Highly selective production of hydrogen peroxide on graphitic carbon nitride (g-C3N4) photocatalyst activated by visible light

Article abstract of DOI:10.1021/cs401208c

Photocatalytic production of hydrogen peroxide (H₂O₂) on semiconductor catalysts with alcohol as a hydrogen source and molecular oxygen (O₂) as an oxygen source is a potential method for safe H ₂O₂ synthesis because the reaction can be carried out without the use of explosive H₂/O₂ mixed gases. Early reported photocatalytic systems, however, produce H₂O₂ with significantly low selectivity (～1%). We found that visible light irradiation (λ > 420 nm) of graphitic carbon nitride (g-C ₃N₄), a polymeric semiconductor, in an alcohol/water mixture with O₂ efficiently produces H₂O₂ with very high selectivity (～90%). Raman spectroscopy and electron spin resonance analysis revealed that the high H₂O₂ selectivity is due to the efficient formation of 1,4-endoperoxide species on the g-C ₃N₄ surface. This suppresses one-electron reduction of O₂ (superoxide radical formation), resulting in selective promotion of two-electron reduction of O₂ (H₂O₂ formation).

Full text of DOI:10.1021/cs401208c

Research Article
pubs.acs.org/acscatalysis
Terms of Use
Highly Selective Production of Hydrogen Peroxide on Graphitic
Carbon Nitride (g‑C N ) Photocatalyst Activated by Visible Light
3
4
,
†
†
†
†
†
Yasuhiro Shiraishi,* Shunsuke Kanazawa, Yoshitsune Sugano, Daijiro Tsukamoto,
Hirokatsu Sakamoto, Satoshi Ichikawa, and Takayuki Hirai
‡
†
†Research Center for Solar Energy Chemistry and Division of Chemical Engineering, Graduate School of Engineering Science, Osaka
University, Toyonaka 560-8531, Japan
‡
Institute for NanoScience Design, Osaka University, Toyonaka 560-8531, Japan
*
S Supporting Information
ABSTRACT: Photocatalytic production of hydrogen peroxide (H O ) on
2
2
semiconductor catalysts with alcohol as a hydrogen source and molecular
oxygen (O ) as an oxygen source is a potential method for safe H O
2
2
2
synthesis because the reaction can be carried out without the use of
explosive H /O mixed gases. Early reported photocatalytic systems,
2
2
however, produce H O with significantly low selectivity (∼1%). We found
2
2
that visible light irradiation (λ > 420 nm) of graphitic carbon nitride (g-
C N ), a polymeric semiconductor, in an alcohol/water mixture with O efficiently produces H O with very high selectivity
3
4
2
2
2
(
∼90%). Raman spectroscopy and electron spin resonance analysis revealed that the high H O selectivity is due to the efficient
2
2
formation of 1,4-endoperoxide species on the g-C N surface. This suppresses one-electron reduction of O (superoxide radical
3
4
2
formation), resulting in selective promotion of two-electron reduction of O (H O formation).
2
2
2
KEYWORDS: photocatalysis, hydrogen peroxide, visible light, surface chemistry, reduction
INTRODUCTION
selectivity for the amount of H O formed relative to the
2 2
8
■
−11
amount of alcohol consumed is less than 1%.
Several kinds
13
Hydrogen peroxide (H O ) is a clean oxidant that emits only
water as a byproduct and is widely used in industry for organic
synthesis, pulp bleaching, and disinfection. At present, H O is
commercially produced by the anthraquinone method, which
has some nongreen features, such as high energy utilization,
because of the multistep hydrogenation and oxidation reactions.
Recently, direct production of H O with H and O gases has
2
2
12
of TiO catalysts modified with fluoride or loaded with Au
2
14
or Au−Ag alloy nanoparticles have been proposed; however,
2
2
1
all of these systems exhibit insufficient H O selectivity. This is
2
2
because one-electron reduction of O occurs predominantly
2
15
and produces a superoxide anion (eq 3). This suppresses two-
electron reduction of O (eq 2), resulting in insufficient H O
2
2
2
2
2
2
2
2
−4
selectivity.
been studied extensively with Pd
or Au−Pd bimetallic
This direct synthesis can be an alternative process
for H O synthesis from the viewpoint of green chemistry,
5
−7
catalysts.
+
−
•
O + H + e → OOH
(3)
2
2
2
although some care is necessary for operation because of the
explosive nature of H /O mixed gases.
Another reason for the low H O selectivity is that the formed
2
2
2
2
H O is subsequently decomposed by absorbing UV light
2
2
Photocatalytic H O production on semiconductor materials
16
2
2
(
∼400 nm). Efficient H O production, therefore, requires
2
2
such as titanium dioxide (TiO ) has also been studied
2
8
−11
selective promotion of two-electron reduction of O and
2
extensively.
The reactions are usually carried out by UV
suppression of subsequent photodecomposition of the formed
H O .
irradiation (λ < 400 nm) of an O -saturated water with TiO in
2
2
2
2
the presence of alcohols as the electron and proton donor.
Graphitic carbon nitride (g-C N ) is a metal-free polymeric
−
3
4
Photoexcitation of TiO produces the electron (e ) and
2
photocatalyst, with a graphitic stacking structure of the C N
+
+
3
4
positive hole (h ) pairs. The h oxidize alcohol and produce
aldehyde and protons, whereas the e promote two-electron
layers consisting of tri-s-triazine units connected through amino
−
17,18
groups (Scheme 1).
This material is successfully activated
reduction of O and produce H O .
2
2
2
by the irradiation of visible light (∼480 nm) and exhibits high
1
9
+
+
photocatalytic activity for H or O evolution from water and
R−CH OH + 2h → R−CHO + 2H
(1)
(2)
2
2
2
20
degradation of organic pollutants. Electrochemical analysis of
g-C revealed that the bottom of its conduction band is −1.3
+
−
N
3 4
O + 2H + 2e → H O
2
2
2
The reaction proceeds without H gas at room temperature and
Received: December 17, 2013
Revised: January 14, 2014
Published: January 21, 2014
2
atmospheric pressure, therefore, a clean and safe H O2
2
synthesis. Its efficiency is, however, very low because the
©
2014 American Chemical Society
774
dx.doi.org/10.1021/cs401208c | ACS Catal. 2014, 4, 774−780

ACS Catalysis
Research Article
Scheme 1. (a) Graphitic Stacking Structure of g-C N Sheet
3
4
a
and (b) Photograph of the Prepared g-C N
3
4
a
The gray and blue spheres are the C and N atoms, respectively, and
the second and third layers of the sheets are green and brown,
respectively.
2
1
V (vs NHE), which is more negative than the reduction
2
2
potential of O (−0.28 V vs NHE). This suggests that, as
2
2
3−26
reported,
under visible light irradiation.
Here, we report that g-C N activated by visible light in water
g-C N is able to promote the reduction of O
3
4
2
3
4
with alcohol and O2 selectively promotes two-electron
Figure 1. (a) Absorption spectrum of g-C N and action spectrum for
3
4
reduction of O while suppressing subsequent photodecompo-
2
H O formation. Apparent quantum yield for H O formation was
2
2
2
2
sition of the formed H O . This thus facilitates highly efficient
2
2
calculated with the equation {Φ
(%) = ([H O formed] × 2)/
photon number entered into the reaction vessel] × 100}. The inset
formation
2 2
production of H O with more than 90% selectivity. Raman
2
2
[
spectroscopy and electron spin resonance (ESR) analysis
shows the change in H O2 selectivity as a function of excitation
2
revealed that selective two-electron reduction of O on the
2
wavelength. (b) Absorption spectrum of H O (420 μM) and action
2
2
photoexcited g-C N is due to the efficient formation of 1,4-
spectrum for H O2 decomposition calculated using the equation
3
4
2
endoperoxide species on its surface, which is selectively
transformed to H O .
{Φ
(%) = [H O decomposed]/[photon number entered
decomposition 2 2
into the reaction vessel] × 100}.
2
2
RESULTS AND DISCUSSION
■
The g-C N photocatalyst was simply prepared by calcination
formed is <0.2 μmol, although a large amount of acetaldehyde
(83 μmol) is produced. The selectivity for the amount of H O
3
4
2
7
of cyanamide as yellow powders (Scheme 1b). As shown in2
2
2
Figure 1a, the obtained powders (BET surface area, 10.3 m
formed relative to the amounts of photooxidation products (=
−1
g ) show strong absorption in the visible region; its band gap
[H O ]/[acetaldehyde + acetic acid + (CO )/2] × 100) is
2
2
2
is 2.62 eV (= 474 nm). The X-ray diffraction (XRD) pattern of
<0.2%, suggesting that two-electron reduction of O does not
2
g-C N4 (Figure S1, Supporting Information) shows a
occur selectively.
As shown by entry 2, TiO loaded with 1 wt % Pt particles
3
distinctive signal assigned to the (002) packing of the C N
3
4
2
28
layers at 2θ = 27.4 (d = 0.325 nm). X-ray photoelectron
(Pt /TiO ) produces a much larger amount of aldehyde (>0.1
1
2
spectroscopy (XPS) of g-C N (Figure S2, Supporting
Information) shows a C 1s peak at 288 eV assigned to the
mmol) because of the enhanced charge separation by the
creation of a Schottky barrier; however, the H O selectivity
2 2
3
4
31
NC−N groups of the triazine ring and a N 1s peak at 398−
scarcely changes (<0.1%). In addition, TiO loaded with Au
2
2
2
29
32
3
99 eV assigned to the sp -hybridized N atoms (CN−C).
particles or modified with fluoride (entries 3 and 4), which
1
2,13
The transmission electron microscopy (TEM) image of g-C N
has been proposed for H O production,
still shows very
3
4
2
2
shows a sheet-like structure (Figure S3, Supporting Informa-
low H O selectivity (∼6.3%). Furthermore, nitrogen-doped
2
2
30
33
tion), as often observed for related materials. These data
suggest that, as shown in Scheme 1a, g-C N indeed possesses a
TiO , a well-known visible light-responsive catalyst, is also
2
ineffective for H O formation (entry 5). In contrast, as shown
3
4
2
2
stacking structure of the C N layers consisting of tri-s-triazine
units.
by entry 6, photoreaction with g-C N produces a very large
amount of H O (30 μmol) with significantly high selectivity
2 2
3
4
3 4
Photocatalytic reactions were performed with ethanol
EtOH) as the electron and proton donor. An EtOH/water
mixture (9/1 v/v, 5 mL) containing catalyst (20 mg) was
(90%). This indicates that g-C N specifically promotes two-
electron reduction of O2 and facilitates selective H O
2 2
3 4
(
production. It is noted that oxidation of water by the
photoformed hole on the semiconductor surface is also a
photoirradiated by a Xe lamp (λ > 420 nm) with magnetic
34
stirring under O atmosphere (1 atm) at 298 K. Table 1
possible mechanism for H O formation. In the present g-
2
2 2
summarizes the results obtained by 12 h of photoreaction with
the respective catalysts. It must be noted that all of the systems
produce acetaldehyde as the main photooxidation product of
C N system, visible light irradiation in pure water with O
3 4 2
scarcely produces H O (<0.2 μmol), although the reaction
2
2
with EtOH produces a much larger amount of H O (30 μmol,
2
2
EtOH, along with minor amounts of acetic acid and CO (mass
entry 6). This clearly suggests that oxidation of water by the
photoformed hole scarcely contributes to the H O production,
2
balance > 95%). With bare TiO (entry 1), the amount of H O
2
2
2
2
2
7
75
dx.doi.org/10.1021/cs401208c | ACS Catal. 2014, 4, 774−780

ACS Catalysis
Research Article
a
Table 1. Results for Photocatalytic Production of H O from EtOH and O with Various Catalysts
2
2
2
b
g
c
d
d
d
e
entry
catalyst
H O /μmol
CH CHO/μmol
CH COOH/μmol
CO /μmol
H O selectivity/%
2
2
3
3
2
2
2
f
1
2
3
4
5
6
7
8
TiO2
Pt /TiO
<0.2
<0.2
3.0
83
188
47
0.1
3.6
3.4
3.0
1.5
1.4
2.1
1.5
1.9
5.6
<0.2
<0.1
6.3
g
1
2
Au /TiO
1
N.D.
N.D.
N.D.
0.7
2
F-TiO2
N-TiO2
g-C N
<0.2
<0.2
30
50
<0.4
<0.8
90
23
32
3
4
g
Pt0.1/g-C N
4.2
38
N.D.
6.7
11
3
4
g
Pt /g-C N
4
1.2
308
0.4
1
3
a
−2
b
Photoirradiation was carried out with a 2 kW Xe lamp (light intensity at 420−500 nm was 26.9 W m ). The subscript texts (M ) denote the
x
c
d
amount of metal loaded on the support (x = wt %). Determined by redox titration with KMnO (detection limit: 0.2 μmol). Determined by GC.
4
e
f
=
[H O ]/([CH CHO] + [CH COOH] + [CO ]/2) × 100. JRC-TIO-4 TiO supplied from the Catalyst Society of Japan (equivalent to Degussa
2 2 3 3 2 2
2
g
−1
P25; particle size, 24 nm; BET surface area, 57 m g ; anatase/rutile = ca. 83/17). Average diameters of metal particles determined by TEM
observations are 3.1 ± 1.1 nm (Pt /TiO ), 3.4 ± 0.9 nm (Au /TiO ), 2.3 ± 0.5 nm (Pt /g-C N ), and 4.9 ± 2.2 nm (Pt /g-C N ), respectively (see
1
2
1
2
0.1
3
4
1
3
4
Figure S4, Supporting Information).
leaving that two-electron reduction of O2 is the main
mechanism for H O formation.
2
2
Figure S5 (Supporting Information) shows time-dependent
change in the amount of products during photoreaction of an
EtOH/water/O system with g-C N . Even after prolonged
2
3
4
photoirradiation (∼24 h), the H O selectivity is unchanged
2
2
(
∼90%), and the rate of H O formation is almost constant. In
2
2
addition, as shown in Table S1 (Supporting Information), use
of other aliphatic or benzylic alcohols also exhibits high H O2
2
selectivity (ca. 90%). These data suggest that g-C N activated
3
4
by visible light with alcohol selectively produces H O .
2
2
Figure 1a (black circle) shows the action spectrum for H O2
2
formation, obtained by photoreaction of an EtOH/water/O2
system with g-C N under irradiation of monochromatic
3
4
3
5
light. The apparent quantum yield for H O formation
2
2
(
Φ
3
formation) almost agrees with the absorption spectrum of g-
C N , indicating that photoexcitation of g-C N promotes
4
3
4
H O formation. The inset of Figure 1a shows the change in
2
2
H O2 selectivity as a function of excitation wavelength.
2
Irradiation with >400 nm light shows high H O selectivity
2
2
(
>90%), but the selectivity decreases with <400 nm light
irradiation. This is because the formed H O is decomposed by
2
2
1
6
absorbing UV light. Figure 1b (white circle) shows the action
spectrum for H O decomposition obtained by photo-
2
2
irradiation of a water (3 mL) containing H O (18 μmol)
2
2
with O2 at 298 K. The apparent quantum yield for
decomposition (Φdecomposition) agrees with the absorption
spectrum of H O and increases significantly with <400 nm
Figure 2. (a) Effect of reaction temperature on the amount of
products and H selectivity during photoreaction with g-C in an
O
N
3 4
2
2
EtOH/water/O system. Reaction conditions are identical to those in
2
2
2
Table 1. (b) Effect of water content in solution on the amount of
light irradiation. The result is consistent with the decreased
H O selectivity by <400 nm light irradiation (Figure 1a, inset).
products and H O selectivity during photoreaction with g-C N in an
2
2
3
4
2
2
EtOH/O system.
2
These data clearly suggest that photoexcitation of g-C N by
3
4
visible light suppresses subsequent photodecomposition of
H O and facilitates selective H O formation.
2
2
2
2
but it increases up to ∼90% by the addition of ≥10% water.
In the present g-C N system, reaction temperature (≤298
3
4
This is probably because the stabilization of H O by the
2
2
37
K) and addition of water to alcohol (≥10%) are very important
hydration with water molecules suppresses decomposition of
H O . These data suggest that photoreaction with water at
factors for selective H O formation. As shown in Figure 2a,
2
2
2
2
photoreaction of an EtOH/water/O system with g-C N at
around room temperature promotes selective H O production.
2
3
4
2 2
≤
298 K produces H O with high selectivity (∼90%), but the
Mechanism for selective H O formation on the photo-
2 2
2
2
selectivity decreases significantly at >298 K. This is due to the
excited g-C N can be explained by Scheme 2. Photoexcitation
3 4
3
6
−
+
thermal decomposition of H O at elevated temperature. In
of g-C N by two photons (a) creates e −h charge-separated
−
2
2
3
4
addition, as shown in Figure 2b, photoreaction of g-C N at
state (b). The formed e are localized at the C1 and N4
3
4
+
2
98 K without water shows very low H O selectivity (63%),
positions of the triazine ring, whereas the h are localized at the
2
2
7
76
dx.doi.org/10.1021/cs401208c | ACS Catal. 2014, 4, 774−780

ACS Catalysis
Research Article
•
42,43
β
Scheme 2. Proposed Mechanism for Selective H O
assigned to the DMPO− OOH spin adduct (α = 13.7 G; α
N
H
2
2
Formation on the Surface of Photoactivated g-C N
= 9.2 G, g = 2.0066).
This suggests that one-electron
3
4
reduction of O indeed occurs on TiO , resulting in very low
2
2
selectivity for H O production (Table 1, entry 1). In contrast,
2
2
as shown in Figure 3b, the reaction with g-C N shows only a
3
4
2
3,44
very small adduct signal. This suggests that, as shown in−
Scheme 2, the superoxo radical is rapidly reduced by another e
d → f), thus suppressing one-electron reduction of O (d →
(
2
e).
The formation of 1,4-endoperoxide species (Scheme 2f) on
the photoactivated g-C N is confirmed by Raman spectrosco-
3
4
py. As shown in Figure 4A, the g-C N itself exhibits three
3
4
38
+
N2 and N6 positions. The h remove α- and β-hydrogens of
3
1
−
alcohol and produce aldehyde (c). The e on either C1 or N4
39
position reduces O and produces a superoxo radical (d). The
radical is usually released as a superoxide ( OOH) radical (e)
2
•
via the reaction with the proton, promoting one-electron
1
5
reduction of O (eq 3). In contrast, on g-C N , the superoxo
2
3
4
−
radical (d) is rapidly reduced by the e at the para position of
the triazine ring, producing a 1,4-endoperoxide species (f).
Protonation of the endoperoxide species produces H O2
40
1
5,41
2
and completes the photocatalytic cycle. These mechanisms
indicate that selective H O formation on the photoexcited g-
2
2
C N is due to the rapid formation of the 1,4-endoperoxide
3
4
species (d → f). This suppresses one-electron reduction of O2
d → e → a) and selectively promotes two-electron reduction
of O2.
(
•
The suppression of one-electron reduction of O ( OOH
2
radical formation) on g-C N is confirmed by ESR analysis with
3
4
5
,5-dimethyl-1-pyrroline N-oxide (DMPO), a spin trapping
reagent. Figure 3 shows the ESR spectra of the solution
recovered after photoreaction of an EtOH/water/O system
2
with catalyst and DMPO. As shown in Figure 3a, the spectrum
obtained after photoreaction with TiO shows strong signals
2
Figure 4. Raman spectra for (A) g-C N and the sample recovered
3
4
1
6
after photoreaction (12 h) of an EtOH/water mixture with (B) O2
and (C) 18O . Calculated Raman shift for (a) the tri-s-triazine unit and
2
16
16
18
18
the units with (b) O− O and (c) O− O 1,4-endoperoxide
species. Calculations were performed at the DFT level with the
B3LYP/6-31g(d) basis set. The white, gray, blue, red, and purple
spheres represent H, C, N, 16O, and O atoms, respectively.
18
−
1
−1
bands at 709, 753, and 982 cm . Both 709 and 982 cm bands
45
are assigned to the breathing modes of triazine ring, and the
1
7
53 cm− band is assigned to the out-of-plane bending mode of
46
graphitic domains. Figure 4a shows the Raman shift of the tri-
s-triazine unit, obtained by ab initio calculation using the
density functional theory (DFT) within the Gaussian 03
program. The obtained three bands (730, 760, and 950 cm )
are similar to the observed bands (Figure 4A), suggesting that
DFT calculation precisely represents the electronic structure of
Figure 3. ESR spectra measured at 298 K for the solutions obtained by
photoirradiation of (a) TiO or (b) g-C N in an EtOH/water/O
2
3
4
2
system. Reactions were carried out for 1 min under conditions
identical to those in Table 1 with DMPO (0.1 mmol). The spectra
obtained by 5 min irradiation are shown in Figure S6 (Supporting
Information).
−1
7
77
dx.doi.org/10.1021/cs401208c | ACS Catal. 2014, 4, 774−780

ACS Catalysis
Research Article
g-C N . Figure 4B shows the spectrum of the g-C N recovered
3
4
3
4
1
6
after photoreaction in an EtOH/water/ O system. A new
2
−1
broad band appears at 891 cm . Figure 4b depicts the
calculated Raman shift of the 1,4-endoperoxide species
adsorbed on the tri-s-triazine unit. The obtained three bands
−1
16
(
876, 912, and 929 cm ) are assigned to C1− O symmetric
16 16 16
vibration, O− O stretching vibration, and C1− O asym-
metric vibration of the 1,4-endoperoxide species, respectively1,
and agree well with the observed broad band at 891 cm−
(Figure 4B).
These data suggest that an endoperoxide species is indeed
produced on the g-C N surface. The endoperoxide formation
3
4
is further confirmed by the spectrum of g-C N recovered after
3
4
1
8
photoreaction with labeled molecular oxygen ( O ). As shown
2
−
1
in Figure 4C, the 891 cm band disappears, and a new broad
band appears at 861 cm . DFT calculation of the O− O
endoperoxide species on g-C N (Figure 4c) revealed that the
−1
18
18
3
4
−
1
isotopic shifts of the three bands are −24, −35, and −4 cm ,
respectively, which are consistent with the observed shift (Δ =
−1
−
→
30 cm ). These data suggest that, as shown in Scheme 2 (d
f), photoexcited g-C N indeed produces 1,4-endoperoxide
Figure 5. (a) Average spectral irradiance of sunlight and (b) time-
3
4
dependent change in the amount of H O and the selectivity during
species via two-electron reduction of O . Almost no formation
of OOH radical (Scheme 2; d → e) is probably because the
2
2
2
•
photoreaction of g-C
N in a 2-propanol/water/O system under
3 4 2
sunlight exposure (orange) without filter and (blue) with filter (λ >
superoxo radical is rapidly reduced (d → f) due to the
stabilization of endoperoxide species on the C1 and N4
positions of the triazine ring, as often observed for related
4
3
20 nm). Photoreaction was performed at 8:00−17:00 (north latitude
4.7°, east longitude 135.5°).
4
7,48
anthracene systems.
This thus suppresses one-electron
CONCLUSION
We found that g-C N activated by visible light irradiation
>420 nm) in water with alcohol and O produces H O with
■
(
reduction of O , thus promoting selective two-electron
2
3
4
reduction.
2 2 2
The proposed photoreaction mechanism (Scheme 2)
involving the formation of endoperoxide species on the g-
C N surface and its transformation to H O is supported by
very high selectivity. Photoexcited g-C N promotes selective
3
4
two-electron reduction of O due to the efficient formation of
2
3
4
2
2
1
,4-endoperoxide species on its surface. This suppresses one-
the photoreaction using g-C N loaded with Pt nanoparticles
3
4
electron reduction of O and promotes selective production of
2
(
Pt /g-C N and Pt /g-C N ). As shown in Table 1 (entries 6,
0.1 3 4 1 3 4
H O . Sunlight activation of g-C N also successfully promotes
2
2
3
4
7, and 8), increased Pt loadings enhance photooxidation of
selective H O2 production. At present, the photocatalytic
2
EtOH because of the enhanced charge separation by the
activity of g-C N is low; the amount of H O formed by 12 h
3
4
2
2
−
31
transfer of conduction band e to Pt particles. The increased
of reaction is only 30 μmol. The activity, however, would be
Pt loadings, however, significantly decrease the H O selectivity
improved by using g-C N materials with high surface area
2
2
3
4
49
(
<11%). This suggests that the g-C N surface indeed behaves
synthesized by polymerization of urea or silica-templated
polymerization of cyanamide. The present photoprocess has
3
4
27
as the active site for two-electron reduction of O . In addition,
2
as shown in Figure S7 (Supporting Information), the intensity
significant advantages: (i) inexpensive metal-free photocatalyst,
−1
(
(
ii) cost-free light source (sunlight), (iii) safe hydrogen source
alcohol), and (iv) mild reaction conditions (atmospheric
of the Raman band for 1,4-endoperoxide species (891 cm ) on
these catalysts decreases with the Pt loadings. These cross-
relationships clearly suggest that, as shown in Scheme 2, the
efficient formation of 1,4-endoperoxide species on the photo-
activated g-C N surface and its transformation to H O
pressure and ambient temperature). The basic concept
presented here, based on selective two-electron reduction of
O on the photoactivated g-C N surface, may contribute to the
2
3
4
3
4
2
2
design of more efficient photocatalytic systems for H O
2
2
promote highly selective formation of H O .
2
2
production and may open a new strategy toward organic
synthesis driven by sunlight.
The g-C N system successfully produces H O even under
3
4
2
2
sunlight irradiation. Figure 5b shows the time-dependent
change in the amount of H O formed and the H O selectivity
2
2
2
2
EXPERIMENTAL SECTION
■
during photoreaction with g-C N under sunlight exposure. As
27
3
4
g-C N . Cyanamide (9.0 g) was added to a porcelain cup
3
4
shown by the blue symbol, irradiation of the entire wavelength
light (>300 nm) successfully produces H O , but the selectivity
and calcined under N flow at 823 K for 4 h with the heating
2
2
2
−1
rate being 2.3 K min . Grinding of the resultant affords yellow
is only about 80% due to the subsequent photodecomposition
powders of g-C N (6.8 g).
16
3
4
of the formed H O by absorbing UV light. In contrast,
50
2
2
Pt /TiO . TiO (1.0 g) was added to water (20 mL)
1 2 2
irradiation with >420 nm light using a glass filter (orange
symbol) produces H O with high selectivity (>90%). These
containing H PtCl ·6H O (26.8 mg). Solvents were removed
2
6
2
2
2
by evaporation at 353 K with vigorous stirring. The obtained
results indicate that the g-C N system is successfully activated
powders were calcined under air flow and then reduced under
3
4
−1
by the visible region light within sunlight and promotes
selective production of H O .
H flow at 673 K. The heating rate was 2 K min , and the
2
holding time at 673 K was 2 h.
2
2
7
78
dx.doi.org/10.1021/cs401208c | ACS Catal. 2014, 4, 774−780

ACS Catalysis
Research Article
3
2
Au /TiO . TiO (1.0 g) was added to water (50 mL)
where the laser power was 100 mW and the total data
accumulation time was 15 s. The samples were prepared as
follows: After photoreaction, g-C N samples were recovered
1
2
2
containing HAuCl ·4H O (22.9 mg). The pH of the solution
4
2
was adjusted to ∼7 with 1 mM NaOH, and the solution was
stirred at 353 K for 3 h. The particles were recovered by
centrifugation, washed with water, and dried at 353 K for 12 h.
The powders were calcined under air flow, for which the
3
4
by centrifugation and dried at room temperature in vacuo. They
were mounted on a microscope slide and subjected to analysis.
ESR Measurement. The spectra were recorded at the X-
band using a Bruker EMX-10/12 spectrometer with a 100 kHz
magnetic field modulation at a microwave power level of 10.5
mW, where microwave power saturation of the signals does not
−1
heating rate was 2 K min and the holding time at 673 K was 2
h.
F-TiO2. TiO (20 mg) was added to an EtOH/water
13
2
−1
15
mixture (9/1 v/v, 5 mL) containing NaF (0.1 mol L ) within a
Pyrex glass tube (φ 12 mm; capacity, 20 mL). The solution was
then used for photocatalytic reaction.
N-TiO2. Tetraisopropyl titanate (35 mL) was added to an
ammonia solution (∼28 wt %, 100 mL) and stirred vigorously
at room temperature for 12 h. The resultant precipitate was
washed thoroughly with water and recovered by filtration. The
obtained powders were dried at 343 K for 12 h and calcined at
occur. The magnetic field was calculated using a using a 1,1-
diphenyl-2-picrylhydrazyl (DPPH) as standard. The measure-
ment was carried out as follows: catalyst (20 mg) was
suspended in an EtOH/water mixture (9/1 v/v, 5 mL) and
DMPO (0.10 mmol) within a Pyrex glass tube (φ12 mm;
capacity, 20 mL), and the tube was sealed with a rubber septum
51
cap. After ultrasonication (3 min) and O bubbling (5 min), the
2
solution was photoirradiated (λ > 420 nm) with stirring. The
catalyst was recovered by centrifugation, and the resulting
solution was subjected to analysis.
6
73 K for 3 h under air flow.
Pt /g-C N . g-C N (1.0 g) was added to water (40 mL)
x
3
4
3
4
containing H PtCl ·6H O [2.7 mg (x = 0.1) or 26.8 mg (x =
Other Analysis. Diffuse-reflectance UV−vis spectra were
measured on a UV−vis spectrophotometer (JASCO Corp.; V-
550) equipped with Integrated Sphere Apparatus ISV-469,
2
6
2
1
.0)]. Solvents were removed by evaporation at 353 K with
vigorous stirring. The obtained powders were reduced under
−1
H flow at 573 K for 2 h with the heating rate being 2 K min .
using BaSO as a reference. XRD patterns were measured on a
4
2
Photoreaction. Catalyst (20 mg) was added to an alcohol/
water mixture (9/1 v/v, 5 mL) within a Pyrex glass tube (φ 12
mm; capacity, 20 mL), and the tube was sealed with a rubber
septum cap. The catalyst was dispersed well by ultrasonication
Philips X’Pert-MPD spectrometer. XPS analysis was performed
using a JEOL JPS-9000MX spectrometer with Mg Kα radiation
as the energy source. TEM observations were performed using
an FEI Tecnai G2 20ST analytical electron microscope
52
for 3 min, and O was bubbled through the solution for 5 min.
operated at 200 kV.
2
The tube was immersed in a temperature-controlled water bath
and photoirradiated at λ > 420 nm using a 2 kW Xe lamp
ASSOCIATED CONTENT
Supporting Information
Photoreaction results with various alcohols (Table S1); XRD
pattern (Figure S1), XPS chart (Figure S2), and TEM image
■
3
5
(
4
USHIO Inc.) with magnetic stirring. The light intensity at
*
S
−
2
20−500 nm was 26.9 W m . Sunlight reactions were
performed on October 14, 2013 at 8:00−17:00 at the top of
the laboratory building (north latitude 34.7°, east longitude
(
(
(
Figure S3) of g-C N ; Size distribution of metal particles
3 4
1
35.5°). The light intensities at 300−500 nm and at 420−500
Figure S4); Time-dependent change in H O formation
Figure S5); ESR spectra for spin adduct (Figure S6); Raman
spectra for endoperoxide species on various catalysts (Figure
S7); Cartesian coordinates for tri-s-triazine unit and the unit
with 1,4-endoperoxide species. This material (PDF) is available
free of charge via the Internet at http://pubs.acs.org.
2
2
−2
nm were 8.0 and 4.1 mW cm , respectively. The highest
temperature of the solution during reaction was 303 K. After
the reaction, the gas-phase product was analyzed by GC−TCD
(
Shimadzu; GC-14B). The catalyst was recovered by
centrifugation, and the liquid-phase products were analyzed
by GC−FID (GC-2010 system). The H O amount was
2
2
4
3
determined by redox titration with KMnO . Spectral
4
AUTHOR INFORMATION
■
*
irradiance for the light source was measured with a
spectroradiometer USR-40 (USHIO Inc.).
Action Spectrum Analysis. Photoreactions were carried
Corresponding Author
E-mail: shiraish@cheng.es.osaka-u.ac.jp.
out using an EtOH/water mixture (9/1 v/v, 3 mL) and g-C N4
Notes
3
(
12 mg) within a Pyrex glass tube (φ 12 mm; capacity, 20 mL).
The authors declare no competing financial interest.
After ultrasonication and O bubbling, the tube was photo-
2
irradiated using a Xe lamp for 3 h, for which the incident light
ACKNOWLEDGMENTS
■
was monochromated by the band-pass glass filters (Asahi
32
This work was supported by a Grant-in-Aid for Scientific
Research (No. 23360349) from the Ministry of Education,
Culture, Sports, Science and Technology, Japan (MEXT).
Techno Glass Co.). The full-width at half-maximum of the
light was 11−16 nm. Photodecomposition of H O was carried
2
2
out in water (3 mL) containing H O (6 mM) within a Pyrex
2
2
glass tube. After O bubbling, the tube was irradiated by a
monochromated light using a Xe lamp for 3 h.
Calculation Details. All calculations were performed at the
DFT level within the Gaussian 03 program using the B3LYP/6-
2
REFERENCES
■
(
1) Campos-Martin, J. M.; Blanco-Brieva, G.; Fierro, J. L. G. Angew.
Chem., Int. Ed. 2006, 45, 6962−6984.
(
(
2) Lunsford, J. H. J. Catal. 2003, 216, 455−460.
31G(d) basis set. Cartesian coordinates for the tri-s-triazine unit
3) Blanco-Brieva, G.; Cano-Serrano, E.; Campos-Martin, J. M.;
and the unit with 1,4-endoperoxide species are summarized in
Supporting Information.
Fierro, J. L. G. Chem. Commun. 2004, 1184−1185.
4) Melada, S.; Rioda, R.; Menegazzo, F.; Pinna, F.; Strukul, G. J.
Catal. 2006, 239, 422−430.
(5) Choudhary, V. R.; Gaikwad, A. G.; Sansare, S. D. Angew. Chem.,
Int. Ed. 2001, 40, 1776−1779.
(
Raman Spectroscopy. The spectra were measured on a
15
confocal Raman microscope (LabRAM HR-800, HORIBA).
YAG laser (784 nm line) was used as the excitation source,
7
79
dx.doi.org/10.1021/cs401208c | ACS Catal. 2014, 4, 774−780

ACS Catalysis
Research Article
(
6) Landon, P.; Collier, P. J.; Papworth, A. J.; Kiely, C. J.; Hutchings,
G. J. Chem. Commun. 2002, 2058−2059.
7) Edwards, J. K.; Solsona, B.; Ntainjua, E.; Carley, A. F.; Herzing, A.
A.; Kiely, C. J.; Hutchings, G. J. Science 2009, 323, 1037−1041.
8) Kormann, C.; Bahnemann, D. W.; Hoffmann, M. R. Environ. Sci.
(40) Fidder, H.; Lauer, A.; Freyer, W.; Koeppe, B.; Heyne, K. J. Phys.
Chem. A 2009, 113, 6289−6296.
(
(41) Zhang, M.; Wang, Q.; Chen, C.; Zang, L.; Ma, W.; Zhao, J.
Angew. Chem., Int. Ed. 2009, 48, 6081−6084.
(
(42) Harbour, J. R.; Hair, M. L. J. Phys. Chem. 1978, 82, 1397−1399.
(43) Tsukamoto, D.; Ikeda, M.; Shiraishi, Y.; Hara, T.; Ichikuni, N.;
Tanaka, S.; Hirai, T. Chem.Eur. J. 2011, 17, 9816−9824.
(44) Zhang, P.; Wang, Y.; Li, H.; Antonietti, M. Green Chem. 2012,
Technol. 1988, 22, 798−806.
(
(
9) Cai, R.; Kubota, Y.; Fujishima, A. J. Catal. 2003, 219, 214−218.
10) Goto, H.; Hanada, Y.; Ohno, T.; Matsumura, M. J. Catal. 2004,
1
4, 1904−1908.
45) Zinin, P. V.; Ming, L. C.; Sharma, S. K.; Khabashesku, V. N.;
Liu, X.; Hong, S.; Endo, S.; Acosta, T. Chem. Phys. Lett. 2009, 472,
9−73.
46) Papadimitriou, D.; Roupakas, G.; Dimitriadis, C. A.;
Logothetidis, S. J. Appl. Phys. 2002, 92, 870−875.
47) Lauer, A.; Dobryakov, A. L.; Kovalenko, S. A.; Fidder, H.;
Heyne, K. Phys. Chem. Chem. Phys. 2011, 13, 8723−8732.
48) Kupfer, S.; Perez-Hernandez, G.; Gonzalez, L. Theor. Chem. Acc.
012, 131, 1295−1295−14.
49) Zhang, Y.; Liu, J.; Wu, G.; Chen, W. Nanoscale 2012, 4, 5300−
303.
50) Shiraishi, Y.; Sakamoto, H.; Sugano, Y.; Ichikawa, S.; Hirai, T.
ACS Nano 2013, 7, 9287−9297.
51) Higashimoto, S.; Takamatsu, K.; Azuma, M.; Kitano, M.;
Matsuoka, M.; Anpo, M. Catal. Lett. 2008, 122, 33−36.
52) Shiraishi, Y.; Tanaka, K.; Shirakawa, E.; Sugano, Y.; Ichikawa, S.;
2
25, 223−229.
11) Hirakawa, T.; Nosaka, Y. J. Phys. Chem. C 2008, 112, 15818−
5823.
12) Maurino, V.; Minero, C.; Mariella, G.; Pelizzetti, E. Chem.
Commun. 2005, 2627−2629.
13) Teranishi, M.; Naya, S.; Tada, H. J. Am. Chem. Soc. 2010, 132,
850−7851.
14) Tsukamoto, D.; Shiro, A.; Shiraishi, Y.; Sugano, Y.; Ichikawa, S.;
Tanaka, S.; Hirai, T. ACS Catal. 2012, 2, 599−603.
15) Shiraishi, Y.; Kanazawa, S.; Tsukamoto, D.; Shiro, A.; Sugano,
Y.; Hirai, T. ACS Catal. 2013, 3, 2222−2227.
16) Goldstein, S.; Aschengrau, D.; Diamant, Y.; Rabani, J. Environ.
Sci. Technol. 2007, 41, 7486−7490.
17) Wang, X.; Maeda, K.; Thomas, A.; Takanabe, K.; Xin, G.;
Carlsson, J. M.; Domen, K.; Antonietti, M. Nat. Mater. 2009, 8, 76−80.
18) Thomas, A.; Fischer, A.; Goettmann, F.; Antonietti, M.; Muller,
J. O.; Schlogl, R.; Carlsson, J. M. J. Mater. Chem. 2008, 18, 4893−4908.
19) Schwinghammer, K.; Tuffy, B.; Mesch, M. B.; Wirnhier, E.;
(
(
1
(
6
(
(
7
(
(
(
́
́
́
2
(
5
(
(
(
(
(
(
̈
(
̈
Tanaka, S.; Hirai, T. Angew. Chem., Int. Ed. 2013, 52, 8304−8308.
(
Martineau, C.; Taulelle, F.; Schnick, W.; Senker, J.; Lotsch, B. V.
Angew. Chem., Int. Ed. 2013, 52, 2435−2439.
(
20) Cui, Y.; Ding, Z.; Liu, P.; Antonietti, M.; Fu, X.; Wang, X. Phys.
Chem. Chem. Phys. 2012, 14, 1455−1462.
21) Zhang, J.; Chen, X.; Takanabe, K.; Maeda, K.; Domen, K.;
Epping, J. D.; Fu, X.; Antonietti, M.; Wang, X. Angew. Chem., Int. Ed.
010, 49, 441−444.
22) Abe, R.; Takami, H.; Murakami, N.; Ohtani, B. J. Am. Chem. Soc.
008, 130, 7780−7781.
23) Su, F.; Mathew, S. C.; Lipner, G.; Fu, X.; Antonietti, M.;
Blechert, S.; Wang, X. J. Am. Chem. Soc. 2010, 132, 16299−16301.
24) Li, X.-H.; Chen, J.-S.; Wang, X.; Sun, J.; Antonietti, M. J. Am.
Chem. Soc. 2011, 133, 8074−8077.
25) Li, X.-H.; Wang, X.; Antonietti, M. ACS Catal. 2012, 2, 2082−
086.
26) Chen, Y.; Zhang, J.; Zhang, M.; Wang, X. Chem. Sci. 2013, 4,
244−3248.
27) Wang, X.; Maeda, K.; Chen, X.; Takanabe, K.; Domen, K.; Hou,
Y.; Fu, X.; Antonietti, M. J. Am. Chem. Soc. 2009, 131, 1680−1681.
28) Ji, H.; Chang, F.; Hu, X.; Qin, W.; Shen, J. Chem. Eng. J. 2013,
18, 183−190.
29) Li, X.; Zhang, J.; Shen, L.; Ma, Y.; Lei, W.; Cui, Q.; Zou, G.
Appl. Phys. A: Mater. Sci. Process. 2009, 94, 387−392.
30) Xu, H.; Yan, J.; Xu, Y.; Song, Y.; Li, H.; Xia, J.; Huang, C.; Wan,
H. Appl. Catal., B 2013, 129, 182−193.
31) Shiraishi, Y.; Sugano, Y.; Tanaka, S.; Hirai, T. Angew. Chem., Int.
Ed. 2010, 49, 1656−1660.
32) Tsukamoto, D.; Shiraishi, Y.; Sugano, Y.; Ichikawa, S.; Tanaka,
S.; Hirai, T. J. Am. Chem. Soc. 2012, 134, 6309−6315.
33) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science
001, 293, 269−271.
34) Liu, J.; Zhang, Y.; Lu, L.; Wu, G.; Chen, W. Chem. Commun.
012, 48, 8826−8828.
35) Sugano, Y.; Shiraishi, Y.; Tsukamoto, D.; Ichikawa, S.; Tanaka,
(
2
(
2
(
(
(
2
(
3
(
(
2
(
(
(
(
(
2
(
2
(
S.; Hirai, T. Angew. Chem., Int. Ed. 2013, 52, 5295−5299.
(
(
36) Rice, F. O.; Reiff, O. M. J. Phys. Chem. 1927, 31, 1352−1356.
37) Izgorodin, A.; Izgorodina, E.; MacFarlane, D. R. Energy Environ.
Sci. 2012, 5, 9496−9501.
38) Huda, M. N.; Turner, J. A. J. Appl. Phys. 2010, 107, 123703−
23703-5.
39) Nakamura, R.; Imanishi, A.; Murakoshi, K.; Nakato, Y. J. Am.
Chem. Soc. 2003, 125, 7443−7450.
(
1
(
7
80
dx.doi.org/10.1021/cs401208c | ACS Catal. 2014, 4, 774−780