I would like to run spin polarized calculation using quantum espresso. Is it important to add spin polarization in calculating the wavefunction cutoff, K point convergence and geometrical optimization?
Or I can run non polarized calculation for geometrical optimization and add spin polarization in the non-SCF calculation. and also how important the starting magnetization value?
I would like to run a spin-polarized calculation using quantum espresso. Is it important to add spin polarization in calculating the wavefunction cutoff, K point convergence and geometrical optimization?
Ideally, you would start off the convergence tests with varying K-points, wave cutoffs ..etc , with the spin polarization included, i,e ; nspin =2.
But if you're short on time you could start off with the high k value, cutoff and then add spin polarization and fine-tune it. Spin-polarized calculations are in fact more expensive in terms of computational power than normal SCF cycles.
Or I can run non polarized calculation for geometrical optimization and add spin polarization in the non-SCF calculation. and also how important the starting magnetization value?
For geometrical optimization too you need to include spin polarization since the energy tends to be different, you could see that when you do an SCF.
The starting magnetization indicates what the word suggests, for a ferromagnetic substance you would like to start every ion with the same magnetization between -1 to +1, but for Anti- Ferro Magnetic materials you would include two magnetizations (spin up /+1) and spin down (-1) for the corresponding ions.
This can be specified by adding a fake element for example Element_number and then assign a magentization using starting_magnetization(index of that element)
Hope this helps :)
thanks for your answer Anoop. I have few other questions
1. How reliable the results if spin polarized calculations performed on one primitive cell? or one need to consider super cell model? ( In my case I have 2 magnetic ions per one unit cell)
2. what does different band gaps for spin up and spin down signify? the conduction band is almost same in both cases, but the top of the valence band is different, the spin up case has more states there by having a lesser gap.( I found the other way in the literature, valence band will be pretty much same and conduction band edge will be different for spin up&down ).
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running