On 07/19/2011 06:49 PM, Eric Germaneau wrote: > > > On 07/18/2011 10:45 AM, Hongsheng Zhao wrote: >> On 07/19/2011 09:23 AM, Eric Germaneau wrote: >>> Hey Amin, >>> >>> The approach you proposed is ok but you should relax the cell as well. >>> By keeping the cell fix the system does not feel the pressure you wish >>> to apply. >> Not so clear about the meaning of this sentence, could you please give >> some more explanations? ;-( > Under pressure keeping the unit cell fix is useless since the system > won't feel it. > It just makes sense. > So, if you wish to see the effect of the pressure you have to perform > vc-relax calculation. > Take it that way, in thermodynamic you have 2 important ensembles, > isochoric and isobaric. > So either the volume is fix and you get out the corresponding pressure, > or the pressure is fix and you get out the corresponding volume. > Note that the volume of gas for instance decreases inversely > proportionally to pressure. > The pressure is the intensive variable related to the volume (extensive). > You simply can not have both of them constant at the same time.
Yes, you're right. I'm so sorry for my previous misleading posts on this thread. Thanks again. > Hope it helps. >> >>> You can also change the volume, relax the positions only and compute the >>> pressure. >>> The best is to compute the free energy as function of pressure ans so >>> the QHA method is the best way to go. >> I think this method cann't do the job of structure-searching-and-finding >> for unknown phases when we changing the external pressure. > Yes, it can but in the case of structure-searching it's very expensive. > In that case you'd better generate random unit cell parameters, random > positions and run vc-relax. > Even this can be expensive since you have to do it a lot of times .... > Just take a look to this papers, PRL 97 1 (2006) > <http://link.aps.org/doi/10.1103/PhysRevLett.97.045504> and JPCM 23 > 053201 (2011) > <http://stacks.iop.org/0953-8984/23/i=5/a=053201?key=crossref.44087776832bc9a0edd0dc492a67d376>. > It is the main work of Professor Pickard > <http://www.st-andrews.ac.uk/physics/condmat/pickard/Crystals%20ex%20nihilo.html>. > Best. Thanks a lot for this information ;-) Regards > >> >> Regards >> >>> Note that at first you can ignore the thermal effect and compute the >>> free energy from a regular phonon calculations. >>> I'm also very interested in such calculations but don't have experiences >>> doing this by mean of first principal calculations. -- Hongsheng Zhao <zhaohscas at yahoo.com.cn> School of Physics and Electrical Information Science, Ningxia University, Yinchuan 750021, China
