Hi I propose to fix only 3 or 4 first layers and permit another layers to be optimized. When you change the first layer's positions manually, you force the atoms to pass the potential barrier. I think the code's algorithms (such as another codes) can raise small value to the potential, and if the potential barrier is high, it can not change the atoms very high to pass it. It finds only local minimum around it. ? With the Best Regards
Reza Behjatmanesh-Ardakani Associate Professor of Physical Chemistry Address: Department of Chemistry, School of Science, Payame Noor University (PNU), Ardakan, Yazd, Iran. E-mails: 1- reza_b_m_a at yahoo.com (preferred), 2- behjatmanesh at pnu.ac.ir, 3- reza.b.m.a at gmail.com. On Tuesday, December 10, 2013 4:13 AM, "Li, Run" <run.li at my.und.edu> wrote: Hello, ? I have a questions about relax and vc-relax using pw.x. I tried to use them to optimize the geometry of graphite, and the atoms can barely move even though I move them manually I can get lower energy. I?m not sure why. ? For example, I take 7 layers of ABAB stacked grapheme, fix the bottom 6 layers using the geometry parameters of graphite(000 thing), and relax the top layer. Geometry is defined as ibrav=14, celldm(1)=xx, celldm(3)=22.0(should be big enough), nosym=.true. Default bfgs optimization parameters are taken. ?The calculation usually takes a few iterations and stops, with the top layer moved by only 0.0002 A. But if I manually pull the top layer away by 0.7A, the energy lowers by 0.0005Ry. It also didn?t help by increasing trust_radius_max and trust_radius_ini to 3.0 and 1.0. Is this behavior expected? How do I make it converge to a real minimum? ? I can run test calculation example02/al001.mm.in just fine. ? Thank you very much, ? Rain Department of Chemistry University of North Dakota _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://pwscf.org/mailman/listinfo/pw_forum -------------- next part -------------- An HTML attachment was scrubbed... URL: http://pwscf.org/pipermail/pw_forum/attachments/20131209/90214bee/attachment.html
