Dear lammps users
Is it make a difference if i use Lennard-Jones Coeff ,for example, “Fe” (iron ion) with charge +2 instead of “Fe” with charge +2.5 ?
And also, Can i use Lennard-Jones Coeff of “Oxygen” in SiO2 in “Fe2O3” structure ? in both the charge of Oxygen is “-2” !
Dear lammps usersIs it make a difference if i use Lennard-Jones Coeff ,for example, “Fe” (iron ion) with charge +2 instead of “Fe” with charge +2.5 ?
yes.
And also, Can i use Lennard-Jones Coeff of “Oxygen” in SiO2 in “Fe2O3” structure ? in both the charge of Oxygen is “-2” !
no.
Dear Axel
I want to be more specific.
I need “Fe” ion Lennard parameters which the charge is +2 and +3. But i found parameters for +2.1 and +2.5. Can i use Lennard parameters of those +2.1 and +2.5 and then specify charge +2 and +3 in my data file?
I have searched a lot, but i did not find Lennard parameter for Oxygen for Fe2O3 or Fe3O4 when the charge is -2. is there another way? do you have any idea about that?
a) i don’t think that Lennard-Jones plus fixed charges is a good model for either compound
b) you likely need different parameters for the Fe in Fe2O3 than for Fe3O4, same for the oxygen
c) because you are doing a classical model, force field parameters are effective parameters and thus may be different for the same element in different environments. therefore individual charges are not always the formal values, only the net charge for a functional group is usually preserved
d) there are force fields and models specifically parameterized for oxides. many of those tend to use charge equilibration to have the charges for individual atoms adjust to the environment, which is particularly important for surfaces and grain boundaries
this looks to me like you need to do some more research on how to properly model your compounds and also gain some more insight into the parameterization and limitations of classical potentials.
axel.
For the case of Fe2O3 (hematite) there is a set of parameters reported by Vaari (https://doi.org/10.1016/j.ssi.2014.11.027) where the interactions are modeled by a rigid ion Buckingham potential but not Lennard-Jones. You can find the implementation in openkim at https://openkim.org/id/Sim_LAMMPS_Buckingham_Vaari_2015_FeO__SM_672759489721_000 which is fully compatible with LAMMPS.
Maybe it is useful in your research
Regards
Evangelos Voyiatzis
Στις Παρ, 23 Ιουλ 2021 στις 2:15 μ.μ., ο/η Axel Kohlmeyer <[email protected]> έγραψε: