Projectors for B-N

[camelto2]

This is an initial set of projectors for B, C, N, and O. The *.upf files were created from opium, using a modified version that after some cleaning will eventually be pushed into the actual opium code hopefully.

In the meantime, the semi-local version of the pseudopotential has been cross-checked against other codes using semi-local potentials. This indicates that the semi-local potential used to generate the nonlocal is correct. For the reference configuration, the nonlocal and semilcoal evaluation agree, which they must. Then in the report file *.rpt, a number of checks on the nonlocal checks are given, including the ghost states, transferability of the nonlocal projector (against the semilocal version), and the recommended plane wave cutoff.

Although each has not been tested in Quantum Espresso, for the C.upf file, I have tested diamond against the semi-local evaluation with pyscf for diamond primitive at the gamma point. Using a reasonably large gaussian basis, diamond energy is -10.27609 Ha, whereas the plane wave version with the recommended cutoff gives -10.27773 Ha. This is encouraging, and suggests that the projectors are ok at the DFT level.

These should be used with care since they are not thoroughly tested

[prckent]

Thanks. Do you have an H.upf in the works?