19981210: ``nve'' run with cell size from 19981207

GOAL

``nve'' run with cell size from the ``npt'' run in 19981207

Results

I used the cell size from the 19981207 npt run as a sort of optimized value: is this correct? I don't think so.

Anyway, this run is comparable with 19981127 (nve run with polarizable oxygen ions). The following are the differences between the two experiments:

	19981127	19981210
lattice constant $(\mbox{\textit{\AA}})$	5.211	5.400
step size $(ps)$	$0.002$	0.0005
total n. of steps	$105000$	$420000$
total time $(ps)$	$210$	$210$
equilibration steps	$5000$	$20000$
equilibration time $(ps)$	$10$	$10$
cutoff $(\mbox{\textit{\AA}})$	$7.8$	$8.0$

Decreasing the time step by four times (with corresponding increment in the execution time!!) does not seem to make things better. I would say that the MSD plot for 19981127 is better than the present one. The lattice constant seems to be a crucial parameter: how to sort it out in a non-arbitrary way? The attempt made in 19981124 is not encouraging at all!!

Also, the $D_O$ does not seem to be different in two cases. Fitting to a straight line in the interval $[30,140]\;ps$ gives the following results:

$$\begin{eqnarray*} MSD&=&B+6D_Ot \end{eqnarray*}$$

	19981127	19981210
slope $(\mbox{\textit{\AA}}^2/ps)$	$1.66890423\times10^{-2}$	$1.57414584\times10^{-2}$
$D_O\ (cm^2/s)$	$2.78\times10^{-7}$	$2.62\times10^{-7}$