This study combined a bottle-use survey in Britain and Nigeria with ICP-MS measurements of antimony in PET and in bottled water and soft drinks. Antimony in 32 PET materials was similar in the two countries. Table 2 medians are 260 mg/kg (Nigeria, n=14) and 251 mg/kg (Britain, n=18). In freshly purchased British bottled waters and drinks the printed range is 0.033 to 6.61, with one sample above the EU drinking-water value. The text layer renders that unit as mg/L. The page image of the same paragraph prints µg/L, and it prints the EU maximum admissible concentration as 5 µg/L and the US maximum contaminant level as 6 µg/L. The tolerable daily intake on that image is 6 µg/kg/day, and the packaging migration sentence is 40 µg of antimony per kilogram of food. PET concentrations stay in mg/kg because Table 2 states that unit and the values are in the hundreds.
Key numbers
| Analyte | Matrix | Value | Unit | Statistic | N | Where |
|---|---|---|---|---|---|---|
| Sb | drinking water limit as printed | 5 | µg/L | regulatory_limit | p.PDF p.2 (printed 1238) | |
| Sb | drinking water limit as printed | 6 | µg/L | regulatory_limit | p.PDF p.2 (printed 1238) | |
| Sb | EU specific migration limit from PET into food | 40 | µg/kg | regulatory_limit | p.PDF p.2 (printed 1238) | |
| Sb | WHO tolerable daily intake as printed | 6 | µg/kg/day | regulatory_limit | p.PDF p.2 (printed 1238) | |
| Pb | British MAC for Pb until December 2013 | 25 | µg/L | regulatory_limit | p.PDF p.7 (printed 1243) | |
| Pb | EU MAC for Pb after the British reduction | 10 | µg/L | regulatory_limit | p.PDF p.7 (printed 1243) | |
| Sb | water DL as listed in methods sentence | 0.02 | µg/L | lod | p.PDF p.3 (printed 1239) | |
| Pb | water DL as listed in methods sentence | 0.03 | µg/L | lod | p.PDF p.3 (printed 1239) | |
| Cd | water DL as listed in methods sentence | 0.02 | µg/L | lod | p.PDF p.3 (printed 1239) | |
| Zn | water DL as listed in methods sentence | 0.01 | µg/L | lod | p.PDF p.3 (printed 1239) | |
| Co | water DL as listed in methods sentence | 0.16 | µg/L | lod | p.PDF p.3 (printed 1239) | |
| Ti | water DL as listed in methods sentence | 0.02 | µg/L | lod | p.PDF p.3 (printed 1239) |
Methods (brief)
ERM-EC681k polyethylene (certified Sb 99 ± 6 mg/kg) was used because no certified PET was available. Water accuracy was checked with TM-DWS.2. Migration tests used 300 mL deionised water in 500 mL bottles at 40, 60, and 80 °C. The text layer prints a 1 mg/L spike and a 4.08 mg/L migration maximum. Those strings use the same dropped micro prefix as the introduction. In the ledger the water unit is µg/L. The spike sentence itself was not cropped.
Detection and reporting: Methods sentence lists DLs for Sb, Pb, Cd, Zn, Co, Ti, Al, Be and Ge as 0.02, 0.03, 0.02, 0.01, 0.16, 0.02, 0.003, 0.13 and 0.08, based on 3 SD of 10 blanks. The printed unit is not attached to that sentence. Table 3 left-censor flags do not line up with that order. No LOQ is stated.
Basis: PET content is mg/kg of plastic (Table 2). Water and drink results are µg/L on the page image. One of 47 samples exceeded the EU 5 µg/L value, which matches a range whose top end is 6.61 µg/L.
Evidence fitness
This source can support total antimony in British and Nigerian PET at the Table 2 medians, summary antimony in the purchased British bottled waters and drinks on a µg/L basis, and the drinking-water values read from the page image (EU 5 µg/L, US 6 µg/L, intake 6 µg/kg/day, food migration 40 µg/kg).
It cannot support speciation, brand-level tables, or a certification limit. The glyph correction was checked on the introduction image and then applied to the other water rows in this file. Each of those rows still keeps the text-layer unit in the ledger.
Limitations
The text layer has no micro glyph anywhere in this PDF. The introduction image shows µg for the drinking-water values, the intake value, and the 40 µg/kg food sentence. The same correction is applied to other water rows, including the lead values the text layer prints as 25 and 10 mg/L, which are the historical 25 and 10 µg/L lead maxima if the glyph is read the same way. Those lead lines were not cropped separately. Detection limits may be column-scrambled against Table 3. Figure time series were not transcribed number by number. Brand rows were not transcribed.
Related evidence
Update history
No substantive edit history is available in this build. The full commit record is available in git.