From: JC Gumbart (gumbart_at_ks.uiuc.edu)
Date: Sat Jan 28 2012 - 15:47:54 CST
A non-neutral charge normally leads to a divergent term in the in the PME
formulation due to the interactions between the infinite periodic copies.
However, this term, which I believe is the k=0 term in reciprocal space (?),
is typically ignored in MD and thus the simulation doesn't blow up. This is
sometimes called effectively a "neutralizing plasma", if I recall correctly.
So one can simulate non-neutral systems using PME. However, the
self-interaction between the ion and its copies must be accounted for in the
free energy. This is addressed on page 18 of the tutorial and in reference
15. But just as you suspected, and explicitly noted on page 31, this
correction is the same for two legs of the thermodynamic cycle and thus
cancels out.
-----Original Message-----
From: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] On Behalf
Of Thomas Albers
Sent: Saturday, January 28, 2012 1:35 PM
To: namd-l_at_ks.uiuc.edu
Subject: namd-l: Alchemical free-energy methods and electroneutrality
Hello!
I seem to remember that systems simulated with periodic boundary conditions
must be electroneutral, or systematic errors will be introduced through PME
method.
There's the Alchemical Free-Energy tutorial, where one of the examples is
the potassium ion-18-crown-6 complex. In the tutorial, the potassium ion is
made to vanish, but not its countercation. Does the error introduced
through non-electroneutrality cancel itself out through the thermodynamic
cycle, or should for the correct treatment the counter-anion also be made to
vanish?
Thomas
This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:21:09 CST