UC Berkeley

Hydroisomerization and hydrocracking reactions convert long-chain alkanes to more highly branched and shorter analogs, respectively; these reactions occur on bifunctional metal-acid catalysts through kinetic cascades. The metal function is responsible for dehydrogenation of reactant alkanes and hydrogenation of product alkenes, both in equilibrated steps. Reactant alkenes undergo kinetically-relevant reaction on acid sites and thus, measured rates reflect the reactive properties of the acid. Zeolites, zeotypes, and mesoporous aluminosilicates are commonly used as the acid function in industrial catalysts, where they are combined with oxide binders to form extrudates, onto which metal precursors are then applied. Subsequent treatments result in metal clusters whose location and proximity to acid sites cannot be accurately determined. Physical mixtures of a metal function (here, Pt clusters supported on mesoporous SiO2) and different nanoporous solid acids are useful as model catalysts because they define the size of the acid domain, which consists of the contiguous region of acid sites residing between metal sites; these domains are given by the crystallite diameter for zeotype and aluminosilicate acid systems. Additionally, these mixtures can be made with sufficient metal contents to establish equilibrated reactant dehydrogenation, irrespective of the nature and reactivity of the acid function. These bifunctional catalysts thus allow mechanistic interpretation of confinement, acid strength, and site-proximity on reactivity and product selectivity, in the present case for n-hexane and n-heptane isomerization and for 2,4-dimethylpentane isomerization and β-scission reactions.

In this work, we address the effects of site proximity, confinement, and acid strength on alkane isomerization and β-scission reactions by combining experimental kinetic measurements with insights from periodic density functional theory (DFT) calculations. Reaction rates are normalized by the number of protons in each sample, counted either during reaction or using ex situ techniques. Chemical interpretations of measured turnover rates reveal that first-order rate constants (per H+) reflect the free energy difference between the kinetically-relevant ion-pair transition states and the gas-phase alkene precursors with bare protons; zero-order rate constants, in turn, reflect the same transition state free energy but with respect to alkoxides (bound alkenes). The experimental evaluation of rate constants normalized per H+ in each kinetic regime allows rigorous comparison with free energies from structures converged using periodic density functional theory methods that account for dispersion forces on bound intermediates and transition states.

Mesoporous and zeolitic aluminosilicates contain Brønsted acid sites of similar acid strength that reside, however, within diverse topologies. These pores and apertures have sizes on the order of molecular dimensions and stabilize contained reaction intermediates and transition states through van der Waals interactions that depend sensitively on the “fit” between the reactant-derived guest and the inorganic framework. Confinement effects on measured first-order rate constants (per H+) for n-heptane isomerization demonstrate the sensitive nature of the stabilizing van der Waals contacts between the framework and the alkyl-substituted cyclopropyl carbenium ions present at ion-pair transition states.

The same small pores and cages that stabilize transition states also hinder the diffusion of guest molecules in a manner that depends sensitively on the backbone length and degree of branching in reactant and product molecules. This results in significant contribution of secondary reactions to measured product selectivities. We use here experimental techniques to systematically change site proximity on-stream, where we define site proximity as the inverse of the intracrystalline residence time for a given molecule or, more rigorously, its Thiele modulus; such techniques involve the gradual desorption of weakly bound titrants during reaction to systematically to decrease site proximity. These experiments, taken together with rigorous reaction-transport formalisms, illustrate the measured effects of Thiele moduli and the strong effects of diffusion-enhanced secondary reactions on measured product selectivities. They also enable the assessment of intrinsic (single-site sojourn) selectivities that are not corrupted by such diffusional constraints. These approaches, taken together, demonstrate that confinement effects are similar for all isomerization and β-scission transition states involved in n-heptane and 2,4-dimethylpentane isomerization and β-scission pathways. High measured β-scission selectivities during n-heptane reactions, where it is exclusively a product of secondary reactions, therefore only reflect the diffusional hurdles of β-scission precursors (dimethylpentenes) in egressing from acid domains before undergoing these secondary reactions. We also demonstrate here using n-hexane and n-heptane as reactants that equilibrated product isomers can be sensitively differentiated by medium-pore zeolites (MFI, TON, MTT) on the basis of their differing diffusivities. Their measured selectivity ratios thus indicate the apparent preferential formation of one isomer over the others (e.g., 2-methylalkane over 3-methylalkane), while, in fact, these obesrvations merely reflect their different ability to egress from the acid domain where the two isomers are present in thermodynamic equilibrium within acid domains.

β-Scission and isomerization are both primary pathways directly accessible from 2,4-dimethylpentane reactant molecules. In spite of the strong intracrystalline concentration gradients of its alkene regioisomers within MFI zeotypes with different acid strength but similar confining environments, their intrinsic (single-site sojourn) rate ratios can be assessed to determine the effects of acid strength on reactivity and selectivity. Isomerization rate constants decreased with increasing acid strength, defined rigorously as deprotonation energy (DPE), which primarily reflects the less stable conjugate anion present at ion-pair transition states for weaker acids. Isomerization selectivities are invariant with DPE, reflecting the similar amounts and delocalization of charges in the relevant transition state carbocations. In contrast, selectivities to β-scission increase with decreasing acid strength. Carbocations at β-scission transition states localize charge and interact more effectively with the conjugate anion, and their energies are accordingly less affected by the energy required to deprotonate a weaker acid.

These effects of confinement, acid strength, and site proximity were assessed on physical mixtures of Pt/SiO2 with mesoporous and microporous aluminosilicates and zeotype materials. Site proximity here was limited to the size of the acid crystallites. Platinum clusters were introduced into zeolite crystallites in order to decrease the size of acid domains beyond that of the zeolite crystallites. n-Heptane isomerization turnover rates (per H+) on these materials showed increases in rates of factors of 2-8, in spite of the absence of reactant alkene concentration gradients within crystallites, shown by systematic changes in the number of protons and the diffusional characteristic times. rigorous reaction-transport analysis that demonstrated the lack of reactant-alkene concentration gradients within crystallites. We explore here the plausible existence of parallel reaction pathways that become relevant when metal clusters and acid sites are separated by only nanometer distances.

In this work, we have developed correlative and causative relationships between solid acid properties (confinement, acid strength, site proximity) and reactivity and selectivity in bifunctionally catalyzed alkane isomerization and β-scission reactions. The methods elucidating, models describing, and descriptors isolating the catalytic consequences of confinement, acid strength, and site proximity are generalizable to all phenomena in solid acid catalysis. The insights discussed herein can be applied to the design and development of catalytic architectures for desired reactivity and selectivity.