The synergy of in situ-generated Ni⁰ and Ni₂P to enhance CO adsorption and protonation for selective CH₄ production from photocatalytic CO₂ reduction

Abstract

The selective photocatalytic reduction of CO₂ to CH₄ remains a challenge because there is a need for not only strong adsorption sites for the intermediates, but also optimal proton-feeding sites on the photocatalyst surface. Herein, a synergistic dual-site function between in situ-generated Ni⁰ and Ni₂P on carbon nitride nanosheets (CN) for photocatalytic reduction of CO₂ to CH₄ is presented. The highest CH₄ production rate of 69.03 μmol g⁻¹ h⁻¹ is achieved on Ni₂P/CN-0.5 in an aqueous suspension. Detailed analyses show that the promotion of CH₄ is closely correlated with the formation of Ni⁰ sites due to light irradiation, which is confirmed by tracking the compositions of Ni₂P/CN-0.5 through XPS and HRTEM characterization. Density functional theory calculations have been combined with CO-TPD and in situ FTIR spectra to reveal the synergy between in situ-generated Ni⁰ sites and Ni₂P. It shows that the Ni⁰ sites can stabilize the key intermediate *CO, while the Ni⁰–Ni₂P interface can promote *H transfer from Ni₂P to Ni⁰. Therefore, the CO intermediates are rapidly protonated to form CHO* instead of being desorbed from the surface to produce CO, and subsequently CHO* will be converted into CH₄. This work demonstrates a new strategy of designing highly efficient photocatalysts with synergistic catalytic sites for CO₂ conversion to hydrocarbons.

---

1. Introduction

Converting CO₂ into carbon-neutral fuels such as CH₄ is an extremely effective solution to the energy crisis and environmental problems. Among various approaches, photocatalysis could convert CO₂ into value-added chemicals using renewable solar energy without resorting to a secondary energy storage medium.^1,2 However, converting CO₂ to CH₄ is a complex electrochemical process, requiring 8 proton-coupled electron transfer processes (e⁻ + H⁺).^3–5 Along the way, the two-electron reduction products CO and HCOOH may desorb from the catalyst instead of further hydrogenating to CH₄, resulting in low methane selectivity.⁶ Therefore, an efficient photocatalyst for selective CH₄ formation from CO₂ reduction should possess excellent photochemical properties, enriched electrons on active sites with the optimal proton-feeding and adsorption of intermediates.⁷ So far, developing such a highly efficient photocatalyst to selectively reduce CO₂ to CH₄ remains a great challenge. Carbon nitride (CN) is an environmentally friendly semiconductor with excellent photochemical properties and stability.⁸ In addition, the introduction of a metal-based co-catalyst to CN has been approved as one of the effective strategies to improve the photocatalytic activity. The additional metal components on CN not only inhibit the recombination of photogenerated electrons and holes, but also provide the active sites for the adsorption and activation of CO₂, as well as the subsequent hydrogenation steps.⁹

Transition metal phosphides (TMPs), owing to their lower cost, unique electronic structure, and excellent electrical conductivity, have been widely regarded as promising alternatives to Pt for hydrogen production from photocatalytic water splitting.^10,11 Recently, TMPs were also attempted for CO₂ reduction. Wang et al. constructed NiCoOP NPs@MHCF catalysts; the optimized trimetallic oxyphosphide catalysts exhibited considerable activity in photosensitive CO₂ reduction with a CO evolution rate.¹² Calvinho et al. prepared Ni_xP_y for the electrocatalytic reduction of CO₂ at low potentials using a solid-state method, and their results indicated that Ni_xP_y, especially Ni₂P, has an excellent activity in the electrocatalytic reduction of CO₂ to hydrocarbon products.¹³ Li et al. found that the co-modification with Ni₂P and NiO of CN caused the increase of CH₄ yield.¹⁴ These studies demonstrated that Ni₂P could act as a proton-feeding site to promote the kinetics of the protonation step – the key step for CH₄ production in the photocatalytic reduction of CO₂ to CH₄.^15,16

In this work, we develop a dual active site photocatalyst Ni⁰–Ni₂P/CN boosting CH₄ formation from the photocatalytic reduction of CO₂. Firstly, Ni₂P/CN photocatalysts were prepared through a hydrothermal reaction using red phosphorus as the P source. Ni⁰ sites were in situ-generated under light irradiation, which exhibited a close link with the enhanced CH₄ yield. The conditions of introducing Ni₂P to CN have been optimized to maximize the CH₄ production. It is demonstrated that these Ni⁰ sites work synergistically with Ni₂P sites to promote the formation of CH₄. In situ FTIR spectra, CO-TPD and theoretical calculations were used to gain insights into the reaction mechanism of CO₂ reduction on the Ni⁰–Ni₂P/CN photocatalyst.

2. Experimental and computational methods

2.1 Synthesis of CN and Ni₂P/CN-x

Carbon nitride was synthesized by the reported method.¹⁷ Specifically, 10.0 g of urea was dissolved in 20 mL of H₂O. After the temperature stabilized to room temperature, 1.0 g of ammonium oxalate, which generates NH₃ and CO₂ during thermal decomposition, was used as a pore builder in this system to obtain the porous g-C₃N₄, which was added to the above solution while stirring. The mixed solution was stirred at room temperature for 1 h and then raised to 80 °C for 16 h to obtain a white precursor. The precursors were ground and transferred to a 50 ml ceramic crucible and calcined in a muffle furnace at a heating rate of 5 °C min⁻¹ for 2 h at 550 °C to obtain orange carbon nitride nanosheets, named CN.

Ni₂P/CN-x samples were synthesized by a simple hydrothermal method. Typically, 0.3 g of CN and 0.5 mmol of Ni(NO₃)₂·6H₂O were dispersed in 65 mL of H₂O by sonication for 1 h. 5.0 mmol of P was added to the above solution and stirred for 1 h at room temperature. Next, the resulting suspension was loaded into a 100 mL Teflon autoclave and maintained at 140 °C for 10 h. The precipitates were separated by centrifugation and washed several times with water and ethanol, and then vacuum dried at 60 °C for 12 h. The resulting gray solid was labeled as Ni₂P/CN-0.5. By changing the amount of Ni(NO₃)₂·6H₂O and P in a molar ratio of 1 : 10, other Ni₂P/CN-x samples were obtained, where x represents the amount of Ni(NO₃)₂·6H₂O (mmol) used. The preparation process is shown in Fig. 1a.

Fig. 1 (a) Schematic diagram of the synthesis process of Ni₂P/CN-x; (b) XRD patterns; and (c) FTIR spectra of CN and Ni₂P/CN-x; and (d–g) XPS spectra of the elements in CN and Ni₂P/CN-0.5.

2.2 Material characterization

An X-ray diffractometer (XRD, D8-Focus, Bruker AXS) equipped with a Cu Kα (λ = 1.54056 Å) radiation source was used to record the crystalline structure of the samples with a scanning speed of 5° min⁻¹ and a scanning range (2θ) of 10–80°. Fourier transform infrared spectroscopy (FTIR, Magna 670, Nicolet) was used to detect the functional group of the samples, using KBr as a reference. The sample morphology was examined using a field emission scanning electron microscope (SEM, S-4800, Hitachi). A 200 kV field emission transmission electron microscope (TEM, JEM-2100F, Japan Electronics) was used to obtain scanning transmission electron microscopy (STEM) and high-resolution transmission electron microscopy (HRTEM) images and selected area energy dispersive X-ray spectroscopy (EDS) spectra of the samples. An X-ray photoelectron spectrometer (XPS, ESCALAB Xi⁺, Thermo Fisher Scientific) equipped with Al Kα radiation was used to analyze the surface chemical composition of the samples, and the binding energy was calibrated by carbon C 1s (284.8 eV). An ultraviolet-visible diffuse reflectance spectrophotometer (UV-vis DRS, T2600, Youke) was used to acquire the absorption spectra of the samples. The photoluminescence (PL) spectra of the samples were obtained using a steady-state fluorescence spectrometer (Fluorolog 3-21, Jobin Yvon Inc.) with an excitation wavelength of 420 nm. Nitrogen adsorption–desorption isotherms were obtained with a nitrogen physisorption instrument (Tristar 3000 Micromeritics), and CO₂ adsorption isotherms were obtained with a specific surface and pore size analyzer (BELSORP-max, MicrotracBEL, Japan) at 25 °C. All samples were vacuum pretreated at 200 °C for 3 hours prior to the measurement.

2.3 Performance of photocatalytic CO₂ reduction

The photocatalytic CO₂ reduction reaction was carried out in a 250 mL Pyrex reactor with a UV-vis light source (300 W Xe lamp), and the reaction temperature was maintained at 15 °C using a cooling water circulator. Specifically, 25 mg of the as-prepared sample were dispersed into a mixed solution containing 20 mL of 0.1 M NaOH and 10 mL of TEOA. Prior to the illumination, the reaction system was purged with high purity CO₂ for 30 min to expel the air, while allowing the reaction solution to be saturated with CO₂. The reaction was carried out under UV-vis light (320 nm < λ < 780 nm, 200 mW cm⁻²) with continuous stirring. During the reaction, 1 mL of gas was extracted with a syringe at 40 min intervals and the products were analyzed by gas chromatography (GC 9790 II Plus, Fuli), and the results were averaged from three repeating tests (more details can be seen in the ESI†). A schematic diagram of the photocatalytic reaction system is shown in Fig. S1.†

2.4 Density functional theory (DFT) calculations

DFT calculations were performed using the spin-polarized Vienna Ab initio Simulation Package (VASP) with projector-augmented wave pseudopotential (PAW).¹⁸ The generalized gradient approximation (GGA)¹⁹ in the Perdew–Burke–Ernzerhof (PBE)²⁰ form was adopted, with a cutoff energy of 400 eV for the plane-wave basis set. The ionic convergence limit was set to 0.02 eV Å⁻¹, while the electronic convergence limit was set to 10^–5 eV. The optimization of gas molecules was performed in a cubic box of 20 Å in each direction and using a single k point (Γ point) for the Brillouin-zone integration. For periodic models, a vacuum of 20 Å in the Z direction was set to prevent the interaction between two periodic units. A k-point grid of 3 × 3 × 1 was set for Brillouin-zone integration of the catalysts.

The Gibbs free energy change (ΔG) of each elementary step was calculated by using the following equation:

ΔG = ΔE + ΔE_ZPE − TΔS, (1)

where ΔE is the electronic energy change between the two intermediates. ΔE_ZPE and ΔS refer to the zero-point energy correction and reaction entropy, respectively. The temperature T corresponds to 298.15 K. The solvent stabilization was taken into account by applying a correction of −0.25 eV for hydroxyl-containing groups (*R–OH) and −0.1 eV for oxygen-containing groups (*CO and *CHO).^21–23 Bader charge analysis was conducted to investigate the charge transfer between intermediates and the catalysts.^24,25 LOBSTER 3.1.0 package was adopted to deal with the electronic structure analysis via the crystal orbital Hamilton population (COHP).²⁶

3 Results and discussion

3.1 Crystal structure and chemical composition

The phase structures of the as-prepared CN and Ni₂P/CN-x were revealed from XRD spectra, and the results are shown in Fig. 1b.

All samples presented the CN characteristic diffraction peaks at 13.7° and 27.2° attributed to the periodic in-plane tri-s-triazine motif stacking and the interlayer stacking of CN.²⁷ In addition, Ni₂P/CN-x samples display the peaks at 40.8°, 44.6°, 47.3° and 54.2°, ascribed to the (111), (201), (210) and (300) crystal planes of the Ni₂P hexagonal crystal structure (PDF #03-0953), respectively. The intensity of the diffraction peaks attributed to Ni₂P increases with the amount of Ni and P precursors, indicating the successful introduction of Ni₂P into the substrate CN.^28,29 From the FTIR spectra (Fig. 1c), the featured absorption peaks of CN can be observed. The peak at 810 cm⁻¹ corresponds to the respiratory vibration peak of the triazine ring in CN, and the band at 1250–1650 cm⁻¹ corresponds to the stretching mode of the CN heterocycle, indicating that the introduction of Ni₂P does not alter the molecular structure of CN. The above results indicated that the composited photocatalysts Ni₂P/CN were successfully prepared.

To investigate the surface composition and chemical states of the samples, XPS spectra of CN and Ni₂P/CN-0.5 were obtained (Fig. 1d–g). For pure CN, the C 1s spectrum displays three distinct peaks at 284.80, 286.20, and 288.27 eV (Fig. 1d), which are assigned to the surface adventitious carbon, C–NH₂ and sp² hybridized carbon (N=C–N), respectively.³⁰ The N 1s spectrum (Fig. 1e) exhibits three peaks at 398.68, 399.54, and 401.17 eV, corresponding to C–N=C, N–(C)₃, and C–NH in CN.^31,32 By comparison, the corresponding peaks for Ni₂P/CN-0.5 slightly shifted, indicating a strong interaction between Ni₂P and CN.³³ The Ni 2p spectrum (Fig. 1f) was spin split into Ni 2p_3/2 and Ni 2p_1/2. In the Ni 2p_3/2 region, three peaks appeared at 852.65, 855.89, and 861.40 eV, corresponding to the Ni^δ+ species in Ni₂P, the Ni²⁺ species and the satellite peak, respectively. In the Ni 2p_1/2 region, there were three peaks, at 870.10, 874.00, and 880.00 eV, which could be attributed to Ni^δ+, Ni²⁺, and satellite peaks, respectively.³⁴ In the P 2p spectrum (Fig. 1g), the peak at 129.78 eV was attributed to the P^δ− species in Ni₂P and the peak at 133.11 eV corresponded to the formation of P–O bonds.^28,35 Note that the reason for the presence of Ni²⁺ and P–O bonds is twofold: the surface oxidation of Ni₂P and the formation of partially hydrated surface phosphates (e.g. Ni₃(PO₄)₂) using a hydrothermal phosphating method.¹³ Therefore, the XPS results confirm the successful loading of Ni₂P on CN and the strong electronic interaction between them.^36,37

3.2 Morphology and microstructure

SEM and TEM images provide the morphology and the microstructure of the samples (Fig. 2). Fig. 2a and b indicate that CN has a thin sheet structure with curled edges. Ni₂P/CN-0.5 maintains the nanosheet morphology (Fig. 2c) and the edges are further curled due to the hydrothermal treatment.³⁸ The TEM and STEM images of Ni₂P/CN-0.5 (Fig. 2d and e) demonstrate that Ni₂P nanoparticles (the bright part) are distributed on the CN sheet.^39,40 In the HRTEM image (Fig. 2f), the lattice spacing of 0.221 nm is indexed to the (111) crystal plane of Ni₂P.^41,42 The EDS maps of Ni₂P/CN-0.5 are shown in Fig. 2g, indicating the presence of C, N, Ni, and P elements.²⁷ The above results confirmed that Ni₂P nanoparticles are successfully dispersed on the CN nanosheets.

Fig. 2 (a and b) SEM and TEM of CN, and (c–f) the SEM, TEM, STEM and HRTEM image of Ni₂P/CN-0.5. (g) The STEM-EDS elemental mapping of Ni₂P/CN-0.5.

3.3 Optical and electrical properties

The UV-Vis diffuse reflectance spectra of the as-prepared samples are shown in Fig. 3a. Compared to CN, the increased visible light absorption of Ni₂P/CN-x is attributed to the intrinsic absorption of Ni₂P.⁴³ There is a slight red shift in the absorption edge of Ni₂P/CN-x, an indication of narrowing of the band gaps (E_g) of Ni₂P/CN-x from that of CN.⁴⁴ The band gaps of CN, Ni₂P/CN-0.3, Ni₂P/CN-0.5 and Ni₂P/CN-1 were calculated to be 2.60 eV, 2.54 eV, 2.50 eV and 2.46 eV, according to the plots of (αhν)^x (x = 2) vs. band energy (hν).⁴⁵ As shown in Fig. S2,† the Mott–Schottky (MS) curves exhibit a positive slope, a characteristic feature of the n-type semiconductor.⁴⁶ The flat-band potential (V_fb) was determined to be −1.15 and −1.07 V (vs. NHE, pH = 7) of CN and Ni₂P/CN-0.5. According to the equation E_CB = V_fb − 0.2 V, the conduction band potentials (E_CB) of CN and Ni₂P/CN-0.5 were −1.35 and −1.27 V (vs. NHE, pH = 7), respectively.⁴⁷ The valence band potential (E_VB) could then be determined based on the equation E_VB = E_g − E_CB.⁴⁸ A summary of these results led to the alignment of the bands of CN and Ni₂P/CN-0.5, as shown in Fig. 3b. Clearly, both CN and Ni₂P/CN-0.5 could drive both CO₂ reduction and H₂O oxidation reactions.³⁷

Fig. 3 (a) UV–vis absorption spectra and band gaps of the as-prepared samples (α, h and ν represent the absorption coefficient, Planck constant, and light frequency, respectively). (b) Band positions of CN and Ni₂P/CN-0.5. (c) PL spectra; (d) photocurrent response curves; and (e) EIS plots of CN and Ni₂P/CN-x. (f) The diagram of the Schottky junction and electron transfer path between Ni₂P and g-C₃N₄.

The PL spectra, photocurrent density profile and EIS plots were used to evaluate the ability of the separation and transport of photocarriers.⁴⁹ As shown in Fig. 3c, the Ni₂P/CN-x composites all show a significant decrease in emission intensity compared to CN, and Ni₂P/CN-0.5 exhibits the lowest emission intensity. The quenching of fluorescence indicates that the complexation of photogenerated electron–hole pairs is effectively suppressed following the successful loading of Ni₂P. In addition, the change in the energy distribution of Ni₂P/CN-x surface states caused a slight blue shift of the PL peaks.^50,51

As shown in Fig. 3d, all Ni₂P/CN composites exhibit a higher photocurrent than CN, demonstrating an effective transfer of photogenerated electrons to the surface. The impedance arc radius of the Ni₂P/CN composites was smaller compared to that of pure CN, indicating less resistance to the charge transport in Ni₂P/CN than in CN (Fig. 3f), with Ni₂P/CN-0.5 showing the smallest radius.⁵² The above results demonstrate that the Ni₂P/CN-0.5 photocatalyst exhibits the most efficient separation of the photogenerated electrons and holes, which is essential to facilitate the photocatalytic CO₂ reduction.

Due to the metallic properties and excellent electrical conductivity of Ni₂P, the enhanced separation and transfer of photogenerated electrons and holes in the composite catalyst can be attributed to the formation of a Schottky junction between CN and Ni₂P (Fig. 3f).⁵³ Previous studies showed that the Fermi level of CN was higher than that of Ni₂P.⁵⁴ When CN and Ni₂P are in contact, the difference in Fermi levels drives the electron transfer from CN to Ni₂P until the equilibrium is reached and a Schottky junction is formed. The existence of the Schottky barrier prevents the electrons to flow back to CN, effectively facilitating the separation and transfer of the photogenerated charge carriers, as illustrated in Fig. 3f. These findings are demonstrated by the density of states (DOS) analysis of C₃N₄ and Ni₂P (Fig. S3 and S4†). As a result, Ni₂P acted as an electron trap and enriched the active sites with a sufficient number of electrons, which drives the reduction of CO₂ to CH₄.

3.4 Photocatalytic performance for CO₂ reduction

The photocatalytic performance of the as-prepared photocatalysts for CO₂ reduction was evaluated in an aqueous solution of NaOH and TEOA under UV-vis irradiation for 200 min. During the reaction, the gas was sampled at an interval of 40 min and detected by the GC method, and the GC chromatograms of gas produced during the photocatalytic reduction are shown in Fig. S5.† The detected gas products included CH₄ and CO from CO₂ reduction as well as H₂ from the competitive H₂O reduction. As shown in Fig. 4a–c, CN exhibits the lowest yields of CH₄, H₂ and CO. All Ni₂P/CN-x photocatalysts show sharply enhanced CH₄ and H₂ yields, but a very low CO yield, suggesting that loading Ni₂P has a positive effect on the adsorption and further hydrogenation of CO. Notably, for the Ni₂P/CN-x photocatalysts, the CH₄ production displays a different trend from H₂ with increasing Ni₂P loading. As shown in Fig. 4d, the CH₄ evolution showed a peak yield of 69.03 μmol g⁻¹ h⁻¹ on Ni₂P/CN-0.5, wherever H₂ yields increase monotonously to the highest on Ni₂P/CN-1. Accordingly, Ni₂P/CN-0.5 provides the highest selectivity of CH₄. It can be explained that the number of reduced protons increases as the Ni₂P loading increases, and when the formation of the reduced protons is beyond their consumption by the hydrogenation of CO to CH₄, the excessive H will recombine and desorb as H₂. Therefore, the loading of Ni₂P has to be optimal to maximize the CH₄ selectivity from CO₂ reduction. Moreover, by a contrastive analysis of the yield-time profiles before and after the irradiation time of 40 min in Fig. 4a and c, it is observed that there is a sharp change of yield for either CH₄ (from a slow yield increase to a fast increase) or H₂ (from a fast yield increase to a slow increase). This interesting finding will be discussed in the next section with regard to the identification of active sites and the mechanism.

Fig. 4 The product yield profiles with the reaction time of (a) CH₄, (b) CO and (c) H₂; (d) the average evolution rates of CH₄, CO and H₂ on CN and Ni₂P/CN-x within a reaction time of 4 hours; (e) the results of the condition controlled experiments; and (f) the gas evolution rates of the reference samples.

Control experiments were conducted on the photocatalyst Ni₂P/CN-0.5 to confirm the source of the resulting CH₄ (Fig. 4e). Either under dark conditions or in the absence of Ni₂P/CN-0.5, no gaseous product was detected, indicating that light irradiation and the photocatalyst Ni₂P/CN-0.5 are essential for the reactions. Only hydrogen was detected if CO₂ was not introduced into the reaction system. Therefore, CO₂ was the source of CH₄ and CO in the products. In addition, without NaOH, the yields of CH₄ and CO decreased significantly, whereas the yield of H₂ increased. The presence of NaOH increased the solubility of CO₂ from the neutral aqueous condition (pH of the different reaction solution as shown in Table S1†).⁵⁵ The role of the sacrificial agent (TEOA) was also tested. In the absence of TEOA, only a small quantity of products was detected due to the high recombination rate of the photogenerated charge carriers when the holes were not consumed efficiently. The test under the visible light irradiation (>420 nm) showed that the photocatalytic activity was substantially lower than that under the light in the wavelength range of UV-vis (320–780 nm). This demonstrated that the energy from the visible light was not enough to generate the sufficient photocharge carriers. The above results indicated that the CH₄ and CO products originated from the photocatalytic reduction of CO₂ on the photocatalyst Ni₂P/CN-0.5 under UV-vis light irradiation.

The additional reference samples were used to eliminate or quantify the effect of the other possible contributions to the performance of the Ni₂P/CN-0.5 catalyst.⁵⁶ As shown in Fig. 4f, hydrothermal treatment of CN nanosheets (Hydro-CN) alone does not improve the catalytic performance for either CH₄ or H₂ production notably.⁵⁷ When adding Ni(NO₃)₂ only for modifying CN, the obtained sample Ni/CN provided a modest increase in CH₄ production. In contrast, H₂ became the dominant product on P/CN, with P being the only promoter. The XRD spectra of both Ni/CN and P/CN in Fig. S6a† demonstrated that the diffraction peaks assigned to the CN nanosheet structure remained well. The very different product distribution with a single component promoter (Ni or P) clearly showed that the Ni species are key to the formation of CH₄. These results also confirmed that the successful loading of Ni₂P on the CN surface was the main reason for the high photocatalytic activity of the composite catalysts. And that P or Ni(NO)₃ alone with CN could not generate high-performance composite catalysts under hydrothermal conditions. Further analysis focuses on understanding the nature as well as the role of the surface species of the Ni₂P/CN-0.5 photocatalyst for catalytic CO₂ reduction.

3.5 Identification of the active sites promoting CH₄ formation: in situ-generated Ni⁰ induced by light irradiation

The average rate of CH₄ and H₂ in every interval and the selectivity of CH₄ along with irradiation time on Ni₂P/CN-0.5 are shown in Fig. 5a. Ni/CN showed a period of activation, but not as noticeably as Ni₂P/CN (Fig. S6b†), suggesting that the new active sites should be relevant to Ni species. In order to probe the new active site due to light irradiation, the XPS spectra of the fresh Ni₂P/CN-0.5 and the light induced generated one (Ni⁰–Ni₂P/CN-0.5) were compared, as shown in Fig. S7† and Fig. 5c. As displayed in Fig. S7a and b,† in comparison with the fresh Ni₂P/CN-0.5, Ni⁰–Ni₂P/CN-0.5 showed positive shifts of the binding energy for both the peak of sp² hybridized C (N=C–N) in the C1s spectrum and the peak of C=N in the N 1s spectrum. It is indicative that CN nanosheets act as the electron donor in the composite Ni⁰–Ni₂P/CN-0.5 under light irradiation.⁵⁸ In addition, note that the peaks corresponding to the P–O bond in the P 2p spectra (Fig. S7c†) decreased and the binding energy negatively shifted. Importantly, the notably different Ni 2p spectra of Ni⁰–Ni₂P/CN-0.5 from Ni₂P/CN-0.5 can be observed (Fig. 5c). For Ni⁰–Ni₂P/CN-0.5, the peaks ascribed to Ni²⁺ decreased significantly, and a new pair of peaks at 850.2 and 867.6 eV corresponding to the metallic Ni (Ni⁰) appeared. Meanwhile, the formation of Ni⁰ in Ni⁰–Ni₂P/CN-0.5 was again observed from the HRTEM image, which exhibited lattice fringes with a spacing of 0.176 nm and 0.221 nm assignable to the Ni (200) and Ni₂P (111) facets, respectively.²⁹ The combined XPS and HRTEM results clearly confirm the in situ formation of the Ni⁰ sites. Considering both the principle of the hydrothermal phosphating method and the decreased peak intensity of Ni²⁺, P–O bond in the XPS spectra of Ni⁰–Ni₂P/CN-0.5, these Ni⁰ sites are attributed to the reduction of the hydrated surface nickel phosphorylate (e.g. Ni₃(PO₄)₂) under light irradiation. The presence of Ni⁰ sites is believed to increase the CH₄ yield as Ni⁰ plays an active role in the methanation of CO.^59,60 Consequently, the Ni⁰–Ni₂P/CN photocatalytic system is in situ generated by light irradiation. The coexisting active sites of Ni₂P and Ni⁰ synergistically work to enhance the production of CH₄.^61,62

Fig. 5 (a) The average rate of CH₄ and H₂ and the selectivity of CH₄ in every interval on Ni₂P/CN-0.5; (b) the product yield profiles in the 3-cycle experiments on Ni₂P/CN-0.5; (c) the Ni 2p spectra of the fresh Ni₂P/CN-0.5 and Ni⁰–Ni₂P/CN-0.5 (b); and (d) HRTEM of Ni⁰–Ni₂P/CN-0.5.

3.6 Understanding the roles of light induced generated Ni⁰ in Ni⁰–Ni₂P/CN-0.5

The CO₂ adsorption capacity of the photocatalysts play an important role in determining their photocatalytic CO₂ reduction performances. The CO₂ adsorption isotherm of CN, Ni₂P/CN-0.5 and Ni⁰–Ni₂P/CN-0.5 at 25 °C is shown in Fig. 6a. Clearly, the CO₂ adsorption capacity of Ni₂P/CN-0.5 was higher than that of CN, indicating that the modification of CN by Ni₂P is beneficial for the absorption of CO₂. Clearly, even higher CO₂ adsorption capacity was achieved over Ni⁰–Ni₂P/CN-0.5, indicative of Ni sites also having the strong affinity of CO₂.⁶³ Moreover, the nitrogen isothermal adsorption–desorption isotherms were conducted, and the results are shown in Fig. S8 and Table S2.† For comparison, the results of CN, Ni₂P/CN-0.5 and Ni⁰–Ni₂P/CN-0.5 are displayed in Fig. 6b; it can be seen that their similar BET areas could exclude the effect of physical structures on the adsorption of CO₂. Therefore, the generated Ni⁰ sites on Ni⁰–Ni₂P/CN-0.5 enhance CO₂ adsorption, thereby improving the ability of activating CO₂ molecules.

Fig. 6 (a) CO₂-adsorption isotherms; (b) N₂ adsorption–desorption isotherms; (c) CO-TPD spectra on CN, Ni₂P/CN-0.5 and Ni⁰–Ni₂P/CN-0.5; and (d) in situ FTIR spectra of CO₂ and H₂O interacting under light irradiation.

Furthermore, CO-TPD was used to evaluate the adsorption properties of CO, a key intermediate in photocatalytic CO₂ reduction, on the CN, Ni₂P/CN-0.5 and Ni⁰–Ni₂P/CN-0.5 samples.⁶⁴ As shown in Fig. 6c, CN displays insignificant CO desorption features compared to the other two samples. This is consistent with the fact that CO is the dominant product on CN (Fig. 4b). The Ni₂P modified CN, i.e., both Ni₂P/CN-0.5 and Ni⁰–Ni₂P/CN-0.5, shows much higher CO desorption peaks, indicating the enhanced ability to adsorb CO. Furthermore, the CO desorption of Ni⁰–Ni₂P/CN-0.5 is increased compared to Ni₂P/CN-0.5, which could be attributed to the appearance of Ni⁰ sites induced by light irradiation. Therefore, the presence of Ni⁰ enhances the adsorption of CO and facilitates subsequent hydrogenation to CH₄. Therefore, the CO-TPD result is indicative of the presence of the light-induced Ni⁰ leading to an enhancement of CO adsorption. Besides, as shown in Table S3,† we compared the results with previous studies. In other systems where transition metal phosphides are used as co-catalysts and the solvent is water or other sacrificial agents, the products are mostly CO with few multielectron hydrocarbons. In contrast, our catalyst has excellent catalytic activity for CH₄ under mild reaction aqueous conditions. Combining the above results and product yields, and the excellent performance for CH₄ is attributed to the generation of Ni⁰ for enhanced CO₂ and CO adsorption.

In addition, in situ FTIR spectroscopy was performed to track the intermediates in the photocatalytic CO₂ reduction process, and the results are shown in Fig. 6d. The peaks corresponding to HCO₃⁻ (1444 cm⁻¹), m-CO₃²⁻ (1373 cm⁻¹) and b-CO₃²⁻ (1309 cm⁻¹, 1530 cm⁻¹) increase with prolonged illumination, demonstrating enhanced CO₂ adsorption and activation due to the appearance of Ni⁰.^65,66 Moreover, the adsorbed *COOH is generally regarded as the initial intermediate in the CO₂ reduction reaction. The bands for C=O stretching and O–H deformation of *COOH can be observed at 1620 cm⁻¹, and 1220 cm⁻¹.⁶⁷ It is noted that the band attributed to *CO (2100 cm⁻¹) is very weak throughout the test, which suggests the enhanced kinetics of the conversion of *CO to the successive intermediates. Further hydrogenation of the adsorbed *CO leads to the formation of intermediates, such as *CHO and CH₃O*, before forming CH₄. As shown in the enlarged view in Fig. 6d, the H–C=O bending mode of *CHO and the C–O stretching mode of the adsorbed CH₃O* species can be observed at 1100 cm⁻¹ and 1125 cm⁻¹, respectively.⁶⁸ These results provide further evidence that Ni⁰ plays an essential role in activating CO₂ and facilitating the adsorption and further hydrogenation of CO. The identification of the CH₃O* intermediate supports that the photocatalytic CO₂ reduction on Ni₂P/CN-0.5 follows the formaldehyde-mediated reaction pathway (CO₂ (g) → *CO₂ → *COOH → *CO → *CHO → *CH₂O → CH₃O* → CH₄ (g)), as shown in Scheme S1.†⁶⁹

DFT calculations were conducted to further evaluate the role of the Ni⁰ site in photocatalytic CO₂ reduction to CH₄ on Ni₂P/CN-0.5. A Ni₄ cluster supported on Ni₂P (111) (namely Ni₄/Ni₂P) was constructed to represent the Ni⁰ active site (Fig. S9 and S10†). The first reduction step of CO₂ was considered to be the formation of the *COOH intermediate. As shown in COHP of Ni₂P(111) and Ni₄/Ni₂P(111) (Fig. 7a and b), the antibonding feature of Ni₄/Ni₂P(111) near the Fermi level was more pronounced than Ni₂P(111). More charges have been transferred between *COOH and the surface on Ni₄/Ni₂P(111) than on Ni₂P (111) based on the Bader charge analysis. The integrated COHP (namely ICOHP) of O–Ni and C–Ni on Ni₄/Ni₂P and Ni₂P (111) surfaces can be used as a quantitative measure of bond strength where a more negative ICOHP value corresponds to a stronger bond.^70,71 As shown in Fig. 7c, the ICOHPs of both O–Ni and C–Ni bonds on Ni₄/Ni₂P (111) are more negative than those on Ni₂P (111), further demonstrating that the presence of Ni₄ can facilitate the adsorption of *COOH.^72,73

Fig. 7 The crystal orbital Hamilton population (COHP) of adsorbed *COOH for the O–Ni and C–Ni bonds on the (a) Ni₂P(111) surface and (b) Ni₄/Ni₂P(111). The insets show the charge density difference of *COOH on these two catalysts. (c) The integrated-COHP (ICOHP) of the O–Ni and C–Ni bonds on Ni₂P (111) and Ni₄/Ni₂P (111), respectively. (d) Reaction pathways for CO₂ reduction to CH4 on Ni₂P (Ni-hol) and Ni₄/Ni₂P catalysts.

The reaction free energy diagrams along the pathway for CO₂ reduction to CH₄ on Ni₄/Ni₂P and Ni₂P are shown in Fig. 7d and Fig. S11.† On the Ni₂P (111) surface, CO₂ prefers to adsorb at the Ni-hollow site with a *COOH formation energy of 0.39 eV, which can be converted to *CO after releasing one water molecule. The formation of *CHO by the hydrogenation of *CO on Ni₂P(111) is endergonic with a value of 0.90 eV and raises the free energy to a level higher than that of the desorbed CO molecule. In contrast, *COOH on Ni₄/Ni₂P (111) is highly stabilized with an adsorption energy of −0.43 eV by combining with two Ni atoms of Ni₄ cluster. The enhanced stabilization of the adsorbed *CO on Ni₄/Ni₂P (111) is consistent with the CO-TPD results. Further hydrogenation will form *CHO with a formation energy of −0.37 eV, which is much lower than that for CO desorption (0.27 eV). Therefore, the CO (g) formation will be suppressed and the hydrogenation of *CO will be facilitated. Among three subsequent reaction pathways considered, two pathways (path 1 and 2) leading to CH₄ are favorable. Bader charge analysis of the intermediates through paths 1 and 2 (Fig. S12†) shows that Ni₄ acts like an electron donor, while larger charge variations are observed on path 1 than on path 2.

On the other hand, the enhancement of *CO adsorption on Ni₄/Ni₂P (111) results in decrease of the free energy barrier of the rate-determining step, which is related to the formation of *CHO on both Ni₄/Ni₂P and Ni₂P catalysts, from 0.90 to 0.69 eV. The lower energy barrier promotes the CO₂ reduction reaction more competitively than the hydrogen evolution reaction (HER) (Fig. S13†), elucidating the selectivity enhanced by the Ni⁰ site formation on Ni₂P. Furthermore, the process for H transfer from the Ni₂P surface to the Ni₄ cluster for *CHO formation on Ni₄/Ni₂P (111) has been calculated (Fig. S14†). It indicates that the emergence of the Ni⁰ site can facilitate the process by overcoming a moderate barrier (∼1.05 eV) and thus enhance the subsequent CH₄ synthesis. To sum up, the DFT results reveal that the synergy of Ni⁰ and Ni₂P contributes to the stabilization of the *COOH and *CO intermediates, which enhance the hydrogenation process, resulting in the efficient formation of CH₄.

Based on the above results and previous reports, the photocatalytic CO₂ reduction reaction over Ni₂P/CN-x catalysts followed a typical co-catalytic mechanism. As shown in Fig. 8, at the initial stage under irradiation, the photogenerated electrons were separated and transferred to the Ni₂P particles, causing the competitive hydrogen evolution reaction (HER) to produce gaseous H₂ as the dominant product, whereas very low yields of CO and CH₄ were obtained from CO₂ reduction. Meanwhile, some photoelectrons on Ni₂P particles were consumed by the reduction of the hydrated surface nickel phosphorylate to generate Ni⁰ sites. Once the ensemble of Ni⁰ sites reaches a critical size, it will act synergistically with Ni₂P to stabilize *COOH and *CO and facilitate subsequent hydrogenation reactions. Since CO₂ consumes *H more rapidly in the reduction process, HER is greatly inhibited. The photocatalyst presents a sharp increase in CH₄ evolution accompanied by the significant decrease of H₂ yield in the initial irradiation period, until CH₄ production becomes completely dominant over H₂ evolution. Therefore, selective production of CH₄ on Ni⁰–Ni₂P/CN can be attributed to (a) favorable photochemical properties in separating and transferring the photogenerated electrons and holes and (b) enhanced CO adsorption and hydrogenation activity due to the in situ generated Ni⁰ sites. The synergy between Ni⁰ and Ni₂P also contributes to the stabilization of the hydrogenation intermediates and facilitates sequential hydrogenation, leading to the efficient formation of CH₄.

Fig. 8 Schematic diagram of the mechanism of the reduction of CO₂ to CH₄ on the Ni⁰–Ni₂P/CN photocatalyst.

4. Conclusion

A dual active site photocatalyst Ni⁰–Ni₂P/CN for selective CH₄ formation has been successfully synthesized. It shows high efficiency in separating and transferring the photogenerated electrons and holes. The conditions of introducing Ni₂P to CN have been optimized to maximize the CH₄ production. Importantly, the in situ generated Ni⁰ sites under light irradiation have been confirmed, which is closely correlated with the promotion of CH₄. In situ FTIR spectra, CO-TPD and theoretical calculations were used to gain insights into the synergy between the in situ-generated Ni⁰ sites and Ni₂P sites. The results reveal that the Ni⁰–Ni₂P interface can promote the H transfer from Ni₂P to Ni⁰, while the Ni⁰ sites can stabilize the key intermediate *CO and reduce the energy barrier for *CO to *CHO. Therefore, the subsequent protonation steps can be facilitated, giving rise to the production of the multi-electron product CH₄ and the suppressed side reaction of hydrogen evolution. This work demonstrated the success of designing a highly efficient photocatalyst with multiple active sites for CO₂ conversion to hydrocarbons.

Author contributions

Xuemei Liu: writing – original draft, methodology, investigation analysis, and data curation. Chaonan Cui: software, methodology, investigation analysis, data curation and formal analysis. Shuoshuo Wei: investigation analysis and data curation. Jinyu Han: writing – review and editing. Xinli Zhu: writing – review and editing. Qingfeng Ge: supervision and writing – review and editing. Hua Wang: conceptualization, supervision, writing – review and editing, and project administration. X. L. and C. C. contributed equally to this work. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

Conflicts of interest

The authors declare no competing financial interest.

Acknowledgements

This work was supported by the Fundamental Research Funds for the Central Universities, the National Natural Science Foundation of China (No. 22178266 and 22272180) and the Beijing Natural Science Foundation (Grant No. 2232035).

References

1. B. Zhao, Y. Zhao, P. Liu, Y. Men, X. Meng and Y. Pan, Jiégòu huàxué, 2022, vol. 41, pp. 12–21.
2. X. Meng, C. Peng, J. Jia, P. Liu, Y. Men and Y. Pan, J. CO₂ Util., 2022, 55, 101844.
3. J. Nai, S. Wang and X. W. D. Lou, Sci. Adv., 2019, 5, x5095.
4. P. Liu, Y. L. Men, X. Y. Meng, C. Peng, Y. Zhao and Y. X. Pan, Angew. Chem., Int. Ed., 2023, 62, e202309443.
5. W. Gao, S. Li, H. He, X. Li, Z. Cheng, Y. Yang, J. Wang, Q. Shen, X. Wang, Y. Xiong, Y. Zhou and Z. Zou, Nat. Commun., 2021, 12, 4747.
6. J. Li, H. Huang, W. Xue, K. Sun, X. Song, C. Wu, L. Nie, Y. Li, C. Liu, Y. Pan, H. Jiang, D. Mei and C. Zhong, Nat. Catal., 2021, 4, 719–729.
7. S. Bai, L. Tan, C. Ning, G. Liu, Z. Wu, T. Shen, L. Zheng and Y. Song, Small, 2023, 19, 2300581.
8. B. Dam, B. Das and B. K. Patel, Green Chem., 2023, 25, 3374–3397.
9. J. Ran, M. Jaroniec and S. Z. Qiao, Adv. Mater., 2018, 30, 1704649.
10. W. Wang, X. Zhao, Y. Cao, Z. Yan, R. Zhu, Y. Tao, X. Chen, D. Zhang, G. Li and D. L. Phillips, ACS Appl. Mater. Interfaces, 2019, 11, 16527–16537.
11. F. Zhang, J. Zhang, J. Li, X. Jin, Y. Li, M. Wu, X. Kang, T. Hu, X. Wang, W. Ren and G. Zhang, J. Mater. Chem. A, 2019, 7, 6939–6945.
12. Y. Wang, S. Wang and X. W. D. Lou, Angew. Chem., Int. Ed., 2019, 58, 17236–17240.
13. K. U. D. Calvinho, A. B. Laursen, K. M. K. Yap, T. A. Goetjen, S. Hwang, N. Murali, B. Mejia-Sosa, A. Lubarski, K. M. Teeluck, E. S. Hall, E. Garfunkel, M. Greenblatt and G. C. Dismukes, Energy Environ. Sci., 2018, 11, 2550–2559.
14. Q. Li, W. Feng, Y. Liu, D. Chen, Z. Wu and H. Wang, J. Mater. Chem. A, 2022, 10, 15752–15765.
15. G. Huang, Q. Niu, J. Zhang, H. Huang, Q. Chen, J. Bi and L. Wu, Chem. Eng. J., 2022, 427, 131018.
16. D. Sun, J. Li, T. Shen, S. An, B. Qi and Y. Song, ACS Appl. Mater. Interfaces, 2022, 14, 16369–16378.
17. S. Zhang, Y. Liu, P. Gu, R. Ma, T. Wen, G. Zhao, L. Li, Y. Ai, C. Hu and X. Wang, Appl. Catal., B, 2019, 248, 1–10.
18. G. Kresse and J. Furthmuller, Phys. Rev. B: Condens. Matter Mater. Phys., 1996, 54, 11169–11186.
19. J. P. Perdew, J. A. Chevary, S. H. Vosko, K. A. Jackson, M. R. Pederson, D. J. Singh and C. Fiolhais, Phys. Rev. B: Condens. Matter Mater. Phys., 1992, 46, 6671–6687.
20. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868.
21. G. Henkelman and H. Jónsson, J. Chem. Phys., 2000, 113, 9978–9985.
22. S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104.
23. C. Cui, H. Wang, X. Zhu, J. Han and Q. Ge, Sci. China: Chem., 2015, 58, 607–613.
24. G. Henkelman, A. Arnaldsson and H. Jónsson, Comput. Mater. Sci., 2006, 36, 354–360.
25. M. Yu and D. R. Trinkle, J. Chem. Phys., 2011, 134, 64111.
26. V. Wang, N. Xu, J. Liu, G. Tang and W. Geng, Comput. Phys. Commun., 2021, 267, 108033.
27. M. Li, S. Zhang, X. Liu, J. Han, X. Zhu, Q. Ge and H. Wang, Eur. J. Inorg. Chem., 2019, 2019, 2058–2064.
28. P. Wen, K. Zhao, H. Li, J. Li, J. Li, Q. Ma, S. M. Geyer, L. Jiang and Y. Qiu, J. Mater. Chem. A, 2020, 8, 2995–3004.
29. Z. Li, X. Zhang, J. Liu, R. Shi, G. I. N. Waterhouse, X. D. Wen and T. Zhang, Adv. Mater., 2021, 33, 2103248.
30. X. Jiang, W. Wang, H. Wang, H. Hong, Y. Yang, K. Wang, L. Zhao and B. Han, Green Chem., 2022, 24, 7652–7660.
31. L. Liu, Y. Qi, J. Yang, W. Cui, X. Li and Z. Zhang, Appl. Surf. Sci., 2015, 358, 319–327.
32. W. Weng, S. Wang, W. Xiao and X. W. D. Lou, Adv. Mater., 2020, 32, 2001560.
33. N. Chen, X. Jia, H. He, H. Lin, M. Guo, J. Cao, J. Zhang and S. Chen, Chin. J. Catal., 2022, 43, 276–287.
34. Y. Yang, P. Zhang, Y. Lei, C. Zhou, J. Liu, S. Guo, S. Li and L. Chen, Int. J. Hydrogen Energy, 2021, 46, 10346–10355.
35. H. Sun, X. Xu, Z. Yan, X. Chen, F. Cheng, P. S. Weiss and J. Chen, Chem. Mater., 2017, 29, 8539–8547.
36. Z. Lin, Y. Zhao, J. Luo, S. Jiang, C. Sun and S. Song, Adv. Funct. Mater., 2019, 30, 1908797.
37. S. Cui, X. Wang, L. Wang and X. Zheng, Dalton Trans., 2021, 50, 5978–5987.
38. C. Chen and J. Wu, ACS Appl. Energy Mater., 2022, 5, 9733–9741.
39. C. Li, Y. Du, D. Wang, S. Yin, W. Tu, Z. Chen, M. Kraft, G. Chen and R. Xu, Adv. Funct. Mater., 2017, 27, 1604328.
40. Y. Guo, Q. Wang, M. Wang, M. Shen, L. Zhang and J. Shi, Catal. Commun., 2021, 156, 106326.
41. A. Wang, C. Wang, L. Fu, W. Wong-Ng and Y. Lan, Nano-Micro Lett., 2017, 9, 47.
42. L. Gao, X. Song, J. Ren and Z. Yuan, Inorg. Chem. Front., 2022, 9, 1964–1972.
43. Y. Yang, P. Zhang, Y. Lei, C. Zhou, J. Liu, S. Guo, S. Li and L. Chen, Int. J. Hydrogen Energy, 2021, 46, 10346–10355.
44. Q. Gai, S. Ren, X. Zheng, W. Liu, Q. Dong and R. Gao, New J. Chem., 2020, 44, 4332–4339.
45. Q. Dong, Y. Fang, Y. Shao, P. Mulligan, J. Qiu, L. Cao and J. Huang, Science, 2015, 347, 967–970.
46. X. Jia, J. Cao, H. Lin, M. Zhang, X. Guo and S. Chen, Appl. Catal., B, 2017, 204, 505–514.
47. Y. Zou, J. Shi, L. Sun, D. Ma, S. Mao, Y. Lv and Y. Cheng, Chem. Eng. J., 2019, 378, 122192.
48. M. Li, S. Zhang, L. Li, J. Han, X. Zhu, Q. Ge and H. Wang, ACS Sustainable Chem. Eng., 2020, 8, 11465–11476.
49. Y. Bai, M. Li, X. Liu, J. Han, X. Zhu, Q. Ge and H. Wang, Ind. Eng. Chem. Res., 2022, 61, 8724–8737.
50. A. Gokarna, N. R. Pavaskar, S. D. Sathaye, V. Ganesan and S. V. Bhoraskar, J. Appl. Phys., 2002, 92, 2118–2124.
51. Y. Tian, W. Tang, H. Xiong, T. Chen, B. Li, X. Jing, J. Zhang and S. Xu, J. Lumin., 2020, 228, 117616.
52. M. Li, H. Wang, X. Li, S. Zhang, J. Han, A. F. Masters, T. Maschmeyer and X. Liu, ChemCatChem, 2018, 10, 581–589.
53. J. Zhou, J. Liang, Z. Fan, W. Tan, X. Sun, R. Ding, P. Gao and Y. Zhang, Int. J. Hydrogen Energy, 2023 DOI:10.1016/j.ijhydene.2023.10.209.
54. Z. Chen, Y. Yu, X. She, K. Xia, Z. Mo, H. Chen, Y. Song, J. Huang, H. Li and H. Xu, Appl. Surf. Sci., 2019, 495, 143528.
55. M. R. Karimi Estahbanati, N. Mahinpey, M. Feilizadeh, F. Attar and M. C. Iliuta, Int. J. Hydrogen Energy, 2019, 44, 32030–32041.
56. H. Li, Q. Song, S. Wan, C. W. Tung, C. Liu, Y. Pan, G. Luo, H. M. Chen, S. Cao, J. Yu and L. Zhang, Small, 2023, 19, e2301711.
57. Z. Gu, Z. Cui, Z. Wang, K. S. Qin, Y. Asakura, T. Hasegawa, S. Tsukuda, K. Hongo, R. Maezono and S. Yin, Appl. Catal., B, 2020, 279, 119376.
58. M. Lei, N. Wang, L. Zhu, Q. Zhou, G. Nie and H. Tang, Appl. Catal., B, 2016, 182, 414–423.
59. K. W. Feng and Y. Li, ChemSusChem, 2023, 16, e202202250.
60. X. Xiong, C. Mao, Z. Yang, Q. Zhang, G. I. N. Waterhouse, L. Gu and T. Zhang, Adv. Energy Mater., 2020, 10, 2002928.
61. C. Lv, L. Xu, M. Chen, Y. Cui, X. Wen, Y. Li, C. Wu, B. Yang, Z. Miao, X. Hu and Q. Shou, Front. Chem., 2020, 8, 32.
62. A. I. Tsiotsias, N. D. Charisiou, I. V. Yentekakis and M. A. Goula, Nanomaterials, 2021, 11, 28.
63. M. Wang, D. Chen, N. Li, Q. Xu, H. Li, J. He and J. Lu, Adv. Mater., 2022, 34, 2202960.
64. M. Muir and M. Trenary, J. Phys. Chem. C, 2020, 124, 14722–14729.
65. J. Sheng, Y. He, J. Li, C. Yuan, H. Huang, S. Wang, Y. Sun, Z. Wang and F. Dong, ACS Nano, 2020, 14, 13103–13114.
66. L. Cheng, X. Yue, L. Wang, D. Zhang, P. Zhang, J. Fan and Q. Xiang, Adv. Mater., 2021, 33, 2105135.
67. Y. Zhang, D. Yao, B. Xia, H. Xu, Y. Tang, K. Davey, J. Ran and S. Qiao, Small Sci., 2021, 1, 2000052.
68. X. Li, Y. Sun, J. Xu, Y. Shao, J. Wu, X. Xu, Y. Pan, H. Ju, J. Zhu and Y. Xie, Nat. Energy, 2019, 4, 690–699.
69. P. Liu, Z. Huang, X. Gao, X. Hong, J. Zhu, G. Wang, Y. Wu, J. Zeng and X. Zheng, Adv. Mater., 2022, 34, 2200057.
70. C. Cui, H. Zhang, R. Cheng, B. Huang and Z. Luo, ACS Catal., 2022, 12, 14964–14975.
71. R. Dronskowski and P. E. Blochl, J. Phys. Chem., 1993, 97, 8617–8624.
72. Y. Li, S. Wang, X. Wang, Y. He, Q. Wang, Y. Li, M. Li, G. Yang, J. Yi, H. Lin, D. Huang, L. Li, H. Chen and J. Ye, J. Am. Chem. Soc., 2020, 142, 19259–19267.
73. P. Liu, L. Niu, Y. Men, C. Peng, Z. Liu, X. Meng and Y. Pan, Appl. Catal., B, 2023, 320, 121985.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d3gc03549b

‡ These authors contributed equally to this work.

---