Photoinduced conversion of methane into benzene over GaN nanowires

Article abstract of DOI:10.1021/ja5004119

As a class of key building blocks in the chemical industry, aromatic compounds are mainly derived from the catalytic reforming of petroleum-based long chain hydrocarbons. The dehydroaromatization of methane can also be achieved by using zeolitic catalysts under relatively high temperature. Herein we demonstrate that Si-doped GaN nanowires (NWs) with a 97% rationally constructed m-plane can directly convert methane into benzene and molecular hydrogen under ultraviolet (UV) illumination at rt. Mechanistic studies suggest that the exposed m-plane of GaN exhibited particularly high activity toward methane C-H bond activation and the quantum efficiency increased linearly as a function of light intensity. The incorporation of a Si-donor or Mg-acceptor dopants into GaN also has a large influence on the photocatalytic performance.

Full text of DOI:10.1021/ja5004119

Communication
pubs.acs.org/JACS
Photoinduced Conversion of Methane into Benzene over GaN
Nanowires
Lu Li,†,‡ Shizhao Fan,‡ Xiaoyue Mu,† Zetian Mi,*,‡ and Chao-Jun Li*,†
†Department of Chemistry, McGill University, 801 Sherbrooke Street West, Montreal, QC H3A 0B8, Canada
‡Department of Electrical and Computer Engineering, McGill University, 3480 University Street, Montreal, QC H3A 0E9, Canada
S
* Supporting Information
large band gap and low optical absorption. Nevertheless, the
ABSTRACT: As a class of key building blocks in the
chemical industry, aromatic compounds are mainly derived
from the catalytic reforming of petroleum-based long chain
hydrocarbons. The dehydroaromatization of methane can
also be achieved by using zeolitic catalysts under relatively
high temperature. Herein we demonstrate that Si-doped
GaN nanowires (NWs) with a 97% rationally constructed
m-plane can directly convert methane into benzene and
molecular hydrogen under ultraviolet (UV) illumination at
rt. Mechanistic studies suggest that the exposed m-plane of
GaN exhibited particularly high activity toward methane
C−H bond activation and the quantum efficiency
increased linearly as a function of light intensity. The
incorporation of a Si-donor or Mg-acceptor dopants into
GaN also has a large influence on the photocatalytic
performance.
widely used metal oxide semiconductors (such as TiO2, ZnO,
and Cu2O) are not suitable supports for the methane conversion
reaction, simply because they are not stable and the lattice oxygen
can be abstracted by the produced hydrogen in the harsh gas−
solid environment.13,17
GaN, a well-known group III nitride semiconductor, has a
direct energy band gap of ∼3.4 eV at rt, which can be further
tuned across the entire solar spectrum by incorporating other
elements.18 Compared with metal oxide semiconductors, the
controlled n- and p-type doping and the inherent chemical
stability, due to the strongly ionic character of the atomic bonds,
make GaN a suitable electronically active support for the
photocatalytic reaction under harsh conditions.19 Furthermore,
compared with conventional powdered photocatalysts, nano-
wires (NWs) are highly desirable due to their large surface-to-
volume ratios, well-defined surface structures, and superior
photoelectrical properties.20 Consequently, we synthesized
nondoped GaN NWs with lengths of 800 nm grown on silicon
(111) substrate by plasma-assisted molecular beam epitaxy
(MBE) under nitrogen-rich conditions.21 Scanning electron
microscopy (SEM, Figure 1a) and transmission electron
microscopy (TEM, Figure 1b and 1c) images of the as-
synthesized GaN NWs revealed that the NWs possess hexagonal
cross sections and are vertically aligned to the substrate, with the
morphology slightly tapered from top to bottom. The electron
diffraction pattern (inset of Figure 1c) indicates that the wires are
of single crystal wurtzite structure with the growth direction
along the c-axis, with the top facet of the c-plane and lateral facet
of m-plane. The diameter distribution of the top facets derived
from SEM measurements fits well in a logarithmic normal
distribution with a mean diameter of dNWs = 100 5 nm (Figure
1d). Figure 1e shows the rt photoluminescence (PL) spectrum of
GaN NWs with an intensive peak around 365 nm, corresponding
to the band gap of 3.4 eV. To evaluate the photocatalytic
performance of the GaN semiconductor comprehensively, we
have also prepared GaN thin films (Supporting Information (SI)
Figure S1) with a thickness of 650 nm grown on sapphire using
AlN as the buffer layer and commercial powdered samples
(Figure S2). As shown in Figure 1f, due to the orientated growth
on the substrates, the only reflections of GaN NWs and thin films
obtained from X-ray diffraction (XRD) measurements are 002
and 004,22 which further confirms that the top facet of GaN NWs
he recent discovery of an enormous amount of shale gas is
T
projected to change the landscape of the chemical
industry,1 since through suitable conversions shale gas may
replace the dwindling petroleum resources as a carbon-based
feedstock.2−7 Unfortunately, due to the inert C−H bonds in
methane, there has been no easy way of turning shale gas into
synthetically useful compounds such as olefins and aromatics,8
being especially hard for the latter.9 This caused a major concern
of a great shortage of aromatic compounds for shale-gas-based
future chemical industry. So far, many efforts have been devoted
to the formation of aromatics directly from methane and several
efficient heterogeneous catalysts have been developed and well
investigated, such as Mo,10 Zn,11 and Re12 supported on ZSM-5
zeolites. However, an elevated temperature (>500 °C), owing to
the large positive Gibbs free energy [eq 1], is required to
promote the equilibrium conversions of methane.
6CH4 → C6H6 + 9H2, ΔG(298 K) = 434 kJ/mol
(1)
Besides a thermal strategy, another promising approach is to
use photoenergy to drive the conversion of methane.13 Recently,
Yoshida et al.14,15 and Chen et al.16,17 developed several
photocatalysts for methane conversion, respectively, such as
SiO2−Al2O3 oxides with highly dispersed Ti, Ga2O3, and Zn+-
ZSM-5 zeolite for the methane coupling reaction. However,
none of these powdered photocatalysts previously reported can
produce aromatic compounds and the methane conversion rate
is still low. It is noted that most of these photocatalysts are
insulator-supported (SiO2 or zeolites) materials which have a
Received: January 14, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja5004119 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
calculated to be 5.7 m2/g by measuring their densities, diameters
and lengths. The proportions of the top c-plane and lateral m-
plane are 3% and 97%, respectively. The Brunauer−Emmett−
Teller (BET) specific surface area of GaN powders is 3.2 m2/g,
and the TEM image (Figure S6) reveals that the amount of
exposed c-planes is at the same level as that of exposed m-planes
on the surface of powdered samples. In the case of GaN thin
films, the exposed surface is entirely c-plane as confirmed by
XRD. Obviously, the surface area of the planar samples is
independent of its thickness and mass, but only depends on the
dimension of the slice used (10.0 cm2 in our test). For further
comparison, a slice of a short GaN NW sample with a length of
400 nm and an average diameter of 80 nm was prepared
(designated as GaN NWs-400) by simply reducing the growth
times (Figure S7 and Table S1, entry 4). Figure 2a presents the
Figure 1. (a) SEM, (b) low- and (c) high-resolution TEM images, (d)
diameter distribution, and (e) room temperature PL spectrum of the as-
synthesized GaN NWs. The inset of (a) shows the top view of one
nanowire. The inset of (c) is the selected area electron diffraction
pattern of the lateral facet of one nanowire. The relatively narrow full-
width-at-half-maximum (fwhm) of the PL peak suggests that the NWs
are strain-free. (f) Powder X-ray diffraction patterns of wurtzite GaN
samples in the forms of nanowire, thin film, and powder (from top to
bottom).
and thin films is the c-plane. The area densities of the NWs and
thin films are measured to be 0.1 and 0.4 mg·cm−2, respectively.
The performances of the nondoped GaN materials in different
forms for the photocatalytic methane conversion reaction were
tested at 5 °C under UV irradiation from a 300-W xenon lamp.
To quantitatively evaluate the photocatalytic activities upon
irradiation at a longer UV wavelength (UVA and UVB
components of sunlight), a set of UV filters (Figure S3) was
carefully mounted in the system to completely block wavelengths
shorter than 290 nm and longer than 380 nm from the xenon
lamp. After the reactions, the GaN NWs were found to show the
highest photocatalytic activity for the methane dehydroaroma-
tization reactions among all the GaN samples tested. The
catalytic reaction led to the formation of benzene and H2 with
nearly 1:9 stoichiometry, as confirmed by using gas chromatog-
raphy (GC) and gas chromatography−mass spectrometry (GC-
MS) (Figure S4 and Table S1, entry 1). The selectivity for
benzene over alternative hydrocarbon products (toluene, ethane,
and ethylene) was measured to be 96.5%. Using a series of UV-
cut-out filters (Supporting Figure S5), we determined that the
minimum photon energy required to drive this reaction over
GaN samples corresponds to a wavelength of 360 nm. No
conversion was detected under darkness due to the thermody-
namic limitation of this reaction (the benzene yield can be
estimated to be ca. 0.000007% in the equilibrium at 298 K in our
case). Besides GaN NWs, GaN powders (Table S1, entry 2) also
showed substantial activity but GaN thin films (Table S1, entry
3) showed negligible photocatalytic activity for methane
conversion. This result is interesting since the strong piezo-
electric field in GaN thin films can induce effective photoexcited
carrier separation, which, together with its single crystallinity, led
to enhanced photocatalytic performance for water splitting
compared to the powdered sample in our previous work.23 It
implies that the photodriven methane dehydrogenation reaction
is governed by not only the photogenerated carriers but also a
surface-assisted process.24
Figure 2. Amount of converted (a) methane and (b) ethylene over
various GaN catalysts in different forms and the total surface area of each
catalyst used in the test. Reaction conditions were as follows: 0.35 mg of
GaN NWs, 0.17 mg of GaN NWs-400, 0.35 mg of GaN powders, and 4.0
mg of GaN thin-films upon 290−380 nm UV irradiation at an intensity
of 7.5 mW cm−2 for 12 h, respectively.
total surface area of each catalyst used in our test and their
corresponding photocatalytic activities for the methane con-
version reaction. We can find that the performance of GaN
materials depends strongly on their exposed m-plane but is not
correlated with the c-plane. GaN thin films have negligible m-
plane exposed and hence are unable to convert methane. For
comparison, if we performed similar experiments with pure
ethylene on the surface of GaN, both the c-plane and m-plane
showed remarkable activities toward the ethylene dehydrogen-
ation reaction (Figure 2b and Table S2). These observations
clearly demonstrate that the photodriven methane C−H
activation is surface sensitive. In addition, due to the closely
spaced nanostructure and large surface area, the equilibrium of
CH4 adsorption on the lateral facets of GaN NWs is not very fast.
Thus, an induction period was observed when the UV
illumination started immediately upon the exposure of GaN
NWs to methane (Figure 3a). For comparison, if GaN NWs were
exposed to pure methane in darkness for 1 day prior to the UV
illumination, the GaN NWs would show a constant conversion
rate during the UV irradiation, and no induction period was
observed.
As shown in Figure 3b, the m-plane of ideal wurtzite GaN is
composed of binary Ga and N atoms tetrahedrally coordinated
with each other, whereas the c-plane contains only one type of
atom (Ga or N). Previously, the polarization of alkane C−H
25
bonds by Ga3+ and adjacent O2− in Ga2O3 and Ga-ETS-1017
materials has been proven to be very strong even at rt. Similarly,
in the exposed GaN m-plane, Ga3+ and N3− can interact with the
methyl group and H atom of the adsorbed methane molecule,
respectively. The GaN lattice consisted of alternating positive-
charged Ga and negative-charged N atomic planes, which can
In this regard, we have comprehensively investigated the
structural effect of GaN materials on the photocatalytic
performances. The specific surface area of GaN NWs can be
B
dx.doi.org/10.1021/ja5004119 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 4. (a) Methane conversion rate and benzene selectivity over a
slice of intrinsic GaN NWs, n-type GaN NWs, p-type GaN NWs, and
pure Si substrate under UV irradiation (290−380 nm) for 12 h. The area
of each slice is 3.5 cm2. (b) Plot of the methane consumption and
quantum efficiency (inset) over n-type GaN NWs in 12 h as a function of
the light intensity.
Figure 3. (a) Methane consumption curve upon the initial irradiation
time. (b) Schematic diagram for methane C−H bond polarization on
the surface of GaN m-plane. The insets are high-resolution TEM images
of the top c-plane and lateral m-plane of GaN.
produce strong electrostatic polarization along the c-direction
and, thus, stretch the methane molecules absorbed on the m-
plane. Note that the length of the Ga−N bond in the m-plane is
1.95 Å, which is significantly longer than the methane C−H bond
(1.09 Å) and, hence, highly beneficial to stretching the C−H
bond. The polarized methane C−H bond is significantly
weakened, promoting its cleavage under UV irradiation.
However, on the other hand, the c-plane is made up of only
Ga (or N) atoms. Owing to the high bond energy of 8.92 eV/
atom, the Ga−N bond is extremely hard to break by photons
with an energyof <4.3 eV (290 nm) with insertion of its Ga (or
N) atom into the methane C−H bond through a classical C−H
bond activation mechanism.8 Consequently, the absence of
dipoles in the exposed c-plane leaves the C−H bond of the
methane molecule intact, which cannot be broken directly by
photons.
The incorporation of a trace amount of silicon or magnesium
dopants into GaN NWs can significantly alter the Fermi level
position and carrier transport properties.26 To evaluate the effect
of doping on the activity of GaN NWs, we have synthesized Si-
doped n-type and Mg-doped p-type GaN NWs (Figure S8). The
doping density is controlled by tuning the effusion cell
temperatures of Si (1350 °C) and Mg (265 °C). The electron
and hole concentrations for the Si-doped n-type and Mg-doped
p-type GaN NWs were estimated to be on the order of n = 5 ×
1018 cm−3 and p = 1 × 1018 cm−3, respectively. Figure S9 shows rt
PL spectra of the as-synthesized nondoped, Si-doped, and Mg-
doped GaN NWs. The PL peak position near ∼365 nm for all
three samples is revealed, indicating the band gap of GaN is
unchanged after doping. As shown in Figure 4a, the n-type GaN
NWs (Table S1, entry 5) exhibited significantly enhanced
photocatalytic activity for methane dehydroaromatization
compared with the nondoped ones. On the other hand, the
activity of p-type GaN NWs (Table S1, entry 6) was severely
reduced.
oxide25 or Ga modified zeolites.27 The photogenerated electron
from the GaN semiconductor will migrate to the surface and
reduce the proton into a H atom, which then combines to release
as H2 gas. Meanwhile, the methyl group will be oxidized by the
photogenerated hole to form a methyl radical, followed by the
C−C coupling reaction, forming ethane. The ethylene produced
from the ethane dehydrogenation reaction can further go
through the cyclization and dehydrogenation process to form
benzene. Under the standard conditions described in Table S1,
the overall quantum efficiency (benzene molecules produced per
photon consumed) between 290 and 380 nm over n-type GaN
NWs was calculated to be ca. 0.72%, which is remarkable for a
static gas−solid system with poor diffusion. To pinpoint the
detailed mechanism, we have tested the catalytic performances of
n-type GaN NWs under varying light intensity (Figure 4b). To
eliminate the photothermal effect, all the reactions were kept at 5
°C by using a powerful cooling system. The experimental results
revealed that the methane conversion rate is proportional to the
square of light intensity and the quantum efficiency increases
linearly as a function of light intensity. This observation indicates
that the process of methane coupling, which is the rate-
determining step, requires two photogenerated holes to produce
two independent methyl radicals simultaneously, which then
couple with each other to form an ethane molecule (a detailed
explanation can be found in the SI).28
To assess the performance of GaN under the UV light without
any filters, Ga2O3, an effective photocatalyst reported previously
for the methane coupling reaction was prepared and tested for
comparison.15 Using 0.35 mg of the catalysts and 150 μmol of
methane, we obtained conversion rates for methane of 148 μmol
h−1 g−1 over Ga2O3 powders (Table S1, entry 9) and 63 μmol h−1
g−1 over GaN powders (Table S1, entry 10) upon UV irradiation
without any filters, respectively. In contrast, a conversion rate of
methane of 725 μmol h−1 g−1 was achieved using n-type GaN
NWs under identical conditions (Table S1, entry 11).
The performance of GaN NWs can be used repeatedly without
noticeable deactivation after several cycles (Figure S11) whereas
the crystal structure remained intact (Figure S12). After a long
reaction time, a small amount of carbon deposits (polyaromatic
and graphitic species) was formed on the surface of GaN NWs
which can be completely stripped by the oxygen plasma
treatment, as judged by Raman spectroscopy (Figures S13 and
S14). More importantly, the lattice Ga3+ remained unchanged
during the methane conversion reaction. As shown by X-ray
photoelectron spectroscopy (Figure S15), neither of the binding
energy peaks with maxima at 19.1 (typical for Ga+) and 17.7 eV
An average methane conversion rate of 325 μmol h−1 g−1 was
achieved after 36-h UV irradiation by using 0.35 mg of n-type
GaN NWs (Table S1, entry 7). During the reactions, acetylene
was not detected, whereas only small amounts of ethane and
ethylene were observed. Thus, a plausible mechanism was
proposed (Figure S10). The dehydroaromatization reaction
starts with the polarization of methane on the exposed m-plane.
The initial interaction between adsorbed methane and GaN
δ‑
results in Hδ+−CH3 polarization. Upon UV irradiation, the
−
polarized methane was first heterolytically split into CH3 and
H+ on the catalyst surface. A similar heterolytic dissociation of
alkane C−H bonds is also evidenced on the surface of gallium
C
dx.doi.org/10.1021/ja5004119 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
(typical for Ga0) was detected after a long reaction time.29
Therefore, GaN materials are proven to be highly stable for the
methane conversion reaction.
In summary, we have discovered a promising approach to turn
shale gas into aromatic compounds powered by photoenergy for
the first time. The GaN semiconductor exhibits superior
chemical stability and photocatalytic activity for C−H activation
of methane at rt.
(16) Li, L.; Li, G. D.; Yan, C.; Mu, X. Y.; Pan, X. L.; Zou, X. X.; Wang, K.
X.; Chen, J. S. Angew. Chem., Int. Ed. 2011, 50, 8299.
(17) Li, L.; Cai, Y. Y.; Li, G. D.; Mu, X. Y.; Wang, K. X.; Chen, J. S.
Angew. Chem., Int. Ed. 2012, 51, 4702.
(18) Han, W. Q.; Fan, S. S.; Li, Q. Q.; Hu, Y. D. Science 1997, 277,
1287.
(19) Schafer, S.; Wyrzgol, S. A.; Caterino, R.; Jentys, A.; Schoell, J. S.;
̈
Havecker, M.; Gericke, A. K.; Lercher, J. A.; Sharp, I. D.; Stutzmann, M.
̈
J. Am. Chem. Soc. 2012, 134, 12528.
(20) Garnett, E.; Yang, P. D. Nano Lett. 2010, 10, 1082.
(21) Nguyen, H. P. T.; Zhang, S.; Cui, K.; Han, X.; Fathololoumi, S.;
Couillard, M.; Botton, G. A.; Mi, Z. Nano Lett. 2011, 11, 1919.
(22) Chuaha, L. S.; Hassana, Z.; Hassana, H. A.; Ahmedb, N. M. J.
Alloys Compd. 2009, 481, 15.
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures and analytical data. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
S
(23) Wang, D.; Pierre, A.; Kibria, M. G.; Cui, K.; Han, X.; Bevan, K. H.;
Guo, H.; Paradis, S.; Hakima, A. R.; Mi, Z. Nano Lett. 2011, 11, 2353.
AUTHOR INFORMATION
(24) Zhong, D.; Franke, J. H.; Podiyanachari, S. K.; Blomker, T.;
̈
■
Zhang, H.; Kehr, G.; Erker, G.; Fuchs, H.; Chi, L. Science 2011, 334, 213.
Corresponding Authors
zetian.mi@mcgill.ca
cj.li@mcgill.ca
(25) Kazansky, V. B.; Subbotina, I. R.; Pronin, A. A.; Schlogl, R.;
̈
Jentoft, F. C. J. Phys. Chem. B 2006, 110, 7975.
(26) Zhang, Z.; Yates, J. T., Jr. Chem. Rev. 2012, 112, 5520.
(27) Luzgin, M. V.; Gabrienko, A. A.; Rogov, V. A.; Toktarev, A. V.;
Parmon, V. N.; Stepanov, A. G. J. Phys. Chem. C 2010, 114, 21555.
(28) Kim, J. D.; Yee, N.; Nanda, V.; Falkowski, P. G. Proc. Natl. Acad.
Sci. U.S.A. 2013, 110, 10073.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
(29) Serykh, A. I.; Amiridis, M. D. Surf. Sci. 2009, 603, 2037.
This work was financially supported by the Canada Research
Chair (Tier 1) foundation, NSERC, FQRNT, Canada
Foundation for Innovation (CFI), and McGill University.
REFERENCES
(1) Richard, A. K. Science 2010, 328, 1624.
(2) Choudhary, V. R.; Kinage, A. K.; Choudhary, T. V. Science 1997,
275, 1286.
■
(3) Lunsford, J. H. Catal. Today 2000, 63, 165.
(4) Periana, R. A.; Mironov, O.; Taube, D.; Bhalla, G.; Jones, C. J.
Science 2003, 301, 814.
(5) Holmen, A. Catal. Today 2009, 142, 2.
(6) Balasubramanian, R.; Smith, S. M.; Rawat, S.; Yatsunyk, L. A.;
Stemmler, T. L.; Rosenzweig, A. C. Nature 2010, 465, 115.
(7) Schwarz, H. Angew. Chem., Int. Ed. 2011, 50, 10115.
(8) (a) Bergman, R. G. Nature 2007, 446, 391. (b) Sun, C.-L.; Li, B.-J.;
Shi, Z.-J. Chem. Rev. 2011, 111, 1293. (c) Yeung, C. S.; Dong, V. M.
Chem. Rev. 2011, 111, 1215. (d) Liu, C.; Zhang, H.; Shi, W.; Lei, A.
Chem. Rev. 2011, 111, 1780. (e) Mkhalid, I. A.; Barnard, J. H.; Marder, T.
B.; Murphy, J. M.; Hartwig, J. F. Chem. Rev. 2010, 110, 890. (f) Lyons, T.
W.; Sanford, M. S. Chem. Rev. 2010, 110, 1147. (g) Young, A. J.; White,
M. C. J. Am. Chem. Soc. 2008, 130, 14090. (h) Wang, D.-H.; Engle, K.-
M.; Shi, B.-F.; Yu, J.-Q. Science 2010, 327, 315. (i) Labinger, J. A.;
Bercaw, J. E. Nature 2002, 417, 507. (j) Crabtree, R. H. J. Chem. Soc.,
Dalton Trans. 2001, 17, 2437. (k) Wencel-Delord, J.; Glorius, F. Nat.
Chem. 2013, 5, 369. (l) Jones, W. D. Top. Organomet. Chem. 1999, 3, 9.
(9) Ahuja, R.; Punji, B.; Findlater, M.; Supplee, C.; Schinski, W.;
Brookhart, M.; Goldman, A. S. Nat. Chem. 2010, 3, 167.
(10) (a) Wang, L.; Tao, L.; Xie, M.; Xu, G. Catal. Lett. 1993, 21, 35.
(b) Wang, D. J.; Lunsford, J. H.; Rosynek, M. P. J. Catal. 1997, 169, 347.
(c) Borry, R. W., III; Kim, Y. H.; Huffsmith, A.; Reimer, J. A.; Iglesia, E. J.
Phys. Chem. B 1999, 103, 5787. (d) Ma, D.; Shu, Y. Y.; Cheng, M. J.; Xu,
Y. D.; Bao, X. H. J. Catal. 2000, 194, 105. (e) Zheng, H.; Ma, D.; Bao, X.
H.; Hu, J. Z.; Kwak, J. H.; Wang, Y.; Peden, C. H. F. J. Am. Chem. Soc.
2008, 130, 3722.
(11) Luzgin, M. V.; Rogov, V. A.; Arzumanov, S. S.; Toktarev, A. V.;
Stepanov, A. G.; Parmon, V. N. Angew. Chem., Int. Ed. 2008, 47, 4559.
(12) Wang, L.; Ohnishi, R.; Ichikawa, M. J. Catal. 2000, 190, 276.
(13) Yuliati, L.; Yoshida, H. Chem. Soc. Rev. 2008, 37, 1592.
(14) Yoshida, H.; Matsushita, N.; Kato, Y.; Hattori, T. J. Phys. Chem. B
2003, 107, 8355.
(15) Yuliati, L.; Hattori, T.; Itoh, H.; Yoshida, H. J. Catal. 2008, 257,
396.
D
dx.doi.org/10.1021/ja5004119 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX