Engineering intrapore environments to tune rates and selectivities for liquid phase catalysis over zeolites

Potts, David Samuel

Permalink

https://hdl.handle.net/2142/125509 Copy

Description

Title

Engineering intrapore environments to tune rates and selectivities for liquid phase catalysis over zeolites

Author(s)

Potts, David Samuel

Issue Date

2024-05-23

Director of Research (if dissertation) or Advisor (if thesis)

Flaherty, David W

Doctoral Committee Chair(s)

Flaherty, David W

Committee Member(s)

• Guironnet, Damien S
• Mirica, Liviu M
• Mironenko, Alex V

Department of Study

Chemical & Biomolecular Engr

Discipline

Chemical Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Zeolites
• Catalysis
• Solvents

Abstract

Covalent and non-covalent interactions among reactive species, solvent molecules, active sites, and intrapore functions influence the rates and selectivities of liquid-phase catalysis in zeolites by orders of magnitude. These effects carry consequences for many chemical processes, including reactions relevant to the conversion of renewable or recyclable feedstocks and the development of distributed chemical manufacturing. Prior literature demonstrates that the epoxidation of alkenes and the ensuing epoxide ring-opening reaction depend sensitively on covalent and non-covalent interactions at solid-liquid interfaces. However, the literature lacks a consensus on how to design more efficient catalysts and solvent systems for these important chemistries. Comparisons among rates and selectivities in previous reports are frequently convoluted by differences among synthesized catalysts, reaction conditions, kinetic regimes, and mass transfer artifacts. Here, we synthesize a suite of well-characterized zeolites with the *BEA framework, incorporated with various active metal atoms and possessing different silanol ((SiOH)x) densities. We utilize a combination of kinetic, thermodynamic, and spectroscopic experimental methods to systematically probe how the solvent composition, reactant structure, and zeolite choice influence the rates and selectivities of alkene epoxidation and epoxide ring-opening. The analysis of these systems provides useful strategies for interpreting and controlling the intrapore interactions that govern liquid-phase catalysis in zeolites. Chapter 1 discusses the importance of catalyst development in the design of next-generation sustainable chemical processes. We introduce a quantitative framework for understanding how interactions at solid-liquid interfaces influence heterogeneous catalysis, which carries relevance for both established and emerging chemistries. An overview of previous alkene epoxidation and epoxide ring-opening literature demonstrates the strong influence of interactions among catalysts, solvents, and reactants on both chemistries, highlighting the utility of these reactions as probes to understand how to manipulate covalent and non-covalent interactions to promote the formation of desired products during catalysis. The size and structure of alkenes significantly influence the rates and selectivities of the adsorption and catalytic transformation of these molecules; however, the physical origins of kinetic differences among substrates remain unclear. Chapter 2 investigates the kinetics of alkene epoxidation with hydrogen peroxide (H2O2) and the thermodynamics of epoxide adsorption in hydrophilic (Ti-BEA-OH) and hydrophobic (Ti-BEA-F) zeolitic materials as a function of alkene and epoxide carbon number (C6-C18). We utilize a combination of experimental and computational methods to describe the significant effects of solvent displacement and confinement on epoxidation catalysis in zeolites. Epoxidation rates over hydrophilic, (SiOH)x-rich Ti-BEA-OH exceed those over hydrophobic, (SiOH)x-poor Ti-BEA-F by 40-450 times across the range of alkenes in acetonitrile (CH3CN) solvent containing water (200 mM H2O). Adsorption measurements with 1H nuclear magnetic resonance (NMR) spectroscopy and grand canonical molecular dynamics simulations (GCMD) show that Ti-BEA-OH adsorbs 10-20 times more H2O than Ti-BEA-F. While rates decrease for each catalyst under anhydrous conditions, rates decrease more significantly over Ti-BEA-OH (10-90 times) than Ti-BEA-F (< 3 times). The rate differences do not result from changes in reaction mechanism or mass transfer constraints, but rather changes in transition state stability described by activation enthalpies (∆H^‡) and entropies (∆S^‡). Values of ∆H^‡ and ∆S^‡ increase systematically with alkene carbon number over both zeolites because longer alkyl chains disrupt more solvent molecules. The disruption of H2O in Ti-BEA-OH also yields greater changes in ∆H^‡ and ∆S^‡ than for CH3CN, which dominates the pores of Ti-BEA-F. Corresponding epoxide adsorption enthalpies, measured experimentally with isothermal titration calorimetry and computed with GCMD, become more endothermic as chain lengths increase and correlate to ∆H^‡, providing evidence that the disruption of solvent molecules depends on adsorbate size. These findings show that the entropic gain for a given enthalpic cost of solvent disruption varies with alkene carbon number, (SiOH)x density, and solvent composition within the pore. These manipulations of non-covalent interactions at zeolite surfaces offer opportunities to increase epoxidation turnover rates. The fractional replacement of organic solvents with H2O as a co-solvent reduces solvent waste but also impacts the kinetics of catalytic reactions. Chapter 3 extends our understanding of solvation effects from the previous chapter by examining how replacing significant fractions of organic solvents with H2O affects alkene epoxidation and the competing H2O2 decomposition pathway. Turnover rates for 1-hexene (C6H12) epoxidation over Ti-BEA-OH and Ti-BEA-F increase while rates for H2O2 decomposition decrease as the mole fraction of H2O increases in CH3CN, methanol (CH3OH), and gamma-butyrolactone co-solvents, leading to significantly higher epoxidation selectivities at higher H2O mole fractions. The mechanisms of epoxidation or H2O2 decomposition do not change across the range of H2O mole fractions and organic co-solvents tested. Instead, differences among rates and selectivities stem from changes in the free energy of reactive species along each pathway, evidenced by ∆H^‡ and ∆S^‡ measurements, epoxide adsorption enthalpies, and rates normalized by reactant activities. Greater epoxidation rates at higher H2O fractions appear to originate from the destabilization of liquid-phase C6H12 and entropic stabilization of the epoxidation transition state through the disruption of hydrogen bonds between H2O. In contrast, decreasing rates of H2O2 decomposition with increasing H2O mole fraction likely result from disproportionate stabilization of reactive intermediates (Ti-OOH, liquid-phase H2O2) relative to the decomposition transition state. Overall, the findings from Chapters 2 and 3 show that creating a more polar reaction environment promotes catalytic pathways with predominantly hydrophobic reactive species while suppressing pathways involving hydrophilic species, providing further guidance for how to control non-covalent reactions to promote more efficient reactions at solid-liquid interfaces. Chapters 4 and 5 demonstrate that the intrapore phenomena that influence alkene epoxidation also influence nucleophilic epoxide ring-opening reactions within both Lewis and Brønsted acidic zeolites, which appear as large differences in turnover rates and regioselectivities. While previous literature reported that active site and solvent environments affect rates and regioselectivities of epoxide ring-opening, a conceptual framework to explain these observations and the design principles necessary to improve these processes remains unestablished. Chapter 4 develops a mechanistic understanding to explain how active metal (Al, Sn, Ti, Zr) in zeolite *BEA affects 1,2-epoxybutane (C4H8O) ring-opening with CH3OH in CH3CN co-solvent. Among Lewis acids, ring-opening rates increase with the Lewis acid strength of the transition metal (r_Ti < r_Zr < r_Sn), based upon reported Mulliken electronegativities and measured C4H8O adsorption enthalpy values. Rates depend on both [CH3OH] (where [x] denotes the concentration of species x) and [C4H8O]. A single kinetic regime dominates at all conditions over Al-BEA, while each Lewis acid shows two kinetic regimes. These results suggest Lewis and Brønsted acidic materials catalyze ring-opening with distinct reactive intermediates or through different mechanisms. Regioselectivities towards the terminal ether product exceed 50% under most conditions but depend on metal identity and reagent concentrations. Values of regioselectivity to the terminal ether generally decrease with increasing [CH3OH] and increase with greater [C4H8O] over each M-BEA. Al-BEA consistently gives the lowest regioselectivities to the terminal ether, while Ti- and Zr-BEA show greater regioselectivities than Sn-BEA among the Lewis acids. Values of ∆H^‡ suggest differences among regioselectivities arise from differences in the free energies of the transition states to form each product. CH3OH uptake measurements support that differences in solvation between transition states contribute to regioselectivity changes. This work demonstrates that covalent (i.e., choice of active metal, acid type) and non-covalent (i.e., solvent environment) interactions in zeolite pores drive differences in regioselectivities for epoxide ring-opening. Chapter 5 reveals that the (SiOH)x density of Lewis (Zr-BEA) and Brønsted acid (Al-BEA) zeolites also carry significance for C4H8O ring-opening with CH3OH in CH3CN co-solvent. Ring-opening rates fall within a factor of two over a high density (SiOH)x Al-BEA (Al-BEA-OH) and a low density (SiOH)x material (Al-BEA-F) at all conditions. In contrast, Zr-BEA-OH provides ten times greater rates than Zr-BEA-F. These rate differences do not originate from differences in reaction mechanisms across materials. Zr-BEA-OH consistently yields more positive ∆H^‡ and ∆S^‡ values than Zr-BEA-F, while Al-BEA-OH and Al-BEA-F show nearly identical activation parameters across the range of [CH3OH]. Liquid and vapor intrapore solvent compositions, quantified by multiple methods, give evidence that Zr-BEA-OH adsorbs greater quantities of hydrogen-bonded solvent molecules than Zr-BEA-F. On the other hand, the pores of Al-BEA-OH and Al-BEA-F contain more similar solvent compositions. Examination of these measurements together with measured rates supports that the displacement of hydrogen-bonded solvent molecules leads to greater rates in Zr-BEA-OH through entropic gains that exceed the corresponding enthalpic penalties, as further corroborated by positive correlations between activation and C4H8O adsorption enthalpies. Product regioselectivities depend strongly on the solvent composition and zeolite acid type but weakly on (SiOH)x density. This study further explains how the interplay between covalent and non-covalent interactions within zeolite pores influence epoxide ring-opening processes. Overall, Chapters 4 and 5 demonstrate that both synthetic control of the active metal, acid type, and zeolite (SiOH)x density and selection of the solvent composition provide opportunities to increase epoxide ring-opening rates toward desired product isomers. Chapter 6 summarizes the lessons learned from Chapters 2 – 5 on manipulating covalent and non-covalent interactions to control rates and selectivities of liquid-phase catalysis in zeolites. Chapter 6 also proposes future directions to develop a further understanding of the interactions that drive catalysis at solid-liquid interfaces. These directions include (1) utilizing batch kinetics and in situ 13C-NMR spectroscopy experiments to gain a deeper understanding of the mechanistic origins for rates and regioselectivities differences in ring-opening, using epichlorohydrin (C3H5ClO) as the epoxide reagent. We also propose to examine the effect of (2) organic co-solvent choice (CH3CN, dioxane, dimethoxyethane) and (3) nucleophile choice (alcohol, amine, thiol) on C4H8O ring-opening. These future directions build on the lessons learned in this thesis and seek to establish further design principles for sustainable catalysis in the liquid-phase. The studies outlined here utilize a combination of kinetic, spectroscopic, and thermodynamic techniques to demonstrate the strong role that covalent and non-covalent interactions play in zeolite catalysis and provide guidance on how to manipulate these effects to facilitate desired reaction pathways. While described here in the context of the industrially relevant alkene epoxidation and epoxide ring-opening processes, these principles apply to other relevant liquid and gas-phase chemistries and give a more general understanding of how to control rates and selectivities through catalyst and reaction design.

Graduation Semester

2024-08

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/125509

Copyright and License Information

Copyright 2024 David Samuel Potts

Owning Collections