Surface reaction of HfCl4 on OH-terminated Si surface(surface is described here with a ||):
||-OH + HfCl4→ ||-O-HfCl3 + HCl
I am using VASP code, when ISPIN=2, structural optimisation shows mag=0 for HfCl4 and mag=28 for OH-terminated Si surface.
If I need to consider spin polarization, any suggestions for the setting of the initial moment (i.e. VASP keyword MAGMOM).
Ultimately, the best answer anyone can give you is to test both and see!
However, spin polarisation is only useful (i.e. not requiring twice the work to give the same answer) in magnetic systems. The easiest way to test this is to produce a spin-polarised DOS; if the two spin channels are identical, you don't need spin polarisation, if they are different then you probably do.
It is entirely possible, however, that a heavy ion such as Hafnium might provide some unexpected/non-intuitive results when you bring the two systems together (spin-orbit coupling is likely to be non-negligible - so probably worth testing too!). This is especially the case if you are performing a transition state search, where the electronic structure of the T.S. may be magnetic even if the reactant and product states are not!
As for initial settings; you can use the moments you have already found as a starting guess (consider whether you have a ferro or anti-ferro alignment present), but it is often worth slightly increasing the initial magmom; it is often easier for the system to "lose" magnetisation than to gain it!
TL;DR : You can't lose accuracy using spin-polarisation, it will just take (approximately) twice as long!
I would also advise using spin polarization for any surface calculation, the dangling bonds present on surfaces can easily result in unpaired electrons even in systems where you wouldn't expect magnetism in the bulk.
- Badges
- Science topic
- Similar topics
- Filing