List of keywords¶
General¶
-
title
: name
¶ Sets the title of the job to name.
-
theory
: name
¶ Sets the theory or exchange-correlation functional to be used. Possible theories are:
- hf (default) - Hartree-Fock method
- hfvwn - HFVWN method, combination of HF exchange with Vosko, Wilk and Nusair correlation functional (VWN) [31]
- slater - Slater exchange functional [28]
- svwn - combination of Slater exchange functional [28] with the VWN (Vosko, Wilk, and Nusair) correlation functional [31]
- slyp - combination of Slater exchange functional [28] with the correlation functional of Lee, Yang, and Parr [11]
- becke - Becke’s 1988 exchange functional [5]
- blyp - combination of Becke’s 1988 exchange functional [5] with the correlation functional of Lee, Yang, and Parr (LYP) [11]
- pw91 - Perdew and Wang’s 1991 exchange-correlation fuctional [24][23]
- pbe - The 1996 functional of Perdew, Burke and Ernzerhof [25]
- b3lyp - Becke 3 parameter hybrid functional, combination of Becke’s 1988 exchange functional, VWN functional for local correlation and LYP functional for non-local correlation [6][29]
- pbe0 - The 1996 functional of Perdew, Burke and Ernzerhof, as made into a hybrid by Adamo [2]
- mpw91 - modified Perdew and Wang’s exchange functional [1] with pw91 correlation functional [24][23]
- mpw91lyp - modified Perdew and Wang’s exchange functional [1] with the correlation functional of Lee, Yang, and Parr (LYP) [11]
- mb3lyp - B3LYP functional, as modified by Tognetti [30]
- b3lyp5 - B3LYP functional with the VWN5 functional for local correlation [6][29][31]
- b3pw91 - Becke 3 parameter hybrid functional with Perdew and Wang’s 1991 functional for non-local correlation (combination of Becke’s 1988 exchange functional, VWN functional for local correlation and Perdew and Wang’s 1991 functional for non-local correlation)
- b2plyp - combination of Becke’s [5] and HF exchange with the correlation functional of Lee, Yang, and Parr (LYP) [11] and a perturbative second order part (MP2) [8]
- mpw2plyp - combination of modified Perdew and Wang’s exchange [1], HF exchange, correlation of Lee, Yang, and Parr [11] and a perturbative second order part (MP2) [27]
- eh - extended Hückel method
-
dispersion
: name
¶ Specifies whether to use additional dispersion correction. Possible name values are:
- none (default) - no additional dispersion correction
- d3 - Grimme’s D3 dispersion correction [9]
-
run_type
: name
¶ Specifies the type of calculation to be done. Possible name values are:
- single_point (default) - single point calculations for a given geometry
- geometry_optimization - geometry optimization
-
charge
: value
¶ Total charge of the system. Default value is 0.
-
multiplicity
: value
¶ The multiplicity of the system. Default value is 1.
-
basis_set
: name
¶ Specifies name of the basis set to be used in calculations. All standard basis sets are available. Full list of defined basis sets is specified in the basis sets section. Additional basis sets can be defined by user in a gaussian (g94) or molcas format.
- sto-3g (default)
-
aux_basis_set
: name
¶ Specifies name of a basis set used during density fitting process. If not specified no density fitting is performed.
-
posthf_aux_basis_set
: name
¶ Specifies name of a basis set used during density fitting process during posthf calculations. If not specified no density fitting is performed.
-
external_electric_field
: x_value y_value z_value
¶ Defines values of the external electric field vector. Default is 0 0 0.
-
units
:
¶ This section controls various units in input and output files.
-
energy
: name
¶ Specifies energy units. Posible units are:
- hartree (default)
- eV
-
length
: name
¶ Specifies length (distance) units. Posible units are:
- angstrom
- bohr (default)
-
storage
: name
¶ Specifies storage units. Posible units are:
- byte
- kilobyte
- megabyte (default)
- gigabyte
-
SCF¶
-
scf
:
¶ Section consists of keywords controlling SCF calculations.
-
method
: name
¶
-
initial_guess
: name
¶ Specifies the initial guess for the Hartree-Fock wavefunction?
- core_hamiltonian - initial guess constructed from core hamiltonian
- fragments (default) - initial guess constructed from promolecular fragments
- from_file - initial guess read from file (filename will be read from
initial_guess_filename
)
-
initial_guess_filename
: name
¶ Name of a file containing initial guess in the HDF5 format. File can be generated by
binary
option inoutput
section.
-
convergence_accelerator
: name
¶ Specifies whether to use convergence acceleration during the SCF
- none - no convergence acceleration
- diis (default) - use Pulay’s direct inversion in the iterative subspace [26].
-
max_no_iterations
: value
¶ Specifies maximum number of SCF iterations. Default value is 50.
-
energy_threshold
: value
¶ Sets the SCF convergence criterion to energy change lower than a value. Default value is 1e-6 (0.000001).
-
density_threshold
: value
¶ Sets the SCF convergence criterion to density? change lower than a value. Default value is 1e-6 (0.000001).
-
occupations
: name
¶ Specifies the way to determine orbital occupations during SCF calculations. Valid names are:
- aufbau (default) - filling orbitals with respect to Aufbau principle
- fermi - filling orbitals with respect to Fermi-Dirac distribution
- maximum_overlap - filling orbitals most similar to the ones which were occupied in the previous SCF cycle
-
degeneracy_threshold
: value
¶ Used only when
occupations
is set to aufbau. Electrons are distributed equally between all orbitals, which energy differs from HOMO by less than value. When not set occupations are allways a whole number.
-
smear
: value
¶ Used only when
occupations
is set to fermi. Then electrons are distributed with respect to the Fermi-Dirac statistics with \(kT=value\). Fermi level is set to \((E_{HOMO} + E_{LUMO})/2\). The default value is 0.001.
-
shift_1
: value
¶ This keyword controlls level shifting. value is a positive number. In RHF and PRHF calculations sets a gap size between virtual and occupied orbitals, in ROHF between single-occupied and double-occupied orbitals and in UHF between virtual and occupied orbitals for spin α.
-
shift_2
: value
¶ Used only during UHF calculations. Sets a gap size between virtual and occupied orbitals for spin β.
-
threshold_linear_independence
: value
¶ Threshold for elimination of (quasi)linear dependence from the basis set. Default value is 0.
-
Geometry optimization¶
-
geometry_optimization
:
¶ This section specifies parameters controlling the geometry optimization process.
-
method
: name
Method to be used during geometry optimization. Valid name values are:
-
coords
: name
¶ Coordinates used during the geometry optimization. Valid name values are:
- cartesian (default) - optimization process performed in cartesian coordinates
- delocalised_internal - optimization process performed in delocalised internal coordinates [4]
-
threshold
: value
¶ Convergence criterium for geometry optimization. Default value is 1e-5.
-
max_no_steps
: value
¶ Maximal number of steps during the optimization. Default value is 100.
-
Integration params¶
-
integration_params
:
¶ Set of integration parameters.
-
engine
: name
¶ Determins which engine is to be used for integration. Valid name values are:
- naive - reference implementation, very slow, not to be used for production runs
- os1 (default) Obara-Saika scheme
-
threshold
: value
Maximum value of integration error. Default value is 1e-11.
-
cache_size
: value
¶ Cache size used for two-electron integrals. Default value is 64.
-
aux_cache_size
: value
¶ Cache size used during density fitting. Default value is 64.
-
storage
: name
Describes the method of integrals storing. Valid name values are:
- none (default) - direct calculations (integrals are not stored and are recalculated when needed)
- local - all integrals are stored on a disk; during MPI calculations every process has its own copy
- shared - all integrals are stored on a disk; during MPI calculations file containg itegrals is shared by all processes
- in_core - all integrals are stored in RAM (uses excessive amount of memory, viable only for small systems)
-
Møller-Plesset¶
-
moller_plesset
:
¶ In this sections parameters controlling Møller-Plesset method are specified.
-
order
: logical_value
¶ Sets an order of perturbation to a logical_value. The default logical_value is 0 (no perturbation analysis). At the moment only 0 and 2 are valid options.
-
dipole_moment
: logical_value
¶ Specifies if dipole moment is calculated. Default logical_value is false.
-
degeneracy_corrected
: logical_value
¶ If true then DCPT2 (near-degeneracy corrected second-order perturbation theory) energy expression is applied instead of standard MP2 expression [3]. The default logical_value is false meaning standard MP2 expression.
-
memory_pool
: value
¶ Maximal amount of memory to be used (in MB). Default value is 256.
-
no_frozen
: value
¶ Specifies number of frozen molecular orbitals (removed from correlation window). Default value is 0.
-
no_deleted
: value
¶ Number of omitted virtual orbitals. Default value is 0.
-
no_frozen_beta
: value
¶ Number of frozen beta manifold occupied orbitals (relevant only for unrestricted case). Default value is 0.
-
no_deleted_beta
: value
¶ Number of omitted beta virtual orbitals (relevant only for unrestricted case). Default value is 0.
-
type
: name
¶ Sets type of calculations. Valid name values are:
- conventional - reference, but very slow implementation of conventional MP2 (restricted case only)
- imp - quite efficient implementation of conventional MP2 (restricted case only)
- minmem - most efficient implementation of conventional MP2 (restricted and unrestricted case)
- sec_imp (default) - slow, but the least memory-demanding implementation of conventional MP2 (restricted case only)
- laplace_mo - LT-MP2 in MO basis, only for reference (restricted case only)
- laplace_ao - LT-MP2 in AO basis (restricted case only)
- laplace_ao_ls - LT-MP2 in AO basis with somehow improved scaling of memory and time requirements (restricted case only)
-
quadrature
: name
¶ Type of quadrature used in LT-MP2, possible values are:
- fitted_simple
- fitted_normalized
- em_static (default)
- em_simple
- em_normalized
-
no_points
: value
¶ Number of points for quadrature used in LT-MP2. Default value is 10.
-
eps
: value
¶ Cutoff for prescreening small contributions. Default value is 1e-8.
-
Excited states¶
-
td
:
¶ This section decribes parameters controlling time dependent calculations.
-
type
: name
Type of time dependent calculations. Possible name values are:
- none (default)
- tda - Tamm-Dancoff approximation (in HF calculations results in CIS method)
- rpa - random phase approximation
-
multiplicity
: value
If value is set to 1 performs calculations only for singlets, 2 results in calculation only for triplets, 3 results in calculations for singlets and triplets. Default value is 3.
-
no_frozen
: value
Number of frozen (core) orbitals. Default value is 0.
-
no_deleted
: value
Number of omitted virtual orbitals. Default value is 0.
-
no_states
: value
¶ Requested number of states. Default value is 0.
-
no_roots
: value
¶ Dimension of Davidson diagonalization subspace. Default value is \(2 \cdot\)
no_states
.
-
no_iterations
: value
¶ Maximum number of Davidson diagonalization iterations. Default value is 50.
-
diagonalization_threshold
: value
¶ Davidson diagonalization threshold. Default value is 1e-4 (Hartree).
-
integral_threshold
: value
¶ Maximum value of integration error. Default value is 1e-10.
-
dress_davidson_threshold
: value
¶ Davidson diagonalization threshold during dressing iterations. Default value is 1e-3 (Hartree).
-
dress_delta_energy
: value
¶ Energy window to scan for doubly excited configurations suitable for dressing. Default value is 0.2 (Hartree).
-
max_no_dressing_iterations
: value
¶ Maximum number of dressing iterations. 0 means that dressing is disabled. Default value is 0.
-
max_no_davidson_dressing_iterations
: value
¶ Maximum number of Davidson diagonalization iterations during dressing iteration. Default value is the same as actual value of
no_iterations
.
-
dress_state
: value
¶ Number of the state to dress. Only states having non-negligible admixture of doubly excited configuration are counted. Default value is 1 (dress the lowest state interacting with doubly excited configuration).
-
Properties¶
-
properties
:
¶ This section specifies which additional properties of analysed system should be computed.
-
max_multipole_moment_order
: value
¶ Maximum multipole moment order to be calculated.
-
population_analyses
:
¶ Set of parameters controlling population analyses.
-
mulliken
: logical_value
¶ Mulliken population analysis [16] [17] [18] [19] (default logical_value is true)
-
lowdin
: logical_value
¶ Löwdin population analysis (default logical_value is true).
-
hirshfeld
: logical_value
¶ Hirshweld population analysis (default logical_value is false).
-
voronoi
: logical_value
¶ Voronoi population analysis (default logical_value is false).
-
bader
: logical_value
¶ Bader population analysis (default logical_value is false).
-
-
bond_order_analyses
:
¶ Set of parameters controlling bond order analyses.
-
Output¶
-
output
:
¶ In this section additional output format can be specified.
-
binary
: logical_value
¶ Specifies if binary output is generated (which can be used as a initial guess
initial_guess
file in order to restart calculations). Default logical_value is false.
-
Grid¶
-
grid
:
¶ Specifies the integration grid to be used for numerical integrations in DFT.
-
type
: name
Type of grid to be used during calculations. Valid name values are:
- simple - simple grid (radial shells with equal number of angular points)
- pruned - simple grid modified with the vicinity of atomic nucleus (less angular points in this area)
- angular_s_adaptive - adaptive angular grid designed to reproduce S matrix
- angular_rho0_adaptive (default) - adaptive angular grid designed to reproduce density of promolecule
- angular_vxc0_adaptive - adaptive angular grid designed to reproduce exchange-correlation potential of promolecule
- full_rho0_adaptive - full adaptive grid (angular and radial) designed to reproduce density of promolecule
-
angular
: name
¶ Specifies type of angular mapping on grid. Valid name values are:
- gauss_legendre - Gauss-Legendre quadrature
- lebedev_laikov (default) - Lebedev-Laikov quadrature
-
radial_mapping
: name
¶ Specifies type of radial mapping on grid. Valid name values are:
- logm (default)
- koester
-
accuracy
: value
¶ Minimal accuracy requested during grid adaptation. Default value is 1e-6
-
max_no_radial_points
: value
¶ Maximal number of radial shells on grid. Default value is 75.
-
max_no_angular_points
: value
¶ Maximal number of angular points on a shell. Default value is 1500.
-
cache_size
: value
Cache size used for storing function values on grid. Default value is 64.
-