Boron Nitride: A New Selective Catalyst for the Oxidative Dehydrogenation of Propane

[ACSGCI]

Contributed by Joseph T. Grant, Graduate Research Assistant, Department of Chemistry, University of Wisconsin-Madison; Ive Hermans, Ph.D., Professor of Chemistry and Chemical & Biological Engineering, University of Wisconsin-Madison

Chemical catalysis plays an important role in employing green chemistry principles, most notably by improving reaction efficiencies. Improvements to reaction efficiency can tangibly come in numerous forms, including reducing the amount of energy delivered to heat a reactor and reducing the energy required to separate a desired reaction product from a reaction’s product mixture.

Emerging “on-purpose” propylene technologies are a contemporary example of catalysis development for the implementation of energy-efficient chemical processes. The recent surge in shale-gas production altered the landscape for the production of small olefins, especially ethylene and propylene, the no.1 and no. 2 most-produced organic chemicals worldwide, respectively.[1] Substitution of oil-derived naphtha by shale gas as the feedstock to steam crackers translates as an economical method to produce ethylene, yet the overall production of propylene from steam crackers has flat-lined as a result. With the ever-increasing demand for propylene, methods to produce propylene “on-purpose” rather than as a byproduct of steam cracking must emerge: The current “on-purpose” propylene technique, non-oxidative propane dehydrogenation, has significant process inefficiencies, including 1) unfavorable thermodynamics (endothermicity), requiring high reaction temperatures and propylene yields limited by equilibrium; 2) reactor down-time to regenerate the catalyst due to carbon deposits constantly forming on the catalyst surface; and 3) use of precious metal (Pt-based) or toxic (Cr-based) catalysts.

The oxidative dehydrogen of propane (ODHP, Scheme 1), however, provides practical solutions to all the process inefficiencies noted above. The inclusion of oxygen in the feedstream makes the reaction exothermic, driving down required reaction temperatures, and additionally eliminates the need for catalyst regeneration. However, inclusion of oxygen also results in the formation of unwanted CO and CO₂ (CO_x) products, lowering the selectivity of the desired propylene product. Despite decades of research dedicated to catalyst development for ODHP, selectivity to propylene remains too low to be implemented in industrial applications. This lack of product control identifies the need for the discovery of alternative materials with the ability to better control this partial oxidation.[2]

While our research team explored an interesting discovery regarding enhanced two-dimensional dispersion of vanadium oxide on SiO₂ [3, 4], we simultaneously sought out an inert material with high thermal conductivity to use as a diluent material in the catalyst bed (the area inside a tube reactor where the catalyst is present). The heat generated by the catalyst during exothermic reactions (like ODHP) must be uniformly distributed throughout a catalyst bed. In such circumstances, it is essential that a reactive catalyst be mixed with an inert thermal conductor to ensure constant temperature throughout a reactor bed. After reading literature describing the thermal conductive properties and chemical resistance of hexagonal boron nitride (hBN) [5], we decided to use it as a diluent in our reactor setup for ODHP. When performing control experiments loading hBN in our reactor to examine its inertness, we were amazed to see that hBN was not only reactive for ODHP, but that it actually yielded high selectivity to propylene and low selectivity to CO_x, something that traditional catalysts for ODHP have consistently struggled to do.

Considering this unexpected yet exciting result, we followed up with many ensuing experiments to both verify these results and further explore hBN as a catalytic material.[6,7] The selectivity to propylene offered by use of this non-toxic, metal-free, and relatively inexpensive material is among the highest reported for ODHP when compared to other state-of-the-art catalysts (Figure 1). An equally exciting observation is that the main byproduct when using hBN is ethylene rather than CO_x – the main byproduct of all previously studied catalysts for ODHP. Use of boron nitride nanotubes (BNNT), best described as a hollow cylinder of hBN, results in almost identical product selectivities to hBN, yet is much more reactive. The increased reactivity of BNNT over hBN is at least partially due to the higher surface area exhibited by BNNT, resulting in an intrinsically higher amount of catalytic active sites.

By altering the concentrations of the propane and oxygen reactants and observing the response to the rate of propane consumption, we determined a rate law that suggested adsorbed oxygen sites on the hBN surface are responsible for propylene formation. Considering this insight, taken with previous literature suggesting stable oxygen sites on the armchair-termination edges of BN [8], we performed density functional theory (DFT) calculations on potential reaction intermediates that form during the ODHP reaction. Our current hypothesis is that following the initial H-atom abstraction by the O-terminated armchair site, the stabilization of the formed propyl radical (•C₃H₇) by an edge nitroxyl radical is the key to the observed high olefin selectivity (Figure 2). Traditional metal oxide catalysts may not as easily stabilize the highly reactive propyl radical and instead remain in the gas phase, free to react with O₂ and contribute to the substantial production of unwanted CO_x.

Reactivity of hBN was confirmed across three separate reactor setups and with three separate gas-chromatographs to quantify reaction products. To be sure potential metal impurities of the hBN material were not responsible, reactivity was reproduced among various hBN batches acquired by different chemical suppliers. We used inductively coupled plasma mass spectrometry (ICP-MS) to search for potential metal impurities in each hBN sample, and revealed that it must be only the boron, nitrogen and oxygen atoms responsible for reactivity.

Efforts are ongoing between our lab and collaborators to perform a techno-economic analysis of ODHP on a large scale using selectivities currently offered when using hBN. This analysis will be used to determine thresholds that must be met in order to realize commercial implementation. Our lab is also in the process of broadening the substrate scope beyond propane and expanding use of hBN as a partial oxidation catalyst for other small alkanes. It is our hope that this discovery of hBN as a catalyst (and the further improvements that will surely follow) will help eventually lead to industrial-scale implementation of these efficient partial oxidation technologies.

[1] Sattler et. al.; Chem. Rev., 114, 2014, 114, 10613-10653.

[2] Cavani et. al.; Catal. Today, 2007, 127, 113-131.

[3] Grant et. al.; ACS Catal., 2015, 5, 5787-5793.

[4] Grant et. al.; US Patent Application #2016/0228851A1, 2016.

[5] Liu et. al.; Nat. Commun. 2013, 4, 2541.

[6] Grant et. al.; Science, 2016, 354, 1570-1573.

[7] Grant et. al.; US Patent Application #15/260,649, 2016.

[8] Lopez-Bezanilla et. al.; J. Phys. Chem. C, 2012, 116, 15675-15681.

“The Nexus Blog” is a sister publication of “The Nexus” newsletter. To sign up for the newsletter, please email gci@acs.org, or if you have an ACS ID, login to your email preferences and select “The Nexus” to subscribe.

To read other posts, go to Green Chemistry: The Nexus Blog home.