UC Berkeley

An emphasis on clean energy and a more efficient use of current fuel sources is important for an environmentally-conscious future energy landscape. This study aims to elucidate the process of growing low molecular weight alkenes, obtained from both dehydration of renewable oxygenate feedstocks (C2-C4 alcohols) as well as conventional refinery sources, to higher molecular weight hydrocarbons in the transportation fuel range. Light hydrocarbons are converted on solid Brønsted acids via oligomerization, isomerization, β-scission, hydride transfer and cyclization reactions, all of which are mediated by ion-pair transition states (TS). The relative stability of these ion-pairs, consisting of organic cations derived from reactants and inorganic anions formed by deprotonation of the inorganic catalyst, relative to their relevant precursors determine the rate of each reaction and their prevalence over the others. The anion contributes to TS stability by (i) van der Waals interactions conferred by the geometry of its surroundings (confinement) and (ii) its willingness to accept the negative charge (acid strength) and to interact electrostatically with the organic cation. The cation, in contrast, determines TS stability through its ability to accept the proton and then interact with the anion. These confinement and electrostatic effects can be probed using reactions mediated by TS that differ in size, shape or charge to various extents from each other and from their relevant precursors; density function theory (DFT) methods that account for the pertinent dispersive interactions further provide detailed information that is inaccessible from experiments alone, such as how the cation and anion conform to each other.

Here, turnover rates (per proton) for C2-C4 alkene oligomerization reactions and for the incorporation of C4-C5 alkanes via hydride transfer on acids of different acid strength and confining environments (TON, MFI, BEA, MOR, FAU zeolites; amorphous silica-alumina (SiAl), Keggin polyoxometalate clusters (POM) clusters on SiO2) are reported. Along with oligomerization and hydride transfer, product distributions indicate extensive isomerization and β-scission. Rapid methyl and hydride shifts generate equilibrated pools of skeletal and regioisomers, as determined from chemical speciation and isotopic labeling. This conclusion that isomerization is equilibrated during oligomerization is made despite disagreement with tabulated gas-phase thermodynamic distribution because even small inaccuracies in this tabulated data yield significant differences in calculated equilibrium constants. While isomerization is facile on all catalysts, the extent of β-scission is different on different catalysts. It does not trend with confinement, but instead with the micropore connectivity of the acid catalyst, where the presence of voids that are larger than the surrounding channels, created from the intersections of the channel network, increases the selectivity to secondary β-scission reactions. The product distribution directly correlates with the difference in size between the intersecting void and the subsequent channel, i.e. products made in voids larger than the apertures through which they must diffuse, will first undergo more β-scission to facilitate egression from the crystal as smaller, more mobile species. This thorough molecular view of the product distribution has not previously been investigated for zeolite-catalyzed oligomerization, despite it having commercial relevance.

Turnover rates for C-C bond formation and hydride transfer show that kinetically-relevant steps involve reactions of an alkene or alkane to alkene-derived alkoxides present at saturation coverage, consistent with in-situ infrared spectra and DFT estimates of activation free energy barriers and of the stability of alkoxide intermediates. These kinetic data, obtained over a very broad range of reactant pressures, allow a systematic comparison between theory and experiment and also accurate estimates of alkoxide adsorption constants. In doing so, this study provides a quantitative assessment of the effects of surface curvature and of alkene size and structure on alkoxide stabilities.

The rate constants for oligomerization and hydride transfer increase exponentially as acid strength, or the stability of its conjugate anion, increases. Here, the acid strength of a Brønsted acid is defined as the deprotonation energy (DPE) required to heterolytically cleave the O-H bond, which is theoretically accessible for known structures. Effects of acid strength on rate constants reflect DFT-derived transition states that differ in charge from their alkoxide precursors. Rate constants also generally increase with increasing TS size for both oligomerization and hydride transfer on each zeolite framework, because of the combined effects of the greater stability of the larger TS carbenium ions and their more effective contact with the void walls. DFT treatments show that zeolite frameworks distort locally, so as to enhance van der Waals contacts at the expense of a slight distortion of the framework lattice, which ultimately becomes too costly as the TS becomes larger than the size of the confining voids. For example, the lattice of TON (0.57 nm channel) locally moves closer to the TS for smaller TS (ethene dimerization; 0.46 nm diameter) and moves away for larger TS (isobutene dimerization; 0.58 nm diameter); the distortions become too costly, however, when the TS is much larger than the void (isobutane-C6 hydride transfer; 0.81 nm diameter), consistent with a lower rate constant for TON than for larger pore environments. These energy compromises are also critical in alkoxide formation. DFT-calculated alkoxide energies indicate bulkier alkoxides distort the framework more in smaller, concave environments (TON) than larger pore ones (MOR, HPW).

This array of transition states and their precursors—formed from a range of reactants and catalysts—exploits the diversity in size, shape and charge in solid acid upgrading of alkenes and alkanes and provides unprecedented clarity of the complex relationship between organic moieties and inorganic catalysts when combined with state-of-the-art theoretical methods. The descriptors developed here that relate transition state properties (intrinsic stability of the organic cation, electrostatic interaction with the anion (acid strength) and stabilization via confinement provided by the flexible inorganic framework) to reactivity and selectivity provide a basis for extrapolation to other transition states and acid catalysts than those studied here, which has implications on overall catalyst efficiency.