For a molecular understanding of coal combustion, realistic coal and char models are pivotal. We propose a novel top-down method for the construction of char models, consisting of stacked graphene-like layers with defects. We start on the mesoscale with corrugated stacked surfaces as a template. Then the carbon atoms are pre-positioned on these surfaces and surrounded by hydrogen atoms. In a constrained reactive MD simulation the carbon atoms are then rearranged into a chemically reasonable state with dangling bonds saturated by hydrogen atoms. Importantly, the initial shape of the template is maintained to a great extent, leading to a substantial amount of defects in the graphitic planes, especially for high curvatures. In a final step, small molecular fragments and unfavorable arrangements are removed and remaining dangling bonds are saturated with hydrogen atoms. The central advantage over other approaches is the ability to choose the shape and size of the templating planes, with the possibility to use experimental input.The method is validated on single sheets with a protrusion and application examples are shown for stacked graphitic layers with up to 100 Å size and more than 20000 atoms. It represents the first step of a longer term project. In the future, heteroatoms and functional groups will be implemented, and by combining several such stacks, the approach will allow the construction of more realistic char models using the data from HRTEM images as an experimental boundary condition. On this basis important questions like the adsorption and diffusion of volatiles under high temperatures can be studied on a molecular level.
Read full abstract