Chapter 1: Introduction to the Application of Nitrides, Carbides, Phosphides and Amorphous Boron Alloys in Catalysis

Catalysts are used to enhance the rates of thermocatalytic, electrocatalytic and photocatalytic reactions.¹  Catalysts function by interacting with reacting species to generate new reaction intermediates that only exist on the catalyst surface, thereby providing alternative, faster reaction pathways to the desired products. Key to the catalytic reaction is the breaking and forming of new bonds between the catalyst and the reacting molecules. Consequently, the efficacy of a particular catalyst is determined by its surface chemistry and new developments in catalysis science and technology are driven in part by the development of new materials with well controlled surface properties. Increasing attention has been given to the discovery of heterogeneous catalysts that have high thermal and chemical stability, using materials that are not strategically limited and are of low cost.²  Catalysts are used in a wide range of applications including motor vehicle emissions control, upgrading and refining of crude oils, and they are responsible for approximately 90% of the chemical processes currently operating world-wide in the chemical industry.^1,3 

Controlling the reactivity of a catalyst surface plays a key role in obtaining catalysts with high activity and selectivity. Sabatier's principle indicates that the interactions between reacting species and a catalyst surface must be of the appropriate strength—interactions that are too strong or too weak yield less active catalysts.¹  Hence, in the classic paper of Levy and Boudart,⁴  C added to W to produce W₂C, was shown to decrease the high reactivity of W toward O and thereby yield an effective catalyst for the H₂+O₂→H₂O reaction at room temperature, a reaction typically catalysed by Pt. Similarly, the W₂C was active for the isomerisation of 2,2-dimethylpropane to 2-methylbutane, a reaction also catalysed by Pt.⁴  This work spawned a large number of studies in which new materials with metallic character, formed by the incorporation of C, N, P and B into the lattices of early transition metals, have been investigated as potential catalysts for various reactions.^5–7  These interstitial alloys adopt simple crystal structures with the carbides and nitrides forming face-centered cubic (fcc), hexagonal close packed (hcp) or simple hexagonal (hex) structures, as summarised in Table 1.1. Many metal phosphides are also known⁷  and their crystal structures as shown in Figure 1.1. The very rich chemistry of these interstitial alloys provides an opportunity for the development of new catalysts with a wide potential for application, especially with the synthesis of these materials as well dispersed nanoparticles.⁵ 

The unique catalytic behaviour of metals bound to C, N, P or S may be attributed to the changes in the electronic properties of the metal surface induced by the ligands and/or the geometry by which the metal and the ligands are arranged at the catalyst surface. Hence, the ligands passivate the metal surface reactivity and these effects have been quantified in some cases using molecular simulation.¹⁰  For example, Liu and Rodriguez¹¹  have shown, by DFT calculation, that the CO adsorption and S adsorption energies on the (001) surfaces of MoN, MoC, MoP and Mo decrease with an increase in their d-band centre energy, as shown in Figure 1.2.

These observations suggest the possibility that metal surface reactivity can be controlled and tuned using interstitial atoms.¹²  Consequently, there exist several detailed reviews describing the incorporation of C,^5,13  N,^6,14  P^7,9 and B¹⁵ into metals and the use of the resulting interstitial alloys for catalysis, with varying degrees of success. Since, in nearly all cases, the interstitial atoms are incorporated into metals, the application of these new materials is mostly focused on reactions catalysed by metals. In some cases, however, the catalysts may also have acidic or basic properties, broadening their potential application.^16–18  In this chapter, the application of these materials is introduced with a view to demonstrating the wide scope of their application in heterogeneous catalysis. The focus is on their use as thermochemical catalysts and electrocatalysts, although they have also been used as catalyst supports. Selected examples are described to illustrate the application. Comprehensive reviews of these materials and their applications are presented in later chapters.

Hydrogenation reactions that use expensive metal catalysts are an essential part of chemical synthesis and product upgrading processes.³  Metal carbides,^19–22  nitrides^23–25  and especially borides^26–28  are also effective hydrogenation catalysts for a wide range of applications and these materials have the potential to replace some metal catalysts.

The hydrogenation of aromatic compounds is possible on metal carbide, nitride, phosphide and boride catalysts. Mo₂C has been shown to hydrogenate naphthalene to tetralin with high selectivity (>90%) at 340 °C and 4 MPa H₂, although catalyst activity declined from 90% to 30% over a period of about 60 h.²⁹  When supported on HY zeolite, the Mo₂C also shows some ring-opening activity with about 10% selectivity to ring-opened products and 80% tetralin selectivity at 300 °C and 3 MPa H₂.³⁰  However, the Mo₂C/HY catalyst was less active than a commercial Pd/HY catalyst operated at the same conditions. The bimetallic nitride Ni₂Mo₃N (unsupported) also has high catalytic activity for benzene hydrogenation²⁵  and promotion with K reduces the deactivation of the catalyst in the presence of S.³¹  Amorphous RuB supported on ZrO₂ hydrogenates benzene to cyclohexene with high yields (46% maximum at 433 K, 5 MPa and 55 min reaction time and liquid phase).³² 

The hydrogenation of toluene on a low surface area (12 m² g⁻¹) unsupported Mo₂C has also been reported at 423–598 K and 2.76 MPa H₂ pressure. Within the temperature range 473–523 K, the yield of methylcyclohexane was 100%, but at higher temperature, ethylcylopentane was also produced as a result of isomerisation reactions.³³  The dehydrogenation of decalin has also been shown to occur on WC,³⁴  while the dehydrogenation performance of a Ni–WC/AC catalyst was better than that of Ni/AC and WC/AC at 101 kPa and 400 °C. Among a series Ni–WC/AC catalysts, decalin dehydrogenation to naphthalene was highest on a 30 wt%Ni–20 wt%WC/AC catalyst. The Ni–WC/AC catalyst showed good stability for decalin dehydrogenation.³⁴  The hydrogenation of olefins and dienes has also been reported on amorphous metal borides,^35–37  nitrides^23,38  and carbides.²³ 

In the past decade, the transformation of residual biomass into fuels and chemicals has been the focus of much research. Hydrogenations of intermediate molecules is an important step in these processes and the Pt-group metals (PGMs) are often the catalyst of choice for the hydrogenation reactions. For example, levulinic acid (LA) has been identified as an important molecule for application in future biorefineries that can be produced from lignocellulosic wastes at low cost. LA can be hydrogenated to γ-valerolactone (GVL), which also has many potential applications. The heterogeneous catalytic hydrogenation of LA to GVL is therefore of interest and Ru supported on activated carbon is usually the catalyst of choice. Recently, Quiroz et al.³⁹  reported that nanostructured β-Mo₂C has a LA to GVL turnover frequency (TOF) of 2.3 min⁻¹, measured at 30 bar H₂ and 180 °C, the same order of magnitude as that reported for Ru/TiO₂ (9.8 min⁻¹)⁴⁰  and a Ni–MoO_x/C (TOF of 3.4 min⁻¹), the latter performed under more severe conditions (50 bar H₂ and 240 °C).⁴¹  The authors also reported that the nanostructured β-Mo₂C was stable under continuous operation. Amorphous metal-boride alloy catalysts have also attracted attention for hydrogenation of bio-based feedstocks. The hydrogenation of furfural, crotonaldehyde and citral on CuB/SiO₂ and Cr–CuB/SiO₂ catalysts, has been carried out with 5 mL of substrate in 75 mL ethanol at 393 K for furfural and crotonaldehyde, and at 373 K for citral, under a constant pressure of 929 kPa.⁴²  The CuB/SiO₂ catalyst was more active than a Cu/SiO₂ catalyst. Citral hydrogenation on a 5%NiB/SiO₂ catalyst has a high yield (∼84%) at 30 °C.⁴³  In the hydrogenation of fructose and fructose/glucose mixtures, bimetallic amorphous CoNiB catalysts operated at 343 K and 6 MPa were more active than NiB, and much more active than CoB and Raney Ni.⁴⁴  In a separate study of glucose hydrogenation over an amorphous Ni–B/SiO₂ catalyst at 373 K and 4.0 MPa, the catalyst had much higher activity than other Ni-based catalysts, including crystalline Ni–B/SiO₂, Ni/SiO₂ and a commercial Raney Ni catalyst.^45,46  The hydrogenation of ethyl lactate derived from fermentation of renewable resources such as carbohydrates, to yield a green propanediol can also be achieved on a RuB/Sn-SBA-15 catalyst operated at 423 K and 5.5 MPa H₂.⁴⁷ 

The carbides, nitrides and borides have also been shown to have high activity and selectivity for the hydrogenation and reduction of nitro groups.^20,48,49  The bimetallic nitride Fe₃Mo₃N promotes selective reduction of –NO₂ in p-chloronitrobenzene to generate p-chloroaniline whereas Co₃Mo₃N favours C–Cl scission with the formation of nitrobenzene.²⁵  Earlier work has also shown that the transition metal nitrides have activity for n-butane dehydrogenation, hydrogenolysis and isomerisation.⁵⁰  Hydrogenation of p-chloronitrobenzene also occurs on La-doped NiMoB,⁵¹  NiFeB⁵²  and NiCoB.⁵³ 

Metal carbide catalysts have also been shown to be effective in the hydrogenation of dimethyl oxalate (DMO) to ethanol.^54,55  This indirect synthesis route to ethanol, uses CO oxidative coupling to produce DMO. The authors report that Cu–Mo₂C/SiO₂ and Mo₂C/SiO₂ have very good stability and activity for the hydrogenation of DMO to ethanol (Figure 1.3) at low temperatures (473 K), compared to conventional Cu/SiO₂, which although very active, degrades during hydrogenation because of agglomeration of the Cu particles.

Hydrotreating refers to a class of reactions in which heteroatoms (S, N, O) are removed from organic molecules by reaction with hydrogen to yield refined products suitable for use as fuels and chemicals.⁵⁶  Conventional hydrotreating catalysts are based on supported MoS₂ promoted with Co or Ni. These catalysts have been used commercially in the petroleum industry for decades, especially for hydrodesulfurisation (HDS) and hydrodenitrogenation (HDN) processes.^3,56  With increased environmental legislation that has reduced allowable S and N levels in fuels, catalysts with higher HDN and HDS activity, especially for the more refractory heteroaromatics present in residue oils, has become the focus of research and in this regard, metal carbides,¹³  nitrides,^13,14  and phosphides^9,57  have been applied to both HDS and HDN. In the HDS of thiophene at 450 °C, both Co₂B and Ni₃B are partially sulfided,⁵⁸  suggesting that the borides are unstable and for this reason there are few studies describing the catalytic activity of metal borides for HDN and HDS. With the growing interest in bio-oil upgrading in which O removal by hydrodeoxygenation (HDO) reactions is critical, the carbides, nitrides, phosphides and borides have also been assessed for HDO.^6,59,60 

Table 1.2 reports data from Sajkowski and Oyama⁶¹  who compared the hydrotreating activity of a commercial Ni–MoS/Al₂O₃ catalyst to that of a Mo₂N and a Mo₂C/Al₂O₃ catalyst. The catalysts were assessed using a residue oil and a gas oil, both derived from coal, so that the feedstocks had high S (810 and 116 ppm S, respectively) and N (4620 and 3580 ppm N, respectively) contents. The apparent first-order rate constants for desulfurisation, denitrogenation and aromatic saturation reactions, normalised for the number of active sites measured by suitable chemisorption experiments, show the Mo₂C to be significantly more active than the commercial catalyst.

However, in a comparison of NiMo sulfide, carbide and nitride catalysts for the HDS of thiophene and Maya crude oil, Villasana et al.⁶²  reported that NiMoS was the most effective catalyst and this trend was the same for residue conversion.⁶²  The HDN and HDS activities of bimetallic Ni–Mo carbide, nitride and sulfide catalysts have also been compared using light gas oil and heavy gas oil derived from Athabasca bitumen at 8.8 MPa in the temperature range 340–370 °C and 375–400 °C, respectively. The supported Ni–Mo sulfide catalyst was more active for HDN and HDS of light gas oil and heavy gas oil than the corresponding carbide and nitride catalysts, on a mass of catalyst basis.^63–65  In another study, P-promoted Mo₂C/Al₂O₃ was shown to be more active for deep desulfurisation of refractory compounds such as 4,6-dimethyldibenzothiophene than a commercial Ni–Mo sulfide.⁶⁶  The metal phosphides are also effective for hydrotreating real feedstocks, as demonstrated by Oyama et al.^57,67  who reported that Ni₂P/SiO₂ had better activity in hydroprocessing than a commercial Ni–Mo–S/Al₂O₃ catalyst when using a real gas oil feed at 593 K and 3.9 MPa and based on equal catalyst weights loaded in the reactor (HDS 85% vs. 80%). Subsequent studies also reported significant activity of bimetallic phosphides using realistic gas oil feedstocks.^68–70  The gradual transformation of Ni–Mo carbide, nitride and phosphide phases into Ni–Mo sulfide phases was observed during the hydrotreating reactions and similar transformations have been reported in other studies.^31,58,62  The impact of changes in composition and morphology on the long-term operation of these alternative catalysts under commercial conditions has not been reported in the literature, but these transformations likely impact the reported relative activity of the materials among the different studies.

Many other studies have demonstrated that the metal carbides, nitrides and phosphides are effective catalysts for HDS and HDN reactions using model reactants. Studies with model compounds allow reaction pathways to be clearly established. For example, in studies of the deep desulfurisation of refractory heteroaromatics on Mo₂C,^66,71  in which 4,6-dimethyldibenzothiophene (4,6 DMDBT) is used as model reactant, it was shown that the reaction proceeds by two parallel routes leading to 3,3′-dimethylbiphenyl (DMBP) by a direct desulfurisation (DDS) pathway or to 3-(3′-methylcyclohexyl)toluene (MCHT) through a hydrogenation (HYD) pathway, as summarised in Figure 1.4.⁷²  On the Mo₂C catalyst, the DDS route is more important (HYD/DDS=0.7) whereas over commercial Ni–Mo–S/Al₂O₃, operating under the same conditions of temperature (613 K) and pressure (4 MPa total pressure), the HDS of 4,6-DMDBT by the HYD pathway is favoured (80% for Ni–Mo–S/Al₂O₃).⁷²  In another important study, a series of transition metal carbides (Mo₂C, WC, NbC) and nitrides (TiN, VN, Mo₂N) were tested in a three-phase trickle-bed reactor using a model liquid feed mixture containing 3000 ppm sulfur (dibenzothiophene), 2000 ppm nitrogen (quinoline), 500 ppm oxygen (benzofuran), 20 wt% aromatics (tetralin), and balance aliphatics (tetradecane).⁷³  The Mo₂C showed higher HDN activity than a commercial sulfided Ni–Mo/Al₂O₃ catalyst (on the basis of unit area of catalyst), while VN was found to exhibit excellent activity and selectivity for the HDO of benzofuran. The HDN activity of WC was found to be comparable to that of a commercial sulfided Ni–Mo/Al₂O₃ catalyst. A comparative study of the metal phosphides using dibenzothiophene as the model reactant for HDS and quinoline for HDN at 643 K and 3.1 MPa, showed the order of activity increased as follows: Fe₂P<CoP<MoP<WP<Ni₂P,⁵⁷  on the basis of an equal number of active sites in the reactor. Few studies of HDN using metal nitrides other than Mo₂N have been reported in the literature.¹⁴  Milad et al.⁷⁴  reported that the activity per unit area of a series of unsupported metal nitrides for pyridine HDN was as follows: Co₄N>Fe₃N>Mo₂N>W₂N>NbN>CrN>VN. The activity of the metal nitrides decreased with the heat of formation of the metal nitrides.

Despite the promising activities of the metal carbides, nitrides and phosphides reported in the literature for the HDS and HDN reactions, these are well established, mature commercial technologies used to produce high quality fuels and it is difficult to displace the highly active Ni(Co)–Mo–S/Al₂O₃ catalysts that have been optimised for these processes over many years. However, with increased interest in the development of biorefinery technology for sustainable green fuels, and given that one major route to green fuels involves O-removal from intermediate bio-oils generated by pyrolysis processes,⁷⁵  there has been a very significant effort to demonstrate the use of metal carbide,^76–78  nitride^79–81  and phosphide^82,83  catalysts for HDO reactions.⁸⁴  There are several examples in the literature where the carbides, nitrides and phosphides are shown to favour reaction pathways for O removal that are different to that which occurs on group 8 metal catalysts.^76,85  For example, with phenol as a model reactant on metals such as Pt, Pd and Rh, hydrogenation reactions dominate, yielding cyclohexanol as product; whereas, on metals with higher O affinities such as Ru, Co and Ni, deoxygenation reactions occur more readily.⁸⁶  On Mo₂C and W₂C, C–O bond cleavage is favoured with minimal hydrogenation occurring especially on oxygen-modified materials.⁷⁶  Studies of the HDO of guaiacol (Figure 1.5) show that a key first step in the HDO over Ru, for example, is via catechol formation (by O–CH₃ bond cleavage) followed by phenol formation (R–OH bond cleavage),^85,87  whereas on Mo₂C direct demethoxylation occurs to yield phenol.⁷⁶  The same demethoxylation reactions occur on transition metal phosphide catalysts⁸²  whereas both paths appear to occur on Mo₂N.⁷⁹ 

Although the unique characteristics of these materials for the HDO of model compounds is well documented, few studies of their application using real bio-oils are available. Recently, however, Guo et al.⁸⁸  compared a series of metal phosphides for the HDO of a pyrolysis oil at 300 °C and 50 bar, using 40 g of pyrolysis oil and 2 g of catalyst reacted for 3 h (Figure 1.6). The authors reported that MoP/AC had HDO activity comparable to a Ru/C catalyst and the HDO activity of the metal phosphides were found to follow the order of NiP/AC>CoP/AC>MoP/AC.⁸⁸  Figure 1.6 also shows that the oil fraction yield was significantly improved by addition of P and the MoP/AC catalyst had the highest oil fraction yield among the Mo-based catalysts.

Hydrotreating vegetable oils mainly composed of triglycerides is another important deoxygenation pathway to prepare high-grade diesel-like hydrocarbons from renewable biomass.⁸⁹  Several studies have investigated metal carbides,^90,91  nitrides⁹²  and phosphides^93,94  for the HDO of vegetable oils. Table 1.3 compares results from various vegetable oils using a Mo₂C catalyst and Table 1.4 gives a comparison among Mo₂N, WN and VN supported on γ-Al₂O₃ used in the hydrotreating of oleic acid.⁹²  The Mo₂C catalyst shows high yields of deoxygenated hydrocarbons for a range of vegetable oils, whereas Table 1.4 shows similar activity for the Mo₂N catalyst, with the VN having significantly lower yield of hydrocarbon products. HDO of methyl palmitate (50% in decalin) over a 5–20 wt% Ni₂P/MCM catalyst at 3 MPa H₂, WHSV=3 h⁻¹, and H₂/oil=1000 at 290–350 °C has also been reported.⁹³  At 290 °C, the conversion was 85% with C₁₅ and C₁₆ selectivities of 55% and 45%, respectively. The HDO of methyl laurate at 573–613 K and 3.0 MPa on Ni₂P on various supports was reported by Shi et al.,⁹⁵  showing that the conversion of methyl laurate and the selectivity to C₁₁ and C₁₂ hydrocarbons was highest on the Ni₂P/SiO₂ catalyst. The effect of the support mainly results from the support acidity and the metal-support interaction that limits reducibility.⁹⁵  The high activity of the Ni₂P for O-removal compared to other metal phosphides has also been observed for HDS and HDN reactions.

In a related study, the bulk metal-rich phosphide, Ni₃P, prepared with high purity using a hydrothermal method followed by annealing at 773 K,⁹⁶  has been shown to have high activity for selective glycerol hydrogenolysis at low temperature, with high selectivity to 1,2-propanediol.⁹⁶ 

The direct catalytic conversion of raw woody biomass into phenols and diols over a carbon supported Ni-W₂C catalyst has also been reported (Table 1.5).⁹⁷  Using various sources of woody biomass, catalyst and 100 mL of water reacted at 235 °C and 6 MPa H₂ for 4 h, the carbohydrate fraction in the woody biomass was converted to ethylene glycol and other diols. A synergistic effect in Ni–W₂C/AC existed, not only in the conversion of the carbohydrate fraction, but also in lignin component degradation and the latter activity was comparable to noble metal catalysts.

In another potential application, the catalytic conversion of Kraft lignin by ethanolysis was reported over a α-MoC_1−x catalyst supported on activated carbon (AC).⁹⁸  The reaction was done in supercritical ethanol at 280 °C for 6 h in the absence of H₂. Products included C₆ monohydric alcohols, C₈–C₁₀ esters, C₇–C₁₀ monohydric phenols, C₇–C₁₀ benzylalcohols, and C₈–C₁₀ aromatic hydrocarbons. No oligomers or char were formed and both the solvent and catalyst affect the molecular yields and product composition. Wheat-straw Kraft lignin from black liquor can also be used as the feed.

In related work, the direct catalytic decomposition of alkaline lignin over WP catalysts in a hot compressed water–ethanol mixed solvent has also been reported (Table 1.6).⁹⁹  The reactor was loaded with alkaline lignin (1.0 g), fresh catalyst (0.3 g) and a mixture of water and ethanol solvent (volume ratio 1 : 1, total 100 mL), pressurised to 2.0 MPa H₂ and reacted at 553 K for 2 h. The products from the WP/Carbon catalyst consisted of 2-methoxy-phenol (guaiacol), 2-methoxy-4-methyl-phenol, 2-methoxy-4-ethyl-phenol, 2-methoxy-4-acetyl-phenol and 2-methoxy-4-propyl-phenol. The highest overall phenol yield was 67.0 mg g⁻¹ lignin.⁹⁹ 

Carbide catalysts have also been used for the conversion of cellulose to polyols, especially to ethylene glycol (EG). The catalytic performance of tungsten carbides, molybdenum carbides and platinum on carbon supports reacted at 518 K and 6 MPa for 30 min is summarised in Table 1.7.¹⁰⁰  Among all the catalysts, tungsten carbide supported on activated carbon, W₂C/AC, showed a higher yield of EG than the Pt, Ni or Mo₂C catalysts and the highest EG yield (61%) occurred with the Ni-promoted W₂C.

The metal carbides, nitrides and phosphides have been shown to be effective for a range of reactions involving molecules with one carbon atom, such as CO₂, CO and CH₄, although most applications have focused on the metal carbides. The production of synthesis gas from CH₄ is possible by steam reforming, dry reforming, or partial oxidation yielding a synthesis gas with a H₂/CO ratio close to 3, 1 or 2, respectively. Depending on the application, additional processing steps are required to adjust the H₂/CO ratio to the desired value. Given the search for new technologies for the conversion and capture of CO₂, there is a growing interest in the dry reforming of CH₄ and Mo₂C catalysts are active for this reaction.¹⁰¹  High activity of Mo₂C on various oxide supports was reported by Brungs et al.,¹⁰²  as summarised in Table 1.8. The relative stability of the catalysts is reported as Mo₂C/Al₂O₃>Mo₂C/ZrO₂>Mo₂C/SiO₂>Mo₂C/TiO₂. Subsequent studies have investigated several other metal carbides as catalysts for dry reforming, as is summarised in Table 1.9.

Combined reforming technologies including bi- and tri-reforming of CH₄ have also attracted interest owing to the elimination of the need for a secondary process to adjust the H₂/CO ratio to the required value.^103–106  Small amounts of Mo₂C (0.5–2 wt%) added to Ni/ZrO₂ catalysts were reported to be beneficial for increasing the catalyst activity of the steam-CO₂ bi-reforming of CH₄ and the increased activity was ascribed to an increased Ni dispersion and a higher Mo(II) content of the catalyst.¹⁰⁶  The best catalyst, with 0.5 wt% Mo₂C–10 wt% Ni/ZrO₂ also had higher stability in comparison with an unmodified 10 wt% Ni/ZrO₂ catalyst.¹⁰⁶  The ability of Ni/Mo₂C to catalyse CH₄ bi-reforming at 950 °C has also been demonstrated.¹⁰⁷  A rapid loss in activity of the catalyst was noted after a certain reaction period, after which Mo₂C oxidation to MoO₂ occurred, but no evidence of coking, the usual limitation in dry reforming reactions, was observed.¹⁰⁷  Tri-reforming in which the three reactions:

occur simultaneously to yield synthesis gas, has also been described using Ni–Mg catalysts supported on SiC.^104,105 

Several studies have focused on Mo₂C for the water–gas shift (WGS) or the reverse water–gas shift (RWGS) reaction: CO+H₂O⇌CO₂+H₂O. When steam reforming of CH₄ is used for the industrial production of hydrogen from CH₄ (CH₄+2H₂O→CO₂+4H₂), the product stream usually contains small amounts (1–5%) of CO as an impurity and the WGS, catalysed by Cu/ZnO/Al₂O₃ catalysts, is used to remove the CO and produce additional hydrogen. Mo₂C has been shown to be a very effective support for Pt catalysts used to catalyse the WGS reaction.^108,109  The synthesised catalyst results in a unique interaction between the Pt and the Mo₂C raft like particles, and as shown in Figure 1.7, the rate of the WGS reaction for the Pt/Mo₂C is significantly higher than that for several other Pt catalysts supported on metal oxides.

In a subsequent study,¹¹⁰  the metals Pt, Pd, Ni, Au, Ag and Cu supported on Mo₂C were also shown to have high activity for the WGS reaction. At 120 °C using a feed gas of 7% CO, 22% H₂O, 8.5% CO₂, 37% H₂, the WGS rate per unit area of catalyst for Pt, Au, Pd and Ni (1.5–2 wt%) supported on Mo₂C was 4–8 times higher than that of the commercial Cu/ZnO/Al₂O₃ catalyst.¹¹⁰ 

Synthesis gas conversion to fuels and chemicals provides a route to clean fuels that are typically S-free. In addition, depending on the catalyst and chosen process, the products from synthesis gas can be varied from heavy waxes to CH₄ and may also include a range of oxygenated hydrocarbons, such as methanol and higher alcohols. The use of non-conventional metal carbides, nitrides, phosphides and the like have all been investigated for synthesis gas conversion.

The Fischer–Tropsch synthesis (FTS) converts synthesis gas into straight chain hydrocarbons by a surface polymerisation reaction with Fe¹¹¹  and Co¹¹²  catalysts being used in industrial processes. Fe carbides, nitrides and carbo-nitrides were first studied in the 1950's,¹¹³  with the nitrided iron catalysts showing significantly higher amounts of oxygenated hydrocarbons than the iron carbide. The catalysts also underwent significant structural and chemical changes upon exposure to the synthesis gas.¹¹³  Other Fe-carbides synthesised by various methods are also known to be active for the FTS.^114,115  During FTS on Fe catalysts, the presence of Fe carbides is well known, with several different Fe-carbides being identified, including ε-Fe₂C, Fe₇C₃, χ-Fe₅C₂ and θ-Fe₃C.¹¹⁶  Recently, the formation of θ-Fe₃C, χ-Fe₅C₂, and ε-Fe₂C was shown to depend on the thermodynamic stability of each phase at the gas phase composition and temperature of the reaction.¹¹⁶  Furthermore, it was shown that a significant part of the Fe carbide phases were amorphous. The χ-Fe₅C₂ was oxidised during FTS conditions, and the θ-Fe₃C and amorphous carbide phases showed lower activity and selectivity than the other phases.

Other metal carbides, especially Mo₂C, are also known to have significant activity and selectivity for CO hydrogenation to methane^117,118  and other FTS products.^119–121  For example, CO hydrogenation at atmospheric pressure, 570 K with a 3/1 : H₂/CO ratio and SV of 10 000 h⁻¹ on unsupported Mo₂C, produced mostly C₁–C₅ paraffins while promotion of the Mo₂C with K₂CO₃ yielded C₂–C₅ hydrocarbons with 80–100% olefins and reduced the methane selectivity.¹¹⁹  Several promoters (K, Na, Co, Ce, Ba) have been added to the Mo₂C catalyst in an attempt to control product selectivity with moderate success.^120–124  The use of Mo₂C catalysts for CO hydrogenation at high pressure to yield alcohols has also been extensively investigated.^10,125  According to Xiang et al.,¹²⁶  Mo₂C catalyst operated at T=573 K, P=8.0 MPa, GHSV=2000 h⁻¹, H₂/CO=1.0 yields mostly light hydrocarbons, whereas when the same catalyst is promoted with K₂CO₃, alcohols are also produced. At a K/Mo ratio of 0.2 the selectivity to alcohols (on a CO₂-free basis) was 53% whereas the hydrocarbon selectivity was 47%.¹²⁷  Adding Fe, Co or Ni to the K/Mo₂C catalysts resulted in an increase in catalyst activity in the order Ni>Co>Fe; whereas, in terms of C₂₊OH selectivity (i.e. selectivity to alcohols above methanol) the increase was Ni>Co>Fe.¹²⁸  Ni-Mo bimetallic carbides operated at H₂/CO=2.0, T=513 K, P=7.0 MPa, GHSV=4000 h⁻¹ also produce a product mix of alcohols and hydrocarbons that depends on the Ni content.¹²⁹  As summarised in Table 1.10, the addition of the Ni significantly increases selectivity to alcohols, with the hydrocarbon fraction dominated by more than 67% CH₄.¹²⁹  In one study, the effect of support was investigated for a K promoted Mo₂C catalyst, with CO conversion reaching a maximum with about 20 wt% Mo₂C loaded onto an active carbon support.¹³⁰  This study also demonstrated the need for the K promoter to be in contact with the Mo for effective promotion of the alcohols synthesis.¹³⁰ 

Recent studies also report on metal nitrides, phosphides and borides as catalysts for synthesis gas conversion. Ultrafine catalysts of CuB and Me-CuB (Me=Cr, Th, Zr) have been investigated for methanol synthesis from CO/H₂¹³¹  and CO₂/H₂.¹³²  In one example, 20% Zr–CuB (approximately 77% Cu by mass) operated at 250 °C, 3.0 MPa and CO₂/H₂ of 1 : 3 in the liquid phase with hexadecane as solvent, produced CH₃OH at a maximum selectivity of 55% and a CH₃OH synthesis rate of 1.5 mol kg_Cu⁻¹ h⁻¹, equivalent to ∼37 g CH₃OH kg_cat⁻¹ h⁻¹.¹³²  At 150 °C, 6 MPa H₂/CO=2 : 1 the CH₃OH synthesis rate was 46 mmol h⁻¹ over the 20% Th-CuB catalyst (2 mmol Cu equivalent) in the presence of methylformate (80 mL) and potassium methoxide (3 g).¹³¹  Synthesis gas conversion to mixed higher alcohols on metal phosphides was first investigated by Zaman and Smith,¹³³  in which synthesis gas conversion at 548 K, 8.3 MPa and a H₂/CO=1 over a 10 wt% MoP/SiO₂ catalyst yielded mostly hydrocarbons with 35% selectivity to methane.¹³⁴  K addition to the MoP–SiO₂ shifted selectivity to oxygenates, with the highest oxygenate space time yield of 147.2 g kg_cat⁻¹ h⁻¹ obtained over a 5 wt% K–15 wt% MoP–SiO₂ catalyst.^133,134  The highest selectivity towards C₂+ oxygenates (76.6 C atom%) and lowest selectivity towards CH₄ (9.7 C atom%) occurred on a 5 wt% K–10 wt% MoP–SiO₂ catalyst. The major oxygenates in the product were acetaldehyde, acetone and ethanol. Promotion of MoP by addition of Co and K has also been shown to improve selectivity to C₂+ oxygenates.¹³⁵  At 548 K, H₂/CO=1.0, 5.0 MPa and GHSV=3600 h⁻¹ the best catalyst (K₁Co_0.75MoP) had a CO conversion of 14.4%, a CH₄ selectivity of 13% and C₂+ oxygenate selectivity of 43%.¹³⁵  The K promotion of Mo₂N for synthesis gas conversion to oxygenates has also been reported recently.¹³⁶  Catalysts with various K loadings (0.45–6.2 wt%) were operated at 275–325 °C, 7 MPa and 60 000 h⁻¹. The highest total oxygenate selectivity (44% at 300 °C) was observed on a 1.3K-Mo₂N catalyst, but the hydrocarbon selectivity on these catalysts remained high.

The ammonia synthesis, N₂+3H₂→2NH₃, is an important industrial process used in the manufacture of fertilisers that occurs at 450–500 °C and 30 MPa on Fe catalysts,³  while the reverse reaction is of current interest because of the potential for NH₃ to be used as a vector for H₂ storage, transportation and supply. Both reactions occur on interstitial metal nitrides and carbides. Ammonia synthesis rates, measured in a stoichiometric H₂/N₂ mixture at atmospheric pressure and 673 K were higher on β-Mo₂C, MoO_xC_y and γ-Mo₂N than on a Ru catalyst, but less than on a K₂O-Fe catalyst.⁵  These promising results led to several studies focused on similar materials.⁶  In particular, the bimetallic nitrides Fe₃Mo₃N, Co₃Mo₃N and Ni₂Mo₃N have been shown to be highly active for the ammonia synthesis^137,138  and a Cs promoted Co₃Mo₃N catalyst is reported to be more active for ammonia synthesis (15 mmol h⁻¹ g⁻¹) at 673 K under 3.1 MPa than a doubly promoted iron catalyst.^139–141  In another recent study, ammonia synthesis under mild conditions (420–500 K, N₂/H₂=1 : 3, WHSV of 60 000 mL g⁻¹ h⁻¹ and 10 bar) has been demonstrated using LiH-transition metal nitride (V to Ni) composite catalysts, in which the LiH acts as strong reducing agent, which removes activated N atoms from the transition metal nitride, yielding LiNH₂ which further reacts with H₂ to yield NH₃ and regenerate the LiH.¹⁴²  A Ba-promoted Ru catalysts, supported on BN, has also been used for the ammonia synthesis, with high activities and catalyst stability reported at 360 and 400 °C and at pressures of 50 to 100 bar.¹⁴³ 

There is significant interest in the production of hydrogen by means other than hydrocarbon reforming/partial oxidation reactions and the use of the interstitial carbides, nitrides and especially the borides as catalysts for these new processes has been proposed in several studies. H₂ production that is free of CO_x is important in the operation of polymer electrolyte membrane (PEM) fuel cells, where the PGM catalysts are readily poisoned by trace amounts of CO_x. Ammonia decomposition is one approach to obtain CO_x-free H₂ and several studies have focused on metal nitrides and carbides for this reaction. High ammonia decomposition activity of MoN_x/γ-Al₂O₃ and NiMoN_y/γ-Al₂O₃ catalysts at temperatures 600–750 °C and a GHSV of 1800–3600 h⁻¹ was reported by Liang et al.¹⁴⁴  who also identified the presence of both Mo₂N and Ni₃Mo₃N phases in the active catalyst. Several other studies of the bimetallic nitrides have since been completed.^145,146  In one case, the activity of a series of bimetallic nitrides was ranked in decreasing order for ammonia decomposition; Co₃Mo₃N>Ni₃Mo₃N>Fe₃Mo₃N>Mo₂N.¹⁴⁷  Zheng et al.¹⁴⁸  report the use of a Mo₂C catalyst with relatively high activity for ammonia decomposition, as compared to a Ru catalyst as shown in Figure 1.8.

Among the novel methods of hydrogen storage and supply, the hydride salts such as sodium borohydride (NaBH₄, NH₃BH₃, LiBH₄, etc.) are seen as safe hydrogen reservoirs that can readily produce CO_x-free hydrogen by hydrolysis or methanolysis at ambient temperature.¹⁵  Catalysts for the hydrolysis of NaBH₄ have been reviewed recently.¹⁵  The most effective catalysts are based on CoB^15,149–151  and for NaBH₄ hydrolysis the Co–B is readily promoted by a second metal.^152,153  Patel et al. reported that the activity of these materials for NaBH₄ hydrolysis decreased in the order Co–Cr–B>Co–Mo–B>Co–W–B>Co–Cu–B>Co–Fe–B>Co–Ni–B>Co–B.^15,154  NaBH₄ methanolysis (NaBH₄+4CH₃OH→4H₂+NaB(OCH₃)₄) also provides high yields of H₂ and is catalysed by Ru/Al₂O₃, Co/TiO₂, FeB, CuB and MoP. Recently, a significantly higher rate of H₂ evolution per mass of catalyst during NaBH₄ methanolysis was reported for a Ni₂P/SiO₂ catalyst, compared to Co and Ru supported on Al₂O₃.¹⁵⁵  Recent efforts to demonstrate H₂ storage and production on a practical scale include the use of a CuB honeycomb catalytic reactor for NaBH₄ hydrolysis that produced 7.55 L min⁻¹ g_Co⁻¹ at 70 °C.¹⁵⁶  At 134 °C and 5 bar outlet pressure H₂ production of up to 32 L min⁻¹ g_Co⁻¹ was obtained.¹⁵⁶ 

The metal carbides have also been the focus of recent developments in electrocatalysis applied to various processes. Water splitting (H₂O→H₂+½ O₂, ΔG=237.1 kJ mol⁻¹, corresponding to a thermodynamic voltage requirement of 1.23 V)¹⁵⁷  is seen as a promising approach to clean hydrogen. Electrochemical water-splitting includes a cathodic hydrogen evolution reaction (HER; 2H⁺+2e⁻→H₂) and an anodic oxygen evolution reaction (OER; H₂O→½ O₂+2H⁺+2e⁻). Both the OER and HER reactions require substantial overpotentials and Pt-based metals have the best activity for the HER, and Ru/Ir-based materials are the benchmark catalysts for the OER. Supply and cost issues have meant that there is a major effort aimed at the development of cost-effective and efficient alternative catalytic materials for water splitting. For example, Yu et al.¹⁵⁸  describe a new porous carbon-supported Ni/Mo₂C (Ni/Mo₂C-PC) composite catalyst derived from the thermal treatment of nickel molybdate nanorods coated with polydopamine, which catalyses the HER and OER. The catalyst affords small overpotentials of 179 mV for the HER and 368 mV for the OER at a current density of 10 mA cm⁻². The home-made alkaline electrolyser, assembled with Ni/Mo₂C-PC as a bifunctional catalyst, can enable a water-splitting current density of 10 mA cm⁻² to be achieved at a low cell voltage of 1.66 V. In another study, the use of a WC catalyst co-doped with Co and N, for both oxygen reduction reaction (ORR) and hydrogen evolution reaction (HER), is reported.¹⁵⁹  In other cases the metal carbide such as TiC has been used as a support for single atom metal catalysts, such as Pt on TiC used for the ORR. In a recent report, Co₂B was shown to be an excellent catalyst for the OER and is simultaneously active for catalysing the HER.^160,161  The catalyst achieves a current density of 10 mA cm⁻² at 1.61 V on an inert support and at 1.59 V when impregnated with nitrogen-doped graphene.^160,161 

Metal carbides have also been used on other electrochemical processes, including Mo₂C which is capable of catalysing CO₂ conversion to CH₄ at low potentials¹⁶²  and a Fe–N–C catalyst, that includes graphitic carbon, graphene, iron carbides, FeN and Fe₂N and that is suitable for SO₂ electrooxidation.¹⁶³ 

The metal carbides, nitrides, phosphides and borides have shown promise as new catalysts for a range of reactions normally catalysed by metals. The behaviour of these new materials is attributable to changes in the electronic properties of the metal surface induced by the ligands and by the metal–ligand surface geometry. The materials are active in catalytic hydrogenations, hydrotreating and synthesis gas production and conversion. In several cases, they provide unique reaction pathways and are more active or selective than conventional metal catalysts. Nevertheless, commercial application remains a challenge. The stability of the materials under the relatively severe reaction conditions encountered in hydrotreating and synthesis gas conversion, for example, is low, resulting in changes to the surface composition of the catalysts. Promising results have been reported in electrocatalytic applications such as water splitting and hydrogen production. The rich chemistry of these materials provides an opportunity for further discovery and development, with the potential to apply these materials to a range of unexplored catalytic reactions.