Methane Dry Reforming Catalysts Based on Pr-Doped Ceria–Zirconia Synthesized in Supercritical Propanol

Abstract

This paper is devoted to the study of active and stable nickel catalysts for methane dry reforming based on Pr-doped ceria–zirconia obtained via the solvothermal continuous method. Studies on the physicochemical and catalytic properties of the 5%Ni/Ce_0.75Zr_0.25−xPr_xO₂ series have showed that Pr introduction leads to an increase in the amount of highly reactive oxygen in the oxide lattice. Praseodymium-based catalysts showed significantly higher reactant conversions. In addition to the nature of support, the method of nickel introduction was also studied; Ni was added both using impregnation and the one-pot procedure with mixed oxide preparation. The method of Ni addition was shown to have significant effect on the morphology of its particles and Ni-support interaction, and, respectively, on catalytic activity and coking stability. The 5%Ni/Ce_0.75Zr_0.15Pr_0.1O₂ catalyst prepared by one-pot method showed stable operation in the MDR reaction for 30 h at CO₂ and CH₄ conversions of ~40% and an H₂ yield of ~18% (T = 700 °C, τ = 10 ms).

1. Introduction

Methane dry reforming (MDR) is a promising way of utilizing two greenhouse gases—CH₄ and CO₂—and converting them into synthesis gas before being used as a feedstock for the chemical industry or as a fuel [1,2,3].

CH₄ + CO₂ = 2CO + 2H₂,     ∆H⁰₂₉₈ = 247.3 kJ/mole

(1)

This reaction can also be effectively integrated into existing hydrogen production chains, taking into account current environmental requirements [4]. MDR catalysts are representatives of typical catalysts for hydrocarbons subject to reforming processes, which consist of a transition metal, primarily nickel, deposited on an oxide support [5]. Studies over the past few decades show that poor stability due to the sintering and carbonization of active species remains a major problem for reforming catalysts [6]. Developing a thermally stable catalyst with superior activity and enhanced resistance to carbon formation and metal sintering is among the major challenges faced by industries and researchers [7,8,9].

The carbonization problem begins at the stage of competition between simultaneously occurring transformation routes of CH_x intermediates on metal particles. If CH_x oxidation occurs rapidly, the formation of desired CO is observed. However, if the rate of CH_x formation is higher than its oxidation, carbon deposits grow. A well-known and highly efficient approach to increase the rate of the desired route is to involve the lattice oxygen of the support into the catalytic cycle [10,11]. The transformation of fuels over catalysts with a high oxygen reactivity occurs via the redox reaction mechanism. During the reaction, the substrate (CH₄ and CO) is oxidized by the lattice oxygen, accompanied by the formation of a vacancy and its replenishment through the activation of the oxidizer (CO₂, H₂O, or O₂) on the oxide surface cations and fast surface oxygen transfer to the direct catalytic zone (reverse oxygen spillover) [12,13]. The activation of highly stable CO₂ molecules can also be a rate-limiting step in MDR, which also becomes an important factor in the effective catalyst design [14].

Establishing an unambiguous relationship between the characteristics of an oxide-supported catalyst and its activity in reforming reactions is still a debated issue. In steady-state reaction conditions, its properties differ from those defined by the preparation procedure [15]. Nevertheless, a huge number of works have been carried out and described in the literature based on a complex of catalytic structural studies and DFT calculations, which showed catalytic activity correlations with the so-called «support oxygen activity» [7,16,17].

The term “support oxygen activity” is complex, including several correlated characteristics of the oxide’s lattice oxygen, such as oxygen storage capacity (OSC), surface and bulk mobility, as well as the number and formation energy of oxygen vacancies. The redox properties of oxide catalysts can be related to the content of oxygen vacancies. However, the mere presence of vacancies is not sufficient; if the vacancy formation barrier is too high, this stage becomes rate-limiting, which casts doubt on the probability of the entire process proceeding along such a route. That phenomenon, for example, was shown and confirmed by DFT calculations for nickel-containing undoped cerium oxide [18]. In the steady-state conditions of the catalytic reaction, the concentration of oxygen vacancies is determined by the ratio of the rates of both oxygen removal with the reducing agent (CH₄, etc.) and reoxidation with the oxidant (CO₂, etc.) [15,19]. A decrease in the energy of vacancy formation can be achieved, for example, by increasing the length of at least one oxygen bond in the lattice or by creating so-called asymmetric vacancies [20,21].

Oxides that are capable of providing cycles of lattice vacancy formation with charge compensation due to cations with a variable valence are mainly transition metal oxides, which are seen as reducible or as oxides with active oxygen [22]. Among the oxides with lattice oxygen sites that take part in the reaction, the most common are perovskites, spinels, and fluorites. Groups of oxides based on ceria doped with a wide range of cations have been extensively studied [16,23,24,25,26]. Detailed kinetic and computational studies have shown that the activity of oxygen in such oxides depends on the concentration and formation energy of these vacancies [27,28]. Moreover, it was also shown that the presence and energy of vacancies on the surface of cerium oxide can determine the oxidation of the surface intermediate -CO into CO₂, thus largely determining the H₂/CO ratio in the resulting synthesis gas [29,30]. One of the most popular doping cations which also increases the thermal stability of CeO₂ oxide is Zr [31,32]. The praseodymium cation has a size similar to that of the cerium cation and is capable of replacing the latter in the oxide crystal lattice. In this case, micro-distortions are formed and the number of anion vacancies can potentially increase, which can favorably affect the catalytic activity. The effective generation of asymmetric vacancies in CeO₂-based oxides upon being doped with Pr, increasing the oxygen catalytic activity, has been shown in a number of studies [18,33,34,35]. Furthermore, the substitution of no more than 25% of cerium has been shown to be effective since the Pr extent would result in the clustering of oxygen vacancies [20,22].

The synthesis of these polycation oxides for catalytic application should provide a high chemical uniformity and oxide dispersion. Coprecipitation [36], hydrothermal method [37], solvothermal method [38], sol–gel–citrate [39,40] and ester polymer precursors (Pechini) [41], microemulsions [42], the microwave method [43,44], the sonochemical method [45], solution combustion [46,47], and spray pyrolysis reactions [48] are all known methods. The use of the solvothermal method for the synthesis of such polycationic oxides provides structural homogeneity, as well as improved textural characteristics and tunable particle size distributions [49]. Moreover, synthesis with the use of supercritical fluids is characterized by a lower energy consumption and simple implementation. The mechanism of oxide formation during hydrothermal synthesis includes two stages: the hydrolysis of metal ions and their precipitation in the form of oxides as a result of dehydration at high temperatures [50,51]. Supercritical alcohols are of interest as solvents because they have lower critical temperatures and pressures than water, and the conditions for reaching the supercritical state that vary depending on the nature of the alcohol. The use of supercritical alcohols contributes to a high rate of nucleation with the slight subsequent growth of the formed nuclei, which makes it possible to obtain highly dispersed oxide particles [52]. Thus, varying the temperature and pressure makes it possible to influence the nucleation rate and synthesize samples with desired characteristics. In our previous study [53], it was shown that synthesis in supercritical isopropanol makes it possible to obtain single-phase mixed Ce-Zr-(Ti/Nb) oxides, in contrast to the Pechini method.

It is important to emphasize the decisive influence of the nickel introduction method. In an oxide–metal system, the dispersion of supported metallic particles and the strength of its interaction with additional support are key properties that determine catalyst stability in reforming reactions on a par with the support oxygen activity [54,55,56]. It was shown that the traditional method of wet impregnation is less efficient than the method of nickel that is introduced into the oxide structure at the stage of support synthesis. In the second case, during the reduction pretreatment, finely dispersed metal particles strongly bound with support are formed [57,58]. At the same time, a part of the active component can remain in the bulk of support that is not accessible for reagents.

Hence, the use of zirconium as a doping cation to ceria increases the stability of the structure and increases the structure defectiveness, while doping with praseodymium increases the concentration of oxygen vacancies and the reactivity of the surface oxygen. Accordingly, in this study, our goal was to synthesize a series of complex metal–oxide catalysts 5%Ni/Ce_0.75Zr_0.25−xPr_xO₂ obtained using the solvothermal method with Ni added by both impregnation and the one-pot method. This study present results on the study of 5%Ni/Ce_0.75Zr_0.25−xPr_xO₂ textural and redox properties and their effect on catalytic activity and stability in the MDR reaction.

2. Materials and Methods

2.1. Synthesis Methods

The catalyst supports were synthesized in supercritical alcohol media using original installation described in more detail earlier [53,58]. Cerium nitrate Ce(NO₃)₃·6H₂O (Vekton, Petersburg, Russia), praseodymium nitrate Pr(NO₃)₃·6H₂O (Vekton, Russia), and zirconium butoxide (80 wt% in n-butanol, Alfa Aesar, Karlsruhe, Germany) were dissolved in required proportions in isopropanol (Reakhim, St. Petersburg, Russia). Then, the solution was fed into the reactor and synthesis was carried out according to the method described earlier [53]. For impregnated catalysts, nickel was supported by the incipient wetness impregnation of supports with the water solution of Ni(NO₃)₂ (Vecton, Russia) [54].

For one-pot catalysts, nickel nitrate Ni(NO₃)₂·6H₂O (Vecton, Russia) was dissolved together with Ce and Zr (Pr) salts, and synthesis in a supercritical medium was carried out, according to the procedure described above. All catalysts were dried at 200 °C and calcined at 700 °C for 2 h. The loading amount of Ni was 5 wt% both for the impregnated and one-pot catalysts. The prepared sample denotations are presented in

Table 1. The sample’s designation, specific surface area (SSA), pore volume (V_pore), and Fm3m lattice parameter (a).

Designation	Ni Addition Method	Composition	SSA, m2/g	Vpore, cm3/g	a *, Å
CZP	-	Ce0.75Zr0.15Pr0.1O2	14	0.055	5.403(1)
N/CZP	Impregnation	5%Ni/Ce0.75Zr0.15Pr0.1O2	9	0.095	5.397(1)
N+CZP	One-pot	5%Ni+Ce0.75Zr0.15Pr0.1O2	14	0.081	5.407(1)
CZ	-	Ce0.75Zr0.25O2	29	0.150	5.369(1)
N/CZ	Impregnation	5%Ni/Ce0.75Zr0.25O2	21	0.144	5.373(1)
N+CZ	One-pot	5%Ni+Ce0.75Zr0.25O2	13	0.126	5.372(1)

*—lattice parameter of the fluorite phase estimated from XRD data.

.

2.2. Characterizations

The specific surface area (SSA) was defined by the BET method using a Quadrasorb evo (Quantachrome Instruments, Boynton Beach, FL, USA) installation. The pore volumes and pore size distributions were determined from the desorption branch of the isotherm using the BJH method.

XRD analysis was performed on a Thermo X’tra diffractometer, in the angle range of 20–85° with a step of 2θ = 0.02° and a speed of 1°/min, using a Mythen2R 1D linear detector (Decstris, Switzerland). The CuKα radiation (λ = 1.5418A) was used. The average sizes of coherent scattering regions (CSRs) were calculated using the Scherrer formula from the 111 fluorite reflection. A description of the diffraction profile was made using the Fityk program with the Lorentz function.

XRD with H₂ in situ analysis was conducted using a Bruker D8 Advance diffractometer (Germany), in the angle range of 20–55° with a step of 2θ = 0.05° and an accumulation time of 3 s at each point using a LynxEye (1D) line detector. The monochromatic CuKα radiation (λ = 1.5418Å) was used. The measurements were carried out using an XRK-900 high-temperature reactor chamber (Anton Paar, Graz, Austria). The 10%H₂/90%He mixture at a flow rate of 100 mL/min was passed through the chamber during heating and cooling to room temperature. Heating was carried out based on the following scheme: 250, 315, 400, 500, 700, and 30 °C, with a heating rate of 12°/min. The lattice parameters and phase relationships were refined with the Rietveld method [59]. In the refinement, the CeO₂ structure was used; only the Ce atoms were in the cationic position.

The Raman spectrometer T64000 (Horiba Jobin Yvon, Kyoto City, Japan) with micro-Raman setup was used to record the Raman spectra. All experimental spectra were collected in the backscattering geometry using the 514.5 nm line of an Ar⁺ laser. The spectral resolution was not worse than 1.5 cm⁻¹. The detector was a silicon-based CCD matrix, cooled with liquid nitrogen. The power of the laser beam reaching the sample was 2 mW. The band at 520.5 cm⁻¹ of the Si single crystal was used to calibrate the spectrometer.

TEM with EDX. Transmission electron microscopy (TEM) micrographs were obtained with a Themis-Z 3.1 instrument (TFS, Durham, NC, USA), equipped with a X-FEG-monochromator and a CS/S double corrector (with an accelerating voltage 200 kV), as well as a JEM-2200FS transmission electron microscope (JEOL Ltd., Akishima City, Japan, Japan, acceleration voltage 200 kV, lattice resolution ~1 Å), equipped with a Cs corrector. Elemental analysis was performed with a Super-X EDS detector (with an energy resolution of about 120 eV) in HAADFSTEM mode. Samples for the TEM study were prepared for ultrasonic dispersion in ethanol and for the subsequent deposition of suspension in a “holey” carbon film supported on a copper grid.

TPR-H₂. Temperature-programmed reduction using hydrogen was carried out from 25 °C to 900 °C in a flow installation with a Tsvet 500 (JSC Tsvet, Kostroma, Russia) thermal conductivity detector using 10 vol. % H₂ in Ar feed at a flow rate of 40 mL/min.

The total H₂ consumption was calculated by integrating the area under the curve.

Ni reducibility was calculated with the formula:

$$\mathrm{R}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{i}\mathrm{b}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{t}\mathrm{y}=\frac{{\mathrm{H}}_2\left(\mathrm{c}\mathrm{a}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{y}\mathrm{s}\mathrm{t}\right)-{\mathrm{H}}_2\left(\mathrm{s}\mathrm{u}\mathrm{p}\mathrm{p}\mathrm{o}\mathrm{r}\mathrm{t}\right)·0.95}{0.85}·100\%$$

where H₂ (catalyst) and H₂ (support) [mmol H₂/g]—the amount of hydrogen consumed for catalyst and support reduction, respectively; 0.95—the support content in the catalyst; 0.85 [mmol H₂/g]—the theoretical amount of H₂ needed to reduce all supported Ni.

2.3. Catalytic Tests

The catalysts were preliminary reduced in a stream of 5% H₂ in He at 600 °C for 1 h. Catalytic activity in methane dry reforming reaction was studied using a tubular quartz plug flow reactor and a gas analyzer with IR sensors for CO, CO₂, and CH₄, and an electrochemical sensor for H₂ (Boner LLC, Novosibirsk, Russia). Studies were conducted with an initial mixture of 15% CH₄ + 15% CO₂ + 70% N₂ in the temperature range of 600–750 °C and a contact time of 10 ms. Long-term stability tests were carried out for 30 h at 700 °C.

The equilibrium values of reactant conversions and product yields were calculated by minimizing the Gibbs energy.

3. Results

3.1. Textural and Structural Features

Table 1 lists the designation, chemical composition, and specific surface area (SSA) of the obtained samples.

As can be seen, the samples obtained did not have a very high specific surface area; the values of SSA for the entire series of samples were about 10–30 m²/g, despite the fact that the nitrogen adsorption–desorption isotherms (Figure 1a) for all samples corresponding to the IV type were characteristic for mesoporous materials [60,61]. The introduction of praseodymium reduced the specific surface area of the initial support from 29 m²/g for CZ to 14 m²/g for CZP. This correlates well with data on the pore’s size distribution for supports (Figure 1b), where the number of mesopores less than 5 nm for praseodymium-doped sample was significantly reduced. In contrast, for samples prepared by the one-pot method (N+CZ and N+CZP), there was no difference in the SSA values. The pore volume was also reduced with the introduction of praseodymium from 0.126–0.150 cm³/g for undoped samples to 0.055–0.095 cm³/g for samples with Pr. The micropore (D ≤ 2 nm) volume was negligible for all samples and did not exceed 0.003 cm³/g.

According to XRD data (Figure 2), for the entire series, the main intense peaks located at 2θ = 28.7, 47.7, 56.6, 59.4, 69.8, 77.0, and 79.4° corresponded to 111, 200, 220, 311, 222, 331, 420, and 422 reflections of the fluorite structure Fm3m (PDF 81–0792). There were no reflections of Zr and Pr individual oxides for the entire series. For the Pr-containing samples, the absence of double peaks and the shift in the fluorite peaks to the lower angles’ region (Figure 2, Table 1) suggest that praseodymium was embedded in the CZ cubic structure, which is consistent with data in the literature [25,62]. The ability of praseodymium to replace cerium without destroying the fluorite structure was ensured by the close radii of cations of the same valencies (Ce⁴⁺ = 1.110 Å; Pr⁴⁺ = 1.100 Å; Ce³⁺ = 1.283 Å; Pr³⁺ = 1.266 Å), and a slightly higher Pr⁴⁺/Pr³⁺ redox potential led to the formation of a structure with more active oxygen compared to undoped ones [62,63]. For N/CZ, N/CZP, and N+CZP, small peaks located at 2θ = 37.47, 43.36, and 62.83° corresponding to 111, 200, and 220 reflections of NiO (PDF 471049) were detected. It is worth mentioning that for N+CZ, a slight increase in the background was observed in the positions of 111 and 200 reflections of NiO, indicating the presence of highly dispersed particles of nickel oxide. The crystallite size of NiO (the average size of a coherent scattering region (CSR)) was estimated using the Scherrer equation for 200 peak analysis. The CSR values of NiO for N/CZP, N/CZ, and N+CZP were 23, 27, and 10 nm, respectively. Hence, one-pot synthesis provided smaller crystallites of NiO than the impregnation route.

Raman spectroscopy was used to investigate vibrations in the oxygen lattice and crystalline symmetry in addition to XRD analysis, since this method is more sensitive to the microstructure of materials. Figure 3 presents the Raman spectra of the investigated samples. The main fluorite structure band in pure ceria was located at 464 cm⁻¹ and corresponded to the F_2g active mode that appears due to the symmetric stretching vibration of O anions around each Ce⁴⁺ cation [40]. This band was very sensitive to any changes in oxygen sublattice; therefore, with a change in the chemical composition or defect’s concentration, the band shifted to lower or higher energies. For our samples, the incorporation of smaller Zr⁴⁺ cation led to the F_2g band being shifted to higher wave numbers, while the incorporation of close-size praseodymium shifted it to lower wave numbers, which is in good agreement with data in the literature and also confirms the incorporation of doped cations in the fluorite lattice [40,62]. Such a strong shift in the case of Pr cation was probably connected with the increased structure distortion [62].

For catalysts obtained by Ni impregnation (N/CZ and N/CZP), the F_2g band practically retained its position. However, Ni addition at the stage of mixed oxide preparation (N+CZ and N+CZP samples) resulted in the incorporation of Ni cations in the fluorite structure and further shifts in this band. In addition to the F_2g mode, there was a band around 570 cm⁻¹ (D band) corresponding to the non-degenerate longitudinal optical (LO) mode due to the presence of defects in the fluorite lattice, resulting from the formation of oxygen vacancies [62,64]. The ratio of D and F_2g band areas ω = D/F_2g can be considered as a parameter that is sensitive to structure defectiveness and the amount of oxygen vacancies [40,58,62]. Its values for investigated samples are presented in Figure 3 in black frames.

It can be seen that ω values increased significantly with the introduction of praseodymium (more than 50 times for supports and 4 times for impregnated catalysts). This confirms that Pr incorporation leads to increased amounts of oxygen vacancies in the fluorite structure. The introduction of nickel with any of the methods used led to an increase in the number of vacancies due to its partial incorporation into the fluorite lattice, and this effect was especially noticeable for both one-pot catalysts, for which ω values were close. The low-intensity band at 230 cm⁻¹ for one-pot N+CZ could be assigned to the transverse acoustic (TA) mode, which was also connected with defects in the fluorite structure [40,64].

The method of adding nickel had a strong effect on the oxide’s morphology. XRD (Figure 2) and TEM data (Figure 4) show that the deposition of nickel via wet impregnation on the formed oxide resulted in the formation of large well-crystallized NiO particles (about 20 nm in size, weakly bound to the support). For these samples, there were peaks at 2θ = 37.4, 43.3, and 62.8°, which could be associated with the 111, 200, and 220 reflections of nickel oxide NiO (Fm3m). The introduction of nickel in the supercritical conditions led to the incorporation of a significant amount of nickel into the fluorite structure and the formation of residual finely dispersed NiO particles firmly fixed on the oxide surface or encapsulated into it.

The images of HAADF-STEM with EDX analysis show the distribution of elements in the particles of the one-pot N+CZP catalyst (Figure 5). As in our previous works [58], the synthesis of one-pot catalysts under supercritical conditions allowed us to obtain a homogeneous distribution of cations. Ce, Zr, and Pr cations of the support, as well as Ni cations, were uniformly distributed in the structure of Ni+Ce_0.75Zr_0.15Pr_0.1O₂.

3.2. Reducibility and Oxygen Reactivity

The oxide’s oxygen activity was estimated by a combination of complementary methods—TPR-H₂ and XRD—with in situ H₂ reduction. The TPR-H₂ results are presented in Figure 6 and

Table 2. Hydrogen consumption calculated from TPR-H₂ results.

Sample	H2 Consumption, mmol H2/gcat	Ni Reducibility, %
CZP	1.21	-
N/CZP	2.14	117
N+CZP	1.84	81
CZ	1.36	-
N/CZ	2.21	108
N+CZ	2.03	87

.

The curves of two supports (CZ and CZP) represent two standard reduction regions for these oxides: the medium-temperature region at 300–600 °C, related to the reduction of more reactive surface oxygen, and the high-temperature region above 700 °C, corresponding to the reduction of bulk oxygen, limited by internal diffusion. It can be seen that for the Pr-doped support, both the start of reduction and the maximum of the first peak shifted by about 200 °C towards lower temperatures, which indicates the significant facilitation of the reduction of surface oxygen with doping.

However, significant changes in the nature of reduction were noted with the introduction of praseodymium and nickel to CZ. For the N/CZP- and N/CZ-supported samples, the most intense sharp peak was located at 250 °C for the Pr-doped one and was shifted by 70 °C relative to the undoped sample, reaching its maximum at 320 °C. In situ XRD data (Figure 7b,c) show that at this temperature, there was an intense change in the fluorite parameter and no formation of metallic nickel particles was observed. This allowed us to attribute this peak to the reduction of the active oxygen of the oxide, located at 435 and 570 °C for the CZP and CZ supports, respectively. A shift in this peak to lower temperatures for Ni-containing samples was observed in most studies and was associated with the facilitation of oxide reduction due to hydrogen spillover in the presence of nickel atoms. Calculating the total amount of hydrogen consumed over the entire temperature range shows that the Ni reducibility (Table 2) for this samples exceeded 100% because the formation of metallic nickel particles led to an increase in the depth of carrier reduction due to hydrogen spillover [65].

For the one-pot N+CZP and N+CZ samples, reduction occurred more uniformly in the range of 200–400 °C and the curves had a complex contour because of the simultaneous reduction of fluorite and nickel from the oxide, which is more strongly associated with the support. Lower Ni reducibility for one-pot samples also indicated a strong interaction of nickel with the support and its partial dissolution in the volume of fluorite.

XRD with in situ H₂ reduction (Figure 7a–c) confirmed that, after reduction at 400 °C, a significant amount of metallic nickel (2θ = 44.2°) was detected for all nickel-containing samples. Moreover, in the temperature range of 400–500 °C, the intensity of nickel reflections for supported catalysts was higher compared to that of one-pot samples, which is consistent with the TPR-H₂ data and the assumption that nickel is partially incorporated into the fluorite lattice in one-pot samples. According to XRD with in situ H₂ reduction data (Figure 7a), the phase composition of the CZP sample did not change under reductive conditions. However, the lattice parameter of oxide increased from 5.403(1) to 5.474(1)Å during heating from RT to 700 °C, while the parameter did not return to the initial value after cooling and was equal to 5.428(1)Å (Figure 7d). An increased value of the lattice parameter can probably be explained by the partial reduction of cations. In the case of N/CZP and N+CZP, a similar effect was observed. The differences in the lattice parameters between the initial state and after reduction were 0.027 Å and 0.037 Å. Such an evolution of lattice parameters could be explained by the exsolution of Ni cations from the structure of a solid solution (Ce⁴⁺ = 1.110 Å; Pr⁴⁺ = 1.100 Å; Ce³⁺ = 1.283 Å; Pr³⁺ = 1.266 Å, Ni²⁺ = 0.69 Å) and/or differences in the reducibility of catalysts prepared by the one-pot method and impregnation. Interestingly, the behavior of lattice parameters differed for CZ and CZP catalysts, indicating the role of Pr addition. For N/CZ and N+CZP, the evolution of fluorite cell parameters almost coincided with temperature. In the initial state, the lattice parameter of the mixed oxide was 5.372(1) Å, and after reduction it was 5.394(1) Å, i.e., the difference between the states was 0.022 Å.

XRD with in situ H₂ (Figure 7d) reduction also showed a non-linear change in the lattice parameter in the temperature range of 200–400 °C, where hydrogen interacted with the most reactive oxide’s oxygen. Stronger parameter changes in this region for Pr-containing catalysts compared to undoped ones confirmed the facilitation of the release of lattice oxygen obtained by the TPR-H₂ method with Pr doping. A noticeable overstating value of the lattice parameter at 300–400 °C coincided with the appearance of nickel metallic reflections on XRD patterns, and indicated a greater depth of reduction due to the well-known phenomenon of hydrogen spillover described in the literature [65,66,67,68]. It is also important to note that after returning to room temperature, the Fm3m fluorite monophase was preserved with an increased parameter; hence, the release of oxygen occurred without destroying the structure.

3.3. Catalytic Activity in MDR

The results of MDR catalytic experiments are shown in Figure 8. Before testing, a blank experiment was carried out with an equivalent to the catalyst’s amount of quartz; no reactants were converted and no products were formed over the entire temperature range. A carbon balance calculated for all experiments was not lower than 95%. For all samples, an increase in the conversions of methane and carbon dioxide was observed with increasing temperature in accordance with the endothermicity of the reaction.

The activity of Pr-containing catalysts N/CZP and N+CZP exceeded the activity of samples based on undoped CZ support, and the difference significantly increased with an increase in temperature. This is consistent with literature data describing an increase in the contribution of support to the reaction with the temperature increase due to the more efficient activation of CO₂ and the transfer of active oxygen forms over the oxide surface to the metal–support interface [29,30].

At the same time, the method of introducing nickel did not noticeably affect the activity in experiments with temperature variation, but strikingly determined stability in the long-term tests. For the N/CZP sample obtained by wet impregnation, reactor plugging occurred within 1 h of a long test at 700 °C (not shown on Figures) due to carbon formation. Figure 8f shows data on the H₂ and CO yields in long-term tests for the N+CZP one-pot catalyst. After 30 h of the reaction, the hydrogen yield decreased from 35 to 32% and the CO yield decreased from 19 to 17%, and the main drop occurred for the first 3 h of the reaction. The catalyst after 30 h of reaction was investigated by TPO and was shown to be free of carbon.

The TEM studies of catalysts after catalytic tests in the temperature range of 600–750 °C showed that the N/CZP sample was characterized by whisker carbon deposits and weak bonds between nickel and the support resulted in the detachment of nickel particles from the surface, while the N+CZP sample demonstrates that metal particles were firmly fixed on the oxide surface and that an insignificant number of thin layers of carbon were directly fixed on the metal particles, which do not block active catalytic centers (Figure 9).

Hence, the catalytic activity of samples modified with praseodymium was significantly higher than that of the unmodified ones. An increase in the conversion of both methane and CO₂ correlated with an increase in the activity of the carrier’s highly reactive oxygen, as shown by the TPR-H₂ method. Moreover, as was shown by Raman spectroscopy, Pr-doped samples contained a higher number of defects and oxygen vacancies (i.e., sites for CO₂ activation). At the same time, stability of the catalyst was dramatically affected by the strength of the interaction of resulting nickel metal particles with the oxide support. Thus, for samples obtained by wet impregnation, the weak bond with the support, shown by TEM and TPR-H₂ methods, eliminated all the advantages of the active support and prevented the effective occurrence of the redox mechanism, leading to rapid deactivation with the formation of a significant amount of fibrous carbon deposits, as confirmed by TEM images. For samples with nickel introduced at the stage of support formation, high stability after 30 h of the reaction at 700 °C and the absence of carbon deposits were shown.

A comparison of data on the catalytic activity in the DRM reaction with literature analogues under similar conditions (

Table 3. A comparison of the methane conversion (X(CH₄)) obtained at 700 °C with literature analogues.

Catalyst	Reaction Conditions	τ, ms	X(CH4) at 700 °C, %	Ref
Ni/Ce0.5Zr0.5O2	CH4 = 20% CO2 = 20% N2 = 60%	120	15	[25]
Ni/Ce0.75Ti0.05Nb0.05Zr0.15O2	CH4 = 15% CO2 = 15% N2 = 70%	10	30	[53]
Ni/ZrO2	CH4:CO2:N2 = 3:3:1	~90	56	[69]
Ni/ZrO2-Al2O3	CH4:CO2 = 1	~150	38	[70]
5%Ni/Y+Zr	CH4:CO2:N2 = 3:3:1	~90	67	[71]
5%Ni/La+Zr	CH4:CO2:N2 = 6:6:1	~90	66	[72]
Ni/Ce0.75Pr0.1Zr0.15O2	CH4 = 15% CO2 = 15% N2 = 70%	10	39	This work

) indicates the significant potential of the praseodymium-modified sample synthesized under supercritical conditions.

4. Conclusions

In this work, catalysts series 5%Ni/Ce_0.75Zr_0.25−xPr_xO₂ (x = 0; 0.1) were successfully synthesized by the solvothermal method in a flow-through setup in supercritical isopropanol. All samples were found to be crystallized in the structural type F_m3m of cubic fluorite with the incorporation of praseodymium and zirconium cations into the lattice. The addition of Pr to mixed ceria–zirconia led to increased reducibility during TPR-H₂ experiments and stronger parameter changes during in situ XRD in hydrogen, which together confirm the higher reactivity of lattice oxygen. The introduction of nickel into the catalyst composition directly at the stage of support synthesis (one-pot) allowed a significant amount of nickel to be incorporated into the fluorite structure, leading to both an increase in the defectiveness of the oxide and the formation of highly dispersed particles of metallic nickel, strongly bound to the support oxide during reduction pretreatment. The introduction of Pr into the oxide support significantly increased the activity of catalysts in syngas production, which is most likely associated with facilitating the oxidant molecule activation, thus promoting the redox mechanism of MDR. The use of Pr-doped ceria–zirconia as a support made it possible to achieve excellent stability due to the high dispersion of metal particles, the optimal strength of the metal–support interaction, and lattice oxygen activity in reactions of oxidant activation and the gasification of carbon precursors. The 5%Ni/Ce_0.75Zr_0.15Pr_0.1O₂ catalyst showed stable performance in the MDR reaction for 30 h at a CH₄ conversion ~40% and a H₂ yield ~ 20% (T = 700 °C; τ = 10 ms).