[gmx-users] implicit solvent model: formulation of non-polar free energy

[Chi-cheng Chiu]

Hi everyone,
I am going through the source code for the GB calculation part of gromacs,
and found the following code in "src/mdlib/genborn.c" :
in the function  "real calc_gb_nonpolar()"

rai       = top->atomtypes.gb_radius[md->typeA[ai]];
rbi_inv   = fr->invsqrta[ai];
rbi_inv2  = rbi_inv * rbi_inv;
tmp       = (rai*rbi_inv2)*(rai*rbi_inv2);
tmp       = tmp*tmp*tmp;
e         = factor*term*(rai+probe)*(rai+probe)*tmp;
dvda[ai]  = dvda[ai] - 6*e*rbi_inv2;
es        = es + e;

so from this source code, energy seems to be (rai+probe)^2 * (rai/rbi)^6
where rbi is the born radius, rai is the atom radius, and probe is 0.14 nm.
I checked the reference Schaefer et al. JMB 284(3):835–848, 1998 which is
mentioned in the manual.
It seems the reference used (rai+probe)^2 * (rai/rbi)
where (rai/rbi) term don't have power of 6 and disagree with the source
code.
Is there a reason using power 6?

Chi-cheng
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110328/0ea411ee/attachment.html>