CCL:G: Pure vs. Cartesian Basis Functions

 Sent to CCL by: Brian Salter-Duke [brian.james.duke\a/gmail.com]
 Fu-Ming Ying fmying * gmail.com wrote:
 
 Sent to CCL by: Fu-Ming Ying [fmying||gmail.com]
 hej
 I don' think the result will differ much for geometry optimization.
 But if you'd better use the same type during all the calculations.
For example, if you use 5D in hf/3-21G, then I think it's better to use 5D in b3lyp/6-31++G(D,P) etc.
I think this misses the point. You should use the appropriate type that the basis set is defined for. Look at the defined defaults in the manual - pg 27 for G03 manual. 3-21G and 6-31G use 6D, while 6-311G uses 5D. If you go from 6-31G to 6-311G you should stick to the defaults, changing 
 from 6D to 5D.
 
 Brian.
 
 gnli gnli,dicp.ac.cn wrote:
 
 *Dear CCLer,*
 *Please forgive me for my simple question.*
*Is it necessary to keep the basis set to be Pure or Cartesian type during all the calculate step in one system?*
*For example, we optimize a molecule using “hf/3-21g hf/6-31g b3lyp/6-31g b3lyp/6-31++g(d,p) b3lyp/aug-cc-pvdz” gradually, need we define the Basis Functions all in 5D 7F or 6D 10F in the route section consistently, or just using default? Is the difference serious? * 
 *Thanks!*
 best wishes
 Guanna Li
 gnli]|[dicp.ac.cn <mailto:gnli%5D%7C%5Bdicp.ac.cn>
 
 --
        Brian Salter-Duke (Brian Duke)    Brian.James.Duke(a)gmail.com
           Post: 626 Melbourne Rd, Spotswood, VIC, 3015, Australia
        http://www.salter-duke.bigpondhosting.com/brian/index.htm
 Honorary Researcher Fellow, Monash Institute of Pharmaceutical Sciences