PCN validation errors: why submissions fail and how to fix them

A failed PCN submission comes back with a validation report, not a rejection letter — a list of rule codes, each pointing at a section of your dossier. The report looks intimidating because one underlying data problem often trips several rules at once, but the rule families are stable and the fixes are mechanical once you know which family you are in. This article maps the families that account for most failures we see, using the actual rule codes from ECHA’s published validation rules list, and ends with a triage workflow that avoids the most common mistake: patching the dossier instead of the data behind it.

How ECHA validation works

Every poison centre notification is checked against two published rule sets: one that runs inside IUCLID when you validate the dossier locally, and one that runs in the ECHA Submission Portal when you upload. Both sets contain two kinds of rules:

• Business rules (BR) — a single BR failure means the submission fails. Nothing reaches the member states until every BR passes.
• Quality rules (QLT)— warnings and reminders about common shortcomings. A QLT finding does not block the submission, but ECHA’s own rules document notes it “can result in further clarification requests from Member state(s) at a later stage.”

Both appear in the same validation report, so read the type prefix before panicking: a report full of QLT lines is a successful submission with homework, while one BR line is a failed submission regardless of how clean everything else is. If you validate in IUCLID before uploading, you catch the first rule set without burning a portal round-trip — the workflow is covered step by step in how to submit a PCN.

Composition totals and concentration ranges

The largest family of blockers lives in the composition section. Three checks catch most first-time submitters:

• Totals. If the declared concentrations sum to below 70%, the dossier is not accepted; between 70% and 90% you get a warning that the full composition appears incomplete; above 105% it is again rejected. Rounding generous ranges up, or listing only hazardous components, are the usual causes.
• Range widths. Annex VIII Part B caps how wide a declared concentration range may be. BR518 enforces the tighter Table 1 limits for components of major concern — the substances poison centres most need precision on — and BR588 enforces the Table 2 limits for everything else. A 1–10% range that is fine for a thickener will fail for ethanol.
• Form of the values. Ranges must carry qualifiers (BR625), only w/w% and v/v% are accepted as units (BR541), components reported at exactly 0% are rejected, and you must give either a typical concentration or a range — not both.

The full field-by-field picture of what a dossier must contain is in PCN required information.

Missing or malformed UFI

Three rules police the UFI: BR528 makes at least one UFI entry mandatory for hazardous submissions, BR549 checks the format against the UFI Developers Manual — including the checksum, so a single mistyped character fails the whole submission — and BR531 requires every product record to have at least one linked UFI. The checksum failures are the cheapest to prevent: run any UFI that has been retyped or pasted between systems through a validator, such as ECHA’s own or our free UFI checker, before it goes anywhere near a dossier or label artwork.

EuPCS category and use type inconsistencies

The product information section must declare at least one use type (BR535) and a main intended use — the EuPCS category (BR534). The rules then cross-check the combination: a limited (industrial-use-only) submission cannot declare professional or consumer use (BR587), and biocidal or plant protection categories cannot appear as secondary uses unless they are also the main intended use (BR589). Beyond the hard rules, a category that plainly contradicts the declared use type invites member-state follow-up even when the dossier passes. Choosing the right category from the ~250 options is its own discipline — see the EuPCS category guide.

MiM references that do not resolve

Mixture-in-mixture components fail in a distinctive way: your data is internally consistent, but the reference points at something ECHA cannot verify. The identification rule (BR527) gives you two routes — provide the supplier’s UFI, or provide a supplier record plus the component substances you know. Each route has its own traps:

• A MiM can carry at most one UFI (BR578), and QLT516 flags a MiM UFI that was never notified by a valid submission in your market placement countries — the supplier generated a UFI but never filed the PCN behind it.
• A MiM UFI notified for industrial use only cannot sit inside your consumer or professional product (BR596); the supplier must upgrade their own notification first.
• If you go the supplier-details route instead, the supplier record must include name, phone and email (BR606), the supplier should be an EU entity, and a hazardous MiM needs at least one substance component listed.

These failures are supply-chain problems wearing a validation-error costume; the negotiation playbook is in mixture-in-mixture (MiM).

Missing toxicological information

A standard submission needs exactly one toxicological information record (BR515), and BR538 requires the free-text toxicological information — section 11 of your SDS, in substance — to run to at least 200 characters in every language the notification declares. The multilingual clause is the one that bites: a dossier covering three markets with three languages needs the text three times. Pasting section 11 verbatim usually satisfies the rule; leaving the field to a translation placeholder does not.

Packaging and label information gaps

Unless the use type is industrial only, every product needs either a linked packaging record or an explicit “product not packaged” selection (BR542), with container type and size filled in (BR536, BR537). Colour and physical state are mandatory (BR529, BR530), and the classification and labelling record must be complete — a classified mixture needs a signal word and hazard statements selected, not just hazard categories. These are rarely conceptual failures; they are fields nobody owned.

Legal entity and role mismatches

The portal checks that the legal entity inside the dossier is the legal entity doing the submitting: the UUID in the “Mixture identity and legal submitter” record must match your portal account (BR570), the countries must match (BR952), and the submitting company must be located in the EU (BR627). Dossiers prepared by a consultant under their own legal entity, or exported from another subsidiary’s IUCLID instance, fail here. Related: QLT571 flags main-mixture UFIs notified by a different legal entity unless there is a valid reason such as a toll-formulation agreement.

A triage workflow that actually converges

1. Split the report by prefix. BRs are the work queue; QLTs are the review queue. Only BRs stand between you and a successful submission.
2. Read the failures bottom-up. Work from the deepest data — substances and MiM components — up through composition to product records. One broken component dataset can fire a dozen rules upstream of it; fixing top-down means fixing symptoms.
3. Fix at source, not in the dossier. A dossier is a snapshot. Hand-editing the failed .i6zclears today’s error and guarantees its return at the next update, because the source records still carry the bad data. Correct the substance, composition and product records, then regenerate — see how the .i6z dossier is built.
4. Revalidate locally, then resubmit. Run the IUCLID validation assistant until the BR list is empty, then upload. Remember BR564: the portal rejects a byte-identical dossier, so a “just try again” resubmission without changes fails by design.
5. Decide on recurring QLT warnings. If the same warning appears on every submission, either the data genuinely falls short — fix it once at source and it disappears everywhere — or your situation is a legitimate exception (a voluntary submission without a UFI, a pH range as wide as your process really runs). In the second case, document the reasoning, because the warning text is exactly what a member state will ask about later.

Most of this article describes failure modes that only exist when dossier fields are filled in by hand. Ufi Number generates the dossier from structured product data and applies these checks before dispatch, which is why the fix-at-source rule comes free: there is nothing to hand-edit.