Auditing a Polycarbonate Supply Chain for BPA Statements

Polycarbonate (PC) is conventionally produced from bisphenol A (BPA; CAS 80-05-7) and either carbonyl chloride via interfacial polymerization or diphenyl carbonate via melt transesterification. The polymer repeat unit contains the bisphenol A bis(carbonate) backbone, and trace unreacted BPA may persist depending on reactor stoichiometry, washing efficiency, devolatilization vacuum, and end-capping. An audit of a polycarbonate supply chain for BPA statements therefore separates three distinct evidentiary layers: polymerization input chemistry, residual unreacted monomer in the resin, and migration of BPA from finished articles. Without that separation, supplier declarations conflate unrelated regulatory and analytical properties.

Interfacial polymerization is typically run as a two-phase reaction in aqueous sodium hydroxide and a chlorinated solvent such as methylene chloride. The organic phase carries the growing oligomer; the aqueous phase controls salt formation and pH. Chain terminators such as p-tert-butylphenol regulate molecular weight. Residual BPA removal is influenced by caustic concentration, phase separation efficiency, and post-reaction acid washing. Melt transesterification plants remove phenol at temperatures between 280°C and 310°C under vacuum below 1 mbar. These process differences alter the evidence needed in a BPA statement because the route determines which purification records and byproduct profiles should be available.

Auditors should treat a statement of residual BPA separately from a statement of migration. A resin can show low residual monomer by HPLC-DAD yet still produce a positive migration result if the article has been processed with excessive residence time or contaminated with BPA-bearing additives. Conversely, a resin containing measurable free BPA may pass a migration limit because the polymer matrix restricts diffusion. Batch-to-batch dispersion observed on extrusion lines frequently reflects dryer efficiency, regrind ratio, and melt temperature rather than base resin alone.

The term “BPA-free” has no standard test definition in the European or United States food-contact framework. An unqualified BPA-free statement is therefore not an acceptable audit endpoint. The auditor must request the exact analytical method, sample preparation, detection limit, and decision rule from the issuing laboratory. If a supplier provides a BPA statement without a test report, the audit should verify whether the statement is a certificate of analysis, a declaration of compliance, or a private specification; these are not interchangeable.

Residual Monomer and Migration Signals in Finished Polycarbonate

Residual free BPA in PC granulate is commonly determined by dissolution in dichloromethane or tetrahydrofuran, followed by precipitation with methanol, centrifugation, and analysis by HPLC with fluorescence or diode-array detection. Migration laboratories use food simulants allocated under Commission Regulation (EU) No 10/2011 Annex III and Annex V. Aqueous simulants are processed directly, while fatty simulants may require solvent extraction. EN 13130-13 is the CEN reference for the determination of BPA in food simulants in many compliance dossiers.

For resin-level audits, LC-MS/MS offers a lower reporting limit than conventional HPLC-DAD and is less affected by co-eluting phenolic additives such as 2,4-di-tert-butylphenol. However, an LC-MS/MS statement must include the internal standard, matrix spike recovery, and quantifier/qualifier transition ratio. A supplier certificate that reports only “< 1 mg/kg total BPA” is incomplete without specifying whether the value is free BPA, conjugated BPA, or total hydrolysed BPA after alkaline digestion.

Analytical verification of a BPA statement begins with the testing laboratory’s scope of accreditation, not with the supplier summary page. A laboratory accredited to ISO/IEC 17025:2017 for EN 13130-13 should provide a test report containing the method revision, analytical balance traceability, calibrated pipette and LC system qualification, and raw data upon request. If the laboratory was not accredited for the exact method, the audit should treat the result as indicative only. For US-bound articles, a supplier may cite FDA 21 CFR 177.1580, but this is a resin listing and not a migration test result.

The auditor confirms that the BPA statement is lot-specific, not grade-specific. A common failure in supply chain audits is a declaration of conformity assigned to a commercial grade family while production lots from different reactors or monomer suppliers are merged. In such cases the statement may be valid for qualification data but not for the lot shipped.

Can a polycarbonate lot be traced through polymerization without breaking the evidence chain?

Traceability requires lot mapping from bisphenol A intake silos, diphenyl carbonate or carbonyl chloride reaction batch, washing/drying line, compounding extruder, and final packaging silo. For resin made by continuous melt transesterification, reactors are usually run as campaign operations; a single lot may contain output from multiple reactor hours. If the supplier defines a lot as a storage silo after compounding, the audit must confirm that all upstream feeds to that silo have the same BPA declaration.

The critical evidence chain is broken when a supplier recertifies inventory without reanalysis under the original method. In that case, the auditor should request the split-lot log, internal release limits, and any customer-quality notification. If a distributor repackages PC pellets, the original lot number and certificate of analysis should remain linked; reissued certificates that omit the original resin producer lot are not equivalent.

Processing history complicates the audit because polycarbonate hydrolyses when wet and thermally degrades at high melt temperatures. Polycarbonate conversion is normally carried out after drying to a moisture content below 0.02% by weight in desiccant dryers with a dew point below −40°C. Published processing guides specify melt temperatures from 280°C to 320°C, with injection molding barrel residence times preferably below 8 minutes and no more than 20 minutes. Above these limits, hydrolysis and thermal chain scission can release BPA and phenol-related compounds at the surface of moulded articles.

Extrusion compounding of PC with additives in co-rotating twin-screw extruders with L/D ratios of 40:1 to 44:1 introduces additional heat history. Regrind ratio is a relevant audit parameter because repeated heating increases carboxyl end groups, hydrolytic sensitivity, and potential release after contact with aqueous food simulants. If a BPA statement covers only virgin granulate, the auditor cannot automatically extend it to moulded articles containing regrind without additional data.

Additive concentrates can carry their own BPA impurities. Anti-drip agents, phosphite stabilizers, and color masterbatches are not necessarily BPA-neutral. A complete audit requests BPA statements for each masterbatch if the final article will be used in a regulated food-contact or medical-device application.

When the end article enters food-contact or pediatric markets

Under Commission Regulation (EU) No 10/2011, the specific migration limit for BPA from plastic materials and articles is 0.05 mg/kg food simulant. Polycarbonate infant feeding bottles are prohibited under Commission Regulation (EU) No 321/2011. The audit must therefore distinguish between a general declaration for PC food-contact articles and a prohibited pediatric article. A supplier statement that simply cites EU food-contact compliance is insufficient if the article shape or intended age group falls within the prohibition.

For medical uses, ISO 10993-18 chemical characterisation can be used to assess BPA as a leachable under clinical conditions. For US articles, FDA 21 CFR 177.1580 remains the polycarbonate resin listing, but specific migration statements are evaluated against end-use conditions in the clearance and the applicable food-contact notifications. The EFSA 2023 BPA hazard reassessment reduced the tolerable daily intake to 0.2 ng/kg body weight per day; this toxicological trigger does not directly set a polymer SML but changes the audit significance of trace BPA in vulnerable applications.

Supplier documentation controls cannot replace direct monomer verification

Supplier questionnaires, ISO 9001:2015 certificates, and statements of compliance are administrative controls. They confirm that a management system exists but do not verify the residual BPA concentration of a specific lot. The audit should review internal specifications against raw material certificates, BPA supplier quality reports, and periodic third-party testing. A supplier questionnaire answer stating “BPA content complies” has no analytical evidentiary weight unless it references a defined method and a defined reporting limit.

A robust audit opens the laboratory data, not just the summary page. It examines sample preparation logs, calibration curve date, reporting limit, instrument injection sequence, and the chain of custody from sample container to final certificate. If the supplier cannot provide those records, the BPA statement is classified as unverified and should not be used for regulatory clearance.

Documentation traceability under ISO 9001:2015 Clause 8.5.2 and ISO 13485:2016 Clause 7.5.9 requires that records remain legible and retrievable for defined retention periods. For polycarbonate resins, the minimum audit record set includes the BPA monomer lot certificate, reaction batch log, devolatilization pressure and temperature trend, resin residual monomer certificate, extrusion line product log, and final packaging lot number. Any discrepancy between the certificate of analysis and the shipping label invalidates the BPA statement.

If a buyer converts PC into articles and generates a new BPA statement for the article, the audit trail should include the purchase specification, incoming resin certificate, processing equipment log, regrind content, and any article-level migration test. Without this data sequence, the statement has no reliable boundary.