# Re: [AMBER] Calculating diffusion Constant D

From: Dickson, Callum <callum.dickson09.imperial.ac.uk>
Date: Thu, 25 Jul 2013 10:39:30 +0000

Hi Vijay,

So from the ambertools manual:

<filename_root>_x.xmgr Mean square displacement(s) in the X direction.
<filename_root>_y.xmgr Mean square displacement(s) in the Y direction.
<filename_root>_z.xmgr Mean square displacement(s) in the Z direction.
<filename_root>_a.xmgr Average mean square displacement(s).
<filename_root>_r.xmgr Total distance travelled.

Assuming your bilayer is aligned along the z-axis, sum the x- and y- plots to get MSD in the xy-plane and take the slope of the resulting graph, then multiply the gradient by (10/4) to get the diffusion coefficient. Also note that you should do simulations in the NVE ensemble if you want to compare resulting diffusion values to experiment.

All the best,
Callum

________________________________________
From: amber-bounces.ambermd.org [amber-bounces.ambermd.org] on behalf of Vijay Manickam Achari [vjrajamany.yahoo.com]
Sent: 25 July 2013 11:07
To: Amber mailing List; Amber Scrops
Subject: [AMBER] Calculating diffusion Constant D

Dear amber experts,

I need some clarification on calculating the diffusion constant D from MD simulation for bilayer system using "diffusion" command in ptraj.

The command I used is :
"diffusion :1-1 5.0 average diffusion-malto-thermo1.dat"

I get to see information saying that
"DIFFUSION
Only the average results will be dumped to diffusion-malto-thermo63.dat_?.xmgr
The time between frames in psec is 5.000.
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: :63"

I get five files. I choose the file with *z extention to calculate the gradient.

For this I used gnuplot with command below:
" f(x) = b + m*x "
" fit f(x) "allxy-mal-cel-iso-bcm-diffusion_z.dat" using 1:2 via b,m "

I get values for "b" and "m" as below:
======================= ==========================

b = 1.12224 +/- 0.003892 (0.3468%)

m = 0.0337373 +/- 0.0001686 (0.4997%)

This means I get the gradient value as "0.0337373" and the y-intercept as "1.12224".

For now I want to calculate the diffusion.
So I times 0.0337373 with (10/4) since mine is bilayer (2 dimensional system).
I get "0.08434325".

Finally should I write my answer as

0.08434325 x 10^(-5) cm^2/s

Is this is the correct way of interpret the D value?

I appreciate anyone's reply to this post.

Vijay Manickam Achari
(Phd Student c/o Prof Rauzah Hashim)
Chemistry Department,
University of Malaya,
Malaysia
vjramana.gmail.com
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber

_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Jul 25 2013 - 04:00:04 PDT
Custom Search