Ni Supported on Pr-Doped Ceria as Catalysts for Dry Reforming of Methane

Abstract

The use of CH₄ and CO₂ as fuels in direct internal reforming solid oxide fuel cells (DIR-SOFCs) is a promising strategy for efficient power generation with reduced greenhouse gas emissions. In this study, Ni catalysts supported on Ce–Pr mixed oxides with varying Pr contents (0–80 mol%) were synthesized, calcined at 1200 °C, and tested for dry reforming of methane (DRM), aiming at their application as catalytic layers in SOFC anodes. Physicochemical characterization (XRD, TPR, TEM) showed that increasing Pr loading enhances catalyst reducibility and promotes the formation of the Pr₂NiO₄ phase, which contributes to the generation of smaller Ni⁰ particles after reduction. Catalytic tests revealed that all samples exhibited low-carbon deposition, attributed to the large Ni crystallites. The catalyst with 80 mol% Pr showed the best performance, achieving the highest CH₄ conversion (72%), a H₂/CO molar ratio of 0.89, and improved stability. These findings suggest that Ni/Ce_0.2Pr_0.8 could be a promising candidate for use as a catalyst layer of anodes in DIR-SOFC anodes. Although electrochemical data are not yet available, future work will evaluate the catalyst’s performance and durability under SOFC-relevant conditions.

1. Introduction

Energy demand is rising due to population growth and increasing industrial activity. However, the extensive use of these fuels has led to higher emissions of greenhouse gases, such as carbon dioxide (CO₂) and methane (CH₄), which are major contributors to global warming [1]. Among the strategies for cleaner energy production, direct internal reforming solid oxide fuel cells (DIR-SOFCs) offer an attractive solution by enabling the conversion of CH₄ and CO₂ mixtures into electrical energy within the cell itself, thus avoiding the release of these gases into the atmosphere while achieving high conversion and system integration simplicity [2].

In DIR-SOFCs, CH₄ and CO₂ are directly fed into the anode compartment, where dry reforming of methane (DRM) occurs, promoted by the catalytic properties of the anode material, generating syngas (H₂ and CO), which is subsequently electrochemically oxidized. This configuration eliminates the need for external reformers and allows for the use of biogas, composed mainly of CH₄ and CO₂, as a sustainable fuel [2,3,4,5,6,7,8,9,10,11,12].

Ni-YSZ cermet is the most used anode material in SOFCs due to its combined electronic/ionic conductivity and mechanical stability at high temperatures [13]. Nevertheless, the high Ni content required for percolation promotes carbon formation from CH₄, leading to deactivation and irreversible damage to the anode [14,15].

To mitigate this issue, recent research has focused on the incorporation of catalytic layers based on Ni supported on cerium-based oxides. Ceria offers high oxygen storage capacity and redox activity, enabling the removal of surface carbon through oxygen transfer near metal sites [16,17]. Doping ceria with cations, such as Gd³⁺, Y³⁺, Pr³⁺, and Zr⁴⁺, enhances its oxygen mobility and conductivity [7,18,19,20,21], improving catalyst performance for DRM. These materials are promising as catalytic layers that promote reforming reactions while protecting the electrochemical anode (e.g., Ni/YSZ).

Some authors have proposed the use of Ni–Ce_0.9Gd_0.1O_2–δ (GDC) and perovskite-type SrTiO₃-based materials containing La, Ce, and Ni (LSCNT) as catalytic layers for the anodes of biogas-fueled DIR-SOFCs [22,23]. However, significant carbon deposition was observed on the GDC catalytic layer. While LSCNT demonstrated high stability during catalytic tests, its CH₄ and CO₂ conversion rates (~40%) were lower than those of other catalysts commonly used in methane reforming reactions [23].

Considering the aforementioned studies, it is evident that it is still necessary to develop catalyst layers for anodes of DIR-SOFCs fueled by CH₄ and CO₂ mixtures that exhibit high activity and strong resistance to carbon formation.

Da Fonseca et al. [24] studied the behavior of Ni supported on ceria containing Zr, Pr and Nb (Me/Ce molar ratio of 1/9; Me = Zr, Pr or Nb) as anodes for SOFC fed by fuels containing methane and CO₂. The catalysts containing 14 vol% of Ni were obtained using a hydrothermal method described in the literature [25] to prepare Ni-based anodes with greater resistance to carbon deposition. The Ni/CeZr, Ni/CePr and Ni/CeNb catalysts were calcined at 800 °C. Catalytic test results in a fixed-bed reactor (with a CH₄/CO₂ ratio of 1) showed that the Ni/CePr catalyst achieved high methane conversion and minimal carbon formation. The reduced carbon deposition was attributed to the CePr support’s enhanced oxygen mobility, which facilitates carbon removal.

Vasiliades et al. [26] studied the performance of Ni (5 wt.%) supported on Ce_1–xPr_xO_2–δ (x = 0.0–0.8) as catalysts for DRM at different reaction temperatures. They observed that the Ni/Ce_1–xPr_xO_2–δ was an active and very promising coke-resistant material for DRM in the 700–750 °C range. According to these authors, the Pr-dopant concentration is very important as it influences the mobility and concentration of lattice oxygen near the Ni-support interface, resulting in the production of gaseous CO. Additionally, the concentration of surface oxygen vacancies is affected by the Pr-dopant, which play a major role in the dissociation of CO₂ into CO(g) and atomic oxygen, aiding in the re-oxidation of the reduced support.

It is worth noting that the DRM catalysts studied by Vasiliades et al. [26] and da Fonseca et al. [24] were calcined at lower temperatures (750 and 800 °C) than the 1200 °C required for ceramic processing of SOFC layers [15,27]. Using high calcination temperatures is crucial for obtaining a catalyst with microstructural properties similar to those used in SOFCs. According to the literature, the Pr loading (up to 80 atom%) does not affect the ceria crystallite size for samples calcined at 750 °C [26]. On the other hand, when ceria doped with Pr (10 and 50 wt.% of Pr) are calcined at higher temperatures (1000 °C), the Pr loading strongly affects the support crystallite sizes [28,29].

Considering the studies cited above, although various dopants such as Gd³⁺, Zr⁴⁺, and Nb⁵⁺ have been investigated to improve the redox properties and catalytic performance of ceria-based supports, Pr³⁺ doping has shown distinct advantages for DRM, including enhanced oxygen mobility, increased surface reactivity, and reduced carbon deposition [24,26]. Therefore, Ni supported on Pr-doped ceria appears to be a particularly promising candidate for catalytic layers in SOFC anodes directly fed with CH₄ and CO₂ mixtures, due to its favorable redox behavior and resistance to carbon formation. However, most studies to date have assessed the performance of catalysts containing Pr after calcination at relatively low temperatures (typically ≤800 °C), which are not compatible with the high-temperature conditions (~1200 °C) required for ceramic processing in SOFC manufacturing. The influence of Pr content on key catalyst properties, such as crystallinity, oxygen mobility, and structural stability, under these high-temperature conditions remains poorly understood. This represents a significant knowledge gap, particularly for the design of catalytic layers capable of maintaining their structural integrity and catalytic activity after thermal treatment during SOFC fabrication.

Therefore, the main objective of this work was to develop a catalyst capable of withstanding the high-temperature conditions typical of SOFC processing while maintaining good catalytic performance for DRM. To this end, we investigated the performance of Ni supported on Pr-doped ceria, evaluating the effects of Pr content on the catalyst’s crystalline structure, Ni and support crystallite sizes, and support reducibility after calcination at 1200 °C.

Assessing the catalytic performance of materials prepared under such conditions in fixed-bed reactors, prior to electrochemical testing in SOFC cells, is a valuable strategy. It helps to screen promising candidates efficiently, reducing the number of cell experiments required, thereby saving time and lowering overall costs. The findings from this study will support the selection of a catalyst with strong potential for future studies focused on DIR-SOFC applications.

For this, the catalysts prepared are characterized using temperature-programmed reduction (TPR), in situ X-ray diffraction (XRD), Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) analyses. The catalytic performance during DRM was evaluated in a fixed-bed reactor. Post-reaction temperature-programmed oxidation (TPO) analyses were performed to investigate the quantity and nature of the carbon formed during the DRM.

2. Materials and Methods

2.1. Catalyst Preparation

The supports containing Ce and Pr (Ce_1–x Pr_x with x = 0, 0.2, 0.5, 0.8) were prepared via co-precipitation of the precursor salts’ aqueous solutions (ammonium cerium nitrate (IV) and praseodymium nitrate hexahydrate) with ammonium hydroxide (NH₄OH). An NH₄OH solution was added in a 4-fold molar excess. After 24 h, the mixtures were vacuum filtered and washed to achieve a neutral pH, then dried at 100 °C for 24 h. The samples were ground and calcined under an air flow (50 mL/min) for 2 h at 300 °C. The calcined supports were sieved (d ≤ 53 µm) for subsequent metal addition.

Ni (10 wt.%) was added to the supports via wet impregnation with aqueous solutions of nickel nitrate hexahydrate (Ni(NO₃)₂·6H₂O). The solutions were kept under stirring for 19 h at room temperature and atmospheric pressure. After evaporation, the obtained materials were dried at 100 °C for 24 h and ground. The samples were calcined at 1200 °C for 5 h (0.5 °C/min) from room temperature to 300 °C, and 2 °C/min from 300 to 1200 °C. This elevated temperature was selected to simulate the conditions necessary for the ceramic processing of SOFC layers, as reported in the literature [15,30]. Then, the following catalysts were obtained: Ni/Ce, Ni/Ce_0.8Pr_0.2, Ni/Ce_0.5Pr_0.5 and Ni/Ce_0.2Pr_0.8.

2.2. XRD

In situ X-ray powder diffraction analyses were carried out at the XRD1 beamline of the National Syncrotron Light Laboratory (LNLS), using a wavelength radiation of 1.03305 Å, a 2θ range from 10° to 100° and acquisition time of 120 s. XRD patterns were obtained from the Ni/Ce_(1–x)Pr_x catalysts, before and during reduction, under pure H₂ at 800 °C for 30 min, at a heating rate of 10 °C/min and a flow rate of 10 mL/min. XRD analysis of Ni/Ce_0.5Pr_0.5 was also performed after CO₂ flow to evaluate the oxidation of Ni particles. For this purpose, after a reduction treatment under H₂ at 800 °C for 1 h and purging under N₂ for 30 min, the catalysts were subjected to a flow of 50 mL/min of CO₂. Subsequently, the system was purged with N₂ for 30 min. Finally, the material was passivated with a bath of liquid nitrogen and isopropyl alcohol under a flow mixture of 5% O₂/N₂.

The crystallite mean sizes were calculated using the Scherrer equation, Equation (1), for CeO₂ and Ni phases.

$$d={\displaystyle \frac{k·\lambda }{\beta ·cos\left(\theta \right)}}$$

(1)

where β = Full Width at Half Maximum (FWHM) calculated from the diffraction line’s Gauss curve, θ = Bragg’s angle, k = 0.9 (spherical particles), λ = radiation wavelength and d = crystallite’s mean diameter.

Lattice parameters of the supports were calculated via Bragg’s equation, considering the (111) plane and pseudo-cubic geometry.

2.3. TEM

STEM analyses and energy-dispersive X-ray spectroscopy (EDS) were conducted on a JEOL JEM 2100F (Tokyo, Japan), with 80 kV to 200 kV voltage, field emission and EDS SSD detector with a high-angle annular dark-field detector (HAADF). For that purpose, small quantities of the catalysts were diluted in ethanol and kept under ultrasound sonification for 40 min for better dispersion. Then, they were dripped onto a TEM mesh (Electron Microscopy Sciences, Hatfield, PA, USA, Cat. # LC200-CU) with Pasteur pipette. EDS allowed for the identification of the different element distribution in the samples and [31], thus, a better understanding of the metal–support interaction.

2.4. SEM

Scanning Electron Microscopy (SEM) was employed to evaluate the formation of filamentous carbon structures. The analyses were performed using a JEOL JSM-7100F field-emission microscope operating at 30 kV and equipped with an energy-dispersive X-ray spectroscopy (EDS) system. Samples were mounted on aluminum stubs using conductive carbon tape.

2.5. TPR

The TPR analysis was performed on an AutoChem II 2920 V5.01 instrument (Micromeritics^®, Norcross, GA, USA), coupled to a thermal conductivity detector. The samples were heated to 1000 °C at a 10 °C/min rate, under a 10% H₂/N₂ mixture and total flow rate of 30 mL/min.

2.6. DRM Reaction

DRM reaction was carried out in a tubular quartz reactor for 24 h, fed with a CH₄:CO₂ equimolar gas mixture at 800 °C, atmospheric pressure and a total flow rate of 100 mL/min. The exit gases were analyzed using a coupled gas chromatograph equipped with a thermal conductivity detector. Samples of around 100 mg were used, diluted with SiC in a weight ratio of 1.5 (SiC/catalyst), to avoid hot spots and temperature gradients. Before each reaction, reduction was carried out under pure H₂ (30 mL/min) at 800 °C for 1 h and heating rate of 10 °C/min, followed by a 30 min N₂ purge, at the same temperature and flow rates. The reactant conversions and product distribution were calculated from the equations below.

$$X_i=\left({\displaystyle \frac{{\left(n_i\right)}_{feed}–{\left(n_i\right)}_{exit}}{{\left(n_i\right)}_{feed}}}\right)\times 100$$

(2)

$$S_p=\left( {\displaystyle \raisebox{1ex}{$n_p$}\!\left/ \!\raisebox{-1ex}{$n_{total}$}\right.}\right)\times 100$$

(3)

where i = CH₄ or CO₂, Xi = conversion of i, ni = moles of i, p = H₂, H₂O or CO, Sp = yield of p, np = moles of produced p, ntotal = sum of the moles of products (CO, H₂ and H₂O).

2.7. TPO

TPO analysis was performed immediately following the catalytic test within the same reaction unit to quantify the carbon deposition during the 24 h reaction. This analysis employed a quadrupole mass spectrometer (Pfeifer D-35614, Memmingen, Germany) to detect the CO₂ signal generated from the combustion of carbonaceous species on the catalyst. After the reaction, the reactor was cooled to room temperature under a He flow. An oxidizing gas mixture (O₂/He ≈ 21.4%) was subsequently introduced at a total flow rate of 70 mL/min. Once the gas signals stabilized, the temperature was ramped to 900 °C at a rate of 10 °C/min, and the CO₂ signal was monitored to track the carbon oxidation process. Upon reaching the final temperature, the feed was switched to bypass mode, and a CO₂ reference signal was obtained at a known flow rate for calibration purposes.

3. Results and Discussion

3.1. Catalyst Characterization

XRD patterns obtained for all calcined samples are presented in Figure 1 and Figure 2. All samples exhibited the diffraction lines characteristic of CeO₂ in cubic phase (ICSD 72155). In the case of Ni/Ce_0.8Pr_0.2, a shift in the CeO₂ lines was detected to lower values of 2θ (Figure 2). This result suggests that there was an expansion in the cerium oxide lattice. No diffraction lines related to praseodymium oxides were observed. On the other hand, for Ni/Ce_0.5Pr_0.5 and Ni/Ce_0.2Pr_0.8 samples, a displacement for higher values of 2θ was detected. In addition, for the sample containing 80 mol% of Pr, in addition to the lines related to CeO₂, the diffraction lines corresponding to Pr₁₂O₂₂ (ICSD 82107) in a monoclinic phase were detected, which is a non-stoichiometric oxide containing Pr⁺³ and Pr⁺⁴. However, the presence of a Pr₆O₁₁ phase cannot be ruled out, since the diffraction lines of this phase overlap with those of ceria in the cubic phase.

Krishna et al. [28] studied the performance of ceria doped with different contents of Pr as catalysts for soot oxidation. They observed a shift in the peaks corresponding to the ceria cubic phase to higher 2θ values and a PrO_x phase segregation in XRD analyses of Pr-doped ceria. According to these authors, the higher 2θ values indicate that most of the praseodymium in a solid solution is present in the 4⁺ oxidation state.

Concerning the Ni species, all samples exhibited NiO diffraction lines. Nevertheless, in both Ni/Ce_0.5Pr_0.5 and Ni/Ce_0.2Pr_0.8 catalysts, a reduction in intensities of NiO lines was observed. Furthermore, the main diffraction line of ceria became less symmetric, indicating the formation of another phase, such as Pr₂NiO₄, in samples with a higher Pr content.

To confirm the possibility of this phase formation, analyses were carried out for a sample containing only Ni and Pr (Figure 1 and Figure 2) prepared using the same method as the other samples. For the Ni/Pr sample, diffraction lines corresponding to NiO were not identified. Instead, only the diffraction lines of Pr₂NiO₄ in the orthorhombic phase (ICSD 81577) were detected. Thus, in this work, the XRD results suggest that increasing Pr loading favored the formation of Pr₂NiO₄ and PrO_x phases, in addition to the ceria phase.

The values of the ceria lattice parameter obtained for Ni/Ce, Ni/Ce_0.8Pr_0.2, Ni/Ce_0.5Pr_0.5 and Ni/Ce_0.2Pr_0.8 are shown in

Table 1. Values of lattice parameter of CeO₂ and crystallite size of cerium oxides and Ni⁰ after calcination at 1200 °C and after reduction at 800 °C for 1 h.

Catalyst	Lattice Parameter of Calcined Samples (Å) 1	CeO2 Crystallite Sizes (nm) 2	NiO Crystallite Sizes (nm) 2	Ni0 Crystallite Sizes (nm) 3
Ni/Ce	5.4137	95	81	116
Ni/Ce0.8Pr0.2	5.4155	81	90	98
Ni/Ce0.5Pr0.5	5.4095	116	78	73
Ni/Ce0.2Pr0.8	5.4050	107	88	63

¹ Calculated by using the (111) CeO₂ (experimental error = ±0.0008). ² After calcination at 1200 °C. ³ After reduction at 800 °C for 1 h.

.

A comparison between the values of lattice parameters of ceria for Ni/Ce_0.8Pr_0.2, Ni/Ce_0.5Pr_0.5 and Ni/Ce_0.2Pr_0.8 with those obtained for Ni/Ce and for CeO₂ from ICSD 72155 (5.4124) showed that the addition of 20 mol% of Pr led to an increase in the lattice parameter. The increase in the lattice parameter and the shift to lower values of 2θ suggest the insertion of praseodymium ions in the CeO₂ crystalline lattice [32]. On the other hand, for Ni/Ce_0.5Pr_0.5 and Ni/Ce_0.2Pr_0.8, there was a decrease in the values of the CeO₂ lattice parameter, indicating a contraction of the crystalline lattice of ceria.

Chiba et al. [33] reported that the insertion of Pr in the crystalline lattice of CeO₂ can cause both expansion and contraction effects, depending on the oxidation state of Pr ions. This effect can be explained by the significantly larger radius of Pr³⁺ ions (1.128 Å) compared to that of Ce⁴⁺ ions (0.97 Å). On the other hand, the ionic radius of Pr⁴⁺ (0.96 Å) is slightly smaller than that of Ce⁴⁺ [34,35].

The XRD results presented above indicated the formation of a solid solution between ceria and praseodymium oxides for all doped samples. Furthermore, increasing Pr content resulted in a phase segregation, which is consistent with the results reported by Krishna et al. [28], showing a segregation phase for CePrO_x catalysts, when Pr loading was higher than 50%.

Table 1 also shows the crystallite size of CeO₂ obtained through XRD analyses for calcined samples.

Compared with the results obtained for the undoped sample, the addition of 20 mol.% of Pr slightly decreased the support crystallite size. However, increasing Pr loading to 50 and 80 mol.% of Pr increased support crystallite sizes. Some authors reported that the crystallite size of CePrO_x calcined at high temperatures depends on the amount of praseodymium dopant. They observed that after calcination at 1000 °C, a lower amount of Pr (10 wt.%) decreased the crystallite size of CeO₂, whereas higher quantities of this dopant (50 wt.%) increased the support crystallite sizes [28,29]. All catalysts showed similar NiO particle sizes (∼78–90 nm).

Figure 3 and Figure 4 show the results obtained through TEM and EDS analyses for Ni/Ce and Ni/Ce_0.8Pr_0.2, respectively. Micrographs of the other samples are presented in the (Supporting Information Figures S1 and S2). TEM and EDS analyses revealed that all samples presented large CeO₂ and NiO particles, which are in agreement with XRD analyses. In addition, isolated NiO and CeO₂ crystallites were observed, with no significant difference in their sizes. On the other hand, in the samples containing Pr, smaller NiO particles dispersed on the support were observed, especially in the sample containing 80 mol% of Pr, indicating greater contact between the Ni particles and the support for this catalyst.

Figure 5 and Figure 6 show the diffractograms obtained for in situ XRD experiments for Ni/Ce and Ni/Ce_0.2Pr_0.8 during reduction treatment from room temperature to 800 °C. The results obtained for the other doped cerium samples (Ni/Ce_0.8Pr_0.2 and Ni/Ce_0.5Pr_0.5) are presented in Figures S3 and S4 (Supporting Information).

The diffractogram of the samples Ni/Ce, Ni/Ce_0.8Pr_0.2, Ni/Ce_0.5Pr_0.5 and Ni/Ce_0.2Pr_0.8 at room temperature shows the diffraction lines of CeO₂ and NiO.

For Ni/Ce (Figure 5), we observed a shift in the CeO₂ diffraction lines to lower values of 2θ after heating to 560 °C. This displacement is likely due to the distortions of the crystallite structure that are caused by heating and reduction of the oxide [33,36]. At 579 °C, a diffraction line around 27.90° was detected (Figure 5b), which is attributed to the presence of non-stoichiometric oxide Ce₁₁O₂₀ (ICSD 88758). According to the literature [36,37,38], the ceria-phase transitions observed with increasing temperature are as follows: CeO₂ → Ce₁₁O₂₀ → Ce₇O₁₂ → CeO_1.67.

Bekheet et al. [36] observed the transformation of Ce₇O₁₂ into Ce₁₁O₂₀ for ceria-based materials at approximately 530 °C during cooling under H₂. Increasing the temperature to 786 °C led to the appearance of a diffraction line at 2θ = 27.53°, which can be attributed to the formation of the Ce₇O₁₂ phase (ICSD 88754) with rhombohedral geometry. This phase is more stable at high temperatures, even though it has the same quantity of vacancies as the previous phase [37]. After 1 h at 800 °C, the CeO_1.67 phase with a bixbyite structure (ICSD-88753) was also detected.

Regarding the nickel species, the appearance of Ni⁰ lines is observed around 280 °C. At 474 °C, NiO lines are no longer detectable, and only the Ni⁰ lines are observed, indicating a complete reduction of NiO. After this temperature, the intensity of Ni lines increased as the temperature increased.

For the doped samples, a slight displacement of CeO₂ diffraction lines for smaller values of 2θ is also observed. In addition, increasing temperature reduction also led to the appearance of the Ce₁₁O₂₀ and Ce₇O₁₂ phases. The first one was observed at 490, 289 and 296 °C for Ni/Ce_0.8Pr_0.2, Ni/Ce_0.5Pr_0.5 and Ni/Ce_0.2Pr_0.8, respectively. The Ce₇O₁₂ phase was detected at 712 °C (for Ni/Ce_0.8Pr_0.2), 310 °C (for Ni/Ce_0.5Pr_0.5) and 296 °C (for Ni/Ce_0.8Pr_0.2). The formation of the Ce₁₁O₂₀ phase was also observed for samples with Pr during heating. However, the addition of Pr increased the temperature at which this phase was detected. This effect is attributed to the reduction of Pr⁴⁺ ions, leading to a valence state of Pr ions similar to that in Pr₆O₁₁ (mean valence value = +22/6) [33]. This contributes to the abrupt increase in the cell volume of the ceria cubic structure and the concentration of oxygen vacancies [33].

In addition to the ceria phases, the appearance of cubic Pr₂O₃ in bixbyite form (ICSD 647290) was detected from 747, 611 and 340 °C for Ni/Ce_0.8Pr_0.2, Ni/Ce_0.5Pr_0.5 and Ni/Ce_0.2Pr_0.8, respectively. However, the lines corresponding to this phase are more intense for the sample containing 80 mol% of Pr (Figure 6b,c). The Pr₂O₃ phase is formed by the reduction of the Pr₁₂O₂₂ phase, which occurs simultaneously with the reduction of the Ce₁₁O₂₀ phase. After reduction at 800 °C for 1 h, the presence of Ce₁₁O₂₀ was no longer detected for the samples containing 20, 50 and 80 mol % of Pr. Only the presence of the Ce₇O₁₂ and Pr₂O₃ phases was observed. The intensity of the Pr₂O₃ lines was higher for samples with 50 and 80 mol % of Pr.

Figure 7 presents the reduction temperature at which the different phases were first detected as a function of Pr loading. It is noticed that increasing Pr content decreased the temperature at which the presence of the Ce₁₁O₂₀, Ce₇O₁₂ and Pr₂O₃ phases was detected.

Thus, these results indicate that the reduction of CeO₂ occurs at lower temperatures for samples containing higher Pr loading. This suggests that an increase in Pr content led to an increase in the reducibility of the catalysts. Ahn et al. [32] reported that the addition of Pr to ceria favors the ordering of oxygen vacancies, creating paths for better diffusion of O₂, which increases the reducibility.

Regarding nickel species, all doped samples exhibited lines corresponding to NiO at room temperature. For the sample containing 80 mol% of Pr, in addition to the diffraction lines corresponding to NiO, diffraction lines related to Pr₂NiO₄ were also detected. The intensity of the lines corresponding to NiO decreased, and the lines related to Ni⁰ were detected during reduction at around 300–319 °C for Ni/Ce_0.8Pr_0.2, Ni/Ce_0.5Pr_0.5, and Ni/Ce_0.2Pr_0.8. The Pr content did not have a significant influence on the initial reduction temperature of NiO (Figure 7). Furthermore, it was not possible to determine at what temperature the Pr₂NiO₄ phase was reduced, since its diffraction lines overlap with the diffraction lines of the other phases detected at higher reduction temperatures.

Then, to determine the reducibility of the Pr₂NiO₄ phase, the in situ XRD analysis of Ni/Pr sample was carried out (Figure S5). At room temperature, the lines of Pr₁₂O₂₂ and Pr₂NiO₄ were detected. Increasing temperature decreased the intensity of the lines corresponding to Pr₁₂O_22, while the diffraction lines related to Pr₂O₃ with the cubic phase (ICSD 647290) were detected. At 464 °C, the lines associated with Pr₁₂O₂₂ were no longer observed. When the temperature was increased to 523 °C, the presence of lines of Pr₂O₃ with a hexagonal phase (ICSD 61179) was observed. The formation of this phase was followed by a decrease in the lines of Pr₂NiO₄. At 620 °C, only Pr₂O₃ with cubic and hexagonal phases was detected. No changes in XRD patterns were observed at temperatures higher than 620 °C. Since the results obtained showed the Pr₂NiO₄ reduction was followed by Pr₂O₃ formation, the absence of the diffraction lines of Ni⁰ could be attributed to the formation of very small Ni crystallites dispersed on Pr₂O₃.

Tarutin et al. [39] reported that, when Pr₂NiO_4+δ is submitted to reducing atmospheres, Ni is completely reduced, and a Ni/Pr₂O₃ cermet is formed, exhibiting small Ni particles.

Osinkin [40] also investigated the use of Pr₂NiO_4+δ as a catalyst to simultaneously enhance oxygen reduction and hydrogen oxidation reactions in solid oxide electrochemical devices. A double-layer electrode based on Sr₂Fe_1.5Mo_0.5O_6-a was coated with a symmetric catalyst using the wet impregnation method, applying a precursor solution for synthesizing Pr₂NiO_4+a. XRD analysis, after a 10 h heat treatment at 800 °C under various atmospheres, indicated that in in an oxidizing atmosphere, NiO and Pr₆O₁₁ are present, while Ni⁰ and Pr₂O₃ were detected in a reducing atmosphere.

Table 1 presents the crystallite sizes of Ni⁰ after reduction at 800 °C for 1 h. The Ni/Ce catalyst exhibited the largest Ni crystallite size (116 nm). The smallest Ni crystallite size was obtained for the sample with 80 mol% of Pr (63 nm). This result could be related to the presence of Pr₂NiO₄, which led to the formation of smaller Ni particles. The formation of the Pr₂NiO₄ phase is favored in the sample with the highest amount of Pr, as revealed by XRD analyses. Since metallic Ni is the active phase for methane dissociation in DRM, and smaller Ni particles are known to exhibit higher activity, the formation of Pr₂NiO₄ during calcination can be considered beneficial. In this context, Pr₂NiO₄ acts as a precursor that facilitates the generation of smaller Ni⁰ particles.

The TPR profiles are presented in Figure 8. The Ni/Ce exhibited H₂ consumption at 400, 450, 597 and 830 °C. The peaks in the region between 400 and 597 °C can be attributed to the reduction of NiO and superficial CeO₂ [41,42].

The consumption of H₂ at higher temperatures is related to the reduction of bulk CeO₂ [43]. For samples containing Ce and Pr, similar TPR profiles were obtained. However, increasing Pr content increased the intensity of the peaks at temperatures below 700 °C and decreased H₂ consumption at higher temperatures. The increase in consumption at lower temperatures may be associated with an increase in the reducibility of CeO₂ that is due to the formation of the solid solution between Ce and Pr [44]. This agrees with XRD analyses during reduction treatment, which showed that the increase in Pr content led to an increase in the reducibility of the catalysts. Moreover, for the samples containing 50 and 80 mol% of Pr, it is not possible to rule out the reduction of the Pr₂NiO₄ phase at temperatures lower than 750 °C, since the results obtained via XRD analyses suggest that these samples also exhibited this phase. In the case of the sample with 80 mol% of Pr, in addition to the Pr₂NiO₄ phase, the reduction of Pr₁₂O₂₂ cannot be ruled out either.

The total H₂ consumption and the degree of reduction of Ni species and support for all catalysts are presented in

Table 2. Amount of H₂ consumed and reduction degree of all catalysts obtained during TPR analyses.

Catalyst	Reduction Degree at Low Temperature Region (T < 750 °C) (%)	Reduction Degree at Low Temperature Region (T < 750 °C) (%)	Total H2 Consumption (mol/g Catalyst)
Ni/Ce	99	32	2.36 × 10−3
Ni/Ce0.8Pr0.2	131	12	2.37 × 10−3
Ni/Ce0.5Pr0.5	162	13	2.91 × 10−3
Ni/Ce0.2Pr0.8	231	12	2.77 × 10−3

.

It is important to note that the presence of the Pr₂NiO₄ phase does not impact the calculation of the degree of reduction for nickel, as the reduction stoichiometry for NiO reduction is the same as that for Pr₂NiO₄ reduction. Therefore, the degree of reduction of nickel can be calculated as if all nickel was in the NiO phase. The reduction of Ni/Ce at temperatures lower than 750 °C was close to 100%. However, for samples containing Pr and Ce, the values of the degree of reduction were higher than 100%, mainly for the sample containing 80 mol% of Pr. This suggests that, in addition to Ni species, there would also be a reduction in the support at temperatures lower than 750 °C for these samples. Furthermore, the reduction degree increased as the Pr content increased. These results agree with those reported by Zoellner et al. [45]. They conducted TPD of O₂ analyses for Ce_xPr_(1-x) oxides (x = 0.2–0.6) and observed that the increase in Pr content led to an increase in the O₂ storage capacity. The reduction of the supports at higher temperatures was not affected by Pr loading.

3.2. CO₂ Reforming of Methane

Catalytic tests were carried out at 800 °C and 1 atm, using a CH₄/CO₂ molar ratio of 1.00 (Figure 9 and Figure 10). The value of initial CH₄ conversion obtained for Ni/Ce was 58%. The highest and lowest values of initial CH₄ conversion were obtained for the samples with 80 mol% of Pr (72%) and 50% of Pr (45%), respectively (Figure 9). When comparing the initial methane conversion values obtained in this work with those reported in the literature for catalysts tested at 800 °C and calcined between 400 and 800 °C (

Table 3. Crystallite size of Ni⁰, initial methane conversion, and rate of carbon formation on Ni catalysts calcined at 400–1200 °C and tested in DRM at 700–800 °C with a CH₄/CO₂ molar ratio of 1.0 (this work and literature data).

Catalyst	Tcalc. (°C) a	Treaction (°C) b	Ni0  Crystallite Size (nm)	Initial XCH4 (%) c	mgC/ (gcat·h)	Reference
5%Ni/CeO2	400	800	10	~66	1.6	[38]
10%Ni/CeO2	400	800	20	~63	20.9	[38]
10%Ni/CeO2	600	800	28	~71	10.6	[38]
10%Ni/CeO2	800	800	46	~61	5.9	[38]
5%Ni/CeO2/YSZ	500	750	15	~100	2.9	[46]
10%Ni-CeAl	800	800	11	~70	2.0	[47]
10%Ni@CeO2	400	800	19	~60	1.6	[48]
10%Ni/CeO2	400	800	33	~60	9.7	[48]
10%Ni/Ce0.2Pr0.8	1200	800	63	72	0.36	this work
10%Ni/Ce0.5Pr0.5	1200	800	73	46	0.20	this work
10%Ni/Ce0.8Pr0.2	1200	800	98	56	0.29	this work
10%Ni/Ce	1200	800	116	58	0.33	this work
NiO	1200	800	132	37	0.0	[38]
5%Ni/CeO2	750	750	31	80	14.8	[26]
5%Ni/Ce0.8Pr0.2O2	750	750	34	75	12.0	[26]
5%Ni/Ce0.5Pr0.52	750	750	30	70	0.4	[26]
5&Ni/Ce0.2Pr0.8O2	750	750	36	50	0.06	[26]
18%Ni/CeNb	800	800	32	66	6.5	[24]
18%Ni/CeZr	800	800	24	74	9.68	[24]
18%Ni/CePr	800	800	33	66	0.42	[24]
5%Ni/CeO2	750	800	24	~55	4.8	[49]
5%Ni/Zr-DC (Zr-doped ceria)	750	800	9.2	~55	3.2	[49]
5%Ni/Sm-DC (Sm-doped ceria)	750	800	19	~38	0.6	[49]
5%Ni/La-DC (La-doped ceria)	750	800	21	~45	0.4	[49]
10%Ni/Ce/Al	800	800	4.6	~40	n.d.	[16]
10%Ni/Ce0.75Zr0.25O2/Al	800	800	6.6	~70	n.d.	[16]
10%Ni/Ce0.50Zr0.50O2/Al	800	800	5.2	~70		[16]
10%Ni/20%CeO2-Al2O3 Co-precipitation method	600	800	22.6	~75	200	[50]
10%Ni/20%CeO2-Al2O3 Wet impregnation method	600	800	11.1	~55	138	[50]
10%Ni/20%CeO2-Al2O3 Sol-gel method	600	800	9.5	~55	50	[50]
10%Ni/20%CeO2-Al2O3 Citric acid method	600	800	5.8	~70	12.5	[50]
20%Ni/CeO2-NR (nanorods morphology)	500	700	20.5	~45	0.54	[51]
5%Ni/CeO2-NR (nanorods morphology)	500	700	6.8	~70	0.21	[51]
3%Ni/CeO2-NR (nanorods morphology)	500	700	6.2	~55	0.50	[51]
26%Ni/CeO2-ac25 Combustion method (red./oxi. ratio = 2.5)	750	750	28	~50	n.d.	[52]
13%Ni/CeO2-ac25 Combustion method (red./oxid. ratio = 2.5)	750	750	15	~41	n.d.	[52]
7%Ni/CeO2-ac25 Combustion method (red./oxid. ratio = 2.5)	750	750	16	~35	n.d.	[52]
13%Ni/CeO2-ac10 Combustion method (red./oxid. ratio = 1.0)	750	750	18	~65	n.d.	[52]
13%Ni/CeO2-ac15 Combustion method (red./oxid. ratio = 1.5)	750	750	12	~90	n.d.	[52]

^a calcination temperature (°C). ^b reaction temperature (°C). ^c Initial methane conversion (%). n.d. not determined.

), it is observed that our results (46–72%) are within the reported range, which goes from around 37% to 75%. This shows that the catalysts developed in this study perform well under similar reaction conditions, even though they were calcined at a much higher temperature (1200 °C). The initial CO₂ conversions followed the same tendency as initial CH₄ conversion (Figure S6). However, the CO₂ conversions were higher than CH₄ conversions. The main products obtained were H₂ and CO (Figure 10). A small amount of water was also observed. Figure S7 shows the catalytic test of Ni/Ce with error bars.

The H₂/CO molar ratio values also had a similar behavior to that observed for the conversion of CH₄ during the reaction (Figure 10). Ni/Ce_0.2Pr_0.8 exhibited the highest H₂/CO ratio (0.89), which could be related to its highest initial CH₄ conversion. The lowest value of the H₂/CO molar ratio (0.65) was observed for the material with 50% of Pr.

Thus, all catalysts exhibited CO₂ conversion higher than CH₄ conversion and the H₂/CO molar ratios below 1.0, which is due to the reverse water gas shift reaction [53].

Regarding catalyst stability, continuous deactivation was observed over 24 h of time on stream (TOS) for Ni/Ce. In the case of the sample containing 20 and 80 mol% of Pr, after a slight loss of catalytic activity in the first few hours of reaction, the methane conversion remained practically unchanged.

For the sample containing 50 mol% of Pr, more pronounced initial deactivation was observed, followed by a gradual activation phase starting after 3 h of time on stream (TOS), which continued until the end of the reaction. Moreover, for the catalyst with 50 mol% Pr, which exhibited the most significant loss of catalytic activity, the deactivation was accompanied by an increase in CO and H₂O formation and a decrease in H₂ production (Figures S8 and S9). As the reverse water gas shift reaction is in equilibrium under operating conditions, the formation of CO increases with the loss of catalytic activity for reforming reactions [53].

Some authors [16,17] also observed a period of activation in the DRM reaction. Faria et al. [16] studied the DRM over Ni/Al₂O₃ and Ni/Ce_xZr_1–xO₂/Al₂O₃ (x = 0.5; 0.75; 1.0) catalysts. The activation period was observed for Ni/Al₂O₃ and Ni/CeO₂/Al₂O₃ catalysts. An XRD experiment after exposition of the Ni/Al catalyst to CO₂ stream was carried out in order to investigate the evolution of the phases present in the catalyst. After the exposition of the sample to CO₂, the line characteristic of Ni⁰ was no longer observed, whereas a line attributed to the NiO phase appeared. This result suggested that Ni particles were partially oxidized at the beginning of the reaction for Ni/Al and Ni/Ce/Al catalysts. Rabelo-Neto et al. [17] investigated the performance of catalysts obtained from perovskites (LaNiO₃, LaNiO₃/Al₂O₃ and LaNiO₃/CeSiO₂)_. for DRM. In situ XPS experiments carried out under reaction conditions showed that metallic Ni particles were oxidized by CO₂ in the feed for both LaNiO₃ and LaNiO₃/Al₂O₃ catalysts. Takanabe et al. [54] also reported a loss of activity at the beginning of the DRM at 750 °C over Co supported on TiO₂. This was attributed to the oxidation of metallic Co particles by the CO₂ of the feed, as revealed by XRD and XPS.

In the present work, to evaluate the oxidation of metallic Ni particles by CO₂ from the feed, XRD analysis was performed after exposing the Ni/Ce_0.5Pr_0.5 catalyst to a CO₂ stream for 1 h (Figure 11). The results showed that the characteristic lines of Ni⁰ were still present. However, the appearance of the lines was related to the NiO phase. Additionally, the diffraction line corresponding to the Ce₁₁O₂₀ phase was detected. These findings suggest that both the metallic Ni crystallites and the support were partially oxidized when exposed to a CO₂ stream.

Thus, in this work, the partial oxidation of Ni⁰ to NiO by CO₂ from the feed could contribute to the initial deactivation period observed for Ni/Ce_0.5Pr_0.5 during the DRM reaction. As hydrogen is formed during the reaction, the NiO particles are reduced, which could explain the observed activation period for this sample.

Some authors [55,56] studied the kinetics of the oxidation of metal particles and reported that the rate of oxidation increases with decreasing metal particle size. Then, the smaller metallic Ni particle observed for Ni/Ce_0.5Pr_0.5, when compared to the undoped sample and sample with 20 mol% of Pr, may explain why this behavior is more pronounced for this sample.

On the other hand, although the sample with 80 mol% Pr exhibited the smallest Ni⁰ particles, no activation period was observed. This may be attributed to its highest reducibility, as indicated by TPR analyses, which suggest that under reaction conditions, the support was preferentially oxidized, thereby preventing the oxidation of the Ni particles [18].

3.3. Characterization of Used Catalysts

In order to evaluate the carbon formation during DRM, TPO analyses were carried out after 24 h of TOS (Figure S10). All catalysts exhibited two main peaks of CO₂ at 594–619 °C and 821–830 °C. For Ni/Ce_0.2Pr_0.8, in addition to these peaks, CO₂ formation was also detected around 400–500 °C.

According to the literature [57], peaks of CO₂ at temperatures between 400 and 500 °C are related to the formation of amorphous carbon, while peaks above 700 °C are attributed to graphitic carbon. The formation of CO₂ at intermediate temperatures (from 600 °C) is attributed to carbon filaments [58]. Da Fonseca et al. [38] investigated the nature of carbon formed during DRM over Ni/CeO₂ catalysts. They performed TPO analyses of well-defined carbonaceous structures: amorphous carbon, multi-wall carbon nanotube (MWNT) and graphite. For the amorphous carbon, oxygen uptake was observed between 350 and 550 °C, with a peak at 538 °C. The TPO profile of MWCNT exhibited a peak at 600 °C, whereas the oxidation of graphite took place at high temperature (above 650 °C).

Thus, the results obtained in this work indicate a predominance of carbon nanotubes and the presence of graphitic carbon in all catalysts studied. In addition, amorphous carbon appears to be present in the Ni/Ce_0.2Pr_0.8 sample.

SEM images of Ni/Ce and Ni/Ce_0.2Pr_0.8 showed the formation of carbon filaments after DRM at 800 °C (Figure S11), corroborating the results obtained through TPO analysis.

Previous studies [38,59,60] have demonstrated a correlation between Ni particle size and the nature of carbon deposits formed during methane decomposition and DRM over Ni-based catalysts. These authors observed that amorphous carbon tends to form preferentially on smaller Ni particles, while carbon nanotubes are predominantly produced on larger ones. Da Fonseca et al. [38] reported that Ni/CeO₂ catalysts with Ni particle sizes larger than 10 nm mainly exhibited the formation of carbon nanotubes after DRM. Moreover, a small amount of graphitic carbon was also detected.

A similar trend was observed in this work. All catalysts exhibited large metallic Ni crystallite sizes (63–116 nm). Consequently, carbon nanotubes were the predominant form observed for all samples. The presence of amorphous carbon in the Ni/Ce_0.2Pr_0.8 sample could be associated with the formation of the Pr₂NiO₄ phase during calcination, as suggested by XRD analyses. Upon reduction, this phase could facilitate the generation of smaller metallic Ni particles, which may in turn favor the formation of amorphous carbon.

The amount of carbon formed over 24 h of TOS obtained for the catalysts is shown in

Table 4. Rate of carbon formation obtained by TPO analyses after DRM at 800 °C.

Catalyst	Rate of Carbon Formation mgC/(gcat·h)	Rate of Carbon Formation (mgC·gcat−1·h−1·molCH4conv−1)
Ni/Ce	0.33	0.23
Ni/Ce0.8Pr0.2	0.29	0.19
Ni/Ce0.5Pr0.5	0.20	0.15
Ni/Ce0.2Pr0.8	0.36	0.19

.

All samples showed a low rate of carbon formation between 0.20 and 0.36 mgC·g_cat⁻¹·h⁻¹. Table 3 also presents the carbon formation rates reported in the literature for various Ni catalysts supported on ceria [38,46,47,48]. These studies show that carbon deposition during DRM is influenced by multiple factors, such as Ni particle size, catalyst composition, synthesis method, and reaction conditions. Due to these variations, a direct comparison with literature data is not straightforward. To allow for more consistent analysis, only catalysts with similar initial methane conversions (between 66% and 75%) were considered. Within this range, carbon formation rates still vary significantly from as low as 0.4 mgC·gcat⁻¹·h⁻¹ (for 5%Ni/Ce_0.5Pr_0.5O₂ calcined at 750 °C) to as high as 200 mgC·gcat⁻¹·h⁻¹ (for a Ni/Ce–Al₂O₃ catalyst calcined at 600 °C).

The Ni/Ce_0.2Pr_0.8 catalyst developed in the present study showed a comparable initial CH₄ conversion (72%), while maintaining one of the lowest carbon formation rates (0.36 mgC·gcat⁻¹·h⁻¹), despite being calcined at higher temperature. These results demonstrate high resistance to carbon deposition under DRM conditions.

Given that the initial conversion values for the catalysts studied in this work varied among the samples, the carbon formation rate was normalized by the number of moles of methane converted (mol CH₄ conv) to allow for a more accurate analysis. When considering the carbon formation normalized in this way, it is evident that, despite the small amounts of carbon formed overall, the sample containing 50 mol% Pr exhibited the lowest carbon formation (0.15 mgC·g_cat⁻¹·h⁻¹. molCH₄conv⁻¹). The undoped sample and those containing 20 and 80 mol% Pr showed similar carbon formation values (0.23–0.19 mgC·g_cat⁻¹·h⁻¹·mol_CH₄conv⁻¹), which were slightly higher than that observed for the 50 mol% Pr sample.

According to the proposed mechanism for carbon formation on hydrocarbon reforming reactions [61,62], initially, highly reactive carbon species (C_α) are produced from the dissociation of methane on the nickel surface. C_α species can undergo gasification, forming CO. However, if the gasification rate is low or excess C_α species are formed, the carbon can undergo polymerization, generating C_β species. These species are less reactive and can lead to the occurrence of two processes: (i) encapsulation of the particle, causing the deactivation of the catalyst, and (ii) formation of carbon filaments, which occurs through the dissolution of the C_β species in the nickel particles. The formation of carbon filaments does not lead to an immediate deactivation of the catalyst, since the surface of the metallic particle remains free for the reaction. While filamentous carbon does not immediately block active sites, its progressive accumulation may lead to mechanical degradation (e.g., cracking or delamination of the catalytic layer) and hinder mass transfer by obstructing pore networks or displacing active metal particles. Therefore, controlling its formation is important to ensure long-term structural and catalytic stability.

The size of nickel crystallites plays an important role in the nucleation and growth of carbon layers. The formation of these layers is favored over a minimum number of atoms. Thus, there is a critical size for the nickel particles, above which the accumulation of carbon occurs. Then, to avoid carbon formation, the size of the Ni crystallites should be kept at a value lower than this critical size [63]. The presence of large Ni crystallites would favor carbon production.

Some authors [64,65] showed that very large metallic particles can also prevent the formation of carbon. Chen et al. [65] studied the behavior of Ni catalysts supported on hydrotalcites, CaAl₂O₄ and α-Al₂O₃ during the decomposition of methane. They observed that the formation of carbon is favored when the size of the metallic particles is between 20 and 40 nm. They also concluded that carbon production is inhibited for particles smaller than 10 nm and for very large particles. In addition, Toebes et al. [66] did not detect the formation of carbon filaments during the decomposition of methane, using bulk NiO.

Most studies that explore how the size of Ni particles affects carbon buildup during DRM [48,59,67,68] do not consider catalysts with very large particles (over 100 nm), which are typically found in catalytic layers of SOFC anodes [15].

Da Fonseca et al. [38] investigated the effect of the metallic Ni crystallite size on carbon formation during the dry reforming of methane under the same reaction conditions used in the present work. The results showed that the rate of carbon formation initially increases with Ni crystallite size, reaching a maximum around 20–30 nm, and then gradually decreases with further growth in the metallic particles, up to 133 nm. This behavior indicates the existence of a critical Ni crystallite size for carbon formation on Ni/CeO₂ catalysts, at which the highest carbon deposition occurs. In that study, Ni/CeO₂ catalysts with Ni⁰ crystallite sizes ranging from 10 to 132 nm were evaluated. It was observed that the carbon formation rate decreased from 5.9 to 0.0 mgC·gcat⁻¹·h⁻¹ as the crystallite size increased from 46 to 132 nm.

In the present work, the catalysts exhibited a large Ni⁰ crystallite size (63 to 113 nm) and low carbon formation rates (0.20–0.36 mgC·gcat⁻¹·h⁻¹). These values are consistent with the previously observed trend and reinforce the correlation between larger Ni crystallite sizes and reduced carbon formation under DRM conditions. Regarding the Ni/Ce_0.5Pr_0.5 sample, the minimum carbon formation observed suggests that the presence of NiO and Ni⁰ species, resulting from the oxidation of metallic Ni particles by CO₂, may also have contributed to the reduced carbon formation in this sample.

For catalysts with larger Ni crystallites, the CH₄ dissociation rate is likely lower, which may help suppress carbon accumulation [38]. However, the presence of oxygen vacancies in ceria does not significantly aid in removing carbon from the Ni surface due to the weaker metal–support interaction associated with larger Ni particles [38].

Since all samples exhibited low carbon formation, the most promising material for catalytic anodic layers in DIR-SOFCs fueled by CH₄ and CO₂ is the sample containing 80 mol% Pr, likely due to its higher methane conversion and greater stability during the catalytic tests.

4. Conclusions

This study investigated the effect of Pr content on the catalyst’s crystalline structure, support reducibility, and stability during DRM. The reducibility of the catalysts was enhanced by the increase in Pr loading. However, higher Pr content led to the formation of PrO_x, indicating phase segregation. Furthermore, increasing Pr loading favored the formation of Pr₂NiO₄, in addition to the NiO phase after calcination, especially in the sample with the highest Pr content. The reduction of Pr₂NiO₄ could have contributed to the decrease in the metallic Ni crystallite size in this catalyst.

DRM stability tests and the characterization of spent catalysts revealed that all catalysts exhibited very low carbon formation, likely due to the large metallic Ni particles. The lowest carbon deposition was observed for the sample containing 50 mol% Pr. However, this catalyst was not stable during the reaction, exhibiting significant deactivation followed by continuous activation. XRD analysis after exposure to a CO₂ stream suggests the occurrence of metallic Ni particle oxidation by CO₂ from the feed, which leads to the initial deactivation followed by activation observed for this sample. This behavior is consistent with the results obtained in the literature for Ni catalysts [16,17,54]. The addition of 80 mol% Pr led to higher methane conversion and improved catalyst stability. The lack of an activation phase in this sample is attributed to its enhanced reducibility, as evidenced by TPR analyses. These results indicate that, under reaction conditions, the support underwent preferential oxidation, which likely helped preserve the metallic Ni phase from oxidation, as previously reported in the literature for Ni supported on ceria-based oxides [16].

This study stands out from previous research by demonstrating that Ni supported on ceria doped with 80 mol% of Pr has potential as a catalyst layer for anodes in DIR-SOFCs fueled by CH₄ and CO₂ mixtures. Although the catalyst exhibited high performance and stability under DRM conditions in a fixed-bed reactor, its electrochemical behavior under actual DIR-SOFC operating conditions has not yet been evaluated. Future studies should include long-term stability tests under realistic feed compositions, as well as electrochemical testing in single-cell SOFC configurations to confirm the material’s suitability for practical applications.