COMB10 potential for quartz and cristobalite

LAMMPS users,

I have been reading the dissertation by Dr. Shan from 2011 to try and understand more of COMB10 potential and how to use it. In his dissertation, he has a table reporting structural properties of quartz and cristobalite (Table 3-1).

I get exactly the same results he reports for BKS and TTAM.

For COMB10, I cannot reproduce the results.

I am using a 0.2 fs timestep.

I am using 5,184 atoms.
I am using fix qeq with error of 1.0e-3
I update self-charge every timestep

Any ideas why I might not get the same values?

Like I said, BKS and TTAM are in agreement which confirms the data file is sound.

I am using NPT ensemble at 1 atm and 300 K and have equilibrated for 100 ps before collecting the data.



Why not contacting Ray directly? Slim chances anyone else here knows his thesis in such detail.