On projected Density of States

1. May 27, 2005

leoant

Dear all:
I am frustrated in the concept of the PDOS, may someone be kind to refer me to the origin of this concept? Thanks.

Regards

2. May 27, 2005

marlon

Don't you mean partial density of states ? When plotted out, they express the relative contribution to the total DOS (of a molecule built out of two atoms for example) wrt each other. I mean, the DOS of each atom is plotted in one graph. When you also add the total DOS you can clearly see which atom delivers the biggest contribution to the DOS of the molecule. These data are always plotted with respect to the energy (the horizontal axis if you will). They are the direct analogon of COOP's (crystal orbital overlap population) in chemistry. The origin is just the idea of COOP and more specifically, the Mulliken population analysis...Do you know this ? When two electronic levels overlap, you say that there is an electronic density between two atoms : a chemist calls this the bond, right ? Now, what if you persist in wanting to devide up the electronic density between the two atoms that are involved ? I mean, how to assign the overlap density to the two centers ? Well, Mulliken suggested that you should just deivide this overlap density by a actor two and then assign it to each atom...

marlon

3. May 27, 2005

leoant

Thank you for your reply, marlon. Yeah I am reading the paper by R. S. Mulliken, however, what I really do is calculation with plane-wave pseudopotential ab initio method and I want to interpret my results. Thus I wonder how can one derive the same meaning for plane waves as AOs in Mulliken's realm? And further more, as to crystals, which are different from moleculars, how can one use, say, the Mulliken Population Analysis correctly and make sense? And I still want to know if there is any difference between "Projected DOS" and "Partial DOS", or there's no "Projected DOS" and only "Partial DOS"?

4. May 28, 2005

Gokul43201

Staff Emeritus
I think the terminology is almost interchangeable, but not quite. Partial DOS refers to both Gross Population DOS as well as Projected DOS.

I do not know enough about this but I too would have recommended Mulliken's paper on population analysis.

Also, you could look for calculations for crystals in the framework of the Extended Huckel model (a modified form of molecular orbital theory) using Overlap Population DOS.

Last edited: May 28, 2005