Epoxy resin coatings based on bisphenol A diglycidyl ether (DGEBA) remain a primary internal barrier lining for steel and aluminium food-contact packaging because they combine adhesion, flexibility, and resistance to aggressive retort conditions. The current regulatory assessment for these products is organised around EU Food-Contact Epoxy Coatings: BPA Migration Limits & Compliance Path, since residual free bisphenol A migrates from incompletely polymerised films into aqueous and fatty food simulants. Commission Regulation (EU) 2018/213 fixed the BPA specific migration limit at 0.05 mg/kg in food and food simulants, while Regulation (EU) No 10/2011 Annex I and Annex III govern overall migration and simulant selection. The technical problem is not solely the resin chemistry but the processing history: undercured coatings retain low-molecular-weight fractions that are analytically detectable at concentrations close to the legal limit.

What Limits BPA Migration from Epoxy Can Linings: Free Residue, Crosslink Density, or Simulant Polarity?

Migration is controlled primarily by the concentration of unreacted BPA in the dried film and by the diffusion coefficient through the crosslinked epoxy network. Liquid DGEBA grades with epoxide equivalent weight of 170–190 g/eq contain residual free BPA as a consequence of incomplete bisphenol A epichlorohydrin reaction. High-purity resin lots typically specify free BPA below 500 mg/kg before compounding, but batch-to-batch variation between 80 mg/kg and 400 mg/kg appears in supplier certificates of analysis. Crosslink density after thermal cure with phenolic hardeners at 200–230 °C determines diffusional resistance. Overbaking above 240 °C can degrade the cured matrix by oxidative chain scission and raise apparent migration, whereas underbaking below 190 °C leaves unreacted oxirane functions and mobile low-molecular-weight molecules. Published data for a direct correlation between film glass transition temperature and BPA migration in production coatings remains limited because the migration test result also depends on partition coefficients with ethanol-based simulants and on the degree of network plasticisation by absorbed simulant.

On high-speed coil lines, internal spray application of solvent-borne epoxy-phenolic coatings at 120–180 m/min uses airless nozzles at 120–180 bar; dry film weight is controlled to 8–12 g/m². The substrate enters a three-zone oven with peak metal temperature of 205–225 °C for 4–10 min depending on line speed. Process stops longer than 30 s in the heated zone are a recognised failure mode: partially cured coils show higher BPA extraction because the low-molecular-weight fraction remains solvent soluble and migrates into 3% acetic acid after retort. Manufacturers therefore quarantine any splice section with peak metal temperature below the specified lower cure boundary and test migration before release. On aluminium drawn containers, similar criticality occurs at the flange area where film thickness can fall below 5 g/m² during deep drawing, producing localised migration hotspots.

Simulant Assignment and Routine Migration Test Protocols Under EN 1186 and CEN/TS 13130-13

Compliance testing for epoxy resin coatings follows a tiered protocol. The overall migration limit of 10 mg/dm² is measured gravimetrically according to EN 1186-1:2002 and EN 1186-3:2002 using food simulants assigned by Regulation (EU) No 10/2011 Annex III. Specific migration of BPA is measured after food simulant exposure using high-performance liquid chromatography with fluorescence detection according to CEN/TS 13130-13:2005. The method reporting limit should be ≤0.01 mg/kg to allow reliable judgement against the 0.05 mg/kg SML. Standard exposure conditions for long-term ambient storage are 40 °C for 10 days for aqueous simulants, while hot-fill and retort applications may require 70 °C for 2 hours or 121 °C for 1 hour in pressurised cells. Simulant selection depends on food type, contact area, and the intended thermal processing.

For two-piece beverage cans filled with carbonated soft drinks, simulant B is often used because acidic fillings have historically shown the greatest extraction of BPA from epoxy linings. For three-piece food cans containing oil-packed fish, simulant D2 or its authorised substitute is required. When a coating is intended for infant food contact, additional legal prohibitions may apply beyond the numerical migration limit, and the formulation must be verified against the specific restriction laid down in Regulation (EU) 2018/213.

When BPA Concentration Approaches 0.05 mg/kg, Batch-Specific Confirmation Becomes Mandatory

At BPA concentrations between 0.04 mg/kg and 0.05 mg/kg, the compliance path shifts from routine certificate of analysis to batch-specific migration validation because HPLC method reproducibility at the SML can have expanded measurement uncertainty of ±20% to ±35% depending on simulant matrix. In this range, laboratories use duplicate migration cells, method blanks, and fortified simulant recovery at 0.025 mg/kg and 0.05 mg/kg. The reference method CEN/TS 13130-13:2005 requires recovery between 80% and 110%, but real food simulant emulsions may reduce recovery below that range. Under EU law, non-compliance is judged against the migration limit without subtraction of measurement uncertainty, so manufacturers typically set release limits at 0.035–0.045 mg/kg to maintain an analytical margin. Batch release then requires at least three independent migration cells from the start, middle, and end of a coil, because tank-side viscosity and oven temperature profile drift along a 5,000 m coil.

The documented compliance path therefore requires raw-material declarations in which free BPA content in the epoxy resin is below 500 mg/kg, process temperature records from each coil, and migration test reports from EN 13130-1:2004 and CEN/TS 13130-13:2005. Supporting documentation should include batch numbers, line speed, oven set points, measured peak metal temperature, film weight, and the simulant exposure condition used. For repeated production of the same recipe on a validated line, reduced testing frequency may be justified under the conformity presumption provisions of Regulation (EU) No 10/2011, but any change in resin lot, hardener stoichiometry, or cure oven configuration invalidates the prior migration data.

Increasing Crosslink Density Through Phenolic Hardeners Without Adding Free BPA

Blending high-molecular-weight DGEBA with resole phenolic or amino-phenolic hardeners at stoichiometric ratios of 1:0.8 to 1:1.2 oxirane-to-phenolic hydroxyl equivalents increases crosslink density and reduces BPA diffusion. The reaction is catalysed with phosphoric acid or blocked sulphonic acid at 0.1–0.5 wt% solids, and cure windows are narrow. Differential scanning calorimetry data from production batches commonly show peak cure exotherm between 180 °C and 210 °C, with vitrification below 150 °C limiting conversion. Adding free amine-based catalysts or amine-functional silane adhesion promoters to acid-catalysed phenolic systems can cause premature gelation in the wet state and should be avoided. Pre-drying of formulated coating at relative humidity above 60% is required because absorbed water slows oxirane conversion and increases residual BPA after baking.

Replacing liquid DGEBA with solid high-molecular-weight epoxy resins of EEW 700–950 g/eq reduces free BPA content by approximately one order of magnitude but raises melt viscosity and requires solution or hot-melt application. In powder coating manufacture, these solid resins are melt-compounded with phenolic hardeners in twin-screw extruders at L/D 40–48, screw speed 300–600 rpm, and barrel temperature 90–110 °C; output is 10–35 kg/h depending on screw diameter. The cured film shows lower migration because the longer resin chains reduce the number of chain ends and the free volume available for diffusional transport. However, films formulated with high-molecular-weight epoxy resins exhibit lower crosslink density and may require higher cure temperatures or supplemental catalyst to maintain retort resistance in acidic food simulants.

The substitution of BPA-based epoxy resins with polyester or acrylic alternatives is not a direct compliance shortcut. BPA-free coatings may still require confirmation of overall migration, sensory inertness, and mechanical integrity after retort at 121 °C for 60 min. In high-acid tomato cans, published data for long-term organoleptic stability of some BPA-free alternatives remains limited, and the coating must be qualified using the same simulant and test method framework applied to epoxy-phenolic systems. The compliance decision therefore rests on migration data generated under EN 1186 and CEN/TS 13130-13:2005, not on a marketing classification of the resin chemistry.