Kwant

Official Resources

• Homepage: https://kwant-project.org/
• Documentation: https://kwant-project.org/doc/
• Repository: https://gitlab.kwant-project.org/kwant/kwant
• License: BSD 2-Clause License

Overview

Kwant is a powerful, open-source Python package for numerical quantum transport calculations. It allows for the construction of tight-binding models with arbitrary shapes and dimensionality and the calculation of their transport properties using the scattering matrix formalism (Landauer-Büttiker). Kwant is widely considered the community standard for mesoscopic transport due to its flexibility, ease of use, and "Builder" pattern which decouples the physics (Hamiltonian) from the geometry.

Scientific domain: Mesoscopic Physics, Topological Matter, Quantum Transport Target user community: Theorists and experimentalists simulating quantum devices (QPCs, Hall bars, nanowires)

Theoretical Methods

• Tight-Binding: Discrete lattice models with arbitrary hopping terms.
• Scattering Matrix ($S$): Calculated via the wave-function matching method (or recursive Green's functions) for open systems connected to semi-infinite leads.
• Landauer Formalism: Conductance $G = \frac{2e^2}{h} \sum_n T_n$.
• Green's Functions: Integration for local quantities like density of states (LDOS).

Capabilities

• System Construction:
  • Arbitrary lattices (1D, 2D, 3D) and complex shapes.
  • Symmetries (translational, rotational) handling.
• Observables:
  • Conductance and Shot Noise.
  • S-matrix elements (transmission/reflection amplitudes).
  • Wavefunctions $\psi(\mathbf{r})$ in the scattering region.
  • Local currents and spin densities.
• Physics:
  • Quantum Hall Effect (magnetic Peierls phases).
  • Superconductivity (Bogoliubov-de Gennes).
  • Topological insulators and Majorana modes.
  • Spin-Orbit Coupling.

Key Strengths

• Flexibility: The "Builder" interface is extremely intuitive: syst[site] = potential.
• Performance: While the interface is Python, the heavy lifting is done by efficient C/Cython cores and sparse linear algebra (MUMPS, UMFPACK).
• Ecosystem: Highly extensible (e.g., Tkwant for time-dependent transport) and integrates perfectly with the SciPy stack.

Inputs & Outputs

• Inputs: Python scripts defining the lattice, shape functions, and Hamiltonian values.
• Outputs:
  • S-matrices (NumPy arrays).
  • Scalar fields (LDOS) mapped to sites.
  • Band structures of leads.

Interfaces & Ecosystem

• Visualization: Built-in plotting using Matplotlib (kwant.plot).
• Tkwant: Extension for time-dependent transport.
• Qsymm: Symmetry analysis of Hamiltonians.

Performance Characteristics

• Scaling: Efficient for 2D/3D systems using sparse direct solvers ($O(N^{1.5...2.0})$ typically).
• Parallelism: Comparisons over parameters (energy, field) are embarrassingly parallel (e.g., kwant.parallel).

Comparison with Other Codes

• vs. OMEN/NEMO5: Kwant is for model Hamiltonians (physics concepts); OMEN/NEMO5 are for atomistic material simulations (device engineering).
• vs. Quantica.jl: Quantica is a Julia-based spiritual successor/alternative to Kwant with higher performance for creating Hamiltonians but a smaller ecosystem.

Application Areas

• Majorana Fermions: Simulating signatures of topological superconductivity in nanowires.
• Quantum Hall: Edge state transport and interference in interferometers.
• Graphene: Veselago lensing and p-n junction transport.

Community and Support

• Development: CEA Grenoble (Xavier Waintal) and TU Delft (Anton Akhmerov).
• Source: GitLab.

Verification & Sources

• Website: https://kwant-project.org/
• Primary Publication: C. W. Groth et al., New J. Phys. 16, 063065 (2014).
• Verification status: ✅ VERIFIED
  • The Gold Standard for mesoscopic quantum transport.