Solar-Driven H2O2 Generation From H2O and O2 Using Earth-Abundant Mixed-Metal Oxide@Carbon Nitride Photocatalysts

Article abstract of DOI:10.1002/cssc.201600705

Light-driven generation of H₂O₂ only from water and molecular oxygen could be an ideal pathway for clean production of solar fuels. In this work, a mixed metal oxide/graphitic-C₃N₄ (MMO@C₃N₄) composite was synthesized as a dual-functional photocatalyst for both water oxidation and oxygen reduction to generate H₂O₂. The MMO was derived from a NiFe-layered double hydroxide (LDH) precursor for obtaining a high dispersion of metal oxides on the surface of the C₃N₄ matrix. The C₃N₄ is in the graphitic phase and the main crystalline phase in MMO is cubic NiO. The XPS analyses revealed the doping of Fe³⁺ in the dominant NiO phase and the existence of surface defects in the C₃N₄ matrix. The formation and decomposition kinetics of H₂O₂ on the MMO@C₃N₄ and the control samples, including bare MMO, C₃N₄ matrix, Ni- or Fe-loaded C₃N₄ and a simple mixture of MMO and C₃N₄, were investigated. The MMO@C₃N₄ composite produced 63 μmol L^?1 of H₂O₂ in 90 min in acidic solution (pH 3) and exhibited a significantly higher rate of production for H₂O₂ relative to the control samples. The positive shift of the valence band in the composite and the enhanced water oxidation catalysis by incorporating the MMO improved the light-induced hole collection relative to the bare C₃N₄ and resulted in the enhanced H₂O₂ formation. The positively shifted conduction band in the composite also improved the selectivity of the two-electron reduction of molecular oxygen to H₂O₂.

Full text of DOI:10.1002/cssc.201600705

DOI: 10.1002/cssc.201600705
Full Papers
Solar-Driven H O Generation From H O and O Using
2
2
2
2
Earth-Abundant Mixed-Metal Oxide@Carbon Nitride
Photocatalysts
[a]
[a]
[a]
[b]
[b]
Ruirui Wang, Kecheng Pan, Dandan Han, Jingjing Jiang, Chengxiang Xiang,
[b, c]
[a, b]
[a]
Zhuangqun Huang,
Lu Zhang,
and Xu Xiang*
Light-driven generation of H O only from water and molecular
including bare MMO, C N matrix, Ni- or Fe-loaded C N and
3 4 3 4
2
2
oxygen could be an ideal pathway for clean production of
a simple mixture of MMO and C N , were investigated. The
3 4
ꢀ
1
solar fuels. In this work, a mixed metal oxide/graphitic-C N4
MMO@C N composite produced 63 mmolL of H O in 90 min
3 4 2 2
3
(
MMO@C N ) composite was synthesized as a dual-functional
in acidic solution (pH 3) and exhibited a significantly higher
rate of production for H O relative to the control samples. The
3
4
photocatalyst for both water oxidation and oxygen reduction
to generate H O . The MMO was derived from a NiFe-layered
2
2
positive shift of the valence band in the composite and the en-
hanced water oxidation catalysis by incorporating the MMO
improved the light-induced hole collection relative to the bare
C N and resulted in the enhanced H O formation. The posi-
2
2
double hydroxide (LDH) precursor for obtaining a high disper-
sion of metal oxides on the surface of the C N matrix. The
3
4
C N is in the graphitic phase and the main crystalline phase in
3
4
3
4
2
2
MMO is cubic NiO. The XPS analyses revealed the doping of
tively shifted conduction band in the composite also improved
the selectivity of the two-electron reduction of molecular
oxygen to H O .
3
+
Fe in the dominant NiO phase and the existence of surface
defects in the C N matrix. The formation and decomposition
3
4
2
2
kinetics of H O on the MMO@C N and the control samples,
2
2
3
4
Introduction
Hydrogen peroxide (H O ) is a prevailing chemical oxidant,
approaches have been proposed to address these issues, such
as photocatalytic generation of H O , which has distinctive ad-
2
2
widely used in a variety of industrial fields owing to its low
2
2
[
1]
cost and environmentally benign characteristics. For fuel
cells, H O is a clean energy source because the exhausts are
vantages because solar energy is utilized and converted to
a chemical fuel in this process without pollutant and carbon
2
2
[4]
only water and oxygen. H O is superior to hydrogen or other
emission.
2
2
fuel gasses because of its convenient and safe storage and
transportation in liquid form. In addition, a single-compart-
ment configuration is possible for H O based fuel cells, which
The key to establishing solar-to-fuel conversion as a viable
alternative is to develop efficient, stable, and scalable photoca-
talysts that could operate at mild conditions. TiO has been
2
2
2
is structurally simpler and scales better than the two-compart-
used as a photocatalyst for H O production. However, it pro-
2
2
[
2]
ment H based fuel cells. Despite these advantages, the pro-
duced a very low yield (several micromoles in hours) owing to
its wide band gap (>3 eV), or transparency to most of the visi-
ble light spectrum, and poor charge separation characteristics
2
duction of H O usually involves organic substances, produces
2
2
undesirable by-products, and is highly energy-demanding
[
3]
[5]
(
e.g., the anthraquinone oxidation process). Many alternative
caused by defects and surface states. To improve the per-
formance of low-cost TiO catalysts, the incorporation of noble
2
[6]
metals or surface modification has been reported. For exam-
[
a] R. Wang, K. Pan, D. Han, Dr. L. Zhang, Prof. Dr. X. Xiang
State Key Laboratory of Chemical Resource Engineering
Beijing University of Chemical Technology
ple, Shiraishi et al. prepared a bimetallic AuAg-loaded TiO2
photocatalyst that facilitated the charge separation of the pho-
1
5 Beisanhuan Dong Lu
togenerated carriers and the decomposition of H O was sup-
2
2
Beijing 100029 (P.R. China)
E-mail: xiangxu@mail.buct.edu.cn
pressed owing to the weakened adsorption of H O on the
2
2
[
6a]
TiO surface. Maurino et al. demonstrated that the modifica-
2
[
b] J. Jiang, Dr. C. Xiang, Dr. Z. Huang, Dr. L. Zhang
Joint Center for Artificial Photosynthesis
California Institute of Technology
ꢀ
tion with F ions could effectively increase the catalytic activity
[6e]
toward H O . Zhao and his colleagues reported that phos-
2
2
Pasadena, California, 91125 (USA)
phate-modified TiO decreased the aggregation and decompo-
2
[6f]
[
c] Dr. Z. Huang
AFM Business Unit
Bruker Nano Surfaces
Santa Barbara, California, 93117 (USA)
sition of H O on the surface. In addition to TiO , a graphitic
2
2
2
carbon nitride (g-C N ) photocatalyst has also shown high se-
3
4
lectivity toward H O in the presence of organic proton
2
2
[7]
donors. Carbon nitride has been demonstrated to be a photo-
catalyst for oxygen reduction by photoexcited electrons or se-
Supporting Information for this article can be found under:
http://dx.doi.org/10.1002/cssc.201600704.
ChemSusChem 2016, 9, 1 – 11
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
lective oxidation of organic substances by surface-reaching
high dispersion of metal oxides on C N , a NiFe-layered double
3
4
[
8]
holes. Lately, new C N S polymer-based materials have been
hydroxide (LDH) was utilized as a precursor for preparing the
MMO. This strategy allowed more efficient exposure of the cat-
3
3 3
[
9]
developed as efficient photocatalysts.
Particularly, the
Cd (C N S ) coordination polymer showed robust photocata-
alytically active sites on the MMO relative to the conventional
3
3
3 3 2
[
9a]
[12]
lytic H O production from a methanol/water solution.
solid-state synthesis.
The H O2 was photocatalytically pro-
2
2
2
Despite the recent advancements, it is still a great challenge
to produce H O efficiently at a large scale using solar energy
duced only from water and dioxygen without any sacrificial
carbon-containing organic substances such as alcohols. This
work provides a promising way for clean and sustainable pro-
duction of H O owing to the facile carbon-free synthesis, inex-
2
2
[
10]
and earth-abundant elements. Sacrificial hole acceptors such
as alcohols have often been used to reduce the probability of
electron–hole recombination, and thus facilitate the targeted
2
2
pensive chemical sources, and extremely accessible feedstocks
(H O and O ).
[
10b–d]
reaction of dioxygen reduction into H O .
These systems
2
2
2
2
were not sustainable and resulted in the loss of chemical
energy, the introduction of side products, and an increase in
cost owing to the use of chemicals other than water. An ener-
getically and economically wiser strategy would be to direct
Results and Discussion
The morphology of C N and MMO@C N were characterized
3
4
3
4
[
11]
the photogenerated holes to oxidize water. For instance, Fu-
kuzumi and co-workers reported sustainable production of
H O from H O and O by using a combination of a water oxi-
by TEM and AFM. Figure 1A shows a typical structure of the
bare C N matrix. In contrast, the MMO@C N composite exhib-
3
4
3
4
ited particulate features with an average size of approximately
5.5 nm (Figure 1B). Figure 1C shows the high-resolution TEM
(HRTEM) image of the MMO@C N composite, in which the d-
2
2
2
2
3
+
dation catalyst (WOC), a photosensitizer, and a Lewis acid (Sc
[
11b]
ions), without using any sacrificial species.
Shiraishi and co-
3
4
workers have also demonstrated the production of H O using
spacing in the (111) plane of NiO was measured to be
0.236 nm (inset). Furthermore, the MMO consisted of Ni, Fe,
and O and the elements of the matrix were carbon and nitro-
gen, as shown in the EDS mapping (Figure 1D). The iron
oxides were also confirmed to be highly dispersed in the NiO
phase.
2
2
g-C N4 photocatalyst incorporated with pyromellitic diimide
3
[
10a]
(
PDI) units.
In this work, an all-earth abundant, mixed-metal oxide/
graphitic-C N4 (MMO@C N ) was synthesized as a dual-func-
3
3
4
tional photocatalyst for both the water oxidation and the
oxygen reduction reactions. The two catalytic reactions were
The surface topography was studied by AFM in PeakForce
[13]
well-coupled on the MMO@C N4 composite. To ensure the
tapping mode. At the micrometer scale, the C N matrix in
3 4
3
Figure 1. TEM images of (A) C
.236 nm, which is identified with the (111) plane of NiO. (D) EDS mapping of MMO@C
tion of C, N, Ni, Fe, and O. The dark contrasts were oxide particles. The particles showed a Ni-rich feature. AFM images of (E) C
3
N
4
, (B) MMO@C
3
N
4
, and (C) HRTEM image of MMO@C
3
N
4
; the HRTEM lattice image in the inset of (C) shows a d-spacing of
(scale bar 500 nm). The elemental mappings showed the distribu-
and (F) MMO@C with
0
3 4
N
3
N
4
3 4
N
a surface roughness of 0.199 and 0.438 nm, respectively. On the scale of <500 nm, (E) and (F) show the surface morphologies of the flat surface regions of
these samples. The scale bars on the AFM images are 40 nm.
&
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
both the C N and MMO@C N samples showed hill and valley
tected (Figure 2G). The molar ratio of Ni/Fe in MMO was mea-
sured by inductively coupled plasma atomic emission spectros-
copy (ICP-AES) to be 2.8:1 (deviation <10%), as expected from
the nominal value (3:1) in the initial feedstock. The monome-
tallic nickel and ferric samples after calcination were also ana-
lyzed to confirm the phase composition (Figure S2 and S3). To
exclude the possibility that Ni and Fe directly coordinate with
C N , FTIR measurements were conducted and compared. The
3
4
3
4
features with average heights of hundreds of nanometers (Fig-
ure S1A and S1B, Supporting Information). However, at the
sub-micrometer scale, the C N4 had a fairly smooth surface
3
(
Figure S1C), whereas the MMO@C N composite had a much
3 4
rougher surface and showed particulate features (Figure S1D).
Figure 1E and 1F show detailed views of the C N4 and
3
MMO@C N surfaces on a 200 nm range. The surface rough-
3
4
3
4
ness of the C N and MMO@C N was 0.199 nm and 0.438 nm,
FTIR spectrum of MMO@C N exhibited the same absorption
3 4
3
4
3
4
respectively. The surface morphology obtained from the AFM
measurements was consistent with the TEM observations, in
which the MMO@C N showed particles with an average size
bands to that of C N (Figure S4). No new bands appeared for
3 4
MMO@C N . The results verified that the direct coordination of
3
4
Ni and Fe with C N is less likely.
3
4
3
4
of approximately 5.5 nm dispersed onto the surface of the
C3N4 matrix. The specific surface area of MMO@C N was
The valence and binding information of the elements were
determined by XPS. The core level spectra of Ni2p in
MMO@C N were deconvoluted into two components in the
3
4
2
ꢀ1
1
8.6 m g , approximately 50% larger than the pristine C N
3 4
3
4
2
ꢀ1
(12.6 m g ). This could be ascribed to high dispersion of
Ni2p3/2 region (Figure 3A). The component peak at 855.4 eV
MMO on the C N matrix.
was higher than that of pure phase NiO (at ꢁ854.0 eV), partial-
3
4
3
+
Figure 2A–E shows the typical XRD patterns of C N ,
ly because the high doping of Fe with large electronegativity
3
4
2
+ [16]
LDH@C N , MMO@C N , Ni@C N , and Fe@C N All the reflec-
led to a higher valence state of Ni
.
The peak at 861.5 eV
3
4
3
4
3
4
3
4.
2
+
tion peaks were assigned to the graphitic carbon nitride
g-C N ) with (100) and (002) lines, which come from in-plane
was assigned to the satellite of Ni (sat.). The XPS analysis
verified that the divalent NiO was a dominant species, and the
(
3
4
3
+
repeating tri-s-triazine and stacking of the conjugated aromatic
Ni
was below the detection limit. For the Fe2p spectra,
[
14]
units, respectively. No diffractions from the oxides or LDH
were observed, indicating that they were highly dispersed into
the C N matrix. The formation of single phase LDH precursor
three deconvoluted peaks appeared in the Fe2p3/2 region (Fig-
2
+
ure 3B). The peak at 710.2 eV is characteristic of Fe , and the
3
+ [17]
neighboring one at 712.1 eV is the contribution of Fe
.
The
3
4
was also verified by XRD (Figure 2F). The diffraction lines were
indexed to a hydrotalcite-like structure (JCPDS 40-0215), which
co-existence of divalent and trivalent ferric species indicated
2
+
that Fe -containing species such as Fe O4 could exist. The
3
[
15]
was consistent with the literature report.
The pattern of
analyses of the N1s spectra confirmed the existence of surface
defects in the C N matrix. Both N1s spectra of the C N and
MMO derived from an LDH precursor was indexed to a cubic
NiO phase (JCPDS 71-1179). No other crystalline phase was de-
3
4
3
4
MMO@C N4 samples were deconvoluted into three compo-
3
nents (Figure 3C and D). For the C N sample, the peaks at
3
4
2
3
98.9, 400.2, and 401.3 eV were assigned to sp -hybridized N
atoms of the melem units (Npyridine), trigonal N atoms of the
melem center (Ncenter) and tertiary amine N atoms (Ntertiary) of
the melem terminal, and the primary (Nprimary) and secondary
[18]
amine(Nsecondary) N atoms of the melem terminal, respectively.
The peak positions of these three components were within
.1 eV of those in MMO@C N . However, the contribution of
0
3
4
the component (Nprimary +Nsecondary) decreased in the N1s spec-
tra of MMO@C N . This was because the sites of the primary
3
4
and secondary amine moieties were occupied by and coupled
with the MMO particles. XPS is a surface-sensitive technique
with a detection depth of several nanometers. The signal in-
tensity of the specific sites on the surface decreases if these
sites are partially occupied by foreign species.
The UV/Vis diffuse reflectance spectra of the composites and
the controls are shown in Figure 4. The spectra were obtained
by converting the reflection data measured using the Kubelka–
2
[19]
Munk equation F(R)=(1-R) /2R, where R is the reflectance.
The C N sample exhibited strong absorption below 420 nm,
3
4
[8b]
which was consistent with the previous reports.
The
MMO@C N composite presented enhanced absorption below
3
4
4
00 nm relative to the C N . The Ni@C N sample also exhibited
3 4 3 4
stronger absorption in the UV wavelength region. In contrast,
the mixture (MMO/C N -Mix) showed a small decrease in the
3
4
absorption below 400 nm compared to the C N . This finding
Figure 2. XRD patterns of the samples (A) C
3
N
4
, (B) LDH@C
3
N
4
,
3
4
(
C) MMO@C , (D) Ni@C , (E) Fe@C , (F) LDH precursor, and (G) MMO.
3
N
4
3
N
4
3
N
4
indicated that the improved absorption of MMO@C N likely
3 4
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
3 4 3 4 3 4
Figure 3. XPS core level spectra of (A) Ni2p and (B) Fe2p in MMO@C N , N1s in (C) C N , and in (D) MMO@C N .
Figure 4. UV/Vis diffuse reflectance spectra of C
Fe@C , and MMO/C -Mix.
3 4 3 4 3 4
N , MMO@C N , Ni@C N ,
3
N
4
3 4
N
originated from the strong interaction among the components
[
20]
rather than a simple physical mixing effect.
The formations of H O on various photocatalysts were con-
2
2
ducted and monitored in an O -equilibrated acidic aqueous so-
2
ꢀ
2
lution (pH 3) under irradiation with a Xe lamp at 100 mWcm .
As shown in Figure 5A, the H O was rapidly generated over
2
2
the MMO@C N photocatalyst within the first 30 min, with an
3
4
Figure 5. (A) The light-driven H O generation in O -equilibrated conditions
2
2
2
ꢀ
1
initial rate that exceeded 4.0 mmmin . The concentration of
the photogenerated H O reached a plateau value of 63 mm
over MMO@C
3
N
4
, Ni@C
3
N
4
, Fe@C
3
N
4
, and MMO/C
3
N
4
-Mix; the semi-hollow
photocatalyst.
3 4
circles (blue) represents the repeated use of the MMO@C N
The experimental conditions were as follows: 1 gL photocatalyst, pH 3, illu-
2
2
ꢀ
1
after approximately 90 min. In the absence of photons and O2
ꢀ2
mination intensity 100 mWcm . (B) Formation rate constant (K
f
) and decom-
in the system, no H O2 generation was observed (Figure S5
2
position rate constant (K ) for H O production.
d
2
2
and S6). The content of MMO present in the MMO@C N was
3
4
optimized; the use of more or less MMO led to decreased H O2
2
productivity (Figure S7). Table 1 shows the sustainable produc-
experiments using the C N , MMO, Ni@C N , Fe@C N , and
3 4 3 4 3 4
tion of H O2 from water and oxygen without any sacrificial
a mixture of MMO with C N as the photocatalysts were also
3 4
2
hole acceptors. The photocatalytic activity of the MMO@C N4
performed. First, both C N and MMO samples produced unde-
3 4
3
composite is among the best-reported photocatalysts to date.
To elucidate the origin of the photocatalytic species, control
tectable H O (<0.01 mm after 210 min) under the same condi-
2 2
tions. Second, as summarized in Table 1, the yield of H O on
2
2
&
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
depended on the K rather than K in the initial stage,
f
d
Table 1. Comparisons of H
reported values in the literature.
2
O
2
production on the as-prepared photocatalysts and the
during which the formation greatly outperformed
the decomposition. This is inconsistent with the
latest results observed for a reduced graphene oxide
[
a]
[c]
Photocatalysts
Initial rate
Rate constants
2 2
H O yield
ꢀ
1
ꢀ1
ꢀ1
[b]
ꢀ1
[
mmmin
]
K
f
[mmmin
]
K
d
[min
]
K
d
’
[min
]
[mm]
(
rGO)/TiO2 photocatalyst and its control samples,
MMO@C
MMO/C
Ni@C
Fe@C
g-C
rGO/TiO
3
N
4
4.0
0.33
0.57
0.61
–
5.3094
0.4629
0.5542
0.7019
–
0.0836
0.0566
0.0184
0.0333
–
0.0054
–
0.0032
0.0021
–
–
63
whereby the catalyst with the smallest Kd had the
3
N
4
-Mix
8.3
24.7
19.0
31
[11a]
best overall production of H O .
2
2
3
N
4
To study the stability of the MMO@C N photocata-
3
4
3
N
4
[
d]
lyst, the used catalyst sample was separated from the
solution after the reactions, washed, and dried. Sub-
sequently, the photocatalyst was re-used in a fresh
solution for monitoring the generation of H O under
3
N
4
/PDI
/CoPi
[
e]
2
–
–
–
66
[
a] Initial rate was calculated using the yield of H
concentration: C =5 mm. [c] Yield at 90 min. [d] Data from Ref. [10a]. [e] Data
from Ref. [11a].
2 2
O after the first 10 min. [b] Initial
H
2
O
2
0
2
2
the same conditions. As shown in Figure 5A (semi-
hollow circles in blue), the performance of the recy-
cled MMO@C N4 photocatalyst exhibited marginal
3
the MMO@C N composite after 90 min was 3.3 times higher
degradation (<5%) as compared to that of the fresh sample.
The performance was almost unchanged even after it was re-
cycled 5 times (Figure S9). The structural characterizations by
IR, XRD, and HRTEM proved that the MMO@C N was stable
3
4
than Fe@C N , 2.5 times higher than Ni@C N , and 7.6 times
3
4
3
4
higher than the mixture MMO/C N -Mix with the same MMO
3
4
loading on C N . These control experiments provided two im-
3
4
3
4
portant pieces of evidence that the catalytic enhancement re-
sulted from the chemical interaction between MMO and C N
after it was used in catalytic reactions (Figure S10, S11, and
S12). In addition, the elemental analyses by ICP indicated that
the etching of Ni and Fe in the MMO@C N was below 5%
3
4
in the MMO@C N composite. First, the sum of the yields after
3
4
3
4
9
0 min from both Ni@C N and Fe@C N were 43.7 mm. The
after the reactions. This high repeatability of photocatalytic ac-
tivity is not only a strong evidence of the stability of the mate-
rial but also has a great value for practical use.
3
4
3
4
sum of the yields was still 31% lower than that of the compo-
site. Second, the simple mixture only resulted in the initial pro-
ꢀ
1
duction rate of 0.33 mmmin and 12% yield (8.3 mm) after
0min relative to the composite. These two pieces of evidence
Suppression of the decomposition of H O2 is an efficient
2
9
way to increase the yield of light-driven H O production. The
2
2
suggested that the strong interaction between MMO and C N4
use of TiO nanoparticles as a photocatalyst, for example, the
3
2
ꢀ
induced by in situ hydrothermal growth and subsequent calci-
nation is favorable to endow the photocatalyst with excellent
activity for H O production.
surface modification of TiO by F ions, enhances the activity
2
[6e]
toward H O production. In this work, the decomposition be-
2
2
havior of H O in the presence of photocatalysts was further in-
2
2
2
2
The formation and decomposition of H O on the photocata-
vestigated. The experiments were conducted in an acidic solu-
2
2
lysts proceeds through two competitive pathways owing to
the thermodynamic instability of H O at room temperatur-
tion (pH 3) with an initial H O concentration of 5 mm. The
2
2
concentration of H O2 decreased by <7% over C N after
2
2
2
3
4
[
6e,11a]
e.
To further understand the mechanism of H O genera-
90 min illumination (Figure 6). In contrast, the H O2 decom-
2
2
2
tion, a well-established kinetic model shown in Equation (1)
was used to understand the behavior of the H O concentra-
posed approximately 40% over MMO@C N4 after 90 min,
3
higher than those over the control Ni@C N4 (25%) and
2
2
3
tion as a function of time.
Fe@C N (19%). In the absence of any catalysts, the decompo-
3
4
sition of H O was less than 3% after 90 min illumination.
2
2
Kf
d
K t
Owing to the large quantity of H O (5 mm) and the con-
2 2
stant amounts of the catalysts, the decomposition is expected
to follow a pseudo-first-order reaction [Eq. (2)]:
½
H O ꢂ ¼ ð1 ꢀ e
Þ
ð1Þ
2
2
Kd
In Equation (1), K and K are the formation and decomposi-
f
d
[
21]
tion rate constants, respectively. In this model, the formation
rate follows zero-order kinetics owing to the continuous O2
purge, whereas the decomposition rate follows first-order ki-
netics. The dashed lines in Figure 5A are the fitting results
using Equation (1). According to the parameters fitted (Fig-
ure S8), K and K were compared in Table 1 and plotted in Fig-
dCðtÞ
0
ꢀ
¼ k00 or CðtÞ ¼ ꢀk t þ C
ð2Þ
0
0
dt
where C(t) is the H O2 concentration at time t, k ’ is the
pseudo-first-order rate constant, and C is the initial concentra-
2
0
0
tion of H O (5 mm) in this work. To more accurately derive the
f
d
2
2
ure 5B. The formation rate, i.e., K of MMO@C N was approxi-
rate constants, we fit these decomposition plots using a first-
f
3
4
mately one order of magnitude larger than the other three
order reaction equation, where k1 is the first-order reaction
ꢀ
1
control samples (5.3 vs. 0.4–0.7 mmmin ). The MMO@C N
rate constant. The results indicated that the order of k fol-
3
4
1
ꢀ
1
ꢀ1
composite also exhibited a higher decomposition rate, approx-
imately 1.5–4.5 times higher relative to the control samples.
However, the decomposition rate of H O even for this most
lowed MMO@C N4 (0.0054 min )>Ni@C N (0.0032 min )>
3 3 4
ꢀ
1
ꢀ1
Fe@C N4 (0.0021 min )>C N (0.00075 min ) (Figure S13).
3
3
4
Compared to the decomposition rate in the photocatalytic
system, MMO@C N remained the most active among all the
2
2
active MMO@C N composite was fairly slow, i.e., approximate-
3
4
3
4
ꢀ
1
ly 0.0836 min . Consequently, the generation of H O mostly
catalysts, i.e., had the largest decomposition rate constant. Fur-
2
2
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
2 2 0 3 4
Figure 6. The photocatalytic decomposition of H O (C =5 mm) over C N ,
MMO@C
3
N
4
, Ni@C
3
N
4
, and Fe@C
3
N
4
. The experimental conditions were as
ꢀ
1
ꢀ2
follows: 1 gL photocatalyst, pH 3, illumination intensity 100 mWcm
.
thermore, all the rate constants were one order of magnitude
lower than those for the photocatalytic reactions used for the
generation of H O . This indicated that photons activated
2
2
these catalysts to reduce kinetic barrier so that both the for-
ward and backward reactions were accelerated. However,
owing to the sluggish kinetics, the impact of illumination on
the back reactions only played a trivial role in the total genera-
tion of H O .
2
2
The production of H O also depended on the pH of the so-
2
2
lution, as the generation of H O is a proton-coupled electron
2
2
[
11a,b]
transfer process.
In this regard, we tested three different
Figure 7. (A) Linear sweep voltammetry (LSV) curves of catalyst-loaded elec-
pH values (1, 3, and 5) in an aqueous solution of HClO (Fig-
4
trodes for C
3 4 3 4
N and MMO@C N toward oxygen-evolving reaction, (B) LSV
ure S14). At pH 5, the MMO@C N only produced approximate-
curves of catalyst-loaded electrodes for C N and MMO@C N under N - or
3
4
3
4
3
4
2
ly 7 mm of H O after 210 min, which was one order of magni-
O
2
-equilibrated conditions, respectively. N
2
-equilibrated conditions: Before
for 12 h. O -equilibrated condi-
2
purge during the experiments. Acidic solution (pH 3),
2
2
reaction, the solution was purged with N
tions: continuous O
scan rate 10 mVs
2
2
tude lower than at pH 3 and pH 1. In the case of pH 3 and
pH 1, the plots showed a similar formation rate up to approxi-
mately 30 min and subsequently separated from each other
with yields of 66 mm at pH 3 and 88 mm at pH 1 after 210 min
ꢀ1
.
(Figure S14). This indicated that a higher pH results in faster
H O decomposition. In addition, to exclude the effect of urea
(LSV) curves of the electrodes with the catalysts from this
work. The catalysts were dispersed in Nafion solution and cast
onto the FTO conductive glass as working electrodes. The
2
2
hydrolysis on the H O2 generation over C N , the control
2
3
4
sample C N (urea) was prepared (see details in the Experimen-
3
4
tal Section). This control showed negligible activity toward
H O generation under the same operational conditions (Fig-
onset potential of C N4 for water oxidation was at approxi-
3
mately 2.1 V vs. reversible hydrogen electrode (RHE) in the
dark, which had a 0.9 V overpotential because the thermody-
namic potential was 1.23 V vs. RHE. In addition, the slow in-
crease of the current density (J) with bias (E) indicated a high
kinetic barrier. However, for the MMO@C N , the onset poten-
2
2
ure S15).
We have extensively characterized the MMO@C N photoca-
3
4
talysts that were suspended into the solution for the genera-
tion of H O . All the spectroscopic and photocatalytic studies
2
2
3
4
above indicated that the chemical interaction and structural
optimization favored enhanced charge separation, facilitating
the formation of H O . Furthermore, these studies also sug-
tial negatively shifted to approximately 200 mV, and the cur-
rent density was greatly enhanced compared to the C N . For
3
4
example, at 1.75 V (vs. Ag/AgCl), the current density of
MMO@C N4 had 4.6 times increase (0.00723 vs.
0.00161 mAmg ). Also, MMO@C N showed semiconductor/
2
2
gested that the MMO@C N acted as a dual-functional catalyst
a
3
4
3
ꢀ
1
for both water oxidation and O reduction. To confirm this, we
2
3
4
also performed the studies focused on the half reactions, i.e.,
water oxidation and oxygen reduction. We prepared catalyst-
casted fluorine-doped tin oxide (FTO) electrodes and these
electrodes were used in a three-electrode electrochemical cell.
Such a setup has been used to evaluate transition metal
liquid junction characteristics featured by a leaky diode behav-
ior of the J–E curve. In addition, the MMO@C N was compared
3
4
with the control LDH@C N in oxygen-evolving reaction (OER).
3
4
Although the LDH@C N has a comparable OER performance
3
4
to the MMO@C N (Figure S16), the former is not applicable in
3
4
[
16a,22]
oxides serving as WOCs.
For example, in relation to this
the present acidic solution (pH 3) because of the alkaline fea-
work, Fe-doped NiO or NiFeO have been reported to be effi-
ture of LDH. The LDH@C N4 also showed inferior activity
x
3
[
23]
cient WOCs. Figure 7A shows linear sweep voltammogram
toward H O generation (Figure S17).
2
2
&
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
We also investigated the catalytic activities of the FTO elec-
trodes with MMO@C N and the C N control for O reduction.
conductor materials. The flat band potentials (E ) of C N and
fb
3
4
MMO@C N in contact with an electrolyte at pH 3 were esti-
3
4
3
4
2
3
4
As shown in Figure 7B, in the absence of O , C N had a current
mated by electrochemical Mott–Schottky measurements at
varied frequencies. The Mott–Schottky plots of both samples
exhibited positive slopes, suggestive of their n-type semicon-
2
3
4
ꢀ
1
density <ꢀ0.05 mAmg at a bias of ꢀ1.0 V vs. Ag/AgCl,
which was likely owing to proton reduction. Although
MMO@C N roughly doubled this value, it remained less than
[25]
ductor features (Figure 9). The E value was estimated by ex-
3
4
fb
ꢀ
1
ꢀ
0.1 mAmg when the potential was more positive than
ꢀ
1.0 V vs. Ag/AgCl. In the continuous O purge, these current
2
ꢀ
1
density values were greatly improved, e.g., ꢀ0.138 mAmg
ꢀ
1
for C N and ꢀ0.22 mAmg for MMO@C N at ꢀ1.0 V vs. Ag/
3
4
3
4
AgCl. The persistent increase of the reduction current and the
absence of a current plateau indicated a two-electron reduc-
ꢀ
[24]
tion of O to HOO . This 60% enhancement in current densi-
2
ty of MMO@C N relative to C N confirmed that the composite
3
4
3
4
was also a good oxygen reduction catalyst (ORC) in addition to
a WOC. This proved that MMO@C N served as both a WOC
3
4
and an ORC, i.e., a dual-function catalyst in the case of a sus-
pension solution. In addition, comparing Figure 7A,B, the cur-
rent density of water oxidation was approximately one order
of magnitude lower than that of O reduction. This was expect-
2
ed as O evolution is a proton-coupled four-electron process
2
whereas O reduction only requires two electrons.
2
The chronoamperometric traces of catalyst-loaded electro-
des were further measured under chopped light to compare
the photoresponsive behaviors. The C N showed a steady-
3
4
ꢀ
1
state photocurrent density of 0.11 mAg for water oxidation
upon light-on (Figure 8). In contrast, the MMO@C N had a pho-
3
4
Figure 9. Mott–Schottky plots of (A) C
tion (pH 3) in the dark. The frequency used was 1.0 kHz, 1.5 kHz, and
.0 kHz.
3 4 3 4
N and (B) MMO@C N in acidic solu-
2
trapolating the linear portion of the curve to intercept the x-
Figure 8. Current density–time traces of catalyst-loaded electrodes under
chopped light in acidic solution (pH 3) for C , MMO@C , Ni@C , and
Fe@C . The potential applied was 1.5 V vs. Ag/AgCl. Back illumination, in-
tensity 100 mWcm
[26]
axis. A flat band potential of ꢀ0.71 V and ꢀ0.026 V vs. RHE
3
N
4
3
N
4
3 4
N
was obtained for C N and MMO@C N , respectively. The con-
3
4
3
4
3
N
4
ꢀ
2
duction band could be 0–0.1 V more negative than the flat
band position. Hence, the corresponding conduction band
edge of C N and MMO@C N was estimated to be ꢀ0.81 V
.
3
4
3
4
ꢀ
1
tocurrent density of 0.28 mAg , which was 2.5 times higher
than the bare C N . Furthermore, the photocurrent was stable
and ꢀ0.126 V vs. RHE, respectively. A positive shift of 0.684 V
from the bare C N to the MMO@C N composite was ob-
3
4
3
4
3
4
and had little decay under illumination. As for the controls,
Ni@C N and Fe@C N had a much lower photocurrent density
served. The positive shift of the conduction band in the com-
posite was likely to improve the selectivity of the oxygen re-
3
4
3
4
ꢀ
1
[10a]
(
0.13–0.14 mAg ) than MMO@C N . This indicated that the
duction to H O .
3
4
2
2
MMO (Fe-doped NiO) exhibited higher water oxidation capabil-
ity than the respective monometallic counterparts. As a conse-
quence, the photoelectrochemical (PEC) measurements con-
firmed the superiority of MMO@C N for water oxidation.
The Tauc plots of C N and MMO@C N were derived accord-
3 4 3 4
2
ing to the relationship of (F(R)·hu) =A(hu-E ), where h is
g
Plank’s constant, u is the frequency, A is a constant, and E is
g
[19,27]
the material band gap (Figure S18).
From the Tauc plots,
3
4
The activity and selectivity of the oxygen reduction reactions
are highly dependent on the band edge positions of the semi-
the MMO@C N has a band gap of 2.89 eV, which is 0.01 eV
3
4
larger than that of C N (2.88 eV). The band gap value of C N
4
3
4
3
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[28]
was close to that of bulk g-C N reported in the literature.
eration of H O in an acidic solution. The MMO particles con-
3
4
2
2
The valence band (VB) positions were estimated according to
the band gap and the conduction band (CB) positions. Both
the CB and VB were more positive in MMO@C N (Figure 10).
sisting of Fe-doped NiO were chemically incorporated into the
graphitic phase C N matrix. The formation and decomposition
3
4
kinetics studies of H O indicated that the MMO@C N compo-
3
4
2
2
3
4
site possessed much higher productivity of H O relative to the
2
2
controls. The production of H O also depended on the pH of
2
2
the solution. Lower pH resulted in higher yield of H O . The en-
2
2
hanced productivity of H O was ascribed to the improved ca-
2
2
pability toward water oxidation because the valence band of
the MMO@C N catalyst was positively shifted as compared to
3
4
the bare C N . Both the cyclic voltammetry (CV) measurements
3
4
in the dark and the photoelectrochemical current density–time
J–t) verified the superiority of MMO@C N for water oxidation.
(
3
4
Simultaneously, the positive shift of the conduction band in
the composite improved the selectivity of the two-electron re-
duction of dioxygen toward the formation of H O . Tuning of
2
2
the band edges was achieved by chemical coupling of MMO
and C N on the nanoscale, which could be an important path-
3
4
Figure 10. Scheme of energy levels and charge transfer pathways of C
3 4
N
way for the design of C N -based composites with improved
3
4
and MMO@C . The positive shift of the valence band in the MMO@C N
3
N
4
3 4
oxygen reduction performance. This study dictates a promising
way for clean and sustainable production of H O using only
promoted hole-involved water oxidation and the positive shift of the con-
duction band improved the selectivity of the oxygen reduction through
2
2
a two-electron pathway. The potentials of O
cated.
2 2 2 2 2
/H O and H O/O are also indi-
water and oxygen over an earth-abundant catalyst.
Experimental Section
Because the water oxidation is a rate-limiting step, the positive
shift of the VB in MMO@C N facilitated to overcome the kinet-
Materials
3
4
Melamine (1,3,5-triazine-2,4,6-triamine, C H N , 99%), Ni(NO ) ·6H O
3
6
6
3 2
2
ic barrier of the sluggish O evolution. In addition, its CB level
2
(nickel nitrate hexahydrate), Fe(NO ) ·9H O (iron nitrate nonahy-
3 3 2
was 0.816 V more negative than the reduction potential of O /
2
drate), urea, H SO (sulfuric acid), HClO (perchloric acid), KH PO
4
(potassium dihydrogen phosphate), and K HPO ·3H O (potassium
2 4 2
2
4
4
2
[
29]
H O (0.69 V vs. RHE). This finding was very similar to that of
2
2
g-C N /PDI, in which both the CB and VB levels of g-C N were
3
4
3
4
hydrogen phosphate, trihydrate) were purchased from Sinopharm
Chemical Reagent (Beijing Co. Ltd.). All reagents were of analytical
grade and were used without further purification. Deionized water
was used throughout the experiments.
[
10a]
shifted positively with increasing PDI units.
It was confirmed
that the positive shifts of VB enhance the capability of g-C N4
3
for water oxidation and provide sufficient potentials for O re-
2
duction.
According to the results and the energy levels analyzed
above, we proposed a tentative mechanism for H O produc-
Preparation of graphitic-Carbon Nitride (g-C N )
3
4
2
2
The g-C N was prepared by directly heating of a melamine precur-
tion over the MMO@C N photocatalyst. The light irradiation
3
4
3
4
[14]
sor, as described by Chen et al. In a typical synthesis, melamine
3.0 g) was added to a porcelain cup and calcined at 5208C for 4 h
on MMO@C N led to enhanced charge separation owing to
3
4
(
the reduced band gap and the presence of MMO as a WOC.
The photogenerated holes were involved in the water oxida-
tion promoted by the chemical incorporation of MMO owing
to the positive shift of the valence band level. The separated
electrons reduced dioxygen by a proton-coupled two-electron
pathway to generate H O . The more positive potential of the
ꢀ
1
with a heating rate of 48Cmin . Grinding of the product resulted
in yellow powders.
Synthesis of LDH@C N and MMO@C N composites
3
4
3
4
2
2
The LDH@C N4 precursor was synthesized by a hydrothermal
3
valence band elevated the driving force to climb the large bar-
rier for water oxidation, efficiently promoting photogenerated
hole separation and involvement in surface oxidation reactions.
Simultaneously, the sufficient potential difference (0.816 V) be-
tween the conduction band of MMO@C N (ꢀ0.126 V vs. RHE)
method using urea as a precipitator. In a typical procedure,
Ni(NO ) ·6H O (0.3 mmol), Fe(NO ) ·9H O (0.1 mmol), and urea
3
2
2
3 3
2
(2 mmol) were dissolved in 50 mL of water and stirred to form
a clear solution. The as-synthesized C N powder was added to the
3
4
3
+
solution and sonicated for 1 h. The molar ratio of C N to the Fe
3 4
3
4
ions was 300:1. The suspension was transferred to a Teflon-lined
stainless-steel autoclave and maintained at 1208C for 12 h. The re-
sultant suspension was filtered and washed with water until the
pH of the filtrate reached 7. The wet products were dried at 708C
in a vacuum oven overnight. The obtained LDH@C N composites
and O reduction potential (0.69 V vs. RHE) guaranteed the oc-
2
currence of O reduction reactions.
2
Conclusions
3
4
ꢀ
1
were calcined at 3008C for 1 h in air (heating rate of 58Cmin ),
during which LDH was transformed into MMO. The products were
denoted as MMO@C N . To exclude the effect of urea hydrolysis on
A dual-functional MMO@C N4 composite photocatalyst was
3
synthesized from an LDH@C N precursor for light-driven gen-
3
4
3
4
&
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
the structure of C N , the as-synthesized C N powder was subject
sequently, the used MMO@C N4 photocatalyst was reused in
3
4
3
4
3
to the same urea hydrolysis reaction (at 1208C, 12 h) without the
addition of metal ions feedstock. The product was calcined at
a fresh solution under the same conditions. To investigate the de-
composition behavior of H O over the photocatalysts, a sample of
1 gL was dispersed in H O solution (initial concentration: 5 mm,
2 2
2
2
ꢀ
1
3
008C for 1 h. The sample was denoted as C N (urea). Single phase
3
4
LDH was prepared using the same conditions with no addition of
C N . In addition, MMO was produced by calcining the LDH precur-
sor at 3008C for 1 h in air.
pH 3) and irradiated for 90 min under continuous stirring.
The concentration of hydrogen peroxide was determined accord-
3
4
[11]
ing to the literature. The stock solutions of N,N-diethyl-1,4-phen-
ylene-diamine sulfate (DPD), peroxidase (POD) horseradish, and po-
tassium phosphate buffer were prepared as follows. DPD (0.1 g)
was dissolved in 10 mL of 0.1N H SO solution, and POD (5 mg)
For comparisons, monometallic nickel or ferric oxides and the cor-
responding composites with C N were prepared by the same pro-
3
4
cedure only adding single metal salts during the synthesis. The
2
4
2
+
molar ratio of C N to the Ni ions was 100:1 and C N to the
Fe
was dissolved in 5 mL of water. The POD solution was kept in a re-
frigerator (48C) and was prepared once every five days.
K HPO ·3H O (1.4 g) and KH PO (6 g) were dissolved in 100 mL of
3
4
3
4
3
+
ions was 300:1, which was the same as the ratio in
MMO@C N . The products were named Ni@C N and Fe@C N . In
3
4
3
4
3
4
2
4
2
2
4
addition, a mixture of MMO and C N was prepared by a grinding-
water to make potassium phosphate buffer solution. Two milliliter
aliquots were removed by a syringe during the irradiation at fixed
times and filtered through a 0.45 mm polytetrafluoroethylene
(PTFE) filter. Phosphate buffer (0.8 mL), 2 mL of the sample ali-
quots, DPD (0.1 mL), POD (0.1 mL), and 2.24 mL of water were
mixed and stirred vigorously for 90 sec. Depending on the concen-
tration of H O , the ratio of the sample aliquot/water was changed
3
4
heating method. The C N4 powder was mixed with LDH in an
3
agate mortar. The mixture was calcined at 3008C for 1 h in air. The
product was named MMO/C N -Mix. The ratio of C N to MMO was
3
4
3
4
the same as that in MMO@C N .
3
4
2
2
Characterizations
to avoid exceeding the detection limit of this method. The calibra-
tion curves of H O concentration were made based on the range
2
2
HRTEM images were recorded using a JEOL JEM-3010 microscope.
For TEM observations, the samples were ultrasonically dispersed in
ethanol, and then a drop of the suspension was deposited onto
a carbon-coated copper grid followed by evaporation of the sol-
vent in air. Energy dispersive X-ray spectroscopy (EDS) mapping
analysis was performed by using a JEOL JEM-2010F microscope
of the H O concentration. The absorbance at 551 nm was mea-
2
2
sured using a UV/Vis spectrophotometer for quantitative analyses.
Electrochemical tests
N
combined with an EDX (OxFord X-Max 80-TLE) spectrometer. Sur-
The electrochemical water oxidation of the samples was studied
using a three-electrode cell connected to a potentiostat (CHI660E,
CH instrument Co. USA). The Ag/AgCl electrode and Pt wire were
used as reference and counter electrode, respectively. The working
electrode was fabricated by immobilizing samples on the FTO con-
ductive glass by a casting method. The as-prepared samples
face topography imaging was conducted on a Bruker Dimension
ICON AFM. The mode for surface mapping was PeakForce Quanti-
tative Nano Mechanics. Bruker MPP-12120-10 probes were used.
The normal mechanical specifications of the rectangular cantilever
were a spring constant of 5 N/m and a resonance frequency
1
50 kHz. For AFM imaging, powder samples were cast onto FTO
(10 mg) were dispersed in a solution (ethanol: 800 mL, Nafion:
electrodes as described in the “Electrochemical tests” section
below. The XRD patterns of the samples were collected on a Shi-
madzu XRD-6000 diffractometer (40 kV, 30 mA, graphite-filtered
CuKa radiation, l=0.15418 nm). Elemental analyses were per-
formed using ICP-AES on a Shimadzu ICPS-7500 spectrometer. The
XPS spectra were recorded on a ThermoVG ESCALAB MK II X-ray
2
00 mL), and sonicated for 3 h. Subsequently, the resultant suspen-
sion was cast dropwise on the surface of the FTO. The FTO electro-
des immobilized with the samples were connected to a copper
tape and used as the working electrode. The LSV measurements
were obtained in an aqueous solution of HClO (pH 3) at room
4
ꢀ1
temperature at a scan rate of 10 mVs . Mott–Schottky plots were
measured in an acidic solution (pH 3) at a frequency of 1 kHz in
the dark. The potentials measured were converted to reversible hy-
drogen electrode (RHE) based on the formula [Eq. (3)]:
ꢀ
9
photoelectron spectrometer at a pressure of about 2ꢁ10 Pa by
using AlK X-ray as the excitation source (1486.6 eV). The positions
a
of all binding energies were calibrated by using the C1s line at
2
84.8 eV. Solid-state UV/Vis absorption spectra were recorded at
room temperature using Shimadzu UV-3000 spectrometer
a
ERHE ¼ EAg=AgCl þ EAg=AgCl vs: NHE þ 0:059 pH
ð3Þ
equipped with an integrating sphere attachment using BaSO as
4
a background.
in which EAg/AgCl vs. NHE is the correction factor for Ag/AgCl with re-
spect to the normal hydrogen electrode (NHE), and equals 0.197 V
at 208C.
Photocatalytic production and decomposition of H O2
2
ꢀ
1
The catalyst (1 gL ) was added to water (30 mL) within a borosili-
cate glass bottle (f=35 mm, capacity 50 mL) and was dispersed
by ultrasonication for 10 min. The pH of the suspension was ad- Acknowledgements
ꢀ
1
justed to pH 3 by adding HClO solution (1 molL ). The bottle was
4
sealed with a rubber septum cap with a gas inlet and outlet. The
solution was first purged by oxygen bubbling while stirring in dark
for 30 min. The solution was illuminated using a 300 W Xenon
This work was supported by the 973 Program (Grant
014CB932104), the National Natural Science Foundation of
China (NSFC), Beijing Natural Science Foundation (Grant
152022), and the Fundamental Research Funds for the Central
2
ꢀ2
lamp at 100 mWcm under magnetic stirring. The oxygen purging
was maintained throughout the reaction. The solution was sam-
pled at certain times to determine the concentration of H O . After
2
Universities (YS1406). This work was also partly supported by the
Joint Center for Artificial Photosynthesis, a DOE Energy Innova-
tion Hub, supported through the Office of Science of the U.S. De-
partment of Energy under Award Number DE-SC0004993. AFM
2
2
the first photocatalytic H O2 formation run for 210 min, the
2
MMO@C N photocatalyst was removed by centrifugation and
3
4
washed thoroughly with water and dried under a flow of N . Sub-
2
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
data were collected at the Molecular Materials Research Center
of the Beckman Institute of the California Institute of Technology.
1568–1576; e) W. He, R. Wang, L. Zhang, J. Zhu, X. Xiang, F. Li, J. Mater.
Chem. A 2015, 3, 17977–17982.
13] Z. Q. Huang, J. J. Jiang, Y. M. Hua, C. Z. Li, M. Wagner, H. J. Lewerenz,
M. P. Soriaga, C. Anfuso, Energy Environ. Focus. 2015, 4, 260–277.
14] J. Chen, S. H. Shen, P. H. Guo, P. Wu, L. J. Guo, J. Mater. Chem. A 2014, 2,
[
[
Keywords: hydrogen peroxide · mixed metal oxide · oxygen
reduction · photocatalysis · sustainable chemistry
4605–4612.
[
[
15] G. Fan, X. Xiang, J. Fan, F. Li, J. Mater. Chem. 2010, 20, 7378–7385.
16] a) J. Qi, W. Zhang, R. Xiang, K. Liu, H.-Y. Wang, M. Chen, Y. Han, R. Cao,
Adv. Sci. 2015, 2, 1500199; b) G. Natu, P. Hasin, Z. J. Huang, Z. Q. Ji, M. F.
He, Y. Y. Wu, ACS Appl. Mater. Interfaces 2012, 4, 5922–5929; c) L. Troto-
chaud, S. L. Young, J. K. Ranney, S. W. Boettcher, J. Am. Chem. Soc. 2014,
[1] R. Hage, A. Lienke, Angew. Chem. Int. Ed. 2006, 45, 206–222; Angew.
Chem. 2006, 118, 212–229.
[2] a) Y. Yamada, S. Yoshida, T. Honda, S. Fukuzumi, Energy Environ. Sci.
2011, 4, 2822–2825; b) Y. Yamada, M. Yoneda, S. Fukuzumi, Chem. Eur.
1
36, 6744–6753.
J. 2013, 19, 11733–11741; c) L. Han, S. J. Guo, P. Wang, S. J. Dong, Adv.
Energy Mater. 2015, 5, 1400424.
[
[
[
[
[
17] R. Dedryve‘re, M. Maccario, L. Croguennec, F. L. Cras, C. Delmas, D.
Gonbea, Chem. Mater. 2008, 20, 7164–7170.
18] Y. Shiraishi, Y. Kofuji, H. Sakamoto, S. Tanaka, S. Ichikawa, T. Hirai, ACS
Catal. 2015, 5, 3058–3066.
19] L.-F. Gao, T. Wen, J.-Y. Xu, X.-P. Zhai, M. Zhao, G.-W. Hu, P. Chen, Q.
Wang, H.-L. Zhang, ACS Appl. Mater. Interfaces 2016, 8, 617–624.
20] B. Li, Y. Zhao, S. Zhang, W. Gao, M. Wei, ACS Appl. Mater. Interfaces
[
3] J. M. Campos-Martin, G. Blanco-Brieva, J. L. G. Fierro, Angew. Chem. Int.
Ed. 2006, 45, 6962–6984; Angew. Chem. 2006, 118, 7116–7139.
4] a) Y. Yamada, A. Nomura, T. Miyahigashi, S. Fukuzumi, Chem. Commun.
[
2012, 48, 8329–8331; b) V. Maurino, C. Minero, G. Mariella, E. Pelizzetti,
Chem. Commun. 2005, 2627–2629; c) T. Hirakawa, Y. Nosaka, J. Phys.
Chem. C 2008, 112, 15818–15823.
2
013, 5, 10233–10239.
21] a) A. J. Hoffman, E. R. Carraway, M. R. Hoffman, Environ. Sci. Technol.
994, 28, 776–785; b) C. Kormann, D. W. Bahnemann, M. R. Hoffman,
[
[
5] H. Park, Y. Park, W. Kim, W. Choi, J. Photochem. Photobiol. C 2013, 15, 1–
20.
1
6] a) D. Tsukamoto, A. Shiro, Y. Shiraishi, Y. Sugano, S. Ichikawa, S. Tanaka,
T. Hirai, ACS Catal. 2012, 2, 599–603; b) M. Teranishi, S. Naya, H. Tada, J.
Am. Chem. Soc. 2010, 132, 7850–7851; c) H. Sheng, Q. Li, W. H. Ma,
H. W. Ji, C. C. Chen, J. C. Zhao, Appl. Catal. B 2013, 138–139, 212–218;
d) A. H. Michael, Surf. Sci. Rep. 2011, 66, 185–297; e) V. Maurino, C.
Minero, G. Mariella, E. Pelizzetti, Chem. Commun. 2005, 2627–2629; f) H.
Sheng, H. Ji, W. Ma, C. Chen, J. Zhao, Angew. Chem. Int. Ed. 2013, 52,
Environ. Sci. Technol. 1988, 22, 798–806.
[
22] a) D. Friebel, M. W. Louie, M. Bajdich, K. E. Sanwald, Y. Cai, A. M. Wise,
M. J. Cheng, D. Sokaras, T. C. Weng, R. A. Mori, R. C. Davis, J. R. Bargar,
J. K. Nørskov, A. Nilsson, A. T. Bell, J. Am. Chem. Soc. 2015, 137, 1305–
1
313; b) R. D. L. Smith, M. S. Prꢂvot, R. D. Fagan, Z. P. Zhang, P. A.
Sedach, M. K. J. Siu, S. Trudel, C. P. Berlinguette, Science 2013, 340, 60–
3; c) R. D. L. Smith, M. S. Prꢂvot, R. D. Fagan, S. Trudel, C. P. Berlingu-
6
9686–9690; Angew. Chem. 2013, 125, 9868–9872.
ette, J. Am. Chem. Soc. 2013, 135, 11580–11586; d) A. Indra, P. W. Me-
nezes, N. R. Sahraie, A. Bergmann, C. Das, M. Tallarida, D. Schmeißer, P.
Strasser, M. Driess, J. Am. Chem. Soc. 2014, 136, 17530–17536.
[
[
7] Y. Shiraishi, S. Kanazawa, Y. Sugano, D. Tsukamoto, H. Sakamoto, S. Ichi-
kawa, T. Hirai, ACS Catal. 2014, 4, 774–780.
8] a) X. C. Wang, S. Blechert, M. Antonietti, ACS Catal. 2012, 2, 1596–1606;
b) Y. Wang, X. C. Wang, M. Antonietti, Angew. Chem. Int. Ed. 2012, 51,
[
23] a) K. Fominykh, P. Chernev, I. Zaharieva, J. Sicklinger, G. Stefanic, A. Dçb-
linger, M. Mꢃller, A. Pokharel, S. Bçcklein, C. Scheu, T. Bein, D. F. Rohlf-
ing, ACS Nano 2015, 9, 5180–5188; b) J. Landon, E. Demeter, N. C. Ke-
turakis, I. E. Wachs, R. Vasi c´ , A. I. Frenkel, J. R. Kitchin, ACS Catal. 2012, 2,
68–89; Angew. Chem. 2012, 124, 70–92; c) R. C. Pawar, V. Khare, C. S.
Lee, Dalton Trans. 2014, 43, 12514–12527; d) R. C. Pawar, S. Kang, S. H.
Ahn, C. S. Lee, RSC Adv. 2015, 5, 24281–24292; e) R. C. Pawar, Y. Pyo,
S. H. Ahn, C. S. Lee, Appl. Catal. B 2015, 176–177, 654–666; f) Y. Cao, Z.
Zhang, J. Long, J. Liang, H. Lin, H. Lin, X. Wang, J. Mater. Chem. A 2014,
1
793–1801; c) Y. F. Li, A. Selloni, ACS Catal. 2014, 4, 1148–1153; d) J. A.
Bau, E. J. Luber, J. M. Buriak, ACS Appl. Mater. Interfaces 2015, 7, 19755–
9763.
24] S. J. Guo, S. Zhang, L. H. Wu, S. H. Sun, Angew. Chem. Int. Ed. 2012, 51,
1770–11773; Angew. Chem. 2012, 124, 11940–11943.
25] C. Zhao, H. Luo, F. Chen, P. Zhang, L. Yi, K. You, Energy Environ. Sci.
014, 7, 1700–1707.
26] R. L. Spray, K. J. McDonald, K.-S. Choi, J. Phys. Chem. C 2011, 115, 3497–
506.
1
2, 17797–17807.
[
[
[
[
[
[
[
9] a) H. Zhuang, L. Yang, J. Xu, F. Li, Z. Zhang, H. Lin, J. Long, X. Wang, Sci.
Rep. 2015, 5, 16947; b) J. Xu, L. Luo, G. Xiao, Z. Zhang, H. Lin, X. Wang,
J. Long, ACS Catal. 2014, 4, 3302–3306.
1
2
[10] a) Y. Shiraishi, S. Kanazawa, Y. Kofuji, H. Sakamoto, S. Ichikawa, S. Tanaka,
T. Hirai, Angew. Chem. Int. Ed. 2014, 53, 13454–13459; Angew. Chem.
3
2
014, 126, 13672–13677; b) H. Goto, Y. Hanada, T. Ohno, M. Matsumura,
J. Catal. 2004, 225, 223–229; c) R. Cai, Y. Kubota, A. Fujishima, J. Catal.
003, 219, 214–218; d) Y. Shiraishi, S. Kanazawa, D. Tsukamoto, A. Shiro,
27] J. Liu, Y. Liu, N. Liu, Y. Han, X. Zhang, H. Huang, Y. Lifshitz, S.-T. Lee, J.
Zhong, Z. Kang, Science 2015, 347, 970–974.
28] Y. Gong, J. Wang, Z. Wei, P. Zhang, H. Li, Y. Wang, ChemSusChem 2014,
2
Y. Sugano, T. Hirai, ACS Catal. 2013, 3, 2222–2227.
7, 2303–2309.
[
[
11] a) G. Moon, W. Kim, A. D. Bokare, N. Sung, W. Choi, Energy Environ. Sci.
29] J. Chen, P. Wagner, L. Tong, G. G. Wallace, D. L. Officer, G. F. Sweigers,
2014, 7, 4023–4028; b) S. Kato, J. Jung, T. Suenobu, S. Fukuzumi, Energy
Angew. Chem. Int. Ed. 2012, 51, 1907–1910; Angew. Chem. 2012, 124,
Environ. Sci. 2013, 6, 3756–3764; c) N. Kaynan, B. A. Berke, O. Hazut, R.
Yerushalmi, J. Mater. Chem. A 2014, 2, 13822–13826.
1943–1946.
12] a) X. Xiang, H. I. Hima, H. Wang, F. Li, Chem. Mater. 2008, 20, 1173–
1
182; b) X. Xiang, L. S. Xie, Z. W. Li, F. Li, Chem. Eng. J. 2013, 221, 222–
29; c) X. Xiang, L. Bai, F. Li, AIChE J. 2010, 56, 2934–2945; d) W. He, Y.
Yang, L. Wang, J. Yang, X. Xiang, D. Yan, F. Li, ChemSusChem 2015, 8,
2
Received: May 26, 2016
Published online on && &&, 0000
&
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
10
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
Just add water: An all-earth abundant
R. Wang, K. Pan, D. Han, J. Jiang,
mixed metal oxide/graphitic-C N com-
C. Xiang, Z. Huang, L. Zhang, X. Xiang*
3
4
posite is constructed and acts as a dual-
functional photocatalyst for solar-driven
H O generation without any sacrificial
&
& – &&
2
2
Solar-Driven H O Generation From
2
2
organic substances, which provides
a promising pathway for clean and sus-
tainable production of a green oxidant
H O and O Using Earth-Abundant
2 2
Mixed-Metal Oxide@Carbon Nitride
Photocatalysts
(
H O ) from only water and dioxygen.
2 2
ChemSusChem 2016, 9, 1 – 11
www.chemsuschem.org
11
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ