Role of Perovskite Phase in CeXO₃ (X = Ni, Co, Fe) Catalysts for Low-Temperature Hydrogen Production from Ammonia

Abstract

The drive to utilize ammonia as a carbon-free hydrogen source necessitates the development of effective, non-precious metal catalysts for ammonia decomposition. We successfully synthesized a series of Ce-based perovskite oxides (CeXO₃; X = Co, Ni, Fe) via combustion method using citric acid. These catalyst precursors were tested for NH₃ decomposition to study the effect of the perovskite structure on catalytic activity. The results were directly compared to corresponding impregnated catalysts, X/CeO₂, which had similar metal concentrations. A remarkable enhancement in catalytic performance was observed with the perovskite catalysts, particularly at lower temperatures, relative to their impregnated counterparts. The exception was the CeFeO₃ catalyst, which exhibited lower activity, likely due to the formation of metal nitrides. Both CeNiO₃ and CeCoO₃ showed good NH₃ decomposition activity, but CeNiO₃ emerged as the most active catalyst at lower temperatures. This superior performance attributed to the presence of oxygen vacancies—confirmed by Raman and XPS analyses—and enhanced metal reducibility at lower temperatures, both of which accelerate NH₃ decomposition. Furthermore, CeNiO₃ also displayed a high surface metal concentration. These Ce-based perovskite materials are cost-effective, easily synthesized, and highly stable; hence, they are attractive candidates for large-scale hydrogen production.

1. Introduction

Hydrogen (H₂) is expected to play a pivotal role in the future of sustainable energy. Nevertheless, effective storage and transportation remain critical challenges that must be addressed for the advancement of the hydrogen energy industry. Among various hydrogen carriers, ammonia stands out as one of the most promising options due to the following advantages: low-cost production, high energy density (4 kWh kg⁻¹), ease of liquefaction for storage and transportation (liquefies at 8 bars at room temperature), and carbon-free participation throughout the entire process. The production of ammonia is a proven technology that allows for economical, large-scale synthesis and benefits from established infrastructure for its liquid storage and transportation [1,2,3,4].

So far, a wide variety of metal catalysts have been explored for ammonia decomposition, such as ruthenium (Ru), nickel (Ni), cobalt (Co), iron (Fe), and copper (Cu) [5,6,7,8,9,10,11,12,13]. Noble metal ruthenium (Ru) catalysts such as Ru/CNTs [14], Ru/MgO [15], and Ru/CeO₂ [16] exhibit outstanding catalytic activity even at low temperatures. However, their high cost and limited availability hinder large-scale industrial applications. Consequently, significant efforts are being directed toward developing non-noble metal catalysts for practical use. Among non-noble metals, cobalt (Co), nickel (Ni), and iron (Fe) are extensively studied due to their low price and comparatively good catalytic activity for ammonia decomposition [2,10,17,18].

On the other hand, the selection of catalyst support is undeniably a key factor in catalytic processes. A variety of materials are used as catalyst supports for H₂ production from NH₃, including metal oxides (MgO, SiO₂, Al₂O₃, TiO₂, ZrO₂), various forms of carbon (such as activated carbon and multi-walled carbon nanotubes or MWCNTs), and both mesoporous and microporous materials [14,19,20,21,22,23,24,25]. There is a growing interest in employing cerium-based materials as prominent catalyst supports [2,7,26]. Ceria-based catalysts boast distinctive qualities, including oxygen storage capacity and effective dispersion of surface-active components. These properties together enable their exceptional catalytic performance, particularly at low temperatures, in a wide range of oxidation reactions [26,27,28].

In addition to the choice of active metal and support, the method of catalyst preparation plays a crucial role in determining the chemical and morphological heterogeneity of the resulting material, which directly influences catalytic performance. Conventional catalyst synthesis methods, such as impregnation or co-precipitation, often lead to incomplete utilization of the support surface area due to pore blockage and limited dispersion of the active phase. Recent advancements in synthesis techniques, such as high-intensity power (HIP) processing, microwave-assisted synthesis, and plasma-assisted activation, have emerged as effective non-equilibrium methods to overcome these limitations [29,30,31,32]. These methods enable the formation of catalysts with extensive heterogeneity featuring enhanced oxygen vacancies, increased porosity and surface area, and variable catalyst/support or oxygen/catalyst ratios within the porous framework. Such heterogeneity can even induce quantum effects in certain nanostructures, leading to superior catalytic performance. Furthermore, to preserve these heterogeneous structures, extensive post-synthesis heat treatments should be minimized, as they can diminish the unique non-equilibrium morphology and associated activity [31,32].

Perovskite-type mixed oxides have garnered considerable attention as efficient catalysts for ammonia decomposition due to their structural stability and reversible oxygen storage capacity. In the perovskite structure (ABO₃), the larger A-site cation occupies the corners of the octahedron, while the smaller B-site cation resides at its center. This configuration endows the material with excellent thermal stability and tunable catalytic properties, making it highly suitable for high-temperature reactions [33]. Typically, lanthanides or rare-earth metals (La, Ce, Nd, Sm, etc.) occupy the A site, whereas transition metals (Fe, Ni, Co, etc.) are incorporated at the B site [34,35,36]. After activation and exsolution, the formed B⁰ metal particles and AxOy alkaline metal oxides, respectively, provide active metal and support for NH₃ decomposition reaction. Pinzon et al. [37] prepared LaNiO₃ perovskite catalysts and studied ammonia decomposition. They also studied effect dopants, such as Ce and Mg in LaNiO₃ perovskite, and found enhanced catalytic activity. Our group also reported the influence of 50% Ce substitution in LaXO₃ (X = Ni or Co) perovskites in catalyst activity enhancement for ammonia decomposition [38]. Osama et al. [33] studied cobalt catalysts supported on perovskite-type XCeO₃ (X: Mg, Ca, Sr, Ba) oxides and reported that these catalysts are active for ammonia decomposition. Based on the above considerations, it is interesting to study the Ce-based catalysts, which possess a perovskite structure for ammonia decomposition.

In this paper, we describe the synthesis of CeFeO₃, CeCoO₃, and CeNiO₃ perovskite catalysts. They were successfully synthesized via the solution combustion method, utilizing citric acid (CitAc) as the fuel source. The synthesized materials were used as catalyst for ammonia decomposition for H₂ production. The effect of X site atom in CeXO₃ (X = Co, Fe, Ni) perovskites on the catalyst performance was systematically investigated.

To assess the effect of the perovskite structure on catalyst performance, we prepared another series of X/CeO₂ (X = Co, Fe, Ni)-supported catalysts by impregnation and tested for NH₃ decomposition. A thorough literature search suggests that the use of CeXO₃ perovskite catalysts for environmentally friendly hydrogen production from ammonia has not yet been reported.

2. Results and Discussion

N₂ adsorption–desorption experiments were carried out to determine the textural properties of the catalysts (Figure 1).

Table 1. BET surface area, average pore volume, and pore width of the studied catalysts.

Catalyst	BET	Pore volume	Pore width
	(m2 g−1)	(cc g−1)	(Å)
CeO2	73	0.254	78
CeNiO3	29	0.09	48
CeCoO3	24	0.07	40
CeFeO3	14	0.047	40
Ni/CeO2	44	0.147	78
Co/CeO2	39	0.135	78
Fe/CeO2	40	0.148	78

presents the key findings, including the BET surface area (S_BET), total pore volume (V_total), and average pore width. The corresponding isotherms and pore-size distribution for the bare CeO₂ support are provided in the supplementary material (Figure S1). The CeO₂ support showed type IV isotherm with H1 type hysteresis indicates mesoporous material possessing well defined pores. All catalysts showed lower surface area than the pure CeO₂ support (Table 1). Figure 1a shows N₂ isotherms of all perovskite catalysts. According to the IUPAC classification, the isotherms in Figure 1a belong to the type IV isotherms with H₃ hysteresis loops that indicate the presence of mesoporous [39]. The CeNiO₃ catalyst showed the highest surface area value among the other perovskite catalyst. The surface area of CeCoO₃ catalyst is slightly lower than that of the CeNiO₃ catalyst, where the CeFeO₃ catalyst showed the lowest surface area among all catalysts. The order of surface area for perovskite catalysts is as follows: CeNiO₃ > CeCoO₃ > CeFeO₃.

On the other hand, the variation in surface area of impregnated catalysts is low; however, the Ni/CeO₂ showed slightly higher surface area value than the Co/CeO₂ and Fe/CeO₂ catalysts. All M/CeO₂ (M: Ni, Co, Fe) catalysts showed type IV isotherms with H1 type hysteresis, indicating mesoporous materials (Figure 1b).

The hysteresis loop in the N₂ adsorption–desorption isotherm for the Co/CeO₂ catalyst appeared at relative pressures (P/P₀) between 0.5 and 1.0. In contrast, the loops for Ni/CeO₂ and Fe/CeO₂ were shifted to a higher P/P₀ range (0.7 to 1.0), which also corresponded to an increased total pore-volume in these catalysts. The results clearly indicate presence of pores in the lower meso-pore region in Co/CeO₂ catalysts. The results clearly showed no big change observed in catalysts prepared by impregnation process.

Figure 2 shows the histograms of pore size distribution pattern of all catalysts. The textural properties of the synthesized perovskite materials indicated that changing the metal (Co, Ni, Fe) in CeMO₃ slightly altered the pore size distributions within the meso range. CeFeO₃ exhibited a pore size distribution in the 20–50 Å range, indicating low porosity with a small pore volume.

The average pore width of all impregnated catalysts is the same as that of its support (Table 1). The pore size distribution of the CeO₂ support was in the range of 50–150 Å (Figure S1b). The Ni/CeO₂ catalyst showed a similar pore size distribution as that of its support, whereas the Co/CeO₂ and Fe/CeO₂ catalysts showed a pore size distribution in the range of 10–150 Å. The results indicate an increase in micropores in cobalt and iron catalysts, which might be due to partial pore blockage.

2.1. XRD Studies

The XRD patterns of all the catalysts are shown in Figure 3 and Figure 4. The “MATCH” software 1.1 was used to analyze the XRD data with PDF database. The synthesized CeXO₃ catalysts exhibited well-defined peaks in their X-ray diffraction (XRD) patterns (Figure 3), a clear indication of their high crystallinity. Furthermore, these diffraction patterns closely matched those previously reported in the literature for perovskite-structured CeMO₃ [40,41]. The characteristic peaks of metal oxide such as Fe₂O₃ [PDF:00-033-0664], Co₃O₄ [PDF:03-065-3103], and NiO [PDF:00-047-1049] were observed with very low intensity in the CeFeO₃, CeCoO₃, and CeNiO₃ catalysts, respectively. This might be due to low concentration of metal oxide or low crystalline nature of metal oxides. Overall, the XRD results confirm the successful formation of the intended perovskite catalysts. The average crystal size of the perovskite phase in each catalyst was calculated using the Scherrer equation based on the dominant reflection peak observed at a 2θ value of approximately 28 degrees. The resulting crystal sizes are summarized in Table S1 (Supporting Information). Among the perovskite catalysts, CeFeO₃ displayed the smallest crystal size, while the CeCoO₃ catalyst exhibited the largest crystal size.

The highly crystalline nature of the Ni/CeO₂ catalyst is clearly evidenced by the intensity of the diffraction peaks in its XRD pattern in Figure 4. The pattern contains peaks identifying the CeO₂ phase (2θ = 33.2°, 38.61°, 55.74°, and 66.48°) and peaks identifying the NiO phase (2θ = 43.54°, 50.72°, 74.53° and 74.60°). The phases were matched to the respective PDF cards: 00-034-0394 for CeO₂ and 00-044-1159 for NiO. In Fe/CeO₂ catalyst along with CeO₂ phase observed peaks at 2θ values of 33.18, 35.64, 49.52, and 54.14 corresponds to the Fe₂O₃ phase [PDF: 00-033-0664]. Similarly, the Co/CeO₂ catalyst exhibited both the CeO₂ and Co₃O₄ phases.

2.2. H₂-TPR Studies

The reducibility of the CeXO₃ perovskite catalysts (where X is the B-site cation) was analyzed using temperature-programmed reduction (TPR). The resulting profiles varied significantly based on the B-site cation. Notably, the CeNiO₃ catalyst exhibited two slightly overlapping reduction peaks between 240 °C and 420 °C. These two peaks are indicative of a two-step reduction process for the nickel (Ni) species within the CeNiO₃ structure, as detailed by the following equations [41]:

CeNiO₃ + H₂→NiO + CeO₂,

(1)

NiO + H₂→Ni + H₂O.

(2)

For the CeCoO₃ catalyst, the reduction peaks are present at a temperature range between 210 °C and 480 °C, which corresponds to the stepwise reduction processes of Co³⁺ → Co²⁺ and Co²⁺ → Co⁰, respectively [42]. It seems that there are some non-interacted Co₃O₄ species present in the CeCoO₃ catalyst. According to the literature, the reduction profiles of the CeFeO₃ samples included three steps at 300–440 °C (peak1), 511–650 °C (peak2), and above 750 °C (peak3), which were not observed in Figure 5 [43,44]. These peaks correspond to the sequential reduction steps of Fe₂O₃ → Fe₃O₄, Fe₃O₄ → FeO, and FeO → Fe.

The reducibility of the impregnated catalysts (X/CeO₂) is illustrated by their profiles in Figure 5. For reference, the H₂-TPR profile of pure CeO₂ support is presented in the Supplementary Material (Figure S2). As established in the literature, pure CeO₂ exhibits two characteristic reduction peaks: one near 480 °C, attributed to the reduction in surface oxygen, and a second peak above 700 °C, corresponding to the reduction in bulk oxygen within the ceria lattice [41,45]. From Figure S2, it is clear that the CeO₂ reduction is very low. In the case of the Fe/CeO₂ catalysts, the H₂-TPR profile showed two reduction peaks: one at 365 °C, and the other at 600 °C, which corresponds to the reduction of iron species, i.e., Fe₂O₃ → Fe₃O₄ and Fe₃O₄ → Fe.

The literature indicates that the reduction in Co₃O₄/CeO₂ is strongly governed by three factors: the overall catalyst composition, the specific synthesis method employed, and the degree of cobalt oxide dispersion [46,47]. The smaller Co₃O₄ crystallites in the catalyst shows that the three distinct peaks at different temperatures correspond to a gradual reduction in (1) Co₃O₄ to CoO; (2) reduction in weakly interacted CoO with CeO₂; and (3) reduction in strongly interacted cobalt oxide with CeO₂ support, which hinders their reduction due to migration of oxygen ions from ceria to cobalt. Meanwhile, larger Co₃O₄ crystallites in the catalyst leads to an almost direct reduction in Co₃O₄ to Co⁰ [47]. The H₂-TPR profile for the Co/CeO₂ catalyst (Figure 5) featured the main reduction peak at 366 °C, preceded by a shoulder at 318 °C. This suggests that cobalt oxides weakly bound to the CeO₂ support are reduced first, forming metallic cobalt (Co⁰). This metallic cobalt then facilitates the reduction in the remaining cobalt species through an H₂ spillover effect, which effectively shifts the overall reduction to lower temperatures.

According to the literature, the reduction behavior of nickel-based catalysts supported on CeO₂ depends on the strength of the interaction between NiO and CeO₂ [28,46,48]. Stronger interaction results in a reduction at higher temperatures. TPR peaks below 200 °C are ascribed to the reduction in surface NiO, which does not interact with the support. In contrast, the sharp and intense peak below 400 °C corresponds to the reduction in the NiO phase weakly interacting with ceria. The broad peak above 400 °C is attributed to the reduction in well-dispersed NiO that interacts strongly with ceria [41,49]. From the TPR results, it is clearly inferred that the Co-based and Ni-based catalysts reflected easier reduction in their surface high-valent metal oxides under low-temperature conditions compared to Fe-based catalysts.

2.3. CO₂–TPD Studies

The surface basicity of CeMO₃ (M = Co, Ni, Fe) catalysts and M/CeO₂ (M = Co, Ni, Fe) were investigated by CO₂-TPD, and the results are shown in Figure 6. The CO₂ desorption profiles exhibited similar peak shapes but differed in intensity. In general, desorption peaks below 300 °C corresponded to weak basic sites, which were associated with surface hydroxyl groups, while desorption peaks between 300 and 500 °C were attributed to medium basic sites.

The peaks above 500 °C corresponded to strong basic sites, which were related to the low level of coordinated surface oxygen (O²⁻) [50]. It is clearly understood from Figure 6 that all catalysts possess weak and moderate basic sites. The Ni-based catalysts showed higher basicity compared to Fe-based and Co-based catalysts. According to the literature, moderate basic sites are favorable to ammonia decomposition reaction [2,38]. The CO₂ TPD results of pure CeO₂ support are presented in Figure S3. The CeO₂ support also showed weak and moderate basic sites. Therefore, the results clearly indicate that the basic sites are not much changed either in the catalysts or the impregnation catalysts.

2.4. Raman Studies

Raman spectroscopy is a powerful tool for probing chemical bonds and molecular symmetry. Figure 7 presents the vibrational modes of perovskite nanomaterials as revealed by their Raman spectra. According to the literature, CeO₂ exhibits a characteristic Raman band at 460 cm⁻¹, which corresponds to the first-order F2g symmetric stretching of oxygen atoms surrounding cerium ions within the fluorite lattice of CeO₂ [51,52,53].

In comparison, CeCoO₃ displays three distinct Raman bands at 186, 474, and 656 cm⁻¹, whereas Co₃O₄ exhibits major peaks at 194, 481, and 686 cm⁻¹. These bands in Co₃O₄ are attributed to the F2g, Eg, and A1g vibrational modes, respectively. The A1g mode originates from vibrations at octahedral sites, whereas the F2g and Eg modes arise from a combination of tetrahedral-site and octahedral oxygen motions [54,55]. Notably, a systematic downshift of Raman bands is observed when comparing Co₃O₄ to CeCoO₃: from 686 to 656 cm⁻¹, 481 to 474 cm⁻¹, and 194 to 186 cm⁻¹. These downshifts suggest the successful incorporation of cobalt species into the ceria lattice and, importantly, the generation of oxygen vacancies [56].

The Raman spectrum of CeNiO₃ reveals a peak at 460 cm⁻¹, corresponding to the F2g symmetric stretching of oxygen atoms around cerium ions, along with a broad weak band spanning 497–718 cm⁻¹. This broader feature may originate from Ni–O vibrations, particularly a one-phonon longitudinal optical (LO) mode at 560 cm⁻¹ and a two-phonon overtone (2TO) mode at 740 cm⁻¹ [57]. This feature is associated with oxygen vacancies, indicating the presence of a certain amount of oxygen vacancies. Similarly, CeFeO₃ shows a prominent Raman band at 457 cm⁻¹ with a low intense band around 680 cm⁻¹. These results closely resemble that of CeNiO₃.

The Raman spectra of metal-impregnated catalysts are shown in Figure S4. The Co/CeO₂ catalyst displays features similar to those of CeCoO₃, with contributions from both CeO₂ and Co₂O₃. In contrast, the Fe/CeO₂ catalyst exhibits a markedly different spectral pattern compared to CeFeO₃. According to the literature, Fe₂O₃ exhibits four bands at 222, 288, 407, and 605 cm⁻¹, respectively. The Fe/CeO₂ catalyst shows bands associated with Fe₂O₃ and CeO₂. Finally, the Ni/CeO₂ catalyst shows a strong band at 466 cm⁻¹, corresponding to the F2g mode of CeO₂. The absence of bands in the 550–650 cm⁻¹ region indicates that nickel cations do not substitute cerium cations, and no Ce_1−xNi_xO_γ solid solution is formed, which indicates a lack of oxygen vacancies [58].

2.5. XPS Studies

X-ray photoelectron spectroscopy (XPS) was employed on all calcined catalysts to provide a detailed surface analysis. This technique allowed for the investigation of the chemical environment and oxidation state of the transition metal and oxygen components on the catalysts’ surface. The core level spectra of the Ce 3d spectrum and the O1s spectrum for all synthesized perovskite catalysts are shown in Figure 8. The Ce 3d spectrum (in Figure 8A) consists majorly of eight characteristic peaks located at 916.5, 907.4, 903.2, 207, 900.9, 898.3, 889.0, 885.3, and, 882.5 eV, which are labeled as u‴, u″, u′, u, v‴, v″, v′, and v, respectively. Among these, the u′ and v′ peaks were assigned to the Ce³⁺ species, while the remaining peaks corresponded to the Ce⁴⁺ species. Studies have shown that the Ce³⁺ species were generally associated with surface oxygen vacancies, structural defects, and unsaturated chemical bonds [51,59,60].

The O1s spectra obtained via XPS for the calcined perovskite catalysts are presented in Figure 8B. The observed broad signal in the O1s region was subjected to deconvolution, revealing three primary components. These components are chemically assigned based on their binding energies: the peak around 529.0 eV corresponds to lattice oxygen (O²⁻) within the perovskite structure. A second peak, centered near 530.6 eV, is attributed to surface-adsorbed oxygen species (O⁻/O²⁻). Finally, the high-energy component at approximately 532.0 eV indicates the presence of surface hydroxyl (OH⁻) or carbonate (CO₃²⁻) groups.

Figure 8B clearly demonstrates that, across all perovskite catalysts analyzed, the concentration of lattice oxygen (O²⁻) on the surface is lower than that of the adsorbed oxygen species (O⁻/O²⁻). The surface-adsorbed oxygen species are high in CeNiO₃ compared to other perovskite catalyst (

Table 2. XPS surface atomic concentrations.

Catalysts	Surface Atomic Concentration (%)
	Ni2p	Co2p	Fe2p	Ce3d	O1s
CeNiO3	29.0	-	-	9.94	29.07
CeCoO3	-	28.14	-	9.80	27.80
CeFeO3	-	-	24.92	10.93	27.93
Ni/CeO2	28.62	-	-	6.74	35.10
Co/CeO2	-	29.14	-	7.45	28.93
Fe/CeO2	-	-	31.0	7.58	29.25

). According to the literature [59,60,61], the peak at 530 eV and the peak at 531.3 eV in the O1s XPS spectrum correspond to Ce⁴⁺ and Ce³⁺, which indicates the coexistence of both oxidation states in CeO₂. The binding energies of O1s are slightly shifted to a lower energy in the perovskite catalysts and the surface-adsorbed oxygen, majorly corresponding to Ce³⁺. This indicates formation of adsorbed oxygen species was closely related to the presence of oxygen vacancies. The oxygen vacancy increases the electron density of the adjacent metal site and enhances the Lewis basic site [62]. On CeO₂ surfaces, oxygen ions located near oxygen vacancies can serve as Lewis base sites [63]. The increase in oxygen vacancies could increase the basic sites on the surface of CeO₂ to a certain extent. The oxygen vacancies can improve the interaction between the active metal (e.g., Ru) and the support material, creating a synergistic interface that is crucial for the reaction [62,64,65].

The core-level Co2p, Ni2p, and Fe2p XPS spectra of calcined CeXO₃ catalysts are presented in Figure 9. The Co2p of the CeCoO₃ catalyst exhibited a main peak at 780.1 eV (2p_3/2) with a low-intensity satellite peak at a higher binding energy (785.2 eV). The Co2p_1/2-Co2p_3/2 spin–orbit splitting was observed at a binding energy difference of 15.0 eV, i.e., at 795.0 eV. These results indicate the presence of Co³⁺ and Co²⁺ species on the catalyst surface [12,51,60]. The Ni2p of the calcined CeNiO₃ catalyst exhibited a major peak at 855.2 eV and a minor peak 853.5 eV, which were ascribed to Ni³⁺ of Ni₂O₃ and Ni²⁺ of NiO, respectively. The two prominent peaks at 855.2 and 872.3 eV, along with a satellite peak at 861.0 eV, correspond to the Ni2p_3/2 and Ni2p_1/2 energy levels. Therefore, the results clearly suggest the presence of Ni³⁺ species as the major component and Ni²⁺ species as the minor one [51,59]. The Fe2p of the calcined CeFeO₃ catalyst showed peaks at 711.1 and 724.8 eV, suggesting energy levels of Fe2p_3/2 and Fe2p_1/2. The Fe2p spectra also showed a weak peak at 718.5 eV, which is a characteristic shake-up satellite peak for Fe³⁺. These results indicate that Fe in CeFeO₃ is present predominantly as Fe³⁺ species [66,67]. The table shows that the surface metal content in the CeXO₃ catalysts follows the order of Ni > Co > Fe. The nickel (Ni) catalyst has a higher amount of surface metal than the cobalt (Co) and iron (Fe) catalysts. The surface oxygen content follows the same trend. Conversely, the Fe catalyst contains the highest amount of surface cerium.

The core-level XPS spectra of the impregnated catalysts are presented in the Supplementary Information. Compared to perovskite catalysts, the impregnated catalysts showed lower surface concentration of Ce. The Fe/CeO₂ catalyst showed higher Ce concentration than the other impregnated catalysts but the same level as the perovskite catalysts. The Co2p spectra of the Co/CeO₂ catalyst clearly showed two major peaks at 782.2 eV and 797.5 eV (Figure S5). The obtained binding energies—Co2p_3/2 (782.2 eV) and 2p_1/2 (797.5 eV)—are close to those reported for Co²⁺ ions in Co–O bonding, with an energy difference between Co2p_3/2 and Co2p_1/2 of 15.5 ± 0.2 eV [26,60,68]. The results clearly suggest the presence of Co²⁺ species.

Similarly, the Fe2p spectra of the Fe/CeO₂ catalyst (Figure S5) showed two major peaks at 711.6 eV and 752.2 eV, with weak peak at 718.9 eV. The results clearly suggest the presence of Fe³⁺ and Fe²⁺ species on the surface Fe/CeO₂ catalyst [12,67,69].

The Ni2p spectra of the Ni/CeO₂ catalyst showed two major peaks at 852.9 and 855.0 eV, which are assigned to the Ni2p_3/2. The remaining peaks in the range of 870–885 eV and the peaks in the range of 860–864 eV suggest energy levels of Ni2p_3/2 and Ni2p_1/2. Thus, the results clearly suggest the presence of Ni²⁺ and Ni³⁺ species [70,71,72,73]. The Ni³⁺ species might result from the interaction between Ni and CeO₂.

The O1s spectra of all impregnated catalysts are presented in Figure S6. All catalysts consist majorly of a peak around 529 eV, which is associated with metal–oxygen bond or lattice oxygen. There is another one around 532.0 eV, which is due to surface-adsorbed, low-coordinated oxygen. This surface oxygen creates a number of defect sites in the catalysts [71,72,73]. Figure S6 clearly suggests that the binding energy values shifted toward a lower energy in the Ni/CeO₂ catalyst compared to other catalysts. Also, it is clearly noticed that there is an increasing amount of lattice oxygen at the surface of Ni/CeO₂ compared to the other catalysts.

The surface oxygen concentration in the impregnated catalysts was high compared to their corresponding perovskite catalysts. In Ni-based catalysts observed higher surface oxygen concentrations than the Co-based and Fe-based catalysts. The order of surface metal concentration in the impregnated catalyst was the opposite to that observed in the perovskite catalyst.

The XPS results validate that the perovskite phase alters the electronic interactions among the catalyst components compared to their corresponding impregnated catalysts.

2.6. Activity Studies

Figure 10 demonstrates the catalytic performance of both perovskite-type and impregnated catalysts for the ammonia decomposition reaction. As expected for an endothermic reaction, all samples demonstrated a clear increase in catalytic activity with a rise in temperature across the range of 300 to 600 °C. The data provides a solid foundation for evaluating the performance of the prepared catalysts under these conditions.

The results in Figure 10 clearly demonstrate the effect of the perovskite phase on the ammonia decomposition reaction for Ni-, Co-, and Fe-based catalysts. Each catalyst exhibited maximum conversion at different temperatures. The perovskite-based catalysts achieved their highest conversion at relatively lower temperatures, except for CeFeO₃. The CeNiO₃ catalyst showed complete conversion at 550 °C. The perovskite structure promotes a synergistic interaction between NiO and CeO₂, leading to an extended Ni–support interfacial area [51].

From Figure 10, it is clear that the CeFeO₃ catalyst exhibited increasing activity up to 500 °C, with little further improvement at higher temperatures. The CeFeO₃ catalyst showed lower activity than the Fe/CeO₂ catalyst. According to the literature, the metallic Fe phase is active for ammonia decomposition. Due to the high enthalpy of the Fe–N bond, this can lead to the formation of surface nitrides such as α-Fe(N), α″-Fe₁₆N, γ′-Fe₄N, ξ-Fe_xN, and ζ-Fe₂N [74,75,76,77,78]. The accumulation of these nitrides on the active surface effectively poisons the catalyst, resulting in a significant reduction in the reaction rate and loss of performance [79]. Therefore, the perovskite phase promotes electronic interactions between the metal and CeO₂ during reduction. These interactions favor the CeNiO₃ and CeCoO₃ catalysts for their higher activity and promote the formation of iron nitrides in the CeFeO₃ catalyst.

The metal–nitrogen binding energy plays a vital role in designing efficient catalysts for ammonia decomposition. According to the literature, among non-noble metals, Co-based catalysts exhibit nitrogen binding energies closest to the ideal value, second only to Ru [26,38]. However, the overall ammonia decomposition activity is also strongly influenced by the choice of catalytic support and the nature of metal–support interactions. Although the impregnated catalysts have higher surface areas than perovskite catalysts, they showed low catalytic performance. In the present study, it is clear that perovskite catalysts are more active than the impregnated catalysts.

The metal reducibility occurred at lower temperatures in the perovskite catalysts than in the impregnated catalysts. Among all catalysts, the lowest metal reducibility temperature was observed in the CeNiO₃ catalyst. From the Raman and XPS results, it was concluded that a higher concentration of oxygen vacancies was observed in the perovskite catalysts compared to impregnated catalysts. These oxygen vacancies can help in the overall reaction pathway by mediating the evolution of reaction intermediates and promoting nitrogen recombination [80]. This indicates that the perovskite catalyst surfaces are electron-rich in nature, especially in CeNiO₃. According to the literature [9,81,82,83], metal atoms in close contact with electron-rich surface sites tend to donate electrons and facilitate nitrogen desorption, which is the rate-determining step in the ammonia decomposition reaction. The catalytic activity trends for all tested catalysts are summarized as follows: CeNiO₃ > CeCoO₃ > Co/CeO₂ > Ni/CeO₂ > Fe/CeO₂ > CeFeO₃.

The superior activity of the perovskite catalysts (from Figure 10) is clearly evident from the temperatures required to achieve 50% NH₃ conversion:

Perovskite: CeNiO₃ (433 °C), CeCoO₃ (457 °C) and CeFeO₃ (483 °C);

Impregnated: Ni/CeO₂ (510 °C), Co/CeO₂ (478 °C) and Fe/CeO₂ (483 °C).

The significant shift to lower reaction temperatures for the perovskite samples, particularly CeNiO₃, to reach the same conversion level represents the key evidence supporting the ABO₃ structure’s advantage.

Figure 11 presents the apparent activation energy (Ea) values determined for the NH₃ decomposition reaction over the synthesized perovskite-derived and impregnated catalysts. These values were calculated from the Arrhenius plots (ln(k) versus 1/T), using data collected between 400 and 500 °C at a gas hourly space velocity (GHSV) of 6000 h⁻¹. All prepared catalysts—CeNiO₃ (38.22 KJ mole⁻¹), CeCoO₃ (50.14 KJ mole⁻¹), CeFeO₃ (104.59 KJ mole⁻¹), Ni/CeO₂ (82.8 KJ mole⁻¹), Co/CeO₂ (65.43 KJ mole⁻¹), and Fe/CeO₂ (72.49 KJ mole⁻¹)—showed clear activation energies. The calculated activation energy (Ea) values are notably lower than those reported in the literature for many comparable Ni- and Co-based NH₃ decomposition catalysts. For instance, the Ea of the current catalysts falls significantly below that of numerous benchmark systems, including 5%Co/Mg-La (67 KJ mole⁻¹) [9], Ni/Ce/SiO₂ (68.4 KJmole⁻¹) [84], CoNi/MgO-CeO-SrO₂ (56.8 KJmole⁻¹) [85], 5%Co/CeO₂ (81.29 KJ mole⁻¹) [2], 5%Co/CaO-CeO₂ (65.8 KJ mole⁻¹) [2], 20%Co/La-MgO (167 KJ mole⁻¹) [86], 5%Co/CNTs (92 KJ mole⁻¹) [87], 40%Ni/Mg-La (54 KJ mole⁻¹) [8], 20%Ni/La-Mg (181 KJ mole⁻¹) [86], 10%Ni/Al₂O₃ (53.9 KJ mole⁻¹) [8], 4%Ni/Al₂O₃ (99.5 KJ mole⁻¹) [88], and 4%Ni/CeO₂ (71.0 KJ mole⁻¹) [88], among others.

The perovskite phase formed by Ni and Ce proved crucial for enhancing catalytic performance, resulting in a Ni-based catalyst with a lower activation energy (Ea) than the corresponding Co-based catalyst. The substantial increase in activity, reflected by the reduction in Ea, clearly indicates that the perovskite structure significantly improves the catalyst’s effectiveness for NH₃ decomposition [83]. Among all the materials synthesized, CeNiO₃ was the most intrinsically active, exhibiting the lowest Ea.

The hydrogen formation rate for all the prepared catalysts was calculated at 550 °C with a space velocity of 6000 h⁻¹. The results, displayed in Figure 12, show a higher hydrogen formation rate for the perovskite catalysts than for the impregnated catalysts, except for the Fe-based systems. The difference in activity between the perovskite and impregnated catalysts was most pronounced in the Ni-based samples compared with the Co- and Fe-based ones. An approximately 1.3-fold increase in activity was observed from CeNiO₃ to Ni/CeO₂.

The long-term stability of the CeNiO₃ catalyst was evaluated through a 50 h time-on-stream (TOS) test at 550 °C and a gas hourly space velocity (GHSV) of 6000 h⁻¹. The performance data, presented in Figure S7, indicate that the CeNiO₃ catalyst demonstrated excellent stability, maintaining nearly constant NH₃ conversion over the entire 50 h duration at 550 °C. The negligible variation in conversion efficiency throughout the stability test clearly demonstrates that the catalyst experienced no significant deactivation under the specified reaction conditions.

3. Experimental

3.1. Preparation of Perovskite Type Oxide

The CeXO₃-type perovskites were synthesized using a solution combustion method, with CitAc serving as the organic fuel. Specifically, to prepare 5 g of CeNiO₃, 20.24 milli moles of Ce(NO₃)₃·6H₂O (Aldrich, San Diego, CA, USA) and 20.25 mmol of Ni(NO₃)₂·6H₂O (Merck, Darmstadt, Germany) were dissolved in 50 mL of deionized water to form a metal nitrate solution. Separately, 44.3 milli moles of CitAc (Aldrich, San Diego, CA, USA) was dissolved in 50 mL of deionized water. The metal nitrate solution was then added dropwise to the CitAc solution, maintaining a CitAc-to-metal salt molar ratio of 1:1. The resulting mixture was heated at 90 °C on a hot plate with continuous stirring for over 6 h until a light green gel was formed. The gel was aged in a water bath at 90 °C for 24 h and subsequently dried in an oven at 110 °C for another 24 h. The dried mass was ground into a fine powder and calcined in a muffle furnace with controlled temperature ramping. The temperature was first increased to 350 °C at a rate of 3 °C min⁻¹ and held for 3 h, followed by a further increase to 700 °C at 5 °C min⁻¹, maintained for 6 h to obtain the desired perovskite phase. The resulting catalyst was designated as CeNiO₃. The same procedure was used to synthesize Ce–Co and Ce–Fe perovskites, denoted as CeCoO₃ and CeFeO₃, respectively.

3.2. Preparation of Supported Catalysts

The supported catalysts were synthesized using the traditional wet impregnation approach. The CeO₂ obtained from Rhodia (Gent, Belgium) was used as the support material. To maintain a consistent molar ratio of X/Ce = 1 (X = Co, Fe, Ni), 30 wt% of the respective metal was loaded onto the CeO₂ support. For instance, to prepare 5 g of 30 wt% Ni/CeO₂, 7.432 g of Ni(NO₃)₂·6H₂O was dissolved in 50 mL of deionized water and added to 3.5 g of CeO₂ in a round-bottom flask. Impregnation was performed using a rotary evaporator under vacuum, with continuous rotation at 150 rpm and 80 °C, until the solvent was completely evaporated. The resulting solid was dried at 110 °C for 24 h and calcined at 550 °C for 5 h. The obtained catalyst was designated as Ni/CeO₂. The same procedure was followed for the preparation of the 30 wt% Co/CeO₂ and 30 wt% Fe/CeO₂ catalysts, which were denoted as Co/CeO₂ and Fe/CeO₂, respectively.

3.3. Catalyst Characterization

Catalyst textural properties, specifically surface area and pore-size distribution, were determined by N₂ physisorption using a Quanta chrome (NOVA-2200e) analyzer (Quantachrome Instruments, Boynton Beach, FL, USA). To prepare the samples for testing, they were first degassed by heating under vacuum at 200 °C for two hours. Adsorption measurements were carried out with Nitrogen at −196 °C.

Powder X-ray diffraction (XRD) was used to identify the crystalline phases of the calcined catalysts. The measurements were conducted on an INEL Equinox 1000 (INEL, Artenay, France) instrument, utilizing a Cu Kα X-ray source (λ = 1.54 Å) operated at 30 kV and 30 mA. Data acquisition was performed in real-time across the 2θ range of 10–110°, with phase identification achieved by comparison against the PDF library.

Hydrogen temperature-programmed reduction (H₂-TPR) experiments were conducted to assess the reducibility of the catalysts using a Micromeritics Autochem HP-2950 (Micromeritics Instrument Corp., Norcross, GA, USA) instrument. A 0.15 g catalyst sample was placed in a quartz U-tube reactor. Before testing, the sample was pretreated by flushing it with He at 150 °C for 30 min to remove moisture. Subsequently, the sample was heated to 800 °C at a ramp rate of 10 °C min⁻¹ under a flow of 10% H₂/Ar (50 mL min⁻¹).

To examine the basicity of the samples, CO₂-–TPD analysis was carried out using a Micromeritics Auto Chem instrument (Micromeritics Instrument Corp., Norcross, GA, USA). The protocol began with a two-hour reduction under 10%H2/Ar (20 mL min⁻¹) at 550 °C. After reduction, the sample was purged with He for one hour and cooled to 40 °C. After stabilizing at 40 °C, the adsorption step involved exposing the sample to 10%CO₂/Ar at 40 °C for one hour. Desorption was then monitored as the temperature was linearly ramped to 850 °C (10 °C min⁻¹) in a flowing He atmosphere (50 mL/min).

The surface elemental composition and chemical states were studied by XPS using a SPECS High Vacuum system equipped with a Magnesium Kα source (1253.6 eV). Sample preparation included 16 h of vacuum degassing inside the load lock. For charge correction and accurate quantification, the binding energies were normalized to the C 1s peak from residual carbon, set at 284.8 eV.

Raman spectroscopy was utilized to analyze the vibrational modes of the samples and confirm the formation of the perovskite structure. The Raman spectra were recorded using a Renishaw spectrometer equipped with a Centrus 0LL756 detector (Renishaw, Gloucestershire, UK). A 532 nm excitation laser served as the light source, providing a high spectral resolution of 4 cm⁻¹. The measurements were conducted over a wavenumber range of 101.17 to 1601.45 cm⁻¹ to capture the characteristic vibrational features associated with metal–oxygen bonding and lattice dynamics within the perovskite framework.

3.4. Activity of Catalysts

Catalytic activity tests for ammonia decomposition were carried out in a fixed-bed microreactor with an outer diameter of 9 mm. Approximately 0.2 g of catalyst was loaded into the reactor for each experiment. Prior to testing, the catalyst had been reduced under a gas mixture of H₂/N₂ (1:1 by volume) at 550 °C for 5 h. After reduction, the reactor was cooled to 300 °C and purged with nitrogen to remove residual gases. Ammonia decomposition was then performed by feeding 100% NH₃ at a gas hourly space velocity (GHSV) of 6000 mL·g(_cat)⁻¹·h⁻¹. The reaction temperature varied from 300 to 600 °C in 50 °C increments. At each temperature, data were collected after the system reached steady-state conditions. The effluent gases from the reactor were analyzed using a Varian 450 gas chromatograph equipped with a Porapak Q column and a thermal conductivity detector (TCD). The ammonia conversion (X) was calculated using the following equation:

$$Ammonia Conversion\left(X\right)={\displaystyle \frac{n_{{NH}_3}^i-n_{{NH}_3}^o}{n_{{NH}_3}^i}}\times 100,$$

where nⁱ_NH3 is the number of moles of NH₃ at the reactor inflow and n^o_NH3 is the number of moles of NH₃ at the reactor outflow.

4. Conclusions

Ce-based Ni, Co, and Fe perovskite catalysts (CeXO₃, X = Ni, Co, Fe) were synthesized and employed as catalyst precursors for ammonia decomposition. For comparison, impregnated catalysts with the same active metal concentration were also prepared. Overall, the perovskite catalysts exhibited superior catalytic activity compared to their impregnated counterparts, with the exception of CeFeO₃. Among the tested samples, CeNiO₃ demonstrated the highest activity at lower temperatures. The perovskite structure of CeXO₃ played a crucial role in enhancing catalytic performance, particularly for Ni- and Co-based systems, by improving NH₃ conversion. Notably, the CeNiO₃ catalyst exhibited the best activity among all synthesized materials. The superior catalytic performance of CeNiO₃ can be attributed to its high oxygen vacancies, which favors the active metal (Ni), conducive to the ammonia decomposition. Facile reducibility and appropriately distributed basic sites also helped to achieve CeNiO₃’s high performance. Furthermore, the homogeneous inter-dispersion of metal oxide constituents ensured uniform nickel distribution, thereby contributing to the high performance of the catalyst.