Migration of Tinuvin P (UV stabiliser) and Irganox 3114 (antioxidant) from high-density polyethylene (HDPE) was studied. HDPE pieces were soaked in either milk (1.5% or 3.5% fat content) or 50% (v/v) ethanol-water mixture - the food simulant for milk as specified in Regulation No. 10/2011/EC. The obtained extracts were analysed by LC-MS/MS. For statistical assessment variography was used. It proved to be a useful tool for making a distinction between the early migration range and the equilibrium, despite the variance of the data. Regulation No. 10/2011/EC specifies 10 days of contact time for milk at 5°C. Our experiments with the food simulant with 24 dm(2) kg(-1) surface/mass ratio showed that both Tinuvin P and Irganox 3114 need less than 1 h to reach equilibrium. Furthermore, 10-day experiments with daily sampling showed that these additives are stable in milk, as well as in the food simulant. The effect of the concentration of the additives in HDPE was studied in the 0.01-5% (m/m) range. For both Tinuvin P and Irganox 3114 and all three extractants the migrated amount became independent of the concentration of the additive in the HDPE approximately at 1% (m/m). For Tinuvin P the food simulant gave a close estimate for the milk samples. However, using the food simulant for modelling the migration of Irganox 3114 into milk gave an overestimation with a factor of minimum 3.5. In the case of Tinuvin P special care must be taken, since the recommended amount in the HDPE can result in additive concentrations near or even over the specific migration limit (SML). However, Irganox 3114 cannot reach the SML either in milk or in the food simulant.

Retention in gas chromatographic systems has a central role in the identification of compounds even if detectors providing spectral information are used. But linear retention indices (LRI) of a single compound originating from multiple sources tend to vary greatly, probably due to differences in the experimental settings of the determinations. The effect of gas chromatographic parameters on LRI has been investigated using 41 compounds - previously identified from food contact plastics - and n-alkanes (n-C₇-n-C₄₀) used as reference series. As the reproducibility of LRIs under the same conditions is generally very good, the smallest changes in the settings often caused statistically significant, though irrelevant changes in the LRI values. Therefore, a multicriterial scoring-ranking system has been worked out to highlight the LRI value differences. Our results highlight that column length, heating rate, and film thickness can all be the reasons of the varying published LRI values. We also demonstrated that for the reproduction of LRI data, the chemistry (and not simply the polarity) of the stationary phase is crucial.

The coupling of large volume injection (LVI) with comprehensive two-dimensional gas chromatography (GC × GC) can be a powerful technique in the analysis of trace-level complex samples. The coupling of LVI and GC × GC using a cost efficiently operable pneumatic modulator based on capillary flow technology has been examined. The aim was to optimize the LVI parameters in the case of samples with compounds covering a wide boiling range. For the optimization of LVI 25 microliters of a solution containing 27 target compounds modelling the composition and the boiling range of diesel oils was used. The injection parameters were evaluated for peak shapes, reproducibility and peak volumes relative to peak volumes obtained using cold splitless injection. For all GC × GC experiments a non-polar first column (Rxi-5ms) and a polar second column (HP-INNOWax) were applied. Through extensive method optimization solvent vent proved to be an unsuitable technique for the injection of compounds covering a wide boiling range: at lower vent times peaks split, while higher vent times caused severe losses of highly volatile compounds. Therefore, a split-splitless LVI method was optimized. Injection speed, split vent time, splitless vent time and vent flow during split vent have been optimized. The developed method is suitable for the coupling of LVI with flow modulated GC × GC. Using the optimized split-splitless LVI parameters no peak distortion of the target compounds was observed. The relative peak volumes were between 60-120% for all compounds (80-120% for 13 compounds).