Fe-Co-Zr Magnetic Materials: A DFT and Machine Learning Dataset

This dataset contains the key computational data, structural information, and validation files supporting the findings presented in the manuscript: "Accelerated discovery and design of Fe-Co-Zr magnets with tunable magnetic anisotropy through machine learning and parallel computing." The data were generated using a high-throughput, machine learning-guided discovery framework (exa-AMD) coupled with first-principles calculations. This dataset is provided to facilitate open science, reproducible research, and further exploration of the Fe-Co-Zr ternary system for rare-earth-free permanent magnets. This repository includes: Crystallographic Information Files (CIFs):The 9 newly discovered thermodynamically stable Fe-Co-Zr compounds. The 2 engineered phases, Fe₅Co₁₈Zr₆ and Fe₅Co₁₆Zr₆Mn₄, derived from local atomic substitutions 3. The 81 unique, low-energy metastable Fe-Co-Zr compounds (with formation energies within 0.1 eV/atom above the convex hull). Dynamic Stability Data:Raw data files for the phonon dispersion calculated for the 11 key stable and derived compounds. Raw data files from the ab initio molecular dynamics (AIMD) simulation of FeCo₅Zr₁₂, including energy trajectory and radial distribution function (RDF) data. Model Validation Data: Raw data used to generate the comparison plot of ML-predicted vs. DFT-calculated formation energies for the converged structures.