SMEAR
Controls smearing of MO occupation numbers for systems with small or zero gaps. [Edit on GitHub]
Keywords
Keyword descriptions
- SECTION_PARAMETERS
Type: logical
Default: F
Lone keyword: T
Usage: &SMEAR ONDescription: Controls the activation of smearing
- ELECTRONIC_TEMPERATURE
Type: real
Default: 3.00000000E+002 [K]
Aliases: ELEC_TEMP, TELEC
Usage: ELECTRONIC_TEMPERATURE [K] 300Description: Electronic temperature used for Fermi-Dirac smearing.
Mentions: ⭐How to make a SCF run converge
- EPS_FERMI_DIRAC
Type: real
Default: 1.00000000E-010
Usage: EPS_FERMI_DIRAC 1.0E-6Description: Accuracy checks on occupation numbers use this as a tolerance
- FIXED_MAGNETIC_MOMENT
Type: real
Default: -1.00000000E+002
Usage: FIXED_MAGNETIC_MOMENT 1.5Description: Imposed difference between the numbers of electrons of spin up and spin down: m = n(up) - n(down). A negative value (default) allows for a change of the magnetic moment. -1 specifically keeps an integer number of spin up and spin down electrons.
- LIST
Type: real[ ]
Usage: LIST 2.0 0.6666 0.6666 0.66666 0.0 0.0Description: A list of fractional occupations to use. Must match the number of states and sum up to the correct number of electrons
- METHOD
Type: enum
Default: GAUSSIAN
Usage: METHOD Fermi_DiracDescription: Selects the smearing method to apply.
Valid values:
FERMI_DIRACFermi-Dirac distribution defined by the keyword ELECTRONIC_TEMPERATURE. Use this method if the temperature equivalence is important for you, e.g. if you want to compute some properties based on the occupations. If you use this method without interest in electronic temperature, it’s suggested to use extrapolated result from finite ELECTRONIC_TEMPERATURE to ELECTRONIC_TEMPERATURE = 0. Note the forces and stress are consistent with the free energy and not with the extrapolated energy.ENERGY_WINDOWEnergy window defined by the keyword WINDOW_SIZE.LISTUse a fixed list of occupations.GAUSSIANGaussian broadening with width SIGMA; should work well in most cases. With this method you have to use extrapolated results from finite SIGMA results to SIGMA = 0, but usually this value would not be quite accurate without systematically reducing SIGMA. Note the forces and stress are consistent with the free energy and not with the extrapolated energy.METHFESSEL_PAXTONFirst-order Methfessel-Paxton distribution with width SIGMA. Don’t use it for semiconductors and insulators because the partial occupancies can be unphysical and thus lead to wrong results.MARZARI_VANDERBILTMarzari-Vanderbilt cold smearing with width SIGMA.
Mentions: ⭐How to make a SCF run converge
- SIGMA
Type: real
Default: 2.00000000E-003 [hartree]
Usage: SIGMA [eV] 0.2Description: Smearing width sigma (in energy units) in the case of Gaussian, Methfessel-Paxton or Marzari-Vanderbilt smearing.
Mentions: ⭐How to make a SCF run converge
- WINDOW_SIZE
Type: real
Default: 0.00000000E+000 [hartree]
Usage: WINDOW_SIZE [eV] 0.3Description: Size of the energy window centred at the Fermi level