# GST-GAP-22: Ge-Sb-Te Phase-Change Memory Material DFT Training Set, ColabFit Conversion

## Facts

- Provider: Yuxing Zhou, Wei Zhang, Evan Ma, Volker L. Deringer; ColabFit conversion
- License: CC-BY-4.0 (https://zenodo.org/records/8208202)
- Size: 4 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/colabfit/GST_GAP_22_main
- Access: public, readable without an account
- DOI: https://doi.org/10.60732/f2d6e02c
- Category: Chemistry & Materials
- Source: https://doi.org/10.1038/s41928-023-01030-x
- Page: https://scigantic.com/archive/ac6813ab-1eb7-549a-87a9-e43c5fb4aea3
- Keywords: Chemistry & Materials, DFT, Parquet, Hugging Face

## About

The main training set of GST-GAP-22, the Gaussian approximation potential (GAP) that Zhou, Zhang, Ma and Deringer built for device-scale atomistic modeling of phase-change memory materials (Nature Electronics 2023). Configurations lie on the quasi-binary GeTe-Sb2Te3 (GST) line of germanium-antimony-tellurium compositions and were used to simulate a range of GST compositions under realistic device conditions. Labels are PBEsol DFT energies, atomic forces and Cauchy stresses. The ColabFit card counts 2,690 configurations and about 341,000 atoms; the repository is about 30 MB (measured 2026-11-19). The extended and refitted sets are separate ColabFit repositories. This is the ColabFit Exchange conversion: configurations are Parquet files under co/ (structure, positions, cell, computed properties and metadata per row) with dataset-level information in ds.parquet; read them with pyarrow, pandas, polars or Hugging Face datasets, following the ColabFit configuration schema at materials.colabfit.org/docs. The original Zenodo record is CC-BY-4.0; cite Zhou et al., Nature Electronics 2023.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("colabfit/GST_GAP_22_main")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `co/co_0.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/ac6813ab-1eb7-549a-87a9-e43c5fb4aea3
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/ac6813ab-1eb7-549a-87a9-e43c5fb4aea3
- Over MCP: `get_schema_card` with id `ac6813ab-1eb7-549a-87a9-e43c5fb4aea3` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Chemistry & Materials datasets](https://scigantic.com/topics/chemistry-and-materials)
- [DFT datasets](https://scigantic.com/topics/dft)
