OECD eChemPortal

Datadory delivers OECD eChemPortal commodity chemicals data covering 683,840 substance records and 1,136,073 endpoint records aggregated from 34 government databases — ECHA REACH dossiers, US EPA hazard resources, Japan's J-CHECK, Canada's Categorization Results — spanning physicochemical properties, environmental fate, ecotox, toxicology and GHS classifications in one record.

What is the OECD eChemPortal?

OECD eChemPortal is the OECD's aggregation of government chemical data — 34 national and regional databases queried in one pass, spanning substance identity, physicochemical properties, environmental fate, ecotoxicology, toxicology and GHS classifications. The contributor roster runs from ECHA's REACH dissemination and classification inventory to the US EPA's hazard resources, Japan's J-CHECK, Canada's Categorization Results, Australia's AICIS assessments and Germany's ETOX. The portal's own published census counts 683,840 substance records, 1,136,073 endpoint records and 805,418 synonyms.

Three query paths cut across all of it: substance lookup by name or CAS Registry Number, property queries against a structured endpoint tree, and classification lookups against the harmonized GHS code catalogue.

Get a sample of this dataset to see real records with their complete field dictionary before you commit.

What do the first rows look like?

Four records exactly as the portal returns them — minus the pointers that lead back into each agency's own dossier system. Two describe the same substance, methanol (CAS 67-56-1): once as a classification entry from ECHA's C&L inventory with no study records attached, once as an endpoint-bearing record from Canada's Categorization Results. The third falls out of a boiling-point property query — formaldehyde, CAS 50-00-0, sourced from ECHA REACH study records. The fourth is an entry from the GHS classification catalogue itself.

SUBSTANCE SEARCH — methanol
| SUBSTANCE | CAS RN  | PARTICIPATING SOURCE | MATCH LEVEL | ENDPOINT DATA |
| --------- | ------- | -------------------- | ----------- | ------------- |
| Methanol  | 67-56-1 | ECHA C&L inventory   | 1           | no            |
| Methanol  | 67-56-1 | CCR                  | -           | yes           |

PROPERTY SEARCH — boiling point
| SUBSTANCE    | CAS RN  | CHAPTER       | ENDPOINT KIND | SOURCE     |
| ------------ | ------- | ------------- | ------------- | ---------- |
| formaldehyde | 50-00-0 | Boiling point | BoilingPoint  | ECHA REACH |

GHS CLASSIFICATION CATALOGUE
| CODE | ABBREVIATION | TEXT               | CATEGORY                                              |
| ---- | ------------ | ------------------ | ----------------------------------------------------- |
| 7007 | Flam. Gas 1A | Flammable gases 1A | Flammable gases (including chemically unstable gases) |

Same identifier, different depth per contributor. That asymmetry is the portal's defining texture — and the reason one normalized copy beats thirty-four separate lookups.

What fields does each record carry?

Ten fields ride along with every substance record, from the contributor's own spelling of name down to substance_id, the portal's internal handle for a record inside a given source. The level field grades how literal the name-to-number binding is — 1 marks a high-confidence pairing, 2 a broader match. An eleventh field carries a pointer back to the contributing agency's own dossier view; because it resolves inside each agency's systems rather than the shared record, it ships as one of the additional fields available on request.

The dictionary below is the complete core set.

How far does the property taxonomy reach?

Property queries walk a structured endpoint tree in four families. Physical and chemical properties run from melting and boiling point through density, granulometry, vapour pressure, water solubility, octanol-water partitioning, flash point, auto flammability and viscosity. Environmental fate and pathways cover phototransformation, hydrolysis, biodegradation, bioaccumulation, adsorption and desorption, and Henry's Law constants. Ecotoxicological endpoints span aquatic, sediment and terrestrial compartments. Toxicological endpoints climb the whole ladder — acute toxicity by route, irritation, sensitisation, repeated dose, genetic toxicity, carcinogenicity, reproductive toxicity.

Depth is uneven by design: four contributors submit searchable endpoint studies — Canada's CCR, ECHA CHEM, Japan's J-CHECK and the OECD SIDS IUCLID collection — and ECHA CHEM alone accounts for 911,173 of the portal's 1,136,073 endpoint records, spread over just 20,906 substances.

What ground does the portal cover?

  • Geography: global aggregation across OECD members and partners — EU/ECHA, US EPA, Canada, Japan, Australia, Germany, France, the Nordic countries, New Zealand and the UK sit side by side in one result set.
  • Temporal: revision activity documented through May 2026, when the portal's news log recorded an ECHA REACH data refresh; per-source counts are published as a census snapshot rather than running totals.
  • Granularity: substance identity records plus individual endpoint study records, one row per substance per contributing database — so a single CAS number fans out across as many sources hold it.

How is the data delivered?

Your call, not ours: API, files, or your warehouse. Daily, weekly, or hourly.

Every delivery ships the complete field dictionary above, the sample rows, and the coverage profile — mapped and ready, with no integration archaeology on your side.

Who uses this data?

  1. Market researchers mapping chemical portfolios and hazard profiles across a dozen jurisdictions without stitching together national websites.
  2. Data scientists joining property and classification attributes onto an internal substance master, keyed on CAS number.
  3. Competitive intel and product teams tracking which substances carry which GHS classifications where — a leading indicator of reformulation pressure.
  4. Investors and quants reading classification churn across jurisdictions as a regulatory-cost signal for chemical producers.
  5. Journalists and academics citing hazard classifications with a provenance trail back to the contributing agency, not an anonymous aggregator.

Where should you go next?

eChemPortal tells you what a substance is and how it is classified; the quantitative layer lives elsewhere. The Toxics Release Inventory reports what US facilities actually emit, the European Industrial Emissions Portal mirrors that for Europe's largest plants, and the TSCA inventory fixes the boundary of the US commercial substance list. Pair the classification picture with emission volumes and you get the full exposure story.

Which datasets pair well with this one?

Notes and adjacent reading:

Browse the whole vertical on the Commodity Chemicals industry hub or the ranked shortlist of the best commodity chemicals datasets.

Field dictionary

Every field below is documented against real records. The full dictionary ships with the sample.

Field dictionary — the ten fields carried by every substance record
FieldTypeDefinitionExample
namestringSubstance name as recorded by the contributing source (IUPAC or common name)Methanol
numberstringChemical identifier value, typically the CAS Registry Number67-56-1
number_typeenumLabel for the identifier typeCAS Number
levelintegerSpecificity of the name match (1 = high confidence, 2 = broader match)1
participant_acronymstringAcronym of the contributing database holding the recordECHA REACH
participant_types_localizedtextInformation categories the source contributes: property information, exposure and use information, GHS classificationsProperty information
endpoint_databooleanWhether the source has submitted searchable endpoint (study result) records for this substancetrue
chapterstringEndpoint study chapter title returned in property-search resultsBoiling point
endpoint_kindstringMachine key of the endpoint type used by the property-query builderBoilingPoint
substance_idintegerPortal identifier for the substance record within a given source5961632

Questions buyers ask

How many substances does OECD eChemPortal cover?

683,840 substance records and 1,136,073 endpoint records, plus 805,418 synonyms, across 34 participating databases. ACToR contributes the most identities at 269,971 substances, while ECHA CHEM dominates depth with 20,906 substances carrying 911,173 endpoint study records between them.

Which contributors provide searchable property data?

Four of the thirty-four: Canada's Categorization Results, ECHA CHEM, Japan's J-CHECK and the OECD SIDS IUCLID collection. The remainder contribute classification results, exposure and use information, or reference material — so identity coverage runs wide while study-record depth concentrates in those four.

What endpoint types can a property query target?

A structured tree of four families: physical and chemical properties (melting point, boiling point, density, vapour pressure, water solubility, flash point and more), environmental fate and pathways (hydrolysis, biodegradation, bioaccumulation, Henry's Law constant), ecotoxicology, and toxicology down to genetic toxicity and carcinogenicity.

Is coverage even across jurisdictions?

No — and the unevenness is informative. More than ten jurisdictions contribute identities and classifications, but endpoint study records cluster heavily in ECHA REACH submissions, with Canada's CCR and Japan's J-CHECK supplying the next deepest pools. Treat per-source depth as a coverage map, not a uniform baseline.

How current are the records?

Revision activity runs through 2026 — the portal's news log documents an ECHA REACH data refresh in May 2026 — while the published record census is a snapshot, so per-source counts should be read as indicative magnitudes rather than live totals. Every delivered extract carries its own extraction date.

Can Datadory scope a pull to my substance list?

Yes. Send a CAS number list or a named substance set and the sample comes back scoped to it — mapped to the field dictionary above and delivered by API, files, or your warehouse on the schedule you pick.

See the rows before you pay anything.

Name this dataset and we send real records from it — scoped to the fields you asked for.

See pricing