I am trying to reference 2 different Reax files with this code:
# LAMMPS input script for simulation with Au wall and EMI-BF4
units real # Use real units
atom_style charge # Atom style for ReaxFF, requires charges
boundary p p p # Periodic boundary conditions in all directions
read_data eq_SYST_1000.dat # Reading data file with Au wall and EMI-BF4 molecule
# Define ReaxFF potential
pair_style reax/c NULL # ReaxFF with control file
pair_coeff * * ffield.reax.AuO C H N B F # Atom types for Au and EMI-BF4
pair_coeff * * ffield_Il Au # Atom types for Au and EMI-BF4
# Group definitions
group Au_wall type 1 # Assuming type 1 is for Au
group EMI_BF4 type 2 3 4 5 6 # Assuming types 2-6 are for EMI-BF4
# Velocity settings (if the system initially is at rest, no velocity commands are needed)
velocity Au_wall set 0.0 0.0 0.0 units box # Au wall is stationary
velocity EMI_BF4 set 0.0 0.0 -0.1 units box # Initial velocity for EMI-BF4
# Integration and thermodynamics
fix 1 all nve # NVE ensemble for the entire system
thermo 100 # Output thermo info every 100 steps
thermo_style custom step temp pe ke etotal press # Custom thermo output
# Dump settings for output
dump 1 all custom 100 dump.wall_col id type x y z vx vy vz # Dump atom info
# Running the simulation
run 5000 # Number of timesteps, adjust as needed
# End of simulation
print "Simulation completed"
I am getting an error in pair_coeff * * ffield.reax.AuO C H N B F saying that there are incorrect args for pair coefficients however I am referencing files that have been used prior and have worked correctly.
It is not obvious to me that you can “mix” ReaxFF parameter sets (for example, this capability is not mentioned in the documentation).
Moreover, it is not obvious to me that a ReaxFF treatment is justifiable for an ionic liquid-gold interface – I’m not sure what chemical reactions you’re hoping to observe.
With electrostatics, you mainly need to account for the polarisability of the ions (using either explicit polarisability or scaled charges) and the gold (which can be done with constant potential molecular dynamics, like in the ELECTRODE package); charge transfer between ions, or between ions and gold, is pretty minimal compared to those effects.
ReaxFF parameterizations are very system specific general many-body parameters that cannot be easily mixed and most definitely not treated like pairwise additive potentials where you can mix and match parameters more liberally (whether it makes sense at a physical level is still questionable).
Technically, you could bypass the limitation with using pair style hybrid, but using ReaxFF twice in a hybrid system has multiple ambiguities with respect to performing the charge equilibration. A more approachable setup would be to use EAM for the gold and ReaxFF for the rest, but even then you have multiple problems: a) you need to use pairwise additive potentials (like Lennard-Jones or Morse) to describe the interactions between the atom types of the two subsystems, b) the EAM part is lacking the contribution to the embedding energy from the atoms treated with ReaxFF, c) the atoms handled with ReaxFF don’t “see” the gold atoms and thus the charge equilibration within range of the surface atoms will be incorrect/incomplete.
Thus, having a parameterization for that encompasses your entire system is the most reasonable choice. If that doesn’t exist, it is probably the best approach to collaborate with someone that has the knowledge about how to parameterize ReaxFF.
BTW: using a ReaxFF parameter file called ffield.reax.AuO on the elements C H N B F looks highly suspicious. Are you sure those parameters are suitable for your molecules?