Two questions about the COMB potential

Hi Jun,
Precision is the maximum allowed difference between the electronegativity
(or the derivative of energy w.r.t. charge) of atom i to the mean value.
"Yu, Sinnott and Phillpot, PRB 2007" has detailed description on how comb
does qeq.

Hi, Ray,
Thanks so much for your explanation. I have understood the meaning of “precision” in the command of “fix qeq”.
I have one further question to consult you. This is related your paper: “Atomistic simulation of copper oxidation and Cu/Cu2O interfaces using charge-optimized many body pontentials, PRB, 2011”. The simulation results are very attractive and interesting, especially about the oxidation of Cu surfaces. I want to realize this simulation by myself, but I am not very sure about how to use fix command correctly in this system.
In my simulation, I divide all of the atoms into 4 groups:
(1) g_oxygen: the group of oxygen atoms
(2) g_surface: the group of Cu atoms on the surface of Cu substrate

(3) g_heat: the group of Cu atoms used to control the temperature of Cu substrate
(4) g_fix: the group of Cu atoms used to fix the position of Cu substrate
I update the positions, velocities and charges of atoms using the following 4 fix commands:
fix 1 all nve
fix 2 all qeq/comb 1 0.001

fix 3 g_heat temp/rescale 10 300 300 0.01 1.0
fix 4 g_fix setforce 0.0 0.0 0.0

In this way, due to the “fix 3”, the velocities of Cu atoms belong to g_heat group will be changed after “fix 1” and “fix 2”; due to “fix 4”, the velocities and positions of atoms belong to g_fix group keep invariant.
My question is: Should I also change the charges of the atoms belong to groups of g_heat and g_fix after “fix 1” and “fix 2”? or should the “fix qeq” command be used excluding g_heat and g_fix groups?
I am a little confused. Your help will be highly appreciated.