PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE WE DATE DUE [RTE DUE nr my 2000 JUIJI 3 230'] Jar-ta 1) 152003 1/98 chlHC/DfiDuoaGS-p.“ EVALUATING FUNCTIONAL BARRIERS TO PREVENT MIGRATION OF CONTAMINANTS FROM POST CONSUMER RECYCLED PLASTIC FABRICATED INTO SINGLE LAYER & LAMINATION FILMS By Weerasak Lertsiriyothin A THESIS Submit to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of Packaging 1997 ABSTRACT EVALUATING FUNCTIOAL BARRIER TO PREVENT MIGRATION OF CONTAMINANTS FROM POST CONSUMER RECYCLED PLASTIC FABRICATED INTO SINGLE LAYER AND LAMINATION FILMS By Weerasak Lertsiriyothin A study involving the potential migration of contaminant substances from a recycled high density polyethylene (HDPE) film into a selected food simulant was carried out. Isopropanol and tetracosane, which are chemical contaminants recommended by the FDA as representatives of a polar-volatile group and a nonpolar-nonvolatile group, respectively, were selected as the chemical surrogates. For isopropanol, the diffusion coefficient in HDPE film was determined by both isostatic and quasi-isostatic permeation methods with good agreement. The partition distribution of isopropanol between HDPE film and water was also determined from a sorption experiment. For tetracosane, the diffusion coefficient was determined by monitoring the migration of tetracosane from impregnated HDPE film, which was fabricated fi'om a tetracosane spiked HDPE resin by film extrusion, into hexane. The efl‘ectiveness of a functional barrier to prevent the migration of both compounds from a laminated film composed of a contaminated HDPE layer and a virgin HDPE layer into food was evaluated by two migration models. Modeling the migration of the two chemical surrogates from a single layer film was also carried out. A comparison of, and optimization of the respective models for estimation of migration rates was also concluded. To my dear Mom and Dad, big brother and sisters iii ACKNOWLEDGEMENTS First of all, I would like to thank my mom and dad for their love, encouragement and support which are always my motivation to do better things. My wonderful big brother and sisters, their dedication made my dream come true. For my best friend, Varoonvam Svangsopakul, who always understands what I have done and encourages me to make things happen. She also bring the cheerfulness back to my life and always teach me how to look on a bright side. Thank you for your love. For Prof. Jack R. Giacin, your guidance has been valuable things since we started this work. He is the most concerned and nice professor I ever met. He taught me how to think wisely and be patient . He so understood in every situations and gave a lot of help in the time I need. It would be better to say what a wonderful person he is. For Prof. Susan Selke and Prof. Kris Berglund, I appreciated all of your suggestions and comments throughout this work. I also would like to give a special thank to my fi'iends, Christopher Barr, Lee Youn Suk, Nade Riddle, Shu Jung Huang, Bob Htu'witz for their help. I also would like to say thank you to all of my friends whom I might forget to mention here. I would love to say thank you to the Thai government for the financial support. iv TABLE OF CONTENTS LIST OF TABLES .......................................................................... viii LIST OF FIGURES .......................................................................... xi INTRODUCTION ........................................................................... 1 LITERATURE REVIEW Recycled Plastics in Food Packaging Applications ............................ 4 Food and Polymeric Packaging Materials Interactions ........................ 6 Diffusion in Polymers ............................................................. 7 Solubility in Polymers ............................................................. 10 Permeation of Gases and Organic Vapor through Polymers ................. 11 Migration in Polymers ............................................................. 13 Role of Partition Coefficient, K in Migration Rate ............................ 17 Role of Sample Thickness in Migration Rate ................................... 17 Sorption in Polymers ............................................................... 20 Analytical Techniques for Determining Diffusion, Solubility and Permeability Coefficient Values .................................................. 23 Permeation measurement ................................................. 24 Isostatic or dynamic permeability method ..................... 24 Quasi-isostatic method ........................................... 26 Analytical Techniques for Sorption and Migration Measurements .......... 27 Thermal Stripper and Thermal Desorption ............................. 28 Solid phase microextraction ............................................. 29 Predictive Models for Migration from Lamination Film ..................... 31 The Laoubi and Vergnaud model ....................................... 32 The Begley and Hollifield model ....................................... 35 MATERIALS AND METHODS Materials ............................................................................ 36 Polymers ................................................................... 36 Solvents .................................................................... 36 Solutes ..................................................................... 37 Methods ............................................................................. 37 Determination of diffusion coefficient of Isopropanol in HDPE film ................................................................. 37 Isostatic technique ............................................... 37 Quasi-isostatic technique ....................................... 41 Determination of partition coefficient of Isopropanol in HDPE film and water ..................................................... 42 Analysis technique ............................................... 43 Migration study of Isopropanol from HDPE film into water ....... 46 Spiked film sample preparation ................................ 47 Analysis for migration study ................................... 47 Determination diffusion coefficient of Tetracosane in HDPE film .......................................................................... 49 Spiked film making ............................................. 49 Migration test ................................................... 50 Evaluating the migration amount of Isopropanol and Tetracosane from Single layer HDPE film and lamination HDPE film .......... 52 RESULTS AND DISCUSSIONS Permeation Measurements of Isopropanol through HDPE Film ............. 53 Difi‘usion coefficient of isopropanol in HDPE film .................. 53 Permeability coefficient of isopropanol through HDPE film ....... 55 Solubility coefficient of isopropanol in HDPE film ................. 56 Determination of the Equilibrium Partition Coefficient of Isopropanol between HDPE Film and Water ................................................. 60 Migration Studies of Isopropanol fi-om HDPE Film into Water ............. 61 Determination of Diffusion Coefficient of Tetracosane in HDPE Film ...... 64 Estimating Migration Levels of Isopropanol and Tetracosane from a Single Layer HDPE Film and a Lamination HDPE Film ................... 67 SUMMARY AND CONCLUSIONS ................................................... 89 RECOMMENDATIONS FOR FUTURE STUDIES ................................. 92 APPENDICES Appendix 1: Standard Calibration Curves ...................................... 93 Appendix II: Statistical Analysis for Permeation Experiments .............. 98 Appendix III: Migration Study of Isopropanol Spiked HDPE Film into Water by the SPME ......................................... 104 Appendix IV: Migration Study of Tetracosane Spiked Film into Hexane by the GC-FID .................................................... 105 Appendix V: The Comparison of the Predictive Migration Levels by Models H and III .................................................. 116 BIBLIOGRAPHY .......................................................................... 119 vii LIST OF TABLES Tables Pages 1. Diffusion, Permeability and Solubility Coefficient values for Isopropanol in HDPE film by the Isostatic technique ............................... 55 2. Diffusion, Permeability and Solubility Coefficient values for Isopropanol in HDPE film by the Quasi-isostatic technique ....................... 55 3. Equilibrium partition coefficient of Isopropanol between an aqueous phase and HDPE film by the sorption method at 23 °C .................................... 60 4. Migration of Isopropanol from spiked HDPE film into aqueous phase ( water) at 23 °C ......................................................................... 63 5. Initial amount of tetracosane in coated pellet and extruded film .................. 65 6. Partition coefficient and diffusion coefficient of tetracosane in HDPE/Hexane at 23 °C ................................................................. 67 7. Comparing difl‘usion coefficient of tetracosane in HDPE/Hexane to n-C18 and n-C32 in HDPE/corn oil .............................. 67 8. Comparison of the migrated amount of isopropanol from a single layer and a lamination film based on predictive Models 1, II and III ......................... 74 9. Comparison of the migrated amount of tetracosane from a single layer and a lamination film based on predictive Models 1, 11 ad III ........................... 75 10. The migration data of isopropanol for a lamination of 1.15 mil contaminated HDPE layer/ 1 mil HDPE barrier layer by Model 111 ............. 80 1’1 . The migration data of tetracosane for a lamma' tion of 1.15 mil contaminated HDPE layer/ 1 mil HDPE barrier layer by Model III ............. 80 12. Standard calibration data of isopropanol for HP 5890A set 1 .................... 93 13. Standard calibration of isopropanol for HP 5890A set 2 ......................... 94 14. Standard calibration of isopropanol for GC-MS .................................... 95 15. Standard Calibration of isopropanol for HP 6890 by SPME ...................... 96 16. Standard calibration of hexane for HP 5890A ...................................... 97 17. Statistical descriptive of diffusion coefficient determined by isostatic .......... 98 18. Statistical descriptive of permeability coefficient determined by isostatic ...... 98 19. Statistical descriptive of solubility coefficient determined by isostatic .......... 98 20. Statistical descriptive of diffusion coefficient determined by quasi-isostatic. . . 99 21. Statistical descriptive of permeability coefficient determined by quasi-isostatic ............................................................................ 99 22. Statistical descriptive of solubility coefficient determined by quasi-isostatic ............................................................................ 99 23. The ANOVA for diffusion coefficient determined by isostatic .................. 100 24. The AN OVA for permeability coefficient determined by isostatic .............. 100 25. The ANOVA for solubility coefficient determined by isostatic .................. 100 26. The ANOVA for diffusion coefficient determined by quasi-isostatic ........... 101 27. The AN OVA for permeability coefficient determined by quasi-isostatic ....... 101 28. The ANOVA for solubility coefficient determined by quasi-isostatic ........... 101 29. Two way AN OVA of test techniques and concentration effects for diffusion coefficient ............................................................................... 102 30. Two way ANOVA of test techniques and concentration effects for permeability oefficient ................................................................. 102 31. Two way ANOVA of test techniques and concentration effects for solubility coefficient ................................................................... 103 32. The comparison of migration of tetracosane between the experiment and the model fit for sample 1 ........................................................ 106 33. The comparison of migration of tetracosane between the experiment and the model fit for sample 2 ........................................................ 107 34. The comparison of migration of tetracosane between the experiment and the model fit for sample 3 ........................................................ 107 35. The comparison of migration of tetracosane between the experiment and the model fit for sample 4 ........................................................ 108 36. The comparison of migration of tetracosane between the experiment and the model fit for sample 5 ........................................................ 108 37. The comparison of migration of tetracosane between the experiment and the model fit for sample 6 ........................................................ 109 38. The migrated amount of isopropanol from a laminated film with the barrier thickness of 1,2 and 4 mil based on the predictive Models H and HI ............ 116 39. The migrated amount of isopropanol fiom a laminated film with the barrier thickness of 6,8 and 10 mil based on the predictive Models H and HI .......... 117 40. The migrated amount of tetracosane from a laminated film with the barrier thickness of 1,2 and 4 mil based on the predictive Models H and III ............ 117 41. The migrated amount of isopropanol from a laminated film with the barrier thickness of 6,8 and 10 mil based on the predictive Models H and HI .......... 118 LIST OF FIGURES Figures 1. Difi‘usion coefficient of Isopropanol in HDPE film at 23 °C by Quasi-isostatic and Isostatic permeation techniques ................................ 2. Permeability coefficient of Isopropanol through HDPE film at 23 °C by Quasi-isostatic and Isostatic permeation techniques ................................ 3. Solubility coefficient of Isopropanol in HDPE film at 23 °C by Quasi-isostatic and Isostatic permeation techniques ................................ 4. Migrated amount of Isopropanol from spiked HDPE film into aqueous phase (water) at 23 °C .......................................................................... 5. The profiles of the migrated amount of isopropanol from a single layer and a laminated film based on predictive Models 1, H and HI .......................... 6. The profiles of the migrated amount of tetracosane from a single layer and a laminated film based on predictive Models I, H and HI .......................... 7. The migration profile of isopropanol for a lamination of 1.15 mil contaminated HDPE layer/ 1 mil HDPE barrier layer by Model 111 .............. 8. The migration profile oftetracosane for a lamination of 1.15 mil contaminated HDPE layer/ 1 mil HDPE barrier layer by Model HI .............. 9. The migration profile of isopropanol for a lamination of 1.15 mil contaminated HDPE layer/ 1, 2, 4 mil HDPE barrier layer by Models H and IH .................................................................................. 10. The migration profile of isopropanol for a lamination of 1.15 mil contaminated HDPE layer/ 6,8,10 mil HDPE barrier layer by Models Hand HI ................................................................................. 11. The migration profile of tetracosane for a lamination of 1.15 mil contaminated HDPE layer/ 1, 2, 4 mil HDPE barrier layer by Models 11 and III ................................................................................. Pages 57 58 59 64 76 77 81 82 83 84 85 ' 12. The migration profile of tetracosane for a lamination of 1.15 mil contaminated HDPE layer/ 6, 8, 10 mil HDPE barrier layer by Models H and HI ............................................................................... 86 13. The relation of time to reach maximum allowable level as a frmction of a barrier thickness for isopropanol as initial concentration of 0.84ppm ......... 87 14. The relation of time to reach maximum allowable level a function of a barrier thickness for tetracosane as initial concentration of 0.84ppm .......... 88 15. Standard calibration curve of isopropanol for HP 5890A set 1 ................. 93 16. Standard calibration curve of isopropanol for HP 5890A set 2 ................. 94 17. Standard calibration curve of isopropanol for GC-MS ........................... 95 18. Standard calibration curve of isopropanol for HP 6890 by SPME .............. 96 19. Standard calibration curve of hexane for HP 5890A .............................. ' 97 20. A profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 1 .......................................... 110 21. A profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 2 .......................................... 111 22. A profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 3 .......................................... 112 23. A profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 4 .......................................... 113 24. A profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 5 ........................................... 114 25. A profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 6 ........................................... 115 xii INTRODUCTION The idea of recycling of materials came from our concern with the environment and problems faced in the handling of solid waste (209 million tons in 1994). Plastic packaging materials account for a large percentage of solid waste (19.8 million tons), and at present there is considerable interest in developing a plastics recycling process, especially intended for food packaging applications, where plastics have been significantly employed (EPA, 1994). For the purpose of using recycled plastics in food packaging applications, it must comply with the Food, Drug and Cosmetic Act and the Food Additive Regulations. In other words, it must be shown that the recycled plastics will not adulterate the food or otherwise compromise its safety. The level of 0.5 ppb (wt/wt), at which dietary exposures to substances of unknown toxicology can be considered a negligible risk, has been proposed by the Food and Drug Administration (FDA, 1992). To focus on the use and applicability of recycled plastics, the National Food Processors Association (NFPA) and the Society of Plastics Industry (SP1) have cooperated to develop “The Guidelines for the Safe Use of Recycled Plastics in Food Packaging Applications” (NFPA, SPI, 1994). In achieving this application, three critical aspects of recycled plastics were addressed, namely: (1) the some of recycled material; (2) the nature of the recycling process; and (3) the conditions of use which needed to be considered in order to meet FDA requirements. 2 A major concern of recycled plastics for food-contact use is remaining contaminants from post-consumer plastic. Unlike metal or glass, which are impervious to contaminants, polymeric materials have the potential to sorb contaminants into their structure, which are difficult to remove. As a result of this, development of analytical methods to evaluate the adequacy of a recycling process, or to measure trace amounts of remaining contaminants and potential contaminant migration into food, has been the focus of considerable interest. In addition to evaluating the recycling process, the study of migration can be employed to design a multilayer food package, which has recycled material as a non-food contact layer. Based on these considerations, the present study will focus attention on developing analytical techniques which will allow solution of mathematical expressions, which describe the transfer of contaminants from single and two layer laminate structures, together with migration and sorption studies involving selected migrant/polymer/food simulant systems. Here, two chemical surrogates, based on FDA suggestion, will be used to conduct experiments. One is isopropanol, a representative of a polar-volatile species, which is commonly found in paint strippers, household cleaners and windshield washing fluid etc. The other surrogate is tetracosane, a representative for nonpolar-nonvolatile species, which is an example of an intermediate molecular weight hydrocarbon. High density polyethylene, HDPE, is the polymer sample studied and the food sirnulants are water and hexane. The systems to be investigated included isopropanol/HDPE/water and tetracosane/HDPE/hexane. The objectives of the present study include: 3 1. To determine the diffusion coefficient of isopropanol (polar volatile) and tetracosane (nonpolar nonvolatile) in single layer HDPE film as a function of permeant concentration, by permeation and migration techniques, respectively. 2. Carry out actual migration studies on “doped isopropanol” single layer HDPE in contact with water. 3. Carry out sorption studies on virgin single layer HDPE in contact with spiked isopropanol in water as the food simulant, and determine the partition coefficient. 4. Based on the mathematical expression for migration fi'om single layer film, estimate the level of migration to a contact phase. 5. Evaluate mathematical models for a two layer laminate of recycled HDPE/ virgin HDPE, where the virgin HDPE layer is the food contact surface, and predict levels of migration to a food contact phase. A similar study involving two additional chemical smrogates; namely, xylene and 2,4 dichlorophenol, which are representatives of nonpolar-volatile and polar-non volatile substances respectively, has been reported by Shanna (1997). LITERATURE REVIEW Recycled Plasties in Food Packaging Applications In this decade ’90, attempts at increasing the amount of post-consumer plastics recycled around the world have continued to grow. For example, in the United States, recycling increased from about 20,000 tons in 1980 to about 680,000 tons in 1993 (Selke, 1996). The main reasons for the growth in plastic recycling are environmental concerns and consideration of handling municipal solid waste (MSW). Plastics were classified as the third largest contributor (19.8%, by volume, of MSW in 1994), but only 4.7% of plastic waste generation was recovered (EPA, 1994). Further, the development of plastic recycling processes which enhance the application of recycled plastic allow the use of tertiary recycled PET, produced via glycolysis and methanolysis processes, in food contact applications (Selke, 1996). Even though there is a Significant effort being directed to recycling plastic, the ratio of recovery to generated plastic waste is still quite low, compared to the ratio of paper and metal materials. To help accelerate the amount of recycled plastic, especially for packaging usage, the applications of recycled plastics need to be expanded. Among the wide range of recycled plastic applications, which vary depending on the nature of each resin type, use of recycled plastics in food packaging applications is one significant possible area. In dealing with food packaging applications, safety issues are of much more concern with plastics than with other packaging materials. This is due to the fact 5 that the post-consumer recycled plastics have a potential to be contaminated with toxicologically or health dangerous substances. Like any other material intended for use in food packaging, recycled plastics may only be used in a manner that complies with the Federal Food, Drug and Cosmetic Act and the Food Additive Regulations (21C.F.R. Part 170). In general, basic food packaging materials are regulated by FDA on a “ generic basis “, but there are no regulations that deal specifically with the use of recycled material for food packaging. However the existing regulations can work to ensure the safe use of recycled plastic in food packaging in two ways (NF PA, SP1, 1994), namely: 1. Any substance used as a component in a food package must be of a purity suitable for use in contact with food, such that it will not contaminate or adulterate food by, for example, imparting a taste or odor to food (as set forth in 21 CFR part 174.5), the so-called good manufacturing practices section of the regulations dealing with food contact materials ); and 2. Any package component made from recycled materials must comply with the food additive regulations applicable to that component, including any specifications for or limitations on the use of the material or any adjuvants in the polymer. With respect to public health safety concerns, the FDA proposed an upper limit dietary exposure level to chemical contaminants of 0.5 ppb (wt/wt)as the threshold of regulation for substances used as a food contact article (FDA, 1993). In other words, dietary exposures to contaminants from recycled food contact articles on the order of 0.5 ppb or less generally are negligible risk. 6 Food and Polymeric Packaging Materials Interactions Polymeric materials have been widely used for food packaging due to benefits derived from their chemical structures, physical properties and economical attraction. These various properties make plastic packaging an excellent protection for food against deterioration factors and for the preservation of food quality. However, material properties are not equivalent for all types of polymers and each type of polymer package is not appropriate for all types of food products. In order to develop an appropriate food packaging design, an indepth understanding of the interaction of food and polymeric packaging materials is required. There are three fundamental processes related to food packaging interactions, namely: (i) permeation, (ii) sorption, and (iii) migration. Permeation is the process by which molecules are transported from one side of the packaging material to the other as a result of sorption, diffusion and desorption processes. For example, gases and vapors in the external storage environment can permeate through the packaging material into the package internal environment and interact with the packaged food, resulting in die development of off-favor, and off-odor, as well as the deterioration of the food product. The permeation process can also occur in the opposite direction, resulting in the loss of water from the food product, as well as the loss of flavor volatile. Sorption is the process by which a low molecular weight organic moiety is taken up by one medium or phase, flour a second contacting phase, the process being driven by a difference in molecular concentration. Flavor scalping, or the sorption of food components by a packaging material, is one such example. Migration is a sorption process in which the transfer of substances proceeds fiom the packaging material to its contents. Such a case would be 7 the migration of residual monomer or contaminants fi'om post-consumer recycled plastic into foods. These three fundamental processes are the result of differential equilibrium between the polymeric packaging, food components or a surrounding environmental phase and introduces a phenomenon of molecular transport that tends to move to the equilibrium state. The molecular transport in polymers is controlled by the diffusivity and solubility of the penetrant molecules. Thus, the process of diffusion and solubility will be reviewed prior to discussing permeation, migration and sorption processes. Diffusion in Polymers Small molecules of gases, water vapor or organic vapors can diffuse through a polymer film or sheet to equilibrate their chemical potential or reduce thermodynamic forces. The phenomena of diffusion through polymeric packaging materials occurs via microvoids, which are referred to as the “free volume ” or “void volume” of the polymer matrix. The fiee volume or voids between polymer chains is very dependent on the chemical structure of the polymer and its morphology, which result in differences in polymer chain segmental mobility. The concept of molecular transport regulated by free volume was first introduced by Cohen and Tumball (1959). The development of a number of models based on the free volume concept, such as the theory of Vrentas and Duda (Duda and Zielinski, 1996), have been described in the literature (Kumins et al., 1968, Rudnick et al., 1979). 8 The diffusion behavior of polymers can be broadly described as “F ickian” or “Non-Fickian”. In general, diffusion in rubbery polymers follows F ickian behavior, while glassy polymers Show anomalous or non-Fickian behavior (Crank 1975). For “F ickian” behavior, F ick’s first and second laws of diffusion accurately describe the diffusion process. Fick’s first law assumes that the rate of transfer of the diffusing substance through a unit area of a section is proportional to the concentration gradient measured normal to the section. For an isotropic media (structure and diffusion properties at any point are the same relative to all directions), Fick’s first law is described by the expression: F = -D-——- (1) where F is the rate of transfer per unit area of section, C is the concentration of diffusing substance, x is the space coordinate measured normal to the section and D is the diffusion coefficient. Based on the concept of the conservation of substances, the total concentration change across the slab bulk phase with time is directly preportional to the change in concentration gradient with permeant permeation depth (Neogi 1996). For an isotropic media, therefore: 6C 0". 9: = “9:1; (2) where C is the concentration of diffusing substance, t is the time and j, is the flux in x direction, which is along the thickness of a membrane. From Fick’s first and second law and under the boundary condition that the solute is dilute or, when the vohrme averaged reference velocity is being used or assumed, the flux in x direction can be written as: . 5C .1: — —D fix (3) 5C 5 5C Thus, ; = 5:;[DE] (4) Equation 4 is therefore the mathematical expression describing Fick’s second law of diffusion. When this differential equation is solved under appropriate initial and boundary conditions, for related experimental theory, it describes the. measured quantities in terms of diffusivity D, where D can be calculated. Crank (1975) solved and described D in the form of mathematical solutions for a number of experimental situations. l 0 Solubility in Polymers The solubility coefficient, S, is a parameter that describes the equilibrium amount of substances sorbed by the polymer per unit sorbate partial pressure. For organic vapors, the solubility coefficient is normally presented in weight per weight of polymer in equilibrium per unit vapor pressure. For low partial pressures of gasses, or for dilute concentrations of organic vapors, the solubility follows Henry’s law; C = Sp (5) where; C = the concentration of sorbate in polymer S = the solubility coefficient p = the vapor pressure ofsorbate In a thermodynamic sense, the solubility coefficient can be described in terms of the. heat of solution for the specific sorbate / polymer system and temperature by equation 6. sm-s 5‘1] 6 _ 08X RT () where AH, is the heat of sohrtion, T is temperature in Kelvin degrees, and So is a constant which can be derived by measurement of 5(7) at different temperatures. The heat of solution, for dissolved gases above their critical point, can be calculated fiom: 11‘ AH. =V.(> K15 Vf), the amount of analyte sorbed by the coating at equilibrium is independent of sample volume, thus: n = K,V,C, (29) With proper calibration, SPME can be used to determine the concentration of the analyte in the sample. The speed of extraction is controlled by the mass transfer of analyte from the sample to the coating. Since sorbate diffusion governs the mass transfer rate in polymers for most gaseous samples, extraction equilibrium will be reached in less than 1 min, since gases usually have large diffusion coefficient values as compared to coating materials. While for an aqueous sample, the equilibrium time is much longer, since the analytes must diffuse through the aqueous phase, and specifically the static layer of water adjacent to the coating before they can be sorbed by the coating. If SPME needs to be directly applied to extract analytes from an aqueous phase, vigorous agitation of the sample, such as sonication, may improve the equilibrium time. The SPME device consists of a 1 cm. length of fused silica fiber, coated onto the outer surface of the stationary phase and bonded to a stainless steel plunger. The firsed silica fiber can be drawn into a hollow needle, which frmctions to pierce the sampling septum of a GC by using a fiber holder. This needle liked device allows the exposed 31 fiber to be easily introduced into the injection port of the GC where the sorbed analytes are thermally desorbed and delivered to the GC column. Presently, there are several commercial coating phases that are claimed to work for compounds ranging from non-polar to polar, such as polydimethysiloxane for non- polar compounds, polydimethylsiloxane/divinylbenzene polymer for polar volatile compounds, and polyacrylate for polar semivolatile compounds. Boyd-Boland et a1. (1995) applied polyacrylate SPME to collect nitrogen- containing herbicides in water samples. Shirey (1997) compared the extraction limits of different coated fiber phases for polar analytes in water samples. The results show that carboxen/polydimethylsiloxane gave superior extraction capability as compared to other coating phases, even a more polar phase such as carboxen/divinylbenzene and polyacrylate fiber, because small pores of the carboxen/polydimethylsiloxane coating, optimized the extraction capacity of the coated fiber phase. Predictive Models for Migration from Lamination Film The use of a functional barrier layer, i.e. virgin film layer, to prevent the migration of contaminants from a lamination incorporating recycled film has a potential for food packaging, as long as it is proven that no significant level of contaminants migrate into the food contents. Presently, the use of recycled HDPE resin has been considered by the FDA for application in food packaging, specially for consideration of its compliance with all related regulations for food safety. Due to the fact that migration studies may require a significant time frame and budget to complete, attempts at developing a method for estimating migration rates have been made, based on solution of two basic differential 32 equations describing diffusion in a polymer matrix (Fick’s first and second law). Even though migration models may not be fitted for all food-polymer systems, they can provide an estimation for the migration from representative food packaging systems and a worse case level of migration, which can be employed as a safety factor in film production design. Two predictive models describing migration from laminate films are presented below. 1113 Laoubi and Vemgud model The solution of migrant transfer from film into a food contact phase, described by the Laoubi and Vergnaud model (Laoubi et.al., 1995), was based on the diffusion process for a laminate film containing a recycled layer and a virgin polymer layer and was derived fiom Fick’s second law (eqn 4) under initial and boundary condition corresponding to the following assumptions: 1. The migrant transfer process is controlled by F ickian diffusion with a constant diffusion coefficient. 2. The migrant is initially in the recycled layer and only migrates through the virgin layer into the food contact phase. 3. There is no transfer of food components into the virgin polymer (food contact phase) and no effect of film thickness associated with migrant transfer, due to the very low concentration of migrant. 4. The concentration of migrant in food contacting the virgin polymer layer remains negligible by assuming a large volume of surrounding food contact phase. 33 For a laminate film (thickness = L) comprised of a recycled layer (thickness = H) and a virgin layer (thickness = L-H) and given the above assumptions, the initial and boundary conditions can be described in mathematical equations as: Initial conditions: t = 0 , at 0 < x < H C = C," (30) H 0 , at x = 0 2;; = 0 (32) 5 C x=L —D-0;;;-=h(CL—Cm) (33) where h is the convective coefficient of mass transfer ifh—roo x=L CL=Cw (34) and for all cases Cm = 0 (35) Using the separation of variables method, the authors solved the expression: 5ij 0" 6C” é‘t - 0"): [D 6x ] (36) The solution for determining the concentration of migrant at position x and time t, C,“ is presented below as a fraction of the initial concentration in the recycled layer : 34 C” 1% in(but-1):: H cos(27H- l)7r x ex (2n + 027:2 Dr] 0,, z: ,2, (2 + 1)sm 2L 2L 4L2 (37) By integrating with respect to any position between 0 and L, the solution for the amount of migrant remaining in the film at time t, M,,, is expressed as a fraction of the initial amount of migrant in the recycled layer, Mi" : M,, _ LZ_______ (-1)" s1n(2"+l)” Hexp[— (2n+1) 7:2 0t] 38 M, ”HH,,,(2,,+1) 2L 4!.2 ( ) By integrating with respect to position between 0 and H, the solution for the amount of migrant m in the barrier layer at time t, M5,, is expressed as a fraction of the initial amount of migrant in the recycled layer, M, : M, __ Mi sm2(2n+1)7rH x (2n+1)7r2Dt] (39) Min n=0(2n+1)zs 2L 6 . 4L2 The amount of migrant transferred into food could be calculated from: M. = Mm - M” (40) 35 The Begley and Hollifield model Begley and Hollifield (1993) applied a solution, which was developed by Crank (1975) for a uniform initial distribution and different constant surface concentration at both sides, to estimate the amount of migrant migrating fi'om a recycled layer through a virgin layer into a food contact phase. The assumptions make by Begley and Hollifield are as follows: 1. The migrant transfer process is controlled by Fickian diffusion with a constant diffusion coefficient. 2. The initial migrant concentration is an infinite source contained only in the recycled layer. In this case, at the surface of the virgin layer adjacent to the recycled layer side, the concentration of migrant remains constant, while the migrant concentration on the surface contacting to the food is always zero. 3. The food is an instantaneous and infinite sink for the contaminant. For a laminate film made up of a recycled layer (negligible thickness) and a virgin layer (thickness = I), the total amount of migrant, M, transferring through the membrane at time t can be calculated fi'om the equation : tn _ n f 2 2 £2-2£_l_%2( 1) -D" ’7’) (41) 11:1 n K 12 Where C1 is a constant concentration of migrant at the surface of the virgin layer adjacent to the recycled layer, t is time and D is the diffusion coefficient. MATERIALS AND METHODS Materials Polymers Resin and film samples of high density polyethylene (HDPE) were provided by Tredegar, Inc (T erre Haute, Indiana, USA). The HDPE film, Monax®, was used for permeability, sorption and migration studies with isopropanol as a penetrant, while the HDPE resin samples were spiked with tetracosane and extruded as film to study the migration propensity of tetracosane. Monax film: Lot # TI-49611, size 43 in. x l m., thickness 1.15 mil, and density 0.96 g/ cc. The resin pellets were the resin grade used to produce Monax film. SELL/ELIE Acetonitrile CH3CN (HPLC grade) FW 41.05, specific gravity 0.783, bp 81.6 ‘C‘, mp -41 °C from EM Industries, Inc (Gibbstown, NJ, USA). Hexane CH3(CH2)4CH3 (96 %) F W 86.18, specific gravity 0.659, bp 69 °C, mp - 94 °C and Water H2O (HPLC grade) FW 18 specific gravity 1, bp 100 °C, mp 0 °C from J.T. Baker (Phillipsburg, NJ, USA). 36 37 015 2-propanol (CH3)2CHOH (99.5 %) FW 60.1 specific gravity 0.785, bp 82.4 °C, mp -89.5 °C Tetracosane CH3(CH2)22CH3 (99 %) FW 338.66 specific gravity 0.779, bp391 °C, mp 49-52 °C. Both from Aldrich Chemical Company, Inc (Milwaukee, WI, USA). Methods Det§_rrn_1na_' tion of diffusion coefficient of Isopropanol in HDPE film The diffusion coefficient of isopropanol in HDPE film was determined by the permeability method, which was based on two different procedures, namely the isostatic and quasi-isostatic (Hernandez, et.al., 1986) techniques. The methodology involved with the respective procedures is presented below. Isostatic technique A commercial instrument, the MAS 2000TM organic vapor permeability tester from Testing Machines Inc., (Amityville, NY, USA) was used for measuring the permeability of isopropanol vapor through I-H)PE film. The MAS 2000 is designed for continuous flow of organic vapor through one permeability cell chamber and the permeated vapor level monitored by N2 carrier gas continually conveying the vapor from the low concentration cell chamber to a flame ionization detector (F ID) for quantification. The test system is equipped with a flow controller, temperature controller and a flame 38 ionization detector, all of which are connected to and controlled by computer. The instrument software is not only designed for controlling the parameters of the system, but also for determining the permeability, diffirsion and solubility coefficient values of the test penetrant/polymer system from permeation rate data. The precision of flow control is within :1: 0.1 ml/min and temperature control is within :1: 0.05 ‘C. The instrrunent sensitivity is at 1 ppb (volume by volume) or 120 picogram/mz/sec. The isopropanol vapor was generated by bubbling N2 gas through liquid isopropanol contained within a Pyrex brand gas-washing bottle with fiitted cylinder and 29/42 stopper, 250 ml (Aldrich Chemical Company, Inc., Milwaukee, WI, USA). The gas washing bottle was placed in a water bath maintained at 23 :1: 0.05 °C (MW 1110, Blue M Electric Co., Chicago, Illinois, USA) and the vapor activity or concentration of isopropanol was set up by mixing the saturated vapor stream fiom the gas washing bottle with pure N2 gas. Both N2 strearrrs were generated from the same gas cylinder with the regulator pressure set at about 20 psi before it was delivered to the MAS 2000 unit. The outlet N2 from the MAS 2000 was separated into two gas streams for bubbling and mixing. Each N2 stream flow was controlled by a needle valve (Nupro “M” and “S” series, Swaglok Co., Detroit, MI, USA) with rotameters indicating the flow rate before it was delivered into the gas washing bottle and was mixed with a pure vapor stream. The concentration of the vapor stream was determined by gas chromatography (GC) analysis. A 100 pl sample of the vapor stream was injected directly into the GC and the detector response from the GC (area unit) was used to determine the permeant vapor pressure by the following equation: 39 AxRxT poFwaxMW (42) where p = vapor pressure of permeant at T [mmHg] A = area response [A U] R = gas constant (62.36) [1. mmHg/moI.K] T = temperature of vapor stream [K] CF == calibration factor [A U/g] V2,.) = volume injection [1] MW = molecular weight of permeant [g/mol] a = £30— (43) where a = vapor activity of permeant p = vapor pressure of permeant at T [mmHg] pa = saturated vapor pressure of permeant at T [mmHg] Standard solutions of isopropanol were prepared by a serial dilution of isopropanol in acetronitrile from 1000 ppm to 20 ppm (volume by volume). A 1 ,ul sample of each concentration was withdrawn by a 10 pl liquid syringe (801N, Hamilton liquid syringe) and manually injected into the GC. A Hewlett Packard model 5890A gas chromatograph with flame ionization detector interfaced to an integrator, PIP-3395, was used (Avondale, PA, USA). The calibration factor was obtained fi'om the slope of the standard calibration curve, which was a plot of area response as a function of sample 40 weight injected (for each concentration). The calibration curve and calibration factor are presented in Appendix I. The column for this separation was Supelcowax 10 (60 m x 0.25 mm, 0.25 ,u m, Supelco, Bellefonte, PA, USA). The gas chromatographic conditions of separation were as follows: Mode: splitlcss Oven: 110 0C (10 min) to 220 °C at 5 °C/min and hold for 5 min Carrier: helium 1 ml/min Injection temp.: 220 °C Detector temp.: 250 °C Gases: He 40 psi (1 ml/min), H2 19 psi (20 ml/min) , air 35 psi (300 ml/min), N2 30 psi (30 ml/min) Elution time: 5.60 min The saturated vapor pressure of isopropanol at room temperature was determined by injection of a 100 pl headspace sample flom a sample vial (30 m1, crimp seal, Supelco, Inc, Bellefonte, PA, USA) to which was added 3 ml liquid isopropanol and the sealed vial stored in a constant temperature chamber (23 :1: l f). Samples of film were prepared by cutting test samples of about 13 cm x 13 cm and mounting the test film on a flame which was designed to have an exposed surface area of 0.0081 m2 . This was the area assumed for permeation determination. For calibration of the MAS 2000 test system, a clean surface aluminum foil sample was mounted on the mounting flame and placed between the front and back cell chambers. A 100 pl sample of a preset vapor stream was injected directly into the sampling port on the system chassis and the injected vapor sample conveyed by carrier gas directly to the FID detector for 41 analysis. Afler the instrument was calibrated, permeability tests were performed on a series of HDPE film samples. During a test run, the amount of isopropanol which had permeated was measured until a steady state rate was attained. Several vapor activity levels (0.1-0.4) were tested to evaluate the effect of penetrant concentration on the permeability, diffusion and solubility coefficient of the isopropanol / HDPE system. Quasi-isostatic technique The quasi-isostatic permeation runs were carried out with a system similar to that described by Hernandez, et a1. (1986). The test system consists of a gas supply (N2), a vapor generator system, flow controllers and a permeability cell. The vapor generator system was similar to that described aboves. The N2 from a cylinder was split into two streams by a T-value and connected directly to flow controllers, one for the gas washing bottle and the other for mixing of pure N2 with the vapor outlet stream fl'om the gas washing bottle. The permeability cell is made of stainless steel and consists of three cylindrical shape parts; a bottom and top cell chamber with an inside volume of 50 cc and a hollow center ring mounted between the two cell chambers, with a separation volume of 50 cc. This permeability cell creates an exposed surface area of the test film of about 50 cm). Sampling ports are affixed to each cell chamber and vapor inlet and outlet ports are configured to the center ring. For permeation measurements, a film sample was cut and mounted between the top cell chamber and the center ring, while the bottom chamber and center ring was isolated by an aluminum foil. The vapor stream was allowed to continuous flow through the center cell chamber during test and the quantity of 42 isopropanol permeated to the low concentration cell chamber was determined at predetermined time intervals by a GC procedure. For this analysis, a 100 pl sample was withdrawn fl'om the low concentration cell chamber (1750 , Hamilton 500 pl gas tight syringe). The conditions of GC separation and the standard calibration factor were the same as previously described for the isostatic technique. Between each sample test, the permeability cell was washed and dried in a vacuum oven. For vapor activity selection, the center cell chamber was isolated flom the top and bottom cell chambers with the aluminum foil and the organic vapor stream was flowed through the center cell chamber or high concentration cell chamber. A 100 pl headspace of sample was removed fl'om the center cell chamber with a 500 pl gas tight syringe (1750, Hamilton 500 pl gas tight syringe) and the vapor concentration was determined by GC analysis. The gas flew through the vapor generator and diluant gas were then adjust to provide the desired vapor pressure. Determination of partition coefficient of Isopropanol in @PE film and water The equilibrium partition coefficient of isopropanol between HDPE film and an aqueous solution was determined by a sorption technique. The sorption apparatus was designed based on the description presented in ASTM D4754-87, a standard test method for two-sided liquid extraction of plastic materials using the FDA migration cell. Sorption was monitored flour a a low concentration solution of ispropanol in water (100 ppm , volume by volume). Test specimens were prepared fl'om HDPE film in the form of round disks (OD 21.69 mm) by cutting with a cork borer. A stainless steel wire was formed as a support 43 stand, next 30 film disks were weighed and alternately threaded with glass beads onto the stainless wire, thus creating two sided contact. Each set of mounted disks were then placed into 40 ml precleaned amber vials, filled with 40 ml sorption solution (100 ppm , v/v isopropanol in water) and the vials closed with open-top screw cap-teflon / silicon septa (vials and caps fi'om Supelco, Inc, Bellefonte, PA, USA). Four sets of sorption cells and one blank cell (no film disks) were prepared and stored in a constant temperature chamber (23 i 1 ‘C) for a period of 60 days to ensure equilibrium state. This storage time was estimated flom a survey of equilibrirun sorption times reported by Bauer (1993), who reported that equilibrium sorption times for aqueous solutions of a variety of aromas into HDPE and polypropylene (PP) films were reached in about 20 days at 20 ”C. During the course of the study, each of the sorption cells was shaken by hand every seven days. At the end of the storage time, the quantity of isopropanol remaining in solution and the concentration sorbed by the HDPE film disks were measured by application of a two step process, namely: sample preparation by a thermal stripper (TS) step and analysis with interfaced thermal desorption-gas chromatography-mass spectrometry (TD-GC-MS). Analysis technique The thermal stripper and thermal desorption-GC-MS analytical techniques eliminated solvent extraction, which avoids loss of isopropanol during transfer and allowed for sufficient sensitivity for detecting very low levels of isopropanol sorbed by the HDPE film. 44 For the analytical procedure, the instrument parameters were optimized and a standard calibration curve for the TS and TD-GC—MS procedures was constructed by analysis of a series of standard solutions of known concentration. The optimized conditions for the TS and TD-GC-MS procedures are presented below. Thermal Stripper Instrument: Sorbent material: Carrier gas: Flow rate: Temperature: Thermal Desorption Instrument: Carrier gas: Flow rate: Temperature: Model 1000 (Dynatherm, Kelton, PA, USA) Carbotrap 302 with glass flit at sample inlet ID 4 mm (Supelco Inc, Bellefonte, PA, USA) He 50 psi Preheat state 50 ml/min for 5 min Purge state 150 ml/min for 2 min Dry state turn off Block 199 T, oven 110 °C, tube 75 T Model 890 (Dynatherm, Kelton, PA, USA) He 60 psi Desorption at 7 ml/min for 6 min Clean at 35 mein for 30 min Desorption 250 T, clean 280 T, transfer line 230 T. Gas Chromatography and Mass Spectrometer Instrument: GC is a HP 5890A 45 Detector: MS 5970 with Chemstation analysis sofiware (Hewlett Packard, Avondale, PA, USA) Column: Supelcowax 10 (60 m x 0.25 mm, 0.25 p m) Carrier gas; He Mode: SIM (Selected Ion Monitoring) Selection m/z = 27, 31, 43, 59 Oven: 40 °C (5 min) to 200 at 5 T/min and hold for 10 min Elution time: 12.30 min The TD unit was interfaced to the GC-MS unit. Thus, samples were thermally desorbed from the sorbent tube and delivered directly to the GC for separation and detected by MS. The calibration curve and calibration factor are presented in Appendix 1. Following a predetermined storage (equilibrium) time, the amount of isopropanol sorbed by the film disks (polymer phase) and that remaining in solution (liquid phase) were determined. Quantification of isopropanol in the liquid phase of the sorption cell and the analysis of standard solutions of known concentration was carried out by the same procedure. First, the TS unit was preheated until the oven temperature reached 110 T, next a precleaned sorbent tube was affixed to the sample receiver port. A 3 pl of liquid sample was then withdrawn and injected into the side port of a 10 ml hang down tube mounted in the oven of the TD unit. The isopropanol concentration in the liquid sample was then thermally transferred to the sorbent tube with a He flow of 150 ml/min for 2 min . The sorbent tube was then removed and transferred to a glass storage tube with screw cap closure. The sample sorbed by the sorbent tube was immediately analyzed by the TD-GC-MS procem. 45 Desorption conditions were identical to those described above. The total sorbed amount of isopropanol was conveyed to the GC column which allowed for separation and quantification by MS operated in the SIM mode. The quantity of isopropanol in the liquid phase was further calculated in the form of concentration (wt/wt). For determining the amount of isopropanol sorbed by the HDPE film disks at equilibrium, the stack of film disks was removed flom the sorption cell and the film disks were collected and rapidly wiped flee of remaining solution on the surface to reduce loss of isopropanol due to volatilization. Next, all of the dry film disks (about 0.35 g) were transferred to the sample hangdown tube of the TS unit and the analysis performed as described for the liquid phase. Again, the quantity of isopropanol in the polymer phase was further described in term of concentration (wt/wt). Finally, the equilibrium partition coefficient of isopropanol between HDPE film and water was calculated by taking a ratio of the equilibrium concentration of isopropanol in the water and film, respectively. Mi tion f I ro ol flo HDPE film into wat A preliminary experiment for the migration of isopropanol from film into water was designed and tested by employing a solid phase microextraction (SPME) procedure. The preliminary experiment involved preparation of a spiked film sample followed by a migration experiment. The experimental details are described below. 47 Spiked film sample preparation Isopropanol was spiked into cleaned HDPE film samples by a sorption technique. The cleaned HDPE film sample was exposed to a high concentration of isopropanol vapor in a closed chamber until reaching an equilibrium state. First, a stainless steel chamber (H 23 cm x ID 13 cm) equipped with gas inlet and outlet port and sampling port was connected to a vapor generator and a vapor stream of isopropanol was allowed to continuously purge through the assembled chamber. The vapor concentration was determined by sampling 100 pl of vapor from the gas outlet port with a 500 pl gas tight syringe (1750 , Hamilton 500 pl gas tight syringe) and analysis by GC-FID. The vapor generator system was similar to the one described for the permeation experiments. Next, four sets of mounted film disks were prepared as described in the previous section (on page 42) and fihn disks of each set were weighed. The respective sets of film disks were then hung on a stainless steel flame inside the sorption chamber and the chamber assembled. The samples were maintained in the chamber for more than 70 days to ensure an equilibrium sorption of isopropanol by the HDPE film disks. Analysis for migration study The migration apparatus was designed as described in ASTM D4754-87, a standard test method for two-sided liquid extraction of plastic materials using the FDA migration cell. The mounted series of isopropanol spiked film disks was placed into a 40 ml precleaned amber vial filled with 40 ml water and closed with an open-top screw cap- teflon/silicone septa. At predetermined time intervals, a 1.3 ml aliquot of solution was withdrawn and transferred to a 2 ml vial which was closed with a screw top hole cap 48 PTFE/silicone septa (Supelco, Inc, Bellefonte, PA, USA). At the time of sampling, a disk of film was removed florn the stack to maintain a constant ratio of contacting surface area and volume of water (1 ml/inz). The isopropanol in the liquid sample transferred to the 2 ml vial was extracted by SPME and simultaneously stirred with a magnetic stirring bar for 10mm. Following extraction, the amount of isopropanol sorbed by the fiber of the SPME apparatus was analyzed by direct thermal desorption into the GC-FID via the injection port. The fiber type used for the SPME and the GC-FID conditions are shown below. SPME fiber: 65 um Polydimethylsiloxane/divinylbenzene polymer (Supelco, Inc, Bellefonte, PA, USA) GC-FID conditions Instrument: GC model HP 6890 with Integrator model HP 3396 (Hewlett Packard, Avondale, PA, USA) Column: HP-5 (crosslinked 5% PH ME siloxane), (30 m x 0.32 mm, 0.25 p m) (Hewlett Packard, Avondale, PA, USA) Mode: splitlcss Oven: 50 °C (5 min) to 90 T at 2 T/min and to 220 T at 10 T/min hold for 5 min Carrier: helium 1 ml/min Injection temp: 220 T Detector temp: 280 T 49 Gases: He 40 psi (1 ml/ min), H2 50 psi (33 ml/ min) , air 60 psi (400 ml/ min), N2 40 psi (25 ml/ min) Desorption: 30 min Elution time: 3.08 min Between each sample analysis, the fiber of the SPME apparatus was cleaned thermally by maintaining the fiber inside the injection port of the GC for a period of 30 minutes. The migrating amount of isopropanol in water was further calculated. The calibration curve and calibration factor are presented in Appendix 1. Determination diffusipn coefficient of Tetracosane in HDPE film Tetracosane being a solid at room temperature and exhibiting an extremely low vapor pressure negated the ability to determine the diffusion coefficient of tetracosnae throuugh HDPE film by either a permeability method or by the vapor sorption method. The migration method could be applied to determine the diffusion coefficient under an experimental design corresponding to the boundary conditions of the mathematical relationship expressed in equation 11. With fruther assumptions, an estimation of the diffusion coefficient could be made, based on inclusion of tetracosane into HDPE film and monitoring it migration. Spiked film making Tetracosane spiked pellets were prepared by a coating technique, and subsequently used to fabricate tetracosane spiked HDPE film. A level of tetracosane in the pellets was set at about 500 ppm by weight. First, the HDPE pellet and tetracosane 50 were weighed at levels of 150 g and 0.075 g, respectively. Next, the tetracosane was dissolved in 100 ml of hexane and transferred to a 1000 ml round bottom flask. The HDPE pellets were then mixed into this solution and the process of coating was started together with solvent evaporation by using a Buchi Evaporator model RE-l 11A (Fisher Scientific, Pittsburgh, PA, USA). The round bottom flask was immersed in a water bath which was heated to 65 T with a Thermolyne Nuova H hotplate (Fisher Scientific, Pittsburgh, PA, USA) and it was slowly rotated under water asperator pressure, until the solvent was fully evaporated. Three batches of coated pellets were prepared for extrusion to film. The coated pellets were fabricated to films using a Killion Extruder model KLB- 100 (Killion Extruder, Inc., Cedar Grove, NJ, USA). The films were extruded under the following conditions. Barrel zone temp.: zonel 355 ‘F, zone2 380 ”F , zone3 380 ‘77 Die temperature: 385 ‘77 Screw speed: 14.5 rpm Chill roll temp.: 65 3“ Chill roll speed: 9.8 rpm With the above conditions, the thickness of the film was between 00015-00020 inch. Migration test The migration apparatus was designed following the procedure described in ASTM D4754-87, a standard test method for two-sided liquid extraction of plastic 5 1 materials using the FDA migration cell. The migration of tetracosane flom spiked film into hexane, water and acetonitrile was monitored under the same test conditions. First, the initial concentration of tetracosane in the spiked film was determined; 30 disks of spiked film (about 0.39 g / 0.022 m2) were weighed and extracted with hexane by soxhlet extraction technique for 24 hours, and the extracted solution analyzed for tetracosane by GC-FID. Eight sets of mounted spiked film disks and three sets of blanks were prepared following the procedure described in the previous section for determination of the partition coefficient of isopropanol. The weight of the spiked film disks for each set was recorded. The stacks of spiked film disks were then placed separately into 40 ml precleaned amber vials, filled with 40 ml hexane for four cells, water for two cells and acetonitrile for two cells. The remaining three blank cells were separately filled with these solvents. All migration cells were closed with open-top screw cap-teflon/silicone septa. At predetermined time intervals, a 1 pl aliquot of solution was analyzed by GC- FID, under the following conditions. GC-FID conditions Instrument: GC model HP 5890A with Integrator model HP 3395 (Hewlett Packard, Avondale, PA, USA) Column: SPB-S (poly 5‘Vo-diphenyl-95‘Vo—dimethylsiloxane), (30 m x 0.32 mm, 0.25 p m), (Supelco, Inc, Bellefonte, PA, USA) Mode: splitlcss Oven: 250 °C hold for 30 min Carrier: helimn 1 ml/min 52 Injection temp: 300 T Detector temp: 250 T Gases: He 45 psi (1 ml/min), H2 20 psi (33 ml/min) , air 40 psi (300 ml/min), N2 35 psi (25 ml/min) Elution time: 6.63 min Standard solutions of tetracosane in hexane were prepared by a serial dilution procedure and used to determine the linearity and sensitivity of the analytical procedure. The calibration curve and calibration factor are presented in Appendix I. Evalgting thg migppting amount of ISOpropanpl and Tetracome flom single layer HDPE film and lamination HDPE fihn By knowing the diffusion coefficient of isopropanol and tetracosane in HDPE film, the migration rate of these compounds into a food contact phase, which is assumed to be controlled by the diffusion process, can be calculated fl'om suitable equations derived for specific boundary conditions. Here, a migration rate of both compounds flom a single layer film was estimated by equation 1 1. In addition, the functional barrier for preventing the transfer of migrants into a food contact phase flom a laminate film contaning a contaminated HDPE layer and a virgin HDPE layer was evaluated by computing the migrating amount of both compounds from film into food contact phase with the Laoubi and Vergnaud migration model (Laoubi et al., 1995) and the Begley and Hollifield migration models (Begley et al., 1993). The results of migration from a single layer and a lamination film were compared and the functional barrier was evaluated, based on the capacity of remaining contaminants level inside the film. RESULTS AND DISCUSSIONS Permeation Measurements of Isopropanol through HDPE Film The results of permeation measurements determined by both the isostatic and quasi-isostatic techniques are summarized below. Diffusion coefficignt of isopromol in HDPE fil_rp The diffusion coefficient of isopropanol in HDPE film was determined hour the permeation data for isopropanol vapor through a I-HDPE film, measured by both the isostatic and quasi-isostatic techniques. Both techniques employ the same concept to determine the diffusion coefficient of vapor in film. The isostatic or dynamic permeation measurement allows the diffusion coefficient to be calculated fi'om the ta 5 value, which is related to the flactional change in mass flux fl'om time zero to steady-state flux, where the to; value is the time required to reach one-half of the steady state transmission rate, as shown in equation 24. Whereas the difl‘usion coefficient obtained from the quasi-isostatic procedure is related to the intersection of the linear portion of the transmission rate profile curve with the time axis, where the obtained lag time value (6) is used to calculate the diffusion coefficient by substitution into equation 26. The permeation measurements were performed at several vapor activity levels, at room temperature (23 :1: l T), to evaluate the effect of vapor concentration on the diffusion coeflicient. Since the test temperature was well above T8 for HDPE, about 195-200 K, the HDPE film is in the 53 54 rubbery state. Thus, the measrued permeation of isopropanol vapor is assumed to be governed only by the diffusion mechanism. The diffusion coefficient values for isopropanol through HDPE fihn by both isostatic and quasi-isostatic permeation method are presented in Tables] and 2, respectively. For the isostatic technique, the diffusion coefficient was between 3.6x10'14 to 3.9 x 10'” mZ/s for the vapor activity range between 0.1 and 0.38, while for the quasi-isostatic technique, D values were between 3.6 x 10'" to 4.5 x 10'14 mz/s, for the vapor activity levels between 0.15 to 0.5. The quasi-isostatic procedure showed greater scatter in the data as shown in Figure l, where the diffusion coefficient values are plotted as a function of vapor activity for both the isostatic and quasi-isostatic procedures. The results flom the quasi-isostatic permeation method show much more fluctuation of the diffusion coefficient than the values obtained by the isostatic method. This is attributed to the fact that the calculation of diffusion coefficient from the quasi-isostatic procedure requires a very precise linear portion of the steady state transmission rate to define the time lag by extrapolation. The error was not considered due to extrapolation of the steady state portion of the cmve to the x axis, but rather that the experimental method required manually sampling and analysis during the permeation experiments. However, in terms of statistical analysis, there was no statistically significant difference in the population mean of diffusion coefficient values observed between either technique, at a significance level of 0.5 (Appendix H). It was also found that there was no statistically significant difference in population mean of the diffusion coefficient between concentrations levels for each technique. This leads to the conclusion that the diffusion coefficient of isopropanol in HDPE fihn is likely independent of concentration, within the experimental range of activities studied. Table 1 Diffusion, Permeability and Solubility Coefficient values for Isopropanol in HDPE film by the Isostatic technique“) Vapor activity Diffusion Permeability Solubility Coefficient Coefficient Coefficient (NZ/8) (ll g.m/m’.s.Pa) (g/g.Pa) 0.10 3.6E-14 4.6E-9 1.6E-7 0.21 3.8E-l4 4.6E-9 1.5E-7 0.38 3.9E-l4 4.6E-9 1.5E-7 Average_ 3.7E-14 4.6E-9 1.6E-7 SD i2.9E-15 i4.2E-10 :tl .8E-8 (I). The results are the average of four samples. Table 2 Diffusion, Permeability and Solubility Coefficient values for Isopropanol in HDPE film by the Quasi-isostatic technique“) Vapor activity Diffusion Permeability Solubility Coefficient Coefficient Coefficient (mzls) (It g,m/m’.s.ra) ML. 0.15 3.6E-14 3.7E-9 1.3E-7 0.19 4.2E-l4 3.9E-9 1.2E-7 0.27 4.5E-14 4.5E-9 1.3E-7 0.41 3 .9E-l4 4.4E-9 1.4E-7 0.50 3.9E-14 3.8E-9 1.2E-7 Average 4.1E-l4 4.1E-9 1.3E-7 SD i5.7E-15 i8.4E-10 :t3.0E-8 (2). The results are the average of at least two samples. Pmeability coefficient of ispmppppol through I-H)PE film The permeability coefficient values for isopropanol through HDPE film, determined by the isostatic and quasi-isostatic procedures are summarized in Tables 1 and 2, respectively. For better illustration, the permeability coefficient values determined are plotted on a function of permeant concentration in Figure 2. As shown, the permeability 56 coefficient values agreed well over the vapor activity range evaluated. Further, there was good agreement between the P values determined by the two procedures Splubilig coefficient of isopropanol in HDPE film In addition to determining the diffusion and permeability coefficients, the permeation measurement also allows for determination of the solubility coefficient fl'om the relationship P= DS, assuming the diffusion coefficient is independent of the penetrant concentration and Henry’s law is obeyed. Here, the diffusion coefficient was constant and did not vary statistically with concentration and Henry’s law was assumed, due to the low vapor activity range studied. Thus, the solubility coefficient values determined flom permeability data by both the isostatic and quasi-isostatic techniques were assumed to be acurate (see Tables 1 and 2). As Shown, the solubility coefficient values obtained from both techniques were in agreement with s falling between 1.2 x 10'“ and 1.6 x 10'“ g/g Pa. The relationship between the solubility coefficient values determined by the respective permeability experiments and vapor activity is shown in Figure 3. 57 5.00E-14 -:- —— 5.005-14 7 4.50544 «~ 4.50544 ".5. 1 D 4.005-14 n n .- 4.00514 : x 3.50E-14 -I- 6 -_ 3,505.14 g I uQuasi-isostatic 3.00E-14 '5' xlsostafic “* 3.00E-14 2.50E-14 4+ 7‘ + if 2.50544 0.1 0.21 0.38 0.5 0.15 0.19 0.41 Vapor activity, a Figure 1 Diffusion coefficient of Isopropanol in HDPE film at 23 °C by Quasi-isostatic and Isostatic permeation techniques. 58 5.00509 -;- 4. 5.00509 6: 4.50E-09 -. 9 Q : x X E 4.00509 -- -. 4.50509 E, 2 D a ' D 2; 3.50509 -:- S 3.00509 1» 4.00509 2.50509 =3 : 3 2-005'09 nQuasi-isoatatic “‘ 3-505-09 g : xlsoctatic ._ 1.50509 -:- 1.00509 15 1 J. 1 3.00509 0.1 0.21 0.39 0.5 0.15 0.19 0.41 Vapor activity, a Figure 2 Permeability coefficient of Isopropanol through HDPE film at 23 °C by Quasi-isostatic and Isostatic permeation techniques. Solubility coefficient (a I g.Pa) 59 1.90507 -»— 1.90507 1.70507 + 1.70507 : x 1.50507 -:- x x .. 1.50507 :1 1.30507 a .- 1.30507 : n 1: 1.10507 .- 1.10507 9.00509 -:- ~— 9.00509 ; a Quasi-isostatic 7.00509 X 'Wc ~- 7.00509 5.00509 49 t . + 5.00509 0.1 0.21 0.39 0.5 0.15 0.19 0.41 Vapor activity, a Figure 3 Solubility coefficient of Isopropanol in HDPE film at 23 °C by Quasi-isostatic and Isostatic permeation techniques. 60 Determination of the Equilibrium Partition Coefficient of Isopropanol between HDPE Film and Water The equilibrium partition coefficient of isopropanol between HDPE film and water was determined by measuring the equilibrium sorption of isopropanol in HDPE film and the level of isopropanol remaining in the aqueous solution at equilibrium. The equilibrium partition coefficient values determined are reported as a ratio of the equilibrium concentration of isopropanol in the aqueous solution divided by the sorbed concentration of isopropanol in HDPE film at equilibrium (KI/p). Three samples and one blank or control were tested and the results are summarized in Table3. Table 3 Equilibrium partition coefficient of Isopropanol between an aqueous phase and HDPE film by the sorption method at 23 °C Sample ct. (wt/we“) Cr (wt/wt) “7 Kn?”— blank 7.5E-5 - - 1 7.3E-5 1.2E-7 6.0E+2 2 7.5E-5 1.3E-7 5.6E+2 3 7.5B-S 1.2E-7 6.3E+2 average - - 6.0E+2 standard deviation - - 136 71). CL is a concentration of isopropanol in aqueous phase. (2). Cp is a concentration of isopropanol in HDPE film. 0’. Kup is a partition coefficient of isopropanol between aqueous phase and HDPE film. A partition coefficient value of 600 clearly indicates a very low amount of isopropanol sorbed by the HDPE film from the aqueous solution and supports the expectation of low solubility of isopropanol in HDPE film. Here, the partition coefficient was determined by the equilibrium sorption procedure and was determined at only one sorbate concentration (100 ppm, wt/wt). This partition coefficient value was determined 61 at an initial isopropanol concentration of 100 ppm (wt/wt) and it was assumed to be a constant, with no concentration dependence over a wide range of sorbate concentrations. The analytical method developed to determine the concentration levels of isopropanol in the HDPE phase, the TS-TD-GCMS procedure, has a sensitivity at a sufficiently low level for the analysis of relatively low sorbed amounts of isopropanol in the polymer. Gavara et al. (1996) was successful in using TS-TD-GCFID procedure to determine the partition coefficient of toluene in water/polystyrene, as well as for binary mixtures of toluene and d-limonene, toluene and ethyl acetate in water/polystyrene (Gavara, et al., 1996). In this case, partition coefficient (KP/L) values ranging between 1.9 and 5100 were determined. Migration Studies of Isopropanol from HDPE Film into Water The migrated amount of isopropanol from spiked HDPE film into water was measured as a function of time by the SPME procedure until a constant amount of migration was reached. Next, the actual level of migration in mass per unit area was plotted as a function of time and the predicted migration levels were calculated from equation 12, where the diffusion coefficient was varied until the linear portion of the prediction curve was superposed over the actual migration curve. The results are summarized in Table 4 and presented graphically in Figure 4, where the migrated levels of isopropanol (g/mz) is plotted as a function of contact time. As shown in Figure 4, the predictive migration levels agreed well with the actual migration levels for the contact time from 0 to 10,800 sec. However, since the initial concentration of migrant in the film was assumed to be constant, for application of equation 12, but in practice the migrant 62 concentration is gradually decreasing until an equilibrium concentration is obtained, application of equation 12 may not be applicable over the total time range to equilibrium. Since the assumption of a constant migrant concentration is not valid over the total time range. The diffusion coefficient by this superposition method to give the best fit to the test data was 3.0 x 10'“ mz/s, which the average the diffusion coefficient determined by the isostatic technique was 3.7 x 10‘” mz/s. A complete calculation of the predictive curve is presented in Appendix III. The SPME is theoretically excellent for qualitative analysis and good for quantitative analysis, if it is well calibrated. In this experiment, the SPME with PDMS/divinylbenzene fiber was applied to extract isopropanol in water. In this case, the stagnant layer of water around the fiber may interfere with the partitioning of isopropanol between water and fiber, even with the use of magnetic stirring. This can limit the sensitivity of the fiber to extract analyte, particularly small amounts of polar compounds such as isopropanol from the water phase. Theoretically, the SPME affords an alternative, and highly sensitive extraction technique, provided the fiber has a strong afi'mity for analytes such as isopropanol in water and reduces the chance of loss of isopropanol during analysis. Nevertheless, the SPME could be a fast extraction instrument for survey experiments. In general, analytical methods for determination of very low levels (ppb) of volatile polar compounds in aqueous solution are difficult. Many commonly used solvent extraction techniques are not suitable for extraction of polar compounds in water. Even though the purge and trap technique has worked well for some solvents, the sensitivity is limited to part per million, due to their solubility in water (Shirey 1997). Hence, for migration studies, it may be preferable to determine the actual 63 contaminant levels in the film, which are normally very low, rather than to determine trace levels of migrants in an aqueous contact phase. Table 4 Migration of Isopropanol from spiked HDPE film into aqueous phase (water) at 23 °C Time Actual M' ration Predicted ration ‘0 (sec) (g/m) (24m ) 0.00 0.00 0.00 9.01=.+2 0.011 0.012 3.6E+3 0.021 0.024 1.11=.+4 0.034 0.041 2.2E+4 0.039 0.059 9.0E+4 0.039 0.12 m. The migrated amormt are predicted by using equation 12. 0.12 x Experiment 9 A J T .o 8 Migrated isopropanol in aqueous (glm’) p 8 0.04 . x 0.02 « 0 + : 1 1 0 20000 40000 60000 80000 100000 Time (ace) Figure 4 Migrated amount of Isopropanol from spiked HDPE film into aqueous phase (water) at 23 °C Determination of Diffusion Coefficient of Tetracosane in HDPE Film The tetracosane spiked HDPE film samples were prepared as described in the materials and methods section and were extracted with hexane by soxhlet extraction to determine the initial concentration of tetracosane present. The extraction solvents were analyzed by GC-FID and the results are summarized in Table 5. The initial tetracosane concentration of coated HDPE resin expected was about 500 ppm (wt/wt), based on the mixing ratio. As shown in Table 5, the calculated level of tetracosane in the spiked HDPE pellets was in good agreement with the mixing ratio concentration. This verified the validity of the coating method for impregnation of tetracosane onto HDPE resin pellets. As also shown in TableS, the initial amount of tetracosane in the extruded film 65 appeared to be constant and with minimal losses during the extrusion step, providing further supportive evidence of the validity of using this average initial concentration as the initial concentration for the migration experiments. Table 5 Initial amount of tetracosane in coated pellet and extruded film Sample") Weight of sample Weight of Cull” Concentration (K) (It) L filml 0.3944 0.00016 0.00039 film2 0.4041 0.00016 0.00039 film3 0.43 86 0.00018 0.00042 coated pellet 0.423 8 0.00021 0.00049 (1). Sample films were cut from several area of film roll. The diffusion coefficient of tetracosane in HDPE film was estimated by solving for M/Mw from equation 12, with varying diffusion coefficient values and comparing the calculated and actual migration data for the best fit. The actual migration experiments were carried out with three selected solvents, namely: hexane, acetonitrile and water. For acetonitrile and water, there was no measurable migration of tetracosane within experimental time, while for hexane, a rapid rate of migration of tetracosane was observed. The actual migration data obtained and the predicted migration values are presented in Appendix IV. The diffusion coefficients estimated for each sample are summarized in Table 6. Also summarized in Table 6 are the determined partition coefficient values. No attempt was made at evaluating the effect of tetracosane concentration on the partition distribution and the diffusion coefficient was assumed to be a constant and independent of sorbate concentration 66 The average diffusion coefficient for tetracosane in HDPE film was 1.1 x 10’16 mz/s. This values agreed with the diffusion coefiicient values for n-C18 and n-C32 hydrocarbon in HDPE, as determined by Limm and Hollififield (1996) and shown in Table 7. Even though the migrant, the solvent and testing condition were different fi'om these employed in the present study, comparison provided the trend of the diffusion coefficient value. The diffusion coefficient of tetracoane in hexane was less than the n-32 hydrocarbon in corn oil since hexane is a stronger solvent than corn oil and the testing temperature was also lower. It is important to note that since hexane is an polymer swelling solvent for HDPE, the calculated diffusion coefficient value is considered an upper bound value. For a non-interactive contact phase, the diffusion coefficient for tetracosane in HDPE may be several order of magnitude lower. A potential source of error associated with the above method for determination of the diffusion coefficient is lack of agreement with the assumption that tetracosane migration is a F ickian diffusion controlled process, since the prediction model (equation 12) is a solution for Fick’s second law of diffusion. However, by observing the migration curve (a plot of M/Mw as a function of square root of time as shown in Figure 20-25, Appendix IV), which showed a linear region in the initial stage (but less than 60 % of Mao) followed by the curve being concave to the abscissa axis, it was assumed that the migration process was well described by equation 12 and the estimated difiusion coefficient would not be in serious error. For a prediction of migration of tetracosane in other food systems, the high value of this diffusion coefiicient, compared to food solvent would be considered as a safety factor of the packaging design. 67 Table 6 Partition coefficient and diffusion coefficient of tetracosane in HDPE/Hexane at 23 °C Sample Partition coefficient Diffusion coefficient Km In2 / s 1 0.0048 5.0E-17 2 0.0075 1.5E-16 3 0.0080 1.6E-16 4 0.0069 6.5E-17 5 0.0062 1313-16 6 0.0034 2.0E-l7 Average 0.0061 1.0E-l6 SD 0.0017 i6.9E-17 Table 7 Comparing diffusion coefficient of tetracosane in HDPE/Hexane to n-Cl8 and n-C32 in HDPE/corn oil Polymer Migrant Temperature Solvent Diffusion coefficient 1°C) (m’ls) HDPE n-18 30 corn oil 1.3514m HDPE n-24 23 hexane 1.1E-16 HDPE n-32 30 corn oil 1.2516(2) (”(2). The data were from Limm and Hollifield (1996). Estimating Migration Levels of Isopropanol and Tetracosane from a Single Layer HDPE Film and a Laminated HDPE Film Based on the assumption that migration of contaminants into a food contact phase (liquid food) is a solely diffusion controlled process, the migration of isopropanol and tetracosane from a polymer film to a fluid contact phase was estimated separately, for a single layer HDPE film and laminated HDPE fihn by using the diffusion coefficient values determined for the respective migrants in HDPE film. Three migration models, namely: (i) a simple model (Model I) for a single layer film; (ii) the Begley and Hollifield 68 model (Model 11); and (iii) the Lauobi and Vergnaud model (Model HI), were evaluated to estimate the amount of migration. The simple model (Model I) was used to estimate migration levels fiom a single layer structure, while Models 11 and III were used to estimate migration from a two layer laminate structure. An evaluation of these models was made and the effectiveness of a functional barrier layer (virgin HDPE layer) to prevent the migration of both compounds was also determined. The three migration models considered are discussed below. Single layer of HDPE film (Model I): D t ”2 M, = 2C” [-fi—J from eqn. 11 Double layer lamination model: Begley and Hollifield model (Model 11) M Dr 1 2 °° -1" D»2 2t __L=___-;—2—Z( 2) exp[— Ir J fromeqn.4l 69 Lauobi and Vergnaud model (Model HI) M.,. _ 8L i (-1)" S. (2n+1)7rH (_(2n+1)’zr2 Dr) fi'om eqn. 38 M, " n’H... (2n+1)2 m 2L ““1 4L2 114,, 9L °° 1 . 2(211“)er ( (2n+1)21r2 J ’ = -——— 1———-—-—-Dt fro .39 M, ”mg (2,. +1)2 8‘“ 2L ””1 4L2 m “l“ M, = M," — M” from eqn. 40 It should be noted here that the single layer would be called the recycle layer or contaminated layer and the virgin layer is called the barrier layer. The single layer was a recycled HDPE film, while the laminate was a HDPE/HDPE structure, with the recycled HDPE forming the external surface, and the barrier or virgin HDPE layer providing the food contact layer. From the Plastics Recycling Task Force document, “Guidelines for the Safe Use of Recycled Plastics for Food Packaging Applications” (NFPA, SP1, 1994), a study of the effectiveness of plastics reclamation processes was made. In these guidelines, seven compounds which are 2,4 dichlorophenol, ethylene glycol, isopropanol, methyl salicylate, methy stearate, 2,2,4-trimethylpentane and xylene were selected to represent possible contaminants present in post consumer recycled plastics. It was stated in the guidelines that “a level of 1% of each surrogate in plastics was an effective contamination rate, which was believed to exaggerate the level of contaminated container that can be found in 70 the recycled stream.” The guidelines involved spiking virgin HDPE flakes with 1 % of each of the contaminants and determining the levels of the respective contaminants remaining in the spiked HDPE flakes after they had been through a cleaning step in the reclamation process. The remaining contaminant levels were then assumed to be representative of the initial concentration levels of contaminants in recycled HDPE plastics. By assuming a 1% level of the contaminant in post consumer HDPE plastic, followed by the cleaning process, the retained isopropanol level was 0.21 ng/cm’ which is lower than the minimum initial concentration level of 0.91 ng/cm’ required to attain the maximum allowable contaminant level in food. In addition, there was no data report for tetracosane in the guideline. Thus, here, the initial concentration level of isopropanol and tetracosane in HDPE film was calculated based on a threshold of regulation dietary exposure level, 0.5 ppb in food, as shown below. For estimation of migration levels, an initial concentration of migrants in HDPE film was calculated, based upon a threshold of regulation dietary exposure, 0.5 ppb, together with a consumption factor of HDPE, 0.33,and a ratio of food weight and contacting package area, 1.55 g/cm’ , as recommended by the FDA (FDA, 1992). The density of the HDPE film and the thickness of the recycle layer were 0.965 g/cm3 and 0.002921 cm (1.15 mil), respectively. From the regulation dietary exposure and the consumption factor, the concentration of migrant in the food contact phase was obtained by the following expression; 71 CF- M = 05 ppb (44) where; CF == consumption factor M = the concentration of migrant in food Thus, for recycled HDPE plastic, the maximum allowable migrant concentration level in the food was 1.5 ppb (gram migrant/gram food). Next, the minimum initial concentration in HDPE film yielding 1.5 ppb in the food was calculated fiom the expression; W - M p bl = 4 M F C ( 5) where; M = the concentration of migrant in food [gm /g,i WP = the weight of polymer for 1 cm2 [gp/cmz] Mn. = the initial concentration of migrant in polymer [g,,, /g,,] F C = the ratio of food to contacting package area [gf/cmz] By substituting all variables in equation 45, the minimum initial concentration level of migrant in HDPE film yielding 1.5 ppb in the food was 0.84 ppm (wt/wt) or 8.1 x 10'7 g/cm’. This minimum initial concentration level of migrant in film represents the lower bound of migrant level that can result in a migrant concentration level of 1.5 ppb in food, assuming the migrant completely transfers to the food. Furthermore, 72 concentration levels of migrant which are above this minimum concentration level in the film would be expected to show higher migration rates, due to the higher concentration gradient and thus exceed the restriction migrant concentration level more rapidly. Even though using a minimum initial concentration level of migrant of 0.84 ppm (wtbvt)in the film to estimate migration levels to a food contact phase by Model I, II and III will equal but not exceed the regulation limit of 1.5 ppb. (assuming total migration), this minimum initial concentration level of migrant in film was still selected to estimate the migration levels of isopropanol and tetracosane to food migrating from a single layer and a laminated HDPE film, based on the following considerations. First, the migrant level in food will never reach the regulation limit if the migrant concentration level is lower than the selected minimum initial concentration level of migrant in film (0.84 ppm). Second, if the estimated migrant levels in the food contact phase, calculated by Model H and 111 show an ‘ineffectiveness’ of the functional barrier layer to prevent the migration of contaminant, such that the time for the migrant level in the food to equal or exceed the regulation limit is less than the expected product shelf life, it can be concluded that the migration of higher initial concentration levels of migrant in film can not be prevented by the same functional barrier layer. On the contrary, if the estimated migrant level in the food, calculated by Models 11 and III, shows an ‘effectiveness’ of a functional barrier layer to retard the migration of contaminant, such that the time for the migrant level in the food contact phase to approach the regulation limit is much longer than the expected product shelf life, it can not be directly concluded, as to whether it is safe to use the functional barrier to prevent the migration levels from being exceeded at higher initial concentration levels. However, it still demonstrates the potential use of a functional 73 barrier to restrict migration, by reducing the initial concentration level of contaminant in film to as low a value as possible. The diffusion coefficient values determined for the respective migrants and used to solve for the levels of migrant transfer with time were an average value for each of migrant, namely: 3.7 x 10'") cmZ/s and 1.1 x 10'12 cmz/s for isopropanol and tetracosane respectively. For both migrants, the estimated migration levels from the recycle layer by Model I were compared to the results obtained from Models 11 and III. In the latter cases, a functional barrier layer of HDPE was considered. The results of the solution of the respective migration models are summarized in Tables 8 and 9 and presented graphically ill Figures 5 and 6. For Models I and II, the presented results which were above 1.5 ppb as shown in Tables 8 and 9 and Figures 5 and 6 are invalid numbers and are only used to illustrate the deficiencies of the two Models. As shown in Figures 5 and 6, migration model I gave the fastest time to reach the maximum allowable level in the food, 1.5 ppb, since no functional barrier layer was considered. From Tables 8 and 9, it becomes apparent that a barrier layer retards the rate of migration, especially for Model IH. Model II also predicted a rate of migration similar to that of Model 1, since in Model 11, the initial concentration of migrant was assumed to be constant at the surface of the barrier layer located next to the recycle layer and there was no consideration given to the migration rate of migrant within the recycle layer. In Model 111, the thickness of the recycle layer and the gradual reduction of the initial migrant concentration are considered, making this model more realistic. Nevertheless, for a very thin barrier thickness such as 1 mil, the results from Model III showed a failure of the functional barrier to totally prevent 74 the migration of isopropanol. The total amount of isopropanol was transferred into the food in less than five days. In contrast to isopropanol, the results from Model 111 showed an effectiveness of the barrier layer to retard the migration of tetracosane. It would take more than one and a half years to attain about 95% migration at the same initial concentration of migrant. Table 8 Comparison of the migrated amount of isopropanol from a single layer film and a laminated film based on predictive Models 1, 11 and III“) Time Model 1“) Model “in“ Model 111 (sec) (ppb) (ppb) (0pr 0.00 0.00 0.00 0.00 3.6E+3 0.68 0.09 0.04 7.2E+3 0.96 0.33 0.16 1.1E+4 1.2 0.60 0.30 1.4E+4 1.4 0.97 0.42 1.8E+4 1.5 1.1 0.53 2.2E+4 1.7 1.4 0.64 2.3E+4 1.7 1.5 0.67 2.9E+4 1.9 2.0 0.91 7.2E+4 3.0 5.2 1.3 8.6E+4 3.3 6.3 1.4 1.8E+5 4.9 13.4 1.5 (I). (21.13). Asinglelayerfilrnisal.15milcontaminatedl-IDPEfilm,whilealaminatedfilmisa laminationofl.15 mil contaminated HDPE film and 1 mil virginHDPE film. The values above 1.52 are an invalid numbers andjust presents the defect ofModel I and II. 75 Table 9 Comparison of the migrated amount of tetracosane from a single layer film and a laminated film based on predictive Models 1, 11 ad 111“) Time Model rm Model 11?” Model 111 (sec) (ppb) (ppb) (0pr 0.00 0.00 0.00 0.00 5.0E+6 1.3 0.86 0.41 6.0E+6 1.5 1.1 0.50 8.0E+6 1.7 1.5 0.66 1.013+7 1.9 1.9 0.90 1.2E+7 2.1 2.4 0.92 2.01~:+7 2.7 4.1 1.2 5.01=.+7 4.2 10 1.5 9.01m 5.4 17 1.5 9.013+7 5.7 19 1.5 1.01::+9 6.0 21 1.5 1.512+9 7.4 32 1.5 (I). (21.13). Asinglelayerfilmisal.15milcontaminatedHDPEfilm,whilealaminatedfilmisa lamimtion of 1.15 mil contaminated HDPE film and 1 mil virgin HDPE film. The values above 1.52 are an invalid numbers and just presents the defect of Model I and II. 76 6 —-+—-Modell 5 .. —x—Model ll 4 ~— - -- Model Ill -———Max. allowable level Migrated amount of Isopropanol In food (ppb, wtlwt) (a) 200000 Figure 5 The profiles of the migrated amount of isopropanol from a single layer and a laminated film based on predictive Models 1, II and III Migrated amount of tetracosane in food (ppb. Wt) Figure 6 The profiles of the migrated amount of tetracosane from a single layer and 77 l ‘f 4L —-+—Mode| l —x—Model II - -— Model lll -—- Max. allowable level L j 2E+07 4E+07 6E+07 8E+07 1E+08 1E+08 1E+08 2E+08 Time (see) a laminated film based on predictive Models 1, II and III 78 A profile of the amount of migrant in each layer was determined, which considered the remaining amount of migrant in the recycle layer, incoming amount of migrant into the barrier layer, and outgoing amount of migrant to the food contact phase as a function of contact time. The results were summarized in Table 10 and presented graphically in Figure 7 for isopropanol and in Table 11 and Figure 8 for tetracosane. The effect of the functional barrier thickness on the predictive results from Models II and III was determined for the barrier thickness ranging fi'om 1 mil to 10 mil. The results were summarized in Tables 3841 (Appendix V) and presented graphically in Figures 9-12 where the migrated amounts of isopropanol and tetracosane were plotted as a function of time for the respective barrier thickness levels. As shown, for the thinner barrier layers (1 and 2 mil), the migration levels calculated from Model 11 were always higher than these obtained from Model 111. However, when the barrier thickness was equal or greater than 4 mil, the migration levels estimated timing the early stages of storage time (i.e. less than 40 % of total migration time) were greater with Model 111 than for Model H. The time for total transfer of migrant from Model ID was still greater than for Model 11. This phenomenon was observed for both migrants. Thus, if Model III was used to estimate migration levels for a laminate fabricated with a barrier layer greater than 4 mil, during the early stages of migration the result would be an over estimation of transfer levels. By varying the barrier thickness, 9 relationship between the time to reach maximum allowable migration levels as a function of barrier thickness was determined for both migrants, as shown in Figure 13 and 14, respectively. The relationship between barrier layer thickness and the time to reach the maximum allowable level of migrant can be described by the following polynomial expressions. 79 For isopropanol: T = (2 ><10'5)t5 -—(1)<10")i4 +(2 x10”)!3 +(4 x109)t2 +(9 x107)t+112381 (46) For tetracosane: T=(3x10”)t5 -(2 ><10"5)t4 +(4 x10“)!3 +(2 x1012)!2 +(5x10m)t+3x107 (47) where T is the time (sec) to reach maximum allowable level and t is the thickness (cm) of barrier layer. These two equations allow the prediction of the time to reach the maximum allowable level of migrant at any barrier thickness above 0, for the same initial migrant concentration level. Based on the results discussed, it can be concluded that Model H is not the most appropriate migration model for describing the migration rate from a two layer laminate film. Further, while Model IH afforded a certain level of satisfaction as a prediction model, it is important to note that this expression was derived based upon several important assumptions, namely: (i) that the convective coefficient of mass transfer was set as infinite; (ii) that the diffusion coefficient was independent of concentration and time; (iii) and there was no concern of the interaction between food content and polymer. The validity of the application of this model to estimate or predict migration levels should therefore be evaluated on a case by case basis. 80 Table 10 The migration data of isopropanol for a lamination of 1.15 mil contaminated HDPE layer/ 1 mil HDPE barrier layer by Model HI Time Migrant in recycle layer Migrant in barrier layer Migrant in food (sec) (ycm’) Jg/cm’) (gzcm’) 0.00 2.4E-9 0.00 0.00 3.6E+3 1.8E-9 4.6E-10 6.8E—1 1 7.2E+3 1.6E-9 4.9E-10 2.6E-10 1.0E+4 1.5E-9 4.8E-10 4.1E-10 2.0E+4 1.1E-9 3.7B-10 9.2E-10 3.0E+4 7.9E-10 2.7E-10 1.3E-9 4.0E+4 5.8E-10 2.0E-10 1.6E-9 5.0E+4 4.3E-10 1.7E-10 1.8E-9 6.0E+4 3.2E-10 1.1E-10 1.9E-9 7.0E+4 2.3E-10 7.9E-1 l 2.0E-9 8.0E+4 1.7E-10 5.8E-11 2.1E-9 Table 11 The migration data of tetracosane for a lamination of 1.15 mil contaminated HDPE layer/ 1 mil HDPE barrier layer by Model III Time Migrant in reczycle layer Migrant in barrier layer Migrant in food (sec) (gem) (g/cmz) (g/cmz) 0.00 2.4E—9 0.00 0.00 5.0E+6 1.3E-9 4.3E-10 6.4E-10 6.0E+6 1.2E—9 4.0E-10 7.8E-10 8.0E+6 9.9E-10 3.4E-10 LOB-9 1.0E+7 8.3E-10 2.8E-10 1.2E-9 1.2E+7 7.0E-10 2.4E-10 1.4E-9 2.0E+7 3.5E-10 1 .2E-10 1 .9E-9 S.0E+7 2.6E-1 1 8813-12 2.3E-9 8.0E+7 1.9E-12 6.5E-13 2.4E-9 9.0E+7 7.9E-12 2.7E-13 2.4E-9 1.0E+8 7.9E-13 LIE-13 2.4E-9 81 2.5E-09 1. 2.0509 b“ «1‘ “‘t‘ 5 15509» ' B E ‘ ------ In recycle layer 2 '\ --—-In barrier layer ; 1-05'09 ‘ lnfood 5.0510 ~~ ,-- g “ . i ‘\.‘ 0i ‘ *--._.,_ _. ...,. .- 00500 e 1 4 ’7 “" " 0 20000 40000 60000 60000 100000 Tlmelaee) Figure 7 The migration profile of isopropanol for a lamination of 1.15 mil contaminated HDPE layer/l mil HDPE barrier layer by Model 111 82 2.5E-09 2.0509 "5 1.5509 3 o E .5 10E—09 : ------ Inreeyelelayer ---- lnbarrlerlayer In food 5.0E-10 1 0.0E+00 5 4 .~?.=.é‘r :. T at i 4 0.0E+O 2.0E+O 4.0E+O 6.0E+O 8.0510 1.0E+O 1.2E+0 1.4E+O 1.6640 0 7 7 7 7 8 8 8 8 Time (see) Figure 8 The migration profile of tetracosane for a lamination of 1.15 mil contaminated HDPE layer/l mil HDPE barrier layer by Model 111 83 3.5 +1 mll Model ll 3 3"T --D--1milModellll .E -— Max. allowable level '2 2-5 r --o——2 mil Model ll 3‘... --o--2milModellll 5g 2 +4 mil Model ll 14- , [ --a--4mllModelll § E 15 -.- = ..... W 3 V ’. . ..a """ 5 o a" E 1 '1 .g .0. g D 5 0.5 0 1 1 1 1 1 1 1 0 200000 400000 500000 900000 100000 120000 140000 150000 0 0 0 0 Tlme(aee) Figure 9 The migration profile of isopropanol for a lamination of 1.15 mil contaminated HDPE layer/1, 2, 4 mil HDPE barrier layer by Models H and III 2.5 ." .g. 0. . . -—x—6 mil Model ll - - -x- - -6 mil Model ill . Max. allowable level i‘ —x—9 mil Model ll 0 ---x---8milModellll ,1 ,1 10 mil Model ll 5 ------ 10 mil Model Ill .5 5 Q” a .0 a: Migrated amount of isopropanol in food (ppb. MM) 5; l 0 1000000 2000000 3000000 4000000 5000000 6000000 Tlme(aee) Figure 10 The migration profile of isopropanol for a lamination of 1.15 mil contaminated HDPE layer/ 6,8,10 mil HDPE barrier layer by Models II and III 85 -—u—-—1 mil Model ll --l:i--1milModellll -——Max. allowable level —-o—2 mil Model ll --o--2mil model Ill +4 mll Model ll --e--4milModellll (Mimi Migranted amount of tetracosane in food I l r T I I I 0.0 +00 1.0E+08 ZOE-+08 3.0E+08 4.0E+08 5.0E+08 6.0E+08 7.0EL+08 Time (see) Figure 11 The migration profile of tetracosane for a lamination of 1.15 mil contaminated HDPE layer/ 1, 2, 4 mil HDPE barrier layer by Models II and [H Migrated amount of tetracosane in food 86 1.8 1 1.61~ 1.4 .. 1.2 1 (Pill). MM!) 0.0 = +00 1.0E+09 2.0E+09 3.0E+09 4.0E+09 5.05 [+09 —x—6 mil Model ll - - -x- - -6 mil Model Ill Max. allowable level —x—6 mil Model ll -- x- - -8 mil Model ill 10 mil Model II ------ 10 mil Model Ill l L l I I I 'l’lme (ace) Figure 12 The migration profile of tetracosane for a lamination ofl.15mil contaminated HDPE layer! 6, 8, 10 mil HDPE barrier layer by Models 11 and III 87 1.2E+07 '1; T=2E+1515-1E+14t‘+2E+12t’+4E+09t’+9E+07t 112391 ; 1.0507 .. R2“ 3 8.0E+06 1» I E a A 5 § 9.0154094 3 4.0E+06 «r 3 2 0E+06 q x Model III E ‘ I -—Polynomial headline 1: 0.0E+00 a 1 1 1 1 0.0500 5.0503 1.0502 1.5502 2.0502 2.5502 3.0502 Barrier thickness (cm) Figure 13 The relation of time to reach maximum allowable level as a function of a barrier thickness for isopropanol as initial concentration of 0.84ppm 88 5.0E+09 T = 35171“ - 25160 + 451.1413 + 2E+1212 + 55101 + 3E+07 ' R2 = 1 4.5509 . 4.0509 1» 3.5E'l‘09 «1 511 11 “1‘13 2 raae maximum allowable level (see) x Model Ill 1.0509 .. 3 —— Polynomial trendline E50509 .. i: 0.0E+00 1 1 1 1 1 0.0500 5.0503 1.0502 1.5502 2.0502 2.5502 3.0502 Barrier thickness (cm) Figure 14 The relation of time to reach maximum allowable level a function of a barrier thickness for tetracosane as initial concentration of 0.84ppm SUMMARY AND CONCLUSIONS For isopropanol, the diffusion coefficient in HDPE film was determined by the permeation method as a function of vapor activity. For tetracosane, which is an extremely low vapor pressure solid at room temperature, the diffusion coefficient was determined via the migration method. The diffusion coefficient of isopropanol in HDPE film at room temperature showed a rather constant value within the vapor concenuation range evaluated. An assumption of a constant diffusion coefficient, which was defined beforehand, would therefore not be in serious error for use in migration modeling expressions. The partition coefficient of isopropanol between HDPE film and water at room temperature showed a very low level of isopropanol sorbed by HDPE film, giving an equilibrium partition coefficient value of K [/p = 600. For the tetracosane, the diffusion coefficient was determined fiom a migration study rather than from a permeation method. Only one concentration level of tetracosane in HDPE film was used to perform a migration experiment and to estimate the diffusion coefficient. The calculated diffusion coefficient was an estimation value or an integral diffusion coefficient, since it was determined by assuming a constant diffusion coefficient value. In addition, since hexane is a polymer swelling solvent for HDPE, the calculated diffusion coefficient is assumed to be an overestimated value and would be considered an upper bound value. For a non- interactive contact phase such as water, the diffusion coefficient for tetracosane in HDPE may be several order of magnitude lower. By using the migration study to determine the 89 90 diffusion coefficient, the partition coefficient was determined simultaneously. Further, the migration study also provided a real condition of food simulant/packaging interaction. For tetracosane, even using a polymer swelling solvent as a food simulant, the partition coefficient still afforded a significant level of migrant distributed to the HDPE film. By knowing the diffusion coefficient and maximum allowance of contaminants level in food content, the migration amount of isopropanol and tetracosane from HDPE film into a food contact phase was estimated, based upon a diffusion controlled process, by appropriate mathematical expressions for each system. For a single layer of contaminated HDPE film, the total amormt of isopropanol in film would migrate out in less than 6 hours, while it would take about 2 months for tetracosane. Thus, this contaminated film would be considered as unacceptable for use direct contact with food. In addition to a simple single layer structure, application of a functional barrier to prevent migration was also evaluated. The laminate considered, was composed of a contaminated HDPE layer and a virgin layer of the same HDPE resin. Based on the assumption that migration of contaminants into a food contact phase (liquid food) is a solely diffusion controlled process, the migration of isopropanol and tetracosane from a polymer film to a fluid contact phase was estimated separately, for a single layer HDPE film and a laminated HDPE film by using the diffusion coefficient values determined for the respective migrants in HDPE film. Three migration models, namely: (i) a simple model (Model I) for a single layer film; (ii) the Begley and Hollifield model (Model II); and (iii) the Lauobi and Vergnaud model (Model 1H), were evaluated to estimate the amount of migration. The simple model (Model I) was used to estimate migration levels from a single layer structure, while Models II and III were used to estimate migration flour a two 91 layer laminate structure. An evaluation of these models was made and the effectiveness of a functional barrier layer (virgin HDPE layer) to prevent the migration of both compounds was also determined. For the isopropanol, a barrier thickness of 8 mil is required to retard complete migration of the minimum initial concentration of 0.84 ppm for up to 3 months. Such a thickness would be considered a polymer sheet, rather than a film phase. In contrast, the functional barrier showed an effectiveness to prevent migration. For example, only 1 mil of the barrier thickness could increase the totally migration time to 3 month. Even though, isopropanol showed a rapid migration rate from the laminate containing a functional barrier, the concept of the application of a functional barrier could have validity if the virgin layer was a high barrier resin such as a polyamide. This study demonstrated the potential of the application of recycled plastic for use as a food packaging material, as a laminate film, if the food contact layer is a good barrier for organic substances. However, the migration of a variety of contaminants to various food systems and end use conditions has to be studied, especially actual migration studies from lamination films, since the interaction of food components with the polymer and severe or abusive storage conditions could accelerate the migration rate. In such cases a highly compatible solvent could swell the polymer matrix, resulting in a change in the mechanism for the migration of contaminants. Further, higher temperature conditions may increase the diffusion coefficient value, which is a major mechanism involved with the migration process. It would be a risk to make a decision regarding the safety of such structures for food contact without a good understanding of such variables. RECOMMENDATIONS FOR FUTURE STUDIES In order to investigate whether the use of recycled HDPE plastic, as a food packaging material, is safe, several factors need to be considered. Based on this study, there are a number of questions and concerns related to the migration of dangerous substances from post consumer recycled plastic. One important subject is that the selected chemical surrogates are appropriate or representative of possible contaminant substances in recycled HDPE plastic. This leads to a suspicion of the method for evaluating the use of recycled HDPE plastic, if its safety will be evaluated only fi'om migration studies of selected surrogates. This study also showed that recent migration models can only provide an estimation of the migration rate since there is a limit due to assumptions made in their solution. Future studies proposed would include actual migration studies involving recycled HDPE obtained from several sources, to a variety of actual food systems. For the application of recycled HDPE plastic in the form of a lamination film, the effectiveness of a barrier layer to prevent the migration should be evaluated based on the migration studies of real laminated film. This actual migration studies will also be used to validate the accuracy of mathematical expressions. The tasks of the actual migration studies will be useful for the future development of mathematical models. 92 APPENDICES 95 Table 14 Standard calibration of isopropanol for GC-MS Concentration Quantity injection“) Area response (2) (ppm) (31 (AU) 0 0.0 0 5 4.4E-09 745908 10 8.7E-09 1555413 50 4.4E-08 7260250 100 ' 8.7E-08 14966860 m. For 3 ul volume injection (2). Average values 16000000 14000000 4* y = 6.3E+13x 2 = 5‘ 12000000 . R 1.05+00 < " 10000000 3 8000000 r § 6000000 «~ g 4000000 x Experiment 2°°°°°° ‘” ———Trendline o I. : a T 0.0E+00 5.05 1.0E-v07 1.5Ev07 ZOE-07 2.5E-07 Quantlty Injection (9) Figure 17 Standard calibration curve of isopropanol for GC-MS 96 Table 15 Standard calibration of isopropanol for HP 6890 by SPME Concentration Mass Area response (ppm 1 (3) (AU) 0.5 4.0E-07 471“) 5 4.0E-06 4572(2) 50 4013-05 45320 100 7.9E-05 92024 200 1.6E-04 182664 (13.0). Estimation values 200000 180000 .. .. 160000 .. = 140000 4 1200001 100000 - T r “Reopens“ 60000 1» 40000 ~- 20000 4. y = 9.1E+02x R’ = 1.0E+00 x Experiment ——Trendline 100 ConcenmfloMppm) 150 200 Figure 18 Standard calibration curve of isopropanol for HP 6890 by SPME 97 Table 16 Standard calibration of hexane for HP 5890A Concentration Quantity injection“) Area response (2) (PPM) (3) (AU) 0 0.00 0 0.5 3.9E-10 312.5 5 3.9E-09 6186.5 10 7.8E-09 14534.5 25 2.0E-08 32747.5 50 3.9E-08 67102 ‘1’. For 1 ul volume injection ‘2). Average values 70000 s 1. y 9"" 1.7E+12X 5 50000 .. R2 = 1.0E g 40000 4 a 3‘ “ n "' «>- 20000 1 X _ 1°°°° ‘* ——Trendline 0 5E-09 1E—08 1.5E-08 2E-08 2.5E-08 3E-08 3.5E—08 4E-08 Quantity of injection (9) Figure 19 Standard calibration curve of hexane for HP 5890A APPENDIX II Statistical Analysis for Permeation Experiments Statistic descriptive of determined diffusion, permeability and solubility coefficient either by isostatic or quasi-isostatic procedure were presented below. Table 17 Statistical descriptive of diffusion coefficient determined by isostatic Concentration Sample number Mean Standard Standard Error Deviation 0.10 4 3.6E-14 2.7E-15 1.4E-15 0.21 4 3.8E-14 1.7E-15 8.6E-16 0.38 6 3.8E-14 3.6E-15 1.5E-15 Total 14 3.7E-14 2.9E—15 7.7E-16 Table 18 Statistical descriptive of permeability coefficient determined by isostatic Concentration Sample number Mean Standard Standard Error Deviation 0.10 4 4.6E-09 6.0E—10 3.0E-10 0.21 4 4.6E-09 4.7E-10 2.3E-10 0.38 4 4.6E-O9 2.25-10 1.1E-10 Total 12 4.6E-09 4.2E-10 1.2E-10 Table 19 Statistical descriptive of solubility coefficient determined by isostatic Concentration Sample number Mean Standard Standard Error Deviation 0.10 4 1.6E-07 1.6E-08 8.2E-O9 0.21 4 1.5E-07 2.3E—08 l .lE-08 0.38 4 1.5E-07 1.7E-08 8.4E-O9 Total 12 1.6E-07 1.8E-08 5.1E-09 98 Table 20 Statistical descriptive of diffusion coefficient determined by quasi-isostatic 99 Concentration Sample number Mean Standard Standard Error Deviation 0.19 3 4.2E-l4 4.6E-15 2.6E-15 0.27 2 4.4E-14 1.3E-14 8.9E-15 0.41 4 4.0E-l4 LIE-30 5.3E—31 Total 9 4.1E-14 5.4E-15 1.8E-15 Table 21 Statistical descriptive of permeability coefficient determined by quasi-isostatic Concentration Sample number Mean Standard Standard Error Deviation 0.19 3 3.9E-09 1.7E-09 9.7E-10 0.27 2 4513-09 3.5E-10 2.5E-10 0.41 4 4.4E-09 2.6E-10 1.3E-10 Total 9 4.2E-09 9.1E-10 3.0E—10 Table 22 Statistical descriptive of solubility coefl'icient determined by quasi-isostatic Concentration Sample number Mean Standard Standard Error Deviation 0.19 3 1213-07 6.0E-08 3.4E-09 0.27 2 1.3E-07 3.0E-08 2.2E-08 0.41 4 1.4E-07 8.1E-09 4.0E-09 Total 9 1.3E-07 3.3E-08 1.1E-08 100 For each test procedure, the analysis of variance method (ANOVA)was used for evaluating whether there was a difference of a mean of determining parameters (diffusion, permeability and solubility coefficient) among each testing concentration or not. The results of AN OVA for each parameter were presented below. Table 23 The ANOVA for diffusion coefficient determined by isostatic Group Sum of squares Degree of Mean square F F table freedom, 111' Between PIE 1.265E-29 2 6.232E-30 0.724 3.98 Within group 9.473E-29 11 8.612E-30 Total 1.074E-28 13 Table 24 The AN OVA for permeability coefficient determined by isostatic Group Sum of squares Degree of Mean square F F table freedom, df Betweermoup 8.867E-21 2 4.423E-21 0.021 4.26 Within group 1.889E-18 9 2.098E-19 Total 1.897E-l8 11 Table 25 The ANOVA for solubility coefficient determined by isostatic Group Sum of squares Degree of Mean square F F table - freedom, df Between group 2.462E-16 2 1.231E-l6 0.349 4.26 Withingroup 3.179E-15 9 3.532E-l6 Total 3.425E-1 5 1 l 101 Table 26 The ANOVA for diffusion coefficient determined by quasi-isostatic Group Sum of squares Degree of Mean square F F table freedom, df Between group 3.401E-29 2 1.701E-29 0.510 5.14 Within Qgroup 2.000E-29 6 3.334E-29 Total 2.340E-28 8 Table 27 The ANOVA for permeability coefficient determined by quasi-isostatic Group Sum of squares Degree of Mean square F F table freedom, df Between group 6.414E-19 2 3.207E-l9 0.325 5.14 Within group 5.92813—18 6 9.880E-19 Total 6.569E-18 8 Table 28 The AN OVA for solubility coefficient determined by quasi-isostatic Group Sum of squares Degree of Mean square F F table freedom, df Between group 6.608E-16 2 3.304E-16 0.242 5.14 Within group 8.206E-15 6 1.368E-15 Total 8.867E-15 8 102 A two way analysis of variance method (AN OVA)was also used for evaluating whether there was a difference of a mean of determining parameters (diffusion, permeability and solubility coefficient) due to test procedures and concentration or not. In order to have an analysis, it was assumed that a concentration of quasi-isostatic was similarto the one for isostatic procedure. The results of two way ANOVA were presented below. Table 29 Two way ANOVA of test techniques and concentration effects for diffusion coefficient Sum of df Mean F Sig Squares Square Main effects ( Combined ) 8.503E-29 3 2.834E-29 4.170 .025 Concentration 6.805E-29 2 3.403E-29 5.006 .022 Techniques 8.332E-30 1 8.332E-30 1.226 .286 2-Way Interactions Concentration 1.840E-29 2 9.198E-30 1.353 .288 " Techniques Model 9.459E-29 5 1.892E-29 2.783 .057 Residual 1.019E-28 15 6.797E-30 Total 1.965B-28 20 9.827E-30 Table 30 Two way AN OVA of test techniques and concentration effects for permeability coefficient Sum of (If Mean F Sig Squares Square Main effects ( Combined ) 2.254E-l 8 3 7.513E-l9 1.464 .264 Concentration 5.427E-19 2 27141349 .529 .600 Techniques 1.962E-18 l 1.962E-18 3.824 .069 2-Way Interactions Concentration 3.765E-19 2 1.883E-19 .367 .699 " Techniques Model 2.396E-18 5 4.793E-19 .934 .487 Residual 7.697E-18 15 5.131E-19 Total 1 .009E-17 20 5.047E- l 9 103 Table 31 Two way ANOVA of test techniques and concentration effects for solubility coefficient Sum of df Mean F Sig Squares Square Main effects ( Combined ) 3.5 1913-15 3 1.173e-15 1.682 .213 Concentration 2.863E-16 2 1 .43 115-16 .205 .817 Techniques 3.098E-15 1 3.098E-15 4.443 .052 2-Way Interactions Concentration 6.188e-16 2 3.094E-l6 .444 .650 " Techniques Model 3.881E-15 5 7.761E-l6 1.113 .395 Residual 1.046E-14 15 6.974E-16 Total 1.434E-14 20 7.170E-16 APPENDIX [11 Migration Study of Isopropanol Spiked HDPE Film into Water by the SPME All parameter for prediction a migration amount of isopropanol from spiked film into water were presented below. Referring to equation 12 M! M' = ’ aKC; = (1 - exp(Z2 )erfi'Z) Calculation for the thickness of food content a: V L a " A where: V1, = Volume of food content ( water) 40 cm3 A = Food / film contacting area ( surface area of film 30 disks ) 2.217 cm2 So, a = 0.18 cm Calculation for an initial concentration of isopropanol in spiked film C 2 K — C, Where: K = Partition coefiicient (600) CL = Concentration of isopropanol in water at equilibrium of migration (1.8 x 10'3 g/cm’) So, Cp = Concentration of isopropanol in polymer at equilibrium of migration (2.9 x 1045 g/cm3) The initial concentration of isopropanol in film is a sum of a concentration of isopropanol in liquid and polymer at equilibrium (1.8 x 10‘3 g/cm3). 104 APPENDIX IV Migration Study of Tetracosane Spiked HDPE Film into Hexane by the GC-FID All parameter for prediction a migration amount of tetracosane fi'om spiked film into hexane were presented below. Referring to equation 12 M = aKC; = (1 - exp(22)erch) M ,' Calculation for the thickness of solvent ( hexane ), a: _ KL. 0 " A where: VL = Volume of solvent (hexane) in migration cell A = Food / film contacting area (surface area of film disks) So, a = thickness layer of solvent The equilibrium partition coefficient of tetracosane for spiked film/hexane system could be determined by equation: K = -— CP (MmP / M P ) Where: K = Equilibrium Partition coefficient C1, = Concentration of migrant in liquid phase at equilibrium Cp == Concentration of migrant in polymer phase at equilibrium MM = Mass of migrant in liquid phase at equilibrium MM == Mass of migrant in polymer phase at equilibrium ML = Mass of liquid phase Mp = Mass of polymer phase The initial concentration of tetracosane in film is a ratio of a sum of the amount of migrant in liquid phase and polymer phase at equilibrium to a volume of film disks. 105 106 A comparison of experiment data of migration and model fit data, by equation 12, have to make to estimate the diffusion coefficient of tetracosane in HDPE film. The results of matching data were presented below. Table 32 The comparison of migration of tetracosane between the experiment and the model fit for sample 1 Square root of time M. I M... m M. / M... ‘2’ seem (experiment data ) ( model fit data ) 0 0.00 0.00 37.95 0.042 0.051 55.78 0.070 0.073 70.65 0.093 0.091 100.99 0.13 0.13 110.62 0.15 0.14 148.55 0.21 0.18 248.51 0.25 0.27 324.24 0.29 0.33 785.41 0.34 0.56 993.97 0.36 0.61 1402.42 0.37 0.70 1463.76 0.39 0.69 1525.30 0.40 0.69 1603.34 0.40 0.64 ““2" The ratio ofthe amount ofmigration at time t to the initial amount in film. Table 33 The comparison of migration of tetracosane between the experiment 107 and the model fit for sample 2 Square root of time M. l M... (I) M. l M... (2) seem (experiment data ) ( model fit data) 0 0.00 0.00 40.99 0.038 0.034 75.61 0.063 0.060 89.77 0.069 0.070 101.75 0.074 0.079 1 15.99 0.085 0.089 143.06 0.094 0.11 185.67 0.14 0.14 195.26 0.21 0.14 202.13 0.22 0.15 282.86 0.24 0.20 "1 (2" The ratio of the amount of migration at time t to the initial amount in film. Table 34 The comparison of migration of tetracosane between the experiment and the model fit for sample 3 Square root of time M. / M... (I) M. I M... (2) seem (experiment data ) (model fit data 1 0 0.00 0.00 34.64 0.043 0.03 264.56 0.21 0.21 398.79 0.30 0.29 875.63 0.32 0.49 1010.53 0.33 0.53 "“2" The ratio of the amount of migration at time t to the initial amormt in film. Table 35 The comparison of migration of tetracosane between the experiment 108 and the model fit for sample 4 Squarerootoftime Mrlem M./M...m see"2 (experiment data 1 1 model fit data 1 0 0.00 0.00 42.43 0.058 0.041 264.61 0.22 0.22 572.22 0.38 0.39 644.44 0.42 0.42 792.80 0.42 0.48 895.87 0.42 0.51 983.92 0.45 0.53 1028.04 0.45 0.54 (11(2) The ratio ofthe amount ofmigrationattimetto the initial amormt in film. Table 36 The comparison of migration of tetracosane between the experiment and the model fit for sample 5 Square root oftime M./M...“’ M./M...m secm (experiment data ) (model fit data) 0 0.00 0.00 397.95 0.38 0.30 494.34 0.37 0.36 572.29 0.38 0.39 648.48 0.36 0.42 725.99 0.39 0.46 824.58 0.42 0.49 (”‘2’ The ratio ofthe amount ofmigration at time t to the initial amount in film. 109 Table 37 The comparison of migration of tetracosane between the experiment and the model fit for sample 6 Square root oftime M./M...m M./M...a) sec"2 (experiment data ) ( model fit data ) 0 0.00 0.00 263.94 0.18 0.17 398.27 0.28 0.24 494.23 0.29 0.28 572.58 0.32 0.32 770.04 0.32 0.39 846.33 0.40 0.41 938.61 0.42 0.44 1355.99 0.42 0.54 (11(2). The ratio of the amormt of migration at time t to the initial amount in film. 110 0.3 0.25 4 0.2 .. r 3 0.15» 0.1 4 0.05 1~ 0 50 100 150 200 250 Square rootoftlme ( a“) Figure 20 The profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 1 111 0.25 0.2 1~ 0.15 1~ It’ll“ 0.1 «r 0.05 ~~ 0 it 1 1 e. a 0 50 100 150 200 250 300 Square rootottlmeta'") Figure 21 The profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 2 112 0.5 0.45 «~ 0.4 4 0.35 -. 0.3 «~ 3 0.25 «~ 1’ 0.2 « 0.15 ~~ 0.1 .. 0.05 1* 0 200 400 600 800 1000 Squarerootoftlmeu‘“) Figure 22 The profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 3 113 0.6 0 200 400 600 800 1000 Square rootoftlme ( a“) Figure 23 The profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 4 114 0.5 0.45 1r 0.4 1+ 0.35 ~~ 0.3 ~~ g 0.25 4» 0.2 1- 0.15 J» 0.1 ._ 0.05 .1 0 1 1 1 1 1 1 1 0 100 200 300 400 500 600 700 800 Squarerootoftlmfls'”) Figure 24 The profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 5 115 0.6 0.5 -r 0.4 .. 0.3 -- "Jun 0.2 1~ x Experiment —Model fit 0.1 ~- O 200 400 600 800 1000 1200 1400 Square root oftime ( a“) Figure 25 The profile of the comparison of migration of tetracosane between the experiment and the model fit for sample 6 APPENDIX V The Comparison of the Predictive Migration Levels by Models II and III The predictive migration levels of isopropanol and tetracosane from a laminated film of a 1.15 mil contaminated HDPE film and virgin HDPE films of 1,2,4,6,8 and 10 mil were calculated by using Model 11 and III as shown in the Tables 38-41. Table 38 The migrated amount of isopropanol from a laminated film with the barrier thickness of 1,2 and 4 mil based on the predictive Models 11 and III Time Barrier of 1 mil Barrier of 2 mil Barrier of 4 mil (sec) Model II Model 111 Model [1 Model 111 Model 11 Model III (11pr (11pr (ppb) (0pr (ppb) (0pr 0.0 0.00 0.00 0.00 0.00 0.00 0.00 2.2E+4 1.4 0.63 0.23 0.20 0.00 0.01 4.3E+4 3.0 1.1 0.69 0.53 0.05 0.08 6.5E+4 4.7 1.3 1.2 0.79 0.13 0.20 1.0E+5 7.4 1.4 1.9 1.1 0.30 0.41 3.0E+5 22 1.5 6.3 1.5 1.4 1.1 3.5E+5 26 1.5 7.3 1.5 1.6 1.2 5.0E+5 38 1.5 11 1.5 2.5 1.4 8.0E+5 60 1.5 17 1.5 4.1 1.5 1.0E+6 76 1.5 21 1.5 5.2 1.5 1.2E+6 91 1.5 26 1.5 6.3 1.5 1.5E+6 1.1E+2 1.5 32 1.5 7.9 1.5 116 117 Table 39 The migrated amount of isopropanol from a laminated film with the barrier thickness of 6.8 and 10 mil based on the predictive Modeb II and III Time Barrier of 6 mil Barrier of 8 mil Barrier of 10 mil (sec) Model 11 Model 111 Model 11 Model 111 Model 11 Model III (ppb) (ppb) (0pr (ppb) 1011b) (ppb) 0.0 0.00 0.00 0.00 0.00 0.00 0.00 2.5E+5 0.36 0.56 0.12 0.27 0.04 0.12 5.0E+5 0.96 1.0 0.43 0.70 0.20 0.43 7.0E+5 1.4 1.2 0.70 0.93 0.36 0.65 8.0E+5 1.7 1.3 0.84 1.0 0.45 0.74 9.0E+5 1.9 1.4 0.97 1.1 0.53 0.83 1 .0E+6 2.2 1.4 1.1 1 .2 0.62 0.90 2.0E+6 4.6 1.5 2.5 1.5 1.5 1.3 3.0E+6 7.0 1.5 3.8 1.5 2.4 1.5 Table 40 The migrated amount of tetracosane from a laminated film with the barrier thickness of 1,2 and 4 mil based on the predictive Modeb II and III Time Barrier of 1 mil Barrier of 2 mil Barrier of 4 mil (sec) Model 11 Model 111 Model 11 Model III Model 11 Model 111 (ppb) Ippb) (ppb) (MM (99'!) (1)pr 0.0 0.00 0.00 0.00 0.00 0.00 0.00 5.0E+6 0.85 0.41 0.10 0.08 0.00 0.00 8.0E+6 1.5 0.66 0.25 0.22 0.00 0.01 1.0E+7 1.9 0.80 0.37 0.31 0.1 0.02 3.0E+7 6.2 1.4 1.6 0.98 0.22 0.32 1.3E+8 28 1.5 7.8 1.5 1.8 1.2 1.5E+8 32 1.5 9.0 1.5 2.1 1.3 3.0E+8 64 1.5 18 1.5 4.4 1.5 4.0E+8 86 1.5 24 1.5 5.9 1.5 5.0E+8 1.0E+2 1.5 31 1.5 7.5 1.5 6.0E+8 1.3E+2 1.5 37 1.5 9.0 1.5 7.0E+8 1.5E+2 1.5 43 1.5 10 1.5 118 Table 41 The migrated amount of tetracosane from a laminated film with the barrier thickness of 6,8 and 10 mil based on the predictive Models 11 and III Time Barrier of 6 mil Barrier of 8 mil Barrier of 10 mil (sec) Model 11 Model [11 Model 11 Model 111 Model 11 Model 111 (ppb) (ppb) (ppb) (ppb) (ppb) (ppb) 0.0 0.00 0.00 0.00 0.00 0.00 0.00 5.0E+6 0.00 0.00 0.00 0.01 0.00 0.03 8.0E+6 0.00 0.00 0.00 0.02 0.00 0.08 1.0E+8 0.44 0.64 0.16 0.34 0.05 0.16 2.0E+8 1.1 1.1 0.52 0.78 0.25 0.51 3.0E+8 1.8 1.3 0.90 1.1 0.49 0.79 5.0E+8 3.2 1.5 1.7 1.3 0.98 1.1 8.0E+8 5.2 1.5 2.8 1.5 1.7 1.4 1.5E+9 10 1.5 5.5 1.5 3.4 1.5 2.0E+9 13 1.5 7.5 1.5 4.7 1.5 3.0E+9 20 1.5 11 1.5 7.2 1.5 5.0E+9 34 1.5 19 1.5 12 1.5 BIBLIOGRAPHY BIBLIOGRAPHY Baner, A.L., 1993, “Partition coefficient of aroma compounds between polyethylene and aqueous ethanol and their estimation using UNIF AC and GCFEOS,” Doctoral Thesis, Michigan State University, East Lansing, MI. Begley, T.H., Hollifield, HQ, 1993, “Recycled polymers in food packaging migration considerations,” Food Technology, November, 109-112 Boyd-Boland, A.A., Pawliszyn, J .B., 1995, “Solid-phase microextraction of nitrogen-containing herbicides” J. Chromatogr. A, 704, 163-172 Crank, J ., 1975, The Mathematics of Dzfiitsion, 2"d ed.,Clarendon Press, Oxford. Duda, J .L., Zielinski, J. M., 1996, “Free-Volume Theory,” Dlflitsion in Polymer, Marcel Dekker Inc., New York, 143-169. Dynatherm Analytical Instruments, Inc, 1989, “Manual for Thermal Desorption Unit 890,” Kelton, PA. FDA, 1992, Point to consider for the use of recycled plastics in food packaging: chemistry considerations, Food and Drug Administration, Indirect additive branch, HFS-216, Washington, DC. FDA, 1993, “Food additives: Threshold of regulation for substances used in food-contact articles: 21 CFR.,” Federal Register, 58, 52 71 9-52 729. Figge, K., 1988, “Dependence of the migration out of mass plastics on the thickness and sampling of the material,” Food Additives and Contaminants, 5,(1), 397-420. Gavara, R., Hernandez, R.J., Giacin, J .R., 1956, “Methods to determine partition coefl‘icient of organic compounds in water/polystyrene systems,” J. of Food Science, 61, (5), 947-952 Hernandez, 11.1., Giacin, JR, and Bauer A.L., 1986, “The evaluation of the aroma barrier properties of polymer film,” J. of Plastic Film & Sheeting, 2, 187-211 119 120 Huang, S.J., 1996, “Evaluating isostatic and quasi-isostatic procedures for determining the organic vapor barrier properties through polymer membranes,” Master Thesis, Michigan State University, East lansing, MI. Komolprasert, V., Hargraves, W. A., Armstrong, D. J ., 1994, “Determination of benzene residues in recycled polyethylene terepthalate (PETE) by dynamic headspace-gas chromatography,” Food Additives and Contaminants,11, 605-614. Kumins, C.A., Kwie, T.K., 1968, “Free volume and Other Theories,” Dzfiitsion in Polymers, Academic Press, London, 107-139. Laoubi S., Vergnaud, J.M., 1995, “Process of contaminant transfer through a food package made of a recycled film and a functional barrier,” Packaging Technology and Science, 8, 97-110 Limm, W., Hollifield, HQ, 1996, “Modeling of additive diffusion in polyolefins,” Food Additives and Contaminants, 13, 949-967 Nalan, L., 1996, “Solid phase extraction IHPLC analysis of acidic herbicides in drinking water,” The Reporter, Supelco, 15, 8-9. Neilsen, T., 1994, “Aroma sorption by food packaging polymer”, Dissertation, University of Delaware, USA. Neogi, P., 1996, “Transport phenomena in polymer membrane,” Diflitsion in Polymer, Marcel Dekker Inc., New York, 1 73-205. NFPA, and SP1, 1994, Guide for the safe use of recycled plastics for food packaging application, National Food Processors Association,Washington, DC. Profelhof, R. C., Throne, J. L., 1993, Polymer Engineering Principles, Hanser, Munich, 364-366. Reid, R. C., Sidman, K. R., Schwope, A.D., Till, D.E., 1980, “Loss of Adjuvant from Polymer Films to Food Simulants, Efi‘ect of External Phase,” Ind Eng. Chem. Prod. Res. Dev., 19, (4),580-587. Rogers, C.E., 1985, “Permeation of gases and vapors in polymers,” Polymer Permeability, Elsevier, New York. Rudnick, J ., Taylor, P.L., 1979, “Theory of free volume in polymers,” J. of Polymer Science, Polymer Physics,17, 31 1-320. 121 Selke, S., 1996, “Recycling of plastic materials,” Hand book of plastics, elastomers and composites, McGraw-Hill Inc., New York, 11.1-11.31. Sharma, R., 1997, “Determination of diffusion and prediction of migration of xylene and 2,4-dichlorophenol through high density polyethylene,” Master Thesis, Michegan State University, East Lansing, MI Shirey, R., 1997, “SPME/capillary GC analysis of solvents from water at low ppb levels,” The Reporter, Supelco, 16, (2), 1997, 6 US. Environmental Protection Agency, 1994, Characterization of Municipal Solid Waste in the United States: 1994 Update, EPA530-R-94-042, Washington, DC. Van Krevelen, D.W., 1990, Properties of Polymers, 3d ed., Elsevier, New York. Zhang, Z., Yang M.J., Pawliszyn J., 1994, “Solid-phase microextraction” J. Anal. Chem, 66, (17), 844-852 Zhang, Z., Yang, M.J., Pawliszyn, J., 1994, “Solid-phase microextraction,” Analytical Chemistry, 66, 844-853 "I11111111111111.7110