Battery Materials · foundry-ml
dataset_li_conductivity
Experimental lithium-ion solid electrolyte conductivity data for 372 materials to support machine learning discovery.
What dataset_li_conductivity is
This dataset compiles experimentally measured solid-state lithium electrolyte conductivities across 372 materials. Published by Cameron J. Hargreaves, Matthew S. Dyer, Matthew J. Rosseinsky, and collaborators through the Materials Data Facility and Foundry-ML, it brings together physical measurements and crystallographic groupings for computational analysis.
Developing high-performance solid electrolytes is a key requirement for next-generation solid-state lithium batteries. By providing standardized experimental conductivity measurements paired with compositional and structural descriptors, this collection enables the training and evaluation of machine learning models for materials screening and electrolyte discovery.
How it was made
The dataset is organised in a tabular format where each entry represents an individual solid electrolyte material. For every entry, it captures material composition, reduced chemical formulae, structural family, chemical family, structure group encodings, and the temperature at which conductivity was measured in degrees Celsius.
The primary prediction targets comprise experimental lithium-ion conductivity on a logarithmic scale (measured in Siemens per centimetre, S/cm) for regression tasks, alongside a categorical indicator separating high and low conductivity materials for classification tasks. Cluster identifiers and original data citations are also provided.
Key facts
- Dataset size
- 372 materials
- Data format
- Tabular (Parquet)
- Supported tasks
- Tabular regression, tabular classification
- Target properties
- Li conductivity (log scale, S/cm) and high/low conductivity classification
- Temperature unit
- Degrees Celsius (°C)
- Publication year
- 2023
- DOI
- 10.18126/k9tn-3965
- Publisher
- Materials Data Facility
What you can use it for
- Continuous conductivity regression
- Train regression models to predict log-scale lithium ionic conductivity from chemical composition, structural groups, and measurement temperature.
- Conductivity regime classification
- Develop classification algorithms to rapidly distinguish candidate solid electrolytes exhibiting high conductivity from those with low conductivity.
- Structure-property relationship mapping
- Examine the correlation between structural or chemical families and ionic transport properties in solid-state lithium conductors.
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/dataset_li_conductivity at the moment you asked. Nothing is cached or stored — every row above came from that request.
Strengths and limits
Good at
- Compiles experimentally measured conductivity values rather than theoretical or simulated approximations.
- Supplies both continuous log-scale values and discrete binary labels to support multiple machine learning problem formulations.
- Includes enriched structural and chemical family groupings alongside standard composition formulae.
Watch out for
- The dataset contains 372 materials, presenting a modest sample size for deep learning architectures.
- All entries are supplied under a single training partition, requiring practitioners to generate their own validation and test splits.
Questions
Can I look inside dataset_li_conductivity 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 dataset_li_conductivity under?
Its authors declare other. Read the terms on the original page before reusing the data.
What targets are available for supervised learning?
The dataset contains log-transformed lithium conductivity in S/cm for regression tasks and a classification target distinguishing low from high conductivity.
What structural and chemical descriptors are included?
Each material entry includes composition strings, reduced formulae, structural group encodings, chemical family classifications, and structural family classifications.
In what units is temperature provided?
Measurement temperatures are recorded in degrees Celsius.
How to cite it
Hargreaves, Cameron J., Gaultois, Michael W., Daniels, Luke M., Watts, Emma J., Kurlin, Vitaliy A., Moran, Michael, Dang, Yun, Morris, Rhun, Morscher, Alexandra, Thompson, Kate, Wright, Matthew A., Prasad, Beluvalli-Eshwarappa, Blanc, Frédéric, Collins, Chris M., Crawford, Catriona A., Duff, Benjamin B., Evans, Jae, Gamon, Jacinthe, Han, Guopeng, Leube, Bernhard T., Niu, Hongjun, Perez, Arnaud J., Robinson, Aris, Rogan, Oliver, Sharp, Paul M., Shoko, Elvis, Sonni, Manel, Thomas, William J., Vasylenko, Andrij, Wang, Lu, Rosseinsky, Matthew J., Dyer, Matthew S. (2023). A database of experimentally measured lithium solid electrolyte conductivities evaluated with machine learning. Materials Data Facility. https://doi.org/10.18126/k9tn-3965
Written from dataset_li_conductivity'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.