DFTB+

(contributed by Patrick Melix)

DFTB+ is a density-functional tight-binding implementation. More information about DFTB can be found on its official website.

PLAMS offers a simple and incomplete DFTB+ interface. It is so far capable of handling molecular calculations. The relevant classes are DFTBPlusJob and DFTBPlusResults.

Preparing a calculation

Preparing an instance of DFTBPlusJob follows the general principles for SingleJob. Information adjusting the input file is stored in the myjob.settings.input branch. The geometry of your system can be supplied via the class Molecule. If the Atomic Simulation Environment (ASE) is installed, the molecule is written using its engine. Otherwise this interface provies a basic routine. Note that the molecule is in this case transformed into the GenFormat with the C (cluster) or S (supercell) option, meaning the internal routine can handle clusters and supercell systems. The option F for fractional coordinates is not available! See the manual for further information on the different geometry-input types.

Input

Input files for DFTB+ are either in HSD (human-friendly structured data) or XML format. This interface will produce HSD format input files. See the manual for further information on keywords and structure. The input file must be named dftb_in.hsd and is therefore created using this name. Note that many values have standard settings, those will all be printed to dftb_pin.hsd when you start a calculation. Check both files for errors when having problems.

HSD input files are organized using different properties. Each property is represented by a key. The key can be of type logical, integer, real, string, property list or method type. The last two begin and end with curly brackets. Since property lists and method types are very similar in notation we require a way of representing that style in the tree-like structure of Settings. This is done by automatically creating either property lists or method types whenever a key has a subkey.

  • If a key has a subkey and a key _h, the string assigned to _h will be used as a method name.

  • If a key has one or more subkeys it will be created as a property list:

    #sets the hamiltonian property to be of a method type named DFTB
myjob.setting.input.hamiltonian._h = 'DFTB'
#sets the key parserversion of the property list parserversion to 4
myjob.setting.input.parseroptions.parserversion = '4'

Empty method types can be created by not giving any subkeys to a key except _h.

Runscript

The runscript will call the binary dftb+, so make sure it is in your $PATH. No options are supported. The standard output is redirected to $JN.out, errors to $JN.err.

Results extraction

DFTB+ creates multiple outputfiles, none of them are renamed. See detailed.out for the results of your calculation. Resulting geometries are saved in .xyz and .gen format by DFTB+. Other files might be created depending on your calculation type.

General text processing methods from Results can be used to obtain data from results files. At the moment only three functions for result extraction are defined:

Example

common = Settings()

common.input.driver._h = 'ConjugateGradient'
common.input.hamiltonian._h = 'DFTB'
common.input.hamiltonian.scc = 'yes'
common.input.hamiltonian.mixer._h = 'Broyden'
common.input.hamiltonian.mixer.mixingparameter = '0.2'
common.input.hamiltonian.slaterkosterfiles._h = 'Type2Filenames'
common.input.hamiltonian.slaterkosterfiles.prefix = '"~/SLAKO/mio-1-1/"'
common.input.hamiltonian.slaterkosterfiles.separator = '"-"'
common.input.hamiltonian.slaterkosterfiles.suffix = ".skf"
common.input.hamiltonian.slaterkosterfiles.lowercasetypename = 'No'
common.input.hamiltonian.maxangularmomentum.c = '"p"'
common.input.hamiltonian.maxangularmomentum.h = '"s"'
common.input.parseroptions.parserversion = '4'

mol = Molecule(filename='mol.xyz') # read Molecule from mol.xyz

job = DFTBPlusJob(name='plamstest', molecule=mol, settings=common)
jobres = job.run()

energy = jobres.get_energy(string='Fermi energy', unit='ev')
print(energy)

mol = jobres.get_molecule()
print(mol)

atomic_charges = jobres.get_atomic_charges()
print(atomic_charges)

API

class DFTBPlusJob(molecule=None, name='plamsjob', settings=None, depend=None)[source]

A class representing a single computational job with DFTB+. Only supports molecular coordinates, no support for lattice yet.

get_input()[source]

Transform all contents of setting.input branch into string with blocks, keys and values.

Automatic handling of molecule can be disabled with settings.ignore_molecule = True.

get_runscript()[source]

dftb+ has to be in your $PATH!

check()[source]

Returns true if ‘ERROR!’ is not found in the output.

class DFTBPlusResults(job)[source]

A Class for handling DFTB+ Results.

get_molecule()[source]

Return the molecule from the ‘geo_end.gen’ file. If there is an Error, try to read the ‘geo_end.xyz’ file.

get_energy(string='Total energy', unit='au')[source]

Return the energy given in the output with the description string, expressed in unit. Defaults to Total energy and au.

get_atomic_charges()[source]

Returns a dictonary with atom numbers and their charges, ordering is the same as in the input.