-
|
Hello, I recently started working with PySCF to calculate ionization potentials with IP-EOM-CCSD, but I have a question regarding the use of partition (matrix-partitioning for the doubles-doubles block). Which method would be considered more accurate, "full" (full diagonal elements) or "none"? Thanks! |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
The three partitioning options are |
Beta Was this translation helpful? Give feedback.
-
|
I see, now I understand! Thank you for your help! |
Beta Was this translation helpful? Give feedback.
The three partitioning options are
None,'full', and'mp'. If using CCSD for the ground state, there is not a huge incentive to use partitioning for the EOM problem, so I suggest usingNone. If using MP2 for the ground-state, partitioning is worth considering. Although technically'full'makes fewer approximations, I suspect'mp'is more accurate due to some cancellation of errors. But these aren't rigorous statements, and I don't know of many benchmarks in the literature to provide statistical guidance.