TURBOMOLE Users Forum

TURBOMOLE Modules => Ridft, Rdgrad, Dscf, Grad => Topic started by: toni on April 30, 2008, 10:10:13 AM

Title: False vdW Parameter R_0 for Iodine in Dispersion Interaction (DFT+D)?
Post by: toni on April 30, 2008, 10:10:13 AM

In the literature, the van der Waals parameter for I is 1.892 Å [S. Grimme, J. Comp. Chem. 27, 1787 (2006)].
The Turbomole output, however, claims that the value used is 3.2503290 (1.720 Å) in the DFT+D.
Which one is correct,
the literature value (which seems to be the same as for element Te)
or the implemented value (which seems a bit small when compared to series F-Cl-Br: 1.287-1.639-1.749)?

Title: Re: False vdW Parameter R_0 for Iodine in Dispersion Interaction (DFT+D)?
Post by: cml on May 06, 2008, 09:27:54 AM

Hello all,

 let me shortly explain the discrepancy between the vdW radii in the
second DFT-D paper (S. Grimme, J. Comput. Chem. 2006, 1787) and
in the Turbomole output:

The paper states that the calculated radii are scaled in the DFT-D
method with a factor of 1.10 (1.22 in the previous work). Unfortunately,
Turbomole prints the unscaled values before it evaluates the dispersion
energy while the paper lists the scaled values.
When you compare 1.892 with 1.72 Angstrom, the ratio is 1.10.

The energy and gradient calculations are always done with the
correct (scaled) vdW radii of DFT-D.

Best regards,

Christian