step2_dftb

libra_py.workflows.nbra.step2_dftb.do_ovlp(snap, params)

Compute the overlap matrix in the AO basis for two geometries, i and i+1

Parameters

• snap (int) – index of the time step to be used from the trajectory file

• params (dictionary) –

  the control parameters of the simulation

  • params[“EXE”] ( string ): path to the DFTB+ executable [ default: dftb+ ]

  • params[“md_file”] ( string ): the name of the xyz file containing the trajectory - the

    file should be in the xyz format produced by the DFTB+ program. [default: “md.xyz”]

  • params[“ovlp_gen_file”] ( string ): the name of the .gen file that is listed in the

    DFTB+ input file and contains the geometry of the system at two time steps (the content of this file will be updated for every i value). [default: “x2.gen” ]

  • params[“syst_spec”] ( string ): the string that is a part of the DFTB+ .gen file and defines

    whether the system is a non-periodic/cluster (“C”) or periodic (“S”). [default: “C”]

  • params[“ovlp_in_file”] ( string ): the name of the file containing the template for running

    calculations that construct H and S matrices and print them out by the DFTB+ calculations.

    • It should use the geometry file defined by params[“ovlp_gen_file”].

    • It should have the section: “WriteHS = Yes” to initialize the writing of the H and S matrices

    [default: “dftb_in_overlaps.hsd”]

Returns

the matrix of the AO overlaps for two geometries, where A - is the size of the AO basis

Return type

CMATRIX(A, A)

libra_py.workflows.nbra.step2_dftb.do_step(snap, params)

Runs a single-point SCF calculation for a given geometry along a trajectory

Parameters

• snap (int) – index of the time step to be used from the trajectory file

• params (dictionary) –

  the control parameters of the simulation

  • params[“EXE”] ( string ): path to the DFTB+ executable [ default: dftb+ ]

  • params[“mo_active_space”] ( list of ints or None ): indices of the MOs we care about

    The indexing starts from 0, not 1! If set to None - all MOs will be returned. [default: None]

  • params[“md_file”] ( string ): the name of the xyz file containing the trajectory - the

    file should be in the xyz format produced by the DFTB+ program. [default: “md.xyz”]

  • params[“sp_gen_file”] ( string ): the name of the .gen file that is listed in the

    DFTB+ input file and contains the geometry of the system (the content of this file will be updated for every i value). [default: “x1.gen” ]

  • params[“syst_spec”] ( string ): the string that is a part of the DFTB+ .gen file and defines

    whether the system is a non-periodic/cluster (“C”) or periodic (“S”). [default: “C”]

  • params[“scf_in_file”] ( string ): the name of the file containing the template for running regular

    SCF calculations for a single-point DFTB+ calculations.

    • It should use the geometry file defined by params[“sp_gen_file”].

    [default: “dftb_in_ham1.hsd”]

  • params[“hs_in_file”] ( string ): the name of the file containing the template for running

    calculations that construct H and S matrices and print them out by the DFTB+ calculations.

    • It should use the geometry file defined by params[“sp_gen_file”].

    • It should have the section: “ReadInitialCharges = Yes” to use the previously-converged charge density

    • It should have the section: “WriteHS = Yes” to initialize the writing of the H and S matrices

    [default: “dftb_in_ham2.hsd”]

  • params[“do_tddftb”] ( bool ): the parameter to control the type of energies to use in the

    calculations. The DFTB+ input files should be setup accordingly

    • Truecompute and read TD-DFTB energies. In this case, only the first entry of the returned

      results ( energies ) is meaningful. Other properties are just None for now

    • False : compute and read only the single-prticle energies [ default ]

Returns

(Ei, MOi, Hi, Si), where:

• Ei ( CMATRIX(M, M) ), the matrix of the converged Hamiltonian eigenvalues (at given geometry)

  Here, M = len(params[“mo_active_space”]) - we output only the MO energies that are of interest to us

• MOi ( CMATRIX(A, M) ), the matrix of the converged Hamiltonian eigenvalues (at given geometry)

  Here, M = len(params[“mo_active_space”]) - we output only the MO energies that are of interest to us. A - is the number of AOs in this calculation

• Hi (list of CMATRIX(A, A) ): the Hamiltonian matrices in the AO basis, for each k-point

• Si (list of CMATRIX(A, A) ): the overlap matrices in the AO basis, for each k-point

Return type

tuple

libra_py.workflows.nbra.step2_dftb.run_step2(params)

Calculate the overlaps, transition dipole moments, and vibronic Hamiltonian matrix elements in the AO basis

Currently under development 1/30/2020 - This function is currently not compatable with the parameters

“get_midpoint_energy” or “do_tddftb”.

Parameters

params (dictionary) –

the control parameters of the simulation

• params[“dt”] ( double ): nuclear dynamics timestep - as encoded in the trajectory [ units: a.u., default: 41.0 ]

• params[“isnap”] ( int ): initial frame [ default: 0 ]

• params[“fsnap”] ( int ): final frame [ default: 1 ]

• params[“out_dir”] ( string ): the path to the directory that will collect all the results

  If the directory doesn’t exist, it will be created [ default: “res” ]

SeeAlso: the description of the parameters in `do_ovlp(i, params)` and in `do_step(i, params)`

Returns

but generates the files (S, St, and hvib) indexed in the range [isnap, fsnap), for instance, if isnap = 0, fsnap = 3, we will have files hvib_0, hvib_1, hvib_2

Return type

None

libra_py.workflows.nbra.step2_dftb.run_step2_lz(params)

Calculate the diagonal vibronic Hamiltonian matrix elements (energies) in the AO basis and (optionally) the TD-DFTB energies

Parameters

params (dictionary) –

the control parameters of the simulation

• params[“dt”] ( double ): nuclear dynamics timestep - as encoded in the trajectory [ units: a.u., default: 41.0 ]

• params[“isnap”] ( int ): initial frame [ default: 0 ]

• params[“fsnap”] ( int ): final frame [ default: 1 ]

• params[“out_dir”] ( string ): the path to the directory that will collect all the results

  If the directory doesn’t exist, it will be created [ default: “res” ]

• params[“get_midpoint_energy”] ( bool ): if True, compute and read energies as Hvib = 0.5*(E_prev + E_curr) (As is done in Pyxaid NBRA)

  if False, compute and read energies at everytime timestep Hvib = E_curr

SeeAlso: the description of the parameters in `do_step(i, params)`

Returns

but generates the files (hvib) indexed in the range [isnap, fsnap), for instance, if isnap = 0, fsnap = 3, we will have files hvib_0, hvib_1, hvib_2. Only diagonal elements of the matrix are populated.

Return type

None