About nvt system explosion

[111323]

Dear all experts

I am trying to make a minimize and nvt to let the model to balance at 300k.

but when I try to do that ,initially the model is half of circle

the final is all explosion and fill out all simulation

I want that final is the same as initial that half of circle and the same place.

Is that my potential file wrong or another problem with this ?

thanks for helping me!!

below is my in

3d friction simulation

dimension 3
boundary p p p

atom_style charge
neighbor 0.3 bin
neigh_modify every 1 delay 0 check yes

create geometry

read_data SiO2sphere.lmp

pair_style tersoff
pair_coeff * * SiO.tersoff O Si

min_style cg
minimize 1.0e-12 1.0e-12 1000 1000

intial velocities

compute 1 all pe/atom
thermo 10
thermo_style custom step temp pe etotal dt time
dump 1 all custom 100 W.xyz id type x y z

velocity all create 300 666 mom no rot no dist gaussian
fix 1 all nvt temp 300 300 0.05
timestep 0.001
run 5000

[Will_Pisani]

I don’t have any experience with the Tersoff potential, but I do have some experience with ReaxFF. If this is your initial structure, you may need to warm it up to 300 K slowly rather than start at 300 K. Start at 0.1 K (or something else really low) and ramp to 300 K over 100 ps or so.

[akohlmey]

Please verify that the potential parameters you are using are suitable for non-bulk systems. I would suspect they are not, since the physics of oxide surfaces are much more complicated than that of semiconductor surfaces (e.g. pure Si) and I would be surprised that the Tersoff formalism is suitable for that. You would need something more complex like COMB (LAMMPS has the original COMB and COMB3 included) for that.

Axel.