Calculate the free energy of desorption?

[jianxl...@gmail.com]

Dear users.

My name is Lian and I am a new PLUMED user. I am simulating the desorption of an alkyl chain from graphene surface with the GROMACS 2.6 software package. So, during the MD simulation, the alkyl chain is moving away from graphene surface.

I already calculated the free energy profile using the Umbrella simulation method (potential of mean force) within GROMACS. However, this method takes a lot of time, and requires to do plenty of lengthy MD simulations. In addition, the Umbrella simulation does not guarantee that the simulation time is long enough to have the free energy profile well converged.

Now, I would like to know how to plot the desorption free energy using PLUMED.  I guess that the collective variable of interest would be the distance between the center of mass of the molecule and the surface.
What I know from the tutorial is that the sum_hills executable allows for building the free energy surface (FES). However, I don't know how to extract the free energy profile (free energy vs distance) from the FES.

More specifically my question is. "How to plot the free energy profile using all the FES.DAT files?"

Thanks a lot for your help !

Lian

[jianxl...@gmail.com]

There is another point I would like to know.
To study the desorption of a molecule, is there any other CV I should take into account (apart from the distance between the molecule and the surface)?

Since I am just interested in the profile of the free energy with respect to the distance, I suppose that only one CV is sufficient. Am I right?

Thanks.
Lian

[joshua....@gmail.com]

It sounds like you're using only one collective variable,  hence the FES.DAT file contains only two columns and can be plotted using any appropriate graphing software.  How about something simple like 'gnuplot'?

[jianxl...@gmail.com]

Dear Joshua.

First, thank you for your reply.
I thought the same as you but I am not sure that the first column in FES.DAT corresponds to the CV values.

The reason is that I calculated the distance between my molecule and the surface using G_DIST program included in GROMACS package. I also monitored the distance within plumed (in COLVAR file), and found that the distances correspond more or less to those calculated within GROMACS (going from 0.3 to 3 nm). However, these values do not correspond to the first column of FES.DAT at all. In FES.DAT, the distances go from -1 to 4 !

When I plotted the data in FES.DAT with gnuplot, I obtained a symmetric U-shaped curve.
On the other hand, my previous calculations with Umbrella simulation have shown that  the free energy profile should look like a "Lennard-Jones potential" curve. I mean I observed a steep decrease of the potential corresponding to a repulsive region before the equilibrium distance, followed by an increase of the potential (attractive part).

So my question is, what means the first column in FES.DAT? Which options I should use for SUM_HILL in order to have the proper distance in the FES.DAT file?

Thanks a lot for the help. 

Best regards. 
Lian

[jianxl...@gmail.com]

Dear users.

Can anyone of you say something about the free energy with respect to DISTANCE between two different molecules?

I changed my MD simulations. Now, the molecule is moving toward (periodic) graphene surface and I obtained a more reasonable free energy profile.
(Why moving the molecule toward or away from the surface leads to two different FES.DAT ?)
However, as the distance is going to zero, the free energy is not diverging (see attached file). This means that the molecule is able to pass through the surface.


I tried PLUMED with either a lower WALL at LIMIT = 0 with KAPPA up to 50000, and/or ALIGN_ATOMS for graphene since the surface is periodic (infinite) along x-y plane, but they did not change anything.

Please, can someone tell me which directive should be used instead? Do I need to define another CV in order to study the adsorption/desorption of a molecule on a surface?

Thanks for your consideration.
Lian