2 For Br and I, I recommend to modify radii as mentioned in my blog article http://sobereva.com/613. For other elements, simply use the default radii in SMD should be OK.

For charged ions, uESE performs notably better than SMD, see http://sobereva.com/593. If you need better solvation energy of ions, using uESE instead is recommended.

3 I didn't take care of this parameter and thus cannot comment on it.

I have two questions on the G16 input parameters for the SMD model, and one separate question on CPCM.

1. The original SMD paper (J. Phys. Chem. B 2009, 113, 6378-6396) states that:

"... In the SMD model, the boundary is defined to enclose a superposition of nuclear-centered spheres with radii rho_k, which are called intrinsic Coulomb radii. The values depend only on the atomic numbers of the atoms. This boundary forms a so-called solvent-accessible surface (SAS). ..."

also,

"... The options surface=sas and noaddsph were used to generate the solvent-accessible surface of the solute molecule defined only by the superposition of nuclear-centered spheres. ..."

Do these mean that, in order to reproduce the results of SMD optimized by the authors, I should use both SCRF=(read,SMD) and Surface=SAS? I am not sure if this is true, or if this is no longer needed after the SMD keyword is implemented in Gaussian.

2. I notice that several papers assert that Pauling atomic radii option is preferred for the highest accuracy of IEFPCM. I also notice one paper (J. Phys. Chem. A 2019, 123, 9498-9504) that asserts that SMD model plugged with Bondi radii performs better on charged ions.

Do you have an opinion on the generally preferred atomic radii option for SMD model (such as default SMD radii, Pauling radii, or Bondi radii)?

3. This is a separate question on CPCM. The paper (JCTC 2015, 11, 4220-4225), which is quoted by your blog article (http://sobereva.com/327), provides the desired x_feps parameters for maximum agreement with IEFPCM.

However, to my understanding, it is not clear (from the Methods section of this paper) which atomic radii and solute cavity surface options are commonly used. Should I assume that the paper used the default UFF radii and vdW cavity?

Additionally, do you think that the same x_feps parameters recommended by this paper cannot be applied to ORCA because ORCA uses Bondi radii by default?

Thank you very much in advance.