Hello,
I am simulating Kapton sputtering by atomic oxygen. I want to describe the Kapton and atomic oxygen interaction with the ReaxFF potential and the desorbed gases after reaction with a classic potential that does not allow further chemical reactions such as lj/cut.
My idea was to have 8 types of atoms. 1 to 4 are respectively the atoms : C, H, O, N, and 5 to 8 are respectively C, H, O, N. The idea is that types 1 to 4 are controled by ReaxFF and types 5 to 8 are controlled by Lennard-Jones. In a loop inside the bombardment part of the code I change the type of the atoms that are in the upper part of the supercell. 1 becomes 5, 2 becomes 6 etc…
To do so I use the following commands to define the pair_style :
pair_style hybrid reaxff/kk lmp_control lj/cut 10.0
pair_coeff 5*8 5*8 lj/cut 1.0 1.0 2.5
pair_coeff * * "/stck/dfernand/ATOX/2-LAMMPS/0-ReactiveForceFields/CHONSSiNaFZr_Rahnamoun_van-Duin.ff.txt" C H O N NULL NULL NULL NULL
But LAMMPS gives me the following error :
pair_style hybrid reaxff/kk lmp_control lj/cut 10.0
pair_coeff 5*8 5*8 lj/cut 1.0 1.0 2.5
pair_coeff * * "/stck/dfernand/ATOX/2-LAMMPS/0-ReactiveForceFields/CHONSSiNaFZr_Rahnamoun_van-Duin.ff.txt" C H O N NULL NULL NULL NULL
ERROR: Expected hybrid sub-style instead of /stck/dfernand/ATOX/2-LAMMPS/0-ReactiveForceFields/CHONSSiNaFZr_Rahnamoun_van-Duin.ff.txt in pair_coeff command (src/pair_hybrid.cpp:530)
Last command: pair_coeff * * "/stck/dfernand/ATOX/2-LAMMPS/0-ReactiveForceFields/CHONSSiNaFZr_Rahnamoun_van-Duin.ff.txt" C H O N NULL NULL NULL NULL
Does any of you know how to solve this please ? I would also like to have your opinion about my idea of having 8 types of atoms. I still don’t know how tot tell LAMMPS that atom types 5 to 8 are C, H, O, N.
There is a syntax error in your pair_coeff statement and the error message already tells you where the problem is.
Furthermore, your setup is a bogus model. ReaxFF is not well suited for mixing with other potentials. You cannot apply the required charge equilibration to the non-ReaxFF part, but having static charges there creates a bogus polarization of the ReaxFF atoms and bad energy conservation. The same phenomenon well known from QM/MM simulations and requires very sophisticated electrostatic embedding of the more complex calculation into the simpler one.
Classical molecular force fields don’t work that way. There is a simple introduction in https://doi.org/10.1021/acs.jpcb.3c08419 (preprint at: Type Label Framework for Bonded Force Fields in LAMMPS | Theoretical and Computational Chemistry | ChemRxiv | Cambridge Open Engage). You have to assign types based on the topology and chemical nature of the molecules, not only based on element.
P.S.: you got lucky that the Materials Project maintainer has been recategorizing your posts recently. When will you learn to post in the LAMMPS categories?? Otherwise, your posts will go unanswered since they are off-topic there.
There is a syntax error in your pair_coeff statement and the error message already tells you where the problem is.
I know it is telling me the error, what I find difficult is how to write it in the hybid sub-style. It is probably written somewhere but I did not find it, either in the pair_coeff, pair_style or ReaxFF explanations in LAMMPS site.
Classical molecular force fields don’t work that way. There is a simple introduction in https://doi.org/10.1021/acs.jpcb.3c08419 (preprint at: Type Label Framework for Bonded Force Fields in LAMMPS | Theoretical and Computational Chemistry | ChemRxiv | Cambridge Open Engage). You have to assign types based on the topology and chemical nature of the molecules, not only based on element.
Thanks for your help I understand that my idea will not work. The only thing i want to do is to prevent the desorbed gases to interact with Kapton after being reflected in the upper part of the simulation cell, So I will just try to find a way to delete these atoms if it is possible to do so in lammps.
P.S.: you got lucky that the Materials Project maintainer has been recategorizing your posts recently. When will you learn to post in the LAMMPS categories?? Otherwise, your posts will go unanswered since they are off-topic there.
Thank you for your help. I really appreciate the time you dedicate to supporting the community. I understand it can be frustrating to correct repeated mistakes, but I believe there are more constructive ways to provide feedback. This applies especially to many responses from akholmey, which can come across as harsh. Many users here, including professionals, aren’t experts in LAMMPS or matsci.org, so of course we will do some basic errors. I apologize for my mistake and will be more careful in the future.