this message means that the dynamical matrix is not hermitian as it should... it may be due to poor self-consistent convergence in the phonon calculation this may also be the reason for the negative (imaginary) frequencies you get. try to tighten the tr2_ph threshold. stefano
Zhen Huang wrote: > Dear friends, > > I am running a single-wall carbon nanotube case (5,0). The result of phdos > shows > Max |d(i,j)-d*(j,i)|/|d(i,j)|: 210.0909% > Max |d(i,j)-d*(j,i)| = 0.062230 > Max |d(i,j)-d*(j,i)|/|d(i,j)|: 266.6923% > Max |d(i,j)-d*(j,i)| = 0.062230 > Max |d(i,j)-d*(j,i)|/|d(i,j)|: 253.1387% > Max |d(i,j)-d*(j,i)| = 0.062230 > Max |d(i,j)-d*(j,i)|/|d(i,j)|: 233.9506% > Max |d(i,j)-d*(j,i)| = 0.062230 > Max |d(i,j)-d*(j,i)|/|d(i,j)|: 208.3492% > Max |d(i,j)-d*(j,i)| = 0.062230 > Max |d(i,j)-d*(j,i)|/|d(i,j)|: 213.2651% > Max |d(i,j)-d*(j,i)| = 0.062230 > Max |d(i,j)-d*(j,i)|/|d(i,j)|: 256.8845% > Max |d(i,j)-d*(j,i)| = 0.062230 > > Does any one know what is wrong with it? > > On the other hand, I used C.pz-rrkjus.UPF for PP. It works well with graphite > and give reasonable dispersion. However, there are several negative value in > nanotube results. I use 1,1,40 in relax calculation and nq=1,1,32 for ph.x > calculation. > > --------------- > Zhen Huang > Ph.D. Student > Nanoscale Transport Research Group > School of Mechanical Engineering > Purdue University > Tel: 765 426 9302 > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum >
