Investigation of model scale effects on coal pyrolysis using ReaxFF MD simulation

ReaxFF MD is a promising method for exploring complex chemical reactions, allowing large coal molecular systems simulated at high temperature. Due to the amorphous and diverse nature of coal chemical structure, model scale effect on the simulation of coal pyrolysis can be a big issue, which was investigated by comparing heat-up ReaxFF MD simulations of three Liulin coal models with 2338, 13,498 and 98,900 atoms. ReaxFF MD simulation results show its consistency in the observed trends of weight loss profiles and product generations lumped by C number for different model scales. The small-scale coal model facilitates observation of reaction sites, but can hardly access reasonable evolving trends of products and reactions. ReaxFF MD simulations of the coal model in middle scale can obtain much better evolving trends of pyrolyzates and reproduce some reaction pathways. But certain fluctuations or randomness of reactions still exist in the evolution trends of representative products. The diversity of chemical reactions in coal pyrolysis can be much more accessible by employing the large-scale coal model, thus permitting distinguishable evolving trends in pyrolyzates.