Testing Packaging for PFAS and Heavy Metals: A Practical Programme

A defensible testing programme is tiered. Tier one, screening: total fluorine (by combustion ion chromatography or equivalent) across food-contact materials to find the PFAS question; XRF for the four metals across components. Tier two, confirmation: targeted PFAS analysis (LC-MS/GC-MS panels) where screening or known use requires identification and sums; laboratory metals determination where XRF is positive or borderline. Tier three, assurance: periodic re-testing keyed to supplier and formulation change, plus surveillance sampling of incoming goods for high-risk categories.

The programme's economics depend on the tiering. Screening everything and confirming selectively costs a fraction of panel-testing everything - and produces better evidence, because it covers the whole range at the level each line's risk justifies.

Total fluorine measures all fluorine in the sample, from any source. Its logic in this regime is the regulation's own: above 50 ppm in a food-contact material, the burden moves to demonstrating the fluorine is not PFAS - because in packaging, fluorine at that level without PFAS is uncommon (fluoropolymer processing aids and PFAS grease-proofing being the usual sources, both inside the family). A result comfortably below 50 ppm is strong negative evidence, filed with its method and detection limits. Results near the threshold deserve the confirmation tier, because screening methods carry uncertainty bands that straddle lines drawn this fine - and because the legal question is the PFAS content, which targeted analysis plus, for polymers, total oxidisable precursor and polymer-specific methods can characterise.

One caution against a known error: total organic fluorine and total fluorine are different measurements with different numbers. Reports must name the method exactly as the laboratory reports it; files that paraphrase methods create findings from thin air.

Test results describe the sample, and the sample must represent the line: current production, the actual substrate-and-coating combination as placed, components separated where their materials differ. The classic programme failures are sampling failures - testing the base board but not the coated article, testing last year's specification, testing a supplier's showroom sample rather than delivered goods. For importers, sampling from delivered consignments on a risk rotation converts testing from a supplier's promise into the importer's own evidence, which is what the verification duty in the roles station actually asks for.

Frequency follows change: a stable line from a stable supplier earns longer intervals; any reformulation, site change or new decoration reopens the question. Record the sampling rationale in the programme document - an auditor reads the rationale as the programme's intelligence.

Each result enters the technical file with: the sample's identity and specification version, the laboratory and its accreditation for the method, the method named precisely, the result with units and uncertainty where reported, and the conclusion drawn against the threshold (not merely the number). Results ageing beyond their specification's currency are flagged by the change-control gate, not discovered there. Retained to the regulation's periods, this record set is the difference between a testing programme and a drawer of certificates: the first answers questions, the second merely proves money was spent.

Brief the laboratory like a witness, because that is what its report becomes. The brief names: the legal question (which threshold, which article), the sample's specification identity, the methods expected, and the reporting form - results per analyte with detection and quantification limits, not summary verdicts. Accreditation for the specific methods matters more than institutional reputation. Keep the raw reports, not transcriptions; conclusions drawn from them belong in your file in your words, so the distinction between the laboratory's facts and your judgements survives staff changes. And re-brief on method developments: PFAS analytics is a fast-moving field, and a programme frozen on a 2026 panel will quietly narrow as the family's measurement widens.