Hello LAMMPS users,
This may be a more focused question for Ray. I am looking to use the COMB potential for the HfO2 system, and wanted to replicate the data shown in the cited paper from ffield.comb. In “Charge Optimized many body potential for the hafnium/hafnia oxide system” there is a cohesive energy vs volume plot for a variety of crystal structures. I have attempted to replicate this data and have run into multiple issues in doing so.
I started from the example file under comb in.comb.HfO2 and adapted my script as needed. As is shown I am not moving the atoms, only running a fix/qeq (I also tried without this fix and setting the charge to the value in the paper). I am only showing the cubic structure as it is the calcium fluorite structure and is the simplest of the structures (FCC sites for Hf and all tetragonal sites are filled for O)
The other structures I used materialsproject from Lawrence Livermore for fractional coordinates
The minimum energy/volume in the paper is (-30.39 eV, 30.76 A^3)
The minimum energy/volume using my script with a fix/qeq is (-30.35 eV, 27.22 A^3) the fix changed the charge to 3.22 for Hf
The minimum energy/volume using my script with a set charge of 3.44 for Hf (-29.77 eV, 30.78 A^3)
units metal
atom_style charge
dimension 3
boundary p p p
read_data data.fine.cubic.csv.5.025
mass 1 178
mass 2 16.0
pair_style comb
pair_coeff * * ffield.comb Hf O
group type1 type 1
group type2 type 2
timestep 0.0002
compute eng all pe/atom
compute charge1 type1 property/atom q
compute charge2 type2 property/atom q
fix 1 all qeq/comb 1 0.00001 file fq.out
thermo 1
thermo_style custom step temp etotal vol press pxx pyy pzz pxy pxz pyz
dump 1 all custom 1 dump.in.fine.1.cubic.10.* id type x y z c_eng c_charge1 c_charge2 fx fy fz
run 1
Even when going to SiO2 structures, I have run into issues with the energy/vol, only being able to replicate a good energy/vol curve for pure silicon diamond
Thank you,
Jonathan