The oxidation of alcohols to carbonyl compounds is a fundamental and widely used reaction in both academic and industrial settings. Traditional methods often employ inorganic oxidants, but aerobic oxidation using green oxidants like molecular oxygen and hydrogen peroxide is increasingly preferred for its economic and environmental benefits. Transition metals such as Pd, Pt, Au, and Ru are frequently used in heterogeneous catalysts for these oxidation reactions. An alternative approach, acceptorless dehydrogenation, produces both valuable carbonyl compounds and H₂, although it is less thermodynamically favorable than oxidation. This approach minimizes over-oxidation to carboxylic acids and is significant for H₂ production from biomass-derived alcohols and H-transfer reactions. Homogeneous catalysis using Ru, Ir, and Os complexes has shown success, but suffers from drawbacks like complex recycling and the need for additives, hindering large-scale applications. Heterogeneous catalysis offers a solution, with various transition metals (Ru, Pt, Pd, Re, Ag, Au, Fe, Mn, Cu, Co, Ni, and alloys) showing potential. However, a combined experimental and theoretical approach for efficient catalyst screening is lacking. This study aims to fill this gap by using a DFT-based micro-kinetic study to predict the activity of transition metals for alcohol dehydrogenation to ketones, building an experimentally verified activity-descriptor relationship to efficiently screen catalysts and guide experimental design.

Extensive research exists on the use of transition metals in both homogeneous and heterogeneous catalysis for alcohol oxidation and dehydrogenation. Homogeneous catalysts based on Ru, Ir, and Os complexes have demonstrated high efficiency in acceptorless alcohol dehydrogenation, but their industrial scalability is limited by recycling challenges and the requirement for additives. Heterogeneous catalysts, offering advantages in terms of catalyst reusability and easier separation, have been explored using various transition metals such as Pd, Pt, Au, and Ru. Studies have investigated the impact of support materials and metal particle size on catalytic activity. Theoretical studies using Density Functional Theory (DFT) have provided mechanistic insights into alcohol dehydrogenation, particularly focusing on the adsorption and dehydrogenation of methanol and ethanol on various metal surfaces. Brønsted-Evans-Polanyi (BEP) correlations have been used to predict activity trends based on C-H and O-H bond cleavage energies. However, a comprehensive, experimentally validated, combined DFT and micro-kinetic study for efficient in silico screening of catalysts for aliphatic alcohol dehydrogenation remains scarce.

The researchers employed a combined experimental and computational approach. DFT calculations were performed using the Vienna Ab Initio Simulation Package (VASP) with periodic slab models to explore the methanol dehydrogenation mechanism on various transition metal catalysts. The generalized gradient approximation with the Perdew-Burke-Ernzerhof formalism (GGA-PBE) was used for exchange-correlation energy. Two pathways were considered: the alkoxy path and the hydroxyalkyl path. The adsorption energies of atomic carbon and oxygen were used as descriptors for the micro-kinetic modeling, facilitated by the CatMAP software. Scaling and BEP relations were applied to acquire the energetic information. A 0.1% conversion at 453 K and 101 kPa was used for kinetic modeling. The theoretical predictions were validated experimentally using a continuous fixed-bed reactor with alumina-supported Ni, Co, Ru, Pd, and Pt catalysts, and silica-supported Pd and Pt catalysts. 2-octanol dehydrogenation was used as a benchmark reaction, with the experimental TOF values compared to the theoretically predicted values. Further screening was performed on 294 dilute alloys and a series of carbides and nitrides using the established activity volcano map. The stability of various facets of Mo2N was considered using Wulff construction, and the catalytic activity of selected Mo2N samples was experimentally evaluated.

DFT calculations revealed that the preferred reaction pathway for methanol dehydrogenation varied across different transition metals, with some favoring the alkoxy path and others the hydroxyalkyl path. A volcano plot of turnover frequency (TOF) as a function of carbon and oxygen adsorption energies was generated, showing two high-activity zones corresponding to the hydroxyalkyl and alkoxy pathways. Experimental results from 2-octanol dehydrogenation showed good agreement with the theoretically predicted activity trends for several metals (Ru, Co, Ni). While the experimental TOF values for Pt and Pd were lower than predicted, potentially due to coking, the overall agreement validated the theoretical approach. The volcano map enabled efficient screening of potential catalysts. 294 dilute alloys were screened, identifying 12 promising candidates with potentially high activity, although many contain noble metals. Screening of metal carbides and nitrides suggested β-Mo₂N and γ-Mo₂N (exposing (001) and (100) facets respectively) as potential cost-effective alternatives to noble metals. Experimental results with a synthesized Mo₂N sample confirmed its catalytic activity, although lower than predicted, likely due to limited (001) facet exposure and contamination. The impact of H₂ partial pressure on the TOF was also investigated experimentally for both Pt and Mo₂N catalysts, with differing responses suggesting potential advantages for Mo₂N under low H₂ pressures.

The strong agreement between the theoretically predicted activity trends and experimental results for 2-octanol dehydrogenation validates the combined DFT and micro-kinetic modeling approach. The volcano plot provides a powerful tool for catalyst screening and design, offering guidance for the identification of potentially more active catalysts beyond those currently known. The identification of β-Mo₂N and γ-Mo₂N as promising non-noble metal candidates represents a significant advance towards developing economically viable catalysts for alcohol dehydrogenation. The discrepancy between predicted and experimental TOF values for some metals highlights the importance of considering factors like coking and facet exposure in catalyst design. Future work should focus on optimizing the synthesis of Mo₂N to maximize the exposure of (001) and (100) facets to enhance its catalytic performance.

This study successfully establishes a framework for identifying active catalysts for acceptorless alcohol dehydrogenation. The combined DFT and micro-kinetic approach, validated experimentally, enables efficient in silico screening of catalysts. The identification of β-Mo₂N and γ-Mo₂N as potential cost-effective replacements for noble metals represents a major contribution. Future research should focus on synthesizing phase-pure Mo₂N with high (001) and (100) facet exposure, and further exploring the identified promising alloy catalysts.

The study acknowledges limitations, including the potential impact of coking on the activity of some catalysts, particularly Pt and Pd. The experimental TOF values for Mo₂N were lower than predicted, possibly due to a limited proportion of the most active (001) and (100) facets and surface contamination. The theoretical calculations assumed a constant number of active sites, which may not perfectly reflect real-world conditions. The synthesis and evaluation of the identified alloy catalysts require further experimental effort to fully assess their performance.