Dear Lammps Users,

I am trying to compute the thermal accommodation coefficient of different gases on Nickel surface. In this type of simulations, the interaction potential coefficient between gas and solid has an important effect on the final results. Therefore, as the ReaxFF force field includes interaction coefficients for different gas on Nickel surface which are computed based on quantum chemistry, I decided to use it. The problem is using ReaxFF my simulations become much slower which I know is normal for ReaxFF. I went through LAMMPS forum to find out how can I accelerate my simulations and I noticed that many people are using ReaxFF in combination with a nonreactive force field. Therefore, I used LJ potential between Nickel atoms as follows and now my simulation is much faster

Dear Lammps Users,

I am trying to compute the thermal accommodation coefficient of different gases on Nickel surface. In this type of simulations, the interaction potential coefficient between gas and solid has an important effect on the final results. Therefore, as the ReaxFF force field includes interaction coefficients for different gas on Nickel surface which are computed based on quantum chemistry, I decided to use it. The problem is using ReaxFF my simulations become much slower which I know is normal for ReaxFF. I went through LAMMPS forum to find out how can I accelerate my simulations and I noticed that many people are using ReaxFF in combination with a nonreactive force field. Therefore, I used LJ potential between Nickel atoms as follows and now my simulation is much faster

###############################

pair_style hybrid reax/c NULL checkqeq no lj/cut 10

pair_coeff * * reax/c ffield.reax.CHONSSiPtZrNiCuCoHeNeArKrXe Ni Kr

pair_coeff 1 1 lj/cut 15.22 2.239

################################################

My only concern is that I read in some discussions in LAMMPS forum that ReaxFF is a stand-alone force field and combining it with other force fields is absolutely not recommended and leads to producing garbage results!

I will be really thankful if someone could tell me if what I am doing here makes any sense or not.

the setup from above makes no sense. you could just as well drop reaxff completely, keep the Ni atoms immobile and use only lj/cut. if you want (e.g.) Ni-Kr interactions described with reaxff you *MUST* use reaxff for *ALL* Ni and Kr atoms, you cannot replace only Ni-Ni with lj/cut since reaxff is a manybody potential. overriding single pairs of atom types is only possible with pair-wise additive potentials.

besides, the choice of a force field should be motivated by the required degree of accuracy, not the speed of the simulation.

axel.