Project
HyCAN-DB: A FAIR Database of Experimental Hydrogen Uptake in Carbon Sorbents
A systematically screened, FAIR-compliant database of experimental hydrogen-sorption measurements in carbon nanomaterials, built for meta-analysis.
Motivation
Hydrogen storage on carbon has a credibility problem with a history: extraordinary uptake claims around early nanotube work later failed to reproduce, and the surviving literature spans inconsistent units, conditions, and reporting standards. HyCAN-DB is built to establish what the experimental record actually supports, by screening papers against explicit criteria before a single number is extracted, and by storing what survives in a FAIR-compliant, versioned, citable form.
Methods
Candidate papers enter a tracked corpus with DOIs and screening metadata. Inclusion requires passing a four-point relevance test applied to the primary record: an experimental measurement of hydrogen uptake rather than a simulation, on a carbon sorbent (nanotubes, graphene and its oxides, activated, templated, or carbide-derived carbons, nanofibers, doped or functionalized carbons), at stated temperature and pressure, with at least one physical characterization, BET surface area at minimum. Every screening decision and exclusion reason is recorded, so the corpus is reproducible rather than curated by feel. Extraction targets uptake values with their temperature, pressure, and surface-area context, under a shared schema with validators and unit converters.
Current status
The pilot screen is complete: 30 primary studies processed and 28 included under the four-point test, with roughly 25 physisorption records extracted so far. The shared schema, validators, and converters are in active build-out. I treat 30 studies as a pilot corpus, sufficient for protocol validation and early meta-analysis, and deliberately too small to support machine-learning claims yet.
Future work
Full-precision extraction across the included papers, then cross-study statistical synthesis. Versioned releases with a DOI per dataset version, plus registration in research-data indexes such as re3data and FAIRsharing, are the path to the database being cited as infrastructure. HyCAN-DB’s schema and converters are also scheduled deliverables for CHASM’s evidence atlas.