Journal of Colloid and Interface Science 341 (2010) 23–29

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

Preparation of spherical silica particles by Stöber process with high concentration

of tetra-ethyl-orthosilicate
Xiao-Dong Wang a, Zheng-Xiang Shen a, Tian Sang a, Xin-Bin Cheng a, Ming-Fang Li b, Ling-Yan Chen a,
Zhan-Shan Wang a,*
a
Institute of Precision Optical Engineering, Department of Physics, Tongji University, Shanghai 200092, China
b
Department of Chemistry, Tongji University, Shanghai 200092, China

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, Stöber process with high concentration of tetra-ethyl-orthosilicate (TEOS) up to 1.24 M is
Received 25 May 2009 used to prepare monodisperse and uniform-size silica particles. The reactions are carried out at
Accepted 14 September 2009 [TEOS] = 0.22–1.24 M, low concentrations of ammonia ([NH3] = 0.81[TEOS]), and [H2O] = 6.25[TEOS] in
Available online 18 September 2009
isopropanol. The solids content in the resulting suspension achieves a maximum value of 7.45% at
1.24 M TEOS. Various-sized particles in the range of 30–1000 nm are synthesized. The influences of TEOS,
Keywords: NH3, and H2O on the size and size distribution of the particles are discussed. A modified monomer addi-
Silica
tion model combined with aggregation model is proposed to analyze the formation mechanism of silica
Solid content
Hydrolysis
particles.
Condensation Ó 2009 Elsevier Inc. All rights reserved.

1. Introduction reaction, the elemental silicon must be activated by washing with

aqueous hydrofluoric acid, pure water, alcohol and ether consecu-
In recent years, there has been increasing interest and demand tively to remove silicon dioxide film from the particle surface and
for silica nanoparticles not only in scientific field [1,2], but also in to expose a clean silicon surface [15]. Stöber process [11,18] is re-
industry fields, such as catalysis [3], stabilizers [4], pigments [5] garded as the simplest and most effective route to prepare mono-
and chemical mechanical polishing [6]. The size and uniformity disperse silica spheres because the reactants are normal and
of the silica particles have great influence on the quality of these reaction condition is controllable and is easy to be carried out. Seed
products. The demand for silica nanoparticle of narrow size distri- growth method derived from Stöber process is mainly used to pre-
bution and monodispersity has increased [7,8]. pare silica particles of larger than 1 lm. A certain amount of small
As for synthesis process of silica particles, three main types of particles prepared from Stöber process are introduced into the
methods were proposed [9–19]. Pileni [9] prepared silica particles alcoholic solution at the presence of NH3 and H2O. Then, the mea-
in the size range of tens–hundreds nanometers by using micro- sured TEOS is added into the growth solution above mentioned,
emulsions method. This process works particularly well for parti- and the hydrolyzed TEOS reacts with seed particles to form uni-
cles having a size of 30–60 nm, yielding silica spheres with better form larger particles rather than new particles through adjusting
average monodispersity in comparison with particles synthesized the parameters of reaction [16,17,19].
by the Stöber process [10,11]. However, it requires large amounts Since Stöber prepared silica particles in the range of 0.05–2 lm
of surfactants (typically several times more by weight than silica) through the hydrolysis and condensation of alkyl silicates under
to form inverse micelles [12]. Therefore, the particles need to be alkaline conditions in alcoholic solutions [11], many researches
further cleaned so as to remove the remained surfactant. Moreover, were conducted in this area [16–19]. Chen prepared 73–730 nm-
it is still a great challenge to sustain uniform micelle compart- sized particles with 0.068–0.22 M TEOS [16]. Wang et al. synthe-
ments in soft and sensitive phases [13]. Balthis and Mendenhall sized silica particles ranging from 20 nm to 880 nm with 0.05–
[14] reported the production of monodisperse small silica particles 0.67 M TEOS [18]. Bogush et al. [19] presented a correlation which
through hydrolysis finely-divided elemental silicon with water can be used to predict final particle sizes over concentrations of
using ammonia as catalyst in the temperature range of 20–90 °C. 0.1–0.5 M TEOS in experiments. In addition, they found that at-
Meanwhile, notable disadvantage of this method is that before tempts to increase the solids weight fraction above 3% by simply
increasing the initial TEOS concentration resulted in heterodis-
perse particle size distribution. Therefore, they proposed a seeded
* Corresponding author. Fax: +86 21 65984652.
growth technique to increase mass fraction up to 17%, but this
E-mail address: wangzs@tongji.edu.cn (Z.-S. Wang). process required long reaction time and multi-addition steps

0021-9797/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcis.2009.09.018
24 X.-D. Wang et al. / Journal of Colloid and Interface Science 341 (2010) 23–29

[19]. But high TEOS concentration in the reaction solution means addition; low ratio of NH3/TEOS; high TEOS concentration. In order
high production of silica particles, which counts most for industry to attain the highest mass fraction, we choose the excess water
production. Hence, it is of both fundamental and practical impor- addition to make the hydrolysis of TEOS as thorough as possible.
tance to build a simple and readily available reaction system allow- Note that NH3 cannot only accelerate the hydrolysis and condensa-
ing for high TEOS concentration. tion of TEOS, but also supply the OH adhering on the particle,
In this paper, Stöber process with high concentration of TEOS up which can stabilize the suspension by static repulsion against
to 1.24 M is implemented to prepare submicrometer particles, and van der Waals attractive force. Thus coagulation may be formed
we successfully increase mass fraction up to 7.45%, which is 1.85 if the concentration of NH3 is too high, and the reaction will need
times as large as that can be obtained in Wang et al.’s experiment long time or even no silica particles can be obtained if the concen-
[18]. The reaction conditions including TEOS, NH3, H2O, solvents, tration of NH3 is too low. To solve this problem, an appropriate NH3
temperature, and the formation mechanism of particles are also concentration should be chosen, and we choose 0.81[TEOS] for the
investigated. concentration of NH3 by multiple experiments.
The detailed synthesis conditions of our reaction system are
presented in Table 1. As can be seen in Table 1 and Fig. 1, mono-
2. Material and methods
disperse particles can be achieved at 0.67–1.115 M TEOS and bi-
modal final particle size distribution appears at 0.22–
2.1. Reagents
0.56 M TEOS, and then coagulation occurs at higher concentration
at 20 °C. This phenomenon has never reported for the preparation
Ammonium hydroxide (NH3, 25%, A.R.), TEOS (98%, A.R.), etha-
of silica particles in the past. Furthermore, the diameter of the
nol (99.9%, MOS), and isopropanol (IPA, A.R.) are purchased from
particles also increases gradually as the concentration of the TEOS
Sinopharm Chemical Reagent Co., Ltd., China. Ammonium hydrox-
increases (Fig. 2a), and the TEOS quantity has a significant influ-
ide is used as a catalyst for the hydrolysis and condensation of
ence on the particle diameter. Still, the fact that decreasing of
TEOS. Water (>18 M X) is obtained from a Milli-Q water purifica-
temperature contributes to large particle formation reported by
tion system (Nanjing high-pure water Co., Ltd., China). The follow-
Tan et al. [21] can be confirmed by our experiment below. The
ing densities (q) and molecular weights (MWS) are used to
monodisperse particles in the range of 920–940 nm can be ob-
calculate the concentrations of the various chemicals: qTEOS =
tained for 1.24 M TEOS at 5 °C (TE9 in Table 1), but coagulation
0.934 g/cm3, MWTEOS = 208.33 g/mol; MWwater = 18 g/mol;
can be formed at 20 °C (TE8 in Table 1). Low temperature may
qammonia = 0.907 g/cm3 (25 wt.% NH3), MWammonia = 17 g/mol;
not only slow down TEOS hydrolysis and condensation reactions,
qSiO2 = 1.9 g/cm3 [20], MWSiO2 = 60.09 g/mol.
but also reduce vibration of oligomers, which may avoid aggrega-
tion among oligomers to some extent and result in formation of
2.2. Experimental procedures
large particles at high TEOS concentration. To the best of our
knowledge, this is the first time to give the detailed preparation
Silica particles in the size range of 30–1000 nm are synthesized
condition for 900 nm-sized silica particles by one-step process
by hydrolysis and condensation of TEOS in alcohol solvent at the
although many researchers claimed that 900 nm-sized silica par-
present of ammonium hydroxide [11]. The reactions are carried
ticles can be prepared by Stöber process [11,19]. Fig. 3 shows the
out at [TEOS] = 0.22–1.24 M, low concentrations of ammonia
particle size distribution of 920–940 nm-sized silica particles (TE9
([NH3] = 0.81[TEOS]), and [H2O] = 6.25[TEOS] in isopropanol. First,
in Table 1).
a certain amount of deionized H2O, NH3 and isopropanol were
Thirty nanometers-sized silica particles are achieved for
blended for 10 min, and then a known volume of TEOS was added
0.22 M TEOS at 20 °C with the mass fraction equal to 1.32%, which
dropwise into the above solution at certain temperatures under
is 3.14 times of the highest results reported [18]. In addition, the
magnetic stirring. After the addition of TEOS, the clear solution
7.45% solid content of resulting suspension for 1.24 M TEOS at
gradually turned opaque due to the formation of a white silica sus-
5 °C is about 1.85 times of maximum results of Wang et al. [18].
pension, and it was continuously stirred for 5 h. The silica particles
Table 2 presents the comparison on mass fraction between our
were centrifugally separated from the suspension and washed with
work and the highest results reported.
water and ethanol for four times, and then they are dried in an
In our experiment, methanol, ethanol, and isopropanol are used
oven at 150 °C for 5 h before the characterization. The adding
to compare and to observe the influence of the solvents on the par-
quantity of reactants was measured by volume. All concentrations
were calculated based on final concentration in the reaction mix-
ture. All the vessels were cleaned with water under sonication at
Table 1
40 kHz and 80 kHz before using. The vessels were coved by plastic The silica particles prepared by one-step process ([NH3] = 0.81[TEOS],
membrane to prevent NH3 volatilization so as to achieve good [H2O] = 6.25[TEOS]).
reproducibility.
Number Solvent Temperature TEOS Particle SD (±nm)
The size and morphology of the resulting particles were deter- (°C) (M) diameter
mined with a field emission scanning electron microscope (FE- (nm)
SEM, JEOL JFC-1600, Japan). The size and size distribution of silica TE1 IPA 20 0.22 30 Bimodal
samples were measured by dynamic light scattering (DLS, Horiba distribution
LB-550, Japan). The composition of the silica samples were ana- TE2 IPA 20 0.45 275 Bimodal
lyzed using Fourier Transform Infrared (FT-IR) spectrometer (Bru- distribution
TE3 IPA 20 0.56 403 Bimodal
ker-Tensor27). distribution
TE4 IPA 20 0.67 680 50
TE5 IPA 20 0.75 783 42
3. Results and discussion
TE6 IPA 20 0.90 820 33
TE7 IPA 20 1.115 860 25
3.1. Reaction system TE8 IPA 20 1.24 Aggregation
TE9 IPA 5 1.24 930 30
S1 Methanol 5 1.24 Aggregation
Our reaction system ([NH3] = 0.81[TEOS]), [H2O] = 6.25[TEOS],
S2 Ethanol 5 1.24 Aggregation
isopropanol as solvent) has three major features: the excess water
 X.-D. Wang et al. / Journal of Colloid and Interface Science 341 (2010) 23–29 25

Fig. 1. SEM images of silica particles synthesized under our reaction system: (a) TE1, (b) TE2, (c) TE3, (d) TE4, (e) TE5, (f) TE6, (g) TE7, (h) TE8, (i) TE9 in Table 1.

ticle size. As shown in Table 1 and Table 3, coagulation can be ob- ferent dielectric constants (methanol = 32.6, ethanol = 24.3, isopro-
tained when we choose methanol and ethanol as solvents at panol = 19.9), which affects the particles size in the growth process
1.24 M TEOS, and small particles can be formed with the increment [18]. The static repulsive force between nuclei is smaller than the
of polarity of solvent. This is because these three solvents have dif- van der Waals attractive force when the solvent’s dielectric con-
26 X.-D. Wang et al. / Journal of Colloid and Interface Science 341 (2010) 23–29

Fig. 3. Particle size distribution of 920–940 nm-sized silica particles (TE9 in Table
1).

Table 2
The comparison on mass fraction between our work and the highest results reported.

Number Particle TEOS Preparation conditions Mass

diameter (M) NH3 (M)/H2O fraction (%)
(nm) (M)/solvent/
temperature
TE1 30 0.22 0.178/1.375/IPA/20 °C 1.32
N1 [18] 23.80 0.07 0.26/0.55/ethanol/40 °C 0.42
TE9 930 1.24 1.004/7.750/IPA/5 °C 7.45
TEOS5 [18] 882.7 0.67 0.52/8.32/ethanol/40 °C 4.02

Table 3
Experimental data for 0.75 M TEOS at 20 °C in different preparation conditions.

No. NH3 H2O Solvent Particle SD (±nm)

(M) (M) diameter (nm)
A1 0.15 4.7 IPA 445 Bimodal distribution
A2 0.46 4.7 IPA 508 75
A3 0.61 4.7 IPA 783 25
A4 0.76 4.7 IPA 873 30
A5 1.06 4.7 IPA 616 Aggregation
W1 0.61 1.9 IPA 218 Polydispersity
W2 0.61 3 IPA 313 Polydispersity
W3 0.61 4.7 IPA 783 25
W4 0.61 6 IPA 799 40
W5 0.61 9 IPA 753 50
S3 0.61 4.7 Methanol 105 Aggregation
S4 0.61 4.7 Ethanol 170 50

3.2. Effect of H2O and NH3 concentration

Fig. 2. Effects of the amounts of TEOS (a), NH3 (b), and H2O (c) on the particle Fig. 2b shows that the particle size rises first and then aggrega-
diameter. tion (A5 in Table 3) is formed with the increment of NH3 concen-
tration. This result is inconsistent with that of Wang et al. [18].
stant is smaller. Therefore, it promotes the reaction between nu- NH3, as a catalyst, could accelerate TEOS hydrolysis and condensa-
clei, and the formation of larger particles is easier [18]. tion reactions, resulting in producing more oligomers to form lar-
Based on the above consideration, three major parameters con- ger particles, but at higher NH3 concentration, aggregation will
tribute to the best result (7.45% mass fraction): (1) low tempera- be formed if condensation rate is larger than TEOS hydrolysis rate.
ture stabilizes the solution; (2) low ratio of NH3/TEOS avoids the Fig. 2c illustrates that the influence of the water concentration on
coagulation; (3) low polarity of IPA benefits the formation of large particle size follows the similar trend as NH3, that is, the particle
particles, and steric effect of IPA avoids the aggregation. size increases with increasing H2O concentration and reaches the
We can infer that much higher mass fraction of particles in the maximum value at about 6 M, and then descends at higher concen-
range of 30–1000 nm may be achieved by optimizing reaction tration. This result shows good agreement with the work of
parameters (temperature, reactants), selecting multi-alcohol (gly- [16,18,19]. H2O could accelerate TEOS hydrolysis to contribute to
col and glycerol) to increase the effect of steric hindrance, and formation of larger particles, while at higher H2O concentration,
choosing appropriate high-polarity alcohol to reduce viscosity. H2O dilutes the oligomers in reaction solution, resulting in the for-
 X.-D. Wang et al. / Journal of Colloid and Interface Science 341 (2010) 23–29 27

mation of smaller particles. Much detailed discussion about effect particles. This can be proven by TE1–TE8 and TE8–TE9 in Table 1,
of NH3 and H2O on the particle size could be seen in Section 3.3 and respectively. The ions concentration reduces the double layer
4. Moreover, Bougush et al. [19] found that near the maximum thickness making repulsive potential decline and leads to a disper-
achievable size for any given TEOS concentration, monodispersity sion that is more easily flocculated [26]. An increase in the concen-
could hardly be achieved and bimodal final particle size distribu- tration of NH3 and H2O will not only stabilize the suspension
tions were commonly observed. However, in our experiments, near (through the production of more OH ) by increasing the net nega-
the maximum achievable size (873 nm) for 0.75 M TEOS, monodis- tive charges on the particles, but also destabilize the dispersion
persity is often achieved and only the bimodal final particle size resulting from an increase in the ionic strength through the pro-
distribution can be observed at 0.15 M NH3 and 4.7 M H2O (A1 in duction of ions. That is, at certain concentrations of H2O and
Table 3). Further investigations on this phenomenon are underway NH3, the destabilizing effects will influence the size distribution
to understand completely the formation mechanism of particles in more remarkably comparing with the increase in surface charges,
this reaction system. and aggregations can be formed [26]. As a proof, A5 particles (in
Table 3) with a diameter of 616 nm aggregated at 1.06 M NH3.

3.3. Size distribution of particles

3.4. FT-IR analysis

Particle formation is rate-limited by the slowest step, the

Fig. 4 shows that FT-IR spectra of silica spheres (a) TE1, (b) TE2,
hydrolysis of the TEOS. The balance between monomer addition
(c) TE3, (d) TE4, (e) TE5, (f) TE6, (g) TE7, (h) TE9 in Table 1. The very
and nucleation determines the size distribution of particles and fi-
intense and broad band appearing at 1082–1114 cm 1 is assigned
nal particle size [22–27]. The ratio of hydrolysis rate and conden-
to asymmetric stretching vibrations of Si–O–Si [28], which reveals
sation rate directly influence the final particle size and size
that a dense silica network was formed. On the other hand, one can
distribution. As long as hydrolyzed monomers react much more
see the other important infrared vibrations of silica in Fig. 4: bend-
quickly than they are produced, monodisperse particles can be ob-
ing mode of Si–O–Si appearing at about 467 cm 1 [29], and
tained (TE4–TE7 in Table 1, A2–A4 and W3–W5 in Table 3); other-
stretching of O–H appearing at 3352–3477 cm 1 [30]. This sug-
wise, particles with bimodal distribution and polydispersity will be
gests that no significant structure differences exist among TE1,
achieved (TE1–TE3 in Table 1, A1 and W1, W2 in Table 3). In addi-
TE2, TE3, TE4, TE5, TE6, TE7, and TE9 particles in Table 1.
tion, according to the result of previous reports [26], the reason for
aggregation includes two aspects: (1) the distance between vibrat-
ing particles; (2) ions concentration. Low concentration of TEOS 4. Mechanism of formation
and low temperature can increase the average distance between
vibrating particles to avoid the overlapping of double electrical lay- Despite many investigations have been done about the mecha-
ers of particles, which benefits the formation of well-dispersed nism of silica particle formation, growth, size distribution, particle
size, and shape [20–27,31–35], but no general agreement ac-
counted for them was achieved yet. Differences that are found in
the literatures are best illustrated by two extremes (monomer
addition and controlled aggregation model). The monomer addi-
tion model argues that after a limited period time for nucleation
generation, growth occurs through the addition of hydrolyzed
monomers to the oligomers surface and becomes gradually the
dominant process [22,23]. By contrast, controlled aggregation
model claims that the aggregation of subparticles leads to particle
formation [24–27,31]. We believe that the monomer addition
model is reasonably responsible for our experimental results.
Herein we present a modified monomer addition model to explain
our experimental results as below.
We use the Q ji notation where Q denotes a tetrafunctional sili-
con (Si) site, the subscript i is the number of siloxane bridges at-
tached to the Si atom, and the j represents the number of silanol
bonds on the Si atom [35]. As we know, the hydrolysis of TEOS is
a step-by-step process, which has four kinds of hydrolyzed prod-
uct: Q 40 ; Q 30 ; Q 20 ; Q 10 . The order of the hydrolysis velocity is
Q 30 < Q 20 < Q 10 <TEOS. Si(OH)4(Q 40 ), the fully hydrolyzed species, is
three-dimensional open and fractal in structure and has four equal
functional group (Si–OH), which can give rise to uniformly com-
pact and spherical oligomers. Thus Q 40 may be the primary nuclei
or the core of the particles, which had been also described by Kee-
fer [34] using the Monte Carlo model. The reason that some
researchers can not detect Q 40 may be that it is unstable and will
quickly condense with another Q 40 into Q 31 .
As the concentration of TEOS in the solution decreases, Q 40 is
more difficult to be formed. Thus, no enough Q 40 reacts with oligo-
mers. A few Q 30 and Q 20 begin gradually to react with oligomers. At
the end of reaction, few Q 10 could react with oligomers and thus
poison (do not polymerize) the active group (Si–OH). This forma-
Fig. 4. FT-IR spectra of silica spheres (a) TE1, (b) TE2, (c) TE3, (d) TE4, (e) TE5, (f) tion mechanism of silica particles can be verified by the experi-
TE6, (g) TE7, (h) TE9 in Table 1. mental results of Blaaderen and Kentgens [27], that is, silica
28 X.-D. Wang et al. / Journal of Colloid and Interface Science 341 (2010) 23–29

particles made from the hydrolysis and condensation of TEOS con-

tain about 60% Q4, 30% Q3 and several percent Q2, and that several
percent of the ethoxy groups never leave the TEOS molecules. Of
course, similarly to other polymerizations, this reaction has side
products, such as Q 01 and Q 12 , which have already been detected
by Lee et al. [36].
The formed active oligomers can also react with other active
ones, which is aggregating described by Bogush and Blaaderen
[24–27,31]. When two oligomers react with each other, the forces
including chemical bond and van der Walls force may cause two
oligomers generating elastic and plastic deformation [37,38],
which may make two oligomers engulf into each other to keep
spherical shape. As a proof, we conduct a seed growth experiment
in which TE9 is used as seeds, the growth condition is:
0.22 M TEOS, 6 M H2O, and 2 M NH3. As illustrated in Fig. 5, small
particles react with seed particles at 10 s after reaction begins,
and then small particles become almost invisible protuberances
by adhesion-induced deformation at 20 s.
Generally speaking, Q 40 is difficult to be obtained as the quantity
of H2O or NH3 reduces remarkably. But a large amount of
Q 30 and Q 20 emerge [33], and they can also react with oligomers.
Participation of relatively larger amount of Q 30 and Q 20 may result
in irregular shape and rough surface. This is because Q 30 is unsym-
metrical in structure and Q 20 is open in only one dimension. This
mechanism can be verified by Fig. 6, A1 (Fig. 6a) and W1
(Fig. 6b) show irregular shape and rough surface in appearance
due to lack of NH3 and H2O, respectively.
Using this modified monomer addition model, we can also ex-
Fig. 6. SEM images of A1 (a) prepared at 0.15 M NH3, 4.7 M H2O, 0.75 M TEOS and
plain the seed growth experiments: the detailed process of the W1(b) prepared at 0.61 M NH3, 1.9 M H2O, 0.75 M TEOS in Table 3.
seed growth method can be seen in Refs. [16,17,19]. The seeds in
the growth solution can be activated by NH3 and H2O, which can
make Q 10 ; Q 20 ; and Q 30 on the particles further hydrolyze, and thus hydrolyzed product of new adding TEOS again, and larger particles
open and fractal structure on the seed particles is formed. Some can be formed.
hydrolysis product may diffuse into the solution, while others still
cling to the particle surfaces. Thus the seeds can react with the 5. Summary

We successfully design a reaction system characterized by high

TEOS concentration to attain high mass fraction up to 7.45%. More-
over, it is inferred that by optimizing the system parameters, it is
possible for one to improve the solid content much higher than
that of the best result within the diameter range of 30–1000 nm.
We generalize the monomer addition model proposed by
Matsoukas and Gulari [22,23] to include aggregation model [24–
27,31] by the mechanism of adhesion-induced deformation
[37,38] between oligomers. We believe that the Q 40 is the core or
the nucleus of the particles, Q 10 ends the reaction of adding hydro-
lyzed monomer by poisoning the active group (Si–OH), and partic-
ipation of relatively larger amount of Q 30 and Q 20 results in irregular
shape and rough surface. This is because Q 30 is unsymmetrical in
structure and Q 20 is open in only one dimension. Moreover, it is
speculated that in the seed growth solution, Q 10 ; Q 20 ; and Q 30 on
the outer surface of the particles can be further hydrolyzed, result-
ing in refreshing the particles open and fractal in structure at the
presence of NH3 and H2O, which may benefit the seed growth.

Acknowledgments

This work was supported by National Natural Science Founda-

tion of China (Grant No. 10825521), by National 863 Program
and by the Shanghai Committee of Science and Technology, China
(Grant No. 07DZ22302).

References

Fig. 5. Sample obtained at 10 s (a) and 20 s (b) in seed growth experiment (TE9 as [1] Y. Kobayashi, J. Imai, D. Nagao, M. Konno, Colloids Surf. A 326 (2008) 109–114.
seeds, growth condition: 0.22 M TEOS, 6 M H2O, 2 M NH3). [2] H. Wang, M. Yu, C.K. Lin, J. Lin, J. Colloid Interface Sci. 300 (2006) 176–182.
 X.-D. Wang et al. / Journal of Colloid and Interface Science 341 (2010) 23–29 29

[3] R.K. Iler, The Chemistry of Silica, Wiley, New York, 1979. [22] T. Matsoukas, E. Gulari, J. Colloid Interface Sci. 124 (1988) 252.
[4] H. Hassander, B. Johansson, B. Törnell, Colloids Surf. 40 (1989) 93–105. [23] T. Matsoukas, E. Gulari, J. Colloid Interface Sci. 132 (1989) 13.
[5] K. Unger, J. Chromatogr. 359 (1986) 61. [24] G.H. Bogush, C.F. Zukoski IV, J. Colloid Interface Sci. 142 (1991) 1.
[6] T. Suratwala, R. Steele, M.D. Feit, L. Wong, P. Miller, J. Menapace, P. Davis, J. [25] G.H. Bogush, C.F. Zukoski IV, J. Colloid Interface Sci. 142 (1991) 19.
Non-Cryst. Solids 354 (2008) 2023–2037. [26] A. van Blaaderen, J. van Geest, A. Vrij, J. Colloid Interface Sci. 154 (1992) 481.
[7] S.K. Park, K.D. Kim, H.T. Kim, Colloids Surf. A 197 (2002) 7–17. [27] A. van Blaaderen, A.P.M. Kentgens, J. Non-Cryst. Solids 149 (1992) 161.
[8] S. Sridhar, D.H. Rasmussen, F.P. Chen, R.K. Kannabiran, Colloids Surf. A 132 [28] A. Duran, J.M. Fernandez-Navarro, P. Casariego, A. Joglar, J. Non-Cryst. Solids 82
(1998) 45. (1986) 69.
[9] M.P. Pileni, Nat. Mater. 2 (2003) 145. [29] A. Bertoluzza, C. Fagnano, M.A. Morelli, V. Gottardi, M. Guglielmi, J. Non-Cryst.
[10] F.J. Arriagada, K.J. Osseo-Asare, J. Colloid Interface Sci. 221 (1999) 210. Solids 48 (1982) 117.
[11] W. Stöber, A. Fink, E. Bohn, J. Colloid Interface Sci. 26 (1968) 62–69. [30] C.J. Brinker, G.W. Scherer, Sol–Gel Science, The Physics and Chemistry of Sol–
[12] K.S. Finnie, J.R. Bartlett, C. Barbe, L. Kong, Langmuir 23 (2007) 3017. Gel Processing, Academic Press, New York, 1990. pp. 581–585.
[13] K.D. Hartlen, A.P.T. Athanasopoulos, V. Kitaev, Langmuir 24 (2008) 1714–1720. [31] S.L. Chen, P. Dong, G.H. Yang, J.J. Yang, Ind. Eng. Chem. Res. 35 (1996) 4487–
[14] J.H. Balthis, P. Mendenhall, Preparation of sols from finely divided silicon, USP 4493.
2614994, 1952. [32] S. Sadasivan, A.K. Dubey, Y. Li, D.H. Ramussen, J. Sol–Gel Sci. Technol. 12 (1998)
[15] J.J. Guo, X.H. Liu, Y.C. Cheng, Y. Li, G.J. Xu, P. Cui, J. Colloid Interface Sci. 326 5.
(2008) 138–142. [33] D.L. Green, S. Jayasundara, Y.F. Lam, M.T. Harris, J. Non-Cryst. Solids 315 (2003)
[16] S.-L. Chen, Colloids Surf. A 142 (1998) 59–63. 166–179.
[17] S.L. Chen, P. Dong, G.H. Yang, J. Colloid Interface Sci. 189 (1997) 268–272. [34] K.D. Keefer, Mater. Res. Soc. Symp. Proc. 73 (1986) 79.
[18] H.C. Wang, C.Y. Wu, C.C. Chung, M.H. Lai, T.W. Chung, Ind. Eng. Chem. Res. 45 [35] C.J. Brinker, G.W. Scherer, Sol–Gel Science, Academic Press, Boston, 1991.
(2006) 8043–8048. [36] K. Lee, J. Look, M.T. Harris, A.V. McCormick, J. Colloid Interface Sci. 194 (1997)
[19] G.H. Bogush, M.A. Tracy, C.F. Zukoski IV, J. Non-Cryst. Solids 104 (1988) 95– 78.
106. [37] D.S. Rimai, L.P. DeMejo, R.C. Bowen, J. Appl. Phys. 68 (1990) 6234.
[20] D.L. Green, J.S. Lin, Y.F. Lam, M.Z.-C. Hu, D.W. Schaefer, M.T. Harris, J. Colloid [38] D.S. Rimai, L.P. DeMejo, W.B. Vreeland, R.C. Bowen, S.R. Gaboury, M.W. Urban,
Interface Sci. 266 (2003) 346–358. J. Appl. Phys. 71 (1992) 2253.
[21] C.G. Tan, B.D. Bowen, N. Epstein, J. Colloid Interface Sci. 118 (1987) 290.