TightBinding++ (TB++)

Official Resources

• Homepage: https://tightbinding.github.io/
• Repository: https://github.com/TightBinding/tbpp
• License: MPL-2.0

Overview

TightBinding++ is a modern C++11 framework for efficient simulation of tight-binding models. Designed with modularity and performance in mind, it provides a unified interface to compute a wide range of electronic properties, from standard band structures to advanced topological invariants like Berry curvature and Chern numbers. It supports parallel execution via OpenMP and HDF5 I/O, making it suitable for both rapid prototyping and data-intensive research.

Scientific domain: Band Theory, Topological Matter Target user community: C++ developers needing a TB library backend

Theoretical Methods

• Tight-Binding: General orthogonal tight-binding models.
• Linear Response: Kubo formula for optical conductivity $\sigma_{\alpha\beta}(\omega)$.
• Topology:
  • Berry connection and curvature $\Omega(\mathbf{k})$.
  • Chern numbers via integration over the Brillouin Zone.
• Green's Functions: Iterative methods for DOS and transport.

Capabilities

• Model Construction:
  • Simple API for adding sites and hoppings.
  • 1D, 2D, 3D lattices.
• Observables:
  • Band structures (complex and real).
  • Density of States (DOS).
  • Berry phase properties.
• I/O:
  • HDF5 support for storing large result sets.
  • Python bindings (experimental) allow scripting.

Key Strengths

• Performance: Written in highly optimized C++11 with OpenMP, ensuring fast execution on multi-core machines.
• Modern Design: Uses modern C++ paradigms (RAII, smart pointers) making the library safer and easier to use than legacy Fortran/C codes.
• Topology Native: Unlike many older TB codes, topological quantities (Berry curvature) are first-class citizens in the API.

Inputs & Outputs

• Inputs: C++ driver code or input parameter files.
• Outputs: HDF5 files containing bands, curvature, etc.

Interfaces & Ecosystem

• Dependencies: Eigen3, HDF5, FFTW.
• Visualization: Helper scripts for plotting results.

Performance Characteristics

• Efficiency: Matrix operations are vectorized via Eigen.
• Parallelism: OpenMP threading for k-point loops.

Comparison with Other Codes

• vs. TBTK: Both are C++ TB libraries. TightBinding++ has a slightly stronger focus on standard band/topology tasks, while TBTK focuses on general graph-based models and Green's functions.
• vs. PythTB: PythTB is Python (slower, easier); TightBinding++ is C++ (faster, steeper learning curve).

Application Areas

• Topological Photovoltaics: Calculating shift currents (via Kubo).
• Material Screening: Fast band structure generation for databases.

Community and Support

• Development: TightBinding team (GitHub).
• Source: GitHub.

Verification & Sources

• Website: https://tightbinding.github.io/
• Verification status: ✅ VERIFIED
  • Valid C++ framework.