Nine PET bottled waters bought in Arizona were measured for antimony by ICP-MS, with a local tap water for comparison. As purchased, bottled-water antimony ranged from 0.095 to 0.521 ppb, with means of 0.195 ± 0.116 ppb at the start and 0.226 ± 0.160 ppb after three months at 22 °C. Heating one bottle to 80 °C raised antimony well above the 6 ppb USEPA maximum contaminant level, reaching 14.4 ppb after seven days, while pH from 6.3 to 8.3 and freezing did not. Microwave digestion of that bottle’s plastic found 213 ± 35 mg/kg antimony. Brand names appear in the methods and are omitted here; the paper’s brand ID numbers are used where a result is tied to one bottle.
Key numbers
| Analyte | Matrix | Value | Unit | Statistic | N | Where |
|---|---|---|---|---|---|---|
| Sb | drinking water limit (USEPA/Ontario/Health Canada) | 6 | ppb | regulatory_limit | p.PDF p.2 (printed 552) | |
| Sb | drinking water limit (Germany) | 5 | ppb | regulatory_limit | p.PDF p.2 (printed 552) | |
| Sb | drinking water limit (EU) | 5 | ppb | regulatory_limit | p.PDF p.2 (printed 552) | |
| Sb | drinking water limit (Japan) | 2 | ppb | regulatory_limit | p.PDF p.2 (printed 552) | |
| Sb | drinking water limit (WHO) | 20 | ppb | regulatory_limit | p.PDF p.2 (printed 552) | |
| Sb | ICP-MS water | 0.004 | ppb | lod | p.PDF p.3 (printed 553) | |
| Sb | nine Arizona PET bottled waters, as purchased | 0.095-0.521 | ppb | range | 9 | Fig. 1 p.PDF p.1 (printed 551) |
| Sb | nine Arizona PET bottled waters, start of study | 0.195 ± 0.116 | ppb | mean | 9 | Fig. 1 p.PDF p.1 (printed 551) |
| Sb | nine Arizona PET bottled waters, after 3 months at 22 °C | 0.226 ± 0.160 | ppb | mean | 9 | Fig. 1 p.PDF p.1 (printed 551) |
| Sb | Tempe, Arizona tap water | 0.146 ± 0.002 | ppb | mean | p.PDF p.3 (printed 553) | |
| Sb | Brand 9 bottled water, Fig. 3 maximum cited | 13.9 | ppb | max | Fig. 3 p.PDF p.6 (printed 556) | |
| Sb | sunlight-exposed Brand 9, 7 days, water about 35 °C | 0.37-0.61 | ppb | range | triplicate | Fig. 5 p.PDF p.5 (printed 555) |
Methods (brief)
Bottles purchased summer 2006. Water transferred to acid-washed HDPE, acidified with HNO3, indium internal standard. Incubations at 22, 40, 60, and 80 °C. Screening matrix on brand ID 9: pH 6.3, 7.3, 8.3; control 48 h; frozen −20 °C 48 h; 80 °C 48 h; UV-PEN 254 nm for 6 h. Outdoor sun exposure 9-16 August 2006 with foil-wrapped controls. Piece leaching: 9.4 cm × 9.4 cm plastics in 1 L nanopure water at 80 °C up to 10 days.
Detection and reporting: Limit of detection 0.004 ppb; method reporting level 0.028 ppb. Handling blank <0.005 ppb. Precision within 3% (1s) on NIST 1640, NIST 1643e, and NRC SLRS-4.
Basis: Sb in water as ppb. PET content as mg/kg plastic and mg/cm2. Migration experiments are release into the bottled water or into nanopure water, not total content. The paper equates ppb with mg/L in the MCL sentence; the water concentrations are in the μg/L range that drinking-water MCLs use, and are transcribed in the printed unit ppb.
Evidence fitness
This source can support: Total Sb in southwestern US PET bottled water as purchased, as a range and mean; Temperature-dependent Sb release from PET into water; Total Sb and a few other elements in one PET bottle after digestion; Cited drinking-water Sb limits.
It cannot support: A full brand-level concentration table; Sb speciation; A certification limit; Sunlight as a large independent driver (the paper found only a small, non-significant difference versus a temperature-matched control).
Limitations
Per-brand as-purchased concentrations exist only as Fig. 1, so they were not transcribed bottle by bottle; the printed range and the two study means were entered, and individual rows were not all transcribed. One pH 7.3 control cell is a dash in Table 1 and was omitted. The 376 ppb figure is a complete-leach calculation, not a measurement. Chromium in the digest is total chromium, not chromium(VI). The paper prints the MCL unit as ppb and once as mg/L in the same sentence.
Related evidence
Update history
No substantive edit history is available in this build. The full commit record is available in git.