ABINIT Discussion Forums

Hi all,

Is it possible to get the contributions to the energy bands from each atoms with norm-conserving pp? Just like the fatbands in PAW. From the manual, fatbands are not support by ncpp. However, is there another way to get the information?

Thanks

Xiaoming Wang
Postdoc
Rutgers

Hi!

You can do it yourself using Xmgrace, for example. You just plot the contribution as errors represented as symbol sizes.

Best regards,

Igor

ilukacevic wrote:Hi!

You can do it yourself using Xmgrace, for example. You just plot the contribution as errors represented as symbol sizes.

Best regards,

Igor

Thanks for your reply. In fact, I'm not concerning the plot of the fatbands, but how to get the contribution data of each atom for the NCPP.

Best,
Xiaoming

I don't know if there is a variable which will 'directly' give you the atom species decomposed DOS. Some of the 'elders' probably know better. I do it by using prtdos 3 to obtain the l decomposed DOS for each atom. Then I add all the l contribution for each atom and get what you asked for. Expect 'errors' here, however, since you are not using PAW datasets (read prtdos 3 description).

Best regards,

Igor

ilukacevic wrote:I don't know if there is a variable which will 'directly' give you the atom species decomposed DOS. Some of the 'elders' probably know better. I do it by using prtdos 3 to obtain the l decomposed DOS for each atom. Then I add all the l contribution for each atom and get what you asked for. Expect 'errors' here, however, since you are not using PAW datasets (read prtdos 3 description).

Best regards,

Igor

I have another question with regard to the projected dos calculations. In QuantumEspresso, for projected dos calculation, it will output the contribution or the weight of each angular momentum l for each k point and each band. Would that be possible in ABINIT? Actually, I need the k point resolved DOS (in fact the fatbands), which sounds weird but the DOS is the summation over all the k points. So I think ABINIT also has that information.

Best,
Xiaoming