If you are just interested in the energy of the system and this system has Fe atoms, will it matter to set the MAGMOM tag? Will it affect the computed energies or affect the convergence speed if set wrongly? Thanks.
I found some short and nice explanation on spin polarization and magmom setting from this site: http://kitchingroup.cheme.cmu.edu/dft-book/dft.html#sec-2-6
Also, in their example at section 3.8.1.1 tp 3.8.1.2 (http://kitchingroup.cheme.cmu.edu/dft-book/dft.html#sec-3-8-1-1), they showed that adding spin polarization improved the computed dissociation energy of oxygen molecule.
may not，if the magmon value large enough.
http://blog.sciencenet.cn/blog-695784-1066908.html
very detail and easy introduction for setting of magmon，
If you turn on spin-polarization, the magnetic moment will converge to its closest solution. In most cases I've seen different MAGMOM values give the same results, as long as your values are not very far away (say +-1 per atom). In my experience in VASP, sometimes different values in your MAGMOM tag yield different results. My recent calculation on FCC Ni using one-atom unit cell shows that if I set MAGMOM to 0, the final magnetic moment on the Ni atom is around 0 B. When I set MAGMOM to 1 and run with the same settings of the other tags and procedures, the final magnetic moment on the Ni atom is around 0.64 B. The latter one is more realistic compared with the experimental data of FCC Ni.
It is non-trivial to deal with 3d transition metals since their orbitals are quite soft. The structure of the system may change the electronic structures on some lattice sites.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science