Dear Bertrand: 1) The MSD is quadratic in time at the beginning if you start with finite velocities. If you start with zero velocities the MSD will be even slower (t**4 in principle). For a fluid, the asymptotic behavior is linear in t, for a solid it converges to a finite value (which can be connected to the Debye-Waller factor of x-ray intensities) in a time scale comparable to the vibrational time scale, so not too slow. 2) The MSD computed in cp.x are 3-D. 3) MSD's are computed in subroutine "ions_displacement" which can be found inside "ions_base.f90" in directory "Modules".
Your simulation looks ok, except perhaps for the linear dependence in t at the beginning, but the quadratic (or other) initial laws are generally seen only in the first very few steps, 10-20 steps are more than enough to spoil the ideal behavior typically. Sandro > > > I am running CP dynamic with fixed temperature for a given solid > system, using cp.x. The microcanonical calculation went very well with > a conserved energy.Now the energy is still conserved. But my emphasis > is on the mean square displacement(MSD) of atoms. They are calculated > and plotted as a function of time. > At the begining, the dependence seems to be linear at after, the MSD > seem to oscillate around certain values which can be taken as a mean > value after a sufficiently long time. My questions are: > 1) According to the litterature, the dependence at the beginning > should be quadratic and became linear after a long time. But for > condensed system as a solid it is, the MSD should saturate after a > very long time. What is happening for my case? > 2) The MSD computed with cp.x are tree dimensionnal. Please, confirm > it or not. > 3) In which file should I see how those MSD are computed? > > > Thanks very much for your help > > > *************************** > Bertrand SITAMTZE YOUMBI > Laboratory of Material Sciences > Department of Physics > University of Yaounde I-Cameroon > **************************** > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -- Sandro Scandolo, ICTP | http://www.ictp.it/~scandolo -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.democritos.it/pipermail/pw_forum/attachments/20100430/e9bfd199/attachment.htm
