Hi, Thanks for your help. According to the suggestions, I have tested the convergence of dielectric constants and effective charges with respect to k mesh and they in fact converged with the size of k mesh.
As Prof. Marzari indicated, the phonon dispersion including LO-TO splitting might have the discontinuity becase it depends how q vector appoaching gamma point. It should be corrent. I also found some example of this kind unusal phonon dispersion in literature. Thanks, Changpeng Marzari Nicola <[email protected]> 于2018年11月14日周三 下午4:32写道： > > > Note you can converge to different limits depending on direction, so a > discontinuity in a plot where the direction changes angle is possible. > Nicola > > Sent from a tiny keyboard... Contact info: > http://theossrv1.epfl.ch/Main/Contact > > On 14 Nov 2018, at 09:29, Lorenzo Paulatto <[email protected]> wrote: > > >> I found that optical branches are not continuous at gamma point (See > figure below). I guess this is becasue the wrong LO-TO spiliting relation I > got. How can I check it? > > > > It is definitely because of LO-TO splitting, if it is wrong or not, well > you can check how it converges with k-points. Dielectric constants and > effective charges can be VERY slow to converge with k-points. Just repeat > the phonon calculation at Gamma, no need to redo all the grid each time. > > > > Or if you think the should be no LO-TO splitting at all (typically > because of doping), just edit the first dyn matrix file (the one at gamma, > i.e. dyn1) and remove all the section about epsilon and zeu, from > > "Dielectric Tensor:" > > included, down to > > "Diagonalizing the dynamical matrix" > > excluded. > > > > hth > > > > > > -- > > Lorenzo Paulatto - Paris > > _______________________________________________ > > users mailing list > > [email protected] > > https://lists.quantum-espresso.org/mailman/listinfo/users > _______________________________________________ > users mailing list > [email protected] > https://lists.quantum-espresso.org/mailman/listinfo/users >
_______________________________________________ users mailing list [email protected] https://lists.quantum-espresso.org/mailman/listinfo/users
