Carbothermally generated copper–molybdenum carbide supported on graphite for the CO₂ hydrogenation to methanol

Abstract

The carbothermal synthesis of monometallic and bimetallic molybdenum carbide and copper, supported on high surface area graphite (H), has been studied by in situ XRD, XPS, D₂-TPD, TEM/STEM, TG-mass spectrometry, and N₂ adsorption. The catalysts were prepared using H₂ at 600 °C or 700 °C and tested in the hydrogenation of CO₂ to methanol. Molybdenum carbide and oxycarbide phases were obtained, as well as hydride species, at 600 °C on both monometallic Mo_xC/H and bimetallic CuMo_xC/H in a similar proportion. Upon increasing the temperature up to 700 °C, the formation of metallic Mo is favourable. Although this is observed on supported Mo_xC and CuMo_xC, the bimetallic sample is less affected by the formation of the hydride, and molybdenum carbide is also observed upon treatment at 700 °C. With regards to the catalytic performance, supported monometallic copper was not active, but copper increased the activity and selectivity of the molybdenum carbide. The yield of methanol per catalyst's weight increases upon increasing the copper loading, indicating that a cooperation reaction takes place between the smallest Cu particles in contact with the molybdenum phase. The catalysts synthesized at 700 °C are less active and less selective to methanol favouring the reverse water gas shift under the studied conditions. Interestingly, the catalysts are stable under the reaction conditions, and the detected phases by XRD of the spent catalysts suggest that the hydride species favoured transformations involving MoO_xC_yH_z ↔ β-Mo₂C.

---

1. Introduction

The need for diminishing carbon dioxide emissions to comply with the target of not surpassing 2 °C increase in the Earth temperature during this century has prompted several strategies to achieve such an objective while maintaining our current industrial model. The transformation into other energy models is specially challenging for many industries such as petrochemical, steel or cement which are among the largest CO₂ emitters. Besides the direct use of CO₂, attention has been paid to its transformation into other chemicals so that it can be considered as a valuable chemical platform. This would be an interesting alternative for industries that cannot avoid CO₂ emissions in their production and they can use the CO₂ by-product as a raw material in their own facilities. In this sense, the hydrogenation of CO₂ to CH₄ or CH₃OH presents an optimal solution. Several heterogeneous catalysts have been studied, and many of them are based on the traditional system used for CO hydrogenation, Cu/ZnO, and others based on noble metals.^1,2

During the last few years, transition metal carbides have been attracting great attention as excellent hydrogenation catalysts, with catalytic performances that can reach or even surpass those of noble metals, with the advantage of their larger availability.³ For the transformation of CO₂ into methanol, molybdenum carbide has been successfully employed in the liquid phase, the bimetallic copper–molybdenum carbide being more selective to methanol than metallic iron and cobalt, and just slightly less selective than palladium.⁴ Other bimetallic systems formed with molybdenum carbide together with Au or Co have been compared to Cu–molybdenum carbide and it has been found that Cu promotes the methanol synthesis.^5,6

Several methods have been reported to synthesise molybdenum carbide using hydrocarbon/hydrogen mixtures at temperatures of up to 700–800 °C. However, the resulting catalysts are usually covered by graphitic carbon, and the surface areas are low, i.e. around 30 m² g⁻¹, for catalytic applications.⁷ To overcome these limitations, other alternative synthetic procedures have been reported in which organic precursors are employed, but still the coverage of the active phase with polymeric carbon appears to occur.^8–10

In order to find a suitable solution, carbon supports have been employed and a hydrogen atmosphere was used directly to obtain the corresponding carbide, using the support as a carbon source. This method, known as carbothermal reduction, has been employed with supports having different structures such as activated carbon, carbon nanofibers or carbon nanotubes, and different results regarding the molybdenum carbide phase have been reported.^11–13 These catalysts have been tested for several relevant industrial reactions such as methanol steam reforming, hydrodeoxygenation of stearic acid and also CO₂ hydrogenation, and correlation between the active phase and the catalytic results has been attempted.

Nonetheless, the identification of the carbide phase and quantitative determination of the stoichiometry is a challenging task in these carbon-based systems, and different results are reported in the literature. However, carbon is a very attractive support since it can provide carbon atoms for carburization, eliminating the need for an additional carbon source. The precise structure of the support also influences the structure of the resulting carbide.^11–13

Despite the research performed so far, less literature is found related to supported molybdenum carbide catalysts applied on this reaction. Hence, we have studied the synthesis of copper–molybdenum carbide catalysts on a commercial graphitic support with a high surface area and evaluated the effect of copper and carburization temperature on the physicochemical and catalytic performance of these catalytic materials.

2. Results and discussion

2.1 Active phase characterization

In situ XRD in the range of 2θ = 33–60° is shown in Fig. 1 and the whole pattern is shown in Fig. S1.† The XRD patterns of Mo_xC/H and CuMo_xC/H treated at 600 °C under H₂ display diffraction peaks with maxima at 2θ of 36.9–37.1° and 53.6°, and the wideness of both peaks suggests that they might include more than one contribution. According to reference patterns, the first peak may contain contributions from either MoO_xC_y and/or MoO₂, and the second peak from MoO₂.¹⁴ However, the possibility of these peaks including diffractions of hydride species cannot be disregarded as will be later explained. As expected, CuMo_xC/H also displays diffractions at around 43.1° and 50.0°, which are assigned to the (111) and (200) planes of metallic copper (JCPDS 04-0836). However, the peak at 2θ of 50.0° shifted to lower angles, which could be due to the additional contribution of MoO₃ whose (002) plane appears at 49.1° (JCPDS-PDF 76-1003). The absence of additional diffractions corresponding to Cu_xC that would appear at around 2θ of 36.0° indicates that under the employed conditions its formation is not favourable.¹⁵ This was also verified by XPS analyses, which showed no contributions attributed to Cu_xC, as shown in Fig. S2.†

Fig. 1 In situ XRD patterns of the catalysts under H₂. a) Mo_xC/H; b) 1CuMo_xC/H; c) 3CuMo_xC/H; d) CuMo_xC/H; e) Mo_xC/H-700; f) CuMo_xC/H700. XRD patterns of the spent catalyst after the reaction at 230 °C: g) Mo_xC/H-PR; h) CuMo_xC/H-PR; i) Mo_xC/H700-PR; j) CuMo_xC/H-700-PR. β-Mo₂C; ♦ MoO_xC_y; MoO₂/MoO_xC_yH_z; ♠ MoO₃; X Mo⁰; ΦCuO ■ Cu(0).

Interestingly, upon heating at 700 °C, the most intense diffraction peaks of Mo_xC/H appeared at 40.6° and 58.7° which correspond to the (110) and (200) planes of the cubic structure of Mo⁰ (JCPDS 42-1120), respectively. Under these conditions, the peak at 2θ of 37.1° is still observed, but its intensity is clearly lower than that of the sample carburized at 600 °C. Similarly, the catalyst CuMo_xC/H prepared at 700 °C led to a pattern different to that observed at 600 °C, since the diffraction peaks at 2θ of 36.9°, 37.7° and 39.5° are now present. In this case, it appears that the sample contains contributions attributed to MoO_xC_y, the first peak, and attributed to the (002) and (101) planes of the β-Mo₂C hcp phase, the last two peaks (JCPDS-PDF 77-0720). As observed for the Cu-free catalyst treated at 700 °C, diffractions at 2θ of 40.6° and 58.7° assigned to Mo⁰ appear but their relative intensity compared to those assigned to MoO_xC_y and β-Mo₂C are lower.

The nature of the molybdenum phases was further studied by XPS of the carburized samples using the Mo 3d region, as shown in Fig. S2,† from which the atomic contributions were obtained and are shown in Table 1. The Mo 3d_3/5 region displayed several contributions which can be assigned to the following oxidation states: Mo⁰ (227.7–227.9 eV), Mo²⁺ (228.5–228.6 eV), Mo^δ+ (229.4–229.6 eV), Mo⁵⁺ (231.2–231.6 eV) and Mo⁶⁺ (232.1 eV). Mo²⁺ corresponds to Mo in Mo–C bonds, while Mo⁵⁺ and Mo⁶⁺ are the species of MoO₃ and Mo₂O₅, respectively.¹⁶ Although MoO₃ appears due to incomplete carburization, the contribution of Mo₂O₅ is likely due to the reduction of the Mo⁶⁺ species under the high vacuum of the XPS analysis chamber. In addition, Mo^δ+ is an intermediate oxidation state between +4 and +2 that has been ascribed to MoO_xC_y species, which agrees with the phases identified by XRD.^17,18

Table 1 XPS Mo 3d region binding energies: species and percentage (%) in brackets

Catalyst	MoOx	MoOxCy	MoxC	Mo^0
	Mo^5+/6+	Mo^δ+	Mo^2+	Mo^0
MoxC/H	231.6 (5)	229.4 (38)	228.5 (57)	—
CuMoxC/H	231.1 (10)	229.4 (39)	228.6 (51)	—
MoxC/H700	232.1 (23)	229.6 (10)	228.5 (22)	227.7 (45)
CuMoxC/H700	232.1 (19)	229.6 (13)	228.5 (22)	227.9 (44)

---

The XPS spectra indicated that a similar MoO_xC_y/Mo_xC ratio is found on Mo_xC/H and CuMo_xC and that, in fact, the concentration of molybdenum carbide is more significant than what the XRD patterns showed. This is not surprising as the TEM images in Fig. 2 and S3† proved that a large fraction of molybdenum carbide is highly dispersed. Also, the XPS results confirm the presence of Mo⁰ on the samples carburized at 700 °C. Moreover, a similar relative proportion of Mo species estimated by XPS on the catalysts prepared at 700 °C agrees quite well with the comparable decrease of methanol yield obtained with these catalysts compared to their counterparts synthesized at 600 °C, which is in the range of 22–25% (Fig. 7).

Fig. 2 Representative TEM/HRTEM images of CuMo_xC/H a) and b) higher magnification and the corresponding fast-Fourier transform (FFT) pattern; c) and d) STEM images of CuMo_xC/H with the mapping of Cu (green) and Mo (red).

In order to obtain more insight into the catalytic system, D₂-TPD experiments of the carburised samples were performed. Before the experiments, the catalysts were subjected to D₂ pulses at room temperature. The profiles of desorbed gases from Cu/H and the support alone, as shown in Fig. S4,† did not show any desorption in the whole temperature range. However, the profiles of the samples Mo_xC and CuMo_xC exhibited the evolution of m/z = 2 at temperatures of ca. 500 °C. This result suggested that during catalyst activation, metal hydride nanoparticles are formed. The hydride formation was further verified by performing the D₂-TPD experiment on a sample previously reduced under D₂, as shown in Fig. 3. The reduction with deuterium followed by the desorption experiment proved the presence of the hydride since m/z = 3 ascribed to HD is also observed as a result of the hydrogen atom exchange.

Fig. 3 MS profiles after D₂ adsorption on CuMo_xC reduced under D₂ (solid) and H₂ (dashed).

These experiments showed that a hydride is formed on Mo_xC or CuMo_xC at temperatures above 400 °C and that neither Mo_xC nor CuMo_xC chemisorbed H₂ at room temperature. The presence of hydride species cannot be accurately detected by XRD or XPS since perturbations due to H atoms are negligible, and their contribution would be included in the observed species.

The synthesis of molybdenum carbides by the carbothermal method using H₂ and carbon supports takes place through several steps that we have previously identified by XANES.¹⁴ After the decomposition of the molybdate precursor into MoO₃, oxycarbide starts to be formed at ca. 300 °C, which is then converted into β-Mo₂C thanks to the contribution of carbon atoms from the support (reaction (1)). When the synthesis is carried out using hydrogen, it may help in the reduction process thanks to its diffusion coefficient. The higher Mo/C ratio, as shown in Table 2, for the samples treated at 700 °C agrees with the breakage of the crystallites during the synthesis, reported recently for Re carbides also supported on carbon.¹⁹ Such small nanoparticles cannot be accurately detected by TEM and the average particle size distribution, as shown in Table 1, is quite similar for the catalyst.

MoO₃ + C* + H₂ → MoO_xC_y + H₂O → β-Mo₂C + CO (1)

One possibility to explain the reduction to Mo⁰ could be that not enough reactive carbon, C* as shown in reaction (1), is available on the support, e.g. CH₄, to carry out the carburization of the molybdenum oxides and oxycarbides. In this context, the formation of Mo⁰ could be favourable in a mechanism involving the previously formed hydride.

Table 2 Textural and physical–chemical properties of the catalysts and XPS metal ratios

Catalyst	Mo dp TEM (nm)	S BET (m^2 g^−1)	XPS ratio
			Mo/C	Cu/C
Cu/H	—	307	—
MoxC/H	1.8 ± 0.2	328	0.007	—
CuMoxC/H	2.1 ± 0.3	205	0.013	0.231
MoxC/H-700	1.9 ± 0.3	304	0.015	—
CuMoxC/H-700	2.0 ± 0.3	202	0.036	0.327

---

The carburization process under H₂ was also studied using mass spectrometry to follow the evolved masses and the results of m/z = 18, 17 and 15 are shown in Fig. 4. It can be observed that the profiles for the Mo_xC/H and CuMo_xC/H samples are quite different.

Fig. 4 MS profiles during H₂-TPR of the impregnated samples. Mo_xC/H (dashed) and CuMo_xC/H (solid).

On the one hand, for the sample Mo_xC/H, the ratios of m/z = 18 and 17 indicate that they are due to water and ammonia desorption, representative of chemical surface changes such as precursor decomposition and Mo oxide reduction. The absence of m/z = 15 in the whole range of temperatures, which should correspond to CH₄ evolution, is not conclusive regarding the formation of oxycarbide and carbide species which are certainly identified by XPS. The evolution of CO, m/z = 28, from the support, as shown in Fig. S5,† could also be responsible for metal carburization although the intensity is also low. Nevertheless, the intense peak of m/z = 18 due to H₂O at temperatures above 600 °C, along with the XRD and XPS results, suggests that the previously formed molybdenum species were transformed into Mo⁰ in that temperature range. Reactions of the type MoO_xH_z → Mo⁰ + H₂O can be suggested. So, in agreement with XRD, the reduction to Mo⁰ is favourable on these samples above 600 °C.

On the other hand, for the sample CuMo_xC/H, the ratios of m/z = 18 and 17 can be ascribed mainly to water desorption at temperatures below 500 °C. The most interesting feature of the mass profile of the CuMo_xC/H sample compared to Mo_xC/H is the appearance of m/z = 15 in a proportion that can be undoubtedly ascribed to CH₄ evolution. This mass is already envisaged in the range of 300–500 °C, with maxima at around 400 and 450 °C, and then at 600 °C for CuMo_xC/H, an intense m/z = 15 signal is observed. This suggests that different proportions of intermediates and/or different paths are followed for Mo_xC/H and CuMo_xC/H, and that copper possibly aids in the carburization process by activating hydrogen and catalysing the support's activation to produce more available reactive carbon. Nevertheless, the MS profile of Cu/H, as shown in Fig. S5,† did not show the evolution of m/z = 15 which could indicate that the carburization is enhanced by a synergy between Cu and Mo. Alternatively, m/z = 15 could emerge from the reactions involving MoO_xC_yH_z species to give Mo⁰ and CH₄.

Still, the estimated XPS concentration of molybdenum carbide, oxycarbide and metallic species in both Mo_xC/H and CuMo_xC/H activated at 600 °C or at 700 °C is quite similar and differences are only observed in the XRD patterns. These results seem to indicate that the carburization of the largest particles is favourable on the bimetallic sample, CuMo_xC/H, upon increasing the temperature up to 700 °C.

Surprisingly, the XPS results of the samples synthesized at 700 °C display a larger concentration of MoO₃ than those treated at 600 °C. Considering the XRD patterns of the in situ carburized and spent catalysts and D₂-TPD results, it could be inferred that some hydride species are transformed into oxide upon increasing the temperature. This transformation can take place in the presence of the as-formed water and would be also in agreement with the conclusions from the MS profiles regarding the reactions at temperatures above 600 °C.

At this point, we can also propose that in the fresh carburized catalysts, small β-Mo₂C nanoparticles coexist along with larger oxycarbide particles detected by XRD. This conclusion agrees with the preferential carburization of the smallest particles in close contact with the support that we recently described.¹⁴

The carbothermal method using H₂ has been used for some carbon supports. For example, it has been reported that the synthesis of molybdenum carbide under H₂ over carbon nanofibers led to the formation of the oxycarbide phase at 550 °C, followed by the transformation into the β-Mo₂C phase starting at 650 °C.¹⁶ Other examples can be found in which the β-Mo₂C phase was formed at 700 °C over carbon nanotubes under H₂ (ref. 12) and over activated carbon, the β-Mo₂C phase was detected at 800 °C.¹³ There are few reports in the literature in which Mo⁰ is present when a carbon support and H₂ as a carburization atmosphere are used. Rodríguez et al.¹³ observed the formation of an unstable intermediate, MoO_xH_y, which was then converted to Mo⁰ upon heating at 400 °C with a heating ramp of 10 °C min⁻¹. Similarly, Xiao et al.²⁰ evidenced a small proportion of Mo⁰ on N, S, and P co-doped carbon nanospheres synthesised at 700 °C under diluted H₂. In contrast, Ochoa et al.¹⁶ did not observe the formation of Mo⁰ when fishbone-like carbon nanofibers were impregnated with the molybdenum precursor and treated up to 750 °C under H₂ atmosphere, with different heating ramps between 1 and 10 °C min⁻¹. The different reactivities of carbon atoms from the support could explain the differences. However, to the best of our knowledge, the carbon nanofibers used should not have a significantly higher proportion of reactive carbon than the other supports that could mask the formation of Mo⁰. In that case, the reason might be also related to other experimental parameters that would require further study, such as the gas hourly space velocity during carburization which may influence the contact time with hydrogen.

2.2 Reaction results

The conversion of CO₂ was evaluated at 20 bar and different temperatures using CuMo_xC (5 wt% Cu and 10 wt% Mo), monometallic Cu (5 wt% Cu) and Mo_xC (10 wt% Mo) supported on a high surface area graphite. Under these conditions, it was observed that the Cu-monometallic catalyst was not active, but it activated at either 250 °C (Cu/H) or 600 °C (not shown) under hydrogen, yielding values which are similar to those of the blank test. The main product obtained with all the catalysts was methanol. The methanol yields, as shown in Fig. 5, increased upon increasing the temperature up to 190 °C and 210 °C for Mo_xC/H and CuMo_xC/H, respectively, and then decreased due to the formation of CO through the reverse water gas shift reaction (RWGS). On the other hand, the yields of products CH₄ and CO, as shown in Fig. S6,† increase in the whole temperature range.

Fig. 5 Methanol formation rate per gram of catalyst at 20 bar, WHSV = 3600 h⁻¹ and CO₂ : H₂ : He (1 : 3 : 6).

These results indicated that monometallic copper supported on the graphitic support employed is not active on the hydrogenation of CO₂ unlike Cu with a well-known behaviour supported on ZnO or ZrO₂ where CO₂ is adsorbed and activated on the oxide.^1,21 This agrees with simulations on pure Cu that proved weak CO₂ interaction with copper surfaces leads to low activity for CO₂ hydrogenation.^22,23 The copper surfaces on monometallic Cu/H that could be identified by TEM, as shown in Fig. S2,† are characterized from the TEM images by fast-Fourier transformation (FFT) and the (111) and (200) crystal planes, which agree with their apparently low ability to activate CO₂ in contrast to the activity of copper defective surfaces such as (311) and they are ascribed to Lewis acidity.²⁴

In addition, the TEM images of CuMo_xC/H in Fig. 2a and b and Fig. S2† show that mainly highly dispersed Mo_xC particles of less than 3 nm are generated. The identification of these particles as Mo_xC is based on the results obtained by XRD and XPS. Moreover, individual large copper particles and/or agglomerates of copper particles, ca. 50–100 nm as shown in Fig. 2c, can be found and barely any physical contact was observed between Cu and Mo. Also, the proportion of the molybdenum phases, i.e. carbide and oxycarbide, on both Mo_xC/H and CuMo_xC/H estimated by XPS, as shown in Table 1, is similar. Hence, a synergy between Cu and Mo_xC is suggested.

Therefore, additional experiments with different Cu loadings were performed at 150 °C, in order to compare under differential conditions, and are shown in Fig. 6. The average yields of CH₃OH, CH₄ and CO increased upon increasing the copper content, and the highest copper loading reached values of ca. 0.21 μmol CH₃OH g⁻¹ s⁻¹. It thus seems plausible that the observed large Cu particles coexist along with small Cu particles, even in the range of single metal atoms, scarcely detected by TEM/STEM or XRD as shown in Fig. 2d. The agglomeration of copper was indeed corroborated by the Cu/Mo atomic XPS ratio, as shown in Fig. S7,† which deviates from the theoretical value more significantly upon increasing the copper loading. Although the large Cu particles do not show any activity on the reaction as the results for the monometallic catalyst showed, the smallest Cu particles are located close to Mo_xC. In this case, the Cu–Mo_xC boundary would be the responsible for the synergy, increasing the conversion and selectivity to methanol.²⁵

Fig. 6 CO₂ conversion rate and CH₃OH, CH₄ and CO formation rate at 20 bar, WHSV = 3600 h⁻¹, CO₂ : H₂ : He (1 : 3 : 6) and 150 °C.

Then, we increased the carburization temperature for Mo_xC/H and CuMo_xC/H to 700 °C aiming at reaching a higher proportion of carbide and better reaction yields. However, the results shown in Fig. 7 are actually the opposite of what is expected, and both the conversion rate of CO₂ and formation rate of methanol are lowered by 20–23% and 22–25%, respectively, compared to the values obtained with the catalysts prepared at 600 °C. Moreover, it can be observed that the selectivity to CO increases for the catalysts synthesized at 700 °C, which is more significant when using Mo_xC/H-700. A similar trend is observed for the selectivity to CH₄.

Fig. 7 CH₃OH formation rates and CO₂ conversion rates and selectivities at 20 bar. WHSV = 3600 h⁻¹, CO₂ : H₂ : He (1 : 3 : 6) and 150 °C.

Stability tests for 24 hours at 230 °C were performed with Mo_xC/H and CuMo_xC/H carburized at 600 and 700 °C. The results, shown in Fig. 8, showed a stable profile for both catalysts prepared at 600 °C and at 700 °C.

Fig. 8 CH₃OH formation rates at 20 bar. WHSV = 3600 h⁻¹, CO₂ : H₂ : He (1 : 3 : 6) and 230 °C.

Overall, the catalytic performance obtained with Mo_xC/H and CuMo_xC/H carburized at 600 °C and 700 °C is in line with their different proportions of Mo phases. The lower activity of Mo⁰ compared to Mo_xC seems to be related to the formation of an HO₂C–Mo intermediate in which the scission of C–O is not energetically favourable.²⁶ Nevertheless, the catalytic results obtained suggest that at least some proportion of the as-formed Mo⁰ particles are active and that they promote the selectivity to CO as the selectivity in Fig. 4 shows.

The XRD patterns of the spent catalysts after the reaction at 230 °C are shown in Fig. 1. The patterns of Mo_xC/H and CuMo_xC/H show diffractions at 2θ of 34.3°, 37.7° and 39.4° attributed to the β-Mo₂C phase and they are not observed on the fresh catalysts. On the other hand, the XRD pattern of the spent catalyst CuMo_xC/H-700 displayed a diffraction at 2θ of 39.9° not accompanied by that at 58°, which can then be attributed to MoO₃ likely resulting from Mo⁰ oxidation.²⁷ With regards to the XRD pattern of spent Mo_xC/H-700, only a broad hump in the range of 35–39° is observed, also reflecting the instability of the Mo⁰ species. The XPS analyses of the spent Mo_xC/H and CuMo_xC/H, as shown in Fig. S2,† only displayed one contribution of MoO₃, due to the unavoidable passivation layer.

The results indicate that once the hydride species are formed at 600 °C, the nanoparticles may follow two routes. Although the hydride particles under reaction conditions are transformed into β-Mo₂C, heating at 700 °C under H₂ leads to the formation of Mo⁰.

As the XRD pattern of the spent catalyst showed, the Mo⁰ particles are not chemically stable under the reaction conditions, and this likely explains why the large particles, which the XRD of the fresh catalyst showed, are not observed by microscopy where the samples are exposed to air. Moreover, the spent catalysts show diffractions at 2θ of 36.4° for CuMo_xC/H and 34.8° and 38.8° for CuMo_xC/H-700 which correspond to copper oxide.²⁸

Also, since the catalysts' methanol yield was stable in the evaluated period, it is reasonable that the active phases are also stable and that the observed changes after the reaction correspond mostly to inactive species.

The synthesis of methanol using CO₂ hydrogenation may proceed by activation of CO₂ and conversion to CO via reverse water gas shift or formate/formaldehyde transformation into methanol. In general, molybdenum carbide suffers from deactivation when molecules with oxygen atoms are in the reaction atmosphere.²⁹ For example, during CO₂ hydrogenation over bulk catalysts, oxycarbides were identified as the most abundant species.³⁰ Under reaction, both phases, i.e. carbide and oxycarbide, coexist and are involved in a redox cycle where CO₂ reacts with Mo₂C through C–C interaction, producing oxycarbide and CO, which could eventually be reduced by hydrogen to obtain Mo₂C again. Based on the results obtained in the present work, we can add that the hydride can aid in the regeneration/formation of the carbide during the reaction.

Concerning the effect of copper, some previous literature regarding the catalytic performance of copper supported on ZnO or ZrO₂ in CO₂ hydrogenation suggested that the activation of hydrogen is boosted by copper; then, activated hydrogen is spilled over and reacts with the activated CO₂ adsorbed on molybdenum carbide.³¹ Despite the D₂-TPD experiments as shown in Fig. S4† indicating that copper does not adsorb hydrogen at room temperature, we cannot rule out this phenomenon under reaction conditions. Nevertheless, the positive effect of Cu is confirmed since higher yields are obtained over CuMo_xC/H and CuMo_xC/H-700 compared to Mo_xC/H.

The methanol production activities obtained in the present work with CuMo_xC/H (see Fig. 5) are higher than the value reported for Cu supported on carbon nanofiber/ZrO₂, for which 0.2 μmol g⁻¹ s⁻¹ at 180 °C and 30 bar was reported,³¹ and to that obtained with codoped nanospheres at 220 °C and 20 bar, which reached 0.24 μmol g⁻¹ s⁻¹. The results obtained with other systems of the type Cu/ZnO or Cu/ZrO₂ offered 0.08–0.09 μmol g⁻¹ s⁻¹ activities to methanol at 250 °C and 20 bar.³²

The obtained results proved that the carburization of molybdenum supported on high surface area graphite with H₂ not only produces active and selective molybdenum phases but also they are stable in the presence of oxygen species. This is relevant since one of the main problems of molybdenum and other transition metal carbides is their transformation into oxycarbides and oxides upon exposure to oxygen-containing molecules. Hence, the carbothermal conditions can be tuned to optimize the stability and regeneration of the catalyst. The results also highlight the difficulties in assessing the active phase by XRD as the most active phases are mainly small nanoparticles.

3. Experimental

3.1 Catalyst synthesis

The support employed was commercial high surface area graphite (HSAG400 from Timcal Graphite), labelled H (S_BET = 399 m² g⁻¹). Metal carbides were prepared by wet co-impregnation on this support with the corresponding metal precursors. An aqueous solution of the Mo precursor, (NH₄)₆Mo₇O₂₄ (99% from Aldrich), and Cu precursor, Cu(NO₃)₂·3H₂O (99% from Aldrich), was used to impregnate the support with an adequate amount to obtain the desired metal loading, left for maturation for 6 h, and dried overnight at 100 °C. The carburization was carried out for 2 h under pure H₂ at 600 or 700 °C (5 °C min⁻¹). The resulting catalysts were labelled according to the metal composition. The Mo loading was 10 wt% in all cases, while the Cu loading was 5 wt% in both the bimetallic and the monometallic samples, except in the catalysts where it is explicitly indicated to be 0.5, 1 and 3 wt%. Monometallic Cu supported on HSAG400 was also prepared by wet impregnation using the same metal precursor and activated at 250 °C (Cu/H) and 600 °C (Cu/H600) under H₂ for 2 h.

In this way, eight samples were prepared on HSAG400 and labelled: Cu/H; Cu/H600; Mo_xC/H; Mo_xC/H-700; 0.5CuMo_xC/H 1Cu/Mo_xC/H; 3Cu/Mo_xC/H; CuMo_xC/H; CuMo_xC/H-700.

3.2 Characterization

Textural properties of the samples were obtained from the N₂ adsorption isotherms at −196 °C, using a 3Flex instrument from Micromeritics. Around 100 mg of sample was previously degassed for 4 h at 120 °C under vacuum, using a SmartVacPrep instrument from Micromeritics. The surface area was calculated from the adsorption branch in the range of 0.02 ≤ p/p₀ ≤ 0.25, using the Brunauer–Emmett–Teller (BET) model.

X-ray diffraction (XRD) patterns of the catalysts were acquired in situ by passing the corresponding atmosphere, H₂, using a reaction chamber (Anton Paar XRK900). The 2θ range was between 4° and 90°, with a step of 0.04° s⁻¹, using a Polycristal X'Pert Pro PANalytical diffractometer with Ni-filtered Cu Kα radiation (λ = 1.54 Å) operating at 45 kV and 40 mA. For the XRD patterns of the spent catalysts, the samples were passivated after the reaction by passing diluted O₂ for 2 hours at room temperature.

Photoelectron spectra (XPS) were recorded using an Escalab 200R spectrometer equipped with a hemispherical analyser and using non-monochromatic Mg Kα X-ray radiation (hν = 1253.6 eV). The samples were treated under the same conditions as those used in the carburization and transferred to an octane solution to avoid oxidation, and transferred to an outgassing chamber. Prior to the experiments, the samples were outgassed in situ for 24 h to achieve a dynamic vacuum below 10⁻¹⁰ mbar. The binding energy (BE) was measured by reference to the C 1s peak at 284.6 eV, with an equipment error of less than 0.01 eV in the energy determination. The surface atomic ratios were estimated from the integrated intensities of Mo 3d, Cu 2p, C 1s and O 1s lines after background subtraction and correction by the atomic sensitivity factors. The spectra were fitted to a combination of Gaussian–Lorentzian lines of variable proportions.

TG-mass experiments were conducted with TA Instruments SDT Q600 TA equipment from 30 °C to 700 °C at 5 °C min⁻¹ in H₂ flow (100 cm³ min⁻¹, STP). The output gases from TGA were monitored by mass spectrometry using a ThermoStar GSD 301 T3 instrument (with a filament at 150 °C, SEM and emission detector at 950 mV). The m/z masses analysed were: 2, 18, 17,16, 15, 28 and 44.

For the D₂-TPD measurements, before each experiment, the catalyst sample (100 mg) was pretreated in a quartz reactor under hydrogen (H₂) flow for 2 hours at 600 °C, passed to a quartz bulb, outgassed under high vacuum and treated again under H₂ or a deuterium atmosphere at 500 °C for 1 hour. At ambient temperature, a deuterium pulse (30 Torr) was applied to the samples until equilibrium was reached and outgassed again under high vacuum. This step was omitted when the sample was treated with deuterium at 500 °C. Once stabilized, the temperature was raised to 500 °C at 10 °C min⁻¹ monitoring the evolved gasses from the samples with a mass spectrometer (SRS-RGA200). Mass to charge ratios (m/e) obtained were 2, 3, 4, 16, 17, 18, 28, 32 and 44.

Information about the supported metal particles was acquired by TEM on a JEOL 2100F field emission gun electron microscope operated at 200 kV and equipped with an energy-dispersive X-ray detector. The samples were ground until powdered and a small amount was suspended in acetone using an ultrasonic bath. Some drops were added to the gold grid (Aname, Lacey carbon 200 mesh) and acetone was evaporated at room temperature before introducing to the microscope. Scanning transmission electron microscopy (STEM) was done using a spot size of 1 nm.

3.3 Reaction

The evaluation of the gas phase catalytic performance was carried out in a stainless steel fixed-bed flow reactor of 3/8′′ internal diameter. The catalysts (0.5 g) were placed in the reactor and in situ carburized in pure H₂, by increasing the temperature up to 100 °C, keeping this temperature for 1 h and then raising it up to 600 °C or 700 °C, with a heating rate of 5 °C min⁻¹. The treatment at this temperature was carried out for 2 h. Finally, the catalysts were cooled under He to the reaction temperature, 150 °C. The reactor was pressurised at 20 bar using a reactant mixture composed of CO₂ : H₂ : He = 1 : 3 : 6 vol, and the total flow during the reaction was 30 ml min⁻¹. The reactants and products were analyzed by gas chromatography (Varian CP 3400) with an FID and TCD, fitted with SupelQ Plot and 60/80 Carboxen-1000 columns, respectively.

The reaction conditions allowed the conversion in all the experiments to be maintained to assume a differential reactor. The carbon balance was over 95% in all cases, and blank tests resulted in a conversion below 0.5%. The conversion and product selectivity were obtained according to the following equations:

where:

n _i: the number of carbon atoms of product i.

mol_i: the number of moles of product i.

molCO₂-un: mol of unreacted CO₂.

Conclusions

Overall, the characterisation showed differences depending on the carbothermal conditions. The XRD and XPS results of the Mo_xC/H catalyst displayed mainly MoO_xC_y/β-Mo₂C phases, indicating that incomplete carburization took place under the studied conditions. Also, D₂-TPD profiles showed that hydride species are formed. Increasing the carburization temperature up to 700 °C did not result in a larger extent of carburization and, in contrast, Mo⁰ was obtained as the main phase. For the analogous samples containing Cu, similar results were obtained, but for the sample treated at 700 °C, the proportion of Mo⁰ identified by XRD was lower compared to MoO_xC_y/β-Mo₂C, which is proposed to be a consequence of the enhanced H₂ activation by copper, favouring the production of CH₄ that would be the carburization source.

In addition, the incorporation of copper improves the catalytic performance, and this seems to happen even in the absence of nearby physical contact between Cu and Mo_xC. This can be explained by the presence of copper particles with few atoms and scarcely identified by microscopy, but the possibility of hydrogen activation on copper and spill over phenomena to β-Mo₂C cannot be ruled out.

Overall, the resulting catalysts are quite stable under reaction conditions although water is formed during the reaction, which is known to oxidize molybdenum carbide. However, the reaction may proceed through a path that implies a redox mechanism in which the molybdenum carbide is oxidized to an oxycarbide phase and further reduced again into the carbide by the hydrogen of the reactant feed. In this proposed mechanism, the exact role of the detected bulk hydride species needs further investigation.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

A. B. Dongil acknowledges financial support from the Fundación General CSIC (Programa ComFuturo and iLink project No 20211 from CSIC (Spain)). Financial support from the Spanish Agencia Estatal de Investigación (AEI) and EU (FEDER) (projects MAT2016-80285-P, CTQ2017-89443-C3-1-R and CTQ2017-89443-C3-3-R) is also acknowledged.

Notes and references

1. I. U. Din, M. S. Shaharun, M. A. Alotaibi, A. I. Alharthi and A. Naeem, J. CO₂ Util., 2019, 34, 20–33.
2. C. Li, X. Yuan and K. Fujimoto, Appl. Catal., A, 2014, 469, 306–311.
3. D. W. Flaherty, S. P. Berglund and C. B. Mullins, J. Catal., 2010, 269, 33–43.
4. Y. Chen, S. Choi and L. T. Thompson, J. Catal., 2016, 343, 147–156.
5. W. Xu, P. J. Ramírez, D. Stacchiola, J. L. Brito and J. A. Rodriguez, Catal. Lett., 2015, 145, 1365–1373.
6. S. Posada-Pérez, P. J. Ramírez, J. Evans, F. Viñes, P. Liu, F. Illas and J. A. Rodríguez, J. Am. Chem. Soc., 2016, 138, 8269–8278.
7. J. Cao, J. Wang, Y. Ma, X. Li, P. Xiaokaiti, X. Hao, A. Abudula and G. Guan, J. Alloys Compd., 2018, 735, 1463–1471.
8. T. Huang, W. Huang, J. Huang and P. Ji, Fuel Process. Technol., 2011, 2, 1868–1875.
9. Z. Yao, J. Jiang, Y. Zhao, F. Luan, J. Zhu, Y. Shi, H. Gao and H. Wang, RSC Adv., 2016, 6, 19944–19951.
10. Z. Yao, F. Luan, Y. Sun, B. Jiang, J. Song and H. Wang, Catal. Sci. Technol., 2016, 6, 7996–8004.
11. P. Liang, H. Gao, Z. Yao, R. Jia, Y. Shi, Y. Sun, Q. Fan and H. Wang, Catal. Sci. Technol., 2017, 7, 3312–3324.
12. L. Souza Macedo, R. R. Oliveira, T. van Haasterecht, V. Teixeira da Silva and H. Bitter, Appl. Catal., A, 2019, 241, 81–88.
13. W. Xu, P. J. Ramirez, D. Stacchiola and J. A. Rodríguez, Catal. Lett., 2014, 144, 1418–1424.
14. D. H. Carrales-Alvarado, A. B. Dongil, J. M. Fernández-Morales, M. Fernández-García, A. Guerrero-Ruiz and I. Rodríguez-Ramos, Catal. Sci. Technol., 2020, 10, 6790–6799.
15. J. Ding, X. Yan, B. K. Tay and Q. Xue, J. Phys. Chem. Solids, 2011, 72, 1519–1523.
16. E. Ochoa, D. Torres, R. Moreira, J. L. Pinilla and I. Suelves, Appl. Catal., A, 2018, 239, 463–474.
17. Y. Qin, L. He, J. Duan, P. Chen, H. Lou, X. Zheng and H. Hong, ChemCatChem, 2014, 6, 2698–2705.
18. A. Tougerti, E. Berrier, A. S. Mamede, C. La Fontaine, V. Briois, Y. Joly, E. Payen, J. F. Paul and S. Cristol, Angew. Chem., Int. Ed., 2013, 52, 6440–6444.
19. E. Blanco, A. B. Dongil, J. L. García-Fierro and N. Escalona, Appl. Catal., A, 2020, 599, 117600–117609.
20. W. Geng, H. Han, F. Liu, X. Liu, L. Xiao and W. Wu, J. CO₂ Util., 2017, 21, 64–71.
21. P. B. Rasmussen, P. M. Holmblad, T. Askgaard, C. V. Ovesen, P. Stoltze, J. K. Nørskov and I. Chorkendorff, Catal. Lett., 1994, 6, 373–381.
22. J. Nakamura, J. A. Rodriguez and C. T. Campbell, J. Phys.: Condens. Matter, 1989, 1, SB149–SB160.
23. Y. Yang, J. Evans, J. A. Rodriguez, M. G. White and P. Liu, Phys. Chem. Chem. Phys., 2010, 12, 9909–9917.
24. K. Jonglack, B. B. Sarma, E. Andrés, N. Pfänder, P. Concepción and G. Prieto, ACS Catal., 2019, 9, 10409–10417.
25. S. Posada-Pérez, P. J. Ramírez, R. A. Gutiérrez, D. J. Stacchiola, F. Viñes, P. Liu, F. Illas and J. A. Rodriguez, Catal. Sci. Technol., 2016, 6, 6766–6777.
26. H. Tominaga and M. Nagai, Appl. Catal., A, 2005, 282, 5–13.
27. M. Rahmani and K. Kalantar-Zadeh, et al. , Sens. Actuators, B, 2011, 145, 13–19.
28. A. Dutta and M. Oezaslan, et al. , J. Catal., 2020, 389, 592–603.
29. A. Kumar, S. Phadke and A. Bhan, Catal. Sci. Technol., 2018, 8, 2938–2953.
30. M. D. Porosoff, X. Yang, J. A. Boscoboinik and J. G. Chen, Angew. Chem., Int. Ed., 2014, 53, 6705–6709.
31. Y. Chen, S. Choi and L. T. Thompson, ACS Catal., 2015, 5, 1717–1725.
32. H. Ren, C.-H. Xu, H.-Y. Zhao, Y.-X. Wang, J. Liu and J.-Y. Liu, J. Ind. Eng. Chem., 2015, 28, 261–267.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/d1cy00410g

---