CO₂ Valorization by CH₄ Tri-Reforming on Al₂O₃-Supported NiCo Nanoparticles

Abstract

CO₂ valorization from real feedstocks through CH₄ tri-reforming (CH₄-TR), combining steam reforming (SR), dry reforming (DR), and partial oxidation (CPO) of methane in a single process, is a desirable strategy for greenhouse gas mitigation and syngas (CO + H₂) production. NiCo/γ−Al₂O₃ catalysts prepared by impregnation at different relative metal contents (Ni₅₀Co₅₀ and Ni₃₀Co₇₀) were investigated for CH₄-TR in a fixed-bed reactor under conventional heating and characterized by XRD, FESEM, and Raman spectroscopy after catalytic runs. This study focused on the role of the Ni/Co ratio and feed composition on selectivity for CO₂ valorization, syngas yield, and deactivation resistance. Both the catalysts showed high activity, with a superior performance of Ni₅₀Co₅₀ confirming Ni metal species as the active sites. While in DR, a slow deactivation occurred due to coke deposition, in CH₄-TR, the addition of small O₂ and/or H₂O amounts stabilized activity and selectivity due to surface carbon removal. Large O₂ and H₂O amounts strongly inhibited CO₂ conversion due to competition with CPO and SR, in the order CPO ≥ DR > SR. Interestingly, the stoichiometric CH₄-to-oxidants ratio favored the DR pathway, giving very high CO₂ conversion. Modulating CH₄ addition into real flue mixtures renders CH₄-TR on NiCo/γ-Al₂O₃ catalysts a favorable strategy for effective valorization of CO₂ industrial or biomass-derived streams.

1. Introduction

CO₂ valorization using real feedstocks, including industrial exhaust gases from power plants and renewable biogas from organic waste, is a desirable approach to mitigate the concentration of this greenhouse gas in the atmosphere, thereby integrating the concepts of carbon emission reduction, reuse, recycling, and the circular economy [1]. Among the catalytic processes for the conversion of CO₂ into value-added products, the tri-reforming (TR) of CH₄ is expected to contribute significantly to the development of renewable energy carriers such as syngas (CO + H₂), accompanied by a reduction in CO₂ emissions [2,3].

The tri-reforming process combines, within a single reactor, steam reforming (SR: CH₄ + H₂O ⟶ 3H₂ + CO), dry Reforming (DR: CH₄ + CO₂ ⟶ 2H₂ + 2CO), and catalytic partial oxidation of methane (CPO: CH₄ + 1/2O₂ ⟶ 2H₂ + CO) reactions to produce syngas [4], using gas mixtures containing CO₂, H₂O, and O₂ co-fed with natural gas. Compared with the individual reactions, the integrated CH₄-TR process brings some advantages by improving the sustainability of the endothermic CH₄ SR and DR reactions due to the CPO exothermicity and by minimizing the catalyst poisoning through the gasifying effects of both O₂ and H₂O. The possibility of using feeds with various compositions also allows modulation of the H₂/CO syngas ratio for specific downstream applications, making this process attractive for the direct utilization of industrial flue gas or biogas, whose composition depends on fuel type or fermentation feedstock [5,6,7].

A number of studies have been devoted to the development and optimization of catalytic systems suitable for methane tri-reforming, which must be thermally stable, resistant to coke deposition, and active for efficient CO₂ and CH₄ conversion in the presence of steam and oxygen [8,9,10,11,12,13,14,15]. Ni-based catalysts, consisting of Ni metal particles supported on various oxides and their mixed oxides (Al₂O₃, CeO₂, ZrO₂, TiO₂, MgO, SiO₂, and La₂O₃) with or without addition of promoters (Pt, Rh, Ru, Ir, Au, K, Li, Mg, and Ca), are the most widely studied for the TR process [12,14,15]. However, the main drawbacks of Ni-based catalysts arise from their rapid deactivation due to carbon deposition and sintering of the active metal. The introduction of a second active metal, such as cobalt, has been shown to enhance activity and stability through synergistic effects, alloy formation, and modifications of metal–support interactions [16]. In particular, supported Ni–Co bimetallic catalysts have been extensively studied for the dry reforming, as summarized in several reviews [17,18,19]. Their superior performances compared with Ni or Co monometallic catalysts are primarily attributed to Ni–Co alloy formation, which leads to smaller metal particles, improves oxygen mobility, and facilitates carbon gasification, thereby mitigating deactivation [20]. In supported Ni-Co based catalysts, both the preparation method and the Ni/Co ratio play a significant role in enhancing catalytic activity and stability, by promoting both alloy formation and the reactivity of surface sites in contact with the CH₄/CO₂ feed. The excellent activity and high resistance to deactivation of Ni–Co catalysts result from the complementary roles of the two metals: Ni efficiently activates CH₄, while Co’s strong affinity for oxygen enhances CO₂ adsorption, thereby boosting the removal of carbon deposited on adjacent Ni sites [21].

Beyond catalyst development, another key challenge for the industrial-scale application of new processes lies in identifying novel approaches for energy input into the catalytic process that reduce greenhouse gas emissions and the global C footprint. Since CH₄-TR operating conditions require high temperatures (700–900 °C) because of reactants’ stability, the heat of the reaction is typically supplied by fuel combustion, which negatively contributes to the overall CO₂ balance. Recently, the electrification of catalytic processes using green electricity from renewable sources has emerged as an innovative strategy for energy input into the system using new methodologies, such as direct electrical heating, magnetic induction heating, electro-catalysis, and plasma activation [22]. Among them, magnetic induction heating offers several advantages, such as high energy efficiency, improved reaction rates, reduced heat transfer limitations, and potential for process intensification [23].

Catalysts combining both magnetic and catalytic properties, mainly based on Fe, Ni, and Co metals, have been employed in various syngas production and conversion processes, focusing on the effect of the application of an external magnetic field on the performance of the magnetic materials [24]. Unsupported ferromagnetic Ni–Co alloy nanoparticles have been found effective under magnetic induction heating for both the industrially relevant CH₄-SR [25] and the innovative CH₄-DR with CO₂ [26]. In these systems, Co, acting as a dissipative agent, enables high temperatures (>750 °C) to be reached through interaction with the electromagnetic field, while Ni provides the active sites for reforming reactions. The Co loading and the Ni/Co ratio proved to be critical factors in catalyst design, influencing both catalytic activity and magnetic heating efficiency. Ni–Co alloy nanoparticles supported on γ-Al₂O₃ (NiCo/γ-Al₂O₃) proved active under magnetic induction heating for syngas production by the SR [27,28], CPO [15], and CH₄ bi-reforming (BR) processes [29], where CO₂ and H₂O reactants performed the SR and DR reactions concomitantly, resulting in desired CO₂ utilization.

In this paper, the supported NiCo/γ-Al₂O₃ catalysts, previously characterized and proven effective for the SR and BR of CH₄ under magnetic induction heating [29], were investigated for the first time for the CH₄-TR in a flow reactor under a conventional thermal heating setup. Considering that similar steady-state catalytic performances have been found under thermal and magnetic activation [30], the results presented here can be reasonably extended to magnetically activated TR, expanding their applicability to different operating conditions.

To elucidate the role of the main parameters influencing the selectivity toward CO₂ conversion, syngas yield, and resistance to deactivation, the catalytic performances of NiCo/γ-Al₂O₃ catalysts were studied using CO₂, H₂O, and O₂ co-fed with CH₄ at different O₂/CO₂, H₂O/CO₂ and CH₄/(oxidants) ratios, with particular emphasis on concentrated feeds whose composition mimic those of industrial off-gases or biogas. Indeed, it is well established that for Ni-based catalysts, both activity and stability in the reforming of CH₄ strongly depend on feed composition [9], as the presence of O₂ and H₂O oxidants, while mitigating carbon deposition, decreases CO₂ conversion and induces metal sites oxidation with subsequent deactivation. Since alloy surface properties are, in principle, dependent on the alloy composition, the reactivity of NiCo/γ-Al₂O₃ catalysts with two different Ni/Co ratios (50:50 or 30:70) was compared. The catalyst properties after catalytic tests were characterized by XRD, FESEM, and Raman spectroscopy to assess how the structural and morphological properties of nanoparticles were affected by the reaction conditions.

2. Results and Discussion

2.1. Characterization of NiCo/γ-Al₂O₃ Catalysts

2.1.1. Fresh Samples

The main features of the fresh samples are summarized hereafter to be compared with those of the samples after catalytic runs (used samples).

The fresh NiCo/γ-Al₂O₃ samples, Ni₅₀Co₅₀ and Ni₃₀Co₇₀, showed a similar total metal loading (~25 wt%) and a relative metal content (Ni/Co) that closely corresponded to the nominal values in the impregnating solutions (

Table 1. Chemical, structural, and morphological parameters of fresh [29] and used samples: specific surface area (SA), pore volume (V_tot), total metal content (%wt_tot), relative metal percentage (Ni/Co), crystallite mean size (d_XRD), and particle mean size (d_SEM) for NiCo alloy.

Sample		S.A. (m2/g) a	Vtot (cm3/g) b	%wttot (AAS)	Ni/Co (AAS)	dXRD/ (nm)	dSEM (nm)
Ni50Co50	Fresh	145	0.51	24.2	51:49	30	37
	Used					37	51
Ni30Co70	Fresh	140	0.51	24.4	31:69	40	42
	Used					26	67

^a ±5 m²/g; ^b ±0.05 cm³/g.

), suggesting a similar metal-support interaction for Ni²⁺ and Co²⁺ species with the γ-Al₂O₃ support. Both samples had similar surface area and pore volume values (about 140 m² g⁻¹ and 0.51cm³/g, respectively) but lower than those of the γ-Al₂O₃ support (250 m² g⁻¹ and 0.84 cm³/g). This result is consistent with the partial pore occlusion of the support by the metal species (Table 1).

FESEM images of fresh samples showed spherical-like particles distributed over the support surface and fairly uniform in size, although slightly smaller and more homogeneous in the Ni₅₀Co₅₀ sample than in the Ni₃₀Co₇₀ one, the mean and maximum diameters being 37 and 70 nm vs. 42 and 90 nm, respectively (Table 1 and Figure S1).

The XRD patterns of fresh samples exhibited the reflections of the γ-Al₂O₃ support and of the NiCo metal alloy (fcc phase). The NiCo metal alloy reflections appeared at intermediate 2θ position with respect to those of pure Ni (fcc) [JCPDS 4-850] and Co (fcc) [JCPDS 15-806]. The mean crystallite size (d_XRD; Table 1) of both the Ni₅₀Co₅₀ and Ni₃₀Co₇₀ samples, calculated by the Scherrer equation for the (111) reflection, was comparable to the mean particle size obtained by FESEM analysis. It can be concluded that the alloy nanoparticles consist mostly of single crystallites strongly interacting with the support surface and, therefore, are resistant to sintering under high-temperature reduction treatment (900 °C). Additionally, a preferential growth of the alloy crystallites along the (111) direction is evidenced by the higher-than-expected relative intensity of this reflection. These findings suggest the formation of fcc crystallites with a truncated-octahedron geometry, preferentially exposing hexagonal (111) planes [31], consistent with the roughly spherical particles imaged by FESEM. Furthermore, in a magnified image of the fresh Ni₃₀Co₇₀ sample (Figure S1c, inset), the presence of a well-defined particle with a regular-hexagonal shape (of a long diagonal length of 87 nm, corresponding to twice the side length of 43 nm), confirms the hypothesis of a truncated-octahedron geometry.

2.1.2. Used Samples

The catalytic reactions (at least 15 catalytic runs) had little influence on the structural features of the samples but somewhat affected their morphological properties. Specifically, the XRD patterns of both used samples correspond to those of the NiCo alloy and the γ-Al₂O₃ support (Figure 1), indicating, as for the fresh samples [29], the retention of the Ni–Co alloy phase without detectable segregation of pure Ni or Co phases nor the appearance of reaction products of metal with the substrate.

The mean crystallite size, calculated from the (111) reflection (Table 1), was slightly larger for the used Ni₅₀Co₅₀ (37 vs. 30 nm), but smaller for the used Ni₃₀Co₇₀ (26 vs. 40 nm).

The morphological features observed in the FESEM images of the used Ni₅₀Co₅₀ and Ni₃₀Co₇₀ samples reveal a more heterogeneous particle size distribution compared with the fresh catalysts, with dimensions of up to 250 nm and larger average values (51 nm and 67 nm, respectively). These results indicate particle aggregation, which is more pronounced in Ni₃₀Co₇₀ than in Ni₅₀Co₅₀ (Figure 2 and Table 1), although both samples exhibit similar metal dispersion (D = 1.1%).

For the larger nanoparticles (Figure S2), EDX elemental analysis confirmed the mixed composition of the alloy, indicating that the aggregation of small particles does not lead to significant segregation of pure monometallic oxides under the catalytic conditions.

The larger increase in particle size observed by FESEM compared with the small variation in crystallite size determined by XRD analysis indicates that particle growth occurs through a coalescence process rather than via the Ostwald ripening mechanism. Since both the Ni₃₀Co₇₀ and Ni₅₀Co₅₀ catalysts did not suffer from significant particle growth after steam- or bi-reforming in rich CH₄ + H₂O or CH₄ + CO₂ + H₂O feeds at temperatures of up to 800 °C [29], the more significant sintering observed in the used samples can be ascribed to the presence of O₂ in the reactant mixture. Accordingly, it has been reported that molecular oxygen at elevated temperatures can bind to surface defects, attenuating particle–support interactions and thereby promoting particle mobility and coalescence [32]. However, since our catalysts were subjected to several oxidative–reductive activation treatments, the effect of these treatments on particle coalescence cannot be ruled out.

The oxidative–reductive activation treatments were sufficient to restore the catalyst activity when the specific catalytic run led to carbon deposition. In fact, for the Ni₃₀Co₇₀ catalysts inspected at an intermediate state of aging after the dry reforming at 750 °C (Section 2.2), for which a slow deactivation was observed (see Section 3.2), FESEM images revealed the presence of small carbon nanotubes and of a thin shading masking the particles, probably consisting of carbonaceous species (Figure 3). However, the amount of carbon-containing species was limited, as suggested by the EDX analysis and by the modest extent of catalyst deactivation.

2.1.3. Raman Characterization

Further insight into the structural properties of the fresh and used catalysts was obtained using Raman spectroscopy. Its high sensitivity to intermediate-range order without long-range periodicity makes Raman spectroscopy complementary to XRD analysis, which discloses only long-range ordered entities, thus providing a detailed picture of local structures.

In both the fresh Ni₅₀Co₅₀ and Ni₃₀Co₇₀ samples (Figure 4a,b, curves 1), Raman spectra analysis of the several broad bands detected in the range of 150–1200 cm⁻¹ disclosed the presence of Co₃O₄- and NiO-like structures. The Raman bands located at about 185, 465, 510, and 670 cm^–1 are in agreement with those expected for the most intense vibrational modes of the Co₃O₄ spinel structure [33]. The band broadening and the apparent shift to lower wavenumbers are attributed to the decrease in particle size and to surface defects. The pronounced and broad band at about 1080 cm⁻¹ (2LO mode) identifies the presence of nanosized NiO [34], whose vibrational modes in the range of 200–750 cm⁻¹ [34,35] are obscured by those of Co₃O₄.

The Raman detection of oxidic phases in the metal alloy-based samples, due to some surface oxidation under air exposure, suggests the formation of a slightly non-homogeneous alloy, with oxidic clusters having a size below the XRD detection limit (about 4 nm).

For the used catalysts (Figure 4a,b), Raman spectroscopy, in addition to the vibrational modes of Co₃O₄- and NiO-like structures (150–1200 cm⁻¹), clearly revealed carbonaceous deposits identified by G and D bands in the range of 1200–2000 cm⁻¹. The G band at about 1580 cm⁻¹ (“graphite band”) is characteristic of a perfect graphite lattice, whereas the D band at about 1370 cm⁻¹ (“defect band”) is related to a defective graphitic lattice. The relative D/G band intensity ratio is indicative of the degree of structural disorder of the carbon deposits [36]. A high D/G ratio indicates a large amount of graphitic edges and defects and/or a high degree of amorphization, typical of relatively small carbon deposits compared with a low D/G ratio. For the used Ni₅₀Co₅₀ sample (Figure 4a, curves 2 and 2′), Raman analysis revealed two distinct pictures, one consisting of sample areas with completely carbon-free deposits (curve 2), while the other one showed non-overlapping small intensities of D and G bands. A low D/G ratio, I_D/I_G = 0.53, was found, suggesting the presence of a few graphitic-like patches. On the contrary, for the used Ni₃₀Co₇₀ sample (Figure 4b, curve 2), D and G bands were detected in all the analyzed areas, yielding a high D/G ratio, I_D/I_G = 1.15, indicative of the prevailing formation of small defective carbon deposits.

2.2. Catalytic Performances for CO₂ Valorization Across Reforming Processes

To identify the most favorable operating conditions for CO₂ conversion to syngas over NiCo/γ-Al₂O₃ catalysts, the specific reforming pathways contributing to methane tri-reforming (TR: CH₄ + CO₂ + H₂O + O₂)—namely, dry reforming (DR: CH₄ + CO₂), oxidative fry reforming (ODR: CH₄ + CO₂ + O₂), and bi-reforming (BR: CH₄ + CO₂ + H₂O)—were first investigated separately. All catalytic experiments were carried out using a reactant CH₄ + CO₂ feed at the stoichiometric ratio for DR (CH₄/CO₂ = 1), while the effect of adding controlled, small amounts of co-oxidants (O₂ and/or H₂O) was systematically evaluated.

A comparison of these CH₄ reforming processes over Ni₅₀Co₅₀ shows that the catalyst is highly active for syngas production under all the investigated conditions (Figure 5a–d).

Specifically, at 750 °C, the highest temperature explored, the conversion of CH₄ and CO₂ always exceeded 90%, and the H₂ yield exceeded 85%, with all values progressively and concomitantly decreasing as the temperature decreased. However, the relative contributions of the reforming reactions varied depending on the oxidant content, affecting both syngas composition and catalyst stability.

For the dry reforming feed (CH₄ + CO₂; Figure 5a), the H₂/CO molar ratio was about ~0.9 at 750 °C, slightly decreasing with decreasing temperature to about 0.7 (see inset). This deviation from the stoichiometric value of the DR reaction (H₂/CO = 1) indicates the occurrence of a parallel side-reaction, probably the reverse water gas shift (RWGS: CO₂ + H₂ ⟶ CO + H₂O), reducing the H₂ yield relative to CO production. This interpretation is further supported by the poor superimposition of the conversions and H₂ yield curves with the following order: CO₂ conversion > CH₄ conversion > H₂ yield. A slight but consistent decrease in activity over time (Figure 5e) was observed, indicating a slow catalyst deactivation. As reported for the DR on Ni-based catalysts [37], the deactivation was likely due to coke deposition from the carbon-rich feed and to active phase sintering.

The addition of a second oxidant, O₂ or H₂O, in a small concentration (0.3% or 0.6%v/v, respectively) to the CH₄ + CO₂ feed increased the selectivity to syngas over the whole temperature range due to catalytic partial oxidation and steam reforming occurrence and to the inhibition of the RWGS side-reaction, as evidenced by the proximity of the curves of CH₄ and CO₂ conversions and H₂ yield in the ODR (Figure 5b) and in the BR (Figure 5c) processes. An even stronger overlap of the curves was observed when both the O₂ and H₂O were added to the feed, also causing the H₂/CO values to be around 1 in the whole temperature range (Figure 5d and inset).

Significantly, unlike DR, the catalytic activity at 750 °C in the ODR, BR, and TR processes remained stable over time-on-stream (Figure 5f–h). Since in these consecutive catalytic runs, the conversions and yields were similar to the initial values for the DR, the deactivation occurring under DR conditions can be considered reversible. Therefore, it can be ruled out that it originates from the limited alloy sintering (SEM characterization); rather, it is attributable to carbon-species deposition on the active sites, which can be gasified during the redox activation carried out between two consecutive runs. In the ODR, BR, and TR processes, the presence of small amounts of O₂ and/or H₂O enhances catalyst stability by increasing the rate of CO formation, which continuously removes the surface carbon intermediates.

In the combined TR process (Figure 5d), the interplay among DR, SR, and CPO becomes highly composition-dependent. In particular, a preliminary analysis of the results at 750 °C shows that, with a limited amount of O₂ and H₂O in the feed, the Ni₅₀Co₅₀ catalyst remains selective toward DR over SR, but exhibits low selectivity toward DR with respect to CPO, which becomes the predominant reaction. Based on simple mass balance, taking into account the CH₄ conversion (97%), and neglecting the minor contribution of RWGS and other side-reactions, the observed H₂ yield (92%) can be mainly attributed to DR (90% CO₂ conversion), with CPO occurring to its maximum extent (100% O₂ conversion) and SR to a limited extent (20% H₂O conversion). Assuming H₂O’s complete consumption (100%), the CO₂ conversion would be expected to decrease to about 72–77% (depending on the calculation procedure adopted; see Table S2), whereas the observed value is significantly higher (90%). The increasing separation between CO₂ and CH₄ conversion curves above 650 °C (Figure 5d) suggests that competition between DR and CPO becomes more pronounced with temperature. These results indicate that the strength of oxidant interaction with the metallic active sites, leading to dissociative chemisorption, follows the order of O₂ ≥ CO₂ > H₂O, in agreement with the observed activity trend CPO ≥ DR > SR and with thermodynamic analysis [38].

2.2.1. Effect of O₂ and H₂O Co-Reactants

To investigate the effect of the concentration of a single co-oxidant (O₂ or H₂O) on the DR reaction, the ODR and BR processes on the Ni₅₀Co₅₀ catalyst were first studied.

In the CH₄ + CO₂ + O₂ (ODR) reactant mixture, increasing the amount of O₂ up to the stoichiometric value (3%) required for the competing CH₄-CPO reaction resulted in a strong inhibition of the CO₂ valorization by CH₄ (Figure 6a). In particular, at 750 °C, the CO₂ conversion markedly decreased (from ~95% to ~45%), whereas the CH₄ conversion remained essentially unchanged, and the O₂ was always completely converted. These results confirm that the DR reaction was partly replaced by CH₄-CPO as well as by CH₄ oxidation, yielding 100% O₂ conversion. Accordingly, the H₂ yield increased from about 90 to 95% at O₂ = 1%, due to the boosting effect of CPO, and it slightly decreased at O₂ = 3%, due to the light-off of total CH₄ combustion, as indicated by the larger H₂O formation (Figure 6a).

For the BR feed (CH₄ + CO₂ + H₂O), increasing the amount of H₂O (from 0.6% to 1.8%) resulted in a slight inhibition of the CO₂ valorization, which was less pronounced than that caused by O₂ (Figure 6b). More specifically, at 650 °C, the inlet H₂O was only partially converted (50% maximum), leading to a slight decrease in CO₂ conversion (from 80 to about 74%) and an increase in H₂ yield (from 63 to 68%), while CH₄ conversion remained unchanged. This experimental evidence confirms that, unlike O₂, only a minor fraction of H₂O competes with CO₂ for interaction with the active sites, causing the SR pathway to occur to a limited extent related to DR and resulting in a slight increase in the H₂/CO ratio from ~1 to about 1.1.

2.2.2. Effect of Feed Concentration and Contact Time

The effect of the feed concentration on CO₂ conversion was determined by studying the activity of the Ni₅₀Co₅₀ catalyst for a selected reforming process (ODR) using mixtures with the same reactant ratio (starting from CH₄:CO₂:O₂ = 2:2:0.1, %v/v) but progressively enriched in reactant amounts, up to 14% of CO₂, which corresponds approximately to the levels found in industrial flue gases [4,39,40]. Up to a 10% CO₂ inlet concentration (CH₄:CO₂:O₂ = 10:10:0.5, %v/v), conversions and yields remained unchanged over time on stream, showing very similar values (see data for 2% and 8% mixtures shown in Figure S3), whereas only a slight decrease occurred for CO₂ concentrations of ≥12%.

The dependence of catalyst performances on the various mixtures was evaluated by calculating the productivity (X_i, molecules per time per catalyst mass) parameters, intended as a measure of the specific amount of product (H₂) or reactant converted (CH₄ and CO₂) over time. Although the physical dimensions of X_i are the same as those of the specific reaction rate (molecules per time per mass), the X_i parameter does not have a kinetic meaning because diffusion effects become significant at the high conversion levels used for its calculation. However, it allows comparison, at the application level, of the overall process efficiency over time under the adopted real operating conditions, as it reflects a cumulative phenomenon occurring on a time scale much larger than that of an intrinsic activity evaluation (hours vs. seconds). As for the ODR on the Ni₅₀Co₅₀, all the productivity X_i parameters increased almost linearly with reactant concentration in the feed (Figure 7). This behavior indicates that, even at 14% of reactant concentration, the catalyst is not operating at the maximum site occupancy, i.e., the active sites have not yet reached full coverage as the residence time on the catalytic bed is much larger than that required for the catalytic cycle.

To clarify whether the type of regime is kinetic or diffusional in our experimental conditions, the performance of the Ni₅₀Co₅₀ sample was evaluated by varying the total flow rate from 150 to 400 cc/min, while maintaining constant the temperature (650 °C) and the feed composition (CH₄:CO₂:H₂O = 6:6:0.6, %v/v). Under these conditions, CH₄ and CO₂ conversion and H₂ yield depended on the catalyst mass to total feed flow rate ratio, W/F (parameter proportional to the residence time, τ, of the feed on the catalyst layer), with a bent trend (Figure S4), far from the linear relationship expected for a reaction under kinetic control. The non-linear behavior reveals that the catalyst operates under diffusion-limited conditions, likely due to its very high intrinsic activity. As a consequence, the productivity values X_i, calculated from conversions under our experimental conditions, are much lower than the true reaction rates. However, although diffusion limitation under our high reactivity conditions precludes a reliable kinetic analysis for active sites and reaction mechanism elucidation, the productivity data allow us to make deductions about the overall performance.

2.3. Catalytic Performances of NiCo Catalysts for CH₄ Tri-Reforming (TR)

2.3.1. Effect of Reactant Mixture Composition

To evaluate the adaptability of NiCo catalysts for CO₂ valorization in real feeds containing CO₂, H₂O, and O₂ in variable amounts, such as tail gases from industrial plants or biogas from waste treatment, the CH₄-TR activity was followed using reactant mixtures at various oxidant contents. Specifically, the O₂/CO₂ and H₂O/CO₂ ratios were changed from 0.3 to 0 (Figure 8a) and from 0.9 to 0, respectively (Figure 8b), the larger values being similar to those in a typical industrial exhaust (13% CO₂, 9% H₂O, and 4% O₂ [4], where O₂/CO₂ ~ 0.3 and H₂O/CO₂ ~ 0.7). The experiments were carried out with a diluted feed (2% in CO₂ and CH₄/CO₂ = 1), as it allows the composition to be stabilized more effectively over time.

In the TR feed with the larger H₂O and O₂ content (CH₄:CO₂:O₂:H₂O = 2:2:0.6:1.8, %v/v) (namely, with oxidant ratios close to real exhausts), CH₄ was almost completely converted, but the CO₂ conversion was low (~35%) (Figure 8a). As for the single ODR and BR processes (see Section 2.2.1), also in the CH₄-TR, the CPO and SR reactions competed with DR, with CPO prevailing over SR, as revealed by the total O₂ conversion (100%) and much lower H₂O conversion value (~20%). Accordingly, when the amount of O₂ in the TR feed was reduced to 0.1% and 0%, CO₂ conversion improved up to 50% and 60%, respectively. The maximum H₂ yield was instead obtained in the presence of a small amount of O₂ (0.1%), suggesting a faster regeneration of the active sites through carbon-adsorbed species gasification under the CPO mechanism.

Keeping this optimal O₂ content (0.1%, relative to CO₂ conversion and H₂ yield), we decreased the H₂O amount to 1%, 0.2%, and 0% in the TR feed, resulting in a further improvement of CO₂ conversion to about 71, 92, and 97%, respectively (Figure 8b). Also in this case, a minimal amount of H₂O was also beneficial for maximizing the H₂ yield, being the highest (91%) at a H₂O content of 0.2%. In all experiments, the activity remained stable over time on stream.

On the whole, for the TR process, the maximum CO₂ valorization occurred with the CH₄:CO₂ = 1:1 mixture co-fed with co-oxidants O₂ and H₂O in a minimal amount, as found for the ODR and BR processes. These conditions minimize the competition of oxidants with CO₂ for reforming, while ensuring stable activity and boosting the H₂ yield.

2.3.2. Effect of Ni/Co Ratio

The effect of the Ni/Co ratio on the CH₄-TR performances of the Ni₅₀Co₅₀ and Ni₃₀Co₇₀ catalysts was evaluated by comparing conversions, since the used samples have a similar metal dispersion (D = 1.1%), assuming homogeneous alloy nanoparticles. The catalysts exhibited a similar selectivity toward the DR with respect to the CPO and SR, shown by the overlapping of the conversion and H₂ yield curves for each catalyst in the whole temperature range (Figure 9a). However, the activity of the Ni₃₀Co₇₀ catalyst (low Ni loading) was lower than that of Ni₅₀Co₅₀’s one (~10% lower CH₄ and CO₂ conversions and H₂ yield), consistent with the lower amount of active Ni sites on the surface of the alloy nanoparticles. This lower activity is in agreement with the widely accepted role of Ni in methane activation [21], suggesting that the Co surface sites contribute less to the rate-determining step of the reforming mechanism.

As for activity stability, the Ni₃₀Co₇₀ catalyst exhibited constant conversion and yield values over time on stream during CH₄–TR (Figure 9b), as well as during the other reforming processes, including DR (Figure S5b). The different behavior of Ni₃₀Co₇₀ with respect to Ni₅₀Co₅₀, which showed partial and reversible deactivation during DR (see Figure 5e and Figure S5a) due to carbon species deposition, is consistent with the lower Ni/Co ratio in Ni₃₀Co₇₀, which is commonly recognized to favor carbon species gasification. Such an effect is attributed to the high oxygen affinity of Co metal sites, which favors CO₂ activation and leads to a higher abundance of CoO_x islands on the alloy surface, thereby enhancing the conversion of carbon species to CO on adjacent Ni sites [16,19,20,21].

For the Ni₃₀Co₇₀ catalyst, the TR activity depended on the O₂ and H₂O contents in the feed (Figure S6) with a trend similar to that of the Ni₅₀Co₅₀ catalyst. The inhibition of CO₂ valorization occurred as the oxidant amount increased due to competition of CPO and SR, but with conversions and yields values lower than those in Ni₅₀Co₅₀, in agreement with the lower content of the Ni active phase.

2.3.3. Effect of CH₄ Content

In TR experiments, all performed with a ratio of CH₄/CO₂ = 1, CH₄ acted as the limiting reactant. To enhance CO₂ conversion on the Ni₅₀Co₅₀ catalyst, the CH₄ amount was increased up to the stoichiometric CH₄-to-oxidants ratio corresponding to DR, SR, and CPO stoichiometries. Under this optimized condition, in the TR feed CH₄:CO₂:O₂:H₂O = 4:2:0.6:1.8 (oxidants relative composition resembling real exhaust gases), the CH₄ conversion was >96%, as in the mixture where CH₄ was limiting (CH₄:CO₂:O₂:H₂O = 2:2:0.6:1.8), but CO₂ and H₂O conversions (~80 and ~75%, respectively) and H₂ yield (~92%) were significantly higher (Figure 10). The presence of some residual O₂ (O₂ conversion ~90%) suggests that the increase in the CH₄ concentration favored DR and SR more than CPO, indicating a CH₄ reaction order higher in the reforming reactions than in the partial oxidation.

Therefore, a real chance to effectively valorize CO₂ by the TR process relies on the possibility to properly modulate CH₄ addition in the real feeds on the basis of the composition of oxidants.

3. Materials and Methods

3.1. Materials

Supported NiCo/γ-Al₂O₃ catalysts were prepared following a conventional impregnation process [29]. Cylindrical γ-Al₂O₃ granules (Alfa Aesar, Ward Hill, MA, USA; specific surface area: 250 m²g⁻¹), evacuated at 300 °C (3 h), were stirred (~12 h) in a 8 M solution containing Ni(NO₃)₂·6H₂O and Co(NO₃)₂·6H₂O (Merck-EMSURE, Darmstadt, Germany) in different relative molar percentages (50:50 or 30:70), filtrated, dried overnight (85 °C), and heated at 600 °C in air flow (1 h). The samples were subsequently reduced in Ar-H₂ (3% v/v) flow at 900 °C (heating ramp: 10 °C/min) for 5 h and referred to as fresh samples.

The Ni and Co contents (wt%; Table 1) were determined by atomic absorption spectroscopy (AAS, Varian Spectra AA-220, Varian, Mulgrave, VIC, Australia). The samples considered in this paper had a total metal content of about 25 wt%, and they were named NiaCob, where a and b represent the relative Ni/Co atomic % in solution.

3.2. Catalysts Characterization

The main structural and morphological properties of the fresh samples (XRD, FESEM, and N₂ physisorption) are summarized in Table 1.

Samples after catalytic tests, referred to as used catalysts, were characterized by XRD, FESEM, and Raman spectroscopy. X-ray diffraction (XRD) patterns were obtained using a Philips PW 1729 diffractometer (Philips, Almelo, The Netherlands) with Cu Kα (Ni-filtered) radiation (40 kV and 20 mA) in the 10–70° 2θ range (step size: 0.02°; step time: 1.25 s).

Field-emission scanning electron microscopy (FESEM) images were obtained by an AURIGA Zeiss 405 HR-FESEM instrument (Carl Zeiss, Oberkochen, Germany) equipped with an energy-dispersive X-ray spectroscopy apparatus (EDXS, Bruker, Billerica, MA, USA) for elemental detection. Particles with a minimum diameter of about 5 nm could be discriminated. Images were processed via the ImageJ free software-version 1.53h [41] to determine particle diameter, d_i, the d_i distribution, and the length–number mean diameter (Σ_in_id_i/Σ_in_i, d_LN [42]). The metal alloy dispersion, D (percent ratio of the exposed metal atoms to the total ones; Table 1), was evaluated using the approximation of spherical particles as D = 100·6·(v_m/a_m)/d_VA [42], where v_m is the atomic volume in the bulk metal, am is the area occupied by a surface atom, and d_VA is the volume–area mean diameter (d_VA = Σ_in_id_i³/Σ_in_id_i²). The v_m and a_m values of the metal alloy at various compositions (Table 1) were calculated as a mean value between those of pure Ni and Co [42], assuming a linear lattice parameter variation with the alloy composition (Table S1).

Raman spectra were recorded at room temperature, in back-scattering geometry, with an inVia Renishaw micro-Raman spectrometer (Renishaw, Wotton-under-Edge, UK), using the 488.0 nm emission line from an Ar ion laser as the exciting source. The power of the incident beam was about 5 mW. Repeated accumulations (10 or 20 scans × 10 s) were generally acquired in at least four regions of each sample using 20× or 5× objectives to check the sample homogeneity. Spectra were calibrated using the 520.5 cm⁻¹ line of a silicon wafer.

3.3. Catalytic Activity Measurements

The NiCo/γ-Al₂O₃ catalysts (25 mg) were tested using a fixed-bed tubular reactor consisting of two coaxial quartz tubes (i.d. 20 and 10 mm), allowing preheating of the feed gas under conventional heating in a tubular oven equipped with a K-type thermocouple positioned adjacent to the catalytic bed. High-purity gas mixtures of CH₄/N₂, O₂/N₂, CO₂/N₂, and H₂/N₂ (Nippon Gases) were used without further purification. The H₂O/N₂ mixture was obtained using a saturator by bubbling a pure N₂ (SOL) stream through distilled water maintained at a constant temperature (40 °C).

Before each catalytic run, all catalysts were subjected to an oxidative–reductive activation procedure, consisting of an oxidative step under a 10% (v/v) O₂/N₂ flow from room temperature to 650 °C (heating rate: 10 °C min⁻¹, held at 650 °C for 1 h), followed by a purging step under N₂ flow with increasing temperature up to 750 °C (heating rate: 10 °C min⁻¹) and an isothermal reductive step under a 10% H₂/N₂ flow (750 °C for 40 min). After activation, the reactor was isolated at 750 °C using a bypass valve, while the reactant mixture was stabilized in a parallel line. Once the feed composition was verified by gas chromatography, the mixture was fed into the reactor to start the catalytic run.

A typical run (generally lasting some hours) consisted of steady-state activity measurements under continuous flow of the reactant mixture performed either (i) at various temperatures, starting from 750 °C and decreasing to 450 °C in a random sequence (each temperature was maintained for at least 15 min, corresponding to three consecutive GC analyses), or (ii) at a selected temperature, monitoring the activity as a function of time on stream for up to 150 min. At the end of each run, the reactor was purged with an N₂ flow for 10 min and allowed to cool to room temperature under static conditions by natural convection. The oxidative step of the activation procedure applied between runs ensured the removal of any possible carbon deposits. For each catalyst, at least 15 catalytic runs were performed, the last one in a tri-reforming mixture; the catalyst recovered from the reactor after this sequence was designated as “used”.

The inlet and the outlet reactant mixture were analyzed by a Varian Micro-GC CP-4900 gas chromatograph equipped with two columns (10 m Molsieve 5A BF, for H₂, O₂, and CO; 10 m Poraplot Q for CH₄, CO₂, and H₂O) and TCD detectors. The conversion of CH₄ and CO₂ and the yields of H₂ and CO were calculated as percentages using the following equations:

$${\mathrm{C}}_{{\mathrm{CH}}_4}=100\times {\displaystyle \frac{F_{tot}^{in}\ast {CH}_4^{in}-F_{tot}^{out}\ast {CH}_4^{out}}{F_{tot}^{in}\ast {CH}_4^{in}}}$$

(1)

$${\mathrm{C}}_{{\mathrm{CO}}_2}=100\times {\displaystyle \frac{F_{tot}^{in}\ast {CO}_2^{in}-F_{tot}^{out}\ast {CO}_2^{out}}{F_{tot}^{in}\ast {CO}_2^{in}}}$$

(2)

$${\mathrm{Y}}_{{\mathrm{H}}_2}=100\times {\displaystyle \frac{F_{tot}^{out}\ast H_2^{out}}{\left({2\ast F}_{tot}^{in}\ast {CH}_4^{in}\right)+({2\ast F}_{tot}^{in}\ast H_2{O}^{in})}}$$

(3)

$${\mathrm{Y}}_{\mathrm{CO}}=100\times {\displaystyle \frac{F_{tot}^{out}\ast {CO}^{out}}{F_{tot}^{in}\ast ({CH}_4^{in}+C{O}_2^{in})}}$$

(4)

where F_totⁱⁿ is the total inlet flow, and F_tot^out is the total outlet flow; CH₄^in/out, CO₂^in/out, H₂^out, and CO^out are the volumetric fractions (ppm) entering and leaving the reactor. The productivity X_i (mol of i-species, converted or produced, per hour per gram) was calculated from Ci or Yi values.

The inlet total flow rate was 150 cm³ (STP) min⁻¹ (space velocity: 3.6·10⁵ NL·kg_cat⁻¹·h⁻¹), except for one experiment (bi-reforming of CH₄), in which it was changed in the range of 150–400 cm³ (STP) min⁻¹.

All experiments yielded a good carbon, hydrogen, and oxygen balance (100 ± 5, as percentage).

4. Conclusions

The NiCo alloy supported on γ-Al₂O₃ proved to be an effective catalyst for the CO₂ valorization into syngas through the tri-reforming of CH₄ under conventional heating using a gas mixture containing up to 14% v/v of CO₂ in the presence of H₂O and O₂. Since similar NiCo alloy-based catalysts have shown comparable reforming performance under thermal and magnetic activation [30], these results highlight the strong potential of NiCo/γ-Al₂O₃ catalysts for the industrial electrification of the CH₄ tri-reforming process via magnetic induction heating.

While the activity for DR of NiCo/γ-Al₂O₃ catalysts gradually decreased with time on stream due to coke deposition, the presence of small amounts of O₂ and H₂O in the TR feed rendered the activity stable over time by improving carbon species gasification, boosted syngas production, and minimally inhibited CO₂ conversion.

The selectivity toward CO₂ conversion in the TR mixture strongly depended on the relative amount of oxidants and on the CH₄-to-oxidants ratio. Large amounts of O₂ and H₂O caused a marked decrease in the selectivity toward CO₂ conversion due to the competition among oxidants for the metal active sites, favoring CPO and SR reactions. The increase in the CH₄ content in the feed up to the stoichiometric value with respect to oxidant amounts preferentially favored the DR pathway, leading to very high CO₂ conversion.

The Ni/Co ratio in NiCo nanoparticles negligibly affects the selectivity toward CO₂ conversion, varying the feed composition. Instead, a high Co content in the alloy decreases the overall activity, confirming Ni metal species as active sites.

From an application perspective, NiCo/γ-Al₂O₃ catalysts render CH₄ tri-reforming a promising approach for CO₂ valorization in industrial off-gases or biogas streams, owing to their ability to maintain high activity across different CO₂ + H₂O + O₂ feeds by adjusting CH₄ addition.