Question to Aidan on Tersoff potential

Hi Aidan,

I have a general question on the Tersoff potential implemented in LAMMPS. If we go through Tersoff paper, fc parameter was calibrated as Sin function for single component system. For multi component system he revised it with Cos function in the paper. In all the paper that I studied, people used Cos function, and I can sent you several paper on it. But, even for multi component system, i.e. SiC or SiCG or etc., you reported using Tersoff-1. To the best of my knowledge it shouldn’t be correct or they should be equivalent, I’m not sure now. Moreover, in the paper people usually report [ Si Si Si ] and [ C C C ] parameters. But, for a two component system, we need 6 other rows, i.e. [ Si Si C ] … [ C C Si ]. How should I fill the parameter for them ? Many people talked about it in the LAMMPS Community. But, at the end I couldn’t find a general rule. I really think it could be a general issue, and it really need to be clarified in the manual directly. I also gone through the Tersoff-2 code that you implemented in LAMMPS. If you double check it you can find you have calibrated it in a way that there is no need to change R and D of the input parameter files for Tersoff-1 and Tersoff-2.

P.S. I Cc you while I send the Email to the LAMMPS community.

Eagerly looking forward to hear back from you,