Potentials

Every potential implements evaluate(grid) and returns V on the grid. Most work in 1D, 2D and 3D. They can be created directly or by name through create_potential, which the YAML runner uses:

from quantumlab.potentials import create_potential, HarmonicTrap

trap = HarmonicTrap((1.0, 1.0, 0.5))                 # anisotropic 3D trap
screen = create_potential('double_slit', V0=60.0, separation=5.0, aperture='pinhole')

Registry name	Class	Description
gaussian_barrier, barrier	GaussianBarrier	1D Gaussian barrier
gaussian_2d_barrier	Gaussian2DBarrier	radially symmetric barrier in 2D (or a cylinder in 3D)
gaussian_3d_barrier	Gaussian3DBarrier	spherical barrier
rectangular_barrier	RectangularBarrier	square barrier
potential_step	PotentialStep	step potential
multiple_barriers	MultipleBarriers	periodic array of barriers
rtd	ResonantTunnelingDiode	double-barrier resonant tunnelling structure
infinite_well, finite_well	InfiniteSquareWell, FiniteSquareWell	square wells
double_well	DoubleWell	quartic double well
harmonic	HarmonicOscillator	isotropic harmonic oscillator
harmonic_trap	HarmonicTrap	anisotropic harmonic trap with an optional centre
aperture_screen, double_slit	ApertureScreen	opaque screen with N slits or circular pinholes (2D/3D)
stadium	StadiumBilliard	soft-walled Bunimovich stadium (classically chaotic)
soft_coulomb	SoftCoulomb	soft-core Coulomb potential −Z/√(r² + a²); hydrogen model for strong-field physics
crystal	CrystalPotential	sinusoidal lattice
disorder	RandomDisorder	cell-based random potential

Sharp edges (screens, billiards) are smoothed over a short length with a tanh profile, so the potential stays resolvable on the FFT grid.

Custom potentials

from quantumlab.potentials import CustomPotential
V = CustomPotential(lambda x, y, z: 0.5 * (x**2 + y**2) + 0.1 * z**4)

Time-dependent potentials

import numpy as np
from quantumlab.potentials import DipoleDrive, HarmonicTrap, TimeDependentCustomPotential, sin2_envelope

# static potential plus a uniform oscillating force, F(t) = F0 env(t) sin(ωt + φ)
drive = DipoleDrive(base=HarmonicTrap(1.0), F0=0.5, omega=0.6, axis='x', envelope=sin2_envelope(40.0))

# any V(r, t)
shaken = TimeDependentCustomPotential(lambda x, t: 0.5 * (x - 0.5 * np.sin(t))**2)

Envelopes: sin2_envelope(duration) and flat_top_envelope(duration, ramp).

Solvers detect these through time_dependent = True and call evaluate_at(grid, t) at every sub-step.