Oxide Properties · foundry-ml
wolverton_oxides_v1-1
DFT-computed thermodynamic and electronic properties of 4,914 ABO3 perovskite oxides for materials machine learning.
What wolverton_oxides_v1-1 is
wolverton_oxides_v1-1 is a tabular materials science dataset compiled by Antoine Emery and Chris Wolverton, published through the Materials Data Facility and Foundry-ML. It contains high-throughput density functional theory (DFT) calculations for 4,914 perovskite oxides.
The dataset addresses the need for curated benchmark data in inorganic chemistry and computational materials discovery. By capturing thermodynamic stability metrics alongside structural and electronic properties, it provides a standardised basis for training and evaluating machine learning models on crystalline oxides.
How it was made
The dataset tabulates 4,914 ABO3 perovskite entries generated via high-throughput density functional theory calculations. Each record combines compositional identity—including the specific A-site and B-site constituent elements—with crystallographic parameters from relaxed structures, such as lattice constants (a, b, c), unit cell angles (alpha, beta, gamma), atomic volume, and lowest distortion structural categories.
Computed electronic and magnetic features include the magnetic moment in Bohr magnetons and the DFT-PBE band gap in electronvolts. Three core thermodynamic targets are provided: DFT-calculated formation energy relative to elemental references (e_form in eV/atom), convex hull energy indicating thermodynamic stability (e_hull in eV/atom), and oxygen vacancy formation energy (e_form oxygen in eV).
Key facts
- Publication year
- 2022
- Dataset size
- 4,914 perovskite oxides
- Format
- Tabular (Parquet)
- Data modalities
- Tabular, text
- Supported tasks
- Tabular regression, tabular classification
- DOI
- 10.18126/h15n-7wu8
What you can use it for
- Perovskite Stability Prediction
- Train classification or regression models to predict convex hull energy (e_hull) and identify thermodynamically stable oxide compositions.
- Formation Energy Estimation
- Develop machine learning models that estimate elemental formation energy (e_form) directly from composition and structural distortion types.
- Oxygen Vacancy Energetics Modelling
- Predict the formation energy of oxygen vacancies across different ABO3 perovskite chemistries for defect engineering studies.
- Materials ML Algorithm Benchmarking
- Evaluate tabular regression and classification architectures on inorganic crystalline oxide features and continuous physical properties.
Explore the data
The schema, real rows, full-text search and a SQL filter — read in your browser, nothing downloaded.
Explore
Read the real rows without downloading anything
Reading rows…
LiveRead from foundry-ml/wolverton_oxides_v1-1 at the moment you asked. Nothing is cached or stored — every row above came from that request.
How the columns are distributed
Over 4,914 rows of default/train
a
float- min
- 3.23
- median
- 4.31
- mean
- 4.61
- max
- 6.89
alpha
float- min
- 49.40
- median
- 90
- mean
- 88.31
- max
- 90
atom a
string_label- B1%
- Tl1%
- Hg1%
atom b
string_label- As1%
- Sr1%
- Be1%
b
float- min
- 3.23
- median
- 4.31
- mean
- 4.69
- max
- 7.37
beta
float- min
- 30
- median
- 90
- mean
- 88.31
- max
- 109
c
float- min
- 3.23
- median
- 4.33
- mean
- 5.40
- max
- 13.16
e_form
float- min
- -3.84
- median
- -1.74
- mean
- -1.66
- max
- 1.88
e_form oxygen
float- min
- -21.79
- median
- -0.71
- mean
- -1.21
- max
- 6.96
e_hull
float- min
- -0.57
- median
- 0.67
- mean
- 0.74
- max
- 3.93
formula
string_text- min
- 4
- median
- 6
- mean
- 5.82
- max
- 6
gamma
float- min
- 49.40
- median
- 90
- mean
- 88.31
- max
- 90
gap pbe
float- min
- 0
- median
- 0
- mean
- 0.36
- max
- 6.22
lowest distortion
string_label- cubic63%
- orthorhombic29%
- rhombohedral6%
mu_b
string_label- 022%
- -19%
- 0.26%
vpa
float- min
- 6.74
- median
- 13.35
- mean
- 13.67
- max
- 25.25
Strengths and limits
Good at
- Supplies multiple DFT-derived thermodynamic targets, including formation energy, convex hull energy, and oxygen vacancy energy.
- Pairs composition with structural relaxation attributes, such as lattice constants, angles, atomic volume, and distortion types.
- Provides aligned electronic and magnetic features, specifically relaxed magnetic moments and PBE band gaps.
Watch out for
- Limited in scope to binary and ternary ABO3 perovskite oxide compositions.
- Dataset size is constrained to 4,914 instances, which may limit data-intensive model training.
- Contains purely calculated DFT properties rather than experimental measurements.
Questions
Can I look inside wolverton_oxides_v1-1 without downloading it?
Yes. This page reads its schema, real rows and per-column statistics in the browser, with full-text search and a SQL filter. Nothing is downloaded to your machine.
What licence is wolverton_oxides_v1-1 under?
Its authors declare cc-by-4.0. Read the terms on the original page before reusing the data.
What target variables are available for prediction?
The dataset contains three primary targets: formation energy relative to elements (e_form in eV/atom), convex hull energy (e_hull in eV/atom), and oxygen vacancy formation energy (e_form oxygen in eV).
What structural features are provided for each oxide?
Features include the A-site and B-site elements, structural distortion category, atomic volume, magnetic moment, and relaxed lattice parameters (a, b, c, alpha, beta, gamma).
What tasks can this dataset support?
It is formatted for tabular regression (such as predicting continuous energy metrics or band gaps) and tabular classification (such as identifying stability thresholds).
How to cite it
Emery, Antoine and Wolverton, Chris. High-throughput DFT calculations of formation energy, stability and oxygen vacancy formation energy of ABO3 perovskites. Materials Data Facility, 2022. https://doi.org/10.18126/h15n-7wu8
Written from wolverton_oxides_v1-1's own dataset card. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.