Skip to content
Heavy Metal Index
A plastic PET water bottle and thermometer illustrating the effect of storage temperature on antimony migration studied by Allafi 2020.

The effect of temperature and storage time on the migration of antimony from polyethylene terephthalate (PET) into commercial bottled water in Kuwait

Allafi (2020) examined how storage temperature and duration affect the leaching of antimony from polyethylene terephthalate (PET) plastic into commercial noncarbonated bottled water purchased in Kuwait. Kuwait’s summer ambient temperatures can approach 50 degrees C, making extreme-temperature storage scenarios practically relevant to that market. Three commercially available brands (9 bottles total) were tested under multiple temperature and time conditions using ICP-MS. The study found that antimony concentrations at room temperature were stable over three months and remained well below international maximum contaminant levels (MCLs), but that heating bottles to 50 degrees C produced concentrations exceeding the USEPA MCL within 24 hours.

Study design

Nine clear PET noncarbonated bottled water bottles (330 mL each) from three commercial brands were purchased in Kuwait City in summer 2018. The bottles were verified as PET by their global resin identification code.

Three experimental protocols were applied.

Storage duration at room temperature: one set of all three brands was held at 25 degrees C for three months to assess duration effects on antimony levels.

Temperature gradient at 24 hours: three sets were stored at -5 degrees C, 25 degrees C, and 50 degrees C for 24 hours to assess temperature effects. Bottles stored at 50 degrees C were held in a Heraeus oven model UT5200 (Kendro Lab Products GmbH, Germany).

Dual temperature-and-time effect: samples from one brand were held at 50 degrees C for 7 days to examine the combined effect of elevated temperature and prolonged exposure.

After incubation, water was transferred to 60 mL high-density polyethylene (HDPE) bottles and acidified with HNO3. Reverse osmosis water (Pure Aqua, CA, USA) containing less than 0.005 ppb antimony was used as the procedural control.

Instrument and methods

Antimony analysis was performed by ICP-MS using a Thermo Electron Element 2 single-collector double-focusing magnetic sector instrument. All reagents were of analytical grade; ultra-clean water was used throughout. Standard solutions were prepared using the method described in Chapa-Martinez et al. (2016). Antimony concentration was measured in triplicate for each brand.

LOD/LOQ: not reported in the paper.

CRM recovery: not reported in the paper.

Statistical analysis used a paired t-test to compare concentrations before and after the 3-month room-temperature holding period, and a three-factor ANOVA to compare concentrations across the three temperature conditions. Significance threshold: p less than 0.05. Analysis was performed in SPSS version 24.0.

Results

Antimony concentrations as purchased and during room-temperature storage

Antimony concentrations in the three brands ranged from 0.247 to 0.511 ppb across the measurement period (Figure 1 of the paper). The mean antimony concentration across brands was 0.390 plus or minus 0.133 ppb at the start and 0.384 plus or minus 0.124 ppb at the end of three months. Paired t-test showed no statistically significant change over the 3-month holding period at 25 degrees C.

Effect of temperature at 24 hours

Freezing at -5 degrees C produced no significant change in antimony concentrations. Heating to 50 degrees C increased concentrations from an initial value of 0.532 ppb to 8.530 ppb after 24 hours (Figure 2 of the paper). This exceeds the USEPA MCL of 6 ppb.

Combined temperature and time effect at 50 degrees C

Storing one brand at 50 degrees C for 7 days increased antimony concentrations to 16.8 ppb (Figure 3 of the paper). This exceeds the USEPA MCL more than twofold. The paper notes that after only 5 days at 50 degrees C, concentrations had exceeded twice the USEPA MCL of 6 ppb.

Key numbers

AnalyteMatrixValueUnitStatisticNSource location
SbPET bottled water, as purchased, range0.247-0.511ppb (ug/L)range3 brandsFigure 1; abstract
SbPET bottled water, mean at start of 3-month study0.390 +/- 0.133ppb (ug/L)mean +/- SD3 brandsResults
SbPET bottled water, mean at end of 3-month study0.384 +/- 0.124ppb (ug/L)mean +/- SD3 brandsResults
SbPET bottled water, initial value before heating0.532ppb (ug/L)value1 brandResults
SbPET bottled water, 50 degrees C for 24 hours8.530ppb (ug/L)value1 brandAbstract; Results; Figure 2
SbPET bottled water, 50 degrees C for 7 days16.8ppb (ug/L)value1 brandAbstract; Figure 3
Sbdrinking water MCL, USEPA / Health Canada / Ontario6ppb (ug/L)regulatory_limitn/aIntroduction
Sbdrinking water MCL, EU / Germany5ppb (ug/L)regulatory_limitn/aIntroduction
Sbdrinking water guideline, WHO20ppb (ug/L)guideline_valuen/aIntroduction
Sbdrinking water standard, Japan2ppb (ug/L)regulatory_limitn/aIntroduction
Sbprocedural blank (reverse osmosis control)less than 0.005ppb (ug/L)controln/aMethods

Regulatory context

The paper cites several antimony MCLs for drinking water. The USEPA, the Ontario Ministry of Environment, and Health Canada all set 6 ppb (mg/L) as the MCL. The German Federal Ministry of Environment and the European Union set 5 ppb. Japan’s standard is 2 ppb. The WHO guideline value is 20 ppb. PET polymerization uses antimony catalysts, primarily antimony trioxide, which accounts for more than 90 percent of global PET production; this is the principal source of antimony in PET bottled water.

Evidence fitness

This source supports: total Sb concentrations in commercial Kuwaiti PET bottled water under room-temperature storage; Sb migration at 50 degrees C for 24 hours and 7 days; the absence of statistically significant change at room temperature over three months; regulatory MCL values as cited by the paper.

It does not support: Sb(III) versus Sb(V) speciation; Sb content in the PET plastic itself (not measured); per-brand identities (brands are unnamed); baby-bottle or food-contact scenarios other than drinking water; a certification threshold.

Limitations

LOD and LOQ are not reported. CRM and spike-recovery figures are not reported. The paper references Chapa-Martinez et al. (2016) for the standard preparation method without fully describing it. The 7-day 50 degrees C test was conducted on only one brand. The number of bottles per condition is not always stated for the temperature experiments; the 3-month study used three bottles per brand. Instrument calibration details are not disclosed. The paper does not distinguish between Sb(III) and Sb(V). Figures rather than tables are used to present most results, so values are read from graphical representations for all but the abstract values and the initial value before heating. The paper’s title contains a typographic error in the published version (“De effect” for “The effect”).

Update history

No substantive edit history is available in this build. The full commit record is available in git.