(679b) Oxidative Dehydrogenation of Propane on Boron Nitride Catalyst: A Computational Investigation of Active Site As Well As Involved Mechanism Using Boron Nitride Nanoribbons

Being one of the most important light olefins in the petrochemical industry, propene has attracted immense research interest in the recent time. Industrially, propene is mainly produced from the cracking of naphtha, which suffers from very high energy demand, poor selectivity and low yield as well as extensive emission of CO₂, and indeed, triggered the continuous efforts from both experimental studies and theoretical simulations to discover more suitable, economic and environment friendly catalysts and reaction mechanisms for propene production. With a favorable thermodynamic and kinetic characteristic, the oxidative dehydrogenation (ODH) of propane to propene is a highly desirable process. In the recent past, for the state-of-art ODH of propane to propene, transition metal oxides such as vanadium oxide, molybdenum oxide, tungsten oxide etc. have been explored extensively as catalyst. However, over-oxidation of the produced propene in the presence of such metal oxide catalyst results in CO_x as the final product, which drastically reduce the selectivity of propene formation.

In this context, hexagonal boron nitride (h-BN) appears as a potential metal-free catalyst for ODHP with high activity but only negligible CO₂ formation. The gold rush starts with the experimental report by Hermans and co-workers (Science 2016, 354, 1570-1573) that the molecular oxygen adsorbed armchair edge of h-BN can offer high activity towards ODHP with negligible CO₂ formation when heated to 460-500 °C under flowing propane, oxygen and nitrogen gas. Few recent studies proclaimed that the high selectivity towards propene under ODHP reaction conditions is not unique for only BN, but rather similar for other boron-containing materials, and hence, concluded that boron is the essential element to achieve high propene selectivity. In spite of these advances, the exact nature and function of the active sites in these highly active catalysts is still a matter of debate and disagreements have been reported regarding the active site as well as mechanism involved. Moreover, we believe that molecular N₂ present in such high concentration at such high temperature reaction conditions can also play some role to provide some kind of nitrogen rich and or NO_x type active site for ODHP, which further complicates the understanding of the actual reaction mechanism. Hence, at this point, clear understanding about the ODHP reaction mechanism on h-BN is becoming very much crucial, not only to develop a more efficient catalyst, rather, to accelerate its commercial implementation. In view of that, combining DFT with microkinetic modeling techniques, we have systematically explored the ODHP reaction on oxygen passivated boron nitride nanoribbons (BNNRs). Our goal is to identify the most stable O₂ adsorbed BNNR edge, which eventually can represent the actual active site of the experimental reaction conditions, and the possible reaction pathways on it.

In the present work, we considered BNNRs with three different edges such as (1) armchair edge (armBNNR), (2) zigzag edge (zzBNNR), and (3) modified zigzag edge (mzzBNNR) with boron dangling bonds in one zigzag edge and nitrogen dangling bonds in other. We find that mzzBNNR is much less stable than armchair and zigzag BNNR in the bare form and the stability order lies as: armBNNR zzBNNR mzzBNNR. After O₂ adsorption, mzzBNNR becomes the most stable structure and the stability order reverses: O@mzzBNNR O@zzBNNR O@armBNNR. Thus, one metastable bare BN edge can eventually become most stable after oxygen adsorption and are likely to occur more readily under experimental reaction conditions. Next, we find that the O adsorbed N-edge of mzzBNNR, with a NO_x type active site, can catalyze the ODHP reaction. The ODHP reaction over this NO_x -type active site can occur through two pathways: Pathway I, with highly stable intermediate but relatively large barrier, and Pathway II, with smaller barrier but higher energy intermediates than the former. The results of our microkinetic model suggest that Pathway II is in isolation faster, with similar kinetic behavior of the reactant gases as observed experimentally (ChemCatChem 2017, 9, 1788-1793). However, the overall rate is dominated by Pathway I that still possess an activity that can possibly explain experimental observations.