Selective Reduction of CO2 to CO via the RWGS Reaction over ZnO-ZrO2-Ga2O3-Supported Catalysts Modified with Keggin-Type Heteropolyacid Precursors

TL;DR

This paper explores how to efficiently convert CO2 into CO using modified catalysts at low temperatures.

Contribution

The study introduces Mo/ZZG catalysts with Keggin-type heteropolyacid precursors for efficient CO2 reduction.

Findings

Mo/ZZG catalysts showed excellent activity in the RWGS reaction at low temperatures.

These catalysts had the lowest activation energy and highest CO2 conversion efficiency.

Characterization techniques revealed strong links between structure and performance.

Abstract

Mo/ZZG and W/ZZG nanomaterials for the catalytic reduction of CO2 were successfully prepared from preformed ZnO-ZrO2-Ga2O3 (ZZG) and HPMo and HPMo heteropolyacids via simple incipient wetness impregnation. To establish the relationship between structural properties and catalytic performance, the prepared catalysts were deeply characterized using XRD, Raman spectroscopy, SEM coupled with EDX, BET, XPS, and H2-TPR techniques. The catalytic performance of the materials was evaluated in the RWGS reaction under atmospheric pressure, using a feed composition of CO2/H2/N2 = 1/3/1 across a temperature range of 250–600 °C. All materials were active in the reverse water gas shift reaction (RWGS) under these conditions, with the Mo/ZZG catalyst exhibiting the best performance, demonstrating excellent catalytic activity at low temperature with the lowest activation energy and the highest CO2 to CO conversion efficiency.

Figures

Full text

1. Introduction

Carbon dioxide (CO_2_), despite being the main anthropogenic greenhouse (GHG) gas and a major contributor to global warming, is currently considered as strategic carbon resource rather than just a GHG that needs to be mitigated. Its valorization as a raw material for the production of fuels and valuable chemical products represents a major environmental and economical challenge [1]. In order to contribute to the reduction in CO_2_ in the atmosphere, several carbon capture, storage and conversion technologies have been developed in recent years [2,3]. Many research projects have focused on converting CO_2_ into high-value-added chemical products, such as methanol, dimethyl carbonate, polymers, fuels and syngas [4,5]. Among the various CO_2_ conversion pathways, the Reverse Water Gas Shift (RWGS) reaction (Equation (1)) occupies a central role in the valorization of CO_2_ into carbon monoxide (CO), an abundant and low-cost C1 feedstock. In this reaction, CO_2_ is reduced by H_2_ to produce a mixture of CO and water vapor. After separation from H_2_O, CO serves as a versatile intermediate for the synthesis of a wide range of chemicals and fuels [4], while syngas constitutes a strategic resource for industrial processes such as Fischer–Tropsch synthesis and methanol production [1].

$$[eqn]$$

Several catalysts have been studied to improve the catalytic activity and selectivity of the RWGS reaction, mainly based on metals and metal oxides [6,7,8,9]. Wu et al. [10] produced 4.0 mol g^−1^ h^−1^ of CO with 90% selectivity using Ru/Mo_2_TiC_2_, while Pajares et al. [11] achieved ~100% CO selectivity and ~11% CO_2_ conversion using the VC(Pr)/Al_2_O_3_ catalyst. Ai et al. [12] tested the mesoporous Cu/γ-Al_2_O_3_ catalyst, which reached a CO selectivity of ~100%. Kipnis et al. [13] studied MoO_3_/γ-Al_2_O_3_, which was able to achieve 6 vol% CO and a CO_2_ conversion of 19.6%. Molybdenum (VI) oxide catalysts have shown strong potential for the RWGS reaction, due to their high CO selectivity and resistance to disactivation [14]. However, their limited activity results in moderate CO_2_ conversion. The incorporation of metals such as Cu, Ni and Pt can significantly improve its catalytic efficiency [15]. In a recent study conducted by Aldajani and co-workers [16], they investigated catalysts based on Keggin-type polyoxometalates as precursors supported on different oxides (γ-Al_2_O_3_, TiO_2_, SiO_2_ and CeO_2_), where H_3_PMo_12_O_40_/SiO_2_ showed the highest activity with CO_2_ conversion of 35% and CO selectivity of 100%.

Polyoxometalates (POMs) are anionic molecular clusters composed of transition metals (W, Mo, V, etc.) and a heteroatom (P, Si, Ge, etc.) linked by oxygen bridges [17]. Heteropolyacids (HPAs) with a Keggin structure are widely adopted as catalysts in a variety of reactions due to their strong acidic and redox properties [18]. However, these HPAs have a low surface area; hence, research has been directed towards supported HPAs which present a higher number of surface acid sites than their bulk counterparts, promoting their use as heterogeneous catalysts [19]. Transition metal oxide compounds have been widely used as supports. For example, ZnO is suitable as a catalyst for CO_2_ activation to CO [20], while ZrO_2_ was found to be active in the synthesis of methanol from CO and H_2_ [21]. In addition, ZnO and ZrO_2_ can form binary oxides, as the sizes of the Zn^2+^ and Zr^4+^ cations are almost identical, enabling them to exchange and form stable crystal structures [22,23]. Furthermore, doping a ZnO-ZrO_2_ solid solution with small quantities of other metal oxides can improve its catalytic properties. For instance, doping Cu-based catalysts with Ga_2_O_3_ increased the selectivity of methanol in the CO_2_ hydrogenation process, making ZnO-ZrO_2_-Ga_2_O_3_ a promising cost-effective candidate to be used as a support [24]. Its multifunctional surface combines Lewis acidic sites, basic sites, and oxygen vacancies, which enhances CO_2_ activation and promotes hydrogen spillover toward adsorbed CO_2_ intermediates. Additionally, the support’s balanced acid-base character and high thermal stability contribute to improved CO selectivity and long-term catalyst stability [25].

In this context, commercial Keggin-type HPAs, namely H_3_PW_12_O_40_ and H_3_PMo_12_O_40_ were impregnated onto a synthesized ZnO-ZrO_2_-Ga_2_O_3_ mixed oxide; the resulting supported materials were characterized by several spectroscopic, structural and surface analytical methods (XRD, RAMAN, SEM-EDX, BET, XPS and H_2_-TPR) before the evaluation of their catalytic performances in the RWGS reaction.

2. Materials and Methods

2.1. Catalysts Preparation

Phosphotungstic acid (H_3_PW_12_O_40_/HPW) and phosphomolybdic acid (H_3_PMo_12_O_40_/HPMo) are commercial products provided by Sigma Aldrich (St. Louis, MO, USA).

ZnO-ZrO_2_-Ga_2_O_3_ (ZZG) support was synthesized using coprecipitation method. Firstly, a mixture of (Zn(NO_3_)2, xH_2_O) and (ZrO(NO_3_)2, xH_2_O) were dissolved in deionized water with a ratio of 13–87% w/w. Under constant stirring at T = 50 °C, a solution of ammonium carbonate (NH_4_)2_CO_3 was added dropwise with a dropping funnel over a period of 30 min thus forming a white precipitate. The obtained suspension was maintained at 70 °C under stirring for 2 more hours before letting it cool to room temperature. The obtained precipitate was filtered and rinsed with deionized water, then a solution of (Ga(NO_2_)3, 8H_2_O) (3% w/w as Ga) dissolved in minimum of deionized water was added dropwise with constant stirring and grinding, forming a dense slurry. The final obtained thick slurry was left at room temperature for an hour and dried at 100 °C overnight. Afterwards, it was grinded and calcined in static air at 500 °C (8 °C/min) for 3 h.

A total of 10% wt% M/ZZG (M = W et Mo) were prepared through incipient wetness technique. The appropriate weight of HPW and HPMo precursors was dissolved in a minimum of deionized water and added to the support ZZG in a dropwise manner. The resulting materials were left at room temperature for an hour and dried at 100 °C overnight. Thereafter, it was calcined in static air using the same conditions as the previous step.

2.2. Characterization

To identify different phases composition, crystallinity, and to determine the crystallite size of the samples present in the obtained catalysts and supports, Powder X-Ray Diffraction (XRD) were carried out by a PANalytical X’Pert PRO MPD Alphal diffractometer (Malvern, UK), using CuKa radiation. All the samples have been analyzed between 2θ = 5° and 100° at 0.017° steps and 50 s for every step. To calculate the crystallite size, the Scherrer equation was used. Scanning Electron Microscopy (SEM) and Energy-Dispersive X-ray Spectroscopy (EDX) were used to study the surface morphology and composition. The equipment used was a FE-SEM JEOL JSM-7001F (FESEM, JEOL Ltd., Tokyo, Japan). Raman spectroscopy was carried out with a Jobin-Yvon LabRam HR 800 (Horiba Jobin Yvon, Longjumeau, France), and used to identify species in the synthetized samples, the laser used had a near UV light of 325 nm and the samples had an exposition time of 60 s for every measurement. The porosity and textural properties were investigated through N_2_ adsorption–desorption isotherms, measured using a Micromeritics Tristar II 3020 analyzer (Micromeritics Instrument Corporation, Norcross, GA, USA). Specific surface areas were determined via multipoint Brunauer–Emmett–Teller (BET) analysis of the obtained isotherms. Prior to analysis, samples underwent a degassing step at 250 °C for 5 h under high vacuum to remove any moisture or contaminants that could affect adsorption. Adsorption isotherms were established by the progressive introduction of known quantities of the adsorbate (N_2_) into the analysis cell, at defined relative pressure steps until equilibrium was reached at each measurement point. X-ray photoelectron spectroscopy (XPS) was carried out using an ESFOSCAN equipment based on the PHI VersaProbe 4 instrument from Physical Electronics (ULVAC-PHI, Chanhassen, MN, USA) with monochromatic 1486.6 eV Al Kα radiation, calibrated using the 3d5/2 line of Ag. Measurements are referenced to the C 1s signal at 284.8 eV of adventitious carbon. The analysis was performed using MultiPak V6.0A software, and the fitting of the spectra was made with CasaXPS program. All measurements were carried out in an ultra-high vacuum chamber at a pressure between 5 × 10^−10^ and 5 × 10^−9^ Torr, without any sample pretreatment. To study the behavior of the samples in a reducing hydrogen atmosphere, as well as to identifying possible species and evaluate synergetic effects, the Hydrogen Temperature-Programmed Reduction (H_2_-TPR) technique was carried out, using a Micromeritics AutoChem II Chemisorption Analyzer device. Approximately 50 mg of the sample was put in a U-shaped quartz reactor, over an inactive layer of quartz wool and introduced into the device. A pretreatment was carried out under a helium atmosphere by applying a temperature ramp up to 120 °C at a heating rate of 10 °C/min, maintaining the maximum temperature for 15 min. Subsequently, the system was allowed to cool down to room temperature, and afterwards, the temperature was increased at a rate of 10 °C/min up until 800 °C under a flow of H_2_/Ar at 12%. The hydrogen consumption at every temperature was measured via a thermal conductivity detector (TCD).

2.3. RWGS Catalytic Activity

The catalytic tests of the prepared materials in the RWGS reaction were carried out in a tubular fixed-bed reactor (316-L stainless steel, 305 mm long, 9 mm i.d.) provided with a thermocouple in direct contact with the catalyst in a Microactivity-Reference unit PID Eng&Tech. For every test, 150 mg of the catalyst was diluted with inactive silicon carbide up to 1 mL and placed between two quartz wool layers. The tests were performed across a temperature range of 250–600 °C with a system pressure of 1 bar. The temperature was increased from room temperature up to 250 °C under N_2_, afterwards the sample was exposed to CO_2_/H_2_/N_2_ = 1/3/1 (molar ratio) flow under a gas hourly space velocity (GHSV) of 3000 h^−1^, and the catalytic behavior was analyzed at 250, 300, 400, 500 and 600 °C. At each selected reaction temperature, the system was maintained under isothermal conditions for 90 min. The first 30 min were allocated to system stabilization, while the subsequent 60 min were used to acquire two independent chromatographic measurements. Temperature transitions between consecutive setpoints were carried out using heating ramps of 30 min, the catalytic behavior was followed for 9.5 h. The products were analyzed on-line using Varian 450-GC equipped with a TCD, two FID detectors, one equipped with a methanizer. The carbon balance between the outlet gas and the reactant inlet was ≤1%. CO_2_ conversion ( $[eqn]$ ) and selectivity to the i product ( $[eqn]$ ) were calculated as follows:

$$[eqn]$$ $$[eqn]$$

where (C_i_) and C_CO_2__ are the molar concentration of the i product (CO or CH_4_) and CO_2_, respectively.

3. Results

3.1. Characterization of Catalysts

3.1.1. XRD Analysis

XRD diffractograms of the prepared materials presented in Figure 1 display intense peaks at 2θ = 30.54°, 35.42°, 50.89°, 60.49° and 63.54° indexed to the (101), (110), (112), (211) and (202) planes, in agreement with the PDF Card 88-1007 of tetragonal phase of ZrO_2_ (space group P4_2_/nmc) [23,26]. The absence of ZnO and Ga_2_O_3_ peaks may be attributed to its low concentration and high dispersion. The similar ionic radii of Zn^2+^ and Zr^4+^ favor the formation of a ZnO–ZrO_2_ solid solution, where Zn^2+^ species are incorporated into the ZrO_2_ lattice rather than forming segregated ZnO phases. Consequently, lattice distortion occurs without the appearance of separate ZnO reflections [27]. Likewise, the low Ga content combined with its high dispersion and possible incorporation into the mixed oxide matrix, renders Ga_2_O_3_ undetectable by XRD, as Ga^3+^ species are likely present in highly dispersed or amorphous forms [28]. Similarly, the presence of W and Mo on the ZZG surface did not alter the crystalline phase, indicating that these species are uniformly dispersed without significantly affecting the crystal structure of the supports the impregnation process.

3.1.2. SEM-EDX Analysis

The surface structure of the HPAs and their dispersions on the ZZG support surface were examined using SEM analysis. Figure 2a,d correspond to ZZG, which has a relatively homogeneous surface with a porous texture and uniform particle distribution. This structure offers good anchoring capacity for impregnated materials. Images (b, e) and (c, f) display significant variations in morphology after dispersion of W and Mo on the ZZG surface. For W/ZZG, dispersed aggregates are formed, with heterogeneous distribution and increased surface irregularities. The increased rugosity suggests an interaction between ZZG and W, which modifies the overall porosity. Mo/ZZG shows a denser, more compact morphology, with more homogeneous coverage of the substrate. The particles appear better integrated, reducing the void zones observed for W/ZZG. Comparing ×10 and ×100 magnifications confirms these observations, highlighting differences in particle distribution and texture evolution after impregnation.

In addition, EDX analysis (Table 1 and Figure S1) was carried out to better understand the elemental composition of the catalysts surface and confirmed the presence of W and Mo with ratios very close to those determined theoretically according to the stoichiometric composition and which were not identified by XRD. Moreover, elemental mapping confirms a uniform and coherent dispersion of W and Mo particles on the ZZG surface (Figure S2).

3.1.3. RAMAN Spectroscopy

Raman spectroscopy was used to further characterize the phase structure of the prepared samples (Figure 3). The Raman spectra recorded for ZZG revealed a narrow band at 148 cm^−1^ and broad bands centered at 268, 318, 464, 602 and 641 cm^−1^ characteristic of tetragonal phase of ZrO_2_ [29]. Both W/ZZG and Mo/ZZG revealed bands associated with the ZZG support. However, a slight shift was detected in the vibrational bands, more pronounced at 268 and 318 cm^−1^, which was clearly evidenced for Mo/ZZG sample. This shift in band position can be interpreted as an indication of new bond formation or oxygen deficiency [30]. Additional bands appeared at 823 and 970 cm^−1^ assigned to M-O (M = W and Mo) stretching and terminal M=O_t_ bond of the dehydrated surface MO_x_ species [31,32,33]. It should be noted that the Keggin structure bands completely disappeared.

3.1.4. BET Analysis

Figure S3 and Table 2 display textural properties for ZZG, W/ZZG and Mo/ZZG. As observed, all compounds showed IV-type isotherms and H2-type hysteresis. ZZG exhibited a surface area of 44.6 m^2^ g^−1^. The surface area progressively decreased with the tungsten and molybdenum impregnation until a value of 32.4 and 22.1 m^2^ g^−1^, respectively. In addition, average pore volume followed a similar trend as the surface area. This decrease in textural parameters may be due to the support surface blocking by the WO_x_ and MoO_x_ addition [34]. Otherwise, the samples displayed pore size distribution in the 5–45 nm range, indicating its mesoporous nature.

3.1.5. XPS Analysis

The XPS spectra of the ZZG-based materials doped with tungsten or molybdenum (Figure 4 and Figure S4) confirmed the presence of Zn, Zr, Ga, W, Mo, P, and O, with no additional elements detected, indicating high chemical purity in agreement with XRD and EDX results. ZZG material display a Zn 2p peak at 1121.9 eV, corresponding to Zn^2+^ in a typical oxide environment ZnO [35]. The Zr 3d region reveals two signals corresponding to Zr 3d_5_/2 and Zr 3d_3_/2 appear, respectively, at 182.1 eV and 184.4 eV that are typical binding energies of Zr^4+^ in ZrO_2_ [36]. The Ga 2p region shows two main signals of Ga 2p_3/2_ and Ga 2p_1/2_ located at 1117.8 eV and 1144.8 eV characteristic of Ga^3+^ in Ga_2_O_3_ [37]. W/ZZG and Mo/ZZG samples show, after deconvolution, similar spectra to ZZG for Zn, Zr and Ga elements. However, a slight shift to higher BE is observed that indicates electronic interactions between support and Mo–O, W–O groups. In addition, low-intensity peaks are observed in the P 2p region at around 134 eV, confirming the presence of phosphorus [38], although the weak signal hinders reliable identification of its chemical state. A peak around 140 eV (characteristic binding energy of Zn 3s core level) was also observed for all materials. W 4f and Mo 3d regions displayed characteristic signals of WO_3_ and MoO_3_, indicating the presence of W and Mo oxide phases on the surface of the samples and effective surface dispersion [38,39].

The deconvoluted XPS spectra of the O 1s region reveals two main peaks at ~530 eV (lattice oxygen, O_α_) and ~532 eV (adsorbed oxygen species associated with oxygen vacancies, O_β_). The calculated O_β_/(O_β_ + O_α_) ratios follow the order Mo/ZZG (23.2%) > W/ZZG (22.8%) > ZZG (20.9%), indicating that the introduction of Mo leads to the highest concentration of surface oxygen vacancies [40].

Surface atomic ratios determined by XPS analysis are compilated in Table 3. The results reveal a clear surface reorganization upon deposition of tungstic and molybdic active phases. Compared to the ZZG support, both W/ZZG and Mo/ZZG catalysts exhibit surface enrichment in Zn accompanied by a slight decrease in Ga content. Furthermore, the W/Zr and Mo/Zr ratios (0.15 and 0.46, respectively) indicate a substantially higher surface accessibility of molybdenum species. These changes suggest partial masking of Ga sites and greater exposure of Zn at the catalyst interface, while the incorporation of Mo species appears more effective than that of W species [41].

3.1.6. H2-TPR Analysis

The H_2_-TPR profiles are illustrated in Figure 5. The ZZG support reveals two distinct reduction peaks at 424 °C and 542 °C due to the consumption of 2.14 and 1.08 cm^3^ g^−1^ STP of H_2_, likely resulting from ZnO reduction. For the W/ZZG catalyst, the H_2_-TPR profile exhibits a relatively low hydrogen consumption at 402 °C (1.34 cm^3^ g^−1^ STP) and 657 °C (0.8 cm^3^ g^−1^ STP). These results indicate that the WO_x_ species, formed during the calcination of the material at 500 °C, are thermally stable and poorly reducible at lower temperatures. In contrast, the Mo/ZZG catalyst displays a markedly different behavior. A strong reduction peak is observed at 433 °C, with a maximum hydrogen consumption of 10.7 cm^3^ g^−1^ STP, followed by a minor peak at 657 °C. This intense low-temperature peak reflects the high reducibility of Mo^6+^ species, which are more easily reduced than their tungsten counterparts. In agreement with previous reports [16,42,43], the high hydrogen uptake suggests that MoO_x_ species are well dispersed and strongly interacting with the support, resulting in a more accessible and reactive surface, which is in line with the relative intensity of the Raman bands in the 100–1010 cm^−1^ range associated to the presence of WO_3_ and MoO_3_ species in the catalysts. The secondary peak at 657 °C may correspond to residual MoO_x_ species or support-related reduction processes, similar to those observed in the tungsten-based catalyst.

3.2. Catalytic Performance

The catalytic performance of ZZG, W/ZZG, and Mo/ZZG materials was evaluated in the RWGS reaction using a fixed-bed reactor under a CO_2_/H_2_/N_2_ atmosphere (1/3/1) across a temperature range of 250–600 °C. Catalyst stability was monitored over a 4 h period. CO_2_ conversion results are presented in Figure 6 and Table 4.

All catalysts demonstrated activity for CO_2_ reduction to CO, with conversion increasing with temperature, consistent with the endothermic nature of the RWGS reaction (Equation (1)). The ZZG support exhibited notable activity, achieving a CO_2_ conversion of 44.6% at 600 °C with a constant CO selectivity of 100%. The introduction of metal phases significantly enhanced catalytic performance at lower temperatures. Mo/ZZG achieved conversions of 29% at 500 °C and 40% at 600 °C, while W/ZZG reached 18% and 38% at the same respective temperatures. In the 250–500 °C range, Mo/ZZG outperformed W/ZZG and ZZG.

CO selectivity remained high for both metal-modified catalysts, although CH_4_ formation was observed between 300 and 500 °C. This effect was more pronounced for Mo/ZZG, reaching up to 16.7% CH_4_ at 300 °C, compared to 6.7% for W/ZZG at the same temperature. CH_4_ selectivity decreased at higher temperatures due to the exothermic nature of methanation (Equation (4)) and Sabatier reactions (Equation (5)).

$$[eqn]$$ $$[eqn]$$

Kinetic data measured using ZZG support exhibited the highest apparent activation energies (E_a_) up to 77.2 kJ mol^−1^ between 300–500 °C. The incorporation of MoO_x_ and WO_x_ significantly reduced this barrier; Mo/ZZG reached a minimum activation energy of 46.47 kJ mol^−1^, while W/ZZG showed an E_a_ of 52.64 kJ mol^−1^ over the same temperature range.

4. Discussion

RWGS reaction was investigated under a CO_2_/H_2_/N_2_ (1/3/1) atmosphere over a temperature range of 250 to 600 °C. The ZZG support and the modified HPA catalysts proved to be active for RWGS under the experimental conditions used. The highest CO_2_ conversion rate (44.6% at 600 °C) and amount of CO (66.1 mmol CO g^−1^ h^−1^) was obtained with the ZZG support, which also exhibited the largest BET specific surface area (44.6 m^2^ g^−1^). This result is consistent with the known synergistic effects of the ZnO-ZrO_2_ material, which enable superior CO_2_ activation. The incorporation of Ga_2_O_3_ is crucial, as it enhances CO_2_ hydrogenation performance by regulating the dispersion and the distribution of active sites throughout catalyst surface [44]. Thus, the designed ZnO-ZrO_2_-Ga_2_O_3_ support effectively combines these properties, which explains the observed improvement in catalytic performance.

Therefore, from this study, the incorporation of molybdenum into the ZZG support is evidenced to significantly enhance the catalytic performance and the reaction kinetics of Mo/ZZG material. Mo/ZZG exhibits the lowest E_a_ (46.47 kJ mol^−1^), indicating a superior ability to initiate the RWGS reaction at lower temperatures. This enhancement is attributed to the introduction of highly reducible MoO_3_ species, which strongly interact with the ZZG support. H_2_-TPR analysis reveals an intense reduction peak at 433 °C with a hydrogen consumption of 10.7 cm^3^ g^−1^ STP, confirming the formation of readily reducible sites. SEM elemental mapping demonstrates the efficient dispersion of MoO_3_ across the support, while XPS confirms strong Mo–support interactions. Mo/ZZG possesses the highest O_β_/(O_β_ + O_α_) ratio (23.2%), reflecting the greatest density of surface oxygen vacancies. These vacancies are known to serve as preferential sites for CO_2_ adsorption and activation [45]. This is further supported by Raman spectroscopy, where a pronounced shift for Mo/ZZG indicates a modified local chemical environment, likely due to new bond formation or oxygen deficiency. This enhanced reducibility explains the high catalytic activity at low temperatures, as well as the formation of CH_4_, which results from deeper H_2_ activation.

In contrast, WO_3_ species are thermally stable and less reducible, limiting the available oxygen vacancy density for CO_2_ activation. W/ZZG displays intermediate kinetic behavior, with an E_a_ of 52.64 kJ mol^−1^. This value indicates improved CO_2_ activation compared to the pristine ZZG support but is less efficient than Mo/ZZG. This trend is consistent with XPS results showing a slightly lower oxygen vacancy concentration (22.8%) and H_2_-TPR data revealing limited hydrogen consumption at low temperature. The reducibility of WO_x_ species occurs predominantly at higher temperatures, as reflected by the reduction peak at 657 °C, implying that active redox sites are progressively generated with increasing thermal input. Consequently, the kinetics of W/ZZG are governed by gradual activation of catalytic sites, resulting in improved performance mainly at elevated temperatures.

Furthermore, the post-reaction structural XRD analysis reveals a critical distinction between the ZZG support and the promoted catalysts (Figure S6). For the pristine ZZG material, distinct diffraction peaks appear at 2θ ≈ 47° and 65°, corresponding to the (102) and (200) planes of hexagonal ZnO, respectively. Their emergence indicates a partial breakdown of the solid solution under the reducing and steam-rich RWGS environment. This segregation underscores the metastability of the undoped ZZG framework when subjected to high-temperature hydrothermal conditions. In marked contrast, the Mo/ZZG and W/ZZG catalysts exhibit no such phase separation, with their XRD patterns remaining identical to those of the fresh samples. This exceptional stability can be attributed to the incorporation of MoO_x_ or WO_x_ species, which likely integrate into the oxide matrix or form stabilizing interfacial structures, thereby suppressing ZnO mobility and crystallization.

For all catalysts, the progressive decrease in apparent activation energy with changing the temperature of catalytic measurement from lower to higher ranges (Figure S5) indicates a shift toward thermodynamically controlled mechanisms, where the role of redox-active sites becomes less dominant. Among the three systems, ZZG shows the highest E_a_ values across all temperature ranges, reaching 77.20 kJ mol^−1^ between 300 and 500 °C. This high barrier reflects its limited intrinsic ability to activate CO_2_ at low temperatures, consistent with the absence of highly reducible redox sites. However, the gradual decline in E_a_ with temperature suggests that thermally activated processes, such as defect formation or partial ZnO reduction, become more accessible, enhancing the support’s reactivity at elevated temperatures [46].

Overall, the comparison of activation energies highlights Mo/ZZG as the most kinetically favorable catalyst, particularly in the temperature range (300–500 °C), where RWGS is thermodynamically favorable but still kinetically constrained. Mo’s ability to significantly lower the energy barrier in this critical zone underscores its potential for moderate-temperature applications. W/ZZG, while less reducible, shows steady improvement with temperature, whereas ZZG, despite its initial inertness, becomes competitive at high temperatures due to thermally activated mechanisms. This observation aligns with the findings of Aldajani et al., [16], who demonstrated that HPMo decomposes under reaction conditions into highly active molybdenum oxide species, positioning it as a promising precursor for RWGS catalysts. These Mo-based systems exhibit excellent CO selectivity and efficient CO_2_ conversion at elevated temperatures, typically operating via the Mars–Van Krevelen mechanism involving the reduction in Mo(VI) to Mo(IV), followed by reoxidation by CO_2_. The Mars–Van Krevelen mechanism itself involves a dynamic surface process, where gas-phase reactants interact with adsorbed species on the catalyst. The reaction proceeds through a well-defined sequence of elementary steps: adsorption of the reactants onto the catalyst surface, surface-mediated transformation into products, and subsequent desorption, which regenerates the active sites. A distinctive feature of this mechanism is the direct participation of the catalyst via the provision of lattice oxygen, subsequently replenished from the gas phase. This cycle sustains continuous catalytic activity and can be described kinetically by accounting for the surface coverage of active sites and adsorbed intermediates, leading to rate expressions that reflect the surface dynamics [47].

Table 5 summarizes a selection of non-noble-metal-based catalysts reported in the literature for the RWGS reaction, together with their corresponding reaction conditions, allowing a direct comparison with the Mo/ZZG catalyst, which exhibits the best catalytic performance at high temperatures in this work. Most of the reported catalysts are based on group 4 and 5 transition metals, particularly V and Mo, reflecting the strong interest in these elements as cost-effective alternatives to noble metals for RWGS catalysis. Mo/ZZG achieves 100% CO selectivity, with competitive performance of previously reported catalysts, while reaching a CO_2_ conversion of approximately 40%, which is close to that obtained with the VC_x_ catalyst, reported to operate under similar conditions.

5. Conclusions

In this study, we successfully synthesized composite materials based on ZrO_2_ and ZnO, doped with Ga_2_O_3_ to form the ternary oxide system ZnO-ZrO_2_-Ga_2_O_3_. Using incipient wetness impregnation, we further modified this support with tungsten and molybdenum oxides derived from Keggin-type HPAs, namely HPW and HPMo. These HPAs were uniformly dispersed onto the ZZG surface, as confirmed by SEM and EDX analyses, which revealed homogeneous distribution of all constituent elements (Zn, Zr, Ga, W, and Mo). XPS analysis revealed the presence of structural defects and oxygen vacancies, particularly pronounced in the Mo-doped samples. The oxidation state analysis confirmed the coexistence of Mo^6+^ and W^6+^ species, along with strong electronic interactions between the metal centers and the ZZG support. H_2_-TPR profiles showed that Mo/ZZG exhibits superior reducibility compared to W/ZZG, with reduction peaks occurring at lower temperatures. This enhanced reducibility suggests easier activation of oxidized species, which correlates with improved catalytic performance at low temperatures. The Mo/ZZG catalyst demonstrated excellent performance in the RWGS reaction in the intermediate temperature range (300–500 °C), characterized by a lower activation energy and enhanced CO_2_ conversion. In contrast, the pristine ZZG support exhibited notable activity only at high temperature (600 °C). The presence of molybdenum promotes the formation of active redox sites, although CH_4_ formation was observed at lower temperatures. Post-reaction XRD analysis revealed that the use of Keggin-type HPAs as precursors for MoO_x_ or WO_x_ species provides an effective route to stabilize the ZZG catalyst against phase segregation under RWGS conditions, ensuring its long-term structural integrity.