Can Polat pronounced “John”
2026 · Zenodo (CERN European Organization for Nuclear Research)

RADII: Radius-Resolved Benchmark of Nanoparticle Structures for Generative Models in Materials Science

Can Polat , Erchin Serpedin , Mustafa Kurban , Hasan Kurban

Resource highlights

RADII seed structures for scale-generalization experiments

Regenerate a radius-resolved nanoparticle benchmark from documented crystal inputs.

Why we investigated it

Reproducible size extrapolation requires a consistent carving procedure and train-test separation by radius.

What the work contributes

The archive provides ten primitive cells and base nanoparticle structures; the generation pipeline constructs the rotation-augmented benchmark.

What the evidence shows

The associated benchmark spans approximately 75,000 structures and 33–11,298 atoms. The archive contains seed inputs rather than the full augmented collection.

Build on this work

Download the seed archive and follow the linked generation pipeline to construct the benchmark locally.

  • Nanoparticle dataset
  • Radius-controlled generation
  • Distribution shift
  • RADII
Abstract
RADII is a radius-resolved benchmark of ~75,000 nanoparticle structures (33–11,298 atoms across ten materials) designed to measure the extrapolation frontier of graph generative models for materials science. Each structure is a spherical truncation of a published crystal lattice; the benchmark treats radius as a continuous scaling knob and provides leakage-free in-distribution and out-of-distribution splits. This archive contains the seed inputs for the benchmark: the ten primitive unit cell CIFs and per-(material, radius) base XYZ structures. The full ~12 GB benchmark (with rotation augmentation) is reproduced locally by running the generation pipeline shipped in the GitHub repository. Paper: "How Far Can You Grow? Characterizing the Extrapolation Frontier of Graph Generative Models for Materials Science" (KDD '26). Code & generation pipeline: github.com/KurbanIntelligenceLab/RADII (MIT license). To reproduce the full benchmark from this archive: python -m create_radii.create_radii --raw-data radii_raw.zip --output radii. Materials included: Ag, Au, CH₃NH₃PbI₃, Fe₂O₃, MoS₂, PbS, SnO₂, SrTiO₃, TiO₂, ZnO.
Cite
@article{polat2026radii,
  title = {RADII: Radius-Resolved Benchmark of Nanoparticle Structures for Generative Models in Materials Science},
  author = {Can Polat and Erchin Serpedin and Mustafa Kurban and Hasan Kurban},
  year = 2026,
  journal = {Zenodo (CERN European Organization for Nuclear Research)},
  doi = {10.5281/zenodo.20431020},
}