TDDFT with Turbomole

Py-ChemShell
1

Hi everyone,

I’m trying to run a TDDFT (S1) QM/MM optimization using the pychemshell version 23.0.2 with turbomole 7.8 as the QM engine.
The ground state (S0) QM/MM optimization works fine, but the excited state jobs fails immediatly with:

ERROR: DL_FIND failed to run callback function of energy/gradients evaluation
Error message: ‘Result’ object has no attribute ‘_E0’

In the chemshell log I only see dscf executed, but not escf or egrad, However my Turbomole installation includes these binaries.

My QM block looks like:

qm = TURBOMOLE(method=‘dft’,
functional=‘B3-LYP’,
basis=‘def2svp’,
rpa=True,
nstates=8,
target_state=1,
self_gradients=True,
)

and the QMMM setup is additive, electrostatic embedding, covalent coupling.

Does the Py-ChemShell/Turbomole interface support TD-DFT excited-state gradients (via ‘escf’/‘egrad’) in 23.0.2? If yes, what is the correct combination of keywords to trigger ‘escf’ + ‘egrad’ instead of ‘dscf’ during DL-FIND optimization? If not, is RICC2 currently the recommended route for excited-state QM/MM optimizations?

Thanks in advance,
Dilara

2

Hi there,

Apologies for the late response with this.

We currently do not support TDDFT in turbomole in py-chemshell.

I will try and push a potential fix for our problem soon so you can run TDDFT.

When I have done this I will give some further instructions, it is unlikely to be a fully tested solution, so I would be grateful if you could check the results carefully to ensure they are what you expect and let me know if there are any problems etc.

Thanks

Joe

3

Hi Joe,

Thanks for your response. Also what I noticed is that with Turbomole the basis set is not taken into the control file. When defining the basis set, it is not formatting it right in the turbomole input and just takes the default basis set.

Best,
Dilara