0.1 11,1. 1,. ‘J' gr 1:... . S‘ m ‘. fi‘fim L .1. p .— '<-'£§'n.".”.€ 1' "it :. r' “raga?- " ‘r v Ag" burr-r . .T M. 19:1; *3 ' E 23". , m ‘ ' A xfilfi‘éfiz‘" y‘x . , : u,»- . fil‘ ‘3? Wm," 25! r w», ”firm '. ‘1 "34‘5”“ ‘ usi‘ifivlolm ”95"“; f _ \ m UBRARIES lll‘lllllllll limit 3 1293 01046 5 W l This is to certify that the ‘ thesis entitled VOLATILIZATION OF VANILLIN AND ORTHO-VANILLIN FROM A CORRUGATED PAPERBOARD SHIPPER AND SORPTION BY A LOW DENSITY POLYETHYLENE PRIMARY CONTAINER presented by SATOSHI MAEKAWA has been accepted towards fulfillment of the requirements for MASM—degree in LAME— Major professor won/x [W 0 Date JUNE 22, 1994 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY Michigan State University PLACE ll RETURN BOXtoromavothbcMckouimnywnoord. To AVOID FINES ntum on or Manda-din. DATE DUE DATE DUE DATE DUE ~ MSU loAn mm Action/Emmi Opportmity Intuition Wt VOLATILIZATION OF VANILLIN AND ORTHO-VANILLIN FROM A CORRUGATED PAPERBOARD SHIPPER AND SORPTION BY A LOW DENSITY POLYEI'HYLENE PRIMARY CONTAINER BY Satoshi Maekawa A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of packaging 1 994 ABSTRACT VOLATILIZATION OF VANILLIN AND ORTHO-VANILLIN FROM A CORRUGATED PAPERBOARD SHIPPER AND SORPTION BY A.LOW DENSITY POLYETHYLENE PREMARY CONTAINER BY Satoshi Maekawa The volatilization of vanillin and ortho-vanillin from a corrugated paperboard shipper and their sorption by a low density polyethylene (LDPE) primary container was investigated. Initial concentration levels of vanillin and o-vanillin in corrugated paperboard, and their equilibrium partition distribution between paperboard and headspace were determined. The solubility coefficients for vanillin and o- vanillin into LDPE, and the equilibrium partition distribution of vanillin and o-vanillin between LDPE and an aqueous phase were also determined. The derived concentration levels and the respective coefficient values were as follows: Initial vanillin and o—vanillin concentration were 0.34 and 0.9 (pg/g), respectively. The average partition coefficient values for vanillin and o-vanillin between paperboard and headspace were 21.6 and 77.5, respectively at 23%:. The average solubility coefficient values for vanillin and o- vanillin in LDPE were 11.3 and 0.3 (Kg/m3 Pa), respectively at 23°C. The average diffusion coefficient value for o-vanillin estimated from the weight up taken by LDPE was 8.5 (m2 /sec x 10-13) . The average partition coefficient values for vanillin and o-vanillin between LDPE and an aqueous phase were 3.8 and 574, respectively at 23°C. ‘ These parameters were incorporated into expressions for estimating the equilibrium.concentration of vanillin in the paperboard and aqueous phase,as well as in the shipper headspace and LDPE primary package system. Estimated values of the paperboard, headspace, polyethylene and aqueous phase concentration for o-vanillin were 1.5 x 10'2 (pg/g), 7.0 x 10' ‘ (pg/cc), 1.4 x 10'1 (pg/cc), and 2.2 x 10" (pg/cc), respectively, and 0.59 (pg/g), 7.7 x 10-3 (pg/cc), 3.7 x 10-2 (pg/cc), and 5.9 x 10-3 (pg/cc), respectively for vanillin. The times to reach equilibrium in the system for this study are very long and are on the order of 103 days for a 20 mil wall thickness LDPE ampule. ACKNOWLEDGMENTS Firstly, I would like to express sincere thanks and appreciation to Dr. Jack R. Giacin, Professor in the School of Packaging and my major adviser, for his great kind counsel and assistance throughout the course of this graduate program. I would also like to express sincere thanks and appreciation to Dr. Ruben J. Hernandez, Assistant Professor in the School of Packaging, and Dr. Alec B. Scranton, Assistant Professor in Chemical Engineering, for being a committee member and their professional advises. I'm.very grateful to Dr. Heidi Hoojjat for teaching how to operate the laboratory equipment and to solve the problems I acknowledge to Dr.Peter Lim, Stanford Research Institute, for measurement support of this study. I also acknowledge financial support for this study provided by Dey Laboratory Inc. and Suntory Ltd. Lastly, I would like to thank to graduate students in the school of packaging for the sharing of knowledge and skills. TABLE OF CONTENTS 1. LIST OF TABLES viii 2. LIST OF FIGURES ix 3 . INTRODUCTION 1 4. LITERATURE REVIEW 3 4.1 Corrugated board making 3 4.1.1 Linerboard making process 3 4.1.2 Corrugating medium 6 4.1.3 Corrugated board making process 7 4.2 WOod Chemistry 8 4.2.1 Chemical composition of wood 8 4.2.2 Fibers 8 4.3 Vanillin 10 4.4 Review of mass transport phenomena 12 4.4.1 Permeation 12 4.4.2 Migration 16 4.4.3 Sorption 16 4.4.4 Partitioning 18 4.4.5 Effect of temperature 19 4.4.5.1 Diffusion process 19 4.4.5.2 Sorption process 20 4.4.5.3 Permeability process 21 4.5 Group-contribution method 22 5. MATERIALS AND METHODS 24 5.1 Materials 5.1.1 velatiles 24 5.1.2 Sorbate 24 5.1.3 Film. 25 5.1.4 Corrugated board 25 5.2 Methodology 5.2.1 Determination of the Initial Concentration of O-vanillin and Vanillin in Paperboard 26 5.2.2 Determination of the equilibrium partition distribution (Ks/a) of O-vanillin and Vanillin between corrugated board and headspace 29 5.2.3 Determination of the equilibrium.solubility of O-vanillin and Vanillin in LDPE 32 5.2.4 Determination of the equilibrium partition distribution (KL/A) of o-vanillin and Vanillin between LDPE and aqueous phase 34 6. RESULTS AND DISCUSSION 6.1 Initial concentration of O-vanillin and Vanillin in paperboard 36 6.2 Partition coefficient of O-vanillin and Vanillin between Headspace and Paperboard 39 6.3 Solubility coefficient of O-vanillin and Vanillin into Low Density Polyethylene film. 43 6.4 Partition Coefficient of O-vanillin and Vanillin between LDPE and an Aqueous phase 51 6.5 Predicted solubility parameter values by group- contribution method 53 6.6 Modeling the Equilibrium Distribution of Vanillin between the Paperboard shipper and a LDPE Primary vi 7. 8. 9. 10. Packaging System 54 SUMMARY AND CONCLUSION 65 PROPOSAL FOR FUTURE RESEARCH 69 APPENDIXES 9.1 Appendix A. Standard calibration by Stripping Thermal Desorption method (Part I) 70 9.2 Appendix 8. Standard calibration by Stripping Thermal Desorption method (Part II) 75 9.3 Appendix C. Standard calibration by Gas Chromatography 81 9.4 Appendix D. Saturated Vapor pressure 86 BIBLIOGRAPHY 91' vii LIST OF TABLES Table page 1 Range of O-vanillin levels in corrugated paperboard samples 38 2 Range of Vanillin levels in corrugated paperboard samples 38 3 Partition coefficient (Ks/a) of O-vanillin between headspace and paperboard at 23%: 40 4 Partition coefficient (Ks/a) of Vanillin between headspace and paperboard at 23°C 40 5 Solubility coefficient and diffusion coefficient between O-vanillin and LDPE film at 23 W: 46 6 Solubility coefficient between Vanillin and LDPE film 46 7 Partition coefficient between aqueous solution of o- vanillin and LDPE film 52 8 Partition coefficient between aqueous solution of vanillin and LDPE fihm 52 9 Predicted solubility parameter values 53 10 Prediction of equilibrium.concentration in the paperboard and in an aqueous phase 60 11 Time to steady state for the Diffusion of o-vanillin into LDPE film as a function of ampule wall thickness 63 viii LIST OF FIGURES Figure page 1 Cell wall structure 9 2 Chemical structure of Vanillin and Ortho-vanillin 10 11 12 11 A schematic diagram.of the vanillin generation system 31 Schematic diagram of sorption test system. 33 Partition coefficient of O-vanillin between headspace . and paperboard at 23 °C 41 Partition coefficient of Vanillin between headspace and paperboard at 23 °C 42 Solubility coefficient of O-vanillin into LDPE at 23 W: 47 Sorption curve of O-vanillin into LDPE (o-vanillin driving force: 0.155 Pa) 48 Sorption curve of O-vanillin into LDPE (o-vanillin driving force: 0.301 Pa) 49 Diffusion coefficient of O-vanillin into LDPE at 23 W: 50 VOlatilization from the paperboard through package into contents 55 Diffusion time of o-vanillin vs. wall thickness of LDPE film. 64 O-vanillin standard calibration curve by thermal C-1 C-2 D-l D-2 stripping thermal desorption procedure 73 Vanillin standard calibration curve by thermal stripping thermal desorption procedure 74 O-vanillin standard calibration curve by thermal stripping thermal desorption procedure 79 Vanillin standard calibration curve by thermal stripping thermal desorption procedure 80 O-vanillin standard calibration curve by GC 84 Vanillin standard calibration curve by GC 85 Saturated vapor pressure of O-vanillin 89 Saturated vapor pressure of Vanillin 90 INTRODUCTION In the packaging and distribution of food and pharmaceutical products packaged in polymeric package systems, there is considerable concern regarding the contamination of the contained product as a result of permeation of organic volatiles from the external environment. A potential source of such organic volatiles may be derived from shelf cartons or corrugated shipping containers used in the distribution of the primary package system. For example, organic volatiles present in shelf cartons or corrugated shipping containers may contaminate a product through a process which involves volatilization or evaporation of the organic from.the carton shipper wall, followed by permeation through the primary package which is a polymeric structure. The present study addresses this problem, where the volatilization of vanillin and ortho-vanillin from corrugated board and its subsequent permeation through a low density polyethylene ampoule will be investigated. Vanillin is a decomposition product of the lignin in corrugated paperboard. Ortho-vanillin is also a natural product which is present in corrugated board. Vanillin and ortho-vanillin are proposed as model or probe compounds for this study. 2 A schematic of the mechanism.of vanillin and ortho— vanillin volatilization / permeation and associated equilibria is as follows: Ck) amaiboani shirprgigng container —’. Headspace —" Polyethylene (LDPE): Aqueous Phase The objectives of this study include: 1. Determine equilibrium partition distribution (Kpl) of vanillin and o-vanillin between corrugated board and headspace. 2. Determine solubility of vanillin and o-vanillin in low density polyethylene at the vapor concentration level established from partition distribution studies. 3. Determine equilibrium partition distribution of vanillin and o-vanillin between polyethylene and an aqueous phase. The assumptions of this study include: 1. Volatilization of vanillin and o-vanillin from corrugated board to headspace is inward. 2. Vanillin vapor activity within the corrugated shipper is constant. 3. No headspace in the polyethylene / aqueous phase, package / product system. LITERATURE REVIEW 1 . Corrugated board making A corrugated board is made from.two or more sheets of linerboard and one or more fluted sheets of corrugating medium. In the United States, almost all of the linerboard used to make corrugated board is kraft and a great majority of the corrugating medium.is made by the semi-chemical pulping process ( Fiber box handbook, 1989 ). Three-fourths of all the trees harvested from the forests each year are A used as lumber, plywood and furniture. These processes have a 15 percent waste factor of sawdust and small chips. Previously burned, this material is now used as one of the major sources of fibers for making kraft linerboard ( Fiber box handbook, 1989 ). The other quarter of the annual forest harvest is used directly by the paper industry. The remaining portion of the fiber used in making corrugated board comes from the recycling of paperboard material. The ratio of this portion is presently increasing because of the demand of environmental concerns. 1.1 Linerboard making process In the kraft process, chemicals and heat are used to dissolve the material which holds the wood fibers together to recover the fiber itself. This cementing material, called lignin, must be removed in sufficient amounts to allow clean 3 4 separation of the fibers. Compared with the mechanical process, the fiber strength of the kraft process is stronger, making this kraft linerboard suitable for packaging containers (James, 1982). The kraft linerboard is made from virgin fibers which are about 93% softwood,approximately 0.03 mm in diameter and up to 8 mm long (James, 1982). However, it may contain up to 20% hardwood fibers or secondarily fibers which are cuttings from the corrugating plant or other waste corrugated board. The softwood fiber is longer and larger in diameter than hardwood fiber (James, 1982). This results in the softwood fibers giving the linerboard the necessary mechanical strength. After harvesting, bark is removed from the softwoods and the wood is subjected to the chipping process. The logs need to be reduced to small chips, about 1/2" wide and 3/4" long, to allow the chemical or cooking liquor, for a period of approximately 2 hours, to penetrate the fibers and dissolve the lignin (Norbert 1971). Sodium hydroxide (NaOH) and sodium sulfite are used in the chemical process. The sodium hydroxide contributes to size reduction of lignin molecules, or breaks the lignin down into its basic components. The sodium sulfide contributes to the maintenance of the desired pH level and helps to buffer the reaction of the caustic with the wood, to prevent or reduce damage to the pulp. The sodium sulfide also provides sulfur to react with the lignin building blocks, making them.more soluble. Both the caustic and the sodium sulfide contribute sodium ions, which help in the removal of the lignin reaction products from the wood (James, 1982). After chemical processing, wood pulp is washed to remove the chemical solution by using water and vacuum. The pulp is then refined to get equal fiber size and is separated from foreign materials by using a screen. As a result, clean and fine fibers are obtained through these processes (James, 1982). Paperboard is formed from.the pulp on a flat-wire fourdrinier, in a basis weight range of from.26 to 90 lb/1000 ft2 . It is then dried and calendered through a series of highly polished rolls, where the desired finish is applied by pressure and friction. Steam, water and starch solution is ‘ applied. From the calender stack the sheet passes to reels where it is wound into full width rolls. (James, 1982, Norbert, 1971). During this process, some additives are used to improve water resistance and to get better bonding and so on. Cellulose, which is made from the fiber, is very hydroscopic and will absorb water from the air. There are two major methods of obtaining water resistance, which include internal sizing and surface sizing. The most common method of internal sizing is with rosin and alum. Rosin, a natural organic acid obtained from trees, is emulsified in water and added to the fibers before they are sent to the paper machine. The rosin is slightly anionic and will tend to stick to the fibers. After the rosin is mixed with the fibers, alum is added to the stock. Alum.is a water solution of aluminum sulfate, with some of the aluminum also in the form of aluminum hydroxide. Alum.flocculates with the rosin 6 and with itself, thus creating flocks that adhere to the fibers. The alumprosin flocks are water resistant after drying (James, 1982). Starch is the most common surface sizing material. Since starch is not too water resistant itself, it is believed that it functions primarily by improving bonding at the surface, to plug the surface and retard penetration of water . 1 . 2 Corrugating medium The predominant raw material used in the manufacture of corrugating medium is semichemical hardwood. Hardwood pulp is used rather than softwoods because the hardwood is less - costly and contribute to the particular type of strength needed in corrugating medium. The long fibers of the softwoods are generally considered to make stronger paper, but in this application the long fibers are not needed. The pulp for medium is what would be called a raw cook, meaning that the lignin is not totally removed. Furthermore, the pulp is not washed thoroughly, or in some cases, is hardly washed at all. By not washing or using a strong cook, the lignin and other hemicelluloses are left with the fibers and ‘will be formed into the web of paper. When this web goes through the corrugator, these chemicals help form the rigid fluted shape needed. The short hardwood fibers are less flexible than softwoods and therefore contribute to the stiffness of the fluted structure (James, 1982). The processing starts with the hardwood chips, which sometimes contains some bark when whole-tree chipping is 7 used. These chips are cooked by using neutral sulfite. The corrugated medium.don't use any fillers and sizing agents, unless wet strength agents are added to make special water resistant containers. 1.3 Corrugated board making process Linerboard and corrugated medium.are bonded together by pressure with a starch adhesive (corrugated handbook). The linerboard for the corrugated medium passes through preheated rolls and steam.showers to plasticize it, and make it pliable. It then travels through the fluted rolls of the single facer where the corrugations are formed and adhesive is applied to the tips of the flutes. At the same time, the. first liner ( usually the inner liner) passes though a series of heated rollers and is stuck to the flute tips. This single face board is then laminated with a second liner which is preheated. The bonding between linerboard and corrugated medium is completed under heat and pressure (Paine, 1991) 2 wood Chemistry 2.1 Chemical composition of wood WOod consists of about 50 % carbon , 6 % hydrogen, and 44 % oxygen under dry conditions. The compounds of wood are categorized as cellulose, hemicellulose, lignin, and extractives. Sven (1965) defined the cellulose and lignin terms as follows: Cellulose: the linear polysaccharide, of sufficient chain length to be insoluble in water or dilute alkali and acids at room temperature, containing only anhydroglucose units linked together with 1:4-B-glucosidic bonds and possessing a well-ordered structure. _ Lignin: the aromatic polymer of wood, consisting of four or more substituted phenylpropane monomers per molecule. 2.2 Fibers Fibers for papermaking are mainly composed of cellulose. Cellulose is formed from the elements carbon, hydrogen and oxygen, obtained from the atmosphere by the method known as photosynthesis. The chemical formula is usually written as (C 5 H 10 O 5):: because the cellulose molecule dose not exist as a single unit, but as a number of units. Associated with cellulose in the plant fibers, are other materials known collectively as hemicellulose (Robert, 1963). Fibers have a layer like structure which are covered by and bonded to the middle lamella which is mainly composed of lignin ( the remove of lignin is very important to obtain fibers for papermaking ). A schematic representation of the various layers of fiber wall is shown in Figure 1. (a) Figure 1. Cell wall structure. M middle lamella, P primary wall, 51 transition lamella, 52 main layer of secondary wall, S3 tertiary wall (Sven,1965) The outermost layer of the fiber is called the primary wall (P). The secondary wall, formed during the maturation of the fiber, is not homogeneous but subdivided into three layers such as transition lamella (51), main secondary wall (52), and tertiary wall (S3). The relative thickness of these layers are: P7-14%, 515-11%, 32 74-84%, and S3 3-4% (Sven,1965). The bonding between fibers is associated with hydrogen bonds, which involve the OH groups of the long chain cellulose molecules. It is the number of these bonds formed during drying on the paperboard-making machine that will determine the ultimate overall strength of the paperboard (Robert, 1963). 10 3. Vanillin The use of vanillin as a flavor can be traced back to about 1520 when Cortez, the Spanish conqueror of Mexico, was served a chocolate drink flavored with vanilla, which contains vanillin, by the Aztec Indians. Cortez took the flavor back to Europe which started its spread around the world (Van, 1983). Vanillin occurs in nature as a glucoside, which hydrolyzes to vanillin and sugar. It has been identified in many oils, balsams, resins, and woods. The best known vanillin source is the vanillin plant, a member of the orchid family. The vanillin is extracted from ripe vanillin plant beans with ethanol. Vanillin beans are grown in such countries as Mexico, Malagsy Republic, and Indonesia. However, the majority of vanillin used today comes mainly from lignin, which is a waste product from the manufacture of paper pulp. A smaller amount of vanillin is also made synthetically from.guaiacol (2-methoxyphenol) (Van, 1983). Nowadays in addition to food flavoring, it appears to be useful as a starting material for pharmaceutical compound such as L-dopa, which is used in the treatment of Parkinson's disease (James, 1980). Vanillin is a common name for 3-methoxy-4- hydroxybenzaldehyde. Another type of vanillin is referred to as ortho—vanillin or o-vanillin, which is the name for 2- methxy-4-hydroxybenzaldehyde (Van ness,1983). The structure of both vanillin isomers are shown in Figure 2. 11 CH0 CH0 C}1 OCH3 OH OCH3 vanillin onho-vanillin Figure 2. Chemical structure of vanillin and ortho- vanillin Vanillin is a low-molecular-weight oxidative degradation product from lignin, which is a component of wood. Efficient oxidation and recovery processes result in vanillin yields in excess of 10% on lignin (James, 1980). It has been reported. that lignin yield vanillin after oxidizing with nitrobenzene (Brauns,1960). During the pulp making process, lignin is removed by alkali solution to obtain fiber itself. However, there is some residue in the pulp after the pulp making process (James, 1980). O-vanillin is a byproduct from the manufacture of vanillin from guaiacol (S.Coffey,1976). Guaiacol is present in wood tars from which it can be obtained by distribution. Lignin and its derivatives and guaiacum resin are other sources of guaiacol and its alkali derivatives (Coffey,1971). 12 4. Review of mass transport phenomena 4.1 Permeation Permeability is defined as the transmission of gasses or vapors through a resisting material with no macroscopic pores (Paine,1983). The permeant molecule has to undergo the following processes in succession: (1) absorption of the penetrant molecule into the surface of the polymer, (2) diffusion of the penetrant molecule through the polymer, and (3) desorption of the penetrant molecule from the other side of the film. Permeation of polymers, in general, is a function of two variables, one is related to permeant diffusion between polymermolecular chains and the other to the solubility of the permeant in the polymer (Wangwiwatsilp, 1993). The rate of permeation of the substance through the material is usually expressed by the permeability constant, which is dependent on temperature and is described by the relationship: P =ID-S (1) where: P: permeability constant diffusion coefficient (DU solubility coefficient Diffusion theory states that the penetrant molecule transfers across a polymer membrane as a result of a driving force based on the partial pressure or concentration gradient between the high concentration and low concentration environments. The "holes" or free volume accessible to the 13 penetrant molecule, and which allow penetrant diffusion, reside primarily within the amorphous regions of the polymer bulk phase. Therefore the size and frequency of “hole" formation and the size of the penetrant molecular is a critical element for diffusion. The amplitude and motion of the space or voids between polymer molecules are directly related to the temperature, chemical composition, and morphology of the polymer (Rogers, 1964). The diffusion process is described by Fick's first and second laws of diffusion and solubility is described by Henry's law. (Crank,1975). Eq.2 is known as Fick's first law and states that the rate of diffusion is directly proportional to the concentration gradient between the two contact phases. The negative diffusion coefficient indicates that the concentration of the penetrant within the polymer bulk phase decreases with distance from the surface. dc J =-D a (2) where: J: flux, rate of transmission per unit area of permeant through the polymer D: diffusion coefficient c: concentration /unit volume x: direction of diffusion Eq.3 is known as Fick's second law and applied under circumstances where the diffusion coefficient is constant and independent of concentration (Crosby,1981). d)(2 (3) where: At room temperature and atmospheric pressure, diffusion coefficients for gases or vapors lie in the range of 0.1 to 1 cmR/s. In liquids of about 1 centipoise viscosity, diffusivities are of the order 10" to 10-5 cmZ/s (Sybil,1993). Diffusion coefficients in a solid phase such as a polymer are much lower and range from 10'11 cmzlsec to 10-14 cmz/sec (Crosby,1981). Where the permeant penetrates into the polymer, a boundary layer is created. The boundary phase of the polymer is, in effect, swollen and the diffusion coefficient in this phase is not equivalent to that in the unswollen phase. The diffusion coefficient in the swollen polymer is larger than that in unswollen phase. In the swollen phase, Fick's laws of diffusion do not apply because the boundary phase is changing with time (Crosby,1981). At low vapor pressure levels, the vapor pressure of the penetrant is directly proportional to its concentration in the polymer. This is described by Henry's law (Hernandez,1986) and is shown by: C=S'p (4) where: Cu concentration of the penetrant in the polymer phase S: solubility coefficient 15 p: partial pressure of the penetrant in the gas phase At steady state conditions, the rate of permeation (J) is described as the amount of penetrant passing (Q), during unit time (t), through unit surface area (A) and is shown by: Q J = ——— A" (5) The rate of permeation is also described by Eq.2 at the steady state = [no—c.) J I (5) where: Ch C2 : concentration of gas or vapor at the surface x=o, x=l , respectively ( C1>C2) l : thickness of the polymer The amount of penetrant passing (Q) per some unit time is derived from Eq.5 and 6 = D(Ct-C2)'A't o I (7) The permeability constant (P) is derived from.Eq.1, 4 and 7 = Q-l A°t(pt-p2) (8) There are four assumptions made in describing P by the above expression (Robertson, 1992) (1) D and S are independent of the concentration (2) D is independent of the thickness (3) No interaction among gases and vapors (4) Ideal gas law is applicable for vapors 16 4.2 Migration Migration is the term used to describe the transfer of substances from the package into a contacting phase. Substances that transfer to the food as a result of contact or interaction between the food and the packaging material, are often referred to as migrants (Crosby, 1981). While the process of transfer from the package into food is complex, diffusion is considered to be the main controlling mechanism. 4.3 Sorption Sorption is a phenomena describing the dissolution or transfer of a substance from the contacting phase to a polymeric packaging material. Sorption measurements are usually carried out at equilibrium vapor pressure by using a gravimetric technique. The electrobalance is commonly used to record continually the gain of weight by a test sample. Early techniques involved placing a specimen in an organic vapor atmosphere and then removing it at intervals to measure its weight gain (Baner,1987). The solubility coefficient can be calculated from Eq.(9) (Hernandez,1986). S==|M. W—'5 (9) where: S : the solubility coefficient (kg/kg-Pa or kg/m3-Pa) M“: the total amount of vapor absorbed by the polymer at equilibrium for a given temperature (kg) w': the weight (or volume) of the polymer sample under 17 test (kg or m?) b : the penetrant driving force in units of concentration or pressure (ppm or Pa) The diffusion equation appropriate for the-sorption of the penetrant by a polymer film was described by Crank (1975) as: 'Ililt = ’Rgiiex" (ZDTZfl%%°Xp(-golénz.t)] (10) where: Mt and M” = the amount of penetrant sorbed by the polymer film.sample at time (t) and the equilibrium sorption level after infinite time, respectively t : the time to attain Mt l : the thickness of the film sample The sorption diffusion coefficient (EL) can be calculated from Eq.(10) when Mt / M, is equal to 0.5 o. = 0.049:2 Rm (11) where: ‘Ris = the time required to attain the value, Mt / Mno = 0.5. 18 4.4 Partitioning When mass transfer equilibrium.and thermal equilibrium between two immersible phases have been achieved, the equilibrium.distribution or partition distribution of a solute between the two phases can be detenmined (Grant et al.,1990). The equilibrium.distribution of a solute between two solvents at constant temperature is a constant described by Nernst's law (Shinoda,1978). The partition coefficient (Kp) is defined as the ratio of a solute's concentration in the polymer phase (Cp) to its concentration in the contact phase (CL), in Eq.(12) 0p Kp=— CI. (12). For the linear model, the equilibrium.concentration of the penetrant in the packaging material, resulting from contact with the contact phase, is directly proportional to the initial penetrant concentration and can be estimated by solution of equation (13) (Imai, 1988) Cap _ CIL ’ we !!5+VVP KP (13) where : cgp: concentration of penetrant in polymer phase at equilibrium (Hr: initial concentration of penetrant in the contact phase 19 fig and W5: weight of contact phase and polymer, respectively Kp‘ partition coefficient 4.5 Effect of temperature 4.5.1 Diffusion process In general, as the temperature increases, the diffusibility of a gas or vapor in the polymer increases. This is related to morphological changes within the polymer. Above the glass transition temperature (Tg), polymer chains experience segmental mObility. In addition, the polymer molecule experiences torsional oscillation and/or rotation around covalent single bonds, which is called micro-brownian' motion. As the temperature increases, the size and frequency of “hole" formation increase, as a result of increased segmental mobility. Thus, creating a more accessible pathway for permeant molecules to diffuse through the polymer bulk phase. The temperature dependence of the diffusion coefficient, within small temperature ranges and at a constant vapor concentration, is described by an Arrhenius relation: D = Do exp -(Ea / RT) (14) where E3 is the activation energy of diffusion, which will vary from one system to another (Crosby,1981); the pre- exponential factor, (Do), is thought of as being related to the frequency and magnitude of the holes in the polymer. R is the gas constant; and T is absolute temperature. 20 4.5.2 Sorption process In general, as the temperature increases, the solubility of a gas or vapor in the polymer decreases. The temperature dependence of the solubility coefficients can also be described by an Arrhenius type equation: S = so exp —(Hs / RT) (15) where H, is heat consumed by dissolving a mole of penetrant in the polymer; So is the pre-exponential constant. H; can be described as the sum.of the molar heat of condensation(Hc) and the partial molar heat of mixing (H1). The heat of mixing is indicated as a positive number and the heat of condensation can be positive or negative, depending on whether the sorbate molecule is a gas or vapor. For permanent gases, H, is slightly positive, therefore the solubility coefficient of the gas increases slightly with increasing temperature. However, for the more condensable vapors such as an organic compound, H, is negative, due to the relatively large heat of condensation. Therefore, the solubility coefficient of a vapor decreases with increasing temperature (Rogers,1964). 21 4.5.3.Permeability process The temperature dependence of the permeability constant within a small temperature range and at a constant vapor concentration is derived from Eq.1,14 and 15: p = b- s = D. exp —(3. / RT) so exp -(Hs / RT) = Do So exp (—Ed - Hs)/ RT) (16) where: P0 = Do. So E9 = E. + as Values of ED are in the range of 0~20 kcal/mole for the commonly used plastics. 22 5 . Group-contribution method There are numerous chemical compounds of potencial interest. However, the number of structural and functional groups making up these compounds is much smaller. Group- contribution methods assume that a physical property of a fluid is the sum of contributions made by the structural and functional groups in the molecule. By using this assumption, it is possible to develop correlation techniques for a large number of fluids using a much smaller number of parameters which characterize the contributions of the individual groups (Fredenslund et a1., 1975). Any group-contribution method is necessarily an approximation because the contribution of a given group in one molecule is not necessarily the same as that in another molecule (Fredenslund et a1., 1975). The fundamental assumption of a group-contribution method is additivity (in series), i.e. the contribution made by one group is assumed to be independent of that made by another group. This assumption is valid only when the influence of any group in a molecule or in a structural unit of a polymer is not affected by the nature of the other groups (Fredenslund et al, 1975). According to solubility theory developed by Hildenbrand, the enthalpy of mixing is given by: AH". = AME-6p) (18) where: AHm : enthalpy of mixing per unit volume ¢,land ¢b : volume fraction of compounds (i.e. solvent and polymer) 23 5, and 6p : solubility parameters of the components According to Hildenbrand, AHm has to be equal to or near zero for solubility to take place. Equation 18 indicates that AHn =0, if 65 = 6p. This is equivalent to saying that if the polymer and solvent have equal solubility parameters they should be matually soluble or, that substances with similar chemical structures are quite likely to dissolve in each other. MATERIALS AND NETNODOLOGY MATERIALS 1. Volatiles In this study, vanillin and ortho-vanillin were evaluated as volatiles derived from corrugated paperboard. Research grade vanillin and ortho-vanillin were purchased from.Aldrich Chemical Co. Inc.(Milwaukee, WI). Selected properties are shown below. Vanillin Molecular weight 152.15 Melting temp. 81~83 °C Boiling temp. 170 °C (at 15 mmHg) Purity 99 % Ortho-vanillin Molecular weight 152.15 Melting temp. 40~42 °C Boiling temp. 265-266 °C (at 760 mmHg) Purity 99 % 2. Sorbate Chromatography grade acetonitrile (Aldrich Chemical Co. Inc.,Milwaukee, WI) was used to prepare calibration curves for gas chromatography analysis. HPLC grade water (Aldrich Chemical Co. Inc.,Milwaukee, WI) was used to determine the partition coefficients for the 24 25 LDPE / aqueous phase / vanillin systems. 3 . Film A.50 pm.LDPE film.was used in this study. Films were made by a compression molding procedure, using a Carver laboratory press (Carver Inc. Menomone Falls, Wis.). LDPE resin pellets, obtained from Dey Laboratories, Inc (Napa, CA), were placed between polyethylene terephthalate (PET) film.and the platens of the press were heated to 150 °C and maintained at this temperature. Those platens were pressed with 30,000 pounds pressure for 5 minutes. 4 . Corrugated board Kraft corrugated board with the specification of D/F 200 lb. and C flute (Midwest Papermart, Lansing, MI). 26 METHODOLOGY 1. Determination of the Initial Concentration of Vanillin and o-vanillin in Paperboard The initial concentration of the vanillins was determined by two different procedures namely: a thermal stripping/thermal desorption technique and a chloroform extraction method. 1.1 Dynamic Thermal Stripper-Thermal Desorption (TS/TD) Method In the thermal stripping/thermal desorption method, two . small pieces of paperboard, 3x60 mm in dimension, for a total surface area of 3.6 cm?*were put into the sparging tube (Supelco Co., Bellefronte, PA) of the thermal stripper instrument (Model 1000, Dynatherm.Analytical Instrument, Inc. Kelton, PA). The sparging tubes were then mounted in the oven of the thermal stripper instrument (Model 1000, Dynatherm Analytical Instruments, Inc., Kelton, PA) and connected to sorption tubes positioned outside the oven. The sorption tubes, containing Carbotrap 300 multi-bed materials (Supelco Co., Bellefronte, PA) were covered with sleeve heaters. During the thermal stripping procedure, the oven temperature was maintained at 80 °C, the block temperature was held at 105 °C and the sorption tubes were maintained at 50 W3. The samples were preheated for two minutes, and then purged for 30 minutes with a helium stream to drive the sorbed volatiles from the paperboard to the sorption tubes. 27 After purging the samples, a two minutes drying time was selected to remove any residual water vapor from the sorbant tubes. The carrier gas pressure was maintained at 40 psi and the flow rate through the sample and heated block was set at 100 ml/min. during purging, and 50 ml/min. through the heated block during drying step. After the sample preparation, the Carbotrap sorption tubes containing the trapped volatiles were transferred to the tube chamber of the thermal desorption unit (Model 890, Dynatherm Analytical Instrument, Inc. Kelton, PA). The sorbed volatiles were then desorbed by heating for 8 minutes at 340 °C, with the valve and the transfer line held at 230 °C to maintain the desorbed compounds in the vapor phase, while being transferred to the gas chromatograph. Helium.was used as carrier gas through the thermal desorption unit at a flow rate of 7.0 ml/ min., at 40 psi. After sample desorption, the sorbant tubes were conditioned at 340 °C for 10 minutes prior to reuse. The flow rate of helium through the sorbant tube to the side port was 10 ml/min., at 40 psi. Gas chromatographic analysis was carried out with a Hewlett-Packard 5890A Gas chromatograph, equipped with a dual flame ionization detector (Avondale, PA). The GC conditions were as follows: Column, fused silica capillary column (30 m x 0.32 mm ID) intermediate polarity stationary phase SPB-20 (Supelco Co., Bellefronte, PA), helium.carrier at 7.0 ml/min., H5 at 40 ml/min., air at 400 ml/min., makeup gas (N;) at 30 ml/min., column oven initially held at 40 °C for 5 minutes and then increased t01150 °C at 7.5 °C/min.,and detector at 250 °C, to give a retention time of 13.9 min. for o-vanillin and 15.8 min. for vanillin. The 28 procedure for constructing a calibration curve of response vs. vanillin and o-vanillin concentration is described in detail in Appendix A. The respective calibration curves are also presented in Appendix A. Another procedure for constructing a calibration curve is described and also presented in in Appendix B. More linear relationship between response vs. o-vanillin concentration is obtained by the procedure in Appendix A than by that in Appendix B; therefore the procedure in Appendix A is used in this study. 1.2 Solvent Extraction Method In the chloroform.extractive method, vanillin and o- vanillin in the paperboard were extracted with chloroform, as outlined in the Code of Federal Regulation (21CFR 176.170), except that chloroform was substituted for heptane as the extraction solvent due to vanillin's higher solubility in chloroform. The solution was concentrated by Rotavapor (Buchi laboratoriums-technik) to a volume of 10 ml and then analyzed by gas chromatograph (HP5890A, Hewlett Packard, Avondale, PA). The GC conditions were as follows: Column, fused silica capillary column (60m x 0.32 mm ID) intermediate polarity stationary phase Supelcowaxm (Supelco.Co. , Bellefronte, PA), helium.carrier at 1.7 ml/min., 85 at 40 ml/min., air at 400 ml/min., makeup gas (N2) at 30 ml/min., column oven initially held at 60 °C for 1 minutes and then increased to 240 °C at 5 °C/min.,and detector at 250 °C, to give a retention time of 35.4 min. for o-vanillin and 44.4 min. for vanillin. The calibration curve and procedure are presented in Appendix C. 29 2. Determination of the equilibrium partition distribution (Rpl) of Vanillin and o-vanillin between corrugated board and headspace. The procedure to determine the partition coefficient for the corrugated board / headspace / o—vanillin system.was as follows: 14 Strips of paperboard samples, 1x6.5 cm in dimension, for a total surface area of 91.0 cm2 were cut from paperboard which had been exposed to o-vanillin vapor to provide corrugated board samples of known migrant concentration levels. The surface to volume ratio is the same as found with the designed shipping carton. The strips of paperboard were placed into 150 ml crown cap amber glass vials (Aldrich Chemical Co. Inc., Milwaukee, WI) with teflon faced septa. The vials were stored at 23°C for 48 hours to allow equilibration and then sampled. The equilibrium concentration of o-vanillin in the headspace was determined by removing a 400 pl sample with a gas-tight syringe (Hamilton Co., Gastight #1750) through the septa and analysis by gas chromatography, as described in Appendix C. O-vanillin in the paperboard was determined by taking paperboard samples of approximately 0.3 grams and analyzing for o-vanillin by the thermal stripper thermal desorption method. .A schematic diagram of the o-vanillin generation system is shown in Figure 3. The procedure to determine the partition coefficient for the corrugated board / headspace / vanillin system was as follows: 30 42 Strips of paperboard samples, 1x6.5 cm in dimension, for a total surface area of 273.0 cm2 were cut from paperboard which had been exposed to vanillin vapor, to provide corrugated board samples of known vanillin concentration levels. The surface to volume ratio is the same as found with the designed shipping carton. The strips of paperboard were placed into uncoated two-piece steel cans, 300 x 406 of diameter and height, respectively (American National Can, La Porte, IN). Prior to sealing a small hole (1 mm) was punched in the can lid to provide a sampling port, which was covered (interior surface) with teflon pressure sensitive tape (Supelco Co., Bellefronte, PA). A self sealing silicone rubber septum was affixed to the outer surface of the sampling port (Mocon, Minneapolis, MN). The cans were seamed with a hand seamer (Model 23H, Dixie Canner Equipment Company, Athens, GA). The sorption cells were stored at 23°C for 14 days to allow system equilibration and then sampled. The equilibrium concentration of vanillin in the headspace was determined by removing 400 pl sample from the headspace with a gas-tight syringe through the sampling port and analysis by gas chromatography, as described in Appendix C. Vanillin in the paperboard was determined by taking paperboard samples of approximately 0.3 grams and analysis by the thermal stripper thermal desorption method. 31 Glass tube 39—. Stainless containel / To hood }———AD> Glass fiber \ Nitrogen gas O-vanillin Corrugated paperboarc Figure 3. A schematic diagram of the vanillin generation system 32 3. Determination of the equilibrium solubility of vanillin and o-vanillin in LDPE Solubility measurements for o-vanillin was Carried out on a Cahn 2000 Electrobalance by the continuous flow method (Cahn Instruments Inc., Cerritos, CA). The electrobalance and sample film.were maintained at a constant temperature of 2311 W3. .A schematic diagram.of the test apparatus is shown in Figure 4. A constant concentration of o-vanillin vapor was produced by continuously flowing nitrogen gas through a bed of the solid sorbate. This was achieved by assembling a vapor generator system.consisting of a glass tube containing the o-vanillin, which was fitted with gas inlet and outlet ports. In order to obtain a range of vapor concentration levels, the sorbate vapor stream.was mixed with another stream.of pure carrier gas (nitrogen). Flow meters (Cole Parmer, IL) and needle valves (Nupro"M"series, Nupro Co.,OH) were used to adjust the flows and to indicate constant flow rates. The weight gain of the polymer sample was continuously recorded on a Linseis Model L6512 flat-bed recorder (Linseis, Inc., NJ). Experiments were allowed to continue until steady-state was reached. The composition of the vapor stream was determined by taking samples with a gas- tight syringe through the sampling port of the hangdown tube and analysis by gas chromatography, as described in Appendix C. 33 Flow meter Sampling pori Vapor generator Figure 4 Schematic Diagram of Sorption Test System To determine the equilibrium solubility of vanillin in the LDPE film samples, the following procedure was employed. Approximately 3 grams of vanillin were placed in the bottom of a 500 ml jar which was covered with aluminum foil. The sample films were suspended from the cap of the jar. The sealed jars were stored at 23 and 32°C, respectively for 10 days and sampled. The equilibrium concentration of vanillin in the headspace was determined by taking samples with a gas-tight syringe through the sampling port affixed to the cell lid and analysis by the gas chromatography procedure, as described in Appendix C. Vanillin levels in the sample films were determined by the TS/TD procedure. 34 4. Determination of the equilibrium partition distribution (Kp2) of Vanillin and O-vanillin between LDPE and aqueous phase. The procedure to determine the partition coefficient for the LDPE / aqueous phase / vanillin systems was as follows: 14 Round polymer disks, 23 mm diameter, for a total surface area of 8.3 cm2 were cut from LDPE film by a punch. The surface to volume ratio was the same as found with the designed LDPE formsfill seal ampoule. The films of approximately 50 um.thickness were made from the LDPE ampoule resin by a compression molding procedure. The polymer disks were weighed and threaded onto a stainless steel wire frame with 2 mm high and 4 mm diameter glass rings used to separate the polymer disks. The wire frames with the mounted disks were then placed into 40 ml screw cap amber glass vials (28 mm diameter, 98 mm high) with teflon coated silicone septa (Supelco,PA) and 35 ml of the respective sorbate solutions were added to the vials. Concentrations of 1 ppm (pg/ma, wt/v) of an aqueous solution of o-vanillin and 0.1 and 1 ppm (pg/ml, wt/v) of an aqueous solution of vanillin were evaluated in the sorption studies. The sorbate solutions were prepared by dissolution of the individual sorbates in HPLC grade water. The sorption vials were stored at 23,32, and 40 °C for o—vanillin samples, at 23 °C for vanillin samples, and sampled after 30 days to insure equilibrium sorption conditions. Aliquots of the aqueous solution were removed from the sorption vials and the equilibrium concentration of o- 35 vanillin determined by an HPLC procedure. HPLC analysis were carried out by the SRI International Life Science Division (MenloPark, CA). The levels of o-vanillin sorbed by the LDPE film.were determined by difference. For the vanillin sorption samples, polymer samples were removed from the aqueous phase were taken at the end of the storage time. The polymer disks were then rinsed two times with aliquots of HPLC grade water and the disks blotted dry with a lab tissue and placed in 40 ml crimp cap vials. The sorbed vanillin in the polymer was extracted by adding 15 ml of acetonitrile, capping the vials with teflon coated silicon septa and placing them in 35 °C oven for 30 days. The vials were shaken by hand twice during extraction and after 30 days - removed from the oven. Samples of the acetonitrile were concentrated by Rotavapor (Buchi laboratoriums-technik) to a volume of approximately 0.2 ml and then analyzed by gas chromatography (HP5890A, Hewlett Packard, Avondale, PA). The GC conditions were the same as the methodology 1. The calibration curve and procedure are presented in Appendix C. The level of sobed vanillin was determined by substitution into the following equation. . . _ _ 1 concentrated volume (ml) Sorbed mm" ("9) - 3'“ ""“s " calm ia—cfor(area units—7T9) " injected volume (ml) (19) RESULTS AND DISCUSSION 1. Initial Concentration of O-vanillin and Vanillin in. Paperboard Initial concentration levels of o-vanillin and vanillin determined in corrugated paperboard samples are summarized in Tables 1 and 2. From the data of o-vanillin, the average concentration levels in the paperboard showed a range which was dependent upon the location of the board within a stack. The number 1 paperboard sample, which showed the lowest concentration level, was located on the top of the paperboard stack on the pallet shipper. Whereas, the second and third paperboard samples which showed relatively higher o-vanillin levels were taken from the middle of the stack. It is assumed that o-vanillin in the top paperboard test (sample no 1) sample was volatilized into the surrounding air during the storage and resulted in the lower initial concentration. The corrugated paperboard samples taken from the center of the stack (samples no 2 and 3) would be expected to experience minimal losses of o-vanillin and would represent the worst case or upper bound limit for residual o-vanillin in paperboard. The measurements of o-vanillin were carried out by two different procedures, namely, the thermal stripping thermal desorption technique and the chloroform extraction method. To determine if an oven temperature of 80°C and a purge time 36 37 of 30 min would provide quantitative desorption of residual vanillin and o-vanillin from the corrugated paperboard samples, the thermal stripping/ thermal desorption procedure was repeated with the same samples. A comparison of the chromatograms obtained from the first and second analysis showed no detectable levels of vanillin and o-vanillin present where the thermal stripping/ thermal desorption procedure was repeated. It was therefore assumed that with an oven temperature of 80°C and a purge time of 30 min, greater than 90 % of the residual vanillins were desorbed from the corrugated board during the initial heating period. As shown it Table 1, good agreement was attained between these two procedures and both methods are suitable for measurement of residual vanillin and o-vanillin in corrugated paperboard. The initial concentration levels of o-vanillin and vanillin in the paperboard, which was located in the middle of the stack on the pallet were 0.34 and 0.9 (pg/g paperboard) , respectively . Table 1 38 Range of o-vanillin levels in corrugated paperboard samples Sample O-vanillin concentration Procedure NO. Number of Average S.D. replicates (pg/g) t l (c) 3 0.05 0.01 (a) 2 (d) 3 0.34 0.02 (a) 3 (d) 4 0.24 0.02 (a) 4 0.27 0.04 (b), (a): Thermal desorption procedure (b): Solvent extraction procedure (c): paperboard sample from top of stack (d): paperboard sample from center of stack Table 2 paperboard samples Range of vanillin levels in corrugated Sample Vanillin concentration Procedure No. Number of Average S.D. replicates (pg/g) t 1 (b) 3 0.9 0.03 (a) (a): (b) Thermal desorption procedure paperboard sample from center of stack 39 2. Partition Coefficient of O-vanillin and Vanillin between Headspace and Paperboard Partitioning studies of o-vanillin between headspace and paperboard were carried out over a range of initial 0- vanillin concentrations between 0.15 and 0.31 (pg/g), at 23°C. The results of partitioning studies are summarized in Table 3. The average partition coefficient value was 21.6, where the partition coefficient (Ram) is defined as the equilibrium.concentration of o-vanillin in paperboard divided by the equilibrium.concentration in headspace. The relationship between the partition coefficient and initial 0- vanillin concentration is illustrated graphically in Figure 5, where the partition coefficient (Ram) is plotted as a function of initial o-vanillin concentration. Partitioning studies of vanillin were carried out at initial vanillin concentration levels of 1.3 and 2.9 (pg/g), at 23°C. The results of partitioning studies are summarized in Table 4. The average partition coefficient was found to be 77.5. The relationship between the partition coefficient and initial vanillin concentration is illustrated graphically in Figure 6, where the partition coefficient (Ram) is plotted as a function of initial vanillin concentration. 40 Table 3 : Partition coefficient (Ks/n) of o-vanillin between headspace and paperboard at 23°C Paperboard] lrl'tial 0 -vanillin Storage Final o-vanillin conc. Material Partition weight (g) Conc. Total tine In headspace In paperboard balance coefficient (hf) (pg/ml x 103) (119/9) (°/°I (Kn/a) _24—'—'675'— 0.22 122.7 33.2 5.6 0.24 1.3 24 6.6 0.10 110.2 15.1 5.4 0.23 1.2 48 5.9 0.15 130.2 25.3 5.6 0.20 1.1 24 4.7 0.10 106.3 20.3 5.6 0.15 0.8 48 4.0 0.06 104.8 14.3 Average 212.5. S . D . P32 Table 4 : Partition coefficient (Ks/n) of vanillin between headspace and paperboard at 23°C Paperboard Initia vanillin Storage Find vanillin conc. Material Partition weight (g) Conc. 1 Total time in headspace In paperboard balance coefficient (119/9) ‘ (.09) (daY) (pg/mlx103) (119/9) (96) (Kn/n) 15.9 2.9 46.1 7 24.2 2.0 89.5 82.1 15.9 2.9 46.1 14 23.8 2.2 96.5 92.4 15.9 1.3 20.7 14 13.4 0.8 87.5 59.7 15.9 1.3 20.7 14 11.9 0.9 92.2 75.7 Average 8 ET 41 y=ZL6 100 I 75 " x 50 I CI 25 -‘ D ——-o n D 0 l r I 0 0.1 0.2 0.3 0.4 Initial o-vanillin concentration in paperboard (pg/g) Figure 5. Partition coefficient of o-vanlliln between headspace and paperboard at 23°C 42 y=77.5x 150 100 ‘1 U D 5 TI X U i) —i 0 I I I I - 0 l 2 3 4 5 Initial vanillin concentration in paperboard ([1919) Figure 6. Partition coefficient of vanillin between headspace and paperboard at 23°C 43 3. Solubility coefficient of o-vanillin and Vanillin into a Low Density Polyethylene film. Sorption studies of o-vanillin by the electrobalance procedure were carried out over a range of partial pressure values between 0.155 and 0.317 Pa, at 23 W3. Solubility coefficient values determined by substitution into Equation 9 are summarized in the Table 5. The results of sorption studies indicate that the solubility coefficient was independent of o-vanillin concentration over the range of vapor pressures evaluated. The mean solubility coefficient value determined was 11 . 3 Kg/m3 Pa, with a standard deviation of i 1.5. The relationship between the solubility coefficient and the o-vanillin concentration is presented graphically in Figure 7. As shown, a linear plot was attained, indicating that the sorption process follows Henry's Law, over the vapor pressure range studied. The half-time diffusion coefficient (D,) values, calculated from the sorption curve for each sorption run by substitution into Equation 11, are also summarized in Table 5. The relationship between the diffusion coefficient (D,) and o-vanillin vapor pressure is presented graphically in Figure 10. As shown, the D, values exhibited no trend as a function of o-vanillin vapor pressure, but rather a random variation, with D, values ranging from 6.2 to 11.5 a? /sec x 10'”, with a mean value of 8.5 :i: 2.1 m2 /sec x 10‘“. The half-time diffusion coefficient (D3) is an integral diffusion coefficient, which is derived from the transient state portion of the sorption curve. Because of the steepness of 44 the non-steady state portion of the curve, toJ values cannot be determined with the same degree of accuracy as the weight uptake values obtained at steady state. This can account for the variation in the DI values observed, and for the good agreement obtained between the solubility coefficient values. From the definition of the integral diffusion coefficient, the integral diffusion coefficient represents an average of the diffusion coefficients throughout the polymer thickness. Thus, the integral diffusion coefficients are assumed to be independent of time effects and are defined for the given boundary conditions of the sorption experiment, even though the concentration distribution of sorbate in the polymer is not explicitly known. The integral diffusion coefficient is Fickian in nature, meaning that the diffusion coefficient is independent of time effects. A representative plot of Mt/M, vs t1/2 for sorption of o- vanillin in the LDPE test film, obtained at an o-vanillin vapor pressure of 0.155 Pa, is shown in Figure 8. A plot similar to that in Figure 8 is presented in Figure 9, for an o-vanillin vapor pressure of 0.301 Pa. It can be seen that the initial portion of the curve is approximated by a straight line. Therefore, at these low partial pressure or concentration levels, mass transport is assumed to follow Fickian type behavior. Sorption studies of vanillin into LDPE film.were carried out at temperatures of 23 and 32 °C, respectively. Solubility coefficient values determined by substitution into Equation 9, are summarized in Table 6. The mean solubility 45 coefficient value at 23 °C was 0.3 t 0.08 Kg/m3 Pa and 0.23 1: 0.05 Kg/m3 Pa at 32 °C. Table 5 : 46 Solubility coefficient and diffusion coefficient of o-vanlllln between headspace and LDPE film at 23°C o-vanilli driving force (Pa) n solubility coefficient (Kg/m3xPa) diffusion coefficient (m2/sec x1013) 0.155 10.6 11.5 0.187 10.9 10.5 0.207 9.4 6.7 0.301 13.8 7.4 0.317 12.0 6.2 Ave. : 11.8 Ave. :8.5 S.D. :1.5 S.D. 22.1 Table 6 : Solubility coefficient of vanillin between headspace and LDPE film storage vanillin solubility coefficient temperature driving force (Pa) (Kg/m3xPa) (0C) no. of ave. S.D. ave. S.D. replicates 23 4 0.142 0.01 0.30 0.08 32 4 0.422 0.05 0.23 0.05 47 y=11.8x 20 76‘ D. so 15" E n 3, c: ‘x" D .- u E 1° 0 .2 E a 5‘ 27:: 2 ‘3 0 l l l O 0.1 0.2 0.3 0.4 Driving force (Pa) Figure 7. Solubility coefficient of o-vanillln into LDPE at 23°C 48 1.25 0.75 -i Nit/Mo, I I I I F I 0100200300400500600700800 110 . Time"2 (see) Figure 8. Sorption curve of o-vanlllin Into LDPE (o-vanlllln driving force: 0.155 Pa) 49 1.25 0.75 ‘ Mthoo 0.5-re . . .. . I! 0L I 'I I I I I I 0 100200300400500600700800 170 Time"2 (sec) Figure 9. Sorption curve of o-vanlllln Into LDPE (o-vanillln driving force: 0.301Pa) 50 y=8.5x 20 5'27 b 1- 3 N .5, I E El ,_ 10 ~ c .9 .9 D o 5 ‘ c .9 (I) 3 E D 0 I I I 0 0.1 0.2 0.3 0.4 Driving force (Pa) Figure 10. Diffusion coefficient of o-vanlllin Into LDPE at 23°C 51 4. Partition Coefficient of O-vanillin and Vanillin between LDPE and Aqueous phase. Partitioning studies of O-vanillin and vanillin between LDPE and an aqueous phase were carried out at initial concentrations of O-vanillin in water of 1 ppm (wt/v) at storage temperatures of 23, 32, and 40 °C, and at initial concentrations of vanillin in water of 0.1 and 1 ppm.(wt/v) at storage temperatures of 23 °C. All samples were equilibrated for storage times of 30 days. The results of partitioning studies are summarized in Tables 7 and 8. The average partition coefficient value was 574 for o-vanillin and 3.8 for vanillin at 23 °C, where the partition coefficient (Rum) is defined as the equilibrium concentration- of sorbate in the LDPE phase divided by the equilibrium concentration in the aqueous phase. 52 Table 7 : Partition coefficient between aqueous solution of o- vanlllln and LDPE film at 30 days storage * concentration of storage | partition coefficient (KL/A) I o-vanillin solution temperature (ppm) (°C) Table 8 : Partition coefficient between aqueous solution of vanillin and LDPE film at 30 days storage concentration of storage partition coefficient (KL/A) ] vanillin solution temperature (ppm) (°C) 53 5. Predicted solubility parameter values by group- contribution method The component group-contributions method of Van Krevelen (1990) was used to estimate the solubility parameter for the polymer and vanillin. The solubility parameter (6) values of vanillin, LDPE and water are shown in Table 27. This would predict that vanillin has a higher solubility in water than in LDPE. Table 9: Predicted solubility parameter values Solubility parameter (6 ) vanillin 23.3 LDPE 16.9 water 19.2 units = (J/cm3)1/2 54 6. Modeling the Equilibrium Distribution of Vanillin between the Paperboard shipper and a LDPE Primary Packaging System As illustrated in Figure 11, for the packaging situation described, volatiles in the paperboard migrate into the headspace of the corrugated shipper and then permeate through the LDPE container into the aqueous phase. VOlatiles derived from the corrugated shipper are assumed to equilibrate between the corrugated paperboard/headspace/LDPE/and the contained aqueous phase. In describing such a system, The following assumptions are proposed. 6.1. Assumption (1) Volatilization of vanillin from corrugated board to headspace is inward. (2) The aqueous phase does not react with volatiles (3) The initial concentration of the volatile in the system is assumed by: cin = on = on = 0 This means that volatile is initially present only in the paperboard. (4) At equilibrium, the mass balance of the volatile in the system is described by: Wa-Ca=Wa-C.9+VH-C.H +VL-C.L +VA'CQA (20) 55 A ueous Paperboard Headspace LDPE ptIase The CIH CiL _CiA WB VH VL VA . _J KBkI S . Kan Ce CeH CeL CeA Figure 11. VOlatilization from the paperboard through package into contents. C1, C“, C“. C“ : initial concentration of the volatile (i.e. vanillin) in the shipper paperboard, in the headspace, in the LDPE container, and in the aqueous phase, respectively (pg/g, pg/cc) Ca, C‘, C“, C“ : equilibrium concentration of the volatile (i.e. vanillin) in the paperboard, in the shipper headspace, in the LDPE container, and in the aqueous phase, respectively (pg/g, pg/cc) WB V3 VL VA : weight or volume of the paperboard, of the shipper headspace, of the LDPE container, and of the aqueous phase, respectively (g, cc) Ram : partition coefficient between paperboard and headspace S : solubility coefficient between sorbate vapor and LDPE Rum : partition coefficient between LDPE and aqueous phase 56 6.2 Theory The partition coefficient between paperboard and headspace phase is defined by : K... = g: 008 = KB/H ' Con (21) Solubility coefficient between sorbate vapor and LDPE is defined by: Ca. = S ' D." (22) where : pd : penetrant vapor pressure in the headspace at the system.temperature. From the Kinetic theory of gases, we have that ””491 (23, where: pressure mole number gas constant temperature <8WIJ'O volume Number of mole n can be calculated by gas weight and molecular weight as shown in Eq 24. _rn n.” (24) where: m : gas weight 57 M : molecular weight From Eq.23 and 24 "1. . p=lVI R T=m,Fi-T=C,Fi-T —V— V ‘M" “M— (25) p=C-Bfil where c is concentration in mass/ volume. From.Eq.25, penetrant vapor pressure in the headspace is shown by: pm=Cw5VI (26) Eq 27 is derived from Eq 22 and Eq.26 c.=s-c..-EN'|I (27) The partition coefficient between LDPE and aqueous phase is defined by: CeL K = LIA C.A CeL '-" KLIA ' Colt (28) Eq.20,21,27, and 28 are rewritten as a system of linear equation: M'Cea + VH'Cu-I + VL'Ca + Va‘Cga = Wa'CIa Cg-KB/H°C¢g+0'C1+O'C“=O (29) 0°Cea+ S‘B'Cat "' Cu. + 0°C“ = 0 O‘Ca + O'Cuq + Ca - KL/A'CQA = 0 where: fi__R-1' ‘T 58 By using the definition of matrix multiplication, the system of linear equations can be written as a linear transformation Ax=y (30) where : VW; \& Vi VA 1 -Ke/H 0 0 A 0 es —1 0 0 0 1 -KLIA Cge WH'CB _ Cit _ 0 X Ca Y— 0 Cu 0 Rank [A]... = n was obtained based on the independency of the ' rows in matrix A. Therefore, expressions for C.B ,C.El ,Cd, , and Cd , which were obtained by Cramer's rule (Kreyszig, 1979), are presented below. 0.. = WB'C" , , (31) (w8 + IEBIH RBI IH BIH. IA “5 - CB 8 BV (32) (Kent W3 +VH +8 8 VL-I- T—L/A‘) c .. ,' WQIFL (33) (TI—M" "- w " 57‘s ”L * 123:) .. .. .w "H (B/H .IA 8+ LIA H+K.IA' VL+VA) CaH where : 59 In this study, values for C1,, Ram: S, and Km: for o- vanillin and vanillin were determined These parameters for o-vanillin and vanillin are as follows: o-vanillin C1, : 0.34 (pg/g) RM, : 21.6 (cc/g) S : 11.3 (kg/m3 Pa) RNA : 574 vanillin C13 : 0.9 (pg/g) Km : 77.5 (cc/g) S : 0.3 (kg/m3 Pa) Kala : 3.8 Values for WB,VB, V1,! and VA were obtained from the designed packaging system. These values are as follows: W13 : 90.3 (g) (corrugated shipper) Va : 1797 (cc) ( 2614 cm3 of total package volume - 217 cm3 , V1. / 0.92 g/cm3v - 600 cm3, VA x 1 ml/g) (corrugated shipper headspace volume) V1. : 200 (cc) (2 cm3/piece x 100 pieces) (LDPE ampules total mass) Vi : 600 (cc) (6 cc x 100 pieces) (total ampule volume) Values for Ca ,Ca ,C.L , and C“ for o-vanillin and vanillin determined by substitution into Equations 31,32,33 and 34, are summarized in Table 10. 60 Table 10 : Prediction of equllibrium concentration in the paperboard and in an aqueous phase ca ca 0,,L C... ug/g ug/cc uglcc pg/ cc O-vanillin 1.5 X 10'"2 7.0 X 10" 1.4 X 10'1 2.2 X 10" Vanillin 0.59 7.7 x 10-3 3.7 x 10'2 5.9 x 10"3 61 The migration of volatiles such as vanillin and o- vanillin from a corrugated paperboard shipper and their subsequent permeation and associated equilibria is a complex process, depending in part ( assuming no chemical reaction ' takes place) on the diffusivity of the migrating species. Diffusivity, or the diffusion coefficient (D) is defined as the tendency of a substance to diffuse through a resisting phase. Migration of volatiles such as vanillin and o- vanillin from a corrugated paperboard shipper, therefore, is considered a mass transport process under defined test conditions to include: (i) time, (ii) temperature, and (iii) the nature and volume of the respective phases associated ‘with the system equilibrium. The driving force for the migration is the concentration gradient, where the volatile species of interest transfer from a region of higher concentration (i.e. corrugated paperboard) to a region of lower concentration (i.e. primary package contents). The rate of diffusion of the migrant in the present system.is related to the resistance against the movement within the bulk phase of the corrugated paperboard, shipper headspace, polyethylene ampule wall and contained aqueous phase. Thus, if migration of vanillin or o-vanillin from the corrugated paperboard shipper to the contained aqueous product is to occur, the migrant has to undergo the following equilibrium processes. Corrugated board _, —> shipping container <— HeadSpace ‘— Polyethylene (LDPE)—"_ Aqueous Phase Further, the equilibrium distribution of vanillin or o- vanillin as indicated above can be considered a function of 62 migrant-contact phase interaction affinity and diffusion. The affinity or interaction of the migrant for the corrugated paperboard, LDPE ampule and aqueous phases will determine the equilibrium.amount of migrant transferred to the respective phases. Such migrant-contact phase interaction affinity may become increasingly more important as the initial migrant concentration in the corrugated paperboard decreases. While the affinity of the migrant in the respective phases will determine the equilibrium.concentration levels, the rate of diffusion of the migrant through the paperboard, LDPE ampule and aqueous contact phase will affect the time to attain equilibrium. Assuming that the rate limiting step in the equilibrium process is the rate of diffusion of vanillin or o-vanillin within the bulk phase of the LDPE ampule wall, and the diffusion coefficient of o-vanillin in LDPE is 8.5 mP/sec x 10*“, substitution into equation 11 can be provide an estimate of the time for the system to read steady-state (i.e. equilibrium) as a function of ampule wall thickness. The estimated equilibrium.time values are summarized in Table 11, and illustrated graphically in Figure 12 where the estimated time to steady state is plotted as a function of ampule wall thickness. As shown, application of equation 11 indicates that the times to reach equilibrium are very long and are on the order of 103 days for a 20 mil wall thickness ampule. 63 Table 11 : Time to steady state for the Diffusion of o-vanlllln Into LDPE film as a function of ampule wall thickness Wall thickness Diffusion time of LDPE film (days) (mil) 0 0 1 0.02 5 10 1 0 1 60 1 5 811 20 2562 25 6254 30 12968 64 15000 9 'EIIXXX)‘ 0) .§ H c .C a £50004 0 01: WTIIIIIII 05101520253035404550 Thickness (mil) Figure 12 . Diffusion time of o-vanlllln vs. wall thickness of LDPE film SUMMARY AND CONCLUSIONS The volatilization of vanillin and ortho-vanillin from a corrugated paperboard shipper and their sorption by a low density polyethylene primary container was investigated. (1) The initial concentration in the paperboard was measured by a thermal stripper-thermal desorption method and by a solvent extraction method. Both methods gave equivalent values for o-vanillin levels in corrugated paperboard. Average values of 0.34 and 0.9 (pg/g) of o-vanillin and vanillin, for the initial concentration in the paperboard were derived, respectively. (2) Partition coefficient values for vanillin and o-vanillin between paperboard and headspace were determined by gas chromatography and a thermal stripper-thermal desorption procedure. Partition coefficient values for vanillin and o- vanillin were 77.5 and 21.6 respectively at 23 °C, where the partition coefficient is defined as the equilibrium concentration in the paperboard over the equilibrium concentration in the headspace. (3) Sorption studies indicated that for o—vanillin, the solubility coefficient was independent of sorbate concentration over the range of vapor pressures evaluated. The mean solubility coefficient value determined was 11.3 11.5 Kg/mP Pa at 23 °C. The relationship between the solubility coefficient and the initial o-vanillin 65 66 concentration, indicated that the sorption process follows Henry's Law. Sorption studies of vanillin into LDPE yielded solubility coefficient values of 0.3 i 0.08 and 0.23 i 0.05 Kg/m3 Pa, respectively at 23 and 32 °C. (4) The half-time diffusion coefficient (D,)'was determined from the o-vanillin sorption studies and showed no trend as a function of o—vanillin vapor pressure, but rather a random variation, with D, values ranging from 6.2 to 11.5 mP Isec x 10-13, with a mean value of 8.5 t 2.1 m2 /sec x 10'”. (5) Partition coefficient values for vanillin and o-vanillin between LDPE and an aqueous phase were determined. Partition coefficient values for vanillin and o-vanillin were 3.8 and 574 respectively at 23 °C, where the partition coefficient is defined as the equilibrium.concentration in LDPE over the equilibrium concentration in an aqueous phase. (6) The equilibrium distribution of vanillin and o-vanillin between the corrugated board/shipper headspace/LDPE/and an aqueous contact phase was modeled. The model was based on the assumptions that: ° volatilization of vanillin from the corrugated board is only inward to the shipper headspace. - The aqueous phase does not react with the vanillin or o-vanillin. - The initial concentration of the volatile in the system is assumed by: (RH = Cug==(hA = 0 This means that volatile is initially present only in the paperboard. 67 - At equilibrium, the mass balance of the volatile in the system.is described by: VVB'CB=VIb'C.a + VH'Cu-I + VL'Cq, + VA'CaA Taking into account KB ,3, S and KL,“ expressions were derived to allow estimation of the equilibrium.concentration of volatiles in the paperboard and in the aqueous phase as well as in the shipper headspace and polyethylene ampule. The equations are as follows: 008 = we ' CIB W S'fi'WA VV +— SB L +-———————) ( B KV—sm+ KB/H KBIH' KL/A _ WE ° C; CG'I - S'B'VA (KB/H' We + Va + 8 "fi V), + —R_L/A ) Ca = ”CV—'3 V T('e/H A ‘TB— “5‘? ”L i m) Cie Co = A (K_—e/H'SB KL/A We + KL/A VH +KL/A WL +WA) S 5 where: Can Can CeL Ca : equilibrium concentration of the volatile (i.e. vanillin) in the paperboard, in the shipper headspace, in the LDPE container, and in the aqueous phase, respectively (pg/g, pg/cc) fi-V In this study, estimated values of CeB, CeH, CeL, and CeA for o-vanillin were 1.5 x 10'2 (pg/g), 7.0 x 10" (pg/cc), 1.4 x 104'(pg/cc), and 2.2 x 104*(pg/cc), respectively. Estimated values of CeB, CeH, CeL, and CeA for vanillin were 68 0.59 (pg/g), 7.7 x 10'3 (pg/cc), 3.7 x 10‘2 (pg/cc), and 5.9 x 10-3 ( pg/ cc) , respectively . The equilibrium.distribution of vanillin or o-vanillin, as indicated above, can be considered a function of migrant- contact phase interaction affinity and diffusion. The affinity or interaction of the migrant for the corrugated paperboard, LDPE ampule and aqueous phases will determine the equilibrium.amount of migrant transferred to the respective phase. Such migrant-contact phase interaction affinity may become increasingly more important as the initial migrant concentration in the corrugated paperboard decrease. While the affinity of the migrant in the respective phases will determine the equilibrium.concentration levels, the rate of diffusion of the migrant through the paperboard, LDPE ampule and aqueous content phase will affect the time to attain equilibrium. Assuming that the rate limiting step in the equilibrium process is the rate of diffusion of vanillin or o-vanillin within the bulk phase of the LDPE ampule wall, and the diffusion coefficient of o-vanillin in LDPE is 8.5 m?/sec x 11r43, substitution into the following equation can provide an estimate of the time for the system to reach steady—state (i.e. equilibrium), as a function of ampule wall thickness. D.==(104QF 105 Application of the “half-sorption time" expression indicates that the times to reach equilibrium are very long and are on the order of 103 days for a 20 mil wall thickness ampule. PROPOSAL FOR FUTURE RESEARCH A number of studies can be proposed for future investigation which can lead to an increased understanding of the equilibrium partition distribution of volatiles derived from a corrugated shipping container and their subsequent transport through a contained polymeric primary package system. Studies proposed for future investigation include: (1) Determination of the equilibrium partition distribution (RB/3) of vanillin and o-vanillin between corrugated board and headspace as a function of temperature. (2) Determination of the permeability constant of vanillin and o—vanillin through a low density polyethylene film.as a function of temperature. (3) Determination of the equilibrium distribution of vanillin and o-vanillin in the designed package system, which consists of a low density polyethylene ampoule and an aqueous phase, as a function of vanillin and o-vanillin concentration. And to compare the experimentally determined concentration levels of vanillin and o-vanillin in the contained aqueous phase to the calculated values based on the proposed mathematical expressions. 69 APPENDIX APPENDIX A Standard Calibration by Stripping Thermal Desorption method (Part 1) Materials: Ortho-vanillin 99% (Aldrich chemical company, Inc.) Vanillin 99% (Aldrich chemical company, Inc.) Thermal desorption unit (Model 890, Dynatherm.Analytical Instrument, Inc. Kelton, PA) Carbotrap 300 sorbent tube (Supelco Co., Bellefronte, PA) Gas syringe (Hamilton) Gas Chromatograph (HP5890A, Hewlett Packard, Avondale, PA) Procedure: 1. Sample preparation Liquid standard of concentrations 5, 10, 15, and 20 ppm (wt/v) were prepared using Aoetonitrile as a solvent. 2. A 1.5 pl sample of each standard solution was injected into carbotrap directly using a 10 pl liquid syringe. 3. The sorbent tube was put into the tube chamber of the thermal desorption unit. Vanillin and O-vanillin was desorbed directory into the GC column for analysis. 4. The conditions of Thermal desorption unit and Gas Chromatograph were as follows: 70 71 Thermal desorption unit Desorption time: 4 min Temperature: Transfer line - 230 h: Desorber - 340 °C Valve - 230 °C Gas Chromatggraph column: SPB-l: 0.32 mm i.d. x 30 m fused silica capillary column conditions: Initial temperature: 40 °C Initial time: 5 min Rate: 7.5 °C/min The retention time of O-vanillin and Vanillin were 13.9 and 15.8 minutes. Calibration factors (C.F.) were determined by the slope of the calibration curve between area units and quantity. The calibration data and standard curve for O-vanillin and Vanillin are shown in Table A-1, A-2 and Figure A-l, A—2 respectively. The derived calibration factor is as follows: C.F of O-vanillin: 534.5 (Area units x103 / Quantity pg) C.F of Vanillin: 830.8 (Area units x103 I Quantity pg) 72 Table A-1: O-vanillin calibration data by sorption thermal desorption method concentration quantity of area response of o-vanillin o-vanillin injected average S'D° solution x 106 (g) Pb 0 0 0 0 100 0.15 85050 4035 500 0.75 439500 7882 1000 1.5 797795 12458 Table A—2: Vanillin calibration data by sorption thermal desorption method concentration quantity of area response of o-vanillin vanillin injected average 8°D' solution x 106 (g) All” 0 0 0 0 100 0.15 100363 6558 500 0.75 676000 10689 1000 1.5 1225286 25689 73 y=534.5x r=0.999 800 600"I E X (D 400‘ a: C D 8 h - < 200 01;! I I I 0 0.5 l 1.5 2 Quantity (pg) Figure A-1. O-vanillln standard calibration curve by sorption thermal desorption method 74 y=830.8x r=0.998 1500 (a) 1000‘ ,_ x (D :2: 5 (U SIX)" 0 h < 0L I T f 0 05 l 15 2 Quantity ( lug) Figure A-2 . Vanillin standard calibration curve by sorption thermal desorption method APPENDIX B Standard Calibration by Stripping Thermal Desorption method (Part II) Materials: Ortho-vanillin 99% (Aldrich chemical company, Inc.) Thermal stripper instrument (Model 1000, Dynatherm Analytical Instrument, Inc. Kelton, PA) Thermal desorption unit (Model 890, Dynatherm.Analytical Instrument, Inc. Kelton, PA) Carbotrap 300 sorbent tube (Supelco Co., Bellefronte, PA) Sparging vial (Supelco Co., Bellefronte, PA) Gas syringe (Hamilton) Gas Chromatograph (HP5890A, Hewlett Packard, Avondale, PA) Vial Procedure: 1. Sample preparation Solid materials of O-vanillin was put into the vial and crimped by teflon/silicon septa and aluminum cap. Vials were stored at 70°C oven. 2. Gas phase of volume 100, 200, and 400 pl was obtained from the vial by using gas syringe after 3 times rinses and injected into sparging vial in the oven of the thermal 75 76 stripper instrument. 3. Carbotrap 300 sorbant tube absorbs O-vanillin by helium. purge. 4. The sorbent tube was removed from the thermal stripper and put into the tube chamber of the thermal desorption unit. 0- vanillin was desorbed directory into the GC column for analysis. 5. Gas sample of volume 400 pl was injected to Gas Chromatograph directly under the method of Appendix A to get o-vanillin quantity in the gas. 6. The condition of Thermal stripper instrument, Thermal desorption unit, and Gas Chromatograph were as follows: Thermal strippgr instrument: Temperature: Block - 160°C; Oven - 80°C; Tube - 50°C; Time: Preheat - 2 min; Spurge - 5 min; Dry - 2 min; Gas flow rate: Spurge flow - 100 cc/min Dry flow - 50 cc/min Thermal desogption unit Desorption time: 4 min Temperature: Transfer line - 230 %: Desorber - 340 °C Valve - 230 °C Gas Chromatograph column: SPB-l: 0.32 mm i.d. x 30 m fused silica capillary column 77 conditions: Initial temperature: 40 °C Initial time: 5 min Rate: 7.5 °C/min The retention time of O-vanillin from stripping thermal desorption data was 13.9 minutes. The data and curve between area units and quantity injected are shown in Table B-1 and Figure B-l, respectively. The derived factor between area units and quantity injected is 1.314 (Area units x103 / Quantity injected pl) O-vanillin quantity in 400 pl sample was shown in Table B-2 by using calibration factor of standard calibration by Gas Chromatography in Appendix C. Calibration factor was determined by O-vanillin quantity in 400 pl sample and the factor between area units and quantity injected. The calibration data and standard curve are shown in Figure B-2. The derived calibration factor is 511.0 (Area units x103 / Quantity pg). 78 Table 8-1 : o-vanillin data by sorption thermal desorption method (part II) area response Injected quantity x103 (”1) average S.D. 0 ' 0 o 100 139 11 200 231 28 400 530 89 Table H-2: Area response of 400 pl vanillin vapor I Sample number I Area response=J 1 764300 2 698800 3 723800 4 7411 00 5 684000 ave. 722400 S.D. 28770 79 y=1.314x 500‘ 400‘ 300- 200" Area Units ( x 103) 100‘ 0 I I I I 0 100 200 300 400 500 Quantity injected (pl) Figure B-1. O-vanlilln standard calibration curve for stripping thermal desorption method (part ll) 80 y=511x 600 ”—1 ca) 400- f. x 9 300- "E D m 2004 2 < 100- 01'..- I f I I 0 0.25 0.5 0.75 1 1.25 Quantity(pg) Figure B-2 . O-vanlllln standard calibration curve by stripping thermal desorption method (part II) APPENDIX C Standard Calibration by Gas Chromatography Materials: Ortho-vanillin 99% (Aldrich chemical company, Inc.) Vanillin 99% (Aldrich chemical company, Inc.) Acetonitrile (Aldrich chemical company, Inc.) Volumetric flask with stopper 10 pl liquid syringe (Hamilton Co.,Reno, NV) Gas chromatograph (HP5890, Hewlett Packard, Avondale, PA) Procedure 1. Sample preparation Liquid standard of concentrations 5, 10, 15, and 20 ppm (wt/v) were prepared using Acetonitrile as a solvent. 2. A 1.5 p1 sample of each standard solution was injected into gas chromatograph for analysis using a 10 pl liquid syringe. 3. The conditions of Gas Chromatograph were following: column: Supelcowax 10: 0.32 mm i.d. x 60 m fused silica capillary column conditions: Initial temperature: 60 °C Initial time: 1 min Rate: 5.0 °C/min 81 82 Injector temp. 220 °C Detector temp. 250 °C The retention time of O-vanillin and Vanillin were 35.4 and 44.4 minutes, respectively. Calibration factors (C.F.) were determined by the slope of the calibration curve between area units and quantity. The calibration data and standard curve for O-vanillin and Vanillin are shown in Table C-1, C-2 and Figure C-1, C-2 respectively. The derived calibration factor is as follows: C.F of O-vanillin: 0.725 (Area units x103 / Quantity x10-9 gms) C.F of Vanillin: 0.812 (Area units x103 / Quantity x10-9- 91118) 83 Table C-l: O-vanillin calibration data by CC concentration of o-vanillin solution (ppm) quantity of o-vanillin injected average area response x 109 (g) 0 0 0 5 7.5 3502 10 15 8813 15 22.5 15509 20 30 21210 Table C-2: Vanillin calibration data by GC concentration of quantity of average area vanillin vanillin injected response solution (ppm) x 109 (g) 0 0 0 5 7.5 9001 10 15 15147 15 22.5 17827 20 30 26021 84 y=0.725x r=0.995 25 U 20 " v- 15 - x (D :0: C 10 ' D (U G) h- < 5 _ U 0 I.- I I I 0 10 20 30 40 Quantity ( x 10'9 gms) Figure C-1 . O-vanlllln standard calibration curve by G.C. 85 y=0.812x r=0.987 30 25—1 ‘35 2°" ‘- D X en 15~ El .1: c D 10" “i D 9 < S- 01;- I I I 0 10 20 30 40 Quantity ( x 10'9 gms) Figure C-2 . Vanillin standard calibration curve by G.C. APPENDIX D Saturated Vapor Pressure The saturated vapor pressure was used to determine driving force of organic penetrant at the test condition of the sorption studies. The procedure is as follows: approximately 2 grams of solid was put into a 50 ml vial and stored for one week at 23 and 32 °C to get equilibrium. 400 pl of gas in the headspace was placed and injected into the GC directly. The saturated vapor pressure values of O-vanillin and Vanillin are shown Table D-1 and Figure D-l. Vapor pressure of the organic penetrant is calculated from the ideal gas law: p=nRT=WRT v MV where: p: vapor pressure (Pa) W: weight of organic penetrant (g) R: gas constant T: temperature (°K) M: molecular weight of organic penetrant V: injection volume The vapor activity is the ratio of actual vapor pressure to saturated vapor pressure at the temperature observed and 86 87 showed by: =&-. a p. where: a: vapor activity pa: actual vapor pressure (Pa) p,: saturated vapor pressure (Pa) 88 Table D-1: Saturated vapor pressure of O-vanillin and Vanillin saturated vapor pressu re temperature (Pa) (°C) average S.D. 23 0.41 0.07 O-Vanillin 32 1.73 0.2 70 48.2 2.0 23 0.14 0.02 Vanillin 32 0.42 0.05 70 3.88 0.4 89 y = 0.056 *100'042" r = 0.993 50 40d ’11? Q: Q 30- h ‘2’ 9 Q 20-l 8 (U > 10« 0' I I 0 20 40 60 80 Temp. (°C) Figure D-1 . Saturated vapor pressure of O-vanillin 90 y = 0.037 * 100929" r = 0.989 ’4? Q: G) L— 3 8 2‘ G) h D. 8 (U 1‘ > 0 r I r 0 20 40 60 80 Temp. (°C) Figure D-2 . Saurated vapor pressure of Vanillin BIBLIOGRAPHY BIBLIOGRAPHY Baner, A.L.(III),"The measurement and analysis of the diffusion of toluene in polymeric films." Master Thesis, School of Packaging, Michigan State University, East Lansing, MI, 1987 Baner, A.L.(III),"Partition coefficients of aroma compounds between polyethylene and aqueous ethanol and their estimation using Unifac and Gcfeos." Doctor Thesis, Food Science and Human Nutrition and Agriculture Engineering department, Michigan State University, East Lansing, MI, 1992 Code of federal regulations, 21., The national archives of the united states, 1987 Corrugated box manufacture's handbook, 3rd ed., S&S corrugated paper machinery co., inc., (1965) Crank, J., The mathematical of diffusion,2nd ed., Clarendon Press., Oxford, England, (1975) Crosby, N.T., Food Packa in Materials, Applied science publishers LTD., London, (1981) Fiber Box Handbook, Fiber box association Rolling Meadows, IL, 1989 Friedrich Emil Brauns, The chemistgy of Lignin, Academic press Inc., New York and London, (1960) Fredenslund, A., Jones, R.L., Prausnitz, J.M., Group contribution estimation of activity coefficients in nonideal liquid mixtures., AICHe Journal.1975 Grant J.W.David, Higuchi Takeru, Solubility behavior of organic copppund,John wileyasons, New York, (1990) Hernandez R. J.,Giacin J.R.,Baner A.L., ” The evaluation of the aroma barrier properties of polymer films.", Journal of plastic film & sheeting, vol. 2, July 1986 Imai takayuki, "Absorption of organic volatiles by sealant films." Master Thesis, school of packaging, Michigan State University, East Lansing, MI, 1988 91 92 James E. Kline, Pa r and Pa rboard, Miller Freeman Publications Inc., San Francisco, section 11,111, (1982). James P. Casey, Pulp and pappr chemistgy and chemical technolo , 3rd ed., V01. 1, John wiley & sons New York, 1980 Krevelen Van, Pro rties of l rs,3rd., Elsevier science publishing company Inc., New York., 1990 Kitti wangwiwatsilp,"The effect of surface sulfonation on barrier properties of polymer films." Master Thesis, School of Packaging, Michigan State University, East Lansing, MI, 1993 Kreyszig Erwin, Advanced engineering mathematics, John Wiley & sons, Inc., New York., 1979 Norbert G. Rennicke, The manufacture of papgrboard, NOrbert Rennicke, Hinsdale,Ill., 1971 Paine F.A., A handbook of food cka in , Leonard Hill, London, 1983 Paine F.A., The packaging user's handbook, Avi, New York, 1991 Robert R. A. Higham, A handbook of a rmakin , Oxfod university press, New York Toronto, 1963 Robertson, G.L., "Structural Properties of Plastic Polymers", Food Packa in : Princi al and Practice, 1992 Rogers, C.E., Permeability and chemical resistance of pplypp;§,Rheinhold, New York, 1964 S. coffey, Rodd’s chemistgy of carbon copppunds,2nd ed., Elsevier scientific publishing company, New York,vol. III partA,1971 S. coffey, Rodd's chemistpy of carbon copppunds,2nd ed., Elsevier scientific publishing company, New York,vol. 111 part D,1976 Shinoda kozo, Principles of solution and solubility, Marcel Dekker, Inc., New York and Basel, 1978 Sven A. Rydholm, Pul in rocesses, Interscience publishers, New York, Part II, 1965 Sybil P. Parker, Encyclopedia of chemistry, 2nd ed. McGraw-Hill, New York, 1993 93 Van Ness, J.H. In encyclopedia of chemical technology, 3rd ed.,willy, New York, 1983, vol. 23