how to calculate electronic occupation of certain orbital in VASP?

Message

linglin · #1 Post by **linglin** » Mon Aug 31, 2009 5:32 am

Dear Sir,

I'm wondering how to calculate the occupation number of certain orbital in VASP, for example, the occupation number of d orbital for a metal.

regards

sincerely,
tom
08/31/09

alex · #2 Post by **alex** » Mon Aug 31, 2009 9:03 am

bulk metal or metal atom?

linglin · #3 Post by **linglin** » Mon Aug 31, 2009 9:39 am

dear alex, bulk metal

alex · #4 Post by **alex** » Mon Aug 31, 2009 4:06 pm

There are no orbitals left. It's bands in solids.

linglin · #5 Post by **linglin** » Tue Sep 01, 2009 12:20 am

right, i mean d-band occupation, is there any way to do the calculation?

panda · #6 Post by **panda** » Thu Sep 03, 2009 3:12 pm

read the user manual regarding use of LORBIT tag and analysis of the PROCAR file. You can calculate partial occupancies after performing a SCF calculation, there is a section in Chapter 7 of the manual outlining how partial occupancies are calculated and Ch. 9 outlining how to set up the calculation. Once you have the partial occupancy of the bands you can count up the orbitals to give the number of d orbitals. The PROCAR even nicely breaks down the contribution of each ion corresponding to the occupation, so in say a binary alloy of TiAl you could obtain the contribution of Ti d-electrons and Al-d electrons to the overall d-band occupation.

boris · #7 Post by **boris** » Mon Sep 14, 2009 4:30 pm

Dear Tom
You can use a trick: you activate the LDA+U formalism for the d-orbitals and you set U and J values to zero. This will be like a GGA calculation, and by setting LDAUPRINT to 1, you'll be given the occupation matrices of the d-orbitals.

Regards