My Community

I am working with cluster and trying to know local magnetic moment of each atom!

is there any method (other then trial n error) to set RWIGS?B)

else,is there any other way to calculate magnetization (without using RWIGS?)

i am using PAW

if you use the PAW method, you can
1) either use LORBIT>=10 and the pre-defined projection operators (and RWIGS) of the POTCAR file
2) or set LORBIT < 10 and set RWIGS explicitely in INCAR according to some appropriate criterion: the spheres of NN atoms of course should be non-overlapping, furthermore you can choose the ratios of the RWIGS according to the ratio of the atomic (or ionic) radii
3) the total magnetic moment of the cell (independent on any local quantities) is given in the last line of OSZICAR

Either you can recalculate everything with N different RWIGS values, or you can calculate the values by hand from your CHGCAR. I'd go for the latter, i.e. integrate over a volume you think corresponds to an atom, and you get a magnetic moment. Or even better, treat the integration volume as a parameter and play around until you get the value you wanted

Bear in mind that whatever 'atomic' magnetic moments you get are arbitrary. The absolute values have no meaning. At the end of the day, it doesn't really matter what RWIGS you use, because all you can do is compare the relative magnitudes of magnetic moments calculated within the same RWIGS.