CCL:G: Spin Contamination in Gaussian

["Theodore S. Dibble" <tsdibble#syr.edu>]

 Sent to CCL by: "Theodore S. Dibble" [tsdibble!^!syr.edu]
   The E(PMP2) results probably better, but "large" spin contamination
 can mean that the MP2 results are unreliable.  What molecular system (including
 spin, charge, and symmetry) are you doing calcuations on, and are you doing MP2
 optimizations or just single points?  What is the spin before an after
 annihilation?
    CCSD calculations automatically eliminate the effect of the first spin
 contaminant.
   - Ted Dibble
 Theodore S. Dibble
 Chemistry Department
 SUNY-Environmental Science and Forestry
 1 Forestry Drive
 Syracuse, NY 13210
 (315) 470-6596
 (315) 470-6856  (fax)
 http://www.esf.edu/chemistry/faculty/dibble.htm
 > "Decai Yu decaiyu23===yahoo.com"  wrote:
 >
 > Sent to CCL by: Decai Yu [decaiyu23() yahoo.com]
 > Dear All,
 >
 > I am doing MP2 and CCSD calculations using Gaussian.
 > I found there is large spin contamination in the calculations.
 > But after annihilation, the spin contamination is smaller.
 > For the final energy, should I use E(PMP2) or EUMP2?
 >
 > For CCSD calculation, is the E(CORR) before or after annihilation?
 >
 > Anyhelp would be greatly appreciated!
 >
 > Decai Yu
 >
 >