Highly Efficient and Selective Photocatalytic Nonoxidative Coupling of Methane to Ethylene over Pd-Zn Synergistic Catalytic Sites

Photocatalytic nonoxidative coupling of CH4 to multicarbon (C2+) hydrocarbons (e.g., C2H4) and H2 under ambient conditions provides a promising energy-conserving approach for utilization of carbon resource. However, as the methyl intermediates prefer to undergo self-coupling to produce ethane, it is a challenging task to control the selective conversion of CH4 to higher value-added C2H4. Herein, we adopt a synergistic catalysis strategy by integrating Pd-Zn active sites on visible light-responsive defective WO3 nanosheets for synergizing the adsorption, activation, and dehydrogenation processes in CH4 to C2H4 conversion. Benefiting from the synergy, our model catalyst achieves a remarkable C2+ compounds yield of 31.85 μmol·g−1·h−1 with an exceptionally high C2H4 selectivity of 75.3% and a stoichiometric H2 evolution. In situ spectroscopic studies reveal that the Zn sites promote the adsorption and activation of CH4 molecules to generate methyl and methoxy intermediates with the assistance of lattice oxygen, while the Pd sites facilitate the dehydrogenation of methoxy to methylene radicals for producing C2H4 and suppress overoxidation. This work demonstrates a strategy for designing efficient photocatalysts toward selective coupling of CH4 to higher value-added chemicals and highlights the importance of synergistic active sites to the synergy of key steps in catalytic reactions.


Introduction
Under the reality of insufficient coal and oil stockpiles, conversion of methane (CH 4 ), which is the predominant component in natural gas, biogas, shale gas, and combustible ice, to value-added chemical feedstocks is an intriguing approach for sustainable development [1][2][3][4]. However, as a nonpolar molecule with tetrahedral symmetry, CH 4 has a high C-H bond energy which requires high energy input (i.e., high operating temperatures and pressures) to cleave the C-H bond [5][6][7]. Additionally, such hash reaction conditions commonly lead to the production of undesired but thermodynamically favorable overoxidized products (i.e., CO and CO 2 ) [8][9][10][11]. Given such circumstance, there are giant economic and environmental incentives for developing efficient sustainable approaches to achieve selective CH 4 conversion toward the target products.
Photocatalysis, employing inexhaustible solar energy instead of thermal energy, provides an attractive alternative route to sustainable CH 4 conversion under ambient reaction conditions [11][12][13][14][15]. Among various methane conversion schemes, nonoxidative coupling of CH 4 to ethylene (C 2 H 4 ) along with simultaneous production of H 2 is a preferable pathway, as C 2 H 4 is the high value-added key chemical feedstock, while H 2 is an important clean energy carrier. Nevertheless, it is still a grand challenge to achieve efficient and selective conversion of CH 4 to C 2 H 4 , mainly because photocatalysts often lack efficient active sites for activation of C-H bond, and the generated methyl radicals upon activation prefer to undergo self-coupling toward production of less valued ethane (C 2 H 6 ) [16][17][18][19][20]. In this regard, there is an urgent need to rationally engineer active sites on the photocatalyst surface for synergizing the adsorption, activation, and dehydrogenation processes to enable in achieving the efficient and selective photocatalytic nonoxidative coupling of CH 4 to C 2 H 4 .
Among various reported active sites, the Zn + −O − pairs in ZnO have been well recognized as efficient active sites for photocatalytic CH 4 activation and coupling, and as such, ZnO has been extensively applied to construct photocatalysts for CH 4 conversion [21][22][23]. However, due to the wide band gap, ZnO can only absorb ultraviolet light, which severely restricts their catalytic efficiency in practical application under sunlight. In addition, the insufficient dehydrogenation capability for Zn + sites limits the efficiency of C 2 H 4 production. To achieve the goal of CH 4 to C 2 H 4 nonoxidative coupling conversion, it is greatly desired, yet challenging, to implement the Zn + −O − pairs in other photosensitive semiconductor materials with the light harvesting capacity in broad spectral range and to simultaneously introduce another active site with strong dehydrogenation capability.
Herein, we aim to adopt a synergistic catalysis strategy by integrating multiple active sites on a visible lightresponsive substrate for harnessing the adsorption, activation, and dehydrogenation processes to achieve highly efficient and selective CH 4 to C 2 H 4 conversion. Taking the visible light-responsive defective WO 3 as a model substrate, Zn + −O − paired active sites are implemented into WO 3 nanosheets with large specific surface area and suitable energy band position through a doping method [24,25]. Furthermore, considering the strong dehydrogenation capability toward C-H bond, Pd sites are introduced on WO 3 nanosheets by depositing Pd nanoparticles via a selfreduction process. Benefiting from the synergy of two active sites, our model catalyst achieves remarkable activity and selectivity for CH 4 to C 2 H 4 conversion as well as a nearly stoichiometric H 2 evolution, benchmarked against the state-of-the-art photocatalysts. In situ spectroscopic studies reveal that the Zn sites promote the adsorption and activation of CH 4 molecules, while the Pd sites facilitate the dehydrogenation of methoxy intermediates and suppress the overoxidation [26][27][28]. Moreover, we unravel the reaction pathway for CH 4 to C 2 H 4 , in which the adsorbed CH 4 is activated and dehydrogenated to in turn generate methyl, methoxy, and methylene intermediates, and finally, the methylene radicals undergo self-coupling reactions to produce C 2 H 4 . This work provides a new perspective for designing the photocatalyst through leveraging synergistic active sites and highlights the key role of strong dehydrogenation capability in enhancing the selectivity for CH 4 to C 2 H 4 conversion.

Results
As illustrated in Figure 1(a), the Pd-Zn comodified WO 3 nanosheets are constructed through a two-step protocol, in which Zn is in situ doped during hydrothermal synthesis of defective WO 3 nanosheets, and subsequently, Pd nanoparticles are integrated by self-reduction. The prepared model catalyst is denoted as Pd 5 /Zn 0.35 -WO 3 (5% and 0.35% refer to the theoretical mass fractions of Pd and Zn in the composite). The actual mass fractions of Pd and Zn determined by ICP-MS results are 4.46% (10.27 mg/L) and 0.31% (0.71 mg/L) in Pd 5 /Zn 0.35 -WO 3 (230 mg/L). The x-ray diffraction (XRD) patterns ( Figure S1a) show that the prepared defective WO 3 substrate is monoclinic phase (JCPDS Card No. 83-0950), and the crystalline phase structure remains unchanged after Zn doping and Pd modification [25]. Such a result is consistent with that from UV-vis diffuse reflectance spectra (DRS, Figure S1b), showing that the samples maintain the inherent light absorption characteristics of defective WO 3 substrate after the Zn doping and Pd modification. Taken together, the results indicate that the Pd-Zn comodification strategy has no negative effect on the basic physical properties of WO 3 substrate.
Scanning electron microscopy (SEM) images ( Figure S2a and S2b) reveal that the bare WO 3 substrate has a nanosheet structure with a diameter of approximately 400 nm. Furthermore, transmission electron microscopy (TEM) also demonstrates the nanosheet morphology of the prepared Pd-Zn comodified WO 3 (Pd 5 /Zn 0.35 -WO 3 ). The clear lattice fringes with the spacings of 0.379 nm and 0.397 nm at an angle of 90°in a high-resolution TEM (HRTEM) images can be assigned to the (020) and (002) crystal planes of monoclinic phase WO 3 (Figure 1(b)), which are consistent with the XRD results [25]. Meanwhile, such a nanosheet morphology remains unchanged after Zn doping and Pd modification ( Figure S2c and S2d). Additionally, the aberration-corrected high-angle annular dark-field scanning TEM (HAADF-STEM) image of the Pd-Zn comodified WO 3 nanosheets (Figure 1(c)) shows that Pd nanoparticles with an average diameter of 5 nm are highly dispersed on the WO 3 substrate surface. The lattice fringes with an interplanar distance of 0.228 nm ascribed to the (111) plane of Pd nanoparticles are clearly observed in HRTEM image (Figure 1(d)) [28]. The corresponding energy-dispersive x-ray spectroscopy (EDS) elemental mapping (Figure 1(e)) demonstrates that Zn dopants and Pd nanoparticles are uniformly distributed on the surface of the WO 3 substrate. Moreover, x-ray photoelectron spectroscopy (XPS, Figure S3) reveals that isolated Zn 2+ ions have been successfully doped into the WO 3 lattice while the zero-valent state of Pd proves that Pd nanoparticles are indeed formed by the self-reduction method. Note that low-valence W 5+ species is resolved along with the binding energy of 531.8 eV for adsorbed oxygen species at the defects in the WO 3 substrate by XPS, manifesting the existence of oxygen vacancies as defects [29,30]. Taken together, the characterizations confirm that Zn sites and Pd nanoparticles have been successfully introduced onto WO 3 substrate, constructing the Pd-Zn comodified WO 3 nanosheets.
After confirming the formation of Pd-Zn comodified WO 3 nanosheets, we then evaluate their performance as a photocatalyst for CH 4 conversion under full-spectrum 2 Research illumination. The catalytic performance of the synthesized Pd-Zn comodified WO 3 (denoted as Pd x /Zn y -WO 3 , x%, and y% refer to the mass fractions of Pd and Zn in the composite) in reference to control samples are summarized in Figure 2(a). The CH 4 coupling products of the optimized Pd 5 /Zn 0.35 -WO 3 are C 2 H 6 , C 2 H 4 , and C 3 H 6 ( Figure S4, S5), in which the C 2+ compounds yield reaches 31.85 μmol·g -1 ·h -1 with a C 2 H 4 selectivity of 75.3% (57% in total carbonaceous products). The control experiments demonstrate that there is no thermal-catalytic contribution ( Figure S6a), indicating that the coupling of CH 4 is a photocatalytic reaction rather than a photothermal catalytic reaction. Furthermore, the Pd 5 /Zn 0.35 -WO 3 catalyst can still effectively realize selective photocatalytic nonoxidative coupling of CH 4 to C 2 H 4 outside the laboratory under condensed sunlight ( Figure S6a), indicating its great potential for practical application. Such performance well exceeds the activity and selectivity of the state-of-the-art catalysts for photocatalytic nonoxidative coupling of CH 4 to C 2 H 4 ( Table S1). The yields of C 2 H 4 and C 2 H 6 for Pd 5 /Zn 0.35 -WO 3 are 19-fold and 5.5-fold higher than that of bare WO 3 , confirming that the introduced Zn and Pd give a boost to the activation and coupling of CH 4 . More notably, the yields of H 2 and carbon-containing products almost conform to the stoichiometric ratio during the CH 4 conversion process, indicating that the hydrogen atoms are derived from CH 4 with high atom economy ( Figure S6b and  3 Research nanoparticles are explored by investigating the catalytic performance for the control samples with only Zn doping and Pd loading, as well as assessing the effects of their loading amounts on the performance. Compared to bare WO 3 , the Zn-doped WO 3 exhibits remarkably enhanced C 2 products yield and C 2 H 4 selectivity, and the C 2 H 4 yield increases with the amount of doped Zn within a certain range (Figure 2(a)), suggesting that the doped Zn can facilitate CH 4 activation and coupling. However, excessive Zn doping leads to the decrease in C 2 products yield, most likely due to the increased work function and reduced lattice oxygen content by excessive doped Zn [31]. The activation of CH 4 heavily depends on the O − centers in Zn + −O − pairs. In the absence of Pd, the O − centers in Zn 0.35 -WO 3 serving as strong oxidants cause the serious overoxidation of activated ·CH 3 to CO 2 ( Figure S6b). The Pd nanoparticles-modified WO 3 (Pd 5 /WO 3 ) also exhibits substantially enhanced C 2 products yield and C 2 H 4 selectivity, while a considerable amount of C 3 H 6 emerges. The production of C 3 H 6 indicates that the introduced Pd nanoparticles serving as   4 Research active sites have a strong dehydrogenation capability to further dehydrogenate methyl intermediates to generate methylene and methyne radicals, which undergo crosscoupling reaction to produce C 3 H 6 . Furthermore, after modifying Zn 0.35 -WO 3 with Pd nanoparticles, the production yield of C 2 H 4 further increases significantly with the loading amount of Pd (Figure 2(a)). This verifies that the modified Pd nanoparticles are conducive to dehydrogenating methyl radicals and further promoting the generated radicals to undergo self-and cross-coupling reactions, which eventually dramatically suppresses the overoxidation of carbon intermediates to CO 2 as compared with Zn 0.35 -WO 3 ( Figure S6b). Moreover, excessive Pd loading, such as Pd 7 /Zn 0.35 -WO 3 , leads to the reduced yields of C 2 H 6 and C 2 H 4 without production of C 3 H 6 , most likely because the agglomeration of small Pd nanoparticles weakens their dehydrogenation effect. The results above demonstrate the significant roles of synergistic Pd and Zn on WO 3 in promoting the CH 4 coupling and C 2 H 4 production as well as suppressing the overoxidation.
To evaluate the stability of the model photocatalyst, the physicochemical properties of Pd 5 /Zn 0.35 -WO 3 after reaction are investigated. After the reaction, the Pd 5 /Zn 0.35 -WO 3 sample exhibits slight changes in color, crystallinity, and light absorption properties, but no obvious change on morphology can be observed ( Figure S7a-d). Such changes are related to the consumption of a small amount of lattice oxygen in the sample during the reaction, as evidenced by O 1 s XPS spectra ( Figure S7e). The production of CO 2 ( Figure S6b) under oxygen-free reaction conditions also proves that the lattice oxygen could be consumed in the process of CH 4 overoxidation. Nevertheless, such lattice oxygen consumption would not limit the long-term application of photocatalyst; the consumed lattice oxygen can be effectively replenished after photo-oxidation treatment under air conditions by seizing oxygen atoms from the environment. The color, crystallinity, and light absorption properties are almost restored to the state of fresh sample after the treatment ( Figure S7). These results suggest that the model photocatalyst can maintain the recyclability through the batch reaction mode. To further assess the recoverability and recyclability of our model photocatalyst, the cycling tests are performed on the Pd 5 /Zn 0.35 -WO 3 catalyst, during which a photo-oxidation pretreatment of 30 min is performed on the recycled catalyst under air conditions before each cycle (Figure 2(b)). After the six cycles, the recycled catalyst well retains the activity and selectivity for photocatalytic coupling of CH 4 to C 2 H 4 , manifesting the eminent recyclability and practicability of our model photocatalyst.
To further verify the carbon source of produced C 2 H 4 and C 2 H 6 in photocatalytic CH 4 conversion, isotope labeling experiment is performed by using 13 CH 4 as the reactant. The 13 C 2 H 4 , 13 C 2 H 6 , and 3 CO 2 products as well as various intermediates derived from isotopic 13 CH 4 can be observed by gas chromatography-mass spectrometry (GC-MS, Figure 2(c) and S8). Additionally, no products are detected during the control experiments without catalyst or under dark condition (Figure 2(a)). These results confirm that all coupling products are derived from CH 4 , rather than the release of any residual organic matters in raw materials during the preparation process.
To elucidate the origin of the superior activity by our designed photocatalyst, we further survey its charge dynamics behavior, which is a key factor to the photocatalytic efficiency. The time-resolved surface photovoltage (TR-SPV), as an advanced characterization method to comprehend the behavior of charge separation and transfer, can qualitatively estimate the charge separation efficiency and photogenerated carrier lifetime by analyzing the signal intensity and duration. To put it simply, the stronger TS-SPV signal, the better charge separation; the wider TS-SPV signal, the longer charge lifetime. For the Pd 5 /Zn 0.35 -WO 3 with Zn doping and subsequent Pd modification on WO 3 , both the intensity and duration of the TS-SPV signal gradually increase (Figure 3(a)), suggesting the enhanced charge separation efficiency and photogenerated carrier lifetime by introducing Zn and Pd. Such a positive effect on charge dynamics is also confirmed by the steady-state surface photovoltage spectroscopy ( Figure S9). Moreover, the transient-state photoluminescence (TS-PL) spectroscopy (Figure 3(b) and Table S3)  Since the promoting effect of Pd-Zn comodification on charge separation and transfer is clarified, in situ electron paramagnetic resonance (EPR) technology is used to understand the specific charge migration process. As shown in the EPR spectra (Figure 3(c) and S10), the intensity of Zn + signal at g = 1:968 for Pd 5 /Zn 0.35 -WO 3 increases upon light irradiation, manifesting that the photogenerated electrons are transferred from WO 3 to doped Zn 2+ to produce Zn + sites [32]. In addition, the signals at g = 2:005 attributed to unpaired electrons trapped in surface defects (V O + or O − ) become stronger for both WO 3 and Pd 5 /Zn 0.35 -WO 3 upon light irradiation. Considering that Zn + and O − are always generated in pairs, the enhanced signals at g = 2:005 for Pd 5 /Zn 0.35 -WO 3 suggest the emergence of photogenerated holes-enriched lattice oxygen sites (O − centers) [22,28]. Meanwhile, the enhanced signals at g = 2:005 for WO 3 can be attributed to the unpaired photogenerated electrons trapped in oxygen vacancies, as there is no obvious signal for lattice electron trapping sites. Upon introducing CH 4 , the signal intensity of O − centers at g = 2:005 slightly decreases for Pd 5 /Zn 0. 35  Upon ascertaining the charge dynamics behavior, we further examine the role of introduced Zn and Pd in CH 4 activation and coupling process. The CH 4 temperature programmed desorption (CH 4 -TPD) measurements are first performed to explore the CH 4 adsorption behavior, which is the essential prerequisite for activation process. As shown by the CH 4 -TPD curves (Figure 3(d)), two wide gas desorp-tion peaks, corresponding to physical adsorption and chemical adsorption, appear, respectively, in the low temperature range (100-150°C) and the middle temperature range (200-400°C) for WO 3 . After doping Zn sites, Zn 0.35 -WO 3 shows similar physical adsorption peak in low temperature range, while the chemical desorption peak area in middle temperature range increases significantly as compared with WO 3 , indicating that more CH 4 is firmly adsorbed on the catalyst surface by chemisorption. More importantly, a new CH 4 chemical desorption peak appears around 325°C after Zn doping, suggesting that the doped Zn sites can promote the CH 4 adsorption capacity, thereby contributing to the enhanced CH 4 conversion performance. When the Pd nanoparticles are incorporated into the catalyst, the CH 4 chemical desorption temperature for Pd 5 /WO 3 and Pd 5 /Zn 0.35 -WO 3 shifts toward higher temperature by about 25°C, and the peak area further increases. This indicates that the additional Pd nanoparticles can further enhance the CH 4 adsorption capacity through increasing the binding strength for CH 4

Research
technique is further used to analyze the related species derived from CH 4 dissociation. As the essential product in CH 4 conversion, the peak area for H 2 is significantly enhanced after Zn doping ( Figure S11), especially after Pd-Zn comodification as compared with WO 3 , demonstrating that the modified Pd nanoparticles exhibit a stronger effect on dehydrogenation of the intermediates, contributing to the increased H 2 production. The results above indicate that the doped Zn plays a major role in promoting the CH 4 adsorption, while the modified Pd nanoparticles play a dominant role in facilitating the dehydrogenation of the intermediates.
To further decode the reaction mechanism, in situ diffuse reflectance infrared Fourier-transform spectroscopy (DRIFTS) and in situ near ambient pressure x-ray photoelectron spectroscopy (NAP-XPS) are employed to track the evolution of intermediates during the CH 4 activation and coupling process. As shown in the in situ DRIFTS spectra for WO 3 , Zn 0.35 -WO 3 , and Pd 5 /Zn 0.35 -WO 3 (Figures 4(a)-4(c)), when the photocatalysts are fully immersed in the CH 4 atmosphere for 40 min under dark conditions (Figure 4(a)), the absorption peaks at 1428, 1473, and 3015 cm -1 , corresponding to the symmetrical and asymmetrical deformation vibration of the C-H bond in CH 4 molecules, emerge and increase with the adsorption time [21,27,[33][34][35]. This implies that the CH 4 molecules are increasingly adsorbed on the photocatalyst surface. After modification of Zn and Pd, the intensity of each CH 4 -related absorption peak for Zn 0.35 -WO 3 , and Pd 5 /Zn 0.35 -WO 3 in the same adsorption time period is significantly promoted as compared with WO 3 . This observation is particularly emphasized on the peak for the symmetric vibration of the C-H bond in CH 4 molecules (1541 cm -1 ), proving that more CH 4 molecules are captured on photocatalyst surface attributed to the modified Zn and Pd. In addition, when the adsorption time exceeds 30 min, the peak intensity for CH 4 has no obvious increase, suggesting that a saturated adsorption has been achieved.
Furthermore, the light irradiation is introduced to examine the CH 4 coupling reaction process (Figures 4(b) and 4(c)). With the increased light irradiation time, the in situ DRIFTS spectra for Pd 5 /Zn 0.35 -WO 3 show that the intensities of various CH 4 -related peaks have no obvious change, indicating that the adsorption sites can continuously capture CH 4 from the environment to maintain the adsorption saturation state during the CH 4 coupling process (Figure 4(c)). In stark contrast, multiple peaks at 870, 890 cm -1 , and in the range of 3200-3600 cm -1 corresponding to the C-C bond and -OH groups of intermediates appear upon light irradiation, and their intensities gradually increase with the light irradiation time, indicating that the adsorbed CH 4 has undergone a cleavage and coupling process to generate C 2 H 6 [36]. Remarkably, the peak at 1641 cm -1 in the in situ DRIFTS spectra for Pd 5 /Zn 0.35 -WO 3 , assigned to C=C of methylene radicals, suggests the process that the adsorbed CH 4 is dehydrogenated, and coupled to produce C 2 H 4 [37].
In situ NAP-XPS studies are further performed to supplement the information for the evolution of intermediates during the CH 4 coupling reaction. The corresponding high-resolution C 1 s XPS spectrum of Pd 5 /Zn 0.35 -WO 3 displays a peak at 284.8 eV under vacuum condition (Figure 4(d)), due to the exogenous residual carbon on photocatalyst surface [21,37]. Although such residual carbon cannot be completely purged by the Ar flow, the peak intensity basically maintains unchanged and thus has no interference to the subsequent measurement. With the continuous CH 4 adsorption and accumulation, the intensity of the peak at 286.3 eV assigned to methoxy intermediates increases significantly, indicating that the generated methyl intermediates are adsorbed on lattice oxygen sites. The peak intensity for the methoxy intermediates significantly decreases upon light irradiation (Figure 4(e)), suggesting that the methoxy intermediates are consumed by participating in the following reaction. With the continuous light irradiation, the peak intensities for the methoxy intermediates are recovered, indicating that the methoxy intermediates can be continuously supplemented for maintaining the CH 4 coupling reaction [38][39][40].
Based on the information gleaned above, the roles of modified Zn and Pd on charge dynamics and reaction intermediates evolution have been elucidated. In addition, the photogenerated holes in the valence band maximum of WO 3 are thermodynamically feasible for driving the oxidation of CH 4 to methyl radical ( Figure S12) [41]. As such, a reasonable reaction pathway including the photogenerated charge transfer path can be proposed as illustrated in Figure 5. Upon light irradiation, the photogenerated holes are enriched at lattice oxygen (O 2-) sites to form O − centers, while the photogenerated electrons are transferred to the nearby doped Zn 2+ sites via W 6+ , forming Zn + −O − pairs. The formed Zn + sites are available for capturing the CH 4 molecules through donating the single electron to the empty C−H σ * -antibonding orbital of CH 4 molecule, while the O − centers have a strongly attractive force to abstract the H atoms from CH 4 [22,23]. Subsequently, the adsorbed CH 4 is activated by Zn + −O − pairs to generate methyl intermediates, which are then stabilized by the Zn + sites. The formed methyl intermediates can either follow the C 2 H 6 pathway after desorbing from the catalyst to generate methyl radicals, or undergo the C 2 H 4 pathway by diffusing onto the adjacent O 2sites and being further dehydrogenated by Pd nanoparticles to generate methylene radicals. Finally, the produced free methyl radicals and methylene radicals undergo self-coupling reactions to produce the products of C 2 H 4 and C 2 H 6 . Simultaneously, the H atoms dissociated from the activated CH 4 can couple with each other to produce H 2 . The holes enriched at O − centers can recombine with the electrons trapped at Zn + sites to generate O 2and Zn 2+ . It should be noted that the lattice oxygen participates in the photocatalytic CH 4 conversion in two different pathways. One is producing O − centers as active sites to activate CH 4 molecules for photocatalytic nonoxidative coupling of CH 4 with a stoichiometric H 2 evolution, during which the lattice oxygen will not be consumed. The other is serving as strong oxidants leading to the over oxidation of CH 4 to CO 2 , during which the lattice oxygen will be consumed similarly to the Mars-van Krevelen mechanism [42].

Research
Nevertheless, the consumed lattice oxygen during the overoxidation process can be effectively supplemented by the photo-oxidation treatment under air conditions, realizing the recycling of the model Pd 5 /Zn 0.35 -WO 3 photocatalyst.

Discussion
In summary, we have designed Pd-Zn comodified WO 3 nanosheets as advanced photocatalysts for efficient CH 4 nonoxidative coupling to C 2 H 4 with high selectivity under ambient conditions and moderate light irradiation. The opti-mized Pd 5 /Zn 0.35 -WO 3 nanostructure, featuring the Zn and Pd as synergistic active sites, achieves superior catalytic performance as compared to the state-of-the-art catalysts for photocatalytic nonoxidative coupling of CH 4 to C 2 H 4 . Importantly, a nearly stoichiometric yield of valuable H 2 also has been obtained, suggesting the giant economic incentives. Based on structural characterization and in situ spectroscopic analysis, the remarkable performance of our model photocatalyst is ascribed to the synergized adsorption, activation, and dehydrogenation of CH 4   8 Research a major role in promoting the adsorption and activation of CH 4 molecules, while the Pd sites play a dominant role in facilitating the dehydrogenation of the intermediates and suppressing overoxidation. The synergistic functions dramatically increase the selectivity toward C 2 H 4 and efficiently suppress the CH 4 overoxidation to CO 2 . This work provides insights for designing highly efficient photocatalyst for selective photocatalytic nonoxidative coupling of CH 4 toward the high value-added hydrocarbon products through leveraging synergistic catalytic sites to synergize the adsorption, activation, and dehydrogenation processes.   Figure S4. GC data for photocatalytic CH 4 conversion over (a) WO 3 , (b) Zn 0.35 -WO 3 , and (c) Pd 5 /Zn 0.35 -WO 3 . Figure S5. Time-dependent production yield of C 2 H 4 , C 2 H 6 , and C 3 H 6 in photocatalytic CH 4 conversion over Pd 5 /Zn 0.35 -WO 3 photocatalysts under light irradiation of 2 h. Figure S6. Products yields and theoretical H 2 yield for CH 4 Figure S8. GC-MS data of 13 CO 2 produced in photocatalytic 13 CH 4 coupling by Pd 5 / Zn 0.35 -WO 3 . Figure S9. SS-SPS responses of WO 3 , Zn 0.35 -WO 3 , and Pd 5 /Zn 0.35 -WO 3 . Figure S10. In situ EPR signals of WO 3 collected under different conditions. Figure S11. TPD-MS profiles of H 2 over WO 3 , Zn 0.35 -WO 3 , and Pd 5 / Zn 0.35 -WO 3 . Figure S12. Illustration of the band structures of WO 3 and the redox potentials for oxidizing CH 4 to·CH 3 . Table S1. The comparison of catalytic performance with representative state-of-the-art photocatalysts for photocatalytic coupling of CH 4 to C 2 compounds.