Advice how to create new parameter set for REAXFF?

The problem I have right now is the only REAXFF solvated protein parameter set with F was actually created for studying protein interaction with Au. I’m starting to suspect that the parameters for F are nowhere near optimal, small molecules with a sulfonyl fluoride functional group are not behaving at all like i would expect. It looks like the parameter set was optimized solely for Au without much regard to S and F by Monti(2016) Simulation of Gold Functionalization with Cysteine by Reactive Molecular Dynamics.

so im starting with that parameter set C/H/O/N/S/Mg/P/Na/Ti/Cl/F/Au, dropping Au and Ti, and creating some DFT training data with INQ and crosschecking with ADF using the same density functional HYB_GGA_XC_WB97X from libxc to optimize for S / F properties.

my question is: does anyone know of a non-commercial alternative to ParAMS ? we have a license for the Amsterdam Modeling Suite but it’s only available on one cluster which is getting shut down for replacement on dec 7.

i also had a look at van Duin’s website but the pdf manual there is from 2002.

ill have a few more questions shortly …

I know a former Phd student from my lab wrote to van Duin about Reaxff in the recent years. He wrote back. Maybe he’s the best person to reach for that matter.

1 Like

Please note that this is about ReaxFF, not LAMMPS and thus off-topic for any of the LAMMPS categories, so I am closing this thread. Within MatSci.org it is probably most suitable to post questions like this in “Science Talk”.