I am having trouble running MD with a water molecule using the COMB3 potential. I have a Zr substrate and am attempting to deposit a single water molecule onto the Zr surface.
I first thermalise the Zr substrate and then restart a lammps run, using velocities from the thermalisation, with the addition of a water molecule with some velocity towards the substrate (same velocity added to each atom in the molecule). However, on the first step, the water molecule appears to be unstable as it splits and fires the hydrogen out of the system rather than depositing as a single molecule.
Note: I have done separate calculations to determine the ideal angles and distances within the water molecule for the potential.
An example setup is shown below. I have also tried polar_on option and various qeq/comb parameters.
pair_style comb3 polar_off
pair_coeff * * ffield.comb3 Zr H O
fix 5 all qeq/comb 1 0.001