Projected DOS with PAW

[kuollut]

Hello,
I found some inconsistencies in comparing the DOS curves with the sum (on all atoms and on all (s,p,d) angular momenta) of projected DOS. This seem to hold especially when using a PAW basis. In particular, everything looks fine for low-energy orbitals, and the two curve (DOS and sum of PDOS) overlap pretty good. However, at higher energies, the curve with the sum of PDOS seem to be "rescaled" (some time 1/10 smaller) with respect to the DOS curve. It looks like some contribution (of d orbitals?) is missing in the PDOS. I tried with a hands-on example as well (3_7 for CO on NI(111)) and it seems to have the same problem. I tried with different RWIGS values, but it didn't change this behavior. Moreover, I tried with both LORBIT turned either off and on, but it was again the same. Any help or suggestion will be appreciated, best regards.

[oja]

Of course, I don't know what your system is like, but in general you would think that the highest energy states cannot be expressed as simple atomic orbitals. If pDOS is calculated just by projecting the plane wave states on localized orbitals (which do not form a complete basis!), there is bound to be some electron density missing?

[sumy]

hello,I have the same problem now, and how do you solve with this problem? thank you very much!

[sumy]

hello,I have the same problem now, and how do you solve with this problem? thank you very much!