[gmx-users] epsilon_rf

[Volker Knecht]

Dear GROMACS users,

in the past I tended to use PME to calculate electrostatic interactions, 
but to be able to reach larger timescales I would like to use reaction 
field now whenever appropriate. I am wondering about an appropriate 
value for the dielectric constant for a reaction field correction for 
simulations of peptides in bulk SPC water or an SPC/bilayer interface. 
Whereas some (group of van Gunsteren) have used the self-consistent 
value for SPC water in simulations of pure SPC water, 54, most people 
seem to use the experimental value of the dielectric permittivity of 
water at room temperature, 78 (leading to an inconstent model). Often, 
authors using reaction field do not even give the value for epsilon_rf 
chosen. Anezo et al. have shown that both epsilon_rf = 54 and 80 yield 
similar properties for lipid bilayer systems, and (besides PME) reaction 
field was one of the recommended options for calculating electrostatic 
interactions for bilayer systems. Strictly spoken, reaction field is 
only appropriate for a bulk system, but it is widely used for 
simulations also of bilayer systems for practical reasons. For a bilayer 
system, in principle, it is even less clear which value for epsilon_rf 
should be chosen, since molecules in the interface see environments with 
both very high and very low (close to 1) dielectric permittivity.

I have the feeling that very different opinions on the usage of reaction 
field and epsilon_rf for biomolecular simulations, ranging from rigorous 
to pragmatic, coexist in the community, and it would be nice if some of 
you could provide their opinion.

Kind regards,
Volker

-- 
Dr. Volker Knecht
Department of Theory and Bio-Systems
Max Planck Institute of Colloids and Interfaces
Research Campus Golm
14424 Potsdam, Germany
http://www.mpikg-golm.mpg.de/th/people/vknecht/index.html

Phone: +49-331-567-9610
Fax: +49-331-567-9612