all-electron calculations on actinides in Turbomole

[Dempsey]

Dear Users,

I have been calculating actinide systems using the standard 60e ECP but would now like to run an all-electron single-point energy calculation too. I see that the DKH and ZORA Hamiltonian, I see that DKH is implemented by using $soghf but I am not sure how to set up these calculations with the basis sets available in Turbomole. If I attempt to use the SARC-ZORA basis set from basis set exchange there is no EHT data, if I attempt to use the SARC-DKH2 basis set then in define I am asked if I want to supply the 60e ECP and then when I say no define fails.

I could not find a tutorial or much in the manual about how to set up calculations like this. Could you please share a step-by-step method to set up an all-electron relativistic single point using the DKH Hamiltonian for actinide systems?

Best,
Dempsey

[yannickf]

Dear Dempsey,

you have probably found another way round. But I would still like to answer the question (even if it is too late) for future applications.

If you know the occupation from an ECP run, you can manually set the number of occupations (closed-shell) or (UHF). Delete all ECP stuff and add the X2C/DKH/BSS kexywords (X2C is recommended nowadays). You do not need soghf for DKH runs, you can use 1c and 2c DKH. The first option only needs $rdkh and the order. X2C needs $rx2c.

After setting the occupation, you can use the hcore option, as this starts a calculation from a core Hamiltonian guess (1e integrals only). This does not need Hückel vectors or eht data. In the upcomming V7.8, you can use hcore -a or hcore -atoms to use a superposition of atomic densities as initial guess. As there are no Hückel vectors, you need to set the atomdens keywords for the orbital basis set guess. Then, the basis set is used to construct the atomic density.

Best,

Yannick