COMMUNICATION
Efficient visible light-active N-doped TiO2 photocatalysts by a
reproducible and controllable synthetic route
Suil In, Alexander Orlov, Felipe Garc´ıa, Mintcho Tikhov, Dominic S. Wright and Richard M. Lambert*
Received (in Cambridge, UK) 19th July 2006, Accepted 7th August 2006
First published as an Advance Article on the web 31st August 2006
DOI: 10.1039/b610316b
measurement by XPS difficult and possibly misleading.2,12
Alternative routes such as ion implantation2 can in principle
A reproducible and controllable method allows the synthesis of
practical quantities of efficient, visible light active TiO2(N)
permit some degree of control over the level of N-doping, albeit
with unknown dopant depth distribution, but they are costly and
unsuited to scale-up.
photocatalysts in which the nitrogen content may be varied to
achieve optimum performance.
Photocatalytic degradation of organic compounds on semicon-
ductor surfaces is an important area of current research with
respect to both fundamental understanding1 and potential
practical applications.2 TiO2, a promising, low-cost, robust
photocatalyst is the most extensively investigated material in this
field. However, its band gap (3.3 eV for anatase) is such that it can
only be activated by UV radiation (l , 376 nm). Given that such
radiation corresponds to only y5% of the incident solar flux at
ground level, very large gains in efficiency of practical solar-
powered photocatalytic systems are to be expected if they could be
operated under visible light. Such materials would also be effective
under artificial ambient light, for example, within buildings. As a
result, intensive effort has been directed towards achieving this
goal. An attractive strategy involves doping TiO2 with low levels of
Here we describe a straightforward, low-cost and reproducible
method for the synthesis of TiO2(N) by which the level of
N-doping is readily controlled to yield effective visible light active
photocatalysts.
A series of TiO2(N) catalyst precursors with nominal nitrogen
doping levels in the range 0.2–1.0 wt% was prepared by a simple
procedure involving the sequential reaction of H2O with a small
known excess of TiCl4 in toluene (step 1) under dry, O2-free argon,
followed by stoichiometric reaction of the remaining TiCl4 present
with a standard solution of NH3 in dioxane (step 2).
step 1
TiCl4z2H2O DCCA TiO2z4HCl
step 2
,
3TiCl4z4NH3 DCCA Ti3N4z12HCl
nitrogen [hereafter TiO2(N)],
a process that red-shifts the
absorption spectrum of titania.2–8 However, it is important to
note that the visible light photocatalytic activity of TiO2(N) is
highly sensitive to the synthetic route used in its preparation: not
infrequently, although such materials absorb visible light they are
It is unlikely that either step in this procedure will give rise to
pure components. For example, in step 1 oxo-chlorides will also be
formed [TiOxCl422x], the Ti–Cl groups of which will then react in
the second step. Although we cannot fully exclude the possibility
of some unreacted Ti–Cl or Ti–NHn (n = 1–3) groups remaining in
samples after reaction, few, if any, of these groups are likely to
remain after final calcination. All reactions were performed under
dry, O2-free argon on a conventional glass vacuum line using a
three-necked 2 dm3 round-bottomed flask, equipped with an
overhead magnetic stirrer and water-cooled reflux condenser.
Greaseless fittings were used throughout in order to prevent
contamination of samples. This included teflon sleeves on all
joints, use of Youngs taps and a stirrer guide with chemically
resistant o-rings. The stirrer guide was also equipped with a small
(ca. 6 cm) water condenser in order to prevent chemical
degradation of the rubber mountings during the reaction. In a
typical reaction (here illustrated for a nominal 0.56 wt% sample), a
solution of TiCl4 (4.773 g) in toluene (500 mL) was reacted with
H2O (0.883 g) at room temperature. The mixture was then stirred
(24 h). A solution of NH3 in dioxane (0.5 mol dm23, 1.60 mL) was
then added and stirred at room temperature (24 h), yielding a
bright yellow suspension. This mixture was then brought to reflux
(24 h), giving a light yellow suspension. The catalyst precursor was
filtered off using a porosity 3 filter frit, washed with n-hexane then
dried for 10 h under vacuum (1022 bar) at room temperature.
In order to determine the optimum procedure for generating the
final photocatalysts, samples of each precursor produced by the
nonetheless totally inactive with respect to photocatalysis.9
A
possible explanation for this is that the photogeneration of
electrons and holes is not in itself enough to impart photo-
activity—in addition, the electron–hole recombination rate must
also be low enough to permit a sufficient flux of these active species
to reach the catalyst surface where they initiate reactions with
adsorbed species. If practical applications are to be achieved, it is
also necessary for the synthetic method to be (i) controllable, (ii)
reproducible and (iii) low-cost. A variety of methods for N-doping
of titania have been reported,2–8 none of which meets all these
conditions. Thus, although the hydrolysis of titanium precursors in
the presence of a nitrogen source provides an inexpensive method
of nitrogen doping,5,8,10,11 the composition of the resulting
TiO2(N) is difficult to control and cannot be systematically varied
so as to establish the optimum composition for photoactivity.
Thermal treatment of TiO2 in an ammonia atmosphere results in
significant reduction of Ti and substitutional doping with N.12
Although the apparent N concentration can be varied by
subsequent annealing, inhomogeneous nitrogen distribution within
the sample impairs reproducibility and renders N concentration
Department of Chemistry, University of Cambridge, Cambridge,
England. E-mail: rml1@cam.ac.uk; Fax: +44 (0)1223 336362;
Tel: +44 (0)1223 336467
4236 | Chem. Commun., 2006, 4236–4238
This journal is ß The Royal Society of Chemistry 2006
In, Suil
