Application of QM simulations and multivariate analysis in the study of alkene reactivity in the zeolite H‐ZSM5

Abstract

Application of QM simulations and multivariate analysis in the study of alkene reactivity in the zeolite H-ZSM5 Duangkamol Gleeson a *Reported herein are the results of an investigation into the effect of the extended framework of the zeolite ZSM-5 on the reaction energetics and structures of (a) the physisorbed complex formed between the zeolite and six alkenes, (b) the corresponding chemisorbed alkoxide intermediate and (c) the transition states (TS) connecting the two. For this, quantum mechanical (QM) simulations of ZSM-5 in the presence and absence of the zeolite framework have been employed. A 46T density functional theory (DFT) cluster model and a 3T:46T DFT:UFF ONIOM model are used to represent the former scenario and a simple 3T DFT cluster model for the latter. The structural implications of neglecting the zeolite framework have been rigorously compared using the multivariate statistical method principal components analysis (PCA). This method allows one to assess the correlated nature of the changes in structure along the reaction coordinate, for multiple different alkenes, in a facile, reliable way. . ZSM-5 is a well-studied medium pore-size zeolite from the MFI family. It is characterized by a three-dimensional pore system consisting of sinusoidal 10T ring with intersecting 10T ring channels. While the Brønsted acid site locations can be similar in different 10T ring zeolites, they can display considerable selectivity differences, including the production of isobutene from linear butanes [1]. This is because the framework topologies can alter the catalytic reactions through (a) moderation of the pKa of the Brønsted acid site, (b) hinder/allow the formation of key intermediates due to the local pore volume, (c) stabilize stationary points on the potential energy surface through long range interactions as a result of high Si/Al ratios or the presence of heavy metal additives or (d) impose diffusional limitations as a result of coking/time on stream.Quantum mechanical (QM) calculations [3][4][5][6][7] are often used to model the reactivity and selectivity of zeolites since through a consideration of the relative energies and structures of the zeolite-substrate complexes, transition states (TS) and products, one can often get insights into chemical processes not possible by experimental techniques. A common feature of zeolite simulations is their consideration of acidity, focusing either on the non-bonded 'physisorption' of various small molecules to the Brønsted acid site [8], or more complex reactions that occur at it. Alkoxide intermediate formation involves the transfer of the Brønsted proton on the zeolite O1 atom to the C1 of the alkene with the subsequent formation of a bond between the C2 and O2 atoms of the substrate and zeolite, respectively (Figure 1). Theoretical studies on alkoxide intermediates have often been restricted to relatively small cluster models [9][10][11] Studies comparing gas phase cluster models to those that encode information about the active site are generally restricted to ...

Structural Fluctuations in Enzyme-Catalyzed Reactions: Determinants of Reactivity in Fatty Acid Amide Hydrolase from Multivariate Statistical Analysis of Quantum Mechanics/Molecular Mechanics Paths

“…To extract meaningful information from QM/MM calculations, we have employed a range of statistical approaches: (a) principal component analysis (PCA), (b) partial least-squares (PLS) regression analysis, , (c) multiple linear regression (MLR) analysis . These are widely used statistical techniques in quantitative structure−activity relationships (QSARs). − Although PCA and/or PLS has been applied to the field of QM calculations on small model systems, − to the best of our knowledge this is the first time where several multivariate statistical techniques have been employed to relate the changes in energy to distances and angles in QM/MM calculations of an enzyme. The analyses revealed the presence of two fully independent geometrical effects directly correlated with substrate binding and TS stabilization controlling the rate of the reaction.…”

DFT based microkinetic modeling of confinement driven [4 + 2] Diels–Alder reactions between ethene and isoprene in H-ZSM5

“…Shen W. found the intrinsic energy barrier for the chemisorption of ethene from the physisorbed state to be 70–98 (kJ mol −1 ); meanwhile Chu et al have found barriers between 68 and 80 (kJ mol −1 ). Work by Gleeson D. 87 using DFT based cluster models calculated an intrinsic barrier between 78 and 119 (kJ mol −1 ). Based on these results, our calculated intrinsic barrier of 64 (kJ mol −1 ) was slightly underpredicted.…”