5
$\begingroup$

Disclaimer: Assume the questioner is a novice in DFT

Is PAW an improvement over the pseudopotential method? In terms of Perdew, Burke, and Ernzerhof (PBE) parameterization, what is the difference between PAW-PBE and GGA-PBE?

Improve this question
$\endgroup$
5
  • 2
    $\begingroup$ Related answer: mattermodeling.stackexchange.com/a/1986/5. Particularly the part starting with "There is no systematic manner for deciding which pseudopotential is ‘better’ as a pseudopotential..." $\endgroup$
    – Nike Dattani
    Oct 12, 2020 at 18:15
  • $\begingroup$ That was a nicely detailed answer. That still leaves the PBE part out. Also, I've started putting these disclaimers so answers would become novice-friendly. No harm in having the same question twice on the SE; one for seasoned DFTers, another for rookies. $\endgroup$
    – Hitanshu Sachania
    Oct 12, 2020 at 18:33
  • 1
    $\begingroup$ This question could use more clarity. "Is PAW an improvement over the pseudopotential method" would be a good standalone question for the site (although it has a partial answer already as @NikeDattani notes). I'm not sure what the second part is asking - perhaps you could elaborate and ask it as a separate question? $\endgroup$
    – wcw
    Oct 13, 2020 at 17:42
  • $\begingroup$ @wcw, your suggestion will make this a better question but will make Jack's answer seem misplaced. Should I edit his answer to include the original question? $\endgroup$
    – Hitanshu Sachania
    Oct 14, 2020 at 11:48
  • $\begingroup$ I am not a moderator so I don't want to overstep, but here is one idea. Maybe this question could be something like "what is the relationship between PAW/pseudopotentials and XC functionals?", assuming I've understood your intent correctly. And then you could ask "Is PAW an improvement over the pseudopotential method?" as a separate question since Jack's answer doesn't really address that part anyway. $\endgroup$
    – wcw
    Oct 14, 2020 at 20:05

1 Answer 1

Reset to default
3
$\begingroup$

PAW is proposed to deal with the interaction between electrons and ions. However, PBE or GGA (PBE just one kind of GGA) is just the exchange-correlation functional. You can think PAW and PBE are related to the two different terms in the Kohn-Sham equation. So your question is totally meaningless because you are comparing two completely different concepts.

Improve this answer
$\endgroup$
5
  • $\begingroup$ Understood. I have a doubt, though. When a PP file says PAW-PBE, does it mean it uses the PAW method for the ion-electron interaction and GGA-PBE as the xc functional? $\endgroup$
    – Hitanshu Sachania
    Oct 13, 2020 at 11:56
  • 2
    $\begingroup$ Yes. For example, the POTCAR for each element in VASP contains information for pseudopotential and exchange-correlation. $\endgroup$
    – Jack
    Oct 13, 2020 at 14:38
  • 1
    $\begingroup$ PAW involves more than just the electron/ion interaction. For example, it allows determination of all-electron kinetic energies. $\endgroup$
    – wcw
    Oct 13, 2020 at 17:42
  • 1
    $\begingroup$ @wcw What's information contained POTCAR of VASP? Except for exchange-correlation functional and pseudopotential. $\endgroup$
    – Jack
    Oct 14, 2020 at 5:27
  • 1
    $\begingroup$ Projectors, all-electron basis functions, pseudized basis functions, core densities, augmentation charge info, etc. And this is a little pedantic, but strictly PAW isn't a pseudopotential method - it's all-electron (usually with a frozen core). So the PAW transformation affects every term in the energy functional. $\endgroup$
    – wcw
    Oct 14, 2020 at 10:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .