wannier keywords

Wannier function calculator

wannier.bandi

Band index range for the inner energy window. If wannier.bandi is given, wannier.innwin will be generated using wannier.bandi. If both wannier.bandi and wannier.innwin aren’t set, RESCU will compute the Wannier function of all bands fully comprised in the outer window.

type: array

size: [1,2]

example: wannier.bandi = [2,6]

wannier.innwin

The inner energy window, also know as the frozen window, determines which Bloch states to be preserved identically in the projected manifold. If wannier.bandi, wannier.innwin will be generated using wannier.bandi. If both wannier.bandi and wannier.innwin aren’t set, RESCU will compute the Wannier function of all bands fully comprised in the outer window.

type: array

size: [1,2]

example: wannier.innwin = [-0.4 -0.1]

wannier.outwin

The outer energy window determines which Bloch states are included in the projected manifold. The states comprised in the outer window but not the inner window can be discarded such as to yield a maximally smooth subspace for the Wannier function generation. If wannier.outwin isn’t set, RESCU will set it equal to the interval given by wannier.bandi and wannier.innwin.

type: array

size: [1,2]

example: wannier.outwin = [-0.4 -0.1]

wannier.maxit

Maximal number of iterations (used during the subspace smoothing procedure and the spread functional optimization).

type: scalar

default: 100

example: wannier.maxit = 50

wannier.mixbeta

Mixing fraction of the Wannier function calculator (used during the subspace smoothing procedure).

type: scalar

default: 0.5

example: wannier.mixbeta = 0.45

wannier.niterSD

Number of steepest-descent iterations prior to the conjugate gradient algorithm during the spread functional optimization.

type: scalar

default: 1

example: wannier.niterSD = 1

wannier.omgTol

Optimization tolerance. The optimization procedure will stop once the spread functional changes by less than the prescribed tolerance (used during the subspace smoothing procedure and the spread functional optimization). The tolerance is relative, which means for example that if wannier.omgTol = 1e-4 the calculation will stop once the functional is precise to 4 digits.

type: scalar

default: 1e-05

example: wannier.omgTol = 1e-4

wannier.optAlgo

Algorithm employed to perform the spread function minimization. SD and CG stand for steepest-descent and conjugate gradient respectively.

type: string

default: 1e-05

allowed: 'SD','CG'

example: wannier.optAlgo = 'CG'

wannier.orbi

Array containing the indices of the orbitals used to construct an initial guess for the maximally localized Wannier gauge.

type: array

default: 1e-05

size: [inf,1]

example: wannier.orbi = [1 3:5]

wannier.sdalpha

Step scaling for the steepest-descent algorithm (used during the spread functional optimization). 0.5 is usually stable, 1 is marginally stable.

type: scalar

default: 0.5

example: wannier.sdalpha = 0.5

wannier.shiftEf

Shift the energy windows by the Fermi energy, i.e. the energy windows are relative to the Fermi energy.

type: scalar

default: false

allowed: true,false

example: wannier.shiftEf = true

wannier.status

Compute Wannier functions if wannier.status = true.

type: scalar

default: false

allowed: true,false

example: wannier.status = true