Force field benchmark of asphalt materials: Density, viscosity, glass transition temperature, diffusion coefficient, cohesive energy density and molecular structures

The molecular dynamics research method is helpful to improve the understanding and prediction of various microscopic behavior mechanisms of asphalt binders. However, there are few studies on the force field benchmark of asphalt materials, and the selection of force fields is often empirical. A smaller number of atoms would cause the predicted asphalt properties to become more unstable and have greater errors. In this study, we evaluate the ability of different force fields and model sizes to predict the density, viscosity, glass transition temperature, diffusion coefficient, and cohesive energy density of asphalt systems. In addition, the suitability of the GAFF, CHARMM, GROMOS and OPLS force fields for describing asphalt molecule systems was described by comparing minimized structures of asphalt molecules in the force field with the results of quantum chemistry calculations. The results show that different force fields have their own advantages and disadvantages in predicting different properties. In general, the GAFF force field is the worst force field for predicting the properties of asphalt condensed phase systems, and other force fields have their own advantages and disadvantages for predicting the different properties of asphalt condensed phases. Conversely, the GAFF and CHARMM force fields are the best to describe asphalt molecules. These conclusions help to promote more accurate selection of molecular dynamics force fields in different asphalt research scenarios.