Re: Cutoff radius for electrostatic embedding in QM/MM calculations

From: Agisilaos Chantzis (agisilaos.chantzis_at_gtn.ai)
Date: Mon Sep 10 2018 - 12:41:25 CDT

Dear Marcelo,

Thank you very much for your feedback. It makes sense to apply the same
cutoff for the QM/MM non-bonded interactions as the one used for the
classical calculation. So from what you said, in order to include all
the interactions of the unit cell explicitly then the value of the
cutoff (cutoff keyword in the NAMD input file) should be adjusted so as
to take all the molecules of the unit cell, is that right? Moreover, the
electrostatic interactions that are beyond the cutoff value are taken
into account using PME when solving the QM electronic structure problem,
right?

Thank you once more!

Agisilaos

On 10/09/18 17:53, Marcelo C. R. Melo wrote:
> Dear Agisilaos,
>
> The cutoff used in the QM/MM code  in NAMD is the same as the cutoff
> defined for non-bonded interaction in the classical calculation, so
> changing that value will change the cutoff for both MM non-bonded
> interactions and QM/MM point charge determination.
> All QM codes we used so far take some sort of point charge information
> in order to run their calculations in an electrostatic embedding. your
> question regarding the cutoff is very pertinent, and all I can tell
> you is that it will depend on the error you tolerate. One of the
> methods used to reduce the error of a hard cutoff is applying a
> smoothing function of the point charges (which NAMD provides, with
> several options). Another one is to simply increase the cutoff and
> allow the natural 1/r part of the potential do the smoothing for you
> (some people add *all* of the classical system in their QM calculations).
> In this field, it is hard to find one recipe for all applications, but
> I hope this helps.
> Best,
> Marcelo
> ---
> Marcelo Cardoso dos Reis Melo
> PhD Candidate
> Luthey-Schulten Group
> University of Illinois at Urbana-Champaign
> crdsdsr2_at_illinois.edu <mailto:crdsdsr2_at_illinois.edu>
> +1 (217) 244-5983
>
>
> On Wed, 5 Sep 2018 at 10:22, Agisilaos Chantzis
> <agisilaos.chantzis_at_gtn.ai <mailto:agisilaos.chantzis_at_gtn.ai>> wrote:
>
> Dear all,
>
>
> In the QM/MM implementation of NAMD it is documented that in the
> electrostatic embedding of QM part an effective cutoff radius is used
> (rmax) to decide which MM partial charges will be included in the QM
> calculations. What I could not really find documented is how this
> cutoff
> is chosen and whether it can be modified so as to asses it's
> influence
> on the accuracy of the calculations. Since the long range
> electrostatic
> interactions should not be abruptly cut in general, how is the use of
> this rmax cutoff justified in the description of the electrostatic
> environment of the QM part?
>
>
> Sincerely,
>
>
> Agisilaos Chantzis
>

-- 
Dr. Agisilaos Chantzis, Ph.D.
Computational Quantum Chemist
GTN LTD
338 City Road, London
e-mail: agisilaos.chantzis_at_gtn.ai
United Kingdom

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2018 - 23:21:23 CST