CCL: What's the reasonable range for Mulliken charges on metallic cluster?

[Mikael Johansson <mikael.johansson__helsinki.fi>]

 Sent to CCL by: Mikael Johansson [mikael.johansson%helsinki.fi]
 Hello Tsung-Wen!
 On Tue, 8 Dec 2009, Tsung-Wen Yen peter308:-:gmail.com wrote:
> My caluculation of Mulliken partial charge on noble
>  metallic
>  clusters(N=38) has some peculiar results. Some of the atoms will show
>  5~6 electrons. Is this reasonable? I used (RECP19|SD) basis set for
>  Gold
>  and Copper. Could you tell me on which range should be reasonable for
>  MPA charges? I have the feeling my electron structure may have
>  converged
>  to a wrong state. I appreciated for your advice.
>  
Mulliken charges for metal systems can give you almost any random
 numbers.
 You should not use them for analysing anything important. Thus, a
 charge
 of 5-6 is not an indication of problems in your calculations, only in
 the
 way Mulliken charges are defined.
The most hilarious failure I've encountered is a charge of -78 on
 tungsten
 in WAu12:
   PCCP 6 (2004) 11-22, http://xlink.rsc.org/?DOI=B310395A
 Have a nice day,
     Mikael J.
     http://www.iki.fi/~mpjohans/