CP2K

class CP2K

This interface supports energy and gradient evaluations with the CP2K electronic structure package (version 8.1.0 or later)

As a QM interface, the options described in QM interfaces are available, as well as the CP2K specific options outlined below.

A CP2K Theory object can be specified as follows:

my_frag = Fragment(...)

my_theory = CP2K(frag=my_frag,
                 functional='PBE',
                 basis='DZVP-GTH-PADE',
                 ...)

which would then be called as theory=my_theory in subsequent tasks (see Tasks).

General options

basis

(default: 'DZVP') CP2K basis set name

Allowed types of value:

A Python str for basis set name, for example: 'DZVP-GTH-PADE'
A Python dict of element—to—basis-set mapping, for example: {'Pd':'SZV-MOLOPT-SR-GTH', 'O':'TZV2P-MOLOPT-GTH'}

basis_file

(default: '') CP2K basis set filename: when this is not specified, ChemShell will try to search into the CP2K data library

Allowed types of value:

A Python str for filename, for example: 'BASIS_SET'
A Python list of multiple filenames, for example: [ 'BASIS_SET', 'BASIS_MOLOPT' ]

diag_method: Algorithm for diagonalization. Allowed values are 'DAVIDSON', 'FILTER_MATRIX', 'LANCZOS', 'OT', and 'STANDARD' (default)

functional: Allowed values are 'B3LYP', 'BEEFVDW', 'BLYP', 'BP', 'HCTH120', 'LDA', 'NONE', 'NO_SHORTCUT', 'OLYP', 'PADE' (default), 'PBE', 'PBE0', and 'TPSS'

grid_cutoff: (default: 280.0 Rydberg) The cutoff of the finest grid level. See: CP2K Manual

grid_real_cutoff: (default: 40.0 Rydberg) The grid at which a Gaussian is mapped

guess: (default: 'ATOMIC') Initial guess for SCF if the job is not a restarted one

input: (default: '_cp2k.inp') Filename of CP2K input script

kpoints_method: (default: 'MONKHORST-PACK') K-points scheme to be used. Available options are 'NONE', 'GAMMA', 'MONKHORST-PACK', 'MACDONALD', and 'GENERAL'

lattice_scale: (default: 1.0000) Used for scaling the non-periodic vector in QM/Me only

maxiter_outer: (default: 50) Maximum number of iterations for outer SCF

mixing_alpha: (default: 0.3) Fraction of new density to be included. The default value is 0.4 in CP2K, see: CP2K Manual

mixing_method: Allowed values are 'BROYDEN_MIXING', 'BROYDEN_MIXING_NEW', 'DIRECT_P_MIXING' (default), 'KERKER_MIXING', 'MULTISECANT_MIXING', 'NONE', and 'PULAY_MIXING'. For the 2nd QM energy evaluation (without adsorbate) in QM/Me, only 'DIRECT_P_MIXING' can be used

mixing_gmix_p: (default: False) Toggle the mixing of the density matrix, see CP2K Manual

ngrids: (default: 4) The number of multigrids to use

nmos_added: (default: 10) Number of additional MOs added for each spin. For the QM energy evaluation in QM/Me, smearing should be switched on in order to avoid numerical instabilities due to discontunuous occupation numbers around the Fermi level. When smearing is used, make sure to use a basis set which provides enough vacant orbitals so that nmos_added does not exceed the maximum unoccupied orbitals. In CP2K it is defaulted to 0, see: CP2K Manual

outer_scf: (default: False) Toggle outer SCF loop

output: (default: '_cp2k.out') Name of CP2K output file

periodic: (default: 'XYZ') For QM/Me calculations 2D periodicity is imposed and the value for periodic is automatically detected (XY, XZ, or YZ) according to the system’s orientation. This value also determines the Poisson solver

poisson_solver

(default: 'PERIODIC') 'PERIODIC' is only available for 3D systems. For QM/Me calculations, ChemShell automatically determines which method to use:

'ANALYTIC': When periodic is 'XY' or 'YZ'
'WAVELET': When periodic is 'XZ'

potential_file: (default: '') Name of pseudo potential file. ChemShell will find one according to pseudopotential from the CP2K data library if the value is ''

precision_all: (default: 1.0E-10) A universal EPS energy criterion for Quickstep (QS)

precision_gvg: (default: 1.0E-8) Precision of the real-space Kohn-Sham matrix element integration for Quickstep (QS)

precision_rho: (default: 1.0E-10) Precision of the density mapping on the grids for Quickstep (QS)

smear: (default: True) Toggle DFT SCF smearing, see: CP2K Manual

smear_method

Allowed values:

'ENERGY_WINDOW': (default) See also smear_temperature and smear_window
'FERMI_DIRAC': This is the default method for QM/Me calculations
'LIST': Use a fixed list of occupations

smear_temperature: (default: 300.0 Kelvin) Electronic temperature in the case of 'FERMI-DIRAC' smearing method

smear_window: (default: 0.02 Hartree) In CP2K it is 0.0 by default but we use 0.02 for QM/Me

symmetry

Allowed values:

'auto': (default) Let ChemShell automatically detect the system’s symmetry. Available symmetries are: CUBIC, HEXAGONAL, MONOCLINIC, MONOCLINIC_GAMMA_AB, ORTHORHOMBIC, RHOMBOHEDRAL, TETRAGONAL_AB, TETRAGONAL_AC, TETRAGONAL_BC, TRICLINIC, and NONE
'' or None or False: No symmetry

wavefunction: (default: '_cp2k.wfn') Wavefunction restart filename

xc_derivative: (default: 'SPLINE2_SMOOTH') The method used to compute the XC derivatives. Defaulted to 'PW' in CP2K, see: CP2K Manual

xc_finer_grid: (default: False) Uses a finer grid only to calculate the XC

xc_smooth_rho: (default: 'SPLINE2') Density smoothing method used for the XC calculation. 'NONE' by default in CP2K, see: CP2K Manual