Symm for states of EOM-CCSD using RECP from DIRAC

More
7 years 4 months ago #335 by wangf
Dear Mihaly,


I am trying to run EOM-CCSD calculations with two-component spinors through interface to DIRAC from RECP with MRCC. As you know I already have EOM-CCSD for excitation energies using RECP with SOC included in CC calculation in CFOUR. I would like to evaluate error of my approach on neglecting SOC in the SCF step.


Unfortunately, we still could not work the symmetry out with MRCC. I attached with this email some of the results we get from MRCC and from CFOUR. We are calculating excitation energies of Zn, Cd, Hg, Xe, I- etc. and the excited states are 3P and 1P states. With SOC, the states will be 3P_0, 3P_1, 3P_2, 1P_1 states. For Zn, Cd, Hg, the lowest state is 3P_0, while the lowest state for Xe, I- is 3P_2. This means we should see a non-degenerate state for 3P_0, three-fold degenerate state for 3P_1 and 1P_1, while fifth-fold degenerate state for 3P_2. However, we could not see such degeneracy from output files of MRCC. Output files for Zn are also attached where Dirac-Coulomb Hamiltonian is used.


Furthermore, MRCC only output total energy without any additional information on transitions, which makes assignment rather difficult, or impossible. Presently assignments are made based on agreement with CFOUR results. Another strange problem with MRCC is that sometimes (not always), some states could not be obtained. As listed in the attached output file, we could not obtain the 1P_1 state of Cd or the 3P_0 state of Xe, while they can be obtained for Zn and I-.


By the way, I would like to remind you that in calculations with spin-orbit coupling, symmetry of the spinors is different from symmetry of the states. Spinors should transform according to irreducible representation of Fermion representations of the double point group, while the states should transform according to Boson representations, or those representations same to single point group.


I am sending this email to know whether this is because we did not run the code correctly or there is still some problems with implementation of MRCC for spin-orbit coupling.


with best regards,


DongDong
Attachments:

Please Log in or Create an account to join the conversation.

Time to create page: 0.037 seconds
Powered by Kunena Forum