magnon.magnons¶

Classes¶

MagnonSpectrum

A class to build and diagonalize the Linear Spin Wave Hamiltonian from exchange interactions.

Functions¶

_single_q_energy_worker(args)

Compute and return sorted real magnon energies at a k-point.

Module Contents¶

class magnon.magnons.MagnonSpectrum(atoms, interactions, calc=None, supercell=(1, 1, 1), name=None, num_threads=1, **kwargs)¶

A class to build and diagonalize the Linear Spin Wave Hamiltonian from exchange interactions.

Parameters:

atoms (ase.Atoms object) – The Atoms object of the system to which the magnon spectrum pertains.
interactions (magnon.interactions.InteractionList object) – The InteractionList object of the exchange interactions to use for the spectrum calculation.
num_threads (int) – The number of threads to use for parallel calculations.
mag_order_vect (numpy.ndarray, dtype: float) – Wavevector for Q-commensurate spin vector ordering.
mag_order_axis (numpy.ndarray, dtype: float) – The rotation axis around which the Q-commensurate spin vector ordering rotations are to occur
ham_prefactor (float) – Choice of Hamiltonian prefactor convention.
spins_are_unit (bool) – Whether to use the unit spin model, or dimensionful spins.
supercell (array-like) – Size of the calculation cell with respect to the primitive cell
name (str) – The name of the calculation

u¶

The u vectors for each site. See magnon.magnons.MagnonSpectrum._generate_uv_vectors().

Type:: numpy.ndarray, dtype: float

v¶

The v vectors for each site. See magnon.magnons.MagnonSpectrum._generate_uv_vectors().

Type:: numpy.ndarray, dtype: float

S¶

Spin quantum numbers.

Type:: numpy.ndarray, dtype: float

usm_factors¶

Unit spin model conversion factors.

Type:: numpy.ndarray, dtype: float

See also

This

_num_threads = 1¶

interactions¶

atoms¶

cell_vectors = None¶

interaction_matrix = None¶

calc = None¶

name = None¶

supercell = (1, 1, 1)¶

spin¶

_build_interaction_matrix()¶

static _generate_uv_vectors(spin_vectors)¶

Compute u and v vectors from spin orientation vectors for LSWT calculation

Parameters:: spin_vectors (numpy.ndarray) –
Return type:: (numpy.ndarray, numpy.ndarray)

calculate_eigen_energy(k)¶

Compute the eigenvalues of the Hamiltonian at wavevector k. The Hamiltonian in the rotated frame approach [1] is built and diagonalized using the treatment for generalized quadratic bosonic Hamiltonians [2].

Parameters:: k (ndarray, shape (3,)) – Wavevector in reciprocal space.
Returns:: hamiltonian_eigs – Eigenvalues of the Bogoliubov-transformed Hamiltonian.
Return type:: ndarray

References

get_band_structure(path)¶

Gets the magnon bandstructure along a path given by an ASE BandPath object. This is a wrapper for band_structure() which enables handling of k-space paths and bandstructures using ASE objects.

Parameters:: path (ASE.dft.kpoints.BandPath object) – The path along which to compute the bandstructure.
Returns:: An ASE BandStructure object containing the magnon bandstructure.
Return type:: ase.spectrum.band_structure.BandStructure

band_structure(path_kc)¶

Computes the bands energies for a set of reciprocal space points.

Parameters:: path_kc (array-like) – The reciprocal space path along which to compute the bandstructure, given in Cartesian coordinates.
Returns:: The band energies along the specified path.
Return type:: numpy.ndarray

get_dos(kpts=(10, 10, 10))¶

Computes the magnon density of states.

Parameters:: kpts (array-like, dtype=int) – The dimensions of the Monkhorst-Pack grid to sample.
Return type:: ase.spectrum.dosdata.RawDOSData object

magnon.magnons._single_q_energy_worker(args)¶

Compute and return sorted real magnon energies at a k-point.

This is a multiprocessing-friendly worker function for use with MagnonSpectrum. It must be defined at module level to be pickleable.

Parameters:

args (tuple) –

Tuple of (instance, q_c) where:

instanceMagnonSpectrum: Instance of the MagnonSpectrum class.
q_carray_like: k-point in scaled Cartesian coordinates.

Returns:

energies – Sorted real eigenvalues of the magnon Hamiltonian at the given k-point.

Return type:

ndarray

Notes

The k-point is converted to cartesian space (multiplied by 2*pi). Only the real parts of the eigenvalues are returned.