| 5:31:48 AM PDT - Thu, Apr 4th 2013 |
|
When running COSMO calculations with NWChem I noticed that the molecular surface and volume decrease with increasing the solvent radius which should not be observed. The obtained values are comparable to other software packages (GAMESS, DMol3) only if rsolv = 0 is used. Calculations for H2 molecule with different rsolv values demonstrated that a number of points close to the boundary between atoms is missed if rsolv>0. I guess this is because those points lie inside the cavity if r=rat + rsolv, but should be again on the surface or should be substituted if r=rat. Since the calculations of surface area and volume are based on a number of points, the calculations fail.
Does anyone know how to solve this problem if rsolv is not equal to 0?
|
|
|