A confinement-stabilized Cu⁰–Cu⁺ redox pair on silica and its catalytic role in the water–gas shift reaction

Abstract

The construction and stabilization of Cu–O–Si interfaces in copper-based catalysts remain critical yet challenging for the water–gas shift reaction (WGSR), as conventional strategies fail to mitigate copper nanoparticle (NP) aggregation and interfacial instability. This study innovatively proposes a spatial confinement strategy by anchoring Cu NPs inside (Cu_in/SBA-15) or outside (Cu_out/SBA-15) the mesoporous channels of SBA-15, leveraging nanoscale confinement to optimize Cu–O–Si interfaces. The confined Cu_in/SBA-15 catalyst demonstrated exceptional WGSR performance, achieving a reaction rate of 5.4 μmol_CO g_cat⁻¹ s⁻¹, significantly surpassing conventional Cu/SBA-15 (3.6 μmol_CO g_cat⁻¹ s⁻¹) and surface-loaded Cu_out/SBA-15 (3.1 μmol_CO g_cat⁻¹ s⁻¹), along with remarkable stability. This enhancement originates from the SBA-15 channels enabling the in situ formation of stable Cu–O–Si interfaces, which regulate the dynamic equilibrium between Cu⁰ (active for H₂O dissociation) and Cu⁺ (critical for CO adsorption) species. In situ studies revealed their synergistic dynamic interconversion during the WGSR, while domain confinement effects suppressed sintering by maintaining interfacial integrity. Kinetic and mechanistic analyses further identified the associative pathway, with HCOO* intermediate dissociation as the rate-determining step, facilitated by the stabilized Cu–O–Si interfaces. By resolving interfacial instability through confinement engineering, this work provides a paradigm for designing robust Cu-based catalysts, advancing both fundamental understanding and practical applications in the WGSR and related heterogeneous catalysis.

---

Huihuang Fang

Huihuang Fang is an Associate Professor at the College of Chemical Engineering, Fuzhou University. He received his PhD in Physical Chemistry at Xiamen University and conducted joint doctoral research at the University of Oxford. He has been recognized as a Minjiang Youth Scholar (2024), Talents in Quanzhou City (2022), and Qishan Scholar (2021), and serves on the Youth Editorial Board of Chin. J. Struct. Chem. His research interests primarily focus on energy catalytic materials, reaction mechanisms, modelling and process technologies for ammonia-hydrogen energy conversion, including ammonia catalysis, hydrogen production, and ammonia/hydrogen fuel cells.

1. Introduction

Copper-based catalysts are widely employed in heterogeneous catalysis, including ethanol dehydrogenation,^1,2 methane reforming^3–5 and WGSR,^6–8 due to their natural abundance, low cost, excellent catalytic performance and high selectivity. However, copper nanoparticles in these catalysts exhibit a relatively low Tammann temperature (T_m = 405 °C),⁹ which renders them prone to aggregation, leading to structural deactivation. In addition, impurities in the feedstock and unsaturated products formed during the reaction tend to deposit on the catalyst surface, resulting in poisoning and subsequent deactivation. Furthermore, the surface structures of some supports, such as Si–OH and Al–OH,^10,11 are prone to dehydration and reconstruction, altering the surface energy of the catalyst. This can lead to the pulverization of the copper-based catalyst and a sharp increase in reaction rates, causing physical deactivation of the catalyst.

Typically, the deactivation of copper-based catalysts manifests as a disturbance in the balance between Cu⁺ and Cu⁰ in the catalyst. As a result, the copper particles aggregate and grow by migrating across the silica surface. Two forms of migratory sintering of copper species have been identified in the literature;^12–14 one mechanism involves the migration of entire particles across the support, resulting in aggregation. The other mechanism involves the detachment of metal species from one particle, followed by their transport through the support or in the gas phase, and subsequent attachment to another particle, and this process is commonly referred to as Ostwald ripening. These phenomena have limited the widespread application of copper-based catalysts. To spatially restrict the migration of copper nanoparticles, some researchers have stabilized copper-based catalysts using spatial immobilization techniques involving oxides,^15,16 carbon materials,¹⁷ and molecular sieve encapsulation.^18–20 Li et al.²¹ prepared Cu@SiO₂ catalysts with high Cu dispersion that effectively prevented particle growth and aggregation during the reaction. Ma et al.²² reported a carbon nanotube-alumina strip (CAS)-loaded Co–Cu catalyst for the CO₂ methanation reaction, where CAS exhibited a limiting effect, prolonging the residence time of the reactants within the mesoporous pores compared to single carbon nanotubes. Cui et al.²³ used an HKUST-1 encapsulated Cu precursor and successfully avoided sintering of homogeneous substances. Despite these advances, maintaining the dynamic Cu⁺/Cu⁰ balance under WGSR conditions remains unresolved. Conventional approaches often fail to stabilize multivalent Cu species while ensuring reactant accessibility, limiting their long-term catalytic robustness.

Hydroxyl groups (–OH) on the surface of SiO₂ can form Cu–O–Si bonds with copper species, and their formation mainly depends on the interaction between the metal and the support. It has been shown that the Cu–O–Si interface can significantly enhance catalytic activity, as it not only provides more active sites but also stabilizes reaction intermediates. For example, in the selective hydrogenation of dimethyl oxalate (DMO) to ethylene glycol (EG), the presence of the Cu–O–Si interface facilitates ester hydrogenation because Cu–H⁺ and Si–OH⁺ can stabilize the hydrogenation transition state, thus lowering the reaction energy barrier.²⁴ The ordered mesoporous structure of SBA-15 has been identified as a promising approach for enhancing the catalytic performance of Cu-based catalysts due to its distinctive pore structure. A large number of isolated silanol (Si–OH) groups on the surface undergo dehydration and condensation with Cu species after heat treatment or reduction, resulting in the formation of a stable Cu–O–Si bonding structure at the interface. In addition, the unique pore structure promotes the dispersion of active centers and facilitates the transfer of reactants and products within the confined space, without leading to unwanted side reactions.^25–27 The confined nanospace thus provides a novel platform for promoting catalytic reactions.

Although the metal confinement by mesoporous or zeolitic frameworks has been widely studied, pure-silica SBA-15 with uniform mesopores to form highly dispersed and redox-stable Cu species is still attractive. Unlike aluminosilicates such as HZSM-5 or HUSY, SBA-15 allows the formation of stable Cu–O–Si interfaces and offers sufficient pore space for both molecular transport and in situ characterization. These features make SBA-15 an ideal platform for understanding the dynamic evolution of Cu⁺/Cu⁰ during the WGSR. Therefore, a detailed mechanistic understanding of confined Cu species and their redox dynamics under reaction conditions is both timely and crucial. Meanwhile, the multivalent state of Cu implies that the catalytically active sites, Cu⁰ and Cu⁺, undergo stochastic and dynamic changes in the reaction environment. In recent years, studies on the synergistic effect of Cu⁺ and Cu⁰ in copper-based catalysts for the WGSR have intensified.^28–30 It has been well documented that synergy between Cu⁺ species, which play a key role in CO adsorption and activation, and Cu⁰, which promotes H₂O dissociation and OH* species generation, is essential for achieving high activity and selectivity. Although these studies have made progress in material design, most studies are limited to static valence characterization and fail to reveal the in-depth reaction mechanism, particularly the dynamic valence transformation of Cu species under reaction conditions and its direct correlation with catalytic performance. In addition, there is still a lack of systematic approaches to observe how the Cu⁺/Cu⁰ ratio is stabilized by the microstructure during the actual WGSR process and how its evolution affects the intermediate conversion and the rate-determining steps. As a result, identifying the active Cu species and further elucidating the catalytic mechanism of Cu-based catalysts remain challenging.

Herein, we address these challenges by anchoring Cu nanoclusters within SBA-15 mesochannels to construct a stable Cu–O–Si interface. Systematic characterization reveals that the confined environment enhances Cu dispersion and interfacial integrity, and the catalyst with stable and balanced Cu⁺ and Cu⁰ species exhibits excellent activity and stability in the low-temperature WGSR. Microstructural characterisation confirms that the mesoporous material provides spatial confinement effects, resulting in uniform dispersion of copper nanoparticles within the SBA-15 pores. This structure stabilizes the Cu⁺ and Cu⁰ species through the formation of a large number of Cu–O–Si interfaces. In addition, in situ and transient isotope-switching experiments reveal the rate-determining steps of the WGSR process over the domain-confined Cu-based catalysts and provide insight into the state of the active Cu species and the dynamic evolution of the catalytic process.

2. Experimental methods

2.1 Materials

Analytical-grade copper(II) nitrate (Cu(NO₃)₂‧3H₂O), urea (CO(NH₂)₂, 99%), and tetraethyl orthosilicate (TEOS, 98%) were purchased from Sinopharm Chemical Reagent Co., Ltd. Triblock copolymer Pluronic P123 (EO₂₀PO₇₀EO₂₀, molecular weight 5800), methanol (CH₃OH, 99%), and hydrochloric acid (HCl, 38%) were obtained from Shanghai Aladdin Biochemical Technology Co. Ltd.

2.2 Methods

2.2.1 Preparation of the support. Mesoporous SiO₂ (SBA-15) was prepared using the classical method reported by Zhao et al.³¹ First, 8.0 g of Pluronic P123 triblock copolymer was dissolved in 200 mL of 1.6 M HCl aqueous solution and stirred at 38 °C until the polymer was completely dissolved. Then, 17 g of tetraethyl orthosilicate (TEOS) was added dropwise, and the mixture was stirred at 38 °C for 24 hours. The resulting solution was then transferred to a Teflon-lined stainless steel autoclave and heated at 100 °C for 24 hours. The sample was filtered and washed with deionized water and ethanol, followed by overnight drying in an oven at 60 °C. Subsequently, the dried sample was calcined at 600 °C for 2 hours in a static air atmosphere with a heating rate of 4 °C‧min⁻¹ to produce the SBA-15 support.

2.2.2 Preparation of catalysts. The catalysts were prepared by a urea-assisted impregnation method. In a typical impregnation method, copper nitrate and urea were dissolved in deionized water, followed by impregnation of the SBA-15 support with the above mixed solution of copper nitrate and urea, and left to stand for 24 hours away from light, to obtain the wet Cu/SBA-15 sample, and the surface Cu species of the above wet sample of Cu/SBA-15 was washed off with deionised water to obtain the wet sample of Cu_in/SBA-15. In addition, prior to the above impregnation, the SBA-15 support was wetted in one step first, followed by impregnation with a mixed solution of copper nitrate and urea, referred to as Cu_out/SBA-15, and the wet samples were dried in an oven at 60 °C for 12 hours and then calcined in hydrogen gas at an elevated rate of 4 °C‧min⁻¹ for 1 hour. Subsequently, the obtained drying samples were reduced at 350 °C for 2 hours under a constant hydrogen flow of 40 mL‧min⁻¹ to achieve the final Cu/SBA-15, Cu_in/SBA-15 and Cu_out/SBA-15 catalysts.

2.3 Catalyst characterization

The powder X-ray diffraction (XRD) patterns of the catalysts were recorded using a PANalytical X'Pert Pro powder diffractometer equipped with an X'Celerator detector. The measurements were performed with a Cu K_α radiation source (wavelength of 0.15406 nm) at a tube voltage of 40 kV and a tube current of 40 mA. The specific surface area and pore structure of the precursors and catalyst samples were measured using a Micromeritics gas adsorption analyzer (ASAP 2020) at a temperature of −196 °C in liquid nitrogen. The pore structure and specific surface area of the catalysts were calculated using the BJH and BET methods. X-ray photoelectron spectroscopy (XPS) experiments were implemented on an ESCALab 250Xi electron spectrometer with Al K_α radiation (1484.6 eV) as the X-ray source and all binding energies were corrected against the C 1s peak at 284.8 eV. The temperature-programmed reduction (TPR) tests were conducted using a fully automated chemisorption analyzer (AutoChem 2920) from Micromeritics equipped with a TCD detector. The dispersion of Cu in the Cu/SBA-15 catalysts was determined using a chemisorption analyzer (Micromeritics, ASAP 2020). The samples were first pre-treated to the state prior to activity testing, followed by the introduction of 30 mL min⁻¹ of 2 vol% N₂O/He at various temperatures to oxidize surface Cu to Cu₂O. After purging with high-purity Ar, H₂ pulse reduction was performed, and the dispersion was subsequently calculated.

2.4 Catalyst performance evaluation

The catalytic performance evaluations of the WGSR were performed on a CO-CMAT9002 apparatus under a gas atmosphere of 15% CO, 55% H₂, 7% CO₂, and 23% N₂ (vol%). Before the activity test, 0.5 g catalysts (40–60 mesh) diluted with quartz sand to the fixed bed height were pre-reduced in a 5 vol% H₂/N₂ reducing atmosphere at 350 °C for 2 hours, followed by switching to the feed gas for testing. The reaction products were analyzed via online gas chromatography (GC). The test temperature ranges from 200 °C to 400 °C, and the molar ratio of steam to dry feed gas was 1 : 1, with a space velocity of 4500 mL g⁻¹ h⁻¹. The post-reaction gas was condensed to remove unreacted water and then passed into a gas chromatograph equipped with a thermal conductivity detector (TCD) using hydrogen as the support gas for online detection. CO conversion was calculated using the following equations:where V_CO and are the volume fractions of CO in the dry feed gas and the outlet gas.

Kinetic evaluations were performed in the same continuous flow fixed bed with a quartz reactor as previously described. In order to eliminate internal and external diffusion limitations, measurements were carried out at a temperature of 175 °C, where the CO conversion was kept below 20% by adjusting the weight of the catalyst or the flow rate of the mixed gas. The reaction rate and the turnover frequency (TOF) were calculated as follows:³²

3. Results and discussion

3.1 Characteristics of domain-limited catalysts

To investigate the influence of domain-limiting effects on the structure of copper-based catalysts, three distinct Cu/SBA-15 catalysts were synthesized via urea-assisted impregnation^33,34 using SBA-15 as the support. These catalysts exhibited different copper distribution patterns: Cu_in/SBA-15 with predominant intraporous Cu species, Cu_out/SBA-15 featuring mainly extraporous Cu species, and Cu/SBA-15 with unregulated copper distribution. A comparison of the small-angle XRD patterns (Fig. 1(b)) for the as-reduced catalysts revealed a prominent diffraction peak at 2θ ≈ 0.9°, attributed to the (100) plane of the 2D hexagonal mesostructure of SBA-15,³⁵ indicating that metal Cu loading had no obvious effect on the symmetry of the mesoporous framework. Additionally, two diffraction peaks were also observed at 2θ ≈ 1.6° and 1.8°, corresponding to the (110) and (200) planes,^36,37 respectively. However, subtle differences in the (100) peaks of the three catalysts suggest variations in unit cell parameters and framework wall thickness. Notably, the (100) peak of Cu_in/SBA-15 exhibited a slightly rightward shift and decreased intensity, indicating that Cu species were primarily confined within the mesopores, thus causing minimal disruption to the SBA-15 framework. High-resolution transmission electron microscopy (Fig. 1(c and d) and ESI Fig. S1†) clearly revealed ordered mesoporous channels (∼4 nm diameter) in all catalysts with comparable average Cu particle sizes. Cu_in/SBA-15 displayed uniformly dispersed Cu clusters (0.9–2.7 nm) within channels, while Cu_out/SBA-15 contained larger aggregates (2–7 nm) primarily localized outside pores. Wide-angle XRD analysis (ESI Fig. S2†) showed significant post-reaction intensification of Cu (JCPDS: 04-0836) and Cu₂O (JCPDS: 65-3288) peaks for Cu/SBA-15 and Cu_out/SBA-15, contrasting with stable peak intensities for Cu_in/SBA-15. This differential behavior highlights the role of the confinement effect in suppressing Cu migration and agglomeration through both spatial restriction and potential Cu–O–Si interface formation with silanol groups.

Fig. 1 Structural characterization of Cu/SBA-15, Cu_in/SBA-15 and Cu_out/SBA-15 catalysts. (a) Schematic illustration of the catalysts; (b) small-angle XRD patterns of the catalysts; HAADF-STEM images of (c) Cu_in/SBA-15 and (d) Cu_out/SBA-15; (e) TEM images and (f) HAADF-STEM image and (g) EDS elemental mappings of the reduced Cu_in/SBA-15 catalyst.

A similar trend was observed in dissociative N₂O adsorption. The dispersion of Cu over the three catalysts was investigated utilizing N₂O as a probe, with the results of the N₂O chemisorption measurements summarized in Table 1. The data revealed superior Cu dispersion in Cu_in/SBA-15 (27.8%) versus counterparts, confirming the enhanced distribution of active sites due to spatial confinement. Nitrogen physisorption analysis (ESI Fig. S3†) maintained type IV isotherms with H1 hysteresis for all materials, characteristic of mesoporous systems. Comparative textural properties (Table 1) showed sequential reductions in the BET surface area and pore volume relative to pristine SBA-15. This trend suggested gradual pore occupation by Cu species and indicated that the apparent decrease in the surface area for Cu_out/SBA-15 may be attributed to pore blockage by larger, externally located particles.

Table 1 Textural properties of the as-prepared catalysts

Sample	Cu content^a (wt%)	A BET (m^2‧g^−1)	V Pore (cm^3‧g^−1)	d Pore (nm)	D Cu (%)	d Cu (nm)	Ratio^e of Cu^+/Cu^0 (%)
a Determined by ICP-OES analysis. b BET specific surface area. c Cu metallic surface area per gram of catalyst determined by the N2O titration method. d Calculated on the basis of TEM images. e Quantitative by Auger.
SBA-15	—	762.7	1.02	6.48	—	—	—
Cu/SBA-15	4.35	503.6	0.76	6.63	25.1	2.71	41
Cuin/SBA-15	5.51	492.7	0.75	6.69	27.8	1.87	35
Cuout/SBA-15	5.45	440.7	0.74	7.14	19.1	4.41	52

---

The morphology and details of the reduced state of the Cu_in/SBA-15 catalyst are shown in Fig. 1(e and f). The high-resolution TEM image of Cu_in/SBA-15 revealed lattice fringes with a spacing of 0.209 nm, corresponding to the Cu(111) planes.³⁸ This observation indicated that the predominant exposed facet of the confined Cu nanoparticles was Cu(111), which was well-recognized for its balanced activity in both CO adsorption and H₂O dissociation.³⁹ The prevalence of Cu(111) planes, stabilized by SBA-15 confinement, may thus synergistically contribute to enhanced catalytic performance.

The co-existence of fcc Cu and cubic Cu₂O phases was confirmed by XRD (Fig. S2†), with no evidence of amorphous or mixed-phase transformations. This crystalline dual-phase structure remains stable under reaction conditions, particularly in the Cu_in/SBA-15 catalyst, where spatial confinement helps prevent Cu nanoparticle agglomeration and surface oxidation. In contrast, the external Cu in Cu_out/SBA-15 undergoes partial structural degradation, as evidenced by broadened diffraction peaks and reduced crystallinity. The ability to retain a well-defined Cu⁰/Cu⁺ crystalline interface is considered a key factor in maintaining redox activity and catalytic efficiency in the WGSR process. The dark-field TEM image (Fig. 1(f)) showed numerous high-density bright spots corresponding to Cu clusters uniformly dispersed within the inner wall of the mesopores, indicating small and well-dispersed Cu particles. Furthermore, elemental mapping (Fig. 1(g)) confirmed the uniform distribution of Si, O, and Cu elements, further supporting the role of SBA-15 mesopores in facilitating the anchoring and dispersion of Cu species through a confinement effect.

3.2 Catalytic performance

To further investigate the confinement effect under WGSR conditions, Fig. 2(a) shows CO conversion at 240 °C and WGSR rates at 175 °C across catalyst variants. The reaction rates were calculated at a relatively low conversion (20%) to minimize the limitations of internal and external diffusion.

Fig. 2 (a) CO conversion (T = 240 °C) and WGSR rate (T = 175 °C) of the Cu/SBA-15 series catalysts; (b) TOF values (T = 175 °C), dispersion of Cu, and Cu⁺/Cu⁰ ratios of the Cu/SBA-15 series catalysts; (c) CO conversion of Cu/SBA-15 series catalysts and some typical WGSR catalysts: 1–5 wt.%Cu/SBA-16,⁴⁰ 2–9 wt.%Cu/SiO₂,⁴¹ 3–5 wt.%Cu/SiO₂, 4–5 wt.%Cu/Silicalite-1, 5-Cu/SBA-15, 6-Cu_in/SBA-15, and 7-Cu_out/SBA-15; (d) stability test and (e) process of associative and regeneration and (f) start–stop cycle tests of Cu_in/SBA-15; reaction conditions: WHSV = 4500 mL‧g_cat⁻¹‧h⁻¹, steam : gas = 1 : 1, and T = 240 °C.

Among all the prepared catalysts, even at a relatively low temperature of 240 °C, the Cu_in/SBA-15 catalyst with 5% Cu loading achieved a CO conversion of 75%, whereas the Cu_out/SBA-15 catalyst exhibited only 30% to 40% conversion under the same conditions. This activity disparity aligned with structural characteristics, as the Cu_in/SBA-15 catalyst exhibited a superior reaction rate of 5.4 μmol_CO‧g_cat⁻¹‧s⁻¹, approximately 1.7 times higher than that of the Cu_out/SBA-15 catalyst. To further rationalize the structure and performance relationship among the three catalyst configurations, a simplified quantitative model was introduced to evaluate the additive behavior of copper distribution, as illustrated in Fig. S5.† Based on structural and catalytic data, the performance of the three catalysts can be described using a semi-quantitative expression, suggesting that the activity of the randomly distributed Cu/SBA-15 catalyst arises from a weighted combination of confined and non-confined copper contributions. In the absence of spatial control, the Cu species cannot fully benefit from either confinement-induced stabilization or external accessibility, leading to diluted and suboptimal catalytic performance. These results clearly indicated that the internal confinement of copper species within the mesopores of SBA-15 (Cu_in/SBA-15) provides a more favorable structural environment compared to external deposition (Cu_out/SBA-15) or random distribution (Cu/SBA-15). The enhanced performance arose from higher copper dispersion, an optimized Cu⁺/Cu⁰ ratio, and enhanced accessibility of dual active sites. Collectively, these factors synergistically contribute to enhanced catalytic activity. These results conclusively demonstrated that spatial confinement enhanced catalytic activity by stabilizing highly dispersed active species and optimizing pore architecture, thereby maximizing active site accessibility.

Kinetic analysis (Fig. 2(b)) further elucidated performance drivers through turnover frequency (TOF) correlations. TOF exhibited an inverse dependence on Cu⁺/Cu⁰ ratios but scaled positively with Cu dispersion, establishing low oxidation states and high metal dispersion as critical for catalytic efficiency. The nanoconfined Cu_in/SBA-15 system achieved a TOF 1.16 times greater than its non-confined counterpart, highlighting interfacial synergy between Cu nanoparticles and the SiO₂ framework. This synergistic interaction likely stabilizes metastable Cu species while facilitating electron transfer at Cu–O–Si interfaces, thereby enhancing reaction dynamics under confinement conditions.

To more comprehensively evaluate the catalytic performance of the Cu_in/SBA-15 catalyst developed in this study, we have summarized and compared its CO conversion and TOF values with those of representative Cu-based WGSR catalysts reported in the literature, as shown in Fig. 2(c) and Table S1.† Catalysts supported on conventional SiO₂ or silica zeolite frameworks (e.g., Cu/Silicalite-1 and Cu/SBA-16) typically exhibit lower CO conversions in the range of 16–45% at 240 °C, even when the copper loading is comparable.

In contrast, the Cu_in/SBA-15 catalyst achieved a significantly higher CO conversion of 75.8% under similar reaction conditions and metal content, outperforming both the randomly dispersed Cu/SBA-15 (55.6%) and the externally deposited Cu_out/SBA-15 (38.5%) catalysts. Notably, Cu_in/SBA-15 also exhibited a markedly higher TOF compared with several benchmark systems, including 10Cu/CeO₂, 5Cu5Ni/CeO₂, Cu/ZnO, and Cu/ZnO/La, despite the use of reducible oxide supports in those systems. These results clearly demonstrate that the superior catalytic performance of Cu_in/SBA-15 is not solely attributable to copper loading or support type, but rather to the precise spatial confinement of Cu nanoclusters within the mesoporous SBA-15 channels.

Such confinement effectively suppresses particle migration and sintering, while simultaneously enabling the formation of stable Cu–O–Si interfacial structures within the pores. These interfaces help to dynamically stabilize the Cu⁺/Cu⁰ redox pair and promote a balanced interplay between H₂O dissociation and CO adsorption, which are two essential steps in the WGSR pathway.

In summary, although a wide variety of Cu-based WGSR catalysts have been reported (Table S1†), most studies lack precise control and in situ monitoring of the copper valence state and interfacial dynamics during the reaction. The domain-confinement strategy proposed in this work effectively stabilizes both the reactive redox state and interfacial structure, while maintaining a high dispersion of active species.

3.3 Catalytic stability

Long-term stability evaluation (Fig. 2(d)) demonstrated remarkable durability, with Cu_in/SBA-15 maintaining stable performance over 50 hours of operation at 240 °C. Deactivation pathways of copper-based catalysts are generally discussed from two distinct perspectives.⁴² One hypothesis is that the catalyst agglomerates, resulting in a reduction in active sites, and the other is that the active valence state of the active sites changes during the reaction. Mechanistic investigation of deactivation pathways (Fig. 2(e)) revealed reversible activity loss upon water vapor withdrawal, excluding permanent structural degradation. In addition, post-reaction XRD analysis confirmed the absence of copper sintering, supporting valence state alteration as the primary deactivation mechanism. Specifically, the Cu⁺ sites tended to become overly stabilized by adsorbed CO, disrupting the dynamic Cu⁺/Cu⁰ equilibrium and leading to decreased activity. From an applied perspective, long-term durability and resistance to thermal cycling are crucial. In the cyclic start-stop test, the reaction temperature was increased to 240 °C and maintained for three hours, followed by cooling to 25 °C for one hour. As illustrated in Fig. 2(f), the Cu_in/SBA-15 catalyst demonstrated a relatively high level of catalytic activity after ten cycle tests, even slightly higher than the initial activity, demonstrating its potential industrial application. Overall, the comparative analysis clearly demonstrated that copper distribution patterns critically affect both activity and stability, where confinement within SBA-15 mesopores facilitates optimal metal-support interaction, redox balance, and dispersion, leading to superior catalytic performance over externally loaded or randomly dispersed systems.

3.4 The role of copper species in the WGSR

Fig. 3(a) presents the H₂-TPR profiles of different catalysts. For the Cu/SBA-15 catalyst, a main reduction peak at 197 °C and a shoulder peak at 173 °C were observed. The shoulder peak at 173 °C may be attributed to the reduction of CuO particles, while the main peak at 197 °C may be ascribed to the reduction of bulk aggregates of CuO.^43,44 In contrast, the Cu_out/SBA-15 catalyst showed a broad reduction region between 200 and 250 °C, with the primary reduction peak occurring at 213 °C, indicating the presence of relatively less uniform and larger CuO particles on the external surface. Notably, the Cu_in/SBA-15 catalyst displayed a narrower reduction peak centered at 194 °C, shifted to a lower temperature compared to the other two samples. This behaviour suggested a higher degree of dispersion and smaller Cu cluster size, consistent with the uniform intrachannel distribution of Cu species confirmed by TEM observations.

Fig. 3 (a) H₂-TPR profile and (b) CO₂-TPD profile and (c) in situ Cu 2p XPS spectra and (d) in situ Cu LMM Auger spectra of Cu/SBA-15, Cu_in/SBA-15 and Cu_out/SBA-15 catalysts.

CO₂-TPD profiles (Fig. 3(b)) elucidated surface basicity variations critical for WGSR equilibrium management. All catalysts exhibited three desorption regimes: α-peak (bidentate carbonate decomposition), β-peak (monodentate carbonate on moderately basic sites), and γ-peak (monodentate carbonate on strongly basic sites),⁴⁵ respectively. The Cu/SBA-15 catalyst exhibited a greater number of weakly basic sites, while the Cu_out/SBA-15 catalyst displayed a greater number of moderately basic sites. The presence of weakly basic sites on the Cu_out/SBA-15 catalyst may result in the formation of carbonate-based reaction products that were challenging to desorb within the specified temperature range. This could potentially cover the active sites of the catalysts,⁴⁶ leading to a reduction in their overall activity. This was consistent with the results shown in Fig. 2(a) above, indicating that the limiting effect of SBA-15 can effectively avoid the coverage of some active sites.

To further investigate the existence state of copper species in copper based catalysts, XPS characterization was performed on the Cu/SBA-15, Cu_in/SBA-15 and Cu_out/SBA-15 catalysts. The XPS energy level spectra of the Cu 2p orbitals for each catalyst are shown in Fig. 3(c), with two primary peaks centered at binding energies of 932.4 eV corresponding to Cu 2p_3/2 and 952.3 eV corresponding to Cu 2p_1/2, attributed to copper species in the reduced state (Cu⁰ or Cu⁺).⁴⁷ Notably, Cu_in/SBA-15 exhibited positive binding energy shifts versus counterparts, signaling enhanced Cu-support electronic interactions. As previously reported, the Cu oxidation state played a crucial role in WGSR performance,^48,49 but since Cu⁰ and Cu⁺ had very similar binding energies, further distinction was made using X-ray excited Auger electron spectroscopy (XAES). The Cu LMM Auger spectra (Fig. 3(d) and Table 1) identified two components, located at 918.12 eV (Cu⁺) and 914.37 eV (Cu⁰). The relative proportions of Cu species for the three samples were also listed and the Cu_in/SBA-15 catalyst had the highest molar ratio of Cu⁰/(Cu⁰+Cu⁺) (74.0%) compared with the other three samples, which was responsible for the shift of the core energy level binding energy of the Cu 2p orbitals of the catalyst towards a higher binding energy, contrasting sharply with the Cu⁺-dominant profile (52% Cu⁺) of Cu_out/SBA-15. The critical Cu⁺/Cu⁰ ratio emerged as a performance-determining parameter. While Cu⁺ facilitates CO adsorption and Cu⁰ promotes H₂O dissociation,^48,50 the synergistic balance governs the overall activity. The Cu⁺/Cu⁰ ratio of Cu_in/SBA-15 optimally reconciles these complementary functions, and sufficient Cu⁺ ensures CO activation without overstabilizing intermediates, while abundant Cu⁰ accelerates water splitting. Based on these results, it is reasonable to believe that on catalysts in this study, a suitable Cu⁺/Cu⁰ ratio is the predominant factor to control the reaction performance.

3.5 Dynamic evolution of the active sites

Operando XPS and Auger spectroscopy tracked dynamic copper speciation under catalytic conditions (Fig. 4(a and b) and ESI Fig. S8†). Post-water treatment at 200 °C, Cu⁺ content increased from 26% to 36% in Cu_in/SBA-15, accompanied by kinetic energy shifts (−1.87 eV for Cu⁰ and −1.14 eV for Cu⁺), indicative of hydroxylation via H₂O dissociation.⁵¹ The larger energy shift for Cu⁰ suggested preferential hydroxylation at metallic sites,⁵² corroborated by faster CO consumption at Cu⁰ observed in CO-DRIFTS. These findings demonstrated water-induced surface restructuring that preferentially activates metallic sites while maintaining structural integrity.

Fig. 4 (a) In situ Cu LMM Auger spectra and (b) changes in the Cu⁺/Cu⁰ ratio under different conditions of Cu/SBA-15, Cu_in/SBA-15 and Cu_out/SBA-15 catalysts; (c) in situ DRIFTS spectra of the Cu_in/SBA-15 catalyst adsorbing CO at 280 °C and introducing water vapor; (d) in situ DRIFTS spectra of the desorption process of the Cu_in/SBA-15 catalyst after CO and H₂O co-adsorption at 280 °C; (e) illustration of the functions of Cu⁺ and Cu⁰ sites and the pathway on the Cu_in/SBA-15 catalyst.

Subsequently, the addition of CO led to the formation of metallic Cu⁰ during the WGSR process. It was worth noting that this process involved both CO-induced reduction of Cu⁺ to Cu⁰ and the participation of H₂ produced during the WGSR process in reducing Cu⁺, thereby inhibiting CO conversion. As observed in Fig. 4(b), after 1 hour of the WGSR, the Cu⁺/Cu⁰ of the Cu_out/SBA-15 catalyst decreased the most. This was because H₂ molecules directly participated in the reduction of Cu⁺ to Cu⁰ on the surface of SBA-15, competing with the main WGSR pathway and resulting in a decreased CO conversion. Therefore, the Cu_out/SBA-15 catalyst exhibited the worst catalytic activity for the WGSR. In contrast, the confined architecture of Cu_in/SBA-15 moderated excessive Cu⁺ depletion, maintaining optimal redox equilibrium through spatial regulation of metal–support interactions.

In situ DRIFTS investigations (Fig. 4(c)) elucidated site-specific behaviours and the Cu_in/SBA-15 catalysts were pre-treated with CO at 280 °C, followed by in situ DRIFTS experiments. Fig. 4(c) shows the co-adsorption process through H₂O after the pre-adsorption of CO. The peaks at 2111 cm⁻¹ and 2130 cm⁻¹ corresponded to the vibrations of Cu⁰–CO and Cu⁺–CO,^53,54 respectively. The intensity of the CO absorption band on the Cu⁰ surface at 2111 cm⁻¹ decreased rapidly and finally disappeared after the introduction of water vapor to simulate the feed gas of the reactants. In contrast, the intensity of the CO absorption band at 2130 cm⁻¹ on the Cu⁺ surface decreases slowly, and the spectral band generated between 3000 and 3500 cm⁻¹, attributed to the stretching vibrations of the O–H bond, gradually increased.⁵⁵ This suggested that H₂O can dissociate more easily on the Cu⁰ surface and form CO₂ by the WGSR with adsorbed CO molecules on the Cu⁰ surface, whereas the slowly decreasing intensity of the CO–Cu⁺ band suggested that H₂O can also dissociate on the Cu⁺ site, but the dissociation process was slower. Fig. 4(d) shows the desorption process after co-adsorption at the same temperature. The spectral bands near 2090, 2111, 2130, and 2170 cm⁻¹ were detected corresponding to the vibrations of Cu⁰–CO, Cu⁺–CO, and gaseous CO, respectively, confirming the co-existence of the Cu⁰ and Cu⁺ sites, but the adsorption of CO at the Cu⁰ site was unstable. The Cu⁰–CO and gaseous CO peaks disappear rapidly within the first 4 minutes but the Cu⁺–CO peak decreased much more slowly and reached stability after 12 minutes. Thus, Cu⁺ species may be the main active site for adsorption and activation of CO molecules. Based on these findings, a dual-site mechanistic framework can be shown in Fig. 4(e). The H₂O molecule was adsorbed and activated at the Cu⁰ site and then produced *OH and *H, while the CO molecule was adsorbed and activated mainly at the Cu⁺ site and then gradually converted to become CO₂, which was the representative intermediate of the mechanism of the reaction, and a number of other intermediate products. The confined architecture in Cu_in/SBA-15 enabled spatial orchestration of these complementary processes, with Cu⁰ sites driving rapid hydroxyl generation while Cu⁺ sites sustain CO activation, and their synergy was moderated through controlled electron transfer at Cu–O–Si interfaces.

3.6 Proposed reaction mechanism over the Cu–O–Si interfaces

For Cu-based catalysts, two principal catalytic mechanisms have been established: the redox mechanism and the associative mechanism.^56,57 The redox pathway mechanism⁵⁸ involves sequential oxidation and reduction steps where H₂O dissociates into surface-bound OH and H species. These hydrogen species subsequently combine to form H₂, while adsorbed CO reacts with oxygen from water dissociation to generate CO₂. In contrast, the associative mechanism proposed by Armstrong and Hildtich⁵⁹ initiates with H₂O dissociation into reactive H and OH species. These intermediates then interact with CO to form transient species that ultimately decompose into CO₂ and H₂. This associative pathway further branches into formate and carboxylate mechanisms based on the specific intermediates involved, with the distinction lying in the structural characteristics of these transitional complexes.

To unravel the dominant pathway in Cu/SBA-15 systems, we focused on the high-performance Cu_in/SBA-15 catalyst. Fig. 5(a) shows the in situ DRIFTS profiles of CO and H₂O co-adsorption over the Cu_in/SBA-15 catalyst at different temperature points, simulating the process of the WGSR. It can be observed that the bands in the range of 1000–1700 cm⁻¹ correspond to both symmetric and asymmetric stretching vibrations of OCO, but both carbonates and formates contain OCO stretching vibrations.^60,61 In addition, peaks in the region of 1550–1650 cm⁻¹ were detected, indicating the formation of a single bidentate formate species,⁶² while the peaks at 2700–3000 cm⁻¹ were associated with the C–H stretching vibration of the formate species.^63–65 The intensity of all the peaks increased with decreasing temperature, which may be due to the fact that formate species with COO vibrations increased the stability of the active metal with the support at lower temperatures, and strongly and irreversibly adsorbed carbonate species are retained on the metal surface, leading to a decrease in the catalytic activity, a result consistent with the activity results. Therefore, it can be tentatively concluded that the catalyst followed the associative mechanism.

Fig. 5 (a) In situ DRIFTS study of the Cu_in/SBA-15 catalyst under a WGSR gas mixture; (b) normalized MS signals for CO₂ and H₂ evolution during CO + H₂O-TPSR on the equilibrated Cu_in/SBA-15 catalyst; (c) illustration of two mechanisms of the WGSR.

On the other hand, as the temperature increases, the carbonate species decomposes more readily into CO₂. The results showed that at higher temperatures, there were two distinct absorption bands at 2322 and 2369 cm⁻¹ indicating the presence of asymmetric stretching of the CO₂ molecule and that the intensity of these absorption bands decreased with decreasing temperature.^66,67 Thus, it can be described that the WGSR of formic acid species as intermediates and the decomposition of formic acid were considered as rate-determining steps. Briefly, the WGSR process over the Cu_in/SBA-15 catalyst can be outlined as the production of reactive surface OH groups on silica sites and the formation of formate species (HCOO–) following a reduction process, followed by the decomposition of the bicarbonate formate species into bidentate formate species, which were finally decomposed to produce H₂ and CO₂, resulting in a decrease in the intensities of the bands associated with formate species and a significant increase in the bands associated with CO₂. Complementary TPSR experiments (Fig. 5(b)) employing mass spectrometry provided mechanistic validation.⁶⁸ The Cu_in/SBA-15 catalyst was reduced and then treated with CO and H₂O for 2 hours to obtain the real catalytic reaction interface structure. After purging with helium, the catalyst was exposed to H₂O for 2 hours by exchanging helium with water vapour at 280 °C (H₂O dissociation experiment). It was then treated at 280 °C for 1 hour under a helium atmosphere to remove excess adsorbed water from the surface. After cooling to room temperature, the helium was converted to CO and reacted at 280 °C for 2 hours (CO-TPSR experiment). The output gas was analysed with m/z = 44 (CO₂) and 2 (H₂).

In the redox mechanism, CO reacts with reactive oxygen species at the catalytic interface to form CO₂ and O vacancies, and H₂O dissociates at the oxygen vacancies to form hydrogen. The process of the associative mechanism is shown in Fig. 5(c), including the dissociation of H₂O molecules directly at the O vacancies to form hydroxyl groups and the reaction of CO with the hydroxyl groups to produce CO₂ and H₂ in a yield of approximately 2 : 1. There was no direct appearance of the H₂ signal, suggesting that the reaction process did not involve the redox mechanism. In the following CO-TSPR experiments, the concentration ratio of CO₂ to H₂ remained close to 2, further demonstrating that the reaction followed the associative mechanism.

Isotopic substitution experiments employing D₂O as a probe molecule (Fig. 6(a)) yielded critical kinetic insights. By comparing the reaction in the presence of H₂O or D₂O, the difference in the reaction rates (k_H/k_D), known as the kinetic isotope effect (KIE), reveals whether protons are involved in the rate-determining step (RDS) of the reaction.⁶⁹ As shown in Fig. 6(b), by comparing the rate constants of water oxidation under H₂O and D₂O, a KIE (k_H/k_D) value of 1.46 was obtained for Cu_in/SBA-15, indicating a first-order kinetic isotope effect.¹ This significant KIE value suggested that the HCOO* breakage was involved in the rate-determining step of the WGSR.⁷⁰ The microscopic reaction pathway of the WGSR is shown in Fig. 6(c), where the HCOO* intermediate was formed at the Cu–O–Si interface, followed by slow decomposition to form carbon dioxide and hydrogen gas.

Fig. 6 (a) In situ DRIFTS spectra exposed to the stream of CO + H₂O and subsequently switched to CO + D₂O and (b) KIE test of Cu_in/SBA-15; (c) illustration of the reaction pathway of the WGSR at the Cu_in/SBA-15 interface.

4. Conclusion

In conclusion, a Cu_in/SBA-15 catalyst with a domain-limited structure was prepared for the first time, which exhibited a reaction rate of 5.4 μmol_CO‧g_cat⁻¹‧s⁻¹ in the WGSR, and furthermore, the Cu confined in the channels showed excellent durability for more than 50 h and 10 start–stop cycles. In situ characterisation confirmed the presence of Cu species in the form of Cu⁰ and Cu⁺, where Cu⁰ species were mainly involved in the dissociation process of H₂O, while Cu⁺ species had strong adsorption capacity for CO and participated in the activation process. The microstructures illustrated that the mesoporous materials provided a spatial confinement effect. This structure stabilises the Cu⁺ and Cu⁰ species in the nanoparticles through the formation of a large number of Cu–O–Si interfaces, effectively avoiding the aggregation and sintering of copper species. It reveals that the domain-limited Cu-based catalysts follow an association mechanism with *HCOO dissociation as the rate-determining step during the WGSR by TPSR and in situ DRIFTS and transient isotope switching experiments. This work provides an effective strategy for the construction of Cu⁺ and Cu⁰ catalytic systems with equilibrium stability, which has great potential for practical applications.

Data availability

All relevant data are within the manuscript and its additional files.

Author contributions

Yue Chen: writing – original draft, investigation, and formal analysis. Yu Ma: validation and formal analysis. Chen Liao: validation and data curation. Jiacheng You: writing – review & editing and data curation. Huihuang Fang: supervision and writing – review & editing. Chongqi Chen: writing – review & editing, supervision, and funding acquisition. Yu Luo: supervision and project administration. Lilong Jiang: supervision, project administration, and conceptualization.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (Grant No. 22078062) and the Science Fund for Creative Research Groups of the National Natural Science Foundation of China (22221005).

References

1. T. K. Phung, Int. J. Hydrogen Energy, 2022, 47, 42234–42249.
2. J. Pang, M. Zheng, C. Wang, X. Yang, H. Liu, X. Liu, J. Sun, Y. Wang and T. Zhang, ACS Catal., 2020, 10, 13624–13629.
3. S. Bepari, M. Khan, X. Li, N. Mohammad and D. Kuila, Top. Catal., 2023, 66, 375–392.
4. J. Lin, C. Hu, X. Xu, M. Shao, Y. Hu and C. Ma, Chem. Eng. Technol., 2021, 44, 1121–1130.
5. M. Tommasi, D. Ceriotti, A. Gramegna, S. N. Degerli, G. Ramis and I. Rossetti, Catalysts, 2024, 14, 2073–4344.
6. M. Saito, K. Tomoda, I. Takahara, K. Murata and M. Inaba, Catal. Lett., 2003, 89, 11–13.
7. S.-Y. Ahn, K.-J. Kim, B.-J. Kim, J.-O. Shim, W.-J. Jang and H.-S. Roh, Appl. Catal., B, 2023, 325, 122320.
8. C. H. Lee, S. Kim, H. J. Yoon, C. W. Yoon and K. B. Lee, Renewable Sustainable Energy Rev., 2021, 145, 111064.
9. D. Li, X. Li and J. Gong, Chem. Rev., 2016, 116, 11529–11653.
10. Y. Chen and Y. Huang, Int. J. Hydrogen Energy, 2023, 48, 1323–1336.
11. Y. Zhang, H. Zhu, T. Zhang, J. Li, J. Chen, Y. Peng and J. Li, Environ. Sci. Technol., 2022, 56, 1917–1926.
12. M. Li, A. Borsay, M. Dakhchoune, K. Zhao, W. Luo and A. Züttel, Appl. Surf. Sci., 2020, 510, 145439.
13. P. Wynblatt and N. A. Gjostein, Acta Metall., 1976, 24, 1165–1174.
14. T. W. Hansen, A. T. DeLaRiva, S. R. Challa and A. K. Datye, Acc. Chem. Res., 2013, 46, 1720–1730.
15. Y. Gu, C. Han, J. Huang, V. A. Vinokurov and W. Huang, Fuel, 2022, 322, 124111.
16. X. Fang, C. Chen, X. Lin, Y. She, Y. Zhan and Q. Zheng, Chin. J. Catal., 2012, 33, 425–431.
17. X.-S. Zhang, H.-T. Zhao, Y. Liu, W.-Z. Li, N. Luo and J. Luan, Dalton Trans., 2022, 51, 17319–17327.
18. J. Zhang, W. Liu, Q. Wang, F. Ning, Q. He, G. Li, C. Liu, Z. Li and H. Peng, Sep. Purif. Technol., 2024, 346, 127544.
19. P. Cao, L. Lin, H. Qi, R. Chen, Z. Wu, N. Li, T. Zhang and W. Luo, ACS Catal., 2021, 11, 10246–10256.
20. C. S. Chen, Y. T. Lai, T. W. Lai, J. H. Wu, C. H. Chen, J. F. Lee and H. M. Kao, ACS Catal., 2013, 3, 667–677.
21. X. Li, R. Zhu, B. Sun, L. Chen, H. Yang, X. Liu, Y. Guo and Y. Wang, Chem. Eng. J., 2024, 487, 150652.
22. M. Ma, Y. Zhang, C. Gao, G. Liu, C. Cui, Duoni, Q. Hu, A. S. Hunaidy, M. A. Moniee, Y. A. Dawsari, A. H. Alahmed and W. Qian, Catal. Today, 2024, 437, 114781.
23. W.-G. Cui, Y.-T. Li, L. Yu, H. Zhang and T.-L. Hu, ACS Appl. Mater. Interfaces, 2021, 13, 18693–18703.
24. C. Xu, G. Chen, Y. Zhao, P. Liu, X. Duan, L. Gu, G. Fu, Y. Yuan and N. Zheng, Nat. Commun., 2018, 9, 3367.
25. H. Zhang, H. Sun, D. Zhang, W. Zhang, S. Chen, M. Li and P. Liang, Chem. Eng. J., 2019, 361, 139–147.
26. V. J. Margarit, M. Osman, S. Al-Khattaf, C. Martínez, M. Boronat and A. Corma, ACS Catal., 2019, 9, 5935–5946.
27. S. Liang, L. Zou, L. Zheng, F. Li, X. Wang, L. Song and J. Xu, Adv. Energy Mater., 2022, 12, 2103097.
28. H. Yan, X.-T. Qin, Y. Yin, Y.-F. Teng, Z. Jin and C.-J. Jia, Appl. Catal., B, 2018, 226, 182–193.
29. J. Ma, L. Mao, H. Du, J. Zhong, L. Jiang, X. Liu, J. Xu, X. Xu, X. Fang and X. Wang, Chem. Eng. J., 2024, 496, 154195.
30. X.-C. Li, J.-H. Wang, T.-T. Huang, Y. Hu, X. Li, D.-J. Wang, W.-W. Wang, K. Xu, C.-J. Jia, H. Dong, G. Li, C. Li and Y.-W. Zhang, Nat. Commun., 2025, 16, 736.
31. D. Zhao, J. Feng, Q. Huo, N. Melosh, G. H. Fredrickson, B. F. Chmelka and G. D. Stucky, Science, 1998, 279, 548–552.
32. J. Yu, X. Qin, Y. Yang, M. Lv, P. Yin, L. Wang, Z. Ren, B. Song, Q. Li, L. Zheng, S. Hong, X. Xing, D. Ma, M. Wei and X. Duan, J. Am. Chem. Soc., 2024, 146, 1071–1080.
33. N. Liu, S. Cui, Z. Jin, Z. Cao, H. Liu, S. Yang, X. Zheng and L. Wang, Processes, 2023, 11, 1353.
34. F. Lan, J. Aarons, Y. Shu, X. Zhou, H. Jiao, H. Wang, Q. Guan and W. Li, Appl. Catal., B, 2021, 299, 120651.
35. P. H. K. Charan and G. R. Rao, J. Chem. Sci., 2015, 127, 909–919.
36. G. D. Mihai, V. Meynen, M. Mertens, N. Bilba, P. Cool and E. F. Vansant, J. Mater. Sci., 2010, 45, 5786–5794.
37. J. A. Colín-Luna, G. E. Zamora-Rodea, A. K. Medina-Mendoza, L. Alvarado-Perea, C. Angeles-Chávez, J. Escobar, J. G. Pacheco-Sosa and J. C. García Martínez, Catal. Today, 2020, 353, 63–72.
38. Z. Zhang, S.-S. Wang, R. Song, T. Cao, L. Luo, X. Chen, Y. Gao, J. Lu, W.-X. Li and W. Huang, Nat. Commun., 2017, 8, 488.
39. B. Zhang, Y. Chen, J. Li, E. Pippel, H. Yang, Z. Gao and Y. Qin, ACS Catal., 2015, 5, 5567–5573.
40. C. Marras, D. Loche, D. Carta, M. F. Casula, M. Schirru, M. G. Cutrufello and A. Corrias, ChemPlusChem, 2016, 81, 421–432.
41. Z. Bian, W. Zhong, Y. Yu, B. Jiang and S. Kawi, Int. J. Hydrogen Energy, 2020, 45, 27078–27088.
42. D. Li, Z. Wang, S. Jin and M. Zhu, Sci. China:Chem., 2023, 66, 3645–3652.
43. F. Morales-Anzures, P. Salinas-Hernández, C. Ornelas-Gutiérrez, F. J. Tzompantzi-Morales and R. Pérez-Hernández, Catal. Today, 2020, 349, 228–234.
44. L. Ma, M.-F. Luo and S.-Y. Chen, Appl. Catal., A, 2003, 242, 151–159.
45. Y. Shao, J. Wang, K. Sun, G. Gao, M. Fan, C. Li, C. Ming, L. Zhang, S. Zhang and X. Hu, ACS Appl. Nano Mater., 2022, 5, 5882–5894.
46. Y. Shao, K. Sun, L. Zhang, Q. Xu, Z. Zhang, Q. Li, S. Zhang, Y. Wang, Q. Liu and X. Hu, Green Chem., 2019, 21, 6634–6645.
47. H. Choi and M. Kang, Int. J. Hydrogen Energy, 2007, 32, 3841–3848.
48. C. Chen, H. Ren, Y. He, Y. Zhan, C. Au, Y. Luo, X. Lin, S. Liang and L. Jiang, ChemCatChem, 2020, 12, 4672–4679.
49. H. Yan, X.-T. Qin, Y. Yin, Y.-F. Teng, Z. Jin and C.-J. Jia, Appl. Catal., B, 2018, 226, 182–193.
50. Z. Cui, S. Song, H. Liu, Y. Zhang, F. Gao, T. Ding, Y. Tian, X. Fan and X. Li, Appl. Catal., B, 2022, 313, 121468.
51. S. Navarro-Jaén, M. Á. Centeno, O. H. Laguna and J. A. Odriozola, J. Mater. Chem. A, 2018, 6, 17001–17010.
52. Y. Fang, C. Zhao, Z. Gong, S. Wang, X. Ye, Q. Pan, G. Li, Y. Cui, Y. Yao and W. Luo, Chem. Phys. Lett., 2022, 794, 139496.
53. S. Poulston, P. M. Parlett, P. Stone and M. Bowker, Surf. Interface Anal., 1996, 24, 811–820.
54. Z. Liu, F. Yu, K. Pan, X. Zhou, R. Sun, J. Tian, Y. Wan, J. Dan and B. Dai, Chem. Eng. J., 2021, 422, 130099.
55. X. Li, S. Ren, Z. Chen, L. Chen, M. Wang, L. Wang and A. Wang, Fuel, 2023, 347, 128435.
56. Y. Shen, Z. Wang, S. Ge, L. Wang, W. Zhan, Q. Dai, Y. Guo, Y. Guo and A. Wang, Appl. Catal. B: Environ., 2024, 353, 124094.
57. M. Zhu and I. E. Wachs, ACS Catal., 2016, 6, 2827–2830.
58. D. Li, Y. Li, X. Liu, Y. Guo, C.-W. Pao, J.-L. Chen, Y. Hu and Y. Wang, ACS Catal., 2019, 9, 9671–9682.
59. P. Yoo and P. Liao, Appl. Surf. Sci., 2023, 630, 157501.
60. E. F. Armstrong and T. P. Hilditch, Proc. R. Soc. London, 1920, 97, 265–273.
61. J. Yu, M. Yang, J. Zhang, Q. Ge, A. Zimina, T. Pruessmann, L. Zheng, J.-D. Grunwaldt and J. Sun, ACS Catal., 2020, 10, 14694–14706.
62. J. Wu, S. Yang, Q. Liu, P. He, H. Tian, J. Ren, Z. Guan, T. Hu, B. Ni and C. Zhang, Environ. Sci. Technol., 2016, 50, 5370–5378.
63. A. Razeghi, A. Khodadadi, H. Ziaei-Azad and Y. Mortazavi, Chem. Eng. J., 2010, 164, 214–220.
64. N. Zhang, L. Li, Y. Jing, Y. Qian, D. Chen, N. Maeda, T. Murayama, T. Toyao and K. Shimizu, J. Catal., 2024, 433, 115500.
65. K. Sun, M. Kohyama, S. Tanaka and S. Takeda, J. Phys. Chem. C, 2017, 121, 12178–12187.
66. Z. Li, N. Li, N. Wang, B. Zhou, P. Yin, B. Song, J. Yu and Y. Yang, ACS Omega, 2022, 7, 3514–3521.
67. S. Kattel, P. J. Ramirez, J. G. Chen, J. A. Rodriguez and P. Liu, Science, 2017, 355, 1296–1299.
68. W. Wu, Y. Wang, L. Luo, M. Wang, Z. Li, Y. Chen, Z. Wang, J. Chai, Z. Cen, Y. Shi, J. Zhao, J. Zeng and H. Li, Angew. Chem., Int. Ed., 2022, 61, e202213024.
69. H. Yan, C. Yang, W.-P. Shao, L.-H. Cai, W.-W. Wang, Z. Jin and C.-J. Jia, Nat. Commun., 2019, 10, 3470.
70. T. Shido, K. Asakura and Y. Iwasawa, J. Chem. Soc., Faraday Trans., 1989, 85, 441.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5ta03435c

---