Figure 4 presents the XRD patterns of the modified Ni/MgO-MgAl₂O₄ catalysts before and after reduction as well as after the reaction. All samples exhibit characteristic diffraction peaks associated with Ni-, Mg- and Al-containing phases. Strong diffraction peaks at 42.9° and 62.4° can be assigned to the MgO crystalline phase (JCPDS 45-0946), while the weak peaks located at 36.8°, 44.9°, 59.7° and 65.2° correspond to the MgAl₂O₄ spinel phase (JCPDS 21-1152). In addition, diffraction peaks at 31.4°, 37.0°, 45.0°, 59.7° and 66.5° are attributable to the NiAl₂O₄ phase (JCPDS 10-0339). As shown in Figure 4(a), characteristic peaks of cubic NiO appear at 37.2°, 43.3° and 62.9°, partially overlapping with those of MgO. The Ce-, La- and Zr-modified catalysts exhibit similar diffraction features to the unmodified 12.5%Ni sample, while no distinct diffraction peaks corresponding to CeO₂, La₂O₃ or ZrO₂ are detected, indicating that these promoter oxides are either highly dispersed or present in an amorphous state.

Notably, the diffraction peak intensities of both MgO and MgAl₂O₄ phases reduce after promoter incorporation, suggesting that Ce, La and Zr doping decreases the crystallinity of the support. Meanwhile, the decreased intensity of NiO diffraction peaks implies enhanced dispersion of Ni species and reduced Ni particle size, which is beneficial for achieving a more uniform distribution of Ni.

Figure 4(b) shows the XRD patterns of the reduced catalysts. All samples retain the mixed MgO-MgAl₂O₄ support structure, whereas the diffraction peaks of NiO disappear and those of NiAl₂O₄ significantly weaken. Meanwhile, characteristic peaks of metallic Ni appear at 44.5° and 51.8° (JCPDS 04-0850). Among them, the unmodified 12.5%Ni sample exhibits markedly stronger Ni⁰ diffraction peaks than the three modified catalysts. It should be emphasized that although improved metal dispersion is generally favorable for catalytic activity, the activity decrease observed in DRM is not attributable to a reduction in the amount of Ni, but is more likely related to promoter-induced regulation of reaction pathways or the promotion of side reactions. Figure 4(c) presents the XRD patterns of the catalysts after 20 h of reaction. All samples display diffraction peaks corresponding to Ni, MgO, NiAl₂O₄ and MgAl₂O₄ phases. The diffraction intensity of metallic Ni is significantly enhanced compared with the reduced state, indicating that Ni particles undergo sintering to varying extents during the reaction. Notably, the increase in Ni peak intensity for the modified catalysts is much less pronounced than that for the 12.5% Ni-loaded sample, suggesting that interactions between Ce, La, Zr and Ni effectively inhibit the metal sintering. In addition, new diffraction peaks appearing at 22.9°, 34.8° and 39.2° can be assigned to hydrotalcite-like carbonates (HTLcs, JCPDS 70-2151), which are probably formed via surface hydration-carbonation-induced structural reconstruction of metal oxides under the combined influence of water generated from the RWGS reaction and CO₂^[36]. Furthermore, a diffraction peak at 25.9° corresponding to Graphite-2H is detected, originating from graphitic carbon formed during the DRM process^[37]. Compared with the 12.5%Ni catalyst, the modified catalysts exhibit much weaker Graphite-2H diffraction peaks. For the 12.5%Ni7%Ce sample, this peak is nearly invisible, further demonstrating the significant suppression effect of the promoters on carbon deposition.

Table 1 summarizes the key elemental contents and pore structure parameters of the four catalysts after calcination. The actual Ni loadings were determined to be 10.25% for 12.5%Ni, 10.56% for 12.5%Ni7%Zr, 10.68% for 12.5%Ni7%La and 10.25% for 12.5%Ni7%Ce. While the actual contents of the Zr, La and Ce promoters were 5.58%, 5.67% and 5.67%, respectively. All measured values are slightly lower than their corresponding theoretical values (Ni: 12.5%; promoters: 7%), which is mainly attributed to the partial incorporation of Ni²⁺ and promoter cations into the MgO or MgAl₂O₄ lattices during calcination at 750 °C, leading to the formation of stable NiO-MgO solid solutions and NiAl₂O₄ spinel phases. The introduction of promoters effectively regulates the dispersion of the active metal and the interfacial structure. The specific surface areas of the 12.5%Ni, 12.5%Ni7%Zr, 12.5%Ni7%La and 12.5%Ni7%Ce catalysts are 57.46, 60.62, 60.49 and 61.98 m²/g, respectively. Figure 5 presents the SEM-EDS images and elemental mapping of the calcined catalysts. As can be observed, the unmodified 12.5%Ni catalyst exhibits a relatively dense surface with small crystallites and a limited number of pores, whereas the Zr-, La- and Ce-modified catalysts display uniformly distributed particles with clear grain boundaries and well-developed porous structure. EDS and elemental mapping analyses further reveal the compositional distribution characteristics of the catalysts: Ni species in the unmodified catalyst appear relatively aggregated, while those in the Zr-, La- and Ce-modified samples are uniformly dispersed and form embedded interfacial structures with the promoters on the support surface, thereby constructing abundant and homogeneous Ni-X (X = Zr, La, Ce) interfaces. This structural feature is in good agreement with the XRD results, confirming that the incorporation of Zr, La and Ce effectively enhances the Ni dispersion.

To further elucidate the morphology and dispersion of metal particles in the reduced catalysts, transmission electron microscopy (TEM) images are recorded (Figure 6). Dark particles composed of Ni or Ni-X (X = Zr, La, Ce) species can be clearly observed in the reduced catalysts. Particle size distributions obtained using Nano Measurer 1.2 software reveal that the sizes of Ni or Ni-X particles follow a normal distribution and exhibit systematic variations depending on the type of the promoter introduced.

The unmodified catalyst exhibits an average particle size of 8.41 ± 0.98 nm, whereas the average particle sizes of the 12.5%Ni7%Ce, 12.5%Ni7%La, and 12.5%Ni7%Zr catalysts decrease to 6.23 ± 0.82, 6.52 ± 0.98 and 6.72 ± 1.12 nm, respectively. These results indicate that the promoters can form stable interfaces with Ni, effectively restricting metal migration and aggregation during the calcination and reduction.

Furthermore, the carbon deposition morphology and changes in metal particle size of the spent catalysts are investigated by TEM analysis (Figure 7). After 20 h of reaction, both filamentous carbon and encapsulating carbon are observed on the catalyst surfaces. Filamentous carbon has relatively minor impact on catalytic activity, whereas encapsulating carbon blocks metal active sites and leads to catalyst deactivation. The unmodified 12.5%Ni catalyst generated significantly larger amounts of filamentous and encapsulating carbon than the Zr-, La- and Ce-modified catalysts, indicating that the construction of Ni-X (X = Zr, La, Ce) interfaces effectively suppresses carbon deposition, which is consistent with the XRD results.

Statistical analysis of the metal particle sizes after reaction reveals that they still follow a normal distribution. After 20 h of reaction, the average particle sizes increase to 13.39 ± 2.75, 8.07 ± 1.22, 7.95 ± 1.72 and 10.69 ± 3.76 nm, respectively, representing an increase of 1.43−4.98 nm compared with those of the reduced catalysts. The results indicate that varying degrees of metal migration, aggregation, and sintering occur during the DRM process, with the unmodified catalyst exhibiting the most severe sintering. In contrast, the Ni-X interfaces can partially inhibit excessive growth of metal particles during reaction, thereby enhancing the structural stability of the catalysts.

Figure 8 presents the O 1s XPS spectra of the reduced 12.5%Ni catalyst and its Ce-, La- and Zr-modified counterparts. The O 1s peak can be deconvoluted into lattice oxygen (O_α, 530.0−530.5 eV) and adsorbed oxygen/surface −OH species (O_β, 531.8−532.0 eV)^[38]. Because O_β species exhibit higher mobility and reactivity, their relative abundance is generally positively correlated with the concentration of surface oxygen vacancies, thereby influencing the adsorption and activation of CO₂ during the DRM reaction^[39]. Based on the integrated peak areas, the unmodified 12.5%Ni catalyst shows the lowest O_β content of 46.5%, while the Ce-modified sample exhibits the highest content of 67.6%. The La- and Zr-modified catalysts display the O_β contents of approximately 60%. The results indicate that the introduction of Ce, La and Zr effectively regulates the structure and reactivity of surface oxygen species.

Figure 9 present the Ni 2p_3/2 XPS spectra of the calcined, reduced, and spent catalysts. In the calcined samples (Figure 9(a)), the characteristic peak at ~856.0 eV accompanied by satellite features is assigned to Ni²⁺ species, mainly originating from NiO and Ni²⁺ strongly interacting with the MgO-MgAl₂O₄ support. Upon modification, the Ni 2p_3/2 binding energies shift overall toward higher values, indicating that Ce, La and Zr possess high oxygen affinity and can induce the formation of oxygen vacancies and redistribution of oxygen charge density. This effect strengthens the Ni-support interaction and places Ni in a more electropositive environment.

In the reduced catalysts (Figure 9(b)), the 12.5%Ni sample exhibits a typical metallic Ni⁰ peak at ~852.0 eV, while residual Ni²⁺ signals associated with hardly reducible Ni species in the NiAl₂O₄ structure are still observed. In contrast, no distinct peak appears in the Ni⁰ region for the Ce- and Zr-modified samples. This phenomenon is not due to stronger anchoring effects that hinder NiO reduction, but is more plausibly attributed to two other reasons: (i) Ce/Zr modification enhances NiO dispersion, leading to extremely finely divided Ni⁰ particles with weakened XPS signals; and (ii) strong electronic interactions between Ni and Ce (or Zr) alter the electron density of Ni⁰, shifting its characteristic peak toward higher binding energies (approximately 853.5−855 eV). This interpretation is corroborated by the weak Ni⁰ diffraction peaks observed in the XRD patterns of reduced samples.

Figure 9(c) shows the Ni 2p_3/2 spectra of the spent samples. For all catalysts, both Ni⁰ and Ni²⁺ signals are markedly attenuated, mainly due to: (i) high-temperature-induced Ni sintering or inward migration, which decreases the surface Ni concentration; and (ii) coverage by carbon deposits that limits XPS detection. Nevertheless, the intensified Ni⁰ peaks in XRD indicates the growth of bulk Ni⁰ particles and increased reduction degree, revealing surface-bulk structural reconstruction under prolonged reaction conditions. Notably, the Ce- and La-modified catalysts retain relatively higher Ni 2p_3/2 intensities, suggesting that the modifiers more effectively suppress Ni sintering and carbon deposition, thereby imparting superior structural stability. For the La-modified catalyst (Figure 9(d)), the peak at ~851.5 eV in the calcined sample arises from overlap between the Ni 2p_3/2 and La 3d_3/2 signals. After reduction, the peak shifts to ~852.5 eV, which can be attributed to the combined contribution of metallic Ni⁰ formation and La 3d signals.

The reducibility of Ni²⁺ species in the calcined catalysts was investigated by H₂-TPR (Figure 10(a)). All samples exhibit two characteristic reduction features. The weak peak in the range of 450−500 °C can be attributed to NiO species with moderate interaction with the support, whereas the broad high-temperature peak between 550 and 850 °C arises from the sequential reduction of highly dispersed NiO (600−620 °C) and NiAl₂O₄ spinel (710−740 °C). The introduction of Ce and La have limited influence on the reduction peak positions, while Zr modification markedly shifts the high-temperature reduction peak to lower temperatures by approximately 15 °C. This indicates that the superior oxygen ion mobility of ZrO₂ and its ability to modulate the Ni-support interaction facilitate the reduction of NiO.

Based on H₂ consumption, the reduction degrees of the catalysts range from 45% to 55%, with the Ce- and La-modified samples exhibiting the lowest values, suggesting incomplete reduction of Ni²⁺ species. This is mainly attributed to the relatively low reduction temperature employed (650 °C) and the presence of the hardly reducible NiAl₂O₄ phase.

The basic properties of the catalysts are characterized by CO₂-TPD (Figure 10(b)). All samples display two CO₂ desorption peaks: the peak at 200−300 °C corresponds to medium-strength basic sites, primarily associated with metal-O²⁻ groups, while the peak at 450−480 °C is assigned to strong basic sites related to low-coordinated lattice oxygen. Compared with the unmodified 12.5%Ni catalyst, Ce-, La- and Zr-modified samples show desorption peaks shifted to lower temperatures to varying extents, particularly for the medium-strength basic sites, indicating that the promoters reduce the surface basic strength and modify the distribution of basic sites. It has been reported that rare-earth oxides can enhance the basic characteristics of MgO or MgAl₂O₄ supports by increasing oxygen coordination and surface oxygen vacancy concentration, which is consistent with the results in this study^[40]. Table 2 summarizes the desorption peak positions, basic site contents, and their relative proportions for the reduced catalysts.

In the 12.5%Ni sample, strong basic sites dominate (62.7%), with a total basicity of 1.012 mmol/g. The incorporation of Ce, La or Zr significantly increases the proportion of medium-strength basic sites and further shifts the desorption peaks to lower temperatures. Specifically, for 12.5%Ni7%Ce, the proportion of medium-strength basic sites increases to 71.5%, the peak temperature decreases by approximately 70 °C, and the total basicity rises to 1.281 mmol/g. The 12.5%Ni7%La catalyst exhibits the most pronounced increase in total basicity (1.647 mmol/g), with the medium-strength basic site content reaching 1.093 mmol/g. In contrast, 12.5%Ni7%Zr shows a slightly lower total basicity (0.930 mmol/g), while the proportion of medium-strength basic sites increases to 65.0%, indicating that Zr modification mainly enhances CO₂ adsorption by improving oxygen mobility and Ni dispersion rather than increasing the total number of basic sites. Overall, the introduction of Ce, La and Zr favors the formation of a higher fraction of medium-strength basic sites, which is closely related to promoter-induced oxygen vacancy formation and local structural modulation. The oxygen storage capacity of CeO₂ and the intrinsically strong basicity of La₂O₃ further enhance the total surface basicity, whereas Zr doping predominantly strengthens CO₂ adsorption and activation through regulation of oxygen mobility and Ni-support interactions.

Although the four catalysts exhibit similar macroscopic deactivation behaviors during 20 h of reaction at 750 °C, their microscopic structural stability and carbon deposition characteristics may differ significantly. To verify this assumption, the carbon deposition behavior of the spent catalysts was systematically analyzed (Figure 11). The TG curves of all catalysts show three typical regions: the low-temperature weight loss below 200 °C is mainly attributed to the desorption of physically adsorbed water and gases; the intermediate-temperature weight gain at 300-400 °C originates from the oxidation of Ni⁰; and the high-temperature weight loss between 400 and 700 °C corresponds to the oxidative decomposition of surface carbon species. Therefore, the weight loss in the high-temperature region is used to quantitatively evaluate the amount of carbon deposition.

Within the investigated temperature range, the 12.5%Ni catalyst exhibits the most severe carbon deposition, with the carbon content reaching as high as 32% at 650 °C. As the reaction temperature increased, carbon deposition decreased markedly, dropping to 14.62% at 750 °C. The suppression of carbon deposition at elevated temperatures can mainly be attributed to: (i) higher temperatures promoting the DRM reaction while thermodynamically suppressing side reactions such as CH₄ cracking and the Boudouard reaction; and (ii) accelerated oxidation gasification and surface migration of carbon species at high temperatures, which kinetically reduce the carbon formation rate. Compared with the benchmark catalyst, the 12.5%Ni7%Ce catalyst exhibits the best resistance to carbon deposition in the temperature range of 650−750 °C, with carbon contents maintained at 10%−12%. This behavior is primarily attributed to the excellent redox properties and abundant oxygen vacancies of CeO₂, which continuously promote the oxidation of carbon species. However, at 800 °C, the anti-coking performance of the catalyst declines, possibly due to the sintering or phase evolution of CeO₂, leading to reduced oxygen mobility, together with further sintering of Ni particles that promotes the CH₄ cracking and thus enhances carbon formation. The carbon content of the 12.5%Ni7%La catalyst is maintained at 15%−18% after reaction, exhibiting an overall moderate anti-coking performance, with the lowest carbon deposition observed at 750 °C. The optimal performance may originate from the La promoter enhancing surface basicity and improving CO₂ adsorption, while the synergistic conversion efficiency of CH₄ and CO₂ and the oxidation rate of carbon species reach an optimal balance. For the 12.5%Ni7%Zr catalyst, the carbon content continuously decreases with increasing temperature, from 19.37% at 650 °C to 10.85% at 800 °C. This trend is closely related to the enhanced oxygen mobility and oxygen storage capacity of ZrO₂ at high temperatures, which promote CO₂ activation and in situ oxidation of surface carbon species. In addition, elevated temperatures can accelerate the carbon-oxygen cycle on Ni sites, further improving the overall resistance to carbon deposition.

Raman analysis is conducted to identify the surface carbon species on the catalysts after 20 h of reaction at 750 °C (Figure 12). All samples exhibited a D band at approximately 1350 cm⁻¹ (disordered carbon) and a G band at approximately 1580 cm⁻¹ (graphitic carbon). The D/G intensity ratio (I_D/I_G) is commonly used to characterize the degree of graphitization of carbon species, with higher values indicating more disordered carbon that is easier to oxidize and remove. The I_D/I_G values of the four catalysts follow the order: 12.5%Ni7%La (1.21) > 12.5%Ni7%Ce (1.20) > 12.5%Ni7%Zr (1.17) > 12.5%Ni (1.06).

The unmodified Ni catalyst exhibits the lowest I_D/I_G value, indicating the highest degree of carbon graphitization, which is consistent with its most severe carbon deposition observed in the TG analysis. In contrast, the La- and Ce-modified samples possess the most disordered carbon structures, further explaining their superior anti-coking performance. Combined with SEM/TEM observations, fibrous carbon is identified as the dominant carbon species formed on the catalyst surface after reaction. Such carbon readily covers active sites and destroys the pore structure, constituting one of the main causes of catalyst deactivation.

Overall, the results demonstrate that Ce, La and Zr modification alters the dynamic balance between carbon formation and removal, enabling carbon to exist in a more dispersed and readily oxidizable form, while effectively suppressing excessive growth of Ni particles during reaction. From a structural perspective, this significantly reduces the potential risk of catalyst deactivation.

Figure 13 presents the XPS spectra of Ce 3d, La 3d_5/2, and Zr 3d on the catalyst surfaces before and after reduction. Overall, after 20 h of reaction, the XPS signals of the elements are significantly weakened and exhibit a certain degree of spectral roughness. This is mainly attributed to carbon deposition, migration of metal or promoter species, surface structural reconstruction, and charge accumulation during the reaction, which collectively reduce the effective surface concentration of the elements and increase the heterogeneity of the chemical environment, thereby deteriorating the quality of the photoelectron signals.

Figure 13(a) shows the Ce 3d XPS spectra of the catalysts after reduction and reaction. Typical Ce⁴⁺ features (3d_5/2: V, V2, V3; 3d_3/2: U, U2, U3) and Ce³⁺ features (V1, U1) can be clearly identified. Deconvolution analysis indicates that the Ce⁴⁺/(Ce⁴⁺+Ce³⁺) fraction decreases from 0.774 in the reduced sample to 0.686 after reaction, indicating a decrease in the proportion of Ce⁴⁺ and a corresponding increase in Ce³⁺. This variation suggests that CeO_x participates in O²⁻ migration and release during the reaction. The reversible Ce⁴⁺/Ce³⁺ redox cycle provides surface active oxygen species, which are beneficial for in situ oxidation and removal of carbon deposits. Moreover, the Ce 3d spectra shift toward higher binding energy after reaction, further implying changes in the surface electronic environment, possibly associated with an increased concentration of oxygen vacancies and reduced electron density.

Combined with the CO₂-TPD results, it can be inferred that Ce incorporation increases the number of medium-strength basic sites and enhances oxygen storage-release capability, thereby significantly promoting CO₂ activation and carbon-oxygen reactions, which accounts for the excellent anti-coking performance of the 12.5%Ni7%Ce catalyst.

Figure 13(b) and 13(c) display the La 3d_5/2 and Zr 3d spectra of the catalysts before and after reaction, respectively. The peaks located at 838.8±0.5 and 835.5±0.5 eV correspond to the characteristic signals of La³⁺. Peak fitting results show that the relative content of La³⁺ is 56.82% after reduction and 55.76% after reaction, indicating only a slight change and suggesting that La maintains a relatively stable chemical valence during the reaction. Similarly, both the binding energies and peak areas of Zr 3d_3/2 and 3d_5/2 exhibit negligible changes before and after reaction, confirming that Zr species are chemically stable. The results demonstrate that La and Zr mainly act as structural modifiers, with their oxidation states remaining essentially unchanged under reaction conditions. Although XPS indicates stable chemical states for La and Zr, the CO₂-TPD results reveal that their incorporation significantly increases the proportion of medium-strength basic sites and shifts the desorption peaks toward lower temperatures, indicating enhanced CO₂ adsorption and activation. This enhancement may originate from lattice distortion of the support induced by La or Zr incorporation, leading to the formation of new surface defects or oxide interfacial structures, which optimize CO₂ adsorption-conversion behavior. The increased basic sites and improved CO₂ activation capability are conducive to promoting the oxidation and removal of carbon deposits.