[gmx-users] Coulomb term calculation in gromacs

[Justin Lemkul]

On 11/26/19 5:54 AM, nikolaev at spbau.ru wrote:
> Dear Justin,
>
> Thank you very much for your answer. I have a few more questions.
>
>
> 1)
>
>>> I have a question about the calculation of the Coulomb term in Gromacs.
>>> As I understand, two terms are involved: Coulomb 1-4, which is
>>> calculated
>>> among the 1st and 4th atoms, and the term for the rest of the
>>> interactions.
>>>
>>> In the manual the standard Coulomb formula is given with f = 138.935...,
>>> is it applied for the Coulomb 1-4 interactions?
>> Electrostatics are calculated the same way everywhere (within the
>> short-range cutoff), unless the force field applies a scaling factor to
>> this interaction (fudgeQQ).
>>
> So, within the short-range cutoff all interactions between non-bonded
> atoms are calculated with the Coulomb formula with f = 138.935.., am I
> correct? I.e. the interactions between 1-4 atoms are calculated in the
> same way? Why does .psf file include the "pairs" list for the 1-4 atom
> pairs?

You can look up the exact constants in src/gromacs/math/units.h

Everything follows the normal Coulomb Equation. As I said before, 
[pairs] are 1-4 interactions that may or may not be scaled by some value 
that is intrinsic to the force field. The value (fudgeQQ in the case of 
electrostatics) is a prefactor for the calculation of the charge-charge 
interaction via Coulomb's Law.

> 2)
>
>>> If I use the electrostatic scheme Cut-off, then the usual Coulomb
>>> formula
>>> is applied for all atoms within the cutoff, excluding the 1-2, 1-3 and
>>> 1-4
>>> bonded atoms?
> If I apply the Cut-off scheme, how is the electrostatics calculated beyond
> the cut-off?

A plain cutoff? Everything is zero beyond the cutoff. Doing this is 
wildly inaccurate and shouldn't be done in modern simulations.

> 3)
>
>>> If I use the PME scheme, then interactions between all atoms within the
>>> rlist is calculated, excluding the 1-2, 1-3 and 1-4 bonded atoms, by
>>> applying the Ewald sum?
>> Interactions within rlist (direct space) are computed by the normal
>> (shifted) Coulomb potential.
>>
> So, does it mean that in case of PME scheme only the electrostatics beyond
> the cutoff is calculated with PME formulae?

The "PME formulae" may refer to both parts of the calculation so to be 
clear: the real-space is within rcoulomb and is the normal Coulomb 
Equation with a shifted potential (see the GROMACS manual). The Ewald 
summation is done in Fourier space (beyond rcoulomb).

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Assistant Professor
Office: 301 Fralin Hall
Lab: 303 Engel Hall

Virginia Tech Department of Biochemistry
340 West Campus Dr.
Blacksburg, VA 24061

jalemkul at vt.edu | (540) 231-3129
http://www.thelemkullab.com

==================================================