Buculei et al. 2012 — Tin and iron migration from tinplate cans into canned foodstuffs over 1080-day storage (Romania)

This Romanian packaging-migration study measured Fe and Sn concentrations in four canned product types (peas, tomato paste, pork in own juice, and pork liver pate) packed in three-piece tinplate cans with different internal varnish systems and held at ambient temperature (~20 °C) for up to 1080 days (≈36 months). For each product/varnish combination, samples were drawn at 0, 180, 360, 720, and 1080 days and analysed by atomic absorption spectrometry for iron and by inductively coupled plasma atomic emission spectrometry for tin. The authors report that Fe and Sn concentrations rise monotonically with storage time across all products and varnish systems, that acidic matrices (tomato paste) accumulate Sn substantially faster than near-neutral matrices (peas, pork, pate), and that the relative migration ordering between varnish types is product-dependent. Numerical values are presented exclusively as scatter-plus-regression-line graphs (Figures 1–8) and short conclusion-section qualitative statements; no underlying tabulated values are printed. The authors cite the Turkish Food Codex 15 µg/g Fe limit as the relevant ceiling and the Romanian Ministry of Health Order No. 975/1998 as the in-country reference; they do not specify a Sn ceiling.

Key numbers

All concentrations are reported in mg/kg of product as sampled (the source uses the Romanian axis label “Concentrația [mg/kg]”; basis is not explicitly stated as wet or dry but the product is sampled directly from the can without drying, so the values are functionally wet-weight / as-consumed basis for canned foods). Storage time x is on the figure x-axis in days (“Perioada de depozitare [zile]”) with sample points at 0, 180, 360, 720, and 1080 days. Raw per-point values are not tabulated in the source; values below are (a) the linear-regression equations printed inside each figure and (b) approximate endpoint reads from the plotted points where the regression form alone would understate or overstate the visual data. Figures are reproduced from the source PDF as printed; precision is limited to the visual resolution of the embedded plots.

Iron migration (Fe), mg/kg, regression of concentration on storage time x in days:

Authors’ notes on the Fe values: the iron content of canned peas, tomato paste, and pork rests between ~12 and ~21 mg/kg across products and varnish systems, near or just above the Turkish Food Codex ceiling of 15 µg/g (= 15 mg/kg) the authors cite as the relevant Fe limit for canned food. Liver pate Fe is roughly an order of magnitude higher (~125–155 mg/kg), consistent with the matrix being an organ-meat product naturally rich in heme iron rather than reflecting packaging migration alone.

Tin migration (Sn), mg/kg, regression of concentration on storage time x in days:

Authors’ notes on the Sn values: tomato paste accumulates Sn far faster than the other matrices, reaching approximately 23 mg/kg under the universal varnish system at 1080 days, with the sulfur-resistant Varnish 3 the slowest accumulator (~9 mg/kg at 1080 days). For peas, pork, and pate, Sn remains below ~1.2 mg/kg across the full 1080-day window. The authors state that “the migration of Sn content preserved covered with lacquer, sulfur-resistant is more pronounced in the first storage period” for pork in own juice and that for liver pate “a significant increase in the levels of Tin in the second period preserved in storage-covered varnish sulfur-resistant to varnish preserved covered with universal” is observed — phrasings reproduced as published from the source’s non-native-English text. The negative intercepts on several Sn regressions reflect that day-0 values are at or below the LOQ rather than literally negative; the figures show plotted points at or just above zero for the corresponding day-0 measurements.

Conclusion-section qualitative endpoints (p. 302):

• Peas: after 6 and 12 months under Varnish 1 and Varnish 2, heavy-metal contents change “do not very significantly” and remain within the scope of Romanian Ministry of Health Order No. 975/1998. After 24 months, metal content “increases significantly”.
• Tomato paste: after 6 and 12 months, Fe and Sn vary “significantly”, with progressive increases under all three varnish systems through both time points.
• Pork in own juice and liver pate: not addressed in the conclusion text beyond the figure presentations; the conclusion final bullet reports a generic “significant increase in heavy metal content of the product packed in cans of film, protected by varnish 1 and varnish 2”.

Visual inspection of can interiors (p. 302): the authors report that on opening representative cans at the end of storage, no visible varnish damage was observed under normal ambient-temperature storage for either varnish system.

Methods (brief)

Sample preparation: 1 g aliquots of each canned product were digested with 6 mL concentrated HNO₃ (65%, Suprapur, Merck) plus 2 mL concentrated H₂O₂ (30%, Suprapur, Merck) in a Milestone Ethos D closed-vessel microwave digestion system (maximum pressure 1450 psi, maximum temperature 300 °C) using a stepped program (2 min @ 250 W → 2 min @ 0 W → 6 min @ 250 W → 5 min @ 400 W → 8 min @ 550 W → 8 min vent). Digests were diluted to 10 mL with Milli-Q water (18.2 MΩ·cm resistivity). A reagent blank was carried through the digestion in parallel.

Analytical instruments: iron and most metals were determined on a Perkin-Elmer Analyst 400 atomic absorption spectrometer with deuterium-lamp background correction, using an HGA graphite-furnace programme with argon as inert gas for Fe (the source text contains an internal inconsistency, additionally stating that “the elements were determined by using air–acetylene flame,” which is the flame-AAS configuration for the same instrument; the source does not resolve which configuration was used for which analyte). Tin was determined on a separate ICP-AES spectrophotometer (vendor and model not specified). Operating parameters were used as recommended by the instrument manufacturer; the source does not print element-specific wavelengths, calibration ranges, recoveries, LODs, or LOQs.

Speciation: total Fe and total Sn only; no inorganic/organic speciation. No other metals from the HMTc analyte vocabulary (Pb, tAs, Cd, MeHg, tHg, iAs, Ni, Al, Cr-VI) were reported in this paper.

Experimental design: four canned product types (peas, tomato paste, pork in own juice, pork liver pate) were packed in three-piece tinplate cans coated internally with one of three varnish systems — Varnish 1 = universal varnish; Varnish 2 = epoxyphenolic varnish; Varnish 3 = sulfur-resistant varnish, applied to the two meat products. Cans were held at ambient temperature (~20 °C) and sampled at 0, 180, 360, 720, and 1080 days. The source does not report can dimensions, fill weight, headspace, replicate counts per time point, or whether the time-zero sample was taken before or after retort processing.

Limitations

• Regional Romanian technical journal (Journal of Agroalimentary Processes and Technologies, 2012, 18(4)) without a DOI; located via the journal’s host site at journal-of-agroalimentary.ro. Tier C reflects the publication route and the methodological limitations below, not a judgment on the underlying experimental design.
• Numerical results are presented exclusively as scatter-plus-regression-line plots embedded in the article. No tabulated mean ± SD values, no per-time-point sample counts, no LODs, no recovery percentages, and no reference-material results are printed. All numerical values transcribed to this page are read either from printed regression equations or by visual inspection of the plotted points and carry the corresponding precision limits.
• Sample size per product/varnish/time-point cell is not stated; the regressions appear to be fitted to single mean values at each of the five time points, giving N ≈ 5 (time points) per regression line and no replication estimate at any individual time point.
• The Methods section contains an internal inconsistency regarding AAS configuration: it states both “HGA graphite furnace using argon as inert gas” and “the elements were determined by using air-acetylene flame” for iron and the unspecified non-tin metals. The source does not resolve which atomisation mode applied to iron specifically. Whichever configuration was used, no quality-assurance data (NIST reference materials, certified reference materials, intra-/inter-day precision) is reported.
• The source-language English contains transcription artefacts (“conservability”, “white Lake”, “yellow Lake”, “the canned lacquer coated universal”, inverted figure descriptions); the descriptive text in the Results and Conclusion sections occasionally states the opposite of what the corresponding figure shows (e.g., the Conclusion bullet on tomato paste references “white” and “yellow” lacquer films not introduced in Methods). Where the figure and the prose disagree, the figure has been treated as primary in this transcription.
• Day-0 negative intercepts on the Sn regressions for tomato paste and liver pate reflect linear fits to data clusters that start at or below the analytical limit and rise sharply afterwards; the regression equations should not be extrapolated to negative time and should not be read as evidence of pre-canning Sn contamination.
• Storage was at ambient temperature (~20 °C) only; no elevated-temperature accelerated study was conducted, and migration trajectories at temperatures encountered in real-world retail and household storage (e.g., 25–30 °C in non-air-conditioned environments) are not extrapolated by the authors.

Implications

• Certification (HMTc): Of the two analytes measured, only tin is on the HMTc 10-analyte certification list. The strongest signal in the data is the acid-driven Sn migration in canned tomato paste, which reaches roughly 9–23 mg/kg at 36 months depending on the internal varnish system; this is well below the Codex CXS 193-1995 general-food Sn maximum of 200 mg/kg and the canned-food guidance level of 100 mg/kg but well above day-0 values, demonstrating that storage time and varnish chemistry together are first-order determinants of Sn occurrence in acidic canned foods. For low-acid matrices (peas, pork in own juice, pork liver pate), Sn remains below ~1.2 mg/kg throughout the 1080-day window. Iron is not on the HMTc analyte list. The paper contributes Sn occurrence data for canned tomato paste, canned peas, canned pork, and canned pork liver pate from a Romanian production-and-storage context; the data should be pooled into product-category Sn distributions only with a note that the values are storage-time-resolved rather than market-snapshot.
• Courses: Useful as a worked example of the storage-time dimension of metal migration from tinplate cans, the role of internal varnish chemistry in determining migration rates, and the difference acidic versus near-neutral matrices make for tin uptake from the can wall. Also a useful illustration of how regional technical journals can report directionally informative results while withholding the methodological transparency a Cochrane-grade audit requires.
• App: Adds Sn data points for canned-peas / canned-tomato-paste / canned-meat composites from a Romanian storage-experimental context. The values are storage-time-dependent; an as-purchased market sample would correspond to an intermediate time point rather than the 1080-day endpoint. Fe values for the four products are reported but Fe is not part of the app’s contamination-likelihood model.
• Microbiome: Not addressed.

Wiki pages this source may touch

• tin
• iron
• tomato-paste
• tomato
• meat
• organ-meats
• canned-vegetables
• tomato-paste
• canned-tomatoes
• processed-meats
• pork-product
• food-packaging-cans-lids
• codex-cxs-193-1995-tin-canned-foods

Verification notes

• 2026-06-02 — Fresh ingest from raw/manual-fetch/Kimi_Agent_Download Corruption Issue/condiments2_papers/05_Snacks_Canned_Prepared/Tin_Iron_Migration_Canned_Food_Romania.pdf. All five pages of the PDF (abstract, introduction, methods, results-and-discussion with embedded Figures 1–8, conclusion, references) were read in full.
• Numerical values in the Key numbers tables are transcribed directly from the regression equations printed inside Figures 1–8; the day-0 / day-1080 approximations are visual reads from the scatter points in the same figures, rounded to the nearest unit (Fe) or 0.1 mg/kg (Sn). No tabulated per-time-point values are printed in the source.
• The source’s Methods section contains the internal inconsistency about AAS atomisation mode (graphite furnace vs flame) noted in the Limitations section; both statements are quoted from p. 300 of the PDF and the source does not reconcile them.
• The source’s Conclusion bullet on tomato paste introduces “white Lake” and “yellow Lake” varnish identifiers not defined elsewhere in the paper; these have been treated as transcription artefacts of the source’s English and have not been propagated to this page.
• Brand-firewall (Part 12): the source does not name any commercial brands of the four canned products tested; the cans appear to have been produced experimentally or by the authors’ affiliated facility. No brand-firewall redactions were required.
• Wiki/HMTc firewall (Part 2): the Implications section reports the literature observation that tomato Sn migration reaches ~9–23 mg/kg at 36 months under the three varnish systems tested and notes the relationship to the Codex CXS 193-1995 ceilings without proposing an HMTc threshold value.
• 2026-06-02 — Audit subagent (general-purpose) flagged Check 1 (numerical fidelity) with seven ⚠️ concerns on regression-equation digits in Figures 3, 4, 5, 6 where the subagent’s read disagreed with the initial transcription. Re-read Figures 3 and 4 (PDF p. 3) and Figures 5 and 6 (PDF p. 3) at higher attention; the subagent was correct on six of seven and false-positive on one. Corrections applied: Figure 3 Tomato V1 Fe intercept 12.064 → 12.456; Figure 3 Tomato V3 Fe R² 0.9425 → 0.9725; Figure 4 Tomato V2 Sn equation 4.036·x − 4.503 → 4.426·x − 4.901 and R² 0.6595 → 0.7104; Figure 5 Pork V1 Fe R² 0.7794 → 0.7796; Figure 5 Pork V2 Fe intercept 11.127 → 10.127; Figure 6 Pork V1 Sn R² 0.6312 → 0.8327; Figure 6 Pork V2 Sn equation 0.215·x + 0.179 → 0.242·x + 0.179 and R² 0.8427 → 0.827. False positive (subagent uncertain, re-read confirmed original): Figure 8 Liver pate V2 Sn R² remains 0.93; the rendered figure label reads 0.93, not 0.63. Day-0 / day-1080 visual reads were not affected; the corrections are to the printed regression equations themselves, not to the plotted scatter points used for the endpoint approximations.
• 2026-06-02 — Audit subagent Checks 2 (slug vocabulary), 3 (speciation and methods), 4 (Part 12 brand firewall), and 5 (Part 2 wiki/HMTc firewall) returned ✅. No changes required.

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.