I need to create a pseudopotential of Calcium in which is considered the valence electrons of orbitals 3p6 and 4s2. Usually, we considered only orbital 4s2 as valence and the input is
pe -- file generated from Ca ps-file
tm2
Ca pb
0.000 0.000 0.000 0.000 0.000 0.000
5 4
4 0 2.000 0.000 #4s
4 1 0.000 0.000 #4p
3 2 0.000 0.000 #3d
4 3 0.000 0.000 #4f
2.98000 2.98000 2.66000 2.66000 0.00000 2.60000 Core corrs
#23456789012345678901234567890123456789012345678901234567890 Rule
if I add a line of type
3 1 6.000 0.000 #4p
the result is a error about VPSOUT and VPSFMT. So, someone has a tip?
Hi Pedro.
If you have a sharp boundary between core and valence electrons, then it is a simple practice as you may already know. However, sometimes you need to treat some semicore states explicitly as valence ones.
Foe example, if "2s" is a smicore state snd "3s, 3p" are your valence, and you would like to have valence configuration as "2s, 3s, 3p".
Then in the pseudopotential generation you have to use "2s, 3p" configuration and find two solutions for 2s (one is nodeless, the other with one node) which the second (excited) one gives eigenvalue of 3s. You must test for ghost states.
HTH.
Bests,
Mahmoud
Dear Mahmoud,
Thank you for your tips. After some research about this, I found the following topics:
1) https://personales.unican.es/junqueraj/JavierJunquera_files/Metodos/Pseudos/Semicore/Semicore.pdf
2)Presentation PAO Basis Sets in SIESTA
These topics were very useful to me. In some cases, states of semicore better describe the band structure of your system.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Electrodes