The most popular choice of photocatalyst is titanium dioxide (TiO2) and much of the published work on photocatalysis uses this material. TiO2 is a promising photocatalyst due to its easy access, low cost, photoactivity, high stability, non-toxicity, hydrophilicity and high refractive index. It can be used for a variety of applications. Recent applications of TiO2 have involved self-cleaning, anti-bacterial, air purification and water treatment.
TiO2 exists in different phases such as amorphous, anatase, brookite
and rutile. It has been generally accepted that amorphous TiO2 has
no or negligible photocatalytic activity. Crystal structure of TiO2
is known to exist in a stable phase called rutile and two metastable phases
called anatase and brookite. Anatase and brookite are irreversibly transformed
into rutile. However, phase equilibrium does not exist for the transformation
and therefore, no specific temperature can be assigned to the phase transition.
Although the rutile phase has a wide variety of applications primarily in the
pigment industry, most work has shown that the anatase phase with a band gap
of 3.2 eV is usually preferred because it has proven to be the most active crystal
structure that possesses higher levels of photoactivity due to its favourable
energy band positions and high surface area (Nishimoto et
Addition of dopants such as iron (Fe) into the lattice sites of TiO2
could improve some of its properties for instance it can reduce the bandgap
energy of TiO2 to encourage them to absorb light at higher wavelength
for indoor applications (Zhang and Lei, 2008). Other
than that, dopants can help to promote the phase transition from anatase to
rutile hence reduce the phase transition temperature (Zhang
and Lei, 2008; Ghosh et al., 2001). It is
also proven in this study that addition of Fe dopant can promote the phase transition
from amorphous to anatase. It is vital to produce crystalline TiO2
before they can be used for applications because amorphous TiO2 is
not photocatalytically active. Apart from that, although anatase phase is preferred
over rutile phase for applications, it is also widely accepted that mixed phase
of TiO2 results in better photocatalytic activity than just anatase
phase alone. It is therefore important to understand the role played by the
dopant in the TiO2 lattice sites.
A glance through the literature reveals that the synthesis of TiO2
has been widely studied using various approaches such as pulsed laser deposition
(Paily et al., 2002), diffusion flame reactor
(Jang and Kim, 2001), water in oil emulsion (Mori
et al., 2001), hydrothermal synthesis (Wu
et al., 1999; Andersson et al., 2007)
and metal organic chemical vapour deposition (MOCVD) (Li
et al., 2002). Compared with the other methods, MOCVD is a promising
technique for catalyst manufacture due to the ease and simplicity of the process.
All the traditional steps in catalyst preparation such as drying, reduction,
centrifuge, or hydrothermal processing which critically affect catalyst performance
can be eliminated. Another advantage of MOCVD technique is that dopants can
be easily introduced into the reactor either through a solid source, separated
from precursor, or mixed in with the precursor. Thus, it is possible to produce
doped TiO2 via a one step MOCVD process. Furthermore, the control
of size distribution is easily accomplished by simply controlling the deposition
temperature and the flow rate of the carrier gas.
The aim of this study was specifically to investigate the effect of Fe doping
on phase transition in TiO2 systems. Furthermore, the effect of Fe
doping on the size of nanoparticles produced was also studied. In this work,
undoped and Fe-doped TiO2 nanoparticles were produced at 400 and
700°C deposition temperatures with different concentrations of Fe. To the
best of our knowledge, very limited research has been published regarding the
effect of Fe doping on the TiO2 phase transition from amorphous to
anatase. Only one study has been done to study the effect of Fe on the phase
transition from anatase to rutile of Fe-doped TiO2 produced using
MOCVD (Zhang and Lei, 2008). However that particular
study only investigated the effect of deposition temperature on the crystal
structure of Fe-doped TiO2 and not the effect of different concentrations
of Fe on the crystal structure of Fe-doped TiO2. This study will
hopefully provide a better understanding of the amorphous to anatase and anatase
to rutile phase transitions in TiO2 as well as the role of Fe in
MATERIALS AND METHODS
Synthesis of undoped and Fe-doped TiO2 nanoparticles: The MOCVD setup consists of stainless steel gas flow lines, mass flow controllers, a bubbler and a quartz tube (52 mm o.d. and 800 mm long) in a split tube furnace as the reactor. The heating zone was 300 mm long. The precursor, titanium (IV) butoxide (TBOT) obtained from Aldrich was used as received. The TBOT precursor was stored in a stainless steel bubbler located on a hot plate and maintained at 175°C. To produce different Fe dopant concentration of TiO2 nanoparticles, different amount of Fe precursor which was ferrocene (0.005, 0.01, 0.03 and 0.05 g) was mixed directly with 20 mL TBOT precursor inside the bubbler. The quartz tube was purged by nitrogen (400 mL min-1) during the initial heating step towards reaching the desired deposition temperature (400 and 700°C). Once the reaction temperature was achieved, the precursor was introduced into the quartz tube using nitrogen as the carrier gas (400 mL min-1) along with an oxygen feed (100 mL min-1). Undoped and Fe-doped TiO2 nanoparticles would be thermophoretically deposited inside the quartz tube upon thermal decomposition of the precursor gas. After 3 h of reaction time, the flow of the precursor gas and the flow of the oxygen gas inside the reactor were stopped by switching off the carrier gas and oxygen gas valves. The reactor was allowed to cool to room temperature under 400 mL min-1 nitrogen flow. The nanoparticles were then collected from the quartz tube wall using a spatula and kept in a small container.
Characterization method: The morphology of undoped and Fe-doped TiO2 nanoparticle samples was studied by Transmission Electron Microscope (TEM) using LEO 912AB Energy Filter TEM. The nanoparticle size in terms of diameter was determined from TEM micrographs using the ImageJ software. The error or tolerance in nanoparticle size determination using the software is ±0.005 nm. The nanoparticle size distribution was presented in histograms that were fitted using Gaussian distribution. The phase of the samples were determined by x-ray diffraction (XRD) using Philips X'pert Pro PW3040 with a Cu Kα radiation source (λ= 0.15406 nm) operated at 40 kV and 30 mA. Energy dispersive x-ray (EDX) analysis via LEO 1455 VP SEM was carried out to determine the elemental composition of the samples and to prove the presence of Fe in the TiO2 crystal structure.
RESULTS AND DISCUSSION
Undoped TiO2 nanoparticles: Figure 1a and
b compare the TEM micrographs of undoped TiO2 nanoparticle
samples deposited at 400 and 700°C. The respective size distribution histograms
of the nanoparticles in terms of diameter are also shown. Mean nanoparticle
diameter and standard deviations are also included. The micrographs show that
the nominal diameter of the nanoparticles were less than 50 nm and that the
nanoparticles appeared to be homogeneous and uniform in size. From Fig.
1, it is clear that increasing deposition temperature has a direct effect
on the nanoparticle size. The mean nanoparticle diameter decreased from 35.75
to 10.03 nm when the deposition temperature was increased from 400 to 700°C.
The decrease in nanoparticle diameter as the temperature was increased can
be attributed to the improvement of nucleation rate and/or suppression of growth
(Shi et al., 2000). If there is an increment
in the nucleation rate, a greater percentage of the precursor vapor can be used
to form small particles quickly which would lead to less vapor being available
for condensation and further growth (Nolan et al.,
2006). According to Nakaso et al. (2003),
at lower reactor temperatures only small amounts of the precursor reacts to
||TEM micrographs (left) and respective nanoparticles size distribution
histograms (right) of undoped TiO2 nanoparticle samples deposited
by MOCVD at (a) 400°C and (b) 700°C. The histograms were fitted
using Gaussian fittings
The unreacted precursor condenses out forming partially oxidized Ti that would
become TiO2 upon thermal decomposition (Nakaso
et al., 2003). This results in large and amorphous particles at lower
reaction temperatures. Furthermore, Fig. 1 demonstrates that
the increase in temperature causes a decrease in size distribution. This shows
that a higher deposition temperature generates smaller nanoparticles with a
narrower size distribution.
The crystal structure of undoped TiO2 nanoparticle samples obtained
by XRD is shown in Fig. 2. The XRD pattern for the sample
deposited at 400°C showed no detectable peaks, indicating that the sample
was amorphous. The peaks at 2θ values of 25.3, 37.8, 48.0, 53.8 and 54.9
for the samples deposited at temperature 700°C correspond to the diffractions
of the (1 0 1) (0 0 4) (2 0 0) (1 0 5) (2 1 1) planes of anatase respectively
(Li et al., 2002; Zhang et
al., 2006). These peaks confirmed that undoped TiO2 nanoparticle
samples deposited at 700°C were in the anatase phase. From the XRD data,
there were no other detectable peaks to suggest the presence of rutile phase.
||XRD patterns of undoped TiO2 nanoparticle samples
produced by MOCVD (a) 400°C amorphous and (b) 700°C anatase. A:
Figure 3a and b compare the EDX spectra
of the undoped and Fe-doped TiO2 nanoparticles (0.03 g Fe). Peaks
around 0.4, 4.5 and 5.0 keV correspond to TiO2 in the sample. The
peaks due to iron are clearly distinct in Fe-doped TiO2 (Fig.
3b) at around 0.7, 6.4 and 7.0 keV thus proving the presence of Fe in the
crystal structure of TiO2. Apart from emission peaks of Ti, O and
Fe, the peaks of C and Au were also observed most probably due to the carbon
in the TBOT precursor and from the gold coating during sample preparation for
EDX analysis, respectively.
||EDX spectra of undoped and Fe-doped TiO2 (0.03
g Fe) nanoparticle samples produced by MOCVD at 700°C
||Effect of amount of Fe dopant on the percentage of anatase
and rutile crystal structure and average nanoparticle size of sample deposited
|aDetermined using EDX, bCalculated using
Eq. 1, cCalculated using Eq.
The elemental composition of the samples was also determined via EDX analysis.
Fe/Ti ratio (at. %) values are tabulated in Table 1. As expected,
it can be seen that Fe/Ti ratio increased with increasing amount of Fe dopant.
Figure 4a-d compare the TEM micrographs of Fe-doped TiO2
nanoparticle samples deposited at 400 and 700°C with different Fe dopant
concentrations. The respective size distribution histograms of the nanoparticles
in terms of diameter are also shown. The micrographs show that the nominal diameter
of the nanoparticles was still less than 50 nm and that the nanoparticles appeared
to be homogeneous and uniform in size. From Fig. 4, it can
be seen that the effect of deposition temperature on nanoparticle size of Fe-doped
TiO2 samples is similar to undoped samples. Furthermore, Fig.
4 also shows that the nanoparticle diameter of samples deposited at a certain
temperature is about the same no matter how much Fe dopant was introduced. However,
the histograms demonstrated that the mean nanoparticle diameter decreased with
an increase of the Fe dopant concentration. This is probably due to the fact
that Fe doping slowers the growth of the TiO2 (Naeem
and Ouyang, 2010). This finding is also supported by Zhou
et al. (2005). Nonetheless, the difference is relatively small.
The crystal structure of undoped and Fe-doped TiO2 nanoparticle samples obtained by XRD are shown in Fig. 5. No characteristic peaks of iron oxide phases appeared in the XRD spectra for all the samples. This was probably attributed to the fact that the Fe dopant concentration was very low and hence could not be detected by XRD.
From Fig. 5a, the XRD pattern for the undoped sample deposited
at 400°C showed no detectable peaks indicating that the sample was amorphous.
However, as the Fe dopant concentration was increased, the peaks at 2θ
values of 25.3, 37.8, 48.0, 53.8 and 54.9 started to appear which correspond
to the diffractions of the (1 0 1) (0 0 4) (2 0 0) (1 0 5) (2 1 1) planes of
anatase respectively (Li et al., 2002; Zhang
et al., 2006). It is known that from the effective radius of ions
for a coordination number of 6, Fe3+ has an ionic radius (0.645Å)
comparable to Ti4+ (0.605 Å) (Shannon,
1976). Therefore, Fe3+ can easily substitute Ti4+
in the TiO2 lattice sites and distort the crystal structure of the
host compound. The oxidation state of Fe in the samples is currently unknown
but can be determined from x-ray photoelectron spectroscopy (XPS). However,
the transition from amorphous to anatase crystal structure as can be seen from
Fig. 5a seems to suggest that Fe dopant does indeed induce
crystallization. This finding is consistent with the finding of Zhang
and Reller (2002) who used sol-gel technique to prepare homogeneously doped
TiO2. Their study demonstrated that Fe doping could cause phase transition
from amorphous to anatase at lower calcination temperature. Furthermore, the
peak intensity corresponding to the anatase crystal structure increased as the
Fe concentration was increased. This indicated that increasing the Fe dopant
concentration can promote the formation of anatase phase. Naeem
and Ouyang (2010) established the same trend of results for lower Fe dopant
Figure 5b shows that undoped samples deposited at 700°C
were in the anatase phase. Nonetheless, as the Fe dopant was introduced into
the TiO2 lattice, new apparent peaks started to appear. These new
peaks at 2θ values of 27.4, 36.1 and 41.23 corresponds to the diffractions
of the (1 1 0) (1 0 1) and (1 1 1) planes of rutile, respectively (Jitputti
et al., 2007). These peaks indicated that the sample had a mixture
of anatase and rutile crystal structure. The fractions of anatase and rutile
crystal structure were calculated according to the following equation (Yan
et al., 2005):
||TEM micrographs (left) and respective nanoparticles size distribution
histograms (right) of TiO2 nanoparticle samples deposited by
MOCVD at (a) 400°C-0.01 g Fe dopant, Mean = 32.92 nm, STD = 6.84 (b)
400°C-0.05 g Fe dopant, Mean = 30.53 nm, STD = 7.56 (c) 700°C-0.01
g Fe dopant, Mean = 9.98 nm, STD = 1.63 and (d) 700°C -0.05 g Fe dopant,
Mean = 9.71 nm, STD = 1.58. The histograms were fitted using Gaussian fittings
where, fA is the fraction of anatase crystal structure in the samples,
IA and IR are the X-ray intensities of the anatase (1
0 1) and rutile (1 1 0) diffraction peaks, respectively and K is a constant.
The percentage of anatase and rutile crystal structure was tabulated in Table
Table 1 illustrates that the percentage of rutile increases
with the increase in Fe dopant concentration. This finding is supported by Fig.
5b where there seems to be a decrease in the peak intensity for (1 0 1)
plane of anatase as the amount of Fe dopant increases. A decrease of the anatase
phase with the increase of the Fe dopant concentration compensates the increase
of rutile fraction (Alexandrescu et al., 2007).
||XRD patterns of undoped and Fe-doped TiO2 nanoparticle
samples produced by MOCVD (a) 400°C and (b) 700°C. A: Anatase, R:
Hence it is concluded that Fe doping into the TiO2 lattice can promote
the anatase to rutile phase transition. Although the oxidation state for Fe
in the samples are unknown, it is speculated that considering the ionic radius
of Ti4+ and Fe3+ is almost the same, Fe3+ could
replace the Ti4+ in lattice sites and therefore forming a solid solution
of Fe in the TiO2 matrix. Consequently, this causes oxygen vacancies
because of charge compensation and their movement would favor the rutile nucleation
in anatase particles (Zhang and Lei, 2008). Further
characterization of the samples using XPS would be able to support the findings.
Note that the anatase to rutile transition is a metastable to stable transformation.
However, it is generally accepted that phase equilibrium does not exist for
the anatase to rutile transformation and therefore, no specific temperature
can be assigned to the phase transition. The Gibbs free energy for the two phases
is as follows (Liao et al., 1997):
where, G is the Gibbs free energy and T is the temperature. The energy difference, ÄG between the anatase and the rutile phases at 700°C is 6674.56 J/mol, which is the driving force for rutile nucleation.
Shannon and Pask (1965) hypothesized that oxygen vacancies
reduces the energy that must be overcome before the rearrangement of the Ti-O
octahedra can occur to cause phase transition from anatase to rutile. Thus,
it is speculated that the mechanism responsible to promote the phase transition
from anatase to rutile is the oxygen vacancies due to Fe doping. Gosh et
al. (2001) also reported that the addition of Fe could promote the transition
of anatase to rutile.
Moreover, Fig. 5b shows that Fe doping has a direct albeit
small effect on the size of the Fe-doped TiO2 nanoparticles. It can
be seen from Fig. 5b that the greater the concentration of
Fe, the wider the width of anatase (1 0 1) diffraction peak and therefore the
smaller the grain size of Fe-doped TiO2 nanoparticles. The mean nanoparticle
diameter of the samples was calculated by applying the Debye-Scherrer formula
(Yan et al., 2005) on the anatase (1 0 1) diffraction
where, D is the mean nanoparticle diameter, K is a constant which is taken as 0.89 for spherical crystalline shape, λ is the wavelength of the X-ray radiation (Cu Kα = 0.15406 nm), β is the corrected band broadening at full width at half-maximum (fwhm) and θ is the diffraction angle.
The results are also tabulated in Table 1. The results demonstrate
that the mean nanoparticle diameter decreased with an increase of the Fe dopant
concentration although the difference is very small (.2 nm). This finding seems
to support the TEM results that were discussed earlier. Regardless of the deviation
in the values calculated using Debye Scherrer formula, the results are consistent
with the values obtained from the TEM micrographs. Note that the Debye-Scherrer
formula is not accurate at estimating nanoparticles size because the assumption
of an underlying crystal structure (translational symmetry) is often invalid
(Hall et al., 2000).
In this study, undoped and Fe-doped TiO2 nanoparticles were successfully
synthesized via a one step MOCVD method at 400 and 700°C deposition temperatures.
Major effects of Fe doping found in this study are: 1) promote the phase transition
from amorphous to anatase and from anatase to rutile 2) nanoparticle size decreases
with an increase in Fe dopant concentration. TEM results disclosed that in general,
the average nanoparticle size decreases with increasing temperature. XRD data
revealed that the undoped TiO2 nanoparticle sample deposited at 400°C
was amorphous but the addition of Fe dopant induces crystallization of the sample.
Increasing the amount of Fe dopant promoted the phase transition from amorphous
to anatase phase. Meanwhile, the undoped TiO2 nanoparticle sample
deposited at 700°C was in anatase crystal structure. Fe doping causes phase
transition from anatase to rutile crystal structure. Increasing the amount of
Fe dopant reduces the mean nanoparticle size and reduces the anatase fraction.
Fe doping also promoted the phase transition from anatase to rutile due to the
oxygen vacancies in TiO2 lattice sites.
This study was financially supported by Fundamental Research Grant Scheme, University Putra Malaysia (Grant No. 5523426).