PROJECTED DOS

[abc]

Hi,

this is probably a dumb question.
I wish to calculate Partial DOS for a ternary system (ie 3 chemical species).
Although I use PAW potentials I wish to calculate the PDOS using the RWIGS values for each atomic species.

I can check the volume of each species type in OUTCAR after the run:
ie OUTCAR gives

volume of typ 1: 139.7 %
volume of typ 2: 91.2 %
volume of typ 3: 0.3 %

Should I adjust the RWIGS values until volume of typ 1+ type 2+type 3 = 100%
If so, there is no unique way to do this so what is the best way to choose the set of RWIGS values?

[forsdan]

The RWIGS values should be choosen so that the total volume will be as large as possible, but with the condition that the spheres don't overlap each other. The covalent radii can be a hint as a starting point but there is no unique way of doing this and the optimal settings (if such a concept can be defined) will differ between different geometries.

Best regards,
/Dan

<span class='smallblacktext'>[ Edited Wed Feb 06 2008, 11:06PM ]</span>

[abc]

Thanks for your reply,

I have another question. How do you check whether spheres are overlapping?
Is this just the condition that the sum of the volumes of each species type cannot be greater than the cell volume?
(ie in OUTCAR file check that vol of type 1+vol of type 2 + ... < 100% )
OR is there some other way of checking this?

Thanks for any help you can provide.

[admin]

1) you will never be able to fill any volume completely with spheres. if you rescale RWIGS such that it fills 100% of space, the spheres will overlap and a substantial part of the cell volume will be counted (at least) twice.
2) add up RWIGS of NNs and check whether this sum is larger than the distance between the ions.