[Debichem-devel] Bug#705746: cp2k: wrong basis set definition for 6-31++G** basis sets

[Michael Banck]

Package: cp2k
Version: 2.2.426-1
Severity: serious
Tags: patch

A problem has been been found today with the 6-31++G** version of the
standard EMSL basis sets, which will lead to wrong results.

The problem has been fixed in upstream revision 12850:

http://sourceforge.net/p/cp2k/code/12850/


Michael

On Fri, Apr 19, 2013 at 10:49:14AM +0200, hutter at pci.uzh.ch wrote:
> Dear Frank
> 
> thank you very much for pointing out this problem. All basis
> sets in EMSL_BASIS_SETS of the 6-31++G** type are missing the 
> difuse set, meaning they are identical to 6-31G**. 
> I think the 6-311++G** sets are correct. We will have to
> correct this.
> 
> best regards
> 
> Juerg 
> 
> --------------------------------------------------------------
> Juerg Hutter                         Phone : ++41 44 635 4491
> Physical Chemistry Institute   FAX   : ++41 44 635 6838
> University of Zurich               E-mail:  hutter at pci.uzh.ch
> Winterthurerstrasse 190
> CH-8057 Zurich, Switzerland
> ---------------------------------------------------------------
> 
> -----cp2k at googlegroups.com wrote: -----
> To: cp2k at googlegroups.com
> From: Frank Uhlig 
> Sent by: cp2k at googlegroups.com
> Date: 04/19/2013 09:56AM
> Subject: [CP2K:4380] Inconsistency in EMSL basis sets (missing basis function in oxygen 6-31++G** basis set)
> 
> Dear CP2K'lers,
> 
> I was performing a  test calculation on a water dimer with CP2K/B3LYP/6-31++G** setup and observed a 40% relative error in the interaction energy compared to different "traditional" quantum chemistry package. After a long and tedious search for the inconsistency in my setup, I found that there is a basis function function missing in the CP2K provided basis set file (SP, exponent 0.0845000) compared to the one directly provided on the EMSL web page (see below). After including latter in the CP2K basis set definition, the relative error in the interaction energy for the water dimer is smaller than 1%, which is of course more than acceptable.
> 
> Is there any specific reason for this inconsistency? I have not checked all the basis sets, but also for the fluorine 6-31++G** basis set one basis function is missing compared to EMSL. 
> 
> Cheers,
> 
> Frank
> 
> from EMSL_BASIS_SETS:
> 
> O  6-31++Gxx 6-31++G**
>   4
>   1  0  0  6  1
>        5484.67170000          0.00183110
>         825.23495000          0.01395010
>         188.04696000          0.06844510
>          52.96450000          0.23271430
>          16.89757000          0.47019300
>           5.79963530          0.35852090
>   1  0  1  3  1  1
>          15.53961600         -0.11077750          0.07087430
>           3.59993360         -0.14802630          0.33975280
>           1.01376180          1.13076700          0.72715860
>   1  0  1  1  1  1
>           0.27000580          1.00000000          1.00000000
>   1  2  2  1  1
>           0.80000000          1.00000000
> 
> directly from EMSL basis set exchange (NWChem format):
> 
> BASIS "ao basis" PRINT
> #BASIS SET: (11s,5p,1d) -> [4s,3p,1d]
> O    S
>    5484.6717000              0.0018311        
>     825.2349500              0.0139501        
>     188.0469600              0.0684451        
>      52.9645000              0.2327143        
>      16.8975700              0.4701930        
>       5.7996353              0.3585209        
> O    SP
>      15.5396160             -0.1107775              0.0708743        
>       3.5999336             -0.1480263              0.3397528        
>       1.0137618              1.1307670              0.7271586        
> O    SP
>       0.2700058              1.0000000              1.0000000        
> O    SP
>       0.0845000              1.0000000              1.0000000        
> O    D
>       0.8000000              1.0000000        
> END
> 
> 
> 
>   
>   -- 
>  You received this message because you are subscribed to the Google Groups "cp2k" group.
>  To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
>  To post to this group, send email to cp2k at googlegroups.com.
>  Visit this group at http://groups.google.com/group/cp2k?hl=en.
>  For more options, visit https://groups.google.com/groups/opt_out.
>   
>   
>  
> 
> -- 
> You received this message because you are subscribed to the Google Groups "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+unsubscribe at googlegroups.com.
> To post to this group, send email to cp2k at googlegroups.com.
> Visit this group at http://groups.google.com/group/cp2k?hl=en.
> For more options, visit https://groups.google.com/groups/opt_out.
>