) at 960 s, and (d) at complete monolayer coverage of surface by a dense, homogeneous and close-packed film.
(
calculated by the box-counting method, following the fast algo-
particle meets another, they form a critical nucleus with less
mobility, i.e., an island or cluster. Deposition inside a cluster is
initially neglected since mostly single particles exist on the
electrode surface. An island grows when a particle meets a cluster
and attaches to its edge (due to the total interaction energy barrier
8
rithm of Liebovitch and Toth (software, FD3, written by John
9
Sarraille and Peter DiFalco ).
III. Results and Discussion
(the result of ionic and van der Waals forces) (Fig. 1(b))). A
particle arriving on top of an existing island will diffuse until it
falls off the edge of island (or it stays on the upper terrace due to
the total interaction energy barrier). Particles on the cluster surface
have a surface mobility that allows clusters to attain minimum
energy and the equilibrium configuration of close-packed struc-
tures. The surface mobility of particles is due to (1) Brownian
A plot of cluster fractal dimensions versus fractional surface
area coverage during submonolayer formation is shown in Fig. 2.
At low surface coverage (Fig. 1(a)), particles arrive on the
substrate surface, deposit, and diffuse randomly; i.e., the deposi-
tion and diffusion processes simultaneously take place. When one
7
7
motion and/or (2) electro-osmotic flow. Solomentsev et al.
reported that electro-osmosis around a charged, nonconducting
particle, near or on a flat conducting surface, creates a fluid flow
adjacent to the particle. The fluid is drawn laterally (at 90° to
electrical field) toward the particle near the electrode due to
electro-osmosis and is pushed outward from the particle further
away from the electrode (above the particle). This lateral convec-
tive fluid flow results in either particle–particle or cluster–cluster
aggregation over length scales comparable to their respective sizes.
7
Solomentsev et al. also demonstrated that this long-range electro-
osmotic flow causes surface mobility and reorientation of the
particles in a given cluster, resulting in a close-packed (equilibri-
um) structure.
Figure 1(a) shows that, at short times and low surface coverage,
the size of isolated islands is small and their number density
increases more rapidly than single particles. When the number
density of islands and single particles becomes comparable, the
rate of increase of the number density of islands slows (Fig. 1(b)).
When the size of the islands is Х the distance between them, a
rapid decrease of number density of single particles is observed as
Fig. 2. Fractal dimension of clusters in Figs. 1(a) through (d) as a
function of surface coverage during submonolayer formation of silica
particles.

June 2000
Mimicking Nanometer Atomic Processes on a Micrometer Scale
1401
well as an unchanged number density of islands (Fig. 1(c)). No
further islands form in this regime.
virtually constant with time (since EPD is conducted at constant
current) and the change of particle concentration in the large
volume of suspension is negligible. DЈ for particles is also assumed
to be approximately constant with time at room temperature if the
particle flux is constant. Thus, Figs. 1(a) through (d) show the
variation of island size and number, as a function of surface
coverage, for DЈ/F Х constant.
Thus, the mechanism of nucleation, growth, and aggregation
during the submonolayer formation of micrometer-size particles
mimics the classical DDA model of the atomic, thin-film growth
process as per MBE. The results provide insight into the growth
kinetics and the microstructural manipulation of films of
micrometer-size particles as practiced in the semiconductor, ce-
ramic, polymer, and other coating industries.
As time progresses, the islands grow and become consolidated
by deposition of particles inside them. As a result, the individual
islands grow into close-packed structures with minimum configu-
rational energy and distinct domain boundaries. When the consol-
idated islands touch each other, their growth stops. Finally, the
close-packed islands, forming a monolayer where the distinct
domain boundaries of the islands are preserved, cover the entire
surface (Fig. 1(d)). Cluster formation on top of this monolayer now
becomes relevant.
The mechanism of nucleation, growth, and aggregation during
submonolayer formation of micrometer-sized particles can be
classified into two regions as per the classical DDA (deposition,
diffusion, and aggregation) model of atomic thin-film growth
during MBE. At low surface coverage (region A in Fig. 2 where
deposition and diffusion of particles take place simultaneously),
the mechanism of growth is similar to the diffusion-limited
aggregation (DLA) of atomic, thin-film growth where atoms
deposit outside clusters, diffuse randomly, and stick to a cluster on
touching. Figure 2 shows that the fractal dimension of the
two-dimensional ramified clusters in this region is Х1.65. In the
classical DLA mechanism, only single atoms diffuse and this is the
only source of aggregation. The fractal dimension of two-
IV. Summary
Submonolayer formation as a function of deposition time has
been observed during the colloidal film growth of silica particles
(diameter Ϸ 0.5 ␮m) on a silicon wafer substrate by electro-
phoretic deposition. A close-packed, homogeneous, dense mono-
layer of silica particles evolves at the end of submonolayer
formation. Analysis of the fractal dimensions of the particle
clusters versus fractional surface-area coverage by particles re-
veals that the film growth of silica particles mimics the classical
DDA model of the atom-by-atom, thin-film growth process via
MBE.
1
0
dimensional clusters in the classical DLA mechanism is Х1.7.
1
1
Meakin reported that, when the clusters also diffuse, the fractal
dimension (x) is 1.45 Ͻ x Ͻ 1.5. In the present experiments, single
particles and clusters diffuse, and long-range, electro-osmotic
8
flow plus diffusion results in particle aggregation. The fractal
dimension (x Х 1.65) in this DLA region is 1.45 Ͻ x Ͻ 1.7.
Subsequently, as clusters grow, deposition of particles inside them
becomes important and the latter become more consolidated and
form a close-packed structure (region B in Fig. 2). This region
constitutes the consolidation process of clusters (like the consoli-
dation of atom clusters in the atomic thin-film growth process) and
Fig. 2 shows a change of fractal dimension (1.65 toward 2) with
increasing surface coverage.
References
1
C. A. Murray and D. G. Grier, “Video Microscopy of Monodisperse Colloidal
Systems,” Annu. Rev. Phys. Chem., 47, 421–62 (1996).
A. E. Larsen and D. G. Grier, “Like-Charge Attraction in Metastable Colloidal
Crystallites,” Nature (London), 385 [16] 230–33 (1997).
Z. Zhang and M. G. Lagally, “Atomistic Processes in the Early Stages of
Thin-film Growth,” Science, 276 [18] 377–83 (1997).
H. Roder, E. Hahn, H. Brune, J. Bucher, and K. Kern, “Building One- and
2
3
4
Following the classical DDA model of atomic thin-film growth,
the time scale of the two competing processes, i.e., deposition and
diffusion (or nucleation and growth) of particles, depends on the
flux of deposition (F) and the surface diffusion coefficient (D) for
particles. As F increases, the deposited particles must diffuse over
a shorter distance before meeting others; i.e., the time scale for
diffusion of a particle is proportional to 1/D. In the formation of a
submonolayer of micrometer-size particles, diffusion is enhanced
by long-range electro-osmotic flow and results in particle aggre-
gation. At a fixed surface coverage, an identical surface morphol-
ogy is expected as long as the ratio DЈ/F (DЈ is the enhanced
diffusion coefficient) remains constant during the film growth
process, since deposition and enhanced diffusion processes are in
competition in the early stages of submonolayer formation by
micrometer-size particles. The flux of the particles (F) remains
Two-Dimensional Nanostructures by Diffusion-Controlled Aggregation at Surfaces,”
Nature (London), 366, 141 (1993).
5
H. H. Wickman and J. N. Korley, “Colloidal Crystal Self-Organization and
Dynamics at the Air/Water Interface,” Nature (London), 393 [4] 445–47 (1998).
6
M. B o¨ hmer, “In Situ Observation of Two-Dimensional Clustering during Elec-
trophoretic Deposition,” Langmuir, 12 [24[rsqb] 5747–50 (1996).
7
Y. Solomentsev, M. Bohmer, and J. L. Anderson, “Particle Clustering and Pattern
Formation during Electrophoretic Deposition: A Hydrodynamic Model,” Langmuir,
1
3, 6058–68 (1997).
8
L. S. Liebovitch and T. Toth, “A Fast Algorithm to Determine Fractal Dimension
by Box-Counting,” Phys. Lett. A, 141, 386–90 (1989).
9
J. J. Sarraille and P. DiFalco, A Program for Calculating Fractal Dimensions, FD3,
version 0.3 (ftp://ishi.csustan.edu/pub/; email: john@ishi.csustan.edu), Department of
Computer Science, California State University, Stanislaus, Turlock, CA 95382.
1
0
A. L. Barabasi and H. E. Stanley, Fractal Concepts in Surface Growth; Ch. 17.
Cambridge University Press, New York, 1995.
11
P. Meakin, “Formation of Fractal Clusters and Networks by Irreversible Diffu-
sion Limited Aggregation,” Phys. Rev. Lett., 51, 1119 (1983).
Ⅺ