Disordered mesoporous TiO2-xNx+nano-Au: An electronically integrated nanocomposite for solar H2 generation

Article abstract of DOI:10.1002/cctc.201300715

We report on H₂ generation by photocatalysis driven by simulated white light by electronically integrated Au nanoparticles with multifunctional, disordered mesoporous TiO_2-xN_x (Au-NT) nanocomposites. Solar H₂ ge

Full text of DOI:10.1002/cctc.201300715

CHEMCATCHEM
FULL PAPERS
DOI: 10.1002/cctc.201300715
Disordered Mesoporous TiO2ÀxNx+Nano-Au: An
Electronically Integrated Nanocomposite for Solar H2
Generation
Kumarsrinivasan Sivaranjani,[a] Sivaraman RajaAmbal,[a] Tanmay Das,[b] Kanak Roy,[a]
Somnath Bhattacharyya,[b] and Chinnakonda S. Gopinath*[a, c, d]
We report on H2 generation by photocatalysis driven by simu-
lated white light by electronically integrated Au nanoparticles
with multifunctional, disordered mesoporous TiO2ÀxNx (Au-NT)
nanocomposites. Solar H2 generation (1.5 mmolhÀ1 gÀ1) from
aqueous methanol has been demonstrated with Au-NT nano-
composites. The water splitting activity of Au-NT is attributed
to the 21.1 ps lifetime of charge carriers observed from fluores-
cence lifetime measurements, which indicates a high electron-
injection efficiency from nano-Au to the conduction band of
TiO2, and hence charge separation as well as utilization. This is
directly supported by the observation of a high photolumines-
cence emission intensity with Au-NT that highlights the energy
transfer from nano-Au to TiO2. The p–n heterojunction ob-
served between the Au (001) and TiO2 (101) facets helps to-
wards the higher charge separation and their utilization. A low
mesochannel depth (<10 nm) associated with disordered mes-
oporous TiO2ÀxNx helps the charge carriers to move towards
the surface for redox reactions and hence charge utilization.
Visible-light absorption, as a result of the surface plasmon res-
onance of nano-Au, is observed in a broad range between 500
and 750 nm, which helps in harvesting visible-light photons. Fi-
nally, electronically integrated nano-Au with TiO2ÀxNx in Au-NT
is evident from Raman and X-ray photoelectron spectroscopy
measurements. All of these factors help to achieve a high rate
of H2 production. It is likely that a higher rate of H2 production
than that reported here is feasible by strategically locating Au
clusters in porous TiO2 to generate hot spots through electron-
ic integration.
Introduction
The visible-light-driven photocatalytic water splitting reaction
(WSR) is projected to be an indispensable solution to meet
a significant part of the clean and sustainable energy demand
on a global level.[1] The production of H2, a clean fuel, from
water and sunlight by the overall water splitting reaction
(OWSR) is an ultimate goal to sustain the globe in a better
way.[1–3] In spite of many efforts in the past, the OWSR by visi-
ble-light-driven photocatalysis is yet to produce a breakthrough
in terms of a quantum efficiency of ꢀ10%, which is the mini-
mum benchmark to consider commercialization. From the rate
of slow progress of the OWSR witnessed in the last few de-
cades, in our opinion, the achievement of 10% efficiency
might be considered as a long-term solution for energy de-
mands. Nonetheless, H2 production through the WSR with
a sacrificial agent, such as methanol or glycerol, might be con-
sidered as a short-term goal, as it avoids the four-electron pro-
cess to produce O2 in the OWSR.[4,5] There are significant num-
bers of reports available on the WSR and OWSR, and more
concerted efforts are necessary to produce a sustainable pho-
tocatalyst system.[1–5]
[a] Dr. K. Sivaranjani, S. RajaAmbal, K. Roy, Prof. C. S. Gopinath
Catalysis Division
Among the materials available, TiO2 is considered to be the
best candidate for photocatalysis for a variety of reasons, such
as photostability, low toxicity, large abundance, cost effective-
ness, and high oxidation potential.[6] However, a large band
gap and fast charge-carrier recombination are the major draw-
backs of TiO2. There have been many efforts in the past to
make TiO2 a visible-light-active photocatalyst through metal
ion doping, anion doping, surface sensitization, etc.[7–12] Nitride
anion doping is considered to be the most suitable path to
reduce the band gap of TiO2 and bring about the absorption
of more visible light.[9,10] It has been demonstrated that nitride
or anionic nitrogen is responsible for the band-gap reduction
and visible-light absorption in TiO2ÀxNx.[9,11] TiO2ÀxNx has been
prepared successfully by various methods.[9,11–16]
CSIR-National Chemical Laboratory
Dr. Homi Bhabha Road, Pune 411 008 (India)
Fax: (+91)20-2590-2633
E-mail: cs.gopinath@ncl.res.in
[b] T. Das, Prof. S. Bhattacharyya
Department of Condensed Matter Physics and Materials Science
Tata Institute of Fundamental Research
1 Homi Bhabha Road, Colaba, Mumbai 400 005 (India)
[c] Prof. C. S. Gopinath
Center of Excellence on Surface Science
CSIR-National Chemical Laboratory
Pune 411 008 (India)
[d] Prof. C. S. Gopinath
CSIR-Network of Institutes for Solar Energy (NISE)
CSIR-NCL Campus
Pune 411 008 (India)
Along with visible-light absorption, the probability of
charge-carrier recombination should be suppressed or mini-
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201300715.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 522 – 530 522

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
mized to increase the quantum efficiency. The introduction of
mesoporosity in TiO2 increases the rate of diffusion of charge
carriers from the bulk to the surface.[14,17] Specifically, disor-
dered mesoporosity is known to reduce the diffusion length of
charge carriers as the depth of the mesopores is a few nano-
meters (<10 nm),[14,18] unlike several hundred nanometers in
conventional mesoporous materials, such as MCM-41 and SBA-
15.[19–21] These types of mesopores are known as pseudo-three-
dimensional (p3D) mesopores.[14,22] This disordered mesopo-
rous framework provides an easy route for the diffusion of re-
actants and products because of low diffusion barriers.
It is known that noble metal cluster deposition on TiO2 re-
sults in an electron sink through selective storage of elec-
trons.[23] Interestingly, nano-Au or Ag deposits on TiO2 cause
more visible-light absorption through surface plasmon reso-
nance (SPR) bands.[24] It is known that the final properties of an
Au/TiO2 system, especially for WSR activity, mainly depend on
the preparation procedure. Bamwenda et al. prepared Au/TiO2
and Pt/TiO2 photocatalysts by different preparation routes and
showed that their WSR activity from aqueous ethanol solution
depends on the preparation procedure.[25] Recently Murdoch
et al. have shown that photocatalytic activity is independent of
the Au cluster size between 3 and 12 nm and Au/anatase gives
a two orders of magnitude higher rate of H2 production than
on rutile TiO2.[26] Subramanian et al.[27] demonstrated a negative
shift of the Fermi level (EF) of a Au/TiO2 nanocomposite;
a higher negative potential shift has been observed with the EF
of Au/TiO2 with decreasing Au particle size. Hence, the result-
ing composite materials are more reductive than pure TiO2.
Furthermore, lifetime measurements indicate that charge injec-
tion occurs from TiO2 to Au. However, exactly the opposite to
the above, electron injection from Au to the conduction band
(CB) of TiO2 was suggested because of the strong localization
of plasmonic near fields close to the Au–TiO2 interface.[28,29] No-
tably, the preparation methods are different in the above re-
ports, and it is very likely that the method of preparation influ-
ences the electron-injection mechanism. Apart from the above
controversies, the mechanistic aspects of the Au/TiO2 system
and the integration of nano-Au with TiO2 is less under-
stood.[27–29] However, H2 production with Au-TiO2 composites
through the WSR demonstrates its high potential,[25–29] and
a better understanding would lead to the application of these
materials with better activity.
Figure 1. a) Wide-angle PXRD pattern and b) UV/Vis absorption spectra of
xAu-NT nanocomposites. Inset in a) is shown the low-angle XRD patterns.
The photograph (top) shows the color variation with increasing Au content
of xAu-NT with sample labels as A=N-TiO2, B=0.01Au-NT, C=0.03Au-NT,
D=0.05AuNT, E=0.10AuNT, and F=0.30Au-NT.
and a significantly long lifetime of electrons (21 ps). These
composite materials show decent activity towards H2 genera-
tion from aqueous methanol solution under simulated white
light. This work is part of ongoing studies in our laboratory on
materials development for visible-light-driven photocataly-
sis.[11–14,18,30–34]
Results and Discussion
Powder X-ray diffraction (PXRD) patterns of all xAu-NT materi-
als, in which x indicates the nominal amount of Au loaded
[at%], are shown in Figure 1a. All the peaks in the XRD pat-
terns could be indexed to the anatase phase of TiO2 (JCPDS
File 21-1272).[34,35] All of the peaks are broad, which indicates
the nanocrystalline nature of the materials. The crystallite size
was calculated by using the Debye–Scherrer equation, and the
results are shown in Table 1. Up to 0.10Au-NT, there is no fea-
ture observed in the XRD patterns that corresponds to metallic
A tandem approach has been employed to synthesize an
electronically integrated composite of nano-Au and disordered
mesoporous TiO2ÀxNx (Au-NT) in
a single step by a simple solu-
Table 1. Physicochemical properties of the xAu-NT materials.
tion combustion method (SCM).
The Au-NT composite material
possesses desirable properties,
such as nano-Au clusters with
Material
Surface
area [m2 gÀ1
Pore Diameter
[nm]
Pore volume
[cm3 gÀ1
]
Crystallite
size [nm]
Bulk Au
[at%][a]
]
N-TiO2 (NT)
0.01Au-NT
0.03Au-NT
0.05Au-NT
0.10Au-NT
0.30Au-NT
234
218
200
213
171
157
7
0.44
0.37
0.4
0.43
0.41
0.27
6.8
6.0
6.2
6.9
8.0
8.8
0
SPR in
a broad visible-light
8.2
8.1
8
9.6
6.8
0.009
0.026
0.045
0.11
0.28
range, TiO2ÀxNx with disordered
mesoporosity, low mesochannel
depth (ꢁ10 nm) and high sur-
face area, electrically intercon-
nected nanoparticles (EINP),[18]
[a] Measured from EDX spectra.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 522 – 530 523

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Au. However, 0.30Au-NT shows a peak at 2q=448 that corre-
sponds to the (200) plane of metallic Au.[35] This supports the
good dispersion of nano-Au clusters on xAu-NT (xꢁ0.1).
The inset in Figure 1a shows the low-angle XRD patterns of
all of the xAu-NT materials. The mesoporous nature of the Au-
NT materials was confirmed by a feature at around 2q=18.
Only one peak was observed, which indicates the disordered
mesoporosity.[14,34] Unlike ordered hexagonal mesoporous ma-
terials, such as MCM-41 and SBA-15, no other low-angle fea-
tures were observed,[19–21] which highlights the disordered mes-
oporous nature that results from the intergrowth of fundamen-
tal particles. A high Au loading (x=0.3) affects the mesoporosi-
ty of the materials significantly. A gradual decrease in the in-
tensity of the low-angle peak at 2q=18 indicates increasing
pore blockage by nano-Au clusters. However, all of the xAu-NT
materials possess disordered mesoporosity along with
nanocrystallinity.
Figure 2. a) N2 adsorption–desorption isotherms and b) BJH pore size distri-
bution of xAu-NT composite materials. c) TEM image of 0.05Au-NT. The SAED
pattern is shown in the inset. No distinct Au particles are directly observed,
which indicates the good dispersion.
The UV/Vis spectra of the xAu-NT nanocomposites are
shown in Figure 1b. The absorption edge of TiO2 remains at
around 380 nm, in spite of N doping.[14] The color photograph
in Figure 1 displays the xAu-NT nanocomposites. N-TiO2 is pale
yellow, and upon Au introduction, a blue to grayish blue color
develops gradually. On increasing the Au loading to 0.1 at%
and above, the color changes increasingly towards dark blue.
The SPR features of nano-Au in these composite materials
cause more visible-light absorption between 500 and
750 nm.[26,27] This light absorption is a result of the collective
oscillation of electrons of the Au particles in response to opti-
cal excitation.[27] Visible-light absorption between 500 and
750 nm demonstrates the surface plasmon state energy, at
least, between 1.65 and 2.5 eV above the EF of Au, which indi-
cates the production of energized electrons. This SPR peak in-
tensity increases with increased Au loading, and the position
of the peak depends on the Au particle size or aggregation
and the surrounding environment.[36,37] Hence visible-light pho-
tons can be harvested by using xAu-NT nanocomposites.
Indeed, TEM studies of 0.1Au-NT show the size distribution of
Au clusters (Figure 3), which suggest that different size parti-
cles lead to a broad absorption in the visible-light regime.
However, the pore volume and surface area decrease with in-
creasing Au content, especially at xꢀ0.1 (Table 1), which indi-
cates increased pore blockage as a result of Au clusters in the
pores. Apart from SPR, nano-Au clusters act as an electron sink
by forming Schottky barriers with TiO2 and the active sites for
H2 evolution. Thus nano-Au in xAu-NT increases the visible-
light absorption significantly.
NT exhibits a surface area of 157 m2 gÀ1, which indicates in-
creasing pore blockage.
A TEM image of 0.05Au-NT is shown in Figure 2c (more TEM
images can be seen in Figure S2). All of the xAu-NT particles
exhibit a spherical or near-spherical shape, and a disordered
mesoporous structure is observed for all xAu-NT materials. This
disordered mesoporosity arises because of the intergrowth of
fundamental particles, which leads to aggregates with a signifi-
cant extra-framework void space. The presence of meso- and
macropores is clearly visible in Figure 2c. These assist the
faster diffusion of reactants and products in heterogeneous
catalysis. Selected-area electron diffraction (SAED) patterns
confirm the crystalline nature of xAu-NT nanocomposites with
anatase phase TiO2. TEM images show that the majority of the
lattice fringes correspond to the (101) crystallographic planes
of the anatase phase (d(101)=0.352 nm). These observations
are in excellent agreement with the XRD results. This disor-
dered mesoporous structure has additional advantages owing
to the small mesochannel depth of ꢁ10 nm; these are referred
to as p3D mesopores.[14,18,22] This is in contrast to micron-size
long mesochannels in ordered mesoporous materials.[19]
Finally, nanocrystallites in xAu-NT are electrically well con-
nected with each other, and this connectivity extends up to
a few micrometers (Figures 2c and S1). As a result of the EINP
nature,[14,18] the excited electron can move from one end to
the other effectively and from the bulk to the surface because
of the small diffusion length under light-harvesting conditions.
This advantageous EINP feature associated with xAu-NT com-
posite materials enhances the diffusion of charge carriers to
the surface.
N2 adsorption–desorption isotherms and the Barrett–Joyner–
Halenda (BJH) pore size distribution of xAu-NT are shown in
Figure 2. All materials show a type IV isotherm with an H2 hys-
teresis loop, which is typical for disordered mesoporous mate-
rials.[14,22] xAu-NT nanocomposites show a narrow pore size dis-
tribution with an optimum pore diameter around 8Æ1 nm.
Compared to TiO2ÀxNx, xAu-NT exhibits a significant decrease in
the BET surface area. Among these composite materials,
0.05Au-NT shows the highest surface area of 213 m2 gÀ1. The
surface area decreases with increasing Au loading, and 0.30Au-
Unlike the reports cited in the introduction, we were unable
to find any Au particles on Au-NT in our TEM studies, likely be-
cause of the similar contrasts of Au and TiO2. To explore vari-
ous aspects of Au, Z-contrast image analysis was performed by
using a FEI-TITAN microscope operated at 300 kV. A represen-
tative Z-contrast image of 0.10Au-NT at low magnification is
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 522 – 530 524

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 4. a) Raman spectra of xAu-NT composite materials. The Raman-active
Eg mode on N-TiO2 at 145 cmÀ1 shows a blueshift to 154 cmÀ1 with increased
line broadening and asymmetry and decreased intensity with increasing Au
content. b) PL spectra for xAu-NT materials. The Au content is given for ma-
terial identification. The dotted lines are a guide to the eye.
which indicates line broadening. This line broadening is ac-
companied with asymmetry for all of the xAu-NT composites,
which is particularly evident for the Eg feature at 640 cmÀ1. The
anatase (101) facet is the most abundant plane (Figures 1a
and 2c), which is corrugated with alternating rows of five-coor-
dinate Ti (likely Ti3+) and bridging oxygen at the edges of the
corrugation.[39] This makes the (101) crystallographic facet
ideal to interact with the deposited Au clusters through charge
transfer, which leads to a blueshift in the Raman spectrum.[39]
Indeed, this electronic interaction on a polar surface is likely to
assist fast electron transfer.
Figure 3. a) Z-contrast image of 0.10Au-NT at low magnification. b) High-res-
olution Z-contrast image of the nanoparticle shown within the white square
shown in a). EDX spectrum recorded c) from the Au nanoparticle shown in
b) and d) on the TiO2 matrix.
shown in Figure 3a, in which small bright regions indicate the
possibility of the accumulation of higher-atomic-number ele-
ments than the TiO2 matrix. On average, the particle size varies
from 5–21 nm (Figure 3a). The high-resolution Z-contrast
image of the bright region within the white square shown in
Figure 3a is shown in Figure 3b, in which the lattice spacing
(ꢂ0.2 nm) matches the projected distance between the Au
(001) planes. The energy-dispersive X-ray (EDX) spectrum re-
corded from the same particle confirms that it is an Au nano-
particle embedded in a matrix that contains Ti and O (Fig-
ure 3c). The EDX spectrum of a TiO2 matrix region proves that
the matrix does not contain any Au (Figure 3d). If we consider
all of the results presented in Figure 3, it can be stated that Au
nanoparticles were grown within a TiO2 matrix and that Au
(001) and TiO2 (101) planes make p–n heterojunctions, which
are expected to help charge separation and hence its better
utilization. Notably, the Au particles are isolated with a mini-
mum average interparticle distance of 200 nm.
The photoluminescence (PL) emission spectra of Degussa
P25 (P25) and xAu-NT nanocomposites are shown in Figure 4b.
All of the xAu-NT composites and P25 show peaks at 419, 442,
and 470 nm. An emission peak as a result of the band-gap
transition appears at 380 nm for all materials (not shown).[40]
The emission band at 419 nm is attributed to the free exciton
emission of TiO2.[41] The appearance of a shoulder around
470 nm is because of the surface state, such as Ti4+ÀOH.[42] The
emission feature at 440 nm originates from the charge-transfer
transition from Ti3+ to oxygen anions in a [TiO6]8À complex[43]
present in the material. These three emission features show
the lowest intensity for P25. N-TiO2 shows a significant increase
in the intensity of all the features, compared to P25. Upon the
deposition of Au on N-TiO2, the intensity of all emission fea-
tures increases, and the maximum intensity is observed with
0.05Au-NT. In particular, the intensity of the feature at 442 nm
increases more significantly than other features, which sug-
gests a better interaction of Au with NT through the unsaturat-
ed (101) facet. These observations reiterate the presence of p–
n heterojunctions in the xAu-NT nanocomposites, which help
the electron injection into the CB of TiO2. On further Au load-
ing, the intensity of all of the emission features starts to de-
crease. Nonetheless, xAu-TiO2 exhibits higher intensity emission
features than P25 and N-TiO2, which underscores the energy
transfer from nano-Au to the CB of TiO2. Indeed this is the criti-
cal observation that highlights the electronic interaction be-
tween nano-Au and TiO2. The maximum intensity emission fea-
Raman spectroscopy is a versatile tool to determine the
structural features of the composites. Raman spectra for all of
the xAu-NT composite materials are shown in Figure 4a. All six
Raman-active fundamental modes are observed at 145 (Eg),
198 (Eg), 398 (B1g), 516 (A1g+B1g), and 640 cmÀ1 (Eg) for the ana-
tase phase.[38] The Raman-active Eg mode of TiO2 at 145 cmÀ1
exhibits a systematic blueshift to 154 cmÀ1 upon nanocompo-
site formation with Au. Furthermore, the full width at half max-
imum (FWHM) measured for Eg (as well as other features)
changes from 15.1 cmÀ1 on N-TiO2 to 25.2 cmÀ1 on 0.1Au-NT,
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 522 – 530 525

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
tures are observed for 0.05Au-NT, which indicates that the life-
time of the charge carriers is likely to be longer and hence the
possibility of electron injection from nano-Au to the CB of TiO2
is significantly high. Indeed, the lifetime measurements fully
support these conclusions (vide infra). The surface plasmon
state energy (1.65–2.5 eV) measured from UV/Vis absorption
spectroscopy (Figure 1b) indicates the production of energetic
electrons (ꢂ3.3 eV) because of PL excitation at l=375 nm.
This ensures the injection of electrons from the surface plas-
mon states of Au to the CB of TiO2 under visible-light irradia-
tion as well as under the PL measurement conditions. The EF
of Au and CBMin of TiO2 appear at around 0 and À0.5 eV, re-
spectively, versus the normal hydrogen electrode (NHE).[2,3,28]
Electron injection from the surface plasmon states of Au to the
CB of TiO2 is definitely possible and in accordance with the
above energy levels. On increasing the Au content ꢀ0.1, the
emission features decreased in intensity, which is attributed to
the larger size or agglomeration of Au particles and hence the
possibility of lower electron injection from Au to TiO2. Larger
Au particles, observed by XRD (Figure 1a), decrease the inter-
action with the TiO2ÀxNx lattice, and thereby the electronic in-
teraction also decreases. Indeed, the aggregation of Au parti-
cles at x=0.3 re-exposes the TiO2ÀxNx surface sites, and hence
the PL emission intensity decreases dramatically. An optimum
quantity of Au around x=0.05 in the xAu-NT composite mate-
rials maximizes the interaction between TiO2 and the Au
clusters.
for NT and TiO2.[11,14] A glance at the Ti2p core-level spectra ob-
tained from xAu-NT indicates a possible nonuniform static-
charge build-up upon photoelectron emission. Nonuniform
static-charge build-up occurs because of the presence of differ-
entially conducting areas on the surface layers, especially be-
cause of the deposition of a small amount of Au on TiO2. A
low-energy electron flood gun was employed to neutralize the
build-up of static charge.[44] With charge neutralization, the XPS
data show typical Ti2p features expected for any TiO2 material;
however, broad core-level peaks are observed for xAu-NT (bold
traces in Figure 5a), which indicate the presence of mixed
valent Ti ions. Indeed, these observations highlight the follow-
ing electronic aspects of xAu-NT. 1) The feature observed at
a low binding energy (BE), after static-charge neutralization, at
around 457 eV corresponds to the Ti3+ oxidation state on xAu-
NT. 2) Even without charge neutralization, the above Ti3+ fea-
ture was observed at 457 eV (thin traces), demonstrating the
conducting nature of part of the xAu-NT material, which is at-
tributed to the electronic interaction of nano-Au with TiO2
through p–n heterojunctions. The observation of the conduct-
ing nature of part of the xAu-NT underscores the equilibration
of Fermi levels and the electronic integration of Au with NT.
3) The Ti3+ oxidation state is observed at the correct BE, even
without charge neutralization, which suggests that Ti3+ is gen-
erated upon the deposition of Au clusters. Indeed, the above
points not only suggest an electronic interaction between Au
and TiO2 in Au-NT but also the polarized nature of the Au–TiO2
interface, which helps to separate electrons from electron–hole
pairs. Small and uniformly sized Au clusters are likely to in-
crease the interactions between Au and TiO2, which is likely to
increase the electronic integration to a higher level and is
worth exploring further.
To know more about the nature of xAu-NT, X-ray photoelec-
tron spectroscopy (XPS) studies were performed by using
a custom-built high-pressure XPS instrument.[44] An interesting
observation was made from the Ti (and other elements) core-
level results (Figure 5a). As-prepared xAu-NT nanocomposites
show three features for the Ti2p core level (thin traces in Fig-
ure 5a); this is in contrast to the typical two core-level spin-
orbit doublets that result from Ti2p3/2 and 2p1/2 as observed
The N1s core level appears at 398.2Æ0.1 eV (Figure 5b).
Compared to NT, there is a marginal increase in the broaden-
ing observed with the N1s core level of xAu-NT. The Au4f core
level was recorded, and the result is shown for
0.05Au-NT in the inset of Figure 5b. The N/Ti atomic
ratio was 0.29Æ0.03 for xAu-NT. Although the signal-
to-noise ratio is very poor, the presence of Au is con-
firmed. Approximately 70 min was required to obtain
this data with repeated scans. The Au4f7/2 core level
appears at 84.2 eV, which suggests the metallic
nature of Au.
To evaluate the efficacy of these Au-NT nanocom-
posites, the photocatalytic WSR was performed
under simulated sunlight in an aqueous methanol so-
lution. The H2 evolution by using 0.05Au-NT nano-
composite continuously for 15 h with evacuation
after every 5 h is shown in Figure 6a. A steady rate of
H2 evolution was observed for at least five of such
cycles with very similar activity, which indicates the
photostability of the nanocomposites. CO2 was also
produced along with H2, which was confirmed by GC
Figure 5. a) XPS core-level spectra recorded for the Ti2p core level for NT and xAu-NT
(thin traces) with a low-energy electron flood gun (bold traces). The low-energy electron
flood gun was employed to neutralize the nonuniform static-charge build-up on xAu-NT.
b) N1s core-level results from NT and xAu-NT, and the inset shows the Au4f core-level re-
sults from 0.05Au-NT.
(not shown). The molar ratio of H2/CO2 is close to
2.92Æ0.1, which suggests a general mechanism of
one molecule each of CH3OH and H2O led to three
molecules of H2 and one molecule of CO2 on TiO2 sur-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 522 – 530 526

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
Figure 6. Photocatalytic H2 evolution activity of a) 0.05Au-NT for 15 h and
b) xAu-NT nanocomposites in visible light with aqueous methanol solution.
The amount of H2 evolution reported is with 100 mg of catalyst. The dotted
lines in a) indicate evacuation after every 5 h.
Figure 7. Emission decay of xAu-NT and TiO2ÀxNx deposited as thin films on
a glass substrate. The excitation and monitored emission wavelength are
375 and 440 nm, respectively. A kinetic fit was obtained by using biexponen-
tial decay, and the results are given in the text.
faces.[5] A H2/CO2 ratio of 3 was maintained even if the metha-
nol content varied between 5 and 25%, which indicates that
the general mechanism holds well. The photocatalytic H2 evo-
lution activity of all of the xAu-NT nanocomposites is shown in
Figure 6b, and 0.05Au-NT shows the highest H2 evolution rate.
Approximately 150 mmol H2 hÀ1 is generated with 100 mg of
0.05Au-NT. Pure TiO2ÀxNx shows the lowest activity towards H2
generation (14 mmolhÀ1) under similar conditions. A half to
one order of magnitude increase in H2 production is observed
with xAu-NT composites compared to TiO2ÀxNx, which demon-
strates the critical role of nano-Au clusters in solar-light har-
vesting. The high rate of H2 production is attributed to the
electronic integration of nano-Au clusters with the disordered
mesoporous TiO2ÀxNx framework.
sion intensity from TiO2 features increased manifold upon the
introduction of nano-Au into the TiO2 matrix. These observa-
tions demonstrate the energy transfer from nano-Au to TiO2 in
an effective manner. This is probably the first report to show
such a long lifetime in a Au-TiO2 system.
Photo-electrochemical (PEC) measurements were preformed
under UV light with xAu-NT nanocomposites, and the results
are shown in Figure 8a. Compared to P25 and TiO2ÀxNx, xAu-
Charge-carrier lifetime measurements were per-
formed to explore the electron injection mechanism.
We analyzed the emission decay of xAu-NT (x=0.05
and 0.1) and TiO2ÀxNx, and the results are shown in
Figure 7. Initially, the materials were excited with
a 375 nm light-emitting diode (LED) source, and the
emission decay was collected at 440 nm with
5000 counts. The emission decay was best fitted by
a biexponential decay for all of the materials. The
first fast decay is a result of radiative and Auger re-
combination processes; this is followed by a slow
decay, which accounts for the energy-transfer mecha-
nism. The lifetime of TiO2ÀxNx, before Au is loaded,
was t1 =1.76 ps (a1 =1) and t2 =925 ps (a2 =0). The
major contribution is from the t1 species with a life-
time of 1.76 ps. Upon the introduction of nano-Au on
TiO2 (0.05Au-NT), the lifetime of the TiO2 species in-
creases to t1 =21.1 ps (a1 =0.17) and t2 =11.4 ps
(a2 =0.83). An order of magnitude increase in the life-
Figure 8. a) The photocurrent generated upon UV illumination is shown to demonstrate
the fourfold increase in photocurrent density with xAu-NT nanocomposites than that
achieved without Au clusters. b) Chronoamperometry measurements made at 0.5 V with
time of t1 from NT (1.76 ps) to Au-NT (21.1 ps) under-
scores the importance of Au deposition on NT. This
result corroborates the PL results, in which the emis- xAu-NT compared with TiO2.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 522 – 530 527

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
NT displays a fourfold increase in photocurrent generation. An
increase in the photocurrent under illumination at positive po-
tentials is typical of n-type conductivity. Careful analysis of the
PEC data revealed that the increase in the PEC amplitude is
the maximum for 0.05Au-NT of these materials. This corrobo-
rates well with the lifetime and PL measurements as well as H2
production data.
parison to other reports, is partially attributed to the material
preparation by the SCM method. Indeed, the different prepara-
tion procedures adopted by different groups[25–29,45] could be
a reason for controversial conclusions about the electron injec-
tion mechanism. This is highly relevant as the surface of
a
nanocomposite, especially the Au–TiO2 interface, plays
a major role in the optical properties, and the nature of the
surface and interface depends on the method of preparation.
Indeed, two different preparation methods would lead to sig-
nificantly different surface characteristics. Nonetheless, differ-
ent preparation methods could also lead to different electronic
integration and hence different characteristics and capabilities.
Chronoamperometry measurements were performed to ex-
plore the photocurrent generation at different voltages, and
a representative measurement made at 0.5 V is shown in Fig-
ure 8b. The photocurrent generation with 0.05Au-NT is ap-
proximately five times higher than that with TiO2 (P25), which
highlights the increase in photon-to-current generation be-
cause of many factors, such as the decrease of defects and the
diffusion of electrons to nano-Au clusters through the mesopo-
rous TiO2 framework.
Experimental Section
Synthesis of xAu-NT materials
We employed SCM that involves a simple synthesis protocol and
requires a short reaction time (<10 min) and cheap starting mate-
rials. All the chemicals employed were of analytical grade and used
without further purification. Titanium nitrate (Sigma–Aldrich) as the
Ti precursor, gold(III) chloride (Sigma–Aldrich) as the Au precursor,
and urea (Merck) as the fuel were used. The required quantity of
aqueous titanyl nitrate, gold chloride, and urea were put into
a 250 mL beaker and introduced into a muffle furnace preheated
at 4008C. Water evaporation takes place in the first few minutes,
followed by a smoldering-type combustion that occurs in the next
1–2 min. Immediately after the combustion process was complet-
ed, the xAu-NT materials were removed from the furnace. A series
of xAu-NT composite materials were prepared by changing the
amount of Au from 0 to 0.3 at% with a fixed urea/Ti4+ ratio of 10.
In xAu-NT, x stands for nominal amount of Au [at%]. We employed
urea as fuel to avoid any carbon impurities so that no further calci-
nation is required. During the combustion process, the in situ gen-
eration of NH3 occurs as a result of urea decomposition, and this
acts as a N source as well as a reduction atmosphere to reduce
Au3+ to Au clusters. Au-NT nanocomposites have been subjected
to detailed characterization and photocatalytic WSR measure-
ments.
Conclusions
We successfully synthesized visible-light-active mesoporous
nano-Au-TiO2ÀxNx composites by using a simple combustion
synthesis method (CSM). The solution combustion synthesis at
4008C employed to prepare the nano-Au on N-doped mesopo-
rous TiO2 (xAu-NT) nanocomposites is predominantly a kineti-
cally controlled process because of the short time of actual
combustion under a high flux of ammonia as a result of
sudden urea decomposition. A greater number of visible-light
photons could be harvested by the surface plasmon resonance
of Au clusters (than TiO2ÀxNx), and charge-carrier mobility is en-
hanced because of the disordered mesoporosity along with
electrically interconnected nanocrystallites of TiO2ÀxNx. This
aspect was further enhanced by preparing well-dispersed Au
particles with p–n heterojunctions, which act as charge-separa-
tion centers. In contrast to the 2 ps lifetime reported for
charge carriers in nano-Au in the literature, we observed
a much longer lifetime (21.1 ps) with our xAu-NT nanocompo-
sites. This helps the electron injection from Au to the conduc-
tion band of TiO2. The electronic integration of nano-Au with
TiO2 was supported by X-ray photoelectron spectroscopy, pho-
toluminescence spectroscopy, lifetime measurements, and
Raman spectroscopy. As (001) nano-Au clusters bind to the
(101) TiO2 facets, a polarized pathway is available through p–n
heterojunctions for charge separation and utilization. In addi-
tion, the presence of disordered mesoporosity greatly reduces
the diffusion length of the charge carriers, and electrically in-
terconnected nanoparticles aid fast charge conduction. These
added advantages as a result of the pseudo-three-dimensional
nature of the mesopores help to utilize the charge carriers effi-
ciently for photocatalysis. The water splitting reaction under
visible
Characterization
PXRD data of TiO2ÀxNx and xAu-NT materials were collected by
using a PANalytical X’pert Pro dual goniometer X-ray diffractome-
ter. A proportional counter detector was used for low-angle experi-
ments. The data were collected with a step size of 0.028 and
a scan rate of 0.58minÀ1. The sample was rotated throughout the
scan to obtain better counting statistics. The radiation used was
CuKa (1.5418 ꢁ) with a Ni filter, and the data collection was per-
formed by using a flat holder in Bragg–Brentango geometry (0.28).
EDX analysis and SEM measurements were performed by using
a SEM system equipped with an EDX attachment (FEI, Model
Quanta 200 3D). EDX spectra were recorded in the spot-profile
mode by focusing the electron beam onto specific regions of the
sample. Calibration of the experiment for N estimation was mea-
sured with several mixtures of GaN and alumina powder mainly to
ensure the reliability of the N estimation. N2 adsorption–desorption
isotherms for the materials were recorded by using a Quantach-
rome autosorb automated gas sorption system (NOVA 1200). The
BET equation was used to calculate the surface area from the ad-
sorption branch. The pore size distribution was calculated by ana-
lyzing the adsorption branch of the N2 sorption isotherm by using
light with 0.05Au-NT generates H2 at 1.5 mmolhÀ1 gÀ1 catalyst,
which is an order of magnitude higher than that of TiO2ÀxNx.
Notably, a comparison of the lifetime of charge carriers and H2
generation under visible light varies linearly between 0.05Au-
NT and TiO2ÀxNx.
We would like to emphasize that the significantly higher life-
time of charge carriers observed in the present study, in com-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 522 – 530 528

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
the BJH method. A FEI TECNAI 3010 electron microscope operating
at 300 kV (Cs=0.6 mm, resolution 1.7 ꢁ) was employed for TEM
measurements of the Au-NT materials. Samples were crushed and
dispersed in isopropanol by sonication before deposition onto
a holey carbon grid. A FEI-TITAN microscope operated at 300 kV
equipped with an FEG source, a Cs (spherical aberration coeffi-
cient) corrector for condenser lens systems, and a high-angle annu-
lar dark-field (HAADF) detector was used to perform scanning
transmission electron microscopy (STEM) experiments. The semi-
convergence angle of the electron probe incident to the specimen
and camera length were maintained 17.8 mrad and 128 mm, re-
spectively, during the experiments, and all images were recorded
with a HAADF detector. All EDX data were measured in spot mode
for 180 s, keeping the other experimental parameters the same as
for the high-resolution Z-contrast imaging, which was acquired for
a dwell time of 20 ms. The specimen for STEM experiments was
prepared by dispersion of the material (0.10 Au-NT) in water and
placing a drop of the dispersion on a Cu TEM grid covered with
a carbon film, which was then dried.
source. The experiments were conducted at around pH 7. The H2
evolved was sampled and analyzed periodically by using a Chemito
8610 GC with a porapak-Q column and a thermal conductivity de-
tector at 353 K.
Acknowledgements
We thank Dr. K. Sreekumar for help in PEC measurements. We ac-
knowledge help from Drs. Jayakannan and Partha Hazra, IISER,
Pune to conduct lifetime measurements at nano- and picosecond
levels and follow-up discussions. K.S., S.R.A., and K.R. thank CSIR,
New Delhi for senior research fellowships. Financial support from
the TAPSUN programme under NWP0056 by CSIR, New Delhi is
gratefully acknowledged. Part of the work is supported by CSISR
under CSC-0404.
Keywords: heterogeneous catalysis · hydrogen · mesoporous
materials · photochemistry · water splitting
Diffuse reflectance UV/Vis measurements were performed by using
a spectrophotometer (Shimadzu, Model UV-2550) with spectral-
grade BaSO4 as the reference material. Raman spectra were record-
ed by using a Horiba JY LabRAM HR 800 Raman spectrometer cou-
pled with a microscope in the reflectance mode with a 633 nm ex-
citation laser source and a spectral resolution of 0.3 cmÀ1. PL meas-
urements were performed by using a Horiba Jobin Yuon Fluorolog
3 spectrophotometer equipped with a 450 W Xe lamp at RT under
the excitation at 280 nm. The conditions were fixed as far as possi-
ble to compare the PL signals directly. The XPS measurements
were performed by using a custom-built ambient-pressure XPS
system from Prevac, Poland, equipped with a VG Scienta SAX 100
emission controller monochromator using an AlKa anode
(1486.6 eV) in transmission lens mode. The photoelectron energy
was analyzed by using a VG Scienta R3000 differentially pumped
analyzer. The spectra were recorded at a pass energy of 50 eV.
[1] S. S. K. Ma, K. Maeda, R. Abe, K. Domen, Energy Environ. Sci. 2012, 5,
8390–8397.
[2] A. Kudo, Int. J. Hydrogen Energy 2007, 32, 2673–2678.
[3] R. M. Navarro, M. C. Alvarez Galvan, J. A. Villoria de La Mano, S. M. Al-
Zahrani, J. L. G. Fierro, Energy Environ. Sci. 2010, 3, 1865–1882.
[4] K. A. Connelly, H. Idriss, Green Chem. 2012, 14, 260–280.
[5] M. Bowker, Green Chem. 2011, 13, 2235–2246.
[6] A. Fujishima, T. N. Rao, D. A. Tryk, J. Photochem. Photobiol. C 2000, 1, 1–
21.
[7] M. Anpo, M. Takeuchi, J. Catal. 2003, 216, 505–516.
[8] W. J. Youngblood, S-H. A. Lee, K. Maeda, T. E. Mallouk, Acc. Chem. Res.
2009, 42, 1966–1973.
[9] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science 2001, 293,
269–271.
[10] S. Sato, Chem. Phys. Lett. 1986, 123, 126–128.
[11] a) M. Sathish, B. Viswanathan, R. P. Viswanath, C. S. Gopinath, Chem.
Mater. 2005, 17, 6349–6353; b) M. Sathish, R. P. Viswanath, C. S. Go-
piath, J. Nanosci. Nanotechnol. 2009, 9, 423–432.
[12] A. Pandikumar, K. Sivaranjani, C. S. Gopinath, R. Ramaraj, RSC Adv. 2013,
3, 13390–13398.
[13] C. S. Gopinath, J. Phys. Chem. B 2006, 110, 7079–7080.
[14] K. Sivaranjani, C. S. Gopinath, J. Mater. Chem. 2011, 21, 2639–2647.
[15] X. Chen, S. S. Mao, Chem. Rev. 2007, 107, 2891–2959.
[16] T. M. Horikawa, M. Katoh, T. Tomida, Microporous Mesoporous Mater.
2008, 110, 397–404.
Fluorescence lifetime measurements were performed by using
a Horiba Jobin Yvon Fluorolog 3 spectrophotometer equipped
with a 450 W Xe lamp. Fluorescence lifetime decays were collected
by using a time-correlated single photon counting (TCSPC) setup
from IBH Horiba Jobin Yvon (USA) equipped with a 375 nm diode
laser (IBH, UK, NanoLED-375 L, lmax =375 nm) that has an FWHM of
89 ps as a sample excitation source.
The PEC properties of the Au/NT materials were measured by
linear sweep voltammetry (LSV) by using an Autolab PGSTAT30
(Eco-Chemie) instrument in a conventional three-electrode test cell
with Ag/AgCl as the reference electrode in 1m NaOH solution
under ambient conditions. To prepare the working electrode, the
F-doped tin oxide (FTO) plates were washed thoroughly with ace-
tone and isopropanol, and the catalyst was slurry made by dispers-
ing catalyst (5 mg) in isopropanol (1 mL), which was drop-casted
on the electrode surface and dried at RT overnight. This electrode
was used as the working electrode for all of the electrochemical
studies. A 400 W medium-pressure Hg vapor UV lamp was used.
[17] X. Chen, T. Yu, X. Fan, H. Zhang, Z. Li, J. Ye, Z. Zou, Appl. Surf. Sci. 2007,
253, 8500.
[18] K. Sivaranjani, S. Agarkar, S. B. Ogale, C. S. Gopinath, J. Phys. Chem. C
2012, 116, 2581–2587.
[19] N. Maity, P. R. Rajamohanan, S. Ganapathy, C. S. Gopinath, S. Bhaduri,
G. K. Lairi, J. Phys. Chem. C 2008, 112, 9428–9433.
[20] P. Maity, C. S. Gopinath, S. Bhaduri, G. K. Lahiri, Green Chem. 2009, 11,
554–561.
[21] S. Basu, H. Paul, C. S. Gopinath, S. Bhaduri, G. K. Lahiri, J. Catal. 2005,
229, 298–302.
[22] T. Mathew, K. Sivaranjani, E. S. Gnanakumar, Y. Yamada, T. Kobayashi,
C. S. Gopinath, J. Mater. Chem. 2012, 22, 13484–13493.
[23] S. Linic, P. Christopher, D. B. Ingram, Nat. Mater. 2011, 10, 911–921.
[24] P. Christopher, H. Xin, S. Linic, Nat. Chem. 2011, 3, 467–472.
[25] G. R. Bamwenda, S. Tsubota, T. Nakamura, M. Haruta, J. Photochem. Pho-
tobiol. 1995, 89, 177–189.
The photocatalytic activity was measured for H2O splitting under
an AM 1.5 source. The reaction was performed under ambient con-
ditions by using
a Borosil photoreactor of 100 mL capacity
[26] M. Murdoch, G. I. N. Waterhouse, M. A. Nadeem, J. B. Metson, M. A.
Keane, R. F. Howe, J. Llorca, H. Idriss, Nat. Chem. 2011, 3, 489–492.
[27] V. Subramanian, E. E. Wolf, P. V. Kamat, J. Am. Chem. Soc. 2004, 126,
4943–4950.
[28] Z. Liu, W. Hou, P. Pavaskar, M. Aykol, S. B. Cronin, Nano Lett. 2011, 11,
1111 – 1116.
equipped with a port for the withdrawal of gas samples at regular
intervals. For each experiment, fresh catalyst (100 mg) was dis-
persed in water (64 mL) and methanol (16 mL) to serve as the sac-
rificial reagent. A 125 W simulated white-light source or a Newport
solar simulator with an AM 1.5 filter was used as the irradiation
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 522 – 530 529

CHEMCATCHEM
FULL PAPERS
www.chemcatchem.org
[29] Z. W. Seh, S. Liu, M. Low, S.-Y. Zhang, Z. Liu, A. Mlayah, M.-Y. Han, Adv.
Mater. 2012, 24, 2310–2314.
[38] T. Ohsaka, F. Izumi, Y. Fujiki, J. Raman Spectrosc. 1978, 7, 321–324.
[39] A. Fujishima, X. Zhang, D. A. Tryk, Surf. Sci. Rep. 2008, 63, 515–582.
[30] B. Naik, K. M. Parida, C. S. Gopinath, J. Phys. Chem. C 2010, 114, 19473–
19482.
[40] Y. Li, N.-H. Lee, D.-S. Hwang, J. S. Song, E. G. Lee, S.-J. Kim, Langmuir
2004, 20, 10838–10844.
[31] a) M. Mapa, C. S. Gopinath, Chem. Mater. 2009, 21, 351–359; b) M.
Mapa, S. RajaAmbal, C. S. Gopinath, Trans. Mater. Res. Soc. Jpn. 2013, 38,
145–158.
[41] K. Nagaveni, M. S. Hegde, G. Madras, J. Phys. Chem. B 2004, 108, 20204–
20212.
[42] K. M. Parida, N. Sahu, A. K. Tripathi, V. S. Kamble, Environ. Sci. Technol.
[32] M. Mapa, K. S. Thushara, B. Saha, P. Chakraborty, C. M. Janet, R. P. Viswa-
nath, C. Madhavan Nair, K. V. G. K. Murti, C. S. Gopinath, Chem. Mater.
2009, 21, 2973–2979.
2010, 44, 4155–4160.
[43] J. C. Yu, J. Yu, W. Ho, Z. Jiang, L. Zhang, Chem. Mater. 2002, 14, 3808–
3816.
[33] M. Mapa, K. Sivaranjani, D. S. Bhange, B. Saha, P. Chakraborty, A. K. Vis-
wanath, C. S. Gopinath, Chem. Mater. 2010, 22, 565–578.
[34] K. Sivaranjani, A. Verma, C. S. Gopinath, Green Chem. 2012, 14, 461–
471.
[44] K. Roy, C. P. Vinod, C. S. Gopinath, J. Phys. Chem. C 2013, 117, 4717–
4726.
[45] V. Subramanian, E. E. Wolf, P. V. Kamat, J. Phys. Chem. B 2003, 107,
7479–7485.
[35] a) A. C. Sunil Sekhar, K. Sivaranjani, C. S. Gopinath, C. P. Vinod, Catal.
Today 2012, 198, 92–97; b) A. C. Sunil Sekhar, C. J. Meera, K. V. Ziyad,
C. S. Gopinath, C. P. Vinod, Catal. Sci. Technol. 2013, 3, 1190–1193.
[36] S. Link, M. A. El-Sayed, J. Phys. Chem. B 1999, 103, 8410–8426.
[37] Y. Tian, T. Tatsuma, J. Am. Chem. Soc. 2005, 127, 7632–7637.
Received: August 26, 2013
Revised: November 7, 2013
Published online on January 21, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemCatChem 2014, 6, 522 – 530 530