patitial DOS

Message

linglin · #1 Post by **linglin** » Mon May 12, 2008 1:30 am

I wanna to draw the patitial DOS. The electron configuration is 5d 6s. But I can see the p and f states density, which is not low. Can you tell me why?

#2 Post by **admin** » Wed May 14, 2008 10:53 am

sorry, your question is not precise enough: what system do you calculate (are there f-electrons in the valence shell such that the spd splitted DOS shows 16 (32) instead of 9 (18) DOS-columns)?

linglin · #3 Post by **linglin** » Thu May 15, 2008 2:31 pm

Yes, the number of column is 32(including f). But the valence electron configuration of one of the atoms is 5d 6s. So I guess the partitial DOS of the atom in f and p is nearly zero, right? But my result is wrong, not zero.

#4 Post by **admin** » Thu May 15, 2008 3:04 pm

5d6s is the starting configuration of the POTCAR of the atom, your are correct. However, after having reached electronic convergence, it may well be possible that the self-consistent charge density gives small contributions of p and f- like projected charge densities as well. Please note that the l-like local DOS is calculated by projecting within the sphere with r=RWIGS located at a certain site. Due to the choice of RWIGS (pre-defined in POTCAR unless you use LORBIT<10 and give it explicitely in INCAR), it may well be possible that e.g. charge of adjactent atoms is counted to a certain atomic charge as well (eg if the distance between 2 adjacent atoms is smaller than the sum of their atomic radii). This may then yield unexpected l-like contributions (Please note that PAW-vasp does not use localized atomic orbitals, but projceted augmented waves!)