Hello, everyone,
recently, I did the Fe3O4/bcc-Fe interface structure optimizing, and selected different interface matching relationships. I set "ISIF=2, ISPIN=2" in INCAR file, also the initial MAGMOM values for different atoms. After structure relaxation, the magmom values of some Fe atoms are reversed. For example, mag= -4 changed to mag= 1, which was unreasonable. I was wondering about the results. Can I consider the interface structure is unstable in this case?
Thanks a lot!
Dear Dr. Ivanova,
thanks for your reply. What does " you should use however DFTs" mean? if there are some other methods identify the interface structure' stability? Or I can confirm it by the current results?
thanks a lot
Dear Zhang,
do you mind expanding what do you mean by "instability" and why the change in magnetisation is unreasonable? You have a surface therefore some of the atom coordinations have changed. Also I would try to look at the atomic positions to see if there has been any reconstruction during the relaxation process. Finally, plotting the electron density can give you an idea of the change in magnetisation.
In conclusion, we need some extra pieces of information before discarding the calculations you have performed. Assuming you have checked convergence with respect your parameters.
Best regards,
Roberto
Dear Roberto,
thanks for your reply. Firstly, there are no vacuum layers in my structure, then I think that during the relaxation, the atoms position is changing till the energy and the force reach the precision. Also whether setting the "ISPIN" affects the results obviously sometimes. And the well-preconditioned initial "MAGMOM value" can help find the minimum energy configuration. If the magmom changed unreasonable, the configuration maybe relaxed far from the initial configuration. In this case, how can we say the initial structure is relative stable? Maybe there're some other methods to identify the stability. What's about "extra information"?
Hope to get your suggestions.
thanks a lot!
Honglin
Dear Honglin,
What I do fail to understand to what you refer saying "instability". Your initial starting point is assigning atomic positions (in a periodic lattice) and then impose some properties on the electrons (for example the magnetisation). You let the system relax (I assume both atoms and electrons) and observe the final result. To say that somewhat you initial configuration is unstable you should see that atoms are moving way away from their initial positions. For the electrons, it is more difficult to say, since you cannot know the exact details of the initial state, or what if indeed the final state, without performing an accurate analysis of all the electronic densities. One thing you can try is to consider the total energy, and see if the states that you consider unstable are high in energy with respect to the other "more physical" state. Indeed, normally electron and atomic relaxation are based upon local minimisation techniques, so they cannot really help in finding the global (or reasonably stable) minima.
Best regards,
Roberto
Dear Roberto,
it really needs some other evidence to confirm the interface's stability, like what you said from the case of total energy. Or from the phonon calculation can final confirm it, but really time-consuming. I then carefully consider it.
thank you!
Honglin
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs