I am working on DFT code having plane wave along xy axis and bspline in z direction. For calculating the properties of TMD materials, spin orbit coupling must be included but i did not find fully relativistic uspp pseudo-potentials of atoms that can be generated from vanderbilt code. Then I get it from Quantum Esspresso UPF format. but not able to parse it. I did not find it in web. can somebody help to convert it uspp format or any txt format so that i can extract parameters from it. Thank you!

  • $\begingroup$ Do you know which version of UPF you need to support? $\endgroup$ Commented Jun 29, 2020 at 1:29
  • $\begingroup$ @PhilHasnip Dear Professor, I want to include Mo.rel-pbe-spn-rrkjus_psl.1.0.0.UPF Origin: PS Library Author: ADC Generated using "atomic" code by A. Dal Corso v.6.3. $\endgroup$ Commented Jun 29, 2020 at 5:54

2 Answers 2


USPP is not a format of pseudopotential. It is a type of pseudopotential that enables you to work with lower cutoffs. On the other hand, upf is a popular file extension for pseudopotentials.

If your question is regarding the availability of fully relativistic Ultrasoft PPs, they aren't very popular but PSLibrary offers that support.

For Norm-conserving and Projector-augmented wave PPs, Pseudodojo is pretty reliable (for both relativistic and non-relativistic cases)

  • $\begingroup$ Yeah, we know USPP is type of pseudo potential in which pseudo wavefuction can be allowed soft as possible within the core so that enable to work on lower cut off energy. In Vanderbilt code it is generated in .uspp format. My question is normally there is USPP to PAW or USPPP to UPF, is there any conversion of UPF format that generated in pslibrary to uspp format. $\endgroup$ Commented Jun 25, 2020 at 1:59
  • $\begingroup$ If you're using vanderbilt's code for uspp, then why do you need to convert it to upf? Quantum ESPRESSO supports both formats. I really see no point in converting formats when it is already supported by QE. But if you want to explore for whatever reason, you need to use the utility called 'upftools'. You will find examples of it in your QE directory. $\endgroup$
    – Xivi76
    Commented Jun 25, 2020 at 2:14
  • $\begingroup$ yeah you are right there is upftool in quantum espresso but there is only uspp2upf not upf2uspp. Actually i want to include relativistic Pseudopotential generated by quantum espresso in UPF format to local code for 2d with bspline. the local code is designed with uspp pseudopotential but need to include spin orbit coupling term on it. Vanderbilt uspp did not support relativistic uspp (which we need to include soc). do you have other ideas how to include spin orbit coupling term (in previously written code that is uspp pseudopotential)? $\endgroup$ Commented Jun 25, 2020 at 6:55
  • $\begingroup$ @RajKumarPaudel I'm afraid my knowledge is limited to what I mentioned earlier.. For slightly more specific questions concerning QE, I would heavily recommend sending a mail to the QE forum. You need to sign up for it, which is free. You will definitely get the required help there. $\endgroup$
    – Xivi76
    Commented Jun 25, 2020 at 7:44
  • $\begingroup$ I have asked in Quantum espresso forum but still did not get. However, I saw in upftool in qe, these routine should read (read_ps.f90 and read_upf.f90) upf but I am not able to parse for UPF file say(Mo.rel-pbe-spn-rrkjus_psl.1.0.0.UPF). could you help me for this $\endgroup$ Commented Jul 3, 2020 at 12:55

I really don't know any engine to convert between pseudopotential formats, but below you can find links to engines where you can generate pseudopotentials according to your own needs:

  • 1
    $\begingroup$ This list misses out the most pertinent one - Davd Vanderbilt's original code which generates the .uspp format requested in the question. physics.rutgers.edu/~dhv/uspp/index.html. I am not sure whether it can generate J-dependent pseudopotentials for spin-orbit though. $\endgroup$ Commented Jul 7, 2020 at 15:06
  • $\begingroup$ Thanks @KeithRefson, I edited the list. $\endgroup$
    – Camps
    Commented Jul 7, 2020 at 15:58
  • $\begingroup$ There is also the oncvpsp (github.com/oncvpsp/oncvpsp) code that is used for the pseudo-dojo website. $\endgroup$
    – fgoudra
    Commented Jul 30, 2021 at 16:31
  • 2
    $\begingroup$ Perhaps the suggestion by @fgoudra could be added here too? Also, since fgoudra hasn't yet had any of their questions answered on this site, is there any way you can help them on this question: mattermodeling.stackexchange.com/q/6473/5 ? Perhaps explaining how to do it manually or semi-manually? $\endgroup$ Commented Dec 27, 2021 at 5:09

You must log in to answer this question.

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