NEGF

References: Rocha2006b, Papior2017

Parameters which control quantum transport calculation based on Non-Equilibrium Green’s Function method. [Edit on GitHub]

Keywords

Keyword descriptions

DELTA_NPOLES: integer = 4

Number of poles of Fermi function to consider. [Edit on GitHub]

DISABLE_CACHE: logical = F

Lone keyword: T

Do not keep contact self-energy matrices for future reuse [Edit on GitHub]

ENERGY_LBOUND: real = -5.00000000E+000 [hartree]

Lower bound energy of the conductance band. [Edit on GitHub]

EPS_DENSITY: real = 1.00000000E-005

Target accuracy for electronic density. [Edit on GitHub]

EPS_GEO: real = 5.29177209E-007 [angstrom]

Accuracy in mapping atoms between different force environments. [Edit on GitHub]

EPS_GREEN: real = 1.00000000E-005

Target accuracy for surface Green’s functions. [Edit on GitHub]

EPS_SCF: real = 1.00000000E-005

Target accuracy for SCF convergence. [Edit on GitHub]

ETA: real = 1.00000000E-005 [hartree]

Infinitesimal offset from the real axis. [Edit on GitHub]

GAMMA_KT: integer = 20

Offset from the axis (in terms of k*T) where poles of the Fermi function reside. [Edit on GitHub]

HOMO_LUMO_GAP: real = 2.00000000E-001 [hartree]

The gap between the HOMO and some fictitious LUMO. This option is used as an initial offset to determine the actual Fermi level of bulk contacts. It does not need to be exact HOMO-LUMO gap, just some value to start with. [Edit on GitHub]

INTEGRATION_MAX_POINTS: integer = 512

Maximal number of grid point for adaptive numerical integration. [Edit on GitHub]

INTEGRATION_METHOD: enum = CLENSHAW-CURTIS

Valid values:

  • CLENSHAW-CURTIS Adaptive Clenshaw-Curtis quadrature method. Requires FFTW3 library.

  • SIMPSON Adaptive Simpson method. Works without FFTW3.

Method to integrate Green’s functions along a closed-circuit contour. [Edit on GitHub]

INTEGRATION_MIN_POINTS: integer = 16

Initial (minimal) number of grid point for adaptive numerical integration. [Edit on GitHub]

MAX_SCF: integer = 30

Maximum number of SCF iterations to be performed. [Edit on GitHub]

NPROC_POINT: integer = 0

Number of MPI processes to be used per energy point. Default is to use all processors (0). [Edit on GitHub]

V_SHIFT: real = 0.00000000E+000 [hartree]

Initial value of the Hartree potential shift [Edit on GitHub]

V_SHIFT_MAX_ITERS: integer = 30

Maximal number of iteration to determine the optimal shift in Hartree potential. [Edit on GitHub]

V_SHIFT_OFFSET: real = 1.00000000E-001

Initial offset to determine the optimal shift in Hartree potential. [Edit on GitHub]