High-Throughput Computational Materials Discovery
ElectrideFlow: An accelerated materials discovery framework combining diffusion-based generative models with hierarchical thermodynamic and electronic structure screening to identify 290 new electron-rich electride compounds, including 13 thermodynamically stable electrides.
Materials Modelling & Informatics Group (MMI) | UNC Charlotte
Browse electrides by formula, space group, energy above hull, or N_excess.
Interactive 3D structures with JSmol and detailed metadata inspection.
Compare DFT and MatterSim energies, band gaps, and thermodynamic stability.
Toggle between binary and ternary systems, or query by keys
| age | Time since creation | |
| band0 | band0 | |
| band1 | band1 | |
| band_gap | band_gap | |
| calculator | ASE-calculator name | |
| charge | Net charge in unit cell | |e| |
| chemsys | chemsys | |
| composition | composition | |
| density | density | |
| dft_e_hull | dft_e_hull | |
| dof | dof | |
| e0025 | e0025 | |
| e05 | e05 | |
| e10 | e10 | |
| electride | electride | |
| energy | Total energy | eV |
| fmax | Maximum force | eV/Å |
| formula | Chemical formula | |
| full_dft_e_hull | full_dft_e_hull | |
| id | Uniqe row ID | |
| magmom | Magnetic moment | μ_B |
| mass | Sum of atomic masses in unit cell | au |
| mattersim_e_hull | mattersim_e_hull | |
| mattersim_energy_per_atom | mattersim_energy_per_atom | |
| N_excess | N_excess | |
| natoms | Number of atoms | |
| pbc | Periodic boundary conditions | |
| pearson_symbol | pearson_symbol | |
| smax | Maximum stress on unit cell | eV/ų |
| space_group_number | space_group_number | |
| structure_id | structure_id | |
| symmetrized | symmetrized | |
| unique_id | Random (unique) ID | |
| user | Username | |
| vasp_energy_per_atom | vasp_energy_per_atom | |
| volume | Volume of unit cell | ų |
| wps | wps |
Accelerated by Generative Models and Hierarchical Prescreening: A four-stage intelligent workflow for high-throughput inorganic electride discovery.
Generative AI models (MatterGen) explore electron-rich composition space with excess valence electrons (0 < Nexcess ≤ 4) from electropositive alkaline, alkaline-earth, and early transition metals.
MatterSim (MLP) pre-screening filters structures by energy above hull to identify thermodynamically promising candidates (Eref-hull-MLP < 0.10 eV/atom).
Coarse DFT relaxation with ELF and PARCHG topological analysis to identify electride candidates with localized interstitial electrons (Bader volume > 20 A^3).
Tight DFT relaxation with refined convex hull analysis for final candidate validation, identifying thermodynamically stable electrides (Ehull-DFT < 0.05 eV/atom).
251 binary and 39 ternary compounds in this release
Diffusion-based generative models for targeted materials discovery
Final stability validation with VASP calculations and hull screening
Broad exploration of electropositive binary and ternary systems
Accelerated Inorganic Electrides Discovery by Generative Models and Hierarchical Screening
Shuo Tao, Qiang Zhu
arXiv:2601.21077 (2026)