correlation energy in TMHF or Lh12ct-SsifPW9

[fsymalla]

Dear Turbomole experts,

I would like to know the correlation energy in calculations using TMHF and Lh12ct-SsifPW9  functionals.
in the $subenergy section of the control file both exchange and correlation contirbutions are listed as zero for both functionals. Is there a way to access the correlation energy some other way?
If it can not be printed, would it give the correct results if i define a functional using xclib which in case of TMHF would only consist of the B95 correlation functional which according to https://doi.org/10.1063/5.0100439 is the one used for the correlation energy in tmhf, and rerun using $scfiterlimit 1 and $scfdamp start=100000000000 and read of the total energy from that run (or for that matter even use any functional which uses B95 for correlation and use the correlation energy printed in $subenergy)?

Thanks for any help,
best regards,
Franz

update:
i tried my own suggestion with functional libxc 240 (MGGA_C_BC95 (id=240): Becke correlation 95)
also no correlation energy is written in $subenergy but i can read of the exchange correlation energy in the normal SCF output
Exc = -1.40427110940     Coul =  312.223154474
 and since i have no exchange functional this should be the correlation contribution only.
now i get some number in a clumsy way. if someone could comment on whether this is the correct one for the tmhf case it would be appreciated, or of course if there is a better way to just write out the energy, because for  Lh12ct-SsifPW9 i did not find the appropriate correlation functional in libxc.

[yannickf]

Dear Franz,

Currently, the individual energy contributions are only printed in the developers' version, directly in the dscf or ridft output. We used that for the development of theTMHF  functional but disabled some output in the official release (I do not know the reason for that anymore). I could add that information for the next version, likely 7.8.1.

Your approach only works for the original B95 correlation. Many local hybrids use a modified version of B95.

Best,

Yannick

[yannickf]

PS: I have implemented it accordingly for V7.8.1. The exchange and correlation energy of local hybrid functionals is now stored as done for the other functionals in both one-component and two-component runs.

[fsymalla]

Dear Yannick,
thank you very much indeed! do you know what date is aimed for, for that release?  Regarding modified versions of B95 in local hybrids: do you know whether it was modified in the case of tmhf?

Best
Franz

[yannickf]

Dear Franz,

TMHF and TMHF-3P use the original B95 correlation (I am one of the authors of the TMHF functional familly). Other local hybrids by the Berlin group use a modified version (lh20t etc.). I guess that 7.8.1 will be available in summer 2024. You could ask the official support for a preversion.

Best,

Yannick

[yannickf]

Dear Franz,

Turbomole V7.8.1 should be available for general users now. So, you can now get the energy contributions for local hybrid functionals.

Best wishes,

Yannick