Skip to content

DFTB (ground state)

Ground-state density-functional tight binding is a first-class OpenQP method, selected with [input] method=dftb. It comes in two families:

  • DFTB2 — the self-consistent-charge model ([dftb] type=ground). This is the default ground-state DFTB.
  • DFTB0 — the non-self-consistent model ([dftb] type=dftb0), useful as a fast zeroth-order reference.

Both deliver single-point energies, analytic gradients, and geometry optimization at tight-binding cost. Excited states are covered by the TD-DFTB and MRSF-TDDFTB manuals; all [dftb] keywords are documented in the [dftb] reference.

Development preview

The DFTB method is provided by the optional OpenQP-DFTB library and is not part of OpenQP 1.2.0; the one-line .oqp format is likewise a development-branch input style (see One-line .oqp). Install and build details are in the [dftb] reference.

Every example below leads with the recommended .oqp form, followed by Python and the legacy .inp form. Two DFTB conventions apply throughout:

  • basis= is a required-but-ignored placeholder (the Slater–Koster minimal basis is always used); functional= must be empty. The .oqp route carries neither.
  • The ground state is state 0 (grad/istate count from 0).

Energy

DFTB2 single-point energy:

.oqp (route dftb; the geometry file sits beside the .oqp file)

dftb
energy
geom="h2o.xyz"

Python

from oqp.openqp import OpenQP

job = OpenQP(project="h2o_dftb")
job.molecule("h2o.xyz")
job.dftb(response_type="ground")
job.workflow.energy()
job.run()

Legacy .inp

[input]
method=dftb
runtype=energy
charge=0
basis=sto-3g
functional=
system=
  O  0.000000  0.000000  0.000000
  H  0.000000  0.757160  0.586260
  H  0.000000 -0.757160  0.586260

[dftb]
backend=native
type=ground

Gradient

Add [properties] grad=0 (the ground state is state 0):

.oqp

dftb
grad
geom="h2o.xyz"

Python

from oqp.openqp import OpenQP

job = OpenQP(project="h2o_dftb_grad")
job.molecule("h2o.xyz")
job.dftb(response_type="ground")
job.workflow.gradient(state=0)
job.run()

Legacy .inp

[input]
method=dftb
runtype=grad
charge=0
basis=sto-3g
functional=
system=
  O  0.000000  0.000000  0.000000
  H  0.000000  0.757160  0.586260
  H  0.000000 -0.757160  0.586260

[dftb]
backend=native
type=ground

[properties]
grad=0

Geometry optimization

runtype=optimize with the native optimizer ([optimize] lib=oqp) on the ground state (istate=0):

.oqp

dftb
opt
geom="h2o.xyz"

Python

from oqp.openqp import OpenQP

job = OpenQP(project="h2o_dftb_opt")
job.molecule("h2o.xyz")
job.dftb(response_type="ground")
job.workflow.optimize(istate=0)
job.run()

Legacy .inp

[input]
method=dftb
runtype=optimize
charge=0
basis=sto-3g
functional=
system=
  O  0.000000  0.000000  0.000000
  H  0.000000  0.757160  0.586260
  H  0.000000 -0.757160  0.586260

[dftb]
backend=native
type=ground

[optimize]
lib=oqp
istate=0

Ground state and the [tdhf] block

A hand-written ground-state deck needs no [tdhf] or [scf] section. The Python builder always emits an inert [tdhf] type=tda block even for response_type="ground"; it is ignored by the ground-state DFTB path.

DFTB0 (non-SCC)

DFTB0 is the same ground-state method without charge self-consistency. Switch [dftb] type=ground → type=dftb0, the .oqp route dftb → dftb0, and the Python response_type="ground" → "dftb0". Energy, gradient, and optimization work identically (state 0 only):

.oqp (aliases dftb-noscc, dftb-nonscc are also accepted)

dftb0
energy
geom="h2o.xyz"

Python

from oqp.openqp import OpenQP

job = OpenQP(project="h2o_dftb0")
job.molecule("h2o.xyz")
job.dftb(response_type="dftb0")
job.workflow.energy()
job.run()

Legacy .inp

[input]
method=dftb
runtype=energy
charge=0
basis=sto-3g
functional=
system=
  O  0.000000  0.000000  0.000000
  H  0.000000  0.757160  0.586260
  H  0.000000 -0.757160  0.586260

[dftb]
backend=native
type=dftb0

parameter_path is optional. To override the bundled parameters, point it to a combined .opdftb file or a directory of Slater–Koster <El>-<El>.skf files. There is no MECI task for a ground-state method (it targets a single state); for conical intersections use MRSF-TDDFTB.