[molpro-user] t1/d1 diag value

[Li Jun]

Hi:
     It should be JPCA-2000-104-9823. Sorry.
     My key point is how reliable is the result from single-reference method ccsd(t). Sometimes, the CCSD(T) calculations are convergent, but with large t1 values, like 0.06, which is larger than the criterion 0.044 suggested by HF Shaefer III (JPCA-2000-104-9823). 
     Thanks.
best
Lijun 
At 2011-09-14 06:11:46,"Li Jun" <lijun830416 at 163.com> wrote:

Dear all:
    I am using the molpro 2010.1 version for uccsd(t)-f12 calculations. I am confused about the t1 or d1 diagnostic values. According to my experience, t1 diag can be reviewed as a criterion to determine how well the wavefunction is described by single-reference determinants. The 0.044 was suggested for open shell systems in JPCA-2000-114-9823. However, in Lee's paper, CPL- 2003-372-362,  t1 and d1 diagnostic, and t1/d ratio, were all recommended. Well, actually, it is still an open problem. Does anyone share some experience on this issue?
    Also, if not convergent in ccsd(t), the results with NOCHECK opti on are reliable?
    Thanks.
 
Li Jun


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20110914/611b3b76/attachment.html>