Co-activation of methane and nitrogen to acetonitrile over MoC_x/Al₂O₃ catalysts

Abstract

Methane and nitrogen are regarded as the most abundant sources of hydrocarbons and nitrogen, respectively. Both compounds exhibit high stability due to the presence of daunting C–H and N≡N bond energies of 439 and 945 kJ mol⁻¹, respectively, leading to their abundance. This study investigates the co-activation of methane and nitrogen using Al₂O₃-supported MoC_x catalysts (1, 5, and 10 wt%) to produce acetonitrile (ACN) at ambient pressure. It was found that the optimum methane conversion (26.1%) and turnover frequency (TOF) of ACN (15.3 h⁻¹) were achieved at 750 °C using 1 wt% Mo loading. To alleviate catalyst deactivation resulting from coking, H₂ co-feeding was implemented and found to effectively sustain the on-stream activity for 50 hours. A plausible mechanism for ACN production, which occurs on the MoC-like surface, was proposed.

---

Introduction

Methane (CH₄), the major component of natural gas, shale gas, and biogas, has long been used as a heat resource by combustion.¹ Because of its abundance, CH₄ is considered an important building block in the production of value-added chemicals.^2,3 For example, CH₄ can be converted to syngas (CO and H₂) by steam or dry reforming^4–7 and further converted to hydrocarbons by Fischer–Tropsch synthesis.^8,9 CH₄ can also be converted to olefins, aromatics, or oxygenated compounds (e.g., formaldehyde and methanol) by non-oxidative pyrolysis, dehydroaromatization, or selective partial oxidation.^9–11 However, the chemical inertness and high C–H bond dissociation energy of CH₄ hinder its conversion to chemicals.⁹Fig. 1 shows the existing routes for catalytic CH₄ conversion, including CH₄ to olefins, aromatics, and hydrogen (MTO, MTA, and MTOAH), oxidative coupling (OCM), and dehydroaromatization (MDA).

Fig. 1 Existing routes of CH₄ conversion to chemicals.²

Besides the conventional reforming processes (e.g., steam and dry reforming), limited studies have been dedicated to the direct conversion of CH₄ to chemicals through non-oxidative routes. McFarland et al. reported the production of H₂ and solid carbon derived from CH₄ catalyzed by molten metal alloys.¹² The catalysts were prepared by dissolving active metals (Ni, Pd, and Pt) in low-melting-temperature metal solvents (Sn, Pb, Bi, In, and Ga) at temperatures below 1000 °C. 27 mol% Ni in Bi was the most active catalyst, achieving a 1.7 × 10⁻⁸ mol_H2 cm⁻² s⁻¹ hydrogen production rate with more than 95% selectivity at 950 °C. The Kopyscinski group studied silica-supported GaN prepared by ammonia nitridation for CH₄ conversion to ethylene.¹³ They found that CH₄ can be converted to ethylene at above 700 °C. Our group recently discovered that GaN made by co-pyrolyzing gallium nitrate and organic nitrogen compounds (e.g., melamine, melem, and g-C₃N₄) could convert CH₄ to acetonitrile (ACN).^14,15 The impact of particle size of GaN and residual CN species (including C=N and C≡N) on the enhancement of ACN production was found to be significant. However, it is worth noting that the residual CN species cannot be regenerated during the conversion process and gradually get depleted.

Currently, ACN is obtained primarily as a by-product (approximately 2–4%) from acrylonitrile synthesis known as the Sohio process.¹⁶ This results in a comparatively low ACN output that may not meet the escalating demand for ACN. In 2021, the global ACN market was valued at USD 264 million and is expected to grow at an annual rate of 5.5%.¹⁷ Therefore, exploring alternative routes for ACN synthesis to meet the growing demand is needed. The advent of the shale gas revolution has led to a significant increase in natural gas production and a subsequent decline in CH₄ prices.^18,19 In 2017, the price of ACN in the United States was approximately nine times higher than that of CH₄.²⁰ Accordingly, the conversion of CH₄ to ACN should be worth exploring.

Taking the Mars–van Krevelen redox cycle (Fig. 2) as an example,²¹ the oxide catalyst delivers its mobile oxygen to partially oxidize the reactant to form an O-containing product, leaving the reduced catalyst in the first half of the cycle. Then, the reduced catalyst is re-oxidized by gaseous oxygen, fulfilling the redox cycle.

Fig. 2 Schematic illustration of the Mars–van Krevelen redox cycle, where MO is a metal oxide, M is the reduced form, R is a reactant, and RO is an O-containing product.

For the production of ACN via CH₄ conversion over supported GaN catalysts, only the first half of the redox cycle is completed. The addition of ammonia as a co-feed was expected to facilitate the remaining half of the cycle. However, the production of ACN was significantly reduced with even small amounts of co-fed NH₃ due to the in situ formed H₂ from NH₃ decomposition.¹⁴ Therefore, using N₂ as a nitrogen source for ACN synthesis seems to be the most promising approach. In other words, a highly effective CH₄ conversion catalyst for ACN production must be capable of activating CH₄ and N₂ simultaneously. Tiwari et al. reported that CH₄ and N₂ could be activated simultaneously over a K-doped Ru catalyst to co-produce ammonia and ethylene.²² The in situ H₂ produced from CH₄ conversion was used as a hydrogen source for ammonia synthesis. However, coke is formed in parallel with hydrogen production, resulting in discontinuous ammonia productivity. Nevertheless, this study introduces the possibility of converting two inert compounds into value-added chemicals.

The crucial step in N₂ activation is the cleavage of the N≡N bond. Mo₂C has been demonstrated to be active in ammonia synthesis.²³ As predicted by the d-band model, transition metal carbides with vacant d orbitals exhibit comparable adsorption behavior towards electron-rich adsorbates.²⁴ In addition, the sp bands of transition metals in metal carbides would hybridize with the d bands from the transition metals and with the s bands from the carbon.²⁵ The extra hybridized orbitals provide more opportunities for back-donation to the π orbitals of the adsorbates, making Mo₂C a promising N₂ activation active site. Moreover, Mo₂C was found as an active site in CH₄ dehydroaromatization.^26–28 Accordingly, the Mo₂C catalyst should have strong potential for simultaneous activation of CH₄ (C–H) and N₂ (N≡N) to synthesize ACN.

In this study, the catalytic conversion of CH₄ and N₂ to ACN using Al₂O₃-supported MoC_x catalysts was investigated. We found that the catalyst could simultaneously convert both CH₄ and N₂ to produce ACN. The optimum operating temperature and Mo loading were investigated. Catalyst deactivation was suppressed by introducing H₂ into the reaction stream, allowing the catalytic activity to be maintained for 50 h. In addition, a plausible mechanism for ACN formation was proposed to occur via the Mo-terminated MoC-like surface structure, which facilitates the reaction between CH₄ and N₂. Kinetics analysis was performed to support the proposed mechanism, in which the surface reaction was suggested to be rate-limiting. The goal of this research is to validate the feasibility of converting two inert compounds (CH₄ and N₂) into a valuable chemical (ACN).

Experimental section

Chemicals

Alumina (Al₂O₃, 99.5%), ammonium heptamolybdate ((NH₄)₆Mo₇O₂₄, 99.95%), molybdenum oxide (MoO₃, 99.95%), and commercially available molybdenum carbide (com-Mo₂C, 99.5%) were obtained from Merck, J. T. Baker, Alfa Aesar, and Sigma-Aldrich, respectively. CH₄ (99.999%), argon (Ar, 99.999%), nitrogen (N₂, 99.999%), and hydrogen (H₂, 99.999%) were purchased from Air Products. Al₂O₃ was calcined at 750 °C before being used as a support. All chemicals were used as received.

Catalyst preparation

Al₂O₃-supported MoC_x catalysts were prepared by incipient-wetness impregnation followed by carbonization. Briefly, a designated amount of ammonium heptamolybdate solution (0.5 M) was added dropwise onto the Al₂O₃ support to obtain xMo/Al₂O₃ (x = 1, 5, or 10 wt%). The impregnated catalyst was dried at 110 °C overnight and was then calcined at 750 °C (5 °C min⁻¹) for 5 h. Afterward, the calcined catalyst was sieved and crushed into 40–80 mesh particles. The carbonization treatment was performed before activity evaluation.

Characterization

An inductively coupled plasma mass spectrometer (ICP-MS, THERMOELEMENT XR) was utilized to quantify the Mo content. X-ray diffraction (XRD) patterns were recorded at 40 mA and 4 kV using a diffractometer (Rigaku D/Max-IIB) equipped with Ni-filtered Cu Kα (λ = 1.5406 Å) radiation. A transmission electron microscope (TEM, JEOL JEM-2010) equipped with an energy dispersive X-ray spectrometer (EDS) was used to analyze the morphology of MoC_x particles. N₂ sorption isotherms were analyzed using an automated physisorption analyzer (ASAP 2020 Plus, Micromeritics). The surface area was evaluated by applying the Brunauer–Emmett–Teller (BET) method at a relative pressure of 0.01–0.1.

X-ray absorption spectroscopy (XAS) at the Mo K-edge was performed at the Taiwan Photon Source (TPS) 44A beamline of the National Synchrotron Radiation Research Center (NSRRC), Taiwan. Mo-foil was used for energy calibration. The XAS spectra were analyzed by using the Athena and Artemis software ver. 0.9.26 included in the Demeter package.²⁹

The CO pulse titration technique was conducted at 196 K to determine the dispersion of MoC_x particles using a chemisorption analyzer (AutoChem II, Micromeritics) and the signals were recorded using a thermal conductivity detector (TCD). Ammonia and nitrogen temperature-programmed desorption (NH₃- and N₂-TPD) were conducted to determine the acidity and the N₂ desorption temperature, respectively, using the chemisorption analyzer. In a typical run, 0.2 g of catalyst was placed in a U-tube quartz reactor. The catalyst was then pretreated at 750 °C (10 °C min⁻¹) for 30 min in a flow of Ar for dehydrating and outgassing. After cooling to 30 °C, the adsorbate (NH₃ or N₂) was charged into the system for 1 h. After that, the sample was purged with He at 30 °C for 1 h to remove the excess gas. TPD was performed in a helium flow from 30 to 700 °C (10 °C min⁻¹). In situ Fourier transform infrared (FTIR) spectra were recorded by using a Thermo Scientific Nicolet iS50 spectrometer equipped with a diffuse reflectance infrared Fourier transform spectroscopy cell (DRIFTS, Praying Mantis, Harrick Scientific). X-ray photoelectron spectroscopy (XPS) spectra were evaluated by using a PHI 5000 VersaProbe spectrometer equipped with a monochromatized aluminum source with a wavelength of 1486.6 eV. The tested samples were placed in a transport chamber from a glove box to the XPS chamber for quasi-in situ analysis to avoid air exposure. The C 1s binding energy of adventitious carbon at 285.0 eV was used to calibrate the energy shift.

Catalytic activity testing

Activity tests were conducted in a horizontal fixed-bed flow reactor. In a typical run, 0.18 g of tested catalyst (no diluent) was sandwiched with quartz wool in the middle of a quartz tube reactor (13 mm OD × 10 mm ID × 220 mm long). The catalyst was pretreated with an Ar stream (20 mL min⁻¹) at 750 °C (10 C min⁻¹) for 30 min to remove moisture. The catalyst was then carbonized at 750 °C using a 10% Ar/CH₄ stream (20 mL min⁻¹) for 15 min to transform MoO₃ into MoC_x. The carbonized catalysts were denoted as xMoC_x/Al₂O₃. After carbonization, 10 mL min⁻¹ of mixed gases, including 45% CH₄, 45% N₂, and 10% Ar (the internal standard), were admitted to the system and allowed to react for 6 h. The outlet products were separated and analyzed using an online gas chromatograph (GC, SRI 8610C) equipped with a flame ionization detector (FID) and a thermal conductivity detector (TCD) for quantification. A PoraPLOT Q-HT capillary column (25 m × 0.53 mm × 20 μm) was used for separation. The CH₄ conversion (%) and turnover frequency of each product (TOF, h⁻¹) were calculated based on the observable products by using the following equations:

“C atoms in product” is the carbon number of each product. F_product is the molar flow rate (μmol h⁻¹) of the product at the outlet stream. F_methane is the molar flow rate of CH₄ (μmol h⁻¹) at the inlet stream.

Results and discussion

Characterization

The bulk structure of the catalysts was determined by XRD, XAS at the Mo K-edge, and TEM, as shown in Fig. 3. The diffraction patterns of all Al₂O₃-supported MoC_x catalysts were similar to that of the pristine Al₂O₃ (Fig. 3a), indicating the presence of well-dispersed Mo₂C or Mo₂C crystals that are undetectable by XRD. Unsupported Mo₂C prepared by carbonization had similar diffraction to that of com-Mo₂C (Fig. S1†), confirming that the Mo₂C phase is successfully synthesized. Fig. 3b shows the Mo K-edge XAS spectra of the tested catalysts. Mo foil, Mo₂C, MoO₃, and 1MoO₃/Al₂O₃ were included for reference. The absorption edge shows an increasing trend as follows: Mo foil (20 003.9 eV) < Mo₂C (20 010.9 eV) < 1MoC_x/Al₂O₃ (20 012.2 eV) < 5MoC_x/Al₂O₃ (20 013.7 eV) < 10MoC_x/Al₂O₃ (20 015.5 eV) < 1MoO₃/Al₂O₃ (20 015.8 eV) = MoO₃ (20 015.8 eV), indicating the increasing oxidation state of Mo species. Moreover, the linear combination result of each XAS spectrum (see Fig. 3c) by using Mo₂C and MoO₃ as representatives showed that the composition of Mo₂C decreased following the order 1MoC_x/Al₂O₃ (73.4%) > 5MoC_x/Al₂O₃ (56.0%) > 10MoC_x/Al₂O₃ (20.0%). This fitted result suggested that the Mo₂C phase is formed after carbonization and is dominant in 1MoC_x/Al₂O₃.

Fig. 3 (a) XRD patterns of Al₂O₃ and tested catalysts, (b) XAS spectra, and (c) linear combination fitting (LCF) results of supported MoC_x catalysts. MoO₃, Mo₂C, Mo foil, and 1MoO₃/Al₂O₃ were included as the standards. Representative dark-field TEM images of (d) 1MoC_x/Al₂O₃, (g) 5MoC_x/Al₂O₃, and (j) 10MoC_x/Al₂O₃ and their EDS mapping results at Mo Lα and C Kα. (e) and (f) show the EDS mappings of 1MoC_x/Al₂O₃; (h) and (i) 5MoC_x/Al₂O₃; (k) and (l) 10MoC_x/Al₂O₃.

The dark-field TEM images (Fig. 3d, g, and j) do not reveal any distinct Mo particles, likely because of the presence of small and evenly dispersed Mo particles. The Mo Lα mapping images (Fig. 3e, h, and k) confirm that Mo is uniformly distributed. Notably, even for 10MoC_x/Al₂O₃, no aggregated Mo clusters were observed. Moreover, the C Kα images (Fig. 3f, i, and l) indicated a higher concentration of carbon species than that of Mo in each catalyst, suggesting the coexistence of deposited carbon species with carbides.

The physicochemical properties of Al₂O₃-supported MoC_x catalysts are listed in Table 1. The Mo content of each catalyst is close to its designated value. The N₂ isotherms of all catalysts showed a type IV isotherm and H1 hysteresis loop, indicating the presence of a mesoporous structure (Fig. S2†).^30,31 The surface areas of the tested catalysts varied in the range of 39 to 139 m² g⁻¹ and they exhibited a pore volume of 0.22–0.32 cm³ g⁻¹. The surface area and pore volume decreased with increasing Mo loading. This is likely due to the occupation of accessible pores by Mo species, making gas adsorption difficult.³² Hence, a higher Mo loading would lead to more available pores being occupied, resulting in a reduced gas adsorption capacity.

Table 1 Physicochemical properties of Al₂O₃-supported Mo₂C catalysts

Catalyst	Mo loading^a (%)	Surface area^b (m^2 g^−1)	Pore volume^c (cm^3 g^−1)	Dispersion^d (%)	Acidity^e (μmol g^−1)
a Determined by ICP-MS. b Estimated by the BET method. c Calculated by the BJH method. d Obtained by CO pulse titration. e Evaluated by the NH3-TPD technique.
1MoCx/Al2O3	1.01	139	0.32	17.8	469
5MoCx/Al2O3	4.74	95	0.24	8.5	393
10MoCx/Al2O3	9.99	39	0.22	2.1	172

---

The dispersion of MoC_x decreased with increasing Mo loading. As seen from the TEM images, the Mo₂C particle size was increased with increasing Mo loading. The increased particle size indicated a higher degree of Mo₂C agglomeration, which reduced the surface-to-volume ratio of Mo₂C. Therefore, the surface availability of Mo₂C at a high Mo content is suppressed. The total acidity of the Al₂O₃-supported Mo₂C catalysts also decreased when the Mo loading was increased (Fig. S3†). Al₂O₃ exhibits Lewis acidity, whereas Mo₂C has Lewis basic sites.³³ By adjusting the Mo loading, it becomes possible to occupy the acidic sites of Al₂O₃, leading to a decrease in acidity. The coverage of acid sites increases with higher Mo loading.

Catalytic evaluation

CH₄ conversion was tested in either a N₂ or Ar stream to verify that gaseous N₂ is activated and reacted with CH₄ to form ACN. Similar profiles of CH₄ conversion and TOF of hydrocarbons (ethylene, ethane, benzene, and toluene) were observed in both N₂ (closed squares) and Ar (open circles) streams (Fig. 4a and S4†). This indicated that the N₂ and Ar atmospheres had little influence on the transformation of CH₄ into hydrocarbons. Moreover, it can be seen that the CH₄ conversion and TOF of hydrocarbons decreased synchronously with time, indicating that the active sites for hydrocarbon formation are deactivated, likely due to coke accumulation.³⁴

Fig. 4 Time on stream (TOS) profiles of (a) CH₄ conversion (black) and TOF of ACN (red) under N₂ (closed squares) and Ar (open circles) atmospheres and the TOS profiles of CH₄ conversion and TOF of ACN during the (b) N₂–Ar–N₂ and (c) Ar–N₂–Ar switching tests. Reaction conditions: 0.18 g of 1MoC_x/Al₂O₃, reaction temperature = 750 °C, GHSV = 1500 mL_CH4 g_cat⁻¹ h⁻¹, feed = 4.5/4.5/1 mL min⁻¹ of CH₄/N₂/Ar or 4.5/5.5 mL min⁻¹ of CH₄/Ar.

Note that ACN production was solely observed when CH₄ was converted in N₂ (Fig. 4a). That is, N₂ activation and N-insertion took place in CH₄ conversion. Since gaseous N₂ is the sole N-containing reagent, the formation of ACN could be attributed to gaseous N₂ activation. To further clarify that ACN can be produced in the presence of N₂, N₂–Ar–N₂ (Fig. 4b) and Ar–N₂–Ar (Fig. 4c) switching tests were conducted. Again, operating in varying streams has a negligible impact on the CH₄ conversion (black scatters in Fig. 4b and c). In the N₂–Ar–N₂ switching test (Fig. 4b), the TOF of ACN was initially high (13.0 h⁻¹) when operating under N₂ (red closed squares). After switching to Ar (red open circles), the TOF of ACN decreased sharply, indicating that ACN cannot be produced in Ar. When the reaction stream was switched back to N₂, the TOF recovered to 8.2 h⁻¹ (red closed squares), underlining that ACN is produced by feeding N₂. Comparatively, in the Ar–N₂–Ar switching test (Fig. 4c), there was no ACN produced at the beginning of the test when feeding Ar (red open circles). After switching the feed to N₂, a significant increase in the TOF of ACN was observed (6.1 h⁻¹, red closed squares), emphasizing that N₂ activation occurs, which can react with CH₄ to form ACN. However, ACN could still be produced when N₂ was replaced by Ar (red open circles). Presumably, there is deposited carbon on the MoC_x surface during the first cycle in Ar. After purging N₂, the converted N₂ could also react with the MoC_x surface, forming a carbonitride-like structure.³⁵ This structure is proposed to act as a N-source for ACN synthesis. It can be concluded at this stage that Al₂O₃-supported MoC_x is active for the co-activation of CH₄ and N₂, and ACN can be produced merely in N₂.

The effects of operating temperature (650, 700, 750, and 800 °C) and Mo loading (1, 5, and 10 wt%) were evaluated for the conversion of CH₄ using the MoC_x/Al₂O₃ catalysts (Table 2, entries 1–4 and Fig. S5†). Mo₂C and MoO₃ were tested, and both showed negligible activities. The CH₄ conversion is increased with temperature from 9.3% (650 °C) to 25.0% (800 °C). The TOFs of C₂ species and coke also increased with reaction temperature. However, aromatics reached the highest TOF at 750 °C and then declined. This is likely caused by the over-oligomerization of aromatics to form coke. The TOF of ACN also increased with reaction temperature, suggesting the progressively enhanced extent of C–H bond and N≡N bond cleavages. At 800 °C, the TOF of ACN was high at the beginning of the reaction, but showed a rapidly decreasing trend. Hence, a temperature of 750 °C was selected for further investigations.

Table 2 Catalytic activity testing for CH₄ conversion over the Al₂O₃-supported MoC_x catalysts. Reaction conditions: 0.18 g catalyst, 10 mL min⁻¹ of the feed mixture

Entry	Catalyst	Temp. (°C)	CH4 conversion^a (%)	TOF^a (h^−1)
				C2	Aromatics	ACN	HCN	Coke^b
a Data were taken from the TOS profiles at 40 min and 6 h (shown in brackets). b Average coke formation rate, determined by TGA.
1	1MoCx/Al2O3	650	9.3 (5.5)	8.6 (5.1)	65.0 (23.3)	1.1 (6.3)	N/D	23.2
2		700	17.2 (6.0)	16.6 (9.6)	123.7 (14.7)	5.1 (9.5)	N/D	26.7
3		750	26.1 (6.0)	30.7 (8.4)	168.8 (1.0)	15.3 (14.3)	N/D	33.6
4		800	25.0 (8.4)	38.6 (11.9)	109.5 (0.5)	20.8 (5.5)	N/D	58.2
5	5MoCx/Al2O3	750	26.5 (5.9)	13.5 (2.3)	72.2 (0.1)	7.8 (4.6)	1.2 (0)	18.1
6	10MoCx/Al2O3		15.4 (3.0)	15.2 (3.1)	80.3 (0.0)	13.2 (5.9)	2.1 (0)	15.3

---

In addition to determining the effect of reaction temperature, CH₄-TPSR experiments were also conducted, as shown in Fig. S6.† CH₄ (m/z = 16) started to be converted at 640 °C with a maximum conversion rate at 730 °C. The onset temperatures of major products, including ACN (m/z = 41), C₂ species (m/z = 29), and benzene (m/z = 78), were approximately 650 °C. Therefore, it could be suggested that the formation reactions of ACN and hydrocarbons proceed in parallel.

The effect of Mo loading (Table 2, entries 3, 5, and 6 and Fig. S7†) was tested at a 750 °C. The CH₄ conversion and TOF of C₂ and aromatics decreased with increasing Mo loading. This is likely due to the suppressed acidity at a high Mo loading since C–H activation is promoted by acidic sites.^36,37 The TOF of ACN had negligible change with respect to the Mo loading. However, a small amount of HCN could be observed when the Mo loading is higher than 5%. Moreover, 1MoC_x/Al₂O₃ exhibited the most stable TOF of ACN during the 6 h testing. Accordingly, 1MoC_x/Al₂O₃ was further investigated.

In comparison, the equilibrium conversion of CH₄ and N₂ to ACN (2CH₄ + 0.5N₂ ↔ CH₃CN + 2.5H₂) was performed from 300 °C to 1000 °C using ThermoSolver. The obtained parameters are presented in Table S1.† It can be seen that the equilibrium constant (K_eq) increases with increasing temperature, indicating the endothermic nature of CH₄ conversion to ACN. However, the values of K_eq at all tested temperatures are very small. That is, the reaction is highly unpreferable, in which low CH₄ and N₂ conversions and a low ACN yield were expected. The equilibrium CH₄ conversion as a function of temperature is illustrated in Fig. S8.† At 750 °C, only 0.55% of CH₄ is converted in the equilibrium state. Note that using a MoC_x-based catalyst resulted in approximately 1.5% CH₄ conversion, suggesting the effectiveness of Mo₂C in the cleavage of C–H and N≡N bonds.

The above activity results showed that the major products were aromatics and coke, possibly due to the strong interaction between Mo₂C nanoparticles and CH₄.³⁸ Moreover, a parallel reaction pathway to produce ACN and hydrocarbons was suggested by the similar onset temperature from CH₄-TPSR. One way to suppress aromatization and coke is through co-feeding hydrogen in the reactant stream.³⁹ Hence, the effect of co-feeding H₂ (10, 20, 25, and 33% v/v) was investigated at 750 °C by using 1MoC_x/Al₂O₃ (see Fig. 5 and S9†). Co-feeding H₂ varies the concentration of CH₄ which can affect reactivity. Accordingly, using Ar as a diluent with the same concentration as that of H₂ was also investigated to validate that the activity is not only affected by the low CH₄ concentration.

Fig. 5 Effects of H₂ co-feeding and Ar dilution on (a) CH₄ conversion and TOF of (b) ACN, (c) C₂, and (d) aromatics by using 1MoC_x/Al₂O₃.

The initial CH₄ conversion and TOFs of hydrocarbons (C₂ and aromatics) and ACN decreased with increasing concentration of diluent. In addition, the initial CH₄ conversion and TOFs under the H₂ stream (magenta bars) were more strongly influenced than those tested in the Ar stream (yellow bars), in which aromatics are the most affected species (see Fig. 5d). In other words, H₂ co-feeding does not only dilute the concentration of CH₄ but also dampens the aromatization activity. Moreover, after 12 h on stream, the CH₄ conversion decreased severely in the Ar stream (22.5%, 19.6%, 18.1%, and 13.7% declined to 0.6%, 0.8%, 0.5%, and 0.4% at 10, 20, 25, and 33% Ar concentration, respectively), compared to those in the H₂ co-fed stream (20.0%, 10.1%, 5.6%, and 1.2% decreased to 4.6%, 4.1%, 2.5%, and 1.0% at 10, 20, 25, and 33% H₂ concentration, respectively). This implied that the reactivity is more stable when H₂ is present.

As seen by the TOFs at 12 h on stream (Fig. 5b–d), the products were still producible in the H₂ atmosphere, while there were almost no products formed under the Ar stream at the 12th hour. Accordingly, it could be claimed that gaseous H₂ also reduces the amount of carbon deposits on the catalyst surface, resulting in a steadier reactivity.³⁹

XPS spectra of the Mo₂C and 1MoC_x/Al₂O₃ catalysts are illustrated in Fig. 6. The photoline of Mo₂C at Mo 3d (Fig. 6a) could be deconvoluted into five responses. The responses at 235.8 and 232.7 eV are attributed to Mo⁶⁺ (red);⁴⁰ 229.5 eV, Mo³⁺ (magenta);⁴¹ 231.3 and 228.4 eV, Mo²⁺ (blue).⁴² The presence of Mo²⁺ signals indicated the existence of Mo₂C species, while Mo⁶⁺ and Mo³⁺ could be assigned as MoO₃ and Mo₂O₃, respectively. The existence of oxidic species could be attributed to the oxidation of the Mo₂C surface.^43,44 The spectrum of 1MoC_x/Al₂O₃ also contained these signals. In addition, another response at 234.4 eV assigned to Mo⁴⁺ species⁴⁵ (cyan) was observed over the 1MoC_x/Al₂O₃ catalyst, which could be referred to the presence of MoC species, likely due to the over-carbonization of small Mo₂C particles. Oxidic molybdenum species were observed, even though the catalyst was in situ carbonized prior to the reaction without exposure to the atmosphere. Therefore, it could be stated that the active phase is composed of carbidic molybdenum species.

Fig. 6 XPS spectra of fresh and spent MoC_x catalysts for (a) Mo 3d and (b) C 1s.

The C 1s spectra (Fig. 6b) of both Mo₂C and 1MoC_x/Al₂O₃ could be deconvoluted into four species, including adventitious carbon (C–C, red) at 285.0 eV, C–O (cyan) at 285.6 eV,⁴⁶ C=O (magenta) at 288.0 eV,⁴⁷ and C–Mo (blue) at 283.7 eV,⁴⁸ respectively. According to the characterization results, it can be concluded that the active sites of the catalysts are predominantly made of the carbide phase.

Table S2† shows the percentage of each Mo and C species obtained from the aforementioned XPS results. 1MoC_x/Al₂O₃ contained 28.1% Mo₂C (Mo²⁺) phase. Additionally, by including all carbide species (MoC, Mo⁴⁺), the relative composition of MoC_x for 1MoC_x/Al₂O₃ is 43.7%, which is much higher than that of its unsupported counterpart (7.4%). This is presumably due to the presence of smaller molybdenum particles that lead to a more rapid and complete carbonization. After the reaction, the C–Mo signals were not observed. The percentage of C–C increased together with the appearance of sharper C 1s signals over the spent catalysts. This suggested the formation of graphitic carbon that dominates the signals at a similar region.⁴⁹

It can be seen in Table S2† that the surface C/Mo ratios of fresh, 6 h spent (Ar), 6 h spent (H₂), 12 h spent (Ar), and 12 h spent (H₂) 1MoC_x/Al₂O₃ are 1.04, 5.66, 5.56, 11.23, and 6.78, respectively. It has been reported that the reactivity of molybdenum carbides depends on the C/Mo ratio, in which the catalyst with a lower C/Mo ratio is more reactive.³⁸ In this case, the C/Mo ratio rapidly increased when the reaction was performed in the Ar atmosphere (from 1.04 (fresh) to 5.66 (6 h, Ar) to 11.23 (12 h, Ar)), while it was more stable in the H₂ stream (from 1.04 (fresh) to 5.56 (6 h, H₂) to 6.78 (12 h, H₂)). Accordingly, the enhanced stability of CH₄ conversion and TOF is likely related to the more consistent surface C/Mo ratio achieved when H₂ is present.

The TOF of ACN and C₂ was more stable when co-feeding H₂. Additionally, it can be seen that the TOF of aromatics is suppressed by ∼87% at 25% H₂. When the concentration of H₂ is further increased to 33%, no aromatics can be initially found. Note that the CH₄ conversion under these conditions is almost negligible, which explains the disappearance of aromatics. Therefore, 25% H₂ was selected for 50 h stability testing.

Fig. 7 shows that the on-stream activity was nearly unchanged in the 50 h duration test, underlining the stability of the reaction process. The CH₄ conversion was maintained at approximately 2.5%. Moreover, three major products, including ACN, ethylene, and ethane, were also sustainably produced at a TOF of approximately 6.5 h⁻¹. Aromatics and HCN were observed as by-products at a TOF of ∼2.5 h⁻¹. Therefore, the presence of 25% H₂ can not only suppress CH₄ aromatization but also enhance the stability of the catalyst during the 50 h durability test.

Fig. 7 50 h stability testing at 25% H₂ co-feeding over 1MoC_x/Al₂O₃.

Proposed mechanism

Our earlier work claimed that ACN is formed via surface CN species over GaN catalysts. The surface CN species are formed by reacting CH₄ with mobile-N of GaN at the surface.⁵⁰ Therefore, CN species may also form over the MoC_x surface by N₂ cleavage. Hence, in situ DRIFTS experiments were performed under a N₂ atmosphere, as shown in Fig. S10.† Regrettably, no CN vibrational band (∼2220 cm⁻¹) could be observed. This might be attributed to (i) the high instability of evolved intermediates or (ii) no interaction being formed between N₂ and surface carbon of MoC_x. The latter could be clarified by performing N₂–H₂ co-feeding experiments. In the N₂–H₂ co-feeding test, HCN should be generated if N₂ and surface carbon react since the surface carbon of MoC_x acts similarly to surface oxygen in the Mars–van Krevelen mechanism.^51,52 However, no HCN could be generated (not shown), indicating that the surface carbon has no interaction with gaseous N₂. Nevertheless, the aforementioned results showed that ACN could be generated in which the only N-source is N₂. That is, the N≡N triple bond of N₂ is activated. We conducted N₂-TPD to further confirm that N₂ can be adsorbed on the Al₂O₃-supported MoC_x. As shown in Fig. 8, the desorption peak of N₂ over the Al₂O₃-supported MoC_x catalysts is located at approximately 107 °C, while no N₂ desorption can be observed over bare Al₂O₃. This implies that N₂ should be firstly adsorbed on the MoC_x surface and then desorbed. Considering no interaction between N₂ and surface carbon (C sites) of MoC_x, it can be inferred that N₂ adsorption occurs on metal (Mo) sites.^53–56

Fig. 8 N₂-TPD profiles of Al₂O₃-supported MoC_x catalysts and unsupported Al₂O₃.

To clarify, the CO and N₂ uptakes were compared since CO is stoichiometrically adsorbed on Mo sites.⁵⁷ As shown in Table S3,† the CO and N₂ uptakes were nearly identical for the tested catalysts. This indicated that the adsorption behaviors of CO and N₂ are similar. Accordingly, it can be concluded that the adsorption of N₂ takes place at the Mo sites.

Although N₂ activation by MoC_x was elucidated, the formation of the C≡N bond and ACN is still unclear. A previous study revealed that CH₄ tends to be completely dissociated on Mo sites exposed on the MoC_x surface.⁵⁸ The MoC_x surface could be further carburized by the dissociated carbon adatoms diffusing into the subsurface layer, forming a MoC-like surface structure. The MoC-like surface could facilitate CH* coupling to form as a potential intermediate.^58,59 The 1MoC_x/Al₂O₃ had its C/Mo ratio estimated by XPS analysis close to unity (Table S2†), similar to that of MoC. That is, the surface of 1MoC_x/Al₂O₃ should have an MoC-like structure. Thus, we propose that the formation of ACN occurs over the Mo sites exposed on the surface of the MoC-like structure, as shown in Fig. 9.

Fig. 9 Proposed co-activation of CH₄ and N₂ to ACN over the MoC_x catalyst (top view of the catalyst surface).

N₂ could be dissociatively adsorbed on the Mo-terminated surface. Then, CH₄ would be activated and dehydrogenated over the adjacent Mo sites, forming CH* species. The CH* species could couple with species that further react with N adatoms, forming ACN. The produced ACN could desorb from the surface, which fulfills the catalytic cycle.

To support this claim, kinetic analysis based on the Langmuir–Hinshelwood (LH) formalism of the proposed mechanism was performed, which is presented in the ESI.† The six possible reactions were grouped into four steps, including N₂ adsorption, CH₄ adsorption, surface reaction, and ACN desorption. The rate expression is derived by assuming that one of the above steps is rate-limiting while the remaining steps are all in equilibrium. The rate expression was fitted to the experimental data (see Table S4†) by nonlinear regression subject to minimizing the residual sum of squares to recover the values of the parameters (see Table S5†). By resorting to the fitted parameters and the parity plot (Fig. S11†), the surface reaction step is the most likely to occur. This result is supported by previous studies showing that the dissociative adsorption of N₂ and the C–H bond cleavage of CH₄ on the MoC surface are facile and unlikely to be rate-limiting.^59,60

Conclusions

In this article, two high-stability compounds (CH₄ and N₂) were simultaneously converted to form ACN over Al₂O₃-supported MoC_x catalysts. ACN could only be produced in the presence of N₂, confirming the activation of the N≡N bond. Due to the strong interaction between MoC_x and CH₄, aromatization and coking were significant, resulting in catalyst deactivation. The aromatization could be diminished by co-feeding H₂, in which the catalytic activity and TOF could be maintained for 50 h. The MoC-like surface structure of MoC_x was proposed to be an active center for ACN synthesis. Kinetic analysis suggested that the surface reaction of adsorbed N₂ and species is possibly the rate-determining step.

Author contributions

Korawich Trangwachirachai: investigation, formal analysis, validation, and writing – original draft. I-Ting Kao: investigation and formal analysis. Wei-Hsiang Huang: XAS data collection and analysis. Chi-Liang Chen: XAS data collection and analysis. Yu-Chuan Lin: supervision, project administration, resources, and writing – review & editing.

Conflicts of interest

The authors declare that they have no competing financial interests.

Acknowledgements

This work was supported by the National Science and Technology Council (projects 109-2628-E-006-011-MY3, 110-2221-E-006-165-MY3, 110-2923-E-006-005-MY3, and 110-2927-I-006-506) and by the Higher Education Sprout Project, Ministry of Education to the Headquarters of University Advancement at National Cheng Kung University (NCKU). The authors gratefully acknowledge the use of XPS (ESCA003700) and TEM (EM000800) of NSTC 112-2740-M-006-001 belonging to the Core Facility Center of NCKU.

References

1. W. Jiang, J. Low, K. Mao, D. Duan, S. Chen, W. Liu, C. W. Pao, J. Ma, S. Sang, C. Shu, X. Zhan, Z. Qi, H. Zhang, Z. Liu, X. Wu, R. Long, L. Song and Y. Xiong, J. Am. Chem. Soc., 2021, 143, 269–278.
2. P. Schwach, X. Pan and X. Bao, Chem. Rev., 2017, 117, 8497–8520.
3. Z. Zhu, W. Guo, Y. Zhang, C. Pan, J. Xu, Y. Zhu and Y. Lou, Carbon Energy, 2021, 3, 519–540.
4. S. M. Kim, P. M. Abdala, D. Hosseini, A. Armutlulu, T. Margossian, C. Copéret and C. Müller, Catal. Sci. Technol., 2019, 9, 5745–5756.
5. M. A. A. Aziz, A. A. Jalil, S. Wongsakulphasatch and D.-V. N. Vo, Catal. Sci. Technol., 2020, 10, 35–45.
6. J. Niu, Y. Wang, Y. Qi, A. H. Dam, H. Wang, Y.-A. Zhu, A. Holmen, J. Ran and D. Chen, Fuel, 2020, 266, 117143.
7. B. C. Ekeoma, M. Yusuf, K. Johari and B. Abdullah, Int. J. Hydrogen Energy, 2022, 47, 41596–41620.
8. J. K. Dahl, A. W. Weimer, A. Lewandowski, C. Bingham, F. Bruetsch and A. Steinfeld, Ind. Eng. Chem. Res., 2004, 43, 5489–5495.
9. Y.-R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC Press, 2007, p. 1688,  DOI:10.1201/9781420007282.
10. W. Taifan and J. Baltrusaitis, Appl. Catal., B, 2016, 198, 525–547.
11. L. Sun, Y. Wang, N. Guan and L. Li, Energy Technol., 2019, 8, 1900826.
12. D. C. Upham, V. Agarwal, A. Khechfe, Z. R. Snodgrass, M. J. Gordon, H. Metiu and E. W. McFarland, Science, 2017, 358, 917–921.
13. K. Dutta, V. Chaudhari, C.-J. Li and J. Kopyscinski, Appl. Catal., A, 2020, 595, 117430.
14. K. Trangwachirachai, C.-H. Chen and Y.-C. Lin, Mol. Catal., 2021, 516, 111961.
15. K. Trangwachirachai, C.-H. Chen, A.-L. Huang, J.-F. Lee, C.-L. Chen and Y.-C. Lin, Catal. Sci. Technol., 2022, 12, 320–331.
16. Z. Yuan, X. Zhang, Q. Yao, Y. Zhang and Y. Fu, J. Anal. Appl. Pyrolysis, 2019, 140, 376–384.
17. Future Market Insights, Acetonitrile Market, https://www.futuremarketinsights.com/reports/acetonitrile-market, 2023.
18. A. Sieminski, US Energy Information Administration, 2015.
19. J. J. Sattler, J. Ruiz-Martinez, E. Santillan-Jimenez and B. M. Weckhuysen, Chem. Rev., 2014, 114, 10613–10653.
20. K. Trangwachirachai and Y.-C. Lin, Dalton Trans., 2023, 52, 6211–6225.
21. A. Daisley and J. S. J. Hargreaves, Catal. Today, 2023, 423, 113874.
22. S. Tiwari, T. S. Khan, P. Tavadze and J. Hu, Chem. Eng. J., 2021, 413, 127501.
23. R. Kojima and K.-I. Aika, Appl. Catal., A, 2001, 219, 141–147.
24. R. Michalsky, Y.-J. Zhang, A. J. Medford and A. A. Peterson, J. Phys. Chem. C, 2014, 118, 13026–13034.
25. H. Cheng, L.-X. Ding, G.-F. Chen, L. Zhang, J. Xue and H. Wang, Adv. Mater., 2018, 30, 1803694.
26. J. Gao, Y. Zheng, G. B. Fitzgerald, J. de Joannis, Y. Tang, I. E. Wachs and S. G. Podkolzin, J. Phys. Chem. C, 2014, 118, 4670–4679.
27. J. Jeong, A. Hwang, Y. T. Kim, D.-Y. Hong and M.-J. Park, Catal. Today, 2020, 352, 140–147.
28. B. Cook, D. Mousko, W. Hoelderich and R. Zennaro, Appl. Catal., A, 2009, 365, 34–41.
29. B. Ravel and M. Newville, J. Synchrotron Radiat., 2005, 12, 537–541.
30. S. Paganelli, R. Tassini, V. D. Rathod, B. Onida, S. Fiorilli and O. Piccolo, Catal. Lett., 2021, 151, 1508–1521.
31. K. S. W. Sing and R. T. Williams, Adsorpt. Sci. Technol., 2004, 22, 773–782.
32. M. J. Ramos, A. Casas, L. Rodríguez, R. Romero and Á. Pérez, Appl. Catal., A, 2008, 346, 79–85.
33. S. K. Bej, C. A. Bennett and L. T. Thompson, Appl. Catal., A, 2003, 250, 197–208.
34. C. H. L. Tempelman and E. J. M. Hensen, Appl. Catal., B, 2015, 176–177, 731–739.
35. P. K. Roy and S. Kumar, ACS Appl. Energy Mater., 2020, 3, 7167–7179.
36. R. Gounder and E. Iglesia, J. Am. Chem. Soc., 2009, 131, 1958–1971.
37. M. C. Cholewinski, M. Dixit and G. Mpourmpakis, ACS Omega, 2018, 3, 18242–18250.
38. M. Figueras, R. A. Gutiérrez, H. Prats, F. Viñes, P. J. Ramírez, F. Illas and J. A. Rodriguez, Phys. Chem. Chem. Phys., 2020, 22, 7110–7118.
39. H. Ma, R. Kojima, S. Kikuchi and M. Ichikawa, Catal. Lett., 2005, 104, 63–66.
40. C. Wu and J. Li, ACS Appl. Mater. Interfaces, 2017, 9, 41314–41322.
41. Y. N. Regmi, C. Wan, K. D. Duffee and B. M. Leonard, ChemCatChem, 2015, 7, 3911–3915.
42. X. Kong, S. Chen, Y. Zou, S. Lyu, X. She, Y. Lu, J. Sun, H. Zhang and D. Yang, Int. J. Hydrogen Energy, 2018, 43, 13720–13726.
43. H. Vrubel and X. Hu, Angew. Chem., Int. Ed., 2012, 51, 12703–12706.
44. J. Li, C. Zhou, J. Mu, E.-C. Yang and X.-J. Zhao, RSC Adv., 2018, 8, 17202–17208.
45. H. Wei, Q. Xi, X. A. Chen, D. Guo, F. Ding, Z. Yang, S. Wang, J. Li and S. Huang, Adv. Sci., 2018, 5, 1700733.
46. L. Diao, J. Qin, N. Zhao, C. Shi, E. Liu, F. He, L. Ma, J. Li and C. He, J. Mater. Chem. A, 2018, 6, 6054–6064.
47. Y. Lei, Y. Yang, Y. Liu, Y. Zhu, M. Jia, Y. Zhang, K. Zhang, A. Yu, J. Liu and J. Zhai, Nanoscale Res. Lett., 2019, 14, 329.
48. D. Wang, J. Wang, X. Luo, Z. Wu and L. Ye, ACS Sustainable Chem. Eng., 2018, 6, 983–990.
49. P. Xiao, X. Ge, H. Wang, Z. Liu, A. Fisher and X. Wang, Adv. Funct. Mater., 2015, 25, 1520–1526.
50. K. Trangwachirachai, A. L. Huang, H. K. Chen, C. L. Chen, J. F. Lee, H. K. Tian and Y. C. Lin, Mater. Today Chem., 2023, 30, 101500.
51. I. Vollmer, B. van der Linden, S. Ould-Chikh, A. Aguilar-Tapia, I. Yarulina, E. Abou-Hamad, Y. G. Sneider, A. I. O. Suarez, J.-L. Hazemann, F. Kapteijn and J. Gascon, Chem. Sci., 2018, 9, 4801–4807.
52. G. Li, I. Vollmer, C. Liu, J. Gascon and E. A. Pidko, ACS Catal., 2019, 9, 8731–8737.
53. M. Grunze, M. Golze, W. Hirschwald, H. J. Freund, H. Pulm, U. Seip, M. C. Tsai, G. Ertl and J. Küppers, Phys. Rev. Lett., 1984, 53, 850–853.
54. S. Dahl, A. Logadottir, R. C. Egeberg, J. H. Larsen, I. Chorkendorff, E. Törnqvist and J. K. Nørskov, Phys. Rev. Lett., 1999, 83, 1814–1817.
55. J. J. Mortensen, L. B. Hansen, B. Hammer and J. K. Nørskov, J. Catal., 1999, 182, 479–488.
56. W. Liao, L. Qi, Y. Wang, J. Qin, G. Liu, S. Liang, H. He and L. Jiang, Adv. Funct. Mater., 2021, 31, 2009151.
57. J. S. Lee, K. H. Lee and J. Y. Lee, J. Phys. Chem., 1992, 96, 362–366.
58. T. Zhang, X. Yang and Q. Ge, Catal. Today, 2021, 368, 140–147.
59. T. Zhang, X. Yang and Q. Ge, Catal. Today, 2020, 339, 54–61.
60. I. Matanovic and F. H. Garzon, Phys. Chem. Chem. Phys., 2018, 20, 14679–14687.

---

Footnote

† Electronic supplementary information (ESI) available: Time on stream (TOS) profiles and N₂ adsorption/desorption isotherms. See DOI: https://doi.org/10.1039/d3cy00585b

---