Solid state materials combustion

Sorry for such a simple question, but I am having trouble in simulating the decomposition/combustion/explosion of some materials, starting in solid state.

The structure of the crystal/molecules is totally changed in the first iteraction.

I am using ReaxFF and building the data file using VMD, after creating the molecule file using avogadro and then creating the crystal structure using packmol (basically I used the crystal lattice parameters and inserted a certain amount of molecules to achieve the near-experimental density). Maybe packmol is creating a high energy distribution?

I even tried to use the fix rigid command, but it keeps the material from decomposing over time.

Thanks for any suggestions.

Sorry for such a simple question, but I am having trouble in simulating the
decomposition/combustion/explosion of some materials, starting in solid
state.

The structure of the crystal/molecules is totally changed in the first
iteraction.

I am using ReaxFF and building the data file using VMD, after creating the
molecule file using avogadro and then creating the crystal structure using
packmol (basically I used the crystal lattice parameters and inserted a
certain amount of molecules to achieve the near-experimental density). Maybe
packmol is creating a high energy distribution?

if your system has a high potential energy, you should see it in your
thermodynamic output.
if that is high, you could run a minimization (as it standard
procedure in MD) as initial part of your equilibration.

another common beginner's mistake with reaxff, is to generate a data
file with bonds/angles/dihedrals etc. those are computed dynamically
and providing them explicitly is an error, as it will change the force
field interactions.

I even tried to use the fix rigid command, but it keeps the material from
decomposing over time.

finally, you have to carefully check the force field parameters. you
need a set that is specifically parameterized for your kind of system.
reaxff is very specific and the resulting parameter sets are not very
transferable.

axel.