[gmx-users] Potential energy for partial system

[Po-Han Lin]

Hi all,

I been working on a thin film system. To avoid the surface effect, I 
would like to collect potential energy data for the central region of 
the thin film by excluding near-surface & near-substrate molecules.  
However, I cannot find any straight forward way to achieve that purpose 
by using energy group and g_energy. One of my colleagues came up with a 
recipe to remedy this problem. For me, this recipe makes sense and gives 
me reasonable quantities. However, I would like to hear comments from 
the community as well as developers in case I miss something. The recipe 
outlines as follows.

1. Create an index.ndx that includes a group consisting of all the 
molecules in the central region. Say the group is called "CENTER".
1. Run a MD run for the full system with a energy group "CENTER".
2. Use g_energy to obtain the LJ and coulomb interactions between 
"CENTER" and the rest system.
3. Use trjconv with index.ndx to create "center.trr" and "center.gro" 
that only contain the trajectory and starting configuration of the 
central region.
4. Modify the topology file and save as "center.top". Specifically, 
change the number of molecules to just the number of those in the 
central region.
5. grompp -f mdpfile -c center.top -p center.gro -t center.trr
6. mdrun -rerun center.trr
7. Use g_energy to obtain the potential energy
8. Total potential energy can be acquired by combining the energies from 
step 3 (intermolecular energy between center and the rest) and step 8( 
all intramolecular interactions & intermolecular interactions among the 
atoms in center region).

Thanks in advance.
Po-Han Lin