Experimental and Computational Investigation of Acetic Acid Deoxygenation over Oxophilic Molybdenum Carbide: Surface Chemistry and Active Site Identity

Ex situ catalytic fast pyrolysis (CFP) is a promising route for producing fungible biofuels; however, this process requires bifunctional catalysts that favor C–O bond cleavage, activate hydrogen at near atmospheric pressure and high temperature (350–500 °C), and are stable under high-steam, low hydrogen-to-carbon environments. Recently, early transition-metal carbides have been reported to selectively cleave C–O bonds of alcohols, aldehydes, and oxygenated aromatics, yet there is limited understanding of the metal carbide surface chemistry under reaction conditions and the identity of the active sites for deoxygenation. In this paper, we evaluated molybdenum carbide (Mo₂C) for the deoxygenation of acetic acid, an abundant component of biomass pyrolysis vapors, under ex situ CFP conditions, and we probed the Mo₂C surface chemistry, identity of the active sites, and deoxygenation pathways using in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), X-ray photoelectron spectroscopy (XPS), and density functional theory (DFT) calculations. The Mo₂C catalyst favored the production of acetaldehyde and ethylene from acetic acid over the temperature range of 250–400 °C, with decarbonylation pathways favored at temperatures greater than 400 °C. Little to no ethanol was observed due to the high activity of the carbide surface for alcohol dehydration. The Mo₂C surface, which was at least partially oxidized following pretreatment and exposure to reaction conditions (possibly existing as an oxycarbide), possessed both metallic-like H-adsorption sites (i.e., exposed Mo and C) and Brønsted acidic surface hydroxyl sites, in a ratio of 1:8 metallic:acidic sites following pretreatment. The strength of the acidic sites was similar to that for H-Beta, H-Y, and H-X zeolites. Oxygen vacancy sites (exposed Mo sites) were also present under reaction conditions, inferred from DRIFTS results and calculated surface phase diagrams. It is proposed that C–O bond cleavage steps proceeded over the acidic sites or over the oxygen vacancy sites and that the deoxygenation rate may be limited by the availability of adsorbed hydrogen, due to the high surface coverage of oxygen under reaction conditions. Importantly, the reaction conditions (temperature and partial pressures of H₂ and H₂O) had a strong effect on oxygen surface coverage, and accordingly, the relative concentrations of the different types of active sites, and could ultimately result in completely different reaction pathways under different reaction conditions.