I have calculated the dielectric tensor using ph.x for the case of PAW pseudopotential with spin-orbit interaction (non-magnetic) and obtained the results.<div><br></div><div>After calculation, I have found the comments in the header of phonon.f90 file as follows:</div>
<div><br></div><div> ! ... Presently implemented: </div><div>... </div>
<div> ! ... dielectric constant NC [5], US [5], PAW [3] </div><div> ! ... born effective charges NC [5], US [5], PAW [3] </div>
<div>... </div><div> ! </div>
<div> ! NC = norm conserving pseudopotentials </div><div> ! US = ultrasoft pseudopotentials </div>
<div> ! PAW = projector augmented-wave </div><div> ! [1] LDA, </div>
<div> ! [2] [1] + GGA, </div><div> ! [3] [2] + LSDA/sGGA, </div>
<div> ! [4] [3] + Spin-orbit/nonmagnetic, </div><div> ! [5] [4] + Spin-orbit/magnetic (experimental when available) </div>
<div><br></div><div>Do they mean that my results for dielectric tensor and born effective charges are not meaningful?</div>