Dear PWSCF Users and Dr. Gabriele. Thank you for your reply.
>By inspecting that file I noticed: > <PP_QIJ> > 8 nqf. If not zero, Qij's inside rinner are computed using qfcoef's > <PP_RINNER> > 1 8.00000000000E-01 > 2 8.00000000000E-01 > 3 8.00000000000E-01 > 4 8.00000000000E-01 > 5 8.00000000000E-01 > </PP_RINNER> >So that maybe you have to apply an interpolation formula to get the correct Qij inside >rinner (which should be 0.8). Have you taken into account this? If not maybe you should >re-read a reply to your post Of course we tried the interpolation formula and the result was same. Basically, the Qfun(r) is correct also in the region (r < rinner ) as already written in your suggested post. Please see the post. > http://www.democritos.it/pipermail/pw_forum/2009-January/011188.html and furthermore, it can also be seen from the UPF file Zn.pz-van_ak.UPF <PP_QFCOEF> section with Qij with i=5, j=5 case (which is the case l(i)=2, l(j)=2 ), or i=6, j=6 case (also l(i)=2, l(j)=2 case), the coefficient of Q^{L=0}_{ij}(r) has non zero coefficient C_{0} = 1.495560*10^{2} (for i=5, j=5) with Q^{L=0}_{ij}(r) = r^{L}*(C_{0} + C_{1}*r^{2} + ..... ) , which has weight around r=0 when L=0. L=0 is possible when the angular momentum of beta_{i} and beta_{j} is same, (Q_{ij}) . I think the origin of weight around r=0 of augumentation charge comes from the L=0 component of Q function, and why C_{0} coefficient is not zero for Q_{ij} with l(i)=l(j)=2 case (Zn case) ? Sincerely. Yukihiro Okuno. Yukihiro Okuno.
