Page 1 of 1
Plotting KB projectors and changing KB energies
Posted: Fri Apr 12, 2013 8:00 pm
by fabiano.corsetti
Dear all,
I have two questions relating to the Kleinman-Bylander projectors that ABINIT calculates from the pseudopotentials it is given:
a) Is there an easy way of getting the code to output the radial part for each l for plotting?
b) Is it possible to change the values of the KB energies ekb that are used in the SCF calculation? In particular I would like to switch off some of the channels.
Many thanks in advance for any suggestions you can give me.
Kind regards,
Fabiano Corsetti
CIC nanoGUNE Consolider
Donostia-San Sebastián
Spain
Re: Plotting KB projectors and changing KB energies
Posted: Sat Apr 13, 2013 12:52 pm
by jzwanzig
I believe both these questions should be approached at the stage of making the pseudopotentials. For example, using the fhipp code to make TM pseudopotentials, you can control which channels are included and at what reference energies, and also plot the various components. The recently developed APE code could be used for this purpose too. Likewise using atompaw to make PAW datasets you can plot the various components as you make a different pseudopotential.
Re: Plotting KB projectors and changing KB energies
Posted: Mon Apr 15, 2013 12:52 pm
by fabiano.corsetti
jzwanzig wrote:I believe both these questions should be approached at the stage of making the pseudopotentials. For example, using the fhipp code to make TM pseudopotentials, you can control which channels are included and at what reference energies, and also plot the various components. The recently developed APE code could be used for this purpose too. Likewise using atompaw to make PAW datasets you can plot the various components as you make a different pseudopotential.
Thank you for your suggestions. The problem is that I am using a pseudopotential generated by SIESTA's atom code (I want to compare total energies in SIESTA and ABINIT, so I need everything to be exactly the same), which only includes the semilocal components and the pseudowavefunctions. The KB projectors are generated during the run both by SIESTA and ABINIT. I can see from the output that the KB energies are almost identical for the occupied shells, but differ significantly for the unoccupied shells:
Code: Select all
(in Ha)
l O (ABINIT) O (SIESTA) H (ABINIT) H (SIESTA)
0 4.626822 4.626827 -1.010971 -1.010970
1 -3.948584 -3.948588 -0.221151 -0.221724
2 -0.004655 -0.961190 -0.000140 -0.070272
3 -0.000000 -0.369138 0.000000 0.000000
This is why I would like to have a look at the KB projectors during the run, so I can try to figure out what is being done differently.