Photothermal direct methane conversion to formaldehyde at the gas-solid interface under ambient pressure

Catalyst screening and characterizations

To screen metallic SAs-modified ZnO (M₁-ZnO) catalysts for the gas‒solid photothermal catalytic selective oxidation of CH₄ to HCHO, density functional theory (DFT) calculations were performed. The key step in the photothermal catalytic selective oxidation of CH₄ to HCHO is the effective desorption of HCHO from the catalyst to prevent its excessive oxidation^11,22,23. Therefore, the ZnO (101) model is selected for modeling HCHO adsorption on metallic SAs modified catalysts. As shown in Fig. 1a, the adsorption energy of HCHO (E_ads) on M₁-ZnO follows the order: Ag₁-ZnO (−1.10 eV) < Au₁-ZnO (−1.41 eV) < Pt₁-ZnO (−1.51 eV) < Pd₁-ZnO (−1.56 eV), indicating Ag SAs incorporation can effectively promote HCHO generation. Based on theoretical simulation results, a series of metal SAs (Ag, Au, Pt, and Pd) modified ZnO (Supplementary Fig. 1) were synthesized for photothermal catalytic oxidation of CH₄. It is found that Ag₁-ZnO with the lowest HCHO adsorption energy shows the highest HCHO generation rate and selectivity (Fig. 1b). There is a good correlation between HCHO/(CO + CO₂) in the photothermal catalytic oxidation of CH₄ on M₁-ZnO and the adsorption energy of HCHO on M₁-ZnO in DFT calculations (Fig. 1c), indicating that metal SAs doping can effectively regulate the adsorption performance of HCHO to control the photothermal catalytic selective oxidation of CH₄ to HCHO. Based on theoretical and experimental results, a specific study is then conducted on the photothermal catalytic selective oxidation of CH₄ to HCHO on Ag₁-ZnO samples in reference to pure ZnO and Ag nanoparticle-modified ZnO (Ag_NP-ZnO).

a Calculated adsorption energy of HCHO (E_ads) on M₁-ZnO. b Gas‒solid photothermal catalytic CH₄ oxidation performance for M₁-ZnO. c HCHO/(CO + CO₂) of gas‒solid photothermal catalytic CH₄ oxidation for M₁-ZnO plotted against the HCHO adsorption on M₁-ZnO. The reaction conditions in (b): CH₄ (45 mL min⁻¹) +  air (5 mL min⁻¹), RH = 20%, external heating at 150 °C, Xe lamp with wavelength range of 200–1200 nm and 1000 mW cm^–2 light intensity. Blue, red, black, white, dark gray, yellow, light gray and turquoise spheres in (a) represent Zn, O, C, H, Ag, Au, Pt and Pd atoms, respectively. Source data are provided as a Source Data file.

The fabricated ZnO exhibits nanoparticles with an average size of approximately 22.7 nm (Supplementary Fig. 2a, b). The high-resolution transmission electron microscopy (HRTEM) image (Supplementary Fig. 2c) shows the lattice fringes of ZnO with an interplanar spacing of 0.247 nm, corresponding to the (101) plane of ZnO. The Ag₁-ZnO and Ag_NP-ZnO samples were obtained through a simple 85 °C bath assisted NaH₂PO₂ reduction method by regulating the amounts of AgNO₃ and NaH₂PO₂. The actual loading amounts of Ag in Ag₁-ZnO and Ag_NP-ZnO are 0.18 wt.% and 0.96 wt.%, respectively, as measured by inductively coupled plasma–optical emission spectroscopy (ICP‒OES). The X-ray diffraction (XRD) spectra of all prepared samples (Supplementary Fig. 3) are well indexed to the typical hexagonal wurtzite structure of ZnO (JCPDS file no. 36-1451)²⁴. No characteristic diffraction peaks of Ag can be observed on either Ag₁-ZnO or Ag_NP-ZnO, demonstrating the small size and good dispersion of Ag species on ZnO. For Ag₁-ZnO, the Ag species remain highly dispersed on the ZnO NPs from the energy-dispersive X-ray spectroscopy (EDS) elemental mapping images (Fig. 2a). The Ag atoms can be observed from the aberration-corrected high-angle annular dark-field scanning TEM (AC-HAADF-STEM) image due to its being much brighter than Zn atoms (Fig. 2b). Furthermore, a lattice spacing of 0.236 nm corresponding to the Ag (111) plane appears in the Ag_NP-ZnO sample (Supplementary Fig. 4a), and the EDS elemental mapping image (Supplementary Fig. 4b) further verifies the formation of Ag NPs in the Ag_NP-ZnO sample. And the average size of Ag NPs on Ag_NP-ZnO sample is around 12.9 nm (Supplementary Fig. 4c, d). Moreover, the ultraviolet–visible diffuse reflectance spectroscopy (UV–Vis DRS) (Supplementary Fig. 5) clearly shows that only the Ag_NP-ZnO sample exhibits a surface plasmon peak, which also proves the formation of Ag NPs and Ag SAs in Ag_NP-ZnO and Ag₁-ZnO samples, respectively.

a HAADF-STEM image and the corresponding EDS elemental mapping signals of Ag₁-ZnO. b Magnified high-resolution AC-HAADF-STEM image of Ag₁-ZnO and the corresponding intensity profile from the marked position. c The normalized XANES spectra at the Ag K-edge. d The k³-weighted Fourier transform EXAFS spectra at the Ag K-edge. e Wavelet transformation for the k³-weighted EXAFS signal at the Ag K-edge. Source data are provided as a Source Data file.

To further explore the local structure of Ag species in Ag₁-ZnO and Ag_NP-ZnO, X-ray photoelectron spectroscopy (XPS) and synchrotron X-ray absorption spectroscopy (XAS) were carried out. The X-ray absorption near-edge structure (XANES) spectrum at the Ag K-edge of Ag_NP-ZnO (Fig. 2c) shows a similar white-line intensity to that of the Ag foil, indicating that metallic Ag is the dominant Ag species in Ag_NP-ZnO. The white line for Ag₁-ZnO is between Ag foil and Ag₂O, indicating that the strong metal-support interaction leads to the formation of partially oxidized Ag species (Ag⁰ <Ag^δ+ <Ag⁺) (Fig. 2c). The states of Ag species of Ag₁-ZnO and Ag_NP-ZnO are also confirmed by the XPS results (Supplementary Fig. 6). The Fourier-transformed k³ weighted extended X-ray absorption fine structure (EXAFS) spectra at the Ag K-edge of Ag₁-ZnO (Fig. 2d) display a prominent Ag–O shell peak at 1.59 Å, and there is no Ag–Ag shell, confirming the atomic dispersion of Ag in Ag₁-ZnO. The Ag_NP-ZnO sample shows a major peak at the Ag–Ag shell, indicating that Ag species appear as metallic NPs. Wavelet transform EXAFS analysis (Fig. 2e) further visualizes structural information. Furthermore, EXAFS fitting analysis was conducted to determine the coordination configurations of Ag₁-ZnO. According to the best-fitted EXAFS (Supplementary Fig. 7, and Supplementary Table 1), the coordination number of Ag–O (2.06 Å) is estimated to be 2 for Ag₁-ZnO. Then, DFT calculations were carried out to acquire the structure of Ag₁-ZnO. In Ag₁-ZnO, each Ag atom binds to two O atoms (Supplementary Fig. 7c), which is consistent with the EXAFS fitting result.

Photothermal catalytic selective oxidation of CH₄ to HCHO

Photothermal catalytic oxidation of CH₄ was investigated in a 50-mL gas–solid flow reactor with irradiation of a 300-W Xe lamp (λ = 200–1200 nm, 1000 mW cm⁻²) under external heating at 150 °C (Supplementary Fig. 8). The obtained HCHO reaches stability after 60 min of illumination (Supplementary Fig. 11 and Supplementary Table 3). For comparison, the products of each sample are analyzed after 180 min of light exposure. A mixture of 50 mL min⁻¹ CH₄/air (25 vol.% O₂/N₂) = 45:5 and 20% relative humidity (RH) was used as the feed gas after screening under preliminary conditions (Supplementary Fig. 12 and Supplementary Tables 4−7). As shown in Fig. 3a and Supplementary Table 8, the yield of HCHO over Ag₁-ZnO sample is 117.8 ± 1.7 μmol h^–1, which is approximately 2.2 times higher than that over pure ZnO. The introduction of Ag SAs increases the HCHO selectivity from 59.6 ± 0.9% to 71.2 ± 0.8%. And the valuable products (CO, CH₃OH and HCHO) generated over Ag₁-ZnO sample reach 136.8 ± 3.1 μmol h^–1 with a selectivity of 82.7 ± 0.3%, and the CH₄ conversion on Ag₁-ZnO is 0.14 ± 0.003%. When the surface temperature of the catalyst is controlled at 228 °C, the apparent quantum yield (AQY) values of HCHO at 365, 400 and 420 nm on Ag₁-ZnO are 20.19 ± 2.32%, 1.98 ± 0.31% and 0.05 ± 0.01%, respectively (Supplementary Fig. 13). The activity of ZnO sample treated with NaH₂PO₂ (ZnO(SM)) is basically consistent with that of pure ZnO, indicating that simple NaH₂PO₂ treatment has no effect on the activity of ZnO. Notably, the HCHO production rate of Ag₁-ZnO can reach 11.8 ± 0.4 μmol h^–1 with a selectivity of 72.1 ± 1.3% under simulated sunlight (AM 1.5 G, 100 mW cm⁻²) when the surface temperature of catalyst is the same as the above temperature (Supplementary Fig. 14 and Supplementary Table 9). However, a bare HCHO yield (8.1 ± 0.5 μmol h^–1) and a large amount of produced CO₂ (80.0 ± 1.6 μmol h^–1) over Ag_NP-ZnO sample are achieved, indicating that the presence of Ag NPs leads to the overoxidation of CH₄. The HCHO yield obtained from the gas‒solid photothermal catalytic oxidation of CH₄ over Ag₁-ZnO under atmospheric pressure is not only higher than all reported photocatalytic systems (Supplementary Table 14)^{11,12,13,14,15,17,18,22,23,25,26}, but also still possesses superiority compare to traditional high-pressure and high-temperature catalytic reaction systems (Supplementary Table 15)^{27,28,29,30,31,32,33,34,35}.

a Activity on ZnO, ZnO(SM), Ag₁-ZnO, and Ag_NP-ZnO under external heating at 150 °C. b Activity on Ag₁-ZnO under external heating at different temperatures. c Activity on Ag₁-ZnO at a catalyst surface temperature of 228 °C under irradiation with different light intensities. d Continuous stability tests on Ag₁-ZnO. e Concentration of the absorption solution (2 mL distilled water at 0 °C) for a 3, 6 and 12 h of reaction. (The insert is a photo of the chromogenic agent added to the absorption solution after dilution 1000 times). f Comparison of the catalytic activity for photooxidation of CH₄ to HCHO over Ag₁-ZnO with other photocatalysts. The reaction conditions in a–d: CH₄ (45 mL min⁻¹) +  air (5 mL min⁻¹), and RH = 20%. Source data are provided as a Source Data file.

The yield and selectivity of HCHO increase with the increasing of external heating temperature (<150 °C) under 1000 mW cm⁻² light intensity irradiation, and the activity cannot be further increased at a higher temperature (Fig. 3b and Supplementary Table 10). The surface temperature of Ag₁-ZnO sample reaches approximately 228 °C under external heating at 150 °C during the reaction (Supplementary Fig. 15). The HCHO yield is almost directly proportional to the light intensity on Ag₁-ZnO when the surface temperature of the catalyst is controlled at 228 °C by heating (Fig. 3c and Supplementary Table 11). Only CO and CO₂ are generated on 250 °C on Ag₁-ZnO in the absence of light (Supplementary Fig. 14). These results fully demonstrate that the HCHO production is driven by light rather than heat. Heat input may promote the separation of photo-generated charge carriers and HCHO desorption, thereby promoting the generation of gas-phase HCHO. The results of comparative experiments demonstrate that no products are detected without photocatalysts or CH₄ feed (Supplementary Fig. 14 and Supplementary Table 9), indicating that all products are derived from photothermal catalytic oxidation of CH₄. Bare HCHO are obtained over Ag₁-ZnO under visible (Vis)-light (λ > 420 nm) (Supplementary Fig. 14), indicating that the HCHO is mainly generated by UV excitation.

A long-term test was carried out to assess the stability of the Ag₁-ZnO sample. The HCHO yield gradually decreases to 92.1 μmol h^–1 after a long-term test of 12 h, and the activity of the catalyst reacted for 12 h can be restored after ultrasonic washing with ultrapure water (Fig. 3d and Supplementary Table 12). The HAADF-STEM image and the corresponding EDS elemental mapping, the magnified high-resolution AC-HAADF-STEM image, and XAFS of spent Ag₁-ZnO (Supplementary Fig. 16) shows that Ag SAs are well retained after the long-term test. The XRD result confirms the stable crystalline structure of Ag₁-ZnO after the long-term test (Supplementary Fig. 17). The XPS result of the spent Ag₁-ZnO (Supplementary Fig. 18) indicates the consumption of lattice oxygen and surface carbon deposition after the long-term test. Additionally, the O₂-TPO results show significant CO₂ generation over the spent Ag₁-ZnO when the temperature rises to around 130 °C (Supplementary Fig. 19). The above results indicate that the activity attenuation of Ag₁-ZnO is mainly caused by carbon deposition and the inability of lattice oxygen regeneration in a timely manner.

To elucidate the advantages of gas‒solid photothermal catalytic selective oxidation of CH₄ to HCHO in a flow reactor, the gas after reaction over Ag₁-ZnO was absorbed for 3, 6 and 12 h with 2 mL of distilled water at 0 °C. The HCHO concentrations in the absorption solution at 3, 6 and 12 h are determined to be 160.9 ± 5.4, 297.1 ± 15.8 and 514.2 ± 33.7 µmol mL⁻¹, respectively (Fig. 3e, Supplementary Fig. 20 and Supplementary Table 13). At the same time, there is a small amount of CH₃OH in the absorption solution, which can effectively prevent the polymerization of HCHO. The 514.2 ± 33.7 µmol mL⁻¹ (1.54 ± 0.10 wt.%) of HCHO solution obtained from the gas‒solid atmospheric pressure photothermal catalytic reaction over Ag₁-ZnO outperforms all state-of-the-art photocatalysts (Fig. 3f and Supplementary Table 14)^{11,12,13,14,15,17,18,22,23,25}, whether it is a gas–liquid–solid batch reaction or continuous flow reaction. It can be observed that this strategy can effectively solve the problems of products not being separated or concentrated in a timely manner in gas–liquid–solid reactions^13,22. In addition, after 12-hour absorption of the gases from reactor, the absorption rates of HCHO and CH₃OH can still reach 81.2 ± 5.1% and 91.1 ± 3.6%, respectively, indicating that it can be expected to synthesize industrial grade HCHO solutions with the extension of reaction time and the expansion of reactor scale. In summary, the gas–solid photothermal catalysis method for CH₄ direct conversion to HCHO is of good industrial application prospects.

Insight into mechanism

To explore the origin of the high HCHO activity and selectivity over Ag₁-ZnO, the charge separation efficiency of the catalysts was studied by photocurrent and electrochemical impedance measurements. ZnO has the lowest photocurrent density (22.5 μA cm⁻²), which is increased by approximately 2.3 and 2.9 times for Ag₁-ZnO and Ag_NP-ZnO, respectively (Supplementary Fig. 21a). The electrochemical impedance results (Supplementary Fig. 21b) show that the electron transmission resistance decreases in the order of ZnO > Ag₁-ZnO > Ag_NP-ZnO. Furthermore, in situ steady-state photoluminescence (PL) was carried out to explore the thermal assistant effect. The PL peak is significantly quenched with increasing temperature in all samples, indicating that heat can improve the internal charge separation of photocatalysts (Supplementary Fig. 22). The weaker PL peaks for Ag₁-ZnO and Ag_NP-ZnO exhibit significantly suppressed charge recombination rates at both 25 and 250 °C (Supplementary Fig. 23), which is in good agreement with the photocurrent and electrochemical impedance analysis. The above results demonstrate that the introduction of Ag species and the heat input greatly improve the charge separation efficiency, which is conducive to the yield of oxygenates in photothermal catalytic CH₄ oxidation.

In situ electron spin resonance (in situ ESR) was carried out to understand the charge transport behavior in the presence of H₂O and O₂ on all samples. The signal at g = 2.000 belonging to oxygen vacancies is only observed on Ag_NP-ZnO (Fig. 4a)^36,37, indicating that Ag NPs contribute to the formation of oxygen vacancies on ZnO. Under illumination, the ESR signal at g = 1.959 (Zn⁺) in all samples increases rapidly³⁷. The order of peak intensity is: Ag₁-ZnO > Ag_NP-ZnO > ZnO (Fig. 4a), which is inconsistent with the photoelectrochemical results. It may be due to the fact that Ag SAs as a hole acceptor promote the transfer of electrons to Zn²⁺, while Ag NPs as an electron acceptor reduce the amount of Zn⁺. To confirm this speculation, in situ near-ambient pressure X-ray photoelectron spectroscopy (in situ NAP-XPS) was then performed. For the Ag₁-ZnO sample, Zn 2p peaks exhibit a negative shift (Fig. 4c), indicating that Zn²⁺ works as the electron acceptor under irradiation. And the Ag 3d peaks also show a negative binding energy shift (Fig. 4d), indicating the increase in valence state of Ag, which is different from most of other elements^38,39. The peak at 531.7 eV (Supplementary Fig. 25) attributed to the oxygen species near oxygen vacancies does not change much⁴⁰. These results show direct evidence that Ag SAs act as the hole acceptor during photocatalytic reaction. For Ag_NP-ZnO sample, the signals for Ag 3d and Zn 2p separately exhibit a positive and a small negative shift (Supplementary Fig. 26a, c), and the peak at 531.7 eV is enhanced (Supplementary Fig. 26b), which support that the electrons on Zn⁺ can transfer to the Ag NPs and the holes transfer to the lattice oxygen for activation. Meanwhile, as temperature rising, the peak intensity of Zn⁺ increases on all samples (Fig. 4b and Supplementary Fig. 24a, b), indicating that heat input facilitates the separation of photogenerated electrons and holes. Compared with those obtained under Ar flow, the peak of Zn⁺ drops and the peak of oxygen vacancies disappears after O₂ addition on Ag₁-ZnO under illumination (Supplementary Fig. 24c), which is mainly due to O₂ consuming the electrons of Zn⁺ and the activated O₂ can fill oxygen vacancies. The intensity of Zn⁺ in the presence of H₂O significantly increase compared with that obtained under Ar or O₂ flow. The holes on Ag atoms are effectively consumed by H₂O, promoting the separation of photo-generated electrons and hole, and thus achieving a large amount of Zn⁺ generation.

In situ ESR spectra of different samples at 75 °C (a), and Ag₁-ZnO at different temperatures (b) in O₂ + H₂O in the dark (dashed line) and after 10 min of light irradiation (solid line). In situ NAP-XPS of Zn 2p (c) and Ag 3d (d) over Ag₁-ZnO under dark and light irradiation. In situ DRIFTS of photothermal catalytic oxidation of CH₄ over different samples under external heating at 150 °C (e) and over Ag₁-ZnO under different external heating temperatures (f). g HCHO-TPD spectra of different samples. h GC-MS of HCHO produced by photothermal catalytic oxidation of CH₄ with  H₂¹⁶O  +  ¹⁶O₂, H₂¹⁸O  +  ¹⁶O₂ or H₂¹⁶O  +  ¹⁸O₂ over Ag₁-ZnO or Ag₁-ZnO with ¹⁸O exchange. In situ ESR spectra of different samples in Ar at 75 °C (i), and Ag₁-ZnO in Ar at different temperatures (j) in the dark (dashed line) and after 10 min of light irradiation (solid line). k COHP and ICOHP for adsorption of *CH₂ intermediates on the lattice oxygen over ZnO and Ag₁-ZnO slabs. The reaction conditions in (a), (b), (d): CH₄ (45 mL min⁻¹) +  air (5 mL min⁻¹), and RH = 20%. Blue, red, black, white, and dark gray spheres in (k) represent Zn, O, C, H and Ag atoms, respectively. Source data are provided as a Source Data file.

To verify the effect of Ag species on the gas‒solid photothermal catalytic oxidation of CH₄, the intermediate steps were dynamically monitored on different samples by in situ diffuse reflectance infrared Fourier transform spectroscopy (in situ DRIFTS). For ZnO and Ag₁-ZnO, the same characteristic peaks are observed. Upon light irradiation, the vibrational modes of *CH₃ deformation (1156 cm^–1)⁴¹, and the stretching vibrations of the C–H bond (2861 cm^–1) and C–O bond (1075 cm^–1) in the absorbed *OCH₃ species appear immediately (Fig. 4e and Supplementary Fig. 27)^42,43,44, indicating that CH₄ is rapidly dissociated into *CH₃ and exists in the form of *OCH₃ over ZnO and Ag₁-ZnO. Furthermore, the peaks of *CH₂ stretching (1430 cm^–1) and *HCHO species (2720 cm^–1) gradually increase with the prolonged light irradiation^45,46. Since the formation of HCHO via the conversion of CH₃OH has been excluded (Supplementary Fig. 28), the results presented above suggest that the target product (HCHO) is generated from the conversion of the *CH₂ intermediate, which comes from the further dehydrogenation of *CH₃. Meanwhile, the growth of *HCOO (1878 and 1773 cm^–1) and *CO₃^2– (1539 and 1239 cm^–1) species is assigned to the overoxidation of CH₄^47,48,49. All peaks show stronger intensity on Ag₁-ZnO, indicating that the doping of Ag SAs is beneficial for the dehydrogenation of CH₄ to HCHO. In contrast, intense *HCOO and *CO₃^2– species and weak *OCH₃ and *HCHO intermediates can be observed for Ag_NP-ZnO (Fig. 4e, Supplementary Fig. 27), demonstrating that CH₄ is more prone to overoxidation with the loading of Ag NPs. As the temperature increases, a significant increase in *OCH₃ and *HCHO species can be observed over Ag₁-ZnO, accompanied by the raise in the likelihood of excessive oxidation (Fig. 4f). Therefore, the selection of an appropriate reaction temperature can effectively promote the selective oxidation of CH₄ to HCHO. In addition, the HCHO-TPD experiment was used to investigate the oxidation and desorption of HCHO on various catalysts. It can be seen that the desorption peak of HCHO located at 96 °C over Ag₁-ZnO is much lower than that of ZnO (117 °C) and Ag_NP-ZnO (178 °C), indicating its easier HCHO desorption feature. HCHO begins to decompose into CO at approximately 40 °C and CO₂ is released as the temperature increases over Ag_NP-ZnO, demonstrating that the loading of Ag NPs effectively promotes the decomposition of HCHO and the activation of lattice oxygen. Hence, the loading of Ag SAs on ZnO facilitates the formation and desorption of *HCHO; while the loading of Ag NPs on ZnO promotes *HCHO adsorption and activation of oxygen, leading to lower HCHO selectivity and the overoxidation of CH₄.

To obtain an in-depth understanding of gas–solid photothermal catalytic selective oxidation of CH₄ to HCHO over Ag₁-ZnO, a series of isotope labeling experiments were carried out. The C–H bond fracture process of CH₄ is elucidated in the reaction with ¹⁸O₂ and D₂¹⁶O. As displayed in Supplementary Fig. 29, the generated water mainly exists in the form of HD¹⁶O, implying that hydrolyzed hydroxyl species mainly contribute to the dissociation of the C–H bond in CH₄. The source of oxygen in the products during the reaction was also determined through isotope labeling experiments. All signals in CH₃OH and HCHO remain basically unchanged with ¹⁶O₂ and H₂¹⁸O (Supplementary Fig. 30a, c), while HCH¹⁸O and CH₃¹⁸OH are observed in the reaction with ¹⁸O₂ and H₂¹⁶O (Supplementary Fig. 30e), indicating that O₂ is the oxygen source for HCHO and CH₃OH. The generated CO₂ is mainly in the form of C¹⁶O¹⁸O with H₂¹⁸O and ¹⁶O₂ or H₂¹⁶O₂ and ¹⁸O₂ (Supplementary Fig. 30b, d, f), demonstrating that H₂O is involved in the overoxidation of HCHO and CH₃OH. To clearly demonstrate that lattice oxygen can participate in the reaction, the reaction over Ag₁-ZnO after ¹⁸O exchange was performed^44,48. CH₃¹⁸OH and C¹⁶O¹⁸O are monitored over Ag₁-ZnO after ¹⁸O exchange, confirming that the surface lattice O in the ZnO participates in the oxidation of CH₄ (Supplementary Fig. 31). In addition, the GC-MS results of HCHO product in isotope experiments (Fig. 4h) further verify that the O in HCHO mainly comes from O₂. And the detection of HCH¹⁸O using H₂¹⁶O + ¹⁶O₂ as the feed gas over Ag₁-ZnO with ¹⁸O exchange (Fig. 4h) also confirms the involvement of lattice oxygen in the generation of HCHO. Moreover, by comparing the effects of different O₂ treatments on catalysts after the photothermal catalytic oxidation of CH₄ under anaerobic conditions (Supplementary Fig. 32), the significance of lattice oxygen in the production of HCHO is confirmed, demonstrating that the photoactivation of O₂ to replenish lattice oxygen is crucial for the reaction. Based on the above results, the photothermal catalytic oxidation of CH₄ over Ag₁-ZnO occurs via a Mars–van Krevelen mechanism. The lattice oxygen on ZnO participates in the reaction, and the consumed oxygen sites are filled with oxygen to continue the reaction.

Based on the analysis above, the activation process of lattice oxygen plays an important role in the gas‒solid photothermal catalytic selective oxidation of CH₄ to HCHO, so the activation effect of heat and Ag SAs on lattice oxygen in ZnO was further investigated. In the dark, the ESR signal of oxygen vacancies disappears after ZnO loading with Ag SAs in Ar, while the ESR peak signal of oxygen vacancies increases after loading Ag NPs in Ar (Fig. 4i). Under light, the peak of oxygen vacancies in Ag₁-ZnO appears clearly, while the signals of ZnO and Ag_NP-ZnO samples show almost no change compared to those in the dark, indicating that the loading of Ag SAs is beneficial for accelerating hole enrichment and thus facilitate the activation of surrounding lattice oxygen. At the same time, the increase in temperature also promotes the activation of lattice oxygen (Fig. 4j). The O₂-TPD results (Supplementary Fig. 33) show that the peak at 153 °C attributed to the physisorption oxygen (100–200 °C) is observed in all samples, and the peak at 350–450 °C in all samples is attributed to the superficial lattice oxygen (350–600 °C)⁵⁰. However, no obvious peak attributed to chemically adsorbed oxygen (200–350 °C) is observed in all samples. Moreover, for the Ag₁-ZnO sample, the superficial lattice oxygen peak shifts to a lower temperature than that of the other samples, which proves that the loading Ag SAs on ZnO contributes to the mobility lattice oxygen (the capacity of lattice oxygen converting to the surface reactive oxygen species). The desorption peak of surface lattice oxygen on Ag_NP-ZnO locates at relatively higher temperature, which is due to the strong adsorption ability of Ag NPs towards oxygen species^51,52. Then, crystal orbital Hamilton population (COHP) are used to analyze the covalent Zn–OCH₂ interactions over ZnO and Ag₁-ZnO. The positive and negative values in the COHP diagram correspond to bonding and antibonding contributions, respectively⁵³. Compared with the ZnO slab, the integrated COHP (ICOHP) value of the Ag₁-ZnO slab decreases and the Zn–O bond length increases, demonstrating that Ag SAs decrease the Zn–OCH₂ bonding state occupancy (Fig. 4k and Supplementary Fig. 34). Therefore, the heat and the introduction of Ag SAs can promote the activation of lattice oxygen, thereby leading to the formation of HCHO.

To obtain a full image of the possible reaction pathways for the selective oxidation of CH₄ to HCHO over Ag₁-ZnO, DFT calculations were performed. ZnO (101) with an Ag atom (Supplementary Fig. 7c) is used for modeling, and the Gibbs free energy is corrected based on the experimental test temperature. *OH is a key species for the activation of the C–H bond in CH₄ according to the isotope labeling experiment. For CH₄ oxidation on the Ag₁-ZnO system, as shown in Fig. 4a, H₂O is absorbed (reaction energy (ΔE) = − 0.47 eV) and activated to form *OH (ΔE = 0.04 eV), and then the generated *OH interrupts the C–H bond in CH₄ to form *(O)CH₃ species and gas H₂O (ΔE = − 1.17 eV). The calculation results of dissociation C–H in CH₄ over Ag₁-ZnO indicate that *OH prefers to break C–H in CH₄ than that ZnO itself (Supplementary Fig. 35). After that, the *(O)CH₃ can either couple with the adsorbed *H to form *CH₃OH or directly continue dehydrogenation to form HCHO. The ΔE of direct *(O)CH₃ dehydrogenation is −1.83 eV, and the competing pathway of *CH₃OH formation must overcome a barrier (E_a) of 0.61 eV. The *CH₃OH desorption to form CH₃OH still requires a ΔE of 0.73 eV. That is, *(O)CH₃ is more inclined to continue dehydrogenation. Then, the dissociated *(O)CH₂ group was attached on the Ag atom with O–Ag bridge to form *HCHO (E_a = 0.20 eV, ΔE = − 0.54 eV). *HCHO undergoes desorption to generate HCHO (ΔE = 0.60 eV). Finally, O₂ is activated and replenishes the consumed lattice oxygen. After comparing with the simulation results of the HCHO formation process on pure ZnO (Supplementary Fig. 37), it can be concluded that Ag SAs doping facilitates the desorption of HCHO and the replenishment of lattice oxygen, which agrees with the above characterization results. In addition, the Gibbs free energy results at different temperatures also indicate that an increase in reaction temperature can effectively promote the generation of HCHO (Fig. 5a, Supplementary Figs. 36 and 37).

a Calculated Gibbs free energy diagrams for CH₄ oxidation to HCHO or CH₃OH at 501.15 K on Ag₁-ZnO. Proposed reaction mechanism for gas–solid photothermal catalytic selective oxidation of CH₄ to HCHO on Ag₁-ZnO from the energy band structure (b) and the molecular level (c). The inset images in (a) are the optimized configurations of the corresponding reaction intermediates. Blue, red, black, white and dark gray spheres in (a) and (c) represent Zn, O, C, H and Ag atoms, respectively. Source data are provided as a Source Data file.

A reaction mechanism for the gas–solid photothermal catalytic selective oxidation of CH₄ to HCHO on Ag₁-ZnO is proposed in Fig. 5b, c. Under thermal action, photogenerated electrons and holes on ZnO rapidly separate to form Zn⁺O⁻ active sites under irradiation. The holes on O⁻ transfer to Ag SAs and then extract the H atom from CH₄ to form *CH₂ species under the action of the surface *OH species. Then, the activation of lattice oxygen is promoted under the joint action of heat and Ag SAs, thereby facilitating the formation and desorption of *HCHO to form HCHO. Finally, the electrons on Zn⁺ activate O₂ to form O²⁻, replenishing the consumed lattice oxygen.