Building a reactive molecular dynamics framework for cellulose pyrolysis studies

Much of the current interest in cellulose pyrolysis stems from technologies that enable the conversion of lignocellulosic biomass into commodity chemicals. Describing these reaction mechanisms is one of the fundamental associated challenges. Computational chemistry methods can complement the experimental knowledge. In our work, we used Reactive Molecular Dynamics (RMD) in conjunction with the ReaxFF force field, enabling formation and breaking of chemical bonds within a classical MD framework with an affordable computational cost. We carried out stochastic RMD simulations to study the high-temperature decomposition of an isolated cellulose molecule [1]. We conducted a total of 16900 simulations for chains with a degree of polymerization (DP) between 8 and 64, for several initial conformations, and in the temperature range of 1400 K to 2200 K. Each simulation was run until the first decomposition event was detected. From this data, the reaction rate constant could be obtained. We observed the decomposition to occur primarily through random cleavage of the glycosidic bonds. An activation energy of (171 ± 2) kJ mol-1 and a frequency factor of (1.07 ± 0.12) * 1015 s-1 were determined for this reaction. These values are within the range of values reported for the global mass loss kinetics of cellulose pyrolysis, and showed no dependence on the DP and on the initial conformation. We also observed the release of some of the characteristic low-molecular-weight products, such as glycolaldehyde, water, formaldehyde, and formic acid. Focusing on isolated molecules contributes to a bottom-up approach, in which we build towards condensed-phase pyrolysis simulations. The next steps may include studying secondary and tertiary decomposition reactions for an isolated chain, adding more chains to the system to include intra-chain interactions, extending the temperature range, and improving the bond information analysis. With these developments, the stochastic RMD approach could be used for detailed studies of cellulose pyrolysis, and used in the design, development and optimization of pyrolytic conversion processes.