Probing the behaviour of lignin derived monomers in catalytic zeolite systems

Abstract

The fundamental dynamical behaviour of lignin pyrolysis derivatives containing aromatic, methyl, hydroxyl, and methoxy functionalities has been investigated, both in their pure forms and adsorbed within three commercial Brønsted acidic zeolite structures. Quasielastic neutron scattering (QENS) experiments combined with Molecular Dynamics (MD) simulations and inelastic neutron scattering (INS) with density functional theory (DFT) calculations were employed to analyse diffusion and adsorption processes, respectively. The study aims to understand how adsorbate-acid site and adsorbate-adsorbate interactions and steric pore hindrance affect local/nanoscale mobility and the nature and strength of such adsorption, to optimise zeolites for the conversion of lignocellulosic biomass feedstocks and for developing accurate computational models to characterise such processes.

Chapter 3 explores the dynamics of the model lignin derivative p-cresol using QENS and MD simulations, which revealed isotropic rotation and jump-diffusion. Two MD force fields were evaluated for their ability to replicate experimental observations, with the OPLS2005 force field providing a better match due to reduced molecular polarity, but still overrepresented hydrogen bonding.

Chapter 4 extends this analysis to a broader range of lignin derivatives, showing that rotation and jump-diffusion rates decrease in the order: anisole > guaiacol ≈ o-cresol > p-cresol ≈ m-cresol, primarily controlled by hydrogen bonding interactions. Simulations accurately reproduced QENS observables for anisole and guaiacol where no/less hydrogen bonding is present. However, simulations of the cresol isomers showed slower dynamics due to increased hydrogen bonding. By mapping the MD output onto the experimental space, we highlight crucial issues in extracting accurate coefficients for both diffusional and rotational dynamics from commonly applied analytical models of QENS data. However, experimentally verified MD models can calculate the true self and rotational diffusion coefficients.

Chapter 5 focuses on the behaviour of p- and m-cresol in H-Y and H-Beta. Only isotropic rotation was observed on the timescale of the QENS instrument (∼54 ps) with confinement to the zeolite micropores and adsorption to acid sites slowing diffusion. QENS observed a larger population of rotationally mobile p-cresol in each catalyst due to its more linear shape, and the larger pores of H-Y allowed for greater mobility of both isomers. An increased rotationally mobile cresol population from H-Beta to H-Y correlated with a decreased rotational rate due to an increase in adsorbate-adsorbate interactions. The MD simulations reproduced the motions and gave further insight into a rapid rattling motion that occurred when bonded to acid sites. Accessing longer timescales in simulations showed extremely restricted diffusion and higher activation energies, but the same trends were observed with pore topology and molecular shape. Relatively fast diffusion was observed for p-cresol in H-Beta due to the longer axis of the molecule inhibiting favourable 180° angled hydrogen bonding to zeolite acid sites. Due to the agreement between experiment and simulation, we modelled dynamics at a higher, catalytically relevant temperature. The study included H-ZSM5 in which diffusion was significantly reduced in its smaller channels. This is attributed to a lower zeolite pore size-to-molecular size ratio rather than increased hydrogen bonding interactions. The relative diffusion rates within the simulations directly correlate with cresol conversion rates before coke formation, suggesting that diffusion is a limiting factor.

In Chapter 6, the dynamics of anisole and guaiacol in H-Y and H-Beta were examined but over longer timescales (up to 340 ps), accessing localised jump-diffusion dynamics along with probing methyl rotations. The proportion of diffusing molecules, diffusion rates, jump distances and the confining radius of diffusion increased from H-Beta to H-Y and as molecular size decreased, or with the absence of a hydroxyl group in anisole. Again, MD simulations reproduced the experiment. When extending the simulations to the nanoscale, unconfined diffusion happened more quickly for anisole in the straight H-Beta channels, compared to more hindered diffusion through smaller apertures between H-Y supercages, which provide a barrier to diffusion beyond the confining region. Guaiacol's long-range diffusion was significantly hindered by hydrogen bonding with acid sites. The QENS/MD studies of zeolite systems highlight the complex interplay between molecular shape, functionality, steric pore hindrance, and acid site interactions, and also emphasise the dependence of diffusion rates obtained on the timescales probed.

Finally, Chapter 7 investigates the adsorption of the derivatives within zeolites (H-Y, H-Beta, H-ZSM5) using INS, where vibrational modes for each sample were assigned by DFT calculations. Specific peaks attributed to molecular bonding group vibrations shifted to lower frequencies and more so for H-ZSM5 compared to H-Y, correlating with decreased adsorbate-adsorbate interactions in smaller pored zeolites, allowing for less restricted vibrational modes. Anisole demonstrates changes in the vibrations of the methoxy group bend upon bonding with zeolite acid sites, due to the lack of strong intermolecular interactions present in the solid anisole samples, compared with the other lignin molecules which can form hydrogen bonds through their hydroxyl groups. Larger changes observed with H-ZSM5 were attributed to stronger bonding, as well as greater molecular separation and higher acid site density encouraging greater frequencies of anisole-acid site bonding. Calculation of adsorption energies from DFT observed stronger adsorption in smaller pored zeolites due to increased stabilising interactions with the pore walls but variations between lignin derivatives were strongly dependent on bonding geometries, influenced by the functional groups present and the acid site location.

Date of Award	22 Jan 2025
Original language	English
Awarding Institution	University of Bath
Supervisor	Alexander O'Malley (Supervisor), Jeff Armstrong (Supervisor) & Asel Sartbaeva (Supervisor)