AMBER Archive (2003)

Subject: diffusion in ptraj

From: Ioana Cozmuta (ioana_at_nas.nasa.gov)
Date: Mon Jul 07 2003 - 16:56:56 CDT


Hi,

I have a question about the message posted in ptraj when doing a diffusion
calculation. It says:

      To calculated diffusion constants, calculate the slope of the
lines(s)
      and multiply by 10.0/6.0; this will give units of 1x10**-5 cm**2/s
      The atoms in the calculation follow: * (All atoms are selected)

This is not clear to me. As far as I know diffusion constants, D, are
calculated from Einstein's equation
MSD = 6*D*t

if log log plotted the eq above becomes
ln(MSD) = ln(6*D) + ln(t)

meaning that one should do a fit with two parameters: the ofset A0 and the
slope A1. If one reaches the diffusive regime A1 should be 1 and the
diffusion constant is calculated from the ofset (and not the slope as
written above):

D = exp(A0)/6 in units of A^2/ps
or
D = exp(A0)*10/6 in units of 1x10^(-5) cm^2/ps

As far as I can see from the code the data dumped in the output files are
the MSD and time.

Why is the preference towards a regular plot instead of log-log plot?

I would appreciate your comments.
Thank you,
ioana