Journal of Sol-Gel Science and Technology 25, 121–136, 2002

c 2002 Kluwer Academic Publishers. Manufactured in The Netherlands.

Preparation of TiO2 Sol Using TiCl4 as a Precursor

D.-S. LEE∗ AND T.-K. LIU

Department of Geosciences, National Taiwan University, 245 Choushan Road, Taipei 106, Taiwan, ROC
ipo02022@mail.moeaipo.gov.tw

Received October 3, 2001; Accepted May 2, 2002

Abstract. A titanium dioxide sol with narrow particle size distribution was synthesized using TiCl4 as the starting
material. The sol was prepared by a process where HCl was added to a gel of hydrated titanium oxide to dissolve it. The
resulting aqueous titanic acid solution was heated to form titanium dioxide sol. The effects of preparation parameters
were investigated. TiCl4 was slowly added to distilled water at 5◦ C. Aqueous solution of sodium hydroxide was added
to adjust the pH of the system to 8–12. After aging for a period of time, the peptized sol was filtered and sufficiently
washed. The filtered cake was repulped in water. Hydrochloric acid was slowly added to the solution with stirring.
After condensation reaction and crystallization, a transparent sol with suspended TiO2 was formed. XRD results
show that the crystalline phase was anatase. The suspended TiO2 particles were rhombus primary particles with the
major axis ca. 20 nm and the minor axis ca. 5 nm. The TiO2 particles prepared at pH 8 had the largest surface area
of 141 cm3 /g and it was microporous. The compositions of the solution which yielded the smallest suspended TiO2
particles were TiO2 :HCl (35% HCl) = 1:1 (molar ratio), concentration of TiO2 = 10%. Hydroxypropyl cellulose
with viscosity of 150–400 cps was added as a dispersant. The sol was excellent in dispersibility and long-term
stability. Transparent thin films could be obtained through dip-coating glass substrate in the sol. The dip-coating
on glass can be less than three times to have one monolayer TiO2 . The transparent TiO2 thin film had strong
hydrophilicity after being illuminated by UV light.

Keywords: titanium dioxide, nano-material, sol, coating sol, photocatalysis

1. Introduction filters, antireflection coatings, and solar energy con-

Metal oxide particles of less than 100 nanometer in
particle size are considered as ultrafine or nanosized
particles and are precursors for the synthesis of new
materials. Titanium dioxide has been attracting much
scientific interest [1–7]. Many researchers have focused
on the preparation and characteristics of titanium diox-
ide nanoparticles as well as their applications in re-
cent years. Ultrafine TiO2 powders are extensively used
in cosmetics for high absorption of ultraviolet light,
pigment, toner, and coating materials. TiO2 thin films
have been very attractive for various applications such
as memory cell capacitors, multiplayer mirrors and
∗ To whom all correspondence should be addressed.
verters [1–8].
To prepare titanium dioxide thin film, a number of
methods have been reported such as sol-gel, CVD, hy-
drothermal and microemulsion method [2–10]. Micro-
emulsion method has been applied to synthesize TiO2
nanometer-sized particles with titanium alkoxide as the
starting material. However, this method is expensive,
because the ratio of aqueous phase to oil phase and sur-
factant in these microemulsion systems are too small
and therefore only a small amount of TiO2 nano-sized
particles can be obtained at the cost of a large amount
of oil and surfactant.
Titanium alkoxide or inorganic salt was used as a
precursor in sol-gel method. After hydrolysis reac-
tion, the transparent coating sol of titanium dioxide
122 Lee and Liu

was formed. The TiO2 sol was then coated on the sub-
strates. The TiO2 film on substrate should be calcined
at high temperature to obtain anatase structure. Since
it should be calcined at high temperature, the choice
of the substrate is limited. In addition, strict control
of the synthesis conditions is required since the alkox-
ide is easily hydrolyzed in air. Furthermore, the high
price of alkoxide limits the commercialization of this
method. The hydrothermal synthesis using the auto-
clave with high temperature and pressure condition pro-
duces high-quality powders but continuous processing
is impossible.
In this study, the acid-hydrolysis method was used
to prepare TiO2 sol with TiCl4 as a precursor. This re-
search was devoted to prepare a suspended titanium
dioxide coating sol with high concentration and nano-
sized TiO2 in anatase crystalline phase. In order to pre-
vent from calcination at high temperature, the acid-
hydrolysis method was developed. Nano-sized TiO2
crystal could be formed at low temperature. The ad-
vantage is that it does not need calcination at high tem-
perature. The following preparation parameters were
studied: pH value, [H+ ]/[Ti] ratio, HCl concentra-
tion, stored temperature and dispersant (hydroxypropyl
cellulose).
Figure 1. Procedure of the preparation of TiO2 sol.

2. Experimental
2.1. Chemicals
Titanium tetrachloride, (TiCl4 , purity >99.9%),
sodium hydroxide (NaOH, >96%) and hydrochloric
acid (HCl, 35%) was from Showa Chemicals Co., Ltd.
Two types of hydroxypropyl cellulose (HPC, viscosity
6–10 cps and 150–400 cps, 2% in water) were from
Merck.
2.2. Preparation of TiO2 Sol
The preparation procedure is shown in Fig. 1. Titanium
tetrachloride was used as a precursor. It was slowly
added to the distilled water at 5◦ C. When TiCl4 dis-
solved in distilled water, the heat of the exothermic
reaction explosively generated the formation of orthoti-
tanic acid [Ti(OH)4 ]. Because the formation of this
species disturbed homogeneous precipitation, 5 ml of
hydrochloric acid (35% HCl) per 100 ml of distilled
water was added to prevent formation of orthotitanic
acid. Concentrated aqueous sodium hydroxide solution
was added to adjust the pH of the system to 8–12. After
aging for a period of time, the peptized sol was filtered
and sufficiently washed to remove chlorine ion from
the cake. The washed cake was repulped in water. Hy-
drochloric acid was slowly added to the solution with
stirring. After condensation reaction and crystalliza-
tion, transparent suspended TiO2 sol was obtained.
TiO2 film on glass was prepared by dip-coating.
Glass slide was used as a substrate. The glass slide
was cleaned in a neutral washing solution, then put
into ultrasonic water bath for an hour. Finally, it was
washed with deionized water to remove the remaining
neutral washing solution. The glass was washed with
NaOH solution, then rinsed by distilled water. After
drying, the glass was used as a substrate. The glass was
dip-coated with TiO2 sol once or several times.
2.3. Characterization
2.3.1. X-ray Diffraction (XRD). The crystalline
structures of the products were analyzed by X-ray pow-
der diffraction by using a Siemens D500 automatic
 Preparation of TiO2 Sol Using TiCl4 as a Precursor
powder diffractometer system. Nickel-filtered Cu Kα
radiation (λ = 0.15418 nm) was used with a generator
voltage of 40 kV and a current of 29 mA. The Bragg-
Brentano focusing geometry was employed with an au-
tomatic divergence slit (irradiated sample length was
12.5 nm), a receiving slit of 0.1 nm, a fixed slit of 4◦
and a proportional counter as a detector. It was operated
in the step scan mode, at scanning speeds of 0.05◦ 2θ /s
and 1 second step time from the range 20◦ to 80◦ .
Scherrer’s equation was used to calculated the par-
ticles size of titanium dioxide crystal:
kλ
D= (1)
β cos θ
where D is crystallite particle size, κ is a constant of
0.94, λ is X-ray wavelength (0.154 nm), β is half max-
imum line breadth, and θ is Bragg angle.
2.3.2. N2 Sorption. The measurement of the sur-
face areas of the samples was achieved by Brunauer-
Emmett-Teller (BET) method, and the pore size dis-
tributions were determined by Barrett-Joyner-Halenda
(BJH) method. Nitrogen sorption measurements were
carried out in a Micromeritics ASAP2010 unit.
2.3.3. Thermogravimetric Analysis (TGA). The
TGA was carried out on a Perkin Elmer TGA-7 unit
with a heating rate from 50 to 600◦ C. The sample was
heated in a nitrogen flow at a heating rate of 10◦ C/min.
2.3.4. Differential Scanning Calorimetry (DSC).
DSC was carried out in a Perkin Elmer DSC-7 with
a heating rate of 10◦ C/min from 50 to 450◦ C.
2.3.5. Transmission Electron Microscopy (TEM).
The suspended TiO2 was deposited on a grid with a
holey carbon copper film and placed several days in
room temperature. It was observed in a JEM-1200 EX
II electron microscope operating at 160 kV.
2.3.6. Scanning Electron Microscopy (SEM). The
samples were dried at 35◦ C, then placed on a stage, spe-
cially mode for SEM. The samples were coated with Au
for 4 minutes for discharge before the experiment and
were observed using a Hitachi S-800 machine. SEM
was operated at 200 kV.
2.3.7. Dynamic Light Scattering (DLS). The particle
size can be measured by DLS by using Zetasizer 3000.
Dynamic light scattering measures Brownian motion
123
and relates it to size. Random movement of particles is
due to bombardment by the solvent molecules that sur-
round them; the larger the particle the slower will be the
Brownian motion. On the other hand, the temperature
is also an important factor in Brownian motion. From
the above step, using the Stokes-Einstein equation to
calculate the hydrodynamic diameter d(H ):
κλ
d(H ) = (2)
3π ηD
where d(H ) is hydrodynamic diameter, κ is
Boltzmann’s constant, T is absolute temperature, η is
viscosity, and D is diffusion coefficient.
2.3.8. Ultraviolet/Visible Absorption (UV-Vis). The
diffuse reflectance UV-vis was measured with a Jasco
UV V-500 spectrophotometer. Powder samples were
loaded in a quartz cell with suprasil windows, and spec-
tra were collected in the range from 300 nm to 800 nm
against quartz standard.
2.3.9. Contact Angle. The contact angle analysis was
measure by using Model CA-D (Kyowa Interface Sci-
ence Co., Ltd. Japan). 40 µl or 140 µl deionized water
was dropped on the glass with TiO2 thin film. The con-
tact angle is based on the equation as follows:
r hρg
γ = (3)
2 cos θ
where γ is surface tension of a capillary, r is radius of
a capillary, h is the height of liquid, ρ is the density of
liquid, g is gravity, and θ is contact angle.
3. Results and Discussion
3.1. Particle Size
3.1.1. Effects of pH Value. After hydrolyzing TiCl4
for 1 h, NaOH was added to adjust pH value of the
solution. Based on literature [11], the isoelectric point
(IEP) of TiO2 is at a pH of about 4.5–6.8. If the pH value
is adjusted to this region, there is no repulsive force
between particles so one can maintain the dispersion
of the particles. The particles collide with each other
and aggregate and precipitate due to van der Waals
interaction. Therefore, the pH value was adjusted to
greater than 8 or to value lower than 3 to have stable
TiO2 sol.
As shown in Fig. 2, the anatase type TiO2 particle
was obtained at pH of 8–11, yet at pH 12 there is no ap-
parent TiO2 crystal peak at 2θ = 25.4. Figure 3 shows
124 Lee and Liu

Figure 2. XRD patterns of TiO2 prepared at various pH values.

Figure 4. SEM image of aggregated TiO2 , but still could see some
dispersed particle <100 nm.

However, some dispersed particles with size less than

Figure 3. XRD patterns of TiO2 prepared at pH 8, [H+ ]/[Ti] = 1,
HCl = 1 M.
the XRD spectrum of TiO2 prepared at pH of 8. The
d(101) peak at 25.4◦ was used to calculate the mean
crystallite size. Scherrer’s equation was used to calcu-
late the mean diameter of a crystal particle. At pH 8,
the mean diameter of the crystallite was the smallest
(Table 1). Figure 4 shows that the TiO2 was aggregated
if the sol was stored at room temperature for long time.
100 nm were still observed.
As shown in the TEM image (Fig. 5), the primary
particles of the sample prepared at pH 8 are rhombus
form with major axis of 20 nm and minor axis of 5 nm.
At pH 11 and 12, the major axis of the rhombus particle
is 40–50 nm and the minor axis is 10 nm. One can
conclude that at pH 8, well-crystallized anatase TiO2
with small size was obtained. The DLS results also
show that the particle distributed in 5–10 nm range
were about 92%, and just only 8% of particles would
aggregate into secondary particles.
The particle size of TiO2 can be determined by XRD,
TEM and DLS [16–20]. The particles sizes from these
three measurements are listed in Table 2. The results
are consistent.

Table 1. Effect of pH value on the mean diameter of TiO2 crystallite Table 2. The particle size of TiO2 from various measurements.
from XRD results.
Method XRD TEM DLS
pH 8 9 10 11 12
Mean diameter 7.8 nm Major axis of 20 nm Main in 5–10 nm
Mean diameter (nm) 7.8 nm 7.9 nm 8.1 nm 8.3 nm 8.7 nm (nm) Minor axis of 5 nm Minor in 30 nm
 Preparation of TiO2 Sol Using TiCl4 as a Precursor 125

Figure 5. TEM image of TiO2 prepared at various pH values. (A) pH 8; (B) pH 9; (C) pH 10; (D) pH 11 (stored temperature 30◦ C).
126 Lee and Liu

3.1.2. Effects of [H+ ]/[Ti] Ratio. In general, while

using inorganic salt to produce TiO2 , most researchers
used strong acid, such as HNO3 or HCl, to carry out
condensation reaction of the gel, and the molar ratio of
[H+ ]/[Ti] was about 1 [21–24]. In this study, [H+ ]/[Ti]
ratios of 0.6, 0.8, 1 and 1.5 were used to elucidate the
influence of this ratio on the sol. When [H+ ]/[Ti] ratio
was 0.6 and 0.8, relatively long time was needed for
condensation reaction and crystal growth. The TiO2
was not well-crystallized and the suspended sol was
not transparent, indicating that the particle size of
TiO2 was quite large. There was some non-dissociated
orthotitanic acid that would form TiO2 . At [H+ ]/[Ti]
ratio of 1 and 1.5, the effects of the amount of acid
on condensation reaction and crystal growth were
observed. At [H+ ]/[Ti] ratio of 1.5, the time needed
for condensation reaction and crystal growth was
about 24 h. At [H+ ]/[Ti] ratio of 1, the time needed
was 48 h. As shown in Fig. 6, the difference in the
crystallinity of TiO2 synthesized at [H+ ]/[Ti] ratio
between 1.5 and 1 was not greater than those at 1, 0.6
and 0.8 (Figs. 3 and 5). Besides, at [H+ ]/[Ti] ratio of
1, the mean diameter of the TiO2 particle was 7.8 nm
(Fig. 7), while at [H+ ]/[Ti] ratio of 1.5, it was 8.0 nm
(Fig. 8). The difference is very small.

3.1.3. Effects of HCl Concentration. After filtration

and sufficiently washing of orthotitanic acid cake, one
can repulp the cake by adjusting the proportion of cake
and water. After stirring for 1 h, various concentrations
of HCl with constant [H+ ]/[Ti] ratio was added to un-
dergo condensation reaction and crystal growth.

Figure 6. XRD patterns of TiO2 prepared at various [H+ ]/[Ti]

ratios. Figure 7. TEM image of TiO2 prepared at pH 8 and [H+ ]/[Ti] = 1.
 Preparation of TiO2 Sol Using TiCl4 as a Precursor 127

Figure 9. XRD patterns of TiO2 prepared at various HCl concen-

trations (pH 9; [H+ ]/[Ti] = 1).

From XRD spectra (Fig. 9), one can see that the crys-
tallinity of TiO2 prepared with 2 M HCl was greater
than that prepared with 1 M HCl. However, the parti-
cle size of TiO2 prepared with 1 M HCl was smaller
than that with 2 M HCl (Fig. 10). In addition, the time
needed for condensation reaction and crystal growth
was 36 h for the sample prepared with 2 M HCl, which
was shorter than that prepared with 1 M HCl. The re-
sults show that the higher the concentration of acid, the
higher is the [H+ ]/[Ti] ratio, and the shorter is the time
needed for condensation reaction and crystal growth.
The highest crystallized nano-sized TiO2 particles were
obtained from the shortest time of crystallization. How-
ever, in stronger acid or higher [H+ ]/[Ti] ratio, the zeta
potential will change and will affect the stability of the
sol, which will be discussed in a later section. If sin-
gle crystal particle with small diameter and well crys-
tallinity is needed, one can immediately titrate basic
solution (ex: ammonia water or sodium hydroxide) af-
ter condensation reaction and crystal growth to change
the original pH value of acidic TiO2 sol to the iso-
electric point (IEP). After precipitation, TiO2 particles
become a little larger, yet its crystallinity is still very
high.

3.1.4. Effects of Storing Temperature on Particle

Diameter. In this study, the synthesis system was
maintained at 5◦ C. After pH adjustment and aging,
smaller orthotitanic acid particles could undergo
condensation reaction and crystal growth, resulting in
small diameter of TiO2 particle. Besides, in filtration
Figure 8. TEM image of TiO2 prepared at pH 8 and [H+ ]/[Ti] = and washing steps, the diameter of orthotitanic acid
1.5. decreased with the number of filtration and washing
128 Lee and Liu

times, especially when 5◦ C distilled water was used in

washing.
The produced TiO2 sol was separated into two parts,
one was stored at 30◦ C and the other at 5◦ C. From TEM
images (Fig. 11), only a small part of the primary par-
ticles sample at 5◦ C would aggregate into secondary
particles or tertiary particles, especially in high con-
centration TiO2 . This was due to smaller kinetic en-
ergy of particles in low temperature, resulting in fewer
collisions, hence double layer (Stern layer and diffu-
sion layer) is larger than van der Waals force between
particles, and stable suspension was obtained.

3.2. Stability of Suspension

3.2.1. Effects of pH Value. The addition of acid is

needed to proceed condensation reaction. However
if the amount of acid is too small, no condensation
reaction would occur. Grishchenko [11] reported that
stable TiO2 sol could be obtained at pH value of 2–3.
The effects of pH value on the properties of sol are
shown in Table 3. The higher the amount of acid is,
the lower the pH value is, the less the stability of the
sol. On the other hand, relatively small amount of
the acid will result in incapability of peptization, and
the pH value will be in its isoelectric point, causing
aggregation due to van der Waals interaction. One
can conclude that at pH value higher or lower than
isoelectric point, there should be a pH value where its
zeta potential is the greatest, sufficient to resist van
der Waals interaction and to hold its own weight, and
no aggregation and precipitation occur.

3.2.2. Effects of HCl Concentration. Adding high

concentration HCl solution into orthotitanic acid cake

Table 3. Effect of pH value on the properties of sol.

Initial pH [H+ ]/[Ti]

value (Molar ratio) End pH value Transparency

2 1 Incomplete Gel
3 1 Incomplete Gel
8 1 0.98 Transparent
9 1 1.11 Transparent
10 1 1.25 Transparent
11 1 1.32 Transparent
12 1 1.40 Transparent
8 0.6 3.9 Opacity
8 0.8 3.2 Opacity
Figure 10. TEM images of TiO2 prepared at various HCl concen-
8 1.5 <0 Opacity
trations. (A) pH 8, 1 M HCl; (B) pH 8, 2 M HCl.
 Preparation of TiO2 Sol Using TiCl4 as a Precursor
Figure 11. TEM images of TiO2 at various stored temperatures. (A) pH 8, 5◦ C; (B) pH 8, 30◦ C.
can accelerate dissolution rate, and, after a while, all
the solids will be dissolved. Therefore, the time for
condensation reaction and crystal growth depend on
the ratios of Ti:acid:water. The high HCl concentra-
tion only accelerates condensation reaction and crys-
tal growth at the beginning of this step. Although the
time for crystal growth became shorter, the TiO2 would
aggregate and precipitate to form more stable type of
rutile. Table 4 shows that at HCl concentration of 4 M,
the time needed for crystal growth was the same if the
[H+ ]/[Ti] ratio was constant at 1.5. This shows that
the effect of [H+ ]/[Ti] ratio on the dispersibility and
stability of TiO2 sol is significant.
3.2.3. Effect of [H+ ]/[Ti] Ratio. As shown in Table 4,
the sample prepared at [H+ ]/[Ti] ratio of 1 had the
best dispersibility.
Table 4. Effect of HCl concentration on the properties of sol.
1 M HCl
The time for crystal growth 48 h
Stored temp. (30◦ C) Stable suspension for 3 months
Stored temp. (5◦ C) Stable suspension for 3 months
The sols contain 5% TiO2 and were prepared at [H+ ]/[Ti] ratio of 1.
129
3.2.4. Effects of Stored Temperature. Two stored
temperature were studied: 5◦ C and 30◦ C. The sample
stored at 5◦ C resulted in low kinetic energy of particles.
The chance to form secondary particles becomes less
(Tables 4 and 5). The best way to preserve TiO2 sol is
to keep them at low temperature.
3.2.5. Effect of TiO2 Solid Content. Table 6 and
Fig. 12 shows the effect of TiO2 solid content on the sta-
bility of sol. As expected, the higher the solid content,
the lower the stability of sol. However, it demonstrated
good stability even for the sample with 10% solid con-
tent. The results show that the samples prepared in this
study had a very good dispersibility and stability.
3.2.6. Effect of HPC Dispersant. Many researchers
[5, 9] reported that adding surfactant could maintain
2 M HCl 4 M HCl
36 h 24 h
Stable suspension for 8 days Stable suspension for 5 days
Stable suspension for 14 days Stable suspension for 11 days
130 Lee and Liu

Table 5. Effect of [H+ ]/[Ti] ratio on the properties of sol.

[H+ ]/[Ti] = 0.6 [H+ ]/[Ti] = 1 [H+ ]/[Ti] = 1.5

The time for crystal growth 60 h 48 h 24 h

Stored temp. (30◦ C) Incomplete crystal growth Stable suspension for 3 months Stable suspension for 3 days
Stored temp. (5◦ C) Incomplete crystal growth Stable suspension for 3 months Stable suspension for 44 h

The sols contain 5% TiO2 and were prepared at HCl of 1 M.

Table 6. Effect of TiO2 solid content on the stability of sol.

2.5% TiO2 5% TiO2 7% TiO2 10% TiO2 a

Stored temp. (30◦ C) Stable suspension for
3 months
Stored temp. (5◦ C) Stable suspension for
3 months
Adding HPC (30◦ C) Stable suspension for
3 months
Adding HPC (5◦ C) Stable suspension for
3 months
a Samples were prepared with 35% HCl.
Stable suspension for 3 months
Stable suspension for 3 months
Stable suspension for 3 months
Stable suspension for 3 months
Stable suspension Stable suspension for 13 h
for 14 h
Stable suspension Stable suspension for 14 h
for 26 h
Stable suspension with Stable suspension for 26 h
a little white color
Stable suspension with Stable suspension for 30 h
a little white color

stable suspension of particles. The surfactants play
steric dispersing agent roles, which can lower parti-
cles diameter to less than µm [5]. The surfactant can
adsorb on TiO2 particle surface, changing its zeta po-
tential, providing stability of the sol, and avoiding the
aggregation of particles [5, 9].
Two kinds of HPC were used as the surfactant in this
study, one with viscosity of 6–10 cps, and the other
with viscosity of 150–400 cps. They were individually
added to TiO2 sol to examine their effects on parti-
cle dispersion. After formation of TiO2 sol, 0.005 g
HPC/ml sol was added. The results show that the HPC
with viscosity of 150–400 cps was better than that with
viscosity of 6–10 cps. Therefore, if TiO2 solid content
was greater than 7% in the sol, the HPC with viscos-
ity of 150–400 cps was added to increase its stability
(Fig. 12). Table 6 shows that adding HPC in the sol
could increase the stability of TiO2 sol. The life time
for the sample with 10% solid content increased from
14 h to 30 h. This confirms that HPC indeed changed
the zeta potential of TiO2 particle and increases its
stability of dispersion.
3.2.7. Effects of Residual Ions. Adding TiCl4 into
distilled water will hydrolyze into titanium hydrate
compound. This step will produce chlorine ion. Re-
lated documents indicate that chlorine ion has disad-
vantageous effect on product [3, 20].
Although chlorine ion can interact with metal ion,
due to low electronegativity, Ti Cl bond was not
formed. Therefore chlorine ion is present in the aqueous
phase. It does not participate in polymerization reac-
tion, but affects the aggregation of gel particles. Hence,
it affects the preparation of TiO2 suspension solution.
If the concentration of chlorine ion is too high, the sta-
bility of sol becomes lower, and white suspension will
be obtained instead of transparent solution.
To obtain stable TiO2 sol, chlorine ions must be re-
moved. AgNO3 titration was used to determine the con-
centration of Cl− in the sol. The chlorine concentration
should be less than 0.001 M to have stable sol, in ac-
cordance with that reported in the literature [13–15].
3.3. Degree of Crystallinity
3.3.1. Effects of pH. XRD patterns show that TiO2 sol
prepared at pH 8 have a higher crystallinity (Fig. 4). The
diffraction plot can be observed from TEM (Fig. 13).
The more orderly the atoms arrangement is, the higher
the degree of crystallization is, and the sharper the SAD
pattern is. The SAD pattern in Fig. 13 shows that the de-
gree of crystallinity of TiO2 decreased with increasing
pH value.
3.3.2. Effects of [H+ ]/[Ti] Ratio. Comparing the
XRD patterns and the SAD patterns of TiO2 prepared at
 Preparation of TiO2 Sol Using TiCl4 as a Precursor 131

Figure 12. TEM images of TiO2 at various solid contents. (A) pH 8, 10 wt% TiO2 ; (B) pH 10, 7 wt% TiO2 ; (C) pH 10, 10 wt% TiO2 .
132 Lee and Liu

Figure 13. TEM images of TiO2 . (A) The sample prepared at pH 8; (A1) TEM SAD pattern; (B) The sample prepared at pH 12; (B1) TEM
SAD pattern.
 Preparation of TiO2 Sol Using TiCl4 as a Precursor 133

[H+ ]/[Ti] ratio of 1 and 1.5 shows that the better crys-
tallinity was obtained at [H+ ]/[Ti] ratio of 1.5. How-
ever, due to its less stability of sol, it is not the best
condition to prepared TiO2 sol. Therefore, [H+ ]/[Ti]
ratio of 1 is the best condition to prepare TiO2 sol.
3.4. Characteristics of TiO2 Particle
The experimental results of this study are summary
in Tables 7 and 8. The sample prepared at pH of
8, [H+ ]/[Ti] ratio of 1 and 1 M HCl had the small-
est particle size. This sample was used for further
characterzation.
3.4.1. TGA. Table 9 shows that increasing tempera-
ture from 50◦ C to 600◦ C can reduce weight to 22%.
In 100◦ C–200◦ C temperature range, weight loss was
about 12%, primarily due to the removal of physically
adsorbed water. The weight loss at 200◦ C–350◦ C range
was about 5%, mainly caused by the removal of acidic

Table 7. Summary of the experimental results.

pH [H+ /Ti] ratio Con. of HCl Stored temp. State HPC TiO2 solid content (wt%)

2, 3 1 1M Can not form sol 5%

8–12 1 1M 5◦ C Stable suspension for 3 months No 5%
8–12 1 1M 30◦ C Stable suspension for 3 months No 5%
8–12 1 1M 5◦ C Stable suspension for 3 months Yes 7%
8–12 1 1M 30◦ C Stable suspension for 3 months Yes 7%
2, 3 0.6 1M Can not prepare 5%
8–12 0.6 1M Unable crystal growth completely 5%
2, 3 0.8 1M Can not prepare 5%
8–12 0.8 1M Unable crystal growth completely 5%
2, 3 1.5 1M Can not prepare 5%
8–12 1.5 1M 5◦ C Stable suspension for 44 h No 5%
8–12 1.5 1M 30◦ C Stable suspension for 3 days No 5%
2, 3 1 2M Can not prepare 5%
8–12 1 2M 5◦ C Stable suspension for 14 days No 5%
8–12 1 2M 30◦ C Stable suspension for 8 days No 5%
2, 3 1 4M Can not form sol 5%
8–12 1 4M 5◦ C Stable suspension for 11 days No 5%
8–12 1 4M 30◦ C Stable suspension for 5 days No 5%
8 1 35% 5◦ C Stable suspension for 30 h Yes 10%
8 1 35% 30◦ C Stable suspension for 26 h Yes 10%

Table 8. Particle size of TiO2 in the sol.

Particles size distribution

pH [H+ /Ti] ratio Con. of HCl Stored temp Con. of TiO2 XRD TEM

8–9 1 1M 5◦ C 7% 7.8–7.9 nm Major axis = 20–25 nm

Minor axis = 5–8 nm
10 1 1M 5◦ C 7% 8.1 nm Major axis = 25–30 nm
Minor axis = 8–10 nm
11–12 1 1M 5◦ C 7% 8.3–8.7 nm Major axis = 40–50 nm
Minor axis = 10–15 nm
8 1 35% 5◦ C 10% 8.6 nm Major axis = 30–40 nm
Minor axis = 10–12 nm
134 Lee and Liu

Table 9. Weight loss of TiO2 powder prepared at various pH

values.

Weight loss

Temperature range 50–100◦ C 100–200◦ C 200–350◦ C

pH 8 4% 12% 5%
pH 9 5% 11% 6%
pH 10 5% 10% 7%

Figure 14. TGA spectra of TiO2 prepared at various pH values.

Figure 15. DSC spectra of TiO2 prepared at pH 8 and pH 12.

HCl. There was no apparent weight loss after 400◦ C.
As shown in Fig. 14, the sample prepared at pH 8 had a
lower weight loss than those prepared at pH 9 and 10.
However, the difference was insignificant.
3.4.2. DSC. Figure 15 shows the DSC spectra of
TiO2 . A broad endothermic peak appeared at 134◦ C,
which is attributed to the removal of physically ad-
sorbed water or chemically unbounded water with TiO2
in the pore. A small endothermic peak appeared at
about 340◦ C, which is caused by the removal of residual
acidic ion on TiO2 surface. There was no endothermic
or exothermic peak after 400◦ C.
Both TGA and DSC results show that at 100–200◦ C
there is structured water formed. At 300◦ C, there is a
removal of acidic ion.
It has been reported [20] hat heating up to 600◦ C, the
anatase crystal will convert to rutile. The XRD results
shown in Fig. 16 demonstrate the same phenomena.
3.4.3. N2 Sorption. Figure 17 shows that the as-
synthesized sample prepared at the best condition af-
ter room temperature drying had a surface area of
141 cm2 /g, and the pore was micropore. The sam-
ple was then heated at temperature 150◦ C, 200◦ C and
300◦ C, respectively. Figure 18 shows that after heating
at 100–300◦ C the pore still retain micropores, and the
surface area increased to 187 cm2 /g (heated at 150◦ C)
and 191 cm2 /g (heated at 200◦ C). After heating at
300◦ C, the microporous structure was collapsed and
the pore became mesopore.
3.4.4. Ultraviolet/Visible Adsorption (UV-Vis). The
samples prepared in this study showed very strong UV
absorption band at 361 nm. In addition, they are fully
transmitted in the range of visible light.
3.4.5. Contact Angle. Deionized water was dropped
on the glass that was coated with TiO2 thin film for the
contact angle measurement. 40 µl and 140 µl deion-
ized water and 20 minutes UV illumination (15 W,
wavelength 360 nm) were used for contact angle
measurements. The results are shown in Tables 10
and 11.
The results show that if the glass substrate was not
washed thoroughly, they might have dirty materials on
them, resulting in high degree of the contact angle.
 Preparation of TiO2 Sol Using TiCl4 as a Precursor 135

Figure 16. (A) TiO2 was calcined at temperature 450◦ C. Sample

was prepared at pH 8, 1 M HCl, [H+ /Ti] = 1. (B) TiO2 was cal-
cined at temperature 650◦ C. Sample was prepared at pH 8, 1 M HCl,
[H+ /Ti] = 1.
Figure 18. N2 sorption isotherms of TiO2 after calcination at vari-
ous temperatures.

The substrate without TiO2 thin film would adsorb

water molecules with its • OH radicals on the surface
[1, 16]. The contact angle would be less than 10 de-
gree. The substrates with TiO2 thin film would have

Table 10. Contact angle of glass with TiO2 thin film

(5 wt% TiO2 coating sol).

Contact angle (degree)

Number of
coating 40 µl water 140 µl water

1 3.12 5.13
2 2.67 3.44
3 2.19 1.98
4 <1 1.91
5 <1 <1
0a 11.23 16.53
Figure 17. N2 sorption isotherm of as-synthesized TiO2 prepared
at pH 8 and was dried at room temperature. a Glass without TiO2 coating.
136 Lee and Liu
Table 11. Contact angle of glass with TiO2 thin film
(7 wt% TiO2 coating sol).
Contact angle (degree)
Number of
coatings 40 µl water 140 µl water
1 2.37 4.08
2 1.97 2.27
3 <1 <1
4 <1 <1
5 <1 <1
smaller contact angle. The contact angle decreased with
an increase of the number of coating. After coating to
a certain amounts, the degree of contact angle would
not decrease any more. The amount of deionized water
and illuminating time by UV also affected the degree of
contact angle. The larger the amount of the water, the
larger is the degree the contact angle. The glass with
TiO2 thin film prepared in this study had a strong hy-
drophilic surface after being illuminated by UV light.
4. Conclusion
A titanium dioxide sol with narrow particle size distri-
bution was synthesized using TiCl4 as the starting mate-
rial. The sol was prepared by a process where HCl was
added to a gel of hydrated titanium oxide to dissolve it.
The resulting aqueous titanic acid solution is heated to
form titanium dioxide sol. The sol is excellent in dis-
persibility and stability. It is excellent in absorbance of
UV light and in transmittance of visible light. The pri-
mary particles of TiO2 were rhombus with major axis
of 20 nm and minor axis of 5 nm. It was anatase.
The concentration of acid, heating temperature and
heating time could affect the morphology of TiO2 par-
ticles. The higher the concentration of acid, the shorter
is the time needed for condensation reaction and crys-
tal growth. The smallest TiO2 crystal was obtained at
concentration of HCl of 1 M and [H+ ]/[Ti] ratio of 1.
The transparent TiO2 thin film on substrate has a very
strong hydrophilic property after being illuminated by
UV light.
References
1. R. Wang, K. Hashimoto, A. Fujishima, M. Chikuni, E. Kojima,
A. Kitamura, M. Shimohigoshi, and T. Watanabe, Nature 388,
431 (1997).
2. G. Sato, Y. Arima, H. Tanaka, and S. Hiraoka, US Patent
5,403,513 (1995).
3. M. Sakamoto, H.O. Yokkaichi, S.K. Suzuka, and Y.Y. Yokkaichi,
US Patent 4,880,703 (1991).
4. H. Takahashi, A. Sakai, and M. Hattori, US Patent 5,536,448
(1996).
5. H.K. Park, D.K. Kim, and C.H. Kim, J. Am. Ceram. Soc. 80,
743 (1997).
6. S.D. Park, Y.H. Cho, W.W. Kim, and S.J. Kim, J. Solid State
Chem. 146, 230 (1999).
7. R.W. Lange and H.G. Sowman, US Patent 4,166,147 (1979).
8. A. Yasumori, K. Ishizu, S. Hayashi, and K. Okada, J. Mater.
Chem. 8, 2521 (1998).
9. G.L. Li and G.H. Wang, NanoStrustured Materials 11, 663
(1999).
10. G.L. Li and G.H. Wang, J. Mater. Res. 14, 3346 (1999).
11. M. Anpo, Stud. Surf. Sci. Catal. 130, 157 (2000).
12. B.L. Bischoff and M.A. Anderson, Chem. Mater. 7, 1772 (1995).
13. P. Chrysicopoulou, D. Davazoglou, C. Trapalis, and G. Kordas,
Thin Solid Films 323, 188 (1998).
14. H. Cheng, J. Ma, Z. Zhao, and L. Qi, Chem. Mater. 7, 663 (1995).
15. L. Elfenthal, E. Klein, and F. Rosendahl, US Patent 5,215,580
(1993).
16. D.L. Foulger, P.G. Necini, and S. Poeri, US Patent 5,630,995
(1997).
17. A.J. Haddow, US Patent 5,599,519 (1997).
18. D.H. Kim and M.A. Anderson, Environ. Sci. Technol. 28, 479
(1994).
19. R.J. Kostelnik and F.C. Wen, US Patent 5,746,819 (1998).
20. H.D. Nan, B.H. Lee, S.J. Kim, C.H. Jung, J.H. Lee, and S. Park,
Japanese J. App. Phy. 37, 4603 (1998).
21. P.M. Oliver, G.W. Waston, K.E. Toby, and S.C. Parker, J. Mater.
Chem. 7, 563 (1997).
22. I. Sopyan, H. Watanabe, K. Murasawa, K. Hashimoto, and A.
Fujishima, J. Photo A. 79 (1996).
23. M. Tunashima, K. Muraoka, K. Yamamoto, M. Mikami, and S.
Sasaki, US Patent 6,113,873 (2000).
24. O. Yoshihisa, H. Kazuhito, and A. Fujishima, J. Phys. Chem. A
101, 8057 (1997).