VMD-L Mailing List

From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Tue Mar 08 2011 - 06:23:07 CST

On Tue, Mar 8, 2011 at 6:21 AM, Sebastian Maximilian Wilhelm
<Sebastian.Wilhelm_at_rwth-aachen.de> wrote:
> Hi,
>
> I am wondering about two different calculated results for the electrostatic energies between two atom sites:
>
> If I am calculating the electrostatic energy via 'measure energy elect {123 124} q1 -0.82 q2 0.41' directly in the VMD console, I receive a value, which I can reproduce manually with a relative dielectric epsilon_r = 1. I know, that this assumption (epsilon_r = 1) isnīt physically correct, especially if there is a great distance and thus probably other molecules between index 123 and 124.

a dielectric constant of 1 is the correct value. why should it be wrong?

> Anyway, if I calculate the same type of energy via NAMDenergy between index 123 and 124 with the same charges as above and the initially loaded values for cutoff-distance = 12 and dielectric = 1, the result differs from the result of VMD, especially for great distances between the analyzed indices. The absolute values of the results are always smaller than the absolute ones from VMD.
>
> To me, the only difference between these two calculation methods (U1=q1*q2/(4*pi*epsilon0*r) and U2=epsilon14*C*q1*q2/r (NAMD Userīs guide)) is the epsilon14. I calculated the ration between U1 and U2 in order to get the epsilon14. The result was: the greater the distance between index 123 and 124, the smaller the epsilon14 and thus the U2. But in the Userīs guide is written, that the epsilon14 has to be a constant.
>
> My question is: why is there a difference between these two methods and which one would you recommend?

the code in VMD does a simple coulomb that is shifted so that the
energy is zero at cutoff.
this would have a discontinuity in the forces around the cutoff and
thus would not be a good
choice for MD. with NAMDenergy, you use the CHARMM smoothing function
that also gives
you a continuous force at the cutoff.

i don't think it makes much of a difference for any analysis, since
whatever you compute
with either method will be tainted by the errors introduced by the
force field in itself, so
must not be overinterpreted in any case. it mostly indicates that you
have to stay consistent.

> By the way: the calculation of vdw-energies via 'measure energy vdw {123 124} rmin1 ... rmin2 ... eps1 ... eps2 ...' and NAMDenergy always produce the same results.

no surprise, since vdw is not as long-range as coulomb, so the shifted
energy/force is not needed.

axel.

> I will appreciate your answer!
>
> Best regards,
> Sebastian
>
> 

-- 
Dr. Axel Kohlmeyer
akohlmey_at_gmail.com  http://goo.gl/1wk0
Institute for Computational Molecular Science
Temple University, Philadelphia PA, USA.