Dry reforming of methane over ZrO₂-supported Co–Mo carbide catalyst

Abstract

The process of dry reforming of methane has the potential to be an effective route for CO₂ utilization via syn-gas production. In the present study, ZrO₂-supported Co–Mo bimetallic carbide catalysts were prepared via a co-precipitation method through a combined reduction and carburization procedure employing a CH₄/H₂ (20/80 %) mixture. All of the as-synthesized materials were tested at 850°, under atmospheric pressure and a CO₂:CH₄ ratio of 1. The importance of the ZrO₂ support became immediately apparent when it exhibited a higher conversion than the corresponding low-surface-area bulk Mo₂C catalyst, which we attribute to lewis acid and base active sites on the surface of ZrO₂. From catalytic tests and pre-and post-reaction X-ray diffraction (XRD) patterns, we observed that different dispersions of the monometallic carbides, caused by varying the pre-heating temperatures on ZrO₂, did not significantly affect conversion or yield. In contrast, incorporation of cobalt atoms into the Mo₂C lattice significantly enhanced the conversion, yield and stability of the catalysts. Post-reaction XRD patterns indicated that the bimetallic carbide had enhanced the resistance to the oxidation effect that is known to deactivate Mo₂C catalysts. In addition, increasing the Co loading in the mixed metal carbides was seen to enhance the resistance of the catalyst to the reverse water gas shift reaction, leading to improved stability of the H₂ yields.

Introduction

The dry reforming of methane $\left(\text{DRM},{\text{CO}}_2+{\text{CH}}_4\to 2\text{CO}+2{\text{H}}_2,\mathrm{\Delta }{\text{H}}_{298}^{\text{o}}=+247{\text{KJmol}}^{-1}\right)$ is an effective method to utilize CO₂, via a reaction with CH₄ to produce a mixture of CO and H₂ known as “synthesis gas (syn-gas)” [2]. This reaction is, however, highly endothermic and generally high temperatures are required both for significant levels of conversion as well as for reducing side reactions. The DRM process provides several advantages over steam reforming of methane $(\text{SRM},{\text{H}}_2\text{O}+{\text{CH}}_4\to \text{CO}+3{\text{H}}_2,\mathrm{\Delta }{\text{H}}_{298}^{\text{o}}=+206{\text{KJmol}}^{-1})$, and perhaps the most important one is the production of syn-gas with a low H₂/CO ratio, which is suitable for use in forming higher level alcohols [10] (in a stoichiometric reaction, the H₂/CO ratio of DRM production is 1:1 while SRM has a product ratio of 3:1). Additionally, of course, DRM does not require the use of water to produce syn-gas.

Metal carbides are a relatively new family of catalysts for DRM. York et al. [24] studied DRM reactions over β-Mo₂C and WC with a direct comparison with noble metal catalysts. Importantly, the results showed that β-Mo₂C has an activity for DRM comparable to certain noble metals, while the price of molybdenum is much cheaper than noble metals. Furthermore, compared to Ni-based catalysts (arguably the earliest-found material to have a high activity in DRM), Mo₂C, has higher stability due to its enhanced resistance to coking.

The dispersion of catalyst onto a support surface is another critical factor which can significantly influence the catalytic activity of materials [1]. The supported catalyst can play an important role in promoting reaction, which has been proven by many researchers after making a direct comparison between bulk and supported materials on a variety of supports, such as TiO₂, Al₂O₃, SiO₂, and ZrO₂ [3, 12, 19].

In recent years, therefore, supported transition metal carbides have gained increased prominence since the support can improve both the efficiency and the stability of carbides in the DRM reaction. Systematic studies on Mo₂C loaded on different supports have been carried out by Brungs et al. [7] and Darujati et al. [11]. Their results revealed that during a long-term DRM test, when Mo₂C is supported by ZrO₂ and γ-Al₂O₃, the catalytic activity and stability were higher than that for materials loaded on other oxide based supports. The advantage of γ-Al₂O₃ appears to be its significantly higher surface area compared to other kinds of support. But the advantage of ZrO₂ is its amphoteric nature [8]. The Lewis acid sites enhance the dispersion of metal due to the preference of metal atoms to reside at Lewis acid sites on the support [6], whereas the Lewis base sites can enhance the adsorption of CO₂ on the support in conjunction with Lewis acid sites [4]. Moreover, on ZrO₂, it is proposed that CO₂ activation takes place at the interface between carbide and support, which leads to a low oxidation effect on carbides [17].

To-date, the activity and stability of catalysts during DRM reaction have been improved only by a small improvement. Moreover, the product selectivity of catalysts (i.e. specifically the H₂/CO ratio) appeared to differ from a supported catalyst because they have the different activities on CO₂ decomposition and CH₄ cracking and/or the existence of side reactions such as reverse Water–Gas shift $(\text{RWGS},{\text{CO}}_2+{\text{H}}_2\to \text{CO}+{\text{H}}_2\text{O},\mathrm{\Delta }{\text{H}}_{298}^{\text{o}}=+41{\text{KJmol}}^{-1})$. Hence, one of the major initiatives for catalyst improvement stem from the need to improve selectivity, by reducing side reactions. Additionally the need to balance the two DRM half reactions, CH₄ cracking and the reverse boudouard reaction, is essential to avoid carbon deposition.

In seeking the high activities (including both the conversion of reactants and the yields of products) and high stabilities, here we prepared the ZrO₂-supported Mo/Co–Mo carbide catalysts for the dry reforming of methane. The incorporation of cobalt into the bimetallic system is because of its high activity in CH₄ decomposition [8] combined with relatively high resistance to coking, when compared to nickel metal [16]. Co–Mo bimetallic carbides have earlier been used for the DRM as an alternative to expensive noble metal [18], and such Co–Mo bimetallic carbides were systematically prepared and characterized by ** rate was set at 10 °C/min. After the calcination process, the sample of MoO₃/ZrO₂ was obtained.

To obtain the carbide materials, 3.0 g of prepared oxides were carburized in a quartz tube under a gas flow of CH₄/H₂ (8 and 32 ml/min respectively) for 3 h at 750 °C, and both the ram** and cooling rate were 3 °C/min. After carburization, the carbide samples were passivated in static air for 48 h before unloading. The carbide samples were labelled with a percentage of Mo₂C loading in terms of their oxide precursors. All the synthesized catalysts were labeled as X %-Mo-Y (X is the loading of MoO₃ in wt% and Y denotes the pre-heating temperature).

Besides the supported Mo₂C catalysts, the pure ZrO₂ support and the bulk Mo₂C carbide were also prepared from zirconia catalyst support (Alfa-Aesor, 43815, 90 m²/g) and ammonium molybdate tetrahydrate (98 %, Sigma Aldrich-A7302), respectively via the same method of calcination and carburization processes.

With the aim of preparing Co–Mo bimetallic carbide samples, the required amount of ammonium molybdate tetrahydrate and cobalt(II) nitrate hexahydrate (98 %, Sigma Aldrich) were dissolved in water using the same method as above and the percentage in weight of the bimetallic oxides were 10 and 15 wt% while the Co/Mo ratio was kept at 0.4/0.6 which has been employed previously and found to possess the highest stability in pyridine hydrodenitrogenation reactions [

$$\begin{array}{c}\text{Conversion}=C_i=1-\frac{X_i^{\text{outlet}}{X}_{N_2}^{\text{inlet}}}{X_i^{\text{inlet}}{X}_{N_2}^{\text{outlet}}}\left(i={\text{CH}}_4\text{or}{\text{CO}}_2\right)\\ \text{Volume}\text{of}\text{gas}\text{in}\text{product}=\frac{X_i^{\text{inlet}}\text{Volume}\text{of}{\text{N}}_2}{X_{{\text{N}}_2}^{\text{outlet}}}\left(i={\text{H}}_2\text{or}\text{CO}\right)\\ \text{Yield}\text{of}{\text{H}}_2=Y_{{\text{H}}_2}=\frac{\text{volume}\text{of}{\text{H}}_2\text{in}\text{products}}{2\ast \text{volume}\text{of}\text{methane}\text{in}\text{reactants}}\\ \text{Yield}\text{of}\text{CO}=Y_{\text{CO}}=\frac{\text{volume}\text{of}CO\text{in}\text{products}}{\text{volume}\text{of}C\text{containing}\text{reactants}}\\ \text{Product}\text{Ratio}={\text{H}}_2/\text{CO}=\frac{\text{volume}\text{of}{\text{H}}_2\text{in}\text{products}}{\text{volume}\text{of}CO\text{in}\text{products}}\\ \text{Carbon}\text{Balance}=B_{\text{carbon}}=\frac{(X_{{\text{CH}}_4}^{\text{outlet}}+X_{{\text{CO}}_2}^{\text{outlet}}+X_{\text{CO}}^{\text{outlet}})X_{{\text{N}}_2}^{\text{inlet}}}{(X_{{\text{CH}}_4}^{\text{inlet}}+X_{{\text{CO}}_2}^{\text{inlet}})X_{{\text{N}}_2}^{\text{outlet}}}\\ \end{array}$$

Characterization of catalyst pre- and post- reaction

All the samples, including ZrO₂ support, the various precursors and carbides, were characterized by high-resolution X-ray diffraction (XRD) using a PANalytical X’Pert PRO diffractometer with CuKα radiation (45 kV, 40 mA). The samples were flat loaded in the custom-built sample holders and scanned from 10° to 70° 2θ with a step size of 0.0084° and a scanning speed at 0.017778° s⁻¹.

Raman spectra were collected (resolution of 2 cm⁻¹) using a PerkinElmer Raman Stage 400F with a 785-nm laser. The samples were pressed onto a microscope slide and the scanning time was set to 20 s (repeated 10 times) within the spectral range from 100 to 2,500 cm⁻¹.

Brunauer, Emmett and Teller (BET) analysis via nitrogen adsorption was carried using a Micrometrics ChemiSorb 2720 to measure the surface area and pore volume of samples both before and after the DRM reactions at the temperature of liquid nitrogen (−195.8 °C).

The carbon deposits of catalysts pre- and post- reaction were measured via thermo-gravimetric analysis (TGA). The instrument employed was a PerkinElmer, TGA-7, using flowing air at 10 ml/min from 50 to 1,000 °C with a ramp rate of 10 °C/min.

Results and discussion

DRM over supported Mo₂C catalysts

Figure 1 shows the XRD patterns of ZrO₂-supported Mo₂C with different loading levels at different pre-heating temperatures. It is apparent that Mo₂C is dispersed differently over the surface of ZrO₂ supports which are pre-heated at 400 and 750 °C, respectively. The peak of Mo₂C at 2θ = 39.49° can be clearly observed on 12 %-Mo-750 carbides while the peaks at the same position can hardly be observed on 12 %-Mo-400, indicating that Mo₂C has a considerably enhanced dispersion on the 400 °C-pre-heated support (this has been shown to have a higher surface area according to Table 2).

Fig. 1

X-ray diffraction patterns of ZrO₂ supported Mo₂C with different loading level (9–24 wt%) and on ZrO₂ supports with different pre-heating temperature (750 °C, 400 °C, as indicated)

In Fig. 2, Laser Raman spectra of Mo₂C 6–24 % loaded on 400 °C-pre-heated ZrO₂ were presented. It was difficult to observe the ZrO₂ support when the loading level was above 12 %, which indicates that the ZrO₂ support had been fully covered by Mo₂C, demonstrating that the availability of the ZrO₂ surface for potential catalytic activation decreased as the loading level increased.

Fig. 2

Laser Raman spectra of Mo₂C 6 %-24 % loaded on 400 °C-pre-heated ZrO₂ (notice: the appearance of MoO₃ in Laser Raman spectra is because the outer layer Mo₂C was oxidized during the passivation step after carburization, and the very thin layer of MoO₃ cannot be observed in the XRD patterns in Fig. 1)

During the process of DRM reaction, CH₄ absorption and subsequent dissociation is usually recognized to be the first step, as well as the rate determining step (1) [13, 20, 21, 23], and this is followed by the oxidation of carbon species (adsorbed carbon and CH _x ) on a catalyst surface by CO₂, forming CO (2):

$${\text{CH}}_4\leftrightarrow {\text{CH}}_x+\left(4-x\right)\text{H}\leftrightarrow \text{C}+2{\text{H}}_2$$

(1)

$${\text{CO}}_2+\text{C}\leftrightarrow 2\text{CO}$$

(2)

Besides the DRM, a side reaction called “reverse water gas shift” (RWGS, (3)) can lead to a higher conversion of CO₂ and the consumption of generated H₂.

$${\text{H}}_2+{\text{CO}}_2\to \text{CO}+{\text{H}}_2\text{O}$$

(3)

The activity test results presented in Table 1 show very poor performance of the bulk Mo₂C to activate and convert both CH₄ and CO₂ at atmospheric pressure, and the conversions are even lower than the ZrO₂ support which is “carburized” under the same conditions to prepare carbide catalysts. This can be attributed to the low surface area of the bulk Mo₂C, which leads to a low availability of Mo₂C active sites on the surface of material (which is different from the high-surface-area metal carbide materials prepared by Claridge et al. [9].). The activity in converting both CH₄ and CO₂ over pure ZrO₂ support is attributed to its amphoteric property [8]. Both Lewis acid sites and base sites can enhance the adsorption of CH₄ and CO₂ on ZrO₂ surface [4] where the reactants can be activated and converted. According to the product ratio, significantly more CO is generated than H₂, and this is due to the RWGS (reaction (3)).

Table 1 Conversion of CH₄ and CO₂, yield of H₂ and CO, product ratio and carbon balance over different samples (bulk Mo₂C, pure ZrO₂ support and ZrO₂ supported Mo₂C catalysts) during DRM reactions (conditions: 850 °C, atmospheric pressure, 4-h duration, CH₄/CO₂ 1, GHSV 4.8 × 10³ ml h⁻¹ g ⁻¹_cat )

Importantly, higher activities and stabilities were observed over supported Mo₂C compared to the bulk Mo₂C. Both reactant’s conversions and product yields were enhanced due to the Mo₂C dispersion effect, and there is the possibility to attribute this improvement to a synergy between Mo₂C and ZrO₂ support.

Comparison of the catalysts with the same pre-treated supports (400 or 750 °C) revealed that a higher Mo₂C loading level led to a lower conversion of CH₄ and CO₂. This may be due to high levels of Mo₂C dispersed on the surface decreasing the availability of the ZrO₂ support acid sites themselves (shown in Fig. 2, the higher loading level of Mo₂C leads to a lower intensity of ZrO₂ peaks) which as started previously is active for the DMR over its Lewis acid and base sites. Despite the lower CH₄ and CO₂ conversions, the samples with higher loading levels (18 %-Mo-400 and 24 %-Mo-750) also had a similar or slightly higher product yields compared to low loading level. It could be concluded that the higher availability of ZrO₂ surface led not only to a higher conversion but also a higher degree of side reactions (e.g. RWGS) which affected the yields of products.

The catalytic performances over 12 %-Mo-400 and 12 %-Mo-750 (Table 1) were shown to be close to each other at 4 h even though the conversions of CH₄ and CO₂ over 12 %-Mo-400 were higher than the other ones. This phenomenon can be explained by a further sintering effect on the materials when reacting temperature reached to 850 °C, and both catalysts reached the similar Mo₂C dispersion level (even though they looked different before the reaction, as shown in Fig. 1).

The carbon balance during reaction over each sample was also displayed in Table 1. It was observed that the carbon balance over Bulk-Mo₂C was above 100 %, revealing that Mo₂C was consumed by CO₂ to form MoO₂ and CO (4). The numbers of carbon balance over ZrO2 support and ZrO2 supported catalysts were slightly below 100 %, which indicates that amorphous carbon was generated during the reaction.

$${\text{Mo}}_2\text{C}+5{\text{CO}}_2\to 2{\text{MoO}}_2+6\text{CO}$$

(4)

The results of the surface area and pore volume of samples both prior to (“pre-”) and post-reaction (“post-”) are presented in Table 2. A noticeable difference between the pre-reaction surface areas of the two samples was observed, indicating a more highly sintered ZrO₂ support after being heated at 750 °C.

Table 2 Surface area and pore volume of samples (ZrO₂ supported Mo₂C catalysts) pre- and post- reactions

Moreover, the difference between the post-reaction surface areas of the two samples is much smaller, which is caused by a further sintering effect when the catalyst was heated up to 850 °C during the reaction, and this sintering effect was possibly one of the main routes of deactivation during the dry reforming of methane.

Comparison of the XRD patterns of the supported carbides pre- and post-reaction is presented in Fig. 3. They showed that the MoO₂ peaks at 2θ = 26.11° and 37.12° were observed in the post-reaction catalysts. This phenomenon indicated the oxidation of Mo₂C to MoO₂ with CO₂ (4), which is known to be another main deactivation mechanism of DRM [9].

Fig. 3

X-ray diffraction patterns of the ZrO₂ supported Mo₂C catalysts before and after DRM (P post-reaction)

A new peak corresponding to the Mo₂C peak at 2θ = 39.49° was observed. This indicated that Mo₂C crystallized when the surface area of ZrO₂ support decreased during the 4-h activity test (shown in Table 2). The crystallite size of ZrO₂ was also increased as calculated with XRD patterns shown in Fig. 3. (Pre-reaction 340–370 nm; post-reaction 570–630 nm), which is another evidence of minimized sintering effect existing on ZrO₂ support.

In order to elucidate whether the catalysts experienced carbon deposition over the catalysts during the dry reforming of methane, TGA (thermo-gravimetric analysis) was undertaken. Results for the pre- and post- reaction 18 %-Mo-400 catalysts are displayed in Fig. 4 with the calculation of weight change at each temperature range exhibited in Table 3. The weight increased between 200 and 470 °C, which indicated the oxidation of Mo₂C to MoO₃. Weight loss was observed starting at 470 and 800 °C and related to the combustion of amorphous carbon [14] and the sublimation of MoO₃ [15], respectively. The graphitic carbon which is expected to be combusted between 600 and 800 °C [5] can only be observed at an arguably low level. Upon comparison of these two TGA curves, more amorphous carbon was observed on the post-reaction catalyst, which means the generation of carbon via the decomposition of CH₄ (reaction (1)) occurred on the surface of catalyst during the dry reforming of methane.

Fig. 4

Thermo-gravimetric analysis of the pre- and post- reaction 18 %-Mo-400 catalysts

Table 3 Weight changes in thermo-gravimetric analysis of the pre- and post- reaction 18 %-Mo-400 catalysts at different temperature ranges

DRM on supported Co–Mo carbide catalysts

As an extension of this work based on the results of mono-metallic carbides, bimetallic carbide catalyst systems were also investigated to enhance the activity and stability of catalysts in the dry reforming of methane. The results can be summarized as follows.

In the XRD patterns after incorporating cobalt with the molybdenum catalyst in Fig. 5, the peaks corresponding to the CoMoO₄ structure at 2θ = 23.33°, 25.50° and 26.50° were observed. In addition, the Co₃O₄ phase was not observed. This indicated that only CoMoO₄ had been formed during the calcination process. After carburisation, phases of Mo₂C at the angle of 2θ = 38.101° and 39.492° were observed clearly and there appeared to be no evidence of Co metal particles or Co carbide in the XRD patterns. According to **ao et al. [22], this phenomenon is attributed to the direct incorporation of cobalt into the lattice of Mo₂C carbide, and due to the comparable atomic radius of Co (r ≈ 126 pm) compared with that of Mo (r ≈ 154 pm), the framework of Mo₂C did not change significantly even though the lattice was partially substituted by Co atoms.

Fig. 5

X-ray diffraction patterns of ZrO₂ supported 10/15 % Co–Mo bimetallic oxides and carbides (C carbide, O oxide)

The TGA results of the pre-reaction bimetallic catalysts are exhibited in Fig. 6 with the calculation of weight change at each temperature range exhibited in Table 4. From the weight loss observed starting around 470 °C (which is related to the combustion of amorphous carbon), it indicates that the increased loading level of the bimetallic carbide enhanced the formation of carbon during the carburization step. The weight loss after 800 °C was much less than that exhibited by the monometallic catalysts (as shown in Fig. 4).

Fig. 6

Thermo-gravimetric analysis of the pre-reaction Co–Mo bimetallic catalysts

Table 4 Weight changes in thermo-gravimetric analysis of pre-reaction Co–Mo bimetallic catalysts at different temperature ranges

The catalyst activity test results over 10, 15 and 20 % bimetallic carbides are shown in Table 5. With the incorporation of Co into Mo₂C, there was a significant increase in both CH₄ and CO₂ conversion: The conversion of CH₄ was nearly complete and the conversion of CO₂ also reached above 95 % when the loading level of the bimetallic carbide reached 15 % or higher. This very large activity improvement was due to the introduction of Co which has been proven to be very active metal in the DRM reaction. In conjunction with the above, the stabilities of both CH₄ and CO₂ conversions were significantly improved with the addition of Co components. In addition, the stability of both H₂ and CO yields were enhanced as the loading level increased.

Table 5 Conversions of CH₄ and CO₂, yields of H₂ and CO, product ratio and carbon balance over bimetallic carbide samples during DRM reactions (conditions 850 °C, atmospheric pressure, 4 h duration, CH₄/CO₂ 1, GHSV 4.8 × 10³ ml h⁻¹ g ⁻¹_cat )

The H₂/CO product ratio over all the three bimetallic carbide catalysts (as shown in Table 5) were above 1, which indicated the CH₄ decomposition (reaction 1) was favoured over the bimetallic catalysts. As mentioned in monometallic carbide catalyst discussion, the availability of ZrO₂ contributes to the conversions of both CH₄ and CO₂, and on the other hand, it also improves the side reactions which can lead to the observed difference between the conversions and the product yields. As exhibited in Table 5, the conversions of CH₄ and the yields of H₂ were arguably at the same level. This phenomenon indicated that the incorporation of Co into the Mo₂C catalysts can significantly depress the Reverse Water Gas Shift (reaction (3)).

The TGA results of the post-reaction bimetallic catalysts are exhibited in Fig. 7 with the calculation of weight change at each temperature range exhibited in Table 6. No weight losses are observed on 10 %-Co–Mo-400 and 15 %-Co–Mo-400, and arguably, there is a reduction of weight loss from 470 °C on 20 %-Co–Mo-400 (the gain of weight from 200 °C over 20 %-Co–Mo-400 is masked by the great weight loss from 470 °C, which means the carbon hasn’t been totally consumed during DRM). This phenomenon on bimetallic catalysts indicated the consumption of carbon by CO₂ (${\text{CO}}_2+\text{C}\leftrightarrow 2\text{CO},$ reaction (3)) during the dry reforming of methane, favouring to the CO production. However, the carbon balances of reaction over bimetallic catalysts (Table 5) were below 100 %, which indicates that carbon-containing product apart from CO and coke was generated during the reaction, and the unexpected product is hypothesized as light hydrocarbons.

Fig. 7

Thermo-gravimetric analysis of the post-reaction Co–Mo bimetallic catalysts

Table 6 Weight changes in thermo-gravimetric analysis of pre-reaction Co–Mo bimetallic catalysts at different temperature ranges

The very high stabilities of Co–Mo bimetallic carbide catalysts were also exhibited in their XRD patterns after being tested in the dry reforming of methane. In Fig. 8, only the peaks corresponding to CoMoC _x were observed on the ZrO₂ support and there appeared to be no evidence of MoO₂ or CoMoO _x generation. Compared to the formation of MoO₂ over ZrO₂ supported Mo₂C catalysts (shown in Fig. 3), these catalysts exhibited significantly enhanced oxidation resistance as the bi-metallic carbide phase, this could be explained by the presence of cobalt in the carbide, which may directly enhance the oxidative resistance of the carbide, or provide a re-carburisation route for the oxidized molybdenum (Table 6).

Fig. 8

X-ray diffraction patterns of ZrO₂ supported 10/15 % Co–Mo bimetallic carbides before and after the dry reforming of methane (P post-reaction)

Conclusions

These investigations have shown that bulk Mo₂C did not exhibit a significantly high activity in the dry reforming of methane (DRM); this was attributed to the very low surface area when using the preparation method as described. On the other hand, pure ZrO₂ support was treated with the same method to give ZrO₂ supported Mo₂C catalyst, and this exhibited a relatively higher activity compared to the bulk Mo₂C. It appeared that Lewis acid and base sites on the support ZrO₂ had contributed to this enhanced activity.

The catalytic activity decreased slightly with time on stream for the supported Mo₂C over a period of 4 h. It was thought that this deactivation was caused by two effects: one was the oxidation effect on Mo₂C by CO₂, and the other one was the loss of surface area caused by sintering effect of the ZrO₂ support.

ZrO₂ supported cobalt and molybdenum bi-metallic carbide catalysts were also prepared and examined for catalytic activity. Significant enhancements were observed for both the activity and the stability of the catalysts in the DRM. It was observed also that the incorporation of cobalt atoms into the Mo₂C lattice significantly improved the materials’ resistance to aerial oxidation. Compared to the mono-metallic carbide, the bi-metallic catalysts clearly have significantly higher CH₄ conversion. Moreover, an increase of Co loading in the mixed carbide also reduced the influence of Reverse Water Gas Shift, which improved the stability of H₂ yields.