Development of Active and Stable Low Nickel Content Catalysts for Dry Reforming of Methane

Abstract

Methane dry reforming (DRM) was investigated over highly active Ni catalysts with low metal content (2.5 wt %) supported on Mg-Al mixed oxide. The aim was to minimize carbon deposition and metal sites agglomeration on the working catalyst which are known to cause catalyst deactivation. The solids were characterized using N₂ adsorption, X-ray diffraction, temperature-programmed reduction, X-ray photoelectron spectroscopy, and UV-Vis diffuse reflectance spectroscopy. The results showed that MgO-Al₂O₃ solid solution phases are obtained when calcining Mg-Al hydrotalcite precursor in the temperature range of 550–800 °C. Such phases contribute to the high activity of catalysts with low Ni content even at low temperature (500 °C). Modifying the catalyst preparation with citric acid significantly slows the coking rate and reduces the size of large octahedrally coordinated NiO-like domains, which may easily agglomerate on the surface during DRM. The most effective Ni catalyst shows a stable DRM course over 60 h at high weight hourly space velocity with very low coke deposition. This is a promising result for considering such catalyst systems for further development of an industrial DRM technology.

1. Introduction

The dry reforming of methane (DRM), which is the strong endothermic reaction of methane with carbon dioxide to syngas (Equation (1)), has gained increasing interest in recent years:

CH₄ + CO₂ → 2CO + 2H₂

(1)

This reaction cannot only be effectively applied for natural gas conversion into liquid fuels via the syngas route, but also may contribute to reducing the emission of CO₂ and CH₄ greenhouse gases to our atmosphere [1,2,3,4]. The syngas from DRM with a low ratio of H₂/CO is suitable for subsequent syntheses of oxygenated chemicals [5] and hydrocarbons via the Fischer-Tropsch synthesis [6]. However, DRM requires a high reaction temperature [7]. This leads to several catalyst deactivation processes resulting in catalytic sites’ agglomeration and coke formation [8]. The deposited carbon is formed during the DRM reaction through CH₄ decomposition (Equation (2)) and Boudouard (Equation (3)) reactions [4,9,10]:

CH₄ → C + 2H₂

(2)

2CO → CO₂ + C

(3)

Therefore, the great challenge for the development of a DRM process is the development of catalysts with high activity for the conversion of CO₂ and CH₄ and strong ability in suppressing the formation of carbon deposits [8]. Ni based catalysts were most frequently reported in DRM studies due to their high activity, availability, and lower price compared to other transition and noble metals [11,12,13]. In those studies, efforts were focused on improving the activity and stability of Ni active sites in the following aspects: screening and modifying the properties of the support, selecting the support and catalyst preparation methods, and the addition of promoters [13].

Suitable supports play a key role in the enhancement of catalytic activity and the suppression of carbon deposition for the DRM reaction [4]. γ-Al₂O₃ has been widely employed as a support for DRM catalysts because of its high availability and low price [4,14]. It was also suggested that an improvement of the support properties for the adsorption and activation of CO₂ would accelerate the gasification of the deposited carbons and prevent the rapid deactivation of the catalyst [13]. In order to achieve that, certain alkaline-earth metal oxides were incorporated into alumina to increase the basicity of the support [13]. MgO was reported to improve the catalytic performance of Ni/Al₂O₃ because of the strong interaction between Ni and the supports and the basic property of the metal oxide could suppress carbon formation in the dry reforming process [15,16]. The cooperation of Mg and Al oxides in same support material can result in many structures, such as MgAl₂O₄ spinel, which is the well-studied support structure for Ni catalysts in several types of reactions due to its good chemical stability and mechanical strength [13,17]. Other structures of MgO and Al₂O₃ mixed oxides could also serve as supports for DRM catalysts in order to disperse and stabilize the metallic Ni species and promote the coking resistance behavior [18,19,20]. However, only a few DRM studies have focused on low Ni content catalysts with such mixed oxide supports, probably due to the partial formation of stable but poorly active phases, such as NiAl₂O₄ spinel or NiO-MgO solid solution [17]. However, Ni catalysts with low metal content are potentially one of the solutions for overcoming the abovementioned problems of DRM. This is due to higher catalyst stability and lower carbon deposition achieved by forming highly dispersed and smaller sized Ni particles [21,22].

Chelating agents are molecules with two or more electron-pair donor atoms which can act as a ligand and attach themselves to metal ions [23]. This causes the formation of metal complexes, which were also studied to modify the preparation route for supported catalysts using the impregnation method [23]. In such studies, uniform distribution and high dispersion of the active component over the support were achieved. Those beneficial properties arise from the diffusion-controlled homogeneous distribution of metal atoms in the precursor structure as well as the enhanced viscosity of the gel-like material inhibiting redistribution of the impregnated species during subsequent drying. Citric acid (CA) is the most common and promising compound among the chelating agents because it is cheap, highly soluble, and environmentally harmless [24].

The aim of this study is to explore the potential of catalysts with low Ni content, according to the literature review [8], on Mg-Al mixed oxide supports (Mg-Al) that can be both stable and active in DRM at high weight hourly space velocity (WHSV). Modifications of the support and catalyst preparation were investigated to enhance the stability of the catalyst without deteriorating the DRM activity. Thus, various Ni catalysts with different supports were prepared via modified routes and tested in a continuous fixed bed reactor. Long-term experiments together with the characterization of fresh and spent catalysts using various techniques, including inductively coupled plasma-optical emission spectroscopy and atomic adsorption spectroscopy (ICP-OES resp. AAS), N₂ physisorption, X-ray powder diffraction (XRD), X-ray photoelectron spectroscopy (XPS), Temperature-programmed reduction (TPR), and UV-Vis diffuse reflectance spectroscopy (UV-Vis-DRS), indicated important attributes that contribute to catalyst activity and stability.

2. Results

2.1. Characterization of the Catalysts

First of all, the crystalline structures of the supports and catalysts can be revealed by X-ray diffraction. Figure 1 depicts the XRD patterns of MgO·Al₂O₃ mixed oxides (denoted as Mg.Al.X with X indicating the calcination temperature) obtained from hydrotalcite material with different calcination temperatures (550, 700, 800, 1000 °C). The patterns of MgO and Al.550 (boehmite calcined at 550 °C) are also displayed as references. When hydrotalcite samples are calcined at temperatures between 550 °C and 800 °C, the XRD patterns of the resulting Mg.Al.X samples show broad reflections at about 2θ = 43° and 63° which have 2θ values higher than those of periclase (cubic form of magnesium oxide, ICDD (International Centre for Diffraction Data) file No. 01-071-1176) but lower than those of γ-Al₂O₃ (ICDD file No. 00-010-0425). The patterns of periclase and γ-Al₂O₃ structures are clearly observed in MgO and Al.550 samples, respectively. This highlights the incorporation of smaller Al³⁺ cations into the bulk of the periclase (MgO) structure lattice, establishing a solid solution between MgO and Al₂O₃ (Mg-Al-O) [25,26]. A higher calcination temperature modifies such an incorporation which can be observed as a shift of reflections to lower 2θ values. Up to 800 °C, the pattern of such a solid solution is still observable. Besides, Mg.Al.700 and Mg.Al.800 show reflections (2θ = 31.3°; 37°; 45°; 55.5°; 59°; 65°) assigned to the formation of the MgAl₂O₄ phase (ICDD file No. 00-021-1152) at higher calcination temperatures [27]. At 1000 °C, no reflection of the solid solution can be observed while peaks assigned to periclase MgO and MgAl₂O₄ phases are sharp and characteristic, indicating the high crystallinity of those phases. Those results are in accordance with other studies [25,28,29] focusing on the thermal decomposition behavior of hydrotalcite, producing Mg-Al oxides and mixed oxides in solid solution or spinel structures, particularly when a low heating rate was applied.

The XRD patterns of impregnated low content Ni samples (not shown) show no significant differences from those of their corresponding supports. This reflects the formation of well-dispersed Ni²⁺ species leading to very small Ni particles that were not detected by the XRD measurements [30,31]. This might be also due to Ni diffusion into the bulk of the supports, especially when a high calcination temperature was applied during the preparation [31].

The XRD patterns of spent Ni catalysts are presented in Figure 2. Ni metal species (Ni⁰), which are the result of the pre-reduction process before DRM, can be detected with reflections at 2θ = 44.5° and 51.8° (ICDD file No. 01-071-3740). Spent Ni/Al.550 (Figure 2b) shows a pattern with additional reflections at the mentioned 2θ values compared to the fresh one. This is assigned to the formation of highly crystalline Ni⁰ domains due to agglomeration during the reaction. On the other hand, the XRD pattern of spent Ni/Mg.Al.550 (Figure 2a) displays a similar profile as the fresh sample, proving the stability of the material during the reaction. This result is probably due to the incorporation of Ni into the solid solution or/and the MgAl₂O₄ spinel framework since such stable phases prevent the structural transformation at the high reaction temperature [17,18,21].

The N₂ adsorption-desorption isotherms together with the BJH (Barrett-Joyner-Halenda) pore size distributions (PSD) of the Mg.Al.550.800 support and corresponding Ni catalysts are shown in Figure 3 and Figure 4, respectively. All isotherms exhibit type II curves with H3 type hysteresis loops according to the IUPAC (International Union of Pure and Applied Chemistry) classification, which reflects the contribution of the mesopores in the materials.

Regarding the pore size distribution (Figure 4), both the Ni samples show a shift toward higher pore diameter compared to their support. This feature most likely indicates the re-structuring of the support material in the impregnation step. However, in both cases, no limitation by internal diffusion is expected for the feed molecules, which are small compared to the size of the pores.

The textural parameters of the calcined Ni samples and the supports are summarized in

Table 1. Textural properties of calcined catalysts and their supports.

Catalyst	SBET (m2/g)	Total Pore Volume (cm3/g)
Mg.Al.550	180	0.24
Mg.Al.550.800 1	147	0.24
Mg.Al.1000	85	0.37
Al.550	234	0.54
MgO	50	0.35
Ni/Mg.Al.550	174	0.40
Ni/Mg.Al.550.CA 2	156	0.27
Ni/Mg.Al.1000	82	0.36
Ni/Al.550	178	0.47
Ni/MgO	45	0.33

¹ Mg.Al.550 calcined again at 800 °C; ² Citric acid was used during impregnation.

.

While all bare supports were calcined once at 550 °C and the resulting catalysts at 800 °C, one particular sample denoted as Mg.Al.550.800 was calcined at 550 and 800 °C as a reference support without impregnation.

Support Al.550 displays the highest surface area and pore volume among the studied supports. However, Ni/Mg.Al.550 and Ni/Al.550 show similar texture parameters. Both Ni/Mg.Al.550 and Ni/Mg.Al.550.CA expose larger surface areas than their corresponding support (Mg.Al.550.800) which might be explained by the mentioned re-structuring of the support surface due to the dissolution of alumina and magnesia into the acidic solution (containing Ni²⁺ and/or citric acid) during the impregnation step [30]. Ni/Mg.Al.1000 possesses relatively low surface area because its support was calcined at high temperature. Ni/Al.550 and Ni/MgO show textural parameters which are in agreement with their supports’ attributes.

The reducibility of the materials was examined by TPR measurements. It should be noted that Ni²⁺ species are the predominant reducible ones in those samples. The TPR profile of the unsupported NiO sample as the reference is exhibited in Figure 5 with an intense peak at 350 °C, illustrating the bulk NiO reduction. Generally, all studied supported Ni samples display poor reducibility, possibly due to their low Ni loading and high sample calcination temperature. Those effects might cause strong interaction of Ni²⁺ and the support, suppressing free NiO that can otherwise easily react with hydrogen [32,33]. On Al₂O₃, the Ni²⁺ species show high-temperature reduction peaks at 700 °C and the reduction is incomplete below 1000 °C (Figure 5). This behavior is typical for the reduction of nickel aluminate species. The mentioned samples were prepared at high calcination temperature (800 °C) and low Ni content (2.5 wt %), leading to unavoidable poor reducibility due to the diffusion of Ni²⁺ into the Al₂O₃ lattice, suppressing the formation of bulk NiO species [34]. The TPR profile of Ni/MgO discloses peaks at somewhat lower temperatures mainly assigned to the reduction of Ni²⁺ species in the outer face or subsurface layers of the MgO lattice. The remaining shoulder at higher temperature of the profile is related to the reduction of the species in solid solution, resulting from the diffusion of Ni²⁺ species into the periclase lattice due to high calcination temperature (800 °C) [32]. Ni/Mg.Al.550 shows a broad peak starting at 600 °C, reflecting the reduction of Ni²⁺ species mainly in the solid solution [35] and/or in the spinel structure when Ni is embedded in Mg and Al mixed oxides prepared from the hydrotalcite precursor [36]. The reducibility of that sample is slightly better than that for Ni/MgO or Ni/Al.550. Indeed, Ni/Mg.Al.550 displays a weak band at 500 °C in its TPR profile, attributed to the reduction of NiO species formed on the surface that have weak interaction with the support [37]. Additionally, it shows higher H₂ consumption than the single oxide supported samples (

Table 2. H₂ consumption in TPR for the catalysts.

Sample	H2 Consumption (µmol/g)
Ni/Al.550	397
Ni/MgO	181
Ni/Mg.Al.550	480
Ni/Mg.Al.1000	404
Ni/Mg.Al.550.CA	437

). This is similar to literature reports [20] concluding that the availability of Mg in Ni-Al containing samples facilitates the solid solution formation and stabilizes the Ni species on the surface. This benefit avoids the Ni diffusion into the alumina lattice, and therefore suppresses the formation of the poorly reducible NiAl₂O₄ spinel structure.

In comparison to Ni/Mg.Al.550, both Ni/Mg.Al.550.CA and Ni/Mg.Al.1000 show a shift in reduction peak maxima in their TPR profiles to higher temperature (900 °C) (Figure 5). Besides, they consume less hydrogen in the TPR experiment (Table 2), illustrating the formation of poorly reducible Ni²⁺ species that have higher interaction with the support frameworks [35,38].

XPS measurements were carried out to characterize the surface composition of the materials. The binding energy (BE) values were also measured to verify the oxidation states of the atoms as well as their chemical and physical environments. All of the samples show binding energies (BE) corresponding to the Ni 2p_3/2 region within the range around 855–857 eV (

Table 3. XPS (X-ray photoelectron spectroscopy) and ICP-AES (Inductively coupled plasma atomic emission spectroscopy) results for the catalysts.

Sample	Binding Energy (eV)	Surface Molar Ratio 1		Bulk Molar Ratio 2
	Ni 2p3/2	Ni/(Mg + Al)	Mg/Al	Ni/(Mg + Al)	Mg/Al
Ni/Al.550	855.5	0.029	-	0.022	0
Ni/MgO	- 3	- 3	- 3	0.017	-
Ni/Mg.Al.550	856.9	0.063	0.4	0.019	1.3
Ni/Mg.Al.1000	856.1	0.057	0.3	0.019	1.3
Ni/Mg.Al.550.CA	855.2	0.071	1.6	0.019	1.3
Mg.Al.550	-	-	1.2	-	1.3
Mg.Al.550.800	-	-	1.1	-	1.3
Mg.Al.1000	-	-	1.5	-	1.3

¹ XPS results; ² ICP results; ³ Ni content is below the XPS detection limit.

) [39]. Those are slightly higher than the values for free NiO (854.5 eV) [39,40] and close to that observed for Ni₂O₃ (856 eV) [35,39], NiAl₂O₄ (856 eV) [35,39], NiO-MgO solution (855.7 eV [40] or 856 eV [41]), and solid solution Mg(Ni.Al)O (855.5 eV) [35].

In any case, it could be suggested that the transfer of electrons from nickel to electron-poor magnesium and/or aluminum species in the structure causes the BE to shift to a higher value than that of free NiO [35]. Those results indicate a strong interaction of Ni with the supports, as seen in the TPR results, lowering their reducibility compared to pure NiO (Figure 5).

Regarding the Mg/Al surface ratio (Table 3) of samples with the Mg.Al support, there are differences between the Ni catalysts and their corresponding supports, suggesting the mentioned alumina and magnesia dissolution and re-dispersion of the surface components after preparation in some cases. On one side, impregnating Ni²⁺ onto the support in the presence of CA slightly increases the Mg/Al ratio from 1.1 to 1.6, but still maintains the Mg-rich surface. On the other hand, Ni/Mg.Al.550 and Ni/Mg.Al.1000 show Al-rich surfaces. Those observations are not fully understood, but are probably due to the Mg and Al leaching rate and re-dispersion in different precursor solutions during catalyst preparation. However, the surface Ni concentrations of those samples are still comparable.

Comparison of the surface Ni/(Mg + Al) ratio (from XPS) and bulk composition (from ICP) (Table 3) highlights the preferred location of Ni in different supports. Regarding the reference samples, Ni/MgO exposes almost no Ni on the surface and Ni/Al.550 shows a slightly higher surface Ni/(Mg + Al) ratio (0.029) compared to the bulk values (0.022). On the other hand, Ni/Mg.Al.550, Ni/Mg.Al.1000, and Ni/Mg.Al.550.CA expose Ni on the surface with significantly higher atom ratios (0.063, 0.057, and 0.071) compared to their bulk values (0.017–0.019). These data reflect the ability of Mg.Al supports to augment active Ni species on the outer shell of materials. Such an effect contributes to the higher reducibility of the Mg.Al supported samples compared to Ni/MgO and Ni/Al.550, as shown in the TPR profiles (Figure 5). Increasing the support calcination temperature from 550 to 1000 °C produces samples with less Ni²⁺ domains on the surface available for reduction [38]. As a result, Ni/Mg.Al.1000 provides poorer reducibility compared to Ni/Mg.Al.550 (Figure 5). This illustrates the ability of the Mg.Al.550 support with Mg-Al-O solid solution domains to enrich the surface Ni species and thus to enhance sample reducibility.

Introduction of CA during preparation increases the fraction of surface Ni²⁺ species. However, the reducibility of Ni/Mg.Al.550.CA as mentioned is not as high as that of Ni/Mg.Al.550 (Figure 5). The sample shows a sharp and symmetric reduction peak with a maximum temperature around 900 °C. These observations suggest that the low reducibility of Ni/Mg.Al.550.CA is primarily due to the higher dispersion of Ni species on the support surface [23], leading to a strong interaction with the Mg.Al material.

The coordination of the nickel (mostly Ni²⁺) species in the samples was examined by UV-Vis-DRS (Diffuse reflection spectroscopy). This technique is helpful for the identification of the nickel phases. No absorption can be seen in the support samples (not shown), suggesting that only Ni²⁺ species are sensitive at the chosen analysis conditions. The UV-Vis-DRS spectra of the Ni impregnated samples on different supports are shown in Figure 6. Those spectra show intense signals in the UV region of 250–350 nm, which relate to the O²⁻ → Ni²⁺ charge transfer (CT) in the octahedral NiO lattice [42]. NiO shows strong absorption over the whole mentioned range which can be assigned to nonstoichiometric NiO containing some Ni³⁺ domains [42]. Mg.Al supported samples show shoulders in the range of 250–350 nm which cannot be found in the spectra of Ni/MgO and Ni/Al.550, suggesting weaker interaction of Ni species with the Mg.Al supports [30]. In comparison to Ni/Mg.Al.550.CA and Ni/Mg.Al.1000, a red-shift of the CT band of NiO is observed for Ni/Mg.Al.550, which points to the weakest interaction of Ni species and support among the studied samples. This explains the TPR results, which proved that Ni/Mg.Al.550.CA and Ni/Mg.Al.1000 have lower reducibility compared to Ni/Mg.Al.550 (Figure 5).

A magnification of the UV-Vis-DRS spectra in the visible region (350–800 nm) of the mentioned samples can be seen in Figure 7. The signals in this region represent d-d transitions of Ni²⁺ ions mainly in octahedral (O_h) and tetrahedral coordination (T_d) [43]. Pure NiO presents intense signals at 380, 420, and 720 nm which illustrate the Ni²⁺ (O_h) in the cubic (rock-salt) NiO lattice [44]. Ni/Al.550, on the other hand, shows a doublet signal at 600–650 nm which is related to the Ni²⁺ (T_d) species in the nickel aluminate spinel lattice [44,45]. Besides, small absorption bands at 380 nm can be assigned to Ni²⁺ (O_h) in the NiAl₂O₄ spinel [46]. Ni/MgO displays a spectrum similar to that of the pure NiO sample. It shows an intense signal at 410 nm and a wide band at 650–800 nm which suggests a majority of Ni²⁺ (O_h) species. Those bands belong to the solid solution of Ni and Mg when Ni species migrate into the MgO parent lattice, which has a cubic rock-salt structure as well [47]. The absence of intense signals at 380 nm and 720 nm is probably due to the lack of bulk NiO in Ni/MgO [46]. This suggestion is supported by the XRD pattern of Ni/MgO (not shown), in which no crystalline structure other than periclase MgO can be seen. Indeed, the low content of Ni in the sample and the strong interaction between NiO and MgO results in the high dispersion of Ni species, thus suppressing the agglomeration of NiO in the calcination process. Ni/Mg.Al.550 shows a wide band from 380 nm to 410 nm with the highest intensity at around 400 nm, which is located between the corresponding signals of Ni/MgO and Ni/Al.550. This reveals the formation of octahedral coordination of the Ni domains which diffuse into the support framework of the solid solution Mg-Al-O. This kind of structure containing the three elements was also discussed in other studies [29,48] in which Ni²⁺ and Al³⁺ were substituted for the Mg²⁺ site in the parent MgO (periclase) structure (denoted as Mg(Ni,Al)O). Furthermore, it should be noted that the presence of Ni²⁺ (O_h) in the NiAl₂O₄ spinel can be stated for the Ni/Mg.Al.550 sample. In [30], the authors claimed the formation of Ni²⁺ (O_h) species in NiAl₂O₄-like “surface spinel”, identified by bands in the range of 405–410 nm. These surface spinel species were suggested to more likely exist in the samples with higher nickel content on the surface, similar to the surface features revealed by XPS in this study (Table 3). Briefly, Ni²⁺ (O_h) species are formed in both types of structures but the individual species are hard to identify and localize because of their overlapping absorption ranges in the UV-Vis-DRS spectra.

No significant difference between the spectra of Ni/Mg.Al.550.CA and Ni/Mg.Al.550 in the range of 350–800 nm can be found. However, the intensity of the broad signal at 720 nm in the spectrum of Ni/Mg.Al.550.CA (Figure 7b) is lower than that of Ni/Mg.Al.550, indicating less Ni²⁺ species in the domains of bulk NiO-like species which are reducible at low temperature (475 °C, Ni/Mg.Al.550, Figure 5). This suggests that higher dispersion of Ni species can be obtained when CA is added during the preparation. Compared to Ni/Mg.Al.550, a red-shift of the band at around 403 nm can be seen in the Ni/Mg.Al.1000 spectrum (Figure 7b). This shift toward the position of the corresponding intense signal of Ni/MgO illustrates the behavior of the NiO-MgO solid solution. Such a phase formation might be in accordance with the blue-shift of the CT band of the NiO lattice in this sample, displaying the higher stability of the Ni species (Figure 6).

2.2. Catalyst Performance in the DRM Reaction

Blank tests were carried out in the absence of catalysts or only with the supports in the inert material bed (not shown). No conversion of the feed gases was observed at the favored conditions. Figure 8 shows the CH₄ and CO₂ conversions in DRM during 8 h on stream over the supported Ni catalysts at 650 °C with various supports. The reaction conditions (650 °C, 1 bar, CH₄:CO₂ = 1, WHSV = 85 L/(g_cat × h)) were chosen to differentiate the activity of the catalysts. Higher activity can be obtained with Mg.Al supported Ni catalysts compared to the single oxide supported samples. Addition of CA during preparation plays a minor role regarding the activity. In contrast, supporting Ni²⁺ on Mg.Al.1000 causes a deterioration of conversion compared to Ni/Mg.Al.550. The latter catalyst exposed the highest activity with the conversions of CH₄ and CO₂ being very close to the thermodynamic limitations for CH₄ (71%) and CO₂ (82%) [49].

In order to verify such high activity of Ni/Mg.Al.550, additional DRM tests at lower temperature with the same WHSV were investigated. Figure 9 shows the excellent performance of Ni/Mg.Al.550 in comparison with the thermodynamic limitations [49]. The catalyst activates CH₄ and CO₂ already around 500 °C. It should be noted that DRM requires high temperatures to convert the reactants to syngas. According to the literature, the reaction starts at 350 °C, referred to by the thermodynamic calculations [49,50]. Moreover, remarkable conversions of methane and CO₂ were observed only above 500 °C [13]. Some studies claimed the catalysts to be active in DRM at very low temperature (400 °C [51] or 450 °C [50]). However, in those studies more beneficial conditions for high catalytic activity compared to this study were applied (lower WHSV, higher content of active sites, or usage of noble metals). Therefore, the activity of Ni/Mg.Al.550 in this study is outstanding at low temperature considering its low Ni content and the high WHSV (85 L/(g_cat × h)) in comparison with other investigations [3,8,52]. Figure 9 displays the effect of temperature on the H₂/CO ratio. This ratio is always lower than unity at any set reaction temperature, which indicates the occurrence of the reverse water gas shift (RWGS) reaction, lowering H₂ selectivity through the reaction with CO₂ to form CO and H₂O [1]. This reaction plays a more important role at lower temperature due to the thermodynamic equilibrium [1].

The coke deposition behavior in DRM was examined for the Ni catalysts by analyzing the carbon content in the catalysts after 8 h on stream (Figure 10). According to a previous DRM review [49], carbon formation occurs due to methane decomposition (MD) above 550 °C as well as the Boudouard (BD) reaction below 700 °C. Therefore, the reaction temperature in this study was set to 600 °C to study the coking behavior, which is one of the DRM main issues. Carbon deposition during DRM is known to cause catalyst deactivation and reactor plugging [53]. High WHSV (170 L/(g_cat × h)) was applied to clearly discriminate the coking resistance among the catalysts. Longer catalyst pre-reduction time (1.5 h) was applied in order to avoid the instant deactivation due to re-oxidation of Ni at a high flowrate of CO₂. The carbon amount found on spent Ni/Mg.Al.550 is higher than that of Ni/Mg.Al.1000 and Ni/Al.550 (Figure 10). Those amounts appear to be proportional to the conversions in DRM (Figure 8). The addition of CA during the preparation significantly contributes to designing a coking resistant catalyst. Indeed, the carbon amount on Ni/Mg.Al.550.CA is almost 8 times lower than that on Ni/Mg.Al.550 at the same reaction conditions and with similar activity. After 8 h on stream, the carbon content is as low as 0.77 wt % of the sample mass, which is in the range of carbonaceous remainders from the hydrotalcite precursor after calcination. This indicates the high coking resistance of the sample, which is a remarkable result compared to previous research [17,54], especially considering the Ni, Al, and Mg containing catalyst systems.

Long-term DRM was also carried out with Ni/Mg.Al.550.CA (Figure 11). The reaction was implemented at a considerably high WHSV of 170 L/(g_cat × h) compared to the literature data [6]. The conversions of CH₄ and CO₂ are very close to the thermodynamic limitations for CH₄ (82%) and CO₂ (86%) [49]. The catalytic performance is almost unchanged after 60 h on-stream, representing high stability of the catalyst in the reaction. At any TOS, the CO₂ conversion is about 8% higher than the CH₄ conversion, illustrating the contribution of the RWGS reaction. The CO selectivity achieves a constant value around 100% (not shown) which is consistent with negligible carbon deposition (<1 wt %) measured after the reaction (Figure 10).

3. Discussion

3.1. Activity of Mg.Al Supported Catalysts

Mg.Al supported catalysts show higher activity compared to single oxide supported samples. This is related mainly to their excellent reducibility and Ni enriched surface. This behavior has been also studied elsewhere [20,55], reporting that the coexistence of both Mg and Al in the supports reduces the formation of inactive species (NiO-MgO solid solution, NiAl₂O₄ spinel), stabilizes the Ni species on the surface, and subsequently enhances the DRM performance.

Ni/Mg.Al.550 is the most active catalyst for DRM among the studied samples because it possesses a high surface concentration of Ni in combination with the large surface of the support, forming species that are well reduced by H₂. Those active sites are proposed to be Ni²⁺ (O_h) in cooperation with the solid solution of MgO and Al₂O₃ (periclase structure) or surface NiAl₂O₄. Such phases are formed and stabilized at high calcination temperature (800 °C), but are still active at high WHSV and low temperature conditions even with low loading of Ni. The modification of such samples by using more severely calcined Mg.Al.1000 or adding CA during impregnation causes poorer reducibility and thus lower CH₄ and CO₂ conversions.

3.2. Coking Resistance and Stability of Mg.Al Supported Catalysts

It is well-known that coking during DRM is related to the large metal particles [4,56]. This is the result of the metal mobility on the surface and the subsequent agglomeration leading to less dispersed particles or clusters [49]. Ni/Mg.Al.550 presents the highest carbon deposition after 8 h on stream (Figure 10). However, the conversion is almost the same, the H₂/CO ratio is constant at 0.78, and the CO selectivity is almost 100% over the complete TOS. Therefore, such carbon deposition can be considered as negligible for catalyst performance. This high coking resistance can also be correlated to the low Ni content and high surface area of the support, by which both the metal and carbon species can be highly dispersed. Such dispersion results in less formation of big Ni particles and lower pore plugging by coke deposition, suppressing fast catalyst deactivation [53].

Ni/Mg.Al.550.CA shows the lowest carbon deposit amount after 8 h and even after 60 h, which is likely to be due to the stable dispersion of Ni species on the support surface. It was suggested above that the Ni/Mg.Al.550.CA sample has a stronger metal-support interaction and less NiO-like domains on the surface, contributing to a better coking resistance of the sample. This strong interaction also contributes to maintaining the catalyst performance in the long-term DRM experiment (Figure 11), by preventing Ni agglomeration and the loss of active sites.

There is another possible explanation for the coke resistance of Ni/Mg.Al.550.CA when comparing the Ni 3p_3/2 BE in the XPS results. The BE values, which are recorded at 856.91 eV and 855.18 eV (Table 3, Figure 12), are assigned to the surface NiAl₂O₄ in the Ni/Mg.Al.550 sample and the Ni-Mg-O solid solution in Ni/Mg.Al.550.CA.

Additionally, the surface atom Mg/Al ratio of Ni/Mg.Al.550 (Table 3) exposes an excess amount of Al species on the surface, also suggesting surface NiAl₂O₄ formation in Ni/Mg.Al.550 [38]. In contrast, the surface Mg concentration is highest in Ni/Mg.Al.550.CA, revealing the possibility of preferable Ni-Mg interactions on the surface. Our current explanation is that most of the Ni species on the surface of Ni/Mg.Al.550.CA might interact with Mg in the support, providing an effective carbon gasification by CO₂ (reverse Boudouard reaction) and thus coke deposition is suppressed [20], especially at high reaction temperatures (Figure 10). Therefore, Ni/Mg.Al.500.CA may show two beneficial effects that help to suppress coking during DRM: one correlates with the stable physicochemical properties and the other associates with the carbon gasification behavior.

4. Materials and Methods

4.1. Catalyst Preparation

Mg-Al mixed oxide (calcined Mg-Al hydrotalcite, Pural MG, Sasol) supported Ni catalysts were prepared by wet impregnation. Ni(NO₃)₂·6H₂O (Carl Roth, Karlsruhe, Germany) was used as the precursor for Ni. The Mg-Al hydrotalcite precursor possesses an Mg/Al ratio = 1.3 (data from ICP-OES and AAS measurement). Prior to impregnation, the precursor samples were calcined at different temperatures (550 °C, 700 °C, 800 °C, 1000 °C) for 6 h in air to obtain the corresponding bare supports which are denoted as Mg.Al.X, in which X represents the support calcination temperature. The calculated amount of Ni precursor was dissolved in deionized water and the solution was stirred for 4 h. The supports were then put into the solution and the slurry was kept stirring at 60 °C for 15 h. Water was removed by a rotary evaporator and the samples were dried overnight and calcined at 400 °C for 3 h and then at 800 °C for 6 h, both in air. The final catalysts are symbolized as Ni/Mg.Al.X. Citric acid (CA) was added simultaneously to the Ni precursor solution in the case of Ni/Mg.Al.550.CA preparation to study the influence of the chelating agent on impregnation. The nominal content of Ni in all supported Ni-containing catalysts was 2.5 wt %. Other materials were used as reference materials; those are MgO (FLUKA) and Al₂O₃ (boehmite calcined at 550 °C, Pural SB, Sasol) as well as their corresponding 2.5 wt % Ni samples (Ni/MgO and Ni/Al.550). One Mg-Al hydrotalcite sample was calcined first at 550 °C for 6 h and then at 800 °C for 6 h to obtain a material that provides properties similar to the support of Ni/Mg.Al.550.CA and Ni/Mg.Al.550, because those catalysts were calcined at 800 °C. That material is denoted as Mg.Al.550.800. Pure NiO was prepared by calcining Ni(NO₃)₂·6H₂O (Alfa Aesar) at 800 °C.

4.2. Catalyst Characterization

The metal (Ni, Mg, Al) contents of the samples were determined by inductively coupled plasma-optical emission spectroscopy (ICP-OES) using a 715-ES device (Varian, Inc., Palo Alto, CA, USA) and atomic adsorption spectroscopy (AAS) using an Analyst 300 apparatus (PerkinElmer, Inc., Waltham, MA, USA), respectively. The carbon deposition on spent catalysts was analyzed using a Truespace CHNS analyzer (LECO Corporation, Saint Joseph, MI, USA). After the reaction, the reactor was cooled to room temperature under Ar flow. Then the spent catalysts were separated from the inert material by sieving and transferred to the CHNS device.

The nitrogen physisorption method was used to calculate the specific surface area and pore volume distribution according to the BET theory and BJH calculation, respectively. The measurements were performed on a Micromeritics ASAP 2010 apparatus (Micromeritics GmbH, Aachen, Germany) at −196 °C. The samples were degassed at 200 °C in vacuum for 4 h before the analysis.

The XRD powder patterns were recorded either on a Panalytical X’Pert diffractometer (Almelo, The Netherlands) equipped with a X’celerator detector or on a Panalytical Empyrean diffractometer (Almelo, The Netherlands) equipped with a PIXcel 3D detector system using Cu Kα₁/α₂ radiation (40 kV, 40 mA) in both cases. Cu beta-radiation was excluded by using a nickel filter foil. The Cu K alpha2 radiation contribution was removed arithmetically using the Panalytical HighScore Plus software package (V. 3.0e, Panalytical, Almelo, The Netherlands). The peak positions and profile were fitted with the Pseudo-Voigt function using the WinXPow software package (V3.0.2, Stoe, Darmstadt, Germany). Phase identification was done by using the PDF-2 database of the International Center of Diffraction Data (ICDD).

XPS measurements were carried out with a VG ESCALAB 220iXL instrument (Thermo Fisher Scientific, Inc., Waltham, MA, USA) with monochromatic AlKα radiation (E = 1486.68 eV). The peak areas were determined after background subtraction and fitting with Gaussian-Lorentzian curves. The amount of each component in the near-surface region can be calculated based on these peak areas with division by the element-specific Scofield factor and the transmission function of the spectrometer.

For the H₂-TPR experiments, the measurement was done using a Autochem II 2920 instrument (Micromeritics, Aachen, Germany). A 300 mg sample was loaded into a U-shaped quartz reactor and heated from room temperature (RT) to 400 °C at 20 K/min in 5% O₂/He (50 mL/min) for 30 min at 400 °C, and was then cooled to RT and the sample was flushed with Ar. The TPR run was carried out from RT to 1000 °C in a 5% H₂/Ar flow (50 mL/min) with a heating rate of 10 K/min and was then held at the final temperature for 30 min before being cooled to RT. The hydrogen consumption peaks, indicating the reduction, were recorded using a TCD detector. The amount of hydrogen consumed was calculated based on the peak areas.

UV-Vis-DR spectra were measured over the wavelength range of 200–800 nm using a Cary 5000 spectrometer (Agilent Technologies, Petaling Jaya, Malaysia) equipped with a diffuse reflectance accessory (praying mantis, Harrick). In normal experiments, BaSO₄ is used as a reference white standard and dilute material for pure NiO measurement because of its high Ni content.

4.3. Catalytic Tests

DRM was carried out in a fixed bed continuous flow quartz reactor (ambient pressure, WHSV = 85 or 170 L/(g_cat × h); T = 500–700 °C). All volumetric flow rates given in this study are related to 25 °C and atmospheric pressure. After in situ pre-reduction in H₂ (700 °C, 100% H₂, 50 mL/min) for 1 h (used in the activity and stability tests) or 1.5 h (used for the coking test), the temperature was adjusted and maintained for 8 or 72 h and the reactant mixture (45 vol % CH₄, 45 vol % CO₂, 10 vol % He) was fed to the reactor. Helium was used as the internal standard for volume change estimation in the reaction. The gas compositions were then analyzed by an on-line Agilent 6890 gas chromatograph (Agilent Technologies, Santa Clara, CA, USA) equipped with a flame ionization detector (HP Plot Q capillary, 15 m × 0.53 mm × 40 µm) and a thermal conductivity detector (Carboxene packed, 4.572 m × 3.175 mm) for the analysis of hydrocarbons and permanent gases, respectively. Pure components were used as references for peak identification and calibration. Carbon balances were calculated from the gas products reaching more than 95% in this work. Conversions (X) and the H₂/CO ratio were calculated using the formulas given below:

$${\mathrm{X}}_{{\mathrm{CH}}_4}(\%)=\frac{{\mathrm{moles}\mathrm{of}\mathrm{converted}\mathrm{CH}}_4\text{×}100\%}{{\mathrm{moles}\mathrm{of}\mathrm{CH}}_4\mathrm{in}\mathrm{feed}}$$

(4)

$${\mathrm{X}}_{{\mathrm{CO}}_2}(\%)=\frac{{\mathrm{moles}\mathrm{of}\mathrm{CO}}_2\mathrm{converted}\text{×}100\%}{{\mathrm{moles}\mathrm{of}\mathrm{CO}}_2\mathrm{in}\mathrm{feed}}$$

(5)

$${\mathrm{H}}_2/\mathrm{CO}\mathrm{ratio}=\frac{{\mathrm{moles}\mathrm{of}\mathrm{H}}_2\mathrm{produced}}{\mathrm{moles}\mathrm{of}\mathrm{CO}\mathrm{produced}}$$

(6)