Light‐Driven Charge Redistribution of Pt Cluster/CeO₂ Realizing Dual‐Path Synergistic Catalysis for Low‐Temperature Reverse Water‐Gas Shift

ABSTRACT

Solar‐driven low‐temperature reverse water‐gas shift offers a sustainable route for CO₂ conversion yet suffers from insufficient efficiency and unclear reaction mechanisms. Herein, we demonstrate that light drives the surface charge redistribution of a Pt cluster/CeO₂ catalyst, unlocking synergistic dual pathways for enhanced CO production. Using light as the sole energy input (2.27 W/cm²), the catalyst surface reaches a localized temperature of ∼309°C while the reactor environment remains at only ∼54°C. In a continuous‐flow system with cold inlet gases, this catalyst achieves a CO production rate of 846.9 mmol g_cat ⁻¹ h⁻¹, outperforming conventional thermal systems. Remarkably, comparable performance is achieved using natural sunlight alone, even under outdoor ambient temperature of −21°C. Mechanistic studies reveal that light‐driven interfacial charge redistribution constructs the nonequilibrium Pt^δ+‐O_V‐Ce³⁺ structure, which promotes CO₂ activation, triggering the carboxylate pathway and enhancing the formate route, giving rise to a cooperative effect that significantly accelerates the overall reaction. This stands in stark contrast to the single formate route that dominates conventional thermal catalysis. This work establishes light as a dynamic regulator for engineering catalytic sites, offering a promising strategy for efficient solar energy conversion.

Light‐driven interfacial charge redistribution of a Pt cluster/CeO₂ catalyst reconstructs Pt^δ+‐O_V‐Ce³⁺ sites, enabling concurrent formate and carboxylate pathways, a fundamental shift from single‐path thermal catalysis. Using light only, localized catalyst heating reaches 309°C while the bulk reactor stays at 54°C, delivering a CO production rate of 846.9 mmol g_cat ⁻¹ h⁻¹ in continuous‐flow operation, establishing a new paradigm for solar‐driven catalysis.

1. Introduction

The conversion of CO₂ into value‐added chemical feedstocks is a critical step towards a sustainable energy future. Among various catalytic routes, the reverse water‐gas shift (RWGS) reaction is an effective pathway for converting CO₂ to CO, which can be further upgraded via downstream processes such as Fischer–Tropsch synthesis [1, 2]. However, activating the thermodynamically stable CO₂ molecule remains a key challenge, as high activity typically requires temperatures exceeding 500°C to achieve practical reaction rates. Conventional thermal catalysis demands energy‐intensive heating of the entire reactor environment, leading to high energy costs, operational safety problems, and unavoidable catalyst degradation through sintering and coking [3, 4]. Therefore, the primary goal is to construct catalytic systems that exhibit high activity and selectivity under significantly lower temperatures and more energy‐mild conditions.

Utilizing solar energy to drive CO₂ conversion emerges as a powerful and strategic solution. Light serves as a clean energy source with a unique dual function: it enables highly localized photothermal heating at the catalyst surface, raising active sites to optimal reaction temperatures (200°C–450°C) [5] while the bulk reactor and inlet gases remain near ambient conditions, dramatically reducing energy consumption and enhancing operational safety. More importantly, light excitation generates electron–hole pairs that activate distinct photochemical pathways, which lowers apparent activation energies and creates transient, highly active surface states inaccessible under thermal equilibrium [6]. Recent studies have demonstrated that light excitation can induce and stabilize surface defects, facilitate metal‐support interfacial electron transfer, enhance hydrogen spillover, and drive photo‐induced surface restructuring, enabling efficient light‐driven low‐temperature reactions such as methane dry reforming and plastic upcycling [7, 8, 9]. These advances position light‐driven catalysis as a promising route toward low‐temperature, energy‐efficient RWGS. Recently, Xiong and coworkers demonstrated that atomically dispersed Cu sites tailor oxygen vacancies (O_V) in In₂O₃ to achieve efficient photothermal CO₂ conversion, highlighting the critical role of metal dispersion and O_V engineering [10]. It should be noted that achieving efficient continuous‑flow photothermal RWGS is critical for practical applications. Recently, several studies have demonstrated promising continuous‑flow photothermal RWGS systems, highlighting the feasibility of this approach [10, 11, 12]. Further enhancing the catalytic performance to meet practical demands rational catalyst design that fully harnesses light‐induced effects, thereby enabling rapid and stable conversion of reactant molecules under continuous reaction conditions.

Pt/CeO₂ represents a classic and highly efficient catalytic system that is exceptionally well‐suited to address this challenge. CeO₂ itself is an ideal platform, leveraging its dynamic redox activity (Ce⁴⁺/Ce³⁺) and efficient photon‐harvesting capability. Its low O_V formation energy and localized 4f electrons provide exceptional surface tunability for activating the stable CO₂ molecule [13]. Concurrently, the loaded Pt species, typically in the form of clusters, efficiently dissociate H₂ to supply reactive hydrogen species. This synergy endows Pt/CeO₂ with excellent catalytic performance under high‐temperature thermal conditions. However, under mild temperatures relevant to low‐temperature RWGS, this synergy is often restricted [14]. The generation of O_V is governed by thermodynamic equilibrium, resulting in a limited density that is difficult to regulate under mild conditions, thereby fundamentally constraining the efficiency of CO₂ activation. Moreover, most reported RWGS catalysts to date operate via a single dominant reaction pathway, defining the activity bottleneck of thermal catalysis [15]. Light excitation provides a unique alternative to overcome these constraints. Beyond localized heating, photogenerated charge carriers can dynamically reconfigure the Pt‐CeO₂ interface to modulate catalytic behavior, yet the underlying mechanistic details remain largely ambiguous. Key questions persist regarding metal size optimization and their interactions with the support under illumination, and how these light‐tailored configurations dictate the reaction pathway. Addressing these questions requires a catalyst platform that not only responds efficiently to light but also possesses well‐defined, controllable interfacial sites capable of transducing photogenerated charge carriers into targeted chemical transformations.

Herein, we demonstrate the light‐driven charge redistribution in a Pt clusters/CeO₂ catalyst, switching the RWGS pathway from a single formate route in conventional thermal catalysis to a light‐driven synergistic dual‐path mechanism, giving highly efficient RWGS. Light excitation drives charge redistribution, generating Ce³⁺ and oxygen vacancies (O_v), reconstructing the interface into Pt^δ+‐O_V‐Ce³⁺ ensembles. This nonequilibrium structure promotes CO₂ activation and enables dual‐pathway cooperation, representing a mechanistic shift from the single‐path limitation of thermal catalysis and thereby accelerating the reaction. Using light as the sole energy source, a localized catalyst surface temperature of ∼309°C is achieved while the bulk reactor environment remains at ∼54°C. The catalyst delivers a high CO production rate of 846.9 mmol g_cat ⁻¹ h⁻¹ in a continuous‐flow system with cold inlet gases, achieving performance comparable with high‐temperature thermal catalysis. Importantly, this system maintains its high performance even at an outdoor ambient temperature of −21°C using only natural sunlight. This work establishes a new paradigm for solar energy utilization through dynamically programmable catalytic interfaces.

2. Results and Discussion

2.1. Synthesis and Characterization of the Pt_C/CeO₂

Pt clusters/CeO₂ (Pt_C/CeO₂) catalyst is synthesized via an amine‐assisted anchoring strategy based on the interaction of ethylenediamine (EDA), Pt source, and CeO₂ precursor. The synthetic procedure is schematically illustrated in Figure 1a. Firstly, a spherical CeO₂ precursor (denoted as CeO₂‐P) with abundant hydroxyl groups is prepared by a solvothermal method, and X‐ray diffraction (XRD) patterns exhibit a mixed‐phase composition of CeO₂, Ce(HCOO)₃, and C₃H₆CeO₆ (Figure 1f). Subsequently, the H₂PtCl₆ solution pre‐coordinated with EDA is introduced into the CeO₂‐P suspension. Utilizing EDA as a bifunctional bridging ligand, the Pt‐EDA complexes are uniformly anchored onto the CeO₂‐P surface through coordination between the amine groups and both the Pt source and the surface hydroxyl groups (Figure S1) [16]. After sequential calcination in air and reduction in a hydrogen atmosphere, a Pt_C/CeO₂ catalyst is obtained. During thermal treatment, the strong anchoring effect of the EDA ligand effectively suppresses the migration and aggregation of Pt species on the CeO₂ surface, thereby promoting the formation of the highly dispersed Pt species. Based on this strategy, a single‐atom Pt/CeO₂ catalyst (Pt_SA/CeO₂) can also be synthesized by adjusting the synthesis conditions. A series of characterizations confirms the successful preparation of the Pt_SA/CeO₂ catalyst (Figures S2–S4, Table S1). For comparison, the Pt_N/CeO₂ catalyst is also prepared via direct impregnation of Pt precursor onto pre‐calcined CeO₂ (derived from CeO₂‐P), followed by the same calcination and reduction procedures. Inductively coupled plasma atomic emission spectroscopy (ICP‐OES) confirms that the Pt loadings in the optimal catalyst are approximately 2.1 wt%, as established through subsequent catalytic performance evaluation (Figure S5). Meanwhile, the Pt loading of the Pt_N/CeO₂ and Pt_SA/CeO₂ catalysts is also similar (Table S2).

FIGURE 1.

(a) The synthetic schematic illustration of the Pt_C/CeO₂ catalyst. (b–e) SEM, TEM, and HAADF‐STEM images of Pt_C/CeO₂. (f) XRD patterns of the CeO₂‐P, Pt_C/CeO₂, and Pt_N/CeO₂. (g) CO‐DRIFTS spectra of Pt_C/CeO₂ and Pt_N/CeO₂ catalysts. (h) EDS elemental mapping of Ce and Pt in Pt_C/CeO₂ catalyst.

Systematic characterizations reveal the structure of Pt_C/CeO₂ catalyst. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images show that Pt_C/CeO₂ catalyst maintains the uniform spherical morphology with a diameter of about 450 nm (Figure 1b,c). High‐resolution TEM (HRTEM) further reveals distinct lattice fringes with a spacing of 0.19 nm (Figure 1d), corresponding to the (220) plane of the CeO₂ support. The high‐angle annular dark‐field scanning transmission electron microscopy (HAADF‐STEM) image directly visualizes the Pt species on Pt_C/CeO₂ catalyst. As shown in Figure 1e, Pt clusters are highly dispersed on CeO₂ surface, accompanied by some atomically dispersed Pt species. Elemental mapping further confirms the uniform distribution of Pt, Ce and O elements and the existence of Pt clusters on Pt_C/CeO₂ catalyst (Figures 1h and S6). XRD patterns of Pt_C/CeO₂ and Pt_N/CeO₂ catalyst show only the characteristic diffraction peaks of CeO₂ support with no discernible signals corresponding to Pt species, which is typically attributed to their low content and small size (Figure 1f).

The coexistence of Pt cluster and single‐atom Pt for Pt_C/CeO₂ is further proved by CO adsorption diffuse reflectance infrared Fourier transform spectroscopy (CO‐DRIFTS) (Figure 1g). Generally, the absorption peaks at 2099–2081 and 2080–2035 cm⁻¹ are attributed to linearly adsorbed CO on oxidized and metallic Pt sites, respectively [17]. While the broad peak centered at 1835 cm⁻¹ corresponds to bridge‐bonded CO on metallic Pt, which suggests the presence of larger Pt nanoparticles. For Pt_C/CeO₂, two dominant CO absorption peaks attributed to Pt⁰‐CO and Pt^δ+‐CO alongside a week bridge‐bonded CO signal confirm the coexistence of Pt cluster and single‐atom Pt sites [18]. Notably, the heterogeneous structure may lead to the overlap of multiple types of CO adsorption peaks and the formation of the broad CO‐DRIFTS band [19]. In contrast, Pt_N/CeO₂ exhibits a pronounced bridge‐bonded CO peak, indicating the formation of relatively large Pt nanoparticles. Consistent with the CO‐DRIFTS results, TEM images also reveal the uneven distribution of Pt nanoparticles with sizes of about 4 nm for Pt_N/CeO₂ (Figures S7, S8). These collective results clearly demonstrate the critical role of the amine‐assisted anchoring strategy in the formation of highly dispersed Pt structures.

X‐ray absorption fine structure (XAFS) and X‐ray photoelectron spectroscopy (XPS) measurements are conducted to investigate the electronic structure and coordination environments of Pt species. The normalized X‐ray absorption near‐edge structure (XANES) spectra (Figure 2a) shows that the Pt L₃‐edge white line peak of Pt_C/CeO₂ are located between those of Pt foil and PtO₂ reference. This indicates that Pt species in Pt_C/CeO₂ possess a valence range between 0 to +4, which is attribute to the coexistence of Pt clusters and single‐atom Pt species [20, 21]. The extended X‐ray absorption fine structure (EXAFS) spectra (Figure 2b) provide further insight into the local coordination environment of Pt. For Pt_C/CeO₂, the relatively weak Pt‐Pt scattering peaks, attributed to metallic Pt, are observed in the range of 2.09–3.16 Å, confirming the presence of Pt clusters. In addition, a primary peak centered at 1.55 Å is attributable to Pt‐O coordination, which originates from the combined contribution of single‐atom Pt species and highly dispersed Pt clusters on CeO₂ surface [21]. The wavelet transform (WT) analysis also verify this point. Pt_C/CeO₂ present two WT intensity maximums at about 4.9 and 8.0, corresponding to the Pt‐O and Pt‐Pt coordination, respectively (Figure 2c,d). Moreover, EXAFS fitting curves in k space and R space for Pt_C/CeO₂ are shown in Figure S9 and the detailed fitting parameters are listed in Table S3. The fitting results show the Pt‐O coordination number is 2.9 with the bond length of 1.78–1.97 Å. Besides, the Pt‐Pt coordination number is 6 with the bond length of 2.73 Å, which is similar to Pt foil (2.76 Å). The significantly reduced Pt‐O and Pt‐Pt coordination number than PtO₂ and Pt foil suggests the coordination‐unsaturated feature of Pt clusters for Pt_C/CeO₂. Pt 4f XPS spectra are further employed to elucidate the surface chemical states of Pt. As shown in Figure 2e, both Pt_C/CeO₂ and Pt_N/CeO₂ catalysts exhibit three pairs of peaks corresponding to Pt⁰ (70.6/73.2 eV), Pt²⁺ (71.3/74.9 eV) and Pt⁴⁺ (74.1/77.0 eV) species [22]. Generally, Pt⁰ species come from the Pt clusters or nanoparticles, Pt^δ+ species (Pt²⁺ and Pt⁴⁺) are associated with the single‐atom Pt species or Pt‐CeO₂ interfaces. Pt_C/CeO₂ exhibits a significantly higher proportion of Pt^δ+ species compared to Pt_N/CeO₂, confirming the coexistence of Pt clusters and single‐atom Pt species, while the enhanced Pt⁰ signals in Pt_N/CeO₂ suggest the presence of larger Pt nanoparticles. The above results collectively confirm the coexisting Pt cluster and single‐atom Pt structures in Pt_C/CeO₂ catalyst.

FIGURE 2.

(a) XANES spectra of Pt L₃‐edge of Pt_C/CeO₂, Pt foil, and PtO₂ reference. (b) Corresponding EXAFS spectra. (c, d) Wavelet transform of EXAFS signals for Pt_C/CeO₂ and PtO₂. (e) Pt 4f XPS spectra, (f) H₂‐TPR, (g) Ce 3d XPS spectra, (h) O 1s XPS spectra, (i) EPR spectra for Pt_C/CeO₂ and Pt_N/CeO₂ catalysts, (j) the formation energy of oxygen vacancy of different sites. (k) CO₂‐TPD and (l) diffuse reflection UV–vis absorption spectra for Pt_C/CeO₂ and Pt_N/CeO₂ catalysts.

Generally, The Pt‐O‐Ce structures at Pt‐CeO₂ interfaces can activate Ce─O bond of CeO₂, which can promote the formation of O_V and thus achieve efficient catalysis. First, the redox properties of catalyst precursors are studied using H₂ temperature‐programmed reduction (H₂‐TPR). As shown in Figure 2f, the hydrogen consumption peaks in the region of 150°C–300°C are attributed to the reduction of Pt‐O‐Ce structures, and the high‐temperature hydrogen consumption peaks (350°C–600°C) are attributed to the reduction of surface lattice oxygen of CeO₂ [23]. For Pt_C/CeO₂ catalyst, both of these two types of main reduction peaks shift toward lower temperatures relative to Pt_N/CeO₂ catalyst, which is ascribed to the enhanced H‐spillover effect associated with Pt clusters. For Pt_C/CeO₂, an additional peak is observed at approximately 100°C, which may be associated with the reduction of oxygen atoms associated with single‐atom Pt [24]. Above results suggest the greater potential of forming O_V for Pt_C/CeO₂. Moreover, H₂‐TPR profiles also guide the selection of the catalyst reduction temperature. The reduction temperature of 400°C avoids Pt sintering while promoting the formation of O_V on CeO₂ surface, thereby endowing the Pt_C/CeO₂ catalyst with greater catalytic potential. This is verified by the corresponding control experiments with reduction temperatures of 300°C and 500°C. (Figure S10). XPS and electron paramagnetic resonance (EPR) spectra provide the direct evidences for the O_V‐rich feature of Pt_C/CeO₂ catalyst. Figure 2g displays Ce 3d XPS spectra of Pt_C/CeO₂ and Pt_N/CeO₂ catalysts. The Ce 3d XPS spectra can be deconvoluted into eight peaks, in which the peaks at 885.5 and 902.7 eV are assigned to Ce³⁺ species and the rest are assigned to Ce⁴⁺ species [22]. Notably, Pt_C/CeO₂ exhibits the higher percentages (17.2%) of surface Ce³⁺ species compared to Pt_N/CeO₂ catalyst (16.1%). Generally, the Ce³⁺ species are accompanied by the formation of O_V for CeO₂‐based catalyst. This can be further proved by the O 1s XPS spectra. Figure 2h shows the O 1s XPS spectra of Pt_C/CeO₂ and Pt_N/CeO₂ catalysts, which can be deconvoluted into three sub‐peaks at 529.1, 531.2 and 533.2 eV corresponding to the lattice oxygen (O_L), O_V species [25] and adsorbed OH species [26, 27]. The proportion of O_V species in Pt_C/CeO₂ is 27.9%, higher than that in Pt_N/CeO₂ (25.1%). EPR spectra provide strong evidence for the enhancement of O_V in Pt_C/CeO₂. As shown in Figure 2i, Pt_C/CeO₂ exhibits a stronger EPR signal at g = 2.003 relative to Pt_N/CeO₂, indicating a higher density of O_V [28]. Above characterizations collectively confirm the O_V‐rich feature of Pt_C/CeO₂ catalyst. Furthermore, DFT calculations are used to evaluate the formation energies of O_V at Pt‐CeO₂ interface and CeO₂ surface sites. The formation energy of O_V at the Pt‐O‐Ce interface is 1.91 eV, significantly lower than that on the pristine CeO₂ surface (3.15 eV) (Figure 2j). This result indicates that oxygen atoms within the Pt‐O‐Ce interfacial structures can be more readily removed, thereby facilitating O_V formation. The abundant interfacial sites in Pt_C/CeO₂, in contrast to the limited interfaces in Pt_N/CeO₂, accounts for its enhanced capability to generate O_V under reaction conditions.

An increased abundance of O_V can promote the adsorption/activation of CO₂ molecules, which is crucial for RWGS reaction. CO₂‐TPD (CO₂‐temperature programmed desorption) measurements show that the Pt_C/CeO₂ exhibits a greater CO₂ desorption peak than Pt_N/CeO₂ (Figure 2k). This indicates an enhanced CO₂ adsorption/activation capacity for Pt_C/CeO₂ catalyst, which benefits from the O_V‐rich feature of Pt_C/CeO₂ catalyst and is conducive to the subsequent conversion of CO₂ molecules [27]. Moreover, the light absorption property of the catalysts is also modulated, as investigated by UV–visible diffuse reflectance spectroscopy (UV–vis DRS). As shown in Figure 2l, Pt_C/CeO₂ and Pt_N/CeO₂ catalysts exhibit strong and broad light absorption across the UV–visible region (200‐800 nm), implying high utilization efficiency for solar energy. The corresponding band gaps, derived from Tauc plots, are determined to be 2.28 eV for Pt_C/CeO₂ and 2.43 eV for Pt_N/CeO₂. The narrower band gap of Pt_C/CeO₂ suggests enhanced visible light absorption, which is essential for promoting light‐driven catalysis. Further, photoelectrochemical tests demonstrate that the Pt_C/CeO₂ catalyst possesses a more efficient generation and separation of photogenerated electron–hole pairs (Figure S11). Electrochemical impedance spectroscopy (EIS) Nyquist plots also show the markedly decreased charge transfer resistance for the Pt_C/CeO₂ catalyst (Figure S12). All the above results indicate an improved light‐driven catalytic performance for Pt_C/CeO₂.

2.2. The Light‐Driven Catalytic Performance for Low‐Temperature RWGS

The unique architecture of the Pt_C/CeO₂ catalyst, featuring the coexistence of Pt clusters and single‐atom Pt, enhanced O_V and light absorption, indicating its potential in directly light‐driven CO₂ hydrogenation. The catalytic performance was evaluated in a continuous‐flow reactor under full‐spectrum irradiation from a 300 W Xe lamp, without any external heating. As illustrated in the system schematic (Figure 3a), the reactant gases (CO₂/H₂) are introduced at ambient temperature without preheating, and the reaction is directly driven by the localized photothermal heating of the catalyst under illumination. The temperatures at three key locations are measured: the catalyst surface (localized heating sites), the reactor cavity, and the external reactor. As shown in Figure 3b, all measured temperatures exhibit a positive correlation with light intensity, and the differences are significant. Notably, the localized surface temperature of the catalyst shows the steepest rise. For instance, under a reaction atmosphere, the catalyst surface temperature can reach 309°C at an irradiation intensity of 2.27 W/cm², while the temperature of the reactor chamber remains at only ∼54°C, and the external reactor wall is ∼40°C. To further validate the accuracy of the thermocouple measurements, we simultaneously monitored the catalyst surface temperature using both a thermocouple and a thermal imager under an air atmosphere. The excellent agreement between the two methods (Figures S13 and S14) confirms the reliability of our temperature measurements. This significant temperature difference between the catalyst surface and reactor chamber highlights a key advantage of the light‐driven catalysis: it enables high‐temperature catalysis on the catalyst surface, while maintaining a low bulk reactor temperature, thereby dramatically improving energy efficiency and operational safety.

FIGURE 3.

(a) Light‐driven catalytic device diagram. (b) Temperatures in different regions of the reactor. (c) CO₂ conversion and CO selectivity over Pt_C/CeO₂ catalyst under different light intensities and different temperatures (thermal catalysis). (d) Arrhenius plots of CO generation rates under light‐driven and thermal catalysis for Pt_C/CeO₂ catalyst. (e) CO₂ conversion under different light intensities for Pt_SA/CeO_2, Pt_C/CeO₂ and Pt_N/CeO₂ catalysts. (f) Catalytic stability test for Pt_C/CeO₂ catalyst. (g) The comparison of catalytic performance of Pt_C/CeO₂ catalyst with some catalyst reported in the literatures. (h) Schematic illustration and (i) photographs of sunlight‐driven outdoor reaction system. (j) Infrared thermal imaging image of the reaction device (Harbin, Heilongjiang, December 4, 2025, China). (k) Xe lamp/Sunlight‐driven CO₂ conversion rates for Pt_C/CeO₂ catalyst (GHSV = 96000 mL g_cat ⁻¹ h⁻¹, irradiation intensity: 2.75 W/cm²).

The light‐driven CO₂ hydrogenation performance of Pt_C/CeO₂ catalyst was measured under different irradiation intensities. As shown in Figure 3c, as light intensity increases from 1.18 to 2.75 W/cm², the CO₂ conversion rises progressively from 2.98% to 33.21%, while maintaining high CO selectivity (>97%), exhibiting excellent RWGS reaction performance. The light‐driven CO₂ conversion is significantly superior to that obtained under pure thermal conditions at an identical catalyst surface temperature. In detail, CO₂ conversion can reach 22.20% at the irradiation intensity of 2.27 W/cm² (catalyst surface 309°C), which is approximately 2.1 times higher than that achieved under thermal catalytic conditions (10.82%, 309°C). Notably, the CO₂ conversion (33.21%) exceeds the thermodynamic equilibrium conversion at 2.75 W/cm² (catalyst surface 347°C), underscoring the nonequilibrium nature and the superior efficiency of the light‐driven catalytic process. The CO₂ conversion was calculated using the external standard method and further verified by the internal standard method, confirming the accuracy of the performance data (Table S4). The carbon balance under different conversion levels exceeds 98.5%, demonstrating excellent closure of the carbon mass balance (Table S5). Furthermore, the apparent activation energy (E _a) is measured at relatively low conversion levels (below 15%). The E _a of light‐driven catalytic process is 30.24 kJ mol⁻¹, much lower than that of thermal catalysis process (55.09 kJ mol⁻¹) (Figure 3d). The reduced E _a under illumination confirms that the light does not merely act as a heat source but actively lowers the kinetic barrier through light‐induced electronic effect. The positive correlation between illuminated spot diameter and CO₂ conversion (Figure S15) confirms that a larger effective light area enhances catalytic performance.

The performance comparison of Pt_C/CeO₂, Pt_N/CeO₂, and Pt_SA/CeO₂ catalysts elucidates the critical role of catalyst structure in determining catalytic activity. Under identical illumination conditions, all three catalysts reach essentially the same surface temperature (Figure S16). However, Pt_C/CeO₂ exhibits obviously higher CO₂ conversion and CO generation rate than Pt_N/CeO₂ and Pt_SA/CeO₂. For example, at a light intensity of 2.27 W/cm², the CO₂ conversions are 22.20% for Pt_C/CeO₂, 10.28% for Pt_N/CeO₂, and 9.60% for Pt_SA/CeO₂ (Figures 3e and S17). The superior performance of Pt_C/CeO₂ over Pt_N/CeO₂ can be attributed to three factors. First, the higher Pt dispersion in Pt_C/CeO₂ provides more accessible active sites. Second, the strongly coupled Pt cluster/CeO₂ structure generates abundant Pt–CeO₂ interfaces with unique electronic configurations, which can induce dynamic active states under light excitation. Third, its abundant Pt‐O‐Ce interfacial sites promote oxygen vacancy formation (Figure 2j). Additionally, the coexisting single atoms may further contribute to O_V formation [16]. Compared with the single‐atom catalyst (Pt_SA/CeO₂), the Pt cluster structure offers a decisive advantage in hydrogen dissociation. DFT calculations (Figure S18) show that the hydrogen dissociation barrier on isolated Pt single atoms is 0.94 eV, significantly higher than that on Pt clusters (0.35 eV) [29]. This theoretical prediction is experimentally confirmed by hydrogen‐deuterium (H‐D) exchange experiments (Figure S19), which demonstrate much weaker H₂ activation on Pt_SA/CeO₂. Consistently, the light‐driven RWGS performance of Pt_SA/CeO₂ is substantially lower than that of Pt_C/CeO₂. These results demonstrate that Pt clusters, rather than isolated Pt atoms, are the preferred active sites for H₂ dissociation and efficient RWGS catalysis. Thus, Pt_C/CeO₂ represents an optimal solution: its abundant Pt‐O‐Ce interfacial sites promote oxygen vacancy formation, while its metallic character ensures efficient H₂ dissociation, two functions that are inherently difficult to optimize simultaneously in single‐atom or nanoparticle architectures. This bifunctional optimization underpins the exceptional light‐driven RWGS performance of Pt_C/CeO₂. It should also be noted that for Pt_N/CeO₂, the light‐driven RWGS performance surpasses its thermal catalytic performance at the same temperatures (Figures S20, S21b), demonstrating the universal enhancement effect of light, which is nevertheless structure‐dependent.

Long‐term stability is crucial for practical applications. As shown in Figure 3f, the Pt_C/CeO₂ catalyst exhibits only a slight decrease in CO production rate during a continuous 122‑h light‑driven RWGS reaction. XRD patterns (Figure S22) and XPS spectra (Figures S23–25) of the spent catalyst show no obvious changes compared to the fresh catalyst, confirming its good structural stability and resistance to deactivation. To comprehensively assess the performance potential of the Pt_C/CeO₂ catalyst, the catalytic performance at various gas hourly space velocities (GHSV) and 2.27 W/cm² light intensity conditions is investigated. From Figures S26–S28, the CO production rate can reach 846.9 mmol g_cat ⁻¹ h⁻¹ with the CO₂ conversion of 10.41% and CO selectivity of 99.31% at a high GHSV (792000 mL g_cat ⁻¹ h⁻¹), corresponding to a high turnover frequency (TOF) of 23574 h⁻¹, demonstrating the excellent intrinsic activity. A performance comparison with reported catalysts (Figure 3g, Tables S6, S7) shows that the CO production rate of the Pt_C/CeO₂ catalyst at approximately 300°C is comparable to that of state‐of‐the‐art thermal catalysts and photothermal catalysts [2, 14, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44], indicating the favorable catalytic performance of the Pt_C/CeO₂ catalyst in the RWGS reaction.

To demonstrate the practical application potential, the outdoor RWGS experiment over Pt_C/CeO₂ catalyst under natural sunlight irradiation is conducted in a continuous‐flow reactor. As shown in Figure 3h–k, the outdoor temperature is approximately −21.0°C, and a Fresnel lens is used as a concentrator. The CO₂ conversion rate can reach 30.81% with the CO selectivity of 98.26%. This suggests that sunlight‐driven RWGS performance in natural environments is comparable to that under Xe lamp irradiation (CO₂ conversion: 33.21%, 2.75 W/cm²), demonstrating the great potential for practical applications.

2.3. Dual‐Path Synergistic Catalytic Mechanism for Efficient RWGS

To elucidate the mechanism of light‐driven RWGS reaction, a series of in situ DRIFTS experiments are conducted. As shown in Figures 4a and S29, under light‐driven condition, key intermediates, including formate species (HCOO*, 1594 and 1377 cm⁻¹) [13, 45], carboxylate species (*COOH, 1621 and 1311 cm⁻¹) [46, 47], bidentate carbonate (b‐CO₃ ^{2 −},1541 and 1280 cm⁻¹) [24, 48] and monodentate carbonate (m‐CO₃ ^{2 −}, 1510, 1491 cm⁻¹) [49, 50], are detected on Pt_C/CeO₂. In contrast, under thermal catalysis condition (300°C), only HCOO* and b‐CO₃ ^{2 −} intermediates are observed, with no detectable *COOH signal. To further distinguish the surface species, temperature‐programmed DRIFTS in CO₂+H₂ atmosphere is performed under thermal conditions (Figures 4b, S30). At low temperature (below 150°C), a peak at 1623 cm⁻¹ appears, which is assignable to bicarbonate (*HCO₃ ⁻). Notably, this peak disappears when the temperature is raised above 150°C, accompanied by the emergence of a m‐CO₃ ^{2 −} peak at 1510 cm⁻¹ and HCOO* peaks at 1591 and 1388 cm⁻¹. This behavior indicates that *HCO₃ ⁻ specie is thermally unstable, consistent with literature reports [51]. Under light‐driven conditions, the catalyst surface temperature reaches as high as 309°C (Figure S31). At this temperature, *HCO₃ ⁻ species should have already decomposed. Therefore, under light‐driven conditions, the peak at 1621 cm⁻¹, accomplished with the peak at ∼1311 cm⁻¹ (which is absent under thermal condition), should be assigned to *COOH species. These distinct intermediates clearly indicate a difference in the reaction pathways between light‐driven and thermal catalytic processes. There is only one single HCOO*‐mediated pathway during thermal catalytic processes. For light‐driven processes, two parallel reaction pathways in which HCOO*‐ and *COOH‐mediated routes proceed simultaneously. To further explore the influence of light on CO₂ adsorption/activation, in situ DRIFTS under CO₂ atmosphere of pre‑reduced Pt_C/CeO₂ are conducted (Figure S32). In the dark, only a weak *HCO₃ ⁻ signal is observed. Under light illumination, the intensities of m‐CO₃ ^{2 −} and b‐CO₃ ^{2 −} bands increase significantly, and the formation of both HCOO* and *COOH intermediates is substantially promoted. These results also show that light irradiation enhances CO₂ adsorption/activation and generates dual intermediates. Notably, Pt_N/CeO₂ also displays the similar thermal‐driven single‐path to light‐driven dual‐path transition (Figure S33). To further prove the existence of a photochemical contribution apart from photothermal heating, control experiments under various irradiation conditions are performed using the same Xe lamp equipped with different optical cut‐off filters and reflection sheets. The UV (200–400 nm), Vis (420–780 nm), and IR (>800 nm) bands are isolated using a UV reflection sheet, a visible long‐pass filter combined with a total reflection sheet, and a Cut800 long‐pass filter combined with a total reflection sheet, respectively. Specifically, the light intensity for each band is adjusted to keep the catalyst surface temperature constant at 201°C. Under these isothermal conditions, UV and visible light both give higher CO production rates than pure thermal catalysis (Figure S34b). This clearly demonstrates that photochemical effects participate in the reaction and are not merely a consequence of heating, which is consistent with the observation of dual pathways in the in situ DRIFTS experiments.

FIGURE 4.

(a) In situ DRIFTS spectra of Pt_C/CeO₂ under different conditions (CO₂+H₂ atmosphere, light‐driven and thermal catalytic process). (b) In situ DRIFTS spectra at different temperatures under the thermal catalytic process. (c) ISI‐EPR spectra (d, e) ISI‐XANES spectra of Pt_C/CeO₂ under dark and light irradiation conditions. (f) Ce 3d, (g) O 1s and (h) Pt 4f XPS spectra of Pt_C/CeO₂ under dark and light irradiation conditions. (i) The RWGS reaction mechanism on the Pt_C/CeO₂ catalyst surface under thermal and light‐driven catalysis conditions.

To clarify the mechanism of the enhanced synergistic dual reaction routes in light‐driven RWGS reaction, in situ irradiated XPS, EPR and XANES analysis based on Pt_C/CeO₂ catalyst are conducted. For in situ irradiated EPR (ISI‐EPR), under irradiation, the characteristic signal at g = 2.003 is strengthened, which provides direct evidence for light‐enhanced O_V (Figure 4c). Furthermore, in situ irradiated XANES (ISI‐XANES) under He atmosphere shows the enhanced white line intensity at the Pt L₃ edge and the decreased white line intensity at the Ce L₃ edge upon illumination (Figure 4d,e). This indicates that Pt loses electrons, consistent with Ce gaining electrons [52]. Even under reactive conditions (CO₂ + H₂ atmosphere), the same trend is observed (Figure S35). In situ irradiated XPS (ISI‐XPS) also show this trend. As shown in Figure 4f–h and Table S8, upon light irradiation, the binding energies of both Ce 3d and O 1s show a significant negative shift, while the Pt 4f binding energy shows a positive shift [53, 54]. This suggests the electrons transfer from Pt species to Ce/O species, thereby increasing the electron density of CeO₂ support. Quantitative analysis shows that the proportion of Ce³⁺ increases from 17.2% (dark) to 18.9% (light). Meanwhile, the O_V concentration also increases significantly from 27.9% (dark) to 34.9% (light). Above results reveal that light illumination can induce the charge redistribution on Pt_C/CeO₂ catalyst surface, where the photogenerated electrons migrate to CeO₂ surface and fill its 4f orbitals. This facilitates the reduction of Ce⁴⁺ and the formation of electron‐rich Ce³⁺‐O_V sites, accompanied by an increase in the valence state of Pt species. This light‐induced charge redistribution is further proved by the in situ irradiated CO‐DRIFTS. As shown in Figure S36, under dark conditions, Pt⁰‐CO signal (2058 cm⁻¹) is predominant, accompanied with relative lower peak of Pt^δ+‐CO (2082 cm⁻¹). Upon light irradiation, the intensity of Pt⁰‐CO signal peak decreases, while that of Pt^δ+‐CO signal peak increases. This spectral change directly visualizes the light‐induced increased valence of Pt species. Most importantly, this valence change is reversible: the Pt⁰‐CO and Pt^δ+‐CO bands are restored upon turning off the light (Dark' line in Figure S36). This reversible valence state change confirms the process to be light‐driven dynamic process rather than a permanent chemical transformation. The consistency across XPS, EPR, and XANES, all showing the same direction of change, provides robust evidence for light induced charge redistribution of Pt and CeO₂, leading to Ce³⁺/O_V formation and Pt^δ+ generation. Notably, the observed electron transfer from Pt to CeO₂, which differs from the conventional direction in Pt/CeO₂ systems, may be attributed to the specific conditions in our light‐driven photothermal system, where photoexcitation is accompanied by phonon‑assisted lowering of the Ce(4f) levels, which creates a driving force for electron injection from Pt to the support [55].

Based on the combined experimental evidence, we propose a catalytic mechanism for the light‑driven RWGS reaction on the Pt_C/CeO₂ catalyst (Figure 4g). Under light irradiation, interfacial charge redistribution leads to the formation of a nonequilibrium Pt^δ+‐O_V‐Ce³⁺ structure, which promotes CO₂ activation and concurrently facilitates both the *COOH and HCOO* pathways. As a result, the two reaction routes operate in parallel, giving a cooperative boost to the overall reaction activity. Under conventional thermal conditions, the reaction proceeds predominantly via the formate pathway, limited by the high barrier of formate decomposition [46]. Light excitation overcomes this single‑pathway constraint through efficient interfacial electron redistribution (evidenced by in situ irradiated XPS, EPR, and XANES), generating electron‑rich Ce³⁺‐O_V sites and electron‑deficient Pt^δ+ sites. The resulting Pt^δ+‐O_V‐Ce³⁺ site provides a surface that supports both pathways in a cooperative manner. While establishing a direct causal link between the observed electron transfer and the dual‑pathway activation remains challenging, the existence of a photochemical contribution is clearly demonstrated by wavelength‑dependent control experiments and reduced E _a during the light‐driven process (Figures S34, 3d). Thus, the light‑induced formation of *COOH and HCOO* intermediates, together with the photochemical evidence, supports the proposed dual‑pathway mechanism. By dynamically programming interfacial charge distribution, light transforms a single‐route thermal catalyst into a dual‑pathway photothermal platform, offering a new paradigm for energy‑efficient solar‑driven CO₂ conversion.

3. Conclusion

In summary, we have demonstrated that light‐driven dynamic charge redistribution in a Pt cluster/CeO₂ catalyst unlocks synergistic dual pathways for highly efficient low‐temperature RWGS. The catalyst, rationally designed via an amine‐assisted anchoring strategy, features highly dispersed Pt clusters with abundant Pt‐O‐Ce interfacial sites. This architecture simultaneously optimizes the O_V formation at Pt‐O‐Ce interfaces and efficient hydrogen dissociation on metallic Pt clusters, which is unattainable in individual single‐atom or particle catalysts. Under photoexcitation, light‐induced charge redistribution generates electron‐rich Ce³⁺‐O_V sites while modulating Pt electronic states, reconstructing the interface into active Pt^δ+‐O_V‐Ce³⁺ sites. This nonequilibrium structure enables concurrent operation of both HCOO* and COOH* pathways, which is a fundamental mechanistic shift from the single‐path limitation of conventional thermal catalysis. Using light as the sole energy source, the catalyst achieves localized heating to ∼309°C while the bulk reactor remains ∼54°C, dramatically reducing energy consumption and improving operational safety. A high CO production rate of 846.9 mmol g_cat ⁻¹ h⁻¹ is achieved in the continuous‐flow system with cold inlet gases without preheating, substantially outperforming conventional Pt nanoparticle catalysts under identical conditions. Notably, the practical applicability of this light‐driven system extends well beyond laboratory conditions. Under natural sunlight irradiation, the catalyst maintains comparable performance even at ambient temperatures as low as −21°C, demonstrating its adaptability against changing environmental conditions. This work establishes light as a powerful tool for engineering catalytic architectures, paving the way toward solar‐driven chemical processes for a sustainable energy future.

Author Contributions

Guiyu Huang: data curation, formal analysis, investigation, methodology, validation, writing – original draft, writing – review and editing. Panzhe Qiao: data curation, formal analysis, methodology. Li Fang: formal analysis, writing – review and editing. Congsen Liu: formal analysis, writing – review and editing. Ying Xie: methodology. Aiping Wu: methodology. Dongxu Wang: formal analysis, funding acquisition, methodology, supervision, writing – review and editing. Jiancong Liu: funding acquisition, writing – review and editing, formal analysis, methodology, supervision. Chungui Tian: methodology, funding acquisition. Honggang Fu: conceptualization, formal analysis, funding acquisition, methodology, supervision, writing – review and editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: anie73551‐sup‐0001‐SuppMat.docx.

Data Availability Statement

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