Page 8 - Shimadzu Journal vol.1

Fig. 2 GC×GC-MS chromatogram, relative to the pasta MOAH

fraction. Identification as reported in Table 2.
FA: fatty acid; DINP: diisopropylnaphthalenes

database LRI Fig. 3 GC×GC-MS chromatogram, relative to the icing sugar 1627 MOAH fraction. Identification as reported in Table 2. 1688 1715 1795 1828 1844 1846 1906 1925 1993 2024 2075 2158 2106 2124 2177 2188 2359 2375 2388 Fig. 4 GC×GC-MS chromatogram, relative to the rice MOAH 2475 fraction. Identification as reported in Table 2. 2505 2550 2710 2847 2877 2950

LRI rice ice sugar MS% MS% 1622 95 93 1667 94 93 1703 87 85 1798 82 - 1821 93 94 1846 - 92 1846 96 - 1896 95 - 1929 93 91 1992 94 83 2025 90 90 2085 83 81 2102 83 84 2108 92 - 2130 90 92 2175 - 91 2198 90 93 2302 - 85 2380 - 90 2395 87 88 2477 84 85 2504 83 84 2542 94 93 2697 92 - 2828 - 93 2884 92 - 2968 89 81
3-3. GC×GC-MS results for the MOAH fraction The peaks present in the GC×GC chromatograms, for the three samples, were tentatively-identified on the basis of MS database similarities (≥ 80%) and in accordance with linear retention indices (LRI), contained in the same database. Since a widely-accepted procedure for the calculation of GC×GC LRI values has not been developed, such data were calculated in a one-dimensional mode; furthermore, a rather wide LRI filter window (± 25 units) was applied (to eliminate wrong matches), to compensate for the retention effects of the polar capillary.The tentatively identified compounds, along with experimental and database LRI, are listed in Table 2. Two compounds

Food safety 1 Dioctylether 2 3 Ethyltetradecanoate 4 5 6 7 2-Heptadecanone 8 9 Ethylhexadecanoate 10 11 Abietatriene 12 Octyldodecanoate 13 2-Nonadecanone 14 15 Dodecyloctanoate 16 17 Octyltetradecanoate 18 Tetradecyloctanoate 19 20 21 Octylhexadecanoate 22 23 1-Tetracosanol 24 Squalene 25 1-Hexacosanol 26 27 8

Food safety 7

(p = 0.03) was established running a t-test, at the 5% significance
group. Specifically, for GC×GC-FID quantification, the “polygonal
Asix-point (each point was derived through duplicate applications)
Mandel’s fitting test (Fcalc<Ftab). The significance of the intercept
MOSH and the MOAH fractions, relative to pasta, icing sugar and
Kg range. The least squares method was exploited to estimate the
were evaluated through the regression coefficient (0.9993), and a
approximate estimation of the LoQ was made by considering the
regression line, while the linearity and the goodness of the curve
integration function” was applied, which enabled the definition
MOSH compounds, and the “unknown” peaks from the MOAH
and subtracted from the total area. Quantification information,
saved as well. Thus, the undesired peaks can be easily selected,
visual inspection of the residual plot, and were confirmed using
Measurement of the limit of quantification (LoQ), in mineral oil
rice, were quantified up to C25 (as required by the envisioned
calibration curve was constructed through the FID trace, using
limit), using the aforementioned method; attention was paid,
standard deviation (n=3), calculated at the lowest calibration
solutions of paraffin oil in n-hexane, in the 0.35-24 mg/
analyses, is tightly related to the MW distribution of the
contaminants, hence on the hump width. However an
point, multiplied by 10. The LoQ was estimated to be
min 30.000
MS (/FID) trace
MS-only trace
approximately 1.2 mg/Kg. 3-2. Food analysis during integration, to eliminate the natural alkanes from the of a polygonal area in which all the integrated peaks are automatically summed, and the data relative to each peak is relative to the three foods, is listed in Table 1. 29.975 29.950 MOAH< C25 (mg/Kg) 1.6 1.3 2.2
level. 29.925 Fig. 1 Comparison of raw TIC chromatogram expansions (printing ink analyses), obtained using a GC×GC-MS and a GC×GC-MS/FID system Table 1 Quantification values relative to the MOSH and the MOAH fractions, in samples of pasta, icing sugar, and rice, using Ag-SPE-GC×GC-MS/FID

29.900 MOSH <C25 (mg/Kg) 3.5 8.4 33.8

29.875 29.850 Food Pasta Icing sugar Rice

global w430×h280 Flow division between the FID and MS units was a compromise among different necessities, the main one being the attainment of a satisfactory sensitivity for quantification purposes. Because the detectors employed operate under different pressure conditions, the employment of two branches with equal IDs proved to be a non-ideal choice; the reason was related to the fact that an excessively long “MS” branch was required to generate an adequate flow resistance, to divert the majority of the effluent to the FID. Such a configuration would have led to substantial differences in the second-dimension elution times, between the qualitative and quantitative experiments. A good compromise was found through