Highly efficient light-driven methane coupling under ambient conditions based on an integrated design of a photocatalytic system

Article abstract of DOI:10.1039/d0gc01608j

Direct non-oxidative coupling of methane (NOCM) is an effective way to produce hydrocarbons. However, this process usually requires a high temperature (≥1100 °C) to break the C-H bond of CH4 and suffers catalyst deactivation due to coke formation. Photocatalytic NOCM is an ideal strategy to solve these issues. Herein, we designed a novel photocatalytic methane coupling system consisting of a continuous flow reactor and metal-loaded TiO2 photocatalysts with light-diffuse-reflection-surfaces. It was found that Au/TiO2 was the best catalyst for the system due to the easy transport of photoelectrons from TiO2 to Au particles to inhibit the photoelectron-hole recombination. The yield of C2H6 reached 81.7 μmol gcatalyst-1 h-1 with higher than 95percent selectivity over Au/TiO2 under simulated 1.5G sunlight irradiation and ambient conditions (room temperature and 1 atm), which is 174percent larger than the highest reported value. Furthermore, DFT calculation results revealed that the methyl anion is a possible intermediate species for the formation of ethane.

Full text of DOI:10.1039/d0gc01608j

Green Chemistry
View Article Online
View Journal
PAPER
Highly efficient light-driven methane coupling
under ambient conditions based on an integrated
design of a photocatalytic system†
a,b
Cite this: DOI: 10.1039/d0gc01608j
Junyu Lang,a Yuli Ma,a Xuechen Wu,a Yueyue Jianga and Yun Hang Hu
*
Direct non-oxidative coupling of methane (NOCM) is an effective way to produce hydrocarbons.
However, this process usually requires a high temperature (≥1100 °C) to break the C–H bond of CH4 and
suffers catalyst deactivation due to coke formation. Photocatalytic NOCM is an ideal strategy to solve
these issues. Herein, we designed a novel photocatalytic methane coupling system consisting of a con-
tinuous flow reactor and metal-loaded TiO2 photocatalysts with light-diffuse-reflection-surfaces. It was
found that Au/TiO2 was the best catalyst for the system due to the easy transport of photoelectrons from
TiO2 to Au particles to inhibit the photoelectron–hole recombination. The yield of C2H6 reached
−1
81.7 μmol gcatalyst h−1 with higher than 95% selectivity over Au/TiO2 under simulated 1.5G sunlight
Received 10th May 2020,
Accepted 22nd June 2020
irradiation and ambient conditions (room temperature and 1 atm), which is 174% larger than the highest
reported value. Furthermore, DFT calculation results revealed that the methyl anion is a possible inter-
mediate species for the formation of ethane.
DOI: 10.1039/d0gc01608j
rsc.li/greenchem
(NOCM) is considered an ideal process to avoid CO2 for-
mation (eqn (1)).11
Introduction
Methane, a non-conventional energy resource and a main
component of natural gas and combustible ice (methane
hydrate),1 has been extensively used for the production of
energy and chemicals. Conversion of methane into high-
value chemicals has been intensively explored, including
steam (or CO2) reforming of methane,2,3 partial oxidation of
methane,4 oxidative coupling of methane (OCM),5 and non-
oxidative coupling of methane (NOCM).6 Direct conversion
of methane to ethylene and ethane has attracted much atten-
tion since the first report of OCM in 1982.7 The OCM reac-
tion requires a high temperature to break the strong C–H
bond (434 kJ mol−1) in a CH4 molecule. However, such a
high temperature causes further oxidation of CH4 and C2
hydrocarbon products to CO2, leading to a poor yield of the
desired products. For this reason, enormous efforts have
been made to achieve OCM at a lower temperature.8–10
Furthermore, direct non-oxidative coupling of methane
2CH4 ! C2H6 þ H2; ΔHð298 KÞ ¼ 65 kJ molꢀ1
ð1Þ
However, the endothermic NOCM reaction requires a high
energy input to activate CH4 via breaking a C–H bond. This
has stimulated intensive research on NOCM, such as improv-
ing the catalytic efficiency and developing catalytic systems.12
For example, Lobo and co-workers reported a high selectivity
of ethylene produced by NOCM over a Mo2C/[B]ZSM-5 cata-
lyst.13 They replaced some aluminium atoms by boron atoms
in the ZSM-5 framework to decrease the zeolite acidity, leading
to the change of main products from aromatics to ethylene.
Varma and co-workers synthesized Pt–Bi bimetallic catalysts
on ZSM-5 zeolite to produce the C2 species from CH4 at
600–700 °C.14 However, the coke formation caused the fast de-
activation of the catalyst.
In comparison with thermal catalytic processes, photo-
catalytic NOCM is more sustainable and economically attrac-
tive as fossil fuel prices and greenhouse gas emissions con-
tinue to exacerbate.15 Furthermore, photocatalytic NOCM is
expected to inhibit carbon deposition due to its low operating
temperature. This promoted effort to develop efficient photo-
catalysts for the process. Yoshida and co-workers constructed
silica-supported Ga and Ce oxides as catalysts to couple
methane under light irradiation at room temperature.16,17
Chen and co-workers developed a Ga3+-modified titanosilicate
aSchool of Environmental Science and Engineering, Shanghai Jiao Tong University,
Shanghai 200240, P. R. China
bDepartment of Materials Science and Engineering, Michigan Technological
University, Houghton, Michigan 49931, USA. E-mail: yunhangh@mtu.edu
†Electronic supplementary information (ESI) available: Detailed description of
the experimental methods and additional data and figures. See DOI: 10.1039/
d0gc01608j
This journal is © The Royal Society of Chemistry 2020
Green Chem.

View Article Online
Paper
Green Chemistry
zeolite material (Ga-ETS-10) and Zn+-modified zeolite (Zn-
ZSM-5) photocatalysts to achieve UV-light-driven methane
coupling with 29.8 μmol g−1 h−1 of ethane yield.18,19 Recently,
Long et al. have reported photocatalytic methane coupling over
Au/ZnO, in which CH4 was activated by the synergy of the
plasma resonance of Au nanoparticles and the surface polariz-
ation of ZnO under sunlight irradiation, leading to 11.1 μmol
g−1 h−1 of ethane yield.20 However, so far, the reported largest
yield of ethane from photocatalytic NOCM is still less than
30 μmol g−1 h−1 (Table S1†).
A TiO2 semiconductor is commonly used as a photocatalyst,
but its performance was restricted by the response only to UV
light due to its wide band gap (3.0–3.2 eV) and low efficiency
of light absorbance. Doping metals in the lattice of TiO2 has
been demonstrated as an efficient approach to achieve a
smaller band gap and to increase the quantum efficiency of
TiO2.21,22 Furthermore, the doped metals may induce the posi-
tive Schottky barrier effect that can enhance the separation of
charge carriers. Consequently, Pt, Pd, Au, Rh, and Ir are often
deposited on TiO2 as efficient co-catalysts by a photodeposi-
tion, impregnation, or deposition–precipitation method.23 It
has been found that Au/TiO2 is a promising photocatalyst due
to its attractive interfacial properties.24,25 Furthermore, the
loading of Ag or Au nanoparticles could create surface plasmo-
nic features.26,27
The configuration of reactors plays an important role in
photocatalytic NOCM.28,29 Batch and slurry reactors have
mainly been used in photocatalytic processes.30,31 Batch reac-
tors are often employed for the evaluation of the NOCM activity,
but they suffered low efficiency, difficult separation of products,
and tough industrialization. Furthermore, the application of
slurry reactors was limited by difficult mass transfer among
solid, liquid, and gas. The solvent (water) may participate in the
reaction as an oxidizing agent, resulting in a change from
NOCM to OCM. These drawbacks are expected to be overcome
by using a continuous flow reactor system, in which the pro-
ducts can quickly move away from the catalyst surface and thus
inhibit undesirable reverse reactions and secondary reactions.
Therefore, a continuous flow reactor has been applied to
improve the efficiency and product selectivity in organic photo-
chemistry,32 photocatalytic oxidation of methane,33 and photo-
catalytic treatment of industrial waste water.34 However, since
the flow reactor has a short contact time between feed gas and
catalyst, its application for photocatalytic NOCM requires a
highly active photocatalyst. On the other hand, maximizing the
efficiency of light absorption of a catalyst is important to
increase its photocatalytic performance.
Fig. 1 Schematic of photocatalytic methane coupling system. 5 mg of
catalyst was deposited onto a 1.5 × 1 cm2 light diffuse reflecting glass
holder, which was located at the centre of the reactor.
Experimental section
M–TiO2 was prepared through a photodeposition method.
0.4 g of TiO2 (P25) was dispersed in 20 mL of methanol
aqueous solution (vwater : vmethanol = 4 : 1) under ultrasound.
0.02 mL of HAuCl4·4H2O was added to the suspension, fol-
lowed by 30 min of stirring. The content of Au was 1 wt% of
P25. This mixture was irradiated under UV light for 30 min,
followed by washing with distilled water and drying at 60 °C
for 6 h. As a result, a purple powder was obtained for methane
coupling. The same procedure was applied for other M–TiO2
(M = Ru, Rh, Pd, Ag, Ir and Pt) using RuCl3, RhCl3·3H2O,
H2PdCl4, AgNO3, IrCl3·3H2O, and H2PtCl6·6H2O, respectively.
A schematic of the experimental system for low-temperature
photocatalytic methane coupling is shown in Fig. 1. The flow
rate of reaction gas (10% CH4 in argon) was controlled with a
mass flow controller. A homemade reactor consists of two
parts: an electric heating furnace was used to maintain/alter
the reaction temperature and a quartz inner pipe to offer a her-
metic inert reaction environment and a pass way of light. 5 mg
of catalyst was deposited onto a 1.5 × 1 cm2 diffuse reflecting
holder which is made of glass and located at the centre of the
reactor. The photograph of the assembled unit is shown in
Fig. S1.† The produced hydrocarbons were analysed using a
GC equipped with a 30 m plot-q capillary column and an FID
detector. Hydrogen was analysed using a TCD detector with a
TDX-01 packed column. A 300 W Xe lamp with an AM 1.5G
filter was used as the light source, whose intensity was
100 mW cm−2 (one sun). The selectivity is calculated according
to the following equation.
In this work, to achieve highly efficient NOCM, a novel
photocatalytic system is demonstrated by dispersing a metal
loaded TiO2 catalyst on the light-diffuse-reflection surface of
frosted glass with a specifically designed continuous flow
n
ntotal
n
C2H6
C2H6
C2H6 selectivity ¼
ꢁ 100% ¼
ꢁ 100%
3
ð2Þ
nC þ nC
2
reactor (Fig. 1). A record high yield of ethane (C2H6) with high where nC H and ntotal are the molar amounts of C2H6 and all
2
6
selectivity was obtained over the Au/TiO2 catalyst at ambient carbonic products, respectively. The actual loading content of
temperature under a weak light irradiation (100 mW cm−2). metal was determined by ICP-OES (Thermo Scientific iCAP Q).
Furthermore, the methyl anion is a possible intermediate X-ray diffraction data were collected with a SHIMADZU Lab
species to yield a methyl radical for ethane formation.
XRD-6100 diffractometer using Cu Kα monochromatic radi-
Green Chem.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Green Chemistry
Paper
ation with λ = 1.54 angstrom. A TALOS F200X G2 transmission hybrid B3LYP method41 was used for both geometry optimiz-
electron microscope (TEM) was used to evaluate the mor- ation and frequency/intensity calculations. The D95 basis set
phology of catalysts at an acceleration voltage of 200 kV (field combined with Stuttgart/Dresden ECPs42 for Au and the 6-
emission gun). The elemental mapping and overall structure 311G(3df,3dp) basis set for C and H were used.
of catalysts were examined using scanning electron microscopy
(SEM) in conjunction with energy-dispersive X-ray spectroscopy
(SEM–EDX, FEI Quanta 200F). N2 adsorption was exploited to
determine surface areas with Micro/meritics ASAP 2020 PLUS
Results and discussion
Various metals (M = Ag, Au, Rh, Ru, Pd, Ir, and Pt) were loaded
onto TiO2 as M–TiO2 photocatalysts. The actual contents of the
loaded metals were analysed by ICP-OES and are listed in
Table S2.† The M–TiO2 catalyst (dispersed on a light-diffuse-
reflection surface of 1 × 1.5 cm2 ground glass) was located at
the centre of a continuous flow reactor (Fig. 1) and evaluated
for NOCM under ambient conditions (room temperature and 1
atm) and AM 1.5G irradiation. No hydrocarbon product was
detected for the NOCM process without light irradiation.
Furthermore, when CH4 flow was replaced by argon for the
system, no hydrocarbon was produced even under light
irradiation, eliminating the possibility of the remnant carbon
impurity on the catalyst surface to form hydrocarbons under
light irradiation. In contrast, as shown in Fig. 2a, all M/TiO2
catalysts are active for NOCM under light irradiation. Rh/TiO2
showed the least activity for photocatalytic NOCM, which is
comparable to bare TiO2 (P25). Photocatalytic NOCM over Ru/
TiO2, Pd/TiO2, Ir/TiO2, and Pt/TiO2 produced C2H6 with yields
HD88. Diffuse reflectance spectra (DRS) were obtained using a
Shimadzu UV-2600 UV–vis spectrophotometer (equipped with
a BaSO4 coated integrating sphere). The X-ray photoelectron
spectroscopy (XPS) spectra were recorded on a Kratos Axis
Ultra DLD spectrometer equipped with an Al Kα (1486.6 eV)
radiation. The C 1s peak of hydrocarbon contamination (at a
binding energy (BE) of 284.6 eV) was used as an energy refer-
ence. The in situ DRIFT spectra of NOCM over Au/TiO2 under
light irradiation were recorded using a Thermo Nicolet 6700
FTIR spectrometer equipped with a mercury–cadmium-tellur-
ide (MCT) detector and a Harrick Praying Mantis low tempera-
ture reaction chamber. The IR spectra of the catalyst under
dark conditions were obtained as the background spectrum.
The Vienna ab initio simulation package35 was used for DFT
calculations to evaluate the photocatalytic NOCM mechanism.
The 2s, 2p electrons in carbon and oxygen, the 3d, 4s electrons
in titanium, and the 5s, 4d, 5p electrons in gold were treated
as valence electrons, whereas the kinetic energy cut-off for the
plane wave basis sets was set to 450 eV. The remaining core
electrons were described by the projector augmented-wave
method. The anatase structure was chosen as the model to
present TiO2 according to the XRD analysis and previous
research with optimized lattice parameters of 3.819 and 9.702
angstrom.36 We used a five-layer model with top relaxed layer
for simulating the anatase (1 0 0) surface. To avoid direct inter-
action between the gold particles, we increased the area of the
anatase (1 0 0) surface model by constructing supercells out of
the primitive cell. The surface supercell was (3 × 2) in the cases
of CH4 adsorption, CH4 activation and C2H6 formation, and (6
× 2) for CH3 ion migration with a vacuum layer of 20 ang-
stroms. An Aun (n = 1–10) cluster was used as the model of
gold to build the interface model of Au/TiO2 with full geometry
optimization.37 The average adsorption energy of Aun/TiO2 (n =
1–10) was used to evaluate the Aun cluster size. The climbing
image nudged elastic band (CI-NEB) method was used here to
determine the reaction transition state, and the minimum
energy path of CH3 ion migration. The Monkhorst–Pack
meshes of 2 × 2 × 1 and 5 × 5 × 5 k-point sampling in the
surface Brillouin zone were employed for the slab model and
bulk models of anatase, respectively.38 After the convergence
criteria for optimizations were met, the largest remaining force
on each atom was less than 0.02 eV Å−1. For all calculations,
the spin polarized generalized gradient approximation of the
Perdew–Burke–Ernzerhof functional39 was used. As the stan-
dard DFT functions tended to over-delocalize electrons, DFT+U
was employed with an effective U value of 3.3 eV for Ti 3d-orbi-
tals, as obtained from the linear response. The GAUSSIAN 03
(per gram of metal/TiO2 catalyst) of 18.3, 14.4, 7.5 and
−1
4.1 μmol gcatalyst
h−1, respectively. Furthermore, Ag/TiO2
exhibited excellent performance with a C2H6 yield of 37.9 μmol
−1
gcatalyst
h
−1, which is double that of Ru/TiO2. More impor-
tantly, the C2H6 yield of Au/TiO2 is even double that of Ag/
−1
TiO2, reaching 81.7 μmol gcatalyst
h
−1. This indicates the
remarkable effect of metals on M/TiO2 catalysts for photo-
Fig. 2 (a) Photocatalytic NOCM performance (C2H6 yield per gram of
M/TiO2 catalyst, M = Au, Ag, Ir, Pt, Pd, Ru, and Rh) under AM1.5G sun-
light irradiation, (b) photocatalytic NOCM performance (C2H6 yield per
gram of M) of M/TiO2 catalysts under AM1.5G sunlight irradiation, (c)
C2H6 yield (per gram of catalyst) over Au/TiO2, and (d) NOCM perform-
program40 was used to obtain the theoretical IR spectra. The ance vs. Au content over Au/TiO2.
This journal is © The Royal Society of Chemistry 2020
Green Chem.

View Article Online
Paper
Green Chemistry
catalytic NOCM with an efficiency sequence of Au > Ag > Ru > from TiO2 to Au. However, the photocatalytic activity slightly
Pd > Ir > Pt > Rh. Furthermore, Fig. 2b shows that the specific decreased for 0.78% Au/TiO2, since an excess of Au may
yields per gram of loaded metal follow the same order: Au > Ag obstruct some light from reaching the surface of TiO2.
> Ru > Pd > Ir > Pt > Rh (namely, 16.3 > 5.2 > 4.3 > 4.2 > 1.9 >
Besides the superior performance of the Au/TiO2 catalyst,
0.8 mmol gmetal−1 h−1). The different performances of different the excellent efficiency of photocatalytic NOCM can be attribu-
metals can be explained by CH4 activation, which includes ted to the novel design of the photocatalytic reaction system:
three steps: (1) CH4 adsorption on the metal surface, (2) the (1) The benefit of dispersing Au/TiO2 on the light-diffuse-
transport of photoelectrons to the metal surface, and (3) the reflection-surface of frosted glass was confirmed by its high
reduction of adsorbed CH4 by photoelectrons on the metal ethane yield, which is about 3.4 times that over Au/TiO2 spread
surface. It was reported that CH4 is much easier to be on the common smooth surface of quartz (Fig. S5†). This is
adsorbed on Pt and Pd than on Au and Ag,43 indicating that because the light-diffuse-reflection-surface of a substrate could
CH4 adsorption is not a rate-determining step in photo- greatly increase light absorption of photocatalyst particles.28
catalytic NOCM. It is well-known that the transport of photo- (2) Different from the commonly used batch and slurry reac-
electrons from TiO2 to metal is controlled by the contact resis- tors for photocatalytic NOCM, our continuous flow reactor
tance between TiO2 and metal particles, which is proportional system can allow the desired product to quickly move out of
to the potential drop (ΔVH) across the M–TiO2 interface the reaction region, which reduces the reverse reaction to
region,44 namely, the smaller the potential drop is, the easier increase the reaction rate and avoid side reactions of the
the transport of photoelectrons from TiO2 to metal. From product to increase the selectivity. Indeed, as shown in
Table S3,† one can see that Au and Ag have the lowest and Fig. S6,† the ethane yield from the continuous flow reactor
second lowest ΔVH, which are consistent with the best and the system is 33 and 16 times larger than those from batch and
second best NOCM performances of Au/TiO2 and Ag/TiO2, slurry reactors, respectively. For the continuous flow reactor
respectively. This indicates that photocatalytic NOCM is deter- system, there is an optimum flow rate of feed gas for the
mined by the contact resistance of the M/TiO2 interface. The photocatalytic NOCM, namely, as the flow rate increased, the
lowest contact resistance of the Au–TiO2 interface created the reaction rate increased to a maximum value of 100 μmol
−1
easiest transport of photoelectrons to Au from TiO2 to greatly gcatalyst
h−1 at GHSV (gas hourly space velocity) of
decrease the photoelectron–hole recombination, which was 180 000 mL gc h−1 and then decreased to 93 μmol gcatalyst
−1
−1
a
photoluminescence emission spectrum h−1 at GHSV of 240 000 mL gc h−1 (Fig. S7b†). Because a
−1
confirmed by
(Fig. S2†), leading to the large enhancement in photocatalytic much high flow used for the continuous flow reactor could
NOCM over the Au/TiO2 catalyst. cause insufficient contact time of a reactant with the catalyst
Among all tested catalysts, Au/TiO2 is the best for photo- surface, leading to a low reaction rate. In contrast, the selecti-
catalytic NOCM with an impressive ethane yield (81.7 μmol vity of C2H6 was always improved by increasing the flow rate,
−1
gcatalyst
h
−1), which is 174% larger than the highest reported because side reactions and the reverse reaction were
C2H6 yield without the aid of H2O or any other oxidants suppressed.
(Table S1†). C3H8 was also produced over Au/TiO2, but its yield The Au/TiO2 catalyst was characterized. X-Ray diffraction
−1) was very small, leading to high (XRD) showed both the anatase and rutile phases of TiO2 in
−1
(3.05 μmol gcatalyst
h
selectivity (95%) of ethane. The total conversion of CH4 the catalyst (Fig. S8†). Observations using transmission elec-
reached about 84.8 μmol gcatalyst−1 h−1. Furthermore, the excel- tron microscopy (TEM) and scanning electron microscopy
lent photocatalytic stability of Au/TiO2 for NOCM is shown in (SEM) revealed that the average size of TiO2 nanoparticles is
Fig. 2c, namely, the reaction rate quickly increased to about about 25 nm (Fig. S9†). Furthermore, one can see the interpla-
h−1 in the first 3 hours under light nar lattice spacing of TiO2 from HRTEM image, which corres-
−1
80 μmol gcatalyst
irradiation and then slightly decreased to a constant value at ponds well to the (1 1 0) plane of anatase TiO2 (d = 0.32 nm).
the 38th hour. There is about 21.5% performance degradation Although XRD did not show any diffraction peak of Au due to
during 70 hours of continuous photocatalytic NOCM reaction. its too small content (Fig. S8†), nanostructured Au was clearly
Simultaneously, the equimolar amount of H2 to ethane was observed by HRTEM (Fig. 3a and b), namely, spherical-shaped
produced over Au–TiO2, the ratio of n(H2)/n(C2H6) was trending Au nanoparticles (NPs) with an average particle size of about
to 1 : 1 (Fig. 2c and Fig. S3†). The gas chromatographs for the 20 nm and interplanar lattice spacing of 0.20 nm, which is
products are also presented in Fig. S4.† In comparison with associated with the Au (2 0 0) plane. The dispersion of Au
−1
bare TiO2 (only 1.5 μmol gcatalyst
h−1 ethane yield), the nanoparticles on TiO2 was further supported by elemental
NOCM activity was increased almost 60 times by loading Au mapping (Fig. 3c). The interaction between Au and TiO2 was
on TiO2. This clearly indicates the critical role of Au in the further examined by X-ray photoelectron spectroscopy (XPS).
catalyst for photocatalytic NOCM. Furthermore, Fig. 2d shows The peaks located at 85 eV were deconvoluted into two at 83.09
the significant impact of actual Au content on NOCM activity eV and 86.70 eV, which can be attributed to 4f7/2 and 4f5/2 of
of Au/TiO2, namely, as the Au content increased, the reaction the Au0 species, respectively (Fig. S10†). This indicates that the
rate increased and reached the maximum at 0.50% Au. This gold ions were reduced to zero valence during the photodepo-
happened because the Au–TiO2 interface increased with sition process. The binding energies of Ti 2p of Au/TiO2 were
increasing Au content, increasing the photoelectron transport exactly the same as that of the pure TiO2 (Fig. 3d), indicating
Green Chem.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Green Chemistry
Paper
band (CB) and then transferred to the Au NP, since the work
function of Au is below the CB of TiO2. Consequently, elec-
trons and holes accumulated on the Au nanoparticles and the
surface of TiO2, respectively (Fig. 4b). There are four possible
pathways for dissociative activation of CH4 over Au/TiO2 under
light irradiation (reactions (a)–(d)):
CH4 þ eꢀ ! CH3ꢀ þ H
ðaÞ
ðbÞ
ðcÞ
ðdÞ
•
CH4 þ eꢀ ! CH3 þ Hꢀ
CH4 þ hþ ! CH3þ þ H
•
CH4 þ hþ ! CH3 þ Hþ
−
CH4 may be reduced into a CH3 anion and H atom (reac-
tion (a)) or •CH3 radical and H− anion (reaction (b)) by an elec-
tron accumulated on the Au nanoparticles of Au/TiO2 or a
photogenerated electron on the surface of TiO2 of Au/TiO2. In
+
contrast, CH4 may be oxidized to a CH3 cation and H atom
(reaction (c)) or a •CH3 radical and H+ cation (reaction (d)) only
by a photogenerated hole on the TiO2 surface of Au/TiO2. The
Bader charge analysis based on our density functional theory
−
(DFT) calculations revealed that the CH3 species exist as CH3
+
−
on Au or CH3 /CH3 on TiO2 (Table S4†). Therefore, reaction
(a) instead of reaction (b) would occur at the Au site or TiO2
site of Au/TiO2. Furthermore, as shown in Fig. S20,† the acti-
vation barrier of CH4 into CH3− at the Au site is 0.47 eV, which
is smaller than that (1.1 eV) at the TiO2 site, indicating that
the reduction (reaction (a)) preferentially occurs at the Au sites
of Au/TiO2 rather than at the TiO2 sites of Au/TiO2. Therefore,
Fig. 3 (a) HRTEM image and (b) enlarged image of the green region of
Au/TiO2, (c) EDS mapping of Au/TiO2, (d) XPS Ti 2p spectra of P25 and
A/TiO2, and (e) UV-vis DRS spectra of commercial P25 and Au/TiO2.
that the photodeposition of Au had an ignorable effect on the
valence state of TiO2. This also supports the photostability of
Au/TiO2. The UV–vis absorption spectrum demonstrated that
TiO2 (without Au) showed strong UV absorption (below 400 nm
corresponding to the bandgap energy of 3.35 eV), whereas the
introduction of Au NPs led to an extra visible light absorption
peak at 550 nm (Fig. 3e). The absorption peak at 550 nm can
be attributed to the surface plasma resonance (SPR) effect of
Au, which was well demonstrated.24,25 However, no C2H6 was
produced over the Au/TiO2 catalyst under visible light
irradiation, whereas Au/TiO2 exhibited a large UV-light activity
for methane coupling (Fig. 4a). This indicates that the reaction
cannot be triggered by the Au SPR effect.
−
the dissociative activation of CH4 may generate a CH3 anion
+
on Au (reaction (a)) or a CH3 cation on TiO2 (reaction (c)). We
further exploited DFT calculations to examine whether the con-
version of CH4 into C2H6 follows a mechanism via a methyl
radical intermediate species formed from the CH3 anion or
cation as follows:
adsorbed
2CH4 ꢀꢀꢀꢀꢀꢀ! 2ꢂCH4
(
hþ=eꢀ
eꢀ=hþ
þ
2CH3ꢀ=CH3 ꢀꢀꢀ! 2•CH3 ! C2H6
ꢀꢀꢀ!
2H• ! H2
According to the photoexcitation principle, the excited TiO2
due to light irradiation yielded electrons in the conduction
Fig. 5a shows the two possible reaction mechanisms
(methyl anion-radical path and methyl cation-radical path).
For methyl anion-radical path, the reaction starts by CH4
adsorption on the Au surface with an exothermic energy of
0.31 eV. The barrier for the dissociation of adsorbed CH4 into
−
a CH3 anion is only 0.47 eV, indicating its kinetic feasibility.
−
Then, the CH3 anion migrates to the TiO2 surface and reacts
•
with a photogenerated hole to form a CH3 radical, whose bar-
riers are 1.34 eV and 0.61 eV, respectively. The vital role of
holes was confirmed by hole scavenger experiments
(Fig. S11†). Finally, two •CH3 radicals combine to form an
ethane molecule with an energy barrier of 0.75 eV, followed by
ethane desorption with an endothermic energy of 0.30 eV.
Compared with the methyl anion-radical process, the for-
mation of the CH3 cation on the surface of TiO2 suffers from a
very high barrier of 1.57 eV, which is much larger than that
Fig. 4 (a) The photoexcitation process analysis of Au/TiO2. (a) C2H6
yield over Au/TiO2 under AM 1.5G sunlight and visible light irradiation (λ
> 420 nm). The inset shows the absorption spectra of an AM1.5G filter
and a UV cut filter (λ > 420 nm). (b) Schematic diagrams of photo-
excitation and reactions over Au/TiO2.
This journal is © The Royal Society of Chemistry 2020
Green Chem.

View Article Online
Paper
Green Chemistry
Fig. 6 In situ DRIFT spectra over Au/TiO2 and P25 (gas: 10% CH4 in Ar;
−1
GHSV: 60 000 mL gc h−1; light source: simulated sunlight) and simu-
−
lated IR spectrum of the TiO2-Au-CH3 model (the details of the TiO2-
Au-CH3− model is presented in Fig. S19†).
was detected by the IR peak at around 600 cm−1. Furthermore,
there is another IR peak at 700–800 cm−1, which can be attrib-
uted to CH3 rock in Au-CH3.45,46 The simulated IR response is
very similar to the experimental IR signal in the range of
600–800 cm−1 that confirms the existence of methyl anion
(Fig. 6).
Conclusions
In summary, a novel continuous-flow photocatalytic system
with an efficient Au/TiO2 catalyst dispersed on a light-diffuse-
reflection-surface was designed for NOCM, achieving a high
−1
C2H6 yield of 81.7 μmol gcatalyst h−1 under weak light (only 1
Sun) irradiation and ambient conditions. This is 174% larger
than the highest reported yield for photocatalytic NOCM
without the aid of H2O or any other compounds. Among all
tested catalysts, Au/TiO2 exhibited the best performance due to
the easy transport of photoelectrons from TiO2 to Au to inhibit
Fig. 5 Mechanism of photocatalytic NOCM. (a) DFT-calculated reaction
energy diagram for the NOCM reaction. (b) Illustration of the NOCM
reaction over Au/TiO2 in a CH3− anion process.
(0.47 eV) of the CH3 anion formation. Therefore, the photo- the photoelectron–hole recombination. The dispersion of the
catalytic NOCM reaction over Au/TiO2 would follow the methyl catalyst particles on the light-diffuse-reflection surface can
anion-radical mechanism (Fig. 5b):
greatly increase their light absorption, leading to highly
efficient photoexcitation. Furthermore, the continuous flow
system inhibited the reverse reaction and side reactions by
immediately removing the products from the reaction range,
leading to a high selectivity of above 95%. The photocatalytic
mechanism was deeply evaluated by DFT calculations and
in situ FTIR measurement, revealing that methane transformed
into methyl radicals to form ethane via a methyl anion acti-
vation process. This work provides new insight into how to
design an efficient system for photocatalytic NOCM.
(
hþ
eꢀ
2CH3 ꢀ! 2•CH3 ! C2H6
ꢀ
adsorption
2CH4 ꢀꢀꢀꢀꢀꢀ! 2ꢂCH4 ꢀ!
2H• ! H2
The methyl anion-radical mechanism was further sup-
ported by its comparison with the other two reported types of
mechanisms via DFT calculations (see ESI†). As shown in
Fig. S21,† one can see that the CH3 anion mechanism exhibi-
ted the lowest reaction barrier among all evaluated mecha-
nisms. This demonstrates that the methyl anion-radical
mechanism is superior to others. Indeed, the CH3 anion was
Conflicts of interest
experimentally confirmed by in situ diffuse reflectance infrared
−
Fourier transform (DRIFT) measurement, namely, CH3 anion There are no conflicts to declare.
Green Chem.
This journal is © The Royal Society of Chemistry 2020

View Article Online
Green Chemistry
Paper
23 H. Tada, T. Kiyonaga and S.-i. Naya, Chem. Soc. Rev., 2009,
38, 1849–1858.
Acknowledgements
J. L. thanks the support of China Postdoctoral Science 24 N. Siemer, A. Lüken, M. Zalibera, J. Frenzel, D. Muñoz-
Foundation (2019M651397).
Santiburcio, A. Savitsky, W. Lubitz, M. Muhler, D. Marx and
J. Strunk, J. Am. Chem. Soc., 2018, 140, 18082–18092.
25 T. Whittaker, K. S. Kumar, C. Peterson, M. N. Pollock,
L. C. Grabow and B. D. Chandler, J. Am. Chem. Soc., 2018,
140, 16469–16487.
References
1 Z. R. Chong, S. H. B. Yang, P. Babu, P. Linga and X.-S. Li, 26 Z. Bian, T. Tachikawa, P. Zhang, M. Fujitsuka and
Appl. Energy, 2016, 162, 1633–1652.
T. Majima, J. Am. Chem. Soc., 2014, 136, 458–465.
2 A. Iulianelli, S. Liguori, J. Wilcox and A. Basile, Catal. Rev., 27 W. Hou and S. B. Cronin, Adv. Funct. Mater., 2013, 23,
2016, 58, 1–35. 1612–1619.
3 Y. Wang, L. Yao, S. Wang, D. Mao and C. Hu, Fuel Process. 28 B. Han and Y. H. Hu, J. Phys. Chem. C, 2015, 119, 18927–
Technol., 2018, 169, 199–206. 18934.
4 V. L. Sushkevich, D. Palagin, M. Ranocchiari and J. A. van 29 P. Usubharatana, D. McMartin, A. Veawab and
Bokhoven, Science, 2017, 356, 523–527.
5 A. Galadima and O. Muraza, J. Ind. Eng. Chem., 2016, 37, 1–
P. Tontiwachwuthikul, Ind. Eng. Chem. Res., 2006, 45, 2558–
2568.
13.
30 L. Yu and D. Li, Catal. Sci. Technol., 2017, 7, 635–640.
6 S. C. Oh, E. Schulman, J. Zhang, J. Fan, Y. Pan, J. Meng and 31 K. Dutta, L. Li, P. Gupta, D. P. Gutierrez and J. Kopyscinski,
D. Liu, Angew. Chem., 2019, 131, 7157–7160.
7 G. Keller and M. Bhasin, J. Catal., 1982, 73, 9–19.
8 P. Wang, G. Zhao, Y. Wang and Y. Lu, Sci. Adv., 2017, 3,
e1603180.
Catal. Commun., 2018, 106, 16–19.
32 B. D. Hook, W. Dohle, P. R. Hirst, M. Pickworth,
M. B. Berry and K. I. Booker-Milburn, J. Org. Chem., 2005,
70, 7558–7564.
9 Y.-H. Hou, W.-C. Han, W.-S. Xia and H.-L. Wan, ACS Catal., 33 X. Chen, Y. Li, X. Pan, D. Cortie, X. Huang and Z. Yi, Nat.
2015, 5, 1663–1674.
Commun., 2016, 7, 12273.
10 P. Michorczyk, E. Hędrzak and A. Węgrzyniak, J. Mater. 34 C. McCullagh, P. K. Robertson, M. Adams, P. M. Pollard
Chem. A, 2016, 4, 18753–18756.
11 D. Soulivong, S. Norsic, M. Taoufik, C. Coperet, J. Thivolle-
and A. Mohammed, J. Photochem. Photobiol., A, 2010, 211,
42–46.
Cazat, S. Chakka and J.-M. Basset, J. Am. Chem. Soc., 2008, 35 G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6,
130, 5044–5045. 15–50.
12 P. Schwach, X. Pan and X. Bao, Chem. Rev., 2017, 117, 36 X.-Q. Gong, A. Selloni, O. Dulub, P. Jacobson and
8497–8520. U. Diebold, J. Am. Chem. Soc., 2008, 130, 370–381.
13 H. Sheng, E. P. Schreiner, W. Zheng and R. F. Lobo, 37 A. S. Mazheika, T. Bredow, V. E. Matulis and
ChemPhysChem, 2018, 19, 504–511.
14 Y. Xiao and A. Varma, ACS Catal., 2018, 8, 2735–2740.
15 L. Yuliati and H. Yoshida, Chem. Soc. Rev., 2008, 37, 1592–
1602.
16 L. Yuliati, T. Hattori, H. Itoh and H. Yoshida, J. Catal.,
2008, 257, 396–402.
O. A. Ivashkevich, J. Phys. Chem. C, 2011, 115, 17368–17377.
38 H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State,
1976, 13, 5188.
39 J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson,
M. R. Pederson, D. J. Singh and C. Fiolhais, Phys. Rev. B:
Condens. Matter Mater. Phys., 1992, 46, 6671–6687.
17 L. Yuliati, T. Hamajima, T. Hattori and H. Yoshida, Chem. 40 G. Trucks, H. Schlegel, G. Scuseria, M. Robb, J. Cheeseman,
Commun., 2005, 4824–4826.
18 L. Li, Y. Y. Cai, G. D. Li, X. Y. Mu, K. X. Wang and
J. S. Chen, Angew. Chem., Int. Ed., 2012, 51, 4702–4706.
J. Montgomery Jr., T. Vreven, K. Kudin, J. Burant and
J. Millam, Gaussian 03, Gaussian Inc, Wallingford, CT, 2004.
41 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
19 L. Li, G. D. Li, C. Yan, X. Y. Mu, X. L. Pan, X. X. Zou, 42 D. Andrae, U. Haeussermann, M. Dolg, H. Stoll and
K. X. Wang and J. S. Chen, Angew. Chem., Int. Ed., 2011, 50,
8299–8303.
H. Preuss, Theor. Chim. Acta, 1990, 77, 123–141.
43 C.-T. Au, C.-F. Ng and M.-S. Liao, J. Catal., 1999, 185, 12–22.
20 L. Meng, Z. Chen, Z. Ma, S. He, Y. Hou, H.-H. Li, R. Yuan, 44 C. Yang and Z.-Y. Zhao, Comput. Mater. Sci., 2018, 151,
X.-H. Huang, X. Wang, X. Wang and J. Long, Energy
Environ. Sci., 2018, 11, 294–298.
21 R. Daghrir, P. Drogui and D. Robert, Ind. Eng. Chem. Res.,
2013, 52, 3581–3599.
160–173.
45 A. Antušek, M. Blaško, M. Urban, P. Noga, D. Kisić,
M. Nenadović, D. Lončarević and Z. Rakočević, Phys. Chem.
Chem. Phys., 2017, 19, 28897–28906.
22 Z. Bian, T. Tachikawa, W. Kim, W. Choi and T. Majima, 46 H.-G. Cho and L. Andrews, Dalton Trans., 2011, 40, 11115–
J. Phys. Chem. C, 2012, 116, 25444–25453.
11124.
This journal is © The Royal Society of Chemistry 2020
Green Chem.