ENERGY MINIMIZATION: STOPPING CRITERION=FORCES ARE ZERO and 'NAN's in Log file

Hoping I can get help here.

I am trying to simulate calcium carbonate crystal slab in water using reax/c. My input script is attached below. When I try to simulate the slab in water. I get ‘NaN’s in my log file for temperature and total energy and pressure. In trying to diagnose the issue, I simulated water only in a box with the same input and forcefield file and the I had no issues (guess this rule out forcefield file as the culprit). I then run a simulation with the slab alone in vacuum and I see the issue there. So, I am guessing there is something terribly wrong with the initial crystal slab configuration but I don’t know what exactly. I have tried reconstructing it but I get the same error. Is there any special treatment for slabs like that with the pair_style reax/c? Has anyone experienced this?.

Thanks in advance

Samuel Afari

PS: To the best of my abilities, I have read the LAMMPS manual, gone through published papers and searched the mail list thoroughly before seeking help here. Some of the things I have tried have been commented out in data file.

INPUT SCRIPT:

units real

dimension 3

boundary p p p

atom_style charge

read_data data1.dat

pair_style reax/c NULL safezone 1.6 mincap 100

pair_coeff * * ffield.REAX C Ca O

neighbor 2 bin

neigh_modify every 10 delay 0 check yes page 1000000 one 10000

#fix 1 all nve

fix 2 all qeq/reax 1 0.0 10.0 1e-6 reax/c

#fix 3 all temp/berendsen 500.0 500.0 100.0

#delete_atoms overlap 0.1 all all

#fix 1 all box/relax iso 0.0 vmax 0.1

min_style sd

minimize 1.0e-4 1.0e-6 10000 100000

fix 1 all npt temp 100.0 100.0 10.0 iso 1.0 1. 2000.0

timestep 0.1

thermo_style multi #custom step temp epair etotal press ebond eangle edihed

thermo 1

dump 4 all xyz 5000 dumpnpt.xyz

run 100

unfix 1

fix 1 all nvt temp 100.0 100.0 100.0

thermo_style custom step temp epair etotal press

timestep 0.2

fix 7 all reax/c/species 1 100 100 species.out element C Ca O position 100 species.pos

dump 5 all cfg 100 *.cfg mass type xs ys zs

dump_modify 5 element C Ca O

dump 6 all atom 30 dump.lammpstrj

dump 7 all xyz 1000 dump.xyz

#dump 8 all movie 25 movie.mpg type type &

axes yes 0.8 0.02 view 60 -30

#dump_modify 3 pad 3

run 2000

DATA FILE:

LAMMPS data file. CGCMM style. atom_style charge generated by VMD/TopoTools v1.7 on Thu Feb 07 22:11:41 -0700 2019

2400 atoms

3 atom types

-0.005001 49.994999 xlo xhi

-0.154999 49.845001 ylo yhi

-0.005000 29.995000 zlo zhi

Pair Coeffs

image001.jpg

Hoping I can get help here.

I am trying to simulate calcium carbonate crystal slab in water using reax/c. My input script is attached below. When I try to simulate the slab in water. I get ‘NaN’s in my log file for temperature and total energy and pressure. In trying to diagnose the issue, I simulated water only in a box with the same input and forcefield file and the I had no issues (guess this rule out forcefield file as the culprit). I then run a simulation with the slab alone in vacuum and I see the issue there. So, I am guessing there is something terribly wrong with the initial crystal slab configuration but I don’t know what exactly. I have tried reconstructing it but I get the same error. Is there any special treatment for slabs like that with the pair_style reax/c? Has anyone experienced this?.

which LAMMPS version are you using and which specific force field file (you didn’t include it)? has the force field file been parameterized to be able to represent CaCO3?
there is very little that can be said in general. usually, there is something wrong with the input or the parameters, if your pressure turns NaN.

axel.

image001.jpg

What I think that,
After minimization the velocity of the atoms/molecule have turned to zero. And, you have used
“fix 1 all npt temp 100.0 100.0 10.0 iso 1.0 1. 2000.0”
just after that. But you didn’t used velocity command. After minimization I think you have to initiate atom with some velocity.
Please try it with some velocity. I hope you will not get ‘NaN’.
I am not expert on this. So, may be I will be wrong. But, I have just advice on the basis of little work that I have done on lammps and also currently doing.

My lammps version is 4 Jan 2019. I got the forcefield form literature:

See supplementary information of https://pubs.rsc.org/en/content/articlepdf/2014/cp/c4cp02612h.

it was not explicitly parametrized for CaCO3; using it because it seem to contain all the elements for my needs.

To try trouble shooting the problem, I inserted the delete_atom command with overlap cutoff of 2. The command deletes all my carbon atoms and then runs normally. This makes me think the problem is my initial configuration(I may be wrong). Even if this is true, I don’t know how to deal with it i.e build a better configuration. Any suggestions?

Best

Sam

The paper you quote has a parameterization for a zeolite. There is no mention of carbon is the abstract, so it may just be left in there from a previous, different parameterization.

ReaxFF parameters are not transferable on a per element basis. That is a fundamental property of it and hence there are (and have to be) so many different parameter sets.

Also, your delete_atoms command makes little sense. What is the C-O bond length in carbonate? Thus your reasoning is not valid, either.

Axel