Continuous-Flow Plasma Synthesis of Single-Atom Au₁/CeO₂–Fe for Wide-Temperature-Window Water–Gas Shift Reaction

Abstract

The water–gas shift (WGS) reaction is pivotal for H₂ production and purification, yet conventional processes rely on coupled middle and high temperature stages (180–300 °C). Development of wide-temperature-window catalysts could avoid temperature switching and reduce the energy demand. However, room-temperature WGS is particularly limited by sluggish water splitting. Here, we report a continuous-flow solution plasma (CSP) strategy to construct an oxygen vacancy-rich Au₁/CeO₂–Fe single-atom catalyst (SAC). Fe is uniformly doped into the CeO₂ lattice to increase the density of oxygen vacancies and Au is stabilized as isolated atomic sites on the CeO₂ surface. By photothermal excitation, the catalyst delivers high WGS activity across 25–300 °C. Notably, at 25 °C, the CO conversion rises from near-zero for conventional catalysts to 35%, overcoming the kinetic barrier of room-temperature WGS for a low CO conversion rate. Mechanistic studies reveal that Fe doping selectively promotes H₂O dissociation, and Au SACs enhance CO adsorption and oxidation. Meanwhile, light activates CeO₂ lattice oxygen and engages a Mars-van Krevelen cycle. The as-developed catalyst maintains significant activity from ambient temperature to 300 °C without temperature switching, providing a process toward energy-efficient H₂ production and purification.

1. Introduction

Driven by the global energy transition and carbon neutrality goals, H₂ energy has emerged as a key alternative to traditional fossil fuels due to its efficiency and cleanliness. , However, impurities such as CO present in processes like CH₄ reforming and biomass conversion consistently reduce the purity of H₂. , The WGS serves as a vital link between carbon compound conversion and H₂ synthesis, as it can oxidize CO and produce H₂ (CO + H₂O → CO₂ + H₂, ΔH = −41.1 kJ·mol^–1) through the activation of catalysts across a broad temperature range (180–300 °C). Although significant advancements have been made in WGS research at middle to high temperatures, studies of the WGS reaction at room-temperature still face substantial challenges. This is primarily because traditional metal–support catalysts often feature a single type of active site, leading to competitive adsorption of CO and H₂O at these sites, which prevents sufficient surface active sites from being available to drive the dissociation of H₂O molecules at room-temperature. Additionally, the weak physical adsorption between some transition metals and CO complicates the formation of the adsorbed state of *CO, ultimately resulting in a very low WGS activity at room-temperature. Recent studies have shown that light-assisted methods can effectively activate the lattice oxygen of support materials at room-temperature, thereby inducing the generation of more active sites on the surface. Nevertheless, current light-assisted room-temperature WGS reactions can only achieve approximately 15% CO conversion over single metal–supports catalysts, highlighting the ongoing challenges in this area. Therefore, the development of metal–support catalysts with dual or even multiple active surface sites that are light-assisted is a critical strategy for achieving effective room-temperature WGS reactions.

The continuous-flow chemistry synthesis is well suited for the solution flow process, as it automatically delivers the reaction solution to the reactor, thereby eliminating the repeated feed procedures associated with traditional batch reactors. Meanwhile, the solution plasma process features plasma generated in solution. By coupling continuous-flow solution plasma (CSP), collisions between high-energy electrons in the plasma and reactant molecules improve reaction conversion rates via a clean manner. Compared to photochemical reactions and physical deposition methods, this integrated approach offers significant advantages in the dispersion of metal active sites, minimizing the excessive waste of metal precursors. , In the CSP system, the high-energy environment induced by the plasma facilitates the rapid reduction of various metal ions and their subsequent anchoring to metal oxide supports. This is particularly beneficial in the synthesis of SACs, where the goal is to maximize the number of active sites while minimizing the waste of metal precursors. For instance, we have demonstrated that the CSP method can effectively synthesize SACs in terms of Au₁ or Au₁Rh₁. , Therefore, the CSP method offers a promising pathway for the development of SACs catalysts with high uniformity and multiple active surface sites.

The development of atomic-level dispersed metal catalysts has emerged as a crucial strategy for improving WGS performance. Research by Yao et al. has confirmed that the unique electronic structure of atomically dispersed Au enables it to form strong interactions with CO molecules, resulting in excellent CO adsorption and activation capabilities at low temperatures. Furthermore, the efficiency of the WGS reaction heavily relies on the dissociation of H₂O, which typically occurs at vacancy and defect sites on the catalyst surface. Further theoretical calculations by Wang et al. have validated that H₂O dissociates more readily at surface oxygen vacancies (Ov), leading to the formation of *OH and *O species that can be effectively adsorbed on single-atom Pt and adjacent oxygen atoms. Additionally, research by Li et al. has demonstrated that the initial concentration of Ov in oxide supports plays a critical role in modulating the strong metal–support interaction (SMSI) between Ru and TiO_2–x. Based on these findings, selecting support materials with a high concentration of Ov for SACs can enhance the SMSI mechanism, potentially enabling significant room-temperature WGS reactions. CeO₂ is known for its inherent oxygen vacancies, which can provide lattice oxygen for the WGS reaction and facilitate its progression. However, the Ov content in pure CeO₂ needs to be precisely tuned for controlling the dispersion of SACs on its surface, restricting further improvements in catalytic performance. Therefore, tuning the Ov concentration in CeO₂ and achieving SACs dispersion on its surface could be an effective strategy for enhancing the room-temperature activity of the WGS reaction.

In this work, we employed CSP technology to synthesize an Ov-rich Au₁/CeO₂–Fe single-atom catalyst. Characterization results confirmed the successful anchoring of Au₁ onto the CeO₂–Fe support. The Fe doping increased the Ov content in pure CeO₂. This dual-site synergy between the surface of Ov and SAC in Au₁/CeO₂–Fe not only enhances the dissociation process of H₂O but also promotes the adsorption of CO on the catalyst surface. Consequently, the performance of Au₁/CeO₂–Fe surpassed that of pure CeO₂, achieving a remarkable CO conversion rate at room-temperature, increasing from 0 to 35%. Furthermore, at 300 °C, the catalyst demonstrated a hydrogen production rate of 11.334 mmol·g^–1·h^–1, which is 55.0 times higher than that of pure CeO₂ and 1.5 times that of Au₁/CeO₂. Finally, in situ Fourier transform infrared spectroscopy revealed the evolution patterns of *COO and *COOH intermediates, providing insights into the reaction mechanism of Au₁/CeO₂–Fe during the room-temperature WGS reaction. This highlights the potential of this novel catalyst design in enhancing the efficiency of low-temperature catalytic processes.

2. Experimental Section

2.1. Catalyst Synthesis

As shown in the Supporting Information, Figure S1, continuous-flow solution plasma (CSP) provides an efficient and cost-effective synthetic method. The reaction system utilizes multichannel peristaltic pump equipment for precise transmission of the CeO₂ and HAuCl₄ precursor solutions, a tungsten electrode pulse power supply to produce plasma, and a glassware to collect solution. This design confines the plasma discharge to a channel just 2 mm wide, ensuring uniform exposure of the CeO₂ and HAuCl₄ precursor solutions to the high-density plasma. When synthesizing single-atom catalysts, 2.5 g·L^–1 CeO₂ nanosheets or a CeO₂-M (M = Fe, Co, and Cu) suspension solution was first combined with 0.20 mM of HAuCl₄ and 0.07 M of NaOH solution. The mixed solution of HAuCl₄ precursor and NaOH solutions was introduced into the dispersed CeO₂ or CeO₂-M suspension solution at a flow rate of 50 mL·min^–1. A hollow tungsten electrode was used to direct the combined solutions through a plasma discharge treatment at a uniform flow rate of 100 mL·min^–1. The plasma for the reaction was generated using a bipolar direct current pulsed power source that operated at a voltage of ±2.5 kV, with a frequency of 20 kHz and a pulse width of 2.0 μs. The resultant samples with the loaded single metal atoms (S₁) were immediately collected into sample vials. These samples were then subjected to a series of processing steps, including centrifugation and lyophilization, following which they were marked as S₁/CeO₂ and S₁/CeO₂-M (S = Au and M = Fe, Co, and Cu).

2.2. WGS Evaluations

The detailed WGS evaluation is illustrated in the Supporting Information, Figure S2. The CO conversion rate and H₂ production rate are calculated using eqs and .

$$\mathrm{CO}\%=\frac{{[\mathrm{CO}]}_{\mathrm{in}}-{[\mathrm{CO}]}_{\mathrm{out}}}{{[\mathrm{CO}]}_{\mathrm{in}}}\times 100\%$$ 1

$${\mathrm{H}}_2\mathrm{production}\mathrm{rate}=\frac{{\mathrm{H}}_2(\mathrm{mol}/\mathrm{L})}{g_{(\mathrm{catalysts})}}\times \mathrm{GHSV}(\mathrm{L}/\mathrm{h})$$ 2

[CO]_in and [CO]_out are denoted as the input CO peak area and the output CO peak area, respectively; H₂ (mol/L) is the concentration of hydrogen produced; g_(catalysts) represents the weight of the catalyst necessary for the reaction; and GHSV is gas hourly space velocity.

3. Results and Discussion

3.1. Modified CeO₂ by Au₁ Loading and Fe Doping

The formation of Au₁/CeO₂–Fe is schemed in Figure a. During the solution plasma discharge process, the generated high-energy electrons collide with the HAuCl₄ precursor, triggering the cleavage and dissociation of Au–Cl chemical bonds while rapidly reducing Au³⁺ to atomic Au. The Fe-doped CeO₂ support, which is rich in oxygen vacancies (Ov), enables strong interactions between the reduced Au atoms and Ov, anchoring Au atoms on the support surface and ultimately forming Au₁/CeO₂–Fe. High-resolution transmission electron microscopy (HRTEM) results (Supporting Information, Figure S3a) indicate that the synthesized CeO₂ exhibits a typical porous nanosheet structure. This architecture provides a substantial specific surface area, which not only facilitates the complete exposure of catalytically active sites but also promotes high dispersion of SACs. From the HRTEM image of CeO₂–Fe (Supporting Information, Figure S3b), a crystal plane with a lattice spacing of 0.311 nm can be clearly observed, corresponding to the (1 1 1) crystal plane of CeO₂. To elucidate the doping state and distribution uniformity of Fe atoms, high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) was performed on CeO₂–Fe (Figure b,c). The results reveal that Fe atoms have been successfully embedded into the CeO₂ support lattice while maintaining the CeO₂ crystal structure. Figure d presents the fast Fourier transform pattern of CeO₂–Fe, highlighting the (3 3 1) and (2 2 2) crystal planes of CeO₂. Elemental mapping results (Figure e–h) further confirm the uniform distribution of Fe atoms within the CeO₂ support without Fe agglomeration detected. The experimental results were further validated by the quantitative analysis of the actual Fe content using Inductively Coupled Plasma Mass Spectrometry (ICP-MS) (Table S1), which showed an actual Fe content of 0.9756 wt %. For the Au₁/CeO₂, HAADF-STEM images (Supporting Information, Figure S4a–c) also reveal that Au atoms are atomically distributed on the CeO₂ support surface. This visually demonstrates the effectiveness of the continuous-flow plasma (CSP) synthesis method for preparing SACs. The corresponding intensity spectra of Ce and single-atom Au (Supporting Information, Figure S4d) further corroborate the presence of Au in a single-atom state on the CeO₂ surface. Combined with the corresponding elemental mapping (Supporting Information, Figure S4e–g), this finding verifies the uniform dispersion of Au across the CeO₂ support surface. For the Au₁/CeO₂–Fe sample, the HAADF-STEM image (Figure i) similarly reveals isolated Au atoms uniformly distributed on the CeO₂–Fe support. Local elemental intensity line analysis of the lattice fringe of Au₁/CeO₂–Fe (Figure j,k) shows that the intensity distribution along the blue line exhibits pronounced fluctuations in the broadening of the lattice stripes at Ce sites. Furthermore, due to the higher atomic number of Au compared to Ce, resulting in an increased lattice spacing between Ce and Au to 0.35 nm. The strongest peak corresponds to Au signals that directly overlap with the original Ce signals, indicating that Ce sites on the support surface are directly replaced by adsorbed single-atom Au, forming Au–O–Ce interface interactions. As shown in Figure l-o, the elemental mapping image confirms the uniform distribution of Au on the CeO₂–Fe support, without evidence of strong interfacial adsorption or atomic clustering. Additionally, neither Fe atom agglomeration nor nanoparticle formation induced by plasma bombardment was detected, indicating that the CSP synthesis method preserves both the crystalline structure of the support material and the uniform doping state of Fe atoms. The crystal structure of the catalyst was further confirmed by X-ray diffraction (XRD) characterization. As demonstrated in Supporting Information (Figure S5a), the diffraction peaks of the both CeO₂ and CeO₂–Fe catalysts exhibit a high degree of agreement with the standard fluorite structure (PDF#34-0394), with no significant impurity phases detected, which suggested that Fe successfully forms a unique solid solution structure within CeO₂. These results are consistent with the HAADF-STEM results, confirming the uniform incorporation of Fe atoms into the CeO₂ crystal lattice and precluding the formation of discrete FeO_x phases. As evidenced by Supporting Information (Figure S5b), even when Fe was substituted with other transition metals (Co and Cu) or the Fe doping concentration was adjusted, the XRD patterns of the samples remained highly consistent with that of pure CeO₂, with the absence of detected impurity phases. This observation confirms the excellent structural stability of this doping system. In addition, the XRD patterns of Au₁/CeO₂ and Au₁/CeO₂–Fe SACs synthesized via the CSP method also correspond to the crystal structure of pure CeO₂, with no diffraction peaks observed for Au nanoparticles, Au clusters, or oxides such as Au₂O₃ and Fe₂O₃. This not only corroborates that the CSP synthesis method facilitates the homogeneous distribution of Au on the CeO₂ support surface but also emphasizes the superiority and feasibility of this approach for the preparation of single-atom catalysts.

1.

(a) Schematic diagram of Au₁/CeO₂–Fe SACs via plasma discharge. (b) HAADF-STEM image of CeO₂–Fe. (c) Local high-resolution HAADF-STEM image of CeO₂–Fe. (d) Fast Fourier transform pattern of CeO₂–Fe. (e–h) Element mapping image of Fe, Ce, and O for CeO₂–Fe. (i) HAADF-STEM image of Au₁/CeO₂–Fe. (j) Local high-resolution HAADF-STEM image of Au₁/CeO₂–Fe. (k) The corresponding line profiles from high-resolution HAADF-STEM image intensity analysis between Ce and single-atom Au. (l–o) Element mapping image of Au, Ce, Fe, and O for Au₁/CeO₂–Fe.

To systematically elucidate the electronic structures and coordination environments of Au, Fe, and Ce in the Au₁/CeO₂–Fe catalysts, X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) characterizations were performed with Au, Fe, and Ce foils as well as their corresponding metal oxides used as reference materials. Referring to the visuals in Figure a, the Au L₃-edge XANES spectrum shows that the absorption edge position of Au in Au₁/CeO₂–Fe lies between that of Au foil and Au₂O₃, indicating Au carriers partial positive charge. A thorough analysis of white line intensity in the magnified view of Figure a reveals that Au foil exhibits no distinct white line peak due to its fully occupied 5d orbitals. In contrast, the Au₁/CeO₂–Fe exhibits a prominent white line peak, which is attributed to the 2p3/2 to 5d electronic transition, suggesting the presence of unoccupied 5d orbitals. The EXAFS spectrum of Au₁/CeO₂–Fe (Figure b) displays a characteristic peak at 1.6 Å, corresponding to the specific Au–O bonds formed between the first coordination sphere of Au atoms and the surrounding atoms. This verifies that Au exists in a unique oxidized state structure in the catalysts, rather than as metallic clusters or nanoparticles. The distinctive Au–O coordination structure can enhance metal–support interactions and improve CO adsorption capacity. According to the EXAFS fitting and results (Supporting Information, Figure S6 and Table S2), the coordination number of the Au–O bond is 2.82. The absence of characteristic Au–Au interaction peaks in the WX EXAFS spectrum (Figures c and S7), which further suggests that Au atoms are atomically dispersed on the CeO₂–Fe surface, and unobserved metallic agglomeration or residual Cl^–. Regarding the electronic structure and coordination environment of Fe, the Fe k-edge XANES spectrum (Supporting Information, Figure S8a) demonstrates a negative shift in the absorption edge position of Fe in both CeO₂–Fe and Au₁/CeO₂–Fe compared with Fe₂O₃, thus indicating charge transfer at the Au₁/CeO₂–Fe interface. The Fourier-transformed k ³-weighted EXAFS spectra (Supporting Information, Figures S8b and S9a–d) reveal characteristic peaks corresponding to Fe–O (1.45 Å) and Fe–Fe (2.45 Å) bonds in the R-space for both samples. The WT EXAFS spectra of Au₁/CeO₂–Fe (Supporting Information, Figures S8c–f) similarly detects a characteristic Fe–O bond signal at 5.0 Å^–1, thus confirming the successful doping of Fe atom into the CeO₂ lattice. The Ce L₃-edge XAFNS spectra (Supporting Information, Figure S10a) show a significant blue shift in the absorption edge positions for both CeO₂–Fe and Au₁/CeO₂–Fe. Their k ³-weighted EXAFS spectra (Supporting Information, Figure S10b) show a characteristic Ce–O peak at 1.95 Å, which is typical of pure CeO₂. After Fe doping, the Ce–O peak in CeO₂–Fe and Au₁/CeO₂–Fe shifts to 1.87 Å, which is attributed to Fe atoms replacing Ce sites, causing the contraction of the Ce–O bond. Compared with pure CeO₂ (Table S2), due to the tendency of Fe atoms, upon incorporation into the CeO₂ lattice to form coordination structures with surrounding oxygen atoms, the coordination environment of the Ce atoms changes after Fe introduction. It has been shown to induce lattice stretching, thereby favoring the formation of Ov. The WX EXAFS and k ³-weighted EXAFS measurements of the Ce L₃-edge (Supporting Information, Figures S10c–f and S11) further confirm Fe incorporation into the Ce–O lattice.

2.

Synchrotron radiation analysis of Au₁/CeO₂–Fe SACs. (a) XANES spectra at the normalized Au L₃-edge, (b) Fourier transforms of the k ³-weighted EXAFS spectra at Au L₃-edge in R space, and (c) Wavelet transforms EXAFS at Au L₃-edge. (d) XPS patterns of Au 4f for Au₁/CeO₂ and Au₁/CeO₂–Fe. (e) XPS patterns of Ce 3d for CeO₂, Au₁/CeO₂, CeO₂–Fe, and Au₁/CeO₂–Fe catalysts. (f) Raman spectra for four catalysts and local enlarged spectra of F_2g characteristic peak. (g) XPS patterns of O 1s for CeO₂, Au₁/CeO₂, CeO₂–Fe, and Au₁/CeO₂–Fe. (h) O₂-TPD patterns CeO₂, Au₁/CeO₂, CeO₂–Fe and Au₁/CeO₂–Fe. (i) UV–vis absorption spectra for CeO₂, Au₁/CeO₂, CeO₂–Fe, and Au₁/CeO₂–Fe.

After the analysis of the HAADF-STEM and XANES results, a systematic investigation was conducted on the elemental valence states, the Ov content trends, and the optical properties of the CeO₂-based catalysts. First, in the Au 4f X-ray photoelectron spectroscopy (XPS) patterns (Figure d), the binding energy positions of Au 4f7/2 (86.90 and 85.67) and 4f5/2 (90.90 and 89.37) in Au₁/CeO₂ and Au₁/CeO₂–Fe correspond to Au³⁺, while those for Au⁺ are 4f7/2 (85.06 and 84.61) and 4f5/2 (88.76 and 88.31), respectively. Furthermore, Au was found to be predominantly present in the +1 valence state in both catalysts, which is consistent with the calculated actual valence state of Au (0.92) in Au₁/CeO₂–Fe (Supporting Information, Figure S12). Subsequently, the formation of Ov was verified from the perspective of elemental valence states, with XPS revealing the presence of the Ov content on the catalyst surface. The Ce 3d orbitals can be fitted into multiple characteristic peaks corresponding to Ce⁴⁺ and Ce³⁺ signals from the Ce 3d patterns (Figure e). We find that the Ce³⁺ content on the CeO₂ surface was enhanced after Fe doping and single-atom Au loading, which is attributed to the altered electronic distribution on the support surface induced by Fe and Au, promoting the reduction of Ce⁴⁺ to Ce³⁺. Electron spin resonance (ESR) spectra (Supporting Information, Figure S13) further confirm that the enhanced Ce³⁺ signal in Au₁/CeO₂ is enhanced after single-atom Au loading, thereby indicating that the introduction of single-atom Au effectively increases the Ce³⁺ content within the CeO₂ lattice. Notably, the generation of substantial amounts of Ce³⁺ is usually accompanied by an increase in Ov content, which facilitates the effective dissociation of H₂O and enhances the catalytic efficacy. As shown in the Supporting Information (Figure S14), there is a clear visual indication of an increase in the Ce³⁺ content from 13.7% in pure CeO₂ to 19.5% in Au₁/CeO₂–Fe, which corresponds with the rise in the Ov content. To further investigate the crystal structure and the Ov content of CeO₂-based materials, Raman spectroscopy was employed to characterize CeO₂, CeO₂–Fe, Au₁/CeO₂, and Au₁/CeO₂–Fe, validating the effects of Fe doping and single-atom Au loading on CeO₂ lattice vibrations and the Ov content. From the data presented in Figure f, all four samples exhibited the F_2g vibration peak of CeO₂ at 459 cm^–1, corresponding to the symmetric stretching vibration of the Ce–O bonds. It is noteworthy that the F_2g peak of CeO₂–Fe shows an obvious blue shift in comparison to pure CeO₂, accompanied by a shortening of the Ce–O bond length and a decline in bond energy. For Au₁/CeO₂–Fe, the F_2g characteristic peak exhibits a more pronounced blue shift, indicating that adjacent single-atom Au exists near the Ce–O bonds. These Au atoms can modulate the bonding structure of Ce–O bonds, thereby facilitating the formation of strong metal–support interactions. , The characteristic peak at 598 cm^–1 is typically associated with the Ov signal in CeO₂. It was observed that replacing Ce⁴⁺ with Fe³⁺ caused lattice contraction, leading to a significant increase in the content of the ov of CeO₂. Furthermore, the Au₁/CeO₂–Fe sample showed the strongest characteristic Ov peak, thereby confirming that the synergistic effect of Fe doping and single-atom Au loading enhances the formation of Ov. Changes in the metal–oxygen bond lengths were similarly observed in the O 1s patterns (Figure g). After Fe doping and single-atom Au loading, the lattice oxygen characteristic peaks in CeO₂–Fe and Au₁/CeO₂–Fe shifted. This movement of the peaks can be attributed to the introduction of Fe atoms, which weakens the metal–oxygen bond lengths, causing the characteristic peaks to shift toward higher binding energies. Meanwhile, the O 1s patterns disclose a substantial increase in oxygen species content on the Au₁/CeO₂–Fe surface, which is due to the ability of Au to induce the generation of large quantities of reactive oxygen species on the support surface through interfacial interactions. The oxygen-temperature-programmed desorption (O₂-TPD) experiments were employed to further evaluate the activation of surface oxygen species and lattice oxygen on the catalysts. As demonstrated in Figure h, we observe that the lattice oxygen desorption peak temperatures of Au₁/CeO₂ (707 °C) and Au₁/CeO₂–Fe (605 °C) were distinctly lower than those on pure CeO₂ and CeO₂–Fe. The loading of single-atom Au on the catalyst surface reduces the desorption barrier of oxygen species and enhances the mobility of surface oxygen. These highly mobile oxygen species facilitate rapid participation in the WGS reaction and enable efficient oxygen atoms transfer, ensuring that the lattice oxygen in the catalyst can undergo oxidation–reduction processes within the Mars-van Krevelen (MvK) mechanism, thereby guaranteeing the activation of lattice oxygen and the sustained progression of the catalytic reaction.

Ultraviolet–visible diffuse reflectance spectroscopy (UV–vis DRS) was employed to analyze the effects of Fe doping and single-atom Au loading on the optical absorption properties and bandgap values (Eg) of CeO₂-based catalysts. Analyzing the findings portrayed in Figure i, the UV-vis absorption spectra of CeO₂–Fe exhibit a substantially augmented light absorption intensity in the visible region (390–800 nm) in comparison with pure CeO₂, which is attributed to the introduction of defect energy levels within the CeO₂ bandgap by Fe doping. The change in energy levels broadened the absorption wavelength range, resulting in an increase in the visible light absorption capacity of CeO₂–Fe. Upon loading single-atom Au onto the CeO₂–Fe support, the absorption edge of Au₁/CeO₂–Fe exhibited a slight blue shift relative to CeO₂–Fe, indicating strong electronic interactions between the single-atom Au and the CeO₂–Fe support. The Eg values for pure CeO₂, CeO₂–Fe, Au₁/CeO₂, and Au₁/CeO₂–Fe were calculated using the Tauc formula, with the results illustrated in Supporting Information (Figure S15). Both Fe doping and single-atom Au loading were found to reduce the Eg of CeO₂, which lowers the excitation energy barrier for photogenerated carriers and enhances the catalytic activity in photothermal synergistic reactions.

3.2. Monitoring Ov Variation in Au₁/CeO₂–Fe during WGS Reaction

In situ Raman spectroscopy was employed to investigate the variation law of the Ov content with temperature. Prior to the experiment, the reaction chamber was purged with N₂ in order to eliminate any residual gases, thus avoiding any potential interference with the Raman signal. As shown in Figure a, the Raman characteristic peaks of pure CeO₂ under a N₂ atmosphere are the typical fluorite structure, while no additional signal peaks were detected. In the instance of the introduction of CO into the system at room-temperature, discernible alterations were identified in the characteristic peaks of pure CeO₂. However, as the ambient temperature increases from 25 to 300 °C, the F_2g characteristic peak shifted from 461.1 to 457.4 cm^–1, which is attributed to the lattice structure variation of CeO₂ induced by temperature elevation. The changes in lattice structure lead to the modification of the Ce–O bond length, thereby reducing the lattice vibration frequency. We also observe that the F_2g characteristic peak of Au₁/CeO₂ exhibited a similar temperature-dependent variation trend to that of pure CeO₂ (Figure b), shifting to 454.5 cm^–1 at 300 °C. As presented in Figure c, the Fe doping process into the CeO₂ lattice induces a lattice contraction, consequently leading to the F_2g characteristic peak of CeO₂–Fe shift to 460.6 cm^–1 at room-temperature. Notably, this characteristic peak shifts further at 100 °C, and when the temperature rises to 300 °C, the F_2g peak moves from 460.6 to 454.1 cm^–1, indicating a significant enhancement in the lattice temperature responsiveness. Noting the insights conveyed in Figure d, the F_2g characteristic peak of Au₁/CeO₂–Fe shows a slight blue shift at room-temperature (25 °C), and the shift degree becomes more pronounced as the temperature increases to 300 °C. In summary, the vibration of catalysts lattice can be induced at room-temperature, and its lattice oxygen possesses activation potential under room-temperature conditions, which is crucial for the realization of room-temperature WGS reaction.

3.

In situ Raman spectra of WGS (CO+H₂O) and the local enlarged spectra of F_2g characteristic peak for (a) CeO₂, (b) Au₁/CeO₂, (c) CeO₂–Fe, and (d) Au₁/CeO₂–Fe under the 25–300 °C temperature range.

The trends of the variation of the Ov exhibited marked differences among the four catalysts. As the reactor temperature was elevated from 25 to 300 °C, due to the reductive effect of CO, the Ov intensity of pure CeO₂ at 594 cm^–1 gradually increased, which consumes oxygen atoms in the CeO₂ lattice and consequently generates a large number of Ov. Moreover, a higher reaction temperature is found to accelerate the reduction process, indicating a positive correlation between the content of the Ov of CeO₂ and the reaction temperature. Following the loading of single-atom Au, the Ov content of Au₁/CeO₂ is found to be elevated at room-temperature (25 °C), reaching levels that are even higher than those observed for pure CeO₂ at 300 °C. The enhancement of content is attributed to the electron transfer induced by the SMSI between single-atom Au and CeO₂, which promotes the reduction of Ce⁴⁺ to Ce³⁺ and the concurrent increase in Ov content. Regarding the effect of Fe doping, it is evident that the Ov content of CeO₂–Fe increases obviously at 100 °C, and its Ov content at 300 °C is also higher than that of pure CeO₂. In the case of Au₁/CeO₂–Fe, the synergistic effect of Fe doping and single-atom Au loading resulted in a significant increase in the Ov content after 100 °C, which is a comparison to the content observed in pure CeO₂, Au₁/CeO₂, and CeO₂–Fe. Therefore, the elevated Ov content of Au₁/CeO₂–Fe indicates its superior CO reduction activity at high-temperatures, further substantiating the synergistic promotion effect of Fe doping and single-atom Au loading on the Ov formation and catalytic performance.

3.3. WGS Testing

A continuous-flow fixed-bed reactor was employed to test the CO conversion and H₂ production rate in the WGS reaction, investigating the effects of transition metal doping and single-atom Au loading on the catalytic performance. First, Supporting Information (Figure S16a,b) compares the WGS performance of a series of transition metal-doped CeO₂ catalysts under photothermal synergistic conditions. The results demonstrate that CeO₂–Fe, CeO₂–Cu, and CeO₂–Co exhibit higher CO conversion than pure CeO₂ at 300 °C, accompanied by a substantial augmentation in H₂ production rate increasing from 0.206 to 0.392, 0.414, and 0.366 mmol·g^–1·h^–1, respectively, thus confirming the promotional effect of transition metal doping on the WGS reaction. However, no discernible enhancement in the room-temperature WGS performance is observed. After loading the single-atom Au on the surface of these catalysts (Supporting Information, Figure S16c,d), the WGS performance of all SACs is further enhanced, except for Au₁/CeO₂–Cu. Under identical experimental conditions (Supporting Information, Figure S17), a 5% doping concentration was found to produce superior CO conversion and hydrogen production rates compared to 1% and 10% doping concentrations. Among the catalysts examined, Au₁/CeO₂–Fe even exhibits optimal CO conversion at room-temperature. In consideration of the experimental results, four catalysts (CeO₂, CeO₂–Fe, Au₁/CeO₂, and Au₁/CeO₂–Fe) are selected for the purpose of plotting the curves of the CO concentration versus time over a wide-temperature range under photothermal synergistic catalysis. As shown in Figure a, the four catalysts display different CO conversion rates following 1 h of the isothermal reaction within the low-temperature range. Particularly, at room-temperature (25 °C), Au₁/CeO₂–Fe obtains a CO conversion rate of 35%, which is higher than that of the other three catalysts. In the middle-temperature range (Figure b), the CO conversion rate of CeO₂–Fe is not improved compared with CeO₂ and even exhibits the worst performance. This can be attributed to the excessive oxygen vacancies adsorbing a greater number of H₂O molecules, resulting in H₂O adsorption saturation during the reaction. Currently, the saturated H₂O molecules occupy a large number of active sites on the catalyst surface, thereby hindering the adsorption of CO molecules and forming an obvious competitive adsorption relationship between CO and H₂O on dual active sites, which exerts an adverse effect on the WGS catalytic activity. In contrast, Au₁/CeO₂–Fe maintains optimal WGS performance, a phenomenon that may be due to the charge transfer between the single-atom Au and the support, which facilitates enhanced CO molecule adsorption, consequently weakening the competitive dynamic at surface active sites. In the high-temperature range (Figure c), Au₁/CeO₂–Fe also exhibited considerable WGS performance, indicating that the dual active sites (Ov and single-atom Au) on the catalyst surface are pivotal in realizing the WGS reaction over a wide-temperature range. The dual active sites also promoted the migration of oxygen species on the CeO₂ surface and the effective progress of the Mars-van Krevelen (MvK) mechanism, thereby accelerating the CO oxidation process. Meanwhile, the abundant oxygen vacancies in Au₁/CeO₂–Fe provided sufficient adsorption sites for the H₂O dissociation process, facilitating the adsorption of the formed *OH and *H intermediates on the catalyst surface and their participation in the catalytic reaction. Regarding H₂ production, as illustrated in Supporting Information (Figure S18), the H₂ production rate of Au₁/CeO₂–Fe attained 11.334 mmol·g^–1·h^–1, which is 55.0 times, 28.9 times, and 1.54 times that of pure CeO₂ (7.362), CeO₂–Fe (0.395), and Au₁/CeO₂ (0.206), respectively. The high hydrogen production rate not only verifies the positive effect of Fe doping and single-atom Au loading on improving catalyst activity but also indicates that this catalytic system can efficiently promote the WGS reaction under photothermal synergistic catalysis, achieving high efficiency CO conversion at room-temperature and high H₂ yield at high-temperature. To discuss the synergistic effect between light and heat in the WGS reaction, control experiments are performed on Au₁/CeO₂–Fe. Figure d illustrates the effect of light and dark conditions on the performance of the WGS reaction over Au₁/CeO₂–Fe at room-temperature (25 °C). The results show that under dark conditions, Au₁/CeO₂–Fe exhibits a CO conversion rate of 0% after 1 h of the reaction. In contrast, light exposure results in a marked enhancement of the CO conversion rate for Au₁/CeO₂–Fe, further validating the critical role of light in the room-temperature WGS reaction. Notably, the catalyst exhibited a CO conversion rate of almost 48.5% after 45 min, but this fell to 35% after 1 h. This decrease may be due to the catalyst undergoing pretreatment and preadsorption catalysis during the first 45 min. For photothermal synergistic conditions, as shown in Figure e, the experimental findings indicate that the CO conversion rate of Au₁/CeO₂–Fe is higher than that under heating alone, fully demonstrating a significant synergistic effect between light and heat. − The performance of the prepared SACs was compared to that of single metal–support catalysts in the WGS reaction. In Figure f, the apparent activation energy (E _a) of Au₁/CeO₂–Fe is 4.36 kJ·mol^–1, which is considerably lower than that of CeO₂–Fe (43.07 kJ·mol^–1) and Au₁/CeO₂ (5.66 kJ·mol^–1). On the one hand, light irradiation first excites the CeO₂–Fe support to generate photogenerated charge carriers. Moreover, strong interactions exist between the Au single atoms and the CeO₂ support, which facilitate charge separation and lower the energy barrier for forming reaction intermediates. On the other hand, thermal effects accelerate the adsorption and activation of reactant molecules on the catalyst surface. Meanwhile, photothermal synergy may regulate the reaction pathway, guiding the reaction toward a lower energy barrier. These mechanisms are an apparent reduction in the activation energy to 4.36 kJ·mol^–1. Furthermore, after undergoing 50 h of WGS reaction testing (Figure g), the Au₁/CeO₂–Fe catalyst maintained a CO conversion rate of 35% (at 25 °C), demonstrating the exceptional stability of single-atom catalysts prepared via the CSP method. As shown in Figure h and Supporting Information (Table S3), utilizing T ₅₀ (the temperature required to achieve 50% CO conversion) as the evaluation index, it is determined that the Au₁/CeO₂–Fe exhibited a reduced T ₅₀ value than traditional single metal–support catalysts, indicating its capacity to attain high efficiency CO conversion at a reduced temperature. Moreover, the photothermal reactions show significant advantages compared with thermal reaction alone and plasma-thermal reaction. In the function of photothermal synergism, the excitation of single-atom Au results in its full capacity to react with adsorbed CO molecules. Furthermore, the interaction between photoexcited electron hole pairs and thermally excited lattice vibrations collectively promotes the increase of active sites on the catalyst surface and the activation of CeO₂ lattice oxygen, thereby improving the activity and selectivity of the WGS reaction.

4.

CO concentration–time curves and H₂ production bar chart of the WGS reaction using CeO₂, CeO₂–Fe, Au₁/CeO₂, and Au₁/CeO₂–Fe catalysts under a wide-temperature range: (a) Low-temperature (25–85 °C), (b) middle-temperature (120–190 °C), and (c) high-temperature (225–300 °C). (d) WGS reaction over Au₁/CeO₂–Fe catalysts at room-temperature (25 °C) under light irradiation and dark environment conditions. (e) Relevance of synergistic effects between light and thermal using Au₁/CeO₂–Fe in the WGS reaction. (f) Arrhenius plot for activation energy about CeO₂–Fe, Au₁/CeO₂, and Au₁/CeO₂–Fe catalysts under a wide-temperature range. (g) Stability test of the WGS reaction over the Au₁/CeO₂–Fe catalysts upon extending the time-on-stream (50 h) at 25 °C. (h) Comparative pattern of single-atom catalysts and supported metal catalysts in the WGS reaction (T ₅₀).

3.4. Photothermal Catalytic Mechanism

In the MvK reaction mechanism, the lattice oxygen of the support material plays a pivotal role in the cyclic oxidation process of CO and oxygen. To elucidate the catalytic mechanism of the WGS reaction at room-temperature (25 °C), diffuse reflectance infrared Fourier transform spectroscopy (in situ DRIFTS) was employed to systematically characterize the characteristic peaks of reaction intermediates on the surfaces of CeO₂–Fe and Au₁/CeO₂–Fe catalysts, thereby proposing an efficient room-temperature WGS reaction pathway based on the MvK mechanism. Prior to the experiment, the reaction chamber was continuously purged with N₂ for 30 min to ensure that any residual impurity molecules would be eliminated and, thus, have no effect on the characterization results. As evidenced by the in situ DRIFTS spectra (Figure ), the WGS reaction resulted in the formation of characteristic peaks associated with reaction intermediates on the surfaces of CeO₂–Fe and Au₁/CeO₂–Fe. Specifically, the gaseous CO was adsorbed on the catalyst surface, where it subsequently combined with the lattice oxygen of CeO₂–Fe to form carbonate species (CO₃ ^2–). The CO adsorbed and activated on the catalyst surface was further converted into bridging carbonate (b-CO₃ ^2–) and monodentate carbonate (m-CO₃ ^2–). As shown in Figure a, the characteristic peaks at 1340, 1374, 1554, and 1574 cm^–1 were ascribed to b-CO₃ ^2–, while those at 1456, 1473, and 1539 cm^–1 corresponded to m-CO₃ ^2–. , Notably, only negligible quantities of b-CO₃ ^2– and m-CO₃ ^2– were identified on the CeO₂–Fe surface during the dark-condition adsorption experiment. In contrast, benefiting from its abundant oxygen vacancies and the superior CO adsorption capacity of single-atom Au, Au₁/CeO₂–Fe exhibited distinct characteristic peaks of b-CO₃ ^2– and m-CO₃ ^2– even in the absence of light irradiation. When exposed to light irradiation (Figure b), the intensity of the characteristic peaks for the two carbonate species on Au₁/CeO₂–Fe increased more than those on CeO₂–Fe as the irradiation time is extended. Meanwhile, single-atom Au exhibited the capacity to adsorb more CO molecules, thus ensuring the continuous generation of carbonate species. Additionally, a characteristic peak assigned to bicarbonate (HCO₃ ^2–) was noted at 1417 cm^–1 for both CeO₂–Fe and Au₁/CeO₂–Fe, and its intensity also increased with increasing irradiation time. The appearance of HCO₃ ^2– verified that *OH and *H generated from H₂O dissociation reacted with adsorbed CO, providing a feasible pathway for the formation of CO₂ and H₂. Of particular significance was the observation that the interfacial interaction between single-atom Au and the CeO₂–Fe support promoted CO adsorption and H₂O dissociation, accelerated the formation rate of HCO₃ ^2–, and thus realized an efficient room-temperature WGS reaction. The *COO intermediate was detected at 1361 and 1521 cm^–1, and the *COOH intermediate at 1505 cm^–1, whose formation is closely related to the catalytic mechanism of the WGS reaction. In the contour maps of in situ DRIFTS (Figure c,d), under conditions of continuous light irradiation, the intensities of *COOH and *COO characteristic peaks (red zone) on Au₁/CeO₂–Fe were higher than those on CeO₂–Fe, thus confirming that Au₁/CeO₂–Fe performs better in the efficient WGS reaction.

5.

In situ DRIFT spectra of (a) CeO₂–Fe and (b) Au₁/CeO₂–Fe under dark and light illumination conditions. Contour map of (c) CeO₂–Fe5% and (d) Au₁/CeO₂–Fe for in situ DRIFT spectra at 1300 to 1600 cm^–1. The density of states of (e) CeO₂–Fe and (f) Au₁/CeO₂–Fe catalyst. (g) WGS catalysis mechanism using CeO₂–Fe and Au₁/CeO₂–Fe catalysts.

In this study, we sought to elucidate the enhanced WGS performance exhibited by Au₁/CeO₂–Fe in comparison to CeO₂–Fe, and density functional theory (Supporting Information, Figure S19) was conducted to undertake a more profound analysis of their electronic structures. As shown in Figure e,f, the d-band center of Au₁/CeO₂–Fe was upshifted by 0.25 eV compared with that of CeO₂–Fe, which is closer to the Fermi level, confirming that Ov content increasing and loading of single-atom Au enhanced the separation of photogenerated carriers and the CO adsorption capacity. , In consideration of the results, a detailed catalytic mechanism (Figure g) of Au₁/CeO₂–Fe for the room-temperature WGS reaction is formulated. According to eq , the initial adsorption of gaseous CO and H₂O molecules on the dual active sites of the Au₁/CeO₂–Fe surface was facilitated by the Langmuir–Hinshelwood (L–H) mechanism, resulting in the formation of adsorbed species (*CO and *H₂O). Specifically, H₂O molecules were predominantly adsorbed on the oxygen vacancies and Fe-doped sites that are formed through Fe doping within the CeO₂ lattice. In contrast, single-atom Au generally carried a partial negative charge with increased electron density in its d-orbital, acting as an electron donor, and CO was predominantly adsorbed on the single-atom Au sites. In the Au₁/CeO₂–Fe catalytic system, SMSI between single-atom Au and the support facilitates the effective implementation of the MvK mechanism. Specifically, the Au₁/CeO₂–Fe system followed the MvK mechanism, where *CO adsorbed on single-atom Au sites combined with O atoms provided by lattice oxygen activation to form the *COO intermediate (eq ). The generation of *H from *H₂O dissociation rapidly combined with *COO to form the *COOH intermediate (eqs and ). Subsequently, under the interfacial effect between single-atom Au and CeO₂–Fe, the C–O bond of *COOH was cleaved, releasing a free O atom (eq ). This free O atom exhibits strong adsorption properties with regard to oxygen vacancies and refills the lattice oxygen vacancies, thus maintaining the activation and cycle of lattice oxygen in the support. An additional oxygen atom was formed through a nucleophilic reaction between the *OH and C atoms of *COOH, and the bond energy reconstruction led to the cleavage of *COOH into *COO and one *H atom. Eventually, the generation of CO₂ and the release of *H atoms. These *H atoms migrated into the catalyst surface and combined to form H₂ (eq and ). In summary, the presence of single-atom Au is demonstrated to enhance the adsorption and activation of CO, while the oxygen vacancies present in the support material are found to promote the dissociation of H₂O. The synergistic effect of the dual active sites reduced the reaction activation energy. Furthermore, under light irradiation, the coupling effect between photoexcited electron–hole pairs and thermally excited lattice vibrations further increased the number of active sites on the catalyst surface and accelerated the cycle of CeO₂ lattice oxygen, thus realizing an efficient WGS reaction at room-temperature. The WGS mechanism not only explains the excellent catalytic performance of Au₁/CeO₂–Fe at room temperature but also provides an important theoretical basis for the design of novel photothermal synergistic catalysts.

$$\mathrm{CO}+{\mathrm{H}}_2\mathrm{O}+{}^{*}\to {}^{*}\mathrm{CO}+{{}^{*}\mathrm{H}}_2\mathrm{O}$$ 3

$${}^{*}\mathrm{CO}+{}^{*}{\mathrm{H}}_2\mathrm{O}+{}^{*}{\mathrm{O}}_{(\mathrm{Lattice})}\to {}^{*}\mathrm{COO}+{{}^{*}\mathrm{H}}_2\mathrm{O}$$ 4

$${}^{*}{\mathrm{H}}_2\mathrm{O}\to {}^{*}\mathrm{H}+{}^{*}\mathrm{OH}$$ 5

$${}^{*}\mathrm{COO}+{}^{*}\mathrm{H}\to {}^{*}\mathrm{COOH}$$ 6

$${}^{*}\mathrm{COOH}+{}^{*}\mathrm{OH}\to {{}^{*}\mathrm{COO}}_{(\mathrm{OH})}+{}^{*}2\mathrm{H}+{{}^{*}\mathrm{O}}_{(\mathrm{Lattice})}$$ 7

$${}^{*}\mathrm{COO}+{}^{*}2\mathrm{H}\to {{}^{*}\mathrm{CO}}_2+{{}^{*}\mathrm{H}}_2$$ 8

$${{}^{*}\mathrm{CO}}_2+{{}^{*}\mathrm{H}}_2\to {}^{*}+{\mathrm{CO}}_2+{\mathrm{H}}_2$$ 9

4. Conclusion

In summary, this study focuses on the unique advantages of SACs for WGS reactions at 25 °C and room-temperature. The Au₁/CeO₂–Fe single-atom catalyst was synthesized using a CSP coupling system. The results of the HAADF-STEM, XANES, and EXAFS analyses confirmed the successful preparation of SACs by observing a single-atom Au loading on the CeO₂-based support surface. XPS and Raman spectroscopy verified changes in the Ov content following Fe doping and single-atom Au loading. The MvK mechanism was the foundation for the provision of key evidence for the activation of surface reactive oxygen species and lattice oxygen by the O₂-TPD. In situ Raman spectroscopy was employed to confirm a positive correlation between oxygen vacancies in the CeO₂-based catalyst and temperature. Furthermore, in situ DRIFTS revealed the positive impact of increased oxygen vacancy concentration on the formation and transformation of reaction intermediates, providing insights into the photothermal synergistic WGS catalytic mechanism. The Au₁/CeO₂–Fe catalyst exhibits outstanding catalytic activity across the 25–300 °C temperature range, achieving a remarkable 35% CO conversion rate at 25 °C. This performance stems from the stable oxygen cycling mechanism within the CeO₂ support and the strong interaction between the gold single atom and the support. This study not only provides a novel strategy for developing highly efficient room-temperature WGS catalysts but also reveals the photothermal synergistic effect and dual-active-site mechanism, offering important references for designing other heterogeneous catalytic reaction systems.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.6c00158.

• Additional material synthesis, material characterization, SEM, HAADF-STEM, XRD, XANES, EXAFS, ESR, XPS, bandgap value, WGS reaction, DFT methods, EXAFS fitting results, and comparison of WGS performance (PDF)

The authors declare no competing financial interest.

Published as part of JACS Au special issue “Continuous Flow Chemistry”.