(337b) Catalysis on Microporous Solid Acids: Mechanism and Catalyst Descriptors for the Coupling of Alkenes and Alkanones

The conversions of light alkenes and oxygenates on solid Brønsted acids are mediated by ion-pair transition states (TS) [1,2]. The stability (with respect to their relevant kinetic precursors) of these ion-pairs, consisting of organic cations and inorganic anions formed upon proton transfer, determine turnover rates for these reactions and also those for subsequent conversions of the primary products formed, such as isomerization, β-scission, hydride transfer, cyclization and the formation of unsaturated species that can render protons unreactive. The inorganic catalyst host affects TS stability by the ability of its anion to stabilize negative charge (acid strength) and of its voids to establish van der Waals contacts with the organic guest; such contacts depend sensitively on the matching of the size and shape of host and guest. The stability of the cationic moiety at the transition state also depends on its ability to accept the proton (proton affinity) and then to interact electrostatically with the localized negative charge at the framework and via dispersion forces through van der Waals contacts with the surrounding framework. These confinement and electrostatic effects are probed here for types of reactions mediated by TS structures that differ in size, shape, or charge from each other and from their relevant precursors on solids with diverse acid strength and confining environments. Density functional theory (DFT) methods that include pertinent dispersive interactions are used to provide insights that are inaccessible from experiment alone, such as the nature and magnitude of relevant energetic descriptors in the form of van der Waals confinement energies, proton affinities of organic molecules, and deprotonation energies.

In this work, we report turnover rates (per proton) for C₂-C₄ alkene coupling (503 K, 10- 500 kPa alkene) and for aldol condensation (473 K, 0.1-10 kPa acetone) reactions on acids with different acid strength and confining environments (H-Al-TON, H-X-MFI (X=Al, Ga, Fe, and B isomorphously substituted samples), H-Al-BEA, H-Al-FAU, H-Al-MCM-41, and supported Keggin polyoxometalate clusters (H₃PW₁₂O₄₀/SiO₂)). For both reactions, turnover rates are proportional to the pressure of the respective reactants on all solid acids and the first-order rate constants reflect an energy barrier between bimolecular C-C bond-forming TS and protons that are saturated with less-charged precursors, consistent with infrared spectra and DFT-simulations; these precursors include covalently-bound alkoxides during alkene coupling reactions and H-bonded acetone during condensation reactions.

The measured rate constants are sensitive to changes in acid strength, in the case of unconfined acid sites in polyoxometalates and silica-alumina and of similarly-confined sites within MFI frameworks with Al, Ga, Fe, or B heteroatoms; here, acid strength is described in terms of deprotonation energy, a theoretically accessible value. In both cases, rate constants decreased exponentially with increasing deprotonation energy, consistent with kinetically-relevant energy barriers between charged, ion-pair TS and less charged, precursors.

Measured condensation and alkene coupling rate constants on microporous and mesoporous aluminosilicates, which contain acid sites of similar strength, reach maximum values as the framework voids and the TS structures become of similar size. The consequences of such “good fits” for reactivity are quantitatively described first based on spherical geometry by comparing the diameter of a sphere with volume equivalent to the van der Waals volume of the TS to a characteristic void size defined by the largest included diameter within the pore network. We find, however, that to describe these interactions both the size and the shape of the zeolite void and the TS must be characterized. The incomplete nature of such geometrical descriptors is resolved by using van der Waals interaction energies (E_vdw) obtained by ensemble-averaging Lennard-Jones interactions between each framework O-atom and each atom in the organic moiety (TS or precursor) at each crystallographically distinct T-site for each of the frameworks examined in this study.

The alkene coupling TS structures for alkenes of different chain length vary in size, shape, and proton affinity, thus allowing us to dissect effects of confinement and intrinsic TS stability. Their different size also show that oxide frameworks distort locally upon containment of TS structures and alkoxides as their size approaches that of the confining voids. These distortions are quantified on a geometric basis as the mean O-atom displacement and, energetically as E_dist; both metrics indicate that the distortions enhance van der Waals contacts for the host-guest structures. These distortions, however, bring forth energetic penalties that attenuate and ultimately negate the benefits of more effective van der Waals contacts, consistent with the repulsive component of E_vdw that dominates for the large isobutene coupling TS in the 10-membered ring channel of TON.

Here, we provide a framework via thermochemical cycles to understand the effects of these catalyst and mechanistic descriptors (DPE, E_vdw, E_dist, proton affinity) on the relevant energy barriers. The similar conclusions drawn for both alkene coupling and alkanone condensation reflect the resemblance of chemistry, i.e. C-C bond formation via bimolecular transition states, but, moreover, a commonality of the analysis using energetic descriptors for catalysis on microporous acids. This universality begets a tool to predict reactivity of catalysts and reactants not studied here.

[1] Sarazen, M. L., Doskocil, E. Iglesia, E., J. Catal. 344, 553 (2016)

[2] Herrmann, S., Iglesia, E., J. Catal. 346, 134 (2017)

Financial support from BP through the ICC Program and from the National Science Foundation Graduate Research Fellowship Programs. Computational resources from National Science Foundation’s Extreme Science and Engineering Discovery Environment (XSEDE; ACI- 1053575) are gratefully acknowledged.