Dear Quantum ESPRESSO users, I'm trying to compute the phonon dispersion of bulk and single layer NbSe2. While I succeed for the single layer, I fail for the bulk. I have read previous posts about this issue and tried to reduce alpha_mix, but the non-convergence just moves from one mode to another. Following the recipe from a ICTP school in March 2018 <http://epw.org.uk/School2018/School2018?action=downloadman&upname=Wed.4.Verdi.pdf>, I'm using the inputs in attachment (up to where it fails, anyway). For the single layer, which is successful, I have never obtained this problem (with or without vdW corrections, even without the assume_isolated = '2D' prescription, and with any k-mesh and q-mesh used so far). I tried several setups also for the bulk, with the following behaviour:
no convergence in the bulk for k-mesh 45x45x5, q-mesh 15x15x1 or 9x9x3, convergence in the bulk for k-mesh 9x9x5, q-mesh 3x3x1 (only 3 points). I understand that a system can be far from convergence, but why does this work for its 2D allotrope, and why does it work when I use a coarser grid (without even adjusting the broadening, btw)? More important, what can I do in order to obtain convergence? This system has been already investigated in the bulk, at least in Phys Rev B 80 241108 <https://journals.aps.org/prb/abstract/10.1103/PhysRevB.80.241108> and a more recent PRB <https://journals.aps.org/prb/abstract/10.1103/PhysRevB.87.245111#> (btw, isn't the broadening a bit high in those papers, or should I increase the broadening in my case? which goes back to the question, on why the single layer is not affected by a small broadening ...) Thank you in advance for any help or comment you may have about this issue. Kind regards, Fabrizio -- *Fabrizio Cossupostdoctoral fellow at APCTP (Asia Pacific Center for Theoretical Physics)*, Hogil Kim Memorial Building #501 POSTECH, 67 Cheongam-Ro, Nam-Gu, Pohang-si, Gyeongsangbuk-do, 790-784 (37673), Republic of Korea
uc-scf.in
Description: Binary data
uc-scf_fit.in
Description: Binary data
uc-elph.in
Description: Binary data
_______________________________________________ users mailing list [email protected] https://lists.quantum-espresso.org/mailman/listinfo/users
