Degree of Polymerization of Glucan Chains Shapes the Structure Fluctuations and Melting Thermodynamics of a Cellulose Microfibril

Abstract

A Staggered LATtice (SLAT) model is developed for modeling cellulose microfibrils. The simple representation of molecular packing and interactions employed in SLAT allows simulations of structure fluctuations and phase transition of cellulose microfibrils at sufficiently long and large scales for comparison with experiments. Glucan chains in the microfibril are modeled as connected monomers, each corresponding to a cellobiose subunit, and the surrounding space around the cellulose is composed of solvent cells. Interaction parameters of monomer-monomer interactions were parametrized based on the results of atomistic molecular dynamics simulations. The monomer-solvent interaction was optimized to give a melting temperature of ∼695 K for the 36-glucan chain model cellulose microfibril, which is consistent with the estimation based on experimental data. Monte Carlo simulations of the SLAT model also capture experimentally measured X-ray diffraction patterns of cellulose as a function of temperature, including the region of melting transition, as well as predict the highly flexible regions in the microfibril. Beyond the diameter of ∼3 nm, we found that melting temperature of the cellulose microfibril is not significantly shifted by changing the thickness. On the other hand, a slight decrease in the degree of polymerization of glucan chains is shown to enhance structure fluctuations through the ends of glucan chains, i.e., the defect sites, and thereby significantly reduce the melting temperature. Analysis of the sizes, densities, and lifetimes of defect structures in the microfibril indicates a significant extent of fluctuations on the surfaces even at room temperature and that defect statistics are strong but distinct functions of temperature and solvent quality. The SLAT model is the first of its kind for simulating cellulosic materials, and this work shows that it can be used to incorporate information obtained from atomistic simulations and experimental data to enable the aforementioned findings through computation.