Determination of siloxanes in silicone products and potential migration to milk, formula and liquid simulants

Abstract

A pressurised solvent extraction procedure coupled with a gas chromatography–mass spectrometry–selective ion monitoring (GC–MS–SIM) method was developed to determine three cyclic siloxanes, octamethylcyclotetrasiloxane (D4), decamethylcyclopentasiloxane (D5), dodecamethylcyclohexasiloxane (D6) and three linear siloxanes, octamethyltrisiloxane (L3), decamethyltetrasiloxane (L4), dodecamethylpentasiloxane (L5), in silicone products. Additionally, two different extraction methods were developed to measure these siloxanes migrating into milk, infant formula and liquid simulants (50 and 95% ethanol in water). The limits of quantification (LOQs) of the six siloxanes ranged from 6 ng/g (L3) to 15 ng/g (D6). Silicone nipples and silicone bakewares were extracted using pressurised solvent extraction (PSE) and analysed using the GC–MS–SIM method. No linear siloxanes were detected in the silicone nipple samples analysed. The three cyclic siloxanes (D4, D5 and D6) were detected in all silicone nipple samples with concentrations ranging from 0.5 to 269 µg/g. In the bakeware samples, except for L3, the other five siloxanes were detected with concentrations ranging from 0.2 µg/g (L4) to 7030 µg/g (D6). To investigate the potential migration of the six siloxanes from silicone nipples to milk and infant formula, a liquid extraction and dispersive clean-up procedure was developed for the two matrices. The procedure used a mix of hexane and ethyl acetate (1 : 1, v/v) as extraction solvent and C18 powder as the dispersive clean-up sorbent. For the liquid simulants, extraction of the siloxanes was achieved using hexane without any salting out or clean-up procedures. The recoveries of the six siloxanes from the milk, infant formula and simulants fortified at 50, 100, 200, 500 and 1000 µg/l ranged from 70 to 120% with a relative standard derivation (RSD) of less than 15% (n = 4). Migration tests were performed by exposing milk, infant formula and the liquid simulants to silicone baking sheets with known concentrations of the six siloxanes at 40°C. No siloxanes were detected in milk or infant formula after 6 h of direct contact with the silicone baking sheet plaques, indicating insignificant migration of the siloxanes to milk or infant formula. Migration tests in the two simulants lasted up to 72 h and the three cyclic siloxanes were detected in 50% ethanol after an 8-h exposure and after 2 h in 95% ethanol. The highest detected concentrations of D4, D5 and D6 were 42, 36 and 155 ng/ml, respectively, indicating very limited migration of D4, D5 or D6 into the two simulants.