.37.. 3.5 Air A u 4“— .Rcénux u: .3 z. 1.375.! :4. u .r... . Ia... .. ¢ 1. I3. \r? 4.. :‘nxiic. ). f 5. .hQal . .n: :3. 10.51. .5". .l..r$mx)aa. (a) .. . 2x yr :90. ‘4 «3.... .. ».(a S T , . .1 Y? 3-3;... 4-3..- . . . ,1. ...- i ‘ . . .1 J .(..~v. .V.. ... . . . .. gmvumafiw . . .. .55.....5.%u§§.,.{.s.$ii. :33 . . (talx . v. 3. . . «Mannmu .. . , 555.53" 5.3% 555555 mm 3 . . at... 1 rv- “£538 1 LIBRARY 90 ? Michigan State University This is to certify that the thesis entitled IMPACT OF POLYMER PROCESSING ON SORPTION OF BENZALDEHYDE VAPOR IN RUBBERY POLYPROPYLENE presented by YING QIN has been accepted towards fulfillment of the requirements for the degree In SCHOOL OF PACKAGING MQ/L/WJMM Major Professor's Signature 05/30/2006 Date MSU is an Affirmative Action/Equal Opportunity Institution -‘.- ----'-.-- ' - -.---'9 ‘gg.-.--p-o-.-----.—--.---v--.-1A-v-,_-_.-.-cpx-;-... l-A‘J- PLACE IN RETURN Box to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE W 2211 020707 r‘ 122007 130539.21“? 2/05 p:IClRC/DateDuo.indd-p.1 . __._ IMPACT OF POLYMER PROCESSING ON SORPTION OF BENZALDEHYDE VAPOR IN RUBBERY POLYPROPYLENE By Ying Qin A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of Packaging 2006 ABSTRACT IMPACT OF POLYMER PROCESSING ON SORPTION OF BENZALDEHY DE VAPOR IN RUBBERY POLYPROPYLENE By Ying Qin With the increasing usage of polymers as food and pharmaceutical packaging materials, more information is needed on the barrier properties of these packaging materials. The organic barrier property of polypropylene (PP) was studied in this project. A continuous gravimetric method was developed to study the sorbate uptake and sorption kinetics of benzaldehyde vapor in rubbery PP resin, sheet and polyhedron thermoformed container by using a SGA-lOOR gravimetric analyzer including Rubotherm magnetic suspension microbalance at 25°C, 0% RH. Solubility coefficients for benzaldehyde sorption in PP resin, sheet and container were determined based on the equilibrium uptake. Diffusion coefficients were also calculated on PP resin and sheet in the Fickian diffusion region. The experimental results demonstrated that the polymer conversion process had a great impact on the mass transfer properties of PP. The sorption of benzaldehyde in the PP polyhedron container was twice that of the sorption in PP sheet at lower vapor activities, and the sorption in sheet was almost three times the sorption of the resin at higher vapor activities. The sorption dynamics among the PP forms were different from each other. This finding emphasizes the need to assess mass transfer of formed polymer sheets or containers to accurately determine the barrier properties of packaging systems. Acknowledgements The most appreciation should be expressed to my supervisor, assistant professor Maria Rubino, for her untiring and invaluable supervision and guidance throughout the course of my research; I thank her for awarding me a research assistantship to pursue my degree. Next, I would like to thank my committee members assistant professor Rafael Auras, professor Gregory L. Baker, professor Hugh E. Lockhart for their kind help and suggestions. I also thank all the faculties and staff from SOP for their help. I would like to express my personal appreciation to Pankaj Kumar for his help and cooperation in mold design and polymer processing; Dr. Wu and Li Xiong for their useful suggestions and discussions; Eva for her help in GC analysis; Kuo Chun Yen (Kevin), Xin Li and Xuebing Zhang (Ben) and for their friendship. Also I could not forget all the graduates in School of Packaging, if not for them, the life in SOP will be to too plain. Finally, my deepest gratitude should be expressed to my family: my parents and sisters who gave me so much love and confidence during these years. Last but not least, to my husband, Shan Zhao, he shared all my happiness and sadness during this project. Without his love and encouragement, especially when I felt discouraged, this project would not be possible. iii TABLE OF CONTENTS LIST OF TABLES ....................................................................... LIST OF FINGURES .................................................................... ABBREVIATIONS AND SYMBOLS ................................................ CHAPTER 1 General Introduction .................................................................... General Introduction ..................................................................... 1.1 Mass Transfer ....................................................................... 1.2 Mass Transfer in Polymer Systems .............................................. 1.2.1 Basic Concepts ............................................................... 1.2.2 Sorption Kinetics ............................................................ 1.2.2.1 Constant Diffusion Coefficient .................................... 1.2.2.2 Sorption by a Swelling Sheet ...................................... 1.2.2.3 Concentration-dependent Diffusion Coefficient ................ 1.2.3 Sorption Model ............................................................... 1.2.3.1 Henry’s Law Sorption .............................................. 1.2.3.2 Langmuir Equation ................................................. 1.2.3.3 Dual-Mode Sorption ................................................ 1.2.3.4 Flory-Huggins Mode ............................................... 1.2.3.5 BET Equation ......................................................................... iv viii ix xi 10 11 13 15 15 16 17 18 1.3 1.4 1.5 1.6 1.2.4 F ickian and Non-Fickian Characteristics ................................. 1.2.4.1 Sorption Well Above Tg ........................................... 1.2.4.2 Sorption Below Tg .................................................. 1.2.5 Temperature and Concentration Dependence ............................ 1.2.6 Effect of Polymer Relaxation .............................................. Parameters Affecting the Transport Properties ................................. Methods for Studying the Sorption of Organics by Polymer Materials. . Mass Transfer Studies on Polypropylene ....................................... Objectives of the Project .......................................................... CHAPTER 2 Materials, Instrument and Methods ................................................... 2.1 2.2 2.3 Materials ........................................................................... 2.1.1 Organic Sorbate ............................................................ 2.1.2 Polymer ...................................................................... Polymeric Processing ............................................................. 2.2.1 Casting PP Sheet ............................................................ 2.2.2 Thermoforming PP Containers ............................................ Instrument .......................................................................... 2.3.1 SGA-lOOR Gravimetric Analyzer ....................................... 2.3.2 Operation Procedure ....................................................... 2.3.3 Differential Scanning Calorimetry (DSC) Analysis ................... 19 20 22 23 25 28 31 34 35 37 37 37 37 38 38 39 40 4O 41 41 2.3.4 Gas Chromatography (GC) ................................................ 2.3.5 Polymer Density test ....................................................... CHAPTER3 Measurement ofSorption of Benzaldeh yde in Polymer with Magnetic Suspension Electrobalance ................................................ 3.1 Introduction ........................................................................ 3.2 Experimental ....................................................................... 3.2.1 Gravimetric Instrument .................................................... 3.2.2 Materials ..................................................................... 3.2.3 Quantification of Equipment Variability and Noise .................. 3.2.4 Sorption Test of Benzaldehyde in PP (Resin, Sheet and Polyhedron Shaped Sheet) .............................................................. 3.3 . Results and Discussion ........................................................... 3.3.1 Evaluation of the Instrument Noise ...................................... 3.3.2 Equilibrium Criterion Determination .................................... 3.3.3 Sorption of Benzaldehyde in PP .......................................... 3.4 Conclusion ......................................................................... CHAPTER 4 Impact of Polymer Processing on Sorption of Benzaldeh yde Vapor in Semicrystalline Polypropylene ..................................................... vi 42 42 45 47 49 49 54 56 57 58 58 61 65 67 69 4.1 Introduction ........................................................................ 4.2 Experimental and Method ........................................................ 4.2. 1 Materials ..................................................................... 4.2.2 Polymer Processing ........................................................ 4.2.3 Gravimetric Method ....................................................... 4.2.4 Experimental Procedure ................................................... 4.3 Results and Discussion ........................................................... 4.3.1 Impact of Surface Area and Bulk Properties on the Sorption of Benzaldehyde in PP resin ................................................. 4.3.2 Sorption of Benzaldehyde in Different PP shapes ..................... 4.3.3 Impact of Vapor Activity of Benzaldehyde on the Sorption Behavior ..................................................................... 4.3.4 Benzaldehyde Sorption Isotherm ......................................... 4.3.5 Equilibrium Benzaldehyde Sorption and Sorption Kinetics Measurements ............................................................... 4.4 Conclusion ......................................................................... CHAPTER 5 Conclusion ............................................................................... References ....................... . ........................................................ vii 71 74 74 74 75 76 77 77 81 82 83 87 93 95 97 LIST OF TABLES Table# Title Page 1.1 Different models of sorption and typical interactions associated ....... 14 2.1 Polymer characteristics for different PP forms ........................... 44 3.1 Physical properties of PP resin .............................................. 55 3.2 Regression analysis of the instrumental noise and drift .................. 59 4.1 Hansen Solubility Parameters Data for Benzaldehyde and Polypropylene at 25 °C ....................................................... 86 4.2 % Wt. Gain and calculated solubility coefficient (S) for benzaldehyde- PP at 25 °C ................................................... 91 4.3 Calculated Fickian diffusion coefficient (D) for benzaldehyde—PP at 25 °C ............................................................................ 91 viii Figure# 1.1 1.2 2.1 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 4.1 4.2 4.3 LIST OF FIGURES Title Typical isotherm plots of sorbed concentration vs. vapor pressure. Schematic plan of the PP processing in this project ..................... Male mold (left) and thermoformed polyhedron (right) ................ Schematic of gas flow in Rubotherm SGA-lOOR gravimetric analyzer (top) and the coupling/decoupling action between the electromagnet and suspension magnet under different operation conditions (bottom) ........................................................... Polyhedron container thermoformed by male mode .................... Quantification of equipment variability and noise ...................... Quantification of noise in Rubotherm SGA-lOO gravimetric analyzer ......................................................................... Sorption curve for benzaldehyde vapor in PP resin at vapor activity of 0.5 at 25 °C .................................................................. Sorption curve for benzaldehyde vapor in PP resin at vapor activity of 0.7 at 25 °C .................................................................. Sorption curves for benzaldehyde vapor in PP resin at 25 °C. From top to bottom, the vapor activities are 0.9, 0.7, 0.5, 0.3, and 0.1, respectively ................................................................... Sorption curves for benzaldehyde vapor in PP resin, sheet, and polyhedron container at vapor activity 0.9 at 25 °C ..................... Male mold (left) and thermoformed polyhedron (right). Sorption curve for benzaldehyde vapor in PP resin at vapor activity of 0.7 at 25 °C ................................................................. Sorption curves for benzaldehyde vapor in PP at vapor activity of 0.5 at 25 °C .................................................................... ix Page 1 4 36 40 51 55 59 60 62 63 66 66 75 78 79 4.4 4.5 4.6 4.7 4.8 Sorption curves for benzaldehyde vapor in PP with vapor activity of 0.3 at 25 °C .................................................................... Sorption curves for benzaldehyde vapor in PP at 25 °C at vapor activity of 0.5 (left) and 0.9 (right) .......................................... Sorption isotherm of benzaldehyde by PP resin at 25 °C ................ Sorption curve for benzaldehyde vapor in PP resin at 25 °C with vapor activities of 0.5 ......................................................... Comparison of solubility coefficient (S) and diffusion coefficient (D) at different vapor activities ............................................. 81 83 84 89 92 atm Abbreviations and Symbols standard atmosphere, 1 atm=l.01*105 Pascals Langmuir affinity constant centimeter concentration of permeant at the polymer interface concentration of permeant in Henry’s law mode sorption concentration of permeant in Langmuir mode sorption Langmuir capacity constant concentration of organic in the polymer at equilibrium concentration of organic in the vapor at equilibrium diffusion coefficient diffusion coefficient in an amorphous state value of D at zero concentration diffusion of vapor relative to stationary sheet dew point analyzer differential scanning calorimetry mutual diffusion coefficient apparent energy of activation for diffusion activation energy of diffusion apparent activation energy of permeation ethylene vinyl alcohol rate of transfer xi FDA FFV GPC in M Moo mg min mL mm "a Food and Drug Administration Fractional free volume gram gel permeation chromatography Inch kilogram film thickness bond distance between two atoms sorbate uptake at time t sorbate uptake at equilibrium milligram minute milliliter millimeter number average molecular weight mole weight average molecular weight Avogadro’s number nitrogen universal gas constant relative humidity partial pressure of penetrant permeance xii P0 PP 13le psi QCM RPM S* SCC sp gr STP V 1 VA Vsp VW V0 Wt. Hg vapor pressure saturation vapor pressure polypropylene parts per million pound per square in (1 psi = 6894.76 pascals) quartz crystal microbalance radius of atom I convalently bonded to A pesolutions per minute solubility coefficient solubility coefficient in an amorphous sate second specific gravity standard temperature and pressure time absolute temperature glass transition temperature volume fraction of the permeant in the polymer vapor activity specific volume Van der Waals volume occupied volume weight microgram xiii microliter molar heat of sorption tortuosity factor chain immobilization factor density volume fractio of amorphous phase xiv CHAPTER 1. INTRODUCTION GENERAL INTRODUCTION The use of polymers as food and pharmaceutical packaging materials has experienced rapid and continuous growth in recent years. Polymer packages have evolved from simple monolayer structures to multilayer hybrid structures or from simple wraps to sophisticated containers which have additional demands placed on them. With the increasing usage of polymers as food and pharmaceutical packaging materials, more attention has been paid to the barrier properties of polymeric packaging materials (William, 1990). The interaction of polymer packaging materials with the environment and package contents has been the subject of increasing research and material development. New high barrier and biodegradable polymers are increasingly used in packaging applications (Hernandez,1999, 2000; Lange, 2003) and thus there is a need to assess mass transfer phenomena of multiple organic compounds as applied to polymer and packaging science. 1.1 Mass Transfer The mass transfer process between polymer and product and or environment is normally referred to as packaging interaction. Mass interactions in packaging systems develop from the moment the package and product come into contact during production, and extend throughout the package’s shelf life. Packaging interactions include sorption, migration and permeation processes. In addition, chemical reactions (like oxidation) may develop as a consequence of these processes. Sorption is the uptake of product components by the package material. The extent of a sorption process depends on the initial concentration of the sorbate in the product and the equilibrium thermodynamics between the plastic and product. Migration is the transfer of substances originally present in the plastic material into a packaged product. These particular substances are called migrants, they could be residual monomers, solvents, remaining catalysts, and polymer additives. Migration may affect the product’s sensory quality as well as its toxicological characteristics since it may incorporate into the food undesirable components from the package. The presence of potentially migrating substances in the packaging material is the object of intense legal control by the FDA and other regulatory agencies in other countries. Permeation is the diffirsion-controlled molecular exchange of gases, vapors or liquids (called permeant), across a homogeneous packaging material witch has no perforations or cracks. A permeation process may significantly alter the product’s quality since the product may gain or lose components, and develop unwanted chemical reactions. The migration of organic compounds of product into polymers is termed “sorption” or “scalping”. Sorption is a factor that induces quality alteration during storage. The removal of desirable organic components from a product by plastic package materials has detrimental effects, such as change in the ingredient concentration thus changing medical function, or altering the flavor profile. Furthermore, sorption of organic components can compromise the package integrity or damage the package material by causing delarnination of a multistructure or polymer swelling, thus leading to further sorption, and to a reduction in the shelf-life of the packaged product. Sorption of organic compounds is important to the packaging industry because it is an issue related to consumer safety and consumer acceptability of the packaged project. Therefore, the study of mass transfer, especially the solubility, diffusion and permeation of organic compounds through polymeric packaging structure has practical importance (Risch, 1991). Disregarding obvious discontinuities such as cracks or pin holes in the material, permeants would be transferred through the packaging materials in a specific way depending on several factors: 1) The chemical structure of polymer and permeant, such as polarity of the permeating pair. The old rule of “like dissolves/permeates like” can be applied here. 2) Polymer morphology, such as free volume, orientation, tacticity, and crystallinity. Normally increasing crystallinity, density, polymer orientation and cross linking, decreases permeability. 3) The environmental conditions including temperature and relative humidity (RH). The change in permeability with temperature follows an Arrhenius equation within specific temperature ranges. Relative humidity may increase or decrease permeability depending on polymer and permeant type. For example, oxygen permeability increases along with the increase of RH for EVOH and Nylon 6; however, oxygen permeability decreases with increase RH in amorphous nylon. 4) Finally the polymer processing and manufacturing, such as casting, injection blow molding or thermoforming, usually impact polymer morphology which will influence the mass transfer process. 1.2 Mass Transfer in Polymer Systems The diffusive mass transfer, as observed in practical application of polymeric systems is normally categorized in two ways: permeation of simple gases (such as oxygen, carbon dioxide and water), or of organic vapors (such as flavor, additives, monomer, or solvents left in the polymer during polymer processing). The permeation of simple gases through thin layers of polymers is the most widely studied and the permeation rules are relatively simple. While the mass transfer of heavier organic vapors and liquids as occurred in polymeric constructions such as plastic containers and bottles is more complicated. Very often there is stronger interaction between penetrant and polymer, and between penetrant and penetrant. 1.2.1 Basic Concepts The permeation process can be described as a three-step event. First, collision of the penetrant molecule with the polymer is followed by sorption in the polymer. Next, migration through the polymer matrix by difiusion, and finally, desorption of the permeant from the polymer completes the process. The fundamental driving force that prompts a molecular specie to transfer within the polymer, or between a polymer and a surrounding phase is, according to the solution theory, the tendency to equilibrate the species’ chemical potential. For simple gases, a general relationship between the three main permeation properties permeability coefficient (P), solubility coefficient (S) and diffusion coefficient (D) is almost exactly valid: P = S - D (1) where P is the steady-state transport rate of permeant molecules through a polymer membrane of unit area per unit of thickness and driving force, while D is a kinetic parameter, which represents how fast the permeant molecules move through the polymer bulk phase. S is an equilibrium partition coefficient for distribution of the penetrant between polymer and vapor phase. For all three physical quantities P, S and D, the temperature dependence can be described by a Van’t Hoff-Arrhenius equation (Krevelen, 1997) S (T) = S0 exp(—AHS /RT) (2) M) = D() eXP("E1) /RT) (3) P(T) = P0 exp(—E,, /RT) (4) where AHS = molar heat of sorption ED = activation energy of diffusion Ep = apparent activation energy of permeation As a sequence of equation (1), there are P0=SOOD0 (5) EP=AH8+ED (6) Solubility coefficient (S) and diffusion coefficient (D) are usually determined by observing the weight change of a polymer sample during a sorption process by a gravimetric method. Diffusion coefficient (D) and permeability coefficient (P) values are obtained from permeability studies where the transport of a permeant through a polymer membrane is continually monitored (i.e., isostatic procedure) or by quantifying the amount of the penetrant that has passed through the fihn and accumulated as a function of time (i.e., quasi-isostatic procedure). By the gravimetric method, the equilibrium organic uptakes in sorption experiments are used to calculate the solubility coefficient (S). S is an equilibrium partition coefficient for distribution of the penetrant between polymer and vapor phase. It is a measure of the mass of permeant molecules sorbed by a unit of polymer mass per unit of partial pressure, and it is defined according to equation: C S = —” 7 C. ( ) where Cp is the concentration of organic in the polymer at equilibrium (pg aroma/mg polymer), and Cv is the concentration of organic in the vapor at equilibrium (P/Psm, ppm). For the simplest case of more or less ideal permeation behavior, which includes polymers above their glass transition temperature and at low penetrant pressure, the relation between sorbate concentration and partial pressure can be described by the linear sorption isotherm Henry’s law of solubility: C=kpp (8) where C is concentration of diffusion permeant measured at the polymer interface, p is the partial pressure of penetrant at the contacting phase interface, and kn is the proportionality constant of Henry’s law. In fact, it is the solubility coefficient when there is no concentration dependence. In that case, S is the amount of substance (gas) per unit volume of solvent (polymer) in equilibrium with unit partial pressure. For simple gases S is usually given in cm3 (STP) / (cm3 polymer - bar). For the permeation of simple gases of low molecular weight in rubbery polymers and under relatively moderate pressures, the diffusion mechanism is F ickian and the departures from Henry’s law for the sorption are negligible. The basic equations for describing the diffusion process are Fick’s first and second laws of diffusion (Crank, 1975), F = 4953—9 (9) 6x 2 aC=DaC “m where F denotes the rate of transfer, c is the concentration of diffusing substances, t is time, and x is the space coordinate. Then D is the mutual diffusion coefficient, in (length)2/time. To obtain the flux (F) or the diffusion coefficient (D) from equations (9) and (10), initial and boundary conditions associated with the experimental method are needed, and the expressions would be solved (Hernandez, 1986a). Solution of Equation (10), associated with the initial and boundary conditions, can be performed numerically or analytically. In the latter case, a power-series of solutions usually arises when solving for the unsteady state case. Crank (1975) presented simplified equations related to the first approximation of the power series. It should be recognized that only approximate values will be obtained when the diffusion coefficient is calculated using these equations. More accurate estimation of this parameter can be carried out by using other methods; for example, a nonlinear maximum likelihood sequential method based on the Gauss linearization method (Beck, 1977). Alternative forms of Equation (10) for different co-ordinate systems such as may arise in considering diffusion in a cylinder or sphere are all included in the equation: éa—tq = div(gradC). (11) 1.2.2 Sorption Kinetics 1. 2. 2.1 Constant Diffusion Coefficient The method in terms of the uptake of vapor by a plane sheet is based on the assumptions that the diffusion coefficient is constant and the sheet does not swell. The experimental procedure is to suspend a plane sheet of thickness I in an atmosphere of vapor maintained at constant temperature and pressure, and to observe the increase in weight of the sheet and hence the rate of uptake of vapor. The appropriate solution of the diffusion equation may bewritten M 8 °° 1 —D(2m+1)27rzt ’=1—— ———ex 12 Ma, 7r2,,;:,(2m+1)2 p[ l2 ] ( ) if the uptake is considered to be a diffusion process controlled by a constant diffusion coefficient D (Crank, 1975). Here, where M is the total amount of vapor absorbed by the sheet at time t, and M0, the equilibrium sorption attained theoretically after infinite time, I is the thickness of the film. The application of Equation (12) is based on the assumption that immediately the sheet is placed in the vapor the concentration at each surface attains a value corresponding to the equilibrium uptake for the vapor pressure existing, and remains constant afterwards. The sheet is considered to be initially free of vapor. (t / 12)“; represents the value of t / I2 for which M,/ M00 =1/2, is approximately given by t 1 n2 1 a2 9 (75'); '“nzp‘nlfi‘alfil l “3) the error being about 0.001 per cent. Thus we have 0.049 = (~12)... (14) and so if the half-time of a sorption process is observed experimentally for a system in which the diffusion coefficient is constant, the value of this constant can be determined from Equation (14) (Crank, 1975). 1.2.2.2 Sorption by a Swelling Sheet In deriving Equation (12) the thickness 1 of the sheet is assumed to remain constant as diffusion proceeds. In practice, sometimes, the sheet swells and the thickness increases as the vapor enters. Equation (12) can still be used in such cases, as suggested by Crank (1975) (p.207), by taking a frame of reference fixed with respect to the substance of the sheet, and concentration and thickness are measured. Thus the basic volume of the sheet is taken to be its volume in the absence of vapor and use the unit of length CB such that unit C3 contains, per unit area, unit basic volume of the substance of the sheet B. Then the thickness of the sheet, measured in these units, is constant and equal to the original unswollen thickness, and the diffusion coefficient deduced from Equation (12) by substituting the original thickness for l is that for the diffusion of vapor relative to stationary sheet (denoted by DBA). If there is no over-all volume change on mixing, i.e. if the increase in volume of the sheet is equal to the volume of vapor sorbed at the vapor pressure existing in the experiment, the coefficient obtained by this sorption method is related to the mutual diffusion coefficient DV, deduced by the Matano procedure (Crank, 1975), by DBA = DV (l-volume fraction of vapor)2 (15) 1.2. 2.3 Concentration-dependent Diffusion Coefficient Clearly from Equations (12) and (14) the value oft / l2 for which M,/ Mco has any given value, and in particular the value of (t / 12)1/2, is independent of Mao when the diffusion coefficient is constant. However, Crank and Park (1949) showed a set of curves obtained experimentally for the uptake of chloroform by a polystyrene sheet, each curve corresponding to a different vapor pressure and hence a different Mac. It was evident that (t / 12)1/2 decreased considerably the greater the value of the final uptake MC.o and therefore the diffusion coefficient was not constant but increases as the concentration of chloroform was increased. The problem was to deduce quantitatively how the diffusion coefficient is related to concentration, given the half-time of sorption experiments carried out for a number of different vapor pressures. Application of Equation (14) to each sorption curve yields some mean value, D , of the diffusion coefficient averaged over the range of concentration. Calculations by Crank (1968) have shown that for any one experiment D provides a reasonable approximation to (1 / C0) [1° Ddc , where 0 to C0 is the concentration range existing in the sheet during that experiment. By deducing values of D from the initial gradients for each of a series of sorption experiments and using the approximate relationship D=(1/Co)f°Ddc (16) a graph showing DCO as a function of Co can be drawn and numerical or graphical differentiation with respect to Co gives a first approximation to the relationship between D and C. In many cases the first approximation is sufficiently accurate. If not, a method of obtaining successively better approximations is available (Crank, 1956), but in its most general form the calculations involved in evaluating successive approximations are lengthy. In many polymer systems the diffusion coefficient depends approximately either linearly or exponentially on concentration and so, for these cases, correction curves have been produced by Crank (1956) showing the difference between (1/C0)fo Ddc and D/ Do, where Do is the value of D at zero concentration. Use of these curves usually removes the need for a second approximation. I. 2.3 Sorption Model The term sorption represents the dissolution of the penetrant in the matrix polymer. This term includes in the absorption, the adsorption, and the trapping in microvoids or the clustering of aggregates. It is possible that diffusing molecules may be sorbed according to different sorption modes even in the same polymer membrane. Furthermore, the distribution of penetrant between different modes of sorption may change with changes in temperature, swelling-induced structural state, sorbed concentration, time, and other factors. The equilibrium amount of penetrant sorbed and its sorption distribution mode, in a polymer under given conditions, are governed by the thermodynamics of the polymer-penetrant system, in particular by the nature and the force of the interactions. Basically there are five classic cases of sorption as shown in Figure 1.1 (Klopffer, 2001). The typical interaction associated with each mode was indicated by Klopffer (2001) as shown in Table 1.1. It is to be emphasized that sometimes it is likely that two or more modes of sorption will occur concurrently. I. Henry’s Law 11. Langmuir 111. Dual mode IV. Flory-Huggins V. BET Figure 1.1 Typical isotherm plots of sorbed concentration vs. vapor pressure. Table 1.1. Different modes of sorption and typical interactions associated Sorption mode Main component interactions Henry’s law Langmuir - Dual mode Flory-Huggins BET polymer-polymer polymer-penetrant combination of Henry’s and Langmuir modes penetrant-penetrant combination of Langmuir and Flory-Huggins modes l4 1.2.3.1 Henry ’s Law Sorption The simplest case is that of ideal solution behavior with sorbed penetrant randomly dispersed within the polymer such that Henry’s law is obeyed as follows: C = S p (17) S is the Henry’s law constant, which is the solubility coefficient. The solubility coefficient is then a constant independent of sorbed concentration at a given temperature. The sorption isotherm is a linear relationship of penetrant concentration versus partial pressure (or vapor activity), as showed in Figure 1. This mode is observed essentially for low pressures when the penetrant-penetrant and the penetrant-polymer interactions are weak as compared with the polymer-polymer interactions. 1.2.3.2 Langmuir Equation In terms of the molecular pair distribution approach, this type represents a preference for polymer-penetrant pairs to be formed at relatively small pressures, with a smaller amount of sorption of more nearly ideal solution behavior at higher pressures. In physical terms, this represents initial sorption on some kinds of specific sites in the polymer, i.e. pre-existing microvoids or high-area inorganic fillers. When all the sites are occupied, a small quantity of diflusing penetrant may solubilise and dissolve in the polymer with a more or less random distribution according to the following: = ——C'” b” (18) l+bp CH 15 where Cu is the Langmuir capacity constant [cm3(STP)/cm3] and b is the Langmuir affinity constant (atm'l). 1.2.3.3 Dual-Mode Sorption This model was initially proposed by Barrer (1958) to explain the concentration dependence of the solubility coefficient in glassy polymers, it was then extended by Paul and Koros (1976) for the diffusion coefficient. This mode describes the existence of two distinct populations of diffusing molecules (with local equilibrium between them). At the equilibrium pressure p, the concentration of the dissolved molecules in the polymer by an ordinary mechanism dissolution, CD, obeys Henry’s law (Equation 17), and the concentration of molecules sorbed in a limited number of pre-existing microcavities is given by the equation of Langmuir (Equation 18). In this situation, the sorption corresponds to the combination of the above two modes and is then given by this expression for sorption of non-permanent gases in glassy polymers: C=C,,+C,,=Sp+C—”b3 (19) l+bp where the subscripts D and H represent Henry’ law mode sorption and Langmuir mode sorption, respectively, and b is a constant, characterizing the affinity for the specific sites (microvoids or holes). The Equation 19 is a three parameter model, S, C“ and b, and is known as the dual-sorption isotherm or immobilized dual-sorption model. This model is valid for describing the curves observed in the case of sorption of low-activity gases in 16 glassy polymers under moderate pressures, in the absence of strong interactions. 1.2.3.4 F (my-Huggins Mode The Flory-Huggins mode isotherm represents the situation when the interactions between the diffusing molecules are stronger than the penetrant-polymer interactions and the solubility coefficient increases continuously with pressure. The sorption is then given by the F lory-Huggins equation: ln(P/Po)=1n(V.)+(1-Ifi)+z(1-V.)’ (20) where P and P0 are the vapor pressure and the saturation vapor pressure of the gas at the experimental temperature, individually; V1 is the volume fraction of the permeant in the polymer and X is the interaction parameter between the polymer and the solute. Two principal physical interpretations of this behavior are possible: First molecules sorbed in the polymer tend to loosen the polymer structure locally and make it easier for subsequent molecules to enter in the neighborhood of the first than to go elsewhere. This interpretation implies the effective plasticization of the polymer by the sorbed penetrant. Type IV isotherms are observed when a liquid or vapor penetrant is a strong solvent or swelling agent for the polymer. Although polymer-penetrant interactions are relatively strong, they are not specific in the sense of site-penetrant interactions. Another physical interpretation of type IV behavior is reserved for systems in which penetrant-penetrant interactions are inherently stronger than the corresponding polymer-penetrant interactions. The strong penetrant-penetrant interaction causes the association of stable clusters or aggregation of sorbed penetrant molecules, which would be relatively less mobile in comparison with isolated molecules. Hence, if the proportion of clustered molecules increases with increasing sorbed concentration C as implied by a type IV isotherm, then it would be expected that the diffusion coefficient D of the polymer-penetrant system would decrease with increasing C. This has been observed by Machin and Rogers (1972) in a number of water-polymer systems and contrasts with the behavior of sorbed solvents or swelling agents with D increases with C. 1.2.3.5 BET Equation The last case corresponds to the combination of type Langmuir at low pressures and F lory-Huggins mode at higher pressures. Such isotherms frequently describe the sorption of water by the more hydrophilic polymers (Hernandez, 1992; 1994). Initially, water molecules are strongly sorbed in specific sites corresponding to the polar groups (usually hydroxyl, carboxyl or amide) in the polymer, then at higher relative vapor pressures solution or clustering processes may predominate. The above physical interpretations are probably oversimplified. In particular, it is likely that two or more models of sorption will occur concurrently. So far no information on the degree of overlap of the two modes can be deduced from the simple molecular pair distribution treatment. 18 1.2.4 F ickian and Non-Fickian Characteristics A sorption curve is a plot of the amount of vapor M absorbed in or desorbed (in grams) from unit gram or cubic centimeter of dry polymer as a firnction of the square root of time t. The distribution of a diffusant during sorption is governed by the one-dimensional differential equation for diffusion presented by Fick, with the space co-ordinate taken in the direction of the film thickness. Solutions appropriate to the initial condition of the sorption experiment (i.e., at t= 0 the concentration is uniform in the film) were obtained by Crank (1956), subject to the assumptions that D is a function of concentration C only and that when the ambient pressure is changed from an initial value p, to a final value pf the concentrations at the surfaces of the film instantaneously increase or decrease to the equilibrium value Cf corresponding to pf. Sorption curves having characteristics expected from the above-mentioned assumptions for D and the surface concentration are customarily called F ickian or normal type. At the molecular level, the essential conditions for Fickian sorption are associated with a high mobility of polymer segmental units. The important F ickian sorption features indicated by Crank (1968) include: Both absorption and desorption curves are linear in the initial stage. For absorption, the linear region extends over 60% or more of Moo; Above the linear portions the sorption curves are concave to the abscissa, irrespective of the form of D(c); For absorptions from a fixed C, to different Cf’s, the initial slope of the reduced curve becomes larger as the concentration increment Cf - Ci becomes larger, provided that D increases monotonically with C in the 19 range considered. These are characteristic features of such so-called ‘Fickian’ behavior. Systems which show deviations from this behavior have been termed ‘non-Fickian’ or anomalous. The calculation of diffusion coefficients for the latter systems must account for the concurrent relaxation processes with appropriate caution in subsequent interpretation of the results. 1.2.4.1 Sorption Well Above T 3 Previous work demonstrated that sorption and permeation, which are typical diffusion-controlled processes, of organic vapors in polymer solids exhibit different features in the regions above and below the glass transition temperature, T g, of a given polymer. Normally they are rather simple at temperatures above T g and exceedingly complex at temperature below T g (Crank, 1968). At temperatures well above T , the polymer is in the rubbery condition. Various models developed to describe the diffusion of small gas molecules in polymers generally fall into two categories: molecular models and free volume models (Stern, 1990). Molecular models analyze specific penetrant and chain motions together with the pertinent intermolecular forces based on the energy considerations. Barrer (1937) first showed that the diffusion of molecules in polymers was a thermally activated process. Molecular models commonly assume that fluctuating microcavities or holes exist in the 20 polymer matrix and at equilibrium, a definite size distribution of such holes is established on a time-average basis. A hole of sufficient size may contain a dissolved penetrant molecule, which can jump into a neighboring hole once it acquires sufficient energy. A net diffusive flux arises in a preferred direction in response to a driving force; otherwise molecules will diffuse in random directions since their motion has a Brownian nature. The flux magnitude depends on the concentration of holes which are large enough to accommodate a penetrant molecule. Molecular models include these characteristics largely to describe the Arrhenius behavior of diffusion coefficients observed experimentally, i.e. D = D0 exp(—Eapp /RT) (21) where Eapp is the apparent energy of activation for diffusion, Do is a constant, R is the universal gas constant, and T is the absolute temperature. A correlation is found between Eapp and the molecular diameter of the penetrant, but no theoretical expression for D has been obtained with molecular models. Free volume models attempt to elucidate the relationship between the diffusion coefficient and the free volume of the system, without consideration of a microscopic description. One of the most promising and earliest free volume models was developed by Fujita (1960). Fujita suggested the molecular transport is due to redistribution of free volume and not due to a thermal activation. The free volume models argue that the total free volume is a sum of two contributions. One arises from molecular vibrations and cannot 21 be redistributed without a large energy change and the second is in the form of discontinuous voids. Diffusion in such a liquid is not due to a thermal activation process, but is assumed to result from a redistribution of free-volume voids caused by random fluctuations in local density (Stern, 1990). The basic idea of this theory is that a diffusing molecule can only move from one position to another when, in its neighbourhood, the local free volume exceeds a certain critical value. Based on this theory, the dependence of D with parameters such as the concentration, the penetrant shape and size, the temperature and the glass transition temperature of the polymer can thus be explained. 1.2.4.2 Sorption Below T 8 When the polymers are at temperatures below the glass transition temperature T g, the polymer is in the glassy state. As indicated by Stern (1990), the difference in the mechanism is reflected in the significant differences observed in the dependence of the diffusion coefficient, as well as the permeability and solubility coefficients, on the penetrant gas pressure or concentration in polymers and on the temperature. A polymer in the glassy state has a specific volume Vs bigger than the specific volume of equilibrium V). This difference, due to the non-equilibrium character of the glassy state, is at the origin of the non-linearity of the sorption isotherm (Barrer, 195 8) that is the dependence of the solubility coefficient on pressure. 22 The difference in the transport and solution behavior of gases in rubbery and glassy polymer is due to the fact that the glassy polymers are not in a state of true thermodynamic equilibrium. Glassy polymers have very long relaxation times. Therefore, the motions of whole polymer chains or portions thereof are not sufficiently rapid to completely homogenize the penetrant’s environment. Penetrant molecules can thus potentially sit in holes or irregular cavities with very different intrinsic diffusional mobilities (F risch, 1980). Thus it is possible that there are more than one mode of penetrant absorption and diffusion in glassy polymers exist (Stern, 1990). The dual mode sorption model (Barrer, 1958) is the most usually used model to describe the solution and the diffusion of molecules in glassy polymers. The gas-polymer matrix model was developed by Sefcik (1983) and Rancher (1983) to express the dependence on concentration of sorption and gas transport phenomena in glassy polymers. 1.2.5 Temperature and Concentration Dependence When the diffusant component has a molecular size much smaller than the monomer unit of a given polymer and the thermodynamic interaction between the two components is very weak, a limited rotational oscillation of only one or two monomer units would be sufficient to give a cross-section for the diffusant molecule to jump thermally from one position to a neighboring one. The diffusion of simple gases, such as hydrogen, argon, 23 nitrogen and carbon dioxide, in ordinary polymer substances and of water in hydrophobic polymers probably involves such a molecular mechanism. Molecules comparable with, or larger than, the monomer unit of a polymer require for their diffusion that a co-operative movement by the micro-Brownian motion of several monomer units, i.e. the so-called polymer segment, take place. Such substances include most organic vapors, which are either solvents or swelling agents for ordinary polymers. Many features of their diffusion reflect the micro-Brownian motion of long-chain molecules and thus are unique to polymer systems (Crank, 1968). The rate of diffusion of an organic vapor in a polymer solid is found to be primarily controlled by the mobility of the polymer segmental unit. The polymer chain segmental motion is mainly affected by two factors, temperature and concentration of sorbed penetrant within the polymer. The diffusion coefficients of polymer-organic vapor systems generally increase with increasing diffusant concentration and temperature. It is believed that raising the temperature provides energy for a general increase in segmental motion. If the energy density is sufficient, the polymer may pass through structural transitions, such as the glass and melting transitions, which further affect the solution and diffusion process (Rogers, 1985). The effects of an increase in temperature may also enhance the micro-Brownian motion of segmental units and expands the system (increase in free volume), thus the average inter-chain distance is increased and the molecular interaction between neighboring polymer molecules is weakened (Fujita, 1968). 24 The presence of sorbed penetrant also increases the free volume of the system. If the solution process is ideal, with no volume change on mixing, the change in system free volume with increasing concentration will be proportional to that obtained by an increase in temperature. However, concentration dependence may differ from the temperature due to the possibility of specific component interactions, i.e. modes of sorption, which affect component mobilities and the relative free volume contribution to the mixture by components involved in different modes of sorption (Rogers, 1985). 1.2. 6 Effect of Polymer Relaxation It should also be considered that the ease with which the diffusant molecule can jump by thermal agitations from one position to its neighbor in a polymer may also depend on the physical details at those positions. As postulated by Crank and Park (1951), when a polymer substance absorbs a vapor, the internal stresses may be set up and the polymer molecules rearrange themselves toward a new equilibrium conformation consistent with the sorbed state. This conformational change, however, may not take place instantaneously. Thus the instantaneous conformation of a polymer chain in any given volume element during a sorption process generally differs from the one when the chain was equilibrated with the particular concentration. Thus it indicates that the chain conformation is a volume 25 element as not uniquely determined by the diffusant concentration there, but may depend on the time taken by the element to reach that particular concentration. When a polymer substance absorbs a vapor, the internal stresses may be set up and the polymer molecules rearrange themselves toward a new equilibrium conformation consistent with the sorbed state. It is to be expected the inner, unswollen (or less swelling) part of the solid will exert a compression force on the outer, swelling part, while the swollen part will exert a force on the unattached region that tends to expand the solid. These compression and expansion forces change as sorption proceeds, since the concentration distribution in the solid changes with time and since the polymer chains tend to relieve these stresses by changing their conformations. Thus if the internal stresses do affect the process of diffusion, it allows D to be time-dependent (Crank, 1951). However, at temperatures well above Tg the micro-Brownian motion of polymer molecules is sufficiently active even in the undiluted state of a given polymer to enable equilibrium to be reached rapidly. Therefore, at such temperatures, the chains in any volume element of the polymer may take up almost instantaneously an equilibrium conformation consistent with the sorbed state when a vapor diffuses into the solid. With the micro-Brownian motion of polymer chains being sufficiently active, the time-dependence of D due to internal stress should disappear, since the stress set up by swelling immediately decays by a rapid chain relaxation. In this way we can understand 26 that at temperatures well above Tg the diffusion coefficient of a polymer-organic vapor system becomes free from any time-dependent effect and depends only on the diffusant concentration (Fujita, 1968). This is so called ‘Fickian’ behavior. The purely concentration-dependent D is sometimes called the equilibrium diffusion coefficient, since it is associated with the equilibrium chain conformation. In comparison, sorption and transport in polymers with long relaxation times often exhibit features which cannot be described adequately by any generalized form of Fick’s law with constant boundary conditions and with the diffusion coefficient dependent only on concentration. This so-called ‘non-Fickian’ behavior usually occurs with glassy polymers, with semicrystalline polymers above Tg, and with polymers with more rigid chain conformations and higher internal viscosity when the penetrant swells the polymer. In such cases, D may be a function of concentration, time, the spatial coordinates and history of the sample. The boundary conditions for diffusion depend on time and other variables such that the surface concentrations change with time, applied stress, etc (Rogers, 1985) The type of behavior is characterized experimentally by the concentration gradient profile during sorption and its time dependence. The time dependence is conveniently determined by the slope n of a plot of log M against log time, where M is the amount of penetrant sorbed. In a sheet geometry, Fickian sorption has n = 1/2. Sorption with n between 1/2 and 1 is non-Fickian diffusion (Rogers, 1985) 27 The relaxation times in polymers are related to the many modes of relaxation. All relaxation times decrease with increasing temperature or concentration, with some decreasing more rapidly than others. Rogers (1985) believed that the overall sorption process reflects those relaxation motions of the polymeric matrix which occur on a time scale comparable to or greater than the time scale of the concurrent diffusion process. 1.3 Parameters Affecting the Transport Properties The barrier properties of polymeric materials are determined by the chemical structure of the chain and the polymer morphology. The parameters derived from chemical structure, such as degree of polarity, intermolecular forces, ability to crystallize and chain stiffness, are essentially determined upon the selection of the particular polymer. Generally, orientation and crystallization of polymers improves the barrier properties of the material as a result of the increased packaging efficiency of the polymer chains. As postulated by Michaels et al. (Michaels, 1959; Michaels and Bixler, 1961a, b), the concept that the sorption and diffusion took place exclusively in the amorphous regions has been widely accepted. The crystalline zones act as excluded volumes for the sorption process and are impermeable barriers for the diffusion process. The dispersed crystalline phase presents a resistance to the permeant passage. Moreover, their existence does not seem to influence the sorption mode in the amorphous phase. More exactly, Klopffer 28 (2001) indicated that these crystalline zones have two effects on gas diffusion. On one hand, they increase the effective path length of diffusion; and on the other hand, they seem to reduce the polymer chains mobility in the amorphous phase because the chain ends are trapped in the neighbouring crystalline lamellae, and then lend to a higher activation energy of diffusion. Michaels et a1 (1959, 1961a,b) thus introduced a “tortuosity factor” 1 and a “chain immobilization factor” ,8 to account for these effects. They proposed the following expressions for the coefficients of solubility and diffusion: S=S*m an D * D = __ 23 fit ( ) where S* and D* are the coefficients of solubility and diffusion in an amorphous, hypothetical , completely relaxed state (i. e. a completely amorphous polymer), and (to , the volume fraction of amorphous phase. ,6, the factor relating to chain immobilization, reflects the hindrance of the crystalline zones on the amorphous zone. The tortuosity factor I, characterizes the more tortuous pathway that a diffusing molecule must take in a semicrystalline zones. It is a geometrical term which depends on the crystallites anisotropy degree, the degree of crystallinity and hence, the thermal history. For a constant volume fraction of amorphous phase, fl is correlated to temperature by an exponential relation whereas 1.’ is constant. fl and D* are both supposed to depend on the penetrant diameter (Klopffer, 2001). 29 Certain evidence has shown that the crystalline phase affects the nature of the amorphous phase (Michaels and Bixler, 1961b). More recent studies (Puleo, 1988; Mogri and Paul, 2001; Weinkauf, 1990) have shown that certain crystalline structures will permit small molecule sorption and diffirsion. The results can be explained by a more open structure of these crystals with a density not very different from that of the amorphous phase. Recently, it was also shown that molecules could penetrate into crystals of 6 form in syndiotactic polystyrene (Guadagno,1998; Manfredi, 1997). In the case of amorphous or semicrystalline polymers, the coefficients S and D can be related to the free volume (Vf), which is defined as a space which is not occupied by the macromolecules and its value is equal to a difference between the specific and the occupied volumes of the polymeric system (Danch, 2003). Fractional free volume, FFV [cm3 of free volume/cm3 of polymer] is commonly used to characterize the efficiency of chain packing and the amount of free space available for gas permeation in a polymer matrix (Pixton, 1994; Bondar, 1999; Raharjo, 2005): FFV = V3,. —v0 z 1/p-1.3vW vs}, 1/ p (24) where v,p is the polymer specific volume, and v0 is the occupied volume of the polymer. The occupied volume is typically estimated as 1.3 times the van der Waals volume (vw), which is calculated using group contribution methods (Krevelen, 1997). For an approximate calculation, the Van der Waals volume is assumed to be bounded by the outer surface of a number if interpenetrating spheres. The radii of the spheres are 30 assumed to be (constant) atomic radii for the elements involved and the distances between the centers of the spheres are the (constant) bond distance. The contribution of a given atom A with radius R to the Van der Waals volume is then given by: 4 . vW,A = NA|:§7[R3 _Z7d712 (R ‘%‘)] (25) 2 2 h, ail—+5— (26) 2 21 21. where NA is Avogadro’s number, r, is the radius of atom i covalently bonded to A, and Ii is the bond distance between the atom A and i (Krevelen, 1997). 1.4 Methods for Studying the Sorption of Organics on Polymer Materials Mass transfer of inert gases and water vapor has been well investigated and standardized methods are available for determining transmission rates for these permeants (ASTM E96-66, ASTM D3985-81) (Hernandez, 1986a). In contrast, current knowledge on mass transfer of organic substances is still limited and no standard procedures are available to determine the numerical values of these parameters (Hernandez, 1986a; Zhou, 2004). This is mainly due to the fact that organic substances are capable of interacting with the polymer matrix, leading to swelling of the polymer matrix, resulting in configurational changes and alteration of polymer chain conformational mobility and thus of penetrant diffusivity and permeability. Thus, for penetrants such as organic vapors which can exhibit physicochemical interactions with a 31 polymer matrix, the solubility coefficient (S), diffusion coefficient (D), and permeability coefficient (P) should be determined experimentally in order to describe accurately the mass transfer behavior of penetrant / barrier systems involving organic vapors. Study of the sorption of organics can be categorized to two branches: by polymer film / sheet, which is the most popular way researchers are adopting, and by container systems. Currently several methods have been used for measuring the mass transfer characteristics of polymer films or sheets, including isostatic permeation procedure (Zhou, 2004; Gavara,1996; Van Willige, 2002; Huang,l998), quasi-isostatic permeability method (Gavara,1996; Sajiki,l993); Hernandez,1986), thermal stripping / thermal desorption (TS/T D) procedure (Nielsent, 1994) and gravimetric technique (Baner,l986; Nielsent,l994; Barr, 2000), etc. The studies of mass transfer in container systems were only reported by using sorption cell methods where the permeant was dissolved or suspended in a liquid and brought into contact with polymer (Safa,1999; Demertzis,l997). In that case, components such as aqueous media, or solvents used to disperse organic compounds may affect the partitioning. Among these methods, the gravimetric approach is very suitable and classic for the determination of solubility coefficient (S). The advantages of the gravimetric method are not only the simplicity in the manipulation of the polymer sample, but also continuously recording the weight gain by the test specimen as a function of time. The gravimetric method permits accurate and direct measure of the sorbate uptake at both transient (M) and steady state (M00). The M... value obtained at steady state can be used to 32 calculate S and M and M... values can be used to calculate D (N ielsent, 1994; Barr, 2000). So far, several models of balance used in the gravimetric study of the sorption have been reported, such as Cahn 2000 suspension electrobalance (Barr,2000; Berens, 1989; McDowell, 1999), quartz crystal microbalance (QCM) (Oliverira, 2004), Rubotherm magnetic suspension balance (Kiparissides, 2003; Kruger, 2005), quartz-spring balance (Kriiger, 2005), McBain spring balance (Dhoot, 2001), or Ohaus analytical plus electronic balance (Palamara, 2005). The main difficulty of using conventional gravimetric instruments is the direct connection of the measuring cell (sample atmosphere) and the weighing instrument. The balance can be damaged or disturbed by the measuring atmosphere and the measuring atmosphere can be adversely affected by flushing gases and pollution. Another limitation of those conventional gravimetric instruments is the limitation on the sample shape and size, which are determined by the size of the chamber in the balance. Normally only a small amount of sample in the shape of powder, pellet or film could be tested. A SGA-lOOR gravimetric analyzer that includes a Rubotherm microbalance has been modified to make it possible to weigh samples contactlessly under nearly all environments and to facilitate the test on large sample, such as a container (Qin, 2005). This instrument was used in this project. 33 1.5 Mass Transfer Studies on Polypropylene Several experimental techniques have been developed to measure gas transport in polyolefins at high pressures, especially polypropylene (PP) and polyethylene (PE). PP is a commodity polymer produced on a 10 billion pound per year scale in North America alone (Palamara, 2005). Sato (1999) measured Henry’s constants for carbon dioxide and nitrogen in molten PP. Sato (2001) studied the solubility of propylene in semicrystalline PP. Tsuboi (2001) reported infinitely dilute partition coefficients for ethylene and propylene in semicrystalline PP. Sliepcevich (2000) used packed bed inverse gas chromatography to obtain partition and diffusion coefficients for ethylene and propylene in semicrystalline PP beads. Sorption of various organic solvents in atactic PP and isotactic PP was studied by Ochiai (1971) by using a quartz-spring balance. Palamara (2005) studied the diffusion and solubility of propylene and ethylene in atactic PP with the static sorption technique. The permeation of oxygen, water and limonene through printed and unprinted oriented PP films were studied and compared by Rubino (2001). Although many experimental techniques have been developed to measure mass transfer in polymers, experimental diffusivity and solubility data are still in short supply for many PP-organic systems, such as PP-benzaldehyde. 34 1.6 Objectives of the Project Much of the current permeability data of organic compounds has been obtained only for flat films or sheets, and these values are then used to calculate the barrier properties of containers (Hernandez, 1992; Thalmann, 1990; Nir, 1996). In general, the forming process of a polymer such as blow molding, injection blow molding, thermoforming, by which containers are produced, changes the polymer structure and morphology. It induces localized effect and stresses in the final container that may produce variations in the barrier properties of containers compared with film or sheet (Jasse, 1994; Demertzis, 1997). Therefore, there are substantial differences in the barrier properties between containers and unfonned polymer sheets. Using the permeability data generated from films and sheets for containers will cause some error in the prediction of the container barrier properties. In this project, polypropylene sheets were manufactured and container shapes were produced by casting and thermoforming procedures, individually (Figure 1.2). The goal was to develop a method to study the sorption of organic vapor in polymers. The SGAIOO R gravimetric analyzer was adopted in the study due to the long-term stability and high accuracy benefited from the separation of the balance magnet from the measuring cell. The sorption study was applied on PP resin, sheet and container shape. The correlation between sorption of organic compound and package configurations was studied. The effect of the manufacturing of polymer on sorption was also addressed. 35 Container Resin Sheet ,o. - 0 r:_—_-_—_,-'> I::> . . ' Casting Thermoforming Figure 1.2. Schematic plan of the PP processing in this project. The specific objectives for this project are: To develop a continuous gravimetric method to study the sorption of benzaldehyde vapor in polypropylene via Rubotherm magnetic suspension microbalance To demonstrate the impact of manufacturing on the sorption of benzaldehyde by comparing the sorption of resin, sheet and container To demonstrate the impact of morphology on the sorption of benzaldehyde on resin, sheet, container and atatic PP. 36 CHAPTER 2. MATERIALS, INSTRUMENT AND METHODS 2.1 Materials 2.1.1 Organic Sorbate Benzaldehyde, purified by redistillation, 2 99.5%, was obtained from Sigma-Aldrich Inc. (St. Louis, MO) and adopted as sorbate / permeant in this study. Benzaldehyde has been well identified as a natural flavor compound. It is the character impact flavor for cherry and almonds. It also helps give coffee and cocoa their characteristically pleasant aromas. Benzaldehyde is also widely used in the chemical industry, especially as starting material. Besides, benzaldehyde is safe and remarkably cheap—about 30¢ an ounce at chemical supply houses. Benzaldehyde is easily oxidized. The oxidized product is benzoic acid in white solid. Benzaldehyde is also slightly sensitive to light and moisture. The operation of this liquid should be in the inert protection environment and avoid the long time light exposure (Wiberg, 1955; Hurd, 1929; .Potineni, 2005). 2.1.2 Polymer Polypropylene (PP) resin pellets were used as the starting material in this study. PP resin was obtained from ExxonMobil chemical company (Huston, TX). It is a homopolymer produced with a Ziegler-Nana type catalyst in series bulk stirred tank reactors. It has density of 0.900 g/cm3 and total amount of additives less than 0.3 wt.%. 37 The molecular weight of PP resin by gel permeation chromatography (GPC) has been measured as an Mn of 60,000 g/mol and Mw of 250,000 g/mol, and an Mw/Mn ratio of 4.2. The resin has melt flow rate of between 2.5-3.1 g/ 10 min and an average degree of crystallinity of 39.4 % as determined by differential scanning calorimetry (DSC). The melting point measured by DSC was about 164.9 °C. Atatic polypropylene (PP), a totally amorphous waxy solid with specific gravity 0.85 g/cm3 and Mw of 12,000 g/mol, was obtained from Scientific Polymer products, Inc. (Ontario, NY). Polypropylene was chosen because it is a commodity polymer and has been commercialized in the market for long time. PP is not a very good organic barrier, but it is useful for demonstrating the correlation between configuration and sorption behavior of a PP container shape in limited time range under the selected conditions. 2.2 Polymer Processing 2. 2.1 Casting PP Sheet Polypropylene resin pellets were considered as the starting material in this study. It was then formed to PP sheets in the thickness about 0.46 mm (~ 18 mil) and width of 15 cm (~ 6 in) by extrusion and casting from PP resin by a Killion KLBlOO extruder (Killion Extruder Inc., Cedar Grove, NJ). Casting film is normally homogenous in thickness. The setting of the extruder was: 38 Temp. of zone lSt to zone 3rd Temp. of zone 4th (die) Melt temp. Motor speed Chill roller temp. Chill roller speed Die gap 2. 2.2 T hermoforming PP Containers 530°F 540°F 486°F rlzrrpsr 100°F (issrrpnr 1.79 mm (70 mil) PP polyhedron container shape with surface area (36.3 cmz) was designed. The polyhedron shape represented those containers that have lots of comers and edges. A male mold were chosen for producing the container shape over female mold because male mold can produce shapes with smaller thickness variance, thus improving the thickness uniformity. The mold was designed usingy Rhino software, it is shown in Figure 2.1 left; the thermoformed polyhedron container shape is shown in Figure 2.1 right. The polypropylene container shape was thus thermoformed on a Hydro-Trim 1620 thermoformer (Hydro-Trim Corp., West Nyak, NY) using the male mold. The setting of the thermoformer is: Heating time Forming time Top temp. controller Bottom temp. controller 40 sec 25 sec 410 °F 410 °F 39 . 1%". . ~ ' Figure 2.1. Male mold (left) and thermoformed polyhedron shape (right). 2.3 Instrument 2.3.1 SGA-100R GravimetricAnalyzer Sorption studies were conducted on SGA-100R gravimetric analyzer including Rubotherm magnetic suspension electrobalance (V TI Corporation, Scientific Instruments, Hialeah, Florida) by the continuous flow method (Hernandez, 1986b; Nielsent, 1994). The balance is optimized to have up to 10 g capacity with 1 ug resolution and it has been designed with a chamber in the dimension of 7.5cmX7.5cm><14cm (LXWXH), which is large enough to hold a whole plastic container. The vapor activity (VA) / relative humidity (RH) can be generated and constantly maintained. Because there are three vapor generators in this system, it is possible to produce mixed sorption conditions. The balance is designed to have long-term stability and high accuracy due to the separation of the balance magne and the measuring cell. The detailed instrument information is available in Chapter 3. 40 2.3.2 Operation Procedure Prior to the sorption tests, the polymer sample, either PP resin pellets (approximately 120 mg), or a PP sheet sample in the dimension of 1" Length x 0.75" Width x 0.02" Thickness, or a PP polyhedron container shape was dried by a vacuum oven (40 °C, 30 in vacuum Hg) for 10 hours. After drying the sample was placed on the microbalance basket/hook within the chamber. A complete sorption experiment consisted of a series of runs including the specified conditions for sample drying (optional) and a sequence of vapor activities / relative hurnidities to which the sample was exposed. Vapor sorption was measured by continuously recording the weight uptake of the dry sample. In this study, the pre-dried sample was purged by N2 for 2 hours in the balance chamber in the drying cycle. Five vapor activities (0.1, 0.3, 0.5, 0.7 and 0.9) were tested in the sorption study. 2. 3.3 Differential Scanning Calorimetry (DSC) Analysis TA instrument DSC Q100 was used in this project. DSC was adopted to determine the glass transition temperature (Tg), melting point temperature (T...) and the heat of melting (AH), thus % crystallinity of polymer material. The DSC conditions were as follows: The method used was heat/cool/heat; N2 carrier gas at 50 ml/min. The temperature was first raised at a rate of 20 °C/min to 200 °C and followed by cooling at a rate of 5 °C/min to -80 °C. It was then raised at a rate of 20 °C/min to 200 °C. 41 2. 3.4 Gas chromatography (GC) The composition of the vapor stream generated by gravimetric system was monitored by using GC (HP 6890 series) equipped with a gas flame ionization detector (FID), interfaced with an integrator. The GC conditions were as follows: The column was HP-S; H, carrier gas at 2.2 ml/min; H2 at 30 ml/min; air flow at 300 ml/min; N2 at 25 ml/min; initial oven temperature, 40 °C for 5 min; temperature was raised at a rate of 20 °C/min to 140 °C and kept at this temperature for 1 min. The retention time for benzaldehyde is at ‘ 8.49 min. The calibration samples were prepared by putting 10 ml of benzaldehyde in a 40 ml hermetic vial under nitrogen atmosphere at 40 °C for 3 hours. Samples in the vial headspace (10-70 ul) were injected into the GC port. The calibration curve was thus constructed by plotting the instrument response (area) vs. sorbate quantity (mol). 2.3.5 Polymer Density Test The test procedure followed was in accordance with the ASTM D792-98 “Standard Test Methods for Density and Specific Gravity (Relative Density) of Plastics by Displacement”. Specific gravity (relative density) is the ratio of the mass of a unit volume of the impermeable portion of the material in air at 23 °C to the mass of equal density of an equal volume of gas-free distilled water in air at the same temperature. Density is the mass in air in kilograms per cubic meter of impermeable portion of the material at 23 °C. Specific gravity can be converted to density by use of the following equation: 42 023°C, kg/m3 = sp gr 23/23 °C * 997.6 Test Procedure A (for testing solid plastics in water, specimens 1 to 50g) was adopted in this project. First, the test sample was conditioned at 23 °C :1: 2 °C and 50 i 5 % relative humidity for not less than 40 h prior to test. Second, the water temperature was measured and the sample was weighed in air as a. Third, the immersion vessel was mounted on the support, and the suspended sample and sinkers were completely immersed in water at a temperature of 23 °C :1: 2 °C. The vessel must not touch wire or specimen. Any bubbles adhering to the object were removed. The mass of the suspended sample was determined as b (the mass of the specimen, sinker and the partially immersed wire in liquid). Fourth, the wire and sinker were weighed in water with immersion to the same depth as used in the previous step, this weight was recorded as w (mass of the wire in liquid). The procedure was repeated for each sample. Three samples per type of polymer configuration were measured. The specific gravity of the plastics was calculated as follows: Spgr23/23 °C=a/(a+w-b) The density of the plastic can thus be calculated as follows: 133°C, kg/m3 = Sp gr 23/23 °C * 997.6 In summary, the mass of the solid plastic sample was determined in air. It was then immersed in a liquid, its apparent mass upon immersion was determined, and its specific gravity (relative density) was calculated. 43 The polymer characteristics by DSC and polymer density test are shown in Table 2.1: Table 2.1. Polymer characteristics for different PP forms PP resin PP sheet PP polyhedron container Tg, °C -6.41 -6.13 «7.42 Tm, °C 164.9 160.8 166.7 Crystallinity, % 39.4 40.5 42.3 Density, g/cm3 0.90“ 0.87 0.85 ‘ Obtained from Exxonmobil Chemical Company (Houston, Texas) 44 CHAPTER 3 Measurement of Sorption of Benzaldehyde in Polymer with Magnetic Suspension Electrobalance This chapter is preparing for the submission to Polymer Testing in the section of Equipment and Test method. 45 Measurement of sorption of benzaldehyde in polymer with magnetic suspension electrobalance Abstract A new gravimetric analyzer including a magnetic suspension microbalance was set up and a method was developed to study the sorption of organic and/or moisture in polymer by a continuous gravimetric method. The balance has up to 10 g capacity with 1 pg resolution, and a large chamber with dimension of 7.5cmX7.5cmX14cm (LXWXH). The separation of the balance magnetic and the sample chamber improved the balance stability and accuracy. The gravimetric instrument was calibrated and the noise was quantified before the sorption study. The balance was used to evaluate the sorption of benzaldehyde vapor in polypropylene (PP). The sorption of benzaldehyde in rubbery PP resin exhibited Fickian sorption kinetics at low vapor activities (OJ-0.5) and a two-stage sorption incorporated both Fickian diffusion at the initial sorption stage and protracted polymer relaxation controlled mass uptake kinetics at high vapor activities (0.7-0.9). The sorption behavior of benzaldehyde in different PP forms (resin, sheet and container) was compared. Keywords: gravimetric, sorption, benzaldehyde, polypropylene 46 3.1. Introduction Mass transfer properties have been applied in gas separation membrane industry (Kesting, 1993), in which the economics is largely determined by the membrane’s transport properties - permeability and selectivity for a specific gas in a mixture. In addition, there is an application of the mass transfer and sorption in packaging industry (William, 1990; Risch, 1991), in which high moisture, gas or organic barrier properties of polymeric packaging materials are always expected. Several methods have been used for measuring the mass transfer characteristics of polymer films, including isostatic permeation procedure (Gavara, 1996; Huang, 1998), quasi-isostatic permeability method (Gavara, 1996; Hernandez, 1986 b), thermal stripping / thermal desorption (TS/TD) procedure (N ielsent, 1994) and gravimetric technique (Barr, 2000; Krtiger, 2005). Among them, the gravimetric method was frequently used in the direct determination of permeant solubility and diffusion coefficients. The advantages of gravimetric method are due to its simplicity on the polymer sample manipulation and the continuously recording of the weight gain by the test polymer as a function of time or vapor activity. The gravimetric method permits accurate and direct measure of the sorbate uptake at both transient (M) and steady state (M00). The M... value obtained at steady state can be used to calculate solubility coefficient (S) (N ielsent, 1994; Barr, 2000). 47 Several gravimetric instruments with different balance system have been used to carry out gravimetric study on polymer materials for the determination of sorption of different organic compounds or moisture, such as Cahn 2000 electrobalance (Barr, 2000; Berens, 1989; McDowell, 1999), quartz crystal microbalance (QCM) (Oliverira, 2004), Rubotherm magnetic suspension electrobalance (Kiparissides, 2003; Kruger, 2005), quartz-spring balance (Ochiai, 1971), McBain spring balance (Dhoot, 2001), and Ohaus analytical plus electronic balance (Palamara, 2005). Some of these gravimetric systems were incorporated with vacuum pump for polymer degassing and penetrant removal (Berens, 1989; Dhoot, 2001). The main difficulties in using some of the gravimetric instruments consist in the direct connection of the sample atmosphere (sample in the chamber) and the weighing mechanism. The fact is that if the weighing mechanism is connected to the sample atmosphere it could be damaged or disturbed by the measuring process. Another limitation of those conventional gravimetric instruments is in the small size of the sample chamber and limited sample weight that some of those instrument can weigh, which reduce the field of applications since only small amount of powder, granules or fihns could fit the sample chamber. Our project called for a gravimetric system which can meet our requirements, which includes long-term stability and high accuracy for testing of the high barrier materials; a large chamber to hold different shapes of sample (i.e. containers); various sorption 48 conditions, such as generation of moisture and organic vapor, or a mixture of multiple organic vapors simultaneously. The gravimetric system used in this work was customized to meet our requirement. It has a big sample chamber, a large weighing range (up to 10 g), and the separation of sample chamber and weighing mechanism. Furthermore, this electrobalance can re-zero throughout the measuring process to reduce the drifting and noise. Thus it is necessary to address each part of this gravimetric system in detail. The objectives of this work were to quantify the noise level of the new SGA-lOOR gravimetric analyzer and to develop a methodology to evaluate the sorption of organic and/or moisture vapor by continuously gravimetric method using SGA-100R. For this initial study, the sorption of benzaldehyde vapor on polypropylene resin, sheet and thermoformed sheet into a polyhedron shape have been considered. 3.2. Experimental 3. 2.1 Gravimetric Instrument Sorption studies were conducted on a SGA-lOOR gravimetric analyzer (VTI Corporation, Scientific Instruments, Hialeah, Florida) by the continuous flow method (Hernandez, 1986b; Nielsent, 1994). The gravimetric instrument includes three separate zones: Zone one is the weighing mechanism (microbalance head), zone two is the sample chamber, and zone three includes the vapor generators. Each zone includes its own 49 temperature control system in order to maintain instrument stability and overcome any environmental changes. The flow schematic of the gravimetric system is shown in Figure 3.1. Zone one — Rubotherm electrobalance The core portion of this gravimetric system, the Rubotherm electromagnetic suspension microbalance, is located in zone one. Instead of hanging directly at the balance, the sample to be investigated is linked to a suspension magnet, which consists of a permanent magnet, a sensor core, and a device for decoupling the measuring sample. The electromagnet, which is located at the bottom of the balance head in zone 1, maintains a freely suspended state of the suspension magnet via an electronic control unit. Using this magnetic suspension coupling the measuring force is transmitted contactlessly from the measuring chamber to the microbalance, which is located outside the chamber under ambient atmospheric conditions. The measuring load can be automatically decoupled, as shown in Figure 3.1 (bottom). The load decoupling is carried out by lowering the suspension magnet in a controlled way to a zero point position. This position where only the weight of the suspension magnet is transmitted to the balance, corresponds to an empty balance pan in a normal weighing procedure. Now the balance can be tared and calibrated even when recording measurements under process conditions. This feature increases the measuring accuracy, particularly in the case of long-term measurements. 50 * L. .1 control system F— E Zone one ‘—— . . _ _ set point controller PID controller h T -——>[pos ition transducer i . - Zone three vent V3 V2 V1 13.5 Zone two J_ OFF Tarlng and Calibration Measuring Figure 3.1. Schematic of gas flow in the Rubotherm SGA-IOOR gravimetric analyzer (top) and the coupling/decoupling action between the electromagnet and suspension magnet under different operation conditions (bottom). 51 Zone two - Sample chamber The chamber has been designed in the dimension of 7.5cmX7.5cm><7.5cm>< 14cm (LXWXH), which is large enough to hold a thermoformed sheet / plastic container sample; also, the vapor activity (VA) / relative humidity (RH) is generated and constantly maintained by vaporizer / humidifier. Also there are three vapor generators in this system, which allow the generation of mixed gas atmosphere within the sample chamber; the data is continuously recorded by setting at certain time interval. Besides, the balance is designed to have long-term stability due to the separation of the measuring cell and the weighing system. The automatic decoupling of the measuring load / rezero throughout the testing process increases the measuring accuracy particularly in the long term measurements. Further description of the electrobalance, including instrument details, instrument noise and drifting analyses under different conditions can be found elsewhere (Qin, 2006). 4. 2.4 Experimental Procedure The sorption data was compensated from the drifting of the balance at 25 °C. The balance was calibrated before each experiment and during the measurements and the balance drift was corrected every 10 minutes. These were done to correct for buoyancy 76 on the suspension magnet and any inner parts of the magnetic coupling which were lifted together with the sample. Prior to the sorption tests, the polymer sample was dried in a vacuum oven at 40 °C for 10 hours. It was then purged with N2 for 2 hours in the balance chamber. The sorption tests were conducted at five benzaldehyde vapor activities (0.1, 0.3, 0.5, 0.7 and 0.9). The weight gain of the polymer sample was continuously monitored and recorded at 25°C and 0% relative humidity. The balance equilibrium criterion was the weight change below 0.0001 % in 20 mins. Experiments were continued until the sorption equilibrium was reached or 10,000 minutes had been passed. 4.3 Results and Discussion 4. 3. 1. Impact of Surface Area and Bulk Properties on the Sorption of Benzaldehyde in PP Resin Sorption of organics in polymer films had been studied thoroughly. However, the sorption behavior in polymer resin pellets has not yet been completely elucidated. The sorption took place in the rubbery PP resin pellets at 25 °C, which is well above the glass transition temperature of PP (T8 of -6.4 °C). Unlike what was expected for the sorption of benzaldehyde on rubbery polymer, the PP resin did not just exhibit a simple Fickian kinetics. In our previous work (Qin, 2006), as showed in Figure 4.2, it has been found that at high vapor activities (0.7 and 0.9), the sorption curve exhibited a linear initial slope 77 and it bent to the first equilibrium mass uptake at 1300 mins. At extended time, the mass uptake exhibited a protracted, non-Fickian behavior. Such phenomenon is so called two-stage kinetics and is often observed for organic vapor sorption in glassy polymers (McDowell, 1999; Rossi, 1993). However, it has been shown by Rossi and Mazich (1993) that non-Fickian sorption curves can be obtained for solvent diffusion in rubbery polymers under certain conditions when a spherical geometry was used and the swelling of the polymer and the existence of a moving boundary were properly taken into account. 0.4 . 0.1 Wt. change, °/o d). -1 . . . , J 0 2000 4000 6000 8000 l 0000 Time, min Figure 4.2. Sorption curve for benzaldehyde vapor in PP resin at vapor activity of 0.7 at 25 °C. The black dots are the experimental data, the arrow shows the first Fickan region. Sorption of benzaldehyde vapor in isostatic PP resin pellets was first assessed by studying the sorption in relatively small amount of resin pellets (0.15g), large amount of 78 resin pellets (2.0 g), resin slices (cut from resin pellet), and in atatic PP (totally amorphous). The result is shown in Figure 4.3. 2.000 1 a , , Atatic PP 1.600 ~ +Resin slices e\° 1.200 . +Resin pellets E” (2.0g) , -‘— Resin pellets is 0.800 - : (0.15g) 3 0.400 1 0.000 r I r I ‘ ‘ 0.0 2000.0 4000.0 6000.0 8000.0 10000.012000.014000.0 Time, min Figure 4.3. Sorption curves for benzaldehyde vapor in PP at vapor activity of 0.5 at 25 °C. From bottom to top: sorption in a) resin pellets (0.15g), b) resin pellets (2.0g), c) resin slices and d) atatic PP. The sorption curves a) and b) are the sorption of benzaldehyde in PP resin pellets but in different starting dry weight. The pellets in these tests came from the same batch, they had an average diameter of 0.4 cm. Based on the sorption result of curve a and b, we found that the uptake of benzaldehyde vapor in a different amount of polymer resin had a similar result (weight change is 0.130 % for curve a and 0.131 % for curve b). Thus we conclude that the sorption result does not depend on the polymer starting dry weight. 79 The PP resin pellets were further processed for the purpose of testing the surface effect on the sorption. Each resin pellet was cut into 8 slices. Sorption curve c) in Figure 4.3 represents the sorption in PP resin slices in the similar weight of that in resin pellets (curve a). The result showed that the sorption of benzaldehyde in whole resin pellets (curve a) and in resin slices (curve c) are 0.13 % and 0.50 %, individually. The extra weight gain of resin slices was benefited from the larger surface area compared with whole resin pellets. This study clearly demonstrated the surface area (surface morphology) does affect the sorption. It encouraged us to further explore the other morphology factors control the sorption behavior. The atatic PP (100% amorphous) was thus studied in the sorption experiment. The sorption result of benzaldehyde in atatic PP (curve (1) was compared with that in resin pellets (curve a) in Figure 4.3. Very clear sorption difference between atatic PP (1.90 %) and isostatic PP resin pellets (0.13 %) were observed in the tests. PP resin pellets are semi-crystalline material with crystallinity of 39.4 % and density of 0.90 g/cm3. The atatic PP has 0 % of crystalline and density of 0.85 g/cm3. Since sorption and diffilsion took place exclusively in the amorphous regions and the crystalline zones are impermeable barriers for the diffusion process, it is predictable that atatic PP absorbs more than that of semicrystalline resins, as the sorption result shown in Figure 4.3. As a result, it was found that the surface area as well as morphology of the bulk of the polymer material had major impact on the sorption process. 80 4.3.2 Sorption of Benzaldeh yde in Different PP shapes Figure 4.4 shows the sorption of benzaldehyde by PP resin, sheet, thermoformed sheet and atatic PP studied by continuously recording the weight gain of polymer when exposed to benzaldehyde vapor at vapor activity of 0.3 at 25 °C. Based on the sorption curves in Figure 4.4, we found that the thermoformed sheet gained more weight (0.489 %) than PP sheet (0.271 %) and resin (0.077 %) at same vapor pressure. This demonstrates the fact that any processing, to which the resin was subject, caused an increase of the sorption of benzaldehyde. 1.2 _ + Atatic PP —'- Thermoforrned °\° sheet 55 Sheet :1 a e .5 --~-- Resrn B l 0 4000 8000 12000 16000 Time, min Figure 4.4. Sorption curves for benzaldehyde vapor in PP with vapor activity of 0.3 at 25 °C. From bottom to top: sorption in a) resin pellets, b) sheet, 0) thermoformed sheet and d) atatic PP. Normally when polymer is processed, the internal stresses will be set up and the heat treatment during the processing imparts the molecular chains energy and mobility to rearrange themselves towards the conformational changes. The polymer morphology, 81 including crystallinity, crystal size, the distance between the larnellas, free volume and surface contour will all be changed according to the processing conditions. As mentioned before that sorption behavior of polymer is closely related to the polymer morphology, thus the polymer processing, which causes morphology changes, should cause different sorption behavior for different therrnoprocessed polymers. The processing effect on sorption is clearly demonstrated in Figure 4.4. Furthermore, the sorption of benzaldehyde in atatic PP was used as a reference to compare with that in other PP forms (resin, sheet and thermoformed sheet). The high sorption ability of atatic PP (0.989 %) supports the morphology effect in the organic sorption in polymer. 4. 3.3 Impact of Vapor Activity of Benzaldeh yde on the Sorption Behavior The processing effect on the sorption of benzaldehyde in PP has been demonstrated in Fig.4.4 at vapor activity of 0.3. The similar processing effect was also observed in vapor activity of 0.5 and 0.9. The comparison is shown in Figure 4.5. It is found that for each PP form (resin, sheet, thermoformed sheet or atatic PP), the sorption capability increases as the organic vapor pressure increases. Besides, the higher the VA, the larger the sorption difference was observed among the different PP forms. 82 VA - 0.9 VA = 0.5 3.0 4: .\° -‘- Thermoformed an 2.0 4 sheet .5 ' Sheet ; 10 . . ,5.” * Resm 0.0 3L . . 0 5000 10000 10000 Time, min Time, min Figure 4.5. Sorption curves for benzaldehyde vapor in PP at 25 °C at vapor activity of 0.5 (left) and 0.9 (right). From bottom to top: PP resin, sheet, thermoformed sheet and atatic PP. 4. 3. 4. Benzaldehyde Sorption Isotherm The gravimetric sorption experiments were used to study the sorption of benzaldehyde on PP resin at 25 °C till reaching equilibrium. The position and shape of specific isotherms were determined by the polymer-penetrant interaction parameter and the glass composition of the system. The sorption isotherm for benzaldehyde - rubbery PP resin at 25 °C is shown in Figure 4.6. 83 ,- 0.4 ~ / e\“‘ a . 0.3 ~ ’ .S I ’ a I OD I «3 0.2 .. 3 ’ r 0.1 ~ I ' - y , ’I 0 t t l 0 0.2 0.4 0.6 0.8 1 Vapor Activity [PIP 01 Figure 4.6. Sorption isotherm of benzaldehyde by PP resin at 25 °C. The isolated squares are experimental data; the dashed line is the trend line based on experimental data The isotherm shape in Fig.6 is typical of that for isotherm of organic vapors in rubbery polymers. Thus, the Flory-Huggins equation (1) was used to describe the sorption behavior (Berens, 1989). 1n(P/P.)=ln(K)+(1—K)+z(1-Vl)2 (1) where x is the interaction parameter, V1 is penetrant volume fraction. P and P0 are actual and saturated vapor pressure at certain temperature, respectively. The percent weight gain, (%Wt. Gain), is related to the volume fractions of solvent and polymer, V1 and V2, respectively, by Vldl °/ Wt.Gain = 100- ° Vzdz (2) where d1 and d2 are the solvent and polymer densities. 84 Interaction parameter 1 from Hansen Solubility Parameters (HSP) The F lory-Huggins interaction parameter x, has been used for many years in connection with polymer solution behavior, with the 302 parameter derived from the “New Flory Theory” being currently accepted for general use instead of the older x. It would be desirable to relate the widely used HSP more directly to 302. This would allow estimates of X12 for systems where the HSP are known, but x12 is not (Hansen, 1999). Patterson and co-workers have shown that X12 can be calculated (Biros, 1971). A], = [(6,2 —5,,,)2 + 0.2505,.2 —5,.,)’ + 025(6”, —a,,,)"] (3) and X12 is estimated from: 112 = VA” (4) V where (SD, 6;» and 6” are the Hansen Solubility Parameters representing the contributions from the dispersion interaction, polar interaction, and hydrogen bonding, respectively; and the subscript l and 2 are for solvent and polymer, respectively. V is molecular volume of solvent. R is the gas constant (m3 Pa)/(mol K). T is temperature in Kelvin. The Hansen Solubility Parameters for benzaldehyde and PP are shown in Table 4.1. The 702 is calculated as 1.29 based on the equation (3) and (4) for benzaldehyde and polypropylene pair at 25 °C. Combined the x value with the calculated solubility parameter of benzaldehyde and PP by group contribution method A6T = 93.7 (A6, = [(502 —6,,,)2 + (6,,2 —6,.,)2 +(6H2 —6,,,)2]”2 (Caner, 2004), we can characterize benzaldehyde as a weak swelling agent to PP. 85 Table 4.1. Hansen Solubility Parameters Data for Benzaldehyde and Polypropylene at 25 °C 60 6P 6“ Benzaldehyde l 9.4 7.4 5 .3 Polypropylene l 6. 1 0 0 The experimental equilibrium sorption data for benzaldehyde were compared with the Flory-Huggins equation with 302 = 1.29. It was found that sorption data could not fit well into the Flory-Huggins equation at vapor activity of 0.1 to 0.9 with this x value. As far as the authors’ knowledge, solubility in PP resin as a function of solvent activity for benzaldehyde has not been reported in the literature. Our result suggests that the Flory-Huggins equation, with interaction parameters determined from equilibrium sorption of the pure liquids, could not provide sufficient estimates of sorption isotherms for benzaldehyde vapor — PP resin system. The Flory-Huggins theory is useful in considering the thermodynamics of dilute polymer solutions. The interaction parameter x12 characterizes a polymer-solvent pair and has been used in connection with polymer solution behavior (flory, 1953), but it has some theoretical limitations (Barton, 1991). The interaction parameter 202 is not a constant, and depends on polymer concentration and molecular weight as well as temperature. It is a composite term influenced by factors such as hydrogen bonding, and polymer characteristics. In our study, PP resin pellets were used in the sorption instead of films, which were generally used by sorption studies. The density and chain matrix of PP resin 86 varied from the commercial PP films or the forms Hansen Solubility Parameters had been generated from. Besides, most of the sorption kinetics studied by Flory-Huggins equation have % Wt. Gain in the scale of above ten percent. While in this study, the maximum weight gain was only 0.31%, which is around 30 times less than the reported values for other solvent-polymer systems. The combination of factors described above might cause the deviation of the experimental data away from theoretical Flory-Huggins equation. Further theoretical models need to be explored and adjusted. 4. 3.5 Equilibrium Benzaldehyde Sorption and Sorption Kinetics Measurements These equilibrium organic uptakes in sorption experiments were used to calculate the solubility coefficient (S). S is an equilibrium partition coefficient for distribution of the penetrant between polymer and vapor phase. It is a measure of the mass of permeant molecules sorbed by a unit of polymer mass per unit of partial pressure, and it is defined according to equation (5) (N ielsent, 1994; Barr, 2000): s=M°° (5) V'p where M00 is the total amount (mass) of vapor absorbed by the polymer at equilibrium for a given temperature, v is the volume of the polymer sample, and p is the penetrant driving force in unit of pressure. S, in the unit of kg/m3Pa, was calculated by multiplying the weight gain (kg of sorbate per kg of polymer) by polymer density and divided by vapor 87 pressure. The calculated solubility coefficients for the sorption of benzaldehyde by PP at 25 °C are shown in Table 4.2. The broken line means the solubility coefficients are not available since the equilibrium was not reached in the sorption test. For resin, the equilibrium was reached at all vapor activities except at vapor activity of 0.1, the S data is not available due to the weight gain % is close to the instrument noise (Qin, 2006). As we studied the sorption of benzaldehyde in various PP forms, we found that sorption in resin at high vapor activities exhibited two-stage sorption at 25 °C, the first equilibrium mass uptake showed the sorption kinetics was controlled by Fickian diffusion. Thus it is possible to deduct a diffirsion coefficient (D) in the Fickian region for each curve with vapor activity from 0.1 to 0.9 (Kriiger, 2005). For sorption in PP sheet, one-stage sorption was observed, D could be calculated when the equilibrium was reached. For this purpose each sorption curve is plotted as M/M.o vs. t1/2 as the demonstration of the VA=0.5 step in Figure 4.7. 88 0.15 I I I I l I r P o Weight Gain, % 1° C Ut A A A1 A A 0.00 0 1‘0 20 3A0 40 ‘ T . r . . 02511229“. 1 0 20 40 60 80 100 120 140 160 timeI/z, minl/2 Figure 4.7. Sorption curve for benzaldehyde vapor in PP resin at 25 °C with vapor activities of 0.5. The inner small figure is the plot of M / Moo vs. t”2 in the initial Fickian region The benzaldehyde sorption kinetics for PP resin could be measured by the method applied for polymer sphere or powder. The ratio of the amount of vapor absorbed at any time (I) over the equilibrium sorption level at infinite time (M / M...) for polymer samples of spherical geometry and of diameter (d) is given by the expression (Hernandez, 1986a): M, _1__6_[ex (—4D7r2t)+1ex (—16D7r2t M 7:2 p d2 4 p d2 w )] (6) The diffusion coefficient (D) is readily obtained from the above equation by setting .M / M,O equal to 0.5 and solving to give the expression: 89 d2 D=7.45-10" ._ (7) t0.5 Similarly, the sorption kinetics for PP sheet can be obtained by equation (8) and (9) (Hernandez, 1986a; Nielsent, 1994): i=l—i[ex (—Drr2t)+1ex (907:2: M, a2 p 12 9 p 12 )] (8) I is the thickness of the sheet. The sorption diffusion coefficient (D,) can be calculated from the above equation by setting M / M.o equal to 0.5 and solving to give D5: 2 D, = 0.0491 (9) (0.5 The diffusion coefficient for the initial Fickian sorption stage for sorption of benzaldehy by PP at different vapor activities were calculated and are shown in Table 4.3. The broken line means the diffiision coefficients are not available since the equilibrium was not reached at the end of the sorption test. In general, the diffusion coefficients in resin in the Fickian region are found much larger than that in sheet under testing vapor pressures. 90 Table 4.2. % Wt. Gain and calculated solubility coefficient (S) for benzaldehyde- PP at 25 °C Resin Sheet Therrnoformed sheet V (Polyhedron container) ipf” % Wt. 8 °/. Wt. s % Wt. S “t“"ty Gain Kg/(m3-Pa) Gain Kg/(m3°Pa) Gain Kg/(m3-Pa) [pa/pa] [% [% [% gB/gPP] gB/gPP] 88/398] 0.1 --- --- 0.053 0.030 0.189 --- 0.3 0.075 0.014 0.150 0.028 0.489 «- 0.5 0.085 0.010 0.798 --- 0.899 --- 0.7 0.275 0.023 1.087 m 1.140 «- 0.9 0.440 0.028 1.735 --- 3.135 0.191 Table 4.3. Calculated Fickian diffusion coefficient (D) for benzaldehyde-PP at 25 °C Vapor activity 0.1 0.3 0.5 0.7 0.9 D PP resin --- 33.001 26.489 34.253 20.912 [mz/s] * 10-12 PP sheet 4.873 1.001 --- --- «- 91 0.03 40 .3. ~ 30 q .5" 002 a E .— e 20 Ni” .5. . 0.01 . U1 -- 10 Q 0 i , .9 0 0.3 0.5 0.7 0.9 Vapor Activity, [P /P 0] I S A D —Linear(D) —Linear(S) Figure 4.8. Comparison of solubility coefficient (S) and diffusion coefficient (D) at different vapor activities. The solubility coefficient (S) and diffusion coefficient (D) for sorption of benzaldehyde in PP resin under the test conditions are compared and shown in Figure 4.8. The squares and triangles are the S and D values, individually. The solid lines are the linear trend lines based on the S / D values. It is found that S and D have reverse trend with increasing the vapor pressure, i.e. the higher the vapor activity, the larger the S values, but the smaller the D values. This interesting phenomenon indicates the structural changes within the polymer may occur, such as relaxation effects of the polymer side chains, free volume variations or swelling of the materials, especially at high vapor activities. The crystalline zones might also contribute to the changes by reducing and restraining the polymer chain mobility in the amorphous region because the chain ends 92 are trapped in the neighboring crystalline lamellae, and then lead to higher activation energy of diffusion. That might explain the fact we observed that at high vapor activities (0.7-0.9), where the relaxation likely happened, the lower the D values were found. The exact reason will require future study. 4.4 Conclusion The sorption of benzaldehyde in different PP forms (resin, sheet, thermoformed sheet and atatic PP) was studied at 25 °C. A continuously gravimetric method was adopted in the sorption study by recording the weight gain of dry polymer under certain vapor pressures. It was found that both surface area, polymer bulk properties, geometry and morphology contribute to the sorption properties of polymer. The polymer processing (extrusion and thermoforming) induced surface and morphology changes during the heat treatment, thus changed the sorption of polymer dramatically. It was also found the changes in the morphology due to processing had a large impact on the sorption of benzaldehyde at high vapor activities since the sorption of benzaldehyde on the resin is much lower than that on the sheet, the thermoformed sheet and the ataic PP It was also observed that in PP resin, as vapor activity increased, solubility coefficient (S) increased and diffusion coefficient (D) decreased. The two-stage sorption was observed in resin at high vapor acitivities under the testing conditions. It indicates that upon vapor sorption structural changes within the polymer 93 may occur, such as relaxational effects of the polymer side chains, free volume variations, glass transition temperature lowering and an increase in viscoelastic behavior. The experimental results demonstrated that polymer processing changed the polymer morphology and surface structure, thus had great impact on the sorption behavior, especially at high vapor pressures. Acknowledgement The authors acknowledge helpful discussions with Dr. Gregory Baker in department of chemistry, Michigan State University. Also thank Pfizer for the financial support of this project. 94 Chapter 5 Conclusion The organic barrier property of polypropylene (PP) was studied in this project. A gravimetric method was developed to evaluate the sorption of benzaldehyde by PP via a SGA100R gravimetric system including a Rubotherm magnetic suspension electrobalance. The instrument variability and noise were studied before the sorption study. The drift was observed at the test conditions varied from 2"‘10'8 to 3*10'5. The Noise was observed and the amplitude was in the range of 0.0124% to 0.165%. A continuously gravimetric method was developed to study the sorbate uptake and sorption kinetics of benzaldehyde vapor in rubbery PP resin, sheet and polyhedron thermoformed container via SGA-100R gravimetric analyzer at 25°C, 0% RH. Solubility coefficients for benzaldehyde sorption in PP resin, sheet and container were determined based on the equilibrium uptake. Diffusion coefficients were also calculated on PP resin and sheet in the Fickian difiusion region. Interestingly, as the vapor activity increased, the solubility coefficient increased and the diffusion coefficient decreased in benzaldehyde — PP resin system. It was found that surface area, geometry and morphology contributed to the sorption properties of polymer. The experimental results demonstrated that polymer processing, including extrusion and thermoforming, had a great impact on the mass transfer 95 properties of PP, especially at high vapor activities. The processing changed the polymer morphology and surface structure, thus-induces the changes in organic barrier properties among different polymer forms. It was found that the sorption of benzaldehyde in the PP polyhedron container was twice that of the sorption in PP sheet at lower vapor activities, and the sorption in sheet was almost three times the sorption of the resin at higher vapor activities. The sorption dynamics of the polyhedron container, sheet and resin were found different from each other. The results and conclusions of this work could be used to contribute to the PP mass transfer database. Current finding emphasizes the need to assess mass transfer of formed polymer sheets or containers to accurately determine the barrier properties of packaging systems. The future work of this project includes: 0 To determine how the polymer bulk property and surface property affect the sorption process and how does each factor contribute to the sorption properties. 0 To compare solubility coefficient (S) and diffusion coefficients (D) in different PP forms. 0 Try to determine the method to calculate D for atatic PP and polyhedron containers. 0 To minimize the drift and noise in the SGA 100R gravimetric analyzer. 96 REFERENCES Baner, A.L.; Hernandez, R.J.; Jayaraman, K.; Giacin, JR. 1986. Isostatic and quai—isostatic methods for determining the permeability of organic vapors through barrier membrances. Current Technologies in Flexible Packaging. ASTM STP 912. Barr, C. D.; Giacin, J .R.; Hemandex R. J. 2000. A determinatio of solubility coeflicient values determined by gravimetric and isostatic permeability techniques. Package. Technol. Sci. 13: 157-167. Barrer, RM. 1937. Nature of the difiitsion process in rubber. Nature, 140: 106-107. Barrer, R.M.; Barrie, J .A.; Slater, J. 1958. Sorption and dzflitSion in ethyl cellulose. Part III. Comparison between ethyl cellulose and rubber. J. Polym. Sci. 27: 177-197. Barton, A.F.M. 1991. In CRC Handbook of solubility parameters and other cohesion parameters. Boca Raton, Fla.: CRC Press. Page 266. Beck, J .V.; Arnold, K.J. 1977. Parameter estimation in engineering and science. John Wiley and Son, New York, NY. Berens, AR. 1989. Sorption of organic liquids and vapors by rigid PVC. J. Appl. Polym. Sci. 37: 901-913. Biros, J .; Zeman, L.; Patterson, D. 1971. Macromolecules, 4(1):30-35. Bondar, V.I.; Freeman, ED. 1999. Sorption of gases and vapors in an amorphous glassy perfluorodioxole copolymer. Macromolecules. 32: 6163-71. Buchold, R.;Nak1adal, A.; Gerlach, G.; Herold, M.; Gauglitz, G.; Sahre, K.; Eichhorn, K. J. 1999. Thin Solid Films, 350:178-185. Caner, C.; Hernandez, R.J.; Pascall, M.; Balasubrarnaniam, V.M.; Harte, BR. 2004. J. Packag. Technol. Sci. 17:139-153. Crank, J. 1956. Mathematics of dzfiirsion. 2nd Edition. Oxford University Press, London. Crank, J. 1975. The mathematics of diflitsion. 2nd Edition, Clarendon Press, Oxford, England. 97 Crank, J .; Park, GS. 1949. Trans, Faraday Soc. 45: 240. Crank, J .; Park, GS. 1951. Trans, Faraday Soc. 47: 1072. Crank, J.; Park, GS. 1968. Diffusion in polymers. Academic Press, London and New York. Danch, A.; Osoba, W. 2003. Structural relaxation of the constrained amorphous phase in the glass transition zone. J. Therm. Anal. Calorim. 72: 641-650. Demertzis, P. G; Johansson, F.; Lievens, C.; Franz, R. 1997. Studies on the development of a quick inertness test procedure for multi-use PET containers-Sorption behaviour of bottle wall strips, Packag.Technol. Sci. 10: 45-58. Dhoot, S.N.; Freeman, B.D.; Stewart, M.E.; Hill, A.J. 2001. Sorption and transport of linear alkane hydrocarbons in biaxially oriented polyethylene terephthalate. J. Polym. Sci., Part B: Polym. Phys. 39: 1160-72. Flory, P.J. 1953. In Principles of polymer chemistry. Cornell University Press, New York. Frisch, H.L. 1980. Polym. Eng. Sci. 20:241. Fujita, H. 1968. Organic vapors above the glass transition temperature. Diffusion in Polymers, Crank, J and Park, GS. (eds.). Academic Press, London and New York. Fujita, H.; Kishimoto, A.; Matsumoto, K. 1960. Concentration and temperature dependence of drfiusion coeflicients for systems polymethyl acrylate and n-aIkyl acetates. Trans. Faraday Soc. 56: 424-437. Gavara, R.; Catala, R.; Hemandez—Munoz, P.M.; Hernandez, R.J. 1996. Evaluation of permeability through permeation experiments: Isostatic and quasiisostatic methods compared Packag. Technol. Sci. 9: 21 5-224. Golding, P.D.; Machin, W.D. 1992. The Wagner equation for vapor pressure: obverse and reverse views. J. Chem. Soc., Faraday Trans. 88(14): 1985-91. Guadagno, L.; Baldi, P.; Vittoria, V.; Guerra, G 1998. Sub-T g annealing of the clathrate & Delta; form of syndiotactic polystyrene. Macromol. Chem. Phys. 199: 2671-75. Hansen, CM. 1999. In Hansen Solubility Parameters. A User Is Handbook. CRC Press, Boca Raton, Chapter 2. 98 Hernandez, R.J.; Giacin, J .R. and Baner, A.L. 1986a. The evaluation of the aroma barrier properties of polymer films, J. Plast. Film Sheet. 2: 187—211 Hernandez, R.J.; Giacin, J .R.; Baner, A. 1986b. Measuring the aroma barrier properties of polymeric packaging materials, Packag. Technol. 16: 12-15. Hernandez, R.J.; Giacin, J.R.; Grulke, E.A.. 1992. The sorption of water vapor by an amorphous polyamide. J. Memb. Sci. 65: 187-199. Hernandez, R.J.; Gavara, R. 1994. Sorption and Transport of water in nylon-6 films. J. Polym. Sci., Part B: Polym. Phys. 32: 2367-74. Hernandez, R.J.; Gavara, R. 1999. Plastics Packaging: Methods for studying mass transfer interactions: A literature review. Leatherhead, UK: Pira International. Hernandez, R.; Selke, S.E.M.; Culter, J .D. 2000. Plastic packaging: properties, processing, applications and regulations. Hanser Publishers, Munich, 344-345. Huang, S.J.; Giacin, J .R. 1998. Evaluating the eflect of temperature and vapor concentration on the organic vapor barrier properties of polymer membranes by an isostatic procedure, J. Plast. Film Sheeting. 14: 308-333. Hurd, C.D.; Bennett, CW. 1929. The pyrolysis of benzaldehyde and of benzyl benzoate. J. Am. Chem. Soc. 51: 1197-1201. Jasse, B.; Seuvre, A.M.; Mathlouthi, M. 1994. Permeability and structure in polymeric packaging materials. Food Packaging and Preservation. Mathlouthi, M. (Eds), London, UK: Chapman & Hall, 1-22. Kesting, R.E.; fiitzsche, AK. 1993. Polymeric gas separation membranes. New York, Wiley & Sons. Kiparissides, C.; Dimos, V.; Boultouka, T.; Anastasiadis, A.; Chasiotis, A. 2003. Experimental and theoretical investigation of solubility and drfiusion of ethylene in semicrystalline PE at elevated pressures and temperatures. J. Appl. Polym. Sci. 87: 953-966. Klopffer, M.H.; Flaconneche, B. 2001. Transprot properties of gases in polymers: bibliographic review. Oil Gas Sci. Technol.-Rev. IFP. 56: 223-244. 99 Kriiger, K.M.; Sadowski, G 2005. F ickian and non-Fickian sorption kinetics of toluene in glassy polystyrene. Macromolecules. 38(20): 8408-17. Lange, J .; Wyser, Y. 2003. Recent innovations in barrier technologies for plastic packaging — a review. Packag. Technol. Sci. 16:149-158. Machin, D.; Rogers, CE. 1972. Concentration dependence of drfiitsion coeflicients in polymer-penetrant systems. CRC Critical Reviews in Macromolecular science, CRC Press, Cleveland. 245-313. Manfredi, C.; Nobile, M.A.D.; Mensitieri, G; Guerra, G Rapacciulo, M.J. 1997. Vapor sorption in emptied clathrate samples of syndiotactic polystyrene. J. Polym. Sci., Part B: Polym. Phys. 35: 133-140. McDowell, C.C.; Freeman, B.D.; McNeely, GW. 1999. Acetone sorption and uptake kinetic in poly(ethylene terephthalate). Polymer. 40: 3487-99. Michaels, A.S.; Bixler, H.J. 1961 a. Solubility of gases in polyethylene. J. Polym. Sci. 50: 393-412. Michaels, A.S.; Bixler, H.J. 1961 b. Flow of gases through polyethylene. J. Polym. Sci. 50: 413-439. Michaels, A.S.; Parker, RB. 1959. Sorption and flow of gases in polyethylene. J. Polym. Sci. 41: 53-71. Mogri, 2.; Paul, DR. 2001. Gas sorption and transport in side-chain crystalline and molten poly(octadecyl acrylate). Polymer, 42: 2531-42. Nielsent T.J.; Giacin J .R. 1994. The sorption of limonene / ethyl acetate binary vapour mixtures by a biaxially oriented polypropylene film. Packag. Technol. Sci. 7: 247-258. Nir, M.M.; Ram, A.; Miltz, J. 1996. Sorption and migration of organic liquids in poly(ethylene terephthalate). Polym. Eng. Sci. 36: 862-868. Ochiai, H.; Gekko, K.; Yamarnura, H. 1971. Sorption properties of polypropylene. J. Polym. Sci. 9: 1629-40. Oliverira, N.S.; Oliverira, J.; Gomes, T.; Ferreira, A.; Dorgan, J.; Marrucho, I.M. 2004. Gas sorption in polyflactic acid) and packaging materials. Fluid Phase Equilib. 222-223: 317-324. 100 Palamara, J .E.; Mulcahy, K.A.; Jones, A.T.; Danner, R.P.; Duda, J .R. 2005. Solubility and drfi‘usivity of propylene and ethylene in atactic polypropylene by the static sorption technique. Ind. Eng. Chem. Res. 44(26): 9943-50. Paul, D.R.; Koros, W.J. 1976. Eflect of partially immobilizing sorption on permeability and the dzflusion time lag. J. Polym. Sci., Polym. Phys. Ed. 14: 675-685. Pixton, M.R.; Paul, DR. 1994. Polymeric gas separation membrances. Paul, D.R.; Yampol’skii, Y.P. (Eds) CRC Press: Boca Raton, FL, 83-153. Potineni, R.V.; Peterson, D.G 2005. Influence of thermal processing conditions on flavor stability in fluid milk: benzaldehyde. J. Diary Sci. 88:1-6. Qin, Y.; Rubino, M.; Auras, R. 2005. Efibct of container shape on flavor scalping in polymeric food containers. 230‘h ACS National meeting & Exposition, Aug.27-Sep.01, Washington DC. Qin, Y.; Rubino, M.; Auras, R.; Lockhart, H. 2006. Measurement of sorption of benzaldehyde in polymer with magnetic suspension microbalance. Polym. Test, in preparation. Raharjo, R.D.; Lee, H.J.; Freeman, B.D.; Sakaguchi, T.; Masuda, T. 2005. Pure gas and vapor permeation properties of poly[I-phenyl-2-[p-(trimethylsilyl)phenyl]acetylene] (PTMSDPA) and its desilylated analog, poly[diphenylacetylene] (PDPA). Polymer. 46: 6316-24. Raucher, D.; Sefcik, MD. 1983. Sorption and transport in glassy polymers. ACS Symp. Ser., 223: 111-124. Risch, S.J.; Hotchkiss, J .H. 1991. Food and packaging interactions 11. American Chemical Society, Washington DC. Robinson, J .W.; Frame, E.M.S.; Frame, GM. 2005. Undergraduate instrumental analysis. Sixth ed. Marcel Dekker, New York. Rogers, CE. 1985. Permeation of gases and vapours in polymers. Polymer Permeability. J. Comyn (eds.). Elsvier Applied Science Publishers, Chapter 2, 11-73. Rossi, G; Mazich, K.A. 1993. Macroscopic description of the kinetics of swelling for a cross-linked elastomer or a gel. Phys. Rev. E48: 1182-91. 101 Rubino, M.; Tung, M.A.; Yade, S.; Britt, I. J. 2001. Permeation of oxygen, water vapor and limonene through printed and unprinted biaxially oriented polypropylene films. J. Agric. Food Chem. 49: 3041-45. Safa, H. L.; Bourelle, F. 1999. Sorption-desorption of aromas on multi-use PET bottles. A test procedure. PackagTechnol. Sci. 12: 37-44. Sajiki, T.; Giacin, J .R. 1993. Permeation of ethyl acetate vapor through silica deposited polyethylene terephthalatefilm and composite structures. J. Plast. Film Sheet. 9: 97—111. Sato, Y; Fujiwara, K; Takikawa, T.; Takishima, S; Masuoka, H. 1999. Solubilities and drfiusion coefiicients of carbon dioxide and nitrogen in polypropylene, high-density polyethylene, and polystyrene under high pressures and temperatures. Fluid Phase Equilibr. 162: 261-276. Sato, Y.; Yurugi, M.; Yamabiki, T. Takishima, S.; Masuoka, H. 2001. Solubility of propylene in semicrystalline polypropylene. J. Appl. Polym. Sci. 79: 1134-43. Sefcik, M.D.; Schaefer, J.; May, F.L.; Raucher, D.; Dub, SM. 1983. Drfiusivity of gases and main-chain cooperative motions in plasticized polyvinylchloride. J. Polym. Sci., Polym. Phys. 21: 1041-54. Sliepcevich, A.; Storti, G; Morbidelli, M. 2000. Measurement of drfiusivity and solubility of olefins in polypropylene by gas chromatography. J. Appl. Polym. Sci. 78: 464-473. Stern, S.A.; Trohalaki, S. 1990. Fundamentals of gas drflusion in rubbery and glossy polymers. Banier Polymers and Structure, William, J.K. (eds.). ACS Symposium Series 423. Washington DC. Thalmann, W.R.. 1990. Sorption of solvents in plastics and solvent desoprtion and retention. Package. Technol. Sci. 3: 67-82. Tsuboi, A.; Kolar, P.; Ishikawa, T.; Karniya, Y.; Masuoka, H. 2001. Sorption and partical molar volumes of C2 and C3 Hydrocarbons in Polypropylene copolymers. J. Polym. Sci., Part B: Polym. Phys. 39: 1255-62. Van Krevelen, D.W. 1997 . Properties of polymers. Their correlation with chemical structure; their numerical estimation and prediction fiom additive group contributions. Third edition. Elsevier, Amsterdam. 102 Van Willige, R.W.G.; Linssen, J.P.H.; Meinders, M.V.J.; Van der Stege, H.J. Voragen, A.G.J .2002. Influence of flavour absorption on oxygen permeation through LDPE, PP, PC and PE T plastics food packaging, Food Addit. Contam. 19: 303-313. Vittoria, V.; Rubolo Filho, A. 1993. J. Appl. Polym. Sci. 49(2): 247-256. Vrentas, J .S.; Vrentas, CM. 1998. Integral sorption in rubbery polymers. J .Polym. Sci; Part B: Polymer Phys. 36: 171-180. Weinkauf, D.H.; Paul, DR. 1990. Effects of structural order on barrier properties. ACS Symp. Ser., 423: 60-91. Wiberg, K.B.; Stewart, R. 1955. The mechanisms of permanganate oxidation. 1. The oxidation of some aromatic aldehydes. J. Am. Chem. Soc. 77:1786-95. William, J.K. 1990. Barrier polymers and structures. American Chemical Society, Washington DC. Zhou, Q.; Guthrie, B.; Cadwallader, KR. 2004. Development of a system for measurement of permeability of aroma compounds through multilayer polymer films by coupling dynamic vapour sorption with purge-and-trap/fast gas chromatography: research fiom the University of Illinois, Urbana, IL, USA, Package. Technol. Sci. 17: 175-185. 103 \ E 1111111111111111111111 1293 30284 Y AR _— A I -"'~‘ —' T -‘.'-T:'.——‘ .— ,.._ _