Effect of g-C3N4 loading on TiO2-based photocatalysts: UV and visible degradation of toluene

Article abstract of DOI:10.1039/c4cy00226a

g-C₃N₄ was introduced into a highly active anatase TiO₂-based photocatalyst using a simple yet effective impregnation method. The corresponding composite materials were obtained with a carbon nitride weight content ranging from 0.25 to 4 wt.%, and their surface, morphological and structural properties were characterized using BET and porosity measurements, X-ray diffraction, infrared, UV-vis and photoluminescence spectroscopy, and transmission electron microscopy. They were subsequently tested for toluene photoelimination under UV and sunlight-type illumination conditions. The photocatalytic performance of the materials was analysed quantitatively through calculation of the quantum yield of the reaction under all illumination conditions essayed. We observed that the addition of carbon nitride enhances the photoactivity and selectivity to CO₂ of samples containing 0.5 to 1 wt.% g-C₃N₄. Physicochemical characterization presented evidence that the enhanced behaviour is related to the intimate contact between the two components of the photocatalyst system, which provides two new (i.e. not present in the parent system) charge carrier handling routes affecting photocatalytic properties upon UV and visible light illumination. The Royal Society of Chemistry 2014.

Full text of DOI:10.1039/c4cy00226a

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Effect of g-C N loading on TiO -based
3
4
2
photocatalysts: UV and visible
Cite this: Catal. Sci. Technol., 2014,
degradation of toluene†
4, 2006
M. J. Muñoz-Batista, A. Kubacka* and M. Fernández-García*
3 4 2
g-C N was introduced into a highly active anatase TiO -based photocatalyst using a simple yet effective
impregnation method. The corresponding composite materials were obtained with a carbon nitride
weight content ranging from 0.25 to 4 wt.%, and their surface, morphological and structural properties
were characterized using BET and porosity measurements, X-ray diffraction, infrared, UV-vis and photo-
luminescence spectroscopy, and transmission electron microscopy. They were subsequently tested for
toluene photoelimination under UV and sunlight-type illumination conditions. The photocatalytic perfor-
mance of the materials was analysed quantitatively through calculation of the quantum yield of the reac-
tion under all illumination conditions essayed. We observed that the addition of carbon nitride enhances
the photoactivity and selectivity to CO2 of samples containing 0.5 to 1 wt.% g-C N . Physicochemical
Received 19th February 2014,
Accepted 25th March 2014
3
4
characterization presented evidence that the enhanced behaviour is related to the intimate contact
between the two components of the photocatalyst system, which provides two new (i.e. not present in
the parent system) charge carrier handling routes affecting photocatalytic properties upon UV and visible
light illumination.
DOI: 10.1039/c4cy00226a
www.rsc.org/catalysis
2
light-sensitive phases in intimate contact with TiO has been
shown particularly successful.
1
. Introduction
1–4
Heterogeneous photocatalysis is an advanced oxidation process
that uses nanocrystalline semiconductors broadly applied in
the degradation/transformation of organic pollutants as well
as biological microorganisms. This is essentially based on the
excellent performance and stability of titania, the most promi-
nent photocatalytic material, for the mineralization of typical
pollutants or the inactivation of dangerous microorganisms
typically carried out under mild conditions, e.g. at room tem-
perature and atmospheric pressure, and using oxygen (air) as
In the photocatalytic field, a relatively new sensitizer
material type corresponds to certain C-containing materials
such as carbon nanotubes, graphene, or carbon nitrides.
4
Contrary to classical carbon materials, they show by themselves
interesting photoredox properties. Also they have captured
attention from the point of view of boosting the activity of
titania. As such, they provide not only significant (classical)
adsorption properties but also charge carrier storage after light
5–7
excitation, as well as specific visible light capabilities.
Among the new C-containing materials, nanostructured
carbon nitrides, particularly g-C , the most stable allotropic
1–4
an oxidant.
However, the main limitation of anatase, the
4
stable polymorph of titania at the nanometric level, comes
from its relatively wide band gap energy, ca. 3.0–3.4 eV, which
requires UV radiation for triggering photoredox reactions.
Because UV radiation only accounts for ca. 3–5% of the solar
spectrum, the photocatalytic efficiency of pure titania-based
photocatalysts under sunlight is limited. The quest for fruitful
use of the visible part of the solar spectrum, ca. 45%, has
forced implementation of several strategies aiming at profiting
from the complete solar spectrum, particularly the visible
region. Among the strategies implemented, the use of visible
3 4
N
phase, have displayed interesting properties in several photo-
chemical reactions such as water splitting or dye degrada-
8–10
tion.
3 4
Similarly, the use of g-C N in combination with
titania has been shown to provide outstanding photochemical
1
1,12
activity in hydrogen production
as well as a series of deg-
radation reactions concerning several pollutants such as
1
3
14
15–18
formaldehyde, phenol and organic dyes,
including the
1
5
removal of Cr(VI). In all cases tested, the improvement of
performance against titania is constant throughout the studies,
but the visible light capability is not always better than the
one of carbon nitride. This, however, needs critical assessment
as the comparison was presented using bare reaction rates
and not efficiency parameters. Such an issue can limit the
generalization of the results presented and thus complicate
Instituto de Catálisis y Petroleoquímica, CSIC, C/Marie Curie 2, 28045-Madrid,
Spain. E-mail: ak@icp.csic.es, mfg@icp.csic.es
a,s
†
Electronic supplementary information (ESI) available: Details for e calculation
are included. See DOI: 10.1039/c4cy00226a
2006 | Catal. Sci. Technol., 2014, 4, 2006–2015
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
the interpretation of the effectiveness of the g-C N –TiO com-
at 110 °C for 24 h. The sample names were Ti for the TiO2
reference, and xg-C N /Ti for the composite ones, where x is
the wt.% (0.25, 0.5, 1, 2 and 4) of g-C N with respect to TiO .
3 4 2
This was confirmed with an error below 5% using ICP-AAS
spectrometry.
3
4
2
posite materials. The calculation of efficiency parameters,
particularly the quantum yield of the reaction, involves the
measurement and modelling of the radiation that reached
the surface of the material and the fraction effectively used in
the photocatalytic steps. Such a task typically involves, in the
first place, measurement and modelling of the light source
emission properties and, subsequently, the light absorption
capability of the photocatalysts. The latter requires calculation
3
4
2.2 Characterization details
The Brunauer–Emmett–Teller (BET) surface areas and average
pore volumes and sizes were measured by nitrogen physisorption
of the radiative transfer by calculating in our case the so-called
(
Micromeritics ASAP 2010). XRD profiles were obtained using a
a,s 19–23
local superficial rate of photon absorption, e
.
Seifert D-500 diffractometer using Ni-filtered Cu Kα radiation
with a 0.02° step and fitted using the Von Dreele approach to
Here we will provide quantitative calculation of the toluene
photoelimination quantum yield in order to obtain a reliable
comparison among the g-C N –TiO catalysts. The reaction will
2
8
the Le Bail method; the particle sizes and microstrain were
3
4
2
29
measured by XRD using the Williamson–Hall formalism.
be carried out under UV as well as sunlight-type illumination
conditions. Toluene is a typical pollutant present in urban
atmospheres which demands highly active photocatalysts for
TEM analysis of the materials was carried out using a JEOL
2100F TEM/STEM microscope. Energy-dispersive X-ray spec-
troscopy (XEDS) was performed in STEM mode with a probe
size of ~1 nm using the INCA x-sight (Oxford Instruments)
detector. UV–vis transmission or diffuse-reflectance spectros-
copy experiments were performed on a Shimadzu UV2100
24,25
its elimination.
In addition, and with the aim of inter-
preting the quantum yield or efficiency results on a physico-
chemical basis, we carried out a multitechnique examination
of the materials using X-ray diffraction, infrared, photo-
luminescence and ultraviolet spectroscopy, and transmission
electron microscopy. Applying these tools, we provide evidence
that the underlying activity-enhancement factors with respect
to the titania reference are not the same under UV and visible
light illumination conditions. In spite of this, the composite
provides in all cases a route to improve the performance, e.g.
the quantum yield and selectivity to total oxidation, of the par-
ent materials. We thus observed that the combination of these
apparatus using, for diffuse-reflectance experiments, BaSO
4
as a reference. Photoluminescence spectra were measured at
room temperature on a fluorescence spectrophotometer
(
Perkin Elmer LS50B). The Fourier transform infrared spectra
were taken in a Bruker Vertex 80 FTIR spectrometer using a
MCT detector.
2.3 Photocatalytic experimental details
two materials with a loading of g-C
range yielded photocatalysts overperforming the parent refer-
ence materials under UV, visible and sunlight conditions.
3
N
4
in the 0.5–1.0 wt.%
Gas-phase photooxidation of toluene (Panreac, spectroscopic
grade) was carried out in a continuous flow annular photoreactor
30,31
using a set-up described elsewhere
(details in the ESI†).
The activity and selectivity for gas-phase photooxidation were
obtained in a continuous flow annular photoreactor containing
ca. 40 mg of the photocatalyst as a thin layer coating on a pyrex
tube. The corresponding amount of catalyst was suspended in
1 mL of ethanol, painted on a pyrex tube (cut-off at ca. 290 nm),
2
. Materials and methods
2
.1 Catalyst preparation
Materials were prepared using a single-pot microemulsion
preparation method using n-heptane (Scharlau) as the organic
medium, Triton X-100 (Aldrich) as a surfactant and hexanol
−1
and dried at RT. The reaction mixture (100 ml min ) was pre-
pared by injecting toluene (≥99%; Aldrich) into a wet (ca. 75%
relative humidity) 20 vol.% O /N flow before entering the
(Aldrich) as a cosurfactant. A TiO2 reference sample was
2
2
obtained as a first step using a microemulsion into the aque-
ous phase and titanium tetraisopropoxide as the precursor.
Titanium tetraisopropoxide was introduced into the previously
resulting microemulsion drop by drop from a mixture with
isopropanol (2: 3). Water/Ti and water/surfactant molar ratios
photoreactor, yielding an organic inlet concentration of ca.
700 ppmv. Under such conditions, the reaction rate shows a
zero-order kinetics with respect to the total flow and organic
pollutant/oxygen concentrations. After flowing the mixture for
6 h (control test) in the dark, the catalyst was irradiated by
four fluorescent daylight lamps (6W, Sylvania F6W/D) with a
radiation spectrum simulating sunlight (UV content of 3%;
main emission lines at 410, 440, 540, and 580 nm), symmetri-
cally positioned outside the photoreactor. Similar tests were
carried out using UV lamps (Sylvania F6WBLT-65; 6W, maxi-
mum at ca. 350 nm). Reaction rates were evaluated (vide supra)
under steady-state conditions, typically achieved after ca. 6–10 h
from the start of irradiation. No change in activity was
detected for all samples within the next 24 h. The concentra-
tions of the reactants and products were analyzed using an
on-line gas chromatograph (Agilent GC 6890) equipped with
26,27
were, respectively, 110 and 18 for all samples.
The resulting
mixture was stirred for 24 h and then centrifuged, and the
separated solid precursors were rinsed with methanol and
dried at 110 °C for 12 h. After drying, the solid precursors
−1
were subjected to a heating ramp of 2 °C min up to 500 °C.
The graphitic carbon nitride was obtained by calcination
of melamine (Aldrich) at 580 °C with a heating ramp of 5 °C
−
1
min for 4 h. The incipient wetness impregnation method
was used to obtain the carbon nitride–TiO composites. For
this, an appropriate amount of g-C N was suspended into
2
3
4
methanol and sonicated for 1 h, deposited on TiO and dried
2
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 2006–2015 | 2007

View Article Online
Paper
Catalysis Science & Technology
Table 1 Morphological properties of the samplesa
HP-PLOT-Q/HP-Innowax columns (0.5/0.32 mm I.D. × 30 m)
and TCD (for CO measurement)/FID (organic measurement).
The carbon balance was above 95% in all experiments.
2
BET surface
area (m g
2
−1
3
−1
Sample
Ti
)
Pore volume (cm g
)
Pore size (nm)
91.8
/Ti 91.5
0.123
0.119
0.119
0.118
0.119
0.104
0.103
5.2
5.2
2
.4 Quantum yield or efficiency
0
0
1
2
.25g-C
3
N
4
3
.5g-C N
4
/Ti
89.4
92.6
90.9
88.2
26.8
5.1
5.1
5
The classical formulation of the quantum efficiency requires
the calculation of the ratio between the number of mole-
cules reacting and the number of photons interacting with
g-C
g-C
3
N
4
N
4
/Ti
/Ti
3
5.2
5
4g-C N /Ti
3
4
4.7
15.1
3
2,33
g-C N
3 4
the catalyst:
a
2
−1
Standard error: BET, 1.5 m g ; porosity, 8%.

2
1
Rate

mol m
s

2
Q.E.% 
100
1
(1)
Photon rate

Einstein m
s

The X-ray diffraction patterns of the g-C
are shown in Fig. 1. The technique only detects the presence
3
N
4
–TiO
2
materials
of the anatase (I4 /amd space group) phase. Table 2 lists the
data for the primary particle size, strain and cell parameters
of the anatase phase in all samples.
The nanostructured anatase phase displays a particle size
always close to 12 nm. Furthermore, results concerning strain
and cell parameters strongly suggest that the contact with the
carbon nitride phase does not alter the main structural
parameters of anatase. In particular, the absence of doping
can be foreseen. The constancy of the anatase structural and
To evaluate the denominator of eqn (1), we first evaluate
the local superficial rate of photon absorption, e , of the sam-
1
a,s
ples. The mathematical model of the present sample–reactor
system allowing such calculations is presented in the ESI .† This
provides an “exact” calculation of the number of photons per
surface area and unit time. Such calculation considers the reac-
tor–photocatalyst geometry as well as the optical properties of
a,s
the photocatalyst. Once e is obtained in the desired units
(
eqn (1)) we need to consider the selectivity of the reaction as
two different molecules were obtained in the toluene photo-
oxidation reaction. A general equation for the quantum efficiency
would be obtained using a correction factor (here called γ) of
a,s
the e presented as follows:
TOL

i ni S
i
(2)
where i runs over all products of the reaction, S
i
is the frac-
tional selectivity to product i, and n is the inverse of the num-
i
ber of charge carrier species required to obtain the specific
product i. This correction factor accounts for the number of
charge carrier species (and in turn photons) necessary to pro-
duce one molecule of a specific product and requires mea-
surement of the reaction selectivity. This is possible in the
case of toluene photooxidation due to both its relatively well
established mechanism as well as the limited number of
2
4,25,34
Fig. 1 XRD diffractograms of the Ti and g-C
xg-C /Ti samples.
3 4
N references and
products detected in a gas-phase experiment.
3
N
4
3
. Results and discussion
The physicochemical characterization results for the xg-C
samples are summarized in Table 1.
3 4
N /Ti
Table 2 XRD-derived parameters for the anatase phase of the
a
samples
Although the g-C N reference presents a relatively low
3
4
TiO
2
anatase
2
−1
surface area of 26.8 m g , the composite materials display a
surface area very close to that of the titania reference within
2
1/2 cell parameters (Å)
Size (nm) Microstrain <ξ >
−3
Sample
Ti
TiO
2
(× 10 ) TiO
2
a
b
2
−1
the experimental error (ca. 89.5 ± 1.5 m g ). This is an
expected result due to the relatively mild treatment (in com-
parison with the anatase calcination treatment) carried out to
contact the carbon nitride and anatase components. As
shown in Table 1, this conclusion is extensible to all morpho-
logical properties including average pore volume and size of
the materials.
12.1
/Ti 12.2
2.08
2.10
2.10
2.11
2.09
2.10
5
3.789
3.785
3.790
3.787
3.787
3.787
9.477
9.471
9.475
9.471
9.476
9.474
0
0
.25g-C
.5g-C
3
N
4
3
N
4
/Ti
12.0
12.0
12.0
12.1
5
1g-C
N
3 4
N
3 4
N
3 4
/Ti
/Ti
/Ti
2
4
g-C
g-C
a
Standard errors: 12% (size), 10% (strain), ±0.004 Å (cell parameters).
2008 | Catal. Sci. Technol., 2014, 4, 2006–2015
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
morphological characteristics is important in order to inter-
pret the photocatalytic ability of the composite samples. Such
constancy is, as previously mentioned, a relatively expected
result due to the mentioned mild treatment applied to titania
semiconductor components and/or (less likely) to minor
morphological differences concerning the carbon nitride.
To analyze the interaction between the components, we car-
ried out a TEM analysis of the samples. The g-C N reference
3
4
(
with respect to its calcination) to facilitate the contact with
shows its characteristic platelet-like, layered morphology (Fig. 3).
The presence of carbon nitride is, however, difficult to track in
the composite materials due to both its low weight content in
the samples and its low atomic number constituents. In spite
of this, visual inspection of Fig. 3 evidences the presence of
g-C N platelets surrounded by crystalline oxide (see upper-left
the carbon nitride component. In Fig. 1 we can also observe
that the g-C N nanomaterial displays a XRD profile typical of
its graphitic structure, dominated by the interlayer-stacking
3
4
3
5,36
(002) peak.
As said, no evidence of its presence in the
composite materials can be achieved using XRD.
3
4
To provide information of the carbon nitride component
in g-C N –TiO composite samples, we carried out an infra-
3 4 2
corner of the micrograph as a representative example), indicat-
ing the close relationship established between the components
in the samples. The presence of the carbon nitride phase at the
upper-left corner of the graph was corroborated by XEDS. In
addition, this technique measured an average carbon nitride
contribution corresponding to 3.05 wt.% (reasonably close to
the formal composition) in the micrograph presented in Fig. 3.
The optical properties of the catalysts were examined
using UV-visible spectroscopy (Fig. 4). The composite mate-
rials show spectra dominated by the semiconducting nature
of titania, displaying the intensity decay expected for a band
gap energy of 3.2 eV (Table 3).
red study. Fig. 2 shows the infrared spectra of the materials.
In the case of the g-C N reference, from high to lower
3
4
wavenumbers, we detected the presence of N–H stretching
−
1
vibration contribution(s) at the 3500–2500 cm region. This
is related to residual NH groups, although the presence of
x
1
1
adsorbed water molecules may also need to be considered.
After it, in the 1600–1200 cm region, we can observe several
contributions mostly associated with the N–C stretching modes
−
1
−1
of heterocyclic compounds. Finally, around 850–800 cm we
note additional contributions coming from the breathing
3
7
6 7
modes of tris-s-triazine (C N -based) building blocks. The
same contributions can be observed in composite catalysts
having a carbon nitride content equal or superior to 1 wt.%. Of
course, in some regions, typically water or hydroxyl related
2
ones, they are overlapped with intrinsic TiO contributions.
−1
The region below ca. 1600 cm displays, however, clear carbon
nitride like fingerprints in the composite catalysts. Specifically,
−1
in the 1600–800 cm region, although some of the bands
present differences between the composites and the g-C N
3
4
reference, the constancy of the tris-s-triazine breathing
mode frequency indicates that the carbon nitride component
maintains its main structural characteristics in the g-C N –TiO
2
3
4
materials. Differences in other (N–C stretching) bands are
attributable to the effects of the interaction between the two
Fig. 2 FTIR spectra of the Ti and g-C
3
N
4
references and xg-C
3
N
4
/Ti
3 4 3 4 2
Fig. 3 TEM images of the g-C N (upper panel) and 4g-C N /TiO
samples.
(lower panel) samples.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 2006–2015 | 2009

View Article Online
Paper
Catalysis Science & Technology
3 4 3 4
Fig. 4 UV-vis spectra of the Ti and g-C N references and xg-C N /Ti
samples.
Table 3 Band gap energy of the samplesa
Sample
Ti
Band gap (eV)
3.2
3.0
0
0
1
2
4
.25g-C
3
N
4
/Ti
3
.5g-C N
4
/Ti
3.0
3.0
3.0
3.0
2.7
5
g-C
g-C
g-C
N
3 4
N
3 4
N
3 4
/Ti
/Ti
/Ti
5
Fig. 5 Surface-area-normalized reaction rate for toluene photooxidation.
A. Sunlight irradiation. B. UV irradiation. Lines are only guides.
3 4
g-C N
a
Standard error: 0.05 eV.
samples with activity below that of the Ti reference. In both
cases, UV and sunlight-type excitation, the composite with
4 wt.% g-C N displayed lower activity than the pure g-C N
The band gap was calculated considering titania as an
3
4
3 4
3
8
indirect band gap semiconductor. The composite materials
present a slightly inferior but essentially constant band gap
energy through the series. g-C N , as an indirect band gap
reference.
The presence of the carbon nitride polymer also affects
selectivity although moderately (Fig. 6). All samples generate
3
4
semiconductor, shows a band gap energy of ca. 2.7 eV, in
accordance with literature reports. The UV-vis spectra of the
composite catalysts present intensities above ca. 400 nm; a
fact not observed in titania (see the inset in Fig. 4). Compari-
2
benzaldehyde and CO as reaction products. Total oxidation
7
is favored in some composite samples both under UV and
sunlight-type illumination conditions. Composite samples
with lower quantities of the carbon nitride component dis-
play a reasonable similarity with respect to the parent titania
3 4
son with the g-C N reference makes it obvious that this is
a contribution from the carbon nitride component which
reference, although with a modest increase in CO production.
2
expands the absorption capability of the TiO -based compos-
3 4
The 4g-C N /Ti sample shows however a more similar behav-
2
ite materials into the visible region of the electromagnetic
spectrum.
iour to the carbon nitride reference material. We can thus see a
continuous shifting in selectivity with the increase in the car-
bon nitride component; total oxidation seems continuously
The photochemical properties of the materials under
sunlight-type and UV illumination are presented in, respec-
tively, the upper and bottom panels of Fig. 5. The figure pre-
sents pseudo steady-state reaction rates after 24–30 h, with
constant rate values (within the experimental error) for more
than 20–24 h. This indicates that the catalysts have reason-
able stability. The reaction rates, as a function of the sample's
3 4
favoured up to the 1g-C N /Ti sample. For this sample, the
partial oxidation pathway becomes increasingly important.
As mentioned, here we aim to analyze quantitatively the
differences among our samples. The first step in calculating
quantitative quantum efficiencies corresponds to the numeri-
cal calculation of the local superficial rate of photon absorp-
a,s
g-C
3
N
4
content, present an inverse U shape in both cases. A
tion (e ) at the sample surface. In our set-up (ESI†) this
maximum is always obtained in the intermediate region for
requires calculation on a cylindrical geometry and thus two
independent geometrical variables are needed. This is presented
in Fig. 7 (UV) and 8 (sunlight-type) in Cartesian coordinates.
The two dimensional behaviour of all samples is identical
and thus differences can be expressed by a single number.
0
.5 or 1.0 wt.% samples.
Maximum enhancement factors of ca. 1.4/1.3 were observed
for, respectively, UV and sunlight-type illumination. After the
maximum, the activity decay is sufficiently strong as to obtain
2010 | Catal. Sci. Technol., 2014, 4, 2006–2015
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
Fig.
2
6 Yields toward benzaldehyde and CO production under
sunlight (A) and UV (B) irradiation. Standard error: ±10%.
−
2
−1
Fig. 7 Local rate of photon absorption (Einstein cm
s ) under UV
a,s
Under UV the e coefficient increased with the carbon nitride
irradiation. From top to bottom: Ti, 0.25g-C N /Ti, 0.5g-C N /Ti,
3
4
3
4
1
3 4 3 4 3 4 3 4
g-C N /Ti, 2g-C N /Ti, 4g-C N /Ti and g-C N .
content of the photocatalysts by ca. 5% while under sunlight
the behaviour is more acute and goes up to 50%. The second
factor to be considered is the one related to eqn (2). This
requires the selectivity data from Fig. 6 as well as the constant
values n for benzaldehyde and CO . For the former we can
2
count the hole-related species for the initial steps leading to
benzaldehyde (Scheme 1).
a simpler calculation of the so-called photonic efficiency
(which does not account for the effect of the optical prop-
erties of the samples in light absorption, with their real
33,42,43
surface exposed to reactants and other variables)
pre-
In this case, it is well known that only one charge carrier
species is required to provide the radical species presented
at the first stage depicted in Scheme 1. In the presence of
sents in the present case up to two or three orders of mag-
nitude variation between the true (quantum) and apparent
photonic efficiencies.
(
molecular) oxygen, such an intermediate evolves in the desired
The maximum in our calculation is, in any case, obtained
for the 0.5g-C N /Ti sample both for UV and sunlight-type
3 4
2
5,39–41
product, i.e. benzaldehyde.
For the production of CO
2
the situation is less clear but, on average, four hole-related
excitation. A more pronounced maximum is observed in the
first case while in the second we observed a flat maximum
−
39–41
2
species (and corresponding O ˙ species) are required.
The
ratio between the number of charge carrier species used for
CO2 and benzaldehyde generation is in any case the most
important parameter to obtain an accurate measurement of
the quantum yield throughout the series of samples.
3 4 3 4
region covering two samples: 0.5g-C N /Ti and 1g-C N /Ti.
The plot indicates that optimum profiting of light in the
“whole” UV-visible electromagnetic region is achieved in the
composition region limited by these two photocatalysts. We
thus observed a consistent increase in the efficiency of the
reaction when carbon nitride is contacted with titania in ade-
quate proportions.
Calculation of the quantum yield or efficiency (Q.Y./Q.E.)
rendered the values presented in Fig. 9. First to note is
−
3
that the Q.Y. values are rather small, always below 10 .
Q.Y. values below 1 seem typical for gas-phase photocatalytic
To interpret the physical origin of the enhancement factor,
we carried out a photoluminescence experiment under UV
4
2,43
reactions.
Concerning this issue, we must point out that
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 2006–2015 | 2011

View Article Online
Paper
Catalysis Science & Technology
and visible light excitation. The corresponding experimental
results are presented in the two panels of Fig. 10. UV excita-
tion at 280 nm allows scanning of all potential de-excitation
channels of the titania component. Above 2.0 eV the photo-
luminescence titania samples display the presence of two or
4
4–46
three peaks.
In addition, low energy peaks are less impor-
tant in generating photochemistry and always related to the
4
7
presence of defective Ti (interstitial or lattice) states. The
peaks of about 2.0 eV can be grouped into two types of transi-
tions corresponding to high-energy (sometimes called green)
and low-energy (red) de-excitation channels. Although we lack
a definitive interpretation of the photoluminescence spectra
of anatase materials, the first transition mostly probes the
de-excitation of excited electrons in localized (typically oxygen-
related) states by trap-related holes, while the second probes
the de-excitation of electron localized states by holes at the
44–46
valence band.
In the upper panel of Fig. 10 we can observe an anatase-
type photoluminescence pattern in the composite samples
up to the 1g-C N /Ti sample but with enhanced intensity.
3 4
Due to the fact that this showed enhanced activity, this
strongly suggests that the composite system provides a new
route for de-excitation involving the charge carrier species of
the two components. This fact is further suggested by the
−2
−1
Fig. 8 Local rate of photon absorption (Einstein cm
s
) under
sunlight irradiation. From top to bottom: Ti, 0.25g-C
g-C /Ti, 2g-C /Ti, 4g-C /Ti and g-C
3 4 3 4
N /Ti, 0.5g-C N /Ti,
1
3
N
4
3
N
4
3
N
4
3 4
N .
Scheme 1 Simplified reaction mechanism of toluene photooxidation.
Fig. 9 Quantum yield of the Ti and g-C
3
N
4
references and xg-C
3
N
4
/Ti
3 4 3 4
Fig. 10 PL spectra of the Ti and g-C N references and xg-C N /Ti
samples under UV and sunlight-type illumination. Lines are only guides.
samples. A. λext = 280 nm. B. λext = 420 nm. Note the scale factor of Ti.
2012 | Catal. Sci. Technol., 2014, 4, 2006–2015
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
photoluminescence graph presented in the ESI.† Comparison
of the composite photoluminescence spectra corresponding
to the appropriate g-C N /TiO physical mixture reinforces
3 4 2
de-excitation channel is present in the composite materials.
The de-excitation channel observed under visible light excita-
tion is also depicted in Scheme 2. This channel may withdraw
electrons from the carbon nitride phase and may involve the
localized states of one or both components. The different effi-
ciencies of the de-excitation channel in the different samples
can be related to the specific energy situation of the localized
states with respect to the corresponding Fermi edge of the
the conclusion that composite materials display additional
de-excitation channels not present in the parent materials.
According to the band diagram (see Scheme 2) the compo-
nent interaction would generate a de-excitation channel with-
drawing electrons from the anatase phase and permitting
holes, the active species detected at anatase surfaces in tolu-
4
9
composite. The correct energy position may allow maximum
overlapping with the corresponding final states of the excited
electrons, increasing the likehood of charge carrier transfer
between components. This hypothesis needs however further
assessment. In any case, the interpretation may suggest that
the g-C N surface is active under visible light excitation and
2
3,25,39,40
ene photodegradation,
to be involved in the chemical
steps. Such interpretation is consistent with the higher inten-
sity increase observed in the ca. 2.75–2.25 eV region of
Fig. 10. The new de-excitation channel evidences the creation
of a Z-“enhancement scheme” of the reaction (Scheme 2). This
has been claimed previously to exist in other carbon nitride
3
4
would contribute to the overall activity of the materials. In such
case, the active species for toluene photoattack is not obvious
as it is well known that carbon nitride does not generate holes
with enough chemical potential to produce OH˙ radical
4
8
containing composite samples with oxide semiconductors.
Above the 1g-C N /Ti sample (i.e. samples with higher
3
4
carbon nitride content), we detected a photoluminescence
pattern resembling that of the carbon nitride parent compound
but with decreasing intensity. This may be rationalized in terms
of the previously mentioned reduction of holes at the carbon
nitride component (Scheme 2) although a unique interpretation
of the behaviour is not possible.
50
species. This fact may differentiate the carbon nitride from
its anatase counterpart. The situation is relatively complex to
analyse; on the one hand, the presence of a carbon nitride
“active” surface is in agreement with the majority of previously
reported contributions concerning g-C
rials in contact with photoactive semiconductors.
3 4
N -containing mate-
1
1–18
Upon visible light excitation (420 nm), a different de-excitation
behaviour is reflected in the luminescence spectra (lower panel
of Fig. 10). Here the spectra of the composite samples are
dominated by a broad, featureless peak characteristic of the
On the
contrary, in our case, the UV and sunlight-type chemical
behaviour(s) through the sample series is (are) relatively
similar (Fig. 9), indicating that the hypothetical contribution
of the second de-excitation mechanism may not determine
the series behaviour upon visible light excitation.
7
carbon nitride component. This is an expected result consid-
ering the band gap energies of the two component materials
of the photocatalysts (Table 3).
We finally would like to further stress that the behav-
iour of the photoluminescence spectra depicted in Fig. 10
for the two excitation wavelengths tested indicates the
contact between the two components of the composite sys-
tem. This contact seems critical to maximise/optimise the
photocatalytic properties of samples having a carbon nitride
content between 0.5 and 1 wt.%. So, independent of the
exact electronic pathway used to establish the electronic
contact between components, its effectiveness seems in all
cases (i.e. under UV and visible excitation) strongly related
to the extension of the interface contact. The complete inter-
pretation of the photocatalytic properties is nevertheless
complex and they may not be exclusively driven by the
two de-excitation mechanisms discussed in the composite
materials but can be additionally affected by factors related
to the shadowing of the anatase surface (in the presence
of the carbon nitride component): small, unnoticed but sig-
nificant modification of the interaction between components
as a function of the carbon nitride content, morphology
effects related to the modest variations presented in Table 1,
small enhancement effects of the pollutant adsorption
capability in the presence of carbon nitride in the composite
samples (although the increment in surface area in Table 1
would also indicate the modest potential of this point),
as well as other unknown contributions which may need to
be considered for full interpretation of the photocatalytic
properties.
Upon visible light excitation, a decrease in the photo-
luminescence intensity is observed, of which the minimum
was reached with samples having carbon nitride contents
below or equal to 0.5 wt.%. Above such minimum the intensity
increases. The behaviour can be interpreted as if an additional
Scheme Schematic view of the charge carrier(s) behavior in
2
xg-C N /Ti samples under UV, visible and sunlight-type illumination.
3
4
Solid lines correspond to main de-excitation channels of the single
components observed by photoluminescence (color code indicative of
their characteristic energy). Dashed lines are de-excitation channels
characteristic of the composite samples.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 2006–2015 | 2013

View Article Online
Paper
Catalysis Science & Technology
8
X. Wang, K. Maeda, X. Chen, K. Tanabe, K. Domen, Y. Hou,
X. Fu and M. Antonietti, J. Am. Chem. Soc., 2010, 131, 1680.
Conclusions
In this work we synthesized g-C
3
N
4
–TiO
2
composite materials
9 L. Gei and C. Han, Appl. Catal., B, 2011, 100, 108.
10 S. Chu, Y. Wang, X. Fan and Z. Zou, ACS Catal., 2013, 3, 912.
11 H. Yan and H. Yang, J. Alloys Compd., 2011, 509, L26.
12 S. Obregón and G. Colón, Appl. Catal., B, 2014, 144, 775.
13 J. Yu, S. Wang, J. Low and W. Xiao, Phys. Chem. Chem. Phys.,
2013, 15, 16883.
with the content of carbon nitride in the 0.25 to 4 wt.% range.
A simple impregnation method provides high surface area
materials resembling the anatase component. Physicochemical
characterization indicates that the anatase phase is unaltered
in the presence of g-C N while the minority component
3
4
maintains its main structural and electronic properties.
14 C. Miranda, H. Mansilla, J. Yáñez, S. Obergón and G. Colón,
J. Photochem. Photobiol., A, 2013, 253, 16.
The performance of the g-C N –TiO composite materials
3
4
2
was tested for gas-phase toluene photoelimination under
UV and sunlight-type illumination conditions. Full calcula-
tion of the quantum yield of the reaction allows comparison
of the activity of the samples on a quantitative basis. From
photocatalytic results we provide evidence that the presence
of carbon nitride in the 0.5 to 1 wt.% range enhances both
the activity and selectivity (to total oxidation) of the reaction
irrespective of the illumination conditions. Analysis of
de-excitation after light excitation indicates that the perfor-
mance of the photocatalysts is intimately related to the pres-
ence of de-excitation channels characteristic of the composite
material, i.e. not present in the single parent components.
Under UV excitation the component contact provides a Z-scheme
route for charge carrier recombination which increases the
probability of the anatase active hole species to reach the
surface and be involved in the chemical steps of photo-
degradation. Under visible excitation we observe the predomi-
nance of an alternative channel which would eliminate the
electrons of the carbon nitride component and may boost the
performance of its surface activity during degradation of tolu-
ene. In both cases (UV and visible light), the interface contact
between the two components appears to critically influence
the recombination and thus the overall photoactivity perfor-
mance of the g-C N –TiO materials.
15 K. Shidharan, E. Jang and T. J. Park, Appl. Catal., B,
2013, 142–143, 718.
16 X.-J. Wang, W.-Y. Yang, F.-T. Li, Y.-B. Xue, R.-H. Liu and
Y.-J. Llao, Ind. Eng. Chem. Res., 2013, 52, 17140.
17 N. Boohprokov, N. Wetchakon, D. Waxler and
B. Inceesungrom, J. Colloid Interface Sci., 2014, 417, 402.
18 L. Zhang, D. Jing, X. She, H. Liu, D. Yang, Y. Lu, J. Li,
Z. Zheng and L. Gou, J. Mater. Chem. A, 2014, 2, 2071.
19 G. E. Imoberdorf, A. E. Cassano, O. M. Alfano and
H. A. Irazoqui, AIChE J., 2006, 52, 1814.
20 L. Zhang and W. A. Anderson, Chem. Eng. J., 2010, 65, 1513.
21 Q. L. Yu, M. M. Ballari and H. J. H. Brouwers, Appl. Catal., B,
2010, 99, 58.
22 A. L. Zazueta, H. Destallaits and G. L. Puma, Chem. Eng. J.,
2013, 217, 475.
23 M. J. Muñoz-Batista, A. Kubacka, M. N. Gómez-Cerezo,
D. Tudela and M. Fernández-García, Appl. Catal., B,
2013, 140–141, 626–635.
24 J. Mo, Y. Zhang, Q. Xu, Y. Zhu, J. J. Lamson and R. Zhao,
Appl. Catal., B, 2009, 89, 570–576.
25 M. J. Muñoz-Batista, M. N. Gómez-Cerezo, A. Kubacka,
D. Tudela and M. Fernández-García, ACS Catal., 2014, 4, 63.
26 P. G. De Gennes and C. Taupin, J. Phys. Chem., 1982, 86,
2294.
3
4
2
2
7 M. Fernández-García, X. Wang, C. Belver, J. C. Hanson and
J. A. Rodriguez, J. Phys. Chem. C, 2007, 111, 674.
8 A. Le Bail, H. Duroy and J. L. Forquet, Mater. Res. Bull.,
1988, 23, 447.
Acknowledgements
2
A. Kubacka and M. J. Muñoz-Batista thank the MINECO for
support through, respectively, the postdoctoral “Ramón y
Cajal” and predoctoral FPI programs. Financial support from
MINECO is also acknowledged (project CTQ2010-14872/BQU).
29 G. K. Williamson and W. H. Hall, Acta Metall., 1953, 1, 22.
30 A. Fuerte, M. D. Hernández-Alonso, A. J. Maira, A. Martínez-Arias,
M. Fernández-García, J. C. Conesa, J. Soria and G. Munuera,
Chem. Commun., 2001, 2718.
3
1 A. Fuerte, M. D. Hernández-Alonso, A. J. Maira, A. Martínez-Arias,
M. Fernández-García, J. C. Conesa, J. Soria and G. Munuera,
J. Catal., 2002, 212, 1.
Notes and references
1
2
3
4
M. R. Hoffman, S. T. Martin, W. Choi and D. W. Bahneman,
Chem. Rev., 1995, 95, 69.
O. Carp, C. L. Huisan and A. Reller, Prog. Solid State Chem.,
32 B. Ohtani, Chem. Lett., 2008, 37, 167.
33 C. Baumanis and D. W. Bahnemann, J. Phys. Chem. C, 2008,
112, 19097.
34 M. Sleiman, P. Conchon, C. Ferronato and J.-M. Chovelon,
Appl. Catal., B, 2009, 86, 159.
2
004, 32, 33.
H. Thu, M. Karkmaz, E. Puzenat, C. Guillard and J. M. Herrmann,
Res. Chem. Intermed., 2005, 31, 449.
A. Kubacka, G. Colón and M. Fernández-García, Chem. Rev.,
35 M. Groenewolt and M. Antonietti, Adv. Mater., 2005,
17, 1789.
2
012, 112, 1555.
5
6
A. Kohgkand and P. V. Kamat, ACS Nano, 2007, 1, 13.
W. D. Wang, P. Serp, P. Kalack and J. L. Faria, J. Mol. Catal.
A: Chem., 2005, 235, 199.
36 S. C. Yan, Z. S. Li and Z. G. Zou, Langmuir, 2009, 25, 10397.
37 J. S. Zhang, J. H. Sun, K. Maeda, K. Domen, P. Liu,
M. Antonietti, X. Z. Fu and X. C. Wang, Energy Environ. Sci.,
2011, 4, 675–678.
7
Y. Qu and X. Duan, Chem. Soc. Rev., 2013, 42, 2568.
2014 | Catal. Sci. Technol., 2014, 4, 2006–2015
This journal is © The Royal Society of Chemistry 2014

View Article Online
Catalysis Science & Technology
Paper
3
3
4
4
4
8 M. Fernández-García, A. Martínez-Arias, J. C. Hanson and
J. A. Rodríguez, Chem. Rev., 2004, 104, 4063.
9 J. Mo, Y. Zhang, Q. Xu, Y. Zhu, J. J. Lamson and R. Zhao,
Appl. Catal., B, 2009, 89, 570.
0 M. Sleiman, P. Conchon, C. Ferronato and J.-M. Chovelon,
Appl. Catal., B, 2009, 86, 159.
1 Y. Feng, L. Li, M. Ge, C. Gou, J. Wang and L. Liu, ACS Appl.
Mater. Interfaces, 2010, 2, 3134.
44 T. Tachikawa and T. Majima, J. Am. Chem. Soc., 2009, 131, 8485.
45 X. Wang, Z. Feng, J. Shi, G. Jia, S. Shen, J. Zhou and C. Li,
Phys. Chem. Chem. Phys., 2010, 12, 7083.
46 C. C. Mercado, Z. Seeley, A. Bandyopadhyay, S. Bose and
J. L. McHale, ACS Appl. Mater. Interfaces, 2011, 3, 228.
47 E. Finnazi, C. D. Valentin and G. Pacchioni, J. Phys. Chem. C,
2009, 113, 3382.
48 Z. Jin, N. Murakami, T. Tsubota and T. Ohno, Appl. Catal.,
B, 2014, 150–151, 479.
49 V. Singh, J. Castellanos Beltrán, J. Casamada Ribot and
P. Nagpal, Nano Lett., 2014, 14, 597.
2 M. Peláez, N. T. Nolan, S. C. Pillai, M. K. Severy, P. Falaras,
A. G. Kontos, P. S. M. Dunlop, J. A. Bryne and D. D. Dionysiou,
Appl. Catal., B, 2012, 125, 331.
4
3 A. Kubacka, M. J. Muñoz-Batista, R. Rachwalik, B. Bachiller-Baeza
50 K. Maeda, X. Wang, Y. Nishihara, D. Lu, M. Antonietti and
K. Domen, J. Phys. Chem. C, 2009, 113, 4940.
and M. Fernández-García, J. Catal., 2014, 309, 428.
This journal is © The Royal Society of Chemistry 2014
Catal. Sci. Technol., 2014, 4, 2006–2015 | 2015