I’m not clear about how to write a solvent_data file to specify other solvent in SMD model. In the manual, here says: “The choice Solvent = other requires an additional free-format file called solvent_data that should contain the float-point values of the following solvent descriptors: Dielec, SolA, SolB, SolN, SolG, SolC, SolH” without providing any example.
I tried to create a file named “solvent_data” in the same working dir, but seems not properly work. Hope someone could tell me the right way to use it:) Here are my files
*********Input file:(geometry omitted)
$molecule
0 1
[geometry]
$end
Thank you for your kind answer! I’d like to know that for SMD model, is it necessary to provide so many arguments? For solvent excluded by the internal list and the doc mentioned in the manual, we usually have difficulties getting valid data other than dielectric constant and refractive index. Depending on the task type, like dielectric constant is likely to be the main or even only parameter to influence single point result. Thus, I may choose to use dielectric constant and refractive index of the new solvent but using the other values of a similar solvent which I could get from document for convenience. I wonder if it’s reasonable. Actually, I notice that researchers usually specify only “eps” and “epsinf” which are exactly or derived from dielectric constant and refractive index to use smd model in Gaussian.
The solvent’s static dielectric constant is enough to specify the electrostatic part of the solvation energy, then for some vertical excitation models one also needs the optical dielectric constant, which is equal to the square of the index of refraction. (In Q-Chem, those vertical excitation models are implemented for PCM but not for SMx solvation models.) The remaining parameters are for the nonelectrostatic contributions to the SMx solvation models (cavitation, dispersion, etc.) This is discussed in Section 4 of my review of continuum solvation models: https://doi.org/10.1002/wcms.1519
If you see that only “eps” (presumably, static dielectric constant) and “epsinf” (infinite-frequency or optical dielectric constant) are being specified, then that looks like an electrostatics-only model such as a PCM without any additional nonelectrostatic terms.
For future reference, starting with Q-Chem v. 6.0.1 the ‘solvent_data’ file will be eliminated and the solvent descriptors (for SOLVENT = OTHER) will instead be specified in the $smx section of the input file. I have just committed that change to the code and to the manual for the next release.