First-principles-based kinetic modeling of the zeolite-catalyzed conversion of furans to aromatics
University of Delaware
The development of strategies to convert renewable feedstocks into fuels and chemicals and displace petroleum is motivated from economic, environmental and political reasons and is arguably one of the world's foremost scientific energy challenges. The nascent shale gas revolution may, at best, postpone the transition to renewables. Shale gas, however, does not provide aromatics to the extent that naphtha does and thus the development of renewables-based strategies for aromatics is more crucial than ever. A promising, sustainable route to synthesize aromatics from biomass-derived furans has recently been proposed following the discovery that DMF and ethylene can be selectively converted to p-xylene by dehydration of the Diels-Alder product over zeolites that contain Brønsted sites. The objective of this thesis is to reveal the underlying mechanism and provide insights into catalyst design principles. In order to achieve this, we combine electronic structure calculations and microkinetic modeling to describe how the two reactions - Diels-Alder cycloaddition and dehydration - work in tandem and to investigate the side reactions that affect selectivity. Specifically, we study the Brønsted acid-catalyzed dehydration of the cycloadduct and the ring opening hydrolysis reaction of DMF. We show that Brønsted acids are quite effective at catalyzing dehydration and hydrolysis. We pay particular attention to the ability of zeolites with extra-framework Lewis acid centers to heterogeneously catalyze the Diels-Alder cycloaddition. We elucidate the factors that determine Lewis acid activity and propose reactivity descriptors that encapsulate the underlying physics. In order to accomplish this study, we vary the active Lewis acid site environment from an isolated active site center to an active site cluster model and an embedded active site cluster model to determine the balance between accuracy and computational cost and guide future computational studies about a suitable way of modeling the active site. Even though Lewis acids are known to catalyze Diels-Alder cycloaddition by closing the gap between the frontier molecular orbitals of the addends, our calculations show that alkali-exchanged zeolites Y do not generally exhibit notable Lewis acid activity. Charge screening of the Lewis acid centers, due to significant charge transfer from the framework oxygen atoms to the alkali cations, diminishes their catalytic ability. As a result, the reaction is shown to follow bi-directional instead of normal electron flow and homogeneous catalyzed chemistry is equally effective. We propose that effective heterogeneous catalysis of Diels-Alder cycloaddition should involve zeolite frameworks whereby the Lewis acid centers are less embedded and thus less screened from electron density transfer from the framework. Microkinetic modeling of the conversion of DMF and ethylene over the zeolite HY revealed that HY can only catalyze the dehydration of the Diels-Alder product and that the cycloaddition itself proceeds uncatalyzed, outside the zeolite, i.e., homogeneously. The overall process follows two distinct kinetic regimes as a function of catalyst concentration. At low catalyst loadings, the rate of p-xylene production increases linearly with the concentration of active sites and the rate-limiting step is the Brønsted-catalyzed dehydration. At high catalyst loadings, the reaction rate is maximal and independent of the available Brønsted sites. In this regime, the rate is controlled by the uncatalyzed cycloaddition, despite the fact that the cycloaddition activation energy is higher than that of the catalyzed dehydration.