# G4MP2 Thermochemistry and DFT Solvation Energies for 130,258 Small Molecules (Foundry-ML)

## Facts

- Provider: Ward, Dandu, Blaiszik, Narayanan, Assary, Redfern, Foster, Curtiss (Argonne National Laboratory)
- License: CC-BY-4.0 (https://huggingface.co/datasets/foundry-ml/foundry_g4mp2_solvation_v1-2)
- Size: 3 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/foundry-ml/foundry_g4mp2_solvation_v1-2
- Access: public, readable without an account
- DOI: https://doi.org/10.18126/jos5-wj65
- Category: Chemistry & Materials
- Source: https://doi.org/10.18126/jos5-wj65
- Page: https://scigantic.com/archive/42592bd3-9916-5ee6-b67f-38bde199d7b4
- Keywords: Chemistry & Materials, DFT, quantum chemistry, Parquet, Hugging Face

## About

Quantum chemistry for 130,258 small organic molecules from the QM9 chemical space, produced at Argonne National Laboratory for machine learning of solvation and thermochemistry. Each molecule carries B3LYP/6-31G(2df,p) properties (dipole moment, polarizability, HOMO, LUMO, band gap, zero-point energy, internal energies, enthalpy, free energy, rotational constants, atomic charges), high-accuracy G4MP2 composite energies (0 K and 298 K internal energy, enthalpy, free energy, atomization energy, standard enthalpy of formation), and implicit-solvent DFT solvation energies in water, acetone, acetonitrile, dimethyl sulfoxide and ethanol, with SMILES, InChI and relaxed XYZ coordinates. The data are one Parquet table of about 115 MB (measured 2026-11-15), loadable with Hugging Face datasets, pandas or the Foundry-ML client by DOI 10.18126/jos5-wj65 on the Materials Data Facility. Used for solvation free energy prediction, delta-learning from DFT to G4MP2 and molecular property benchmarks. Released under CC-BY-4.0; cite Ward et al., Materials Data Facility (2022), and Ward et al., MRS Communications 2019.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("foundry-ml/foundry_g4mp2_solvation_v1-2")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/train-00000-of-00001.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/42592bd3-9916-5ee6-b67f-38bde199d7b4
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/42592bd3-9916-5ee6-b67f-38bde199d7b4
- Over MCP: `get_schema_card` with id `42592bd3-9916-5ee6-b67f-38bde199d7b4` 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)
- [Quantum chemistry datasets](https://scigantic.com/topics/quantum-chemistry)
