CCSDT iterations are not converging at stretched geometries

Questions
#1

Hi Team,
I am trying to run my calculations using CCSDT method but i am unable to converge the CC iteration i am stretching the bond length there are several keywords we can use for convergence failure in the manual but i am unable to understand how to use them logically
CC_DOV_THRESH, CC_PRECONV_T2Z, CC_PRECONV_T2Z_EACH, CC_SCALE_AMP, CC_THETA_STEPSIZE, CC_DIIS_START, CC_DIIS
which keywords would be beneficial to use

My input is
$molecule
1 1
N
O 1 2.5994025
$end

$rem
method=ccsdt
basis=cc-pvdz
$end

#2

Most of these keywords are irrelevant to CCSDT T amplitudes. The only relevant ones are CC_CONVERGENCE, CC_DIIS_START, and CC_DIIS_SIZE. I suggest you play with CC_DIIS_SIZE (set it to 20 or 25, for example) and check if CC amplitudes converge. In general, please refer to the keyword groupings in the manual if you have doubts on which keywords are relevant for your use case (B.1.3 $remVariable List‣ B.1 Text Input Summary ‣ Appendix B Q-Chem Quick Reference ‣ Q-Chem 6.2 User’s Manual).

1 Like