Morphological control of zirconia nanoparticles through combustion aerosol synthesis

Article abstract of DOI:10.1111/j.1151-2916.2002.tb00233.x

Ceramic oxide nanoparticles produced by flame-based processes are typically agglomerated, which can limit their use in some applications. In this paper, a novel combustion synthesis method that utilizes the spraying of combustible droplets into a premixed flame to produce nanoscale crystalline particles of agglomerated and unagglomerated morphologies is described. Although the same flame-based experimental setup is used in both cases, variation in peak flame temperatures results in a corresponding variation between fractallike agglomerates and single isolated spherical particles. TEM/ED analysis shows that both classes of particles are the tetragonal crystal phase of zirconia. In the case of the unagglomerated spherical particles, results indicate that each precursor solution droplet, which acts as the feed, produces multiple spherical ceramic nanoparticles with a number mean diameter of 90 nm. The use of an inertial impaction stage in the precursor feed line to eliminate large feed droplets leads to a decrease in the number mean diameter to 60 nm, suggesting that crystalline spherical nanoparticles can be produced in a continuous flame-based process through control of the feed droplet size.

Full text of DOI:10.1111/j.1151-2916.2002.tb00233.x

J. Am. Ceram. Soc., 85 [5] 1127–32 (2002)
journal
Morphological Control of Zirconia Nanoparticles through
Combustion Aerosol Synthesis
Amit U. Limaye and Joseph J. Helble
Department of Chemical Engineering, University of Connecticut, Storrs, Connecticut 06269–3222
gaseous20,22–24 precursors for ceramic oxide particle production
using flames or high-temperature furnaces7,9 has been reported.
Although the chlorides of silicon and titanium are relatively
volatile and have been used as gaseous precursors for oxide
particle formation, the low vapor pressure of zirconium chloride
limits its use as a precursor. Consequently, liquid precursors, such
as aqueous solutions or organometallic compounds of zirconium,
have been used in flame12,21 and spray-drying systems7,25 for
zirconia nanoparticle synthesis. Unagglomerated particles can be
formed using liquid precursors with the final particle size governed
by the precursor droplet size when a spray-drying furnace is used
for processing.7,25 Experimental trends observed in spray pyrolysis
systems have been explained reasonably well by a model reported
in the literature.25 The model compares the rates of solute diffusion
within the droplet and solvent vaporization from the droplet
surface and explains their effect on the particle morphology. Large
particles (micrometer scale) are formed, however, if dilute precur-
sor concentrations are not used in these systems. Although typical
flame-based processes produce agglomerated particles, unagglom-
erated particles also can be formed in a flame-based system using
liquid precursors. The final particle sizes, however, are not directly
dependent on the droplet sizes12,21 in these systems, unlike in
spray pyrolysis. The mechanism of particle formation when liquid
precursors are used in flame-based systems is not yet fully
understood. In processes using fine particles of metals as the
precursor, it is suggested that the oxide particle growth occurs
predominantly by vapor condensation,16 whereas particle growth
occurs by collision and coalescence when gaseous precursors are
used.19
Ceramic oxide nanoparticles produced by flame-based pro-
cesses are typically agglomerated, which can limit their use in
some applications. In this paper, a novel combustion synthesis
method that utilizes the spraying of combustible droplets into
a premixed flame to produce nanoscale crystalline particles of
agglomerated and unagglomerated morphologies is described.
Although the same flame-based experimental setup is used in
both cases, variation in peak flame temperatures results in a
corresponding variation between fractallike agglomerates and
single isolated spherical particles. TEM/ED analysis shows
that both classes of particles are the tetragonal crystal phase of
zirconia. In the case of the unagglomerated spherical particles,
results indicate that each precursor solution droplet, which
acts as the feed, produces multiple spherical ceramic nanopar-
ticles with a number mean diameter of 90 nm. The use of an
inertial impaction stage in the precursor feed line to eliminate
large feed droplets leads to a decrease in the number mean
diameter to 60 nm, suggesting that crystalline spherical nano-
particles can be produced in a continuous flame-based process
through control of the feed droplet size.
I. Introduction
ARTICLES having features with dimensions ranging from a few
P
to a few hundred nanometers are often referred to as nanopar-
ticles. Nanoparticles of zirconia, which have a broad range of
applications, including high-strength ceramics, thermal barrier
coatings, paints, textiles, refractories, transformation-toughened
ceramics, oxygen sensors, and fuel cells,1–4 have been produced
by a variety of techniques. Vapor deposition, mechanical milling,
laser ablation, flame-based methods, spray pyrolysis, sol–gel, and
microwave plasma synthesis have been used to produce ceramic
particles ranging from a few nanometers to a few micrometers in
size and of varied degrees of crystallinity.5–9 Flame-based methods
are of potential commercial importance, because they are relatively
low cost and can be operated as continuous single-step process-
es.10–24 Although very small particles (5–10 nm) can be produced
by these methods, hard agglomerates are generally formed, result-
ing in “fractallike” branched structures that are undesirable in
many applications. Attempts have been made to control agglom-
eration using electric fields,22 but only moderate success has been
achieved, with partial agglomeration always present. In addition,
special burner and gas-flow configurations are often required in
these processes,12 because fuel–oxidizer mixing and flame struc-
ture affect the final particle morphology and size.14,23
In this article, we present the results of a novel droplet
combustion process for the synthesis of nanoparticles with con-
trollable morphology and particle size. Using liquid precursors
sprayed into a premixed methane–oxygen flame, we have elimi-
nated the dependence of particle size on fuel–oxidizer mixing and
flame structure. Details of the process are provided below.
II. Experimental Procedure
An experimental system was designed to provide flames of
varying and controllable temperature profile, resulting in variable
residence time of the combustion gases in the reactor. A schematic
diagram of this experimental setup is shown in Fig. 1. A flat-flame,
water-cooled, porous-plug stainless-steel burner (McKenna Prod-
ucts, Inc., Sebastopol, CA) was used as the heat source and the
main component of the system. Methane, oxygen, and nitrogen
were premixed within the burner, and the resulting two-
dimensional flame was anchored on the porous plug. A shroud of
nitrogen around the porous plug eliminated fluctuations in the
flame caused by external disturbances. Ceramic precursors were
introduced into the flame by spraying a combustible precursor
solution through a feed tube located centrally in the burner.
Zirconium(IV) n-butoxide (Strem Chemicals, Inc., Newburyport,
MA) dissolved in 99.8% anhydrous butanol (Sigma Chemicals,
Inc., St. Louis, MO) was used as the precursor feed stock. To
prepare the feed solution, the zirconium(IV) n-butoxide was added
to butanol with moderate stirring under nitrogen in a glove box to
avoid reaction with water present in the air. The spray was
In high-temperature synthesis methods, such as flame-based
methods, the type of precursor also is important in determining the
final particle characteristics. The use of solid,13,16 liquid,12,15,21 or
K. Bowman—contributing editor
Manuscript No. 187375. Received October 24, 2001; approved February 12, 2002.
Supported by the National Science Foundation under CAREER Grant No.
CTS-9733234.
1127

1128
Journal of the American Ceramic Society—Limaye and Helble
Vol. 85, No. 5
Fig. 1. Schematic diagram of experimental system used for the production of zirconia nanoparticles.
produced using a pressure-driven atomizer (Model 9302, TSI, Inc.,
St. Paul, MN) operating in the pressure range of 5–50 psi
(34.5–345 kPa). The droplet size distribution output of the atom-
izer was characterized at various pressures using a phase doppler
particle analyzer (PDPA; Aerometrics, Inc., Sunnyvale, CA) and
was relatively independent of operating pressure over the entire
range. The size distribution of the droplets produced by the
atomizer had a number mean diameter ranging from 6 to 8 ␮m as
the pressure increased from 5 to 50 psi (34.5–345 kPa). At an
operating pressure of 10 psi (69 kPa) (used in most experiments in
this study) a relatively small fraction of the droplets had a size Ͼ20
␮m (ϳ0.7%), thus indicating the absence of very large droplets in
the spray.
A 1.0 m quartz tower placed downstream of the burner served
to confine the products of combustion and to prevent entrainment
of room air. The tower had sampling ports located axially every 15
cm along its height, providing access for thermophoretic sampling
of particles at locations downstream of the flame. The entire tower
was insulated to maintain high temperatures in the postcombustion
region and avoid thermophoretic deposition of the particles on the
tower walls. Fiberglass wrap was used to insulate the upper 50 cm
of the tower. A steel shroud wrapped around the lower 50 cm of
the tower insulated it against radiative losses while a castable
insulating cement (McMaster, Inc.) insulated the outer surface of
the shroud to minimize the heat loss from the shroud to the
ambient air.
TEM (Model 420, Phillips, Amsterdam, The Netherlands)
equipped for energy dispersive spectroscopy (EDS; EDAX, Inc.,
Stratford, CT).
III. Results and Discussion
To begin an experiment, the burner gases were ignited and the
reactor was then allowed to attain a steady-state temperature
profile by the heating of the reactor walls for 20–25 min. The
precursor solution was then sprayed through the central feed tube
using the atomizer. The premixed flame was run at an equivalence
ratio of one (␸ ϭ 1.0) based on the methane and oxygen content.
The relative contents of methane, oxygen, and nitrogen in the
premixed gases were 1/2/13.2 on a molar basis, respectively,
resulting in a calculated adiabatic flame temperature of 1710 K,
assuming complete species chemical equilibrium. The linear ve-
locity of the hot gases was calculated along the tower height and
used to obtain the residence times of the gases in the tower. The
temperature profile inside the reactor as measured using a K-type
thermocouple and corrected for radiation effects is shown as the
curve labeled “single” in Fig. 2. Figure 2 shows there was a rapid
Particles were sampled by thermophoretic deposition on trans-
mission electron microscopy (TEM) grids attached to the end of
the thermophoretic sampling probe. A solenoid-controlled air-
driven piston was used to introduce the probe through one of the
tower axial sampling ports for 1000 repeated short-duration (700
ms) pulses to collect sufficient samples for analysis. A specially
constructed, movable, water-cooled, nitrogen-quenched collection
probe was also used to collect particles at the tower exit. The
dilution was used to avoid the condensation of water vapor formed
as a combustion product in the lines downstream of the reactor.
Particles were then separated from the gas stream by deposition on
a TEM grid located within a point-to-plane electrostatic precipi-
tator (ESP; Intox Products, Inc., Albuquerque, NM). In some
experiments, a scanning mobility particle sizer (SMPS; TSI, Inc.,
St. Paul, MN) was used to provide on-line measurement of
particle-size distribution in the range of 18–900 nm. Particle
morphology and crystallinity were subsequently analyzed using
Fig. 2. Temperature in the reactor versus gas residence time showing the
two operating conditions used for the production of agglomerated (fractal)
and unagglomerated (single) particles.

May 2002
Morphological Control of Zirconia Nanoparticles through Combustion Aerosol Synthesis
1129
Fig. 5. Comparison of the measured PSD with that predicted from an
assumption of a single particle produced per droplet.
the reactor (overall residence time of 1.4 s) using the collection
probe and the ESP. The particles collected at these three locations
were used to obtain information on the evolution of particle size
and morphology as a function of residence time in the reactor.
TEM was used to analyze particle morphology (size and shape)
and crystallinity. A micrograph of particles collected at 400 ms
residence time at a magnification of 88 600 is shown in Fig. 3.
Isolated spherical particles were produced with no agglomeration
observed. Selected area electron diffraction (SAD) study of these
particles revealed they had the tetragonal crystal structure. A
representative SAD pattern with the Miller indexes of the diffract-
ing planes indicated on it is show in Fig. 4. Although the size
distribution of the particles collected at the residence times of 400
and 850 ms was obtained using image analysis of TEM micro-
graphs, the SMPS was used to obtain the particle-size distribution
(PSD) of particles collected at the exit. This PSD, labeled “base
case” in Fig. 5, indicated that the particle sizes were not uniform
and varied over a wide range. The number mean diameter was 90
nm, but particles as large as 500 nm were also observed. The
number mean diameter and the standard deviation for the PSD
obtained for particles collected at all three locations are presented
in Table I. The mean value and the standard deviation remained
constant with increased residence time, indicating that there was
no change in the particle size beyond a residence time of 400 ms.
Although particle growth can occur by condensation of zirconia
vapors or by coalescence of particles, the results indicated that the
particle formation process was complete within 400 ms and
subsequent particle growth did not occur. Particles collected at all
three locations, as observed using TEM, had the same spherical
shape with no agglomeration.
Fig. 3. Single, spherical unagglomerated particles obtained using a
low-temperature flame (magnification is 88 600 times).
temperature decrease immediately after the flame. This was
beneficial for inhibition of agglomeration resulting from interpar-
ticle collision in the postcombustion region, because colliding
particles are less likely to coalesce at a low temperature.
For production of nanoparticles, a 0.047M solution of the
precursor zirconium(IV) n-butoxide/butanol solution was sprayed
into the core region of the flame using the atomizer operated at 10
psi (69 kPa). The calculated characteristic time required for the
vaporization of a 10 ␮m droplet of butanol was 55 ms, whereas the
calculated residence time of the gases in the flame zone was 20 ms.
Consequently, particle formation was expected to occur within the
flame and in the postcombustion region immediately downstream
of the flame.
Nanoparticles were collected at intermediate locations (resi-
dence times of 400 and 850 ms) in the tower using the thermo-
phoretic sampling probe. Particles were also collected at the exit of
The present process of nanoparticle formation is similar to spray
pyrolysis in that both processes use liquid precursors sprayed into
a high-temperature region to form nanoparticles. In spray pyroly-
sis, however, each precursor droplet produces one particle, thereby
making it possible to predict the final PSD as long as the precursor
concentration in the solution and the size of the solution droplets
are known. Because the concentration and size are known for our
system, the PSD predicted by spray pyrolysis has been calculated
and is shown as the curve labeled “predicted” in Fig. 5. Comparing
Table I. Comparison of Particle-Size Data Collected at
Various Residence Times in the Reactor
Residence time
(ms)
Number mean
(nm)
Standard deviation
(nm)
Particles Ͼ160 nm
(%)
400
850
1400
98
96
90
78
74
80
14
13
11
Fig. 4. Selected-area electron diffraction pattern for spherical particles
shown in Fig. 3. Rings are labeled by the Miller indexes of diffracting
planes: (a) (200), (b) (211), (c) (114), and (d) (312).

1130
Journal of the American Ceramic Society—Limaye and Helble
Vol. 85, No. 5
Fig. 6. PSD of zirconia nanoparticles obtained from experiments in which impactors were used to eliminate the large droplets in the feed stream. Each curve
represents the average of four (base case) or five (impactor) experiments. Error bars represent standard deviation of the mean value of the normalized number
distribution normalized by the total number count (NT) at the indicated point.
the two curves in Fig. 5, it is apparent that spray pyrolysis would
produce much larger particles (mean 727 nm) for the precursor
concentration and droplet sizes used in this study. This clearly
indicates that more than one ceramic nanoparticle results from
each precursor droplet in combustion aerosol synthesis as devel-
oped in this study. This conclusion is of significance because it
indicates that large droplets can be used to produce small particles
without operating at the very low precursor concentrations re-
quired by spray pyrolysis. Such small particles are likely produced
as a result of the competition between solvent vaporization and
solute diffusion within the precursor droplets. If the evaporation
rate is large compared with the solute diffusion rate, the local
concentration of solute increases near the receding droplet surface,
subsequently producing cenospheric particles rather than solid
ones.25 If the evaporation rate is much larger than the diffusion
rate, however, regions of locally high solute concentration are
expected to form. Nucleation occurs in these regions, producing
isolated particles. This mechanism is possible in our system
because of the much higher heating rates (104–105 K/s) compared
with a typical spray pyrolysis process (103–104 K/s). Each droplet
thus acts as a microreactor, producing many small particles.
Further examination of the curve labeled “base case” in Fig. 5
indicates that, although the number of small nanoparticles is high,
large nanoparticles (200 nm or larger) are also present in the final
PSD. This indicates that most of the mass of the sample is
contributed by the larger particles. In general, size as well as
morphology is of importance, and, in many applications, it is
desired that the nanoparticles have a uniform size or narrow PSD.
The presence of large particles in the base-case PSD suggests that
they may be forming from the large droplets in the precursor spray.
Therefore, an inertial impactor has been used to eliminate the large
droplets from the precursor spray.26 In the inertial impactor, the
smaller droplets exit the impactor unaltered while the larger
droplets in the spray are retained in the chamber. Droplets larger
than a desired size are prevented from entering the flame. For this
specific application, two single-stage impactors with different
cut-point sizes (d50 ϭ 4 and 2.3 ␮m) have been designed,
constructed, and utilized to eliminate the large droplets from the
spray. The final PSD of the zirconia nanoparticles obtained when
the spray system is modified by the inclusion of the impactors is
shown in Fig. 6. The curves labeled “impactor 4” and “impactor
2.3” indicate the final zirconia PSD obtained when the 4.0 and 2.3
␮m impactors are used, respectively. The area under the three
curves in Fig. 6 is not the same, because the samples are collected
for a fixed time interval rather than for a fixed mass of particles.
Choosing a particle size of 160 nm to describe the spread in the
size distribution of the zirconia nanoparticles shows that the
fraction of particles Ͼ160 nm decreases as impactors with subse-
quently smaller cut-point sizes are used. For example, 11% of the
total number of particles is Ͼ160 nm when the feed is not
manipulated using an inertial impactor (curve labeled “base case”),
whereas the fraction of particles Ͼ160 nm is decreased to 2%
when the 4 ␮m impactor is used. The use of the 2.3 ␮m impactor
further decreases this fraction to 0.2%, implying that only two of
every thousand particles are Ͼ160 nm. This, therefore, enables the
transformation of the majority of the precursor mass into zirconia
nanoparticles of small and relatively uniform size. The decreased
polydispersity of the PSD is also indicated by the decrease in the
standard deviation of the PSD from 80 nm for the base case to 27
nm for impactor 2.3. The decrease in the spread of the PSD is also
accompanied by a decrease in the number mean diameter from 90
to 60 nm. A summary of the improvement in the nanoparticle size
distribution on using the impactors is provided in Table II. This
manipulation of the atomized precursor droplets has provided an
Table II. Effect of the Sprayed Droplet Size—Comparison
of the Final Particle-Size Distribution
Number
mean
(nm)
Standard
deviation
(nm)
Particles
Ͼ160
nm (%)
Feed conditions
Base case
4 ␮m cutoff impactor
2.3 ␮m cutoff impactor
90
64
63
80
40
27
11
2.0
0.2

May 2002
Morphological Control of Zirconia Nanoparticles through Combustion Aerosol Synthesis
1131
Fig. 8. Mechanisms of liquid- and vapor-phase pathways as applied to
droplet combustion method of nanoparticle synthesis.
in a crystalline final product. Although the tetragonal phase of
zirconia is not thermodynamically favored at room temperature, it
is often found in materials with nanoscale structure, such as those
produced in this work, because of the crystallite size effect.27 This
observation also has been made in this work. The crystallinity of
the zirconia nanoparticles is determined using electron diffraction
and is of the tetragonal phase structure irrespective of the particle
morphology. The tetragonal phase is responsible for the unique
quality of transformation toughening and is the desired phase of
zirconia in applications requiring high fracture strength.4
Fig. 7. Fractallike agglomerated particles obtained with
temperature flame (magnification of 88 600 times).
a high-
effective method of controlling the PSD in the combustion aerosol
synthesis of crystalline ceramic nanoparticles.
Although the discussion thus far has been focused on crystalline
spherical particles, agglomerated fractallike particles also can be
produced using the same precursor and experimental setup. In
general, production of agglomerated particles is facilitated by the
vaporization of the precursors, as evidenced by numerous studies
in the literature where gaseous or volatile precursors have been
used.14,20,22–24 It was speculated that an increase in the flame
temperature would thus result in vaporization of the precursor,
followed by subsequent reaction in the gas vapor phase to form
zirconia nanoparticles. A higher-temperature premixed flame (␸ ϭ
1.0) was obtained by decreasing the nitrogen content such that the
methane, oxygen, and nitrogen content of the premixed gases was
1/2/8.2 on a molar basis. Under these conditions, the calculated
adiabatic flame temperature considering all species at chemical
equilibrium was 2150 K. The radiation-corrected temperature
profile is shown as the curve labeled “fractal” in Fig. 2. The
particles produced under these conditions were composed of small
individual primary particles that formed fractallike branched struc-
tures, as shown in Fig. 7. Although all the primary particles had
comparable sizes of ϳ10 nm, the overall size of the agglomerated
particles ranged from 100 nm to a few micrometers. Such open
structures are characteristic of flame-based systems.20,22–24
The suggested mechanisms of competing vapor-phase and
liquid-phase pathways, as applied to the work reported here, are
shown schematically in Fig. 8. In both the cases, the first two steps
are the same, regardless of whether particles of agglomerated or
unagglomerated morphology are formed. In the liquid-phase path-
way, the reaction of the precursors occurs in the liquid phase
within each precursor droplet. Therefore, zirconium(IV)
n-butoxide undergoes hydrolysis and/or oxidation within the con-
fines of the sprayed liquid droplet. In contrast, in the vapor-phase
pathway, the precursors vaporize/decompose in the high-
temperature region and react in the vapor phase. The oxidized
metal vapors then condense and form small particles that grow by
coagulation. Consequently, hard fractallike agglomerates are more
likely to form in materials with high vapor pressures in flame-
based systems, because they vaporize more easily and react in the
vapor phase to form oxide vapors.
IV. Summary and Conclusions
Combustion aerosol synthesis of zirconia nanoparticles with
controllable morphology has been demonstrated in a single-step
process for the first time. By varying the flame temperature and the
temperature profile in the reactor, particles of either agglomerated
or unagglomerated morphologies were synthesized. Although the
experimental setup and the precursor loading were the same in
both the cases, agglomerated particles resulted when higher flame
temperatures were used. The high temperatures caused the zirco-
nium(IV) n-butoxide precursor to vaporize/decompose and react in
the gas phase to form zirconia particles, which grew by coagula-
tion to form fractallike particles. In contrast, when the process of
droplet to particle transformation occurred entirely in the liquid
phase and at relatively low temperatures, single unagglomerated
particles were formed. TEM analysis revealed that all the nano-
particles had the tetragonal-phase crystalline structure, regardless
of the final particle morphology.
The final PSD, as obtained on spraying the precursors into the
flame, was also polydisperse. Although all the particles were
spherical and unagglomerated, the nonuniformity in their sizes was
undesirable. Removal of large droplets through an inertial impac-
tor caused a decrease in the number of large nanoparticles. When
compared with the case where no impactor was used, the use of
impactors increased the fraction of particles Ͻ160 nm from 89% to
99.8%. Good control on the unagglomerated PSD was thus
obtained through control of the droplet size distribution.
The present process of droplet to particle conversion is com-
pared with the spray pyrolysis method of particle synthesis. In
spray pyrolysis, each particle forms from a single precursor
droplet; however, the results from this work indicate that each
precursor droplet results in multiple ceramic particles of varying
sizes. This is attributed to the higher heating rates in a flame
relative to a typical spray pyrolysis process. The formation of
multiple particles per droplet is important, because it indicates that
it is feasible to obtain small particles from large droplets. Spray
pyrolysis suggests that small particles can be formed from large
droplets only by operating at low precursor concentrations, and,
thus, droplet size and the precursor concentration become control-
ling factors in the process. Although the restriction on the droplet
In general, most nanoparticle synthesis processes require a
postproduction step for crystallization of the final products and
elimination of undesirable byproducts, such as chlorine and
water.8,9 Flame synthesis of nanoparticles eliminates this extra
step, because calcination takes place within the process, resulting

1132
Journal of the American Ceramic Society—Limaye and Helble
Vol. 85, No. 5
12A. Kilian and T. F. Morse, “A Novel Aerosol Combustion Process for the High
Rate Formation of Nanoscale Oxide Particles,” Aerosol Sci. Technol., 34, 227–35
(2001).
size is relaxed in the process reported here, a modest decrease in
precursor concentration can further decrease the size of the final
particles.
13S. H. Ehrman and S. K. Friedlander, “Bimodal Distributions of Two Component
Metal Oxide Aerosols,” Aerosol Sci. Technol., 30, 259–72 (1999).
14S. E. Pratsinis, “Flame Aerosol Synthesis of Ceramic Powders,” Prog. Energy
Combust. Sci., 24, 197–219 (1998).
Acknowledgment
15J. Karthikeyan, C. C. Berndt, J. Tikkanen, J. Y. Wang, A. H. King, and H.
Herman, “Nanomaterial Powders and Deposits Prepared by Flame Spray Processing
of Liquid Precursors,” Nanostuct. Mater., 8 [1] 61–74 (1997).
16I. S. Altman, “On Condensation Growth of Oxide Particles During Gas-Phase
Combustion of Metals,” Combust. Sci. Technol., 160, 221–29 (2000).
17K. R. Venkatachari, D. Huang, S. P. Ostrander, W. A. Schulze, and G. C.
Strangle, “Preparation of Nanocrystalline Yttria-Stabilized Zirconia,” J. Mater. Res.,
10 [3] 756–61 (1995).
The authors would like to thank Dr. Donald Potter, Department of Metallurgy and
Materials Engineering, University of Connecticut, Storrs, for his assistance and
stimulating discussion regarding electron microscopy.
References
18G. D. Ulrich, “Flame Synthesis of Fine Particles” (special report), Chem. Eng.
News, [Aug. 6] 22–29 (1984).
1H. Gleiter, “Materials with Ultrafine Microstructures: Retrospectives and Perspec-
tives,” Nanostuct. Mater., 1, 1–19 (1992).
19B. W. Lee, J. I. Jeong, J. Y. Hwang, M. Choi, and S. H. Chung, “Analysis of
Growth of Non-Spherical Silica Particles in a Counterflow Diffusion Flame Consid-
ering Chemical Reactions, Coagulation, and Coalescence,” J. Aerosol Sci., 32,
165–85 (2001).
2M. Gell, “Application Opportunities for Nanostructured Materials and Coatings,”
Mater. Sci. Eng. A, A204, 246–51 (1995).
3M. J. Mayo, J. R. Seidensticker, D. C. Hague, and A. H. Carim, “Surface
Chemistry Effects on the Processing and the Superplastic Properties of Nanocrystal-
line Oxide Ceramics,” Nanostuct. Mater., 11 [2] 271–82 (1999).
4R. Stevens, “An Introduction to Zirconia,” Pub. No. 113, Magnesium Elektron,
Flemington, NJ, July 1986.
20O I. Arabi-Katbi, S. E. Pratsinis, P. W. Morrison Jr., and C. M. Megaridis,
“Monitoring the Flame Synthesis of TiO2 Particles by in-Situ FTIR Spectroscopy and
Thermophoretic Sampling,” Combust. Flame, 124, 560–72 (2001).
21G. Saracco, F. Geobaldo, D. Mazza, and G. Baldi, “New Method of Catalyst
Powder Manufacturing Based on Solvent Combustion,” J. Therm. Anal. Calorim., 56,
1435–42 (1999).
5T. D. Shen and C. C. Koch, “Formation and Hardening Effects in Nanocrystalline
Ti–N Alloys Prepared by Mechanical Alloying,” Nanostuct. Mater., 5 [6] 615–29
(1995).
22S. Vemury and S. E. Pratsinis, “Charging and Coagulation During Flame
Synthesis of Silica,” J. Aerosol Sci., 27, 951–66 (1996).
6R. J. Munz, T. Addona, and A.-C. da Cruz, “Application of Transferred Arcs to the
Production of Nanoparticles,” Pure Appl. Chem., 71 [10] 1889–97 (1999).
7S. C. Zhang, G. Mulholland, and G. L. Messing, “Synthesis of ZrO2 Nanoparticles
by Spray Pyrolysis,” J. Mater. Synth. Proces., 4 [4] 227–33 (1996).
8L. L. Hench and J. K West, “The Sol–Gel Process,” Chem. Rev., 90 [1] 33–72
(1990).
23S. E. Pratsinis, Z. Wenhua, and S. Vemury, “Role of Gas Mixing in the Flame
Synthesis of Titania Powders,” Powder Technol., 86 [1] 87–93 (1996).
24H. D. Jang, “Generation of Silica Nanoparticles from Tetraethylorthosilicate
(TEOS) Vapor in a Diffusion Flame,” Aerosol Sci. Technol., 30, 477–88 (1999).
25S. C. Zhang and G. L. Messing, “Synthesis of Solid, Spherical Zirconia Particles
by Spray Pyrolysis,” J. Am. Ceram. Soc., 73 [1] 61–67 (1990).
26W. C. Hinds, “Inertial Impaction”; pp. 121–28 in Aerosol Technology: Proper-
ties, Behavior, and Measurement of Airborne Particles, 2nd Ed. Wiley, New York,
1999.
9B. Xia, L. Duan, and Y. Xie, “ZrO2 Nanopowders Prepared by Low-Temperature
Vapor-Phase Hydrolysis,” J. Am. Ceram. Soc., 83 [5] 1077–80 (2000).
10D. E. Rosner, “Combustion Synthesis and Materials Processing,” Chem. Eng.
Educ., [Fall] 228–35 (1997).
11Y. Chen, N. Glumac, B. H. Kear, and G. Skandan, “High Rate Synthesis of
Nanophase Materials,” Nanostuct. Mater., 9, 101–104 (1997).
27R. C. Garvie, “The Occurrence of Metastable Tetragonal Zirconia as a Crystallite
Size Effect,” J. Phys. Chem., 69 [4] 1238–43 (1965).
Ⅺ