twister2d 2.0.1

Twisted 2D material structure generation and analysis (graphene, hBN, TMDs) with LAMMPS integration

Twister2.0

Twister2.0 builds commensurate supercells of twisted 2D materials (graphene, hBN, transition-metal dichalcogenides), generates ready-to-run LAMMPS structure and input files with automatic interlayer-potential configuration, and post-processes relaxed structures into per-layer strain and interlayer-spacing maps. It also includes a parallel commensurate-twist-angle finder for lattice-mismatched heterobilayers and converters to DFT input formats (Quantum ESPRESSO, SIESTA).

Repository layout

Twister2.0/
├── pyproject.toml     # pip-installable package definition
├── twister2/          # Python package (source)
├── potentials/
│   ├── classical/     # KC, SW, Tersoff, REBO, ILP potential files
│   └── ml/            # PANNA (panna/vdw) config + LAMMPS plugin sources
├── docs/              # detailed documentation
└── examples/          # example input files for each workflow

Installation

pip install twister2d      # from PyPI
# or from a checkout:
cd Twister2.0
pip install .              # or: pip install -e .  (editable, for development)
pip install ".[mpi]"       # optional: MPI support for the angle finder

(The PyPI distribution is named twister2d; the installed commands are twister2, twister2-process, etc.)

This installs the library and the command-line tools:

Command	Purpose
twister2	Structure generation (reads twisterase.inp)
twister2-process	Extract layers from a relaxed LAMMPS dump
twister2-analyze	Strain + interlayer-spacing analysis pipeline
twister2-angle	Commensurate twist angle search (reads get_ang.inp)
twister2-cif2qe	CIF → Quantum ESPRESSO input
twister2-cif2siesta	CIF → SIESTA input
twister2-remap-types	Remap LAMMPS atom types to PANNA ordering (manual use)
twister2-restore-tags	Restore original atom tags in an ML dump (manual use)

See docs/installation.md for details, including LAMMPS build requirements.

Quick start

cd examples/tmd-twisted-bilayer-hex
cp ../../potentials/classical/{MoWSSe.KC,tmd.sw} .
twister2                        # -> structure.lammps, lammps.in, process_relaxed.inp
lmp_serial -in lammps.in        # relax with LAMMPS -> dump.Final
twister2-process                # -> layer_N.cif, layer_N_coords.dat
twister2-analyze                # -> StrainMap/, InterlayerSpacingMap/

The full walkthrough is in docs/quickstart.md.

Documentation

Document	Contents
docs/installation.md	Installing the package, dependencies, LAMMPS requirements
docs/quickstart.md	A first end-to-end run, step by step
docs/cli.md	Complete command-line reference
docs/workflow.md	The four-stage pipeline in detail
docs/input_files.md	Full reference for twisterase.inp, layer*.inp, process_relaxed.inp
docs/hbn.md	Automatic hBN substrate / cap / encapsulation
docs/ml_potentials.md	Machine-learned potential (panna/vdw) workflow
docs/angle_finder.md	Commensurate twist angle search (twister2-angle)
docs/analysis.md	Layer extraction and post-processing
docs/converters.md	CIF → QE / SIESTA converters
docs/potentials.md	The bundled potential files and how to use them

Requirements

• Python ≥ 3.10 with numpy, scipy, matplotlib, ase (installed automatically by pip)
• LAMMPS with the MANYBODY and INTERLAYER packages (classical workflow), plus the bundled PANNA plugin for the ML workflow
• Optional: mpi4py + an MPI library to parallelize the angle finder