[gmx-users] Positive potential energy for TFE solvent

[Mark Abraham]

On 11/11/2011 5:07 PM, Harpreet Basra wrote:
> Hi
>
> I am trying to generate an equilibrated box of 216 TFE molecules.To
> generate the 216 TFE molecule box i performed following steps:

A suggested workflow can be found here 
http://www.gromacs.org/Documentation/How-tos/Non-Water_Solvation

>
> 1) I got the tfe.gro file and created a cubic box of edge length =
> 0.516 nm containing 1 TFE molecule (at its center), using the
> following command:
>
>>> editconf -f tfe.gro -c -o tfe_box.gro -bt cubic -box 0.516
>   I chose this length because in the tfe.gro file dimensions of the TFE
> molecule are 0.516 0.516 0.516.

That's not a good reason. Choose a volume and shape that makes sense for 
your target density. Cubic probably doesn't make sense when a 
rectangular shape is possible. Then you'll probably want to choose -nbox 
differently later.

> 2) Then using genconf command i replicated the box to get a bigger box
> with 216 TFE molecules using the following command:
>
>>> genconf -f tfe_box.gro  -o out.gro -rot -nbox 6
> 3) Energy minimization was done using STEEPEST descent
>
> 4) Then I performed 5ns NVT (300K) equilibration and followed by 5ns
> NPT (300K, 1atm) equilibration.
>
> After doing all these steps still I obtain a positive a potentail
> energy. I get positive potential energy of the system (2.45+E04
> kJ/mol) and kinetic energy as 4.03+E03 kJ/mol, thus giving a positive
> total energy of the system. My question is whether obtaining positive
> potential energy indicate some error in my TFE solvent box ? Is it
> because of large Fluorine atoms of TFE ? Does it mean its not properly
> equilibrated ? What can I do to equilibrate it?

You probably have atomic overlaps from your choice of cubic 0.516 box 
earlier. Did you look at the results of genconf before computing with them?

Mark