[gmx-users] Re: What is epsilon_r?

[David van der Spoel]

On Apr 18, 2012, at 8:24 PM, "Andrew DeYoung" <adeyoung at andrew.cmu.edu> wrote:

> Hi,
> 
> Thank you, Justin, for your time!
> 
> By reading the post that you sent
> (http://lists.gromacs.org/pipermail/gmx-users/2004-March/009650.html), it
> seems that epsilon_r is just a scale factor.  I interpret this as meaning
> that if, for example, I use epsilon_r = 2, then Coulomb interactions will be
> scaled down by a factor of 2.  That is, the Coulomb potential
> (1/(4*pi*epsilon0))*(q_i q_j)/|r_i - r_j| will be multiplied by 1/2.
> 
> But even if my interpretation is correct, what does this mean?  epsilon_r
> does not seem to be a reaction field parameter, since the value of epsilon_r
> affects the electric potential that I calculate even though I am not using
> implicit solvent, nor am I using one of the reaction field electrostatics
> methods.  Or, a different way to phrase this is, if most or all modern force
> fields are parametrized based on epsilon_r and if one is not using any sort
> of implicit or coarse-grained solvent or applying a reaction field
> electrostatics method, then why would anyone want to scale down the Coulomb
> potential?
> 



epsilon_r should never be used with a regular force field because the balance between coulomb and other terms will be wrong. maybe we should remove the term.

> Also, I am wondering about the boundary condition for the Ewald method.  It
> seems that, for the Ewald method, one must specify the dielectric constant
> at the boundary.  For example, the text on liquid simulations by Allen and
> Tildesley says (page 157), "... we must specify the nature of the medium
> surrounding the sphere, in particular its relative permittivity (dielectric
> constant).  The results for a sphere surrounded by a good conductor such as
> a metal (epsilon = infinity) and for a sphere surrounded by vacuum (epsilon
> = 1) are different [de Leeuw, Perram, and Smith 1980]."
> 
> My question is, what epsilon parameter in Gromacs serves this purpose.  One
> possibility is epsilon_surface.  However, this does not seem to be the right
> parameter, because the manual entry for epsilon_surface says, "Turn it on by
> setting it to the value of the relative permittivity of the imaginary
> surface around your infinite system."  This wording almost seems to imply
> that it is a dipole correction (not a dielectric boundary condition), and
> that there is some other parameter that, in fact, controls the "relative
> permittivity of the imaginary surface."  Is this true?  If so, what is this
> parameter?
> 
> Thank you for your time! 
> 
> Andrew DeYoung
> Carnegie Mellon University
> 
> Andrew DeYoung wrote:
>> Hi,
>> 
>> What is the parameter epsilon_r mentioned in the manual
>> (http://manual.gromacs.org/current/online/mdp_opt.html#el)?  The manual
> says
>> that it is the relative dielectric constant.  My initial thought was that
>> this would only be relevant for reaction-field electrostatics, or
> somewhere
>> where implicit solvent is used.  However, it seems that epsilon_rf (not
>> epsilon_r) is used for reaction-field purpose.
>> 
>> I am not using implicit solvent in my system, nor am I using a
>> reaction-field method (I am using PME electrostatics); in my system, I am
>> using an all-atom description.  Then, does epsilon_r generate "additional"
>> dielectric, beyond what naturally arises from the all-atom description of
>> the solvent?  
>> 
>> Even though I am not using a reaction-field method, epsilon_r is clearly
>> affecting something, because when I use g_potential to calculate the
>> electric potential, the results vary considerably depending on the value
> of
>> epsilon_r that I use in my .mdp file.
>> 
>> Do you have any thoughts about what epsilon_r does in an all-atom
> descripton
>> of a solvent?  Does it provide "additional" charge screening?
>> 
> 
> http://lists.gromacs.org/pipermail/gmx-users/2004-March/009650.html
> 
> -Justin
> 
> -- 
> ========================================
> 
> Justin A. Lemkul
> Ph.D. Candidate
> ICTAS Doctoral Scholar
> MILES-IGERT Trainee
> Department of Biochemistry
> Virginia Tech
> Blacksburg, VA
> jalemkul[at]vt.edu | (540) 231-9080
> http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin
> 
> ========================================
> 
> -- 
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting!
> Please don't post (un)subscribe requests to the list. Use the 
> www interface or send it to gmx-users-request at gromacs.org.
> Can't post? Read http://www.gromacs.org/Support/Mailing_Lists