magtrans 1.0

MagTrans: Magnetic Transition Estimator

MagTrans — Magnetic Transition Estimator

MagTrans is a fully automated pipeline for predicting magnetic transition temperatures (Curie and Néel) from first principles:

• Enumerates all collinear spin configurations in a crystal (1D, 2D, 3D)
• Performs ab initio DFT relaxations, static, and SOC calculations via VASP + Custodian
• Fits a Heisenberg + anisotropy Hamiltonian
• Executes Monte Carlo simulations to extract transition temperatures

---

🚀 Features

• End-to-end automation: enumeration → DFT → Hamiltonian → Monte Carlo
• Symmetry-aware collinear spin enumeration using KD-Tree + space group reduction
• Lightweight config via a plain-text input file
• Support for high-accuracy functionals, GPU acceleration
• Compatible with PBE, SCAN, R2SCAN, RVV10, DFT+U, and vdW corrections
• Modular execution with flags: --only-mc, --jij

---

📦 Installation

1. Clone the repository:

   git clone https://github.com/your-org/MagTrans.git
   cd MagTrans
   

2. Create & activate a virtual environment:

   python3 -m venv .venv
   source .venv/bin/activate
   pip install -r requirements.txt
   

3. Ensure dependencies such as VASP, Custodian, ASE, and pymatgen are installed and configured properly.

---

⚙️ Input File (input)

Place a plain-text file named input in the working directory. Below is an annotated example for BaNiCl₃:

system_dimension         = 2D
structure_file           = BaNiCl3.vasp
XC_functional            = PBE
DFT_supercell_size       = 1 1 1
VASP_command_std         = mpirun -np 2 vasp_std
VASP_command_ncl         = mpirun -np 2 vasp_ncl
accuracy                 = high

# Magnetic enumeration
mag_prec                 = 0.003
enum_prec                = 1e-7
max_neighbors            = 4
mag_from                 = OSZICAR

# GPU and layer flags
GPU_accel                = True
more_than_2_metal_layers = False

# DFT+U and vdW
dftu                     = false
LDAUTYPE                 = 2
LDAUU                    = Ni 6.0
LDAUJ                    = Ni 1.0
LVDW                     = True

# Plane-wave basis
ENCUT                    = 450
NSIM                     = 4
KPAR                     = 2
NPAR                     = 2
NCORE                    = 1

# k-points
kpoints_density_relax    = 10
kpoints_density_static   = 20

# I/O
log_filename             = BaNiCl3.log
potential_directory      = /home/potential

💡 Tip: Any missing tags are automatically filled with defaults at runtime.

---

🛠️ Usage

From the project root:

# Full workflow: spin enumeration → DFT → fit Hamiltonian → Monte Carlo
./MagTrans

# Run only Monte Carlo (requires existing input_MC + DFT outputs)
./magtrans --only-mc

# Generate exchange interaction file (Jij) for Vampire simulations
./magtrans --jij [--exc_type isotropic|tensorial]

---

📖 How It Works

1. parse_input()
   Parses the input file and sets global parameters.

2. Structure Preparation
   Converts structure with ASE → pymatgen; applies vacuum or strain if needed.

3. Spin Enumeration
   Uses KD-tree + symmetry operators to enumerate unique collinear configs.

4. DFT Execution
   Performs relaxation → static → SOC DFT runs, using Custodian for error handling.

5. Hamiltonian Fitting
   Symbolically solves or fits via least-squares a Heisenberg + anisotropy model (up to 4 shells).

6. Monte Carlo Simulation
   Computes temperature-dependent properties using Metropolis / Hybrid / SSE-QMC.

7. Output & Visualization

   • *_Heisenberg_mc.png — Plots of energy, magnetization, $C_v$, and susceptibility
   • heisenberg_mc_data.txt — Full thermodynamic dataset

---

📂 Project Structure

├── file.vasp              # Input structure file (POSCAR)
├── input                     # Input parameter file
├── MagTrans                      # Main executable script
├── run_CurieD.py             # Core workflow module
├── exchange_generator.py     # Jij generator for MC/Vampire
├── mc_class.py               # Monte Carlo implementation
├── requirements.txt          # Python dependencies
├── README.md                 # ← This file
└── examples/                 # Sample cases and output

---

🧑‍💻 Contributing

We welcome contributions!

1. Fork the repo & clone it locally
2. Create a new feature branch
3. Add tests + documentation for your changes
4. Submit a pull request (PR)

Please follow PEP8 and write clear commit messages.

---

📜 License

This project is licensed under the MIT License. See LICENSE for details.

---

✉️ Contact

Chinedu Ekuma
Department of Physics, Lehigh University
📧 [e-mail] | [e-mail]