API Reference

Public package API

LWW 1D Wigner-Poisson quantum transport package.

class lww_transport.ComputeParams(dense_solver='numpy', kernel_backend='auto', use_numba=True, verbose=False)[source]

Performance and debugging options.

Parameters:

  • dense_solver (str) – Linear solver backend. Default ‘numpy’.

  • kernel_backend (str) – Computational kernel implementation. Options: ‘auto’, ‘cpp’, ‘numba’, ‘python’. Default ‘auto’, which selects the fastest available backend.

  • use_numba (bool) – Enable Numba JIT compilation for hot loops. Default True.

  • verbose (bool) – Print diagnostic information during solve. Default False.

dense_solver: str
kernel_backend: str
use_numba: bool
verbose: bool
class lww_transport.DiscretizationParams(nx=86, n=72)[source]

Spatial and momentum grid parameters.

Parameters:

  • nx (int) – Number of spatial grid points. Default 86 matches the reference RTD configuration.

  • n (int) – Number of momentum grid points. Must be even. Default 72 matches the reference RTD configuration.

nx: int
n: int
class lww_transport.GeometryParams(box=550.0, well=50.0, barrier=30.0, spacer=30.0, pot=0.3)[source]

Double-barrier quantum well (DBQW) geometry.

Describes the device structure: emitter | spacer | barrier | well | barrier | spacer | collector

Parameters:

  • box (float) – Total device length [nm]. Default 550 nm.

  • well (float) – Quantum well width [nm]. Default 50 nm.

  • barrier (float) – Barrier position/height parameter [nm]. Default 30 nm.

  • spacer (float) – Spacer region width [nm]. Default 30 nm.

  • pot (float) – Potential barrier height [eV]. Default 0.3 eV.

box: float
well: float
barrier: float
spacer: float
pot: float
class lww_transport.GeometryRegion(name, start_nm, end_nm, potential_ev, kind)[source]

Named interval in the one-dimensional RTD structure.

Parameters:

  • name (str)

  • start_nm (float)

  • end_nm (float)

  • potential_ev (float)

  • kind (str)

name: str
start_nm: float
end_nm: float
potential_ev: float
kind: str
property width_nm: float
class lww_transport.LWWConfig(discretization=None, material=None, geometry=None, operating=None, solver=None, compute=None, **flat_overrides)[source]

Unified 1D LWW quantum transport simulator configuration.

Parameters are grouped by discretization, material, geometry, operating conditions, solver controls, and compute backend.

Parameters:

  • discretization (DiscretizationParams)

  • material (MaterialParams)

  • geometry (GeometryParams)

  • operating (OperatingConditions)

  • solver (SolverParams)

  • compute (ComputeParams)

discretization

Grid resolution (nx, n).

Type:

DiscretizationParams

material

Semiconductor properties (GaAs defaults).

Type:

MaterialParams

geometry

Device structure (double-barrier QW).

Type:

GeometryParams

operating

Bias voltage and temperature.

Type:

OperatingConditions

solver

Numerical solver settings.

Type:

SolverParams

compute

Backend and performance options.

Type:

ComputeParams

Examples

>>> cfg = LWWConfig.standard_rtd()
>>> cfg = LWWConfig.quick_test(nx=32, n=32)
>>> cfg = LWWConfig(
...     discretization=DiscretizationParams(nx=48, n=48),
...     material=MaterialParams(temp=300.0),
...     operating=OperatingConditions(bias=0.1),
... )
x: ndarray
k: ndarray
discretization: DiscretizationParams
material: MaterialParams
geometry: GeometryParams
operating: OperatingConditions
solver: SolverParams
compute: ComputeParams
property nx: int
property n: int
property c: float
property rmo: float
property rm: float
property temp: float
property rhc: float
property re: float
property rno: float
property ri: float
property box: float
property well: float
property barrier: float
property spacer: float
property pot: float
property irlx: int
property chemp: float
property exchange: bool
property relaxation_tau: float
property hbar_like: float
property convergence_eps: float
property rfa: float
property poisson_vmax_stop: float
property density_stop: float
property dense_solver: str
property verbose: bool
property use_numba: bool
property kernel_backend: str
property rmass: float

Effective mass [eV·ps²/nm²].

property bbeta: float

Inverse thermal energy [1/eV].

property dens: float

Nominal electron density [1/nm³].

property densi: float

Ionized impurity density [1/nm³].

property delx: float

Spatial grid spacing [nm].

property delk: float

Momentum grid spacing [1/nm].

property size: int

Total Wigner phase-space dimension (nx × n).

with_bias(bias)[source]

Set the applied bias voltage.

Parameters:

bias (float)

Return type:

LWWConfig

with_temperature(temperature)[source]

Set operating temperature and return self.

Parameters:

temperature (float)

Return type:

LWWConfig

with_grid(nx, n)[source]

Set grid resolution and return self.

Parameters:

  • nx (int)

  • n (int)

Return type:

LWWConfig

with_exchange(exchange)[source]

Set density-dependent exchange correction.

Parameters:

exchange (bool)

Return type:

LWWConfig

with_verbose(verbose)[source]

Enable/disable verbose output.

Parameters:

verbose (bool)

Return type:

LWWConfig

classmethod quick_test(nx=16, n=16)[source]

Small-grid configuration for unit tests and smoke tests.

Parameters:

  • nx (int)

  • n (int)

Return type:

LWWConfig

classmethod standard_rtd()[source]

Standard resonant tunneling diode at 77K.

Reproduces the reference 1D RTD LWW configuration. - GaAs at 77K - Full resolution (86×72 grid) - Double-barrier QW geometry

Return type:

LWWConfig

classmethod room_temperature_rtd()[source]

RTD configuration at 300 K.

Return type:

LWWConfig

classmethod coarse_grid_rtd(nx=32, n=32)[source]

Coarse-grid RTD configuration for parameter sweeps.

Parameters:

  • nx (int)

  • n (int)

Return type:

LWWConfig

classmethod fine_grid_rtd(nx=128, n=96)[source]

High-resolution RTD configuration.

Parameters:

  • nx (int)

  • n (int)

Return type:

LWWConfig

classmethod narrow_well_rtd()[source]

RTD configuration with a 25 nm quantum well.

Return type:

LWWConfig

classmethod wide_well_rtd()[source]

RTD configuration with a 75 nm quantum well.

Return type:

LWWConfig

class lww_transport.LWW1DSimulator(config=None)[source]

One-dimensional LWW simulation workflow.

Parameters:

config (LWWConfig | None)

config_summary(extra=None)[source]

Return a human-readable summary for the simulator configuration.

Parameters:

extra (dict[str, object] | None)

Return type:

str

print_config_summary(extra=None)[source]

Print a human-readable summary for the simulator configuration.

Parameters:

extra (dict[str, object] | None)

Return type:

None

save_config_summary(output_dir, filename='config_summary.txt', extra=None)[source]

Save the simulator configuration summary next to simulation output.

Parameters:

  • output_dir (str | Path)

  • filename (str)

  • extra (dict[str, object] | None)

Return type:

Path

zeros_state()[source]
Return type:

SimulationState

initial_zero_bias_state(exchange=None)[source]

Compute the zero-bias initial Wigner state.

Parameters:

exchange (bool | None)

Return type:

SimulationState

solve_steady_state(bias=0.0, dbias=0.1, max_iterations=200, initial_state=None, exchange=None)[source]

Run the self-consistent steady-state Wigner-Poisson loop.

Parameters:

  • bias (float)

  • dbias (float)

  • max_iterations (int)

  • initial_state (SimulationState | None)

  • exchange (bool | None)

Return type:

SteadyStateResult

run_transient(state, ivn=45, itn=1000, dbias=0.008, exchange=None, sample_every=10, progress_every=50, output_dir=None, transient_prefix='lww_tcurl', state_prefix='lww')[source]

Run the transient bias sweep.

When cfg.verbose is true, print flushed progress lines at the start and end of each bias point and every progress_every iterations. If output_dir is supplied, transient current trace CSV files are written incrementally as samples are produced, and state checkpoint files are refreshed after each completed bias point.

Parameters:

  • state (SimulationState)

  • ivn (int)

  • itn (int)

  • dbias (float)

  • exchange (bool | None)

  • sample_every (int)

  • progress_every (int | None)

  • output_dir (str | Path | None)

  • transient_prefix (str)

  • state_prefix (str)

Return type:

TransientResult

save_state(state, output_dir, prefix='lww')[source]

Write state arrays in the legacy CSV format.

Parameters:
Return type:

None

load_state(input_dir, prefix='lww')[source]

Load state arrays from the legacy CSV format.

Parameters:

  • input_dir (str | Path)

  • prefix (str)

Return type:

SimulationState

save_transient(result, output_dir, prefix='lww_tcurl')[source]
Parameters:
Return type:

None

class lww_transport.MaterialParams(c=2997.92458, rmo=0.0568567517, rm=0.0667, temp=77.0, rhc=1973.29, re=2.0, rno=1e-06, ri=0.3, chemp=0.0863814)[source]

Physical constants and semiconductor band structure.

Defaults are for GaAs-like material at T=77K.

Parameters:

  • c (float) – Speed of light [nm/ps]. Default 2997.92.

  • rmo (float) – Effective mass ratio. Default 0.0569.

  • rm (float) – Effective mass scaling. Default 0.0667.

  • temp (float) – Temperature [K]. Default 77.0 K.

  • rhc (float) – ℏc product [eV·nm]. Default 1973.29.

  • re (float) – Electron density scaling. Default 2.0.

  • rno (float) – Nominal doping density [1/nm³]. Default 1e-6.

  • ri (float) – Ionized impurity ratio. Default 0.3.

  • chemp (float) – Equilibrium chemical potential [eV]. Default 0.0864 (at 77K).

c: float
rmo: float
rm: float
temp: float
rhc: float
re: float
rno: float
ri: float
chemp: float
class lww_transport.OperatingConditions(bias=0.0, temperature=77.0)[source]

Bias voltage and thermal settings.

Parameters:

  • bias (float) – Applied bias voltage [V]. Default 0.0 V (zero bias).

  • temperature (float) – Operating temperature [K]. Default 77.0 K (liquid nitrogen).

bias: float
temperature: float
class lww_transport.SolverParams(irlx=4, relaxation_tau=525.5074, hbar_like=0.6582186934, convergence_eps=7e-10, poisson_vmax_stop=0.001, density_stop=3e-09, exchange=False, rfa=650.0)[source]

Convergence and numerical method settings.

Parameters:

  • irlx (int) – Relaxation type flag. Default 4.

  • relaxation_tau (float) – Relaxation time parameter [ps]. Default 525.51.

  • hbar_like (float) – Planck constant scaling factor. Default 0.6582.

  • convergence_eps (float) – Density convergence threshold. Default 7e-10. Smaller values tighten convergence and increase runtime.

  • poisson_vmax_stop (float) – Potential convergence criterion [eV]. Default 0.001.

  • density_stop (float) – Density convergence criterion [1/nm³]. Default 3e-9.

  • exchange (bool) – Include optional density-dependent exchange correction. Default False.

  • rfa (float) – Relaxation factor for Poisson solver. Default 650.0.

irlx: int
relaxation_tau: float
hbar_like: float
convergence_eps: float
poisson_vmax_stop: float
density_stop: float
exchange: bool
rfa: float
class lww_transport.SimulationState(rvs: 'np.ndarray', f: 'np.ndarray', fr: 'np.ndarray', fo: 'np.ndarray', density_previous: 'np.ndarray', current_or_density: 'np.ndarray')[source]
Parameters:

  • rvs (ndarray)

  • f (ndarray)

  • fr (ndarray)

  • fo (ndarray)

  • density_previous (ndarray)

  • current_or_density (ndarray)

rvs: ndarray
f: ndarray
fr: ndarray
fo: ndarray
density_previous: ndarray
current_or_density: ndarray
copy()[source]
Return type:

SimulationState

class lww_transport.SteadyStateResult(state: 'SimulationState', density: 'np.ndarray', current: 'np.ndarray', iterations: 'int', converged: 'bool', max_density_change: 'float', max_potential_update: 'float')[source]
Parameters:

  • state (SimulationState)

  • density (ndarray)

  • current (ndarray)

  • iterations (int)

  • converged (bool)

  • max_density_change (float)

  • max_potential_update (float)

state: SimulationState
density: ndarray
current: ndarray
iterations: int
converged: bool
max_density_change: float
max_potential_update: float
class lww_transport.TransientResult(state: 'SimulationState', bias_currents: 'dict[float, pd.DataFrame]')[source]
Parameters:
state: SimulationState
bias_currents: dict[float, DataFrame]
lww_transport.format_config_summary(cfg, extra=None)[source]

Return a human-readable summary of the configuration.

Examples

>>> cfg = LWWConfig.standard_rtd()
>>> print(format_config_summary(cfg))
Parameters:

  • cfg (LWWConfig)

  • extra (Mapping[str, object] | None)

Return type:

str

lww_transport.geometry_potential_profile(cfg=None, points=1200)[source]

Return a piecewise RTD geometry potential profile.

Parameters:

  • cfg (LWWConfig | None)

  • points (int)

Return type:

tuple[ndarray, ndarray]

lww_transport.plot_rtd_geometry(cfg=None, ax=None, show_labels=True, title='Resonant tunneling diode geometry')[source]

Draw the one-dimensional RTD geometry and return the Matplotlib axes.

Parameters:

  • cfg (LWWConfig | None)

  • show_labels (bool)

  • title (str | None)

lww_transport.plot_wigner_phase_space(wigner, cfg=None, ax=None, figsize=(9.0, 6.4), x=None, k=None, title='Wigner phase-space distribution', style='standard', x_unit='um', centered_x=True, scale=1.0, normalize=False, x_lim=None, k_lim=None, z_lim=None, x_buffer=None, k_buffer=None, z_buffer=None, surface_cmap=None, contour_cmap=None, x_projection_cmap='Reds', y_projection_cmap='Blues', contour_levels=40, contour_offset=None, surface_alpha=None, z_projection_alpha=None, x_projection_alpha=None, y_projection_alpha=None, z_projection=True, x_projection=None, y_projection=None, colorbar=None, colorbar_label=None, colorbar_shrink=0.72, colorbar_pad=0.08, colorbar_aspect=18, x_projection_offset=None, y_projection_offset=None, surface_edgecolor='none', surface_linewidth=0.0, transparent_panes=None, show_grid=None, x_label=None, k_label=None, z_label=None, box_aspect=(1.35, 1.0, 0.72), stride=1, elev=None, azim=None)[source]

Draw a 3D Wigner phase-space surface with a contour projection.

wigner may be a flattened simulator vector, a (nx, n) matrix, or a CSV file containing either form. Flattened vectors require cfg so the phase-space dimensions can be restored.

Parameters:

  • wigner (str | PathLike[str] | ndarray)

  • cfg (LWWConfig | None)

  • figsize (tuple[float, float])

  • x (Sequence[float] | None)

  • k (Sequence[float] | None)

  • title (str | None)

  • style (str)

  • x_unit (str)

  • centered_x (bool)

  • scale (float)

  • normalize (bool)

  • x_lim (tuple[float, float] | None)

  • k_lim (tuple[float, float] | None)

  • z_lim (tuple[float, float] | None)

  • x_buffer (float | tuple[float, float] | None)

  • k_buffer (float | tuple[float, float] | None)

  • z_buffer (float | tuple[float, float] | None)

  • surface_cmap (str | None)

  • contour_cmap (str | None)

  • x_projection_cmap (str)

  • y_projection_cmap (str)

  • contour_levels (int)

  • contour_offset (float | None)

  • surface_alpha (float | None)

  • z_projection_alpha (float | None)

  • x_projection_alpha (float | None)

  • y_projection_alpha (float | None)

  • z_projection (bool)

  • x_projection (bool | None)

  • y_projection (bool | None)

  • colorbar (bool | None)

  • colorbar_label (str | None)

  • colorbar_shrink (float)

  • colorbar_pad (float)

  • colorbar_aspect (int)

  • x_projection_offset (float | None)

  • y_projection_offset (float | None)

  • surface_edgecolor (str)

  • surface_linewidth (float)

  • transparent_panes (bool | None)

  • show_grid (bool | None)

  • x_label (str | None)

  • k_label (str | None)

  • z_label (str | None)

  • box_aspect (tuple[float, float, float] | None)

  • stride (int)

  • elev (float | None)

  • azim (float | None)

lww_transport.print_config_summary(cfg, extra=None)[source]

Print a human-readable summary of the configuration.

Parameters:

  • cfg (LWWConfig)

  • extra (Mapping[str, object] | None)

Return type:

None

lww_transport.rtd_geometry_regions(cfg=None)[source]

Return emitter/spacer/barrier/well/collector regions for the RTD.

Parameters:

cfg (LWWConfig | None)

Return type:

list[GeometryRegion]

lww_transport.save_config_summary(cfg, output_dir, filename='config_summary.txt', extra=None)[source]

Save the human-readable configuration summary and return its path.

Parameters:

  • cfg (LWWConfig)

  • output_dir (str | Path)

  • filename (str)

  • extra (Mapping[str, object] | None)

Return type:

Path

lww_transport.save_rtd_geometry_image(cfg=None, path='rtd_geometry.png', dpi=160, show_labels=True, title='Resonant tunneling diode geometry')[source]

Save a PNG image of the RTD geometry and return its path.

Parameters:

  • cfg (LWWConfig | None)

  • path (str | Path)

  • dpi (int)

  • show_labels (bool)

  • title (str | None)

Return type:

Path

lww_transport.save_wigner_phase_space_image(wigner, path='wigner_phase_space.png', cfg=None, dpi=170, **plot_kwargs)[source]

Save a 3D Wigner phase-space image and return its path.

Parameters:

  • wigner (str | PathLike[str] | ndarray)

  • path (str | Path)

  • cfg (LWWConfig | None)

  • dpi (int)

Return type:

Path

lww_transport.save_wigner_phase_space_images(wigners, output_dir, cfg=None, prefix='wigner_phase_space', extension='png', dpi=170, **plot_kwargs)[source]

Save phase-space images for multiple Wigner distributions.

Parameters:

  • wigners (Mapping[object, str | PathLike[str] | ndarray] | Sequence[str | PathLike[str] | ndarray])

  • output_dir (str | Path)

  • cfg (LWWConfig | None)

  • prefix (str)

  • extension (str)

  • dpi (int)

Return type:

list[Path]

lww_transport.wigner_phase_space_grids(cfg=None, shape=None, centered_x=True, x_unit='um')[source]

Return spatial and wave-vector axes for Wigner phase-space plots.

Parameters:

  • cfg (LWWConfig | None)

  • shape (tuple[int, int] | None)

  • centered_x (bool)

  • x_unit (str)

Return type:

tuple[ndarray, ndarray]

Configuration

Grouped configuration objects for the LWW quantum transport simulator.

class lww_transport.config_refactored.ComputeParams(dense_solver='numpy', kernel_backend='auto', use_numba=True, verbose=False)[source]

Bases: object

Performance and debugging options.

Parameters:

  • dense_solver (str) – Linear solver backend. Default ‘numpy’.

  • kernel_backend (str) – Computational kernel implementation. Options: ‘auto’, ‘cpp’, ‘numba’, ‘python’. Default ‘auto’, which selects the fastest available backend.

  • use_numba (bool) – Enable Numba JIT compilation for hot loops. Default True.

  • verbose (bool) – Print diagnostic information during solve. Default False.

dense_solver: str
kernel_backend: str
use_numba: bool
verbose: bool
class lww_transport.config_refactored.DiscretizationParams(nx=86, n=72)[source]

Bases: object

Spatial and momentum grid parameters.

Parameters:

  • nx (int) – Number of spatial grid points. Default 86 matches the reference RTD configuration.

  • n (int) – Number of momentum grid points. Must be even. Default 72 matches the reference RTD configuration.

nx: int
n: int
class lww_transport.config_refactored.GeometryParams(box=550.0, well=50.0, barrier=30.0, spacer=30.0, pot=0.3)[source]

Bases: object

Double-barrier quantum well (DBQW) geometry.

Describes the device structure: emitter | spacer | barrier | well | barrier | spacer | collector

Parameters:

  • box (float) – Total device length [nm]. Default 550 nm.

  • well (float) – Quantum well width [nm]. Default 50 nm.

  • barrier (float) – Barrier position/height parameter [nm]. Default 30 nm.

  • spacer (float) – Spacer region width [nm]. Default 30 nm.

  • pot (float) – Potential barrier height [eV]. Default 0.3 eV.

box: float
well: float
barrier: float
spacer: float
pot: float
class lww_transport.config_refactored.LWWConfig(discretization=None, material=None, geometry=None, operating=None, solver=None, compute=None, **flat_overrides)[source]

Bases: object

Unified 1D LWW quantum transport simulator configuration.

Parameters are grouped by discretization, material, geometry, operating conditions, solver controls, and compute backend.

Parameters:

  • discretization (DiscretizationParams)

  • material (MaterialParams)

  • geometry (GeometryParams)

  • operating (OperatingConditions)

  • solver (SolverParams)

  • compute (ComputeParams)

discretization

Grid resolution (nx, n).

Type:

DiscretizationParams

material

Semiconductor properties (GaAs defaults).

Type:

MaterialParams

geometry

Device structure (double-barrier QW).

Type:

GeometryParams

operating

Bias voltage and temperature.

Type:

OperatingConditions

solver

Numerical solver settings.

Type:

SolverParams

compute

Backend and performance options.

Type:

ComputeParams

Examples

>>> cfg = LWWConfig.standard_rtd()
>>> cfg = LWWConfig.quick_test(nx=32, n=32)
>>> cfg = LWWConfig(
...     discretization=DiscretizationParams(nx=48, n=48),
...     material=MaterialParams(temp=300.0),
...     operating=OperatingConditions(bias=0.1),
... )
x: ndarray
k: ndarray
discretization: DiscretizationParams
material: MaterialParams
geometry: GeometryParams
operating: OperatingConditions
solver: SolverParams
compute: ComputeParams
property nx: int
property n: int
property c: float
property rmo: float
property rm: float
property temp: float
property rhc: float
property re: float
property rno: float
property ri: float
property box: float
property well: float
property barrier: float
property spacer: float
property pot: float
property irlx: int
property chemp: float
property exchange: bool
property relaxation_tau: float
property hbar_like: float
property convergence_eps: float
property rfa: float
property poisson_vmax_stop: float
property density_stop: float
property dense_solver: str
property verbose: bool
property use_numba: bool
property kernel_backend: str
property rmass: float

Effective mass [eV·ps²/nm²].

property bbeta: float

Inverse thermal energy [1/eV].

property dens: float

Nominal electron density [1/nm³].

property densi: float

Ionized impurity density [1/nm³].

property delx: float

Spatial grid spacing [nm].

property delk: float

Momentum grid spacing [1/nm].

property size: int

Total Wigner phase-space dimension (nx × n).

with_bias(bias)[source]

Set the applied bias voltage.

Parameters:

bias (float)

Return type:

LWWConfig

with_temperature(temperature)[source]

Set operating temperature and return self.

Parameters:

temperature (float)

Return type:

LWWConfig

with_grid(nx, n)[source]

Set grid resolution and return self.

Parameters:

  • nx (int)

  • n (int)

Return type:

LWWConfig

with_exchange(exchange)[source]

Set density-dependent exchange correction.

Parameters:

exchange (bool)

Return type:

LWWConfig

with_verbose(verbose)[source]

Enable/disable verbose output.

Parameters:

verbose (bool)

Return type:

LWWConfig

classmethod quick_test(nx=16, n=16)[source]

Small-grid configuration for unit tests and smoke tests.

Parameters:

  • nx (int)

  • n (int)

Return type:

LWWConfig

classmethod standard_rtd()[source]

Standard resonant tunneling diode at 77K.

Reproduces the reference 1D RTD LWW configuration. - GaAs at 77K - Full resolution (86×72 grid) - Double-barrier QW geometry

Return type:

LWWConfig

classmethod room_temperature_rtd()[source]

RTD configuration at 300 K.

Return type:

LWWConfig

classmethod coarse_grid_rtd(nx=32, n=32)[source]

Coarse-grid RTD configuration for parameter sweeps.

Parameters:

  • nx (int)

  • n (int)

Return type:

LWWConfig

classmethod fine_grid_rtd(nx=128, n=96)[source]

High-resolution RTD configuration.

Parameters:

  • nx (int)

  • n (int)

Return type:

LWWConfig

classmethod narrow_well_rtd()[source]

RTD configuration with a 25 nm quantum well.

Return type:

LWWConfig

classmethod wide_well_rtd()[source]

RTD configuration with a 75 nm quantum well.

Return type:

LWWConfig

class lww_transport.config_refactored.MaterialParams(c=2997.92458, rmo=0.0568567517, rm=0.0667, temp=77.0, rhc=1973.29, re=2.0, rno=1e-06, ri=0.3, chemp=0.0863814)[source]

Bases: object

Physical constants and semiconductor band structure.

Defaults are for GaAs-like material at T=77K.

Parameters:

  • c (float) – Speed of light [nm/ps]. Default 2997.92.

  • rmo (float) – Effective mass ratio. Default 0.0569.

  • rm (float) – Effective mass scaling. Default 0.0667.

  • temp (float) – Temperature [K]. Default 77.0 K.

  • rhc (float) – ℏc product [eV·nm]. Default 1973.29.

  • re (float) – Electron density scaling. Default 2.0.

  • rno (float) – Nominal doping density [1/nm³]. Default 1e-6.

  • ri (float) – Ionized impurity ratio. Default 0.3.

  • chemp (float) – Equilibrium chemical potential [eV]. Default 0.0864 (at 77K).

c: float
rmo: float
rm: float
temp: float
rhc: float
re: float
rno: float
ri: float
chemp: float
class lww_transport.config_refactored.OperatingConditions(bias=0.0, temperature=77.0)[source]

Bases: object

Bias voltage and thermal settings.

Parameters:

  • bias (float) – Applied bias voltage [V]. Default 0.0 V (zero bias).

  • temperature (float) – Operating temperature [K]. Default 77.0 K (liquid nitrogen).

bias: float
temperature: float
class lww_transport.config_refactored.SolverParams(irlx=4, relaxation_tau=525.5074, hbar_like=0.6582186934, convergence_eps=7e-10, poisson_vmax_stop=0.001, density_stop=3e-09, exchange=False, rfa=650.0)[source]

Bases: object

Convergence and numerical method settings.

Parameters:

  • irlx (int) – Relaxation type flag. Default 4.

  • relaxation_tau (float) – Relaxation time parameter [ps]. Default 525.51.

  • hbar_like (float) – Planck constant scaling factor. Default 0.6582.

  • convergence_eps (float) – Density convergence threshold. Default 7e-10. Smaller values tighten convergence and increase runtime.

  • poisson_vmax_stop (float) – Potential convergence criterion [eV]. Default 0.001.

  • density_stop (float) – Density convergence criterion [1/nm³]. Default 3e-9.

  • exchange (bool) – Include optional density-dependent exchange correction. Default False.

  • rfa (float) – Relaxation factor for Poisson solver. Default 650.0.

irlx: int
relaxation_tau: float
hbar_like: float
convergence_eps: float
poisson_vmax_stop: float
density_stop: float
exchange: bool
rfa: float
lww_transport.config_refactored.format_config_summary(cfg, extra=None)[source]

Return a human-readable summary of the configuration.

Examples

>>> cfg = LWWConfig.standard_rtd()
>>> print(format_config_summary(cfg))
Parameters:

  • cfg (LWWConfig)

  • extra (Mapping[str, object] | None)

Return type:

str

lww_transport.config_refactored.print_config_summary(cfg, extra=None)[source]

Print a human-readable summary of the configuration.

Parameters:

  • cfg (LWWConfig)

  • extra (Mapping[str, object] | None)

Return type:

None

lww_transport.config_refactored.save_config_summary(cfg, output_dir, filename='config_summary.txt', extra=None)[source]

Save the human-readable configuration summary and return its path.

Parameters:

  • cfg (LWWConfig)

  • output_dir (str | Path)

  • filename (str)

  • extra (Mapping[str, object] | None)

Return type:

Path

Simulator

class lww_transport.simulator.SimulationState(rvs: 'np.ndarray', f: 'np.ndarray', fr: 'np.ndarray', fo: 'np.ndarray', density_previous: 'np.ndarray', current_or_density: 'np.ndarray')[source]

Bases: object

Parameters:

  • rvs (ndarray)

  • f (ndarray)

  • fr (ndarray)

  • fo (ndarray)

  • density_previous (ndarray)

  • current_or_density (ndarray)

rvs: ndarray
f: ndarray
fr: ndarray
fo: ndarray
density_previous: ndarray
current_or_density: ndarray
copy()[source]
Return type:

SimulationState

class lww_transport.simulator.SteadyStateResult(state: 'SimulationState', density: 'np.ndarray', current: 'np.ndarray', iterations: 'int', converged: 'bool', max_density_change: 'float', max_potential_update: 'float')[source]

Bases: object

Parameters:

  • state (SimulationState)

  • density (ndarray)

  • current (ndarray)

  • iterations (int)

  • converged (bool)

  • max_density_change (float)

  • max_potential_update (float)

state: SimulationState
density: ndarray
current: ndarray
iterations: int
converged: bool
max_density_change: float
max_potential_update: float
class lww_transport.simulator.TransientResult(state: 'SimulationState', bias_currents: 'dict[float, pd.DataFrame]')[source]

Bases: object

Parameters:
state: SimulationState
bias_currents: dict[float, DataFrame]
class lww_transport.simulator.LWW1DSimulator(config=None)[source]

Bases: object

One-dimensional LWW simulation workflow.

Parameters:

config (LWWConfig | None)

config_summary(extra=None)[source]

Return a human-readable summary for the simulator configuration.

Parameters:

extra (dict[str, object] | None)

Return type:

str

print_config_summary(extra=None)[source]

Print a human-readable summary for the simulator configuration.

Parameters:

extra (dict[str, object] | None)

Return type:

None

save_config_summary(output_dir, filename='config_summary.txt', extra=None)[source]

Save the simulator configuration summary next to simulation output.

Parameters:

  • output_dir (str | Path)

  • filename (str)

  • extra (dict[str, object] | None)

Return type:

Path

zeros_state()[source]
Return type:

SimulationState

initial_zero_bias_state(exchange=None)[source]

Compute the zero-bias initial Wigner state.

Parameters:

exchange (bool | None)

Return type:

SimulationState

solve_steady_state(bias=0.0, dbias=0.1, max_iterations=200, initial_state=None, exchange=None)[source]

Run the self-consistent steady-state Wigner-Poisson loop.

Parameters:

  • bias (float)

  • dbias (float)

  • max_iterations (int)

  • initial_state (SimulationState | None)

  • exchange (bool | None)

Return type:

SteadyStateResult

run_transient(state, ivn=45, itn=1000, dbias=0.008, exchange=None, sample_every=10, progress_every=50, output_dir=None, transient_prefix='lww_tcurl', state_prefix='lww')[source]

Run the transient bias sweep.

When cfg.verbose is true, print flushed progress lines at the start and end of each bias point and every progress_every iterations. If output_dir is supplied, transient current trace CSV files are written incrementally as samples are produced, and state checkpoint files are refreshed after each completed bias point.

Parameters:

  • state (SimulationState)

  • ivn (int)

  • itn (int)

  • dbias (float)

  • exchange (bool | None)

  • sample_every (int)

  • progress_every (int | None)

  • output_dir (str | Path | None)

  • transient_prefix (str)

  • state_prefix (str)

Return type:

TransientResult

save_state(state, output_dir, prefix='lww')[source]

Write state arrays in the legacy CSV format.

Parameters:
Return type:

None

load_state(input_dir, prefix='lww')[source]

Load state arrays from the legacy CSV format.

Parameters:

  • input_dir (str | Path)

  • prefix (str)

Return type:

SimulationState

save_transient(result, output_dir, prefix='lww_tcurl')[source]
Parameters:
Return type:

None

Core numerical routines

Optimized core routines for the 1D Lattice Weyl-Wigner / Wigner-Poisson model.

Implementation notes: - solve the Wigner linear system directly with LAPACK band storage - avoid converting band matrices to dense Nx*N by Nx*N arrays - vectorize Fermi boundaries, current/density, Poisson charge sums, and potential coefficients - cache kinetic stencil and sine matrix used in the non-local potential

class lww_transport.core.Params(Nx: 'int' = 86, N: 'int' = 72, c: 'float' = 2997.92458, rmo: 'float' = 0.0568567517, rm: 'float' = 0.0667, temp: 'float' = 77.0, rhc: 'float' = 1973.29, re: 'float' = 2.0, rno: 'float' = 1e-06, ri: 'float' = 0.3, box: 'float' = 550.0, well: 'float' = 50.0, barrier: 'float' = 30.0, spacer: 'float' = 30.0, pot: 'float' = 0.3, chemp: 'float' = 0.0863814)[source]

Bases: object

Parameters:

  • Nx (int)

  • N (int)

  • c (float)

  • rmo (float)

  • rm (float)

  • temp (float)

  • rhc (float)

  • re (float)

  • rno (float)

  • ri (float)

  • box (float)

  • well (float)

  • barrier (float)

  • spacer (float)

  • pot (float)

  • chemp (float)

Nx: int = 86
N: int = 72
c: float = 2997.92458
rmo: float = 0.0568567517
rm: float = 0.0667
temp: float = 77.0
rhc: float = 1973.29
re: float = 2.0
rno: float = 1e-06
ri: float = 0.3
box: float = 550.0
well: float = 50.0
barrier: float = 30.0
spacer: float = 30.0
pot: float = 0.3
chemp: float = 0.0863814
property rmass: float
property bbeta: float
property dens: float
property densi: float
property delx: float
property delk: float
property x: ndarray
class lww_transport.core.WignerSolveWorkspace(n, nx, lower_upper, lapack_ab, rhs, rvs_work)[source]

Bases: object

Reusable buffers for the Wigner banded solve.

The LAPACK buffer uses dgbsv storage directly. Its lower view has the same shape as SciPy’s compact band format, so the existing assembly kernels can fill it while LAPACK can factor the full buffer without another matrix allocation/copy inside solve_banded.

Parameters:

  • n (int)

  • nx (int)

  • lower_upper (int)

  • lapack_ab (ndarray)

  • rhs (ndarray)

  • rvs_work (ndarray)

n: int
nx: int
lower_upper: int
lapack_ab: ndarray
rhs: ndarray
rvs_work: ndarray
classmethod create(n=72, nx=86)[source]
Parameters:

  • n (int)

  • nx (int)

Return type:

WignerSolveWorkspace

property band: ndarray
lww_transport.core.band_to_agb(a, params=None)[source]

Convert dense square matrix to LAPACK/SciPy general band storage.

Parameters:

  • a (ndarray)

  • params (Params | None)

Return type:

ndarray

lww_transport.core.agb_to_band(agb, params=None)[source]

Convert LAPACK/SciPy general band storage to a dense square matrix.

Parameters:

  • agb (ndarray)

  • params (Params | None)

Return type:

ndarray

lww_transport.core.integral(x_grid, y, axis=0)[source]
Parameters:

  • x_grid (ndarray)

  • y (ndarray)

  • axis (int)

Return type:

ndarray

lww_transport.core.get_exchange(rho, x_grid=None)[source]
Parameters:

  • rho (ndarray)

  • x_grid (ndarray | None)

Return type:

ndarray

lww_transport.core.get_hartree(rho, x_grid, eps=0.1)[source]
Parameters:

  • rho (ndarray)

  • x_grid (ndarray)

  • eps (float)

Return type:

ndarray

lww_transport.core.aicalc(rvs, ai=None, n=72, nx=86, i=1, params=None)[source]

Vectorized non-local potential coefficient for one 1-based spatial index i.

Parameters:

  • rvs (ndarray)

  • ai (ndarray | None)

  • n (int)

  • nx (int)

  • i (int)

  • params (Params | None)

Return type:

ndarray

lww_transport.core.potential(external_rvs, rho, n=72, nx=86, exchange=False, params=None, use_numba=True, kernel_backend='auto')[source]

Return potential matrix directly in banded storage: shape (4*n+1, n*nx).

Parameters:

  • external_rvs (ndarray)

  • rho (ndarray)

  • n (int)

  • nx (int)

  • exchange (bool)

  • params (Params | None)

  • use_numba (bool | None)

  • kernel_backend (str)

Return type:

ndarray

lww_transport.core.arelax(S, b, n=72, nx=86, params=None, use_numba=True, kernel_backend='auto')[source]

Relaxation/scattering band matrix. No debug printing.

Parameters:

  • S (ndarray)

  • b (ndarray)

  • n (int)

  • nx (int)

  • params (Params | None)

  • use_numba (bool | None)

  • kernel_backend (str)

Return type:

ndarray

lww_transport.core.scattering(b, n=72, nx=86, params=None, use_numba=True, kernel_backend='auto')[source]
Parameters:

  • b (ndarray)

  • n (int)

  • nx (int)

  • params (Params | None)

  • use_numba (bool | None)

  • kernel_backend (str)

Return type:

ndarray

lww_transport.core.tridag(a, b, c, r, nx)[source]

Thomas algorithm for tridiagonal systems.

Parameters:

  • a (ndarray)

  • b (ndarray)

  • c (ndarray)

  • r (ndarray)

  • nx (int)

Return type:

ndarray

lww_transport.core.initial_state(params=None, exchange=True, irlx_value=0)[source]

Compute the default steady-state initialization.

Parameters:

  • params (Params | None)

  • exchange (bool)

  • irlx_value (int)

lww_transport.core.fbndry(*args, **kwargs)[source]
Return type:

ndarray

lww_transport.core.fermi_function(*args, **kwargs)[source]
Return type:

ndarray

lww_transport.core.rvscalc(*args, **kwargs)[source]
Return type:

ndarray

lww_transport.core.kinetic(*args, **kwargs)[source]
Return type:

ndarray

lww_transport.core.wigstd(*args, **kwargs)[source]
Return type:

ndarray

lww_transport.core.poissonb(*args, **kwargs)[source]
Return type:

ndarray

lww_transport.core.poisson(*args, **kwargs)[source]
Return type:

ndarray

lww_transport.core.curcalc(*args, **kwargs)[source]
Return type:

ndarray

lww_transport.core.comp(*args, **kwargs)[source]
Return type:

tuple[int, float]

lww_transport.core.subsrvs(*args, **kwargs)[source]
Return type:

ndarray

Visualization helpers

Visualization helpers for LWW device geometry and phase-space data.

class lww_transport.visualization.GeometryRegion(name, start_nm, end_nm, potential_ev, kind)[source]

Bases: object

Named interval in the one-dimensional RTD structure.

Parameters:

  • name (str)

  • start_nm (float)

  • end_nm (float)

  • potential_ev (float)

  • kind (str)

name: str
start_nm: float
end_nm: float
potential_ev: float
kind: str
property width_nm: float
lww_transport.visualization.geometry_potential_profile(cfg=None, points=1200)[source]

Return a piecewise RTD geometry potential profile.

Parameters:

  • cfg (LWWConfig | None)

  • points (int)

Return type:

tuple[ndarray, ndarray]

lww_transport.visualization.plot_rtd_geometry(cfg=None, ax=None, show_labels=True, title='Resonant tunneling diode geometry')[source]

Draw the one-dimensional RTD geometry and return the Matplotlib axes.

Parameters:

  • cfg (LWWConfig | None)

  • show_labels (bool)

  • title (str | None)

lww_transport.visualization.plot_wigner_phase_space(wigner, cfg=None, ax=None, figsize=(9.0, 6.4), x=None, k=None, title='Wigner phase-space distribution', style='standard', x_unit='um', centered_x=True, scale=1.0, normalize=False, x_lim=None, k_lim=None, z_lim=None, x_buffer=None, k_buffer=None, z_buffer=None, surface_cmap=None, contour_cmap=None, x_projection_cmap='Reds', y_projection_cmap='Blues', contour_levels=40, contour_offset=None, surface_alpha=None, z_projection_alpha=None, x_projection_alpha=None, y_projection_alpha=None, z_projection=True, x_projection=None, y_projection=None, colorbar=None, colorbar_label=None, colorbar_shrink=0.72, colorbar_pad=0.08, colorbar_aspect=18, x_projection_offset=None, y_projection_offset=None, surface_edgecolor='none', surface_linewidth=0.0, transparent_panes=None, show_grid=None, x_label=None, k_label=None, z_label=None, box_aspect=(1.35, 1.0, 0.72), stride=1, elev=None, azim=None)[source]

Draw a 3D Wigner phase-space surface with a contour projection.

wigner may be a flattened simulator vector, a (nx, n) matrix, or a CSV file containing either form. Flattened vectors require cfg so the phase-space dimensions can be restored.

Parameters:

  • wigner (str | PathLike[str] | ndarray)

  • cfg (LWWConfig | None)

  • figsize (tuple[float, float])

  • x (Sequence[float] | None)

  • k (Sequence[float] | None)

  • title (str | None)

  • style (str)

  • x_unit (str)

  • centered_x (bool)

  • scale (float)

  • normalize (bool)

  • x_lim (tuple[float, float] | None)

  • k_lim (tuple[float, float] | None)

  • z_lim (tuple[float, float] | None)

  • x_buffer (float | tuple[float, float] | None)

  • k_buffer (float | tuple[float, float] | None)

  • z_buffer (float | tuple[float, float] | None)

  • surface_cmap (str | None)

  • contour_cmap (str | None)

  • x_projection_cmap (str)

  • y_projection_cmap (str)

  • contour_levels (int)

  • contour_offset (float | None)

  • surface_alpha (float | None)

  • z_projection_alpha (float | None)

  • x_projection_alpha (float | None)

  • y_projection_alpha (float | None)

  • z_projection (bool)

  • x_projection (bool | None)

  • y_projection (bool | None)

  • colorbar (bool | None)

  • colorbar_label (str | None)

  • colorbar_shrink (float)

  • colorbar_pad (float)

  • colorbar_aspect (int)

  • x_projection_offset (float | None)

  • y_projection_offset (float | None)

  • surface_edgecolor (str)

  • surface_linewidth (float)

  • transparent_panes (bool | None)

  • show_grid (bool | None)

  • x_label (str | None)

  • k_label (str | None)

  • z_label (str | None)

  • box_aspect (tuple[float, float, float] | None)

  • stride (int)

  • elev (float | None)

  • azim (float | None)

lww_transport.visualization.rtd_geometry_regions(cfg=None)[source]

Return emitter/spacer/barrier/well/collector regions for the RTD.

Parameters:

cfg (LWWConfig | None)

Return type:

list[GeometryRegion]

lww_transport.visualization.save_rtd_geometry_image(cfg=None, path='rtd_geometry.png', dpi=160, show_labels=True, title='Resonant tunneling diode geometry')[source]

Save a PNG image of the RTD geometry and return its path.

Parameters:

  • cfg (LWWConfig | None)

  • path (str | Path)

  • dpi (int)

  • show_labels (bool)

  • title (str | None)

Return type:

Path

lww_transport.visualization.save_wigner_phase_space_image(wigner, path='wigner_phase_space.png', cfg=None, dpi=170, **plot_kwargs)[source]

Save a 3D Wigner phase-space image and return its path.

Parameters:

  • wigner (str | PathLike[str] | ndarray)

  • path (str | Path)

  • cfg (LWWConfig | None)

  • dpi (int)

Return type:

Path

lww_transport.visualization.save_wigner_phase_space_images(wigners, output_dir, cfg=None, prefix='wigner_phase_space', extension='png', dpi=170, **plot_kwargs)[source]

Save phase-space images for multiple Wigner distributions.

Parameters:

  • wigners (Mapping[object, str | PathLike[str] | ndarray] | Sequence[str | PathLike[str] | ndarray])

  • output_dir (str | Path)

  • cfg (LWWConfig | None)

  • prefix (str)

  • extension (str)

  • dpi (int)

Return type:

list[Path]

lww_transport.visualization.wigner_phase_space_grids(cfg=None, shape=None, centered_x=True, x_unit='um')[source]

Return spatial and wave-vector axes for Wigner phase-space plots.

Parameters:

  • cfg (LWWConfig | None)

  • shape (tuple[int, int] | None)

  • centered_x (bool)

  • x_unit (str)

Return type:

tuple[ndarray, ndarray]

Linear algebra helpers

lww_transport.linalg.band_to_agb(a, n, nx)[source]

Convert a dense matrix to general band storage.

Parameters:

  • a (ndarray)

  • n (int)

  • nx (int)

Return type:

ndarray

lww_transport.linalg.agb_to_dense(agb, n, nx)[source]

Convert general band storage to a dense matrix.

Parameters:

  • agb (ndarray)

  • n (int)

  • nx (int)

Return type:

ndarray

lww_transport.linalg.tridag(a, b, c, r)[source]

Thomas algorithm for a tridiagonal system.

Raises an explicit error if a zero pivot is encountered.

Parameters:

  • a (ndarray)

  • b (ndarray)

  • c (ndarray)

  • r (ndarray)

Return type:

ndarray

CLI module

lww_transport.cli.main(argv=None)[source]
Parameters:

argv (list[str] | None)

Return type:

int