[gmx-users] Re: gmx-users Digest, Vol 43, Issue 49

[Gerrit Groenhof]

What you want to do, a dynamic updating of the QM atoms is not 
implemented yet.
The difficulty with such scheme is energy conservation and force 
conservation when translocating an atom from the MM into the QM region 
and vice versa.
This is a difficult problem. However, there soluitons, see for instance 
Heyden, Lin and Truhlar (J. Phys. Chem. B 2007, 111, 2231-2241).
Therefore, if such a dynamic partitioning is essential for your 
research, you could implement their ideas into gromacs. 

Gerrit



> Message: 1
> Date: Mon, 12 Nov 2007 15:10:08 -0500
> From: "Tandia, Adama" <TandiaA at Corning.com>
> Subject: [gmx-users] QMMM for element with "unknown" forcefield
> To: gmx-users at gromacs.org
> Message-ID:
> 	<73AD45172C37BB49A9BB638AD097159001E6B25B at cvcv0xi06.na.corning.com>
> Content-Type: text/plain; charset=US-ASCII
>
> Dear GMX users,
>
> I'm not very familiar with the QMMM implementation in Gromacs, but I
> have two questions. 
>
> 	1- I would like to run a MD simulation for glass materials that
> contain B and Sb based oxides. Both elements are known to be multivalent
> in the compositions that I'm interested. Is it possible to setup a QMMM
> that assigns a sphere where QMMM is performed with a fixed radius around
> Sb and B, and have the spheres to move with the assigned atoms during
> the simulation? This would be a sort of Learn-On-The-Fly method.
>
> 	2- Has anyone successfully compiled Gromacs 3.3.2 with QMMM set
> to work with Gaussian or CPMD? If yes, can I request the details of the
> compilation procedure?
>
> Best wishes,
>
> Adama
>
>  
> ==========================================
> Adama Tandia
> Modeling & Simulation
> Corning INC
> USA
>  
> Tel:  607 248 1036
> Fax: 607 974 3405
> www.corning.com
>
>