Materials

Family, Tc, pressure and result-evidence filters must match the same extracted result. Unknown pressure does not satisfy a pressure limit. The row shows one Tc result and that result’s conditions.

Pairing and classification filters use source reports, not family priors. They apply to the material reported-summary scope, not a joint Tc/state result. A false classification filter requires an explicit negative report with method and detection conditions; missing data does not match false.

Catalogue eligibility and source tier do not establish scientific approval, independent replication or ML readiness. Operational anomaly checks flag records for review; “no findings” is not scientific validation. Retained values are not capped or rewritten.

Sources count identifiers or legacy links, not independent experiments. Source year comes from the displayed Tc result. Optional classification and structure columns may describe other source results or states. The catalogue maximum remains separate when a filter matches another result.

Sorting uses current source-scoped catalogue projections. Missing data, unavailable sources and pending review remain distinct. Open Evidence for provenance and status details.

Source reference pilot: inspect seven CrB₂ paper items and three independent COD structures, with conditions, source links and downloadable metadata.