Hi Stefano, Thank you very much for sharing your data. After searching the literature, I also see DFT+U is a solution for the elements having f-electrons. One other solution that I think it might also work is to use hybrid functional.
Best, Trinh On 8/31/10 6:09 AM, "Stefano Fabris" <fabris at democritos.it> wrote: > Hi Paul (and Trinh Vo), > > unfortunately I do not have a specific experience with Pr or Yb, but > you are welcome to take as a starting point the parameters I used for > the Ce PP. In attachment you may find some input files for > pseudopotentials we have generated some years ago by using the > Vanderbilt code (http://www.physics.rutgers.edu/~dhv/uspp/, version > 734). I have tested the pseudo on the properties of CeO2-based > materials only. In my experience, semicore states were very important > while adding more projectors to the f channel made little difference. > This may however be very element-dependent. In any case, as noted by > Stefano de Gironcoli, the self-interaction error is a serious issue in > rare-earth materials and the DFT+U approach is one way to cope with it. > > Finally, I could also find some material on Pr that was sent to me > many years ago by Paolo Giannozzi. I do not think that it was ever > used, so again, no guarantee, but maybe you can find it useful. > > You can refer to me if you have questions that are too specific for > the forum, I will be happy to contribute as far as I can. Hope it helps. > > Yours sincerely, > Stefano >
