Oxygen-Deficient BiOBr as a Highly Stable Photocatalyst for Efficient CO2 Reduction into Renewable Carbon-Neutral Fuels

Article abstract of DOI:10.1002/cctc.201600782

A highly facile one-pot ethylene glycol-assisted solvothermal process was employed to fabricate bismuth oxybromide (BiOBr) with oxygen-deficient defects. These defects played an indispensable role for superior photocatalytic CO₂ reduction, in w

Full text of DOI:10.1002/cctc.201600782

DOI: 10.1002/cctc.201600782
Full Papers
Oxygen-Deficient BiOBr as a Highly Stable Photocatalyst
for Efficient CO Reduction into Renewable Carbon-Neutral
2
Fuels
[
a]
[a]
[a, b]
[a]
Xin Ying Kong, W. P. Cathie Lee, Wee-Jun Ong,
Rahman Mohamed
Siang-Piao Chai,* and Abdul
[c]
A highly facile one-pot ethylene glycol-assisted solvothermal
process was employed to fabricate bismuth oxybromide
oxygen–deficiency–induced defect states could effectively trap
photogenerated electrons, thus improving the separation of
the electron–hole pairs and significantly slow down the recom-
bination rate of charge carriers. On top of that, oxygen-defi-
cient BiOBr exhibited long term stability (>50 hours of catalyt-
(BiOBr) with oxygen-deficient defects. These defects played an
indispensable role for superior photocatalytic CO reduction, in
2
which the as-prepared sample demonstrated a remarkable im-
provement of 3.3 and 5.7-fold for CH production over pristine
ic reaction) for CO photoreduction under simulated solar light,
4
2
BiOBr and P25, respectively. The enhancement could be attrib-
uted to the presence of oxygen vacancies, which acted as the
where no reducing agent or any post-treatment was needed
to regenerate the oxygen vacancies.
active sites for CO adsorption and activation. In addition, the
2
Introduction
Our heavy reliance on fossil fuels has not only depleted their
reserves, but has also led to large scale emission of carbon di-
V–VI–VIII ternary oxide semiconductors with special layered
structures (Figure S1A in the Supporting Information), the self-
oxide (CO ), which is the main cause of anthropogenic climate
built internal static electric fields along the layer was proven to
2
[
1]
[14,15]
change. In this regard, photocatalytic CO reduction, an artifi-
favor electron–hole pair separation.
Moreover, BiOX con-
2
cial photosynthesis process, has received immense attention
sists of only p-block elements (Bi, O, and Cl/Br/I), which could
lead to the formation of highly dispersive band structures with
low effective mass of charge carriers (Figure S1B in the Sup-
porting Information), which are hence beneficial for photoin-
owing to its capability to transform abandoned CO into re-
2
newable carbon-neutral fuels by means of clean and inexhaus-
[
2]
tible solar energy. Following the initial discovery of photoca-
[
3]
[16]
talytic CO reduction by Inoue and Fujishima in 1979, tremen-
duced charge carrier transfer. In particular, study of BiOBr is
2
dous effort has been devoted to developing highly efficient
of great interest owing to its appropriate band gap energy
[
4–9]
[17]
and robust photocatalysts.
Recently, bismuth oxyhalides (BiOX, where X=Cl, Br, I),
(2.8–2.9 eV) and visible-light responsiveness.
Previous re-
search has attested to the fact that BiOBr exhibits outstanding
photodegradation activity and, hence, BiOBr is regarded as an
[10]
known to be one of the simplest members of Sillꢀn family,
[18–21]
have garnered considerable attention from worldwide re-
searchers in the field of photocatalysis owing to their remark-
able properties including optical and chemical stability, envi-
efficient catalyst for environmental remediation.
However,
the use of BiOBr in the photochemical energy harvesting field
is still very limited owing to its intrinsic low conduction band
[
11–13]
[16]
ronmental friendliness and inexpensiveness.
As a group of
(CB) potential, along with weak reduction power. To the best
of our knowledge, there is no prior work on the utilization of
co-catalysts or dopant-free BiOBr for photocatalytic CO2
reduction.
[
a] X. Y. Kong, W. P. C. Lee, Dr. W.-J. Ong, Prof. S.-P. Chai
Multidisciplinary Platform of Advanced Engineering, Chemical Engineering
Discipline, School of Engineering
Monash University
Jalan Lagoon Selatan, Bandar Sunway, 47500 Selangor (Malaysia)
E-mail: chai.siang.piao@monash.edu
To realize this goal, the introduction of oxygen-deficient
type defects on the BiOBr surface is proposed. Firstly, oxygen
vacant sites with abundant localized electrons have been
proven to be active sites for adsorption and activation of CO
2
[b] Dr. W.-J. Ong
[22]
molecules. Previous studies divulged that those CO mole-
Institute of Materials Research and Engineering (IMRE)
Agency for Science, Technology and Research (A*STAR)
2
cules adsorbed on the oxygen vacant sites will be spontane-
ously activated to CCO2ꢀ radical ions, which is the rate-deter-
2
Fusionopolis Way, Innovis, 138634 (Singapore)
[23,24]
[
c] Prof. A. R. Mohamed
mining step of CO2 reduction.
Moreover, photoinduced
Low Carbon Economy (LCE) Group, School of Chemical Engineering
Universiti Sains Malaysia, Engineering Campus
Seri Ampangan, 14300 Nibong Tebal, Pulau Pinang (Malaysia)
electrons could be effectively trapped in the oxygen vacancies,
which significantly slow down the recombination rate of elec-
[25]
tron–hole pairs. By virtue of these advantages of surface de-
fects, developing oxygen-deficient defects on BiOBr is
Supporting information for this article can be found under:
http://dx.doi.org/10.1002/cctc.201600782.
ChemCatChem 2016, 8, 1 – 9
1
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
anticipated to be a promising strategy for enhanced CO pho-
2
toreduction. Most notably, BiOBr was found to mainly exhibit
[
26]
{
001} facets. The {001} facets are terminated with a high
density of oxygen atoms (Figure S2 in the Supporting Informa-
tion) and, hence, oxygen vacancies could be easily induced on
BiOBr under very mild conditions.
Herein, we present a highly facile one-pot ethylene glycol
(EG)-assisted solvothermal approach for the synthesis of BiOBr
nanosheets with oxygen vacancy defects, which is denoted as
oxygen-deficient BiOBr. The procedure is relatively simple and
does not employ conventional methods such as metal
[
17,27]
[28]
doping
or electroreduction to induce oxygen-deficient
defects into the sample. In the present study, considerable
amounts of oxygen vacancies were introduced on the {001}
facets of BiOBr by thermal treatment in a reducing atmos-
phere, in which EG played an indispensable role to remove sur-
face oxygen atoms from BiOBr, resulting in selective formation
of oxygen-deficient defects on the sample surface. For compar-
ison, pristine BiOBr was synthesized by using deionized water
as the solvent. As the synthesis procedure in the current work
is a highly facile one-step process and without the incorpora-
tion of any co-catalysts or dopants, we believe that the find-
ings from this study could open up a cost-effective and scala-
ble method to fabricate photocatalysts with robust per-
formance for energy harvesting and related applications.
Figure 1. (A)–(C) FESEM images, (D) HRTEM image (inset shows the schemat-
ic illustration of crystal orientation), (E) SAED pattern, and (F) XRD pattern of
oxygen-deficient BiOBr.
Results and Discussion
Materials characterization
The field emission scanning electron microscopy (FESEM)
images revealed that the as-obtained oxygen-deficient BiOBr
are of 3D flower-like architectures of 0.5–1.5 mm, which densely
assembled in numerous radially grown nanosheets with thick-
nesses about 10 nm (Figure 1A–C). The crystal lattice fringes
with 0.278 nm interplanar lattice spacing are attributed to the
(
(
(
110) atomic plane perpendicular to the (001) plane of BiOBr
Figure 1D). Based on the selected area electron diffraction
SAED) pattern (Figure 1E), the angle of the adjacent spots was
found to be 458, which agrees with the theoretical angle be-
tween the (110) and (200) planes. This set of diffraction pat-
terns could be indexed to the [001] zone axis of tetragonal
BiOBr, hence confirming that the as-prepared sample is en-
closed by dominant {001} facets (inset, Figure 1D). The X-ray
diffraction (XRD) pattern of oxygen-deficient BiOBr (Figure 1F)
further established its tetragonal structure, in which all the dif-
fraction peaks of the sample could be readily indexed to the
tetragonal phase of BiOBr (JCPDS card no.: 01-1004). In addi-
tion, no characteristic peaks of any impurities or other phases
were detected, indicating that the as-prepared sample is of
high purity. The elemental composition of the oxygen-deficient
BiOBr was determined by energy-dispersive X-ray (EDX) spec-
troscopy (Figure 2), which unambiguously revealed the co-exis-
tence of Bi, O, and Br elements that are uniformly distributed
over the sample.
Figure 2. (A) EDX mapping of oxygen-deficient BiOBr, elemental distribution
of (B) O (red), (C) Bi (magenta), and (D) Br (cyan). Inset shows the corre-
sponding FESEM image of oxygen-deficient BiOBr for EDX mapping.
provide fingerprint evidence. As depicted in Figure 3A,
oxygen-deficient BiOBr exhibited a characteristic signal at
about g=2.0, which corresponded to electrons trapped in
[29]
oxygen vacancies, thus confirming the successful introduc-
tion of oxygen vacancy defects on the BiOBr surface by our
facile EG-assisted solvothermal process. In contrast, no
To ascertain the presence of oxygen-deficient defects, elec-
tron spin resonance (ESR) spectroscopy was performed to
&
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
2
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 3. (A) ESR spectra, (B) Raman spectra, (C) high-resolution Bi4f spectra, (D) UV/Vis diffuse reflectance spectra (inset shows the colors of the as-prepared
photocatalysts), and (E) plot of transformed KM function vs. photon energy of pristine and oxygen-deficient BiOBr samples.
significant signal emerged from pristine BiOBr, suggesting that
the formation of oxygen vacancies strongly depends on the
solvent used for the preparation. In this case, EG played an in-
dispensable role to bind and react with surface oxygen atoms
in BiOBr, resulting in the formation of oxygen vacancy defects
as-obtained oxygen-deficient BiOBr and pristine BiOBr were in-
vestigated by X-ray photoelectron spectroscopy (XPS). The
high-resolution Bi4f scan (Figure 3C) of pristine BiOBr exhibit-
ed two peaks at 159.4 and 164.8 eV, which could be attributed
3
+
[33]
to Bi4f7/2 and Bi4f5/2 in the Bi chemical state, respectively.
[
30]
on the sample surface while EG was being oxidized. Further
evidence of the formation of oxygen-deficient defects was pro-
vided by Raman spectroscopy. As shown in Figure 3B, pristine
In contrast, a 0.3 eV shift to lower binding energy and two ad-
ditional peaks at 157.3 and 162.8 eV were observed for
oxygen-deficient BiOBr, which are classically attributed to the
ꢀ
1
(3ꢀx)+
BiOBr exhibited Raman peaks at 57, 91, and 112 cm , which
corresponds to the A stretching mode whereas the peak at
formation of Bi
as a result of the presence of oxygen va-
[
28]
cancy defects. The characterization results above show indis-
putable evidence of the successful introduction of oxygen va-
cancies in the as-prepared oxygen-deficient BiOBr sample, pro-
viding the prerequisite for examining the roles of oxygen va-
cancy defects during photocatalytic reactions.
1g
ꢀ
1
1
61 cm could be ascribed to the E1g internal BiꢀBr stretching
[31]
mode. These sharp peaks imply that pristine BiOBr is of high
crystallinity. After the introduction of surface defects, oxygen-
deficient BiOBr demonstrated several noticeable changes in-
cluding severe weakening and broadening of peaks as well as
peak shifting to lower frequency. This phenomenon confirmed
the formation of oxygen vacancies in which the displacement
of oxygen atoms resulted in lattice distortion and crystallinity
To elucidate the presence of oxygen vacancy defects for en-
hanced photocatalytic performance, the optical properties of
pristine and oxygen-deficient BiOBr were investigated. As re-
ported in the literature, indirect sub-band excitation from
oxygen-deficiency defect states to the conduction band (CB)
[
32]
weakening.
The chemical states of Bi species in the
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
3
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
leads to significant changes in the optical properties of metal
[
34,35]
oxide semiconductors.
As observed in Figure 3D (inset),
pristine BiOBr was pure white in color whereas oxygen-defi-
cient BiOBr was found to be brownish-grey in color. The drastic
change in the optical properties was reflected in the UV/Vis ab-
sorption spectra (Figure 3D), in which more intense continuous
absorption over the UV and visible-light regions was exhibited
by the oxygen-deficient BiOBr. Moreover, a redshift of the ab-
sorption edge was observed for oxygen-deficient BiOBr, which
indicated a band gap narrowing. Based on the Tauc plot of
modified Kubelka–Munk (KM) function (Figure 3E), the band
gap energies of pristine and oxygen-deficient BiOBr were esti-
mated to be 2.85 and 2.70 eV, respectively. This phenomenon
of oxygen-deficiency-induced band gap narrowing was also
[
36,37]
observed in the literature;
it can be ascribed to the over-
lapping of oxygen-deficiency-induced defect states with the
[
38]
band edge of the semiconductor. To ascertain the presence
of such defect states, photoluminescence (PL) spectroscopy
was conducted. As shown in Figure 4, a more intense PL emis-
Figure 5. (A) Transient photocurrent responses and (B) electrochemical impe-
dance spectroscopy (EIS) Nyquist plots of oxygen-deficient and pristine
BiOBr in dark conditions and under simulated solar light irradiation.
demonstrated smaller arc radii compared with pristine BiOBr
under dark conditions and simulated solar light irradiation.
These results indicate that oxygen-deficient BiOBr possesses
lower electron transfer resistance at the surface of electrodes,
which could markedly facilitate the transfer of charge carriers
and lead to a more efficient interfacial charge-transfer process.
These findings further attested to the imperative role of
oxygen vacancy defects for enhanced photocatalytic
performance.
Figure 4. Room-temperature steady-state PL spectra of oxygen-deficient and
pristine BiOBr.
sion peak centered at about 618 nm (2.01 eV) was observed
for oxygen-deficient BiOBr. As the energy of this PL is much
smaller than the band gap of oxygen-deficient BiOBr (2.70 eV),
this PL emission is attributed to the indirect recombination of
electrons trapped in the oxygen-deficiency-induced defect
states with the photogenerated holes in the valence band
Photoreduction of CO and mechanism behind the enhance-
2
ment
[
35]
(VB).
To evaluate the photoreduction activities of as-prepared pris-
It has been well established that transient photocurrent
tine and oxygen-deficient BiOBr, photocatalytic CO reduction
2
measurement is a useful characterization tool to evaluate the
separation efficiency and lifetime of photogenerated charge
with water vapor was conducted at ambient temperature and
atmospheric pressure. The performance of the as-obtained
photocatalysts was studied under both visible light and simu-
lated solar light. A series of control experiments were per-
formed under several conditions including: i) without light irra-
diation; ii) without CO (under N /H O vapor flow); and iii) un-
[
39–41]
carriers.
As shown in Figure 5A, the as-developed oxygen-
deficient BiOBr demonstrated a much higher photocurrent
density than pristine BiOBr, implying that the introduction of
surface oxygen vacancies on BiOBr significantly slowed down
the recombination rate of electron–hole pairs, thereby pro-
longing the lifetime of photoinduced charge carriers. To gain
further insights into the transfer of photogenerated charge car-
riers, electrochemical impedance spectroscopy (EIS) Nyquist
plots have been carried out. Generally, a small arc radius of EIS
Nyquist plot indicates high efficiency of charge carrier trans-
2
2
2
der CO /H O vapor flow and light irradiation but without pho-
2
2
tocatalysts. In all these cases, negligible to no carbon-related
products were detected (Figure S3 in the Supporting Informa-
tion). This showed that the product yields stemmed from pho-
tocatalytic reduction of CO and highlighted the significance of
2
the reactant feeds (CO /H O), photocatalysts, and light source
2
2
[
42]
fer. As can be seen from Figure 5B, oxygen-deficient BiOBr
to drive the reaction.
&
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
4
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 6 shows the total production of CH over pristine and
4
oxygen-deficient BiOBr after 10 h of reaction under visible-light
irradiation. As illuminated by the results, pristine BiOBr exhibit-
ed minimal production of CH4, with
a total yield of
Figure 7. Schematic illustration of the multi-step electron transfer in oxygen-
deficient BiOBr. Step 1: Photoexcited electrons are predominantly transferred
to the oxygen-deficiency-induced defect states and the direct recombination
with the photoinduced holes at ground states is inhibited. Step 2: Electrons
are trapped in the oxygen-deficiency defect states and the indirect electron
2
recombination from the defect states is delayed. Once the CO molecules
adsorbed on the oxygen-deficiency defect states, electrons are transferred
ꢀ
immediately to the adsorbed CO
2
and reduced it to CCO
2
radical ions.
Figure 6. Time course of the photocatalytic CH
light irradiation (l>400 nm).
4
production under visible
implying the imperative contribution of oxygen vacancy de-
fects to CO reduction.
2
ꢀ
1
1
.58 mmolg . By creating oxygen vacancy defects on the
001} facets of BiOBr nanosheets, the as-obtained oxygen-defi-
cient BiOBr demonstrated a significant enhancement for CH
Generally, the oxygen vacant sites in oxide semiconductors
{
are healed gradually by the O atoms released during the CO
2
[47]
dissociation process. This phenomenon was also observed in
4
ꢀ
1
production, achieving a total yield of 4.86 mmolg , which was
.1-fold higher in comparison to pristine BiOBr. The photocata-
the current study, in which the formation of CH gradually de-
4
3
clined and oxygen-deficient BiOBr only maintained a reactivity
of 66% after 10 h of reaction under visible light (Figure S5 in
the Supporting Information). These results imply that the heal-
ing of oxygen vacant sites proceeded simultaneously with the
CO reduction process, which led to the blockage of the CO
lytic enhancement of oxygen-deficient BiOBr was accredited to
several concomitant factors. Firstly, the presence of oxygen-de-
ficient surface defects significantly alters the optical properties
of BiOBr, which guarantees enhanced light absorption to pro-
mote electron–hole pair generation, along with oxygen-defi-
ciency-induced band gap narrowing. In addition, oxygen va-
cancies played an indispensable role to trap photoinduced
electrons, hence improving the electron–hole pair separation
and inhibiting the direct recombination of photogenerated
charge carriers (Figure 7). Some pioneer works on engineering
surface defects revealed that the lifetime of electrons in
oxygen-deficiency defect states were several times longer than
2
2
activation channel as a result of the surface reoxidation of
oxygen-deficient BiOBr. However, these O atoms could easily
be removed from the lattice again under simulated solar light
irradiation owing to the capability of UV light to regenerate
[35]
the surface oxygen vacancies. The occurrence of UV-photoin-
duced oxygen vacancies was attested by the significant color
change of the sample after CO photoreduction (Figure S6 in
2
the Supporting Information). It was observed that oxygen-defi-
cient BiOBr changed to darkish grey after the simulated solar
light driven photocatalytic experiment, owing to the increased
indirect sub-band excitation from oxygen-deficiency defect
[
35,43]
that of free excitons.
On top of that, it has been theoreti-
cally proved that the adsorption of CO molecules is energeti-
2
[
24]
cally favorable on the oxygen vacant sites. The attraction be-
[34]
tween CO molecules and oxygen vacancies may generate un-
states to the CB of oxygen-deficient BiOBr. A similar phe-
2
[48]
expected affinity interactions, which could significantly lower
nomenon was also observed in previous work. Compared
[
22,44]
the energy barrier for interfacial charge transfer.
Once
with commercial TiO (Degussa P25), developing oxygen va-
2
a CO molecule is adsorbed on the oxygen vacancy, the CO2
cancy defects in BiOBr is much simpler as the BiꢀO bonds are
2
molecule is instantly activated to the CCO ꢀ radical ion through
[11]
longer and have lower bond energy. Herein, the photocata-
2
[
45]
accepting an electron from an oxygen vacant site. Next, the
lytic CO reduction experiment was repeated under simulated
2
CCO2ꢀ radicals transform into the desired CH through a multi-
solar light irradiation and it was noted that the photocatalytic
4
[
46]
electron transfer process. To provide clear-cut evidence con-
firming that oxygen-deficient defects played the dominant role
in enhancing the photoreduction activity, oxygen-deficient
BiOBr was annealed in an air-rich atmosphere at 3008C for 5 h
to reoxidize the surface, which was denoted as annealed-BiOBr.
As shown in Figure S4B (in the Supporting Information), the
CH4 yield over annealed-BiOBr was significantly decreased,
production of CH over oxygen-deficient BiOBr was greatly en-
4
hanced (Figure 8, Figure S5 in the Supporting Information).
These results suggest that oxygen vacancies were efficiently re-
freshed and regenerated during the CO photoreduction pro-
2
cess, although UV only constitutes for less than 5% of the
solar spectrum. Notably, oxygen-deficient BiOBr exhibited
ꢀ
1
a total CH production of 9.58 mmolg over 10 h under solar
4
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
5
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
CO reduction over oxygen-deficient BiOBr is still in its infancy,
2
we believe that the oxygen vacancies were the reductive sites
for the conversion CO to CO, which was then further reduced
2
to CH . In this regard, a novel pathway based on engineering
4
surface defects of semiconductors was developed to produce
highly desired products from reduction of CO and the process
2
is schematically illustrated in Figure 10.
Considering its facile preparation, superior performance, and
long term sustainability for photocatalytic CO reduction, it is
2
indisputable to say that oxygen-deficient BiOBr will be a prom-
ising photocatalyst with many potential applications.
4
Figure 8. Time course of the photocatalytic CH production under simulated
solar light irradiation.
light irradiation, which indicated a remarkable improvement of
.2 and 5.7-fold over pristine BiOBr and Degussa P25, respec-
3
tively (Figure 8).
The stability of oxygen-deficient BiOBr was investigated by
evaluating its accumulated CH evolution yield through five
4
consecutive test cycles under simulated solar light irradiation.
As shown in Figure 9, there was no significant decrease in CH4
production after five cycles (>50 hours of reaction) under
2
Figure 10. Schematic illustration of CO photoreduction on the oxygen
vacant sites at the {001} facets of oxygen-deficient BiOBr.
Conclusions
This work has successfully realized the introduction of an
oxygen-deficiency strategy for bismuth-based materials for en-
hanced photocatalytic reduction activity. In this study, oxygen-
deficient BiOBr with abundant oxygen vacancy defects were
synthesized by a highly facile EG-assisted solvothermal process.
The as-prepared oxygen-deficient BiOBr exhibited superior
Figure 9. Stability tests of oxygen-deficient BiOBr under simulated solar light
irradiation.
photocatalytic activity and long term stability for CO reduc-
2
identical experimental conditions, demonstrating the excellent
performance of the as-prepared oxygen-deficient BiOBr as an
efficient photocatalyst with long term stability and reusability.
In addition, the photoreduction activity of oxygen-deficient
BiOBr did not show any decrease after storage in the dark for
six months, implying that the oxygen vacancy defects in
oxygen-deficient BiOBr are highly stable.
tion under simulated solar light, while the recovery and regen-
eration of oxygen vacancies was realized during the CO pho-
2
toreduction process. On top of that, no reducing agent or any
other post-treatment was needed to regenerate the oxygen
vacancy defects on the sample. We believe that our findings
not only present the possibility for the use of oxygen-deficient
BiOBr as a robust material for conversion of CO , but also deliv-
2
It is worth noting that a large amount of carbon monoxide
er a vital concept that developing oxygen vacancy defects on
simple oxides could be a feasible and effective strategy to
design next generation catalysts with remarkable activities.
(
CO) was also produced in the photocatalytic CO reduction
2
over oxygen-deficient BiOBr (Figure S7 in the Supporting Infor-
mation). As mentioned, CCO ꢀ radicals were generated once
2
the CO molecules were adsorbed on the oxygen vacant sites
2
with abundant localized electrons. Previous research has
Experimental Section
shown that CCO2ꢀ radicals are spontaneously dissociated into
[24]
Reagents
CO on a partially oxygen depleted semiconductor surface,
which explains the high yield of CO observed in our current
study. Although the understanding of the exact mechanism for
All chemicals used were of analytical grade and deionized water
(resistivity ꢁ18 MWcm) was used throughout the experiment.
&
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
6
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Bismuth(III) nitrate pentahydrate, Bi(NO ) ·5H O (ꢁ98.0%), potassi-
Photochemical measurements
3
3
2
um bromide, KBr (ꢁ99.0%), ethylene glycol, EG (ꢁ99.0%), and De-
gussa P25 (99.5%) were purchased from Sigma–Aldrich whereas
ethanol, C H OH (Grade AR 96.0%) was purchased from Friende-
The electrochemical impedance spectroscopy (EIS) Nyquist plot
and transient photocurrent responses were determined by using
a CHI6005E electrochemical workstation in a conventional three-
2
5
mann Schmidt Chemical. All chemicals were used as received with-
out any further purification.
electrode quartz cell with 0.5m Na SO4 as the electrolyte. The
2
working electrode was prepared by drop-casting the slurry samples
onto a fluorine-doped tin oxide (FTO) glass slide with an electroac-
2
tive area fixed at 1 cm , while Pt and Ag/AgCl were used as the
Synthesis of pristine and oxygen-deficient BiOBr
counter and reference electrodes, respectively. A 500 W Xenon arc
lamp (CHF-XM-500W) equipped with an AM 1.5 filter was used as
the exciting light source.
BiOBr with abundant oxygen vacancy defects was prepared by
a highly facile one-pot EG-assisted solvothermal process. In brief,
KBr (5 mmol) was dissolved in EG (80 mL) and a stoichiometric
amount of Bi(NO ) ·5H O was added slowly into the solution. The
3
3
2
colorless mixture was stirred continuously for 1 h at room tempera-
ture and then transferred to a stainless steel Teflon-lined autoclave
of 125 mL capacity. The autoclave was sealed and heated at 1608C
for 12 h with continuous stirring. After completion of the reaction,
the autoclave was allowed to cool to room temperature naturally
and the resulting precipitate was collected and washed with etha-
nol and deionized water several times. Finally, the obtained
powder was left to dry overnight at 608C. The sample collected
was denoted as oxygen-deficient BiOBr. On the other hand, the
synthesis of BiOBr without surface defects (pristine BiOBr) followed
the same procedure as that of oxygen-deficient BiOBr, except that
deionized water was used as the solvent instead of EG.
Photocatalytic CO reduction experiments
2
The photocatalytic CO reduction experiments were conducted in
2
a custom-made, continuous gas flow reactor at mild conditions
(ambient temperature and atmospheric pressure). A 500 W Xenon
arc lamp (CHF-XM-500W) equipped with an AM 1.5 filter (to simu-
late solar light) and UV cut-off filter (to provide wavelengths
>
400 nm) was used as the light source. The experimental set-up is
depicted in Figure S8 in the Supporting Information. In brief, high
purity CO (99.999%) was bubbled through H O to generate a mix-
2
2
ture of CO and water vapor and the resultant mixture gas (reac-
2
tant gas) was fed into the reactor. Before each run, the photocata-
lyst was coated on the glass rods and the reactant gas was allowed
to pass through the quartz columns loaded with coated photoca-
ꢀ
1
talyst for 30 minutes at a flowrate of 50 mLmin to evacuate any
excess air and to ensure the adsorption–desorption equilibrium
was achieved on the photocatalyst surface. Then, the flowrate of
Characterization
The surface morphology of the as-prepared samples was studied
by using a Hitachi SU8010 field emission scanning electron micros-
copy (FESEM) and its elemental composition was analyzed by
using energy-dispersive X-ray (EDX) spectroscopy. High-resolution
transmission electron microscopy (HRTEM) images and selected-
area electron diffraction (SAED) patterns were acquired by using
a TECNAI G2 F20 transmission electron microscope with an acceler-
ating voltage of 200 kV. For the sample preparation for TEM analy-
sis, the sample powder was suspended in ethanol and one drop of
the suspension was deposited on a copper grid. Powder X-ray dif-
fraction (XRD) patterns were obtained by using a Bruker AXS D8
ꢀ1
the reactant gas was adjusted to 5 mLmin and maintained at
this flowrate for 10 h under light illumination to study the per-
formance of the photocatalysts by collecting the product gas at
1
h intervals and analyzing it by using gas chromatography
(Agilent 7820A GC equipped with flame ionized detector (FID) and
thermal conductivity detector (TCD)). The yield of products (CH
4
ꢀ1 ꢀ1
and CO, in mmolg h ) was quantified by using Equation (1):
ðCfinal;CH4=CO ꢀ Cinitial;CH4=COÞ ꢂ volumetric flowrate of product gas
Yield ¼
Amount of photocatalyst used
Advance X-Ray diffractometer with CuK radiation (l=0.15406 nm)
a
ð1Þ
ꢀ
1
at a scan rate of 0.02 s over a diffraction angle range of 2q=58–
7
4
08 at an accelerating voltage and applied current of 40 kV and
0 mA, respectively. The electron spin resonance (ESR) spectra of
the samples were recorded with
JES-FA200). The X-ray photoelectron spectroscopy (XPS) measure-
ments were obtained by using a scanning X-ray microprobe PHI
Quantera II (Ulvac-PHI, INC.) with monochromatic AlKa
hv=1486.6 eV) X-ray source. Wide scan and narrow scan analysis
were performed with a pass energy of 280 eV (1 eV per step) and
12 eV (0.1 eV per step), respectively. Prior to deconvolution, all
a JEOL ESR spectrometer
Acknowledgments
(
This work was funded by the Ministry of Higher Education
a
(
MOHE) Malaysia and Universiti Sains Malaysia under the Nano-
(
MITe Long-Term Research Grant Scheme (LRGS) (Reference
number: 203/PJKIMIA/6720009).
1
binding energies were referenced to the adventitious carbon
signal (C1s peak) at 284.6 eV. Raman spectra were recorded at
room temperature by using a micro Raman spectrometer (Horiba
LabRAM HR Evolution) with 514 nm laser excitation. The absorb-
ance spectra of the samples were measured in the range of 200–
Keywords: bismuth oxybromide · CO2 reduction · oxygen
deficiency · photocatalysis · solar energy
[
[
[
[
1] X. Lim, Nature 2015, 526, 628–630.
8
00 nm by using an Agilent Cary 100 ultraviolet-visible (UV/Vis)
2] D. Chen, X. Zhang, A. F. Lee, J. Mater. Chem. A 2015, 3, 14487–14516.
3] T. Inoue, A. Fujishima, S. Konishi, K. Honda, Nature 1979, 277, 637–638.
4] W.-J. Ong, L.-L. Tan, S.-P. Chai, S.-T. Yong, A. R. Mohamed, Nano Energy
spectrophotometer equipped with an integrated sphere. The Ku-
belka–Munk (KM) function, F(R), and extrapolation of the Tauc plot,
1
/2
[
F(R)ꢂhv] vs. hv were employed to determine the band gap en-
2
015, 13, 757–770.
5] W.-J. Ong, L.-L. Tan, Y. H. Ng, S.-T. Yong, S.-P. Chai, Chem. Rev. 2016, 116,
159–7329.
ergies of the samples. Lastly, photoluminescence (PL) measure-
ments of the samples were analyzed by using a LS 55 PerkinElmer
fluorescence spectrophotometer at an excitation wavelength of
[
7
[6] W.-J. Ong, L.-L. Tan, S.-P. Chai, S.-T. Yong, A. R. Mohamed, Nano Res.
2014, 7, 1528–1547.
3
00 nm.
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
7
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[
[
[
7] W.-J. Ong, L. K. Putri, L.-L. Tan, S.-P. Chai, S.-T. Yong, Appl. Catal. B 2016,
80, 530–543.
8] L.-L. Tan, W.-J. Ong, S.-P. Chai, A. R. Mohamed, Chem. Commun. 2014,
0, 6923–6926.
9] L.-L. Tan, W.-J. Ong, S.-P. Chai, B. T. Goh, A. R. Mohamed, Appl. Catal. B
015, 179, 160–170.
[30] J. Jiang, L. Zhang, H. Li, W. He, J. J. Yin, Nanoscale 2013, 5, 10573–
1
10581.
[31] J. Wang, Y. Zhang, L. Tian, F. Liu, Q. Xia, J. Nanopart. Res. 2014, 16,
5
2691–2699.
[32] Y. Yu, C. Cao, H. Liu, P. Li, F. Wei, Y. Jiang, W. Song, J. Mater. Chem. A
2
2014, 2, 1677–1681.
[
[
[
[
10] Z.-Y. Zhao, W.-W. Dai, Inorg. Chem. 2015, 54, 10732–10737.
11] J. Li, Y. Yu, L. Zhang, Nanoscale 2014, 6, 8473–8488.
12] L. Li, L. Ai, C. Zhang, J. Jiang, Nanoscale 2014, 6, 4627–4634.
13] Z. Yang, J. Li, F. Cheng, Z. Chen, X. Dong, J. Alloys Compd. 2015, 634,
[33] P. Madhusudan, J. Zhang, B. Cheng, G. Liu, CrystEngComm 2013, 15,
231–240.
[34] F. Zuo, L. Wang, T. Wu, Z. Zhang, D. Borchardt, P. Feng, J. Am. Chem.
Soc. 2010, 132, 11856–11857.
2
15–222.
14] X. Cao, Z. Lu, L. Zhu, L. Yang, L. Gu, L. Cai, J. Chen, Nanoscale 2014, 6,
434–1444.
[35] L. Ye, L. Zan, L. Tian, T. Peng, J. Zhang, Chem. Commun. 2011, 47, 6951–
[
6953.
1
[36] Y. Liu, B. Wei, L. Xu, H. Gao, M. Zhang, ChemCatChem 2015, 7, 4076–
[
[
15] J. Li, K. Zhao, Y. Yu, L. Zhang, Adv. Funct. Mater. 2015, 25, 2189–2201.
16] J. Shang, W. Hao, X. Lv, T. Wang, X. Wang, Y. Du, S. Dou, T. Xie, D. Wang,
J. Wang, ACS Catal. 2014, 4, 954–961.
4084.
[37] Y. Su, J. Lang, C. Du, F. Bian, X. Wang, ChemCatChem 2015, 7, 2437–
2441.
[
[
17] Z. Liu, B. Wu, Y. Zhao, J. Niu, Y. Zhu, Ceram. Int. 2014, 40, 5597–5603.
18] H. Li, T. Hu, J. Liu, S. Song, N. Du, R. Zhang, W. Hou, Appl. Catal. B 2016,
[38] L. R. Grabstanowicz, S. Gao, T. Li, R. M. Rickard, T. Rajh, D.-J. Liu, T. Xu,
Inorg. Chem. 2013, 52, 3884–3890.
[
39] J. Jiang, K. Zhao, X. Xiao, L. Zhang, J. Am. Chem. Soc. 2012, 134, 4473–
476.
1
82, 431–438.
19] H. Cheng, B. Huang, Z. Wang, X. Qin, X. Zhang, Y. Dai, Chem. Eur. J.
011, 17, 8039–8043.
20] J. Fu, Y. Tian, B. Chang, F. Xi, X. Dong, J. Mater. Chem. 2012, 22, 21159–
1166.
21] J. Xu, W. Meng, Y. Zhang, L. Li, C. Guo, Appl. Catal. B 2011, 107, 355–
62.
4
[
[
[
[
40] J. Di, J. Xia, M. Ji, B. Wang, X. Li, Q. Zhang, Z. Chen, H. Li, ACS Sustaina-
ble Chem. Eng. 2016, 4, 136–146.
2
[
[
41] F. Xiao, J. Miao, B. Liu, Mater. Horiz. 2014, 1, 259–263.
42] M. M. Khan, S. A. Ansari, D. Pradhan, M. O. Ansari, D. H. Han, J. Lee,
M. H. Cho, J. Mater. Chem. A 2014, 2, 637–644.
2
3
[
[
43] M. Sugawara, Phys. Rev. B 1995, 51, 10743–10756.
[
[
22] W. Tu, Y. Zhou, Z. Zou, Adv. Mater. 2014, 26, 4607–4626.
23] C. Yang, B. C. Wood, V. R. Bhethanabotla, B. Joseph, J. Phys. Chem. C
44] G. Xi, S. Ouyang, P. Li, J. Ye, Q. Ma, N. Su, H. Bai, C. Wang, Angew. Chem.
Int. Ed. 2012, 51, 2395–2399; Angew. Chem. 2012, 124, 2445–2449.
45] V. P. Indrakanti, H. H. Schobert, J. D. Kubicki, Energy Fuels 2009, 23,
2014, 118, 26236–26248.
[
[
[
[
[
[
24] L. Liu, C. Zhao, Y. Li, J. Phys. Chem. C 2012, 116, 7904–7912.
25] J. L. White, M. F. Baruch, J. E. Pander III, Y. Hu, I. C. Fortmeyer, J. E. Park,
T. Zhang, K. Liao, J. Gu, Y. Yan, T. W. Shaw, E. Abelev, A. B. Bocarsly,
Chem. Rev. 2015, 115, 12888–12935.
5247–5256.
46] S. Zhu, S. Liang, Y. Wang, X. Zhang, F. Li, H. Lin, Z. Zhang, X. Wang, Appl.
Catal. B 2016, 187, 11–18.
47] J. Lee, D. C. Sorescu, X. Deng, J. Am. Chem. Soc. 2011, 133, 10066–
0069.
48] L. Ye, K. Deng, F. Xu, L. Tian, T. Peng, L. Zan, Phys. Chem. Chem. Phys.
012, 14, 82–85.
[
[
26] L. Ye, Y. Su, X. Jin, H. Xie, C. Zhang, Environ. Sci. Nano 2014, 1, 90–112.
27] M. Gao, D. Zhang, X. Pu, X. Shao, H. Li, D. Lv, J. Am. Ceram. Soc. 2016,
1
99, 881–887.
2
[28] Y. Huang, H. Li, M. Balogun, W. Liu, Y. Tong, X. Lu, H. Ji, ACS Appl. Mater.
Interfaces 2014, 6, 22920–22927.
[29] I. Nakamura, N. Negishi, S. Kutsuna, T. Ihara, S. Sugihara, K. Takeuchi, J.
Mol. Catal. A 2000, 161, 205–212.
Received: June 28, 2016
Published online on && &&, 0000
&
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
8
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
Oxygen vacant sites as active sites:
Oxygen-deficient BiOBr exhibits superior
X. Y. Kong, W. P. C. Lee, W.-J. Ong,
S.-P. Chai,* A. R. Mohamed
performance for CO reduction. This
2
&& – &&
could be attributed to the CO mole-
2
cules tendency to adsorb on the
Oxygen-Deficient BiOBr as a Highly
Stable Photocatalyst for Efficient CO2
Reduction into Renewable Carbon-
Neutral Fuels
oxygen vacant sites with abundant lo-
calized electrons, in which CO is spon-
2
taneously reduced to CO and further re-
duced to CH by a multi-electron pro-
4
cess.
ChemCatChem 2016, 8, 1 – 9
www.chemcatchem.org
9
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ