From: Axel Kohlmeyer (
Date: Wed Jan 08 2014 - 20:28:28 CST


On Mon, Jan 6, 2014 at 4:14 PM, Mike Makowski <> wrote:
> Hello all,
> I have a question that is similar to one posted back in 2011 regarding the
> normalization of the radial distribution function. I'm looking at a number
> of binary solutions with different concentrations. The box for each solution
> is 38 x 38 x 140 and contains between 5000 and 8000 atoms depending on the

i looked at the data your provided and it seems that your initial
density is far off for the thermodynamic state you are simulating.
your system condenses into a lammella. you cannot a g(r) that
approaches 1 for large r when you have a system with homogeneous


> concentration. When I run the RDF utility for these solutions, each one
> converges to a different value. What's more is that they are converging at
> values much higher than one. It appears that the trend is following the
> number density of the solution but when I try to normalize accordingly, it
> doesn't seem to correct the problem adequately.
> Some more information:
> 1. I'm aware that the type of selections are important. Each of the RDFs
> that I am concerned with are those that sel1 = sel2.
> 2. The box has periodic boundary conditions but have tested the result
> without PBC checked in the utility and it doesn't resolve the issue.
> 3. The max r that I'm using is smaller than half the width of my box. I'm
> using the default r = 10 A. Changing this value along with the histogram bin
> size doesn't seem to correct this problem either.
> I'm really just searching for the proper way to normalize these RDFs such
> that they all converge to one. Can anyone help me with this problem? Thanks
> for your time and consideration.
> Regards,
> Mike Makowski
> --
> Michael Makowski
> University of California, Irvine
> Department of Chemistry,
> Chemical and Material Physics,
> Irvine, CA 92617

Dr. Axel Kohlmeyer
College of Science & Technology, Temple University, Philadelphia PA, USA
International Centre for Theoretical Physics, Trieste. Italy.