106 Analyst, February, 1979, Vol. 104, p p . 106-110 Determination of Acrylonitrile Monomer in Plastic Packaging and Beverages by Headspace Gas Chromatography G. 9.-M. Gawell" National Food Administration, Food Research Department, Box 622, S-761 26 Uppsala, Sweden A gas - liquid chromatographic method for determining trace amounts of acrylonitrile in plastic containers and carbonated beverages using a nitrogen- sensitive detector and headspace injection technique is described. The method is suitable for the determination of acrylonitrile at concentrations down to 0.1 mg kg-l in plastics and 0.005 mg kg-l in beverages. Keywords A crylonitrile determination ; headsfiace gas chromatography ; piastic packaging; foodstufls Polymers with acrylonitrile monomer as a component are used for packages and household articles designed for foodstuffs.During the manufacture of acrylonitrile copolymers, a small fraction of unreacted acrylonitrile monomer becomes physically entrapped in the polymer and can migrate slowly during storage and when in contact with food or other materials. Recently, experiments with rats in the USA1 have indicated that acrylonitrile might be carcinogenic. This has led to a proposed ban in the USA on the use of acrylonitrile co- polymer materials for beverage packaging. These packages are also used in Sweden, and this work was initiated to develop a sensitive gas-chromatographic method for the determination of residual acrylonitrile in carbonated beverages and food packages. The methods for the determination of acrylonitrile in food and beverages published so far are often time consuming, require special laboratory equip- ment and are not very sensitive.Jones and Smith2 described a headspace gas-chromatographic method for acrylonitrile in fat with a sensitivity of 0.2 mg k g l . The Nederlandse Vereniging-federatie voor Kunst- stoffen3 has recommended a method that involves distillation into xylene followed by gas chromatography. This method gives a sensitivity of 0.4 mg k g l in aqueous simulants and about 1 mg k g l in fat. The Food and Drug Administration* has recommended a method for the determination of acrylonitrile in aqueous extracts, which involves an azeotropic distillation with methanol, followed by differential-pulse polarography. The sensitivity of this method is reported to be 0.01 mg k g l in water and 0.03 mg k g l in beer.The deter- mination of acrylonitrile in food processing plants using spectrophotometry involving photochemical bromination and the formation of a red pyridine - benzidine complex was described by Kroller.5 The detection limit was reported to be 0.01 mg kg-1. Several methods have been published on the determination of acrylonitrile in plastics,6-8 involving various solvents, gas-chromatographic columns and techniques, The method proposed here is applicable to plastics, carbonated beverages and simulating solvents. The headspace technique and the nitrogen-sensitive detector make the determination of acrylo- nitrile monomer rapid, accurate and sensitive (Fig. 1). Experimental Apparatus Normal laboratory equipment, 20-ml glass vials equipped with natural rubber seals and aluminium caps, a Fermpress H 207 for sealing the bottles and a l-ml gas-tight syringe were used.Reagents All reagents were of analytical-reagent grade. A cry lonitrile. * Present address : Astra Pharmaceuticals AB, S-161 86 Sodertalje, Sweden.GAWELL 107 The propylene carbonate was purified prior to use by the following The propylene carbonate was heated to 120 "C with magnetic stirring and Pvopionitrile. Profiylene carbonate. procedure. nitrogen bubbling for 12 h,* then stored under nitrogen. Gas Chromatography The gas chromatograph was a Varian 2700 with an alkali flame-ionisation detector (rubidium sulphate). Nitrogen was used as the carrier gas at a flow-rate of 23 ml min-1. Air and hydrogen flow-rates were optimised to give the best detector response.The chart speed of the 1-mV recorder was 0.8 cm min-l and the amplifier ranges were 4 x 10-l2 and 8 x 10-12AmV-l. The main column used was a 3 m x 2 mm i.d. glass column packed with 0.2:/, Carbowax 1500 on 60-SO-mesh Carbopack C. The injector temperature was 200 O C , the detector temperature 200 "C and the oven temperature 70 "C, isothermal. In order to confirm the presence of acrylonitrile in a sample, another glass column of 3 m x 2 mm i.d. was used, packed with ZOyo Carbowax 20M on Chromosorb W, 60-80 mesh. The instrumental con- ditions were the same except for the oven temperature, which was 50 "C, and the nitrogen flow-rate, which was 15 ml min-l. The retention times for acrylonitrile and propionitrile on the alternative column were 1.7 min and 2.3 min, respectively.Procedure Preparatiovt of samples of the carbonic acid had been evolved. pieces. Samples of carbonated beverages were shaken in a stoppered Erlenmeyer flask until most Samples of the plastic materials were cut into small Determination of acrylonitrile in beverages A 3-1111 volume of the homogenised sample was weighed into a glass vial, 2.0 ml of d 0 1 2 3 4 5 6 Ti me/m i n Fig. 1. Beer from a glass bottle spiked with 0.03 mg of acrylonitrile per kilo- gram. Propionitrile was added as internal standard. Attenuation 4 x 10-12 A mV-l. I R I I 1 1 I 0 . 1 2 3 4 5 6 Time/m i n Fig. 2. Beer sample from Barex bottle with propionitrile added as internal standard. Attenuation 8 x 10-l2 A mV-l.0 1 2 3 4 5 Time/m in Fig. 3. Plastic sample from Barex bottle with propio- nitrile added as internal standard. Attenuation 8 x 10-l2 A mV-l.108 GAWELL : DETERMINATION OF ACRYLONITRILE MONOMER IN PLASTIC Analyst, VoZ. 104 propionitrile (the internal standard), at a concentration of about 0.5 pg ml-l in distilled water, and 3.0 ml of distilled water were added and the bottle was closed with a rubber seal and cap. The sample was then ready for gas chromatography (Fig. 2). The vial was shaken well and placed in an oven at 90 "C for at least 30 min. Determination of acrylonitrile in plastic containers The plastic sample (0.3-0.5g) was weighed into a glass vial, 4.0ml of propionitrile in propylene carbonate (at a concentration of about 1.5 pgml-l) and 4.0ml of propylene carbonate were added and the bottle was then closed with a rubber seal and cap.The vial was placed in an oven at 90 "C until all the plastic had dissolved arid then left at room temperature overnight. Prior to the gas-chromatographic determination, the glass vial containing the polymer sample was placed in an oven at 90 "C for at least 30 min (Fig. 3). Gas-chromatographic determination (90 "C) gas-tight syringe, and injected into the chromatograph. A l-ml volume of gas was drawn from the gaseous phase in the glass vial using a warm Blank determinations The solutions and solvents were tested by headspace gas chromatography prior to use to ensure freedom from peaks that would interfere with the determination of acrylonitrile (Figs. 4, 5 and 6). Preparation of calibration graphs Acrylonitrile standard solutions of concentrations 0.01, 0.03, 0.05, 0.10 and 0.15 pg ml-l i I I I d 4 I 0 1 2 3 4 5 0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 Tirne/rn i n Ti me/m i n Tirnehin Time/rnin Fig.4. Beer Fig. 6 . Carbonated Fig. 6. Blankdeter- Fig. 7. Purified sample from a glass soft drink from a glass mination of purified propylene carbonate bottle. Attenuation bottle. Attenuation propylene carbonate. spiked with 0.12 pg 8 x 10-12 A mV-'. 8 x 10-12 A mV-l. Attenuation 8 x 10-12 of acrylonitrile per A mV-l. millilitre, correspond- ing t o 0.96mg kg-l in plastic. Propio- nitrile was added as internal standard. Attenuation 8 x 10-la A mV-l.February, 1979 PACKAGING AND BEVERAGES BY HEADSPACE GAS CHROMATOGRAPHY 109 were prepared in distilled water.A 3.0-ml volume of acrylonitrile standard solution, 2.0 ml of propionitrile solution in distilled water at a concentration of about 0.5 pg ml-l and 3.0 ml of liquid (similar to the sample, but which has never been in contact with any plastic, for example, beer from a glass bottle) were pipetted into a vial. The vial was then sealed, shaken and placed in an oven at 90 "C for at least 30 min. The chromatographic deter- mination was then carried out as described above (Fig. 1). Acrylonitrile standard solutions of concentrations 0.2, 0.4, 0.6, 0.8 and 1.0 pg ml-l were prepared in purified propylene carbonate. A 4.0-ml volume of acrylonitrile standard solution and 4.0ml of propionitrile solution in propylene carbonate, at a concentration of about 1.5 pg ml-1, were pipetted into a vial.The vial was then sealed, shaken and heated in the oven at 90 "C for 30 min. A 1-ml volume of the gaseous phase from the headspace was injected into the chromatograph as described above (Fig. 7). Although the peak heights in Fig. 7 are approximately the same as in Fig. 1, the amounts of acrylonitrile and propionitrile vary and proportionally represent substantially different amounts, because the solubilities of acsylonitrile and propionitrile in propylene carbonate are different from the solubilities in water. Duplicate determinations from each acrylonitrile standard solution were made in addition to duplicate injections from each vial. The concentrations of the standard solutions in the vials correspond to 0.01-0.15mgkg1 in the beverages and 2-10mgkg-l in the plastics.For each pair of injections of the standard solutions, the mean of the peak-height ratios for acrylonitrile and propionitrile were calculated (y) and plotted against the mass ratio of acrylonitrile to propionitrile of each standard solution in the same vial (x) in order to con- struct calibration graphs. Quanti$cation and plastics. The linear equation y = kx + I was used to calculate the acrylonitrile content in beverages (y - J)wis WSk Amount of acrylonitrile in sample (mg kg-l) = where y = mean of the two acrylonitrile to propionitrile peak-height ratios from the gas chromatograms of two headspace injections of the sample solution; I = intercept on the y-axis; Wi, = amount of internal standard added (pg); Ws = amount of sample weighed out (g); and k = slope of the straight line.Results and Discussion The proposed method was used to examine twelve samples of beer and four samples of carbonated soft drinks packed in Barex. The residual acrylonitrile in the packages ranged from 2 to 5 mg kg-1 and in some samples of beer and soft drinks trace amounts of acrylonitrile (<0.005 mg k g l ) were found (Fig. 2). The equations used to calculate the acrylonitrile content in beverages and plastics were y = 2.434% + 0.001 and y = 2.254~ - 0.017, respectively. The acrylonitrile standard solutions in distilled water remained stable for approximately 1 week in a refrigerator. The relative standard deviation calculated on ten determinations on Barex plastic from beer bottles was 2.9% with an acrylonitrile concentration of 2.8 mg k g l and 3.5% with an acrylonitrile concentration of 1.8 mg k g l .The method is also suitable for the determination of acrylonitrile in simulating solvents such as water, 3% acetic acid and 10% ethanol, which are often used in specific migration tests. The determination of acrylonitrile with this method is easy, rapid and, due to the use of the nitrogen-sensitive detector, very sensitive. The linear regression was 0.9999 for both calibration graphs. The accuracy of the determinations in plastics was tested. The author thanks Mrs. Margaretha Adolfsson-Erici for her skilful contribution to the experimental work, Miss Marie Kusters for assistance with the English translation and Mr. Bonny Larsson for valuable comments on the manuscript.110 GAWELL References Fd Chem. News, 1977, Jan. 24th, 20. Jones, N. X I . , and Smith, F. J., “Special Report on Monomer Migration from Lustropac 1010 ABS into Soft Margarine and Vegetable Oil,” Monsanto Petrochemicals and Polymers Company, Newport Research, St. Louis, Mo., 1974. Draft Packaging and Utensils Regulation (Food Law), Chapter 12, Section 3, “3.4. Specific Migration. Migration of Acrylonitrile and Methacrylonitrile. 3.4.1.1. Migration in Watery Simulants. 3.4.1.2. Migration in Fat,” Nederlandse Vereniging-federatie voor Kunststoffen , Haarlem, The Netherlands, 1975. Personal communication, Food and Drugs Administration, Washington, D.C., 1976. Kroller, E., Dt. LebensmittRdsch., 1970, 66, 11. Steichen, R. J., Analyt. Chern., 1976, 48, 1398. Bundesgesundheitsblatt, 1977, 20, 162. Personal communication, Lonza AG, Basle, Switzerland. 1. 2. 3. 4. 5. 6. 7. 8. Received August 7th, 1978 Accepted September 1 lth, 1978