Near-infrared photon-triggered CH₄-to-CH₃OH conversion over plasmonic oxyselenides

Abstract

The direct oxidation of methane into methanol exploiting O₂ as oxidant offers an ideal route for methane utilization. Although the reaction is strongly preferred in thermodynamics, conventional catalytic systems always demand intense energy input like high temperatures or high-energy photons (>2.8 eV) to conquer the large kinetic barrier in the conversion process. In this study, we demonstrate that by creation of a suitable plasmonic photocatalyst, namely oxygen-vacancy-rich CuSeO_3-x, the low-energy near-infrared (NIR) photons can serve as the sole energy input to complete CH₄-to-CH₃OH conversion with remarkable activity (Apparent quantum yield of 1.5% at 800 nm with Au cocatalyst) and near unity selectivity (ca. 96%) at 25 °C. Such fascinating performance is attributed to a small activation energy measured at 0.28 eV, enabled by the existence of Cu^II-O_v species in CuSeO_3-x. Our study suggests that the ensemble of Cu^II-O_v constitutes an exceptional active site, which can harness the plasmon-induced hot electrons and meantime brings a kinetically advantageous route for reducing O₂ into •OH radicals greatly favoring methane activation.

Introduction

Methane, as an abundant carbon-based energy resource on earth, its activation and conversion into value-added chemicals represents a crucial process in chemical industry regarding the energy and environmental implications^{1,2,3,4,5,6,7,8}. Although numerous research endeavors have been devoted, great challenges still remain because methane is quite inert, bearing a symmetrical tetrahedral geometry with rather low polarizability, high C-H bond strength and negligible electron affinity. As a result, CH₄ conversion always demands an intense energy input, like high temperatures of 600–1100 °C as in conventional thermal catalytic reactions. In the last decade, photocatalysis, which exploits photons to drive chemical transformations, has shown great promise to enable CH₄ conversion at ambient temperatures^{3,4,9,10,11,12,13,14}. Despite avoiding harsh reaction conditions, UV or short-wavelength visible light (<450 nm) corresponding to high-energy photons surpassing ~2.8 eV, is often essential to activate the stubborn CH₄ molecules^{9,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33}.

However, when looking into the thermodynamics associated with CH₄ conversion, for many oxidative reactions, taking selective oxidation to methanol (an important fuel and platform molecule for synthesis) with oxygen as an example (Eq. 1), the Gibbs free energy is considerably decreased after the reaction,

$$2{{{\rm{C}}}}{{{{\rm{H}}}}}_{4}+{{{{\rm{O}}}}}_{2}\to 2{{{\rm{C}}}}{{{{\rm{H}}}}}_{3}{{{\rm{OH}}}} \; \; \Delta {{{{\rm{G}}}}}^{{{{\rm{\theta }}}}}(298{{{\rm{K}}}})=-\!222.6{{{\rm{kJ}}}}\;{{{{\rm{mol}}}}}^{-1}$$

(1)

which signifies that no thermodynamic constraints need to be overcome by means of external impetus. In other words, the requirements on intensive energy supply are mainly to surmount the large activation energy (E_a) in the course of the reaction⁹. If there is a suitable catalyst permitting substantial reduction in E_a, minor energy input, even near-infrared (NIR) photons, are possible to be leveraged to convert inactive CH₄ into desirable chemicals efficiently. To this end, the key is to develop potent photocatalytic materials beyond the scope of commonly studied semiconductors such as TiO₂, ZnO, Ga₂O₃, BiVO₄, WO₃, In₂O₃, etc.^{9,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33}, to bring forth a distinctive kinetic pathway with much lower energetic barriers for processing CH₄ transformation.

Localized surface plasmon resonances (LSPRs) of noble metals, typically Au nanostructures, have demonstrated prominent advantages in harvesting low-energy photons to promote chemical reaction facilitated by the notably reduced E_a^{34,35,36,37,38}. For example, very recently, Halas et al. have reported that a low E_a of 0.42 eV could be realized in steam methane reforming beneficial from the plasmon-mediated hot carrier contributions³⁷. As an appealing alternative to noble metals which are restricted by their rarity and high cost, degenerately doped semiconductors displaying metal-like LSPRs in visible to NIR region have garnered growing research interests^39,40. Up to now, the developed plasmonic semiconductors are primarily composed of two categories: the metal chalcogenides including Cu_2-xS and Cu_2-xSe, as well as the metal oxides like WO_3-x and MoO_3-x^{39,40,41,42,43,44,45,46,47,48,49,50}. There remain far-reaching opportunities for expanding the library of plasmonic semiconductors to create kinetically more favorable candidates towards CH₄ activation and conversion.

Actually, in specific to the oxidation of CH₄ to CH₃OH, nature has already established a paradigm catalyst, the methane monooxygenase (MMO) that well exemplifies the feasibility in accomplishing the conversion with minimal energy consumption⁸. In particular, it has been uncovered that the particulate form of MMO (pMMO) entails a Cu^IICu^IICu^II active site, which accounts for the activation of molecular O₂ in an initiation step to induce subsequent hydroxylation of CH₄ at ambient conditions^8,51,52. Inspired by these findings, we envision that construction of a Cu^II-based active site featuring strong capability for O₂ activation could be highly conducive to actualize CH₄ oxidation through an energy-sparse process.

Herein, we report a new class of plasmonic semiconductor, i.e., CuSeO_3-x, which can strongly absorb the NIR light via LSPRs arising from a high density of free electrons caused by the abundance of oxygen vacancy (O_v). Upon resonant excitation, the ensemble of Cu^II-O_v sites in CuSeO_3-x accumulate plasmon-induced hot electrons, constituting a kinetically favorable center for reducing O₂ into •OH radicals according to density functional theory (DFT) simulation. This promotes the activation of CH₄ into CH₃• radicals and the highly selective generation of CH₃OH with the aid of deposited Au cocatalysts. As such, an exceptionally low E_a of 0.28 eV is realized, and NIR photons can be directly harnessed to trigger the reaction at 25 ^oC, giving an apparent quantum yield (AQY) of 1.5% at 800 nm and 96.2% selectivity for CH₃OH production.

Results and discussion

Structural and photo-physical characterizations

The crystal structure of triclinic CuSeO₃ with space group P1 is illustrated in Fig. 1a. Two inequivalent Cu sites are involved in one individual CuSeO₃ unit cell. In the first site, the Cu atom is bonded to five O atoms forming corner sharing CuO₅ trigonal bipyramids. In the second site, the Cu atom is bonded to four O atoms in a square co-planar geometry. There is no direct bond between Cu and Se atoms. Their linkage is mediated via O, where each Se atom binds with three O atoms to form a trigonal non-coplanar geometry (SeO₃^2-). It should be mentioned that the successful synthesis of triclinic CuSeO₃ had only been reported until 2020⁵³, by Bera and Chakraverty in their research to develop helimagnetic spin structures toward spintronic applications. However, in their work, the obtained CuSeO₃, although light green-colored, was recognized as a Mott insulator with the band gap evaluated at 3.9 eV.

Fig. 1: Characterizations of CuSeO_3-x.

a Illustration of the crystal structure. b UV-Vis-NIR DRS spectra. c O 1 s, d Cu 2p, and e Se 3 d XPS spectra. f EPR spectra. g Mott-Shottky plots. h Correlation between O_v concentration and free electron density in CuSeO_3-x. Source data for b–h are provided as a Source Data file.

In our study, the triclinic CuSeO_3-x crystals featuring intensive LSPRs absorption in NIR region were accessed using a similar method via a hydrothermal process with modifications (for example, sharply reducing the reaction time from 72 h to 6 h)⁵³. CuCl₂•2H₂O and Na₂SeO₃ were adopted as the precursors and their molar ratio was finely tailored from 1:1 to 1:0.97, 1:0.94 and 1:0.91. According to the results of inductively coupled plasma mass spectrometry (ICP-MS), the chemical compositions of thus prepared samples are CuSeO_2.98, CuSeO_2.94, CuSeO_2.90 and CuSeO_2.82, correspondingly (Table S1). As revealed in Fig. S1, XRD patterns of these CuSeO_3-x are well matched with the structural phase of triclinic CuSeO₃ (JCPDS 48-0375) without any impurities. However, continuing varying the molar ratio between Cu and Se precursors would lead to the generation of a second phase. Transmission electron microscopy (TEM) image in Fig. S2 indicates that the formed triclinic CuSeO_3-x is composed of irregular bulk crystals. Very interestingly, it was found that these series samples manifest a bright green color (Fig. S3). Such observation is in good agreement with the UV-visible-NIR diffuse reflectance spectra (DRS) as shown in Fig. 1b. One can see that in addition to the interband transition absorption below ~600 nm, there is a strong and broad band across the long-wavelength visible and NIR range (~700–1500 nm) centered around 830 nm.

It is well acknowledged that copper chalcogenide nanocrystals, such as Cu_2-xSe, exhibit notable NIR absorption arising from LSPRs due to a high density of Cu vacancies^40,46,47,48. In our case, the stoichiometric ratio between Cu and Se determined by ICP-MS demonstrates that the remarkable absorption beyond band edge observed for CuSeO_3-x should not follow the same mechanism as Cu_2-xSe. Alternatively, the obvious deficiency of oxygen atoms in these samples suggests that the existence of anionic oxygen vacancies is probably an important factor. To validate this assumption, X-ray photoelectron spectroscopy (XPS) measurements were carried out over the prepared CuSeO_3-x. From Fig. 1c, the O 1 s spectra can be deconvoluted into three components: the peak at 531.6 eV attributed to oxygen vacancy, and other two peaks at 530.5 and 532.3 eV associated with lattice oxygen and surface adsorbed species (e.g., OH), respectively^49,54,55,56. As denoted by the alteration in peak areas, decreasing the addition amount of Na₂SeO₃ during synthesis resulted in a gradual increase in the concentration of oxygen vacancies, consistent with the conclusion derived from ICP-MS. For Cu 2p spectra (Fig. 1d), the two peaks at 934.3 and 954.2 eV accompanied by evident satellites are characteristic of Cu²⁺. It can be noticed that along with the increment in O_v concentration, the Cu 2p peaks showed a slight shift to lower binding energies indicative of the enhanced electron density around Cu. Furthermore, the Se 3 d bands at 58.8 and 59.6 eV uncovered that in CuSeO_3-x the valence state of Se is +4 (Fig. 1e)^57,58. The presence of oxygen vacancies was further verified by electron paramagnetic resonance (EPR) characterizations. In Fig. 1f, the same variation trend as XPS spectra was observed, supporting that the concentration of oxygen vacancies in CuSeO_3-x can be readily modulated by adjusting the molar ratio between Cu and Se precursors.

Previous investigations have established that a high charge carrier density on the order of 10²¹ cm⁻³ is essential to give LSPRs response in the NIR spectrum^46,59. To determine the carrier density in CuSeO_3-x, Mott-Schottky measurements were conducted on the basis of Eq. (2):

$$\frac{1}{{C}^{2}}=\frac{2}{e{A}^{2}{\varepsilon }_{0}{\varepsilon }_{A}N}\left(V-{V}_{{fb}}\right)$$

(2)

where C is the charge layer capacitance, e (1.602 × 10⁻¹⁹C) is the elementary charge, A (1 cm² in our case) is the surface area, ε₀ (8.854 × 10⁻¹⁴ F cm⁻¹) is the electrical permittivity in vacuum, ε_r is the relative dielectric constant for CuSeO_3-x (measured to be 6.05 through four probe method, Table S2), N is the charge carrier density, V is the applied potential and V_fb is the flat-band potential^60,61,62. As displayed in Fig. 1h, the positive slopes of Mott-Schottky plots indicate the n-type conductivity of CuSeO_3-x with electrons as the predominant charge carriers_. This outcome is further corroborated by the negative Hall coefficients measured for the materials (Table S3)^63,64. Table S4 lists the calculated electron densities for different CuSeO_3-x in accordance with Eq. 2. It can be seen that the values are varied from 1.0 × 10²¹ to 2.3 × 10²¹cm⁻³ as a function of O_v concentration, indeed capable of causing LSPRs in the NIR domain. These results, together with the observed linear correlation between the amounts of oxygen vacancies and free carriers in CuSeO_3-x (Fig. 1h), substantiate that the strong absorption feature ranging from the red edge of visible to NIR observed for CuSeO_3-x, should originate from LSPRs that are induced by the nonstoichiometric oxygen deficiency.

As illustrated in Fig. 2a, in the light of the different bonding patterns and bond lengths of Cu-O, the oxygen vacancies in CuSeO_3-x could be categorized into four different types: O_v1-1 (1.92 Å and 2.00 Å, three-coordinate), O_v1-2 (2.00 Å and 2.3 Å, three-coordinate), O_v1-3 (2.00 Å and 2.00 Å, three-coordinate), and O_v2 (1.92 Å, two-coordinate). Synchrotron radiation-based X-ray absorption fine structure (XAFS) spectroscopy was employed to distinguish which type of oxygen vacancies dominates in CuSeO_3-x. The obtained X-ray absorption near edge structure (XANES) spectra at Cu K edge are presented in Fig. 2b, where the similar intensities of white lines for CuSeO_2.98 and CuO reveal that the Cu species in CuSeO_2.98 are primarily in the form of Cu²⁺. Differently, the white line intensity for CuSeO_2.82 is slightly lower than CuO, suggesting an enhanced electron density relative to Cu²⁺, in parallel with the findings acquired from XPS as above discussed (Fig. 1d). Figure 2c shows the Fourier transform (FT) of extended X-ray absorption fine structure (EXAFS) spectra for the samples, from which one can see that both CuSeO_2.98 and CuSeO_2.82 display only one major peak at 1.57 Å ascribed to Cu-O coordination. More importantly, the specific Cu-O bond length and Cu coordination numbers were extracted from the R space fitting curves of EXAFS (Fig. 2d and Table 1). It was uncovered that compared to CuSeO_2.98 with less oxygen vacancies, CuSeO_2.82 containing higher density of oxygen vacancies exhibit a marked decrease in the coordination number (from 3.3 to 2.8) exclusively at Cu-O with a bond length of 2.0 Å. Such information enabled us to unravel that the oxygen vacancies in CuSeO_3-x are mainly located at the O_v1-3 site. The wavelet transform (WT) analysis of EXAFS signals was additionally performed, and similar conclusions are reached (Fig. 2e).

Fig. 2: Investigation on the local structure of O_v in CuSeO_3-x.

a Structural models showing different coordination environments for O_v in CuSeO_3-x (Blue spheres represent Cu atoms; green spheres represent Se atoms; red spheres represent O atoms). b Normalized Cu K-edge XANES spectra, c k³-weighted FT-EXAFS spectra, d EXAFS fitting curves in R space and e WT-EXAFS contour maps. Source data for Fig. 2b–d are provided as a Source Data file.

Table 1 Fitting results of Cu K-edge EXAFS Data

It has been reported that some typical oxide plasmonic semiconductors like WO_3-x often possess poor stability, as their oxygen vacancies are easily eliminated in O₂-containing environment causing the loss of LSPRs effect^50,65,66. Therefore, after resolving the origin of LSPRs and the local structure of oxygen vacancies in CuSeO_3-x, we turned to evaluate the stability of O_v and the optical absorption properties for CuSeO_3-x after light irradiation in air. In situ DRS and EPR measurements disclosed that after constant exposure to a Xe lamp (350 mW cm⁻²) for 24 h, no significant changes occurred to CuSeO_2.82, except for a minor enhancement in the LSPRs band accompanied by a slight increment in the O_v concentration (Fig. S4). According to DFT calculations, the O_v formation energies in CuSeO₃ and CuSeO_2.82 are 1.49 and 1.89 eV, respectively, which are considerably lower than those in WO₃ (4.45 eV) and other widely studied oxide photocatalysts (2.36 - 5.7 eV, details see Table S5), explaining the excellent O_v stability observed in our study. Furthermore, we found that by treating the as-prepared CuSeO_2.82 with a mixture of H₂/Ar gases at 300 ^oC, an extra amount of oxygen vacancies can be incorporated to give more pronounced LSPRs absorption with well retained crystallographic phases (Fig. S5, Table S1, the resulting material is denoted as CuSeO_2.71-H₂).

Photocatalytic methane oxidation with O₂

The photocatalytic methane oxidation reaction was performed using molecular oxygen as the oxidant in a watersolution under ambient temperature (Fig. S6a). In specific, 20 mg of the CuSeO_3-x photocatalysts were suspended in 20 mL of distilled water at 25 °C (temperature controlled by a circulating water bath) with CH₄/O₂ (v/v = 2/1, 0.1 MPa) as the feed gas under irradiation of a Xe lamp (350 mW cm⁻²). After 2 h of reaction, the bare CuSeO_3-x material (in specific CuSeO_2.90) showed near one unity selectivity in producing CH₃OH (Fig. 3a), with no other products detected except for a negligible amount of CO₂. To further improve the conversion efficiency, a series of cocatalyst particles were loaded including CuO_x, FeO_x, Ag, Au and Pt (Fig. S7). The results showed that Au/CuSeO_2.90 afforded the highest CH₃OH yield of 4026 μmol g⁻¹ and a selectivity of 93.7% (Fig. 3a). It is noteworthy that for FeO_x/CuSeO_2.90 entirely composed of earth abundant elements, the CH₃OH yield reached 1920 μmol g⁻¹ (90.6% selectivity) comparable to those of Ag/CuSeO_2.90 and Pt/CuSeO_2.90.

Fig. 3: Photocatalytic performance in CH₄ oxidation to CH₃OH with O₂.

Influnences of (a) different cocatalysts and (b) different O_v concentraions. The error bars indicate standard deviation. c GC-MS results of the isotope labeling experiments for Au/CuSeO_2.82 in the presence of ¹³CH₄ or ¹²CH₄. d Impacts caused by the pressure of reaction gas for Au/CuSeO_2.82. e Wavelength-dependent AQY for Au/CuSeO_2.82 under ambient and elevated pressures. f AQY and selectivity at different wavelengths for CH₄-to-CH₃OH conversion over Au/CuSeO_2.82 in comparision with the state of the art (details see Table S8). g Arrhenius plots for Au/CuSeO_2.82 and CuSeO_2.82. h Recyclability measurement for Au/CuSeO_2.82. Source data for a–e, g, h are provided as a Source Data file.

After screening the cocatalysts, the influences of O_v concentration on photocatalytic performance were assayed. According to Fig. 3b, the optimal concentration of oxygen vacancies was given by CuSeO_2.82. With Au cocatalsyts (Au particle size centered at 10.2 nm, Fig. S8), this sample exhibited a CH₃OH yield of 4373 μmol g⁻¹ with 93.4% selectivity. Further increasing the O_v concentration led to a depressed CH₃OH yield, as demonstrated by Au/CuSeO_2.71-H₂ (likely containing an excessive amount of O_v thereby enhancing the probability of charge recombination, Fig. S9). Besides, the impacts of molar ratio between CH₄ and O₂ as well as the amount of H₂O were investigated (Fig. S10). The results showed that under optimal condition (CH₄/O₂ = 2/1, 0.1 MPa, 40 mL of H₂O) the CH₃OH yield over Au/CuSeO_2.82 can be further increased to 4780 μmol g⁻¹ (corresponding to an averaged formation rate of 2.4 mmol g⁻¹ h⁻¹), while the selectivity was nearly unchanged at 95.3% with CH₃OH as the sole liquid oxygenate product. To ensure that the detected CH₃OH was generated from CH₄ oxidation, control experiments were carried out, where no products were generated with the absence of photocatalyst, light, CH₄ or O₂ (Fig. S13). Furthermore, ¹³C isotope labeling experiments using ¹³CH₄ as reactant additionally confirmed the carbon source for CH₃OH collected after the photocatalytic reaction (Fig. 3c). Besides, carbon balance for the reaction system was determined to be 93% (Table S6), strongly confirming the absence of other side products.

When raising the pressure of CH₄ to enhance its solubility in water, the CH₃OH yield over Au/CuSeO_2.82 can be further improved to 6440 μmol g⁻¹ under 2.0 MPa with little perturbed selectivity (Fig. 3d, Fig. S6b). Regarding this sample, the AQY for CH₃OH generation was measured at different excitation wavelengths under atmospheric (0.1 MPa) or elevated pressure (2.0 MPa) with temperature fixed at 25 °C (Fig. S14, Table S7). As displayed in Fig. 3e, under both conditions the action spectra for CH₄-to-CH₃OH conversion basically follows the optical absorption of the material across the UV-Vis-NIR range. At ambient pressure, the AQY was determined to be 1.1% (92.7% selectivity) at 405 nm, and this value was improved to 2.7% (94.6% selectivity) when pressure was increased to 2.0 MPa. Remarkably, the results also showed that with light irradiation at longer wavelengths, for example λ = 800 nm, the AQY was of 0.7% (93.4% selectivity) at 0.1 MPa while reached 1.5% (96.2% selectivity) at 2.0 MPa. Figure 3f summarizes the results of AQY and selectivity with respect to CH₃OH production reported for the state-of-the-art photocatalysts enabling CH₄ partial oxidation by O₂ (25 °C, 0.8–3.0 MPa, details see Table S8)^{10,19,20,21,22,23,24,25,26,67,68,69,70}. It is worth underlining that although significant advances have been achieved, effective CH₄ conversion into C₁ oxygenates driven by photons with wavelengths exceeding 450 nm is still very challenging and has rarely been realized in previous studies^{9,10,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,67,68,69,70,71,72,73}. Here, our results present a stride in this scope, where NIR light can be harnessed to activate CH₄ into CH₃OH with exceptional activity and selectivity. This gives an unambiguous demonstration that the long-wavelength photons are well capable of overcoming the energy barrier in CH₄ oxidation under mild conditions, mapping an attractive route for improving the utilization efficiency of solar energy to transform CH₄ into value-added chemicals.

The apparent activation energy for photocatalytic CH₄-to-CH₃OH conversion was then estimated by fitting the measured temperature-dependent reaction rates with Arrhenius equation, in the temperature range of 0 to 35 °C. The E_a values were determined to be 26.6 kJ mol⁻¹ (0.28 eV) and 34.4 kJ mol⁻¹ (0.36 eV) for Au/CuSeO_2.82 and CuSeO_2.82, respectively, under illumination (Fig. 3g, E_a in the dark cannot be acquired due to the lack of activity). Such low E_a results suggest that the plasmon-mediated hot carriers might greatly assist the activation process of CH₄, as it has been demonstrated that illumination of a plasmonic photocatalyst can dramatically reduce the E_a relative to the thermal case^34,35,36,37. In view of the much higher AQY at 800 nm than that at 550 nm (near the LSPR band of Au nanoparticle) for Au/CuSeO_2.82 (Fig. 3e), as well as the poor activity observed for the bare Au nanoparticles (CH₃OH formation rate of 24 μmol g⁻¹ h⁻¹ under 2.0 MPa, details see Fig. S15), we reasonably assume that the LSPRs of CuSeO_2.82 should make a major contribution in modifying the reaction kinetics. In addition to its excellent catalytic performance, Au/CuSeO_2.82 also demonstrated high recyclability, retaining 92% of its initial catalytic activity after five consecutive reaction cycles (10 h total irradiation, 2.0 MPa, Fig. 3h). The samples collected after photocatalytic reaction exhibited good structural and compositional stability as indicated by combined XRD, XPS and EPR characterizations (Fig. S16).

Mechanism investigation

With regard to plasmonic photocatalysis, the hot carrier-promoted chemical conversion might be entangled by thermal effects. To distinguish the thermal and nonthermal contributions, the surface temperature of the photocatalyst powder was measured first using a thermal imaging camera (Fig. S17a). It was found that the photothermal heating of Au/CuSeO_2.82 was not very significant, with the surface temperature reaching a plateau at 88 °C after light illumination (350 mW cm⁻²) for 300 s. Moreover, when dispersing Au/CuSeO_2.82 into the aqueous reaction solution, the temperature of the system can only be raised to 42 °C by light illumination (Fig. S17b). At this reaction temperature, the yield of CH₃OH was evidently depressed whilst the production of HCHO and CO₂ was promoted in comparison to those at 25 °C (Fig. S18). These outcomes manifest that in our experiments (reaction temperature controlled at 25 °C by circulating water bath) the impacts of photothermal heating on catalytic performance were negligible.

Next, the influences of oxygen vacancies on photogenerated charge separation properties were systematically explored. In Fig. 4a, upon illumination the open-circuit potential (OCP) of the CuSeO_3-x samples all presented a prompt negative shift, in agreement with the n-type conductivity as above discussed^61,74. The saturated OCP response denotes the steady state density of light-excited carriers as a result of the competition between their production and recombination rates. One can see that this value increased in the order of CuSeO_2.98 < CuSeO_2.94 < CuSeO_2.90 < CuSeO_2.82, clearly indicating that the increase in O_v concentration in these samples can lower the rates of photogenerated carrier recombination. The photocurrent as well as electrochemical impedance spectroscopy (EIS) measurements also supported that the optimal O_v density in CuSeO_3-x are propitious for promoting the charge separation and transfer efficiencies (Fig. S9).

Fig. 4: Identification of the active sites and reactive species.

a OCP response under illumination and in the dark. In situ high-resolution (b) Au 4 f and (c) Cu 2p XPS spectra, as well as (d) in situ EPR spectra for Au/CuSeO_2.82 in the dark and under light irradiation. In situ EPR spectra of (e) DMPO-OH and (f) DMPO-OOH over different photocatalysts under light illumination. g Fluorescence spectra of the coumarin aqueous solution dispersed with Au/CuSeO_2.82 under light irradiation in O₂ or Ar atmosphere. Insets: the detection mechanism of •OH by coumarin. h In situ EPR spectra of DMPO-OH and DMPO-CH₃ over Au/CuSeO_2.82 under light illumination in the presence of sacrificial agent or not. TEOA is used for consuming photogenerated holes, and IPA is used for consuming photogenerated holes and •OH radicals simultaneously. i Illustration of the band structure determined for Au/CuSeO_2.82 and the redox potentials for different radicals. Source data for a–h are provided as a Source Data file.

In situ high resolution XPS was exploited to determine the charge transfer direction and chemical nature on the surface of Au/CuSeO_2.82 under light illumination. As shown in Fig. 4b, the binding energy of Au 4 f shifted positively from 84.3 to 84.6 eV upon irradiation, disclosing that the Au cocatalysts were able to accumulate photogenerated holes. On the contrary, the Cu 2p_3/2 and Cu 2p_1/2 signals presented notable negative shifts (to 932.2 and 952.1 eV, respectively) after illumination, along with evident decrease in the intensity of satellite peaks (Fig. 4c). These observations can be rationalized by the reduction of Cu²⁺ under light, which disclosed that the Cu²⁺ ions in CuSeO_3-x played the role of electron acceptors. A similar conclusion can be derived from the in situ EPR spectra exhibited in Fig. 4d. The EPR signals at g_⊥= 2.054 and g_∥= 2.298 attributed to Cu²⁺ gradually decreased as a function of illumination time, indicative of the conversion of Cu²⁺ into EPR-silent species²⁰. Such process accordingly enhanced the amount of oxygen vacancies nearby (g = 2.003) as can be discerned from the inset of Fig. 4d.

Moreover, by means of in situ EPR using 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) as radical-trapping agent, we uncovered that reactive •OH and •OOH radicals were generated over Au/CuSeO_3-x photocatalysts in the presence of O₂ and H₂O under irradiation^20,24,33. Figure 4e, f clearly show that the signal intensities for both of these two species were positively correlated with the density of oxygen vacancies in the materials, with Au/CuSeO_2.82 offering the highest capability for delivering •OH and •OOH. On account of these results as well as the trapping of electrons at the Cu²⁺ sites, we propose that the Cu²⁺ and neighboring oxygen vacancy (Cu^II-O_v) may cooperatively compose an active center for collecting photoexcited electrons and reducing O₂ into •OH and •OOH. However, it should be taken into mind that in an environment containing O₂ and H₂O, water oxidation provides an alternative path to generation of •OH and •OOH radicals. To identify the origin of these radicals in our reaction, fluorescence measurements with coumarin as probe molecules were performed and the results are given in Fig. 4g^22,26. One can see that for Au/CuSeO_2.82 dispersed in H₂O under Ar atmosphere, no fluorescence signal of 7-hydroxycoumain (OH-coumarin) can be detected after light illumination. In contrast, the introduction of O₂ into the system led to a pronounced signal attributed to OH-coumarin. The sharp discrepancy in the presence of O₂ or not strongly substantiated that in our experiments the •OH (and thus •OOH) radicals were dominantly produced via O₂ reduction other than H₂O oxidation. This conclusion was further corroborated by in situ EPR measurements exhibited in Fig. S19, where substitution of O₂ with Ar in the aqueous suspension of Au/CuSeO_2.82 resulted in disappearance of both DMPO-OH and DMPO-OOH signals under light irradiation. Aside from these reactive oxygen species, the generation of H₂O₂ during O₂ reduction process was also confirmed by the iodometric titration method (details see Fig. S20). Combining the information gleaned above, we propose that the photoreduction of O₂ over Au/CuSeO_2.82 should go through a stepwise process via O₂ → •OOH → H₂O₂ → •OH^25,28.

As regards the oxidation of CH₄, we found that in the photocatalytic reaction system, very strong EPR signals corresponding to CH₃• appeared along with •OH upon light illumination, and both signals intensified with increasing irradiation time (Fig. S21)⁶⁸. More significantly, Fig. 4h shows that when introducing triethanolamine (TEOA) into the reaction system to sacrifice photogenerated holes, the intensity of the EPR signals arising from CH₃• was slightly decreased. However, the addition of isopropyl alcohol (IPA), which is able to consume the •OH radicals and holes simultaneously, almost completely ceased the generation of CH₃•. According to these results, we inferred that in our case the in situ generated •OH should function as the primary reactive species that promote CH₄ activation.

On the basis of Tauc plot, valence band XPS (VB-XPS) spectrum and Mott-Schottky measurements (Fig. S22), the conduction band (CB) and valence band (VB) of Au/CuSeO_2.82 were evaluated at 0.34 and 2.35 V vs NHE, respectively (Fig. 4i). Considering the oxidation potential of •OH/H₂O at 2.38 V vs NHE, the VB position of the material cannot provide a large driving force for H₂O oxidation to •OH. Similarly, the activation of CH₄ straightforwardly by photogenerated holes in VB did not possess much advantage in thermodynamics (•CH₃/CH₄ = 2.32 V vs NHE). In sharp contrast, the surface plasmons of excited Au/CuSeO_2.82 with energies 1.49 eV (corresponding to the resonant excitation at 830 nm) above the Fermi level, can afford substantial driving force for reducing O₂ and the involved intermediates as illustrated in Fig. 4i.

To gain deeper insights into the mechanisms underlying O₂ activation and conversion during photocatalytic CH₄ oxidation process, DFT calculations were implemented on CuSeO₃ (111) facet with or without oxygen vacancies (Fig. 5a–e). First, the O₂ adsorption behavior was examined on the two different surfaces. The results unraveled that the incorporation of oxygen vacancies can markedly promote O₂ adsorption, in terms of the enhanced adsorption energy (−0.26 vs −0.11 eV), shortened spatial distance (2.00 vs 2.57 Å) and increased O-O bond length (1.27 vs 1.24 Å, Fig. 5a and Fig. S23). Noteworthily, the O₂ molecules are preferred to be adsorbed at the Cu atom directly adjoining the oxygen vacancy, supporting our assumption that the Cu^II-O_V associative sites are responsible for the activation of O₂ molecules. On the contrary, the presence of oxygen vacancies brings about negligible impacts on the adsorption capacity for CH₄ (Fig. S24). Thereafter, the thermodynamics associated with O₂ reduction during the reaction via proton-coupled electron transfer (PCET) pathway was estimated. As displayed in Fig. 5e, after the adsorption of O₂, both CuSeO₃ and CuSeO_3-x undergo similar conversions through O₂ → *OOH → *H₂O₂ → *OH, in good agreement with our experimental observations discussed earlier. A distinct advantage is that these entire processes are thermodynamically spontaneous with each steps exhibiting a notable decrease in Gibbs free energy, indicative of the exceptional capability for activating O₂ into reactive *OH species. Apparently, compared to the O_v-free CuSeO₃, CuSeO_3-x exhibits even larger driving force for the sequential transformation in view of the steeper energy declines (for example, −0.7 vs −1.47 eV in *H₂O₂ dissociation step). From structural analysis in Fig. 5a–d, this can be attributed to the Cu^II-O_v sites which function as the reaction center to stabilize lower energy states. Such kinetically favorable production of *OH is strongly desired for CH₄ activation and the subsequent coupling reaction to yield CH₃OH in high selectivity. Beyond efficient O₂ to *OH conversion, CuSeO_3-x also demonstrates its superiority in suppressing over-oxidation according to DFT calculations. As manifested in Fig. S25, the substantial energy difference between *CH₃OH desorption and further oxidation (0.58 vs 1.54 eV) makes the generation of higher oxidation products evidently thermodynamically unfavorable.

Fig. 5: Evaluation of reaction pathway.

a–d Illustration of oxygen adsorption and reduction configurations at the Cu^II-O_v active sites in CuSeO_3-x (Blue spheres represent Cu atoms; green spheres represent Se atoms; red spheres represent O atoms; white spheres represent H atoms). All geometry-optimized computational models and corresponding atomic coordinates originate from the file Supplementary Data 1. e Comparison of the calculated energy profiles on the surface of CuSeO_3-x and CuSeO₃ for O₂ adsorption and activiation. f In situ DRIFTS spectra for Au/CuSeO_2.82 during the photocatalytic methane oxidation reaction. Source data for e, f are provided as a Source Data file.

Furthermore, in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) was carried out over Au/CuSeO_2.82 in the presence of CH₄, O₂ and water vapor (Fig. 5f). With constant light irradiation, an evident decrease at 3018 cm⁻¹ ascribed to the antisymmetric stretching of CH₄ was observed^75,76,77, confirming the effective consumption of CH₄ over the photocatalyst. More importantly, several new bands emerged and grew continuously during the irradiation process. In specific, the peak at 1654 cm⁻¹ was probably stemmed from the bending signal of OH*^27,71,76,77, while those at 1412 and 1339 cm⁻¹ could be assigned to the scissoring and symmetric rocking vibrations of CH₃*, respectively^27,73,75,77. These observations are well aligned with the overall picture proposed for photocatalytic CH₄ oxidation on surface of Au/CuSeO_3-x as illustrated in Fig. S26: under light irradiation, LSPRs of CuSeO_3-x are excited. The generated hot electrons are trapped by Cu^II-O_v sites to reduce O₂ into •OH via multiple PCET steps. In the meantime, the photogenerated holes accumulate on Au cocatalysts, generating positively charged Au^δ+ sites where the CH₄ molecules could be activated into CH₃• radicals by in situ produced •OH^17,78. The combination of CH₃• and •OH then yields CH₃OH molecules, which can be easily dissolved in water avoiding overoxidation on the photocatalyst surface.

In summary, a new type of plasmonic semiconductor has been created by manipulating the oxygen vacancies in copper oxyselenide. The resultant CuSeO_3-x can efficiently capture NIR photons to accomplish CH₄-to-CH₃OH transformation at ambient temperature with O₂ as the oxidant. Such fascinating performance is strongly associated with the Cu^II-O_v sites on surface of CuSeO_3-x, which enable a kinetically feasible route to reduce O₂ into •OH radicals, markedly lowering the energetic barrier for CH₄ activation and oxidation. This work showcased the first demonstration that low-energy NIR photons can be utilized as the sole stimulus to drive CH₄-to-CH₃OH conversion, which conventionally suffers from excessive energy depletion due to high activation energies. In addition, our study opened up new possibilities in expanding the currently limited pool of plasmonic semiconductors, to develop superior photocatalysts in response to the worldwide increasing demand of sustainable chemical feedstock and fuels.

Methods

Materials

Copper (II) chloride dihydrate 2H₂O (CuCl₂·2H₂O, 99.99%), chloroauric acid hydrate (HAuCl₄, Au ≥ 47.5%), silver nitrate (AgNO₃, 99.8%), chloroplatinic (IV) acid (H₂PtCl₆, 40% Pt), and anhydrous sodium sulfate (Na₂SO₄, 99%) were purchased from Aladdin Reagent Co., Ltd. Iron (III) nitrate nonahydrate (Fe(NO₃)₃·9H₂O, 98.5%) was obtained from Sinopharm Group Chemical Reagent Co., Ltd. Sodium selenite (Na₂SeO₃) was purchased from Alfa Aesar Reagents Co., Ltd. Other chemicals, including anhydrous ethanol, were purchased from Beijing Tong Guang Fine Chemicals Company. All reagents were used without further purification. Deionized water was prepared using a Merck Direct-Q3 pure water system (18.2 Ω).

Preparation of CuSeO_3-x crystals

The CuSeO_3-x crystals were prepared according to a reported hydrothermal method⁵³, but with modifications to reaction times. The procedure started with dissolving CuCl₂·2H₂O and Na₂SeO₃ (molar ratio adjusted in the range of 1:1, 1:0.97, 1:0.94, and 1:0.91, corresponding to CuSeO_2.98, CuSeO_2.94, CuSeO_2.90 and CuSeO_2.82, respectively) into 60 mL of deionized water within a PTFE-lined autoclave. The mixture was stirred for 5 min under ambient condition, and then the autoclave was sealed and heated to 180 °C in an oven. The reaction proceeded at this temperature for 6 h, after which the product was collected via centrifugation and washed multiple times with a water-ethanol mixture. The final product was thoroughly rinsed with deionized water and dried under vacuum at 60 °C overnight. To achieve even higher oxygen vacancy concentrations, the as-prepared sample (i.e., CuSeO_2.82) was further subjected to a hydrogen treatment under high temperatures. Specifically, 100 mg of the synthesized material was placed in a 3% H₂/Ar atmosphere and heated at 300 °C for 2 h (heating rate: 3 °C min⁻¹). The resulting product was rinsed with deionized water multiple times and labeled as CuSeO_2.72-H₂.

Preparation of cocatalyst-modified CuSeO_3-x

Photo-deposition method was exploited to tailor Au, Ag, Pt and CuO_x cocatalysts on surface of CuSeO_3-x. In a typical synthesis, the as-prepared CuSeO_3-x crystals were thoroughly grounded and dispersed in 50 mL of deionized water with ultrasonication. The reaction system was sealed and purged with high-purity argon gas to completely remove air. Then, 500 μL of HAuCl₄ (13 mg mL⁻¹), or H₂PtCl₆ (15 mg mL⁻¹), AgNO₃ (10 mg mL⁻¹), CuCl₂ (16 mg mL⁻¹) aqueous solution was added, followed by irradiation with a 300 W xenon lamp (PLS-SXE300E, Beijing Perfectlight Technology Co., Ltd) from the top for 0.5 to 1 h. After the photo-deposition procedure, the sample was collected by centrifugation, washed multiple times with deionized water, and dried under vacuum at 60 °C overnight. To prepare FeO_x cocatalyst-modified CuSeO_3-x, 50 mg of CuSeO_3-x crystals were thoroughly grounded and dispersed in 5 mL of deionized water. 11 mg of Fe(NO₃)₃•9H₂O was added and mixed evenly. The resultant mixture was dried at 60 °C using a water bath, then transferred to a tube furnace and calcined in air at 300 °C for 3 h (heating rate: 5 °C min⁻¹). Finally, the sample was collected by centrifugation, washed multiple times with deionized water, and dried under vacuum at 60 °C overnight.

Characterization

Low-resolution transmission electron microscopy (TEM) images were obtained using a HITACHI H-7650 electron microscope with an accelerating voltage of 80 kV. High-resolution transmission electron microscopy (HRTEM) images were performed on a FEI Tecnai G2 F20 S-Twin microscope operating at 200 kV. The dielectric constant (ε_r) was measured using a parallel plate capacitor with the four-probe method, following the standard GB/T 5594.4-2015. X-ray diffraction (XRD) patterns were acquired using a Bruker D8 X-ray diffractometer (5° min⁻¹). UV-visible-NIR diffuse reflectance spectra (DRS) spectra were recorded on a Shimadzu UV-3600i plus spectrophotometer at room temperature. X-ray photoelectron spectroscopy (XPS) spectra were obtained using a PerkinElmer PHI 5300 spectrometer with monochromatic Al Kα radiation (1486.7 eV). All binding energies were corrected based on the C 1 s peak (284.8 eV). In situ solid-state (electron paramagnetic resonance) EPR curves were collected on the ELEXSYS II EPR instrument. The X-ray absorption fine structure spectra data (XAFS, Cu K-edge) were measured at the BL14W1 beamlines at Shanghai Synchrotron Radiation Facility (SSRF) (Shanghai, China). The XAFS data of all the samples were measured at room temperature in transmission mode. All samples were prepared as plates of 10 mm diameter with 1 mm thickness with the addition of graphite powder as a binder.

Photocatalytic methane oxidation with O₂

Photocatalytic methane oxidation reactions were all performed at 25 °C (unless otherwise specified) under irradiation of a 300 W xenon lamp (PLS-SXE300E, Beijing Perfectlight Technology Co., Ltd., 350 mW cm⁻², ~325–1100 nm). For photocatalytic reactions under ambient pressure, typically, 20 mg of photocatalyst was dispersed in 40 mL of distilled water within a glass reactor having a quartz lid (Fig. S6a). After sealing, the suspension was purged with ultra-pure argon (99.999 vol.%) for 20 min to completely remove air. Then, 12 mL of O₂ and 24 mL of CH₄ (0.1 MPa) were introduced into the reactor. The reaction was conducted for 2 h using a circulating water cooling system to maintain the temperature at 25 °C. On the other hand, photocatalytic methane oxidation under high pressures was conducted in a 100 mL high-pressure autoclave with top irradiation (Fig. S6b). Typically, 20 mg of photocatalyst was dispersed in 40 mL of distilled water. After sealing the autoclave reactor, the suspension was purged with ultra-pure O₂ (99.999 vol.%) for 20 min to completely remove air. Then, 0.1 MPa of O₂ was maintained in the reactor, and methane (99.999 vol.%) was injected to achieve a total pressure varied from 0.5 to 2.0 MPa. The reaction was conducted for 2 h using a circulating water cooling system to maintain the temperature at 25 °C. The recycling tests were performed under 2.0 MPa using 50 mg of photocatalyst. After each run, the catalyst was recovered by centrifugation, washed several times with deionized water, and dried at 60 °C before reuse in the next batch.

Photocatalytic products analysis

After photocatalytic reaction, the gaseous products and methanol in the reactor were measured using a gas chromatograph (GC-2014C, Shimadzu Co., Ltd) equipped with a thermal conductivity detector (TCD) and a flame ionization detector (FID). The analysis of CH₃OOH and HCOOH was conducted by means of ¹H nuclear magnetic resonance (NMR), which additionally confirmed the existence of CH₃OH in the system after reaction (Fig. S11). The quantification of aqueous HCHO was performed using the acetylacetone chromogenic method. Specifically, 2 mL aliquots of the sample solution were combined with 2 mL of a freshly prepared 0.25% (v/v) acetylacetone reagent solution. The mixture was subjected to a thermal treatment at 60 °C using a water bath for 5 min to facilitate color development. The absorbance values at 413 nm were measured using spectrophotometric analysis, and formaldehyde concentrations were determined by interpolation against the pre-established calibration curve (Fig. S12).

Carbon source investigation

For the ¹³CH₄ isotope labeling test: 50 mg of Au/CuSeO_2.82 was dispersed in 10 mL of deionized water within the autoclave reactor. The system was sealed and degassed for 30 min to thoroughly remove oxygen, after which 0.1 MPa of O₂ was introduced, together with 0.05 MPa of ¹³CH₄ (99 vol.%) or ¹²CH₄. The reaction proceeded under the irradiation of a xenon lamp for 3 h at 25 °C. The resultant products were analyzed using a headspace gas chromatography/mass spectrometry (HS-GC-MS, Shimadzu GC-MS QP2010 Ultra equipped with a Swiss CTC PAL RTC auto sampler).

Carbon balance determination

Au/CuSeO_2.82 photocatalysts were dispersed in 20 mL of distilled water within a glass reactor having a quartz lid. After sealing, the suspension was purged with ultra-pure argon (99.999 vol.%) for 20 min to completely remove air. Then, 2 mL of O₂ and 5 mL of CH₄ were introduced into the reactor. The reaction was conducted for 2 h using a circulating water cooling system to maintain the temperature at 25 °C. The carbon balance was calculated using the following equation:

$${Carbon\; balance}\left(\%\right)=\frac{{n\left({product}\right)}_{{all}}}{n\left({coverted}C{H}_{4}\right)}$$

(3)

AQY measurements

AQY was measured for the Au/CuSeO_2.82 photocatalyst under irradiation at 400 nm, 405 nm, 550 nm, 700 nm, and 800 nm, using a xenon lamp equipped with different band-pass filters as the light source (Fig. S14). The light intensity was measured with a light intensity meter (PL-MW2000, Beijing Perfectlight Technology Co., Ltd). Since the formation of CH₃OH requires three photogenerated charges, AQY was calculated using the following formula,

$${AQY}=\frac{{numeber\; of\; used\; photons}}{{number\; of\; incident\; photons}}=\frac{n\left(C{H}_{3}{OH}\right)\times 3\times {N}_{A}}{I\times A\times t\times \frac{\lambda }{h\times c}}$$

(4)

where n(CH₃OH) represents the number of CH₃OH molecules formed, N_A represents Avogadro constant, A denotes irradiation area (cm²), t is the irradiation time (3600 s), I stands for the illumination intensity (W cm⁻²), λ gives the wavelength of the irradiation light (m), h and c represent the Planck constant (6.63 × 10^–34 J s) and the velocity of light (3.00 × 10⁸m s⁻¹), respectively.

E_a measurements

E_a measurements were carried out in the high-pressure autoclave reactor under photocatalytic conditions identical to those described above, except for varying the reaction temperature at 0 °C, 15 °C, 25 °C and 35 °C, respectively, using the recirculating water cooling system. The E_a values were calculated using the Arrhenius equation:

$${In}r=-\frac{{Ea}}{{RT}}+{In}A$$

(5)

where r is the reaction rate constant, A is the Arrhenius pre-exponential factor, E_a is the activation energy, R is the universal gas constant, and T is the surface temperature of the photocatalysts.

H₂O₂ concentration quantification

The formation of H₂O₂ during O₂ reduction process was investigated by the iodometric titration method. Typically, 50 mg of Au/CuSeO_2.82 was dispersed in 30 mL of water and loaded into a glass reactor equipped with a quartz cover. After sealing, the system was continuously purged with high-purity oxygen (99.999%) to remove residual air and maintain oxygen pressure at 0.1 MPa. The reaction proceeded for 2 h under irradiation of a 300 W xenon lamp at 25 °C controlled by a circulating cooling water system. After reaction, the solution was collected through a 0.22 μm syringe filter to remove residual photocatalyst prior to iodometric titration. A standard H₂O₂ detection solution was prepared by dissolving 0.045 g of ammonium molybdate tetrahydrate [(NH₄)₂MoO₄•4H₂O] and 2.429 g of potassium iodide (KI) in 150 mL of water under continuous stirring⁷⁹. Subsequently, 2 mL of the standard solution was mixed with 2 mL of the filtered post-reaction solution, forming a yellow dispersion that was allowed to stand for 10 min. The absorbance of the iodine complex at 352 nm was measured using a UV-Vis spectrophotometer, and the H₂O₂ concentration was determined using a pre-established calibration curve constructed from the intensity of the UV-Vis absorption peak at this wavelength.

Photoelectrochemical measurements

The working electrode was prepared by uniformly coating a mixture of photocatalyst, isopropyl alcohol (IPA), and Nafion solution onto ITO conductive glass. Ag/AgCl electrode and platinum plate was used as the reference electrode and counter electrode, respectively, and 0.5 M Na₂SO₄ aqueous solution was used as the electrolyte. The CHI760E electrochemistry workstation (Chenhua, Shanghai, China) was used to record the photocurrent-time curves and open-circuit potential (OCP) under light on/off cycles using a 300 W xenon lamp (PLS-FX300HU, Beijing Perfectlight Technology Co., Ltd). The electrochemical impedance spectroscopy (EIS) Nyquist plots were recorded under illumination with an amplitude of 5 mV and a frequency range of 0.01 Hz to 1.0 kHz. The Mott-Schottky curves were measured in the dark at a frequency of 1.0 kHz.

Photogenerated radical detection

In situ EPR spectra for monitoring the generation of radical species during photocatalytic reaction were recorded on a Bruker A300-10/12 spectrometer using 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) as the radical-trapping agent. Especially, for determining the reactive oxygen species •OH and •OOH produced from oxygen reduction or water oxidation, the photocatalyst was dispersed in water (for detection of DMPO-OH adduct) or 50% (v/v) methanol aqueous solution (for detection of DMPO-OOH adduct), in presence of oxygen under irradiation of a xenon lamp. As for detection of CH₃• radicals, the photocatalyst was dispersed in water in the presence of both oxygen and methane gases followed by irradiation with a xenon lamp (details see below).

To identify the generation source of •OH radicals, a fluorescence probing method was employed using coumarin as the molecular probe, which can selectively react with •OH to yield the fluorescent product 7-hydroxycoumarin (OH- coumarin). In a typical procedure, 10 mg of the photocatalyst was dispersed in 20 mL of coumarin aqueous solution (0.5 mM) under either oxygen or argon atmosphere. Following a 30-min dark adsorption period with continuous stirring, the system was irradiated with a xenon lamp for 5 min. After the reaction, aliquots of 5 mL were sampled and filtered. The fluorescence spectra of the generated OH- coumarin were analyzed using a Horiba fluoromax-4 spectrophotometer.

In order to investigate the generation mechanism of CH₃• radicals, EPR spectra were recorded for different reaction systems with or without the addition of desired quenching agent: (i) standard condition: the photocatalyst was dispersed in deionized water followed by purging with oxygen and methane gases. After 10-min light irradiation using a xenon lamp, 100 μL of the reaction solution was mixed with 100 μL of DMPO aqueous solution (100 mM) for EPR analysis; (ii) in presence of TEOA: TEOA was introduced into the standard reaction system to selectively consume the photogenerated holes, without changing other factors; (iii) in presence of IPA: IPA was introduced into the standard reaction system to sacrifice photogenerated holes and •OH radicals concurrently, while other factors remained the same.

In situ DRIFTS measurements

In situ DRIFTS measurements were conducted using a Thermo Scientific Nicolet IS52 instrument in a CH₄ and O₂ atmosphere with added H₂O, under both illuminated and non-illuminated conditions. A liquid nitrogen-cooled MCT-M detector was employed for signal detection. Approximately 10 mg of catalyst was loaded into an in situ FTIR cell (HVC-DRM-5, Harrick Scientific), with a circulating water system to maintain the sample cell temperature at ambient conditions.

Computational method

We performed spin-polarized density functional theory (DFT) calculations using the Vienna Ab Initio Simulation Package (VASP)^80,81 with the generalized gradient approximation (GGA) under the PBE framework⁸². Projected augmented wave (PAW) potentials^83,84 were used for ionic cores, while valence electrons were treated via a plane wave basis set with a kinetic energy cutoff of 400 eV. The Kohn-Sham orbital partial occupancies were managed using Gaussian smearing with a width of 0.05 eV. Electronic energy was considered self-consistent with an energy change below 10⁻⁵ eV. Geometry optimization was achieved when force changes were less than 0.02 eV/Å. Grimme’s DFT-D3 method⁸⁵ was used for dispersion interactions. The equilibrium lattice parameters of the triclinic CuSeO₃ unit cell were optimized as follows: a = 4.725 Å, b = 8.256 Å, c = 10.292 Å, α = 111.8°, β = 92.2°, γ = 89.2°. Based on these values, we created a CuSeO₃(111) surface model (Model 1) with p(1 × 1) periodicity in both the X and Y directions, and two stoichiometric layers along the Z axis, separated by a 15 Å vacuum layer to prevent interactions between periodic images of the surface slab. Model 1 contained 12 Cu, 12 Se, and 36 O atoms. Model 2 was derived from Model 1 by removing one oxygen atom from the outermost layer. For structural optimizations, a 2 × 2 × 1 k-point grid in the Brillouin zone was used for k-point sampling, with the bottom stoichiometric layer fixed and the top layer allowed to relax.

The adsorption energy (E_ads) of adsorbate A was defined as

$${E}_{{ads}}={E}_{\frac{A}{{surf}}}-{E}_{{surf}}-{E}_{A\left(g\right)}$$

(6)

where E_A/surf is the energy of adsorbate A on the surface, E_surf is the energy of the clean surface, and E_A(g) is the energy of the isolated A molecule in a cubic periodic box with a side length of 20 Å and a 1 × 1 × 1 Monkhorst-Pack k-point grid for Brillouin zone sampling.

The formation energy of oxygen vacancy (E_f) was defined as

$${E}_{f}={E}_{V-O}+\frac{{E}_{O_{2}\left(g\right)}}{2}-{E}_{{surf}}$$

(7)

where E_V-O, E_O2(g) and E_surf are the energy of a surface with one oxygen vacancy, the energy of isolated O₂ molecule, and the energy of pristine surface, respectively.