Confining high-entropy alloys within MOF-derived architectures: a dual-site strategy boosting photothermal CO₂ methanation

Abstract

High-entropy alloys (HEAs) offer immense potential for catalytic transformations, yet they suffer from severe sintering and limited active site accessibility due to their typically dense architectures. Herein, we present a facile molten salt-assisted pyrolysis strategy to confine quinary CoNiCuPtPd HEA nanoparticles within a metal–organic framework (MOF)-derived support. By using UiO-66-NH₂ as a precursor, the introduced amino groups are transformed into pyridinic nitrogen sites during pyrolysis, which, along with the abundant oxygen vacancies generated in the ZrO₂ phase, creates a robust porous support UNH-730. Comprehensive characterization studies (XPS and H₂-TPR) confirm that the pyridinic N sites strongly anchor the metal species, facilitating the formation of homogeneously dispersed HEA nanoparticles (∼15 nm) and preventing agglomeration. In situ DRIFTS and catalytic evaluation reveal a synergistic mechanism: basic N sites and oxygen vacancies significantly enhance CO₂ adsorption/activation, while the HEA sites promote H₂ dissociation. Crucially, we demonstrate that light irradiation functions purely as a thermal source rather than triggering photogenerated electron–hole pairs. Consequently, the optimal CoNiCuPtPd@UNH-730 catalyst delivers a remarkable CH₄ yield of 228.36 mmol g⁻¹ h⁻¹ with 77.8% selectivity at 550 °C under photothermal conditions substantially outperforming the ligand-free counterpart. This work establishes a novel paradigm for stabilizing multi-metal HEAs on functionalized porous carbon/oxide composites for energy conversion applications.

---

Introduction

CO₂ methanation technology is a highly attractive eco-friendly technique, which can not only convert CO₂ into high-value-added fuels but also provide an efficient route for hydrogen (H₂) storage.^1,2 In the field of photothermal catalysis, various catalysts have been developed for CO₂ methanation. For instance, noble metal catalysts (e.g., Ru, Pd, and Rh) exhibit excellent catalytic performance, yet their large-scale application is limited by the exorbitant cost.^3–5 In contrast, some transition metals (e.g., Ni, Co, and Fe) possess superior photothermal conversion efficiency and catalytic activity, emerging as promising alternative materials.^6,7 Among these, Ni-based catalysts are regarded as ideal candidates for CO₂ methanation owing to their high activity, excellent CH₄ selectivity, good stability, and low cost.^8–10 However, Ni-based catalysts are sensitive to reaction temperature as they tend to undergo coking and deactivation at high temperatures, while their activity drops sharply at low temperatures.^11,12 In CO₂ reduction reactions, HEAs have demonstrated remarkable prospects for diverse chemical transformations.^13,14 For example, by integrating metallic copper (Cu) as a cocatalyst with a high-entropy oxide (HEO) precursor, a high-quality heterojunction composite catalyst Cu-(Ga_0.2Cr_0.2Mn_0.2Ni_0.2Zn_0.2)₃O₄ was constructed. Under a 300 W xenon lamp, this Cu-HEO heterojunction exhibited excellent CO₂ reduction activity, with a CH₄ yield reaching 33.84 µmol g⁻¹ h⁻¹.¹⁵ The noble-metal-free FeCoNiCuMn HEA nanoparticles, leveraging the “cocktail effect”, achieved a CH₄ yield of 19.9 µmol g⁻¹ h⁻¹ under sunlight irradiation. The synergistic effect of multiple metals contributes to the enhanced catalytic performance.¹⁶ However, conventional HEA catalysts feature a dense structure,¹⁷ which implies that most potential active sites are embedded in the bulk material, confining catalytic reactions primarily to surface sites.^18,19 Moreover, during high-temperature reactions, the active components tend to migrate, agglomerate, and sinter into a dense structure, which impairs the capability to adsorb and activate gas molecules and thus reduces catalytic efficiency.²⁰ At present, the loading of high-entropy materials relies on supports such as SiO₂ and Al₂O₃.^21,22 These supports can to a certain extent improve the dispersion of the HEA and alleviate issues like migration, agglomeration, and sintering at high temperatures. However, conventional support materials suffer from drawbacks including a single pore structure, lack of basic sites, and weak metal–support interactions. Thus, several studies have explored the use of MOF derivatives as supports for metal loading in catalytic CO₂-to-CH₄ conversion systems. For instance, Li et al.²³ synthesized a MIL-125 (Ti-MOF)-derived Ni–TiO₂ catalyst via a solvothermal method and evaluated its performance in photothermal catalytic CO₂ methanation. The results revealed that at 325 °C under dark conditions, this Ni–TiO₂ catalyst achieved a maximum CH₄ space-time yield of 35.9 mmol g⁻¹ h⁻¹ with a CH₄ selectivity of nearly 99%, which is substantially superior to that of commercial Ni–TiO₂ catalysts. In a related study, Wang et al.²⁴ fabricated MOF-derived Ni/ZrO₂ catalysts with tunable Ni contents for the same photothermal CO₂ methanation reaction. Among these catalysts, the 50Ni/ZrO₂ sample exhibited outstanding catalytic activity, delivering a CH₄ space-time yield of 583.3 mmol g⁻¹ h⁻¹—a value nearly 10 times higher than that of the 50Ni/C–ZrO₂ counterpart under identical irradiation conditions. Despite the remarkable progress achieved by metal–organic framework (MOF)- and MOF-derived supported catalysts in the field of CO₂ hydrogenation, existing MOF-based supported composites are constrained to single-metal or bimetallic metal loading.^25,26 Against this backdrop, the development of a facilely fabricated porous support material featuring multi-metal loading and homogeneous alloying has emerged as an urgent challenge that demands immediate resolution.

Herein, a molten salt-assisted pyrolysis strategy is proposed, where UiO-66-NH₂ is hybridized with KCl–KBr mixtures. Acting as auxiliary agents, KCl–KBr mixtures can in situ form templates in the high-temperature molten pyrolysis atmosphere, effectively suppressing excessive sintering and structural collapse of the UiO-66-NH₂ framework. Subsequent calcination yields a porous support with abundant oxygen vacancies and nitrogen doping, which facilitates the in situ immobilization of quinary metals (Co, Ni, Cu, Pt, and Pd) and formation of a HEA, thus forming the CoNiCuPtPd@UNH catalyst. Owing to the synergistic interplay of oxygen vacancies, basic –NH₂ sites, and multi-metal components, the catalyst achieves efficient light harvesting and photothermal conversion. Strikingly, its catalytic performance under dark conditions (when heated to the equivalent temperature) is identical to that under light illumination, verifying that light functions solely as a thermal source instead of triggering traditional photocatalytic processes dependent on photogenerated electron–hole pairs. As a result, CoNiCuPtPd@UNH demonstrates superior CH₄ production rate and selectivity in the catalytic hydrogenation of CO₂ to methane.

Experimental section

All reagents (analytical grade) were purchased from Aladdin Chemical Co., Ltd (China) and used without further purification: ZrCl₄ (99.99%), zirconium(IV) n-propoxide solution (70%), 2,5-diaminoterephthalic acid (95%), N, N-dimethylformamide (99.8%), acetic acid (99.99%), HCl (37%), KCl (99.5%), KBr (99%), nickel(II) acetylacetonate (95%), copper(II) acetylacetonate (97%), cobalt(III) acetylacetonate (97%), palladium(II) chloride (99.9%), chloroplatinic acid pentahydrate (99.9%), anhydrous ethanol (95%), deionized water, and acetone.

UiO-66 and UiO-66-NH₂ were synthesized following the methods reported by Wang et al.²⁷ and Kandiah et al.,²⁸ respectively. Preparation of UiO-66: acetic acid (2.75 mL) and Zr(OPrⁿ)₄ (0.065 mL) were sequentially added to a three-necked flask containing terephthalic acid (0.025 g) dissolved in DMF (9.75 mL). The reaction flask was sealed and placed in an isothermal oven at 65 °C overnight. The solid was collected by centrifugation (8000 rpm, 5 min), followed by soaking in DMF (5 mL, 3 times) and acetone (5 mL, 3 times) for 24 hours each. Finally, UiO-66 was dried under dynamic vacuum at room temperature overnight. Preparation of UiO-66-NH₂: typically, 1.75 g of ZrCl₄, 1.06 g of 2-aminoterephthalic acid, and 3 mL of acetic acid were dissolved in 80 mL of DMF and placed in a 100 mL three-necked flask. The mixed solution was reacted in an oil bath at 120 °C for 6 h. After cooling, the precipitate was washed four times with an appropriate amount of methanol and acetone respectively, and dried overnight under vacuum at 125 °C.

The as-prepared UiO-66 was uniformly blended with KCl–KBr mixed salts at a weight ratio of UiO-66 : KCl : KBr = 1 : 1 : 35. The mixture was subsequently transferred to a tube furnace, heated to 730 °C at a ramp rate of 5 °C min⁻¹, and held for 3 h under a continuous N₂ atmosphere. The product was rinsed sequentially with 0.2 M aqueous HCl and deionized water, followed by drying at 75 °C, affording a black solid which was designated as UiO-730.

The as-prepared UiO-66-NH₂ was uniformly mixed with KCl–KBr mixed salts at a weight ratio of UiO-66-NH₂ : KCl : KBr = 1 : 1 : 35. The mixture was then placed in a tube furnace, heated to 730 °C at a ramp rate of 5 °C min⁻¹, and maintained for 3 h under a constant N₂ atmosphere. The resulting black product was rinsed thoroughly with 0.2 M aqueous HCl and deionized water, and subsequently designated as UNH-730.

Nickel acetylacetonate (0.0266 g), copper acetylacetonate (0.0271 g), and cobalt acetylacetonate (0.0266 g) were dissolved in 4 mL of a 1 : 1 (v/v) ethanol/water mixture. After stirring for 30 minutes, palladium chloride (0.0184 g) and chloroplatinic acid pentahydrate solution (0.53 mL) were added, followed by continuous stirring for another 30 minutes. Subsequently, 0.5 g of UiO-730 support was added, and the mixture was stirred overnight at 200–300 rpm. The resulting mixture was stirred to dryness at 70 °C, dried in an oven at 80 °C for 12 hours, heated to 300 °C at a rate of 2 °C min⁻¹ under a N₂ atmosphere and held for 2 hours, and finally heated to 650 °C at 5 °C min⁻¹ under a pure H₂ flow (60 mL min⁻¹) and held for 3 hours. The product was named CoNiCuPtPd@UiO-730.

Nickel acetylacetonate (0.0266 g), copper acetylacetonate (0.0271 g), and cobalt acetylacetonate (0.0266 g) were dissolved in 4 mL of an ethanol/water mixed solution (volume ratio 1 : 1). After stirring for 30 minutes, palladium chloride (0.0184 g) and chloroplatinic acid pentahydrate solution (0.53 mL) were added, followed by continuous stirring for another 30 minutes. Subsequently, 0.5 g of UiO66-NH₂-730 support was added, and the mixture was stirred overnight at 200–300 rpm. The resulting mixture was stirred to dryness at 70 °C, dried in an oven at 80 °C for 12 hours, heated to 300 °C at a rate of 2 °C min⁻¹ under a N₂ atmosphere and held for 2 hours, and finally heated to 650 °C at 5 °C min⁻¹ under a pure H₂ flow (60 mL min⁻¹) and held for 3 hours. The product was named CoNiCuPtPd@UNH-730.

Characterization methods

The phase composition of the catalysts was analyzed by X-ray powder diffraction (XRD, Rigaku Ultima IV) with Cu Kα radiation (40 kV, 30 mA, λ = 0.15 nm), operating at a 2θ scanning range of 5–90° and a scanning rate of 5° min⁻¹. The size and morphology of the materials were characterized by transmission electron microscopy (TEM). Brunauer–Emmett–Teller (BET) specific surface area and pore size measurements were performed on a Micromeritics ASAP 2020M instrument at 77 K. The chemical states and valence distribution of elements were evaluated by X-ray photoelectron spectroscopy (XPS, Thermo Scientific K-Alpha), calibrated to the C 1s peak at 284.8 eV. Fourier transform infrared spectroscopy (FTIR) was used to identify the functional groups in the materials. Hydrogen temperature-programmed reduction (H₂-TPR), hydrogen temperature-programmed desorption (H₂-TPD), carbon monoxide temperature-programmed desorption (CO-TPD), methane temperature-programmed desorption (CH₄-TPD), and carbon dioxide temperature-programmed desorption (CO₂-TPD) tests were carried out on a Quanta chrome chemisorption analyzer, with the detailed procedures as follows: H₂-TPD: 50 mg of the sample was first pretreated with a 5% H₂/Ar mixture (30 mL min⁻¹) at 300 °C for 30 min, then switched to an Ar flow (30 mL min⁻¹) and maintained at 400 °C for another 30 min. After cooling the sample to 40 °C, a 5% H₂/Ar flow (30 mL min⁻¹) was introduced for 1 h to ensure complete adsorption, followed by purging with Ar (30 mL min⁻¹, 1 h) to remove the physiosorbed hydrogen. The H₂-TPD was performed under an Ar atmosphere from 40 °C to 650 °C at a heating rate of 10 °C min⁻¹, and the desorbed hydrogen was detected by a thermal conductivity detector (TCD). CH₄-TPD, CO-TPD, and CO₂-TPD: similar procedures were adopted, with the sample first in situ reduced with 5% H₂/Ar at 300 °C and then cooled by a He flow, using CH₄, CO, and CO₂ as the adsorbate gases, respectively. H₂-TPR: 50 mg of the sample was first treated in an Ar flow (30 mL min⁻¹) at 400 °C for 30 min, then reduced in a H₂ atmosphere from 50 °C to 650 °C at a heating rate of 5 °C min⁻¹, and held at 650 °C for 30 min. Diffuse reflectance spectroscopy (DRS) was performed on a Varian Cary 5000 UV-Vis spectrophotometer using BaSO₄ as the reference standard. Electron paramagnetic resonance (EPR) spectra were measured on a Bruker A300 spectrometer at liquid nitrogen temperature, with 10 mg of sample powder loaded into a sample tube for characterization. In situ DRIFT analysis: in situ DRIFT measurements were conducted under thermal conditions on a Thermo Fisher iS50 spectrometer equipped with an MCT/A detector to investigate reaction intermediates. Prior to the experiment, CoNiCuPtPd@UNH-730 was pretreated in 5% H₂/Ar for 30 minutes, followed by cooling to 550 °C and collecting background spectra under a N₂ flow. A gas mixture of 10% CO₂/40% H₂/50% He was introduced into the reaction cell for 15 minutes. Once the gas mixture was fully adsorbed on the sample, the gas supply was terminated, and the exhaust gas was analyzed using a gas chromatography-mass spectrometry (GC-MS) system (Agilent GC 7890-MS 5975C). All density functional theory (DFT) calculations were performed using the Vienna Ab initio Simulation Package (VASP) within the generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE) functional. Projected augmented wave (PAW) potentials were used to describe the ionic cores, and valence electrons were treated with a plane wave basis set employing a kinetic energy cutoff of 450 eV. The DFT-D3 empirical correction method was employed to describe van der Waals interactions. Geometry optimizations were performed with the force convergence smaller than 0.02 eV Å⁻¹ and the energy convergence of 1 × 10⁻⁵ eV. Spin-polarization effect was also considered. Besides, a (1 × 1 × 1) k-point sampling with the Gamma-centered scheme was used for the Brillouin zone integration. All the atoms are relaxed in the calculation.

CO₂ methanation experiments

The photothermal catalytic reaction was carried out in a fixed-bed reactor at atmospheric pressure, and the device schematic is shown in Fig. 1. A 300 W CEL-HXF-T3 xenon lamp (simulating sunlight) was used as a power-adjustable light source, and the light was directly irradiated onto the catalyst surface through the inlet hole. A thermocouple and a TC-1000 temperature controller (JASCO) were used to precisely control the catalyst temperature to maintain the desired reaction temperature. The reactor was a custom-designed quartz tube with a sheet-like central configuration, aiming to enhance the uniform light distribution across the entire catalyst bed. 10 mg of the catalyst was loaded into the quartz tube, and an external thermocouple was placed on the catalyst surface to detect the actual temperature. Before the reaction started, a mixed gas (10% CO₂/40% H₂/50% Ar) was introduced for 10 minutes to purge the air inside the reactor.

Fig. 1 Schematic diagram of the reaction device.

Without illumination, a temperature-increasing program (350–650 °C, heating rate of 5 °C min⁻¹) was set via the fixed-bed temperature controller, with the temperature detected by the thermocouple on the catalyst surface as the reference. After raising the temperature by 50 °C and maintaining it for 30 minutes each time, the reaction off-gas was collected and analyzed by a gas chromatograph (GC) equipped with a thermal conductivity detector (TCD) and a flame ionization detector (FID). With illumination, after turning on the light source and stabilizing for 30 minutes, the actual temperature of the catalyst surface was approximately 69 °C higher than the set temperature of the fixed-bed reactor, as detected by the external thermocouple. The fixed-bed temperature controller was adjusted to make the catalyst surface temperature consistent with the target temperature under the non-illumination conditions, and then the reaction off-gas was collected and analyzed by the same GC mentioned above.

Results and discussion

The crystal structures of UiO-66 and UiO-66-NH₂ were characterized by X-ray diffraction (XRD) (Fig. S1). As shown in Fig. 2a, the diffraction peaks of UiO-730 and UNH-730 match those of the quaternary ZrO₂ phase (JCPDS 96-210-0390, t-ZrO₂), with diffraction angles (2θ) at 30.2°, 35.1°, 50.2°, 60.0°, 62.8°, 74.5°, 81.7°, and 84.9°. Previous reports have indicated that the presence of tetragonal ZrO₂ contributes to enhancing the photothermal catalytic performance for CO₂ methanation.²⁹ From the perspective of high-entropy alloy formation principles, the CoNiCuPtPd system was selected. Its atomic size difference δ ≈ 5.17%, mixing enthalpy ΔH_mix ≈ −1 to −3 kJ mol⁻¹ (within the stable range of −15 to 5 kJ mol⁻¹), and valence electron concentration VEC = 10 all meet the formation criteria for a single-phase FCC structure. Furthermore, Ni, Cu, Pt, and Pd possess an FCC structure, and Co can transform into FCC to achieve crystal structure matching. Subsequent characterization studies including XRD and TEM fully verified this theoretical prediction.^30,31 From the perspective of catalytic functional complementarity, metallic Co and Ni act as active sites for CO₂ reduction and can activate CO₂via electron enrichment.³² Cu weakens surface adsorption and avoids excessive binding of intermediates, which favors the subsequent formation of CH₄.³³

Fig. 2 (a) XRD patterns of UiO-66, UiO-66-NH₂, CoNiCuPtPd@UiO-730, and CoNiCuPtPd@UNH-730; (b) N₂ adsorption–desorption isotherms of UiO-66, UiO-66-NH₂, CoNiCuPtPd@UiO-730, and CoNiCuPtPd@UNH-730; (c) pore size distribution curves of UiO-66, UiO-66-NH₂, CoNiCuPtPd@UiO-730, and CoNiCuPtPd@UNH-730.

The characteristic peaks of CoNiCuPtPd@UNH-730 at 2θ = 41.7°, 48.4°, 70.8°, and 85.7° are assigned to the (111), (200), (220), and (311) crystal planes of the CoNiCuPtPd HEA, respectively, confirming the successful formation and effective loading of the HEA phase.³⁴ For the post-reaction CoNiCuPtPd@UNH-730, no shift was observed in the peak positions of the HEA characteristic peaks (2θ = 41.7°, 48.4°, etc.), indicating that the alloy particles remained stable after the reaction (Fig. S2). The N₂ adsorption–desorption curves from BET tests (Fig. 2b) show that both UiO-66 and UiO-66-NH₂ possess high specific surface areas, which are 1389.66 m² g⁻¹ and 826.21 m² g⁻¹, respectively. During the high-temperature pyrolysis, the KCl–KBr molten salt acts as a template to prevent excessive sintering of the carbon skeleton.³⁵ Nevertheless, the MOF still undergoes a certain degree of framework carbonization, leading to a sharp decrease in the specific surface area of the pyrolyzed material, indicating partial collapse of the framework after pyrolysis. Pyrolyzed UiO-66 is denoted as UiO-730, and UiO-66-NH₂ as UNH-730. BET characterization shows that the specific surface area of CoNiCuPtPd@UiO-730 (149.98 m² g⁻¹) is smaller than that of CoNiCuPtPd@UNH-730 (163.36 m² g⁻¹). This is because, although the presence of amino groups in UiO-66-NH₂ reduces the specific surface area of the original MOF due to steric hindrance, the Zr spectrum of CoNiCuPtPd@UNH-730 in XPS characterization shows a negative shift of ∼1.1 eV in Zr binding energy, confirming that the electron-donating effect of amino groups promotes weak interactions between N atoms in –NH₂ and metal centers (Zr⁴⁺), leading to pyrolysis via gentle decomposition to retain the pore structure. Meanwhile, N-doping creates pores to form porous carbon and inhibits agglomeration by anchoring Zr.^36,37 Additionally, pore size distribution diagrams (Fig. 2c) show that the average pore sizes of UiO-66 and UiO-66-NH₂ are both ∼1.7 nm, while CoNiCuPtPd@UiO-730 (average pore size 6.03 nm) and CoNiCuPtPd@UNH-730 (average pore size 3.78 nm) indicate that after pyrolysis and multi-metal loading, the overall pore size of the materials is significantly increased, which is beneficial for mass transfer in the CO₂ hydrogenation reaction.

SEM images of UiO-66, UiO-66-NH, pyrolyzed UiO-730, UNH-730, and post-reaction CoNiCuPtPd@UiO-730 and CoNiCuPtPd@UNH-730 are shown in Fig. S3. TEM images of the as-synthesized UiO-66 (Fig. 3a) and UiO-66-NH₂ (Fig. 3d) both show distinct octahedral cubic structures. The average particle size of UiO-66 is 413 nm, while that of UiO-66-NH₂ is 327 nm. This is because the electronic effect of the amino groups accelerates the nucleation process, resulting in UiO-66-NH₂ typically having a smaller particle size than UiO-66, which is consistent with previous literature reports.³⁸ Transmission electron microscopy (TEM) images and energy-dispersive X-ray spectroscopy (EDS) analysis (Fig. 3b) demonstrate that UiO-730 retains the morphology of the original MOF well after pyrolysis and loading. As the metal center of UiO-66, Zr element remains uniformly distributed after calcination, indicating that UiO-730 preserves part of the Zr-related structure. It is observed that the multi-metals do not form a single solid solution phase, with a lower metal loading relative to CoNiCuPtPd@UNH-730, and the metal particle size is displayed in Fig. 3c. This is because the organic ligand (terephthalic acid, BDC) in UiO-66 carbonizes to form amorphous carbon, while the metal nodes (Zr) are converted to zirconia (ZrO₂), resulting in a ZrO₂/carbon composite. TEM images of UiO-66-NH₂ after pyrolysis and metal loading (Fig. 3e) show that Zr, its metal center, remains homogeneously distributed post-calcination, confirming that UNH-730 also retains partial Zr-related structures. Energy spectra of C and O reveal the homogeneous distribution of C, verifying that the carbon matrix derived from pyrolyzed MOF ligands is agglomeration-free. The carbon skeleton, as the support matrix, is uniformly dispersed around and in the interstices of ZrO₂ nanoparticles. The homogeneous distribution of O overlaps with the Zr signal, validating the homogeneous dispersion of the ZrO₂ phase with Zr serving as the support framework. Although UNH-730 also forms a ZrO₂/carbon composite during calcination due to the introduction of amino groups, the key difference lies in that the –NH₂ functional groups on the ligand participate in the reaction during high-temperature pyrolysis to form a nitrogen-doped composite when compared with UiO-730. The N 1s spectrum in XPS confirms that the formed nitrogen is pyridinic nitrogen, which can provide lone pair electrons and form strong coordination interactions with metal ions to better anchor metals and inhibit metal migration and agglomeration.³⁹ Therefore, it can be observed that the five metal elements in CoNiCuPtPd@UNH-730 forms a single solid solution (Fig. 3e). The EDS image confirms that this single solid solution phase is uniformly composed of the five metals, which is consistent with the XRD characterization result indicating the formation of a HEA by CoNiCuPtPd. Furthermore, the alloy particles are uniformly dispersed in the UNH-730 matrix without obvious agglomeration, with an average particle size of 15.36 nm (Fig. 3f). After the reaction, the supports of CoNiCuPtPd@UiO-730 and CoNiCuPtPd@UNH-730 still retain the MOF-derived ZrO₂/carbon composite structure.

Fig. 3 (a) TEM image of UiO-66; (b) TEM image and EDS elemental mapping images of CoNiCuPtPd@UiO-730; (c) particle size distribution diagram of CoNiCuPtPd@UiO-730; (d) TEM image of UiO-66-NH₂; (e) TEM image and EDS elemental mapping images of CoNiCuPtPd@UNH-730; (f) particle size distribution diagram of CoNiCuPtPd@UNH-730.

However, the Co, Ni, Pt, Pd, and Cu elements in CoNiCuPtPd@UiO-730 show obvious local segregation, and the metal particles are significantly agglomerated, with the particle size increased by about 3.47 nm. The HEA particles in CoNiCuPtPd@UNH-730 are spherical and uniformly dispersed in the support matrix with no obvious agglomeration. Compared with the particle size of 15.36 nm before the reaction, the difference is only 1.28 nm (Fig. S4).

X-ray photoelectron spectroscopy (XPS) was employed to analyze the electronic structures and valence states of fresh CoNiCuPtPd@UiO-730 and CoNiCuPtPd@UNH-730 (Fig. 4). For CoNiCuPtPd@UNH-730 (Fig. 4a), the main peak in the C 1s spectrum at 284.8 eV is attributed to the C–C bonds of the organic carbon skeleton in the material; the peak around 286–287 eV corresponds to C–O bonds, derived from incompletely decomposed oxygen-containing organic functional groups; the peak near 289 eV is assigned to O=C–OH bonds, indicating the presence of a small amount of oxygen-containing functional groups such as carboxyl groups on the material surface; hydroxyl groups (–OH), acting as weak basic sites, can interact with CO₂ molecules to facilitate the adsorption of CO₂ on the catalyst surface.⁴⁰ The main peak in the N 1s spectrum at ∼400 eV is attributed to C–N bonds, which are the characteristic peaks of nitrogen–carbon bonds retained by amino structures after calcination. The peak around 397–398 eV in the N 1s spectrum corresponds to pyridinic nitrogen, which is formed when nitrogen atoms participate in the formation of pyridine-structured heterocycles or coordinate with metals during high-temperature calcination and loading. This species can donate lone-pair electrons and form strong coordination interactions with metal ions, thereby effectively anchoring the metals and inhibiting their migration and agglomeration.⁴¹ The main peak in the O 1s spectrum at 529.08 eV is related to the lattice oxygen of ZrO₂, and the peak at 530.58 eV is attributed to surface-adsorbed oxygen species (e.g., O⁻) near oxygen vacancies (OVs).^42,43 Such oxygen species are directly associated with OVs and serve as key active sites for CO₂ adsorption and activation. Furthermore, oxygen vacancies greatly reduce the reaction activation energy by optimizing the adsorption energy of intermediates and the efficiency of electron transfer. Studies have shown that oxygen vacancies significantly lower the dissociation barrier of CO₂ on the ZrO₂ surface: the surface barrier of the t-ZrO₂(101) defect decreases from 1.713 eV to 0.352 eV (to 1/4.9 of the original value). OVs also exist in CoNiCuPtPd@UiO-730 as confirmed by EPR characterization. The peak at 532.58 eV corresponds to ordinary surface-adsorbed oxygen.^44,45

Fig. 4 (a) Elemental XPS spectra of CoNiCuPtPd@UNH-730; (b) elemental XPS spectra of CoNiCuPtPd@UiO-730.

In the Zr 3d spectrum, the peaks at ∼180–181 eV (Zr 3d_5/2) and ∼183–184 eV (Zr 3d_3/2) are assigned to Zr⁴⁺. Compared with CoNiCuPtPd@UiO-730, the Zr spectrum peaks of CoNiCuPtPd@UNH-730 show a negative shift of ∼1.1 eV in binding energy. This is because amino groups, as electron-donating groups, alter the coordination microenvironment of Zr in the MOF framework.⁴⁶ In the Co 2p spectrum, the peaks at Co 2p_3/2 binding energy of 781–782 eV and Co 2p_1/2 binding energy of ∼797–798 eV correspond to the characteristic peaks of Co²⁺ indicating the presence of divalent cobalt ions in the sample, which may exist in the form of oxides (e.g., CoO). The peaks at binding energies of ∼778–779 eV (Co 2p_3/2) and ∼793–794 eV (Co 2p_1/2) represent metallic Co⁰, indicating that part of the cobalt was reduced to the elemental form during the loading process. For Ni 2p, peaks at 855–856 eV (Ni 2p_3/2) and 873–874 eV (Ni 2p_1/2) indicate the presence of Ni²⁺, possibly as NiO. At lower binding energies of ∼852–853 eV (Ni 2p_3/2) and ∼869–870 eV (Ni 2p_1/2), additional peaks are assigned to metallic Ni⁰, suggesting that part of the supported nickel is in the zero-valent state. In the Cu 2p spectrum, peaks at ∼932–933 eV (Cu 2p_3/2) and ∼952–953 eV (Cu 2p_1/2) are characteristic of metallic Cu⁰, confirming the zero-valent state of the supported copper. The Pt 4f region shows peaks at ∼71–72 eV (Pt 4f_7/2) and ∼74–75 eV (Pt 4f_5/2), which are indicative of metallic Pt⁰, demonstrating that the supported platinum is in the zero-valent state. For Pd 3d, the peaks at ∼335–336 eV (Pd 3d_5/2) and ∼340–341 eV (Pd 3d_3/2) correspond to metallic Pd⁰, indicating that the supported palladium is in the zero-valent state. Collectively, the X-ray photoelectron spectroscopy (XPS) spectra of the Co 2p, Ni 2p, Cu 2p, Pt 4f, and Pd 3d regions in the Co–Ni–Cu–Pt–Pd alloy all exhibit characteristic binding energies of the metallic (0) valence state, indicating the successful formation of the HEA without significant oxidation. After the reaction, Co⁰, Ni⁰, Cu⁰, Pt⁰, and Pd⁰ remain the dominant valence states in CoNiCuPtPd@UNH-730, with no obvious peak shift or oxidation-state peaks observed, indicating no significant oxidation. The intensity of the characteristic peak at 397–398 eV in the N 1s spectrum shows no obvious attenuation, suggesting that the pyridinic N sites formed by the pyrolysis of amino groups remained structurally intact during the reaction. The surface-adsorbed oxygen peak at 530.58 eV in the O 1s spectrum also shows no obvious change, confirming the stable concentration of oxygen vacancies (Fig. S5). Compared with CoNiCuPtPd@UNH-730, the XPS spectrum of CoNiCuPtPd@UiO-730 (Fig. 4b) shows that Cu and Pd only exhibit a zero-valent state, with only surface oxygen but no lattice oxygen in the material, while Co, Ni, and Pt all exist in oxidized states. This confirms that N in the –NH₂ groups is the core reason for the complete reduction of metals. First, the lone pair electrons on N atoms can form strong coordination interactions with metal cations, constructing stable metal–nitrogen bonds. This “anchoring effect” firmly fixes metal ions on the surface and defect sites of the carbon skeleton, significantly inhibiting their migration and agglomeration during reduction. Notably, highly dispersed metal ions are a prerequisite for complete reduction. Second, nitrogen-doped sites can act as coordination centers to form stable coordination bonds with metal ions. The nitrogen moieties, with their lone-pair electrons, can engage in electronic interactions with metal species, thereby reducing the reduction energy barrier of metal ions. Meanwhile, the electronegativity of nitrogen enables modulation of the electronic states of metals, facilitating their conversion to lower valence states.⁴⁷ Full survey spectra and N 1s X-ray photoelectron spectroscopy (XPS) spectra of CoNiCuPtPd@UiO-730 and CoNiCuPtPd@UNH-730 are shown in Fig. S6.

The ultraviolet-visible-near infrared diffuse reflectance spectroscopy (UV-Vis-NIR DRS) results are shown in Fig. 5a. In the full spectral range, CoNiCuPtPd@UNH-730 exhibits stronger light absorption performance compared with CoNiCuPtPd@UiO-730, which provides more favorable conditions for photothermal conversion during the reaction. In the ultraviolet-visible (UV-Vis) spectral range, CoNiCuPtPd@UNH-730 demonstrates exceptionally strong light-harvesting capability. This superior optical performance originates from two key structural and electronic effects. First, the high-temperature carbonization of the MOF precursor produces a hybrid carbon matrix with abundant delocalized π-electron systems, which substantially augments the probability and intensity of π–π* electronic transitions. Second, the incorporation of pyridinic nitrogen via doping elevates the electronic density of states of the graphitized carbon component, leading to a notable enhancement in its visible-light response capacity.⁴⁸

Fig. 5 (a) DRS spectra of CoNiCuPtPd@UiO-730 and CoNiCuPtPd@UNH-730; (b) EPR spectra of CoNiCuPtPd@UiO-730 and CoNiCuPtPd@UNH-730.

The electron paramagnetic resonance (EPR) spectroscopy results are presented in Fig. 5b, with distinct characteristic signals of oxygen vacancies observed at g = 2.04 for CoNiCuPtPd@UNH-730 and g = 2.03 for CoNiCuPtPd@UiO-730, respectively. This result is in excellent agreement with the O 1s spectral analysis from X-ray photoelectron spectroscopy (XPS), which not only verifies the existence of oxygen vacancies in both catalysts but also reveals that CoNiCuPtPd@UNH-730 has a higher density of oxygen vacancies. The mechanism accounting for this difference in oxygen vacancy abundance is as follows: during the high-temperature pyrolysis of the MOF precursor nitrogen doping tailors the electronic configuration of the carbon matrix, thereby boosting its electron-donating ability. As a consequence, carbon species in the nitrogen-doped carbon framework engage in redox reactions with lattice oxygen in the ZrO₂ phase, sequestering lattice oxygen to achieve self-oxidation. The substantial depletion of oxygen atoms from the ZrO₂ lattice then induces the formation of abundant oxygen vacancies within its crystalline lattice structure.⁴⁹

The results of Fourier transform infrared (FT-IR) spectroscopy are displayed in Fig. 6. The absorption peaks at 499.5 cm⁻¹, 593.3 cm⁻¹, 533.8 cm⁻¹, 619.4 cm⁻¹, and 739.2 cm⁻¹ are associated with metal–ligand bonds in the materials. The absorption peak of CoNiCuPtPd@UiO-730 at 1586.4 cm⁻¹ is attributed to carbon–carbon double bonds, while the vibrational absorption peak at 1625.9 cm⁻¹ arises from the superposition of absorptions of carbon–carbon double bonds and carbon–nitrogen double bonds. The presence of a characteristic absorption peak in the 3432.7 cm⁻¹ region for CoNiCuPtPd@UNH-730 corresponds to the N–H stretching vibration of residual amino functional groups after the calcination of UiO-66-NH₂, a phenomenon confirming that CoNiCuPtPd@UNH-730 retains amino functional groups even after high-temperature pyrolysis.⁵⁰ Compared with the carbon–carbon double bond absorption peak of CoNiCuPtPd@UiO-730 at 1586.4 cm⁻¹, the carbon–carbon double bond vibrational absorption peak of CoNiCuPtPd@UNH-730 at 1625.9 cm⁻¹ shows a certain degree of enhancement and shift. This is because nitrogen atoms from the amino groups in UiO-66-NH₂ are doped into the carbon skeleton during calcination, altering the electron cloud distribution of the carbon skeleton and promoting the formation of carbon–nitrogen double bonds, which further induces the superposition effect of vibrational absorptions of functional groups.⁵¹

Fig. 6 FT-IR spectra of CoNiCuPtPd@UNH-730 and CoNiCuPtPd@UiO-730.

The results of temperature-programmed reduction (H₂-TPR) measurements are displayed in Fig. 7. The temperature corresponding to the main reduction peak of CoNiCuPtPd@UNH-730 is 450 °C, which shifts to the high-temperature region by 100 °C compared with that of CoNiCuPtPd@UiO-730 at 350 °C. This is because pyridinic nitrogen can form stable metal–N coordination bonds with metal cations via lone-pair electrons. Although this anchoring effect can suppress metal agglomeration, it converts the reduction reaction from a single process of breaking metal–oxygen bonds into a two-step process involving the cleavage of metal–oxygen bonds and the rupture of metal–nitrogen bonds, thus increasing the total energy required for the reaction. Characterization results from X-ray diffraction (XRD) and transmission electron microscopy (TEM) demonstrate that the polymetallic species in CoNiCuPtPd@UNH-730 have formed a pentametallic HEA. The alloying process requires the migration, diffusion and agglomeration of reduced metal atoms, which needs to overcome the diffusion energy barrier of metal atoms on the support surface and can only be achieved at higher temperatures. This process is further coupled with the metal reduction process, ultimately leading to the shift of the main reduction peak to the high-temperature region.⁵²

Fig. 7 H₂-temperature-programmed reduction (H₂-TPR) profile.

Carbon dioxide temperature-programmed desorption (CO₂-TPD) tests were performed (Fig. 8a), and the results show that both CoNiCuPtPd@UiO-730 and CoNiCuPtPd@UNH-730 exhibit two desorption peaks, located at ∼150 °C and ∼500 °C, indicating that both materials possess CO₂ adsorption performance. Overall, the desorption peak intensity and peak area of CoNiCuPtPd@UNH-730 are larger than those of CoNiCuPtPd@UiO-730, demonstrating that the former has a stronger adsorption capacity for CO₂. The CO₂ adsorption performance of CoNiCuPtPd@UiO-730 stems from the small amount of oxygen vacancies and metal cations on the support. The Lewis acid metal cations adsorb CO₂ through electrostatic interactions with the oxygen atoms in CO₂ molecules. For CoNiCuPtPd@UNH-730, the desorption peak in the low-temperature region (<300 °C) indicates that CO₂ forms stable chemically adsorbed species on the material surface, such as bidentate carbonates, monodentate carbonates, or activated CO₂-species adsorbed on oxygen vacancies. The reasons for its enhanced CO₂ chemical adsorption can be attributed to a key factor. Specifically, the lone pair electrons of N atoms can form coordination bonds with the polar carbon atoms of CO₂ molecules to promote CO₂ chemical adsorption.^53,54

Fig. 8 (a) CO₂-temperature-programmed desorption (CO₂-TPD) profile; (b) H₂-temperature-programmed desorption (H₂-TPD) profile; (c) CO-temperature-programmed desorption (CO-TPD) profile; (d) CH₄-temperature-programmed desorption (CH₄-TPD) profile.

The strong CO₂ adsorption performance in the high-temperature region (>300 °C) benefits from the abundant oxygen vacancies in CoNiCuPtPd@UNH-730 confirmed by EPR characterization—which is also the key reason for the material's stronger CO₂ adsorption capacity. Strong adsorption and activation are prerequisites for all subsequent reactions, and a high initial activation degree determines a high reactant concentration. In the H₂ temperature-programmed desorption (H₂-TPD) profile (Fig. 8b), CoNiCuPtPd@UNH-730 exhibits remarkable H₂ adsorption affinity, signifying that the high-entropy alloy (HEA) particles can proficiently dissociate H₂ into active hydrogen atoms (*H), thereby facilitating the rapid hydrogenation of reactants.

The CO₂-TPD and H₂-TPD tests reveal the reason why CoNiCuPtPd@UNH-730 has a higher CH₄ yield from the perspective of reactant concentration. The CO-temperature-programmed desorption (CO-TPD) curve (Fig. 8c) shows that CoNiCuPtPd@UNH-730 has a stronger CO adsorption capacity compared with CoNiCuPtPd@UiO-730. This illustrates that CO* persists longer on the catalyst surface, furnishing an ample time window for the subsequent hydrogenation of intermediates such as CHO* and CH₃O* while precluding the desorption and loss of CO* (as byproduct CO) prior to complete hydrogenation.⁵⁵ Furthermore, the desorption peak of CoNiCuPtPd@UNH-730 exhibits a certain shift toward the high-temperature region. This stems from the formation of a HEA by the five metals, which constructs “multimetallic synergistic adsorption sites”. This unique structure induces significant lattice distortion and slow diffusion effects, leading to the upward shift of the metal d-band center. This further strengthens the overlap and interaction between the metal surface and the antibonding orbitals of molecules such as CO, thereby enhancing the chemical adsorption energy. In contrast, CO* on CoNiCuPtPd@UiO66-730 is prone to desorption, rendering it more inclined toward the reverse water–gas shift reaction.⁵⁶

This results in lower selectivity for methanation and increased formation of byproduct CO. CO-TPD elucidates the superior CH₄ yield and selectivity of CoNiCuPtPd@UNH-730 from the perspective of the reaction process. The CH₄-temperature-programmed desorption (CH₄-TPD) curve (Fig. 8d) shows that the material has a high CH₄ desorption peak intensity, and its temperature range highly overlaps with the reaction ranges of CO₂-TPD, H₂-TPD, and CO-TPD.

Prior to the experiment, the feed gas mixture (10% CO₂/40% H₂/50% Ar, volumetric ratio) was sampled from the cylinder and analyzed by gas chromatography (GC). The volumetric concentrations of CO₂ and H₂ were determined to be 9.88 vol% and 39.61 vol% respectively, consistent with the designed ratio. A blank experiment was first performed by feeding the gas mixture into an empty reactor without the catalyst at a flow rate of 25 mL min⁻¹. The temperature was raised from room temperature to 550 °C and tail gas samples were collected from the reactor outlet every 5 minutes for analysis. No gaseous components other than CO₂ and H₂ were detected, ruling out interference from the pipeline and reactor material. Subsequently, 10 mg of the catalyst was loaded into the reactor and the experiment was repeated. Within the first 10 minutes, only the reactants CO₂ and H₂ were detected by GC with concentrations nearly identical to the initial values, indicating that only reactant adsorption occurred during this period without substantial reaction or reactant consumption. At 10 minutes, CH₄ and CO emerged with yields of 106.36 mmol g⁻¹ h⁻¹ and 73.16 mmol g⁻¹ h⁻¹, respectively (Fig. 9a), while the CH₄ selectivity and CO₂ conversion reached 59.3% and 29.33%, respectively (Fig. 9b). As the reaction proceeded, the CH₄ formation rate increased while the CO formation rate decreased, with continuous elevation of CH₄ selectivity and CO₂ conversion corresponding to the stage of catalyst active site activation and establishment of reactant adsorption–conversion equilibrium. All parameters stabilized after reaching their maximum values at 30 minutes, indicating that the system entered a dual thermodynamic and kinetic steady state where catalyst activity, reactant/product concentrations and surface intermediate conversion rates were in dynamic equilibrium. In summary, the reaction involves an adsorption stage before 10 minutes and a main evolution stage between 10 and 30 minutes, and achieves dynamic equilibrium after 30 minutes.

Fig. 9 (a) Rate of CH₄ and CO; (b) selectivity of CH₄ and conversion of CO₂; (c) in situ infrared spectrum of CoNiCuPtPd@UNH-730. During the reaction, a gas mixture of CO₂/H₂ (1 : 4) was used (25 mL min⁻¹), and the HEA particles can efficiently dissociate H₂; (e) free energy profile of key reaction intermediates.

In situ infrared spectroscopy characterization was conducted based on the aforementioned experimental conditions. In situ DRIFTS measurements (Fig. 9c) were conducted under a continuous flow of H₂/CO₂ mixture to identify the involved intermediates and clarify the CO₂ reaction pathway. After introducing the CO₂ : H₂ (1 : 4) mixture at the experimental temperature and stabilizing for 10 minutes, the concurrent enhancement of intermediate/product peaks (HCO-3, *CO, CH₃O⁻, and CH₄) and the stabilization/weakening of precursor peaks (carbonate and COOH*) were observed. This phenomenon indicates that the reaction system is in a dynamic steady state, where the continuous supply of gaseous CO₂ and H₂ sustains the formation of initial adsorbed species and the continuous generation of intermediates (HCO-3, CH₃O⁻, and *CO), which are sequentially hydrogenated to CH₄.^57,58 The peak intensity variation of each species is determined by the relative magnitude of formation and consumption rates, with this balance essentially regulated by the material's structural features (alloying, oxygen vacancies, and Co/Ni mixed valence states) and reaction thermodynamics/kinetics. Characteristic bands of monodentate carbonate (1507 cm⁻¹) and bidentate carbonate (1546 cm⁻¹) are observed in the spectrum of CoNiCuPtPd@UNH-730, with their intensities remaining nearly stable over the subsequent 30 minutes. This is attributed to the strong CO₂ adsorption capacity of the material surface (oxidized Co/Ni sites + oxygen vacancies + amino basic sites), as validated by CO₂-TPD measurements. Under a continuous CO₂ flow, the adsorption rate of gaseous *CO₂ to form m-CO₂-3/b-CO₂-3 approximates the consumption rate of these species via hydrogenation, thereby maintaining a steady-state concentration of carbonate species. Oxygen vacancies strongly anchor carbonate species through electron transfer, reducing their desorption rate and ensuring adsorption stability.⁵⁹

Along with the emergence of characteristic bands for CH₄ (3016 cm⁻¹), species including carboxyl (COOH* 1230 cm⁻¹), methyl (δCH 1380 cm⁻¹), bicarbonate (1248 cm⁻¹, 1430 cm⁻¹, and 1446 cm⁻¹), methoxy (CH₃O⁻ 1643 cm⁻¹), and linear *CO (2110 cm⁻¹) are detected. As the CH₄ band (3016 cm⁻¹) intensifies, the bands corresponding to bicarbonate, methoxy, and linear CO increase proportionally, whereas those of carboxyl and methyl diminish.

This is attributed to the direct hydrogenation of carbonates (m-CO₂-3/b-CO₂-3 + H* → HCO-3), where stable carbonates (m-CO₂-3 and b-CO₂-3) serve as a continuous precursor for bicarbonate. However, in the subsequent hydrogenation step, the conversion of HCO-3 to CH₃O⁻ requires cleavage of the C–O bond, with an activation energy higher than that of the COOH* hydrogenation reaction. This step thus becomes the kinetic-limiting step, leading to the accumulation of HCO-3. The HEA sites enable efficient dissociation of H₂ to generate active hydrogen atoms (H*). During methoxy hydrogenation, therefore, although methoxy is continuously converted to methyl, the high H* accessibility accelerates methoxy formation, resulting in a net accumulation of methoxy. In contrast, methyl is rapidly consumed without significant buildup.^60,61

We have employed the VASP^62,63 to perform all density functional theory (DFT) calculations within the GGA using the PBE functional.⁶⁴ We have chosen the PAW^65,66 potentials to describe the ionic cores and take valence electrons into account using a plane wave basis set with a kinetic energy cutoff of 450 eV. The CoNiCuPtPd@UNH-730 model was established (Fig. 9d), and the DFT results revealed the energy barrier distribution along the reaction pathway (Fig. 9e). H₂-TPD verified that the HEA sites can efficiently dissociate H₂ into active H* species. DFT calculations showed that the Gibbs free energy change (ΔG) of the H₂ dissociation step (H₂ → 2H*) is −1.042 eV, indicating that the multimetallic synergy of the high-entropy alloy significantly lowers the energy barrier for H₂ dissociation. The ΔG for CO₂ activation to form the COOH intermediate (CO₂ + 2H* → H* + COOH*) is −0.249 eV, confirming the facile formation of the carboxyl intermediate.

In situ DRIFTS detected the COOH* signal at 1230 cm⁻¹, verifying its role as a transient intermediate. The ΔG for COOH conversion to CO (H* + COOH* → CO* + H₂O) is −0.711 eV, whereas the CO desorption step (CO* → CO) requires overcoming a high energy barrier, with a ΔG as high as +2.172 eV. This is consistent with the CO-TPD results (Fig. 8c), as CoNiCuPtPd@UNH-730 exhibits stronger CO adsorption and thus inhibits CO desorption. Strong CO adsorption favors methanation by avoiding the reverse water–gas shift (RWGS) side reaction, indicating that CO strongly adsorbed on HEA sites undergoes further hydrogenation rather than desorption. Combined DFT calculations and in situ DRIFTS identified intermediates including CO₂-3, COOH*, and CH₃O⁻, which jointly validate the main reaction pathway that CO₂ is converted to CH₄via carbonate, COOH*, CO*, and CH₃O⁻ intermediates, with the transformation of HCO-3 to CH₃O⁻ as the potential rate-determining step.

Based on the catalytic mechanism analysis in this work and relevant literature reports,^67,68 the formation of CH₄ follows the pathway: CO₂ → *CO₂ → CO₂-3/HCO-3 → COOH* → CO → CH₃O⁻ → CH₄, while the CO formation pathway is: CO₂ → COOH* → CO. The detailed process is as follows: gaseous CO₂ molecules are preferentially adsorbed at the oxidized Co/Ni sites and oxygen vacancies on the surface of the CoNiCuPtPd@UNH-730 catalyst. Through electron transfer, oxygen vacancies bend the CO₂ molecules, weaken the C=O bonds, and facilitate CO₂ activation and coordination with surface metal cations, ultimately leading to the formation of monodentate and bidentate carbonates.⁶⁹ The HEA sites efficiently dissociate H₂ to generate adsorbed H*, which then attacks carbonate ions (CO₂-3) to induce stepwise protonation, forming bicarbonate ions (HCO-3). Under the continuous supply of H* from the HEA sites, the bicarbonate ions are gradually hydrogenated to methoxy groups (CH₃O⁻).⁷⁰ Methoxy undergoes dissociation on the catalyst surface, cleaving the C–O bond to form methyl groups, which ultimately combine with H* to produce gaseous CH₄ that continuously desorbs from the catalyst surface. Partial gaseous CO₂ rapidly combines with surface-adsorbed H* to form highly active carboxyl groups (COOH*) under the synergistic effect of the catalyst's oxygen vacancies.⁷¹ As an extremely unstable transition intermediate in the reaction, COOH* undergoes rapid decarboxylation on the catalyst surface immediately after formation, cleaving its relevant chemical bonds. A small portion decomposes into adsorbed CO (*CO), while the majority further participates in the conversion to methoxy and ultimately to CH₄.⁷²

In the temperature range of 350–650 °C, under light irradiation, the temperature of the fixed bed was adjusted by a temperature controller to ensure that the sum of the light-induced temperature and the fixed bed temperature was equal to the temperature without light irradiation. Catalytic performance evaluations were conducted on the MOF-derived supported CoNiCuPtPd@UiO-730 and the HEA-supported CoNiCuPtPd@UNH-730 catalysts (Fig. 10). The CH₄ formation rate (Fig. 10a) and selectivity (Fig. 10b) were almost identical under light and dark conditions when the same temperature was achieved, indicating that both catalysts convert light into heat rather than relying on the traditional “photogenerated electron–hole” effect. Performance evaluations of CH₄ yield and selectivity under full-spectrum, ultraviolet (UV), visible, and infrared (IR) light irradiation are shown in Fig. S7. Within the temperature range of 350–550 °C, both methane production rates (as shown in Fig. 10a) and selectivities (as shown in Fig. 10b) of the two catalysts increased with increasing temperature, consistent with the kinetic characteristics of methanation reactions where “higher temperatures accelerate reaction rates” (CO₂ + 4H₂ → CH₄ + 2H₂O is exothermic, but low temperatures impose kinetic limitations that can be mitigated by temperature elevation). Beyond 550 °C, the methane production rate and selectivity began to decline due to reduced adsorption capacities of CO₂ and H₂. Within the experimental temperature range, CoNiCuPtPd@UNH-730 exhibited significantly higher methane production rates and selectivities compared to CoNiCuPtPd@UiO-730. At 550 °C, the methane production rate and selectivity of CoNiCuPtPd@UNH-730 reached 228.36 mmol g⁻¹ h⁻¹ (77.8%), far surpassing those of CoNiCuPtPd (31 mmol g⁻¹ h⁻¹, 31%) and CoNiCuPtPd@UiO-730 (65.9 mmol g⁻¹ h⁻¹, 48.5%). Detailed performance evaluations of CoNiCuPtPd@UNH-730, monometallic/bimetallic supported catalysts and various supports are shown in Fig. S8. UiO-67 and MOF-808 (their BET results are shown in Fig. S9) belong to the Zr-based MOF family together with UiO-66-NH₂, and all employ Zr₆O₄(OH)₄ as the metal node. After pyrolysis, they can be converted into tetragonal ZrO₂ (t-ZrO₂) with oxygen vacancies. The CoNiCuPtPd-loaded sample derived from pyrolyzed UiO-67 is denoted as CoNiCuPtPd@U67-730, and that from pyrolyzed MOF-808 is denoted as CoNiCuPtPd@M08-730 (their XRD patterns are presented in Fig. S10). The difference from UiO-66-NH₂ is that these precursors contain no nitrogen. These results indicate that the UiO-66-NH₂-derived template plays a pivotal role in enhancing the methanation activity of multimetallic catalysts. The nitrogen atoms modulate the electronic structure of metals via “electronic effects,” optimizing adsorption and activation of reaction intermediates. Furthermore, the retained amino groups enhance CO₂ adsorption capacity, creating favorable conditions for subsequent CO₂ hydrogenation. Finally, the pentametallic high-entropy alloy (HEA) generates synergistic effects among the constituent metals, allowing key intermediates to undergo rapid hydrogenation at favorable energy barriers. In addition, in comparison with recently reported photothermal CO₂ hydrogenation catalysts, the CoNiCuPtPd@UNH-730 catalyst prepared in this work exhibits outstanding catalytic performance in terms of CH₄ rate (Table S1).

Fig. 10 (a) Methane production rates of CoNiCuPtPd@UiO-730 and CoNiCuPtPd@UNH-730; (b) methane selectivities of CoNiCuPtPd@UiO-730 and CoNiCuPtPd@UNH-730.

Conclusion

In summary, this study developed CoNiCuPtPd-based catalysts using UiO-66 and UiO-66-NH₂ as templates to investigate their methanation performance. The results demonstrated that introducing –NH₂ groups into UiO-66-NH₂ led to the formation of nitrogen-doped porous carbon frameworks post-pyrolysis. Pyridinic nitrogen sites form strong coordination interactions with metals, effectively suppressing the migration and agglomeration of metal particles, thereby promoting the formation of a homogeneously dispersed HEA of Co, Ni, Cu, Pt, and Pd. XPS and TEM analyses confirmed that the electron-donating effect of amino groups not only facilitated complete metal reduction but also strengthened the metal–support interactions, significantly enhancing catalytic stability and active site accessibility. CO₂-TPD and H₂-TPD analyses demonstrated that CoNiCuPtPd@UNH-730 exhibited enhanced CO₂ and H₂ adsorption capacities, attributed to its abundant oxygen vacancies and multimetallic synergy. Furthermore, CO-TPD revealed stronger adsorption and prolonged retention of CO intermediates on this catalyst, facilitating stepwise hydrogenation to CH₄ and thereby significantly improving methane selectivity and production rate.

In situ diffuse reflectance infrared Fourier transform spectroscopy (in situ DRIFTS) detected the formation of intermediates including carbonate/bicarbonate, COOH* and CO*, which constitute the key steps of CO₂ methanation. Combined with DFT calculations, it is demonstrated that the multimetallic synergistic effect of the high-entropy alloy significantly reduces the energy barrier for H₂ dissociation, whereas the CO desorption step (CO* → CO) needs to overcome a high energy barrier. This suppresses CO desorption, prevents the reverse water–gas shift (RWGS) side reaction, and favors the methanation process, revealing the plausible reaction pathway over CoNiCuPtPd@UNH-730. Under photothermal conditions at 550 °C, the catalyst delivered a methane yield of 228.36 mmol g⁻¹ h⁻¹ with 77.8% selectivity, far exceeding that of the unmodified UiO-66-based catalyst.

Author contributions

Chunlin Ke: writing – original draft, writing – review & editing, visualization, validation, software, methodology, investigation, formal analysis, data curation, conceptualization. Siyu Song: conceptualization, supervision, resources, writing – review & editing, project administration. Shenghao Li: supervision, resources, project administration. Fengliang Wang: investigation, data curation, formal analysis, validation, writing – review & editing. Ruiqi Fang: investigation, methodology, software. Xin Zhao: formal analysis, validation, writing – review & editing. Chao Wang: conceptualization, supervision, resources, writing – review & editing, project administration.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supplementary information (SI): experimental details, supporting figures, and tables referenced in the main text. See DOI: https://doi.org/10.1039/d5ta10426b.

Acknowledgements

This work was supported by the Natural Science Foundation of China [No. 22578070 and 22378073] and Guangzhou Basic and Applied Basic Research Foundation [SL2024A04J00939].

References

1. L. Qiu, X. Yao, Y. Zhang, H. Li and L. He, J. Org. Chem., 2022, 88, 4942–4964.
2. R. Abazari, N. Ghorbani, J. Shariati, R. S. Varma and J. Qian, Inorg. Chem., 2024, 63(27), 12667–12680.
3. T. Dong, X. Liu, Z. Tang, H. Yuan, D. Jiang, Y. Wang, Z. Liu, X. Zhang, S. Huang, H. Liu, L. Zhao and W. Zhou, Appl. Catal., B, 2023, 326, 122176.
4. K. Peng, J. Ye, H. Wang, H. Song, B. Deng, S. Song, Y. Wang, L. Zuo and J. Ye, Appl. Catal., B, 2023, 324, 122262.
5. C. Han, X. Zhang, S. Huang, Y. Hu, Z. Yang, T. T. Li and J. Qian, Advanced Science, 2023, 10(19), 2300797.
6. C. Song, X. Liu, M. Xu, D. Masi, Y. Wang, Y. Deng, M. Zhang, X. Qin, K. Feng, J. Yan, J. Leng, Z. Wang, Y. Xu, B. Yan, S. Jin, D. Xu, Z. Yin, D. Xiao and D. Ma, ACS Catal., 2020, 10, 10364–10374.
7. K. Chen, S. Zhou, T. Jiang, X. Li, J. Yu, Q. Wang, X. Xu and L. Zhu, J. Mater. Chem. A, 2021, 9, 22353–22363.
8. G. Zhou, H. Liu, K. Cui, A. Jia, G. Hu, Z. Jiao, Y. Liu and X. Zhang, Appl. Surf. Sci., 2016, 383, 248–252.
9. X. Guo, A. Traitangwong, M. Hu, C. Zuo, V. Meeyoo, Z. Peng and C. Li, Energy Fuels, 2018, 32, 3681–3689.
10. A. Vita, C. Italiano, L. Pino, P. Frontera, M. Ferraro and V. Antonucci, Appl. Catal., B, 2018, 226, 384–395.
11. J. Gao, Q. Liu, F. Gu, B. Liu, Z. Zhong and F. Su, RSC Adv., 2015, 5, 22759–22776.
12. B. Miao, S. S. K. Ma, X. Wang, H. Su and S. H. Chan, Catal. Sci. Technol., 2016, 6, 4048–4058.
13. J. K. Pedersen, T. A. A. Batchelor, A. Bagger and J. Rossmeisl, ACS Catal., 2020, 10, 2169–2176.
14. Y. Ma, Y. Ma, Q. Wang, S. Schweidler, M. Botros, T. Fu, H. Hahn, T. Brezesinski and B. Breitung, Energy Environ. Sci., 2021, 14, 2883–2905.
15. Y. Zhang, Z. Jiang, R. Zhang, K. Wang and X. Wang, Appl. Surf. Sci., 2024, 651, 159226.
16. H. Huang, J. Zhao, H. Guo, B. Weng, H. Zhang, R. A. Saha, M. Zhang, F. Lai, Y. Zhou, R. Z. Juan, P. C. Chen, S. Wang, J. A. Steele, F. Zhong, T. Liu, J. Hofkens, Y. M. Zheng, J. Long and M. B. J. Roeffaers, Adv. Mater., 2024, 36(26), 2313209.
17. J. Zhao, J. Bao, S. Yang, Q. Niu, R. Xie, Q. Zhang, M. Chen, P. Zhang and S. Dai, ACS Catal., 2021, 11, 12247–12257.
18. H. Li, H. Huang, Y. Chen, F. Lai, H. Fu and L. L. T. Zhang, Adv. Mater., 2023, 35(2), 2209242.
19. X. Ding, W. Liu, J. Zhao, L. Wang and Z. Zou, Adv. Mater., 2025, 37(2), 2312093.
20. A. J. Knorpp, M. Mielniczuk, M. Nikolic, A. Vogel, R. Figi, C. Schreiner, A. Borgschulte and M. Stuer, Eur. J. Inorg. Chem., 2024, 27(35), e202400476.
21. X. Liu, X. Wang, H. Sun, Z. Zhang, P. Song and Y. Liu, Ind. Eng. Chem. Res., 2022, 62, 341–354.
22. D. Li, Q. Wu, C. Wang, Z. Sun, Y. Li, Z. Yin and X. Zhang, Adv. Mater., 2025, 38(6), e18092.
23. P. Li, S. Zhang, Z. Xiao, H. Zhang, F. Ye, J. Gu, J. Wang, G. Li and D. Wang, Fuel, 2024, 357, 129817.
24. H. Wang, Q. Li, J. Chen, J. Chen and H. Jia, Adv. Sci., 2023, 10(34), 2304406.
25. H. Wu, Z. Wang, B. Tian, Y. Li, Z. Chang, Y. Kuang and X. Sun, Nanoscale, 2024, 16(6), 3034–3042.
26. Z. Wang, Y. Wang, W. Li, S. Liu, L. Zhang, J. Yang, C. Feng, R. Chong and Y. Zhou, J. Colloid Interface Sci., 2025, 678, 689–702.
27. H. Wang, Q. Li, J. Chen and H. Jia, J. Mater. Chem. A, 2024, 12, 15803–15813.
28. M. Kandiah, M. H. Nilsen, S. Usseglio, S. Jakobsen, U. Olsbye, M. Tilset, C. Larabi, E. A. Quadrelli, F. Bonino and K. P. Lillerud, Chem. Mat., 2010, 22, 6632–6640.
29. M. Yamasaki, H. Habazaki, K. Asami, K. Izumiya and K. Hashimoto, Catal. Commun., 2006, 7, 24–28.
30. Y. Sun, J. Li, Z. Wang and L. Liu, J. Alloys Compd., 2024, 989, 174350.
31. G. U. O. Sheng and C. T. Liu, Prog. Nat. Sci. Mater. Int., 2011, 21(6), 433–446.
32. D. Li, Q. Wu, C. Wang, Z. Sun, Y. Li, Z. Yin and X. Zhang, Adv. Mater., 2025, e18092.
33. Z. W. Chen, Z. Gariepy, L. Chen, X. Yao, A. Anand, S. J. Liu and C. V. Singh, ACS Catal., 2022, 12(24), 14864–14871.
34. J. H. Smith, Q. Luo, S. L. Millheim and J. E. Millstone, J. Am. Chem. Soc., 2024, 146, 34822–34832.
35. X. Zhao, X. Kong, F. Wang, R. Fang and Y. Li, Angew. Chem., Int. Ed., 2021, 60, 10842–10849.
36. S. Xu, A. Dong, Y. Hu, Z. Yang, S. Huang and J. Qian, J. Mater. Chem. A, 2023, 11(18), 9721–9747.
37. C. Han, X. Zhu, J. Ding, T. Miao, S. Huang and J. Qian, Inorg. Chem., 2022, 61(46), 18350–18354.
38. L. Yang, D. Zhang, X. Li, L. Qian, H. Zhang, P. Fang and C. He, Langmuir, 2024, 40, 21395–21406.
39. P. Huang, P. Zhang, C. Wang, J. Tang and H. Sun, Appl. Catal., B, 2022, 303, 120926.
40. B. S. Caglayan and A. E. Aksoylu, J. Hazard. Mater., 2013, 252–253, 19–28.
41. X. Wang, A. Dong, Y. Hu, J. Qian and S. Huang, Chem. Commun., 2020, 56(74), 10809–10823.
42. F. Cui, L. Tang, W. Han, Y. Liu, H. Y. Kim, J. Yu and B. Ding, Compos. Commun., 2020, 22, 100470.
43. Y. Zhou, L. Liu, G. Li and C. Hu, ACS Catal., 2021, 11, 7099–7113.
44. Y. Wu, J. Lin, G. Ma, Y. Xu, J. Zhang, C. Samart and M. Ding, RSC Adv., 2020, 10(4), 2067–2072.
45. J. Niu, C. Zhang, H. Liu, Y. Jin and R. Zhang, Front. Energy, 2023, 17(4), 545–554.
46. H. Kim, K. Choi, K. Lee, S. Lee, K. Jung and J. Choi, Water, 2021, 13, 1869.
47. M. Farrag, J. Alloys Compd., 2022, 920, 165893.
48. B. C. M. Martindale, G. A. M. Hutton, C. A. Caputo, S. Prantl, R. Godin, J. R. Durrant and E. Reisner, Angew. Chem., Int. Ed., 2017, 56, 6459–6463.
49. Y. Zhao, H. Zhou, W. Chen, Y. Tong, C. Zhao, Y. Lin, Z. Jiang, Q. Zhang, Z. Xue, W. Cheong, B. Jin, F. Zhou, W. Wang, M. Chen, X. Hong, J. Dong, S. Wei, Y. Li and Y. Wu, J. Am. Chem. Soc., 2019, 141, 10590–10594.
50. M. Li, X. Liu, Y. Che, H. Xing, F. Sun, W. Zhou and G. Zhu, Angew. Chem., Int. Ed., 2023, 62, e202308651.
51. J. Wu, Y. Song, D. Deng, J. Pu, Y. Li, J. Zhou and X. Liu, Mater. Lett., 2025, 400, 139122.
52. X. Liu, Q. Gu, Y. Zhang, X. Xu, H. Wang, Z. Sun, L. Cao, Q. Sun, L. Xu, L. Wang, S. Li, S. Wei, B. Yang and J. Lu, J. Am. Chem. Soc., 2023, 145, 6702–6709.
53. Y. Yan, Y. Dai, H. He, Y. Yu and Y. Yang, Appl. Catal., B, 2016, 196, 108–116.
54. G. Ren, Z. Wei, S. Liu, M. Shi, Z. Li and X. Meng, Chemosphere, 2022, 307, 136026.
55. W. L. Vrijburg, E. Moioli, W. Chen, M. Zhang, B. J. P. Terlingen, B. Zijlstra, I. A. W. Filot, A. Züttel, E. A. Pidko and E. J. M. Hensen, ACS Catal., 2019, 9, 7823–7839.
56. X. Jia, X. Zhang, N. Rui, X. Hu and C. Liu, Appl. Catal., B, 2019, 244, 159–169.
57. A. Li, Q. Cao, G. Zhou, B. V. K. J. Schmidt, W. Zhu, X. Yuan, H. Huo, J. Gong and M. Antonietti, Angew. Chem., Int. Ed., 2019, 58, 14549–14555.
58. Y. Chen, Y. Zhang, G. Fan, L. Song, G. Jia, H. Huang, S. Ouyang, J. Ye, Z. Li and Z. Zou, Joule, 2021, 5, 3235–3251.
59. Y. Gao, Q. Li, C. Wang, D. Yan, J. Chen and H. Jia, J. Mater. Chem. A, 2022, 10, 16016–16028.
60. Z. Wang, H. Song, H. Pang, Y. Ning, T. D. Dao, Z. Wang, H. Chen, Y. Weng, Q. Fu, T. Nagao, Y. Fang and J. Ye, Appl. Catal., B, 2019, 250, 10–16.
61. X. Zu, Y. Zhao, X. Li, R. Chen, W. Shao, Z. Wang, J. Hu, J. Zhu, Y. Pan, Y. Sun and Y. Xie, Angew. Chem., Int. Ed., 2021, 60, 13840–13846.
62. G. Kresse and J. Furthmüller, Comput. Mater. Sci., 1996, 6(1), 15–50.
63. G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 49(20), 14251.
64. J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77(18), 3865.
65. P. E. Blöchl, O. Jepsen and O. K. Andersen, Phys. Rev. B: Condens. Matter Mater. Phys., 1994, 49(23), 16223.
66. G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater. Phys., 1999, 59(3), 1758.
67. A. Li, T. Wang, X. Chang, Z. Zhao, C. Li, Z. Huang, P. Yang, G. Zhou and J. Gong, Chem. Sci., 2018, 9, 5334–5340.
68. Y. Xie, J. Wen, Z. Li, J. Chen, Q. Zhang, P. Ning, Y. Chen and J. Hao, Green Chem., 2023, 25, 130–152.
69. M. A. A. Aziz, A. A. Jalil, S. Triwahyono, R. R. Mukti, Y. H. Taufiq-Yap and M. R. Sazegar, Appl. Catal., B, 2014, 147, 359–368.
70. S. Kattel, W. Yu, X. Yang, B. Yan, Y. Huang, W. Wan, P. Liu and J. G. Chen, Angew. Chem., Int. Ed., 2016, 55, 7968–7973.
71. X. Jiao, X. Li, X. Jin, Y. Sun, J. Xu, L. Liang, H. Ju, J. Zhu, Y. Pan, W. Yan, Y. Lin and Y. Xie, J. Am. Chem. Soc., 2017, 139, 18044–18051.
72. J. Xu, Z. Ju, W. Zhang, Y. Pan, J. Zhu, J. Mao, X. Zheng, H. Fu, M. Yuan, H. Chen and R. Li, Angew. Chem., Int. Ed., 2021, 60, 8705–8709.

---