<div dir="ltr"><div>The energy of charged cells depends upon a rather arbitrary choice of the G=0 potential. If you compare the energies from CP with those from PW, you will notice that they are quite the same (once the factor 2 between Ry and Ha is taken into account!) for neutral cells, while they differ for charged cells: CP uses a different convention for G=0. In principle, forces and stresses might also depend upon the choice of the G=0. I have never felt or heard that this effect is in practice relevant
for structural optimization. I am less sure about variable-cell optimization, though</div><div><br></div><div>Paolo<br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Apr 30, 2021 at 4:11 PM Hanghui Chen <<a href="mailto:chenhanghuipwscf@gmail.com">chenhanghuipwscf@gmail.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir="ltr">Dear QE developers,<div> I have been using Quantum Espresso for more than a decade and recently a paper (PRB 91, 024107 (2015)) caught my attention, which shows that the total energy, pressure and cell relaxation are ill-defined in charged systems (even if a compensating charge background is present). The paper shows that the total energy and pressure of a charged system depend on the average electrostatic potential convention, and thus is code-dependent (QE, ABINIT and VASP use different conventions).</div><div> I think the paper is theoretically correct. However in practice, we have recently tested doped ferroelectric materials using both QE and VASP. In this charged system, we find both QE and VASP yield qualitatively consistent structural properties.</div><div> Since relaxation of a charged system is very common in literature using QE or VASP (e.g. PRL 109, 247601 (2012) and many others), I am wondering whether the relaxed structure of a charged system from QE is reliable. According to PRB 91, 024107, it is not legitimate to do cell relaxation of a charged system. However both QE and VASP do allow to calculate total energy and perform cell relaxation in a charged system. And from literature, it seems that the relaxed structure of a charge system from QE or VASP calculations is "not too wrong".</div><div> So what is your insight on this issue? Shall we prohibit ourselves from doing cell relaxation of a charged system in QE, or can we still use the standard procedure but with the caution that the optimized structure of a charged system from QE might not be accurate (due to the QE convention of average electrostatic potential)?</div><div> Thank you very much.</div><div><br></div><div>Dr. Hanghui Chen</div><div>Department of Physics </div><div>NYU Shanghai and New York University</div><div> </div></div>
_______________________________________________<br>
Quantum ESPRESSO is supported by MaX (<a href="http://www.max-centre.eu" rel="noreferrer" target="_blank">www.max-centre.eu</a>)<br>
users mailing list <a href="mailto:users@lists.quantum-espresso.org" target="_blank">users@lists.quantum-espresso.org</a><br>
<a href="https://lists.quantum-espresso.org/mailman/listinfo/users" rel="noreferrer" target="_blank">https://lists.quantum-espresso.org/mailman/listinfo/users</a></blockquote></div><br clear="all"><br>-- <br><div dir="ltr" class="gmail_signature"><div dir="ltr"><div><div dir="ltr"><div>Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,<br>Univ. Udine, via delle Scienze 206, 33100 Udine, Italy<br>Phone +39-0432-558216, fax +39-0432-558222<br><br></div></div></div></div></div>