Multiscale molecular behaviour in the zeolite catalysed methanol-to-hydrocarbons process

Student thesis: Doctoral ThesisPhD

Abstract

Classical molecular dynamics simulations and quasi elastic neutron scattering were used to investigate the dynamics of methanol confined within microporous zeolite catalysts, with the aim of understanding fundamental processes underpinning industrially significant reactions such as the methanol-to-hydrocarbons process. Chapter 3 explores the dynamics of methanol in H-ZSM-5 catalysts over varying loadings and Si/Al ratios. The findings indicated that lower methanol loadings promoted higher diffusivity and that self-diffusion coefficients increased with the Si/Al ratio up to a threshold, beyond which further compositional changes had little effect. Detailed analyses on nanosecond timescales, mirroring neutron spin echo experiments, further revealed that both methanol–catalyst and methanol–methanol interactions were crucial in governing diffusion. In Chapter 4, the study extended this analysis by exploring how zeolite pore size and framework topology affected methanol translational and rotational dynamics in CHA,MFI, BEA, and FAU frameworks at different loadings. While diffusion in CHA was negligible on the nanosecond timescale, the MFI, BEA, and FAU systems exhibited a wide, loading-dependent range of diffusion coefficients, with structural differences leading to distinct mobility trends. Additionally, rotational dynamics were generally consistent across larger pore systems, although CHA displayed higher rotation rates at low loadings due to isolated molecular environments. In Chapter 5, quasi elastic neutron scattering experiments combined with molecular dynamics simulations were employed to probe nanoscale methanol diffusion in CHA type zeolites, SSZ-13 and SAPO-34. Analysis of the elastic incoherent structure factor revealed that methanol motion was confined within individual cages, with differences in confined radii and hydrogen bonding interactions influencing localised confined diffusion coefficients. Although the simulations captured the experimental trends providing qualitative agreement between simulation and experiment, they overestimated diffusion coefficients by an order of magnitude underscoring the limitations of idealised models.
Date of Award10 Dec 2025
Original languageEnglish
Awarding Institution
  • University of Bath
SupervisorAlexander O'Malley (Supervisor) & Steve Parker (Supervisor)

Cite this

'