<div dir="ltr">Dear Dorothea, thanks for your comments. I think water at SPC model + Siepmann potential cannot be dissociated. The Siepmann terms are affected only on Pt-O iterations. Dissociation can occur with the separate iterations (Pt-O and Pt-H) and a flexible water models (or dissociations model). Previous calculations (graphite slab+water, metal ions+water) doesn’t show any bugs with the SPC model in cp2k, all intermolecular distance was fixed during calculation. The problem of temperature jumps was fixed by adding REGION MOLECULE in &TERMOSTAT. Now energies and temperature seems quite plausible. Main tendency of previous calculation is a drift Cons Qty. Often the terminations are caused by atoms of water on peripheries of cell. Probably, the termination of all tasks is associated with boundary conditions. On cp2k forum I found useful theme: Energy conservation NVT/NVE simulation <a href="https://groups.google.com/forum/#!topic/cp2k/ZNEiZIFMbKM" target="_blank">https://groups.google.com/forum/#!topic/cp2k/ZNEiZIFMbKM</a> This topic has comprehensive information about stability of the calculation. Dorothea, you were really right at the expense of the MD setup and parameters. Mention above problems is fixed by adding: - - in &constraint SHAKE_TOLERANCE … (more than the default values) ROLL_TOLERANCE … - - in &ewald ALPHA should be increased >0.5 Now all calculations with the siepmann potential are stable. пятница, 13 ноября 2015 г., 0:04:19 UTC+3 пользователь Dorothea Golze написал:<blockquote class="gmail_quote" style="margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div><div><div><div><div><div>Hi Sergey, </div>the Siepmann-Sprik potential assumes intact water molecules at the interface. In your case, some dissociation seems to take place. It is checked with a simple distance criterion if we have 2 hydrogens per oxygen. You get the error message: Error in Siepmann-Sprik part: too many H atoms, i.e. more than 2 hydrogen per oxygen found. </div>The keyword ALLOW_OH_FORMATION ignores that and doesn't calculate the T3 term of the Siepmann-Sprik potential for this MD step. So the program won't stop. This keyword was actually intended for cases where water is described at the QM level, but you can try it as well. </div>It's strange that you have this behavior at the SPC level. However, the temperature development of your MD is very weird and has extreme jumps. So it's maybe not surprising that your O-H bonds are strangely stretched. So better control your MD setup and parameters. </div>btw: examples for the Siepmann-Sprik potential can be found in the regtests: /cp2k/tests/Fist/regtest-6/ </div>Best regards, </div>Dorothea </div><div><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"> </blockquote></div> </div> </blockquote></div>