Charge scaling as a treatment to effectively include the intermolecular charge transfer and polarization effects is widely applied in molecular modelling of ionic solvents. While the charge scaling factor of 0.8 is picked as a generally applicable solution usable in many computational works, in this work based on extremely precise molecular simulations of bulk solvents and nonequilibrium Hamiltonian switching, we highlight the fact that the general 0.8 selection of the charge scaling factor could be a solution specific to the room-temperature condition that most ionic-liquids studies perform. When the environmental condition varies, this ‘general’ recommendation could sometimes lead to huge deviations. The situation is further complicated by the disparate responses of different properties (bulk density and solvation free energy) in the charge scaling space and along the temperature ladder. The driving force triggering the disparate responses could be attributed to the difference between interaction components involved in different thermodynamic properties, i.e., solvent–solvent interactions for bulk density and the additional solute–solvent interaction in the solvation process. A caution hinted by these observations is the necessity of condition- and property-specific parameter adjustment in practical situations.
Read full abstract