Synthesis and photocatalytic oxidation properties of iron doped titanium dioxide nanosemiconductor particles

Article abstract of DOI:10.1039/b301998e

The structure, physical characteristics and selective photocatalytic oxidation properties of quantum confined nanosize iron doped TiO₂ (Q-TiO₂/Fe³⁺) particles were studied by TG-DSC, XRD, DRS, EPR and Selective Oxidation Photocatalytic Measurements. It is shown that the solubility of iron in the obtained Q-TiO₂/Fe³⁺ nanoparticles is 1.0 atom% and the iron doping level has a great influence on the transformation of anatase to rutile (A-R). The quantum confined effect was observed for Q-TiO₂/Fe³⁺ nanoparticles. All of the samples have EPR Bulk-Fe³⁺ and Surf-Fe³⁺ signals, which are located in the bulk and surface of TiO₂ nanoparticles respectively. Quantitative EPR results indicate that the relative EPR intensity of these paramagnetic centers shows regular change with varying corresponding iron modification level. In situ EPR indicates that the photogenerated charge carrier (h⁺, e^-) could be trapped by different Fe ³⁺ sites simultaneously, i.e., trapping of h⁺ is due to Surf-Fe³⁺ sites at g = 4.30, whereas that of e^- is attributed to Bulk-Fe³⁺ sites at g = 1.99. Selective photocatalytic oxidation of cyclohexane into cyclohexanol with higher selectivity has been obtained by molecular oxygen activation over Q-TiO₂/Fe³⁺ nanoparticles under mild conditions. It is thought that the optical surface state of Q-TiO₂/Fe³⁺ nanoparticles play a key role in the selective photocatalytic oxidations.

Full text of DOI:10.1039/b301998e

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis and photocatalytic oxidation properties of iron doped
titanium dioxide nanosemiconductor particles
ab
Xinyong Li,* Po-Lock Yue and Charles Kutal
b
c
a
School of Environmental Science & Engineering, Dalian University of Technology,
Dalian, 116023, China
Department of Chemical Engineering, The Hong Kong University of Science & Technology,
Clear Water Bay, Kowloon, Hong Kong
Department of Chemistry, University of Georgia, Athens, GA 30602, U. S. A
b
c
Received (in Toulouse, France) 19th February 2003, Accepted 14th May 2003
First published as an Advance Article on the web 7th July 2003
The structure, physical characteristics and selective photocatalytic oxidation properties of quantum confined
3
+
nanosize iron doped TiO
Oxidation Photocatalytic Measurements. It is shown that the solubility of iron in the obtained Q-TiO
nanoparticles is 1.0 atom% and the iron doping level has a great influence on the transformation of anatase to
2 2
(Q-TiO /Fe ) particles were studied by TG-DSC, XRD, DRS, EPR and Selective
3
+
2
/Fe
3
+
rutile (A-R). The quantum confined effect was observed for Q-TiO /Fe nanoparticles. All of the samples
2
3
+
3+
2
have EPR Bulk-Fe and Surf-Fe signals, which are located in the bulk and surface of TiO nanoparticles
respectively. Quantitative EPR results indicate that the relative EPR intensity of these paramagnetic centers
shows regular change with varying corresponding iron modification level. In situ EPR indicates that the photo-
+
generated charge carrier (h , e ) could be trapped by different Fe sites simultaneously, i.e., trapping of h is
ꢀ
3+
+
due to Surf-Fe sites at g ¼ 4.30, whereas that of eꢀ is attributed to Bulk-Fe sites at g ¼ 1.99. Selective
3
+
3+
photocatalytic oxidation of cyclohexane into cyclohexanol with higher selectivity has been obtained by
molecular oxygen activation over Q-TiO
optical surface state of Q-TiO
3
/Fe nanoparticles under mild conditions. It is thought that the
+
/Fe nanoparticles play a key role in the selective photocatalytic oxidations.
+
2
3
2
to our knowledge, selective photocatalytic oxidation of alkanes
by using iron doped TiO as photocatalyst with molecular oxy-
Introduction
2
Photocatalysis has received enormous attention in recent years
because of its potential application in environmental treatment
gen as oxidant under mild conditions have not been reported.
Here, this paper deals with the preparation and characteriza-
tion and selective photocatalytic oxidation properties of quan-
1
–9
and fine chemical synthesis. It is known that under aqueous
conditions, organic contaminants in wasted water can be
degraded into CO and H O etc., whereas under dry organic
3
+
tum confined iron doped TiO (Q-TiO /Fe ) nanoparticles. It
2
2
2
2
is shown that selective photocatalytic oxidation of cyclohexane
into cyclohexanol with higher selectivity has been achieved
solvent conditions, the organic substrate may be selectively
oxidized into fine chemical products rather than complete
3
with molecular oxygen over Q-TiO /Fe nanoparticles under
+
2
1
0–12
mineralization.
For the latter case, it may be noticed that
visible irradiation and mild conditions.
C–H bond activation leading to catalytic oxidation of hydro-
carbons is one of the most challenging chemical problems, in
addition to being of great practical importance. In terms of
Experimental
2
the oxidant, molecular oxygen (O ) is clearly the most desir-
Materials
able due to its ready availability and high reduction potential.
Up to now, there has been some progress in the field of
oxidation of alkanes under mild conditions by molecular
3+
Q-TiO
2
/Fe
nanoparticles were prepared by the Sol–Gel
method and were well characterized. The purity of the cyclo-
hexane, cyclohexanol, and cyclohexanone was checked by
gas chromatography. When needed, cyclohexane and aceto-
nitrile were dried respectively over metallic sodium and CaH
or CaCl2
1
3,14
oxygen.
Recently, photocatalytic selective oxidation of
cyclohexane into cyclohexanol with higher selectivity has been
performed by molecular oxygen under UV irradiation and
2
1
5
2
mild conditions by using nano-TiO as photocatalyst. How-
ever, in terms of solar energy utilization, there are still some
shortcomings for the above-mentioned processes, such as low
quantum efficiency and poor photosensitivity of TiO to visible
3+
Q-TiO /Fe nanoparticle preparation
2
2
3
Q-TiO /Fe nanoparticles were prepared by an acid-cata-
+
light. For nanosemiconductor photocatalysts, in order to
increase the quantum yield and the spectra sensitivities, the
2
lyzed Sol–Gel method and subsequently extracted from the
solvent by means of a supercritical CO drying strategy
ꢀ
+
e /h recombination has to be reduced and the band gap
has to be decreased. An effective strategy for achieving this
2
ꢁ
3+
followed by annealing at 400 C for 2 h. The Q-TiO /Fe
2
goal is to introduce some defects into TiO
2
lattice through
nanoparticles had nominal concentration of 0.05, 0.5, 1, 2
and 5 atom% of iron with respect to the molar percentage of
titania (leading to samples denoted TF0.05, TF0.5, TF1,
TF2 and TF5). The iron doped TiO2 sol was prepared as
selective metal ion doping. Up to now, a number of dopants
have been found to be effective to increase the quantum effi-
ciencies of photocatalytic mineralization processes. However,
1
264
New J. Chem., 2003, 27, 1264–1269
DOI: 10.1039/b301998e
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2003

View Article Online
follows: A definite amount of titanium n-butoxide (A. R.
Aldrich) was added dropwise to the solution of 30 ml anhy-
drous ethanol at 0 C under vigorous stirring. Similarly, calcu-
lated iron nitrate (A. R. Aldrich) was dissolved in 30 ml
anhydrous ethanol and it was mixed with the ethanolic solu-
tion of titanium n-butoxide dropwise under vigorous constant
stirring. The mixed sols were kept stirring for 24 h and an aqu-
capacity) at room temperature. Illuminations were provided
by a 150 W Xe lamp. A water filter and 400 nm cut off filter
were used to remove ultraviolet and infrared components
ꢁ
2
respectively. The irradiated cross-section was 6 cm . Q-TiO
2
/
3
+
Fe nanosamples (50 mg), cyclohexane (10 ml), acetonitrile
(10 ml) were placed in the reactor and stirred with a magnetic
stirrer. The oxygen gas was bubbled into the liquid phase at a
ꢀ1
eous solution of HNO
min into the well agitated system at room temperature. The
iron doped TiO aerogels were allowed to age in the mother
liquor for at least twice the gelation time. In supercritical car-
3
was added slowly over the course of 70
flow rate of 20 ml min . After 3 h, the dispersions were
sampled (1 mL), centrifuged, and subsequently filtered through
3
+
2
2
a Millipore filter (pore size, 0.22 mm) to separate Q-TiO /Fe
nanoparticles. The filtrates were then subjected to gas chroma-
tography using a Shimadzu GC-16A gas chromatograph
equipped with a flame ionization detector, using columns
packed with Carbowax 20 M 5% on Chromosorb W-AW.
The reaction products were determined by comparison of their
retention times with those of authentic samples. The selectivity
refers to the part that was oxidized and was calculated by inte-
gration of the three products (i.e., cyclohexanol, cyclohexa-
none and carbon dioxide) from the GC results.
1
6
bon dioxide drying, alcohol from the structure of the alcogel
ꢁ
was replaced by carbon dioxide at 40 C and 100 bar. The
extractor was completely filled with cold alcohol in order to
minimize evaporation of the solvent from the alcogel and to
avoid cracks developing during pressure build up. CO
pumped into the extractor to a pressure above its critical pres-
sure, that is, 100 bar. Before entering the extractor, CO was
2
was
2
heated to a temperature above its critical temperature, that
ꢁ
is, 40 C. During drying, the flow of CO
2
to the extractor
ꢀ
1
was held constant at 0.50 kg h and controlled independently
of pressure. In the separator, the solvent was separated from
carbon dioxide by expansion. When the solvent was comple-
tely replaced, the extractor pressure was slowly reduced to
ambient pressure. The product aerogel was ground to < 100
mesh and heated in a tube furnace in flowing oxygen at 473–
Results and discussion
3
+
The thermal behavior and crystal structure of Q-TiO
nanoparticles
2
/Fe
2
The thermal behavior of iron doped TiO aerogels was inves-
1023 K for 2 h, which was a standard calcination procedure.
The aerogels turn black after few minutes of annealing at
tigated with TG/DSC at temperatures ranging from room
ꢁ
temperature to 500 C. The TG/DSC patterns of 1 atom% iron
doped TiO typical aerogel are shown in Fig. 1. The TG-DSC
4
73–673 K and regain their white color after additional
3
+
annealing for 2 h at 673–1023 K which yields Q-TiO
nanoparticles with different modification levels.
2
/Fe
2
ꢁ
endothermic peak at around 100 C and the weight loss repre-
sents the dehydration and loss of residual solvent. The
endothermic peak at 275 C is attributed to the combustion
of organic substances contained in the aerogel. The exothermic
ꢁ
3
+
2
Characterization of Q-TiO /Fe nanoparticles
ꢁ
peak at 370 C is considered to be the formation of anatase.
ꢁ
The thermal decomposition behavior of the gel-type precursor
was monitored by a Dupont 9900 TG-DSC thermal analyzer.
The gel-type precursor remains amorphous up to 370 C,
which suggested that the crystallization is kinetically hindered
due presumably to the integrity of the porous network arising
3
+
2
The crystal structure of the Q-TiO /Fe photocatalysts was
characterized using X-ray diffraction (XRD). The X-ray dif-
fraction patterns were taken on a Rigaku/D/max g B diffrac-
tometer using Cu Ka radiation. To study the electronic
structure characteristics of nanoparticles, diffuse reflectance
spectra (DRS) were measured in the range of 300–700 nm
2
from the supercritical CO drying. The XRD patterns of
3
Q-TiO /Fe nanoparticles with modification levels ranging
+
2
from 0.05 to 5.0 atom% (TF0.05 to TF5) prepared at low
ꢁ
temperatures (450 C) are shown in Fig. 2. It can be seen that
3
+
2
only the pure anatase phase was detected for Q-TiO /Fe
using a HITACHI UV 200-10 Spectrophotometer. BaSO
was used as the reference. Electron paramagnetic resonance
EPR) was used to investigate the surface chemistry of Q-
4
nanoparticles with modification levels of 0.05, 0.5 and 1
atom% (TF0.05, TF0.5 and TF1). For samples with higher
modification levels (TF2 and TF5), besides the detected ana-
tase phase, two new phases were found. One is a rutile phase,
(
3
+
TiO /Fe nanoparticles and photoinduced redox processes
2
ꢀ
+
as well as interfacial charge carrier (e , h ) transfer properties.
3
+
EPR spectra of all Q-TiO /Fe nanoparticle samples were
the other can be ascribed to a-Fe O . The formation of an a-
2 3
2
2 3 2
Fe O phase may be attributed to the solubility of iron in TiO
performed on a Varian E-115 spectrometer operating in the
X-band (n ¼ 9.2 GHz) with 100 kHz field modulations at 77
K. EPR parameters were calibrated by comparison with a
1
7
matrix which is 1 atom%. More iron doping may induce the
dopant concentration to attain saturation in the inner surface
of the titania framework and is followed by producing an iso-
2
+
standard Mn /ZnS (106 line distance, 34.05 mT) and 2, 2-
1
5
lated iron oxide phase at the surface of the TiO nanoparticles.
2
On the other hand, the emergence of the rutile phase reveals
that doping favors the transformation of anatase to rutile
(A–R), which further supports the conclusion that transition
metal doping has a great influence on the A–R transformation
diphenyl-1-picryldrazyl (DPPH, 9.7 ꢂ 10 spins, g ¼ 2.0036).
For the in situ EPR study of photoinduced redox processes
and interfacial charge carrier (e , h ) transfer properties of
ꢀ
+
3
+
2
Q-TiO /Fe samples, a typical sample of TF1 was degassed
ꢀ
5
at room temperature (10 Torr) in a high vacuum cell fitted
with one quartz window at one side. Irradiation was carried
out with an oriel 150 W Xe lamp fitted with a high efficiency
parabolic reflector. The irradiation from the Xe lamp passed
through a water filter and a 400 nm cut off filter to remove
the ultraviolet and infrared components of the light and was
then focused onto the front face of the EPR sample cavity,
from which the cover plate had been removed.
Evaluation of photocatalytic selective oxidation cyclohexane
3
over Q-TiO /Fe nanoparticles
+
2
The photocatalytic selective oxidation runs were carried out
in a quartz reactor fitted with a gas inlet and outlet (50 ml
Fig. 1 The TG-DSC curve of iron doped TiO aerogel.
2
New J. Chem., 2003, 27, 1264–1269
1265

View Article Online
3+
Fig. 3 Diffuse reflectance spectra (DRS) of Q-TiO
cles: (a) p-25 TiO , (b) TF0.05, (c) TF0.5, (d) TF1, (e) TF2, (f) TF5, (g)
a-Fe
2
/Fe nanoparti-
2
2
3
O .
3+
Fig. 2 X-Ray diffraction patterns (XRD) of Q-TiO
ticles: (a) TF0.05, (b) TF0.5, (c) TF1, (d) TF2, (e) TF5.
2
/Fe nanopar-
2
response of iron doped TiO samples to visible light is very
important for utilizing sunlight in photocatalytic reaction
processes. For samples (TF2 and TF5) with higher iron
modification levels, a broad band centered at ca. 530 nm can
be observed. The spectra show typical features of minor
amounts of a-Fe O , which is in good agreement with our
1
8
temperature of nano-TiO
that higher transition metal modification levels could cause
the isolation of the iron oxide phase from TiO matrix, which
means more iron species may disperse on the surface of the
2
.
The XRD patterns also reveal
2
3
+
2
3
2 2
TiO matrix. The crystallite sizes of Q-TiO /Fe nanoparti-
XRD results.
1
9
cles as calculated from Scherrer’s formula are shown in
Table 1. It is found that the crystallite size in the present inves-
tigation is in the range of 8.5–12.5 nm.
The surface state chemistry properties and light induced redox
3
processes of Q-TiO /Fe nanoparticles
+
2
3
+
3
The electronic structure of Q-TiO /Fe nanoparticles
+
2
EPR measurements of Q-TiO /Fe nanoparticles with differ-
2
ent modification levels as listed in Table 1 reveal that all of
the samples have EPR signals. Typical EPR spectra are shown
in Fig. 4. It can be seen from Fig. 4 that the EPR spectra are
composed of two signals. EPR signal I at g ¼ 4.30 is attributed
Diffuse reflectance spectra (DRS) technique is a useful techni-
que to characterize the electronic structure properties of nano-
3
+
particles. The DRS spectra of Q-TiO
different modification level in comparison with p-25 TiO and
2
/Fe nanoparticles with
3
+
3+
2
to the Fe (Surf-Fe ) dispersed on the surface of the TiO
2
a-Fe
2
O
3
are shown in Fig. 3. It is easily seen that p-25 TiO
2
nanoparticles, whereas EPR signal II at g ¼ 1.99 can be
2
3
+
3+
(Bulk-Fe ) distributed in the bulk TiO
shows an absorption threshold at ca. 380 nm and an O !
assigned to Fe
2
4
Ti charge-transfer band with a maximum at around 325 nm.
+
21
matrix. The corresponding quantitative EPR results show
that the relative EPR intensity of these two signals exhibits a
smooth change with varying corresponding iron modification
level. The quantitative EPR results are shown in Fig. 5. For
3
+
4+
For Q-TiO /Fe nanoparticles, the O ! Ti charge-transfer
2
2
band can be observed at ca. 310 nm. The onset of the absorp-
tion band is shifted towards lower wavenumbers with increas-
ing modification level (460 nm for TF0.05 and 660 nm for
TF5), i.e., a red shift, which is attributed to the charge-transfer
transition between the d electrons of the transition metal and
the conduction band of the TiO2 semiconductors. This red
shift firmly attests the theory of the effect of quantum confine-
ment of the nanoparticles, which could be explained with the
3
+
all of the Q-TiO /Fe nanosamples, it may be seen from
2
3
Fig. 5 that the relative EPR intensity of Surf-Fe on the
+
3
+
TiO2 surface is much lower than that of Bulk-Fe in the
TiO2 matrix. While the iron modification level is less than
1 atom%, with increasing the iron content, the EPR intensity
3
+
of Bulk-Fe
gradually increases and reaches a maximum
2
0
Brus Theory. From Fig. 3, it can be found that the TF0.05
sample with low iron content shows a constant absorption in
value at 1 atom% iron modification level. This reveals that iron
species may diffuse into the crystal lattice of TiO2 to form
a homogeneous solid solution as indicated from the XRD
and DRS results. It may also reveal that, with increasing iron
2
the visible region higher than that of p-25 TiO . It may also
3
be noticed that all of the Q-TiO /Fe nanosamples show
+
2
3
+
enhanced absorption in the range of 400–670 nm, increasing
higher for higher iron doped samples and accompanying the
change of color from white to red brown. The corresponding
modification level, the relative EPR intensity of Surf-Fe
3+
2
Table 1 Characteristics of Q-TiO /Fe nanoparticles prepared by
the Sol–Gel method
D
p
T
cal
b
/
XRLB
D /
a
c
d
Sample (atom%)
K
nm
Phase
Color
TF0.05 0.05
673
673
673
673
673
11.3
12.5
11.3
11.3
8.5
A
A
A
White
Buff
Yellow brown
TF0.5
TF1
TF2
0.5
1.0
2.0
5.0
A + R + F Red brown
A + R + F Red brown
TF5
a
b
Iron doping level. Calcination temperature. Crystallite size calcu-
c
3+
d
Fig. 4 Typical electron paramagnetic spectra (77 K) of Q-TiO
2
/Fe
2 3
lated based on Scherrer’s formula. A: anatase R: rutile F: a-Fe O .
nanoparticles: (a) TF0.05 (b) TF0.5, (c) TF1, (d) TF2, (e) TF5.
1
266
New J. Chem., 2003, 27, 1264–1269

View Article Online
Fig. 7 In situ quantitative EPR characteristics of 1 atom% iron
doped TiO (TF1) nanoparticles with visible light irradiation time from
2
3
+
Fig. 5 The EPR characteristics of Q-TiO
different modification levels.
2
/Fe nanoparticles with
0 to 18 min at 77 K under vacuum conditions and after standing in the
dark at 77 K for another 17 min.
increases correspondingly. This means that more iron species
disperse on the surface of TiO nanoparticles and thus induce
the increase of the relative EPR intensity of Surf-Fe . On
the other hand, when the iron modification level is more than
+
be attributed to the trapping of charge carrier (h , e ) by
ꢀ
2
3
+
19b
3
+
different Fe
sites simultaneously,
for Q-TiO /Fe nanoparticles is considered to be due
i.e., the trapping of
+
3+
h
2
3
+
ꢀ
to Surf-Fe sites at g ¼ 4.30, whereas that of e trapping
1
.0 atom%, the corresponding quantitative EPR results reveal
3
+
3
+
3+
can be though to occur at g ¼ 1.99 of Bulk-Fe sites. The
that the EPR intensities of Surf-Fe and Bulk-Fe signals
decrease with further increasing iron modification level. This
may be caused by the agglomeration and separation of a-
+
fast separation of photogenerated charge carrier (h , e )
ꢀ
3
+
and subsequent trapping simultaneously at different Fe
sites favor the surface–interface charge transfer of nano-
semiconductor particles and thus increase of the photocatalytic
efficiencies.
Fe
these agglomerated and isolated species.
In order to check light induced redox processes and interfa-
2 3 2
O species in the TiO nanomatrix. EPR may not detect
2
2
ꢀ
+
cial charge carrier (e , h ) transfer characteristics, in situ EPR
combined with white light irradiation (400–700 nm) are used
to evaluate the processes. In situ EPR spectra obtained at 77
Selective photocatalytic oxidation of cyclohexane into
3
+
cyclohexanol and cyclohexanone over Q-TiO /Fe
2
nanoparticles under mild conditions
3
+
K on irradiation of a dry Q-TiO
2
/Fe nanosample (TF1)
are shown in Fig. 6. The corresponding quantitative EPR
results are given in Fig. 7. It was found that after performing
irradiation for a period of time of 3, 6, 9, 12, 15 min and 18
min in vacuum (Fig. 6b, c, d, e, f and g), the EPR intensities
The photocatalytic activity of selective oxidation of cyclohex-
3
+
ane over Q-TiO
tion levels in comparison with p-25 TiO
tions are shown in Fig. 8a and 8b. For all samples, it was
2
/Fe nanoparticles with different modifica-
2
under mild condi-
3
+
3+
of Bulk-Fe and Surf-Fe decreased correspondingly (see
Fig. 7). The EPR signals decayed exponentially to 60% of their
original intensity with a half-life of ca. 6 min, but no further
changes occurred on prolonged irradiation. Also, no new
3
+
found from Fig. 8a that Q-TiO
modification levels between 0.5 and 1 atom% iron content are
relatively better catalysts, and the TiO nanosample doped
with 1 atom% iron is the best one. Compared with p-25
TiO , all of the doped nanosamples show higher catalytic
activities for the selective photocatalytic oxidation of cyclo-
hexane. When p-25 TiO was used as the photocatalyst, cyclo-
2
/Fe
photocatalysts with
2
3
+
EPR signals were detected on irradiation of Q-TiO /Fe
2
2
nanosamples. When the irradiation was stopped, the original
spectrum was restored with a half-life of ca. 8.5 min (Fig.
2
6h, 6s). Compared with previous reports, the EPR intensity
decrease of both Surf-Fe and Bulk-Fe EPR signals could
3
+
3+
hexanone was the main product (the selectivity is ca. 82.5%)
as indicated in Fig. 8b, which is in consistent with earlier
2
reports. However the situations are totally different when
3
3
+
2
Q-TiO /Fe nanoparticles were employed as catalysts and
cyclohexanol is the main product (the maximum selectivity is
ca. 81.6%) as shown in Fig. 8b. This reveals that further cata-
lytic oxidation of cyclohexanol into cyclohexanone was greatly
inhibited. Based on our XRD, DRS and EPR results, it is
learned that the solubility of iron in TiO nanoparticles matrix
2
is 1.0 atom%, which is an important transforming point for
the crystal structure, electronic structure and charge-trapping
3
nanoparticles. Both of the Surf-Fe and
+
of iron doped TiO
Bulk-Fe may act as catalytic centers with ‘‘shape selectivity’’
2
3
+
2
4
characteristics similar to that of zeolite. The special surface
3
state of the Q-TiO /Fe nanoparticles could favor the selec-
+
2
tive oxidation of cyclohexane into cyclohexanol instead of
cyclohexanone. It could also be noticed from Fig. 5 and Fig.
8
a that with increasing iron modification level (modification
level from 0.05 to 1.0 atom%), both the photocatalytic activ-
ities of iron doped TiO samples and the EPR intensity of
3+
Fig. 6 EPR spectra of Q-TiO
2
/Fe (TF1) nanoparticles prior to
2
and after visible light irradiation (VI) at 77 K under vacuum:
3+
ꢀ4
bulk-Fe show similar change tendencies, i.e., first increasing
followed by attaining a maximum value and then decreasing
with further increase of the iron doping level (iron content
3
ꢂ 10 Torr: (a) prior to irradiation, (b) VI time (VIT) for 3 min,
c) VIT for 6 min, (d) VIT for 9 min, (e) VIT for 12 min, (f) VIT for
5 min, (g) VIT for 18 min, (h) Stop VIT for 10 min, (S) Stop VIT
for 17 min.
(
1
3
+
from 1.0 to 5.0 atom%). It is initially thought that Bulk-Fe
New J. Chem., 2003, 27, 1264–1269
1267

View Article Online
Scheme 1 Proposed mechanism of photo-induced redox processes
and interfacial charge carrier transfer and mechanism of selective oxi-
3+
dation of cyclohexane by using molecular oxygen over Q-TiO
nanoparticles under mild conditions.
2
/Fe
Fig. 8 The influence of iron doping level on (a) photocatalytic activ-
ity and (b) selectivity of catalytic oxidation of cyclohexane over iron
doped TiO
2
nanoparticles (S Cyclohexanol, X Cyclohexanone).
3
+
Fe nanoparticles. The crystal structure, electronic structure,
EPR characteristics and interfacial charge transfer behavior
were examined with XRD, DRS, and in situ EPR respectively.
could trap the photogenerated electrons and then transfer
them to pre-adsorbed species such as molecular oxygen on
3
+
3
+
Some important results have been obtained. Q-TiO /Fe
2
2
the surface of the Q-TiO /Fe photocatalyst. Here, Bulk-
3
+
nanoparticles with different modification levels have been
synthesized by Sol–Gel methods combined with a supercritical
Fe may act as an ‘‘electron relay’’ and could achieve higher
charge separation and surface–interface transfer efficiency.
Detailed information on photo-induced charge carrier genera-
3
+
2 2
CO drying strategy. The XRD patterns of Q-TiO /Fe indi-
cate that an anatase phase could be produced at low tempera-
ꢁ
2
tion, separation, and interfacial transfer processes in Q-TiO /
3
Fe is given described in Scheme I.
+
tures (i.e., 370 C). The solubility of iron in the TiO matrix is
2
1
atom%. More iron dopant modification level could cause
the isolation of a-Fe and favor the transformation of ana-
tase to rutile. The O ! Ti charge-transfer bands were obser-
While the iron modification level is less than 1 atom%, with
increasing iron content, more and more iron species could dif-
fuse homogeneously into the crystal lattice of titanium dioxide
nanosamples as revealed by XRD. This means that more and
more catalytic active centers and electron relay sites could be
produced. The amount of these active centers would attain a
maximum value on increasing the iron modification level to
2 3
O
4
+
2
3
+
2
ved at ca. 310 nm for Q-TiO /Fe nanoparticles. The quantum
3
confined effect was observed for Q-TiO /Fe nanoparticles.
+
2
All of the samples have EPR signals that could be attributed
and
3
+
3+
3+
to Bulk-Fe
3
and Surf-Fe
signals. The Bulk-Fe
+
Surf-Fe EPR paramagnetic centers are located in the bulk
and surface of TiO2 nanoparticles respectively. Quantitative
1
atom%. Correspondingly, the catalytic activity and selectiv-
ity of selective photocatalytic oxidation of cyclohexane into
cyclohexanol increased. This indicates that the optical surface
3
+
EPR results show that the relative EPR intensity of Bulk-Fe
and Surf-Fe signals exhibit regular change with varying cor-
3
+
3
+
2
state of Q-TiO /Fe plays a key role in selective photocata-
responding iron modification level. In situ EPR irradiation
results indicate that photo-generated charge carriers (h , e )
lytic oxidation of alkanes. On the other hand, with further
increasing the iron modification level, because of the limita-
+
ꢀ
3
could be trapped by different Fe sites simultaneously, i.e.,
+
tion of solubility of TiO
species could not diffuse into the lattice of TiO2 and may
agglomerate at the surface of TiO to form an isolated a-
Fe O phase. The two latter phases may induce the reducing
2
nanoparticles, more and more iron
+ 3+
trapping of h is considered to be due to Surf-Fe sites at
ꢀ
3+
g ¼ 4.30, whereas that of e is ascribed to Bulk-Fe sites at
g ¼ 1.99. Selective photocatalytic oxidation of cyclohexane
into cyclohexanol with higher selectivity has been obtained
2
2
3
of catalytic active centers and electron relay sites, which would
have a negative influence on the photocatalytic selective
oxidations.
3
+
by molecular oxygen activation over Q-TiO
ticles under mild conditions. The photocatalytic activities
show modification level and structure dependence. Bulk-Fe
2
/Fe nanopar-
3
+
could trap the photogenerated electrons and then transfer
them to pre-adsorbed species such as molecular oxygen to
form active oxygen species, whereas Surf-Fe could trap holes.
The charge carrier (h , e ) trapping at different Fe sites
would greatly inhibit the recombination of photogenerated
charge carriers and thus increase the efficiency of charge
separation and interfacial charge transfer.
3
+
Conclusions
+
ꢀ
3+
Selective photocatalytic oxidation of cyclohexane into cyclo-
hexanol with higher selectivity has been achieved by molecu-
lar oxygen activation under mild conditions over Q-TiO /
2
1
268
New J. Chem., 2003, 27, 1264–1269

View Article Online
1
1
1
2
T. Tanaka, T. Ito, S. Takenaka, T. Funabiki and S. Yoshida,
Catal. Today, 2000, 61, 109–115.
A. Maldotti, A. Molinari, A. Bergamini, R. Amadelli, P. Battioni
and P. D. S. Mansuy, J. Mol. Catal. A: Chem., 1996, 113, 147–
Acknowledgements
The author thanks Professor Dr Ralph Orga and Professor
Dr Thomas Johnson for helpful discussions and kind advice.
1
52.
1
3
4
K. I. Shimiz, T. Kaneko, T. Fujishima, T. Kodama, H. Yoshida
and Y. Kitayama, Appl. Catal. A: Gen., 2002, 225, 185–191.
Li Xinyong and C. Kutal, J. Mater. Sci. Lett., 2002, 21, 1525–
References
1
1
527.
1
A. L. Linsebigler, L. Guanquan and J. T. J. Yates, Chem. Rev.,
995, 95, 735–742.
1
1
5
6
J. Li, X. Li and W. Li, J. Jiangxi Normal Univ., 1997, 21, 71–74.
Z. Novak, Z. Knez, M. Drofenik and I. Ban, J. Non-Cryst. Solids,
1
2
3
Y. Mao and A. Bakac, J. Phys. Chem., 1996, 100, 4219–4226.
T. M. El-Morsi, W. R. Budakowski, A. S. Abd-El-Aziz and
K. J. Friesen, Environ. Sci. Technol., 2000, 34, 1018–1022.
M. Miyauchi, A. Nakajima, A. Fujishima, K. Hashimoto and
T. Watanabe, Chem. Mater., 2000, 12, 3–5.
J. Oh-Jin, Proc. Natl. Acad. Sci. USA, 2002, 99, 6482–6486.
C. J. G. Cornu, A. J. Colussi and M. R. Hoffmann, J. Phys. Chem.
B, 2001, 105, 1351–1354.
R. R. Ozer and J. L. Ferry, J. Phys. Chem. B, 2000, 104, 9444–
9448.
A. Maldotti, L. Andreotti, A. Molinari, S. Borisov and V.
Vasil’ev, Chem. Eur. J., 2001, 7, 3564–3571.
H. Einaga, S. Futamura and T. Ibusuki, Appl. Catal. B: Environ.,
2
001, 285, 44–49.
1
7
8
A. R. Bally, E. N. Korobeinikova, P. E. Schmid, F. Levy and
F. Bussy, J. Phys. D: Appl. Phys., 1998, 31, 1149–1154.
G. Cordoba, J. Padilla, V. H. Lara and R. Arroyo, Mater. Lett.,
4
1
2
002, 54, 397–402.
5
6
19
20
21
22
C. C. Wang and J. Y. Ying, Chem. Mater, 1999, 11, 3113–3117.
Louis Brus, J. Phys. Chem., 1986, 90, 2555–2560.
M. Gratzael and R. F. Howe, J. Phys. Chem., 1990, 95, 274–279.
J. A. Navio, M. Macias, M. Gonzalez-Catalan and A. Justoo,
J. Mater. Sci., 1992, 27, 3036–3041.
7
8
9
0
2
3
4
W. Mu, J. M. Herrmann and P. Pichat, Catal. Lett., 1989, 3,
73–76.
M. V. Landau, S. B. Kogan, D. Tavor, M. Herskowitz and
J. E. Koresh, Catal. Today, 1997, 36(4), 497–510.
2
2
002, 38, 215–225.
K. Wada, K. Yoshida, Y. Watanabe and T. Suzuki, Appl. Catal.,
991, 74, L1–L4.
1
1
New J. Chem., 2003, 27, 1264–1269
1269