LaMn₁-_xNi_xO₃ Perovskite Deposited on γ-Al₂O₃ Spheres as Catalyst for Dry Reforming of Methane

1. Introduction

The dry reforming of methane (DRM) occurring via reaction R1 is appealing because it can exploit CO₂ from carbon capture processes, producing syngas with high heating value and suitable for conversion into valuable chemicals, e.g. methanol. The side reaction (R2) of thermal or catalytic decomposition can likely occur when CO₂ is deficient, so that in the reactor the trade-off between the two reactions may take place [1].

CH₄ + CO₂ = 2 CO + 2 H₂    ΔH₀ = 247 kJ/mol

R1

CH₄ = C + 2 H₂        ΔH₀ = 74 kJ/mol

R2

DRM requires temperatures over 600 °C for thermodynamic constraints and heat supply, owing to the endothermic character of reaction R1 [2]. In this respect, peak renewable energy could be exploited in electrified reactors via an easy scalable and intensified process strategy [3,4]. Since the heterogeneous catalysis is strictly necessary to improve DRM kinetics at temperature between 600 and 800 °C, tubular reactors represent the widely adopted option for catalytic reforming of hydrocarbons on a medium plant scale [5]. The catalyst deposited on suitable, mechanically resistant substrates forming a structured catalyst in the form of granules, monoliths or foams, should be effective in exchanging the heat required for chemical reactions. In more advanced systems, such as fluidized-bed, greater thermal performance and easiness in catalyst regeneration can be obtained [6], at the expense of certain back-mixing of reactants and products [7].

Nickel based catalysts are the most diffused for industrial processes, having low cost, large activity and good robustness toward deactivation and carbon deposition [8,9]. Depositing Ni-based catalysts onto structured supports such as honeycomb monoliths significantly improves heat and mass transfer, reduces deactivation, and enables operation near thermodynamic limits, making them highly promising for industrial DRM applications [9,10]. However, nickel oxides are even more reported as toxic and carcinogenic [11], suggesting their substitution with safer elements, such as Cu, Ce, Mn, and Fe, although they are less effective in catalysis [12].

In this context, the total or partial replacement of nickel in catalysts is a current topic of the research, aiming at mitigating the loss of performance using the alternative elements by introducing complex multi-oxide systems, for instance perovskites and high entropy oxides [13,14].

Perovskites are versatile, cost-effective, and thermally stable materials characterized by a tuneable molecular structure [15]. In particular, LaNiO₃ perovskites are valuable for their good thermal stability and ability to undergo controlled reduction under reaction conditions. Upon reduction, they form finely dispersed metallic Ni particles strongly anchored to a La₂O₃-rich matrix, which enhances resistance to sintering and coke deposition. These features make LaNiO₃-derived catalysts effective in DRM, where the in situ–generated Ni provides high activity for CH₄ and CO₂ activation while the lanthanum oxide component promotes CO₂ adsorption and contributes to coke mitigation [16]. Thus, Ni containing perovskites exhibit improved catalytic stability due to strong metal–support interactions and dynamic structural reconstruction under reaction conditions [17]. When Mn is incorporated by substituting Ni or partially occupying the B-site of a perovskite structure, the physicochemical and catalytic properties of the catalyst are improved, leading to formation of smaller and more stable Ni nanoparticles after preliminary reduction. Moreover, Mn increases oxygen mobility and favours carbon coke gasification [18]. Therefore, mixed LaMn_1-xNi_xO₃ perovskites co-doped with Mn and Ni at the B site, are excellent candidates as catalysts for the dry reforming. However, limited results are currently available in the literature on the use of mixed perovskites supported on structured catalysts.

The determination of the elements in the mixed perovskite, apart from chemical analysis, such as inductively coupled plasma (ICP) mass or optical spectroscopy, can be roughly inferred, rather than accurately determined, from X-ray diffraction data. This is due to the small difference in their atomic numbers (i.e., Z = 25 and Z = 28 for Mn and Ni, respectively), which results in similar scattering factors. Accurate determination of the Ni/Mn ratio typically requires neutron diffraction data, owing to the large difference in the neutron scattering lengths (i.e., -0.373·10^-12 cm and 1.030·10^-12 cm for Mn and Ni, respectively). For this reason, Blasco and co-workers studied this family of perovskites also by neutron diffraction [19].

In the present research a mixed LaMn_1-xNi_xO₃ perovskite catalyst having a Ni content less than 60%wt, and Mn as substitute in B site was produced by solution combustion synthesis (SCS). The experimental activities covered various levels of development, including: i) synthesis of a LaMn_1-xNi_xO₃ perovskite powder from pure reactants, ii) physicochemical analysis by chemical and X-ray techniques, iii) soft deposition of the catalyst on γ-alumina supports, and iv) testing in fixed bed reactor of dry reforming under different operating conditions. The results and their discussion provided new insights into the exploitation of catalytic DRM with mixed perovskite at the basic and applied levels.

2. Methods

2.1. Materials and Physicochemical Characterization

A perovskite powder with a nominal stoichiometry LaMn_0.75Ni_0.25O₃ was synthesized via SCS starting from La(NO₃)₃·6H₂O (99.99%, Sigma-Aldrich), Mn(NO₃)₂·4H₂O (97%, Sigma-Aldrich) and Ni(NO₃)₂·6H₂0 (99.99%, Sigma-Aldrich) as metal precursors and citric acid as fuel. The combustion process was carried out with a molar ratio of fuel to total metal cations of 2 and a reducer-to-oxidizer ratio (φ) of 1.5. The pH of the solution was adjusted to 6 using a 28% (v/v) ammonia solution (Merck). NH₄NO₃ (98.0%, Sigma-Aldrich) was incorporated as an oxidizing agent. The mixture was stirred magnetically at 80 °C in a stainless-steel beaker until a uniform gel was obtained. Subsequently, the temperature was raised to around 350 °C, triggering a self-sustaining combustion reaction that yielded a dark-brown powder. The as-burned powder was calcined in static air at 900 °C for 5 h with a heating rate of 10 °C/min and sieved in the particle range d_p=250-400 μm. The as-obtained fresh product is hereafter labelled LMNO, with nominal x=0.25.

Batches of structured catalyst was prepared by soft dry mechanical deposition of the calcined and sieved LMNO powder in a rotating jar for some tenths of hours without milling balls and dispersing wet medium, on γ-alumina porous spheres (1.8 mm outer diameter). The rolling of the jar was repeatedly paused and restarted till the preset loading of perovskite was reached.

X-ray powder diffraction data were collected using a Rigaku RINT2500 rotating-anode diffractometer operating at 50 kV and 200 mA, equipped with a Rigaku D/teX Ultra silicon strip detector. Monochromatic Cu K_α1 radiation (λ = 1.54056 Å) was selected using an asymmetric Johansson Ge(111) crystal, ensuring high angular resolution suitable for ab-initio structure solution. The powders were finely ground and loaded into Lindemann glass capillaries with a diameter of 0.5 mm. Data were collected in transmission mode by spinning the capillary to minimize preferred orientation effects. The diffraction patterns were recorded over the 2θ range of 6–100° with a step size of 0.02° and a counting time of 2 s per step. These conditions allowed collecting high-quality data, suitable for reliable structure solution and subsequent Rietveld refinement.

Temperature programmed reduction (TPR) of LMNO fresh catalyst powder was carried out with Autochem 2950 HP system (Micromeritics, Norcross, GA, USA) equipped with a TCD detector. Before analysis, 0.1 g of the sample was pre-treated with a mixture of 5% O₂/He (v/v, 30mL/min) while heating up to 400 °C, 10 °C/min heating rate; afterwards, it was left at this temperature for 30 min. Upon cooling to room temperature, the gas mixture of 5%vol H₂/Ar (30 mL/min) was introduced into the sample tube, and the temperature increased up to 900 °C, 10 °C/min heating rate. The hydrogen consumption associated with the TPR profile was determined by peak integration applying a calibration curve, obtained at different H₂/Ar ratio. Specific surface area and pore volume were evaluated by N₂ adsorption–desorption analysis at -196 °C using an ASAP 2020 instrument (Micromeritics, Norcross, GA, USA). Prior to analysis, LMNO sample was degassed at 250 °C for 2 h under vacuum. The Brunauer–Emmett–Teller (BET) method was applied over the relative pressure range 0.05–0.30 (P/P₀), to calculate the specific surface area. Pore volume was determined from the desorption branch using the Barrett–Joyner–Halenda (BJH) method. ICP-IOS analyses were conducted with an Agilent 5100 instrument.

Thermogravimetric analysis (TGA) was performed using a TGA 1 Star System (Mettler Toledo, Schwerzenbach, CH). The sample was heated from room temperature to 100 °C and held at this temperature for 1 h. The temperature was then increased to 1100 °C at a heating rate of 10 °C/min under a constant air flow of 30 mL/min.

Raman spectra of the as-prepared fresh and spent catalysts were collected using a ThermoFisher Scientific (Waltham, MA, USA) DXR3 Raman spectrometer using a 532 nm laser at 0.8 mW power. A 50 μm slit and a 20× objective lens was used, with each spectrum acquired over 30 s and averaged from 10 scans. Microstructural characterization of the structured catalyst was carried out by means of a variable pressure SEM (Quanta 200, FEI): surface and fractured sections of the structured catalyst (as deposited and after DRM) were observed without conductive coating to prevent charging.

2.2. Dry Reforming of Methane

A simulation unit, consisting of an electrically heated fixed bed quartz microreactor, 12 mm ID and 200 mm length, was used for preliminary screening of DRM catalytic activity of LMNO powder with 40–60 mesh size diluted 1:5 by weight with silicon carbide of the same mesh size. To clean up the catalyst surface the sample was pre-treated in-situ under flowing O₂ (5% vol. He, 3.0 L/h) at 350 °C for 30'. Upon cooling to room temperature, the sample was reduced under flowing H₂ (5% vol. in He, 1.8 L/h) and increasing the temperature to 800 °C with a 10 °C/min ramp and a holding time of 60 min. The feed gas, consisting of 15% vol. of CH₄ + 15% vol. CO₂ in He, was led over the catalyst (100 mg) at a flow rate of 6.0 NL/h, and the catalytic test was carried out at 700 °C for 25 h on stream. The molar fraction Y of relevant species, were analyzed by GC (Agilent 7890B) equipped with a DB-1 capillary column and a molecular sieve, to follow the evolution of CH₄, CO, CO₂, H₂ and O₂ using Flame Ionization Detector (FID) and Thermal Conductivity Detector (TCD).

DRM tests with the structured catalyst were carried out in a laboratory plant consisting of a tubular reactor made in AISI 316 stainless steel, 21 mm ID, 300 mm length, 48 mL volume (Figure 1). The reactor was filled with 30 mL of structured catalyst comprised between two wads of ceramic wool at the ends. Electronic mass-flow controllers (Brooks SLA5850) supply gas streams from pressurized tanks of N₂, CO₂, CH₄ and H₂ to obtain the desired composition and flow rates of the feed flowing inside the reactor. A continuous gas analyser (Pollutek mod. 3160P) monitored and recorded the composition of the gas stream exiting the reactor, in terms of molar fraction Y of CH₄, CO₂, H₂ and CO. Each test consists in heating the reactor at preset temperature in flowing N₂, conditioning the structured catalyst at 800 °C in H₂/N₂ atmosphere (10%vol./90%vol.), starting the DRM test at fixed operating conditions, namely temperature, residence time and Y_CO2/Y_CH4 feed ratio. Before starting DRM, the reactor is purged with flowing N₂.

Data acquired from the gas analyzer were subsequently elaborated. Conversion ξ and selectivity c of some gaseous species were computed via Equations 1-4.

ξ_CO2 = (n_CO2,in - n_CO2,out) / n_CO2,in

Eq. 1

ξ_CH4 = (n_CH4,in - n_CH4,out) / n_CH4,in

Eq. 2

$${\sigma }_{\mathrm{CO}}=\mathrm{ }{\displaystyle \frac{{\mathrm{n}}_{\mathrm{C}\mathrm{O},\mathrm{o}\mathrm{u}\mathrm{t}}}{{(\mathrm{n}}_{\mathrm{C}\mathrm{H}4,\mathrm{i}\mathrm{n}}-{\mathrm{n}}_{\mathrm{C}\mathrm{H}4,\mathrm{o}\mathrm{u}\mathrm{t}})+{(\mathrm{n}}_{\mathrm{C}\mathrm{O}2,\mathrm{i}\mathrm{n}}-{\mathrm{n}}_{\mathrm{C}\mathrm{O}2,\mathrm{o}\mathrm{u}\mathrm{t}})}}$$

Eq. 3

$${\sigma }_{\mathrm{H}2}=\mathrm{ }{\displaystyle \frac{{\mathrm{n}}_{\mathrm{H}2,\mathrm{o}\mathrm{u}\mathrm{t}}}{2{(\mathrm{n}}_{\mathrm{C}\mathrm{H}4,\mathrm{i}\mathrm{n}}-{\mathrm{n}}_{\mathrm{C}\mathrm{H}4,\mathrm{o}\mathrm{u}\mathrm{t}})}}$$

Eq. 4

Where n_i,in and n_i,out are, respectively, the number of moles of gaseous species i entering and getting out the tubular reactor. The residence time τ was computed as the ratio between empty reactor volume V and volumetric flow rate Q₀ of the reactants at reactor temperature (Eq. 5).

τ = V / Q₀

Eq. 5

The gas hourly space velocity GHSV was computed based on LMNO mass deposited on the structured catalyst and volumetric flow rate of the mixture (Eq. 6).

GHSV = Q₀ / m_cat

Eq. 6

The feeding ratio r₀ was computed by Eq. 7, based on molar flow rate of CO₂ and CH₄ (Eq. 7).

r₀ = n_CO2 / n_CH4

Eq. 7

3. Results

3.1. Characterization of the LMNO Perovskite

A representative SEM micrograph of the calcined and sieved LMNO (Figure 2) clearly shows particle morphology of the synthetized powder along a rather submicron range.

Powder X-ray diffraction (PXRD) data of LMNO were collected to support further physicochemical and catalytic investigations. The crystal structure was determined via ab-initio structure solution process, while structural and microstructural parameters were optimized through Rietveld refinement. The crystal structure of fresh LMNO was successfully solved using the EXPO software package [20]. The structure model well agrees with literature on the La-Mn-Ni-O₃₊_δ series [19], which crystallized in the rhombohedral symmetry (space group R-3c). Rietveld refinement led to optimized unit cell parameters (hexagonal axes), giving the final refined values a=5.5179(9) and c=13.308(5) Å. The refinement procedure also included the site occupancy factor (SOF) for Mn and Ni atoms sharing the B-site, starting from an initial value of 0.5 for both. A. Even though the SOF values for Mn and Ni could not be accurately estimated if refined using X-ray diffraction data (see §1), it was chosen to refine them anyway because the Rietveld results gave useful insights: i) they confirmed, as expected, that the B-site was preferentially occupied by Mn rather than Ni; ii) the refined SOF values (i.e, 0.11(2) and 0.89(2) for Ni and Mn, respectively), were closer to the expected values (i.e., 0.25, 0.75) than the unrefined SOF values (i.e., 0.5, 0.5).

A representation of the asymmetric unit and the crystal packing is shown in Figure 3. The crystal architecture exhibits the characteristic framework of corner-sharing BO₆ octahedra. The Rietveld refinement results for LMNO are visualized in Figure 4, showing good agreement between observed and calculated profiles (the corresponding values of the agreement factors R_p and R_wp are 3.5%, 5.6%, respectively), which supports the reliability of the structural model.

The PXRD pattern of LMNO, as displayed in Figure 5, show the effect of reduction at 800 °C (red curve) and of the dry reforming of methane (post-DRM, blue curve), which is very similar to each other, while both patterns differ significantly from the fresh LMNO perovskite (black curve).

To have insights into the crystalline phase changes, a qualitative phase analysis of the diffraction patterns was carried out using the software QUALX2.0 [22], querying the database POW_COD and revealing the presence of the following three main phases in the reduced catalyst: (1) Lanthanum hydroxide (La(OH)₃, POW_COD entry #4031381), (2) Manganese oxide (MnO, POW_COD entry #1514105), (3) Nickel metal (Ni, POW_COD entry #2100646). The structure models for these phases were retrieved from the Crystallography Open Database (COD) as CIF files and used as starting models in EXPO for quantitative phase analysis. This approach allowed estimation of the weight fractions of the three phases in the sample (the majority phase was Lanthanum hydroxide). Additionally, the average crystallite size of each phase was estimated using the FullProf software package [22]. The fresh LMNO sample exhibited an average crystallite size of approximately 314 Å, whereas the quantitative phase analysis (QPA) results together with the average crystallite size of each phase for the LMNO at 800°C and after DRM are summarized in Table 1.

The phase identification results revealed the complete transformation of the perovskite structure upon high-temperature treatment at 800 °C and after DRM. These results are consistent with the expected structural transformations of LaNi–Mn perovskites under such reactive environments and provide important insight into the relationship between structural evolution and catalytic performance.

The weight fractions shown in Table 1, while providing a useful indication of the majority and minority phases among the main three identified ones, are probably affected by errors due to the possible presence of amorphous component(s) [23,24] and/or further unidentified minority crystalline phases that were not taken into account during QPA.

BET method gave a specific surface area of 11 m²/g for the LMNO powder, a low value typical for non-microporous materials. Pore volume, determined from the desorption branch using the Barrett–Joyner–Halenda (BJH) method, was about 0.05 cm³/g.

ICP analyses reported Mn and Ni content in the LMNO fresh powder equal to 17.1±0.7 and 5.6±0.3 %wt., respectively. Thus, the corresponding formula coefficient is x=0.23, very close to the nominal values 0.25.

In Figure 6 the reduction curve registered by TPR is displayed: the H₂-TPR profile of LMNO exhibits a multi-step reduction behavior characteristic of mixed B-site substituted perovskites, reflecting the progressive reduction of manganese and nickel species together with the removal of lattice oxygen. In the low-temperature region (≈200–400 °C), a broad reduction region with two main peaks was observed, which can be attributed to the reduction of higher oxidation states, namely Mn⁴⁺ to Mn³⁺ and Ni³⁺ to Ni²⁺, typically associated with surface or weakly bound oxygen species and more labile cation environments. Moreover, further reduction of Ni²⁺ to Ni⁰ likely occurred in this region [25]. The relatively low-temperature reducibility of Mn species, so far discussed, was enhanced by the presence of Ni which promotes oxygen mobility and defect formation within the perovskite lattice with respect to Ni-free LaMnO₃ counterpart, previously reported [26]. At intermediate temperatures (≈500–700 °C), a less intense and asymmetric contribution, similar to a shoulder, was detected corresponding to the reduction to the metallic state of more stable Ni²⁺ species well stabilized within the bulk perovskite lattice. In that region reduction of Mn³⁺ to Mn²⁺ cations occupying the B-site of the perovskite with local structural distortions may also take place [27]. Finally, the high-temperature region (≈700–900 °C) was dominated by an intense reduction peak, assigned to the final reduction of all Mn³⁺ to Mn²⁺ species with a stable local environment, likely in the bulk of the perovskite structure, requiring significantly higher activation energy. This step involves the removal of strongly bound lattice oxygen and is accompanied by the structural collapse of the perovskite lattice [25]. The overall experimental hydrogen consumption (110 ±5 mL H₂/g) was higher than the theoretical values expected for the complete reduction of Ni²⁺ to Ni⁰ (~25 mL/g) and Mn⁴⁺/Mn³⁺/Mn²⁺ (~76 mL/g) supporting the presence of some Ni³⁺ species along with the expected Ni²⁺ cations.

A DRM test was performed at 700°C for 25 h with LMNO powder in the simulation unit (Figure 7). The catalyst exhibited stable activity during the test duration, obtaining CO₂ and CH₄ conversion of around 80% and 70%, respectively. CO selectivity was 1.0 while H₂ selectivity was 0.97 because of a small amount of H₂O formation and congruently the H₂/CO ratio was lower than 1.

The catalyst extracted from the reactor after the prolonged test was subjected to TGA characterization for assessing possible carbon deposition. Figure 8 displays the TGA weight change versus temperature in oxidizing atmosphere. No weight loss occurred in the temperature range 600-800 °C that is typical of carbon oxidation, whereas a small increase appeared due to oxidation of reduced phases.

Raman spectroscopy was employed to probe the vibrational modes of the perovskite catalyst and to identify carbon species formed after the DRM reaction. As shown in Figure 9, LMNO as- prepared catalyst exhibited two characteristic bands at 526 and 657 cm⁻¹, which are typical of LaMnO₃-type perovskites [28]. The band at 526 cm⁻¹ is attributed to the bending vibration of Mn(Ni)O₆ octahedra, while the band at 657 cm⁻¹ corresponds to the stretching vibration of the Mn(Ni)–O bond associated with Jahn–Teller distortion of the Mn(Ni)O₆ framework.

After the dry reforming reaction, the Raman spectrum of the spent catalyst showed two additional very weak bands located at 1347 cm⁻¹ and 1583 cm⁻¹, corresponding to the D and G bands typically associated with carbon species [29]. However, thermogravimetric analysis (TGA) of the spent catalyst did not reveal any significant weight loss related to carbon oxidation, suggesting that the amount of carbon deposited during the reaction was negligible. Moreover, the Mn–O stretching band shifted from 657 cm ⁻¹ in the as-prepared LMNO to 643 cm ⁻¹ in the spent catalyst, indicating modifications in the local lattice environment under DRM conditions. The observed shift can be attributed to changes in the manganese oxidation state induced by reduction process occurring during the H₂ pre-treatment reduction, in agreement with the collapse of the original perovskite structure and the formation of MnO [30]. This interpretation is supported by both XRD results (Figure 5, Table 1) and H₂-TPR analysis, evidencing the structural evolution and reducibility of the material, respectively.

3.2. Structured Catalyst: Soft Deposition and Dry Reforming

After the characterization of the as-prepared fresh LMNO powder, the structured catalyst was investigated. It consisted of regular and porous commercial γ-alumina spherules, 1.8 mm outer diameter, coated with the LMNO catalyst, as displayed in Figure 10a.

In Figure 11 an example of a γ-alumina support depleted of catalyst (Figure 11a), and of a coated support before (Figure 11b) and after reforming test (Figure 11c) are shown. The analysis by SEM of the coated support confirmed that the external surface of the spherule is almost fully covered by the LMNO catalyst. Zones of accumulation conferred a surface morphology resembling that of a golf ball, while at a finer inspection scale the partly pristine dimpled appearance gets a more homogeneous texture. The texture of the catalyst coating was studied by SEM, analyzing fractured cross sections of the ceramic support. The thickness of the deposited catalyst was estimated by SEM in the range of 5-10 μm. In particular, the pronounced 3D view of the support in Figure 11c allows to emphasize some zones where the catalyst coating slightly detached after testing.

Figure 12 shows the molar fraction of CO₂, CH₄, CO and H₂ during a typical DRM test carried out at constant flow rate of reactants and three different temperatures. Stable values of all gas components were attained at each temperature; congruent increase of CO and H₂ and decrease of CH₄ and CO₂ with temperature can be noted.

The comparison of the results of three structured catalysts having different nominal content of perovskite, i.e. X=2.5, 5.0 and 7.5%wt., is shown in Figure 13 in terms of CH₄ conversion and H₂ selectivity. Both variables largely increased from X=2.5wt% to X=5.0wt%, whilst the increase was less pronounced, in particular for σ_H2, moving to the largest value X=7.5wt%. The method used for powder deposition led to the formation of an external coating that likely covered the external surface of γ−alumina supports when X exceeded a certain value, making further deposition of the catalyst less effective, as the case of X=7.5%wt. Thus, the value X=5.0%wt. was selected for systematic investigation of DRM in the laboratory plant.

The operating conditions and the results of dry reforming tests, in terms of CH₄ and CO₂ conversion, CO and H₂ selectivity are reported in Table 2. The tests were conducted at three temperatures (T = 700, 750 and 800 °C), different residence time τ and different CO₂/CH₄ feeding ratio r₀. The trends are fairly consistent with the parameter variations: increasing temperature and residence time leads to greater conversion and, conversely, H₂ selectivity increases with CO₂ conversion. The variation in the CO₂/CH₄ feed ratio is also congruently reflected in the decrease in the H₂/CO ratio produced when moving from catalytic decomposition to dry reforming conditions.

The trends of CH₄ and CO₂ conversion, and Y_H2/Y_CO ratio with the temperature are displayed in Figure 14, at constant residence time and CO₂/CH₄ feeding ratio. The increase in conversion of both reactants with T is well evident, and a slight parallel increase in the H₂/CO ratio is noted, as the reduced availability of CO₂ lowers the oxidation of H₂ to H₂O.

The increase of the residence time has a marked positive effect at T=800°C on the conversion of CO₂ and CH₄, as displayed in Figure 15, thanks to the decreased impact of kinetic and diffusion limits on the conversion of reactants at larger τ. In turn, the H₂/CO ratio exhibits a clear increase with the residence time, the H₂ oxidation being hindered as ξ_CO2 increases.

The tests carried out at increasing CO₂/CH₄ feeding ratio (r₀), at constant T and τ, shown in Figure 16a exhibited rather stable conversion of CO₂ close to 1.0, slight increase of CH₄ conversion thanks to major contribution of reaction R1 and decrease of H₂/CO ratio because in the investigated range r₀ = 0 ÷ 1.5 the system moves from CH₄ catalytic decomposition, in which no CO was produced, to dry reforming, with comparable production of H₂ and CO. While the selectivity was practically equal to 1.0 in the tests relating to Figure 14 and Figure 15 (Table 2), in this case (Figure 16b) there is a change of both selectivity values with r₀: σ_CO markedly increased and σ_H2 slightly decreased at larger r₀, because the augmented availability of CO₂ limited carbon deposition and conversely enhanced the oxidation of hydrogen to H₂O.

3.3. SEM Microstructural Analysis of the Structured Catalyst

Figure 17 and Figure 18 display, respectively, SEM micrographs from the structured catalyst before and after DRM. In particular Figure 17 shows how the cross-section of the structured spent catalyst’s coating appears: compared to Figure 17a (i.e., fresh catalyst), the primary particles of the spent catalyst (Figure 16b) underwent a slight grain coarsening though the porous character of the layer was basically retained. The comparison in Figure 17 is representative of the changes in size and shape of the primary particles, where rounded grains in the as-deposited (fresh) catalyst layer on the support progressively evolve into faceted particles. The comparison in Figure 17 also highlights differences in the thickness of the as-deposited catalyst layer: the range of 5–10 μm shows a characteristic “golf ball-like” appearance. (see Figure 11b,c). The imaging mode of Figure 17 enhances the particle morphology more than the average atomic number and its local difference. The corresponding images (not shown) acquired using the back-scattered mode did not show clear signature of coating/substrate interaction. To gain insight into finer microstructures of the spent catalyst, higher magnification portions of the coating were analyzed by SEM (Figure 18).

In particular, special emphasis was put on tiny sub-micrometric dots which were not observed in the as-deposited catalyst layer: some examples of these dots are marked in Figure 18a. The same area was imaged using the backscattered electron mode (Figure 18b): this BSE-SEM micrograph replicates well the porous nature of the coating where dark features are located, but it was not conclusive to assess a possible different chemical nature of such tiny dots which decorate the external surfaces of grains forming its substrate whose typical size in turn tends to fall into a micrometric range.

4. Discussion

The comparison between the tests performed at different scales shows better performances in the case of the quartz microreactor, due to differences in operating conditions and fluid-dynamics of the set-up. The conversions of CH₄ and CO₂ in Figure 7 are around 0.75 and 0.85, respectively, resulting larger than the values achieved in the reforming plant, i.e. ξ_CH4 = 0.61 and ξ_CH4 = 0.79, although the residence time was slightly different. The reduced size of the quartz microreactor and the dispersion of the LMNO perovskite in inert bed of SiC contributed to more efficient reactor operation in terms of temperature control and catalyst availability to reactants. Conversely, the soft deposition of the perovskite on γ-alumina supports, though it is less favourable for a more effective contact between reacting feeds and catalyst occurring over the support surface, it allows to operate with a pressure drop in the bed lower by at least one order of magnitude, according to the Ergun equation [31], being more realistic for an industrial application. The deposition was labelled “soft” because the adopted set-up did not foresee hard mixing media to favour adhesion to the γ-Al₂O₃ support.

A decay of catalytic activity was noted in both cases of quartz reactor and reforming: ξ_CH4 declined from 0.76 to 0.73 (in 20 h) and from 0.73 to 0.72 (in 30 min), respectively. Again, the structured catalyst was likely affected by a temperature gradient inside the reactor in the order of about ±50 °C which however was negligible in the microreactor. To recover (part of) the catalytic activity after DRM, the regeneration by mild oxidation in O₂ (10%vol.) at 800°C, followed by reduction under H₂ (10%vol.) proved to be effective for the structured catalyst.

The microstructural analysis using FullProf software [32] performed on the fresh LMNO catalyst and the two powder samples obtained after reduction at 800 °C and DRM revealed that the MnO phase (i.e., one of the three main crystalline phases resulting from the complete degradation of the fresh LMNO perovskite, see Table 1) was characterized by an average crystallite size larger than that of the original fresh perovskite (~ 314 Å). The increase in the MnO crystallite size can be ascribed to thermally induced sintering and growth, causing partial deactivation of the catalyst. XRD analyses performed on the powder residue collected from the structured catalyst showed diffraction profile similar to that of the pristine LMNO perovskite, confirming that the deposition process onto the spherical γ-Al₂O₃ support does not alter the perovskite structure. In contrast, XRD measurements carried out on the powder residue recovered after DRM experiments in the steel reactor revealed qualitative phase compositions consistent with the complete transformation of the initial LaMn_1-xNi_xO₃ perovskite into new main crystalline phases reported in Table 1.

Table 3 reports CH₄ and CO₂ conversion data, H₂/CO ratios, and selectivity to H₂ based on recent literature articles for Ni supported over Al₂O₃ [33] and LaNi₁₋ₓMnₓO₃ catalysts [24,34,35,36,37,38,39] . Benguerba et al. [33] reported CH₄ conversion equal to 0.91 at 650 °C during DRM in fixed bed reactor with Ni (33% wt.) supported over Al₂O₃, the better performance with respect to present catalyst (Table 2) being likely attributable to the much higher Ni content. As far as perovskites are concerned, although it is difficult to make a direct comparison, as the compositions of the reaction mixtures vary from case to case, a review of Table 3 shows that the LMNO catalyst under study exhibits CH₄ and CO₂ conversion values and an Y_H2/Y_CO ratio that compare very well or are even better than those of similar catalysts reported in literature.

A possible next stage of this research could involve a comprehensive evaluation of the developed structured catalyst for the dry reforming of heavier hydrocarbons and complex residual industrial streams. This line of research aligns closely with the principles of the circular economy, as it aims to transform waste materials into valuable products—such as hydrogen or synthesis gas—thereby reducing environmental impact and improving overall process sustainability. Attention could be paid to low-value by-products such as glycerol derived from biodiesel production, as well as other oxygenated or high-boiling compounds typically found in industrial waste streams [40]. Such an investigation would not only assess the catalyst’s activity, selectivity, and stability under more challenging operating conditions, but also explore its resistance to deactivation phenomena. By extending its application to these more demanding feedstocks, the catalyst’s versatility and robustness could be demonstrated.

5. Conclusions

Single phase LaMn_1-xNi_xO₃ perovskite catalyst was produced by solution combustion synthesis and subsequently treated in furnace at 900 °C. TPR characterization proved higher reducibility of the catalyst with overall hydrogen consumption higher than the theoretical values expected for the complete reduction of Ni²⁺ to Ni⁰ supporting the presence of some Ni³⁺ species along with the expected Ni²⁺ cations.

The structured catalyst obtained by soft mechanical deposition of LaMn_1-xNi_xO₃ perovskite powder over commercial γ-alumina spheres was effective in DRM, i) enabling an easy handling of the structured catalyst and its spatial distribution inside the reactor; ii) favouring the fluid-dynamics inside the reactor with limited pressure drop; and iii) achieving CH₄ conversion and H₂/CO ratio close to 1, at temperature of 800 °C with perovskite loading of 5.0%wt. The tests were performed in a tubular reactor, resembling real plug flow equipment with good reliability.

A decay of catalytic activity was noted after some hours of DRM operation. To restore the initial catalytic activity, the regeneration by mild oxidation in O₂ at 800°C, followed by reduction under H₂ atmosphere, proved to be quite effective. The phase identification results revealed the complete transformation of the perovskite structure upon high-temperature reduction at 800 °C and under DRM conditions.