Radius of Hydrogen used in $esp_fit kollman

[Hauke]

Hi,

I'm using the "$esp_fit kollman" command in TM 6.3.1 to calculate the electrostatic potential.
I know about the differences to the original approach by Kollman (increasing scaling factor of the radii vs. constant additive in TM) but my main concern is about the radius used for hydrogen. In the output is written:

Code: [Select]

Van-der-Waals radii are taken from:
 http://www.ccdc.cam.ac.uk/products/csd/radii/
[...]
  atom  radius/au   charge
   1 c   4.497548 -0.621028
   2 h   3.174740  0.155281
[...]

For carbon I can recalculate the value:  1.70 Ang in the cited table = 3.2125.. au --multiplication with 1.4 as proposed by Kollman --> 4.49754785 au

But for hydrogen I don't get the same value:
1.09 Ang to Bohr = 2.0598.. and times 1.4 = 2.88372186 au vs 3.174740 au used in TM.

The value in the output correspond to a VdW-radius of 2.26767143 au = 1.20 Ang (but according to the cited table it should be 1.09 Ang = 2.05980133 au). Is there a reason for this difference?

To see the effect of the difference radii, I made a test calculation for methane (RI-BP86/def2-TZVP):

Way	q(Carbon)	q(Hydrogen)
just $esp_fit kollman	-0.621	0.155
with explicitly giving default radii1	-0.621	0.155
with radii of the cited table2	-0.684	0.171

¹ using $vdw_radii c 3.21253429 h 2.26767143
² using $vdw_radii  c 3.21253429 h 2.05980133

As expected the charges using the default radii explicitly agree perfectly with the calculation where no radii values where given. But if I use the values of the cited table I get quite some differences. I don't know if there might be a good reason to change the VdW-radius of hydrogen (or maybe the cited table was updated after the values where programmed into TM). Maybe somebody can give a comment on this.
Thanks in advance.