A curated experimental and physics-constrained benchmark dataset for doped zirconia ionic conductivity

This dataset contains a three-layer data resource for doped zirconia ionic conductivity: the source Excel workbook sample_data.xlsx, 1,351 curated experimental reference records, a normalised composition-processing-property relational schema, and a frozen physics-constrained synthetic benchmark with 1,025,619 raw synthetic material-sample records and 1,025,612 compliant synthetic material-sample records. The files are organised as linked tables for material samples, dopants, sintering steps, crystal phases and dictionaries, with source-level provenance, validation outputs, downstream machine-learning validation results, figure source data, a manifest and SHA-256 checksums. The synthetic records are intended as a fixed benchmark for model testing and synthetic-data validation, not as replacements for experimental measurements. Version 1.0.7 adds reproducible Materials Project re-query metadata, raw Materials Project entry records, a reproducible phase-stability table and revised source data for the main-text phase-agreement figure. Version 1.0.8 updates the dataset README citation to match the current published data record. Version 1.0.9 uses the dataset concept DOI in the package README while preserving the version-specific DOI in the Zenodo record metadata. Version 1.0.10 refreshes the manuscript figure files, manifest and source-data package after manuscript streamlining. Version 1.0.11 refreshes the workflow figure, manuscript source and manifest after final manuscript compression.