Onghena et al. 2016 — Chemical migration from non-polycarbonate baby bottles under EU repetitive-use vs real-life duration testing

This Food Additives & Contaminants: Part A paper compares the EU Regulation 10/2011 standardised plastics-FCM migration test (three sequential 2 h fillings at 70 °C with H₂O-EtOH 50:50 v/v milk simulant) against four real-life “duration” treatments (microwave heating, dishwasher washing, steam sterilisation, cooking sterilisation) on six post-BPA-ban alternative-polymer baby bottles (PP brand A, PP brand B, PES, PA, Tritan, silicone). Migration of pre-selected priority organic compounds — additives, monomers, oligomers, degradation products — was quantified by validated GC-QqQ-MS and LC-QqQ-MS methods (Onghena et al. 2016, Food Anal Methods, in press). Under EU repetitive-use conditions, no authorised compound exceeded its specific migration limit (SML), but several non-listed compounds were detected above the EU 10 µg/kg generic-default threshold for non-authorised substances. Under real-life duration testing, repeated use did not increase migration; concentrations were generally lower than or comparable to the reference experiment and became negligible after a number of cycles. Steam sterilisation released fewer compounds at lower concentrations than cooking (boiling-water) sterilisation, particularly for silicone bottles, and the authors recommend steam sterilisation as a pre-use treatment to reduce residual migrants. The paper measures no heavy metals and contributes no occurrence data to the HMI corpus. It is ingested as a methodology and exposure-pathway reference for baby-bottle migration testing — the comparison of EU repetitive-use vs real-life “duration” testing, and the demonstration that EU conditions can over-estimate real-life migration by up to a factor of 10 for some analytes, are reusable framings if/when heavy-metals migration from baby-bottle polymers (or from bottle additives, colorants, or printed graphics) is investigated in future work.

Key numbers

• Six baby bottles tested, one each of: PP brand A, PP brand B, PES, PA, Tritan™ copolyester, silicone (Methods §“Market survey and sampling”, p. 894).
• EU repetitive-use experiment: three consecutive 2 h migrations at 70 °C with 50:50 v/v H₂O-EtOH (EU Reg. 10/2011 milk simulant); analysed third migration step (Methods §“Migration testing: EU repetitive-use conditions”, p. 895).
• Duration tests, per bottle: 5 reference fillings (10 ml preheated simulant, 40 °C, 30 min) used as zero-treatment control; 100 microwave heating cycles (Whirlpool Gusto GT288WH, 500 W to 40 °C; cycles 1-10, 15, 20, 25, 30, 40, 50, 60, 70, 80, 90, 100 analysed); 10 dishwasher cycles (Eco-mode 2 h 55 min at 55-60 °C; cycles 1, 2, 4, 6, 8, 10 analysed); 10 steam sterilisation cycles (Philips Avent 3-in-1 electric steam steriliser, ~10 min, 100 ml tap water; cycles 1, 2, 4, 6, 8, 10 analysed); 10 cooking sterilisation cycles (boiling tap water, 10 min; same cycles analysed). After each treatment, bottles were rinsed with 50 ml Milli-Q water and refilled with 50 ml fresh 40 °C simulant for 30 min at room temperature prior to LLE / direct LC injection (Methods §“Migration testing: duration tests”, p. 895).
• EU 10/2011 thresholds applied as the regulatory reference: authorised-substance SMLs as listed (e.g., TXIB SML 5000 µg/kg; DEHP / bis(2-ethylhexyl)phthalate SML 1500 µg/kg; dibutyl phthalate SML 300 µg/kg; benzophenone SML 600 µg/kg; bisphenol-A SML 600 µg/kg; bisphenol-S SML 50 µg/kg) plus a generic SML of 60 mg/kg for substances without specific limits (Article 11(2), EU 10/2011); for non-listed substances behind a functional barrier, migration must be < 10 µg substance per kg food (Barlow 2009) (Results §“EU repetitive-use experiment”, p. 896).
• EU repetitive-use third-migration results (Table 1, p. 897; µg/kg simulant; ”—” = experimental failure):
  • PP brand A — 2,4-di-tert-butylphenol 12 (no SML); 2-butoxyethyl acetate 15 (no SML); 3,4-dimethylbenzaldehyde 59 (no SML); all other targets < LOQ.
  • PP brand B — 3,4-dimethylbenzaldehyde 11; all other targets < LOQ.
  • PES — acetophenone 3 (no SML); all other targets < LOQ.
  • PA — azacyclotridecan-2-one (PA monomer) 924 (no SML); all other targets < LOQ.
  • Tritan — 2,4-di-tert-butylphenol 8; 4-propylbenzaldehyde 27; dicyclopentyl(dimethoxy)silane 10 (no SML for any); all other targets < LOQ.
  • Silicone — TXIB / 2,2,4-trimethyl-1,3-pentanediol di-iso-butyrate 348 (SML 5000; well below); 2,6-di-tert-butylbenzoquinone 8; benzophenone 9 (SML 600); diisobutyl phthalate 15; dibutyl phthalate 11 (SML 300); 3,4-dimethylbenzaldehyde 15. DEHP, bisphenol-A, bisphenol-S, 4-phenylbenzophenone, p-tert-octylphenol, and 4-n-nonylphenol were not detected in any of the six bottles.
  • Headline: none of the authorised compounds exceeded its SML; several non-listed compounds (notably the PA monomer azacyclotridecan-2-one at 924 µg/kg, silicone TXIB at 348 µg/kg, Tritan 4-propylbenzaldehyde at 27 µg/kg, and PP 3,4-dimethylbenzaldehyde at 59 µg/kg) exceeded the proposed 10 µg/kg threshold for non-authorised substances (Onghena et al. 2016, in press).
• Reference treatment (Table 2, p. 898; µg/kg simulant; five sequential fillings at 40 °C, 30 min, no pre-treatment; ”—” = not detected, “a” = < LOQ):
  • Silicone bottle TXIB across fillings 1-5: 118, 102, 64, 71, 67 (LOQ 6.4). Silicone benzophenone fillings 1-5: 12, 8, 6, 4, 4 (LOQ 3.6). Silicone diisobutyl phthalate: 15, 12, 7, 6, 6 (LOQ 8.0). Silicone dibutyl phthalate fillings 2-5: 9, 7, 7, 7 (LOQ 4.4). Silicone acetophenone: 27, 11, 5, 4, 4 (LOQ 1.7). Silicone 3,4-dimethylbenzaldehyde filling 3: 6 (LOQ 5.6). Silicone 3,5-di-tert-butyl-4-hydroxybenzaldehyde: 4, 4, 4, 4, 4 (LOQ 3.1).
  • PA azacyclotridecan-2-one fillings 1-5: 70, 12, 10, 31, 15 (LOQ 9.6) — the 5-fold drop after filling 1 indicates a leachable surface pool, but the rebound at filling 4 indicates a sustained slow-release matrix pool.
  • PES acetophenone fillings 1-3: 2, 5, 2; fillings 4-5: < LOQ.
  • Tritan dicyclopentyl(dimethoxy)silane fillings 1-5: 2, 1, 1, 1, 1; Tritan 4-propylbenzaldehyde fillings 2-5: 4, 4, 4, 4; Tritan 4-methylbenzaldehyde fillings 3-5: 2, 2, 2. Tritan 2,4-di-tert-butylphenol fillings 1-5: 4, 4, 4, 4, 4 (< LOQ 6.2).
  • PP brand A: all targets either ND or < LOQ across all five reference fillings (3,5-di-tert-butyl-4-hydroxybenzaldehyde LOQ 3.1; no detectable migration).
  • PP brand B: all targets either ND or < LOQ across all five reference fillings (same pattern as PP-A).
  • Reference-treatment headline: maximum concentrations occur in the first filling and decrease over subsequent fillings; silicone is the polymer with the broadest reference-treatment migration profile (TXIB, benzophenone, di-iso-butyl phthalate, dibutyl phthalate, acetophenone, 3,4-dimethylbenzaldehyde, 3,5-di-tert-butyl-4-hydroxybenzaldehyde).
• Microwave heating (Table 3, p. 899; µg/kg simulant; cycles 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 40, 50, 60, 70, 80, 90, 100):
  • PES acetophenone cycles 1-9: 6, 2, 2, 2, 3, 2, 2, 2, < LOQ; gone after cycle 8 (LOQ 1.7).
  • PA azacyclotridecan-2-one: 124, 31, 19, 22, 26, 25, 14, 19, 34, 17, 7, 15, 22, 5, 5, 8, 7, 6, 11, 14, 7 (cycles 1, 2, 4, 6, 8, 10, 15, 20, 25, 30, 40, 50, 60, 70, 80, 90, 100; LOQ 9.6). The only bottle to demonstrate continuous release of a migrant (azacyclotridecan-2-one ≥ 5 µg/kg) throughout the entire 100 cycles.
  • Tritan TXIB: 9, 9, 7, 9, 8, < LOQ, < LOQ, 10, 9, 10, 8, 9, 13, < LOQ, < LOQ, < LOQ, < LOQ, < LOQ, < LOQ, < LOQ, < LOQ (LOQ 6.4) — present at LOQ-13 µg/kg over the first 25 cycles, dropped below LOQ thereafter.
  • Tritan 4-propylbenzaldehyde: cycles 1-7: 8, 4, 3, 3, 2, 1, < LOQ (LOQ 0.6) — disappeared after 8 cycles. Tritan dicyclopentyl(dimethoxy)silane: 1 µg/kg at most cycles 1-50 (LOQ 0.8).
  • Silicone TXIB: 10, 8, < LOQ × 5, 7, 9, 8, 11, < LOQ subsequently; reference-experiment silicone TXIB peak was 118 µg/kg, so the microwave releases far less TXIB than reference-treatment over 100 cycles.
  • Silicone benzophenone cycles 1-8: 14, 10, 6, 6, 6, 4, 5, 6 (LOQ 3.6); < LOQ from cycle 9 onward.
  • Silicone diisobutyl phthalate cycles 1-10: 24, 19, 14, 15, 15, 10, 14, 19, 12, 5; cycles 15-25: < LOQ, 17, 17; < LOQ thereafter (LOQ 8.0).
  • Silicone acetophenone: 43, 24, 14, 14, 13, 7, 7, 8, 5, 6, 2 (cycles 1-10, 15; LOQ 1.7) — versus 27 µg/kg in first reference experiment.
  • Silicone 3,4-dimethylbenzaldehyde cycles 1-7: 10, 6, < LOQ × 5 (LOQ 5.6).
  • PP brands A and B: no compounds detected at measurable concentrations or all < LOQ throughout 100 cycles.
• Dishwasher cleaning (Table 4, p. 900; µg/kg simulant; cycles 1, 2, 4, 6, 8, 10):
  • PA azacyclotridecan-2-one: 98, 18, 13, 55, 22, 39 — drops sharply then partially rebounds, similar to the reference filling pattern.
  • Tritan TXIB: —, 13, 8, 11, 8, 8; Tritan dicyclopentyl(dimethoxy)silane: 1, 3, 3, 3, 3, 3 (LOQ 0.8).
  • Silicone TXIB: —, 36, 34, 27, 23, 22 (well below the 118 µg/kg reference peak). Silicone benzophenone: —, 20, 13, 12, 7, 7. Silicone diisobutyl phthalate: —, 13, 16, 17, 14, 16. Silicone dibutyl phthalate: —, 7, 10, 11, 11, 13 — modest increase over cycles, the only authorised phthalate showing an increasing-with-cycles dishwasher trend. Silicone 2,4-di-tert-butylphenol detected at cycle 6 only (7 µg/kg).
  • PP brands A and B: no compounds detected. PES: no compounds detected.
• Steam sterilisation (Table 5, p. 901; µg/kg simulant; cycles 1, 2, 4, 6, 8, 10):
  • PA azacyclotridecan-2-one: 59, 33, 10, 4, 24, 6 — decreasing then small rebound; well below reference filling 1 (70 µg/kg).
  • Tritan 4-n-nonylphenol cycle 1 only: 2 (LOQ 1.0). Tritan dicyclopentyl(dimethoxy)silane: 1, 3, 3, 3, 3, 3. Tritan acetophenone cycle 1 only: 2.
  • Silicone TXIB: 28, 27, 23, 15, 16, 9 — dramatically lower than the cooking-sterilisation peak (247 µg/kg). Silicone benzophenone: 53, 66, 26, 14, 9, 12; the only silicone migrant initially released at concentrations higher than the reference treatment (12 µg/kg). Silicone diisobutyl phthalate: 10, 9, < LOQ × 4. Silicone dibutyl phthalate: 11, 14, 7, 7, < LOQ, < LOQ. Silicone 2,4-di-tert-butylphenol cycle 1 only: 7. Silicone acetophenone: 14, 8, < LOQ × 4. Silicone 3,4-dimethylbenzaldehyde cycles 1-2: 18, 15 (LOQ 5.6).
  • PP brands A, B and PES: no compounds detected after any steam-sterilisation cycle. PA acetophenone after cycle 2: < LOQ.
• Cooking (boiling-water, 10 min) sterilisation (Table 5, p. 901; µg/kg simulant):
  • PA azacyclotridecan-2-one cycles 1, 2, 4, 6, 8, 10: 51, 13, [b: sample lost], 9, 19, 7 (Table 5 footnote b = “sample lost during the experiment”).
  • Tritan 4-n-nonylphenol cycles 1-6: 6, 5, 2, 2, 1, 1 (LOQ 1.0) — released continuously by cooking sterilisation despite being undetected after the other three treatments. Tritan dicyclopentyl(dimethoxy)silane: 1, 1, 1, 1, 1 (cycle 1 sample lost). Tritan acetophenone cycle 1 only: 12. Tritan 4-propylbenzaldehyde cycle 1 only: 12.
  • Silicone TXIB: 247, 201, 178, 145, 127, 128 — significantly higher than steam (28-9 µg/kg) and reference-treatment (118 max) values; cooking sterilisation releases more TXIB from silicone than any other treatment tested. Silicone benzophenone: 58, 36, 25, 12, 7, 7. Silicone diisobutyl phthalate: 29, 29, 24, 19, 17, 15. Silicone dibutyl phthalate: 10, 12, 11, 9, 8, 8. Silicone 2,4-di-tert-butylphenol cycles 1-2: 12, 7. Silicone 2-butoxyethyl acetate cycle 1 only: 6 (LOQ 5.6). Silicone acetophenone cycles 1-2: 14, 4. Silicone cedrol cycle 1 only: 13 (LOQ 9.7) — not detected in reference, steam, or microwave. Silicone 3,5-di-tert-butyl-4-hydroxybenzaldehyde cycles 1-2, 5: 4, 4, 3. Silicone 3,4-dimethylbenzaldehyde cycle 1 only: 6.
  • PP brands A, B and PES: no compounds detected after any cooking-sterilisation cycle.
• Cross-treatment comparison (Discussion p. 902): the EU repetitive-use experiment over-estimated real-life migration for most compounds by up to a factor of 10 (e.g., silicone TXIB 2-10× higher in EU vs 10-cycle duration tests). For some compounds (notably silicone benzophenone) the 10-cycle duration concentrations were in the same range as the EU repetitive-use value, so the EU test did not under-estimate any of the targeted compounds. The lowest release across all duration tests was generally observed after microwave heating (slightly higher than the reference treatment for some compounds); cooking sterilisation released the highest concentrations for silicone bottles in particular.

Methods (brief)

Materials: six baby bottles purchased on the Belgian market (PP-A, PP-B, PES, PA, Tritan, silicone), corresponding to bottles 2 and 9 for PP and PES 1, PA 1, Tritan, silicone from Onghena et al. 2014. Reagents: ethanol (EMSURE, Reag Ph Eur, absolute), ethyl acetate (LiChrosolv), n-hexane (EDC for GC and FID SupraSolv), formic acid, acetonitrile, ammonia, ammonium sulphate, sodium chloride, sodium sulphate (all Merck KGaA, Darmstadt, Germany); ≥ 28 reference standards from Sigma-Aldrich Chemie GmbH (Steinheim, Germany) at 97-99 % purity covering the target additives, monomers, oligomers, and degradation products (acetophenone, 4-methylbenzaldehyde, 2-butoxyethyl acetate, 3,4-dimethylbenzaldehyde, 4-propylbenzaldehyde, 2-undecanone, 2,4,6-trimethylbenzaldehyde, 2,6-di-tert-butylbenzoquinone, dicyclopentyl(dimethoxy)silane, 2,4-di-tert-butylphenol, oxacyclotridecan-2-one, 2,2,4-trimethyl-1,3-pentanediol di-iso-butyrate (TXIB), p-tert-octylphenol, cedrol, benzophenone, 2,6-di-iso-propylnaphthalene, 3,5-di-tert-butyl-4-hydroxybenzaldehyde, azacyclotridecan-2-one, di-iso-butyl phthalate, dibutyl phthalate, 4-phenylbenzophenone, and methyl oleate); deuterated internal standard 2,6-di-tert-butyl-4-methylphenol-D24 (Campro Scientific GmbH, Berlin); helium ALPHAGAZ™ and nitrogen 99.999 % (Air Liquide). Milli-Q water from Elga Purelab flex (Veolia, Tienen, Belgium).

Bottle pre-treatment: bottles sterilised by filling with boiling water for 10 min, then subjected to three consecutive 2 h migrations at 70 °C with H₂O-EtOH 50:50 v/v (EU Reg. 10/2011 milk simulant); third migration analysed (EU repetitive-use experiment, per Onghena et al. 2016 Food Anal Methods).

Duration tests: reference treatment (5 fillings with preheated 40 °C simulant, 30 min at RT, no other pre-treatment). Four real-life treatments applied to fresh bottles: (1) microwave heating in a Whirlpool Gusto GT288WH oven, 500 W, heated to 40 °C with bottle on a rotating dish; 100 cycles; analysed at cycles 1-10, 15, 20, 25, 30, 40, 50, 60, 70, 80, 90, 100. (2) Dishwasher Eco-mode 2 h 55 min at 55-60 °C, common detergent, bottles inclined for full internal contact with water spray; 10 cycles; analysed at cycles 1, 2, 4, 6, 8, 10. (3) Steam sterilisation in a Philips Avent 3-in-1 electric steam steriliser with 100 ml tap water, ~10 min per cycle per user-manual conditions; 10 cycles; same analysis schedule. (4) Cooking sterilisation: boiling tap water immersion for 10 min per cycle; 10 cycles; same analysis schedule. After each treatment cycle, bottles were rinsed with 50 ml Milli-Q water, filled with 50 ml preheated 40 °C simulant for 30 min at RT, simulant transferred to glass containers, stored at 4 °C until analysis. A glass-bottle blank control was carried through every duration-test schedule.

Analysis: liquid-liquid extraction (LLE) with ethyl acetate/n-hexane 1:1 (three 10 ml extractions of 30 ml simulant), evaporated to ≈ 5 ml, 200 µl injected for GC-QqQ-MS; or direct injection by LC-QqQ-MS (Onghena et al. 2016 Food Anal Methods). QC sample spiked at 50 µg/kg ran in every batch. GC-QqQ-MS: Agilent 7890A GC + Agilent 7000 GC-MS triple quadrupole (Agilent JW Scientific, Diegem, Belgium), 1 µl pulsed splitless on PTV @ 300 °C, DB-5 ms 30 m × 0.25 mm × 0.25 µm column, oven 60 °C → 115 °C (7 °C min⁻¹) → 240 °C (10 °C min⁻¹) → 300 °C (15 °C min⁻¹) hold 15 min, total run 42.36 min, He carrier 1.0 ml min⁻¹ rising to 1.5 ml min⁻¹ at 22 min then back to 1.0, EI ionisation in MRM mode, N₂ collision gas 1.5 ml min⁻¹, quadrupole 150 °C, source 230 °C, multiplier voltages 1537 / 1630 V for gains 10 / 20. LC-QqQ-MS: Waters Acquity UPHLC with degasser, BPGP, thermostated column compartment, autosampler; Waters Acquity UPLC C₁₈ BEH 100 × 2.1 mm 1.7 µm, 0.4 ml min⁻¹, 30 °C, 10 µl injection; ESI(+) mobile phase A H₂O + 0.1 % HCOOH / B AcN + 0.1 % HCOOH; ESI(−) mobile phase A H₂O + 0.1 % NH₃ / B AcN + 0.1 % NH₃; gradient 0 min 5 % B → 0-6 min 95 % B → 6-8 min 95 % B → 8-10 min 5 % B; Waters Xevo TQ-S triple quadrupole MS, ESI capillary 3 kV (+) / 2.5 kV (−), cone 30 V, cone gas 50 l h⁻¹, source 150 °C, desolvation 800 l h⁻¹, collision 0.15 l h⁻¹. Two compounds run in ESI(+) and four in ESI(−); compound-specific MS parameters per Onghena et al. 2016 Food Anal Methods.

Regulatory framing: EU Regulation 10/2011 on plastic FCMs (European Union 2011b) and the 2011 EU ban on polycarbonate baby bottles (Directive 2011/8/EU, European Union 2011a). EU 10/2011 Article 11(2) sets a generic 60 mg/kg SML for substances without specified limits; Barlow 2009 defines the 10 µg/kg generic threshold for non-listed substances behind functional barriers (also acting as the substance-of-very-low-concern threshold). Genotoxicity prioritisation of detected migrants used ToxTree, Derek Nexus™, and Vitotox™ assays per Mertens et al. 2016 plus receptor gene assays (oestrogen, androgen, progesterone, glucocorticoid, thyroid beta, PPAR-γ, AhR) per Simon et al. 2016.

Implications

• Certification (HMTc): Out of core scope. The paper measures only organic migrants from baby-bottle polymers; it does not contribute to the HMTc 10-analyte panel (Pb, tAs, Cd, MeHg, tHg, iAs, Ni, Al, Cr-VI, Sn) on the infant-bottles product category. The methodological scaffolding — six bottles, six polymer types, EU vs four real-life treatments, validated GC/LC-QqQ-MS with deuterated IS, the demonstration that EU 10/2011 conditions over-estimate real-life by up to 10×, and the recommendation that steam sterilisation be used pre-first-use to deplete leachable residuals — is reusable if/when a future study examines metal migration (e.g., Sb from PES, residual Pb from silicone curing catalysts, Ni from stainless-steel manufacturing transfer, Cr from PA polymerisation catalysts) from these same polymer alternatives.
• Courses: Useful as a worked example of FCM migration testing for a brand or QA audience: the EU regulatory framework (Reg. 10/2011), the distinction between authorised SMLs and the 10 µg/kg threshold for non-listed substances behind functional barriers (Barlow 2009), the difference between accelerated standardised testing and real-life duration tests, and the practical “steam-sterilise the bottle before first use” guidance for consumers all sit at an appropriate level for a course module.
• App: Not directly applicable to the heavy-metals consumer app. The paper-internal recommendation that consumers steam-sterilise bottles before first use to reduce migrant release is a behaviour rather than a heavy-metals risk signal, but is consistent with general infant-feeding hygiene guidance and may surface as background context if the app evolves to cover broader FCM exposure.

Wiki pages this source may touch

• infant-bottles — primary routing target (methodology / out-of-core-scope reference; no metal occurrence data to populate the page’s Literature Evidence Summary).

Verification notes

• No heavy-metal occurrence data. metals: [] is correct. The paper measures organic migrants only — additives, monomers (azacyclotridecan-2-one for PA), plasticisers (TXIB, di-iso-butyl phthalate, dibutyl phthalate, DEHP), antioxidants (3,5-di-tert-butyl-4-hydroxybenzaldehyde), bisphenols (BPA, BPS, both non-detected), alkylphenols (4-n-nonylphenol, p-tert-octylphenol), benzophenones (benzophenone, 4-phenylbenzophenone), siloxanes (dicyclopentyl(dimethoxy)silane), aldehydes (3,4-dimethylbenzaldehyde, 4-propylbenzaldehyde, 4-methylbenzaldehyde, 2,4,6-trimethylbenzaldehyde), and ketones (acetophenone, 2-undecanone, cedrol, 2-butoxyethyl acetate, methyl oleate, oxacyclotridecan-2-one). Paper is ingested as out-of-core-scope methodology / FCM-migration reference per the precedent set by aurisano2022-mouthing-exposure-childrens-products.md and lai2025-infant-diaper-phthalate-dna-oxidation.md.
• Brand-firewall compliance. The paper uses anonymised “PP brand A” and “PP brand B” labels throughout the tables and text — the actual bottle brand names are NOT disclosed in the published article. The only proper names present are scientific-method vendor names: Whirlpool Gusto GT288WH microwave, Philips Avent 3-in-1 electric steam steriliser, Agilent 7890A GC / 7000 GC-MS, Waters Acquity UPHLC / Xevo TQ-S, Merck KGaA, Sigma-Aldrich, Campro Scientific, Air Liquide, Veolia/Elga Purelab. These are permitted under Exception 2 (scientific-method vendor/material names) of the strict brand-firewall reading (verification-checklist.md §“Firewall checks”). No brand-attribution-to-contamination-values violations.
• Wiki/HMTc firewall compliance. The Implications section does not propose HMTc threshold values; it describes the paper’s contribution to threshold-relevant methodology only and explicitly notes the paper is out of HMI’s core (heavy-metals) scope. No synthesis claims about other literature, no consumer risk advisories, no new-page suggestions in the body.
• Matrices field empty intentionally. The migration medium is an EU-prescribed simulant (H₂O-EtOH 50:50 v/v) standing in for milk. The HMI matrices vocabulary covers real food matrices (rice, infant-formula, fish, etc.), not regulatory simulants. The products: [infant-bottles] slug carries the routing for this paper; no matrix attribution is necessary or accurate.
• Sample-size and bottle correspondence. Six bottles total: “one bottle of each different polymer type (PP, PES, PA, Tritan and silicone) was selected. Due to the large market share of PP bottles, an additional PP bottle exhibiting another variety of migrating compounds was selected for the duration tests” (Methods §“Market survey and sampling”, p. 894). The six bottles correspond to bottles 2 and 9 (PP-A, PP-B), PES 1, PA 1, Tritan 1, silicone 1 from the authors’ prior market survey Onghena et al. 2014.
• Funding and conflicts. Federal Government Service for Public Health of Belgium grant ALTPOLYCARB (RT 12/10) to Onghena; postdoctoral fellowship from University of Antwerp to Negreira. “No potential conflict of interest was reported by the authors.” Public-funded, no industry conflicts (Funding / Disclosure statement, p. 903).
• License. Taylor & Francis subscription article (Food Additives & Contaminants: Part A). Downloaded by University of Antwerp 12 May 2016 per the watermarked PDF. No CC-BY or open-access designation present. License field set to “Taylor & Francis (subscription)“.
• DOI verified. 10.1080/19440049.2016.1171914 matches the link printed on the title page and the running header. Published volume 33, issue 5, pages 893-904, online publication 20 April 2016 (accepted author version 4 April 2016).
• Audit subagent (2026-05-31) flagged the PP-A / PP-B 3,5-di-tert-butyl-4-hydroxybenzaldehyde reference-filling values as misattributed from the silicone row. Verified against Table 2, p. 898: the “4, 4, 4, 4, 4” sequence belongs to the silicone row only; PP-A and PP-B rows show footnote-”a” (<LOQ) entries across all five fillings. Applied: PP-A and PP-B bullets now read “all targets either ND or < LOQ across all five reference fillings” rather than implying a measured 4 µg/kg concentration.
• Audit subagent (2026-05-31) flagged the PA cooking-sterilisation cycle-4 entry as internally inconsistent (”< LOQ” in the value sequence but “sample lost” in the trailing note). Verified against Table 5, p. 901: cycle 4 is footnote-”b” (sample lost during the experiment), not <LOQ. Applied: rendering changed to “[b: sample lost]” in the value sequence with the footnote definition cited inline.
• 2026-05-31 near-duplicate merge (Claude Opus 4.7, autonomous v2.0 manual-fetch skill): conference-proceedings companion PDF 02_2016_IMEKO-TC23-2016-038.pdf (2nd IMEKO Foods 2016, pp. 191-194, identifier IMEKO-TC23-2016-038) processed from the same _extracted_infantcontact_01_Bottles_Nipples_SippyCups Kimi corruption-folder source bundle. Verified by full-text read against this F&C Part A journal page: identical title (“Evaluation of the migration of chemicals from baby bottles under standardised and duration testing conditions”), identical author list (Onghena M, Van Hoeck E, Negreira N, Quirynen L, Van Loco J, Covaci A), identical six-bottle sample, identical four duration-test protocols (microwave 100 cycles; dishwasher 10 cycles; steam sterilisation 10 cycles; cooking sterilisation 10 cycles), identical EU-repetitive-use reference experiment, identical conclusions. The IMEKO paper is a 4-page conference summary of the same study; its body cites “Onghena et al. 2016” (the separate Food Anal Methods methods paper) for chemicals, procedure, and instrumentation. No new evidence; no separate wiki source page created (would double-count routing and evidence); registered in near_duplicates instead. No re-audit spawned: the IMEKO paper reports a strict subset of the data already audited on this page; the journal-paper page is already at audited-promote per the 2026-05-31 audit. Audit-queue entry for the IMEKO PDF logged with status near-duplicate-merged so the manual-fetch tracker reflects the PDF was processed without spawning a separate audit.
• 2026-05-31 PhD-thesis-superset near-duplicate merge (Claude Opus 4.7, autonomous v2.0 manual-fetch skill): umbrella PhD-thesis PDF 08_133572.pdf (Onghena M. 2016, “Study of the possible migration risks of food contact materials for children under 3 years”, joint UAntwerpen / UJI, 232 pp; supervisors Covaci, Van Loco, Hernández Hernández) processed from the same _extracted_infantcontact_01_Bottles_Nipples_SippyCups Kimi corruption-folder bundle. Verified by direct reading of the table of contents (pp. v-ix) and full reading of the general-discussion chapter (Ch 6, pp. 167-181) plus Table 6.1 summary of investigated baby bottles (p. 171): this F&C Part A journal article appears as Chapter 5.2 of the thesis (§5.2 “Quantitative evaluation of the migration under standardised and real-life use conditions”, pp. 139-156) with identical six-bottle sample, identical EU repetitive-use experiment, identical four duration tests, and identical conclusions. The thesis additionally contains Ch 3 (Belgian-and-Internet market study of infant FCMs), Ch 4.1 (qualitative GC-MS migrant screening, including a stainless-steel bottle not in the journal paper), Ch 4.2 (unknown-migrant elucidation by GC-(APCI)QTOF-MS and LC-QTOF-MS), Ch 5.1 (LLE method development and validation), Ch 5.3 (GC-(EI)TOF-MS degradation-product fingerprinting), and Ch 6 (general discussion / future perspectives). Each of those chapters is the pre-publication chapter form of a separately published Onghena et al. paper. Critically, none of the thesis chapters report heavy-metals occurrence data — the entire 232-page work measures organic FCM migrants only (additives, monomers, plasticisers, antioxidants, bisphenols, alkylphenols, siloxanes, NIAS, degradation products). Same out-of-core-scope determination as the journal-paper page: zero contribution to the HMTc 10-analyte panel. No separate wiki source page created (would double-count routing and evidence with no metals gain); registered as a second entry in near_duplicates instead, paralleling the IMEKO-subset registration above (inverse direction: thesis is the superset, IMEKO is the subset, journal paper is the central audited unit). No re-audit spawned: thesis Chapter 5.2 reports identical data to this already-audited-promote journal page; the additional thesis chapters are out-of-core-scope for HMI and contribute no metals-bearing claims for audit to evaluate. Audit-queue entry for the thesis PDF logged with status near-duplicate-merged. PDF SHA-256 7dcddc150839b2ff8adf46c386adef17dc8eeb8643d9828cf7246c2ee92a6f98.

Ingest log

• 2026-05-31 fresh ingest (Claude Opus 4.7, autonomous v2.0 manual-fetch skill): NEW path. Three identity checks against wiki/sources/ returned no hits: DOI 10.1080/19440049.2016.1171914 not present; raw_handle MFK_01-2015-evaluation-of-the-migration-of-chemicals-f not present; cite-key stem onghena2016 not present. PDF SHA-256 43fd921d1b518d529c4aa1ffb3021a72998df6f6cb26713f7980ac5386671b31. Paper measures organic FCM migrants only — zero heavy metals — ingested as methodology / out-of-core-scope reference per the lai2025 and aurisano2022 precedents. Routed to [[products/infant-bottles]] for discoverability as a baby-bottle FCM-migration reference but metals: [] correctly reflects no metal occurrence data.

Page history

The five most recent substantive edits to this page. The full version history lives in git; when DOI minting comes online (see schema docs), each entry below will also link to a version-pinned DataCite DOI.