Skip to content

TDDFT and TDHF

Source access and public downloads

Repository paths identify files in the licensed engine source; they are not public downloads. For freely downloadable .oqp inputs and results, see Worked Examples.

TDDFT and TDHF response calculations use input.method=tdhf. A non-empty input.functional selects TDDFT; leaving functional empty gives TDHF. Closed-shell examples normally use an RHF reference with scf.multiplicity=1.

Use MRSF-TDDFT when multiconfigurational ground-state character, conical intersections, or spin-flip target-state ordering are central to the study. Use ordinary TDDFT/TDHF for standard single-reference vertical excitations and excited-state gradients.

In Python scripts, TDHF and TDDFT are theory choices. Select the calculation type separately with job.workflow.* only when you need a non-energy workflow.

Energy

.oqp:

tddft(nstate=3)/b3lyp5/6-31g*
geom="h2o.xyz"

For TDHF, use the same layout with the route tdhf(nstate=3)/6-31g*.

Python:

from oqp.openqp import OpenQP

job = OpenQP("h2o_tddft", silent=1)
job.molecule(geometry="water", charge=0, multiplicity=1)
job.theory.tddft(functional="b3lyp5", basis="6-31g*", nstate=3)

mol = job.run()
print("TD energies:", mol.get_td_energies())

Canonical .oqp input:

tddft(nstate=3)/b3lyp5/6-31g* energy(S0)
geom="molecule.xyz"

Runnable .oqp: examples/TDDFT/H2O_B3LYP5-TDDFT_ENERGY.oqp.

For TDHF, use the same structure without input.functional: examples/TDHF/H2O_TDHF_ENERGY.oqp.

Closed-shell triplet states

On a closed-shell RHF/RKS reference, conventional RPA/TDA also gives the triplet excitation energies. Set tdhf.multiplicity=3 and keep the reference a singlet (scf.type=rhf, multiplicity=1). The triplet response has no Coulomb term and uses the f_aa - f_ab exchange-correlation kernel; oscillator strengths are zero. Energies, analytic gradients and geometry optimizations are available (runtype=energy, grad, optimize). Triplet Hessians, NACs, crossings and open-shell references are rejected by the input checker. Use MRSF-TDDFT for those triplet workflows.

.oqp (T0 is the lowest triplet; nstate triplet roots are solved):

tddft(nstate=3)/b3lyp5/6-31g* energy(T0)
geom="h2o.xyz"

Python (multiplicity stays the reference multiplicity):

job.theory.tddft(functional="b3lyp5", basis="6-31g*", nstate=3,
                 state_multiplicity=3)

Canonical .oqp input: the energy example above with

tdhf(nstate=3)/6-31g* energy(T0)
geom="molecule.xyz"

Runnable example: examples/TDDFT/H2O_B3LYP5-TDDFT_TRIPLET_ENERGY.oqp.

Gradient of the lowest triplet in .oqp:

tddft(nstate=3)/b3lyp5/6-31g* grad(T0)
geom="h2o.xyz"

Runnable example: examples/TDDFT/H2O_B3LYP5-TDDFT_TRIPLET_GRADIENT.oqp. Triplet gradients and optimizations require OpenQP with openqp#502; triplet energies alone need openqp#475.

Gradient

In ordinary TDHF/TDDFT, S1 means the first excited state.

.oqp:

tddft/b3lyp5/6-31g* grad(S3)
geom="h2o.xyz"

The state label sets the root count: grad(S3) makes OpenQP solve at least three roots, so the route needs nstate only to request more roots than the driver implies (for example tddft(nstate=6)/b3lyp5/6-31g* grad(S3)). The same rule applies to energy(...), opt(...), and the SF/MRSF labels (grad(root=3), grad(S2)).

Python:

from oqp.openqp import OpenQP

job = OpenQP("h2o_tddft_grad", silent=1)
job.molecule(geometry="water", charge=0, multiplicity=1)
job.theory.tddft(functional="b3lyp5", basis="6-31g*", nstate=3)
job.workflow.gradient(state=3)

mol = job.run()
gradient = mol.get_grad()

Canonical .oqp input:

tddft/b3lyp5/6-31g* grad(S3)
geom="molecule.xyz"

Runnable .oqp: examples/TDDFT/H2O_B3LYP5-TDDFT_GRADIENT.oqp.

TDHF gradient example: examples/TDHF/H2O_TDHF_GRADIENT.oqp.

Notes

  • The internal tdhf.nstate must include the highest state requested by gradients or follow-up analysis; .oqp input widens the root count from the driver's state label automatically.
  • Ordinary TDHF/TDDFT state numbering is excited-state numbering: state=1 in Python, or grad=1 in properties, means the first excited state.
  • SF-TDDFT and MRSF-TDDFT use a different target-state ordering: state 1 is the lowest spin-flip/MRSF target state, which can be the multiconfigurational ground state.
  • Use tdhf.type=rpa or omit type for the default response model. Use tdhf.type=tda when a TDA calculation is desired.
  • For TDHF in Python, use job.theory.tdhf(basis=..., nstate=...).

Scientific references

Linear-response TDHF and TDDFT.