I am new to Quantum ESPRESSO and it's my first calculations with $\mathrm{CsPbBr}_3$ perovskite structure. I am trying to perform vc-relax calculations. The calculation runs fine for pbesol but doesn't converge for pbe-kjpaw, pz-kjpaw, pbe-rrkjus, or pz-rrkjus. See the naming convention here.

The pbesol pseudopotential set was suggested by input file generator hosted over at materials cloud. I noticed that the pseudopotentials are mixed i.e. for $\mathrm{Cs}$ and $\mathrm{Br}$, the PP type is USPP whereas for $\mathrm{Pb}$ I had kjpaw.

Is there any rhyme or reason as to why some PPs work and some don't?

  • 1
    $\begingroup$ +1 adn welcome to the MMSE. Per this policy, I have removed your second question from the post. Also the question what are the guidelines on mixing USPP and PAW indeed is worthy of a separate post. If possible, please add at least one of the input files. You said you generated the PP using materials cloud input file generator but afaik they do not generate PP from the scratch, they just write the name from the SSSP library. So, please state how you obtained your pseudopotentials and how you used them in your input file. $\endgroup$ Commented Feb 29 at 13:03


You must log in to answer this question.