The scalable production and dispersion of 2D materials, like graphene, is critical to enable their use in commercial applications. While liquid exfoliation is commonly used, solvents such as N-methyl-pyrrolidone (NMP) are toxic and difficult to scale up. However, the search for alternative solvents is hindered by the intimidating size of the chemical space. Here, we present a computational pipeline informing the identification of effective exfoliation agents. Classical molecular dynamics simulations provide statistical sampling of interactions, enabling the identification of key molecular descriptors for a successful solvent. The statistically representative configurations from these simulations, studied with quantum mechanical calculations, allow us to gain insights onto the chemophysical interactions at the surface-solvent interface. As an exemplar, through this pipeline we identify a potential graphene exfoliation agent 2-pyrrolidone and experimentally demonstrate it to be as effective as NMP. Our workflow can be generalized to any 2D material and solvent system, enabling the screening of a wide range of compounds and solvents to identify safer and cheaper means of producing dispersions.
Keywords: 2D materials; exfoliation; graphene; molecular modeling; solvent prediction.