[gmx-users] Adsorption energy of a single molecule

[Mark Abraham]

Darrell Koskinen wrote:
> Hi Mark,
> No, I am not trying to build a random number generator and hope that I 
> am not moving in this direction : )
> 
> My definition of adsorption energy would be the difference in energy 
> when two species are at an infinite distance and the energy when these 
> species are at an equilibrium distance from each other.

OK, but I think you need to measure a free energy change, using (for 
example) a PMF generated by the pulling code in GROMACS.

> I think the way in which I can measure this energy difference would be 
> to subtract the LJ (LR) energy between an ammonia molecule and the 
> graphene sheet (where the distance between the molecule and the graphene 
> sheet is large) from the LJ (SR) energy between an adsorbed ammonia 
> molecule and the graphene sheet.

That assumes an unrealistic independence of the forcefield 
contributions. Why would energy not partition over bonded and Coulombic 
interactions also?

Also, the distinction between LR and SR is merely to do with your 
twin-range cut-off choices (read manual!). The adsorption process is a 
transition from no interaction (i.e. outside the LR cutoff) to physical 
contact. The total LJ interaction energy (if that means anything!) will 
be partitioned between LR and SR according to distance from the surface.

> I am not aware that OPLS-AA was parameterized to produce such values, 
> but would appreciate your comments.

I know nothing about OPLS-AA, but would guess that it wasn't 
parameterized for graphite, gaseous ammonia or to reproduce adsorption 
energies. I think you'd have to demonstrate (through simulation or 
literature reference) suitable behaviour of graphite and gaseous ammonia 
  on their own before trying to simulate them together. If there's some 
kind of experimental observable about the behaviour of adsorbed ammonia 
on graphite that you could reproduce, that would be a useful indicator 
that the force field is doing OK. Absent that, you might just have that 
random number generator :-)

Mark

>> Message: 2
>> Date: Thu, 12 Nov 2009 10:42:24 +1100
>> From: Mark Abraham <Mark.Abraham at anu.edu.au>
>> Subject: Re: [gmx-users] Adsorption energy of a single molecule
>> To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>> Message-ID: <4AFB4BE0.1020108 at anu.edu.au>
>> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>>
>> Darrell Koskinen wrote:
>>  
>>> Hi Justin,
>>> So when I create the index file with the molecule and the graphene 
>>> sheet and run g_energy, is the adsorption energy between the adsorbed 
>>> molecule and the graphene sheet the LJ (SR) energy?
>>>     
>>
>> Not necessarily, unless the force field was parameterized to produce 
>> such values. First, seek to define what you mean by adsorption (free) 
>> energy, then consider how a simulation might measure that. Then, 
>> consider whether any force fields exist that might achieve that. Then, 
>> test it on a case where you have experimental data.
>>
>> Other procedures run a much higher risk of producing an expensive 
>> random number generator :-)
>>
>>  
>>> And will this energy be just the interaction energy between the 
>>> adsorbed molecule and the graphene sheet (i.e. does the creation of 
>>> the new index file and executing -rerun cause g_energy to exclude the 
>>> energies between the other molecules and the graphene sheet)?
>>>     
>>
>> Creating an energy group causes mdrun -rerun to report the fraction of 
>> the total nonbonded energies that are due to intra- and inter-group 
>> interactions. See manual section 3.3
>>
>> Mark
>>
>>  
>>>> Date: Sun, 01 Nov 2009 21:27:19 -0500
>>>> From: "Justin A. Lemkul" <jalemkul at vt.edu>
>>>> Subject: Re: [gmx-users] Adsorption energy of a single molecule
>>>> To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>>>> Message-ID: <4AEE4387.40607 at vt.edu>
>>>> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>>>>
>>>>
>>>>
>>>> Darrell Koskinen wrote:
>>>>  
>>>>      
>>>>> Hi Justin,
>>>>> Since I cannot use the original .tpr file, then do I need to run 
>>>>> grompp with the new index file to create a "new" .tpr file, 
>>>>> "mdtopolnew.tpr"? I assume I then need to modify the energygrps 
>>>>> line within the .mdp file to include these new energy groups and 
>>>>> then execute "mdrun -rerun mdtraj.trr -s mdtopolnew.tpr". Is this 
>>>>> correct?
>>>>>
>>>>>             
>>>> Yes, that is the point of the -rerun function.
>>>>
>>>>  
>>>>      
>>>>> Does executing "mdrun -rerun mdtraj.trr -s mdtopolnew.tpr" cause a 
>>>>> molecular dynamics simulation to run once again? If so, then will 
>>>>> the exact same molecules that adsorbed on the first MD run be the 
>>>>> same ones that adsorb on the second MD run?
>>>>>
>>>>>             
>>>> The original coordinates are used to re-calculate the energies.  No 
>>>> new simulation is performed.
>>>>
>>>> -Justin
>>>>
>>>>  
>>>>      
>>>>> Thanks.
>>>>>
>>>>> Darrell
>>>>>
>>>>> *********************************************************************** 
>>>>>
>>>>> *********************************************************************** 
>>>>>
>>>>>
>>>>>
>>>>> Date: Sun, 01 Nov 2009 19:29:36 -0500
>>>>> From: "Justin A. Lemkul" <jalemkul at vt.edu>
>>>>> Subject: Re: [gmx-users] Adsorption energy of a single molecule
>>>>> To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>>>>> Message-ID: <4AEE27F0.50301 at vt.edu>
>>>>> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>>>>>
>>>>>
>>>>>
>>>>> Darrell Koskinen wrote:
>>>>>
>>>>>           
>>>>>>> Hi Justin,
>>>>>>> In my simulation, I have just over 100 ammonia molecules and, of 
>>>>>>>                     
>>>>>> these > molecules, 10 to 20 adsorb onto the graphene sheet.  I 
>>>>>> initially thought > that I would need to tag each one of the 
>>>>>> ammonia molecules, since I > would not know, in advance of running 
>>>>>> the simulation, which of these > molecules would adsorb.
>>>>>>               
>>>>>>>> Are you suggesting that, after the initial simulation run, which 
>>>>>>>>                         
>>>>>> uses an > index file "index.ndx" and is comprised of three 
>>>>>> components {System, > Grph, NH3}, I run make_ndx again to assign 
>>>>>> each adsorbed molecule to its > own index, execute "mdrun -rerun 
>>>>>> mdtraj.trr -s mdtopol.tpr", where > mdtraj.trr and mdtopol.tpr are 
>>>>>> the output of the initial simulation, and > then use g_energy to 
>>>>>> analyze the energies between the groups?
>>>>>>               
>>>>>>>                       
>>>>> Almost.  You can't use the original .tpr file if you are trying to 
>>>>> establish new energy monitoring groups.  Hence the point of making 
>>>>> a new .mdp file.  So, re-run the old trajectory with the new .tpr 
>>>>> file to take advantage of the new groups.
>>>>>
>>>>> -Justin
>>>>>
>>>>> -- ======================================== Justin A. Lemkul Ph.D. 
>>>>> Candidate ICTAS Doctoral Scholar Department of Biochemistry 
>>>>> Virginia Tech Blacksburg, VA jalemkul[at]vt.edu | (540) 231-9080 
>>>>> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin 
>>>>> ========================================
>>>>>