Hi Dr.Shan,

I am trying to reproduce the lattice constant of α-quartz in your 2010 comb potential for si/sio2 paper. I have two questions that I hope you could answer.

My simulation has a timestep of 0.2 fs. The charge is updated every 10 steps with convergence criteria of 0.001. The system has 2250 atoms. The main inputscipt is below:

velocity all create 300 123 dist gaussian mom yes rot yes

fix fix1 all npt temp 300 300 0.02 aniso 0 0 0.02 tchain 10 pchain 10 nreset 100 tloop 5 ploop 5 mtk yes

fix fix2 all qeq/comb 10 0.001

In your paper, α-quartz has a0 = 4.856Å, c0 = 5.316Å, which is in table 2 of the paper. However, my result is a0 = 4.791Å, c0 = 5.304Å. I wonder what caused the difference. I tried with larger system (18000 atoms) and different charge update frequency (n = 1). The change of lattice constant is very small.

The second question is the lattice constant at 0K is even larger than that at finite temperature (a = 4.915). The scipt at 0k is blow.

fix fix1 all box/relax aniso 0 nreset 1

fix fix2 all qeq/comb 10 0.001

min_style cg

minimize 1.0e-25 1.0e-25 1000000 10000000

I would appreciate if you could point out where I did wrong.

Regards,

Wenting