Application Note No. 149/2015 Mineral oil contamination in dry food samples

Transcription

1 Application Note No. 149/2015 Mineral oil contamination in dry food samples SpeedExtractor E-916: Optimization of pressurized solvent extraction (PSE) for the rapid and efficient determination of MOSH and MOAH in dry food samples

2 1. Introduction Many food samples are contaminated with mineral oil from different sources. One possible source of contamination is the migration from cardboard packaging made of recycled fibres [1]. The main sources are the inks from newspapers entering the recycling process and the printed surfaces of the packaging [2, 3]. The extraction and analysis of cardboard samples used as food contact material are described in the Application Note No. 099/2013 Mineral oil contamination in cardboard [4]. Other sources of contamination with mineral oils can be related to the food production process. Food grade mineral oils are used in the food industry as lubricating, release agents, dust suppressants for grain or animal feed, protective coatings for raw fruits and vegetables or as ingredients [5]. Mineral oils are complex mixtures of hydrocarbons. Due to the high number of isomers present, it is not possible to separate individual hydrocarbons by GC and, for this reason GC traces are characterized by humps of unresolved peaks. Nevertheless, mineral oil contamination can be classified as mineral oil saturated hydrocarbons (MOSH), which are open chain paraffins, isoparaffins and cyclic naphthenes and mineral oil aromatic hydrocarbons (MOAH), comprising mainly alkylated 1-3 ring compounds. In humans exposed to MOSH, microgranulomas have been observed in the liver, spleen, lymph nodes and other organs, but these changes have not been associated with inflammatory reactions or other adverse consequences [5]. At the moment there are no legal limits for MOSH and MOAH migrated from packaging into food, however due to the potential carcinogenic effect of MOAH, the Federal Ministry of Food, Agriculture and Consumer Protection (Germany) is currently working on providing some [6]. According to the European Food Safety Authority [5], MOAH should be absent from food. Dry food samples were extracted by pressurized solvent extraction (PSE) using the SpeedExtractor E-916. The MOSH and MOAH fractions were analyzed by LC-GC. A clean-up step after extraction was not necessary. Using PSE is much faster than using classical solvent extraction [7]. This application note describes the extraction and determination of the MOSH and MOAH fractions from dry food samples, including PSE optimization and an application for migration studies. The work in this application note was performed by the group of Prof. Sabrina Moret from the University of Udine [8]. 2. Equipment SpeedExtractor E-916, with 10 ml cells, (BUCHI ) Mixer B-400, (BUCHI ) LC-GC 9000, Brechbühler 3. Chemicals and Materials Chemicals: n-hexane, CHROMASOLV, for HPLC, 97.0 % (GC), Sigma-Aldrich (34859), distilled before use. Ethanol, for HPLC, gradient grade, 99.8 % (GC), Sigma-Aldrich (02854). Acetone, CHROMASOLV, for HPLC, 99.9 % (GC), Sigma-Aldrich (270725), distilled before use. Quartz sand, particle size mm, (BUCHI ), pre-extracted with n-hexane at 100 C, 100 bar with a 5 min hold time. Application Note 149/2015 January /10

3 Mixture of internal/verification standards (Supelco), in toluene for quantification, to check for losses of volatiles and the efficiency of the MOSH and MOAH separation [9] MOSH: n-undecane (C11, 0.3 mg/ml): for detection of loss of volatile components during LC- GC transfer or reconcentration of the sample extract Cyclohexyl Cyclohexane (CyCy, 0.3 mg/ml): standard for quantification, not present in mineral oil products n-tridecane (C13, 0.15 mg/ml): proves no co-elution with CyCy; identification of peaks (peak pair) 5-α-Cholestane (CHO, 0.6 mg/ml): end of MOSH fraction MOAH: n-pentyl benzene (5B 0.3 mg/ml): for detection of loss of volatile components 1-Methylnaphthalene (1-MN, 0.3 mg/ml): standard for quantification; peak pair for easy identification 2-Methylnaphthalene (2-MN, 0.3 mg/ml): standard for quantification; peak pair for easy identification tri-tert-butyl-benzene (TBB, 0.3 mg/ml): start of MOAH fraction Perylene (Per, 0.6 mg/ml): end of MOAH fraction Samples: Dry food samples: pasta The samples were purchased at a local supermarket. 4. Procedure The extraction and analysis of MOSH and MOAH in dry food samples includes the following steps: Preparation of the samples Pre-cleaning Preparation of the cells Extraction with the SpeedExtractor E-916 using method A for superficial contaminations (migrated from the packaging material) and method B for total or deep contaminations (related to different sources): Method A applies to unground samples which allows for the extraction of contamination located on the surface of the sample, such as that migrated from recycled cardboard packaging in dry foods with a low fat content. Method B allows for the extraction of 1) total contamination (when directly applied to a ground sample), or 2) for the extraction of deep contamination (when applied to a sample which has already been subjected to method A to eliminate superficial contamination). MOSH and MOAH separation and quantification was performed using on-line LC-GC Preparation of the sample Superficial contamination related to contamination migrated from the packaging (method A) Directly load small size pasta into the cell. Break larger pasta samples into smaller pieces. Application Note 149/2015 January /10

4 Fig. 1: Different food samples with 10 ml extraction cells Total contamination related to different sources (method B) Homogenize the sample using the Mixer B-400 once for 2 seconds to obtain visual homogeneity 4.2. Pre-cleaning To prevent possible external contamination it is necessary to pre-clean the following parts: Bottom glass fiber filter (BUCHI ): heat for at least 8 h at 450 C Cells: rinse with n-hexane and dry with a hairdryer Metal frit (BUCHI ) and plug (BUCHI ): clean for 5 min in an ultrasonic bath using acetone and/or n-hexane Collection bottles (BUCHI ): rinse with acetone and/or n-hexane and dry with a hairdryer Top cellulose filter paper (BUCHI ): clean for 5 min in an ultrasonic bath using n-hexane 4.3. Preparation of the cells Place a glass fiber filter (BUCHI ) and a metal frit (BUCHI ) at the bottom of the extraction cell and close with a plug (BUCHI ) Turn the cell around and load the sample: Superficial contamination (method A) Load approx. 8 g of sample without dispersing agent into the cell Add 8 µl of the internal standard solution (1 µl/per g of sample) Close the cell with a top cellulose filter (BUCHI ) by pressing down with the plunger (BUCHI ) Total contamination (method B) Load approx. 5 g quartz sand into the cell Weigh approx. 2 g of homogenized sample using an analytical balance into a beaker Mix the sample with approx. 6 g of quartz sand Load the mixture into the 10 ml cell using a funnel (BUCHI ) Add 2 µl of the internal standard solution (1 µl/per g of sample) Fill any void volume with quartz sand Close the cell with a top cellulose filter (BUCHI ) by pressing down with the plunger (BUCHI ) Application Note 149/2015 January /10

5 4.4. Extraction with the SpeedExtractor E Superficial contamination (method A) To carry out the extraction use the parameters in Table 1 [10]. Table 1: Extraction method A for the SpeedExtractor E-916 for contaminations related to the packaging material Parameter Value Temperature 100 C Pressure 100 bar Solvent n-hexane 100 % Cells Vials Cycles 1 Heat-up Hold Discharge Flush with solvent Flush with gas Total extraction time 10 ml 60 ml 1 min 5 min 2 min 1 min 1 min 20 min To separate traces of water leave the extract to rest for 15 min or support the phase separation by centrifugation Deep contamination (method B) To carry out the extraction use the parameters in Table 2 [10]. Table 2: Extraction method B for the SpeedExtractor E-916 for contaminations related to different sources Parameter Value Temperature 100 C Pressure 100 bar Solvent n-hexane 50 % / Ethanol 50 % Cells Vials Cycles 2 Heat-up Hold Discharge Flush with solvent Flush with gas Total extraction time 10 ml 60 ml 1 min / 1 min 5 min / 5 min 2 min / 2 min 0 min 1 min 28 min Add approx. 30 ml water to the extract to separate the ethanol phase from the hexane phase. To enable phase separation cool the extract for 20 min at - 20 C. Application Note 149/2015 January /10

6 4.5. Quantification with LC-GC Table 3: LC-GC parameters LC-GC Column Gradient LC-GC transfer Columns Oven program Data handling LC-GC 9000, Brechbühler Trace GC Ultra, Thermo Scientific PAL LHS2-xt Combi PAL autosampler, CTC, Phoenix 40 three syringe LC pump, four switching valves UV/VIS, UV-2070 Plus detector, Jasco 25 cm x 2.1 mm i.d. packed with Lichrospher Si 60, 5 µm, DGB Start: n-hexane (0.1 min) reaching 30 % of dichloromethane (at 300 µl/min) in 0.5 min to elute MOSH (from 2.0 to 3.5 min) and MOAH (from 4 to 5.5 min) Retention gap technique and partially concurrent eluent evaporation through Y-interface. 10 m x 0.53 mm i.d. uncoated, deactivated precolumn; steel T-piece union connected to solvent vapor exit (SVE) 15 m x 0.25 mm i.d. coated with a 0.15 µm film of PS-255 (1% Vinyl, 99% Methyl Polysiloxane) Gradient: 40 C/min starting from 55 C up to 350 C for GC analysis. FID and SVE were heated at 360 C and 140 C, respectively. Exachrom software, Brechbühler MOSH area was determined by integrating the whole hump of largely unresolved peaks (subtracting the endogenous n-alkanes). All sharp peaks standing on top of the MOAH hump were subtracted from the total area (sharp signals are unlikely to belong to MOAH [11]) [1]. 5. Results Two PSE methods of different selectivity for MOSH and MOAH extraction from dry food with a low fat content have been developed and optimized (see chapter 7 Method development). An example showing the potential of the developed methods has been described. Method A and B were applied to a pasta sample with no superficial contamination before (T0) and after storage (1 month, T1) in direct contact with a recycled paperboard box [10]. The results are shown in Fig. 2. Fig. 2: On the left LG-GC chromatograms (MOSH) of pasta sample extracted with both method A (for superficial contamination) and method B (for total contamination), before (T0) and after the contact (T1) with recycled paperboard. On the right, LC-GC chromatograms obtained by applying method B after method A (to detect deep contamination). Figure 2 the chromatograms on the left show that the surface of the pasta sample at T0 is free of MOSH and MOAH. After one month in contact with the recycled paperboard, a large amount of mineral oil migrated into the dry food sample (and remained on the surface) [10]. The Application Note 149/2015 January /10

7 chromatograms in the center show that the pasta sample before contact (T0) was not free of MOSH contamination. A deep pre-existing contamination was present in the pasta at T0 [10]. The chromatograms on the right show a pasta sample which was extracted first with method A and then with method B. The sample which underwent PSE method A (to eliminate superficial contamination) was recovered from the extraction cell, and re-extracted (after grinding) using PSE method B. In this case, it was possible to discriminate between mineral oil that migrated from the packaging and deep pre-existing contamination. In conclusion, PSE method A was selective towards mineral oil that migrated from the packaging (superficial contamination), while method B enabled extraction of total contamination. By applying the methods in sequence (B after A), it was possible to find evidence of deep contamination. The two methods provided complementary data, useful for migration studies. Both extraction methods showed very good recoveries and repeatabilities, shown in Table 4. Table 4: Recoveries and repeatability of method A and B, n = 6 Method A Recovery (%) Repeatability (rsd %) Method B 6. Comparison to reference method According to the reference method [7] the samples were soaked in hot water (80 C), extracted first with ethanol and then with hexane overnight. Water was added to the combined extracts to separate the ethanol fraction [7]. The optimized PSE method allowed for complete extraction of mineral oil and showed results in agreement to those obtained with the reference method. The extraction using the SpeedExtractor E-916 takes 28 min. The PSE procedure using SpeedExtractor E-916 is much faster than the reference method. 7. Method development The following parameters were optimized during the method development: Solvent Temperature and extraction time Number of cycles 7.1. Method A: superficial contamination (related to migration from packaging) Solvent The solvent used (n-hexane) was the same for the extraction of MOSH and MOAH from cardboard samples [4]. Temperature and extraction time The samples were extracted at different temperatures (80 C, 100 C and 120 C) and times (5, 10 min). Fig. 3: Method A: Overlay of LC-GC chromatograms of MOSH and MOAH fractions at different temperatures: 80 C, 100 C and 120 C (1 cycle, 5 min) and at 100 C (1 cycle, 10 min) [8]. Application Note 149/2015 January /10

8 No appreciable differences in extraction efficiency were observed. The conditions chosen were: extraction at 100 C for 5 min. Number of cycles A single cycle extraction (5 min at 100 C) was not exhaustive, however, a flush with solvent was as effective as a second cycle. Therefore, an additional extraction cycle was not necessary [8]. Fig 4: Overlay of LC-GC chromatograms (MOSH) with one extraction cycle and for an additional cycle [8] Contamination related to different sources (method B): Temperature Three different temperatures were tested (100 C, 120 C and 150 C). Comparable extraction yields were obtained at 100 C and 120 C, while a slight decrease was observed at 150 C. Therefore, 100 C was used for all further extractions [10]. Solvent The following solvents were tested: n-hexane 100 %, n-hexane : acetone (50 % : 50 %) and n-hexane : ethanol (50 % : 50 %). hexane/ethanol (1:1) hexane acetone hexane/acetone (1:1) Fig.5: Method B: Overlay of LC-GC chromatograms (MOSH) of the same pasta samples with different solvents, using a 2 cycle extraction [7]. The mixture n-hexane : ethanol (50 % : 50 %) gave the highest extraction yields [10] min Number of cycles The recoveries with one extraction cycle were low, therefore an extraction with two cycles was performed. Fig. 6 shows that two extraction cycles gave quantitative recoveries. An additional third cycle was not necessary [10]. Application Note 149/2015 January /10

9 Fig. 6: Overlay of LC-GC chromatograms (MOSH) with one, two and three extraction cycles [7]. 8. Conclusion Extraction of mineral oil from dry food samples can be performed by PSE using the SpeedExtractor E-916. A following clean-up step is not necessary. The extracts of method A can be injected directly into the LC-GC system. The extracts of method B requires a phase separation before injection. The results obtained with the SpeedExtractor E-916 show good correlation with those obtained using the reference method, which is based on a hydration step followed by classical solvent extraction with ethanol followed by hexane [7]. Using the SpeedExtractor E-916 six samples can be extracted in parallel in 20 min and in 28 min respectively. Using PSE for total contamination is much faster than the reference method [7]. 9. Acknowledgment We greatly acknowledge Prof. Sabrina Moret, Dr. Marianna Scolaro, Dr. Laura Barp, Dr. Giorgia Purcaro and Prof. Lanfranco S. Conte from the University of Udine, Department of Food Science for their measurements, data and support in developing this application note. 10. References [1] Optimization of pressurized liquid extraction (PLE) for rapid determination of mineral oil saturated (MOSH) and aromatic hydrocarbons (MOAH) in cardboard and paper intended for food contact S. Moret, M. Sander, G. Purcaro, M. Scolaro, L. Barp, L. Conte, Talanta 115 (2013) [2] Rapid and sensitive solid phase extraction-large volume injection-gas chromatography for the analysis of mineral oil saturated and aromatic hydrocarbons in cardboard and dried food - S. Moret, L. Barp, G. Purcaro, L. Conte, Journal of Chromatography A, 1243 (2012) 1-5 [3] Is recycled newspaper suitable for food contact materials? Technical grade mineral oils from printing inks - M. Biedermann, K. Grob, Eur. Food Res. Technol. 230 (2010) [4] AN 099/2013 SpeedExtractor E-916: Optimization of pressurized solvent extraction (PSE) for the rapid and efficient determination of MOSH and MOAH in cardboard [5] Scientific Opinion on Mineral Oil Hydrocarbons in Food, EFSA Journal 2012, 10(6): [6] Fragen und Antworten zu Mineralölbestandteilen in Schokolade aus Adventskalendern und anderen Lebensmitteln, Bundesamt für Risikobewertung (BfR), 30th November 2012 [7] Removal of mineral oil migrated from paperboard packing during cooking of foods in boiling water S. Biedermann-Brem, K, Grob, Eur. Food Res. Technol (2011) 232: [8] Optimization of pressurized liquid extraction (PLE) for rapid and efficient determination of MOSH and MOAH in cardboard and foods - S. Moret; presentation at ExTech 2012 (Messina) [9] Maurus Biedermann Application Note 149/2015 January /10

10 [10] Optimization of pressurized liquid extraction (PLE) for rapid and efficient extraction of superficial and total mineral oil contamination from dry foods S. Moret, M. Scolaro, L. Barp. G. Purcaro, M. Sander, L. S. Conte, Food Chemistry 157 (2014) [11] On-line coupled high performance liquid chromatography-gas chromatography for the analysis of contamination by mineral oil. Part 2: Migrations from paperboard into dry foods: interpretation of chromatograms - M. Biedermann, K. Grob, J. Chromatogr. A 1255 (2012) Operation Manual SpeedExtractor E-916 / E-914 Operation Manual Mixer B-400 Application Note 149/2015 January /10