Two particle density matrix from ccsd(t) calculation

Questions
#1

Hello all,
Can I obtain two-particle density matrix from a ccsd(t) calculation? If yes, How to do it?
Thanks a lot,
Prasanta

#2

To my knowledge it is not implemented.

#3

Unfortunately for me. I am looking for QC programs that can do that. It seems no programs are available except PySCF.

#4

What do you need it for?

#5

Thank you, Prof Krylov. In our group, we are engaged in finding the electron correlation energy from CCSD(t) 2PDM using ESP integrals. However, the 2PDM is very large and we subtract HF 2PDM and Muller approximate 2PDM from it for faster calculation. Please see the latest publication here.
We are engaged with PySCF, but we can’t do it for systems larger than 200 basis functions due to huge memory requirements (~4TB). This is why I am looking for alternating programs, like Qchem.

#6

This is exactly why real quantum chemistry codes do not compute the whole 2PDMs. We compute them by pieces and contract with relevant integrals to get the properties (such as gradients). So to advance your project you will need to go into the code and implement your method in this way.

This said, we do not have CCSD(T) gradients, so we do not have CCSD(T) 2PDM. The best CCSD(T) code is in Cfour. But we have 2PDM for CCSD and for various EOM-CCSD methods.

#7

This is very true that the calculation(s) are not memory-friendly. I was wondering if a well-maintained code can do this more efficiently. However, this seems a long goal to implement. Thank you Prof. Krylov, for your comments and suggestions. I will surely look at Cfour, I am currently looking at NWChem and ACES. Let’s see where this goes… Thanks Again, Professor.