1““? F a; “mu W1“ 1" h ' 15.3" o 3 ' ‘ I .5? 1 5i '4. 45‘. ,2: ._-'. ’3’.le 3 " *1“ a? ' .. ir'l. ". l .. ‘ ‘ n v i :3)- 1. ya 1: ‘t ‘1 A): 2 .h 0.; mv‘ ”V... s rofifi'i‘.’ . . no... ”3225: " flag; ' .. o .- ,.- ;,,".‘.' .x .mfi‘al J“ '3. ‘;'.’:y.:u—- 7 ‘. .'. . ‘ " .’.' ‘) ‘- L ‘W . ”:E‘JE ‘._ .1 , D 1, 3 Vv c; 1 ‘ » mfg: 5‘ ‘ 3. iv: 1,1. . 9 _ _ 3‘1} )X . 1‘ ' . fl ., 9. 3,. . A“: . ‘ , 4 , ih .. c , S ‘1 . 33-22" 9:53.32; . w- - I; "ifh‘f 1x5"?! ,93‘ " ‘ RSITY UBRARIES 3 1111111111111151 ( 3 312930 l? "i J l This is to certify that the dissertation entitled REMOVAL OF PCBS FROM WATER AND PEANUT OIL: STRUCTURAL INTERRELATIONSHIPS OF PCB CONGENERS AND SELECTED POLYMERIC PACKAGING MATERIALS presented by Melvin Arthur Pascall has been accepted towards fulfillment of the requirements for Ph . D. degree m Food Sci ence/Envi ronmental Toxicology Date January 8, 1996 MSUi: an Affirmative Action/Equal Opportunity Institution 042771 LIBRARY Michigan State University V. PLACE ll RETURN BOX to romovo this chockout from your rocord. TO AVOID FINES rotum on or bdoro duo duo. DATE DUE DATE DUE DATE DUE MSU Is An Mariam Aotla'VEquol Opportunity Institution W pus-9.1 REDUCTION OF PCBS FROM WATER AND PEANUT OIL: STRUCTURAL INTERRELATIONSHIPS OF PCB CONGENERS AND SELECTED POLYMERIC PACKAGING MATERIALS BY Melvin Arthur Pascall A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Food Science and Human Nutrition 1995 ABSTRACT REDUCTION OF PCBS FROM WATER AND PEANUT OIL: STRUCTURAL INTERRELATIONSHIPS OF PCB CONGENERS AND SELECTED POLYMERIC PACKAGING MATERIALS BY Melvin Arthur Pascall The kinetic characteristics and uptake capacity of three commercially available polymer (polyethylene (PE), polyvinyl chloride (PVC), and.polystyrene (PS)) films for the removal of polychlorinated biphenyls (PCBs) were evaluated. Solutions of congener specific PCBs in distilled water and peanut oil were prepared. These solutions were contacted with the polymer films within sealed glass vials with no head-space and in the absence of light. Change in concentration of PCB congeners, both in the polymers/aqueous phases and the peanut oil phase alone were determined as a function of storage time. Mass balance determinations showing the ratios of PCBs in the polymeric and liquid phases were used to validate the analytical methodology. From the experimental data generated, partition coefficients, diffusion, solubility and rate loss constants were calculated for the various films and liquids. Results indicated the sorption of PCBs by the polymers generally followed a Fickian diffusion. Polystyrene at 35%: showed a greater sorption than at 25°C, and at an increased material thickness at ZSWZ. The aqueous medium displayed a greater rate loss of PCBs than that of peanut oil. Results also showed that polyethylene exhibited the highest sorption diffusion and partition coefficients when compared to the other materials at similar experimental conditions. Although PVC indicated larger sorption diffusion and partition coefficients for the lower chlorinated congeners than polystyrene, a reversal of this trend was observed for the higher congeners. Lower chlorinated congeners displayed a larger rate loss than congeners with higher chlorination for all polymers in the aqueous phase. lHowever, in the peanut oil medium, the higher chlorinated congeners generally indicated a larger rate loss. Results from this study can be used to model possible transfer of environment contaminants from foods with polar and non polar characteristics to selected packaging materials. DEDICATED TO THE MEMORY OF MY LOVING MOTHER ACKNOWLEDGEMENT I wish to express my gratitude to Dr. Mary E. Zabik my major professor and chair of my guidance committee for her help, support, and nurturing during the course of my studies. I also acknowledge the contributions of the other members of my committee, without whom this study may not have been possible. To Dr. Matthew Zabik for his expertise in analytical chemistry, and Dr. Ruben Hernandez in polymeric packaging I am deeply grateful for your assistance. I am thankful to Drs. Bruce Harte, Denise Smith and Al Booren for their contribution in the areas of food science and for the help with putting together my dissertation. I am also thankful to Dr. Joshua Bagakas for his help with the statistical calculations. I am grateful for the support given to me by the members of my family. Thanks Ellis, Lionel, Eunice, Annette, Lloyd, Lindry and all my other extended relatives who encouraged me through the years while I was here at school. Finally, thank you God for health and strength to accomplish this task. ii TABLE OF CONTENTS CHAPTER 1 . . . . . . . . . . . . . . . . . . INTRODUCTION . . . . . . . . . . . . . . . . CHAPTER 2 . . . . . . . . . . . . . . . . . . LITERATURE REVIEW . . . . . . . . . . . . . . POLYCHLORINATED BIPHENYLS . . . . . . . Properties of Polychlorinated Biphenyls PACKAGING MATERIALS . . . . . . . . . . Polyethylene . . . . . . . . . . . Polystyrene . . . . . . . . . . . . Polyvinyl chloride . . . . . . . . TRANSPORT PROPERTIES THROUGH A MEMBRANE Pore and solute size . . . . . . . Temperature . . . . . . . . . . . . Clustering effect . . . . . . . . . Polarity and charge . . . . . . . . Concentration . . . . . . . . . . . Surface area (dimensional changes) An alternative diffusion equation . PARTITIONING STUDIES . . . . . . . . . . Bioconcentration factor . . . . . . Octanol/water partition coefficient Equilibrium partition coefficient (Kg) CHAPTER 3 EXPERIMENTAL . . . . . . .,. . . . . . . . . EXPERIMENTAL DESIGN . . . . . . . . MATERIALS . . . . . . . . . . . . . . . Glassware . . . . . . . . . . . . . 12 14 14 20 24 26 26 34 37 38 4O 45 46 5O 50 51 51 53 53 55 55 Polymers . . . . . . . . . . . . . Solvents . . . . . . . . . . . . Chemicals . . . . . . . . . . . . EXPOSURE OF FILM TO PCB CONGENERS . . PCB CONGENER ANALYSIS . . . . . . . . PCB EXTRACTION PROCEDURES Extraction of PCB from polymers . . . . . . . Extraction of PCB from oil solution Extraction of PCB from aqueous PCB'S CLEANUP PROCEDURES . . . . solution Gel Permeation Chromatography (GPC) . . Florisil/silica gel . . . . . PCB QUANTIFICATION . . . . . . . . DATA ANALYSIS . . . . . . . . . . . Mass balance determination . . Sorption diffusion coefficient determination . . . . . . Equilibrium partition coefficient (Kg) . Statistical calculations . . . CHAPTER 4 . . . . . . . . . . . . . . . . . . RESULTS . . . . . . . . . . . . . . . . . . . VALIDITY OF THE PCB ANALYTICAL METHODOLOGY . . . . POLYMERIC MATERIALS . . . . . . . . . . POLYETHYLENE . . . . . . . . . . . POLYSTYRENE . . . . . . . . . . . . POLYVINYL CHLORIDE . . . . . . . . EQUILIBRIUM PARTITION COEFFICIENT (k9) . . . STATISTICAL ANALYSIS . . . . . . . DISCUSSION . O O O O O C C O O O C C O O O 0 EFFECTS OF POLYMER TYPE, TEMP. & THICKNESS ON PCB UPTAKE FROM AQUEOUS AND OIL MEDIA CHAPT ER 5 C O O O O O O O O O O O O O O O O 0 SUMMARY AND SIGNIFICANCE OF THE STUDY . . . . iv 55 56 56 57 64 64 65 66 67 67 68 69 72 72 73 74 75 76 76 76 85 85 94 107 113 118 123 123 137 137 APPENDIX.......................142 BIBLIOGRAPHY.....................153 Table Table Table (Table Table Table Table Table Table Table Table Table Table Table Table Table 1. 2. 3. 4. 5. 9. 10. 11. 12. 13. 14. 15. 16. 17. LIST OF TABLES Characteristics of Aroclor" 1254 . . . . . . Properties of polystyrene. . . . . . . . . . Leman Molar Volume, VB . . . . . . . . . . . Leman Molar Volume, Vm . . . . . . . . . . . Calculated molar volumes VB and V3. for PCB congeners used in this study. . . . . . . . . Crystallographic data for polystyrene, PVC and polyethylene polymers used in this study. . . . . . . . . . . . . . . . . . . . The chemical structures of the PCB congeners used in this study. . . . . . . . . . . . . . PVC weights and exposure time at 25°C. . . . Polystyrene weights and exposure time at 25°C............‘......... Polyethylene weights and exposure time at 25°C..................... Polystyrene weights and exposure time at 35°C..................... 5 ply polystyrene weights and exposure time at25°C................... Mass balance ratios for PCB congeners in polyethylene/water. . . . . . . . . . . . . . Mass balance ratios for PCB congeners in polystyrene/water. . . . . . . . . . . . . . Mass balance ratios for PCB congeners in PVC/water . . . . . . . . . . . . . . . . . . Partition coefficient (Kc) for 1 ply materials stored at 25°C. . . . . . . . . . . vi 13 22 31 32 33 34 57 59 60 61 62 63 77 78 79 113 Table Table Table Table Table Table 18. Partition coefficients for polystyrene stored at 25°C. and 35°C., and different thicknesses. . . . . . . . . . . . . . . . . 19. Summary of statistical data for polyethylene, PVC and polystyrene under the same conditions for PCB uptake. . . . . . . . . . . . . . . 20. Summary of statistical data showing areas of significant uptake differences between congener for polyethylene, PVC and polystyrene under the same conditions. . . . 21a. Summary of statistical data for polystyrene exposed at 25 and 35°C. for PCB uptake. . . 21b. Summary of statistical data for 1 ply and 5 ply thick polystyrene exposed at 25°C. for PCB uptake. . . . . . . . . . . . . . . . . . 22. Summary of statistical data for polyethylene exposed to water and peanut oil at 25°C. for PCB uptake. . . . . . . . . . . . . . . . . vii 114 118 119 120 120 121 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure LI§I_Q£_EI§QB§§ 1. Basic structure of polychlorinated biphenyls with the numbered carbon atoms and possible position of substituted chlorine atoms. . . . 2. The monomeric structure of polyethylene. . 3. The monomeric structure of polystyrene. . . . 4. The monomeric structure of polyvinyl chloride. . . . 5. Types of sorption separations . . . . . . . . 6. Gas chromatogram of standard PCB congeners . 7. Gas chromatogram of polyethylene/PCB extract before cleanup. 8. Gas chromatogram of polyethylene/PCB extract after cleanup. 9. Gas chromatogram of polystyrene/PCB extract before cleanup. 10. Gas chromatogram of polystyrene/PCB extract after cleanup. O C O O O O O O O O O O O O O 11. Gas chromatogram of PVC/PCB extract before cleanup. . . . 12. Gas chromatogram of PVC/PCB extract after cleanup. . . . igure 13. Weight of PCBs in water exposed to 1 ply PE Figure Figure Figure Figure Figure at 25°C. . . . 14. Weight of PCBs at 25° C. . . . 15. Weight of PCBs 16. Weight'of PCBs 17. Calculated and uptake by 1 ply 18. Calculated and in water exposed to 1 ply PE uptake by 1 ply PE at 25° C. uptake by 1 ply PE at 25° C. experimental isotherm for PCB PEat25°C. ........ experimental isotherm for PCB viii 10 15 21 24 41 8O 82 82 83 83 84 84 86 86 89 89 9O Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure uptake by 1 ply PE at 25° C. . . . . 19. Comparison of diffusion coefficients congeners by 1 ply PE at 25° C. . . . 20. Comparison of rate loss by congeners plyPEat25°C. 21. Percent PCBs retained in oil exposed plyPEat25°C. 22. Percent PCBs retained in oil exposed ply PE at 25°(L . . . . . . . . . . 23. Comparison of rate loss by congeners water and oil for 1 ply PE at 25° C. 24. Weight of PCBs in water exposed to 1 at25°C. 25. Weight of PCBs at25°C. .............. 26. Weight of PCBs uptake by 1 ply PS at 27. Weight of PCBs uptake by 1 ply PS at 28. Calculated and experimental isotherm uptake by1p1y PS at 25° C. . . . . 29. Calculated and experimental isotherm uptake bylply PS at 25° C. . . . . 30. Comparison of diffusion coefficients in water exposed to 1 congeners by 1 ply PS at 25°<:. . . . 31. Comparison of rate loss by congeners ply PS at 25° C. . . . 32. Percent congeners retained for 35°C................ 33. Percent congeners retained for 35°C................ 34. Weight of PCBs PS at 35. Weight of PCBS PS at 36. Calculated and experimental isotherm uptake bylply PS at 35° C. . . 37. Calculated and experimental isotherm uptake bylply PS at 35° C. . . . . 38. Percent congeners retained for 5 ply uptake by 1 ply uptake by 1 ply ix for 1 PS at PS at 35°<:. 35°<:. for PCB for PCB 93 at. 90 91 91 92 92 93 96 96 97 97 98 98 99 99 100 100 101 101 102 102 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 25° C. O O O O O C C O O O O O O O O O O O O 39. Percent congeners retained for 5 ply PS at 25° C. O O O O C O O O C O O O O O O O O O O uptake by 5 ply PS at 25° C. uptake by 5 ply PS at 25° C. experimental isotherm for PCB PSat25°C. ........ and experimental isotherm for PCB plyPSat25°C. of sorption coefficients for 40. Weight of PCBs 41. Weight of PCBs 42. Calculated uptake by 5 Calculated uptake by 5 and ply 43. 44. Comparison congeners by 1 and 5 ply PS at 259 C. and 1 ply PS at 25°C. 45. Comparison of rate loss for congeners in water for 1 and 5 ply PS at 259 C. and 1 ply PSat25°C................. 46. Weight of PCBs ply PVC at25°C. 47. Weight of PCBs ply PVC at25°C. 48. Weight of PCBs uptake by 1 ply PVC at 25°13. 49. Weight of PCBs uptake by 1 ply PVC at 25° C. 50. Calculated experimental isotherm for PCB uptakebyl PVCat25°C......... 51. Calculated and experimental isotherm for uptakebylplyPVCat25°C. of diffusion coefficients for in water exposed to 1 in water exposed to 1 and P1Y 52. Comparison congeners by 1 ply PVC at 25° C. . . . . . . 53. Comparison of rate loss by congeners for 1 plyPVCat25°C. 54. Comparison of sorption diffusion coefficients for congeners by 1 ply PE, PVC andPSat25°C. 55. Comparison.of rate loss by congeners for 1 PE ' PVC. and PS at 2 5° C O O O O O C O O O O 0 ply 103 103 104 104 105 105 106 106 108 108 109 109 110 110 111 111 112 112 56. Plot of partition coefficient vs molar volume of PCB congeners exposed to PE, PVC and PS stored X at 25°<2 . . . . . . . . . . . . . . . . . . 117 Figure 57. Plot of partition coefficient vs molar volume of PCB congeners exposed to 1 and 5 ply PS stored at 25 and 35°<2 . . . . . . . . . . . . . . . 117 Figure 58. Fickian and Non-Fickian sorption and desorption type curves. . . . . . . . . . . . . . . . . 125 xi QHAPTER 1 W Food safety is a consumer issue of the 1990's. The media often use scare tactics to state the prevalence of food safety problems. . Thus, there is a need for food scientists to understand what issues consumers perceive as important so they can proactively respond and thus be viewed as a credible source of accurate food safety information. According to a 1994 national consumer survey conducted by the Food Marketing Institute (FMI) , 89% of respondents said product safety was a very important or somewhat important factor in Selecting foods. When shoppers were asked to rank a list of potential health hazards in the FMI survey, an overwhelming 79% said residues, such as pesticides and herbicides, are a serious health hazard if taken in with food. This was followed by antibiotics in poultry and livestock (55%), irradiated foods (35%), nitrites in food (35%), additives and preservatives (23%), and artificial coloring (19%) (Food Marketing Institute, 1994). To many people, eating and drinking have become death- defying feats. No wonder sales of "organic" foods and bottled _J_._ 2 waters have surged to new heights (Toufexis, 1989) . This (consumer anxiety is heightened by sensational stories carried lay the news media. Coverage of such issues as Alar, stoked public fears of apples and pesticide use (Newton, 1990) . (:arpenter (1991) reported that the fear'of pesticides has been (exacerbated by frenzied publicity over celebrated cases, such as the "60 Minutes" report on the apple spray - alar (daminozide). In the winter of 1983-84, alarm about residues of ethylene dibromide (EDB) in grain led many states to pull cake mixes and other packaged foods from store shelves. In March 1989, panic over poisoned grapes resulted in a serious public food scare episode. This prompted the FDA to issue a recall of grapes after evidence of cyanide tampering had been discovered (Ravenswaay, 1990). While it may be desirable to shift the debate to a more scientific level, practical considerations dictate that individual processors do all they can to ensure their product's safety and their corporate reputation (Newton, 1990). The use of synthetic pesticides has increased dramatically since World War II. According to the Environmental Protection Agency, in 1988 more than 1 billion pounds of pesticides and related products were used in the 0.8.: herbicides (680 million pounds), fungicides (132 million pounds), insecticides (268 million pounds), and other related chemicals (70 million pounds). This quantity corresponds to more than 4 pounds of pesticides for each person in the U.S. (Huff and Haseman, 1991). ma tr in fEI an. et lac PCE to Con of mem . aanu the 1 3 Industrial contaminants have now become a serious cause of concern because of their prevalence and entry into the food chain. Billings ( 1991) reported that once released into the environment, these chemicals seek out the media (water, air soil, and biota) in which they are the most soluble. For example, trichloroethylene and benzene are most soluble in air and, consequently, inhalation is the principal means of human exposure . Dichlorobiphenyl trichloroethane (DDT) and P°1YChlorinated biphenyls (PCBS) are most soluble in organic matter and tend to be concentrated in biota and thus are transferred to humans through the food chain. PCBs, once used in a broad range of industrial products, are found at Particularly high concentrations in the fatty tissue of fish feeding in polluted waters, such as Lake Michigan, and in animal products from farms with contaminated silos (Jacobson et a1 - , 1989). They also noted that PCBs, polybrominated biphenyls (P835) and DDE are not readily excreted, except by lactatlion, and that placental passage has been documented for PCBS and PBBs, although much larger quantities are transferred to the infant postnatally in breast milk due to its high lipid content. Dioxin (2,3,7,8-tetrachlorodibenzo-p-dioxin, TCDD) is one of the most toxic man-made compounds known. TCDD is not manufactured commercially, it is a contaminant in the mal'lufacture of several chlorinated phenolic products including the herbicide 2,4,S-trichlorophenoxyocitic acid (2,4,5-T) and heMachlorophene (Umbreit et a1. , 1986) . TCDD have been found 4 at low levels in many populations around the world, according to Graham et a1. (1985) . Industrial accidents and environmental contamination have been suggested as possible sources of TCDD, and may be responsible for continued human exposure. Human exposure to polychlorinated dibenzo-p-dioxins ( PCDDs) and polychlorinated dibenzofurans (PCDFs) is a matter of great concern in many parts of the world (Svensson et al. , 1991). In studies of exposure to dioxins and dibenzofurans through the consumption of fish, Svensson et al. (1991) found that fatty fish such as herring and salmon from the Baltic Sea have been found to contain high levels of some PCDDs and PCDFs. They also reported similar levels of TCDD in fish from the Great Lakes . The Great Lakes form the world's largest body of fresh water, and the area around them is home to about 37 million people (Park, 1991). Millions of Canadians and Americans ciez>end on the Great Lakes for bathing and drinking water, fishing and recreation, commerce and other uses. Scientists haVe found increasing evidence that pollutants in the water are causing insidious neurological damage, particularly in c11:1 ldren, and infertility among adults (Park, 1991) . Writing in the Chemical and Engineering News April 23, 1990, Bette Ei-:ll.eman quoted the International Joint Commission's (IJC) IreE>olymerization. There are three common commercial grades of 21 general-purpose polystyrene. These are easy flow, medium flow and high heat grades. High heat polystyrene is used mostly for extrusion applications such as films, whereas the other types are mainly used for injection molding (Hahnfeld and Sugden, 1994). POLYSTYRENE (PS) Figure 3. The monomeric structure of polystyrene. One should not confuse the term crystal polystyrene with the crystalline property of polystyrene. Crystal polystyrene is an amorphous polymer. It is clear and colorless. with excellent optical properties and high stiffness. It is brittle until biaxially oriented, at which time it becomes {comparatively flexible and durable. Twpical properties of laigh heat general purpose polystyrene include the following: 22 Table 2. Properties of polystyrene. Properties Measured values Specific gravity 1.0865 g/cm3 Tensile strength 56.6 MPa Flexural strength 82.75862 MPa Typical shrinkage 0.01143 cm/cm Boiling temperature (1;) 145%: Glass transition temp. (T3) 108%: Izod impact strength notch 24 J/m (Hahnfeld and Dalke, 1986) This type of polystyrene:has several advantages over the other types of polystyrene. Most of its advantages are due to its amorphous nature, which results in its excellent clarity and fabrication ease. This polymer needs little energy input for its fabrication because it has no crystalline melting point, and as such no heat of crystallization. Thus, its transition from glass to a viscous liquid is gradual. Crystal polystyrene finds many uses, especialLy where clarity is desirable. Applications for high heat polystyrene include foam extrusion for consumer electronics, expanded cushioning and food packaging - egg cartons, meat and poultry packaging trays; sheet extrusion and thermoforming for lighting, construction and decoration; film extrusion for oriented food packaging; injection molding for tape reels; and injection blow molding for packaging containers (Hahnfeld and Sugden, 1994) . Impact polystyrene is an amorphous polymer made from the 23 graft polymerization of styrene monomer with an elastomer, or a physical blend of polystyrene with an elastomer (typically polybutadiene). The resulting polymer is tough, usually white, and quite easily extruded and molded. The percentage of the elastomeric phase in such cases directly influences the toughness of the resulting resin. Therefore, the properties of impact polystyrene can vary significantly. For this reason, it is typically divided into three categories - medium impact (Izod <80.07 J/m); high impact (Izod 80.07 J/m); and super-high impact (Izod 138.788 to 266.9 J/m). Typical properties include flexural and tensile strength of 13.7931 to 48.27586 MPa, elongation of 10 to 60%, gloss of 5 to 100%, contact clarity from very good to poor, shrinkage of about 0.0152 cm/cm, and.coefficient.of thermal expansion in the same range as crystal polystyrene. It has a viscous softening of 102.3°C (Hahnfeld and Dalke, 1986). Impact polystyrene has found many applications and industrial uses because of its ease of processing, low cost and performance. Its major uses are for packaging and disposables, appliances, consumer electronics, toys, recreation and buildings materials. Its major food packaging uses.are for dairy containers, vending and portion cups, kits, ‘plates and bowls, and.disposables such as flatware, closures, safety razors and pens (Poloso, 1994) . The most common form of expandable polystyrene is made 13y blowing pentane gas into crystal polystyrene beads. This Inesults in a foamed product. It finds many uses from coffee 24 cups to automobile bumpers. Generally, its major end uses are for disposable drinking cups, cushioned packaging and in thermal insulation (Klepic, 1994). 01 n 1 ch oride The initial step in the industrial production of PVC begins with the reaction of chlorine with ethylene gases to produce ethylene dichloride. This is then dehydrohalogenated to yield vinyl chloride monomer, which is then polymerized to produce polyvinyl chloride, shown below in Figure 4. __l __i __i ___l__ I I I | __ H Cl H c1 n POLYVINYL CHLORIDE (PVC) Figure 4. The monomeric structure of polyvinyl chloride. Commercial PVC is produced by an addition polymerization. In this process, a free radical is formed byrdecomposition of an initiator. This free radical then forms an active center with the vinyl chloride monomer and thus initiates the 25 polymerization process. This continues until the chain is terminated. The average number of molecules in a typical chain has been estimated to be between 950 at a polymerization temperature of 122°F and 480 at a temperature of 158°F (Gabbett, 1994). The polymeric structure of PVC is partly crystalline (approximately 15%), this is mainly due to the regularity of its structure. The chlorine molecules are usually bonded in a head-to-tail alignment and is referred to as being syndotatic. The higher the processing temperature, the higher the level of irregularity of the polymeric structure, and thus the lower its percent crystallinity. In its pure form, PVC is very stiff and rigid (Brittleness - 20 elongation at break point in percent length) with little impact resistance (Izod 8 ft.-lb). To be made soft and pliable, plasticizers must be incorporated within its structure during its manufacture. Liquid plasticizers presumably gather in pockets between adjacent amorphous chain segments, giving them mobility to produce softness and flexibility. The use of plasticizers in PVC manufacture accounts for approximately 70% of the total plasticizer marker (Deanin, 1978). Thus, PVC is seldom used in its pure form. The proportion of plasticizers used directly relates to the Characteristics of the finish polymer. Although the plasticizers increase its machinability, this is at the 'eXpense of its cohesive density, and thus result in a reduction of its barrier properties. Collins and Daniels (L975) reported that the percent plasticizers in PVC can range 26 between 25.9 - 39.4%. PVC is a good barrier for oils, alcohols and petroleum solvents so long as a judicious use of plasticizers was employed in its manufacture. Rigid PVC has been known to retain odors, flavors, oxygen, moisture, and gases in general to a desirable extent, but may react with halogenated hydrocarbons. PVC has a glass-transition temperature (TB) of 87°C and a Melt temperature (Tm) of 212°C. Incorporation of plasticizers can lower these temperatures in proportion to the percentage and type incorporated. TRANSPORT PROPERTIES THROUGH A MEMBRANE Factors affecting the uptake of a compound by an adsorbent include: (1) the physical and chemical nature of the adsorbent, (2) the physical and chemical nature of the compound, (3) the concentration of the compound in contact with the adsorbent, (4) the characteristics of the phase in contact with the adsorbent, and (5) the resident time of the system. Pore and solute size The pore size of a polymer together with the size and shape of the solute will influence the passage of molecules through the membrane. The mobilities of the polymer's Segments and the penetrant molecules in a polymer-penetrant nuixture are primarily determined by the amount of free volume 1r) the system. This can be described by the following: 27 “B M, = Ad exp[—J] V: (1) Where MA is the mobility of the penetrant, vf is the average fractional free volume of the system, and Ad and 8,, are empirical free-volume parameters that are assumed independent of penetrant concentration and temperature. The parameter Ad depends on the size and kinetic velocity of the penetrant, while Bd is equivalent to the critical hole free volume necessary for a penetrant to make a diffusive jump (Cohen and Turnbull, 1959). This expression can be modified by arguing that crystalline material reduces the free volume in direct proportion to the amount of crystalline material present (Kreituss, 1981). Thus, the equation in (1) above will become: -B M, = Ad exp[ q) 3] (2) a f where 0, is the amorphous volume fraction of the penetrant free polymer at zero pressure and the temperature of the system. Other factors that also influence pore size include the flexibility of the polymer chain, the amount of membrane swelling, the extent of cross-linking and the temperature. Bixler and 'Sweeting (1971) studied this phenomenon by investigating the linear relationship between the infinite dilution solubility coefficient and the amorphous fraction in 28 polyethylene at 25°C for various gases and organic vapors. The results from this study supported the notion of the impenetrability of crystalline regions even by gas molecules. Bixler and Sweeting (1971) also showed that the solubility coefficient S1 of the gases and organic vapors are proportional to the volume fraction of the amorphous region 4), and the intrinsic solubility coefficient of the totally amorphous material S, where: = 4),, (3) After the molecule passes through a pore in the membrane surface it must negotiate passage through channels in the membrane. This is referred to as the tortuosity factor. The tortuosity of the corridor is influenced by the density of chain packing, which is decreased by the presence of bulky functional groups. The twofold effect of chain motion analogous to chemical cross-linking and tortuous barrier property has been studied by Michaels and Bixler (1961) who showed that both increase with increasing crystallinity of the membrane. This can be expressed mathematically to show the relationship of tortuosity O to diffusion: 29 D= a (4) where D and D,I are the diffusion coefficients in the actual sample and in a totally amorphous sample, respectively and where D‘ is the immobilization factor. As mentioned earlier, the size of the penetrant molecules can also affect passage through membranes. In the case of non-spherical compounds, the cross-sectional area and volume may be limiting factors. Hwang and Kamermeyer (1975) reported that molecular weight is not an effective description of molecular size and parameters such as the parachlor, permachlor, the space factor or molar volume can be used to correlate molecular properties with passage through a membrane. Heindorf ( 1992) reported that the parachlor (PC) is an empirical relationship between density and surface tension and is defined as: Pc= (—”—)t =2?“ (5) where t is the surface tension, D is the density in a liquid phase and d is the density of the molecule when in the gas phase, which is often negligible. The parachlor is also 30 phase, which is often negligible. The parachlor is also estimated by summing atomic parachlors (Pu) (Banerjee et al., 1995; Tulp and Hutzinger, 1978). Leman (1981) stated that since d in (5) above is much smaller than D, equation (5) can be rewritten as: 1/4 = (M0 ) or PC = Vol/4 (6) where a is surface tension and V is the molar volume. Parachlor values, in units of g1".cm3/(sec”2.mol), are generally in the range of 100-600. 'The permachlor concept that 'relates the polymers structure to its permeability by using a scheme of additive (+/-) structural components descriptive of the polymer was developed by Salame (1962) and Malachowski (1971). In experiments performed by them, they obtained good correlation between membrane permeability and the permachlor values for homologous series of penetrant compounds. The space factor concept was also derived by Malachowski (1971) to allow for the effects of temperature on the polymer. This space factor Q3f can be derived from the following: 31 T ansf=a+b—Tc+c— (7) where in. is the critical absolute temperature for the permeant, d is the molecular diameter of the solute and is temperature dependent, n is the space factor, T is the absolute temperature that.contributes to the solubility of the permeating species in the polymer and a, b, and c are constants that depend on the polymer. The molar volume (V) alluded to by Leman (1981) above, can be used to estimate the size of an organic compound (except organophosphates) from its chemical structure. Tables 3 and 4 list the contributions of selected atoms to the gaseous molar volumes which can be applied to a given chemical structure. Table 3. Gaseous Molar Volume (VB) of selected atoms in aromatic and heterocyclic rings structures. TIIhcrement Atom ( cm3 /mol) Ring size Increment C 16.5 Aromatic & heterocyclic -20.2 H 1.98 rings 0 5.48 CI 19.5 (Adapted from Leman,1981) 32 Table 4. Gaseous Molar Volume (vg')1of selected atoms in heterocyclic rings structures. Increment Atom (cm3/mol) Ring size Increment C 14.8 3-membered -6.0 H 3.7 4-membered -8.5 O 9.1 S-membered -11.5 C1 24.6 6-membered -15.0 Naphthalene -30.0 (Adapted from Leman,1981) From Leman's proposals as shown above, the gaseous molar volumes of the PCB congeners used in this study can be determined” This determination was done and is shown in Table 5. The size of the congeners increase with increasing chlorination of the biphenyl ring from tri-chloro to deca- chlorinated. 33 Table 5. Calculated molar volumes VB and V3' for PCB congeners used in this study. Calculated VB Calculated Va' Leman molar vol Leman molar vol Biphenyl Compound (cm3/mol) (cm3/mol) 2,4',5- Trichlorobiphenyl 229.96 245.00 2,2',4,4'- Tetrachlorobiphenyl 247.48 265.90 2,2',4,5',6- Pentachlorobiphenyl 265.00 286.80 2,2',3,3',4,4'-— Hexachlorobiphenyl 282.52 307.70 2,2',3,3',4,4',6- Heptachlorobiphenyl 300.04 328.60 2,2',3,3',4,5',6,6'- Octachlorobiphenyl 317.56 349.50 2,3,4,5,6,2',3',4',5'- Nonachlorobiphenyl 335.08 370.40 2,3,4,5,6,2',3',4',5',6'- Decachlorobiphenyl 352.60 391.30 (Adapted from Leman,1981) Similar values were also obtained by Spurlock and Biggar (1994) for a range of organic compounds including similar PCB congeners. Crystallographic data on polyethylene, PVC and polystyrene reported by Brandrup and Immergut (1975) in Table 6, is a useful guide to the volumetric space occupied by the polymeric structures of these materials. Due to the presence of the benzene rings in polystyrene, a tightly packed molecular structure) is impossible, hence the reason *why polystyrene is relatively permeable to gases and vapors usually encountered in food packaging and processing.o When compared to PVC or HDPE for example, the regularity, crystallinity and dipole/dipole moments (induced in PVC due to 34 the chlorine atoms) result in a closely packed polymeric structure that is much less permeable than polystyrene to the same penetrants under similar conditions. Table 6. Crystallographic data for polystyrene, PVC and polyethylene polymers used in this study. Crystal Unit Cell Parameter (A) Polymer System a b c polyethylene orthorhombic 7.40 4.93 2.53 monoclinic 8.09 2.53 4.79 triclinic 4.29 4.82 2.54 Polyvinyl orthorhombic 10.60 5.40 5.10 chloride polystyrene 21.90 21.90 6.63 (Brandrup and Immergut 1975) One can thus see that a unit monomeric cell of polystyrene occupies more space than polyethylene and PVC. Even though polyethylene occupies less space, one must remember that commercially used polyethylenes are usually less than 100% crystallinity. Thus, the introduction of amorphous regions are expected to greatly increase its permeability to the usual penetrants which normally come into contact with it. Temperature In cases where temperature is increased, volume dilation and thus an increase in free volume occurs. As a result, the 35 diffusion coefficient increases with temperature. The free- volume fraction may be represented as a linear addition of several variables (Stern et al., 1972; Fang et al., 1975): V. (RP. 01) = vfs(Ts.p5, 0) +0: (T-Ts) -B (p-ps) +10, (8) where vf is the fractional free volume of the pure, penetrant- free, amorphous polymer at some reference temperature T,, usually the glass-transition temperature, and pressure p,, usually 101.3 kPa. The penetrant volume fraction is (a and the coefficients a, [3, and y are positive constants for thermal expansion, compressibility and concentration, respectively. These constants characterize the role temperature, pressure and penetrant concentration play in affecting the free volume in the amorphous phase. They can be derived from the following equations: a “(3?)8 B=(-a—5 5 Y=(-5¢f1)5 (9) where 3 denotes some reference state. The effect of temperature on the rate of diffusion can also be explained by the Activated State Theory. This is based on the assumption that holes covering a spectrum of 36 different volumes and involving segments of several polymer molecules are continuously formed and destroyed due to thermal fluctuations. The rate of diffusion depends on the concentration of transient holes that are sufficiently large to accept diffusing molecules. Assuming a Boltzmann distribution, the concentration of a given set of holes decreases exponentially with the energy associated with its formation. Vergnaud (1986) reported that the activation energy of diffusion generally decreases with increasing temperature above the glass-transition temperature T9, of the polymer. The activation energy is thus related to the amount of energy required to form the transient voids necessary for the process of diffusion. The temperature dependence of the diffusion coefficient based. on the activated state theory can. be described by Arrhenius law and is expressed as: “E D=Do exp(?i?) (10) where Do is the pre-exponential factor and Ed is the energy of activation for diffusion. Temperature may also affect the thickness of the boundary layer at the membrane surface by altering the strength and number of hydrogen bonds in clusters of water or other polar 37 molecules if they are the penetrant under study. In this case, the concentration of this stagnant layer may become elevated through increased random movement of the solute molecules and a resultant decrease in its internal hydrogen bonding. Therefore individual penetrant molecules encounter less self attraction and find it easier to migrate towards the polymer. Clustering effect Clustering in itself can be considered as a factor that can affect the mobility of a polar penetrant through a polymeric matrix. Clustering could reduce the effective mobility of the penetrant since the size of the diffusing group increases. A similar effect has been reported for Kapton by Yang et al. (1985). Koros and Hellumus (1986) reported that if the diffusion process involves primarily jumps by a monomeric, unassociated penetrant, energy must be available both to make a gap to allow the penetrant to jump in or to cause the dissociation of one of the penetrants in a local region adjacent to the gap. As a result, higher activation1energies for'diffusion.and, thus, a lower diffusion coefficient at activities above which clustering occurs are expected. A cluster integral GM proposed by zimm and Lundberg (1956) can be used to discuss the sorption behavior of clustering systems in terms of the partial molar volume V,, the activity a,, and the volume fraction of penetrant 4),: 38 — = -(1 - 4),.) ( 6a, )m. —- 1 (11) When GM/VA is greater than -1, the solute is expected to cluster. GAA can be obtained from equilibrium sorption data as a function of the penetrant partial pressure. Thus, (dJAGM/VA) equals the excess number of 'A' molecules above the number that would exist in the neighborhood of an arbitrarily 'chosen type '8' molecule if the mixing were totally random, that is, no clustering (Koros and Hellumus, 1986). Polarity and charge In considering a situation where an aqueous liquid is in contact 'with. a polymer, it should be visualized as an immiscible organic layer in contact with an aqueous phase. Molecules from the bulk water diffuse and form a thin, highly structured layer at the membrane surface. Partitioning into and through the membrane can only occur if passage through this structured layer is successful. In situations where highly hydrophobic polymers such as polyolefin, vinyl plastics, nylons, epoxy resins, elastomers and other similar materials, are in contact with polar penetrants such as water, alcohols (methanol, ethanol, n-propanol, isopropanol, etc.), and acetic acid, clustering is expected to take place, form a Zless dense liquid layer at the boundary, and thus decrease the 39 equilibrium concentration of the liquid phase on the membrane surface (Lee et al., 1989; Hermann, 1972). Heindorf (1992) reported that for some molecules, variations in pH may influence water solubility, alter the diffusion constant from the bulk phase and the membrane solubility. The presence of dissolved substances such as humic materials may affect partitioning properties in a similar fashion. She also reported that interactions between the molecule and the membrane include dipole/dipole, dipole/induced dipole, and London dispersion interactions can also affect diffusion constants. Derjaguin and Churaev (1981) reported that polymers such as polyolefins have long hydrocarbon segments and therefore are expected to have a strong dispersive character. Thus, if a hydrophobic contaminant in an aqueous phase contacts such polymers, its uptake by the membrane should be expected to be easily facilitated. The dispersive character is greater for saturated hydrocarbons than for similar unsaturated ones, and greater for cyclic than for aliphatic analogs (Derjaguin and Churaev, ( 1981) . Adamson (1990) reported that a polar (nonpolar) adsorbent will preferentially adsorb the more polar (nonpolar) component of a nonpolar (polar) solution. Polarity is used here in the general sense of ability to engage in hydrogen bonding or dipole-dipole type interactions as opposed to nonspecific dispersion interactions. As an extension of this, the Traube's rule states that the adsorption of organic substances from aqueous solutions increases strongly and 40 regularly as we ascend the homologous series of a given compound. Adamson (1990) quoting Holmes and McKelvey (1928), showed that Traube's rule can be modified to show that a reversal of order should occur if a polar adsorbent and a nonpolar solvent were used. Wheeler and Levy (1959) reported that aromatic ring compounds tend to adsorb preferentially to aliphatic ones on carbon. As an explanation for this, they speculated that this may be due to the presence of the It electron interactions within the benzene ring structure, or alternatively, because of the higher polarizability of such rings. Bulky substituents reduce this preference, perhaps because of their preventing a close approach of the ring to the adsorbent surface. Concentration The interaction at the interface between two phases involving the transition of a molecule from one phase to another and driven by different concentration values, can be considered as the sorption of the molecule (Kovach, 1978). Sorption can be divided into two categories, adsorption and absorption. Adsorption is the concentration or accumulation of a penetrant at the surface of a membrane. Absorption on the other hand, is the process by which the molecules of the penetrant interpenetrate almost uniformly within the membrane, forming a solution-like relationship with that of the membrane. The fundamental difference between adsorption and 41 absorption is illustrated in Figure 5 for the reaction in which the molecules of a liquid move towards a solid surface (Weber, 1972). 01 U) 0 H O f0 '0 CD '.> g I 3 " c ; ---------- II “5 I :3 ' x le ,’ III S O I m «u ,I '44 l H H o ’l o a, 0' 1’ 4., "’ 4.1 c: _--- GTH o. '-"‘ :3 o". 'O“ o o x ,,,,, 5S o‘v 4.) Concentration of substance moves from liquid phase to solid phase Figure 5. Types of sorption curves for three penetrants by a given sorbent material (Weber, 1972). Takashi (1994) reported that curves I and III indicate curvilinear dependence of the amount concentrated at the solid surface on the amount remaining in the solution phase for favorable and unfavorable separation patterns, respectively. On the other hand, curve II shows linear adsorption and an absorption which occurs in direct proportion to concentration. He also noted that the adsorption is described by a Langmuir 42 type of isotherm, while the absorption is described by Henry's Law. In certain cases the Flory-Huggins equation instead of Henry's Law can be used to describe an absorption phenomenon (Hernandez, 1989). The adsorption of a penetrant by a membrane, for example, can be described as either physical or chemical in nature. In physical adsorption, the molecules are physically attached to the membrane surface and may be multilayered, where each molecular layer forms on top of the previous layer with the number of layers being proportional to the penetrant's concentration. Thus, the higher the concentration of the penetrant, the greater the number of molecular layer formed. The point should be made that this multilayer diminishes with increasing distance from the membrane due to the inverse relationship that exists between the attractive force of the membrane for the penetrant molecules and/or vice versa. Increases in turbulence increase the number of penetrant molecules that arrive at the boundary layer and thus increase the probability of adsorption. Heindorf ( 1992) reported that if the concentration of an aqueous boundary layer is too great, concentration polarization will limit the efficiency of the transfer. Chemical adsorption occurs when a chemical compound is produced between the adsorbed organic molecule and the adsorbent. This is a one'molecular layer system and is quasi- irreversible. This process requires energy in its formation and so energy would be required for it reversal or desorption. 43 Diffusion is a macroscopic manifestation of a random walk or Brownian motion of individual additive molecules within the polymer lattice (Lim and Hollifield, 1995). Generally, a reference velocity such as mass, molar, or volume average bulk velocity is selected, and movement of the component of interest (gases, vapors, liquids, and solutes) relative to this reference velocity is defined to be true diffusion (Koros and Hellumus, 1986). This mode of molecular transport has been shown in most cases to obey Fick's laws of diffusion (Crank, 1975) which states: Fick’s first law: F = -D 93’ (12) p ax . OC 62 Fick’s second law: ——‘3 = D C" (13) at p 6x2 where F is the rate of transfer of a penetrant per unit area of polymer expressed as mass, molar or bulk volume as described above. Dp is the diffusion coefficient in (length)2/time, t is elapsed time and CI) is the concentration of the penetrant in the polymer expressed in the same units as the diffusant per unit of volume or mass of polymer. One can thus see that Fick's first law is concentration dependent while the second law is independent of concentration. If the diffusion coefficient is time-dependent, the diffusion process 44 is said to be non-Fickian. Chainey (1989) reported that in the steady state, the diffusion flow is constant, and where the diffusion coefficient is independent of concentration, Fick's first law may be integrated to give: L7 = (14) where C1 and Co are the concentrations of penetrant at the high and low concentration faces of the film, respectively. J; is the steady-state flow and L is the film thickness. Chainey (1989) also reported that it is the pressures or partial pressures, p, of gas or vapor above the faces of the film, rather than the surface concentrations, C, which are usually known. These quantities are related by Henry's law, which states that: 0 ll Sp (15) where S is the solubility coefficient. This relationship is obeyed only at low penetrant concentrations. It is often found that where the diffusion is concentration dependent, the solubility coefficient also exhibits a concentration or pressure dependence, although the magnitude of this effect is usually less. It can be seen that a combination of equations 45 14 and 15 can give the permeation equation: DS - J8 = (p1 p°) (16) L where DS is the permeability coefficient P. Thus, J3, the steady-state flow can be rewritten as: J = - (17) Since J3 is the amount of penetrant Q, which has passed through the film in time t, the equation can be written as: :2 J' t (18) Surface area (dimensional changes) As mentioned above in Fick's laws, the flux or rate of transfer of the penetrant is dependent on the exposed area of the membrane. Indeed, the diffusion coefficient itself is expressed in units of area (cm?) per unit time (sec). Gases 46 and vapors can permeate through materials by macroscopic or microscopic pores and pinholes or they may diffuse by a molecular mechanism, known as activated diffusion. In activated diffusion, the gas is considered to dissolve in the membrane at one surface, to diffuse through the packaging material by ‘virtue of a (concentration. gradient, and to reevaporate at the other surface of the packaging material. This phenomenon is referred to as permeability. The permeability constant (P) of a membrane to a given penetrant can be expressed as: JxL P=——-——5 1 AxAP (9) where A is the: exposed surface area and AP is the partial pressure gradient across the membrane. Here again we see the dependence of transport of a penetrant through a membrane on the thickness and surface area of the membrane. One can also see that the dependence of concentration on diffusion, solubility and now permeability coefficients has been dealt with under equations 15 and 17, and the fact that P = 08. An alternative diffusion eguation In cases where the membrane sample is immersed in the penetrant, such that both faces are exposed simultaneously to 47 the same concentration, the penetrant enters the film and diffused towards its center. Chainey (1989) reported that in such cases, the sorption. method for the 'measurement of transport properties should be used. He also stated that the bulk and surface concentrations are usually rapidly attained when compared to the rate of diffusion, and that such conditions are usually satisfied by polymers at temperatures well above their glass-transition ‘temperature. In .such situations the diffusion coefficient is usually independent of concentration and the extent of sorption at time t into the membrane is given by: MC: Dt1/21 " _n' M E (I; {E + 2 (g ( l) lerfc [ma-m1} (20) where Mt and M. are the masses sorbed at time t and at equil ibrium , respectively , and ierfc is the inverse complementary error function. For short times, this equation simplifies to: “1%)” (21) file 48 A plot of M,/M., vs 1/t will give D, the diffusion coefficient, as the initial linear slope. In a case of a concentration dependent diffusion coefficient, the initial slope of Mt vs (It/L gives an average diffusion coefficient 0, over the concentration range of the experiment. Crank (1975) showed that a good approximation of the average diffusion coefficient can be obtained by: Co _ l Dy — “Z" [D(C)dC (22) O In addition to exposing a liquid penetrant to a polymeric membrane in sorption experiments and determining its sorption by spectroscopy and chromatographic techniques, the sorption of a gas or vapor penetrant.can be analyzed by the gravimetric technique. Takashi (1994) reported that this can be done by recording continually the gain or loss of weight of the membrane as a function of time using an electrobalance. Hernandez (1986) reported that the diffusion equation appropriate for the sorption of penetrants by a polymer sample in sheet or film form can be determined from Crank (1975) where: 49 M: _ _ 8 -Dxn2xt 1 -9Dx1t2xt Hernandez et a1. (1986) also reported that the sorption diffusion coefficient (0,) can be calculated from equation 23 by setting Mt/M. equal to 0.5 and solving to give Ds 2 = 0.049 L (24) t0.5 where t“, is the "half-sorption- time" or the time required to attain the value Mth. = 0.5. Crank (1975) also described a similar equation for the sorption of a similar penetrant by a polymer in spherical or powdered form. In this case, L the thickness of the film is replaced by d the diameter of the penetrant particles M.- ___ _ 6 ~4Dxn2xt 1 -160xn2xc T1: 1 ?2-[exp(———d2 )+ zexp( L2 )] (25) In this case, the sorption diffusion coefficient (0,) can be obtained as follows: 50 2 D}3 = 7.45x10'3 d (26) t0.5 the transport of limonene, ethyl acetate and toluene in polyethylene. Choy et al. (1984) and Vergnaud (1986) also used similar techniques when studying the sorption of toluene in {polypropylene and benzyl alcohol by plasticized PVC, respectively. PARTITIONING STUDIES Bioconcentration factor Bioconcentration factor (BFC) is defined as the ratio between the concentration of anlanalyte in an organism and the concentration of the same analyte in the environment. This can be expressed mathematically as: ng compd./g wet: wt. organism (2.7) BFC = 1] g/ ml wa tax This term has been widely used by environmental biological researchers to correlate the chemical and physical properties of xenobiotics to the tendency of biomonitors to accumulate a particular chemical contaminant. As is evident from the equation, BFC can be influenced by the aqueous solubility of the contaminant, the design of the experiment and the 51 biochemical ramifications of the organism involved. Octanollwate; partition coefficient The concept of octanol/water partition coefficient (Km) was first proposed by Fujita et al. (1964). The octanol/water partition coefficient (Km) is.defined.as the ratio between the concentration of a chemical in octanol and the concentration of the same chemical in an aqueous phase. Expressed mathematically, this is describes as: _ cone. 6 the octanol phase (28) °“' cone. 6 the aqueous phase where 8 represents in. This expression was first used to describe the distribution of drugs in a variety of biological systems (Heindorf, 1992). Octanol was used as a standard substance to describe lipophilicity or the potential of the chemical to favor the non polar environment when in an immiscible solution of two polar/non-polar solutions. Egpilibpiup partition coefficient (Kg Instead of using the terms K0,, and BCF, researchers working with polymers express the ratio between the 52 equilibrium concentration of a chemical in the polymeric phase and the equilibrium concentration of the same chemical in an aqueous phase as the equilibrium partition coefficient (K9). This is expressed as: = [Cpl ea K9 [Caqleq (29) where [Cp],Bq is the equilibrium concentration of the compound in the plastic film, and [CM]eq is the equilibrium concentration of the compound in the aqueous solution (Kwapong and Hotchkiss, 1987). The resulting Kg'values can thus be an important parameter in determining the extent of sorption, since it is an indicator of the affinity of the compound for the plastic film. Q§A22§E_§ EXPERIMENTAL EXPERIMENTAL DESIGN The molecular structure, size, uptake capacity, physical, chemical, and kinetic characteristics of commercially available polyethylene, polyvinyl chloride and polystyrene were studied. These characteristics were studied in relation to the ability of these materials to remove or reduce the concentration of PCBs in an aqueous and peanut oil systems. A total of eight (8) PCB congeners, ranging from tri- chlorinated to deca-chlorinated were selected based on their differential.molecular sizeiand structural.arrangement~ .After consultation (J. Gill, Department of Animal Science, Michigan State University, 1993, personal communications) it was determined that three replicates would be sufficient to distinguish.differences in total concentrations1of PCBs in the materials and the liquid phases. Firstly, the relationships between the three materials and each PCB congener in an aqueous system were studied at the same environmental conditions. For the analysis involving the peanut oil, only the polyethylene material was used. In this case the ability of the material to remove the PCB from the oil was compared to its ability to remove the congeners from the aqueous system. 53 54 The study for the polystyrene was repeated but with a material thickness five (5) times that of the original material. This was repeated a third time, but with the original material at an elevated temperature. The results from this part of the study were also compared in order to determine the effect of temperature and material thickness on the potential of the material to remove the different PCB congeners from the aqueous system. The mathematical basis for uptake of the congeners from polymeric phases, was the calculation of sorption diffusion coefficients and partition coefficients for each congener by all materials at all treatment conditions. Rate loss calculations were performed to mathematically compare the loss of the PCBs from the liquid phases. To evalute the results obtained in this study, and to measure its success against the initial objectives, five hypotheses were postulated. Testing of these hypotheses by analysis of variance using SPSS statistical software (1988), assisted in arriving at logical conclusions bases on the calculated results. These hypotheses are: 1. there are no significant differences between the three polymeric materials in their uptake potential for the PCB congeners; 2. there are no significant differences in the removal of the PC85 from water as compared to oil by polyethylene; 3. there is no significant difference in the effect of thickness on the potential of polystyrene to remove PCBs 55 from water; 4. there is no significant difference in the effect of temperature on the potential of polystyrene to remove PCBs from water; 5. there are no significant affinity differences between the congeners for the polymeric materials. MATERIALS Glassware All. glassware (Erlenmeyer’ flasks, beakers, ‘Turbo-Vap tubes, chromatographic columns, Soxhlet condensers, separatory funnels, boiling flasks, 1 mL and 5 mL volumetric flasks) were washed with detergent, rinsed with tap water, distilled water, acetone, and finally with hexane. The glassware was then oven dried. Polymers Polyvinyl Chloride (PVC) - transparent film with a thickness of 1.778 x 10'3 cm (1 ply PVC thickness) and supplied by Reynolds Metals (Grottoes, VA). Polystyrene - transparent film with a thickness of 3.048 x 10‘3 cm (1 ply) and supplied by Dow Chemical (Midland, MI). (Reference to 1 ply polystyrene is equivalent to 3.0481x10'3 cm thick. Reference to 5 ply equals 15.24 x10’3 cm thick.) -Polyethylene - resin supplied-as pellets by Dow Chemical (Midland, MI). The polyethylene pellets were extruded using a single screw 4 zone Killion model KLB-lOO extruder, with 56 a 76.2 cm diameter barrel attached. It was interfaced with a Killion chill roll attached to a model 8412-A sterico steritronic temperature control console. The temperature zones 1 - 4 were set to operate at 110, 121, 132, and 149W3, respectively, while the die temperature was set at 143%:. The screw speed was set to operate at 15 rev/min., the 20.32 cm wide and 15.24 cm diameter chill roll at the dial speed of 100 (xx rpm), with a cooling temperature of 120%:. The percent crystallinity of the polyethylene, after extrusion, was calculated to be 53.56%. It had a thickness of 2.286 x 10'3 cm (1 ply polyethylene thickness). Solvents All solvents were analytical HPLC grade quality. lHexane - 85.04%, EM Science (Gibbstown, NJ). IAcetone - 99.8%, Mallinckrodt Chemicals (Paris, KT). lIsooctane - 99.91%, Mallinckrodt Chemicals (Paris, KT). lPetroleum ether - 99.99% EM Science (Gibbstown, NJ). lToluene - 99.99%, Mallinckrodt Chemicals (Paris, T). lDiethyl ether - 99.99%, Mallinckrodt Chemicals (Paris, KT). lAcetonitrile - 99.8%, Mallinckrodt Chemicals (Paris, KT). cpepicals All chemicals were analytical grade quality. lSodium sulfate Nafiflh - (activated by storing overnight at 130°C), Mallinckrodt Chemicals (Paris, KT). 57 lFlorisil, 60-80 mesh - (activated by storing overnight at 130°C), Mallinckrodt Chemicals (Paris, KT). ISilica gel 60, 70-230 mesh - (activated by storing 16 hrs overnight at 130°C), Mallinckrodt Chemicals (Paris, KT). lPolychlorinated biphenyl congeners - Accustandard Inc. (New Haven, CT) are shown in Table 8 below: lSodium chloride NaCl - Mallinckrodt Chemicals (Paris, KT). Table 8. Systematic names of the PCB congeners used in this study. IUPAC No. Chlorobiphenyl systematic names Internal Standard #30 2,4,6-Trichlorobiphenyl Isomer # 31 47 103 128 171 200 206 209 Biphenyl Compound 2,4',5-Trichlorobiphenyl 2,2',4,4'-Tetrachlorobiphenyl 2,2',4,5',6-Pentachlorobiphenyl 2,2',3,3',4,4'-Hexachlorobiphenyl 2,2',3,3',4,4',6-Heptachlorobiphenyl 2,2',3,3',4,5',6,6'-Octachlorobiphenyl 2,3,4,5,6,2',3',4',5'-Nonachlorobiphenyl 2,3,4,5,6,2',3',4',5',6'Decachlorobiphenyl (Ballschmiter and Zell, 1980) These are based on the occurrence of the congeners in Aroclor 1254 (Ballschmiter and Zell, 1980). EXPOSURE OF FILM TO PCB CONGENERS Deionized distilled water was spiked with a mixed PCB hi mi oi ev in: re: fla pol Ply long weig 9. b' All 58 standard solution to produce a 10 ppb solution of each congener. This was done by initially dissolving equal concentrations of the 8 PCB congeners using acetone as a solvent. After adding the desired quantity to give the 10 ppb solution, the system was heated at 40°C for 1 hr with frequent shaking, to evaporate any residual acetone. After cooling 25W3, the solution was ready for contact with the polymeric materials. A 0.1 ppm stock solution of spiked peanut oil with PCBs was also prepared. This was made by dissolving 1 mL of a 50 ppm mixed PCB in hexane solution in approximately 5 mL of peanut oil in a glass tube. The hexane in this solution was then evaporated for 30 min under nitrogen using a N-Vap Model III instrument (Organomation Assoc., Shrewsbury, MA). The resulting solution was then made up to 500 mL in a volumetric flask. This solution was then ready for contact with the polyethylene material. Approximately 0.160 g of each polymeric film, except the 5 ply thick polystyrene, was prepared by cutting into longitudinal strips approximately 4 x 1.5 cm. The average weight of the 5 ply polystyrene strips was approximately 0.70 9, but the dimension was similar to that of the other samples. All sample numbers, actual weights and duration of exposure are recorded in Tables 9 - 13. 59 Table 9. Weights of PVC samples a storage times in water at 25°C. Exposure time (hrs) Replicate f PVC Sample wt. (g) 0.083 1 0.1565 0.083 2 0.1790 0.083 3 0.1778 0.167 1 0.1534 0.167 2 0.1478 0.167 3 0.1510 0.25 1 0.1540 0.25 2 0.1425 0.25 3 0.1584 0.333 1 0.1648 0.333 2 0.1516 0.333 3 0.1556 0.5 1 0.1665 0.5 2 0.1622 0.5 3 0.1672 0.75 1 0.1710 0.75 2 0.1611 0.75 3 0.1560 1.0 1 0.1597 1.0 2 0.1512 1.0 3 0.1541 1.5 1 0.1425 1.5 2 0.1431 1.5 3 0.1762 2.0 1 0.1453 2.0 2 0.1539 2.0 3 0.1762 4.0 1 0.1476 4.0 2 0.1498 4.0 3 0.1633 8.0 1 0.1581 8.0 2 0.1476 8.0 3 0.1488 12.0 1 0.1582 12.0 2 0.1754 12.0 3 0.1631 60 Table 10. Weights of Polystyrene samples & storage times in water at 25°C. Exposure time (hrs) Replicate # PS sample wt. (g) 0.083 1 0.1514 0.083 2 0.1754 0.083 3 0.1576 0.167 1 0.1821 0.167 2 0.1723 0.167 3 0.1567 0.25 1 0.1449 0.25 2 0.1406 0.25 3 0.1641 0.333 1 0.1703 0.333 2 0.1526 0.333 3 0.1601 0.5 1 0.1600 0.5 2 0.1462 0.5 3 0.1476 0.75 1 0.1391 0.75 2 0.1528 0.75 3 0.1698 1.0 1 0.1575 1.0 2 0.1685 1.0 3 0.1446 1.5 1 0.1507 1.5 2 0.1380 1.5 3 0.1363 2.0 1 0.1367 2.0 2 0.1434 2.0 3 0.1710 4.0 1 0.1526 4.0 2 0.1446 4.0 3 0.1427 8.0 1 0.1491 8.0 2 0.1344 8.0 3 0.1448 12.0 1 0.1375 12.0 2 0.1461 12.0 3 0.1422 water at 25°C. 61 Table 11. Weights of Polyethylene samples a storage times in Exposure time (hrs) Replicate #5 PE sample wt. (9) 0.083 0.083 0.083 0.167 0.167 0.167 0.25 0.25 0.25 0.333 0.333 0.333 P‘H)‘ minu1 NNN o 0 COO abbh GOO 000 000000 12.0 12.0 12.0 1 2 3 MNH UNI-J UNH 0.1634 0.1427 0.1643 0.1641 0.1678 0.1898 0.1565 0.1583 0.1466 0.1529 0.1419 0.1589 0.1878 0.1953 0.1800 0.1767 0.1612 0.1663 0.1828 0.1955 0.1853 0.1848 0.1921 0.1802 0.1835 0.1796 0.1855 0.1575 0.1704 0.1525 0.1456 0.1449 0.1465 0.1608 0.1634 0.1629 water at 35°C. 62 Table 12. Weights of Polystyrene samples & storage times in Exposure time (hrs) Replicate # PS sample wt. (9) 0.083 0.083 0.083 0.167 0.167 0.167 0.25 0.25 0.25 0.333 0.333 0.333 NNN 0.. COO abhnb COO oooooo 0. 00° 12.0 12.0 12.0 UNH MN)“ UNH uNl—J 0.1419 0.1499 0.1463 0.1476 0.1534 0.1630 0.1546 0.1527 0.1534 0.1727 0.1546 0.1759 0.1627 0.1710 0.1631 0.1598 0.1776 0.1434 0.1680 0.1627 0.1595 0.1471 0.1603 0.1560 0.1470 0.1726 0.1482 0.1741 0.1459 0.1523 0.1628 0.1772 0.1691 in water at 25°C. 63 Table 13. Weights.of'5 ply polystyrene samples & storage times Exposure time (hrs) Replicate} PS sample wt. (9) 0.083 1 0.6907 0.083 2 0.7161 0.083 3 0.6336 0.167 1 0.7215 0.167 2 0.7791 0.167 3 0.7588 0.25 1 0.6898 0.25 2 0.7798 0.25 3 0.7064 0.333 1 0.7614 0.333 2 0.7825 0.333 3 0.7784 0.5 1 0.6882 0.5 2 0.7108 0.5 3 0.6839 0.75 1 0.7374 0.75 2 0.7009 0.75 3 0.7186 1.0 1 0.6773 1.0 2 0.6815 1.0 3 0.7142 2.0 1 0.6437 2.0 2 0.7496 2.0 3 0.6576 4.0 1 0.6936 4.0 2 0.6320 4.0 3 0.6203 8.0 1 0.6803 8.0 2 0.7176 8.0 3 0.7694 12.0 1 0.7282 12.0 2 0.6916 12.0 3 0.6779 64 Each sample was then inserted in a 76 mL serum vial supplied by Wheaton Inc., Millville, NJ. Each vial was then brim- filled with the spiked aqueous PCB congener solution; sealed, and stored at 25°C in the absence of light. At given time intervals for a total of 12 hrs for polystyrene and PVC and a total of 72 hrs for polyethylene, samples were removed for PCB congener quantification. The analysis for polystyrene was repeated twice, once at 35%» then again with a film 5 times thicker (5 ply) than the polystyrene used in the initial analysis. The thickness of the 5 ply material was determined to be 16.51 x10"3 cm. This thicker material was manufactured from the same stock as the original one. The polyethylene exposed to the peanut oil was prepared similarly to that of the polyethylene exposed to the water and also stored at 25%: for maximum time of 72 hrs. PCB CONGENER ANALYSIS The PCB analytical method, including the clean-up procedures used in this study, was adapted from electron capture gas chromatographic analyses for PCB as outlined by Price et al. (1986). All samples in this study were extracted in triplicates and subsequently quantified for PCBs. PCB EXTRACTION PROCEDURES Extraction of PCB from polymers After exposure of the samples for the specified time, the 65 polymers were immediately removed from the spiked water. The PCBs were extracted from the materials by Soxhlet extraction using 120 mL hexane as the solvent for 18 hrs. A 1 mL aliquot of a 5 ppm PCB congener 2,4,6 tri-chlorobiphenyl (IUPAC #30) solution was used as an internal standard for determination of concentrations and recoveries of PCBs. This solution was added to each sample prior to extraction. This PCB congener was selected as an internal standard because it does not occur in commercial Aroclor® mixtures nor has it been detected in environmental samples (Williams and Giesy, 1992) . After the extractions were performed, each extract was collected into a 200 mL volumetric flask. These were subsequently reduced in a Turbo-Vap evaporator (Zymark) to approximately 0.5 mL volume at ambient temperature. Care was taken to avoid reducing the extracts to dryness. Emtgactiop of PCB fpom oil solution After exposure of the oil to the polyethylene, the PCBs were extracted from the oil by first adding 25 mL acetonitrile (saturated with hexane) to it in a 250 mL separatory funnel. This was shaken for approximately 2 min, then the lower acetonitrile layer drained into a 500 mL separatory funnel. The remaining oil in the 250 mL separatory funnel was extracted two times more, each time with a fresh aliquot of 25 mL acetonitrile. In each case the acetonitrile layer was drained and combined in the 500 mL flask. To this combined solution was then added 100 mL hexane and 100 mL of a 10% 66 aqueous sodium chloride solution. After shaking for 2 min, three layers were observed. The bottom layer was discarded then the remaining solution was shaken again with fresh 100 mL aliquots of hexane and. the 10% aqueous sodium. chloride solution. If three layers appeared after this second extraction, a third 100 mL aliquot of each of the hexane and sodium chloride was again added then extracted. However, if only two layers were observed, the lower layer was discarded and the upper layer collected into a 200 mL volumetric flask. In cases where three extractions were performed, this upper layer was also collected while the lower was discarded, after it became clear that only two layers were present within the flask. To the solution collected in the 200 mL volumetric flasks 8 g of anhydrous sodium sulfate was added. After allowing half hour for the drying action of the Nafixh, the extracts were reduced in the Turbo-Vap evaporator to approximately 1.0 mL volume at ambient temperature. Care was taken to avoid reducing the extracts to dryness. Each sample was then reduced to approximately 0.1 mL by nitrogen evaporation using the N-Vap Model III instrument then made up with isooctane to exactly 0.5 mL in a volumetric flask. Expragtlgp of PCB fmom agueous solutlon The water solutions were also analyzed for their PCB concentrations. After the polymers were removed from the vials, the spiked water which remained was transferred to 250 mL separatory funnels. To each sample, a 30 mL aliquot of 67 hexane was added and vigorously shaken for approximately 30 sec. After standing for several minutes, the hexane layer rose to the top. This was drawn off and transferred to a 100 mL flask fitted.with a stopperx ‘This extraction procedure was repeated 2 more times on the same water sample, using 30 mL of hexane on each occasion. The extracts were combined, then a 10 9 quantity of granular anhydrous NaZSO, (activated by storing overnight at 130°C) , was added to each flask to remove any traces of water in the extract. .After allowing 30 min for drying with Nafiflx, the extracts were reduced in the Turbo-Vap evaporator to approximately 1.0 mL volume at ambient temperature. Care was taken to avoid reducing the extracts to dryness. Each sample was then reduced to approximately 0.1 mL by nitrogen evaporation using the N-Vap Model III instrument then made up with isooctane to exactly 0.5 mL in a volumetric flask. PCB'S CLEANUP PROCEDURES Gel Permeation Chromatggraphy (GPC) Cleanup by GPC resulted in removal of plasticizers, antislip agents, antistatic agents, flame retardants, monomers, dimers, dissolved polymers, antioxidants, light absorbers, and any other compound removed by the hexane during the Soxhlet extraction, without alteration of residual PCBs. The ' automated GPC system (Waters 712 WISP, Milford, MA) provided for unattended operation with a timed control system to remove the non PCB compounds. Each sample was diluted to 4 mL with [E at ar ir di hid US Sh is 68 HPLC grade hexane and transfered to a 6 mL GPC vial. After the contaminant as well as PCB fractions were collected, an automated washing device would flush the system prior to the next injected sample. This system was used to cleanup the concentrated extracts of polyvinyl chloride (PVC) and polyethylene from the Turbo-Vap evaporator. The following procedure of cleanup by GPC was: a 2 mL aliquot of each extract was injected into the GPC column (Waters 19 x 300 mm Ultrastyragel 500 A) at a solvent flow rate of 5 mL/min. The solvent used was hexane. This was attached to a Waters/590 Programmable HPLC pump(Milford, MA) and a Waters faction collector using a UV detector set at 254 nm. This was interfaced with a Hewlett Packard 3390A integrator (Avondale, PA). The contaminant portion ‘was discarded, while the PCB portion collected, then reduced in the Turbo-Vap evaporator to approximately 1.0 mL. Each sample was reduced to approximately 0.1 mL by nitrogen evaporation using a N-Vap Model III instrument(Organomation Assoc., Shrewsbury, MA), after which the sample was made up with isooctane to exactly 0.5 mL in a volumetric flask. o ' ' i ' e The polystyrene and oil extracts were cleaned up using florisil followed by silica gel low pressure chromatography. The concentrated extracts from the Turbo-Vap were initially eluted through the florisil column with a 6% (by volume) of diethyl ether in petroleum ether solvent. The column (1 cm 69 i.d. x 51 cm), was prepared by placing 1 cm of granular anhydrous NaZSO, on a glass wool plug, followed by 5 g of 60-80 mesh florisil (activated by storing overnight at 130°C) , then another 1 cm of granular anhydrous Na,SO,. The prepared column was then washed by allowing the elution of a 20 mL portion of hexane. When the hexane reached the top of the upper layer of NaZSO“ the Turbo-Vap concentrate was transferred to the column and allowed to drain onto the florisil bed. A 40 mL aliquot of elution solvent (6% diethyl ether in petroleum ether) was then added to the column. Care was taken not to allow the florisil bed to go to dryness. The collected eluent was then reduced to approximately 1 mL volume in the Turvo-Vap evaporator. The silica gel column was prepared similar to that described for the florisil columns, but with silica gel 60 (70-230 mesh) activated at 130°C for 16 hours. A 50 mL quantity of 0.5% toluene in hexane was used as elution solvent. The concentrated eluent from the f lorisil column was eluted through this silica gel column, collected, then reduced to approximately 1.0 mL volume in the Turbo-Vap and to approximately 0.1 mL by nitrogen evaporation using the N-Vap Model III instrument. Each polystyrene sample was then made up with isooctane to exactly 0.5 mL in a volumetric flask. The oil samples were made up to 1 mL in a volumetric flask. PCB QUANTIFICATION The concentration of individual PCB congeners in each 70 extract was determined by gas chromatography (GC) with a 63Ni electron capture detector (Hewlett Packard 5890 series II, Avondale, PA), and equipped with 08-5 capillary column (60 m x 0.25 mm i.d.). PCBs would generally elute in order of chlorination: C12H9Cl first, C,,Cl10 last. The separated peaks from the GC column were detected at different times (retention times). The retention time of each PCB congener was confirmed by comparing the individual peaks with the retention time of the 12 PCB congeners mixed standard. These 12 congeners are found in Arochlor 1254°1and run under the same conditions used for Arochlor 1254° quantitation. The standard PCB congeners used were all the same concentration and are listed in Table 8. All quantification was based on peak areas relative to the area of individual congener standards and congener #30. A standard. curve was constructed for each PCB congener from different concentrations of the standard mixture. The GC conditions for peak separation and quantification were as follows: Detector temperature: 300%: Inject temperature: 160%: Inject volume: 3 uL Carrier gas: Helium Flow rate: 1 mL/min Detector makeup gas: Nitrogen The oven was programmed at an initial temperature of 160%:. The total run time was 60 min with 2 temperature/time levels. 71 Level 1 was set at 8°C/min. to a maximum of 200°C held for 5 min. Level 2 was set at 2°C/min. to a maximum of 280°C held for 10 min. The data were recorded on a HP Level Four integrator. The PVC and polyethylene samples were quantified by comparison of peak area using the following equations: (RA of sample soln internal std.) x 100 (30) R = I 900),er % (RA of Sppm std. No. 30 PCB ) GPC dil. factor x LS of congener x RA of congener Recovery % Cone. (pm) inj. vol. = (31) Wt. P ' 8 (:8 (mg) 112 sample 0.003” 72 GPC vol.: Gas Permeation Chromatography volume = 2 mL GLC vol.: Gas Chromatography injected volume = 3 uL LS: Line Slope RA: Retention Area Sample wt.: Sample weight (g) The polystyrene and the water samples were quantified by comparison of peak area using the following equations: LS of congener x RA of congener (33) c . "- 1" one (pp!!!) 1n] V0 Recovery % Conc. inj. vol. x0.5mLx0.076L Wt. p B ' 1 = C (mg) m “mp ‘9 0.003mL DATA ANALYSIS Mass balance determination A mass balance of the weights of PCBs in the liquid and the polymeric phases were performed. The results of this 73 determination can be used to validate the effectiveness of the method used to quantitate the PCB congeners in the aqueous systems before and after the contact with the polymers and for PCBs in the three polymeric materials. These ratios were obtained by taking the sum of the PCB concentrations in the water and in the plastic and dividing by the initial PCB concentration in the water; This is expressed in the following equation: (Wt. PCB in H20 at timet) +(Wt. PCB in plastic at timec) Wt. PCB in H20 at timeo Mass bal. ratio = (35) In an ideal system, the mass balance ratio should be equal to one (1). In such cases, the concentration of the penetrant removed by the polymer would be equal to the concentration lost from the surrounding medium. However, due to experimental errors one might get small fluctuations above and below this ideal. or o usion coe 'cien de erm’nation The determination of the sorption diffusion coefficients for each material was obtained using Crank (1975) model for the sorption of penetrants by a polymer sample in sheet or 74 film form where: The sorption diffusion coefficient (D,) was calculated from equation 4-2 by setting Mt/M. equal to 0.5 and solving to give D, where: (37) where to.5 is the "half-sorption-time" or the time required to attain the value Mt/M. = 0.5. The values obtained from these calculations provided a mathematical estimate of the equilibrium sorption coefficients for each congener by the polymeric materials. Equilibrium partition coefficient (1(9) The equilibrium partition coefficient (Ke) expresses the ratio between the equilibrium concentration of a given congener in the polymeric phase and the equilibrium concentration of the same congener in an aqueous phase of a un: wh in cc va si fc t1 9! Ci Sc 75 unit sample system. This is expressed as: ___ [c919, K0 [Caqleq (38) where [CF],Eq is the equilibrium concentration of the compound in the plastic film, and [Cgfleq is the equilibrium concentration of the compound in the aqueous solution. The Ke values obtained were used.to determine the extent of sorption, since they are indicators of the affinity of the PCB congeners for the plastic films. Statisticgi calculations The statistical significance of the relationships between the materials, liquid phases, the PCB congeners, and environmental conditions used in this.study, were tested using a one-way analysis of variance (ANOVA) model. This was calculated using SPSS Release 4.1 for IBM (1988) statistical software. The criteria for statistical significance was chosen as 0.05. Where significant differences were found, a contrast analysis was performed. Tukey range test was used to determine the pairs of means which were statistically significant at 0.05 level. CHAPTER 4 B§§ELE§ VALIDITY OF THE PCB ANALYTICAL METHODOLOGY Tables 14, 15 and 16 show the mass balance ratios for PCB congeners concentrations in the polyethylene, polystyrene and PVC, respectively, and the water to which they were exposed. These results can be used to validate the effectiveness of the method used to quantitate the PCB congeners in the aqueous systems before and after the contact with the polymers and for PCBs in the three polymeric materials. In an ideal system, the mass balance ratio should be equal to one (1). In such cases, the concentration of the penetrant removed by the polymer would be equal. to the concentration lost from the surrounding medium. However, due to experimental errors, small fluctuations above and below this ideal might occur. One factor contribution to errors may be the fact that the internal standard used to determine the percent recovery is a tri-chlorinated PCB congener. Since the electron capture detector in the GC instrument is sensitive to chlorine atoms, its sensitivity to the higher chlorinated congeners would be greater than to the lower analogues. .Another factor likely to contribute to these errors may be the low concentration of the PCB analyte. These factors may account for the low ratios 76 77 obtained for the penta-chlorinated congeners shown in Tables 15 and 16, and the high ratios obtained for the tri- chlorinated analogues. Table 14. Mass balance ratios for PCB congeners in polyethylene/water stored at 25%:. Time Tri Tetra Penta Hexa Hepta Octa Nona Deca (hrs) 0.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 ,0.08 1.46 1.20 1.00 1.21 1.20 0.94 0.99 1.12 0.17 1.34 1.23 0.98 1.23 1.21 0.97 0.86 0.95 0.25 .1.13 1.19 0.95 1.08 1.21 0.94 0.80 0.88 0.50 1.15 1.04 0.88 1.09 1.08 0.94 0.77 0.75 0.75 1.21 1.04 0.85 0.96 1.12 0.98 0.84 0.89 1.00 1.21 1.06 0.91 0.99 1.12 1.10 0.88 1.01 1.50 1.25 1.21 1.09 1.14 1.24 1.23 0.96 1.12 2.00 1.26 1.21 1.05 1.06 1.40 1.16 0.92 1.08 4.00 1.23 0.98 0.95 0.97 1.09 1.12 1.16 1.09 8.00 1.36 0.95 0.88 0.88 1.10 1.07 1.06 1.13 12.00 0.98 0.97 0.95 0.91 1.05 1.05 1.06 1.32 24.00 1483 1.22 1.08 1.17 1.17 1.16 0.84 1.20 48.00 1474 1.53 1.51 1.30 1.39 1.40 1.00 1.39 72.00 1484 1.45 1.13 1.19 1.66 1.23 0.90 1.26 78 Table 15. Mass balance ratios for PCB congeners in polystyrene/water stored at 25°C. Time Tri Tetra Penta Hexa Hepta Octa Nona Deca (hrs) 0.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 0.08 1.49 0.99 0.97 1.50 1.58 1.46 1.37 1.30 0.17 1.31 1.01 0.99 1.49 1.56 1.46 1.33 1.26 0.25 1.51 0.94 0.91 1.53 1.55 1.49 1.35 1.23 0.33 1.18 0.85 0.77 1.54 1.59 1.50 1.46 1.24 0.50 1.29 0.86 0.80 1.36 1.40 1.35 1.22 1.18 0.75 1.24 0.80 0.68 1.41 1.47 1.36 1.26 1.19 1.00 1.09 0.75 0.67 1.49 1.56 1.47 1.30 1.28 1.50 1.11 0.68 0.57 1.23 1.38 1.29 1.19 1.16 2.00 0.95 0.67 0.51 1.64 1.69 1.62 1.47 1.43 4.00 1.01 0.56 0.43 1.37 1.44 1.35 1.28 1.26 8.00 1.04 0.54 0.39 1.43 1.48 1.45 1.52 1.35 12.00 1.04 0.55 0.40 1.53 1.56 1.46 1.58 1.42 79 Table 16. Mass balance ratios for PCB congeners in PVC/water stored at 25°C. Time 'Tri Tetra Penta Hexa Hepta Octa (hrs) 0.00 1.00 1.00 1.00 1.00 1.00 1.00 0.08 1.19 1.28 0.87 1.16 1.20 1.01 0.17 1.24 1.27 0.84 1.14 1.37 1.00 0.25 1.51 1.02 0.79 0.92 1.02 0.95 0.33 1.63 1.10 0.73 0.93 1.00 0.83 0.50 1.65 1.03 0.78 1.01 1.14 0.94 0.75 1.27 0.86 0.62 0.86 1.09 0.91 1.00 1.52 0.92 0.66 0.83 0.96 0.92 1.50 1.37 0.65 0.47 0.73 0.91 0.88 2.00 1.39 0.95 0.65 0.96 1.15 1.14 4.00 1.13 0.83 0.52 0.88 1.15 1.08 8.00 0.97 0.80 0.52 0.88 1.16 1.00 12.00 0.87 0.87 0.52 0.90 1.04 0.98 A chromatograph of a standard (10)PCB congeners is shown in 10 ppm solution of the ten Figure 6. Beginning at retention time 9.443 min through retention time 53.402 min, prominent peaks for the mono chlorinated congener to the deca chlorinated congener are seen respectively. These retention times were used as a reference in identifying the PCB congener peaks in sample solutions' chromatographs for all materials and at all storage times. 4469 6% Figure 6. 80 a: u: % % i '3". i cl 9.." :: S c .9. N (D N "‘ 1 » 1 II I H M c: i’ v . .. I co ; | H i !‘ | ‘° f 9% 0 co N ‘7.) v 1‘ ‘13 a) ‘0 ("3 a .'!L'{“:‘.. a, c—a F 0 avg. . b. o l w q '.‘Q C a) . 5‘;N -— ‘ ..fli:" 1. L. r i» Gas chromatogram of standard PCB congeners mono(9.443), di(11.262), tri(14.578), tetra(19.380), penta(21.535), hexa(36.068), hepta(37.924), octa(38.445), nona(50.605) and deca(53.402). pc cc thq se; com; Fiqu PVC/; Chrom for 1 el'ftl‘ac the 80. 3inc e”Pecte ”attire. or these capture identifi, 81 The efficiency of the cleanup procedure for polyethylene is shown in Figure 7 and 8. The chromatograph shown in Figure 7 is that of a polyethylene extract with no cleanup applied. This can be compared with Figure 8 which shows a chromatograph of a cleanup sample of the same polyethylene extract. A chromatograph of polystyrene/10 ppm PCB without cleanup is seen in Figure 9, while that of polystyrene/10 ppm PCB after the cleanup procedure is seen in Figure 10. The polystyrene cleanup was not done by the GPC method because it contained chemical compounds with the same retention time as the PCBs congeners used in this study. Thus, it was unable to separate the PCBs from the other extraneous chemical compounds. These same sequence of presentations are also seen in Figures 11 and 12 for PVC/loppm PCB without cleanup, and PVC/10ppm PCB after cleanup, respectively. Like the chromatographs of the other materials, the cleanup procedure for PVC shows a reduction of the extraneous materials extracted from the polymers together with that of PCB during the soxhlet extraction. Since hexane, the solvent used, is very non-polar, it is expected that all extracted compounds were also non-polar in nature. The presence of chlorinated functional groups in any of these extracts would therefore be detected by the electron capture detector of the GC, and thus would create identification problems alluded to earlier. GPC is a form of liquid chromatography in which solute molecules are separated 82 :63 N8; .2 t.nr.WW.II as s waivefitis . m wasfl'll. ..Czltfib H NI 24.8. 25d“; :3 E...» Sod 2? ' Orr I . okra-o: ~10...‘ nmNAuH 3.5.533" ammuw... “3.....- o ‘2.“ .. . «3.2.1.57 .ubru. ..r. ‘0 .....Q .(.I...|.: Figure 7. Gas chromatogram of polyethylene/PCB extract before cleanup. 898 m8... 4 “@V. r- 8o. ., ~ a.“ sea «8. comma. ham. 1‘ II o. ‘. I v . ..rr uni .OTH 2:... .a w a. . 1- v... .11.. 8Q...a*_. a... s was .5... r... .. 1:. 5.1.5.3! ..PIL s «6. 2 BER. Ill .4“... 11rd a “j ..m \- Figure 8. Gas chromatogram of polyethylene/PCB extract after cleanup. :6: L . 2...: ’V)_’ Guc I 23‘ “F 83 . .. p l‘. o § $ g 'i:. F. ‘ N $ 83 8 <0 . 3 0.2 o i w r O), ; .. j . o— ‘H I '= 23: ‘ O . . g s ‘ e . 1 N u, a c. .. @553: F g a I . ‘3- o ‘3 ..: 8&2? 2:1’: - a"; II :'0 III 0"... - .‘ ‘ ;;;;g::§ar 2! h v 5‘35. Cuiw ’.'.-?'35 ..‘?.’-‘ ‘3?“ C - . r. .:._'.a.' ;?:;i..: .:.-; “"5??- i s: sin? «A. ’1 F‘fiififiiq '2'. 33:21:55 ’:;;§=;,-,..figgz';'§a"‘ k .. :85?) 1 {gig}? gm}; 3‘ ”it"? I W'llfl' Y. '13-'33“ . %{.-'|i!g E‘uhlL ‘ $31.] ; llI J l . ‘ '5'“ . | 1| M: . YmAMMm t' ‘3; E r, 7| 39.61? .I. v 749‘ any —- 6 a ' h ‘ \. ‘33”: r :5 — ,— - 0.- 2:43:33 [3): 3' 93.469 25.992 -\_. fl}; “fig; ‘1.- '0' ." a. g I I ! -—-——" inimi 73445 ~ 7'" 1&3! - 48.308 4.138!) . .3. — '9 '3.” V I Figure 9. Gas chromatogram of polystyrene/PCB extract before cleanup. , V Figure 10. Gas chromatogram of polystyrene/PCB extract after cleanup. 5:2: 5 w: 25:: 3% §E ii «a .i 8 9;? .'. ’2‘; g . . ' #2:; “é = I: __; - g _ . ‘ . A - P l. . ,1. . ‘ H” \‘ i 2? 220 28.594 .i' 13% A: rm 1mm.“ 47.426 ' (A AC! .111 .LIII.I tn All. 57.889 Fawn p11 84 $1 coda . 80. um.“ 0c rib)! 831 1:1 Ls@622] 9%....“ N . as. $3 . $1.». $0. 2 3 8a.»! Figure 11. Gas chromatogram of PVC/PCB extract before cleanup. 1' as 1 $7.. . a? 9.. .5va N8. . a. . Pneptdvxl 8N.. .3651. L11. Ea .. nuns... _ 8W? x. . ..v. - .1.1.1..“ .o . Men...“ ”mew?!“ 0 see .3 _ 8%....” .fi 'I‘Ul Gas chromatogram of PVC/PCB extract after cleanup. _Figure 12. .. .53.. . select The so gel a] Smalle extent molecu of dec column order Separa affini packin facili most . floris The Order the Ch extran. the in Errors ““1121 POLY] The the Do 85 selectively as they permeate the pores in the column packing. The solution molecules larger than the largest pores of the gel are excluded from the pores because of their size. Smaller particles penetrate the gel particles to a varying extent, depending on their size and shape. The solution molecules are therefore eluted from the gel bed in the order of decreasing molecular size. The florisil and silica gel columns on the other hand, used.the principle of1adsorption in order to effect separation of the eluted compounds. Separation in this ion exchange method depends on differing affinities of the eluting compounds for the chromatographic packing material. The use of the different elution solvents facilitated the subsequent removal of the PCBs while leaving most of the extraneous compounds still adsorbed to the florisil and or silica gel. The cleanup procedures used in this study were necessary in order to aid identification of the PCB peaks, and to reduce the chances of overlapping of these peaks with that of the extraneous compounds. If indeed this were to occur, errors in the integration of the peak areas would result, and thus errors in the interpretation of the results. POLYMERIC MATERIALS POLYETHYLENE The weights of PCB congeners (ug)in the water exposed to the polyethylene as a factor of time (hours) are presented graphically in Figures 13 and 14. Figure 13 indicates the noun; 7: wh C. .t. Fic ‘ 8 6 Weight of PCB congeners ln water exposed to 1 ply thick polyethylene stored at 25 degrees C. 0.30000 . ._ o o 06 0.25m 4- 3 E co 5 : 0.20000 . I. < a 2. E 2 Td—chloro g D. c E 0.15000 0 a ‘ 01 0 g Telra-chloro u m 0 0.10000 .- n. '6 Penta-chloro E .9 0 0.05000 1 -chloro L 0.00000 1 : . r . 0.000 2.000 4.000 6.000 0.000 10.000 12000 Time (hrs) Figure 24. Weight of tri to penta-chloro PCB in water exposed to 1 ply PS stored at 25° C. Weight of PCB congeners in water exposed to 1 ply thick polystyrene stored at 25 degrees C. 0.30000 g 025000 .° Deca-chiomPCB___T e . ..‘_____..—————— 3 ". , E Nona-emomece 2 I 5 020000 ;, '5 3 9'3. 1.: E 015000 Octa-chloroPCB g: N» O o . in ———§—.: 2 0.10000 . Hepta-chiomPC ‘6 a S .9 C 3 0.05000 - - 0.00000 1 : : . 1 : 0.000 2.000 4.000 0.000 8.000 10.000 12.000 Time (hrs) Figure 25. Weight of hexa to deca-chloro PCB in water exposed to 1 ply PS stored at 25° C. Figure 97 Weight of PCB congeners in 1 ply thick polystyrene film immersed in water o 3 stored at 25 degrees C. 2 o. E 0.25 -~ 3 .9. ii TL 01 .. .S 3.2" g 0.15 l- g, Tri—chioro 8 \//‘\’ m 0.1 - o c. "6 ’0'} 4.000 0.000 0.000 10.000 12.000 Time (hrs) Figure 26. Weight of tri to penta-chloro PCB uptake by 1 ply P_S stored at 2 5° C . Weight of PCB congeners In water exposed to 1 ply thick polystyrene at 25 degrees C. 0.0000 0.2500 ‘- Welght of PCB congeners (up) in plastic sample 0.0000 ‘ f ; 4' § : 0.000 2.000 74.000 0.000 0.000 10.000 12.000 Time (hrs) Figure 27. Weight of hexa to deca-chloro PCB uptake by 1 ply PS stored at 25° C. Figure 2 hexa-chl Pl. 9 8 Plot of Mtlfdinf vs time for PCBs sorption by 1 ply thick polystyrene film exposed to water at 25 degrees C. 0.5 0.45 «- 0.4 ~- 0.35 -- 0 3 e Trilexp a Tetralexp "é _ P 1110/ Hexale g 025 ‘r + + o 0 exp + :0 2 + 0.2 ‘F 0.15 <- 0 0.1 <~ ° 0.05 1 ‘ ‘ . 3 “ hW—‘n 0 1 0% I I f I 0.000 2.000 4.000 0.000 0.000 10.000 12.000 Time (hrs) Figure 28. Calculated and experimental isotherm for tri to hexa-chloro PCB uptake by 1 ply PS stored at 25° C. Plot of MtIMinf vs time for P085 sorption by 1 ply thick polystyrene film exposed to water at 25 degrees C. 0.5 0.450 eHeptaIexp aOdaIexp 0,4 .. aNona/exp +Oecalup MilMlnf E FD 0.000 2.1!” 4.000 0.1!” 0.1!!) 10.1!” 12.1.!” 11.11. (hrs) Figure 29. Calculated and experimental isotherm for hepta to deca-chloro PCB uptake by 1 ply PS stored at 25° C. -(0 .1... Boy cozacow Figure .5. kw M .6 M «.1. L30: LOQ 000~ NUQ «€00..an o 1 99 Sorption diffusion coefficient (Ds) for PCBs by 1 ply polystyrene immersed in water at 25 degrees C. 14 A ‘ O N A A O _.l Sorption (05) for PCBs x10“10 (cm/sec) T11 Tetra Penta Hexa Hepta Nona Figure 30. Comparison of diffusion coefficients for all congeners by 1 ply PS stored at 25° C. Rate loss of PCB congeners from water exposed to 1 ply polystyrene stored at 25 degrees C. _45 40 .- Percent PCB loss per hour T11 Teh Penta Hexa Hepta Octa Nona Dace Figure 31. Comparison of rate loss for all congeners from water exposed to 1 ply PS stored at 25°C. Fig.1 6 0 ‘ 00:230.. ”mom «COULIQ 20.0 J.- 100 Percent PCBs retained in water exposed to 1 ply thick polystyrene stored at 35' 100.00 _k degrees C. -o—Tfl -o- Tetra + Penta +Hen Percent PCBs retained 10.00 0 0.00 .. _ : 4 : 0.000 4.000 6.000 12.000 16.000 20.000 24.000 Time (hrs) Figure 32. Percent tri to hexa-chloro PCB congeners retained in water exposed to 1 ply PS stored at 35° C. Percent PCBs retained in water exposed to 1 ply thick polystyrene stored at 35 degrees C. 120.00 I +Hepta 100.00 :1‘ +00. J11 +Nona 'o 0 1 g ”-00 N +0000 0— l ( 3 1 2 l .3 O 60.00 < E '\ ‘c‘ o \. 2 g 40.00 .. 1 20.00 0.00 : v _ : : 0.1110 4.000 0.000 12.000 16.000 20.” 24.” Time (hrs) Figure 33. Percent hepta to deca-chloro PCB congeners retained in water exposed to 1 ply PS stored at 35° C. Figu 0.3 Figure 35 a.._._— - 1 0 1 Concentration PCBs In 1 ply thick polystyrene immersed in water at 35 degrees C. 0.3 _ -e—T11-chlo o “'5. +Tetra-chlo :5: 025 " +Penta-chlo .8 0' 02 , u - s d- ’6. a S c 0.15 .- 0 Di c o 0 8 o. 01 .6 _ -f 15, 1 '3 i c o ' . . . . . 0 2 4 6 6 10 12 Time (hrs) Figure 34. Weight of tri to hexa-chloro PCB uptake by 1 ply PS stored at 35° C. Concentration PCBs in 1 ply thick polystyrene Immersed in water at 35 degrees C. 0.3 o r '15. E 0.25 .. a M .2 ‘6 .2 a. 0,2 .. ,5 - c ‘3 3. - a g 0.15 -- c I) 01 -:: 8 8 0.1 q. +Haua-chlo ‘L 4.001.010. .- o E +Nona-chio ca '6 0'05 ‘ +Deca-cflo Z . ll . y : . . 0 2 4 6 6 10 12 Time (hrs) Figure 35. Weight of hepta to deca-chloro PCB uptake by 1 ply PS stored at 35° C. 0.! Figure 3 0.5 0.1 0.4: 03 0.31 0.2: MUM/n! 0-2: 0.15 0.10 1 O 2 Plot of Mthlnf vs time for PCBs sorption by 1 ply thick polystyrene exposed to water at 35 degrees C. 0.5 0.45 h oTrllccp Antwerp 0 ‘ .. a Peru/exp + Hexa/exp 035 J» 0.3 b- .E g 0.25 .. 2 AL * 0 —£— a 0 1‘ 4 4 + : 0.000 2.000 4.000 6.000 0.000 10.000 12.000 Time (hrs) Figure 36. Calculated and experimental isotherm for tri to hexa-chloro PCB uptake by 1 ply PS stored at 35° C. Plot of MtIMinf vs time for P085 sorption by 1 ply thick polystyrene exposed to water at 35 degrees C. 0.5000 0.4500 4- 0.4000 ‘- 0.3500 «- 0.3M 4* 0.2500 ~- MUMlnf 0.2000 ‘- 0.1500 ‘- 0.1000 - - 0.0500 r 0.000 2.000 4.000 0.000 . 0.000 10.000 ' 12.000 ' 11me (hrs) Figure 37. Calculated and experimental isotherm for hepta to deca-chloro PCB uptake by 1 ply PS stored at 35° C. 1 O 3 Percent PCBs retalned in water exposed to 5 ply thick polystyrene film at 25 degrees C. +111 + Tm + Penta u E '3 -O- HOXI 8 a an 0 0. E o 8 o a. 12.000 16.000 20.000 24.000 Time (hrs) Figure 38. Percent tri to hexa-chloro PCB congeners retained in water exposed to 5 ply PS stored at 25° C. Percent PCBs retained ln water exposed to 5 ply thick polystyrene film at 25 degrees C. 120.00 k —o— Hopi: 100.00 “‘ '. + 06.: l‘. + None '3 00.00 .. c -4— Doc- 3 8 8 00 00 d. .0 c O 9. g 40.00 L 20.00 « 0.00 : f : : + v 0.000 4.000 0.000 12.000 ' 10.000 20.000 24.000 Tlme (hrs) Figure 39. Percent hepta to deca-chloro PCB congeners retained in water exposed to 5 ply PS stored at 25° C. 1 0 4 Weight of PCBs in 5 ply thick polystyrene Immersed in water at 25 degrees C. 0.5 +Trl-etio 0.45 0 . Till“ . 0.4 .. +Perls<=hle c mar-Handle E 0.3 0 O a 03 .. .5 A a a 0.25 .. 10 8 a 0.2 0 f/.,a-‘x'“\~“~‘_~- u '5 x/”/’.‘ x-w-M'" E 0.15 <- ,1 e 3 16 20 24 Time (hrs) Figure 40. Weight of tri to hexa-chloro PCB uptake by 5 ply PS stored at 25° C. Weight 01 PCBs in 5 ply thick polystyrene immersed in water at 25 degrees C. 0.5 045 0 +Hspts-dile +Octe-chto :- 0.4 + $0.3 0 +Decs-ehlo 5 °3 ‘ a a a i'é '6 E .9 e 3 0 4 8 12 1e 20 24 Time (hrs) Figure 41. Weight of hepta to deca-chloro PCB uptake by 5 ply PS stored at 25° C. 1 0 5 Plot of Nit/Mint vs time for P083 sorption by 5 ply polystyrene tilm exposed to water at 25 degrees C. , 0.3 e Tnlexp +‘l’rvcel A Tetra/exp +Tetrslcsl 025 -- x Penta/exp +Pentslcsl + "Mexp -e-l~lexslcsl ‘1‘ 5 0 1: I 12 16 20 24 111110 (hrs) Figure 42. Calculated and experimental isotherm for tri to hexa-chloro PCB uptake by 5 ply PS stored at 25° C. Plot of Nit/Mint vs time for PCBs sorption by 5 ply polystyrene fllm exposed to water at 25 degrees C. 0.5 o_45 .. el-leptslexp Loan/exp 0-4 L nNonslexp +Decslexp 0.35 1* 0.3 1r 3 E 035 .. + 02 J /; H 0.15 .. ' 0.1 - «0.05 9 . 4 12 16 20 24 Time (hrs) Figure 43. Calculated and experimental isotherm for hepta to deca-chloro PCB uptake by 5 ply PS stored at 25° C. Fi Figu 106 Sorption diffusion coefficient (05) for PCBs by polystyrene films of various thicknesses and exposed to water at 25 and 35 degrees C. Iiplylsc. I59|y25 C. awryzsc. Sorption (0:) for PCB: x10"10 (cm/sec) Tn‘ Tetra Penta Hexa Hepta Octa Nona Dec- Figure 44. Comparison of sorption coefficients for all congeners by 1 and 5 ply PS stored at 25° C. and 1 ply PS stored at 35° C. Rate loss of PCB congeners irom water exposed to polystyrene films of various thicknesses at 25 and 35 degrees C. _Ilply35C. ‘ ISply 25 c. B1piy 25c. Percent PCB loss per hour A O Hen mm. N None 0.0- Figure 45. Comparison of rate loss for all congeners exposed to 1 and 5 ply PS stored at 25° C. and 1 ply PS stored at 35° C. POLYV The the w of tin 47 sh hepta- water) the s sorpti using polyst EXperi shown . Summar. Congenr t0 thos that th ei'ftent 107 POLYVINYL CHLORIDE The graphical representations of the analysis for PCBs in the water and exposed PVC at 25°C are shown similarly to that of the polyethylene and polystyrene plastics. Figures 46 and 47 show the weights of tri-chloro to hexa-chlorinated and hepta-chloro to deca-chlorinated PCB, respectively, in the water, while Figures 48 and 49 similarly show the weights of the same congeners in the polymer. Calculations of the sorption diffusion coefficients for PVC were also performed using the same mathematical models as for polyethylene and polystyrene. Comparisons of these graphic results with the experimental PCB concentrations were also plotted and are shown in Figures 50 and 51. The bar graph shown in Figure 52 summarized the sorption diffusion coefficients for the PCB congeners in the PVC material. These results are also similar to those obtained for polyethylene and polystyrene in showing that the lower chlorinated congeners were sorbed to a greater extent than the higher chlorinated molecules. chever, in this case, the presence of the nona-chloro and deca- chlorinated congeners were not detected in the material. The rate loss from the water was similar to polyethylene and polystyrene in showing that the lower chlorinated congeners had a greater rate loss than the higher chlorinated ones. This is displayed graphically in Figure 53. A comparison of the sorption diffusion coefficients for each congener exposed to the different polymers can be seen in Figure 54. The corresponding rate loss from the exposed water 108 can be viewed in Figure 55. Weight of PCB congeners in water exposed to PVC film at 25 degrees C. 0.30000 * +Trl +Tetre E. +Pents +Hexs ." Weight of PCB (ug) in 76 ml water sample .0 .o g .0 g 0.00000 . . . , ; - , 0.000 2.000 4.000 6.000 6.000 mm 12.000 Time (hrs) Figure 46. Weight of tri to hexa-chloro PCB in water exposed to 1 ply PVC stored at 2 5° C . Weight of PCB congeners in water exposed to PVC film at 25 degrees C. i 0.15000 «- 0.10000 «- Welght of PCB (ug) in 76 ml water sample -e—Heus +0010 g l +Nons +00“ 0.00000 : : ' r : 4 0.000 2.000 4.000 6.000 _ 6.000 10.000 12.000 . . Time(hrs) . Figure 47. Weight of hepta to deca-chloro PCB in water exposed to 1 ply PVC stored at 25° C. 109 Weight 01 PCB congeners in polyvinyl chloride (PVC) Immersed In water at 25 degrees C. 0.5 Q‘S<> 04-» 035-» 03L Weight 01 PCB (ug) congeners in plastic sample 0.“ 2.“ 4.000 6.“ 0030 10.!!!) 12.“ Time (hrs) Figure 48. Weight of tri to penta-chloro PCB uptake by 1 ply PVC stored at 25° C. Weight of PCB congeners In polyvinyl chloride (PVC) Immersed in water at 25 degrees C. 0.5 0.45»- 0.41» 0.5 u- as.) 0254» 02 .. 0.15 4- W a Weight 01 PCB (119)00an in plastic sample 0.000 2.000 4000 0.000 0.000 10000 12.000 Time (hrs) Figure 49. Weight of hexa to deca-chloro PCB uptake by 1 ply PVC stored at 25° C. MUMInf Figure MUMinf a he 1 1 0 Plot of MtIMinf vs time for PCBs by PVC exposed to water at 25 degrees C. 0.5 eTriIexp 0.45 «- ATetralexp 04 «- oPenta/exp 0.35 <- 0.3 4 - E g 0.25 .. 0 . 2 e 4) A a 1! f 0.000 2.000 4.000 6.000 6.000 10.000 12.000 Time(hrs) Figure 50. Calculated and experimental isotherm for tri to penta-chlo PCB uptake by 1 ply PVC stored at 25° C. Plot of MUMinf vs time for PCBs by PVC exposed to water at 25 degrees C. 0.5 ‘ L eI-iexslexp 0’45 ‘ aHeptalexp 0.4 1» 0001510” 0.35 0 0.3 “- .- .E g 0.25 ..- 2 0.2 r 0.15 ‘ :' ///u ’+ 0.000 0.000 10.000 12.000 Time (hrs) Figure 51. Calculated and experimental isotherm for hepta to octa-chlor PCB uptake by 1 ply PVC stored at 25° C. I! I! II ‘ 1 4 0 ..5 3 Z Z .50: Lou neo. mom «causal Figure I Figure EaflEo. aforx 30¢ 3.39 ccznrow 111 Sorption diffusion coefficient (Us) for PCBs by 1 ply PVC immersed in water at 25 degrees C. 14 .3 N 10- Sorptlon (Ds) for PCBs x10*10 (cm/sec) T11 Tetra Penta Hexa Hepts Octa Nona Decs Figure 52. Comparison of diffusion coefficients for all congeners by 1 ply PVC stored at 25° C. Rate loss of PCB congeners from water exposed to 1 ply PVC stored at 25 degrees C. 45 40-» 35.. 6- :1 O ‘E 30“. 01 a. 3 25+ .9 m 0 20-- 0. fl 8 15.. 2 01 0. 10.. 5‘- 04 T11 Tetra Penta Hexa Hepts " Ode Nona Dee- Figure 53. Comparison of rate loss for all congeners exposed to 1 ply PVC stored at 25° C. Cu i- .ld‘|l d a < 1 5 w 5 w u... m u... w e 530... 500 000. 001 «COULOQ m .1 F < 45» figure .0325. are: amok. .0. .00. Escrow 1 1 2 Sorption diffusion coefficient (05) for PCBs by 1 ply plastic materials Immersed In water at 25 degrees C. Sorption (Os) for PCBs x10‘10 (cmlsec) Figure 54. Comparison of sorption diffusion coefficients for all congeners by 1 ply PE, PVC and PS stored at 25° Rate loss of PCB congeners from water exposed to 1 ply plastic films stored at 25 degrees C. 45W I IPE lllPVC IPS Percent PCB loss per hour ..rlll||"1 "I F “Millie—":5 millage, Tn Tetra Penta Hexa Hepta Nona Ceca Figure 55. Comparison of rate loss for all congeners exposed to 1 ply PE, PVC and PS at 25° C. EQU eac the pol thr POI Cc 113 EQUILIBRIUM PARTITION COEFFICIENT (kg) The results for the partition coefficient determination for each material is shown in Tables 17 and 18. Table 17 shows the results for the polyethylene, PVC and the 1 ply polystyrene stored at 25W3. Table 18 shows the results for the 1 ply polystyrene at 25°C and 35°C, and the 5 ply polystyrene at 25°C . Table 17. Partition coefficient (Kg) for all congeners exposed to all 1 ply materials stored at 25%:. Materials Congeners Polyethylene PVC Polystyrene Tri 115 61 10 Tetra 57 61 13 Penta 53 44 13 Hexa s7 16 39 Hepta 83 12 21 Octa 31 7 12 Nona 12 N/A 10 Deca 11 N/A 5 N/A - not applicable 114 Table 18. Partition coefficients (K;) for polystyrene stored at 25°C and 35°C, and different thicknesses. Materials Congeners 1 ply PS 25°C 1 ply PS 35°C 5 ply PS 25°C Tri 10 222 10 Tetra 13 . 33 16 Penta 13 92 26 Hexa 39 325 17 Hepta 21 40 8 Octa 12' 31 5 Nona 10 15 3 Deca 6 14 1 A comparison of the results for the three different materials in Table 17 shows that polyethylene has the greatest partition coefficient, followed by PVC then polystyrene for congeners tri through penta. For the hexa through octa, the polyethylene still shows the greatest values, but polystyrene shows larger values than that of PVC. Since the nona and deca congeners were not detected in the PVC, no calculations were reported for them. However, the polyethylene also showed a higher partition coefficient than the polystyrene for these congeners. In general, the lower chlorinated congeners showed a higher partition coefficient than the higher analogues. The results for the polystyrene at the different treatments shown ti ti. aq 90 ex} 35c inc con Pig COe] extr 115 in Table 18, also show the general trend where the lower chlorinated congeners have a larger partition coefficient than the higher chlorinated analogues. The polystyrene exposed at 35°C shows a higher coefficient than the 5 ply and the 1 ply stored at 25°C. The 5 ply polystyrene shows higher coefficients for tri through penta chlorination, but lower values for hexa through deca chlorination when compared to the 1 ply polystyrene stored at 25°C. The partition coefficients for the PCB congeners exposed to polyethylene, PVC and polystyrene at 25°C was plotted versus the Leman's molar volume range for these congeners. This plot can be seen in Figure 56. The results of this relationship confirms the principle alluded to earlier, that the potential of the larger chlorinated congeners to penetrate the polymers diminishes with increasing size when in the aqueous medium. It can be seen in Figure 56 that the curve generally decreases with increasing molar volume. A plot of the partition coefficients for the congeners exposed to 1 and 5 ply polystyrene immersed in water at 25 and 35°C, is indicated in Figure 57. Except for the sharp increase in partition coefficient of the hexa-chlorinated congener at 35°C, the relationship is similar to that in Figure 56, where the trend is a decrease in partition coefficients as the molar volume increases. The average percent recovery of the PCB congeners after extraction from the polyethylene, PVC and the polystyrene stored at 25°C was determined to be 69.41%. This was 116 calculated from the area response obtained for the internal standard added to each sample prior to the extraction process, and the area response of the same internal standard injected directly into the GC at the time when the samples were been analyzed. The average percent recovery for the PCBs extracted from the water exposed to the polymeric materials was similarly determined to be 114.80%. The higher percentage observed in the recovery for the congeners in the water phase may be due to errors in measuring the internal standard prior to its addition to the sample, or errors introduced by the automatic sampler of the GC. The lower percentages seen in the polymeric phase may also be due to errors similar to those mentioned for the higher percentages. However, in the case of the lower percentages, the higher partitioning of the congeners for the polymers, increases the difficulty of extracting the PCBs from materials. Thus, this may also be a factor in increasing the errors introduced into the analytical process. This problem is also confirmed by the relatively low recovery for the congeners in the oil which was calculated to be 23.43%. 1 1 7 Comparison of Leman's molar volume with partition coefficients for all congeners exposed to polystyrene, PVC and polyethylene at 25 degrees C. 120.000 4 +W 100.000 ‘- +wc A +1 W PS 25 g 00.000 «- ‘E .2 v3 00.000 4 § \ C O 5 40.000 »» 2 20.000 -» 1 0.000 4 4 4 4 . .. . 230 250 270 290 310 330 $0 Leman's molar volume (cm‘3lmol) Figure 56 . Plot of partition coefficient vs molar volume of the PCB congeners exposed to PE,PVC and PS stored at 2 5°C . Comparison of Leman's molar volume with partition coefficients for all congeners exposed to polystyrene at 25 a 35 degrees C. 350.000 'Ir-x‘ +‘ w PS 25 300.000 »» f' 3.. f ‘. -I-1 w PS 35 250.000 l» , ‘1 +51»; 95 as i J ‘ g a v ,0 e. Partition coefficient (Ks) § § .m/ g s A Y 50.000 4 » 0.” t . t t . t 230 250 270 200 310 330 350 Leman’s molar volume (arms/moi) Figure 57. Plot of partition coefficient vs molar volumes of the PCB congeners exposed to 1 and 5 ply PS stored at 25 and 35°C. 118 STATISTICAL ANALYSIS Full one-way ANOVA tables of all the statistical analyses performed in this study can be seen in appendix 1. A summary of these ANOVA results for the steady state concentrations of the PCB congeners in the three materials (PE, PVC, and PS) stored at 25°C can be seen in Table 19. This table shows the means and standard deviations of these concentrations, the statistical F-test values and the calculated probabilities (p) . for each congener exposed to each material. No data are shown for nona and deca for PVC since the concentrations of these congeners were below the detection limit of the GC. A similar summary for all congeners exposed to 1 and 5 ply polystyrene stored at 25°C and exposed to 1 ply polystyrene at 35°C, can be seen in Table 21. Table 19. Summary of means, std. dev., statistical F and p data for PCBs in water exposed to polyethylene, PVC and polystyrene stored at 25°C. Congeners PVC polyethylene polystyrene F P means SD means SD means SD Tri 0.2164 0.05 0.4213 0.12 0.1078 0.03 53.52 0** Tetra 0.1357 0.05 0.2551 0.05 0.0560 0.01 111.7 0** Penta 0.0953 0.02 0.2232 0.04 0.0391 0.01 135.9 0** Hexa 0.1212 0.07 0.2412 0.04 0.2392 0.05 17.41 0** Hepta 0.1261 0.09 0.2436 0.04 0.1910 0.04 11.72 0** Octa 0.0816 0.06 0.1664 0.03 0.1351 0.03 12.26 0** Nona N/A 0.1221 0.03 0.0607 0.02 32.90 0** Deca N/A 0.1059 0.03 0.0527 0.02 30.24 0** ** significant at 0.001 level, N/A - not applicable. n=37 119 Table 20. Summary of statistical data showing areas of significant.uptake.differences between congener for polyethylene, PVC and polystyrene in water stored at 25°C. Tri, tetra and penta-chloro polystyrene PVC PVC * polyethylene * * * significantly different Hexa, hepta and octa-chloro polystyrene PVC PVC * polyethylene * * significantly different Nona and deca-chloro polystyrene PVC PVC N/A polyethylene * N/A * significantly different 120 Table 21a. Summary of means, std. dev., statistical F and p data for PCBs in water exposed to 1 ply PS stored at 25°C and exposed to 1 ply PS at 35°C. Congeners 25 degrees C 35 degrees C (polystyrene) F means SD means SD Tri 0.108 0.028 0.1901 0.034 45.95 0.0** Tetra 0.056 0.013 0.0770 0.010 18.37 0.0** Penta 0.039 0.009 0.0522 0.015 5.944 0.02 Hexa 0.239 0.049 0.2124 0.019 4.137 0.05 Hepta 0.192 0.039 0.1870 0.014 0.190 0.666 Octa 0.135 0.029 0.1524 0.018 2.941 0.096 Nona 0.061 0.024 0.1972 0.034 97.91 0.0** Deca 0.053 0.017 0.2262 0.066 57.11 0.0** ** significant at 0.001 level; * significant at 0.005 level. n=37 Table 21b. Summary of means, std. dev., statistical F and p data for PCBs in water exposed to 1 and 5 ply PS stored at 25°C. Congeners 1 ply thick 5 ply thick (polystyrene) F P means SD means SD Tri 0.108 0.028 0.1089 0.027 0.008 0.929 Tetra 0.056 0.013 0.0570 0.013 0.035 0.852 Penta 0.039 0.009 0.0660 0.014 25.16 0.0** Hexa 0.239 0.049 0.1753 0.019 23.47 0.0** Hepta 0.192 0.039 0.1506 0.025 13.41 0.001** Octa 0.135 0.029 0.1165 0.026 3.372 0.075 Nona 0.061 0.024 0.1403 0.041 33.27 0.000** Deca 0.053 0.017 0.1238 0.070 9.607 0.004* ** significant at 0.001 level; * significant at 0.005 level. 11:37 121 Table 22. Summary of statistical data for PCB congeners in water and peanut oil exposed to polyethylene at 25 C. Congeners peanut oil Water F P means SD means SD Tri 2.575 0.097 0.081 0.075 15351 0** Tetra 1 . 680 0 . 071 0. 069 0. 065 9952 0** Penta 1.716 0.075 0.063 0.061 10705 0** Hexa 1.819 0.123 0.081 0.074 5820 0** Hepta 1.050 0.071 0.096 0.051 4545 0** Octa 0.790 0.052 0.121 0.042 3769 0** Nona 0.478 0.039 0.202 0.059 474 O** Deca 0.430 0.031 0.190 0.049 524 0** ** significant at 0.001 level. n=7 4 The results for the three materials at the same environmental conditions show that there are significant differences (P<0.05) between the steady statelconcentration.of PCBs in the materials for each congener. The contrasts using the Tukey range test shows where the differences are once significant differences are detected by the ANOVA. These results for the three materials at the same environmental conditions can be seen in Table 20. Significant differences were observed for each material for congeners tri to penta-chlorinated. No differences were obtained for the polyethylene and polystyrene for congeners hexa to octa-chlorinated, while significant difference were detected for the these materials with PVC. Since nona and deca-chlorinated PCBs were not detected in the PVC, only the interrelation of the polyethylene and the 122 polystyrene were tested. Significant differences (P<0.05) were obtained between these materials. Results for the polystyrene at the different temperatures show that there are significant differences (P<0.05) between the steady state concentration of PCBs in the materials for tri, tetra, nona.and.deca-chlorinatedmcongeners. There*were no significant differences between the other congeners. This trend is also the same for the congeners exposed to polystyrene at the different thicknesses. Since only two variables were tested in each case, the Tukey range test was not necessary. ANOVA for the comparing the means for the congeners exposed to polyethylene in the peanut oil medium compared to the water can be seen summarized in Table 22. Significant differences (P<0.05) were obtained for each congener between these liquid media. DISCUSSION EFFECTS OF POLYMER TYPE, TEMPERATURE E THICKNESS ON PCB UPTAKE FROM AQUEOUS AND OIL MEDIA In this study, the polymeric samples were completely immersed in the penetrant, such that both faces were exposed simultaneously to the same concentration. It was assumed that the penetrant entered the film and diffused towards its center. Diffusion behavior differs from polymer to polymer and this paper focuses on two descriptions of this behavior - Fickian and non-Fickian diffusion systems. Fickian behavior is usually associated with rubbery polymers such as polyethylene, while non-Fickian behavior is usually exhibited by glassy polymers such as the PVC and polystyrene used in this study. Polymers in the rubbery state respond rapidly to changes in their condition. A change in temperature for example, causes an almost immediate change to a new equilibrium volume. Glassy polymers on the other hand, tend to be time-dependent. In this case, a change in condition produces an immediate but partial response, followed by a slow, asymptotic approach to equilibrium. Crank (1975) reported that deviations from Fickian behavior are considered to be associated with the finite rates at which the polymer structure may change in response to the sorption or desorption of penetrant molecules. 123 124 Increasing concentrations of many penetrants act similar to that of temperature by increasing the motion of the polymer segments and decreasing relaxation times of the molecular structure. In the case of rubbery polymers such as polyethylene, well above their glass temperature, the polymer chains adjust solquickly to the presence of the penetrant that they do not cause diffusion anomalies (Crank, 1975). In an attempt to classify the diffusion behavior in the different polymeric systems, Alfrey et al. (1966) proposed three descriptive classes: 1. Case I which they referred to as Fickian diffusion. In this instance, the rate ofldiffusion is less than the rate of relaxation of the polymeric chains to the approaching front. In this case, it would be obvious that the process is concentration dependent; 11. Case II in which the rate of diffusion is quicker than the relaxation process. One can see that in this case the diffusion process will be dependent on the relaxation rate of the polymeric chains or as was mentioned above, will be time dependent; 111. Non-Fickian or anomalous diffusion. In this case the diffusion and relaxation processes are described as comparable. Crank (1975) summarizes these behaviors by reference to the types of sorption and desorption curves one should expect for Penetrants with Fickian and non-Fickian systems, these curves are.shown in Figure 58. 125 Fickian Pseudo-Fickian (a) (d l i Sigmoid Two-stage D (b) (c) C 8 A ('——. Figure 58 Fickian and Non-Fickian sorption and desorption type curves. A comparison of the results obtained for the concentration of PCBs in the polyethylene film exposed to water as a function of time, shows that they appear to be Fickian in nature, similar to those in Figure 58 (a) and (b) above. This is not surprising since we know that polyethylene is rubbery at 25°C, the temperature of the study. The departure from an almost perfect Fickian curve by that Obtained for the tri-chlorinated congener in Figure 5 may be ti ex cl: PCB Smaj them. P01 yn Simp 1. 00111011 t0 the produczj lead t o In Can be re a high pa. 126 due to experimental errors. Thus, one should expect that the diffusion process for polyethylene exposed to water was concentration dependent. Similar results were obtained by Stannett and Yasuda (1963) for cyclohexane and benzene through polyethylene. The presence of the PCBs in an aqueous medium also sets up certain conditions which should not be ignored. Water differs from the PCBs and other similar organic compounds in that its molecular size is comparatively small. It can also be described as being in a condensed state with the formation of hydrogen bonds (Chainey, 1989). Thus, in contact with the polyethylene, one should expect clustering of the water molecules to occur. This has been shown in experiments performed by Thornton et a1. (1958) . This clustering will be expected to hinder the transport of the PCBs through the polymer, since only the monomeric species are small enough to diffuse. As a result, these must detach themselves from a cluster before a diffusive jump into the polymers matrix is possible. These processes are by no means simple and clear cut. Chainey (1989) reported that further complications can arise depending on the polymer type exposed to the water, and that the heat of sorption can be large, producing temperature rises of several degrees, and thus can lead to errors in the interpretation of the data. In the case of the polyethylene exposed to the oil, it can be reasonably assumed that both PCB and the oil would have a high partition coefficient for the non-polar polymer. Thus, extensive swelling of the polymer should be expected to occur, t1. PO Chi the fac thrl the that Penet 127 In the case of the polyethylene exposed to the oil, it can be reasonably assumed.that both.PCB and the oil would have a high partition coefficient for the non-polar polymer. Thus, extensive swelling of the polymer should be expected to occur, and both the PCBs and the oil would exert a plasticizing effect on the material. Since the partitioning of PCBs for the oil is also expected to be relatively high, it is not surprising that the rate loss of PCBs from the oil (Figure 23) was quite low for the duration of the study, when compared to that for PCBs in the aqueous phase. Excluding the other complicating phenomena, two major factors can be sighted as influencing the differential sorption diffusion coefficients obtained for the PCB congeners tri to deca chlorinated by all plastic materials used in this study. The first is the molecular size increase of the congeners as we go from tri (229.96 cm3/mol Leman gaseous volume) to deca (352.60 cmfi/mol Leman gaseous volume) chlorinated. The larger the molecular size of a penetrant, the more difficult it is to diffuse through the matrix of a polymeric material (see Table 5 for all congener volumes). Chern (1985) reported that for large non-spherical compounds, the cross-sectional area and volume may be the most important factors that determine the contaminant's ability to pass through an adequately-sized gap. Heindorf (1992) , referred to the use of the parachlor and permachlor concepts in showing that an inverse correlation exists between the size of a penetrant and its diffusion rate through a given polymer. v. (1 inc th reg the not 318( Stru p01); 128 The second major factor influencing diffusion of the PCB congeners is the change in polarity of the biphenyl rings as we go from tri to deca chlorination. Hutzinger et al. (1974) reported that the solubility of PCBs decrease in water with increasing chlorination. This is indeed so because the influence of chlorination disrupts the 11 electron cloud within the biphenyl rings and thus introduces greater non-polar influences. This may well be the reason why a small, but increasing rate loss of the congeners from tri to deca- chlorination in the oil was obtained when exposed to the polyethylene (Figure 23). In addition, the synergistic swelling or plasticizing effect of both the oil and the PCBs on the material would increase the ease in which the higher chlorinated congeners could diffuse into the polymer. In this case, these higher chlorinated congeners exhibited a greater partitioning for the polyethylene than the oil. This pattern was not seen in the case of the polymer/water exposures (Figure 23). This may be due to the effect of increasing molecular size, mentioned above, and its limiting effect on the diffusion of the larger size congeners. This effect is most pronounced when one examines the results obtained for the PVC/water exposure. In this case, the presence of the nona and deca-chlorinated congeners were not detected in the PVC film (Figures 48 and 49) . One should also consider the effect of the tightly packed molecular structure of PVC as compared to both polyethylene and polystyrene. This rigidity and reduced flexibility of its in Co: V81 Sin 901 this 129 structure is also increased by the chain to chain bonding (increased cohersive density) introduced by the presence of the chlorine atoms within the PVC backbone. In addition to these factors, PVC is a glassy polymer with a glass-transition above that of room temperature (25°C) . This factor is expected to decrease the potential for uptake of the PCB congeners, especially the higher chlorinated ones. This result is confirmed by the partition coefficients obtained for the PCB in water exposed to the PVC. Table 4 shows that the lower chlorinated congeners exhibit a higher partitioning for the plastic than the higher analogues. Legett and Parker (1994) obtained similar results for the partitioning of cis-dichloroethane, trans-dichloroethane, 1,1,2-trichloroethane, tetrachloroethane, chlorobenzene, o- dichlorobenzene, m-dichlorobenzene, and p-dichlorobenzene between groundwater and PVC in one instant, and for polytetrafluoroethylene (PTFE) in another instant. They showed that as the size and geometry of these penetrants changed, the partition coefficients for the PVC and PTFE increased with increasing size of the penetrants molecule. Since the use of plasticizers is an essential ingredient in the processing of PVC, its presence introduces further complications to the diffusion process in this study. Vergnaud (1978) reported the diffusion of the penetrant occurs simultaneously with a diffusion of the plasticizer from the polymer towards the liquid environment. He also showed that this phenomenon restricts the diffusion of the penetrant 130 towards the polymer since a loss of the plasticizer induces a tightening and rigidity to the polymeric chains. Thus, time related sorption curves obtained from his results showed an initial uptake of benzyl alcohol by plasticized PVC, which eventually equilibrated but at a lower subsequent concentration value. A similar result for the uptake of the tri-chlorinated congener by PVC was obtained in this study (see Figure 46). Since this is a dynamic process, the net result in a given situation will depend on the plasticizing effect of the penetrant compared to the loss of the migrating initial resident plasticizer within the material. An examination of the results.for the uptake of the other congeners by PVC shows some small deviations from a true Fickian diffusion. This is understandingly so because PVC is a glassy polymer and thus is expected to behave differently from that of polyethylene. Unlike the relatively rapid isotropic swelling response to penetrants by rubbery polymers, glassy polymers exhibit a differential swelling at different geometric parts of the material. Petropoulos and Roussis (1974) reported that during absorption in glassy polymers, the network can open up immediately only to a limited extent since only very small polymer segments can move practically instantaneously. Further opening up of the polymeric structure ‘will occur’ gradually' until the final swelling equilibrium is attained. Fuj ita and Kishimoto (1958) reported two possible explanations for the molecular structural relaxation caused by the penetrating molecules. One is the 131 scission of interchain bonds by a chemical action of the penetrant, while the second is the so called "plasticizing effect" already alluded to above. Thus, if we assume that the fast and slow changes proceed independently of each other, the total change in diffusion.coefficient.D*will be the sum of the two. Crank (1953), on the basis of this assumption, assumed a first-order approach to the equilibrium and expressed the total change in diffusion coefficient by the equation: 60 _ 80.- ac _ ('52) - (y) (36),, T “(De D) (39) where Di is the part of D which can change instantaneously, and is a function of concentration only. De is the equilibrium diffusion coefficient and a is the rate parameter controlling the approach to equilibrium. Other approaches and mathematical models explaining this phenomenon have also been postulated by Petropoulos and Roussis (1974), and Long and Richman (1960). This correction for the diffusion in the glassy PVC and polystyrene were not performed in this study because of the relatively low concentration of the PCB congeners. The low concentration was necessary due to the low solubility of the PCBs in water, being only approximately 12 ppb (Hutzinger, t. or de cor anc Con the (Fig poly: 34 an This 132 1974) . The use of such low concentrations is expected to complicate the situation even further, since it amplifies any experimental errors introduced during the extraction, cleanup and quantification processes for the PCBs. The shapes of the curves for this study are similar to those shown previously by Heindorf (1992), Huckins et al., (1990) and Vergnaud (1986). Vergnaud (1986) reported that curves with such shapes are non Fickian in nature. Since the concentration of the PCBs in this study is so low, they may not result in true non Fickian behavior. Since polystyrene is also a glassy polymer, one should expect similar results to that obtained for the PVC. 'This can be corroborated by reference to Park (1968) who stated that it is not unusual to see an inflection in such non-Fickian curves. 'This is indeed so for the exposure of the polystyrene to the aqueous PCBs at 25°C and at 1 ply thickness. In the case: of the tri-chlorinated. congener (Figure 4-14), the deviation from Fickian diffusion is most pronounced when compared to the other congeners. It appears that the tetra and penta-chlorinated congeners which were lower in concentration in the polymer is more Fickian in behavior than the other congeners which were at a higher concentration (Figures 26 and 27). When the results for the concentration of the PCBs in the polystyrene exposed to the water at 35°C are examined (Figures 34 and 35), one sees a more Fickian type diffusion behavior. This is understandable so because the increased temperature, <31 35 Ch. reg inc siti Th1S the 0018‘ 133 in addition to influencing other factors, would indeed facilitate an increased diffusion rate by lowering the activation energy, and increasing the penetrant molecular motion and thus increasing its concentration at the liquid/polymer boundary. This is also mirrored by the increased rate of loss of PCBs from the water (Figure 45) as compared to that obtained at ZSTL This result is confirmed by the increased partition coefficient obtained for 35°C as compared to the lower value obtained at 25°C (Table 18) . This shows that the PCBs favor polystyrene to a higher extent at 35°C than at 25°C in a aqueous system. Kwapong and Hotchkiss (1987) obtained similar' results for the partitioning of limonene in an aqueous system for LDPE at temperatures of 6, 23 and 3803, showing the highest coefficient at the elevated temperature of 38°C. Although the results obtained for the PCB concentration in polystyrene at the increased thickness of 5 ply show an increase diffusion coefficient, the curves (Figures 43 and 44) are not as Fickian in nature when compared to the those at 35%: (Figures 36 and 37). Chainey (1989) reported that any change in shape of the material due to swelling will be resisted by the unswollen part in glassy polymers. Thus, increasing thickness should not be expected to ameliorate this situation, at least during the initial diffusion process. This is also alluded to by (Chainey 1989) who mentioned that the initial swelling in the material allows further penetrant molecules to enter. Plasticizing effect of these molecules th V0 liar Eat 134 allows viscous creep to occur more rapidly, thus, the rate of entry of penetrant molecules increases, and so on until equilibrium is reached. Thus, as was expected, the increased thickness did not result in a relatively large diffusion coefficient increase by polystyrene when compared to the effect of the increased temperature from 25 to 35%: (Figure 45). This is also confirmed by the results for the partitioning of the PCBs between the 5 ply material and the water as seen in Table 18. One even sees in the case of the hexa to deca chlorinated congeners that the partition coefficient is higher in the 1 ply as compared to the 5 ply material. This result should be considered carefully because there seems to be conflict between it and the results obtained for the sorption.diffusion coefficient for the 5 ply and the 1 ply polystyrene. One should remember that the dimensions of the material pieces were approximately the same. However, by using pieces that were 5 ply (5 times the thickness of the 1 ply), an increase in the weight per piece was obtained. Thus, unless the increase in concentration of the PCBs in the 5 ply material is sufficiently large, errors could be introduced in the interpretation of the result. This could be so because the greater the weight of material per unit volume, the lower would be the weight of PCB per gram (concentration) of material, when.compared to the same weight of PCB in.a thinner material. A comparison of the sorption diffusion coefficients for CI in Sp; acc res IOWQ eithl high, and 135 all the congeners exposed to all the materials under the same environmental conditions, shows that the polyethylene has the largest coefficient and polystyrene the least (see Figure 55). A similar result was obtained for the partitioning of the PCBs between the aqueous environment and these materials (see Table 4). This confirms the fact that polyethylene is more favorable towards the PCB than the other materials under the same environmental conditions. This should not be surprising due to the fact that polyethylene is a rubbery polymer with approximately 55% crystallinityu The fact that the PVC showed a higher diffusion coefficient than polystyrene may be a result of the extent of its plasticization during processing. One should expect that at a relatively low plasticizer concentration, PVC would be expected to show a low rate of diffusion to:most penetrants, even lower than polystyrene. It should be noted that the polystyrene seem to sorb the higher chlorinated congeners almost as well as the polyethylene and indeed better than that of PVC. In this case, the larger pore spaces within the polystyrene :matrix is better able to accommodate the larger molecules than the other more restrictive materials. In summary, one can see that polyethylene removes the lower chlorinated congeners from the water much better than either PVC or polystyrene. However, polystyrene removes the higher chlorinated analogues as well as does the polyethylene; and better than PVC which cannot remove detectable concentrations of these congeners. The polystyrene exposed to 136 the water at 35°C is more efficient in removing the PCBs than polystyrene at 25°C at 1 or 5 ply thickness. It is extremely difficult to remove the PCBs from an oil environment as compared to an aqueous one when exposed to polyethylene at 25°C. be rel po] det watt than rate an 09 253C. teChni. Fab it ihus, Poss ibl E with P011 CHAPTER 5 SUMMARY AND SIGNIFICANCE OF THE STUDY This study investigated the kinetic characteristics and uptake capacity of commercially available polyethylene, polyvinyl chloride and polystyrene films for the removal of polychlorinated biphenyls from aqueous and lipid solutions. The major results of the study indicated that polyethylene removed the lower chlorinated congeners from the water much better than either PVC or polystyrene. However, polystyrene removed the higher chlorinated analogues as well as did the polyethylene, and better than PVC which could not remove detectable concentrations of these congeners. Exposed to water at 35°C, polystyrene was more efficient in removing PCBs than was polystyrene at 25°C at 1 or 5 ply thickness. The rate of removal of PCBs from an oil environment as compared to an aqueous one, is much lower when exposed to polyethylene at 25°C. These results support the initial prediction that the techniques used in this research can facilitate removal of PCBs in media with properties similar to those investigated. Thus, the results from this study can be used to model possible transfer of environmental contaminants from foods with polar and non polar characteristics to selected packaging materials. 137 an t0) the “11m! was of 1» Packa of th Chemic widely edible p°tentj ii” the contamir cage. rn 138 The use of capillary gas chromatography associated with electron.captureidetection, in conjunction with the use of PCB isomers as reference analytical standards, is a viable analytical technique for the analysis of commercial PCB formulations present found in plastic packaging materials. Prior to this quantitation step, extracts of the PCBs can be effectively cleaned up by the use of high pressure gel permeation chromatography and f lorisil / silica gel ion exchange chromatography. Future efforts can address the modification of this technique for quantification of PCBs in other types of packaging materials. Since 209 congeners of PCBs.arejpresently known to exist, and several researchers have reported on the differential toxicity, bioconcentration potential and bioavailability of these congeners both individually and sub-grouped according to number of chlorination, one congener from each such subgroup was selected for this study. Since data exist showing usage of huge annual quantities of several types of polymeric packaging materials, the ones selected were done on the basis of their variabilities in terms of structural, physical and chemical properties. A variety of liquid foods, ranging widely in polarity are daily consumed, thus the water and edible oil selected as the PCB solvents, represent the daily potential real life situation that we are likely to encounter in the types of food we eat. Yes we should. consume contaminant free food, but in reality this is not always the case. Should we do any thing about it? Yes, we should do 139 whatever we can in order to minimize the risk to the ingestion of toxic contaminants. One obvious application of this study is its use in water purification. Since the PCBs possess a higher partition coefficient for plastic than for water as a solvent, engineering applications of selected plastics for use as adsorbent and absorbent materials in water treatment plants can be investigated for possible use as a purification tool. This technology can be used in public water treatment plants, in commercial bottled water processing facilities, and in other situations where water in its purest form is desirable. Another application of the techniques used in this study is in the reduction of contaminants in foods. Since the nature of the penetrant.is an essential factor for the success of good diffusion, solubility, and permeation rates, one would obviously select a product that lends itself to intimate contact with the polymer. Thus liquids, vapors and gases would be products of choice for this technologyu The chemical composition of these contacting products would of course be an important factor influencing their uptake by the sorbent material. For example, non-polar compounds would be more easily removed than polar ones if the polymeric material is a non-polar material such as polyethylene. With these facts in mind, many solid products can be modified in order to increase their contact with that of the polymeric packaging material. Packaging techniques such as vacuum packaging can be used to accomplish this goal, especially if the product is solid of In Sc 140 semi-solid in nature. Other techniques such as gas flushing can also be used. In this case, a gas or vapor could be selected which has a high solubility for the contaminant of interest. Thus, if the packaged product is a powder for example, this gas could be used to dissolve the contaminant, then be absorbed by the polymer, thus reducing the concentration of the contaminant in the food under study. For liquid foods, the design of the packaging can be considered with a view to increasing the product/package contact. Applications such as the use of 'ribs, increase product compartmentalization, the use of smaller packages and selective decorative designs can be considered to achieve the desired end point. The use of this technology can also be adapted for use as a cleanup tool in various analytical laboratory procedures. To increase the sorptive capacity of selected polymers, the use of highly adsorbent fillers, copolymers, laminates or polyblends can be considered. This technique can be modified in the laboratory for use as a concentration step in a given analysis. If for example the concentration of a compound of interest is too low for accurate analysis, exposure of the solution to an.absorbent.polymer followed by extraction of the compound from the polymer, can be a solution to such analytical problems. 1 Additional studies are needed for the investigation of other variables that were not considered during this study. Some of these can include factors such as: 141 Exposure of the samples to several temperatures both above and below room temperature. The use of agitation to increase the arrival of the penetrant species at the sorbent contact surface. A selection of several other liquid contact media with varying polarities, surface tensions and with PCBs of varying concentrations and congener types. The use of solid, semi solid, vapor or gas solvents for contact with the polymer. The use of polymers different from those used in this study. The use of copolymers, polyblends and laminates as sorbent materials. Exposure of the same polymers to other types of contaminants. APPENDIX 142 APPENDIX 1 . 1a. Analysis of variance for aqueous solution of tri-chloro PCB exposed to 1 ply polyethylene, PVC and polystyrene at 25°C. Variable TRI By Variable PLASTIC ANALYSIS OF VARIANCE SUM OF REAR F F SOURCE D.F. SQUARES SQUARES RATIO PROB. BETwEEN Gnours 2 . .5004 .3042 53.5235 .0000 WITHIN GROUPS 33 .1075 .0057 TOTAL I 35 .7959 lb. Analysis of variance for aqueous solution of tetra-chloro PCB exposed to 1 ply polyethylene, PVC and polystyrene at 25°C. Variable TETRA By Variable PLASTIC ANALYSIS OF VARIANCE SUM OF MEAN _ r r SOURCE D.F. SQUARES SQUARES RATIO PR00.- sETwEER GROUPS 2 .2410 .1205 111.6893 .0000 WITHIN GROUPS 33 .0356 .0011 yorAL 35 .2755 143 1c. Analysis of variance for aqueous solution of penta-chloro PCB exposed to 1 ply polyethylene, PVC and polystyrene at 25°C. Variable PENTA By Variable PLASTIC ANALYSIS OF VARIANCE SUM OF MEAN F F SOURCE D.F. SQUARES SQUARES RATIO PROB. BETWEEN GROUPS 2 .2136 .1053 135.8970 .0000 WITHIN GROUPS 33 .0259 .0008 TOTAL 35 .2395 1d. Analysis of variance for aqueous solution of hexa-chloro PCB exposed to 1 ply polyethylene, PVC and polystyrene at 25°C. Variable HEXA By Variable PLASTIC ANALYSIS OF VARIANCE SUM OF MEAN F F SOURCE D.F. SQUARES SQUARES RATIO PROS. BETWEEN GROUPS 2 .1134 .0567 17.4070 .OOOO WITHIN GROUPS 33 .1075 .0033 TOTAL 35 .2208 144 1e. Analysis of variance for aqueous solution of hepta-chloro PCB exposed to 1 ply polyethylene, PVC and polystyrene at 25°C. Variable HEPTA By Variable PLASTIC ANALYSIS OF VARIANCE SUM OF MEAN F F SOURCE D.F. SQUARES SQUARES RATIO PROB. BETWEEN GROUPS 2 .0832 .0416 ll.7232 .OOOl WITHIN GROUPS 33 .1171 .0035 TOTAL ' 35 .2002 If. Analysis of variance for aqueous solution of octa-chloro PCB exposed to 1 ply polyethylene, PVC and polystyrene at 25°C. Variable OCTA By Variable PLASTIC ANALYSIS OF VARIANCE SUM OF MEAN F F SOURCE D.F. SQUARES SQUARES RATIO PROB. BETWEEN GROUPS 2 .0441 .0221 12.2571 .0001 WITHIN GROUPS 33 .0594 .OOlS TOTAL 35 .1036 145 lg. Analysis of variance for aqueous solution of nona-chloro PCB exposed to 1 ply polyethylene, PVC and polystyrene at 25%:. Variable NONA By Variable PLASTIC ANALYSIS OF VARIANCE SUM OF MEAN F F SOURCE D.F. SQUARES SQUARES RATIO PROB. BETWEEN GROUPS 1 .0226 .0226 32.9020 .0000 WITHIN GROUPS 22 .0151 .0007 TOTAL 23 .0377 lb. Analysis of variance for aqueous solution of deca-chloro PCB exposed to 1 ply polyethylene, PVC and polystyrene at 25%:. Variable OECA By Variable PLASTIC ANALYSIS OF VARIANCE SUM OF MEAN F F SOURCE D.F. SQUARES SQUARES RATIO PROB. BETWEEN GROUPS 1 .0170 .0170 30.2418 .0000 WITHIN GROUPS 22 .0124 .0006 TOTAL 23 .0294 146 li. Analysis of variance for aqueous solution of tri-chloro PCB exposed to 1 and 5 ply polystyrene at 25°C. and 1 ply at 35°C. 0 ‘ ‘ A N A L Y S I S O F V A R I A N C E ‘ ‘ ’ TRI by TEMP THICK Sum of Mean Sig Source of Variation Squares 0F Square F of F .053 2 .027 30.237 .000 "a‘TESPfOCtS .041 1 .041 45.953 .000 THICK .000 1 .000 .008 .929 Explained .053 2 .027 30.237 .000 Residual .029 33 .001 Total .083 35 .002 1:] . Analysis of variance for aqueous solution of tetra-chloro PCB exposed to 1 and 5 ply polystyrene at 25°C. and 1 ply at 35°C. ‘ ‘ ‘ A N A L Y S I S 0 F_ V A R I A N C E ‘ ‘ ’ TETRA by TEMP THICK Sum of Mean Sig Source of Variation Squares 0F Square F of F Main Effects .003 2 .002 11.734 .000 TEMP .003 1 .003 18.370 .000 THICK .000 1 .000 .035 .852 Explained .003 2 .002 11.734 .000 Residual .005 33 .000 Total .008 35 .000 147 1k. Analysis of variance for aqueous solution of penta-chloro PCB exposed to 1 and 5 ply polystyrene at 25°C. and 1 ply at 35°C. ‘ ‘ ‘ A N A L Y S I S O F V A R I A N C E ‘ ‘ ‘ PENTA by TEMP THICK Sum of Mean Sig Source of Variation Squares OF Square F of F Main Effects .004 2 .002 12.582 .000 TEMP .001 l .001 5.944 .020 THICK .004 1 .004 25.158 .000 Explained .004 2 .002 12.582 .000 Residual .006 33 .000 Total .010 35 .000 11. Analysis of variance for aqueous solution of hexa-chloro PCB exposed to 1 and 5 ply polystyrene at 25°C. and 1 ply at 35°C. ‘ ‘ ‘ A N A L Y S I S O F V A R I A N C E ‘ ° ‘ HEXA by TEMP THICK Source of Variation Main Effects TEMP THICK Explained Residual Total Sum of Squares .025 .004 .025 .025 .034 .059 OF a...” 33 35 Mean Square .012 .004 .025 .012 .001 .002 F 11.335 4.137 23.469 11.835 Sig of F .000 .050 .000 .000 148 1m. Analysis of variance for aqueous solution of hepta-chloro PCB exposed to 1 and 5 ply polystyrene at 25°C. and 1 ply at 35°C. ‘ ‘ ‘ A N A L Y S I S O F V A R I A N C E ‘ ‘ ‘ HEPTA by TEMP THICK Sum of Mean Sig Source of Variation Squares 0F Square F of F Main Effects .012 2 .006 8.002 .001 TEMP .000 1 .000 .190 .666 THICK .010 1 .010 13.410 .001 Explained .012 2 .006 8.002 .001 Residual .025 33 .001 Total .037 35 .001 1n. Analysis of variance for aqueous solution of octa-chloro PCB exposed to 1 and 5 ply polystyrene at 25°C. and 1 ply at 35°C. ‘ ‘ ‘ A N A L Y S I S 0 F V A R I A N C E ‘ ‘ ‘ OCTA by TEMP THICK Sum of Mean Sig Source of Variation Squares 0F Square F of F Main Effects .008 2 .004 6.309 .005 TEMP .002 1 .002 2.941 .096 THICK .002 1 .002 3.372 .075 Explained .008 2 .004 6.309 .005 Residual .020 33 .001 Total .028 35 .001 149 lo. Analysis of variance for aqueous solution of nona-chloro PCB exposed to 1 and 5 ply polystyrene at 25°C. and 1 ply at 35°C. ‘ ‘ ‘ A N A L Y S I S O F V A R I A N C E ‘ ‘ ‘ NONA by TEMP THICK Sum of Mean Sig Source of Variation Squares OF Square F of F Main Effects .113 2 .056 49.403 .000 TEMP .112 1 .112 97.909 .000 THICK .038 1 .038 33.272 .000 Explained .113 2 .056 49.403 .000 Residual .038 33 .001 Total .150 35 .004 1p. Analysis of variance for aqueous solution of deca-chloro PCB exposed to 1 and 5 ply polystyrene at 25°C. and 1 ply at 35°C. 0 o . A N A L v s I s o F v A R 1 A N c e o o 0 oecx by TEMP THICK Source of Variation Main Effects TEMP THICK Explained Residual Total Sum of Squares .182 .181 .030 .182 .104 .287 DF 33 35 Mean Square .091 .181 .030 .091 .003 .008 28.861 57.107 9.607 28.861 Sig of F .000 .000 .004 .000 1150 1q. Analysis of variance for peanut oil and aqueous solutions of tri-chloro PCB exposed to 1 ply polyethylene at 25°C. - - Analysis of Variance - - Dependent Variable TRI 8y levels of LIOUID Value Label Sum Mean Std Dev Sum of Sq Cases 1.00 PEANUT OIL 70 2.5753 .0974 .2466 27 2.00 WATER 4 .0808 .0751 .2648 48 mm. 5...... To... '"""'§3"""TS§£S"""TBES?"""T§I?§ """" §§ Sum of Mean Source Squares d.f. Square F Sig. Between Groups 107.5228 1 107.5228 15351.0827 .0000 Within Groups .5113 73 .0070 Eta = .9976 Eta Squared = .9953 1r. Analysis of variance for peanut oil and aqueous solutions of tetra-chloro PCB exposed to 1 ply polyethylene at 25°C. - - Analysis of Variance - - Dependent Variable TETRA 8y levels of LIQUID Value Label Sum Mean Std Dev Sum of Sq Cases 1.00 PEANUT OIL 45 1.6798 .070] .1300 27 2.00 WATER 3 .0688 .0651 .1990 48 Within Groups Total 49 .6487 .0671 .3289 75 Sum of Mean Source Squares d.f. Souare F Sig. Between Groups 44.8453 1 , 44.8453 9952.4116 .0000 Within Groups .3289 73 .0045 Eta = .9964 Eta Squared = .9927- 151 ls. Analysis of variance for peanut oil and aqueous solutions of penta-chloro PCB exposed to 1 ply polyethylene at 25%:. - - Analysis of Variance - - Dependent Variable PENTA 8y levels of LIQUID Value Label Sum Mean Std Dev Sum of Sq Cases 1.00 PEANUT OIL 46 1.7158 .0752 .1469 27 2.00 WATER 3 .0625 .0611 .1752 48 Within Groups Total -------- 49- .6577 .0664 .3221 75 Sum of Mean Source Squares d.f. Square F Sig. Between Groups 47.2352 1 47.2352 10704.7163 .0000 Within Groups .3221 73 .0044 Eta = .9966 Eta Squared = .9932 It. Analysis of variance for peanut oil and aqueous solutions of hexa-chloro PCB exposed to 1 ply polyethylene at 25%:. - - Analysis of Variance - - Dependent Variable HEXA 8y levels of LIQUID Value Label Sum Mean Std Dev Sum of Sq Cases 1.00 PEANUT DIL 49 1.8192 .1232 .3949 27 2.00 WATER 4 .0814 .0743 .2596 48 Within Groups Total 53 .7070 .0947 .6545 75 Sum of Mean Source Squares d.f. Square F Sig. Between Groups 52.1817_ 1 52.1817 5820.4383 .0000 Within Groups .6545 73 .0090 Eta = .9938 Eta Squared = .9876 152 1u. Analysis of variance for peanut oil and aqueous solutions of hepta-chloro PCB exposed to 1 ply polyethylene at 25%:. - - Analysis of Dependent Variable HEPTA 8y levels of LIQUID Value Label Sum 1.00 PEANUT OIL 28 2.00 WATER 5 Within Groups Total 33 Sum of Source Souares- d Between Groups 15.7440 Within Groups .2529 Eta = .9921 Eta Variance - - Mean Std Dev Sum of Sq Cases 1.0500 .1317 27 .0955 .1212 48 _-_ .4392" __- .3529 "7; _ Mean .f. Square F Sig. 1 15.7440 4544.8321 .0000 73 .0035 Squared = .9842 IV. Analysis of variance for peanut oil and aqueous solutions of octa-chloro PCB exposed to 1 ply polyethylene at 25%:. - - Analysis of Variance - - Dependent Variable OCTA 8y levels of LIQUID Value Label Sum Mean Std Dev Sum of Sq Cases 1.00 PEANUT OIL ‘ 21 .7901 .0696 27 2.00 WATER 6 .1209 .0803 48 Within Groups Total 27 .3618 .1499 75 Sum of Mean Source Squares d.f. Square F Sig. Between Groups 7.7395 1 7.7395 3768.5408 .0000 Within Groups .1499 73 .0021 Eta = .9905 Eta Squared = .9810 1153 1w. Analysis of variance for peanut oil and aqueous solutions of nona-chloro PCB exposed to 1 ply polyethylene at 25%:. - - Analysis of Variance - - Dependent Variable NONA By levels of LIQUID Value Label Sum Mean Std Dev Sum of Sq Cases 1.00 PEANUT OIL - 13 .4777 .0387 .0390 ‘ 27 2.00 WATER 10 .2022 .0589 .1632 48 Within Groups Total ---— 23 .3014 .0526 .2022 75 Sum of Mean Source Squares d.f. Square F Sig. Between Groups 1.3119 1 1.3119 473.6661 .0000 Within Groups .2022 73 .0028 Eta = .9308 Eta Squared = .8665 1x. Analysis of variance for peanut oil and aqueous solutions of deca-chloro PCB exposed to 1 ply polyethylene at 25%:. - - Analysis of Variance - - Dependent Variable DECA 8y levels of LIQUID Value Label Sum Mean Std Dev Sum of Sq Cases 1.00 PEANUT OIL 12 .4299 .0313 .0254 27 2.00 WATER 9 .1907 .0489 .1122 48 Within Groups Total 21 .2768 .0434 .1376 75 Sum of Mean Source Squares d.f. Square F 519- Between Groups .9885 1 .9885 524.4161 .0000 Within Groups . 1376 73 .0019 Eta = .9369 Eta Squared = .8778 BIBLIOGRAPHY Ba Ber BIBLIOGRAPHY Adamson, A;W. 1990. Solid-liquid interface-Adsorption from solution. Ch. 11. In "Physical Chemistry of Surfaces", 5th edition, p 426-430. John Wiley & Sons, New York, NY. Banerjee, K., Cheremisinoff, P.N., and Cheng, S.L. 1995. Sorption of organic contaminants by fly ash in a single solute system. Environ. Sci. Technol. 29:2243-2251. Beach, D.L. and Kissin, Y.V. 1986. Ethylene polymers. In "Encyclopedia of Polymer Science and Engineering", 2nd ed. Vol. 6. A Wiley-Interscience publications, New York, NY. Ballschmiter, K. and Zell, M. 1980. Analysis of polychlorinated biphenyls (PCB) by glass capillary gas chromatography. Anal. Chem. 20:302. Benning, C.J. 1983. Plastic Films for Packaging Technology, Applications and Process Economics. Technomic Publications Co., Inc. Lancaster, PA. Billings, C. H. 1991. Accessing risk globally. Public Works. No.9 122:150. Bixler, H.J. and Sweeting, C.J. 1971. The Science and Technology of Polymer Films. Vol. 2. John Wiley & Sons, New York, NY. Brandrup, J. and Immergut, E.H. 1975. Polymer Handbook. 2nd ed. A. Wiley-Interscience Publication. New York, NY. Bremer, P.B. 1988. Films, polyethylene, low and high density. In "Packaging's Encyclopedia". McGraw-Hill, Inc. Highstown, NJ. Cady, L.D. and Ratter, C.J. 1988. Films, polyethylene, linear low-density and ultra low-density. In "Packaging's Encyclopedia". McGraw-Hill, Inc. Highstown, NJ. Carpenter, W. D. 1991. Insignificant risks must be balanced against great benefits. Chemical and Engineering News. 69(1):37 154 Chair Chern 155 Chainey, M. 1989. Transport phenomena in polymer films. In "Handbook of Polymer Science and Technology". Vol. 4. ed. Cheremisinoff, N.P. Marcel Dekker, Inc. New York, NY. Chern, R.T., Koros, W.J., Hopfenberg, H.B. and Stannett. 1985. Material selection for membrane-based gas separations. In "Materials Science of Synthetic Membranes". Lloyd, D.R. ed. ACS Symposium Series. American Chemical Society. p 25-45. Choy, C.L, Leung, W.P., and Ma, T.L. 1984. Sorption and diffusion of toluene in highly oriented polypropylene. J. Polym. Sci. Phys. Ed. 22:707-719. Collins, E.A. and Daniels, C.A. 1975. Physical constants of polyvinyl chloride. Ch V. In "Polymer Handbook". Brandrup, J. and Immergut, E.H. A Wiley-interscience publications, New York, NY. Coutant, R.W., Lewis, R.G. and Mulik, J. 1985. Passive sampling devices with reversible adsorption. Anal. Chem. 57:219-223. Crank J. 1975. "Mathematics of Diffusion". 2nd ed. Clarendon press. Oxford, England. Crank, J. 1953. A Theoretical investigation of the influence of molecular relaxation and internal stress on diffusion in polymers. J. Polym. Sci., 11(2):151-168. Deanin, R.D. 1978. Plasticizers. In "Additives for Plastics". Vol 1. p 215. Academic Press, Inc. New York, NY. Derjaguin B.V. and Churaev N.V. 1981. Structure of the boundary layers of liquids and its influence on mass transfer in fine pores. In "Progress in Surface and Membrane Science". Vol. 14. Cadenhead and Danielli. eds., Academic Press, New York. p.69-130. Doak, K.W. 1986. Ethylene polymers. In "Encyclopedia of Polymer Science and Engineering". 2nd ed. Vol. 6. A Wiley-Interscience Publications, New York, NY. Fang, S.M., Stern, S.A. and Frisch, H.L. 1975. A free volume model of permeation of gas and liquid mixtures through polymeric membranes. Chemical Engineering Science. 30:773-780. Food Marketing Institute. 1994. Trends in the United States: Consumer attitudes and the supermarket. p 72. Fujii Fujit Gabbe Giacix H‘rimm 156 - Fujita, T., Iwasa, J. and Hansch, C. 1964. New substituent constant pi, derived from partition coefficients. J. Am. Chem. Soc., 86:5175. Fujita, H. and Kishimoto, A. 1958. Diffusion-controlled stress relaxation in polymers. II. Stress relaxation in swollen polymers. J. Polym. Sci., 28:547. Gabbett, J.F. 1994. Vinyl-based resins. In "Modern Plastics Encyclopedia Handbook". p 70. McGraw-Hill, Inc. Highstown, NJ. Giacin, J.R. 1993. Chemical Structure and Properties of Polymers. Ch. 7. In "Principles of Aseptic Processing and Packaging". Nelson, P.B., Chambers, J.V., And Rodriguez, J.H. (Eds.), p 81. The Food Processing Institute, Washington, DC. Giacin, J.R., and Brozozwska, A., 1985. Analytical measurements of package components for unintentional migrants. J. Plastic Film and Sheeting. 1:292-326. Gonzales, M.J., Ramos, L., and Hernandez, L.M. 1995. Levels of Coplanar PCBs in Human Breast Milk at Different Times of Lactation. Bull. Environ. Contam. and Toxicol. 54:349-356. Graham, M., Hileman, F., Kirk, D., Wendling, J., and Wilson J. 1985. Background human exposure to 2,3,7,8,-TCDO. Chemosphere. 14:925. Grayson, M. A., Pao, P. 8., and Wolf, C. J. 1987. Transport of methylene chloride in poly(aryl-ether-ether-ketone) (PEEK). J. Polym. Sci. 4(25):935. Hahnfeld, J., and Sugden, J. 1994. General purpose polystyrene. In "Modern Plastics Encyclopedia Handbook". p 56. McGraw-Hill, Inc. Highstown, NJ. Hahnfeld, J.L. and Dalke, 8.0. 1986. Styrene polymers. In "Encyclopedia of Polymer Science and Engineering". 2nd ed. Vbl. 16. A Wiley-Interscience publications, New York, NY. Hanlon, J.F., 1992. Handbook of Package Engineering 2nd Ed. Technomic Publications Co., Inc. Lancaster, PA. Heindorf, M. 1992. Two topics in trace analysis: The development of a solvent-filled polymeric membrane as a passive water sampling device. PhD. Dissertation. Michigan State University Library, East Lansing, MI. Herma Herna Hernal Hilema C Huckin S n l b 157 Hermann R.B. 1972. Correlation of hydrocarbon solubility in water with solvent cavity surface area. J. phys. Chem., 76:2754-2759. Hernandez, R.J., Giacin, J.R., Baner, A.L. 1986. The evaluation of the aroma barrier properties of polymer films. Journal of Plastic Film & Sheeting, 2:187-211. Hernandez, R.J. 1989. Modeling the sorption of water, and the effect of sorbed water on the solubility and diffusivity of oxygen in an amorphous polyamide. Ph.D. Dissertation. Michigan State University Library, East Lansing, MI. Hernandez, R.J. and Gavara, R. 1994. Sorption and transport of water in nylon-6-films. J. Polym. Sci., part B: Polymer physics. 32:2367-2374. Hileman, B. 1990. Action on Great Lakes health hazard urged. Chemical and Engineering news. 68(17):5. Huckins, J. N., Tubergen, M. W., and Manuweera, G. K. 1990‘. Semipermeable membrane devices containing model lipid: a new approach to monitoring the bioavailability of lipophilic contaminants and estimating their bioconcentration potential. Chemosphere. 20(5):533. Huckins, J.N., Tubergen, M.W., Lebo, J.A., Gale, R.W. and Schwartz, T.R. 1990b. Polymeric film dialysis in organic media for cleanup of organic contaminants. J. Assoc. Off. Anal. Chem. 72(2):290-293. Huff, J. E., and Haseman, J. K. 1991. Exposure to certain pesticides may pose real carcinogenic risk. Chemical and Engineering News. 69(1):33. Hutzinger, 0., Safe, 8., and zitko, V. 1974. The Chemistry of PCB's. p 22-39. CRC Press Inc. Cleveland, OH. Hwang, S. and Kamermeyer, K. 1975. Techniques of chemistry: Membranes in separations. Vol. II. p 99. John Wiley & Sons, New York, NY. Jacobson, L. J., Humphery, H. E. B., Jacobson, W. S., Schantz, S. L., Mullin, M. D.,and Welch, R. 1989. Determination of polychlorinated biphenyls (PCBs), polybrominated biphenyls (PPBs),and.1dichlorobiphenyl ‘trichloroethane (DDT) levels in the sera of young children. American Journal of Public Health. 79:1401. Klepic, J.G. 1994. Expandable polystyrene. In "Modern Plastics Encyclopedia Handbook". p 58. McGraw-Hill, Inc. Highstown, NJ. Lez Linn 158 Koros, W.J. and Hellumus, M.W. 1986. Transport properties. In "Encyclopedia of Polymer Science and Engineering". 2nd ed. Supplement. A Wiley-Interscience publications, New York, NY. Kovach, J.L. 1978. Gas-phase adsorption and air purification. Ch. 9. In "Carbon Adsorption Handbook". P.N. Cheremisinoff, F. Ellerbuusch (Eds.), p 331. Ann Arbor Science, Ann Arbor, MI. Kreituss, A. 1981. Free-volume estimates in heterogeneous polymer systems. I. Diffusion in crystalline ethylene- propylene copolymers. J. Polym. Sci. 19:889-905 Kwapong, O.Y. and Hotchkiss, J.H. 1987. Comparative sorption of aroma compounds by polyethylene and ionomer food- contact plastics. J. Food. Sci. 52(3):?61-785. Leggett, O.C. and Parker, L.V. 1994. Modelling the equilibrium partitioning of organic contaminants between PTFE, PVC and ground-water. Environ. Sci. Technol. 28:1229-1233. Leman, R. 1981. Handbook of chemical property estimation methods. p 17. Limm, W., and Hollifield, H.C. 1995. Effects of temperature and mixing on polymer adjuvant migration to corn oil and water. Food Additives And Contaminants. 12:609-624 Malachowski, R.A. 1971. Generalized permeability equation. Ph.D. Dissertation. University of Iowa Library, Iowa City, IO. Manders, P.W. 1994. Raw materials. In "Modern Plastics Encyclopedia '95". 71(12):B-3. McGraw-Hill, Inc. Highstown, NJ. McFarland, V.A. and Clarke, J.U. 1989. Environmental occurrence, abundance, and potential toxicity of polychlorinated biphenyl congeners: considerations for a congener-specific analysis. Environmental Health Perspectives. 81:225-239. Michaels, A.S. and Bixler, H.J. 1961. Rheological properties of aqueous clay systems. J. Polym. Sci. 25:1297 Miller, M.A., Madenjian, C.P., and Masnado, R.G. 1992. Patterns of organochlorine contamination in lake trout from Wisconsin waters of the Great Lakes. J. Great Lakes Res. 18(4):742-754. Ra‘ ROd: 159 Newton, 8. 1990. Residue analysis: What every processor should know. Prepared Foods. 159(6):84. Nielson, L.E., 1974. "Mechanical Properties of Polymers and Composites". Vol. 1. p 1-37. Marcel Dekker Publishing CO. Park, P. 1991. Great Lakes pollution linked to infertility. New Scientist. 131:18. Petropoulos, J.H. and Roussis, P.P. 1974. A discussion of Theoretical Models of Anomalous Diffusion of vapors in polymers. In Polymer Science and Technology. Vol.6. Hopfenbery, H.B. Plenum Press, New York, NY. Poloso, A. 1994. Impact polystyrene. In "Modern Plastics Encyclopedia Handbook" p 56. McGraw-Hill, Inc. Highstown, NJ. Price, H.A., Welch, R.L., Scheel, R.H., and Warren, L.A. 1986. Modified multiresidue method for chlordane, toxaphene, and polychlorinated biphenyls in fish. Bull. Environ. Contam. Toxicol. 37:1-9. Ravenswaay, E.V. 1990. Consumer perception of health risks in food. Paper presented at 40th National public policy education conference. Park City, Utah. Rodriguez, F. (1985). Physical states and transition. In "Principles of Polymer Science". McGraw-Hill Book Co., Highstown, NJ. Rundlof, L.C. 1994. Very low density polyethylene. In "Modern Plastics Encyclopedia Handbook". p 30. McGraw-Hill, Inc. Highstown, NJ. Salame, M. and Pinsky, J. 1962. Permeability predictions. Modern Packaging. 36:153-156. Sanders, M. and Haynes, B. 1988. Distribution pattern and reduction of polychlorinated biphenyls (PCB) in Bluefish Pomatomus saltatrix (Linnaeus) fillets through adipose tissue removal. Bull. Environ. Contam. Toxicol. 41:670- 677. Schmidt, L.J and Hesselberg, R.J. 1992. A mass spectroscopic method for analysis of AHH-inducing and other polychlorinated biphenyl congeners and selected pesticides in fish. Arch. Environ. Contam. Toxicol. 23:37-44. Schwope, A.D., Till, D.E., Ehntholt, D.J., Sidman, K.R., Scien Seltz Smitr Sodel Spuri Stan: Ster SWa i Taka 160 Whelan, R.H., Schwartz, P.S., and Reid, R.C. 1987. Migration of BHT and Irganox 1010 from low-density polyethylene (LDPE) to foods and food-simulating liquids Fd. Chem. Toxic. 25(4):317-326. Science/technology concentrates. 1991. Monitor concentrates water contaminants. Chemical and Engineering News. 9(36):24. Seltzer, R. 1989. Great Lakes pollution remains severe. Chemical and Engineering news. 67(35):5. Smith, L.J. and Hesselberg, R. 1992. A mass spectroscopic method for analysis of AHH-inducing and other polychlorinated biphenyl congeners and selected pesticides in fish. Arch. Environ. Contam. Toxicol. 23:37-44. Sodergren, A. 1987. Solvent-filled dialysis membranes simulate uptake of pollutants by aquatic organisms. Environ. Sci. Technol. 21:855-859. Spurlock, F.C. and Biggar, J.W. 1994. Thermodynamics of organic chemical partition in soils. 1. Development of a general partition model and application to linear isotherms. Environ. Sci. Technol. 28:989-995. Stannett, V. and Yasuda, H. 1963. Liquid versus vapor permeation through polymer films. J. Polym. Sci., 28:645. Stern, S.A., Fang, S.M. and Frisch, H.L. 1972. Effect of pressure on gas permeability coefficients. A new application of free volume theory. J. Polym. Sci. A-2. 10:201-219 Svensson, B., Nelson, A., Hansson, M., Rappe, C., Akesson, B., and Skerfving, S. 1991. Exposure to dioxins and dibenzofurans through the consumption of fish. The new England Journal of Medicine. 324(1):8. Swain, W.R. 1983. An overview of the scientific basis for concern with polychlorinated biphenyls in the Great Lakes. In "PCBs: Human and Environmental Hazards". D'ltri, F.M. and Kamrin, M.A. Eds. Butterworth Publishers, Boston, MS. Takashi, U. 1994. Modifying the sorption capacity of polyethylene for organic compounds. M.S. Thesis, Michigan State University Library, East Lansing, MI. Tanabe, S. 1988. PCB problems in the future: foresight from Wt 161 current knowledge. Environ. Pollut. 50:2-28. Thornton, E,R., Stannett, V. and Szwarc, M. 1958. The permeation of vapors and liquids through polymer films. J. Polym. Sci., 28:466. Toufexis, A. 1989. Dining with invisible danger. Time magazine. 133(13):28. Tulp, M. Th. M. and Hutzinger, O. 1978. Thoughts on aqueous solubilities and partition coefficients of PCB and mathematical correlation. between bioaccumulation and physicochemical properties. Chemosphere. 10:849-860. Umbreit, T.H., Hesse, E.J., and Gallo, M.A. 1986. Bioavailability of dioxin in soil from a 2,4,5,-T manufacturing site. Science. 232:497. Vergnaud, J.M. 1986. Matter transport between plasticized PVC and liquids. In" Liquid Transport Processes in Polymeric Materials". Ed. Vergnaud, J.M. p 105-165. Prentice Hall, NJ. Voiland, Jr. M. P., Gall, K. L., Lisk, D. J., MacNeill, D. B. 1991. Effectiveness of recommended fat-trimming procedures on the reduction of PCB and mirex levels in Brown Trout (Salmo-trutta) from Lake Ontario. J. Great Lakes Research. 17(4):454. Waid, J.S. 1986. PCBs and the environment. Vol I, II, and III. CRC Press, Inc. Boca Raton, FL. Weber, W.J. Jr., 1972. Physicochemical Processes for Water Quality Control. John Wiley & Sons, New York, NY. Wheeler, O.H. and Levy, E.H. 1959. Adsorption of sterically hindered phenols on carbon. Can. J. Chem. 37: p 1235. William, D.M. and Koszdin, S. 1991. Speciation and quantitation of Aroclors based on PCB congener data: application to California mussels and white croaker. J. Agric. Food Chem. 39:1457-1467. Williams, L.L. and Giesy, J.P. 1992. Relationships among concentrations of individual 2,3,7,8-tetrachlorodibenzo- p-dioxin equivalents (TCDD-EQ), and rearing mortality of chinook salmon (oncorhynchus tshawytscha) eggs from Lake Michigan. J. Great Lakes Res. 18(1):108-124. Worthy, W. 1989. Ailing Great Lakes need "urgent action". Chemical and Engineering news. 67(44):22. "71.111111111311118