Page 1 of 1

On the calculation of molecules and cluters

Posted: Mon May 11, 2009 6:42 pm
by ganymede
Daer Yambo Developers and users:
I have a bit confusion about the calculation of molecules and cluters use ABINIT,
First, on the groundstate calculation, if the system(molecules or cluters)have N number ofvalence electrons, which number of nband i should choose? or this choose is enough nband = (100% + 20%) of (valence el.)/2 ? or i shoul make a test?
Second, for generation the KSS file, the same confusion , which number of nband2 and nbandkss shoul i choose?
Thanks very much!

Re: On the calculation of molecules and cluters

Posted: Mon May 11, 2009 7:17 pm
by myrta gruning
Hallo Ganymede,

:!: first of all, remember to fill in your signature :!:

In general for the ground state DFT calculation (the scf step) you just need the occupied bands. Usually some more are added, but their number should not affect the calculation. Special care is needed in the case of metalic systems, but for systems with a finite gap should not be any complications and in the case abinit you may even use the default value.
To generate the kss, take nband2 and nbandkss2 equal. You do not know in advance the exact number you need to get a converged calculation. This depends on the systems and the properties you are calculating and you will need to test it.
Choose a pretty large number, if it is not enough you will have to generate a new KSS (note that you can start from the old one).

Cheers
m