<div dir="ltr"><div>Hi Katharina,</div><div><br></div><div>I am not sure if I understand your question correctly and what your computational setup is.</div><div>The basic electrostatic consideration is that the potential should be constant within a conductor. If no external potential is applied, the electrostatic potential is zero within a metal, i.e., V_0=0. That is the default. It can be also set to another value in principle, e.g., maybe to approximate the application of a potential to an electrode. But it seems that's not what you want to do here. Note that the IC-QM/MM does not introduce actual charges, only image charges. You can run with the normal periodic solver when you have a (neutral) molecule on top of the metal (MM).</div><div>If you have a charged system in the QM part, then you need a solver like MT, like in a standard QM calculation.</div><div><br></div><div>Best regards,</div><div>Dorothea<br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">Am Mi., 24. Juni 2020 um 16:23 Uhr schrieb Katharina Doblhoff-Dier <<a href="mailto:k.doblh...@lic.leidenuniv.nl">k.doblh...@lic.leidenuniv.nl</a>>:<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"><div>Dear CP2K community,</div><div><br></div><div>I am confused about the IC-QMMM method as implemented in CP2K. I am not 100% sure whether my question is related to my misunderstanding of the method (as implemented) or on how to use it.</div><div>Originally, my confusion came from my not-understanding of the parameter V_0 (i.e., EXT_POTENTIAL in the CP2K implementation): In periodic boundary conditions, this parameter does not seem to make much sense to me: Depending on the average potential in the quantum region, defining V_0 will make the charge Q on the metal adapt in such a way that the average potential in the cell is zero (I know that the latter is logic, but in view of this, the physical meaning of V_0 becmes unclear to me):</div><div><br></div><p style="text-align:center;clear:both"><img style="margin-left: 1em; margin-right: 1em;" src="cid:125c40b6-eefe-415d-8dda-8cad98055459" alt="tmp1.png" width="320" height="137"></p><div><br></div><div>Alternatively, if a charge is put into the QM part (e.g. an H2+) and V0 is adjusted such that the total charge Q on the metal is exaclty -1, a spurious field will result over the cell (due to Ewald summation) and again, V_0 does not have an obvious physical meaning. <br></div><div><br></div><p style="text-align:center;clear:both"><img style="margin-left: 1em; margin-right: 1em;" src="cid:a4f4e5b1-8692-4b78-af0e-199be08f8ee1" alt="tmp2.png" width="320" height="185"></p><div><br></div><div>Considering that the periodic boundary conditions lead to lots of spurious stuff, I then tried to go to MT boundary corrections or the IMPLICIT poisson solver as implemented in CP2K. This gave me the following results (note that I shifted the green and the blue curve by 0.4V for better comparison)</div><div><br></div><div style="margin-left:40px">red: IMPLICIT poissn solver with Neumann BC at z=0 and z=60 and homogeneous Dirichlet BC at z=30</div><div style="margin-left:40px">green: MT (Martyna-Tuckerman) poisson solver (shifted by 0.4V)</div><div style="margin-left:40px">blue: normal Ewald summation (shifted by 0.4V)<br></div><div><br></div><p style="text-align:center;clear:both"><img style="margin-left: 1em; margin-right: 1em;" src="cid:1e7347fc-fe37-43c4-8d04-11dbca230ab0" alt="tmp3.png" width="400" height="337"></p><p style="text-align:center;clear:both"><br></p><p style="text-align:left;clear:both">The corresponding V_0 (optimized such that Q=-1) were 1.11V for IMPLICIT boundary conditions, 1.24V for MT boundary conditions and -0.37V for periodic boundary conditions. While for periodic boundary conditions (blue line) this can be seen to correspond to the potential in the metal part, this is not the case for MT boundary conditions (green), where the potential in the metal varies from about -0.4 to -0.5 (remember that I shifted the curves by 0.4 volt) and for IMPLICIT boundary conditions. Overall, to me, it looks as if the IC-QMMM method was ignoring the boundary conditions and optimizing the charges for the periodic boundary conditions and then keeping them fixed no matter what I set as POISSON_SOLVER. In the MT boundary condition case (geen) we can thus see the influence of the charges shieding the (spurious) field in periodic boundary conditions (hence the slope in the metal part, which has the same slope as the average field in the periodic solver).</p><p style="text-align:left;clear:both"><br></p><p style="text-align:left;clear:both">Finally, I decided that, in principle, it should be possible to find an energy minimum when Q_image=-Q_QM (as also shown in the original paper by Siepmann and Sprik). With none of the boundary conditions could I find this minimum correctly. However, here comes my non-understanding of Eq. 4 in the paper by Golze, Iannuzzi, ..., and Hutter (<a href="https://pubs-acs-org.ezproxy.leidenuniv.nl:2443/doi/10.1021/ct400698y" target="_blank">https://pubs-acs-org.ezproxy.leidenuniv.nl:2443/doi/10.1021/ct400698y</a>) into play: Here, the energy is written as:</p><p style="text-align:left;clear:both"><br></p><p style="text-align:center;clear:both"><img style="margin-left: 1em; margin-right: 1em;" src="cid:8e5de002-0d00-41b0-8548-abaa80bd37eb" alt="tmp4.png" width="320" height="50"></p><p style="text-align:left;clear:both">I would have thought this to be a grand canonical energy (grand canonical only in the charges on the metal) expression, where the last term accounts for the -N_i*mu_i term. Again, this does not seem physical to me if I think of a capacitor (or a charge+image charge in a box that is periodic in x and y, as I would then expect a correction for the charges in the QM region too (unless the vacuum potential on the QM side is zero and open boundary conditions are used, but likely this is wrong and this may be where my entire confusion starts.<br></p><p style="text-align:left;clear:both"><br></p><p style="text-align:left;clear:both">So summarizing, this boils down to a few questions:</p><p style="text-align:left;clear:both">1.) can the IC-QMMM method be combined with poisson solvers other than periodic? If so, how? I simply set the poisson solver in the MM and the DFT part.<br></p><p style="text-align:left;clear:both">2.) What do I need to do in order to find an energy minimum for Q_image=-Q_QM?</p><p style="text-align:left;clear:both">3.) What is the meaning of V_0 and why is it substracted in the energy expression.</p><p style="text-align:left;clear:both"><br></p><div>Any physical insight is appreciated!</div><div>Thank you and best regards,</div><div>Katharina<br></div></div> <p></p> -- <br> You received this message because you are subscribed to the Google Groups "cp2k" group.<br> To unsubscribe from this group and stop receiving emails from it, send an email to <a href="mailto:cp...@googlegroups.com" target="_blank">cp...@googlegroups.com</a>.<br> To view this discussion on the web visit <a href="https://groups.google.com/d/msgid/cp2k/5774878e-630c-4699-a67e-5040897c1b2do%40googlegroups.com?utm_medium=email&utm_source=footer" target="_blank">https://groups.google.com/d/msgid/cp2k/5774878e-630c-4699-a67e-5040897c1b2do%40googlegroups.com</a>.<br> </blockquote></div>